Contemporary Medical Imaging

Series Editor: U. Joseph Schoepf

U. Joseph Schoepf Editor

CT of the Heart
Second Edition
Contemporary Medical Imaging

Series Editor
U. Joseph Schoepf

More information about this series at:

http://www.springer.com/series/7687
U. Joseph Schoepf
Editor

CT of the Heart
Second Edition
Editor
U. Joseph Schoepf
Department of Radiology
Center for Advanced Imaging Research (CAIR)
Medical University of South Carolina
Charleston, SC
USA

Originally published in the series: Contemporary Cardiology

Contemporary Medical Imaging
ISBN 978-1-60327-236-0    ISBN 978-1-60327-237-7 (eBook)
https://doi.org/10.1007/978-1-60327-237-7

Library of Congress Control Number: 2018966541

© Humana Press 2005, 2019

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is
concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction
on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation,
computer software, or by similar or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not
imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and
regulations and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed
to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty,
express or implied, with respect to the material contained herein or for any errors or omissions that may have been
made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

This Humana Press imprint is published by the registered company Springer Science+Business Media, LLC part of
Springer Nature
The registered company address is: 233 Spring Street, New York, NY 10013, U.S.A.
This book is dedicated to my mother Ursula Schoepf and my late father, Josef
Schoepf.
Foreword to the First Edition

Radiologic technology has made dramatic advances in the last 25 years, and none have been
more impressive than those in computed tomography (CT). The progress in the speed of
obtaining images, computing, postprocessing, and spatial resolution has been incredible. The
result is that CT has moved from displaying purely morphologic information to providing
valuable physiologic data as well. Whether with electron beam or multidetector-row CT,
advances are impressive and nowhere have the applications been more useful and dramatic
than in the heart.
This multiauthored book, CT of the Heart, edited by U. Joseph Schoepf, MD, is a splendid
rendition of the state-of-the-art in CT imaging of the heart; however, where appropriate, it also
features comparisons with other technical approaches, such as magnetic resonance and ultra-
sound. The contributors are leading radiologists, cardiologists, physicists, engineers, and basic
and clinical scientists from Europe, the United States, Israel, and Japan.
The entire contents are meticulous and comprehensive, from the introduction about the
past, present, and future of CT of the heart, through the technical underpinning of the method
and the various clinical, physiologic, and pathologic applications of CT in studying the heart.
This book fills an immense need, particularly at a time when cardiac screening with CT,
whether one agrees with this practice or not, is a reality. Furthermore, with the rapid increase
of aging populations in the industrialized world, noninvasive diagnostic approaches are
increasingly needed. As technology continues to advance and applications of CT to heart stud-
ies expand, it is my hope that the editor will bring this book up to date with a new edition.

Alexander R. Margulis, MD, DSc (HON)

Clinical Professor of Radiology
Weill Medical College of Cornell University,
New York, NY, USA

vii
Foreword

CT imaging of the heart and coronary vessels has emerged over the past two decades as one of
the most important and dynamic advances in medicine. Heart disease is the leading cause of
death in the United States and worldwide, challenging health systems and health providers on
how best to diagnose and manage their patients. CT imaging helps address these questions
through the remarkable and important information it provides about both cardiovascular anat-
omy and function.
The first edition of CT of the Heart: Principles and Applications edited by U.  Joseph
Schoepf was excellent. The book was very well received and widely used. Dr. Schoepf chose
a multiauthor format that allowed him to invite key leaders in each aspect of the subject to
contribute their special knowledge—physicists, engineers, radiologists, cardiologists, and oth-
ers. Their presentations were outstanding—richly illustrated and put in appropriate context
with other methods.
The second edition of CT of the Heart promises to build on the excellence of the first edition
by maintaining the strategy of selecting the most expert people as authors. To this end, Dr.
Schoepf has maintained the multiauthor format in the second edition, now inviting over 170
people from around the globe as contributors. The authorship list is a true “Who’s Who” of
people working in the field.
Two things happen over time that make new editions of even the most classic medical texts
vital and important. These are implicit in the subtitle of the first edition of CT of the Heart and
are advancements in the science and technology underlying the subject and advancements in
the accumulated knowledge about the role and efficacy of clinical applications. In the decade
between editions, it is fair to say that the pace of technology development has steadily increased
and that new applications have been added to clinical practice. Dr. Schoepf and his coauthors
address these important advances in the second edition. The discussion of each topic is
designed to bring it up to date, and several new chapters have been added to cover new areas
of technology development and clinical application. As in the first edition, the discussions are
meticulous and comprehensive.
Another feature of CT of the Heart that will be useful to readers is the separation of chapters
into categories of “Where We Were,” “Where We Are,” and “Where We Are Going.” Too often
in textbooks, there is not a clear separation of material covering topics recognized as well
established versus emerging topics that are important in comprehensive understanding of a
subject but not yet in routine clinical use. The 14 chapters in the section on “Where We Are
Going” bespeak the dynamic advances being made in CT imaging. Topics in this section
include new approaches to anatomic and functional applications such as new approaches to
myocardial perfusion imaging but also the potential roles of radiomics, big data, and machine
learning.

ix
x Foreword

CT of the Heart will be invaluable for students and trainees seeking to learn the subject as
well as established physicians looking for definitive reference information or for ideas about
how to continue to advance their practices. Since the second edition has the same attributes
that made the first edition a trusted resource, it will soon be regarded in the same way. Dr.
Schoepf and his coauthors are to be congratulated for producing such a high-quality and
timely text.

James H. Thrall, MD
Chairman Emeritus, Department of Radiology
Massachusetts General Hospital
Boston, MA, USA
Distinguished Taveras Professor of Radiology
Harvard Medical School
Boston, MA, USA
Preface

Ποταμοῖσι τοῖσιν αὐτοῖσιν ἐμϐαίνουσιν, ἕτερα καὶ ἕτερα ὕδατα ἐπιρρεῖ… – Herakleitos

More than a decade has passed since the publication of the first edition of CT of the Heart. And
what a ride it’s been since then! From the perch of today’s technology, with lightning-fast
acquisition speeds and temporal resolution, massive tube power, yet gentle techniques, the
evolution could not have been more dramatic. Back then we were mostly still living in the dark
ages of 16-slice multidetector-row CT technology, with 64-slice CT faintly on the horizon.
While we have been experiencing the evolution of cardiac CT as a continuum, for many the
introduction of 64-slice CT technology constitutes the pivotal breaking point in history, whence
the clinical use of cardiac CT became more broadly established for the first time. In the subse-
quent years, we all expected a revolution, a wildfire to happen, with cardiac CT ensconcing
itself rapidly, profoundly, and irrevocably in all arenas of cardiovascular medicine. It did not
happen quite as fast as many had betted on, causing a degree of disillusionment in some quar-
ters. After all, it may not be a bad thing that not every latest flash in the pan gets embraced by
mainstream medicine overnight.
But then something quite rare and precious happened; the field of believers in this technol-
ogy came together and, in a manner unprecedented in medical imaging, piece by piece built the
evidence that incrementally drove this test to new heights and today forms the foundation for
the ever-growing importance of cardiac CT. In fact, we submit that cardiac CT may be consid-
ered a beacon, a blueprint, and prime example of how the value of a medical test can be
unequivocally proven and supported via the generation of high-level evidence, a formidable
challenge that the field of medical imaging has mostly unsuccessfully grappled with to date.
This is what this book is about; while in the first edition we mainly investigated a fascinating
new instrument looking for an application, we now have a vast realm of guideline-driven,
robust, and beneficial clinical applications that are enabled by an enormous and ever-growing
field of technology. Accordingly, the focus has shifted from a technology-centric to a more
patient-centric appraisal. While the specifications and capabilities of the CT system itself
remain front and center as the basis for diagnostic success, much of the benefit derived from
cardiac CT today comes from avant-garde technologies enabling enhanced visualization,
quantitative imaging, and functional assessment, along with exciting deep learning, and artifi-
cial intelligence applications. Long have we passed the stage of a mere tool for noninvasive
coronary artery stenosis detection in the chest pain diagnostic algorithms; cardiac CT has
proven its value for uses as diverse as personalized cardiovascular risk stratification, predic-
tion, and management, diagnosing lesion-specific ischemia, guiding minimally invasive struc-
tural heart disease therapy, and planning cardiovascular surgery, among many others.
In the Preface to the first edition of CT of the Heart, we stated that we do not claim to have
all the answers. That is still the case; but we have vastly more answers and enough to know that
cardiac CT is here to stay and bound to occupy the space that we originally envisioned. In
some more regulated and resource-conscious economies, we already see cardiac CT posi-
tioned as the entrance test and gatekeeper to any type of chest pain work-up, invasive or not.
However, also in less progressive, more entrenched, and conflicted healthcare systems around

xi
xii Preface

the globe, this test is now quickly gaining ground and will even more so with newer genera-
tions of healthcare providers who are less enamored with outdated testing strategies of the past.
Like the first edition, the second edition of CT of the Heart is again a snapshot of the status
quo, of the current state-of-the-art, and of a success story in the care for our patients which still
keeps rapidly evolving. Yet, we have a much clearer view now of what we have accomplished,
where we are, and where we are going.
While the first edition was the work of many, the second edition is the result of the work of
even more. An astounding array of the great houses in cardiac imaging, giants in the field,
came together to present our readers with the most comprehensive, coherent, up-to-date, and
in-depth review of cardiac CT principles and applications. We are grateful beyond limits to
this exalted, respected group of experts who poured their genius into this tome. Finally, this
work would not have come to fruition without the invaluable help of Taylor M. Duguay and
Dante Giovagnoli in our lab and Margaret Burns of Springer, who so skillfully and deftly
steered the production of this second edition. We hope that this work will inspire and guide
current and future leaders in healthcare in their quest to optimally harness the powers of a
disruptive, amazing technology, to the benefit of our patients worldwide.

Charleston, SC, USA U. Joseph Schoepf

Contents

Part I Where We Were

1 History of Cardiac CT: A Personal Story. . . . . . . . . . . . . . . . . . . . . . . . . . . . .   3
John A. Rumberger
2 Evolution of Radiation Dose from Cardiac CT. . . . . . . . . . . . . . . . . . . . . . . .   11
Manoj Mannil and Hatem Alkadhi
3 The Long March into Clinical Practice: Cardiac CT
and Its Competitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .   19
Seth Uretsky, Alan Rozanski, and Daniel Berman

Part II Where We Are: Human Requisites

4 Cardiac Computed Tomography: A Team Sport. . . . . . . . . . . . . . . . . . . . . . .   37
Sheldon E. Litwin
5 Cardiac CT: Credentialing and Accreditation. . . . . . . . . . . . . . . . . . . . . . . . .   41
James M. Kofler, Heidi A. Edmonson, Shuai Leng, and Eric E. Williamson

Part III Where We Are: Technical and Operational Requisites

6 Cardiac CT Platforms: State of the Art . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .   51
Bernhard Schmidt, Katharine Grant, Thomas G. Flohr, and Thomas
Allmendinger
7 Principles of Cardiac CT Image Acquisition. . . . . . . . . . . . . . . . . . . . . . . . . .   69
Thomas Henzler, Patricia Carrascosa, Brian S. Ko, and Ronen Rubinshtein
8 Dual Energy and Spectral CT Techniques in Cardiovascular Imaging. . . . .   87
B. Krauss and C. H. McCollough
9 Drugs in Cardiac CT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Sebastian Rogowski, Virginia W. Lesslie, and Ullrich Ebersberger
10 Contrast Media Injection Protocols in CT Coronary Angiography. . . . . . . . 109
Casper Mihl, Madeleine Kok, Joachim E. Wildberger, and Marco Das
11 Cardiovascular CT: Image Reconstruction. . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Annemarie M. den Harder, Arnold M. R. Schilham, and Martin J. Willemink
12 The Challenging Patient. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
Damiano Caruso, Domenico De Santis, Taylor M. Duguay, Sheldon E. Litwin,
and Carlo N. De Cecco

xiii
xiv Contents

13 Cardiac CT: Contemporary Clinical Image Data Display, Analysis,

and Quantification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
Moritz H. Albrecht, Marwen Eid, and Pal Spruill Suranyi
14 Workflow Optimization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Thomas Allmendinger, Andrew N. Primak, and Christian D. Eusemann
15 Defining the Role and Benefits of a 3D Laboratory
for Cardiovascular CT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
Laura J. Pierce, Daniel T. Boll, and Geoffrey D. Rubin
16 Structured Reporting for Cardiac CT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
Anil Attili and Ella A. Kazerooni
17 Integration of CT Data into Clinical Workflows: Role of
Modern IT Infrastructure Including Cloud Technology. . . . . . . . . . . . . . . . . 195
Paul Schoenhagen and Mathis Zimmermann
18 Thoughts on Coding and Reimbursement . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
Adefolakemi Babatunde and Pamela K. Woodard

Part IV Where We Are: Cardiac CT Fundamentals

19 Pathology and Pathophysiology of Coronary Atherosclerotic Plaques. . . . . 211
Hiroyoshi Mori, Frank D. Kolodgie, Aloke V. Finn, and Renu Virmani
20 CT Cardiac Anatomy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
Michael A. Kadoch and Hans-Christoph R. F. Becker
21 Patient Selection: When to Use Cardiac CT Versus Other
Imaging or Non-­imaging Tests. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
Pal Spruill Suranyi, Akos Varga-Szemes, Marques L. Bradshaw,
Richard R. Bayer II, Salvatore A. Chiaramida, Peter L. Zwerner,
and David Gregg
22 Current Guidelines. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
Kweku Appau and Arthur E. Stillman

Part V Where We Are: Risk Stratification and Management

23 Clinical Application of the Coronary Artery Calcium Score
and Implications for Cardiovascular Disease Prevention. . . . . . . . . . . . . . . . 259
Pamela B. Morris and Michael D. Shapiro
24 The Many Uses of Epicardial Fat Measurements . . . . . . . . . . . . . . . . . . . . . . 285
Mohamed Marwan

Part VI Where We Are: Non-invasive Coronary Artery Imaging

25 Nonatherosclerotic Coronary Artery Disease. . . . . . . . . . . . . . . . . . . . . . . . . . 297
Toru Sakuma, Kotaro Ouchi, and Kunihiko Fukuda
26 The Many Faces of Atherosclerosis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
Nandini (Nina) M. Meyersohn, Jan-Erik Scholtz, and Brian B. Ghoshhajra
27 Chronic Chest Pain. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319
Richard A. P. Takx and Csilla Celeng
28 Coronary CT Angiography for Evaluation of Acute Coronary
Syndrome in the Emergency Department . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331
Nam Ju Lee and Harold Litt
Contents xv

29 The Role of Cardiac CT in Patients with Metabolic Disorders . . . . . . . . . . . 349

Gianluca Pontone, Giuseppe Muscogiuri, and Mark Rabbat
30 Use of Coronary Computed Tomography Angiography in
Cardiac Risk Assessment for Non-cardiac Surgery. . . . . . . . . . . . . . . . . . . . . 355
Gregory Jackson and Richard R. Bayer II
31 CT for Guiding Successful Revascularization. . . . . . . . . . . . . . . . . . . . . . . . . 361
Maksymilian P. Opolski
32 Stent Assessment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375
Junjie Yang, Christian Tesche, Taylor M. Duguay, Lucas L. Geyer,
and Yundai Chen
33 Multidetector CT Angiography for Coronary Bypass Graft
Assessment and Reoperative Cardiac Surgery . . . . . . . . . . . . . . . . . . . . . . . . 381
Lloyd M. Felmly
34 Cardiac CT in the Setting of Heart Transplantation. . . . . . . . . . . . . . . . . . . . 391
Gorka Bastarrika and Gregorio Rábago

Part VII Where We Are: CT Assessment of Ventricular Function

35 CT of Cardiac Function and Wall Motion. . . . . . . . . . . . . . . . . . . . . . . . . . . . 407
Prabhakar Rajiah and Suhny Abbara
36 Three-Chamber Function with Cardiac CT. . . . . . . . . . . . . . . . . . . . . . . . . . . 423
Jongmin Lee
Part VIII Where We Are: Cardiac Imaging Outside the Coronaries
37 Diseases of the Myocardium and Pericardium. . . . . . . . . . . . . . . . . . . . . . . . . 443
Ana Paula S. Lima and Karen G. Ordovas
38 CT of Cardiac and Paracardiac Masses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 451
Harold Goerne and Prabhakar Rajiah
39 Valvular Heart Disease and Prostheses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 471
Gudrun M. Feuchtner
40 Cardiac Devices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 487
Ian R. Drexler, Alan C. Legasto, Daniel B. Green, and Quynh A. Truong

Part IX Where We Are: Transcatheter Therapy Planning

41 CT in the Context of Transcatheter Aortic Valve Replacement. . . . . . . . . . . 503
Eli Konen, Orly Goitein, and Arik Wolak
42 CT for Minimally Invasive Repair of Mitral Valve and Other
Structural Heart Diseases. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 519
John F. Mooney, Philipp Blanke, Shaw Hua Kueh, Stephanie Sellers,
and Jonathon A. Leipsic
43 Cardiac CT: Electrophysiological Applications. . . . . . . . . . . . . . . . . . . . . . . . 531
Joan M. Lacomis, Iclal Ocak, Friedrich Knollmann, Andrew Voigt,
and Raveen Bazaaz
xvi Contents

Part X Where We Are: Congenital Heart Disease

44 Special Technique Considerations for Congenital
Heart Disease Imaging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 555
Anthony M. Hlavacek
45 CT of Coronary Artery Anomalies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 565
Long Jiang Zhang, Shahryar M. Chowdhury, and Guang Ming Lu
46 CT Spectrum of Congenital Heart Disease. . . . . . . . . . . . . . . . . . . . . . . . . . . . 579
David Steflik and Anthony M. Hlavacek
47 The Use of Cardiovascular CT in Repaired CHD. . . . . . . . . . . . . . . . . . . . . . 603
B. Kelly Han, Andrew Crean, and John R. Lesser

Part XI Where We Are: The Heart and Beyond

48 CT Imaging of the Heart-Lung Axis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 623
Michelle C. Williams and Edwin J. R. van Beek
49 Ischemic Stroke: The Role of Cardiac CT. . . . . . . . . . . . . . . . . . . . . . . . . . . . 635
Jin Hur and Byoung Wook Choi
50 Incidental Findings on CT Angiography and How to
Manage Them. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 647
Seung Min Yoo, Hwa Yeon Lee, and Charles S. White

Part XII Where We Are: The Bigger Picture

51 Prognosis and Outcome: State of the Evidence. . . . . . . . . . . . . . . . . . . . . . . . 659
Asim Rizvi, Hadi Mirhedayati Roudsari, James K. Min, and Fay Y. Lin
52 Cardiac CT: Comparative Cost-­Effectiveness. . . . . . . . . . . . . . . . . . . . . . . . . 673
Christopher L. Schlett
53 Barriers to Greater Clinical Implementation. . . . . . . . . . . . . . . . . . . . . . . . . . 681
David C. Levin
54 A Test on the Move: Cardiac CT in China as a Case Study. . . . . . . . . . . . . . 689
Bin Lu, Weihua Yin, Xinshuang Ren, and Siyu Chen

Part XIII Where We Are Going: The Genes and the Heart

55 Differences and Disparities in Cardiovascular Medicine
Related to Gender, Race, and Ethnicity: The Role of Cardiac CT. . . . . . . . . 707
John W. Nance
56 Cardiac CT Radiomics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 715
Márton Kolossváry and Pál Maurovich-Horvat
57 Advanced Methods for Coronary Artery Plaque Analysis. . . . . . . . . . . . . . . 725
Pál Maurovich-Horvat and Udo Hoffmann

Part XIV Where We Are Going: Risk Prediction and Management:

The Next Wave
58 Coronary CT Angiography for Screening, Risk Stratification,
and Management of Asymptomatic Patients: State of the Evidence. . . . . . . 739
Felix G. Meinel and Matthias Renker
Contents xvii

Part XV Where We Are Going: Lesion-specific Ischemia, Infarction, and Viability

59 Transluminal Attenuation Gradient and Other CT Techniques
for Gauging Lesion Significance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 749
Yeon Hyeon Choe, Jin-Ho Choi, and Sung Mok Kim
60 CT Angiography-Derived Fractional Flow Reserve . . . . . . . . . . . . . . . . . . . . 767
Adriaan Coenen, Frank Gijsen, and Koen Nieman
61 CT Myocardial Perfusion Imaging: Arterial First-Pass Imaging. . . . . . . . . . 777
Florian Schwarz, Amadeus Altenburger, Michael Gebhard, and Christian Thilo
62 Myocardial Perfusion Imaging: Dual-­Energy Approaches. . . . . . . . . . . . . . . 791
Domenico De Santis, Marwen Eid, Taylor M. Duguay, and Carlo N. De Cecco
63 Dynamic Myocardial CT Perfusion Imaging. . . . . . . . . . . . . . . . . . . . . . . . . . 811
Marly van Assen, Gert Jan Pelgrim, and Rozemarijn Vliegenthart
64 CT’s Role for Myocardial Viability Assessment. . . . . . . . . . . . . . . . . . . . . . . . 829
Ahmed Hamdy and Kakuya Kitagawa

Part XVI Where We Are Going: The Road Ahead

65 Coronary CT Angiography as the Gatekeeper to the Cath Lab:
Where Are We?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 849
Christoph Artzner, Lynne M. Hurwitz, and Fabian Bamberg
66 3D Printing from Cardiac CT Images. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 859
Karin E. Dill, Leonid Chepelev, Todd Pietila, and Frank J. Rybicki
67 Future Technological Advances in Cardiac CT. . . . . . . . . . . . . . . . . . . . . . . . 873
Thomas G. Flohr, Thomas Allmendinger, Herbert Bruder, Chris Schwemmer,
Steffen Kappler, and Bernhard Schmidt
68 Machine Learning and Artificial Intelligence in
Cardiovascular Imaging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 893
Marwen Eid, James V. Spearman, Marly van Assen,
Domenico De Santis, Pooyan Sahbaee, Scott P. Landreth,
Brian Jacobs, and Carlo N. De Cecco
Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 909
Contributors

Suhny  Abbara, MD, FACR, FSCCT Department of Radiology, Cardiothoracic Imaging

Division, UT Southwestern Medical Center, Dallas, TX, USA
Moritz H. Albrecht, MD  Department of Diagnostic and Interventional Radiology, University
Hospital Frankfurt, Frankfurt am Main, Germany
Hatem  Alkadhi, MD, MPH, EBCR  Institute of Diagnostic and Interventional Radiology,
University Hospital Zurich, Zurich, Switzerland
Thomas  Allmendinger, PhD Department of Computed Tomography, Siemens Healthcare
GmbH, Forchheim, Germany
Amadeus Altenburger, Dr. med  Department of Diagnostic and Interventional Radiology and
Neuroradiology, Klinikum Augsburg, Augsburg, Germany
Kweku  Appau, MD Department of Cardiology, Emory University School of Medicine,
Atlanta, GA, USA
Christoph  Artzner, MD Department of Radiology, Duke University Medical Center,
Durham, NC, USA
Department of Diagnostic and Interventional Radiology, University of Tuebingen, Tubingen,
Germany
Anil Attili, MD  Department of Radiology, Division of Cardiothoracic Radiology, University
of Michigan, Ann Arbor, MI, USA
Adefolakemi Babatunde, MD  Cardiovascular Division, Department of Medicine, Washington
University School of Medicine, St. Louis, MO, USA
Fabian  Bamberg, MD, MPH Department of Diagnostic and Interventional Radiology,
University of Tuebingen, Tubingen, Germany
Gorka  Bastarrika, MD, PhD, EBCR Cardiothoracic Imaging Division, Department of
Radiology, Clínica Universidad de Navarra, Pamplona, Spain
Richard  R.  Bayer II, MD  Division of Cardiovascular Imaging, Department of Radiology
and Radiological Science, Medical University of South Carolina, Charleston, SC, USA
Division of Cardiology, Department of Medicine, Medical University of South Carolina,
Charleston, SC, USA
Raveen Bazaaz, MD  Heart and Vascular Institute, University of Pittsburgh Medical Center,
Pittsburgh, PA, USA
Hans-Christoph R. F. Becker, MD  Department of Radiology, Stanford University, Stanford,
CA, USA
Daniel Berman, MD  Department of Imaging and Medicine and the Burns Allen Research
Institute, Cedars-Sinai Medical Center, Los Angeles, CA, USA

xix
xx Contributors

Philipp Blanke, MD  Department of Radiology, University of British Columbia, Vancouver,

BC, Canada
Daniel T. Boll, MD  Department of Radiology, University of Basel, Basel, Switzerland
Marques L. Bradshaw, MD  Department of Radiology and Radiological Sciences, Vanderbilt
University School of Medicine, Nashville, TN, USA
Herbert Bruder, PhD  Department of Computed Tomography, Siemens Healthcare GmbH,
Forchheim, Germany
Patricia Carrascosa, MD, PhD  Department of Cardiovascular Imaging, Diagnóstico Maipú,
Buenos Aires, Argentina
Damiano Caruso, MD  Department of Radiological Sciences, Oncological and Pathological
Sciences, University of Rome “Sapienza”, Latina, Italy
Csilla  Celeng, MD Department of Radiology, Heart and Vascular Center, Semmelweis
University, Budapest, Hungary
Siyu  Chen, BS  Department of Radiologic Imaging, Fuwai Hospital, Chinese Academy of
Medical Sciences and Peking Union Medical College, National Center of Cardiovascular
Diseases, Beijing, China
Yundai Chen, MD, PhD  Division of Cardiovascular Imaging, Department of Radiology and
Radiological Science, Medical University of South Carolina, Charleston, SC, USA
Department of Cardiology, People’s Liberation Army General Hospital, Beijing, China
Leonid  Chepelev, MD, PhD Department of Radiology, University of Ottawa Faculty of
Medicine, Ottawa, ON, Canada
Ottawa Hospital Research Institute, Ottawa, ON, Canada
Salvatore A. Chiaramida, MD  Division of Cardiology, Department of Medicine, Medical
University of South Carolina, Charleston, SC, USA
Yeon  Hyeon  Choe, MD, PhD Department of Radiology, HVSI Imaging Center, Heart
Vascular Stroke Institute, Samsung Medical Center, Sungkyunkwan University School of
Medicine, Seoul, South Korea
Byoung  Wook  Choi, MD, PhD Department of Radiology, Yonsei University School of
Medicine, Severance Hospital, Seoul, South Korea
Jin-Ho  Choi, MD Division of Cardiology, Department of Medicine (J-H C), Samsung
Medical Center, Sungkyunkwan University School of Medicine, Seoul, South Korea
Shahryar  M.  Chowdhury, MD, MSCR  Division of Pediatric Cardiology, Department of
Pediatrics, Medical University of South Carolina, Charleston, SC, USA
Adriaan  Coenen, MD Departments of Radiology and Cardiology, Erasmus University
Medical Center, Rotterdam, The Netherlands
Andrew Crean, BSc, BM, MRCP, MSc, FRCR, MPhil  Department of Cardiology, Ottawa
Heart Institute, University of Ottawa, Ottawa, ON, Canada
Marco Das, MD, PhD  Department of Radiology and Nuclear Medicine, Maastricht University
Medical Center, Maastricht, The Netherlands
CARIM School for Cardiovascular Diseases, Maastricht University Medical Center, Maastricht,
The Netherlands
Carlo N. De Cecco, MD, PhD  Department of Radiology and Radiological Science, Medical
University of South Carolina, Charleston, SC, USA
Contributors xxi

Domenico De Santis, MD  Department of Radiological Sciences, Oncological and Pathological

Sciences, University of Rome “Sapienza”, Latina, Italy
Division of Cardiovascular Imaging, Department of Radiology and Radiological Science,
Medical University of South Carolina, Charleston, SC, USA
Annemarie M. den Harder, MD, PhD  Department of Radiology, University Medical Center
Utrecht, Utrecht, The Netherlands
Karin E. Dill, MD  Department of Radiology, UMass Medical Center, Worcester, MA, USA
Ian  R.  Drexler, MD, MBA  Department of Radiology, Weill Cornell Medicine, New York,
NY, USA
Taylor M. Duguay, BS  Division of Cardiovascular Imaging, Department of Radiology and
Radiological Science, Medical University of South Carolina, Charleston, SC, USA
Ullrich  Ebersberger, MD  Department of Cardiology and Intensive Care Medicine, Heart
Center Munich-Bogenhausen, Munich, Germany
Division of Cardiovascular Imaging, Department of Radiology and Radiological Science,
Medical University of South Carolina, Charleston, SC, USA
Heidi A. Edmonson, PhD  Department of Radiology, Mayo Clinic, Rochester, MN, USA
Marwen  Eid, MD Division of Cardiovascular Imaging, Department of Radiology and
Radiological Science, Medical University of South Carolina, Charleston, SC, USA
Ahmed Hamdy, MD  Department of Radiology, Mie University Hospital, Tsu, Japan
Christian  D.  Eusemann, PhD  Department of Collaborations, Siemens Medical Solutions
USA, Inc., Malvern, PA, USA
Lloyd M. Felmly, MD  Department of Surgery, Division of Cardiothoracic Surgery, Medical
University of South Carolina, Charleston, SC, USA
Gudrun  M.  Feuchtner, MD Department of Radiology, Innsbruck Medical University,
Innsbruck, Austria
Aloke V. Finn, MD  CVPath Institute, University of Maryland, Gaithersburg, MD, USA
Thomas G. Flohr, PhD  Department of Computed Tomography, Siemens Healthcare GmbH,
Forchheim, Germany
Kunihiko Fukuda, MD  Department of Radiology, The Jikei University School of Medicine,
Tokyo, Japan
Michael  Gebhard, MD Department of Diagnostic and Interventional Radiology and
Neuroradiology, Klinikum Augsburg, Augsburg, Germany
Lucas  L.  Geyer, MD Division of Cardiovascular Imaging, Department of Radiology and
Radiological Science, Medical University of South Carolina, Charleston, SC, USA
Institute for Clinical Radiology, Ludwig-Maximilians-University Hospital Munich, Munich,
Germany
Brian  B.  Ghoshhajra, MD, MBA Division of Cardiovascular Imaging, Department of
Radiology, Massachusetts General Hospital/Harvard Medical School, Boston, MA, USA
Cardiac MR PET CT Program, Massachusetts General Hospital, Boston, MA, USA
Frank  Gijsen, PhD  Department of Biomedical Engineering, Erasmus University Medical
Center, Rotterdam, The Netherlands
Harold Goerne, MD  Department of Radiology, Cardiothoracic Imaging, UT Southwestern
Medical Center, Dallas, TX, USA
xxii Contributors

Orly Goitein, MD  Department of Diagnostic Imaging, Chaim Sheba Medical Center, Ramat
Gan, Israel
Katharine Grant, PhD  Siemens Medical Solutions USA, Inc., Malvern, PA, USA
Daniel  B.  Green, MD  Department of Radiology, Weill Cornell Medicine, New York, NY,
USA
David Gregg, MD  Division of Cardiology, Department of Medicine, Medical University of
South Carolina, Charleston, SC, USA
B. Kelly Han, MD  Department of Advanced Congenital Cardiac Imaging, Minneapolis Heart
Institute and the Children’s Hospitals and Clinics of Minnesota, Minneapolis, MN, USA
Thomas  Henzler, MD Institute of Clinical Radiology and Nuclear Medicine, University
Medical Center Mannheim, Medical Faculty Mannheim, Heidelberg University, Heidelberg,
Germany
Anthony  M.  Hlavacek, MD Department of Pediatrics, Division of Pediatric Cardiology,
Medical University of South Carolina, Charleston, SC, USA
Udo  Hoffmann, MD Department of Cardiovascular Imaging, Massachusetts General
Hospital, Boston, MA, USA
Jin  Hur, MD, PhD Department of Radiology, Yonsei University School of Medicine,
Severance Hospital, Seoul, South Korea
Lynne  M.  Hurwitz, MD Department of Radiology, Duke University Medical Center,
Durham, NC, USA
Gregory  Jackson, MD  Department of Cardiology, Medical University of South Carolina,
Charleston, SC, USA
Brian Jacobs, BS  Department of Radiology and Radiological Science, Medical University of
South Carolina, Charleston, SC, USA
Michael  A.  Kadoch, MD  Department of Radiology, University of California, Davis, CA,
USA
Steffen  Kappler, Dr. rer. Nat  Department of Computed Tomography, Siemens Healthcare
GmbH, Forchheim, Germany
Ella A. Kazerooni, MD, MS, FACR  Department of Radiology, Division of Cardiothoracic
Radiology, University of Michigan, Ann Arbor, MI, USA
Sung Mok Kim, MD  Department of Radiology, HVSI Imaging Center, Heart Vascular Stroke
Institute, Seoul, South Korea
Kakuya Kitagawa, MD, PhD  Department of Radiology, Mie University Hospital, Tsu, Japan
Friedrich Knollmann, MD, PhD  Department of Radiology, UCDavis Health, Sacramento,
CA, USA
Brian S. Ko, MBBS (Hons), PhD  Department of Medicine Monash Medical Centre (MMC),
Monash Cardiovascular Research Centre, MonashHEART, Monash Health and Monash
University, Melbourne, Australia
James M. Kofler, PhD  Department of Radiology, Mayo Clinic, Jacksonville, FL, USA
Madeleine  Kok, MD, PhD Department of Radiology and Nuclear Medicine, Maastricht
University Medical Center, Maastricht, The Netherlands
CARIM School for Cardiovascular Diseases, Maastricht University Medical Center, Maastricht,
The Netherlands
Contributors xxiii

Frank D. Kolodgie, PhD  CVPath Institute, Gaithersburg, MD, USA

Márton  Kolossváry, MD Cardiovascular Imaging Research Group, Heart and Vascular
Center, Semmelweis University, Budapest, Hungary
Eli  Konen, MD  Department of Diagnostic Imaging, Chaim Sheba Medical Center, Ramat
Gan, Israel
B. Krauss, PhD  Siemens Healthcare GmbH, Forchheim, Germany
Shaw  Hua  Kueh, MBChB Department of Radiology, University of British Columbia,
Vancouver, BC, Canada
Joan M. Lacomis, MD  Department of Radiology, Thoracic Imaging Division, University of
Pittsburgh Medical Center, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA
Scott P. Landreth  Department of Radiology and Radiological Science, Medical University of
South Carolina, Charleston, SC, USA
Hwa Yeon Lee, MD, PhD  Smile Radiologic Clinic, Seoul, South Korea
Jongmin Lee, MD, PhD  Department of Radiology, Kyungpook National University, School
of Medicine, Daegu, South Korea
Nam Ju Lee, MD  Department of Radiology, Mayo Clinic, Rochester, MN, USA
Alan  C.  Legasto, MD  Department of Radiology, Weill Cornell Medicine, New York, NY,
USA
Jonathon  A.  Leipsic, MD Department of Radiology, St. Paul’s Hospital, Providence
Healthcare, Vancouver, BC, Canada
Shuai Leng, PhD  Department of Radiology, Mayo Clinic, Rochester, MN, USA
John R. Lesser, MD, FACC, FAHA  Minneapolis Heart Institute Foundation, Minneapolis,
MN, USA
Virginia W. Lesslie, BS  Division of Cardiovascular Imaging, Department of Radiology and
Radiological Science, Medical University of South Carolina, Charleston, SC, USA
David  C.  Levin, MD Department of Radiology, Thomas Jefferson University Hospital,
Philadelphia, PA, USA
Ana Paula S. Lima, MD  Department of Radiology and Biomedical Imaging, University of
California San Francisco, San Francisco, CA, USA
Fay  Y.  Lin, MD  Department of Radiology, NewYork-Presbyterian Hospital and the Weill
Cornell Medical College, New York, NY, USA
Harold  Litt, MD, PhD Department of Radiology, Perelman School of Medicine of the
University of Pennsylvania, Philadelphia, PA, USA
Sheldon E. Litwin, MD  Division of Cardiology, Department of Medicine, Medical University
of South Carolina, Charleston, SC, USA
Bin  Lu, MD Department of Radiologic Imaging, Fuwai Hospital, Chinese Academy of
Medical Sciences and Peking Union Medical College, National Center of Cardiovascular
Diseases, Beijing, China
Guang Ming Lu, MD  Department of Medical Imaging, Jinling Hospital, Medical School of
Nanjing University, Nanjing, Jiangsu, China
Manoj Mannil, MD, MSc  Institute of Diagnostic and Interventional Radiology, University
Hospital Zurich, Zurich, Switzerland
xxiv Contributors

Mohamed  Marwan, MD  Department of Medicine, Cardiology and Angiology,

Universitätsklinikum Erlangen, Erlangen, Germany
Pál  Maurovich-Horvat, MD, PhD, MPH  Cardiovascular Imaging Research Group, Heart
and Vascular Center, Semmelweis University, Budapest, Hungary
C. H. McCollough, PhD  Department of Radiology, Mayo Clinic, Rochester, MN, USA
Felix  G.  Meinel, MD Department of Diagnostic and Interventional Radiology, Rostock
University Medical Center, Rostock, Germany
Nandini  (Nina)  M.  Meyersohn, MD  Division of Cardiovascular Imaging, Department of
Radiology, Massachusetts General Hospital/Harvard Medical School, Boston, MA, USA
Casper  Mihl, MD, PhD Department of Radiology and Nuclear Medicine, Maastricht
University Medical Center, Maastricht, The Netherlands
CARIM School for Cardiovascular Diseases, Maastricht University Medical Center, Maastricht,
The Netherlands
James K. Min, MD  Department of Radiology, NewYork-Presbyterian Hospital and the Weill
Cornell Medical College, New York, NY, USA
John F. Mooney, MBB  Department of Radiology, University of British Columbia, Vancouver,
BC, Canada
Hiroyoshi Mori, MD  CVPath Institute, Gaithersburg, MD, USA
Pamela  B.  Morris, MD Seinsheimer Cardiovascular Health Program, Department of
Medicine and Cardiology, Medical University of South Carolina, Charleston, SC, USA
Giuseppe Muscogiuri, MD  Centro Cardiologico Monzino, IRCCS, Milan, Italy
John  W.  Nance, MD Department of Radiology, Medical University of South Carolina,
Charleston, SC, USA
Koen Nieman, MD, PhD  Stanford University, School of Medicine, Cardiovascular Institute,
Stanford, CA, USA
Iclal  Ocak, MD Department of Radiology, Thoracic Imaging Division, University of
Pittsburgh Medical Center, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA
Maksymilian P. Opolski, MD, PhD  Department of Interventional Cardiology and Angiology,
Institute of Cardiology, Warsaw, Poland
Karen  G.  Ordovas, MD, MAS Department of Radiology, University of California San
Francisco, San Francisco, CA, USA
Kotaro Ouchi, MD, PhD  Department of Radiology, The Jikei University School of Medicine,
Tokyo, Japan
Gert  Jan  Pelgrim, PhD Department of Radiology, University of Groningen, University
Medical Center Groningen, Groningen, The Netherlands
Laura  J.  Pierce, MPA, RT (CT) Department of Radiology, Duke University School of
Medicine, Durham, NC, USA
Todd Pietila  Materialise USA Biomedical Engineering, Plymouth, MI, USA
Gianluca  Pontone, MD, PhD, FESC, FSCCT Centro Cardiologico Monzino, IRCCS,
Milan, Italy
Andrew  N.  Primak, PhD Department of Diagnostic Imaging, Siemens Healthineers,
Malvern, PA, USA
Contributors xxv

Gregorio  Rábago, MD Department of Cardiac Surgery, Clínica Universidad de Navarra,

Pamplona, Spain
Mark  Rabbat, MD, FSCCT Department of Medicine, Division of Cardiology, Loyola
University Chicago, Chicago, IL, USA
Department of Medicine, Division of Cardiology, Edward Hines Jr. VA Hospital, Hines, IL,
USA
Prabhakar Rajiah, MBBS, MD, FRCR  Department of Radiology, Cardiothoracic Imaging
Division, UT Southwestern Medical Center, Dallas, TX, USA
Matthias Renker, MD  Department of Cardiology, Kerckhoff Heart and Thorax Center, Bad
Nauheim, Germany
Xinshuang Ren, MD  Department of Radiologic Imaging, Fuwai Hospital, Chinese Academy
of Medical Sciences and Peking Union Medical College, National Center of Cardiovascular
Diseases, Beijing, China
Asim  Rizvi, MD Department of Radiology, Dalio Institute of Cardiovascular Imaging,
NewYork-Presbyterian Hospital and the Weill Cornell Medical College, New York, NY, USA
Department of Radiology, Mayo Clinic, Rochester, MN, USA
Sebastian  Rogowski, MD  Department of Cardiology and Intensive Care Medicine, Heart
Center Munich-Bogenhausen, Munich, Germany
Hadi Mirhedayati Roudsari, MD  Department of Radiology, NewYork-Presbyterian Hospital
and the Weill Cornell Medical College, New York, NY, USA
Alan  Rozanski, MD Department of Cardiovascular Medicine, Gagnon Cardiovascular
Institute, Morristown Medical Center, Morristown, NJ, USA
The Division of Cardiology, Mount Sinai St. Luke’s Hospital, Mount Sinai Heart, New York,
NY, USA
Geoffrey  D.  Rubin, MD, MBA Department of Radiology, Duke University School of
Medicine, Durham, NC, USA
Ronen  Rubinshtein, MD Department of Cardiovascular Medicine, Lady Davis Carmel
Medical Center, The Ruth and Bruce Rappaport School of Medicine, Technion-Israel Institute
of Technology, Haifa, Israel
John A. Rumberger, PhD, MD  Department of Cardiac Imaging, The Princeton Longevity
Center, Princeton, NJ, USA
Frank  J.  Rybicki, MD, PhD Department of Radiology, University of Ottawa Faculty of
Medicine, Ottawa, ON, Canada
Ottawa Hospital Research Institute, Ottawa, ON, Canada
Toru Sakuma, MD, PhD  Department of Radiology, The Jikei University School of Medicine,
Tokyo, Japan
Arnold M. R. Schilham, PhD  Department of Radiology, University Medical Center Utrecht,
Utrecht, The Netherlands
Christopher L. Schlett, MD, MPH  Department of Diagnostic and Interventional Radiology,
University Medical Center Freiburg, Freiburg, Germany
Bernhard Schmidt, PhD  Department of Computed Tomography, Siemens Healthcare GmbH,
Forchheim, Germany
Paul  Schoenhagen, MD  Imaging Institute, Cleveland Clinic, Lerner College of Medicine,
Cleveland, OH, USA
xxvi Contributors

Jan-Erik  Scholtz, MD Cardiac MR PET CT Program, Massachusetts General Hospital,

Boston, MA, USA
Florian Schwarz, PD Dr. med  Department of Diagnostic and Interventional Radiology and
Neuroradiology, Klinikum Augsburg, Augsburg, Germany
Chris Schwemmer, MSc  Department of Computed Tomography, Siemens Healthcare GmbH,
Forchheim, Germany
Stephanie Sellers, MSc  Department of Radiology, University of British Columbia, Vancouver,
BC, Canada
Michael  D.  Shapiro, DO  Department of Medicine, Division of Cardiovascular Medicine,
Oregon Health & Science University, Portland, OR, USA
James  V.  Spearman, MD Department of Radiology and Radiological Science, Medical
University of South Carolina, Charleston, SC, USA
David  Steflik, MD Department of Pediatrics, Division of Pediatric Cardiology, Medical
University of South Carolina, Charleston, SC, USA
Arthur  E.  Stillman, MD, PhD Department of Radiology and Imaging Science, Emory
University School of Medicine, Atlanta, GA, USA
Pal Spruill Suranyi, MD, PhD  Division of Cardiovascular Imaging, Department of Radiology
and Radiological Science, Medical University of South Carolina, Charleston, SC, USA
Division of Cardiology, Department of Medicine, Medical University of South Carolina,
Charleston, SC, USA
Richard A. P. Takx, MD, MSc, PhD  Department of Radiology, University Medical Center
Utrecht, Utrecht, The Netherlands
Christian Tesche, MD  Division of Cardiovascular Imaging, Department of Radiology and
Radiological Science, Medical University of South Carolina, Charleston, SC, USA
Department of Cardiology and Intensive Care Medicine, Heart Center Munich-Bogenhausen,
Munich, Germany
Christian  Thilo, MD Department of Cardiology, Klinikum Augsburg, Herzzentrum
Augsburg-Schwaben, Augsburg, Germany
Quynh A. Truong, MD, MPH, FACC  Department of Radiology, Weill Cornell Medicine,
New York, NY, USA
Seth Uretsky, MD  Department of Cardiovascular Medicine, Gagnon Cardiovascular Institute,
Morristown Medical Center, Morristown, NJ, USA
Marly van Assen, MSc Department of Radiology, University of Groningen, University
Medical Center Groningen, Groningen, The Netherlands
Department of Radiology and Radiological Science, Division of Cardiovascular Imaging,
Medical University of South Carolina Charleston, SC, USA
Edwin J. R. van Beek, MD, PhD  Edinburgh Imaging Facility QMRI, University of Edinburgh,
Edinburgh, UK
Akos Varga-Szemes, MD, PhD  Division of Cardiovascular Imaging, Department of Radiology
and Radiological Science, Medical University of South Carolina, Charleston, SC, USA
Renu Virmani, MD  CVPath Institute, Gaithersburg, MD, USA
Contributors xxvii

Rozemarijn  Vliegenthart, MD, PhD  Department of Radiology, University of Groningen,

University Medical Center Groningen, Groningen, The Netherlands
Department of Radiology and Radiological Science, Division of Cardiovascular Imaging,
Medical University of South Carolina Charleston, SC, USA
Andrew  Voigt, MD  Heart and Vascular Institute, University of Pittsburgh Medical Center,
Pittsburgh, PA, USA
Charles S. White, MD  Department of Diagnostic Radiology, University of Maryland Medical
Center, Baltimore, MD, USA
Joachim  E.  Wildberger, MD, PhD Department of Radiology and Nuclear Medicine,
Maastricht University Medical Center, Maastricht, The Netherlands
CARIM School for Cardiovascular Diseases, Maastricht University Medical Center, Maastricht,
The Netherlands
Martin  J.  Willemink, MD, PhD Department of Radiology, University Medical Center
Utrecht, Utrecht, The Netherlands
Michelle  C.  Williams, MBChB, PhD Edinburgh Imaging Facility QMRI, University of
Edinburgh, Edinburgh, UK
Eric E. Williamson, MD  Department of Radiology, Mayo Clinic, Rochester, MN, USA
Arik Wolak, MD  Department of Cardiology, Shaare Zedek Medical Center, Jerusalem, Israel
Pamela  K.  Woodard, MD Mallinckrodt Institute of Radiology, Washington University
School of Medicine, St. Louis, MO, USA
Junjie  Yang, MD Division of Cardiovascular Imaging, Department of Radiology and
Radiological Science, Medical University of South Carolina, Charleston, SC, USA
Department of Cardiology, People’s Liberation Army General Hospital, Beijing, China
Weihua Yin, MD  Department of Radiologic Imaging, Fuwai Hospital, Chinese Academy of
Medical Sciences and Peking Union Medical College, National Center of Cardiovascular
Diseases, Beijing, China
Seung  Min  Yoo, MD, PhD  Department of Radiology, CHA University Bundang Medical
Center, Bundang, South Korea
Long Jiang Zhang, MD, PhD  Department of Medical Imaging, Jinling Hospital, Medical
School of Nanjing University, Nanjing, Jiangsu, China
Mathis  Zimmermann, Dipl Ing, MBA Digital Health Services, Siemens Healthineers,
Malvern, PA, USA
Peter L. Zwerner, MD  Division of Cardiology, Department of Medicine, Medical University
of South Carolina, Charleston, SC, USA
 Part I
Where We Were
 History of Cardiac CT: A Personal Story
1
John A. Rumberger

As the story goes, Wilhelm Conrad Röentgen, a physicist, with cardiac CT interestingly started nearly at the same time
was working late in his laboratory in Wurzburg, Germany, as the development of commercial CT in general.
experimenting with a vacuum tube made of glass. He was
using this to generate beams of electrons and wrapped the
tube with black paper to avoid viewing the electric discharge In the Beginning, There Was Mathematics
occurring in the gas inside the vacuum tube. When he started
his experiment, he noted that a piece of coated paper lying The story of cardiac CT, and all CT for that matter, begins
near the tube began to glow. He was astonished and did with mathematics that allow us to “reconstruct” the density/
another experiment where he held a thick book between the tissue characteristics of an “unknown” object placed in a
tube and the paper  – however, the “rays” simply passed black box as light [or later X-ray] of known intensity is
through the book, as if totally unobstructed. When Röentgen shown through. Pierre Bouguer, credited in about 1729,
looked at the coated paper it showed a shadowy outline of the noted that the absorption of light through an object is directly
bones in his hand. This was November 8, 1895, and the proportional to its thickness [or path length]. Lambert later
world of the “X-ray” has never looked back. popularized this observation in a paper from 1760. August
The “X-ray” has been intimately linked to the ability to Beer discovered another light attenuation factor in 1852 not-
see “inside” the body since the late nineteenth century; but it ing that light absorbance was proportional to the concentra-
remained a projection image with superposition of all the tions of the attenuating “unknown object.” The modern
densities of the tissue placed between the anode and the cath- Lambert-Beer law combines these two observations and cor-
ode. To separate these various tissue densities, a thin cross-­ relates the changes in light energy to both the concentration
sectional image would be of significant benefit as the various of the attenuating “unknown” object and the thickness of the
organs can be separated from their surrounding tissues of fat, unknown object. Since visible light is part of the electromag-
muscle, and bone. netic energy spectrum, this can apply to the application of
The birth of clinical X-ray computed tomography (CT) X-rays as well. The general application of the Lambert-Beer
was not realized until about 80 years after Röentgen’s dis- law to X-ray imaging is shown in Fig. 1.1.
covery. At the time of this writing, an estimated 90,000 peer Referring to Fig.  1.1, the unknown object of density μ
reviewed scientific articles have been published on or about represents a “pixel” [i.e., picture element] size of 1 × 1. If we
CT.  A dominant majority of these articles deal with body know the incident radiation and the output radiation after
organs and processes that either do not move during the passing through the unknown object, as well as the distance
image acquisition or, in the case of lung imaging, when between these measurements, the Lambert-Beer law then
motion can be suspended long enough to get “static images.” provides an equation with only one unknown, i.e., μ. What if
In the case of cardiac CT imaging, however it is a different I wanted to determine the density of four unknown μ objects?
story and has been a difficult challenge to image an object I can use the Lambert-Beer law to set up four equations in
that is constantly moving in four dimensions and cannot be, four unknowns or a field of view of 2 × 2 pixels as shown in
safely, stopped. Cardiac CT and my personal involvement Fig. 1.2. However, the ability to solve pixel density resolu-
tions of 80 × 80 pixels [as was used on the first-generation
clinical CT scanner] was simply too daunting a task until the
J. A. Rumberger (*) modern development of the computer. The “exact solutions”
Department of Cardiac Imaging, The Princeton Longevity Center, for the individual μ, to speed up the mathematical solutions,
Princeton, NJ, USA
e-mail: [email protected]

© Humana Press 2019 3

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_1
4 J. A. Rumberger

How can we define the density of an  ttainment of Reality in Clinical Medicine:

A
unknown object by x-ray?
The EMI Scanner
It = attenuated energy
Modern X-ray CT was developed by Sir Godfrey Hounsfield
It = Iie-µx Ii = input energy while working for Electronic and Musical Industries Ltd.
[EMI] in England. A prototype scanner using an X-ray tube
x = path length was developed in 1969/1970 and a clinically applicable
scanner installed at the Atkinson Morley Hospital in a
µ = absorption coefficient
London suburb in 1971; the first clinical results were pre-
sented in 1973 [4]. At first only the brain could be imaged
due to very long acquisition times in which the patient was
µ required to be very still [and surrounded by a water phan-
Ii It
tom]. The first clinical CT scanner in the USA was installed
at the Mayo Clinic in Rochester, Minnesota, in 1973.
As an interesting personal anecdote, the engineer install-
ing the scanner at the Mayo Clinic was named David King
x
and had worked closely with Hounsfield in England. David
King, later the founder of “calcium club” [see further discus-
You can then solve for “µ” if you know the other values sion below] and acknowledged as the “father” of coronary
artery calcium scanning, told me a story. When Hounsfield
Fig. 1.1  The Lambert-Beer law applied to X-ray and colleagues at EMI contemplated the “world’s” eventual
needs for such a unique brain imaging device, they estimated
that probably less than a hundred scanners would be eventu-
ally made and sold. After 1 week at the Mayo Clinic, and
Ii
Ii

after performing more head/brain CT scans than had been

Ita=Iie-(µ1 + µ2)x
done in the past 2 years in England, David told Godfrey –
Itb=Iie-(µ3 + µ4)x “Maybe you might want to increase your estimation of the
Ii µ1 µ2
Ita world’s need for the EMI scanner”.
Itc=Iie-(µ1 + µ3)x By 1975 a second-generation EMI scanner, the first proto-
Ii µ3 µ4 Itb Itd=Iie-(µ2 + µ4)x type “body scanner” was introduced. Acquisition times per
slice were about 20  s, and it used iterative reconstruction
techniques although the much faster filtered back projection
4 equations in 4 unknowns: method was now established. Because of the success of the
exact solutions for µ
EMI scanner, many other commercially available scanners
Itc

Itd

were quickly introduced by other manufacturers that, like

Fig. 1.2  The exact solution of the Lambert-Beer law for four unknown EMI itself, eventually went under, while others such as the
objects of density μ Picker, Siemens, and GE survived. But the die was cast. In
1979 both Hounsfield and Cormack received the Nobel Prize
in Physiology and Medicine for the development of the CT
initially involved various “iterative” methods. The final scanner.
­solutions reached in a series of best guesses and comparisons
with the actual data. The original mathematics was part of
the ART [algebraic reconstruction theory] algorithm [1]. The Dynamic Spatial Reconstructor
In 1963 another physicist in South Africa, Allan
Cormack, was working on improving the dose calculations There were several attempts to using “conventional” CT
used in radiation therapy planning, but knowledge of scanning in the early 1980s to study the cardiovascular sys-
cross-sectional density distributions was required. He tem, mainly viewing patency of coronary artery bypass grafts
developed the first concept of image reconstruction from and looking for aortic dissections [5]. Already there was a
projections [2]. This became the basis for another image clear advantage noted in cardiac CT over the conventional
reconstruction called “back projection” [later improved to M-mode/sector echocardiographic examinations and plane
reduce noise at the edges of objects and called “filtered chest X-rays that were “state of the art” at the time for
back projection”] [3]. ­imaging of the chest and heart.
1  History of Cardiac CT: A Personal Story 5

Fig. 1.3  The dynamic spatial

reconstructor. (With
permission of the Mayo
Foundation for Medical
Education and Research. All
rights reserved)

The Biodynamics Research Laboratory at the Mayo approaching 1 mm. However, only 14 sets of X-ray/fluoro-
Clinic had a long and storied history with aviation medicine scopic units were installed due to technical difficulties and
in World War II and under the directorship of Dr. Earl funding. Later the fluoroscopic units were replaced by CCDI
H. Wood did the first human centrifuge experiments studying cameras. Image reconstruction however was arduous and
“G” force [i.e., gravity] effects on pilots during flight and could take as long as several weeks to be completed.
combat. Along with the US government, they developed the Specialized software was developed to review and analyze
first “G-suits,” and after the war Dr. Wood was awarded a the data. Countless contributions to the world of cardiac CT
special commendation by President Harry S. Truman. were introduced by the investigators working on multiple
The idea of using X-ray CT to study the moving heart aspects of the DSR project [6–8]. I can recall at one of my
dated to the first body images produced using the EMI body earliest American Heart Association conventions as a cardi-
scanner; but despite some success as noted by Brundage ology fellow seeing the astounding video presented by Dr.
et al. [5], the spatial resolution and most importantly the tem- Eric Ritman, the then head of the biodynamics research unit,
poral resolution were not sufficient for most clinical cardiac showing detailed anatomy of adult patients in 3-D and, with
work. The idea of developing a CT scanner fast enough to time added, in 4-D. Unfortunately, the DSR was not a com-
make “stop action” images of the beating heart was first real- mercially viable enterprise and was decommissioned in
ized in the Biodynamics Research Laboratory where, under 2002. By that time cardiac imaging using MDCT was devel-
NIH funding, they introduced the Dynamic Spatial oping rapidly.
Reconstructor (DSR) in 1975.
The DSR was imagined as a specialized cardiac CT scan-
ner using conventional X-rays with photomultiplier tubes Electron Beam CT
and fluoroscopic projection imaging applied in a unique
manner (Fig. 1.3). The original design was to use 28 pairs of In the early 1970s, the initial applications to develop a spe-
X-ray sources and 28 direct line visualization fluoroscopic cialized cardiac CT scanner given to the NIH were of two
units. This vast array [requiring literally two floors with gan- distinctly different designs. The X-ray tube/fluoroscopic unit
try and imaging chain] was then rotated at high speed as design as noted above was eventually used for the
images of the beating heart were acquired using intra-arterial DSR.  However, another design using scanning electron
injection of iodinated contrast over a period of about 20 s. beams was discussed.
Using ART reconstruction methods, a temporal resolution of The scanning electron beam approach eventually proved
16 msec/image could be realized as well as spatial resolution to be the superior design for practical clinical applications
6 J. A. Rumberger

and for commercial product success. However, this required moving parts, as with then and current CT scanners, the
many years in development. Two early designs, one by ­limited tube current (650  mA) required slower speeds for
Iinuma et al. [9] at JEOL, a Japanese manufacturer of elec- acceptable mAs values and associated image quality.
tron microscopes, and by Haimson [10] resulted in prototype However, with ultrafast CT temporal resolution speeds of
machines but were abandoned. Published in the same issue 50 msec and 100 msec, this was 10–20 times faster than pos-
as Reference [8], citing the initial DSR results, was a third sible at that time from conventional CT [and frankly would
design, originating with Dr. Douglas Boyd [11, 12]. This not be possible with mechanical CT until 20+ years later].
eventually resulted in the development of what is now called I started my cardiology fellowship in 1981 at the
EBCT [electron beam computed tomography]. University of Iowa under the mentorship of the late Dr.
In 1984 Dr. Boyd and colleagues, working initially Melvin Marcus. Dr. Marcus had already distinguished him-
under the aspics of the University of San Francisco Physics self as a leader in the world of coronary physiology. The ini-
Laboratory, developed a commercially viable electron tial experiments were done in laboratory animals, but he
beam scanner. A for-profit company, Imatron Inc., was longed to be able to study in detail human cardiac and coro-
developed, and then it was time to sell the scanners to aca- nary physiology. Recall this was a time early in the develop-
demic institutions for cardiac research. It was initially ments of nuclear perfusion imaging and the 30° and 60°
called “ultrafast CT.” “sector” two-dimensional echocardiograms. The only way to
The design was radically different from conventional CT evaluate for the severity of coronary artery disease was direct
at the time. The idea behind ultrafast CT was a large bell-­ invasive angiography.
shaped X-ray tube. An electron beam [think back to the ini- Mel had published a paper showing that patients with
tial experiments of Röentgen] emitted from the cathode is severe aortic stenosis and severe left ventricular hypertrophy,
focused into a narrow beam and then, by means of powerful but no evidence for obstructive coronary disease, evaluated
electromagnets, deflected to impinge on a small focal spot on in the operating room using a Doppler coronary flow meter
an annular tungsten target anode. The electron beam [and of placed directly on the coronary arteries, could result in
course the focal spot] was electronically swept along at decreased coronary artery flow reserve and angina related to
semicircular array along 210° of arc (Fig. 1.4). In order to the left ventricular hypertrophy [13]. If only he could study
perform a true cross-sectional image, the beam sweep must such phenomenon in adult patients outside of the operating
be at least 180° plus the width of the “fan beam,” which for room.
the Imatron scanner was 30°. To acquire rapid heart scans Dr. Marcus had heard of the potential for ultrafast CT to
without table movement, ultrafast CT included four anodes do noninvasive human “experiments” of cardiac structure,
and two detector arrays, each offset along the z-axis to function, and flow. The University of Iowa agreed to pur-
acquire eight interleaved slices covering 8 cm of the heart. chase Imatron scanner #3, and I was sent to the UCSF
Although very fast speeds were possible, since there were no Physics Laboratory, along with the late Dr. Andrew

Detectors

Crystal-Photodiodes

Heart
Preamplifiers

47 cm Scan Field
Radiation
Shield
Patient Cross
Section
Source Collimator
Source
collimator
Target Ring
CUUM
Chamber
Vacuum Envelope

Target Rings

Fig. 1.4  Schematic of electron beam tomography. (Personal archives, John A. Rumberger, MD)
1  History of Cardiac CT: A Personal Story 7

J.  Feiring, to begin validation studies using ultrafast CT fast CT installed sites to interest researchers in CAC
for two ­important parameters: accurate definition of global [coronary artery calcification]. At the time, we had two
left ventricular muscle mass and regional myocardial installations of EBT at the Mayo Clinic, and we were ­imaging
perfusion. cardiac physiology [as noted above] and exploring applica-
During our time in South San Francisco, I got to know, tions of fast imaging of the respiratory system [cine angio-
learn from, admire, and befriend the “greats” in early car- graphic imaging of patients with sleep apnea and exploring
diac CT including Dr. Bruce Brundage [cardiology], Dr. the application of ultrafast CT in detecting pulmonary
Charles Higgins [radiology], and Dr. Martin Lipton [radi- emboli].
ology]. After many errors, trials, false endings, and mis- David tried to interest me in CAC in the early 1990s. I had
understandings of the physics of CT and image four comments: (1) We know that CAC is associated with
reconstructions, we were successful with our two primary atherosclerosis, but it does not tell us the severity of coronary
objectives [14, 15]. stenosis, (2) I am happy studying cardiac/coronary physiol-
Over the next 15 years, our laboratory at the University of ogy that I know can be done with EBT, (3) I am not inter-
Iowa and later at the Mayo Clinic, along with scores of other ested, and (4) find somebody else at Mayo that might find
investigators all over the world, used ultrafast CT [later CAC valuable… Should such comments strike me down as I
called simply electron beam CT (EBT/EBCT)] to validate a stand.
number of potential human clinical cardiac situations includ- Dr. William Stanford and I edited the first book on
ing quantitation of left ventricular regurgitant volumes [16], Ultrafast Cardiac CT in 1992 [29]. At that time, we dis-
visualization of coronary artery bypass graft patency [17], cussed the physics of CT imaging and had many of our col-
segmental left ventricular systolic function [18], regional left leagues submit chapters on their ultrafast CT research. One
ventricular diastolic function [19, 20], regional radius to wall such chapter was from Drs. Agatston and Janowitz on the
thickness rations in normal volume overloaded left ventricle clinical applications of CAC scanning. I edited the chapter
[21], right ventricular assessment in patients under consider- and felt that the “suggestions” for applications were a bit
ation for lung transplantation [22], post-infarct left ventricu- “imaginative,” and I asked them to “tone it down” for the
lar and right ventricular remodeling during the first year after eventual publication.
myocardial infarction [23, 24], and revalidation of measure- I was aware peripherally of the research my colleagues
ments of myocardial perfusion in animal models [25] and in Dr. Robert Schwartz in collaboration with a well-known
humans [26]. cardiac pathologist, Dr. William Edwards, in looking at
Imatron, as a stand-alone provider of ultrafast CT, had CAC in autopsy hearts using EBT. At the same time, epide-
many “suiters” during the years beginning with Picker miologists from the University of Michigan, Drs. Pat Peyser
International and later with Siemens. These associations and Lawrence Bielak, were conducting studies in the
later proved to be economically fatal. “Rochester Family Heart Study” looking at CAC in resi-
Despite these and many other publications, imaging of the dents of Olmsted County, MN., in relationship to coronary
heart using CT was considered at the time as a “niche” and angiography [30].
likely not to be applied widely in clinical medicine as mag- At the Mayo Clinic, we required all our internal medi-
netic resonance imaging and two-dimensional [and eventu- cine residents to participate in at least one research project
ally three] echocardiography, both not involving exposure to during their residency. Dr. Brent Simons, who had been
ionizing radiation, gained more and more applications and involved with Drs. Schwartz and Edwards in the pathologic
notoriety. Ultrafast CT/EBT needed a “unique” application studies of CAC versus coronary plaque studies, came to me
to clinical cardiology. to discuss his data and how to best analyze the information.
There had been reports as early as the 1960s using the He basically had a listing of CAC area measurements using
presence of coronary artery calcification, detected at fluoros- EBT made at 3 mm intervals from the ostium of the autopsy
copy, as a noninvasive definition of coronary atherosclerotic coronary arteries and representative coronary histologic
plaque. Early investigations at the University of Chicago atherosclerotic plaque areas from the same coronary seg-
showed a correlation with coronary obstructive disease and ments. I was not his academic advisor but suggested we
non-quantitation of coronary artery calcification using ultra- start with a simple linear correlation of CAC areas versus
fast CT [27]. However, it was the publication of a study from histologic plaque areas. At the time the best program was
Mt. Sinai hospital by Agatston and Janowitz that set the called “Lotus 123.” I took the data and displayed the infor-
course for the use of quantitated coronary artery calcium mation. What I saw was astounding and provocative: there
score [Agatston score] as a surrogate for clinical coronary was a clear, albeit scattered, linear correlation between
artery atherosclerosis [28]. CAC area and comparable atherosclerotic plaque area. I
David King [as identified previously] was then the “scien- turned to Brent and said, quite literally: “Where the hell did
tific director” of Imatron and visited all US and foreign ultra- you get these data?”
8 J. A. Rumberger

From that point on, I had a mission, and we published MDCT

several papers from this dataset and a subsequent dataset
using non-decalcified autopsy coronary segments as com- Conventional CT in the 1980s, as opposed to using a scan-
pared with CAC images using EBT [31, 32]. My initial ning electron beam as noted above, had a disadvantage for
“bias” that CAC could not define percent stenosis was imaging the heart due to the physical time required to move
­subsequently shown to be a consequence of coronary artery a heavy X-ray tube, resulting in temporal resolution on the
remodeling during atherosclerotic plaque development [33]. order of 1000 msec, a spatial resolution of 5–6 mm, and the
As interest, all around the world, began to develop regard- physical attachment to the scanning gantry of the electrical
ing the use of CAC as a “screening” method to define coronary connection receiving the electronic data. This tethering
atherosclerosis, David King formed the “calcium club.” We required that a 230-degree arc for the rotation of the X-ray
would have meetings in the AM prior to the American College tube about the patient had to be “returned” or wound back to
of Cardiology and American Heart Association annual meet- its original position and the scanning table advanced a prede-
ings discussing the latest CAC research from around the termined distance before another tomographic image could
world. Many of the well-known cardiologists of the time be taken.
would attend these meetings only to dismiss such information “Slip-ring” technology introduced in 1989 gave way to
during the true scientific presentations. The “calcium club” the “uncoupling” of the data acquisition system [DAS] from
however was a popular forum to openly discuss research and the X-ray gantry allowing continuous scanning along with
advance ideas among like-minded colleagues. In fact, the orig- synchronized and graduated scanner table motion. Thus,
inal members of the calcium club founded the Society of contiguous X-ray tomograms could be acquired during grad-
Atherosclerosis Imaging [SAI] and later incorporated the uated increments of the imaging table covering more of the
Society of Cardiovascular Tomography [SCCT]. Some of the body in less time than had been previously possible with
“issues” between the calcium club and established conven- prior generations of CT scanners. The CT scanner, unteth-
tional cardiology, interested in obstructive disease and devel- ered, can produce a spiral or helical imaging path through the
opment of angioplasty and stents, somewhat congruous with body. Helical CT provides no obvious slice Z-positions, and
detecting nonclinical atherosclerosis by CAC, are featured in any arbitrarily selected slice plane is equivalent to any other.
the documentary “The Widowmaker” [34]. Thus, once helical data are acquired, slices may be recon-
Although there are likely thousands of publications from structed at any Z-position.
all parts of the world regarding cardiac CT using EBT, the The speed of imaging a body part was improved signifi-
“holy grail” was the potential to image the coronary arteries cantly by the introduction of slip-ring technology, but only
in vivo. EBT had superior temporal resolution and, for a long one slice could be imaged at a single time. Multislice imag-
time, superior spatial resolution over conventional CT.  Dr. ing, looking at multiple slices acquired at the same time,
Stephan Achenbach and colleagues in 1995 first published would afford a better temporal and spatial imaging of the
the use of EBT in quantitative imaging of the coronary arter- heart.
ies [35]. Following that publication, the labeling of cardiac In 1998 Elscint, an Israeli based operation, introduced the
CT as a “niche” changed dramatically. “dual-slice” CT scanning able to acquire two contiguous
Despite the potential advantages of EBT over, at the scans at one sweep of the X-ray source/detector pair – thus
time, conventional CT in terms of cardiac imaging, no more halving the time required for a chest or body scan. Elscint
than 110 Imatron scanners were manufactured/installed. was absorbed by Picker International eventually but then
Many are still installed and operational across the world Picker folded. Siemens and GE followed suit with their
using cannibalized parts from other scanners or via post- introductions – later four-slice CT and eight-slice CT. With
market jerry rigging. They were temperamental, difficult to the introduction of multislice CT scans came the implication
maintain, and very expensive. In 1999 EBT, a niche scanner that these would replace EBT in terms of cardiac imaging.
might cost $1.5 million, while conventional CT used for However, these “wannabees” simply could not duplicate the
thoracic and body imaging might cost <$0.7 million. The spatial and temporal resolution of EBT in terms of cardiac
final curtain was the incorporation of Imatron in 2001 by imaging.
General Electric. The swan song of EBT was the GE In 2002 I reviewed an article detailing the use of Siemens
“e-Speed,” a “stripped down” scanner with improved spa- 16-slice MDCT [multi-detector CT] and its application to
tial resolution to 1.5 mm versus 3 mm, but was not suffi- noninvasive coronary angiography [36]. The spatial resolu-
cient to compete in the world striving to be able to image tion of MDCT was superior to EBT, and with the use of beta-­
the coronary arteries, which are on the order of 3–4 mm or blockade to lower the heart rate, spectacular images as
less in diameter. The e-speed was abandoned by GE within compared with conventional angiography were finally avail-
a few years of its introduction. able. For the first time, I realized that multislice conventional
1  History of Cardiac CT: A Personal Story 9

CT could, indeed, duplicate and improve on the prior studies 7. Ritman EL, Kinsey JH, Robb RA, Gilbert BK, Harris LD, Wood
using EBT. We were ready to “take it up a notch.” EH.  Three-dimensional imaging of heart, lungs, and circulation.
Science. 1980;210(4467):273–80.
Quickly the numbers of simultaneous “slices” were 8. Robb RA, Sinak LJ, Hoffman EA, Kinsey JH, Harris LD, Ritman
increased to 32-, 42-, and finally to the current state-of-the-­ EL.  Dynamic volume imaging of moving organs. J Med Syst.
art 64-slice MDCT. The industry has not stopped at 64 slices, 1982;6(6):539–54. PubMed PMID 7183727.
moving to 256 and even 320 slices acquired at a single sweep 9. Iinuma TA, Tateno Y, Umegake Y, et al. Proposed system for ultra-­
fast computed tomography. J Comput Assist Tomogr. 1977;1:494–9.
of the scanner gantry. Faster rotational speeds, along with 10. Haimson J. X-ray source without moving parts for ultra-high speed
mathematical algorithms, have reduced spatial resolution to tomography. IEEE Trans Nucl Sci. 1979;26:2857–61.
0.4 mm or better as well as duplicating EBT temporal resolu- 11. Boyd DP, Gould RG, Quinn JR, et al. A proposed dynamic cardiac
tion to <100 msec. 3-D densitometer for early detection and evaluation of heart dis-
ease. IEEE Trans Nucl Sci. 1979;26:2724–7.
Scores of publications using 16+ MDCT have duplicated 12. Boyd DP.  Computerized transmission tomography of the heart

the original cardiac research initially done using EBT, and using scanning electron beams. In: Higgins CB, editor. CT of the
64+-slice MDCT is now used routinely in the outpatient and heart and the great vessels: experimental evaluation and clinical
inpatient facilities to define CAC and coronary artery athero- application. Mount Kisco: Futura Publishing Company; 1983.
13. Marcus ML, Doty DB, Hiratzka LF, Wright CB, Eastham

sclerotic plaque severity and type in the outpatient and inpa- CL. Decreased coronary reserve – a mechanism for angina pectoris
tient situations. in patients with aortic stenosis and normal coronary arteries. N Engl
J Med. 1982;307:1362–6.
14. Feiring AJ, Rumberger JA, Skorton DJ, Collins SM, Higgins CB,
Lipton MJ, Ell S, Marcus ML.  Determination of left ventricu-
Conclusions lar mass in the dog with rapid acquisition cardiac CT scanning.
Circulation. 1985;72:1355.
“Those that do not remember the past are bound to repeat it.” 15. Rumberger JA, Feiring AJ, Lipton MJ, Higgins CB, Ell S, Marcus
[George Santayana, 1863–1952  in George Santayana (1905) ML. Use of ultrafast CT to quantitate myocardial perfusion: a pre-
Reason in Common Sense, p.  284, volume 1 of The Life of liminary report. J Am Coll Cardiol. 1987;9:59–69.
Reason] 16. Reiter SJ, Rumberger JA, Stanford W, Marcus ML.  Quantitative
determination of aortic regurgitant volumes by cine computed
I have lived the history of cardiac CT and am still doing tomography. Circulation. 1987;76:728–35.
and teaching it  – the awe has not left me wondering. The 17.
Stanford W, Rooholamini M, Rumberger JA, Marcus
pages to follow in this book will define far better than I can ML. Evaluation of coronary bypass graft patency by ultrafast CT. J
Thorac Imaging. 1988;3:52–5.
the wonder of the physics, mathematics, and computer sci-
18. Feiring AJ, Rumberger JA, Reiter SJ, Collins SM, Skorton DJ, Rees
ence that has, along with the manufacturing of MDCT, come M, Marcus ML. Sectional and segmental variability of left ventricu-
together to establish cardiac CT as a mainstream and in many lar function: experimental and clinical studies using ultrafast com-
instances a primary or initial imaging tool for studying the puted tomography. J Am Coll Cardiol. 1988;12:415–25.
19. Rumberger JA, Weiss R, Feiring AJ, Stanford W, Hajduczok Z,
heart and coronary arteries, going full circle with the initial Rezai K, Marcus ML. Patterns of regional diastolic function in the
efforts to define cardiac size, function, and flow which begun normal human left ventricle: an ultrafast computed tomographic
now nearly 40 years ago. study. J Am Coll Cardiol. 1989;14:119–26.
20. Rumberger JA, Bell MR, Stanson AF, Sheedy PF, Behrenbeck
T.  Assessment of left ventricular diastolic function in man
using ultrafast computed tomography. Am J Cardiac Imaging.
References 1990;4:130–7.
21. Feiring AJ, Rumberger JA. Ultrafast computed tomography analy-
1. Budinger TF, Gullberg GT.  Three dimensional reconstruction sis of regional radius-to-wall thickness ratios in normal and volume-­
in nuclear medicine emission imaging. IEEE Trans Nucl Sci. overloaded human left ventricle. Circulation. 1992;85:1423–32.
1974;NS-12:2–21. 22. Vigneswaran WT, McDougall JC, Olson LJ, Breen JF, McGregor
2. Cormack A.  Representation of function by its line integrals with CGA, Rumberger JA. Right ventricular assessment in patients pre-
some radiological applications. Part I. J Appl Phys. 1964;35:2722. senting for lung transplantation. Transplantation. 1993;55:1051–5.
3. Ramachandran GN, Lakshminarayanan AV.  Three-dimensional 23. Rumberger JA, Behrenbeck T, Breen JR, Reed JE, Gersh BJ. Non-­
reconstruction from radiographs and electron micrographs: appli- parallel changes in global chamber volume and muscle mass during
cation of convolutions instead of Fourier transforms. Proc Natl the first year following transmural myocardial infarction in man. J
Acad Sci USA. 1971;68:2236–40. Am Coll Cardiol. 1993;21:673–82.
4. Hounsfield GN. Computerized transverse axial scanning (tomogra- 24. Hirose K, Shu NH, Reed JE, Rumberger JA. Right ventricular dila-
phy). 1. Description of system. Br J Radiol. 1973;46:1023–47. tation and remodeling the first year after an initial transmural wall
5. Brundage BH, Lipton MJ. The emergence of computed tomogra- myocardial infarction. Am J Cardiol. 1993;72:1126–30.
phy as a cardiovascular technique. Am Heart J. 1982;103:313–6. 25. Wolfkiel CJ, Ferguson JL, Chomka EV, Law WR, Labin IN,

6. Sinak LJ, Hoffman EA, Schwartz RS, Smith HG, Holmes DR, Tenzer ML, Booker M, Brundage BH.  Measurement of myocar-
Bove AA, Robb RA, Holmes DJ, Ritman EL. Three dimensional dial blood flow by ultrafast computed tomography. Circulation.
cardiac anatomy and function in heart disease in adults: initial 1987;76:1262–73.
results with the Dynamic Spatial Reconstructor. Mayo Clin Proc. 26. Bell RM, Lerman L, Rumberger JA. Validation of minimally inva-
1985;60:383–92. sive measurement of myocardial perfusion using electron beam
10 J. A. Rumberger

computed tomography and application in human volunteers. Heart. disease by ultrafast CT: a quantitative pathologic study. J Am Coll
1999;81:628–35. Cardiol. 1992;20:1118–26.
27. Tannenbaum SR, Kondos GT, Veselik KE, Prendergast MR,
32. Rumberger JA, Simons DB, Fitzpatrick LA, Sheedy PF, Schwartz
Brundage BH, Chomka EV. Detection of calcific deposits in cor- RS.  Coronary artery calcium areas by electron beam computed
onary arteries by ultrafast computed tomography and correlation tomography and coronary atherosclerotic plaque area: a histopatho-
with angiography. Am J Cardiol. 1989;63:870–2. logic correlative study. Circulation. 1995;92:2157–62.
28. Agatston AS, Janowitz WR, Hilder FJ, Zusmer NR, Viamonte M, 33. Sangiorgi G, Rumberger JA, Severson A, Edwards WD, Gregoire J,
Detrano R. Quantification of coronary artery calcium using ultra- Fitzpatrick LA, Schwartz RS. Arterial calcification and not lumen
fast computed tomography. J Am Coll Cardiol. 1990;15:827–32. stenosis is highly correlated with atherosclerotic plaque burden in
29. Stanford W, Rumberger JA. In: Stanford W, Rumberger JA, editors. humans: a histologic study of 723 coronary artery segments using
Ultrafast computed tomographic cardiac imaging: principles and non-decalcifying methodology. J Am Coll Cardiol. 1998;31:126–33.
practice. Mt Kisco, NY: Futura Publishing Company; 1992. 34. http://www.nytimes.com/2015/02/27/movies/review-the-widow-
30. Kaufmann RB, Sheedy PF 2nd, Maher JE, Bielak LF, Breen JF, maker-a-heart-care-documentary.html?_r=0.
Schwartz RS, Peyser PA.  Quantity of coronary artery calcium 35. Moshage W, Achenbach S, Seese B, Bachmann K, Kirchgeorg
detected by electron beam computed tomography in asymptomatic M. Coronary artery stenoses: three-dimensional imaging with elec-
subjects and angiographically studied patients. Mayo Clin Proc. trocardiographically triggered, contrast-agent enhanced, electron
1995;70:223–32. beam CT. Radiology. 1995;196:707–14.
31. Simons DB, Schwartz RS, Edwards WD, Sheedy PF, Breen JF, 36. Rumberger JA.  Non-invasive coronary angiography using CT  –
Rumberger JA. Non-invasive definition of anatomic coronary artery ready to kick it up another notch? Circulation. 2002;106:2036–8.
 Evolution of Radiation Dose
from Cardiac CT 2
Manoj Mannil and Hatem Alkadhi

Cardiac imaging with computed tomography (CT) remains Introduction of various CT protocol developments and
a challenging task requiring both a high spatial resolution software modifications such as tube current modulation,
for imaging of the relatively small coronary arteries and a automatic tube potential selection, electrocardiography
high temporal resolution for providing motion-free images (ECG) pulsing, and iterative reconstruction enabled the
of the heart. In parallel to technological advances, tireless decrease of ionizing radiation dose in one order of magni-
efforts to simultaneously reduce radiation dosages were tude over the last two decades [2] (Table 2.1).
implemented. As in any given body region, these measures
are based on the following two principles: restricting the
total number of cardiac CT examinations by close adher- ECG Gating
ence to guidelines and recommendations and reducing the
radiation of each individual CT study according to the “as Cardiac CT imaging must be performed with ECG gating, in
low as reasonably achievable” (ALARA) principle while order to freeze all images in defined time points of the car-
maintaining a diagnostic image quality [1]. This chapter diac cycle. At normal heart rates, least cardiac motion occurs
focuses on the latter. during diastole, when the ventricles are filling. Three

Table 2.1  Abbreviations and terms for cardiac CT imaging

Abbreviations
and terms Description
AEC Automatic exposure control mechanism for dose optimization in CT, which adapts tube current according to the part of the
body being scanned to obtain the desired image quality
CTDI Computed tomography dose index: Radiation dose of a single slice from the primary beam and scatter from surrounding
slices, in milligrays [mGy]
CTDIvol Computed tomography dose index volume: A measure describing the radiation dose factoring in pitch, which represents the
average radiation dose in three dimensions and is typically reported in milligrays [mGy]
DLP Dose length product reflects the total energy absorbed attributable to a complete CT scan acquisition and is determined by
multiplying the CTDIvol by the scan length [mGy*cm]
ECG pulsing Modulation of the tube current according to the electrocardiogram (ECG). Peak tube output during a selected pulsing
window and reduction of the tube output to approximately 25% during remaining parts of the cardiac cycle
IR Iterative reconstruction: Image reconstruction algorithm that localizes and selectively removes image noise. Repetitive
comparison of reconstructed images with a computed expected data set
Pitch Defined as table distance traveled in one 360° gantry rotation divided by beam collimation in single-slice CT or total
thickness of all simultaneously acquired slices in multislice CT
Reconstructed Width of a single slice in the reconstructed image [mm]. The greater the slice thickness, the more X-ray photons are present
slice thickness in the raw data for reconstructing the image, with an effect on image noise similar to that of increasing the tube current
TCM Tube current modulation automatically adapts the mAs delivered during the CT scan based on patient thickness determined
from the topogram

M. Mannil (*) · H. Alkadhi

Institute of Diagnostic and Interventional Radiology, University
Hospital Zurich, Zurich, Switzerland
e-mail: [email protected]

© Humana Press 2019 11

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_2
12 M. Mannil and H. Alkadhi

approaches for ECG gating are currently used: (1) retrospec- below. Radiation dose values of up to 21 mSv have been ini-
tive ECG gating with spiral data acquisition, (2) prospective tially reported for this technique with 64-slice CT.
ECG gating with a sequential (or “step-and-shoot”) data The use of ECG pulsing allows for modulation of the tube
acquisition, and (3) prospective ECG gating with spiral data current according to the ECG. The peak tube output is gener-
acquisition and high pitch. ated during a selected pulsing window, usually during diastole.
During the remaining parts of the cardiac cycle, the tube output
is reduced to approximately 25%. Most scanners allow to
Retrospective ECG Gating adjust the ECG-pulsing window width according to the heart
rate, in order to be as narrow as possible for maximal radiation
In retrospective ECG-gating techniques, partially ­overlapping exposure reduction while being as wide as reasonable to obtain
multi-detector CT projections are continuously acquired in diagnostic image quality at the same time [5]. Modern ECG-
the spiral mode, and the ECG signal is simultaneously pulsing algorithms are able to automatically detect arrhythmia
recorded. Software algorithms are then used to sort the data and switch off ECG pulsing during ectopic heartbeats.
from different phases of the cardiac cycle by progressively When combining retrospective ECG gating with opti-
shifting the temporal window of acquired helical projection mized tube voltage and tube current in combination with
data relative to the R wave (Fig. 2.1). Every position of the modern iterative reconstruction algorithms, radiation doses
heart is covered by a detector row at every point of the car- can be as low as 4–6 mSv per cardiac CT examination with
diac cycle. Therefore, the scanner table continuously moves this technique.
but advances not more than the total width of the active
detectors for each heartbeat [3]. For a gapless volume cover-
age of the heart in each cardiac phase, a low table feed (pitch  rospective ECG Gating
P
<1) is required, which has to be adapted to the heart rate of in the Sequential Mode
the patient: the higher the heart rate, the faster the table can
move. Using a 4-cm detector enables covering the entire Prospective ECG triggering uses the ECG signal to limit
heart volume in 3–4 heartbeats [4]. Due to the low pitch of scanning to time points defined prior to the data acquisition,
this imaging technique, the same anatomic area is repeatedly usually during diastole. Projection data are acquired during
exposed to ionizing radiation during consecutive rotations of more than half a gantry rotation. The total number of slices
the gantry, which results in a relatively higher radiation dose produced per heartbeat during this half rotation of the gantry
as compared to the other ECG-gating techniques described is proportional to the number of rows of active detectors. As

Fig. 2.1 Retrospective Continuous table movement

ECG-gated cardiac CT in the
spiral mode with continuous
table movement. The ECG of
Scan

Scan

Scan

the patient is schematically

shown at the bottom. Data can
be retrospectively
reconstructed at each after
data acquisition at each time
point of the cardiac cycle.
(Modified from Flohr et al.
[4], with permission)

Z
2  Evolution of Radiation Dose from Cardiac CT 13

Fig. 2.2  Principle of

prospective ECG-gated
sequential (or “step-and-­

Scan

Scan
shoot”) scanning. The ECG of Table movement
the patient is schematically
shown at the bottom. The blue
lines indicate the z-positions
of the individual detector
rows of a multirow detector
relative to the patient. Axial
(sequential) scan data are Table movement
acquired with a user-­
selectable temporal distance
to the preceding R wave.
After each axial scan, the
table moves to the next
z-position. Scan data are
typically acquired at every
second heartbeat. (Modified Z
from Flohr et al. [4], with
permission)

axial (rather than helical) scanning is used, the table is not this mode, data is acquired in a spiral fashion, while the table
moving during but only in-between data acquisition. The runs with a high pitch (e.g., 3.2 with third-generation dual-­
table has to move by the total collimation width after each source CT, equivalent to a table feed of 737 cm/s).
acquisition (sequential or “step-and-shoot” mode) (Fig. 2.2). Using this technique yields a “snapshot” scan of the entire
About 12 cm of scanning in the z-axis is required to cover heart within one cardiac cycle, usually during diastole. While
most adult hearts; the total number of heartbeats and number pitch in a single-source CT is limited to values of approxi-
of image stacks for a cardiac CT study depend on the width mately 1.5 to ensure gapless volume coverage along the
of the detector. However, as the scan is sequentially obtained, z-axis, a dual-source CT system contains a second X-ray
there is minimal flexibility in retrospectively choosing dif- tube at 90° offset, which acquires scan data at the same angu-
ferent phases of the cardiac cycle for image reconstruction. lar position at a quarter rotation later. This facilitates a maxi-
This minimal flexibility can be circumvented by using the mum pitch of 3.2–3.4, depending on the scanner generation,
technique of “padding.” Padding allows for acquisition of within a limited scan field of view (FOV) covered by both
images in additional cardiac phases by extending the manda- detectors (Fig. 2.3). A quarter rotation of data per measure-
tory minimum acquisition time. There is a trade-off between ment system is used for image reconstruction, and each indi-
acquisition flexibility and radiation dose. Extensive use of vidual axial image has a temporal resolution of a quarter of
padding is associated with a greater radiation dose. For the gantry rotation time. The ECG is used to prospectively
instance, a 100 millisecond increase in padding results in a trigger the start of table motion and the start of the high-pitch
45% increase in radiation dose [6]. A recent meta-analysis spiral. A requirement for usage of this technique is generally
showed that prospective ECG gating is associated with an a stable sinus rhythm with heart rates ≤60 bpm with second-­
average radiation dose of 2.7 mSv [7]. Currently, prospective generation dual-source CT and  ≤70  bpm with third-­
ECG gating is the most widely used technique for data generation dual-source CT (Fig. 2.4).
acquisition in cardiac CT [2, 8]. The effective radiation dose of a prospectively ECG-
gated high-pitch cardiac CT can be well below 1 mSv [9,
10]. In contrast to prospective ECG gating, where additional
 rospective ECG Gating with Spiral
P X-ray exposure along the entire width of the detector is gen-
Acquisition and High Pitch erated at every slice, the high-pitch mode generates super-
fluous X-rays exposing the entire width of the detector only
Since the commercial launch of dual-source CT technology at the beginning and at the end of the spiral path. This
in 2005, further reduction of ionizing radiation was achieved explains the further reduction of radiation exposure when
through use of prospective ECG gating with high pitch. In using this mode [1].
14 M. Mannil and H. Alkadhi

Fig. 2.3 ECG-gated
high-pitch mode. The ECG of
the patient is schematically
shown at the bottom. The

Scan
table is moved at very high
speed, sufficient to cover the
entire heart volume within a
single diastole. At each
z-position, only the minimum
amount of scan data necessary
for image reconstruction is
acquired. The temporal
resolution per image is Temporal resolution per
therefore close to a quarter of image
the gantry rotation time; and
there is a slight phase shift of
the reconstructed images
along the z-axis. (Modified Scan time
from Flohr et al. [4], with Fast continuous table
permission) Z
movement

rent to account for differences in body shape and attenuation

depending on the body region scanned. Lowering and optimiz-
ing the tube voltage represent another important approach for
dose reduction, as radiation dose varies with the square of the
tube voltage. Initial investigations with 16-/64-­slice CT sys-
tems comparing 120- and 100-kV tube voltage CT protocols
have shown that decreased tube voltages reduce overall radia-
tion dose at the expense of increased image noise. The signal-
to-noise (SNR) and contrast-to-noise ratios (CNR), however,
remain unaffected, as decreased tube voltages approach the
k-edge of iodine (33.2 keV), which results in higher attenuation
of iodinated contrast. With an increase in both parameters, sig-
nal/contrast and noise, the overall ratios for SNR and CNR
remain unaffected [11–13]. Recently, automatic tube potential
selection ­algorithms became available which automatically
lower the tube potential according to patient size, diagnostic
task, and scanner tube current limits. This measurement is
reported to yield a dose reduction between 20% and 50% in
nonobese patient populations. For larger patients, however,
image quality may deteriorate beyond diagnostic quality [14].

Fig. 2.4  3D image of the coronary artery tree with volume rendering of
image data acquired with the high-pitch mode on a third-generation Iterative Reconstruction (IR)
dual-source CT scanner. Estimated effective radiation dose was 0.4 mSv
Iterative reconstruction (IR) is a technique that incorporates
statistical modeling in image reconstruction with the main
 ube Current Modulation and Tube
T aim to reduce image noise. This technique can be also used
Voltage Adaptation for reducing the radiation dose (usually through reduced
tube current) while maintaining the image quality of the
Both tube current modulation and tube voltage adaptation can examination [1, 10]. Similar to the options tube current mod-
be used for all scan protocols described above. Tube current ulation and tube voltage adaptation, IR can be applied for all
modulation is defined as automatic adjustment of the tube cur- scan protocols described above.
2  Evolution of Radiation Dose from Cardiac CT 15

Filtered Back Projection (FBP) photons in the transmission between the X-ray tube, iso-
center, and detector (physical modeling) have been devel-
Until recently, CT images were almost exclusively recon- oped and are collectively referred to as model-based IR
structed with filtered back projection (FBP), largely due to [15]. Current IR algorithms typically combine statistical
the fact that FBP generates diagnostic images at a low level of modeling with model-based IR.  In hybrid IR algorithms,
computational complexity. In FBP the X-ray beam assumes a the imaging data can be blended with traditional FBP at
pencil shape, and the X-ray source is aligned in a parallel certain thresholds. The iteratively reconstructed CT images
fashion to a linear X-ray detector array. For image generation, are generally less noisy and possess an improved image
the X-ray source is rotated over a certain angle, allowing for resolution and edge detection compared to images acquired
intensities to be measured at the detector. These intensities with FBP (Fig. 2.5).
are described as an integral function for a specific angle and Due to this circumstance, diagnostic images can be
the shift in the position of the X-ray tube. After this, the achieved even at lower tube currents when using IR instead
reconstruction process involves solving an integral equation of FBP, and radiation exposure can be significantly
by inversion or so-called back projection. Despite its ability decreased. Many studies have evaluated noise and dose
to rapidly reconstruct images, FBP has various limitations, in reduction with IR algorithms across multiple vendors
particular increased image noise, which is most prominent in (Table  2.2). In summary they demonstrate either (1)
low tube current imaging, poor contrast resolution, and streak improved noise properties with IR compared to FBP when
artifacts. This is primarily due to the inherent failure of the images from the same patients are reconstructed using
FBP algorithm to account for image noise that results from both algorithms (intraindividual analysis) or (2) preserved
Poisson statistical variations in the number of photons across subjective image quality and quantitative noise properties
the imaging plane, leading to an inverse relationship between among patients undergoing IR image reconstruction with
radiation dose and image noise. Until the recent introduction low-radiation dose techniques (reduced tube current and/
of IR, lowering image noise could only be achieved at the or potential) compared to FBP algorithms with standard
expense of increased radiation dose. Attempts at so-called dose techniques (interindividual analysis). For instance,
image-based denoising through smoothing algorithms and fil- performing prospectively ECG-gated high-pitch coronary
ters (convolution kernels) allow for noise reduction but result angiography with third-generation dual-­ source CT and
in compromised spatial resolution and image fidelity [15]. using advanced modeled IT at a 70 kVp tube voltage result
in a reduced image noise and improved overall image qual-
ity compared to standard FBP reconstructions at an aver-
Iterative Reconstruction (IR) age effective dose of only 0.3 mSv [15, 16].

In principle, IR techniques attempt to localize and selectively

remove image noise by frequent comparison with a canvas. Limiting the Scan Length
This is achieved through a process of modeling the imaging
acquisition process, including fluctuations in photon statis- Limiting the scan length, or z-axis coverage, to exactly
tics, the optics system, and various aspects of X-ray interac- encompass the heart and coronary arteries represents
tions to generate an expected data set, which is then compared another important factor for reducing the radiation dose car-
to the actually acquired data set. The differences between diac CT. Traditionally, each cardiac CT examination starts
these two are used to identify and remove noise, and the pro- with acquisition of a scout view for planning the scan length
cess is repeated multiple times until the updated data con- of the examination. It is a common practice to obtain an
verges with or approximates the expected data to maximize unenhanced low-dose calcium scoring scan, whose scan
image optimization. length is usually planned on the scout view. Then,
Initially, statistical IR was introduced in CT imaging. It ­contrast-enhanced CT coronary angiography is performed
deals with noise due to fluctuations in photon statistics by for ­diagnosing or excluding coronary artery stenoses. Due
assigning a relatively higher weighting to data with low to the limited anatomic information provided by the scout
statistical uncertainty (low noise) and lesser weighting to view, the exact cranio-­caudal extent of the coronary arteries
data with high statistical uncertainty (high noise). It oper- cannot be discerned on a scout view. Therefore, the region
ates in either the “raw data domain” with subsequent of the carina is frequently used as the upper border and the
reconstruction using the “noise-reduced” data or in the cardiac apex as the lower border of the scan length of CT
“image domain” after IR.  More recently, IR algorithms coronary angiography.
that model for the physical three-dimensional nature of the Alternatively, scan length can be determined by using
system optics (geometric modeling of the focal point of the axial images from calcium scoring for identifying the
the X-ray tube and pixel detector) and the interaction of table position at the origin of the left main artery and that
16 M. Mannil and H. Alkadhi

a b c

d e f

Fig. 2.5  Multiplanar reformations of the right coronary artery using 2, (d) ADMIRE strength level 3, (e) ADMIRE strength level 4, (f)
(a) filtered back projection (FBP), (b) advanced modeled iterative ADMIRE strength level 5. Note the continuous decrease of image noise
reconstruction (ADMIRE) strength level 1, (c) ADMIRE strength level and the improved vessel sharpness from FBP to ADMIRE images
2  Evolution of Radiation Dose from Cardiac CT 17

Table 2.2  Iterative reconstruction algorithms by vendor radiation doses were formulated, including a linear no-­
Iterative reconstruction threshold model which indicates that each level of radiation
technique Vendor Type is associated with a risk to the patient and that there is no
Adaptive statistical iterative General Electric (GE) Hybrid lower threshold for this risk. Still, the American Association
reconstruction (ASIR) Healthcare
of Physicists in Medicine considers risks of medical imaging
Model-based iterative General Electric (GE) Model-­based
reconstruction (MBIR-Veo) Healthcare
at effective doses <50 mSv for single procedures or <100 mSv
Adaptive statistical iterative General Electric (GE) Hybrid for multiple procedures over short-time periods as too low to
reconstruction (ASIR-V) Healthcare be epidemiologically detectable. While there is no evidence
iDose Philips Healthcare Hybrid between radiation from medical imaging and cancer induc-
Iterative model Philips Healthcare Model-­based tion in a general adult population, it is in the best interests of
reconstruction (IMR) the patients to keep the ionizing radiation exposure ALARA,
Iterative reconstruction in Siemens Healthcare Hybrid
image space (IRIS)
without loss of diagnostic information.
Sinogram-affirmed iterative Siemens Healthcare Hybrid
reconstruction (SAFIRE)
Advanced modeled Siemens Healthcare Hybrid Summary
iterative reconstruction
(ADMIRE)
Adaptive iterative dose Toshiba Hybrid
Recent technological advances in cardiac CT imaging, rang-
reduction (AIDR) ing from automated exposure control, tube voltage optimiza-
Adaptive iterative dose Toshiba Hybrid tion, scan length adjustments, evolution of ECG-gating
reduction 3D (AIDR 3D) techniques, and iterative reconstruction post-processing,
have reduced radiation doses in one order of magnitude over
the last decade. Continuous efforts in the optimal selection of
of the cardiac apex. The scan length for CT coronary angi- each of the techniques help in keeping the radiation exposure
ography can be then planned by adding around 1 cm to the to each individual adult patient so low that the assumed theo-
cranial and around 1 cm to the lower border to account for retical radiation risk from cardiac CT can be neglected.
possible changes in inspiration depth between calcium
scoring and CT coronary angiography. Using this approach
might lead to another reduction of radiation exposure of
References
around 16% [17].
1. Alkadhi H, Leschka S. Radiation dose of cardiac computed tomog-
raphy  - what has been achieved and what needs to be done. Eur
Radiol. 2011;21(3):505–9.
2. den Harder AM, Willemink MJ, de Jong PA, Schilham AM,
 linical Relevance of Dose Reduction
C Rajiah P, Takx RA, et al. New horizons in cardiac CT. Clin Radiol.
and Cancer Risk 2016;71(8):758–67.
3. Desjardins B, Kazerooni EA.  ECG-gated cardiac CT.  AJR Am J
Roentgenol. 2004;182(4):993–1010.
Raised awareness of medical practitioners and stringent
4. Flohr TG, De Cecco CN, Schmidt B, Wang R, Schoepf UJ, Meinel
application of the aforementioned dose reduction techniques FG.  Computed tomographic assessment of coronary artery dis-
have been proven to be effective. Cardiac CT is well estab- ease: state-of-the-art imaging techniques. Radiol Clin North Am.
lished for certain indications, e.g., for ruling out obstructive 2015;53(2):271–85.
5. Leschka S, Scheffel H, Desbiolles L, Plass A, Gaemperli O, Valenta
coronary artery disease at a low-to-intermediate pretest prob-
I, et al. Image quality and reconstruction intervals of dual-source
ability [18]. For determining whether all efforts in minimiz- CT coronary angiography: recommendations for ECG-pulsing win-
ing ionizing radiation are relevant requires comparison of the dowing. Investig Radiol. 2007;42(8):543–9.
clinical benefit of the CT examination with the potential and 6. Labounty TM, Leipsic J, Min JK, Heilbron B, Mancini GB, Lin
FY, et al. Effect of padding duration on radiation dose and image
theoretical cancerogenic risk of the procedure.
interpretation in prospectively ECG-triggered coronary CT angiog-
Reports on cancer risk in medical imaging are mainly raphy. AJR Am J Roentgenol. 2010;194(4):933–7.
derived from extrapolations of risk estimates to low levels of 7. von Ballmoos MW, Haring B, Juillerat P, Alkadhi H.  Meta-
radiation and multiplications with a large number of exposed analysis: diagnostic performance of low-radiation-dose coro-
nary computed tomography angiography. Ann Intern Med.
individuals. However, the data extrapolation is not reliable
2011;154(6):413–20.
below an exposure of 100 mSv. Therefore, various different 8. Menke J, Unterberg-Buchwald C, Staab W, Sohns JM, Seif Amir
hypotheses about extrapolation of radiation risk to lower Hosseini A, Schwarz A. Head-to-head comparison of prospectively
18 M. Mannil and H. Alkadhi

triggered vs retrospectively gated coronary computed tomography tomography in daily practice: impact of different scanning protocols
angiography: meta-analysis of diagnostic accuracy, image quality, on effective dose estimates. Circulation. 2006;113(10):1305–10.
and radiation dose. Am Heart J. 2013;165(2):154–63. e3. 14. Park YJ, Kim YJ, Lee JW, Kim HY, Hong YJ, Lee HJ, et  al.
9. Meyer M, Haubenreisser H, Schoepf UJ, Vliegenthart R, Leidecker Automatic Tube Potential Selection with Tube Current Modulation
C, Allmendinger T, et al. Closing in on the K edge: coronary CT (APSCM) in coronary CT angiography: comparison of image
angiography at 100, 80, and 70 kV-initial comparison of a sec- ­quality and radiation dose with conventional body mass index-
ond- versus a third-generation dual-source CT system. Radiology. based protocol. J Cardiovasc Comput Tomogr. 2012;6(3):184–90.
2014;273(2):373–82. 15. Naoum C, Blanke P, Leipsic J. Iterative reconstruction in cardiac
10. Gordic S, Desbiolles L, Sedlmair M, Manka R, Plass A, Schmidt CT. J Cardiovasc Comput Tomogr. 2015;9(4):255–63.
B, et  al. Optimizing radiation dose by using advanced modelled 16. Hell MM, Bittner D, Schuhbaeck A, Muschiol G, Brand M, Lell
iterative reconstruction in high-pitch coronary CT angiography. Eur M, et al. Prospectively ECG-triggered high-pitch coronary angiog-
Radiol. 2016;26(2):459–68. raphy with third-generation dual-source CT at 70 kVp tube volt-
11. Leschka S, Stolzmann P, Schmid FT, Scheffel H, Stinn B, Marincek age: feasibility, image quality, radiation dose, and effect of iterative
B, et al. Low kilovoltage cardiac dual-source CT: attenuation, noise, reconstruction. J Cardiovasc Comput Tomogr. 2014;8(6):418–25.
and radiation dose. Eur Radiol. 2008;18(9):1809–17. 17. Leschka S, Kim CH, Baumueller S, Stolzmann P, Scheffel H,
12. Hausleiter J, Martinoff S, Hadamitzky M, Martuscelli E, Pschierer Marincek B, et al. Scan length adjustment of CT coronary angiog-
I, Feuchtner GM, et  al. Image quality and radiation exposure raphy using the calcium scoring scan: effect on radiation dose. AJR
with a low tube voltage protocol for coronary CT angiography Am J Roentgenol. 2010;194(3):W272–7.
results of the PROTECTION II Trial. JACC Cardiovasc Imaging. 18. Hausleiter J, Meyer T, Hermann F, Hadamitzky M, Krebs M,

2010;3(11):1113–23. Gerber TC, et al. Estimated radiation dose associated with cardiac
13. Hausleiter J, Meyer T, Hadamitzky M, Huber E, Zankl M, Martinoff CT angiography. JAMA. 2009;301(5):500–7.
S, et al. Radiation dose estimates from cardiac multislice computed
 The Long March into Clinical Practice:
Cardiac CT and Its Competitors 3
Seth Uretsky, Alan Rozanski, and Daniel Berman

In the early twentieth century, control of infectious diseases, e­ lectrocardiographic response were identified as predictors
increasing longevity, and poor lifestyles resulted in chronic of adverse outcomes, including the magnitude, time to onset,
diseases becoming the major cause of disability and death [1]. and postexercise duration of exercise ST segment depres-
By the 1920s, coronary artery disease (CAD) became the sion. In addition, patients’ overall exercise capacity, as
leading cause of death in the United States. As treatments assessed by exercise duration during Bruce protocol exercise
emerged to treat CAD, it became important to develop means ECG, was found to be a potent predictor of patient risk [5].
for identifying those patients at risk for developing the dis- In the mid-1970s, the field of cardiac imaging emerged
ease and its complications. Initial diagnostic work in this due to the introduction of radionuclide imaging techniques.
regard was initiated even prior to the advent of exercise elec- Since then, the field of noninvasive cardiac imaging has
trocardiography. In the 1920s, Master developed his famous grown markedly. Today, many different forms of noninvasive
“two-step” stress test [2], which allowed physicians to assess tests can be used for the work-up and management of patients
patients’ functional capacity in a semiquantitative manner in clinical practice, as listed in Table 3.1. While these tests
and to provoke angina symptoms through physical activity. In vary markedly in their capabilities, they remain focused on
the 1950s, treadmill testing was implemented, and in 1963, three principal needs in cardiology: (1) to screen for latent
Robert Bruce published his protocol for performing graded CAD in asymptomatic subjects, (2) to establish the presence
multistage treadmill exercise with electrocardiographic mon- or absence of CAD in patients presenting with chest pain
itoring [3], a protocol that is still used today. This develop- symptoms (diagnostic testing), and/or (3) to evaluate
ment of the Bruce protocol treadmill exercise was timely patients’ clinical risk for developing future cardiac events
since the 1960s saw the introduction of cardiac catheteriza- (prognostic testing). Herein, we review the development of
tion and coronary bypass surgery. With this dramatic new these imaging modalities and the future competition that will
treatment option, a need emerged to accurately identify those shape how these modalities will be used in clinical practice.
patients who were at risk for cardiac events and thus potential
beneficiaries of coronary artery bypass surgery.
Epidemiological study in the 1960s and 1970s refined the The Advent of Nuclear Cardiology
use of exercise electrocardiography for diagnostic and
prognostic testing [4]. Various aspects of the exercise
­ The first cardiac imaging approach to be applied for diagnos-
tic and prognostic testing was exercise-rest myocardial per-
fusion imaging (MPI). This new imaging technology became
S. Uretsky (*)
Department of Cardiovascular Medicine, Gagnon Cardiovascular possible due to the application of gamma-emitting monova-
Institute, Morristown Medical Center, Morristown, NJ, USA lent cations which could be actively extracted by the myocar-
e-mail: [email protected] dium, such as potassium-43 [6], and its analogues,
A. Rozanski rubidium-81 [7] and thallium-201 [8]. The initial uptake of
Department of Cardiovascular Medicine, Gagnon Cardiovascular these analogues was found to correlate with the distribution
Institute, Morristown Medical Center, Morristown, NJ, USA
of myocardial blood flow over a wide range of flow rates,
The Division of Cardiology, Mount Sinai St. Luke’s Hospital, making these agents well suited for cardiac imaging at stress
Mount Sinai Heart, New York, NY, USA
as well as during rest imaging. Redistribution thallium-201
D. Berman imaging became a useful method for assessing myocardial
Department of Imaging and Medicine and the Burns Allen
viability. Exercise MPI, initially employed using a planar
Research Institute, Cedars-Sinai Medical Center,
Los Angeles, CA, USA technique, was soon shown to have both higher sensitivity

© Humana Press 2019 19

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_3
20 S. Uretsky et al.

Table 3.1  Noninvasive techniques for the work-up of patients of perfusion defects. The quantitation became a standard part
Exercise electrocardiography of the imaging methods.
Stress-rest myocardial SPECT perfusion imaging As investigators and clinicians became seasoned in the
Stress-rest position emission tomography application of nuclear cardiology, they established many
Exercise radionuclide ventriculography
principles regarding the correct validation of testing. Three
Stress echocardiography
Stress magnetic resonance imaging (MRI) principles that are particularly applicable to the tests which
Coronary artery calcium scanning have since followed are reviewed below.
Coronary computed tomography angiography
Carotid ultrasound
Femoral ultrasound
 haracterizing the Full Extent of Test
C
Abnormality
and specificity than its then only competitor for the detection
of CAD – exercise ECG. In the early days of nuclear cardiol- Investigators in nuclear cardiology sought from the onset to
ogy imaging, another radionuclide technique, exercise radio- evaluate all potential information that could be derived from
nuclide ventriculography (RNV), performed by using either the scintigraphic images. Early work established that the
the first pass or multiple-gated equilibrium technique, was extent of perfusion defects (i.e., the area of ischemic myocar-
also found to have higher sensitivity and specificity than dium) and the severity of perfusion defects (ranging from
exercise ECG. Then, after the advent of single-photon emis- very mild abnormality to the complete absence of uptake)
sion computed tomography (SPECT) MPI in the early 1980s, provided complementary and independent information. For
exercise SPECT-MPI became favored over both exercise pla- example, in a follow-up of 1689 diagnostic patients who
nar MPI and exercise RNV, due to this technique’s enhanced underwent exercise thallium-201 MPI, Ladenheim et al. [12]
ability to size the magnitude of stress-induced myocardial found that both the number and severity of perfusion defects
ischemia. Alongside these developments was the creation of within six myocardial regions were related to adverse car-
semiquantitative and quantitative scoring systems to deter- diac outcomes; when both extensive and severe ischemia
mine the severity of perfusion defects on SPECT-MPI as were present, adverse events increased in exponential fash-
well as software that allowed for easy processing, display, ion (Fig. 3.1). Other examples of ischemic parameters that
and analysis of the images. A universal scoring system could be used to characterize the presence of either extensive
helped decrease inter- and intra-observer variability and or severe ischemia for thallium-201 MPI, sestamibi MPI, or
effective communication of results. stress echo are shown in Table 3.2. For example, the identifi-
Other important developments in the 1980s included the cation of transient ischemic dilation was identified as a vari-
introduction of PET imaging for assessing myocardial perfu- able that predicted both extensive and severe CAD [13], with
sion and myocardial viability, the initiation of pharmaco- extensive ischemia needed to induce left ventricular dilation
logic testing with dipyridamole for those patients who could postexercise and severe ischemia needed to maintain dilation
not exercise, and the development of Tc-99m sestamibi and during the postexercise phase.
tetrofosmin, which considerably improved image quality due To convey the information regarding ischemic extent and
to the shorter half-life of Tc-99m versus thallium-201 which severity to physicians, investigators developed routine semi-
allowed for higher injected doses which led to improved quantitative analysis of MPI studies based on the visual scor-
image quality. Subsequently, the development of ECG-gated ing of the degree of radiotracer reduction in each segment of
SPECT imaging in the mid-1990s introduced the assessment the myocardium. This led to the further development of
of left ventricular ejection fraction and regional wall motion summed segmental scores at stress, rest, and their difference
analysis, which could be assessed in myocardial segments [14]. A later development was the conversion of these scores
containing myocardial perfusion defects [9]. into a percent of the myocardium involved for ischemia and
A further key component of the growth of nuclear cardiol- total defect severity [15], first based on a 0–4 scoring of 20
ogy studies was the inclusion of scientists and clinicians in myocardial segments and later, of 17 segments.
the field, with the scientists helping emphasize the impor-
tance of standardization of acquisition and processing meth-
ods and providing development of objective quantitative Combining Clinical and Physiological Data
analysis methods which provided a “second expert opinion”
in interpretation. These methods were first developed for pla- The combined assessment of clinical and imaging data can
nar imaging methods [10] and then subsequently were also improve risk assessment. This point was also well illus-
applied in SPECT-MPI [11]. For the latter, the advent of the trated by the abovementioned study by Ladenheim et al. (see
polar map, with comparison of a patient’s studies to normal Fig. 3.1) [12]. For any given level of ischemia, the cardiac
limits, allowed easy appreciation of the presence and l­ ocation event rate was increased approximately threefold among
3  The Long March into Clinical Practice: Cardiac CT and Its Competitors 21

Prognosis as a Function of
Defect Extent and Severity

Maximal Exercise Sub-Maximal Exercise

% Event % Event
50 50

0
0 3
3 2
Severity 2 Severity
1 1
0 1 2 3 4 5 6 0 1 2 3 4 5 6

Extent Extent

Fig. 3.1  Frequency of cardiac events on the orthogonal z axis accord- maximal predicted heart rate (left graph) and those achieving <85% of
ing to the extent of myocardial ischemia, ranging from 0 to 6 myocar- maximal predicted heart rate (right graph). The extent and severity of
dial zones (x axis); the severity of defects, ranging from none (score of ischemia were both independently correlated to cardiac event rates.
0) to severe (score of 3) on the y axis; and the patients are divided (From Ladenheim et al. [12], with permission)
according to those who showed both for patients achieving >85% of

Table 3.2  Extent and severity of ischemia parameters whether nuclear stress testing added to risk prediction
Study Ischemic when all of the more readily available clinical information
Extent Severity and results of exercise electrocardiography were first fac-
Number of perfusion defects ++++ tored into their risk analysis. Analyses of multiple patient
Severity of perfusion defects ++++ populations consistently demonstrated that radionuclide
Lung thallium-20 uptake +++ + imaging provided substantial additional risk prediction
Delayed thallium redistribution ++++ [19, 20]. Since then, the demonstration of incremental
Stunned myocardium on gated SPECT ++++
information is on the basis by which any new imaging
Transient ischemic dilation of the LV ++++ ++++
modality is evaluated.
Number of stress-induced WMA ++++
Severity of WMA ++++
Finally, a key component of the growth of nuclear cardiol-
Decrease in LVEF +++ ++ ogy was the formation in 1993 of the American Society of
Nuclear Cardiology (ASNC). The organization, formed pre-
those patients who could not exercise to at least 85% of their dominantly by cardiologists, became an advocate for the
maximal predicted heart rate on treadmill exercise. Today, method and laid the ground work for cardiologists employ-
many patients are studied by pharmacologic stress due to an ing the novel methods as part of their own clinical armamen-
inability to exercise. The mere inability to perform treadmill tarium—as occurred with echocardiography. The ASNC had
exercise also increases mortality risk by at least twofold a profound effect on the success of the field.
compared to patients who can exercise (Fig. 3.2) [16]. The Summarizing key components of the success of nuclear
further consideration of CAD risk factors does not generally cardiology might be the following: (1) recognition of the
influence short-term risk assessment, but it plays an impor- need for comprehensive databases containing clinical ECG,
tant role in complementing the results of stress testing for and comprehensive nuclear variables regarding the extent
predicting long-term risk (Fig. 3.3) [17, 18]. and severity of ischemia and scarring and follow-up for car-
diac events, resulting in a large body of published evidence
regarding diagnosis and prognosis and the incremental infor-
The Assessment of Incremental Information mation in large patient populations could be reported; (2)
recognition of the need for standardization of methodology
The early investigators in nuclear cardiology recognized for acquisition and processing and interpretation, along with
that it was insufficient to merely demonstrate that radionu- automated quantitative analytic tools; (3) the development of
clide imaging had superior test sensitivity and specificity vasodilator stress so that the method could be utilized in
compared to exercise ECG.  Rather, they also addressed patients who could not exercise; (4) prominent role played
22 S. Uretsky et al.

Matched Patients Age £ 55 Matched Patients Age 56-65

1.00 Ex

1.00
Ex
0.50 0.60 0.70 0.80 0.90

0.50 0.60 0.70 0.80 0.90

FREE OF ACM

FREE OF ACM
(p<0.0001) Adeno (p<0.0001)
Adeno

2 4 6 8 10 2 4 6 8 10
YEARS OF FOLLOW-UP YEARS OF FOLLOW-UP

Exclusion criteria:
Matched Patients Age > 65

1.00
• Dyspnea
• Peripheral vascular disease

0.50 0.60 0.70 0.80 0.90

FREE OF ACM
• Atrial fibrillation Ex
• LV enlargement
• LVH
• Diabetes (p<0.0001)
• Elevated BMI

Adeno

2 4 6 8 10
YEARS OF FOLLOW-UP

Fig. 3.2  Comparison of cumulative survival among exercise and adenosine patients with normal SPECT-MPI studies who were risk factor-­
matched using propensity analysis. The patients were subdivided by age. (From Rozanski et al. [16], with permission)

Fig. 3.3  Age- and gender-­

adjusted hazard ratios for
all-cause mortality according
to a number of CAD risk 7.29
factors among patients with
normal SPECT-MPI study, (All p’s ≤0.001)
divided according to exercise 8.00
duration. (From Rozanski
et al. [17], with permission) 7.00

6.00
3.78
HAZARD RATIO

5.00
2.53 3.64
4.00 EXERCISE
DURATION
3.00 2.23
1.70 2.57 < 6 min
2.00
1.59 6-8.9
1.00 1.00
> 9 min
0.00
Risk 0 RF’s 1 RF 2-3 RF’s
factors
3  The Long March into Clinical Practice: Cardiac CT and Its Competitors 23

by clinical cardiologists who control the patient referrals; combined with the functional and viability images, gives the
and (5) the development of a professional society that could clinician the ability to assess wall motion, myocardial viabil-
be an advocate for all aspects of the field. ity/scar, and myocardial ischemia in one test.
Despite the broad capabilities of cardiac MRI and its
introduction only a decade after nuclear cardiology methods,
The Advent of Stress Echocardiography it remains used in only a small number of centers in the
United States. In stress testing as an example, while it is esti-
Resting echocardiography has long been a mainstay for the mated that at most a few hundred thousand cardiac MRI pro-
evaluation of cardiac structure and function, valvular heart cedures are performed in the United States per year, the
disease, congenital heart disease, cardiac masses, and pericar- number of stress echocardiography procedures is an order of
dial disease. In the 1980s, postexercise stress echocardiogra- magnitude greater than this and of nuclear cardiology proce-
phy was introduced as a new method for evaluating patients, dures approximately two orders of magnitude greater than
followed later by the application of echocardiography during this. Understanding this lack of growth of cardiac MRI may
dobutamine infusion. Because stress echocardiography provide insight into the development of cardiac CT. MRI uti-
lagged approximately 5–10 years behind the introduction of lizes expensive equipment, costing approximately ten times
nuclear cardiology, the application of this technology was that of a nuclear cardiology scanner and much more than that
able to rely heavily on the principles established by nuclear compared to echocardiography equipment. Given this, it is
cardiology for its own validation. For example, just like with not suitable for office practice. The method has required a
the extent and severity perfusion defects, the extent and sever- great deal of expertise from technologists, to a greater extent
ity of stress-induced echocardiographic wall motion abnor- than with nuclear or echocardiographic methods. The inter-
malities proved to be potent predictors of patient risk [21]. pretation of cardiac MRI is more complex than that of
Generally, stress echocardiography has become employed for nuclear cardiology or echocardiography studies, requiring
similar clinical indications and in the same general popula- greater training of the physician interpreters. Standardized,
tions as stress radionuclide imaging. Among differences validated automated quantitative analytic software has not
between the methods are the commercial availability of vali- become available. While echocardiography and nuclear car-
dated quantitative analytic methods for nuclear cardiology diology became part of the standard curriculum of cardiol-
methods, perceived less dependence on the operator exper- ogy fellows over the last two decades, this training in cardiac
tise, and ease with which the method can be applied to patients MRI was far more limited, being available in a limited num-
who cannot exercise using vasodilator stress. ber of programs. Finally, while echocardiography and
nuclear methods were almost always contained within cardi-
ology divisions, in only a small number of centers in the
 he Advent of Magnetic Resonance
T United States is cardiac MRI within cardiology. Strong col-
Imaging (MRI) laboration between cardiology and radiology in cardiac MRI
has not been common.
MRI applications for cardiovascular disease were first devel-
oped in the 1980s and began to be applied clinically in the
1990s. MRI has a number of clinical utilities. First, MRI has  he Advent of Coronary Artery
T
the ability to differentiate different types of soft tissue from Calcium Scanning
each other and assess and quantify myocardial fibrosis/scar
and edema. Second, because of its ability to image the entire The advent of coronary artery calcium (CAC) scanning with
chest in any imaging plane, MRI has become a gold standard electron beam CT in the later part of the 1980s and 1990s
for assessing left and right ventricular structure and function. ushered in the beginning of a new era in the use of cardiac
Third, MRI can be used for assessment of cardiovascular imaging in asymptomatic patients, which is beginning to
masses, valvular heart disease, and cardiac masses. Fourth, profoundly affect the future course of noninvasive cardiac
the development of myocardial delayed enhancement tech- testing. The test is easily performed, not requiring a great
niques allowed MRI to become a mainstay in assessing myo- deal of technical expertise. An objective, quantitative ana-
cardial viability and the assessment of cardiomyopathies. lytic method has been available since soon after its introduc-
In addition, MRI had also been used as an alternative tion, providing what became described as a single
method to assess patients during pharmacologic stress. Stress number—the Agatston score—which has become the stan-
MRI techniques can be applied in either of two ways: (1) use dard approach since its introduction for expressing the results
of perfusion techniques such as those developed in nuclear of the scan and improved our ability to screen for CAD since
cardiology and (2) use of wall motion techniques as devel- the presence of CAC serves as a highly sensitive and specific
oped for stress echocardiography. Either stress technique, marker for the presence of underlying atherosclerosis.
24 S. Uretsky et al.

Fig. 3.4 Risk-adjusted
cumulative survival among 1.00 0 (n=11,044)
asymptomatic patients
1-10 (n=3,567)
undergoing CAC scanning.
For each increment in the 11-100 (n=5,032)
CAC score, there was a
101-299 (n=2,616)
significant increase in 0.95
all-cause mortality. (From 300-399 (n=561)
Budoff et al. [23], with 400-699 (n=955)

Cumulative Survival
permission)
700-999 (n=514)
0.90
n=25,263
asymptomatic ≥1,000 (n=964)
individuals

0.85

0.80

0.0 2.0 4.0 6.0 8.0 10.0 12.0

Time to Follow-up (Years)

Importantly, patho-anatomic correlations in the 1990s by stratifies risk among all FRS subgroups including those with
Rumberger et  al. established a proportional relationship low and high FRS scores (Fig. 3.5) [27].
between the CAC score and the overall burden of coronary In addition, the CAC score has been shown to provide
atherosclerosis [22]. Consistent study has demonstrated that superior risk stratification compared to other methods used to
cardiac risk increases in direct proportion to increasing screen for CAD. An important analysis in this regard was pro-
degrees of CAC abnormality [23–25]. For example, in a vided by Yeboah et  al. within the MESA population, who
long-term follow-up of 25,263 asymptomatic subjects who assessed 1330 subjects with an intermediate FRS by 6 sepa-
underwent CAC scanning, Budoff et  al. demonstrated an rate measurements: the ankle-brachial index, brachial flow-­
increase in risk in association with each increment of CAC mediated dilation, carotid intima-media thickness,
score increase (Fig. 3.4) [23]. Even a very low CAC score, in consideration of one’s family history of premature CVD,
the range of 1–9, establishes the presence of atherosclerosis. high-sensitivity C-reactive protein, and assessment of one’s
Such low scores suffice to increase the risk for cardiac events CAC score [28]. Over a mean follow-up of 7.5 years, subjects
by twofold compared to a CAC score of zero. were assessed for the occurrence of incident CVD. Other than
As with cardiac stress testing, investigators recognized CAC scanning, the other five screening measurements failed
the need to assess whether CAC scanning added incremental to provide a significant discrimination for predicting incident
information to the assessment of cardiac risk by global algo- CVD. By contrast, CAC scanning was highly predictive. In
rithms, such as the Framingham Risk Score, which incorpo- addition, the net reclassification improvement provided by
rate age, gender, and consideration of CAD risk factors. A CAC scanning in the intermediate risk MESA subjects was
potential limitation of such scores is their reliance on only substantially greater than that provided by the other scanning
current risk factor status, without consideration of the chro- techniques evaluated by Yeboah et al. (Table 3.3) [28].
nicity of the risk factors. Moreover, these algorithms only Several aspects of CAC scanning, however, led to a lag in its
provide a probabilistic assessment of the likelihood of CAD becoming a routine part of clinical cardiology until recent years.
based on populations, whereas CAC scanning provides a These include the manner in which it was introduced as a com-
marker of disease in an individual patient, representing the mercial screening test, its initial perception as being a diagnostic
lifetime effect of all risk factors on the coronary vasculature. test (shown to have low specificity for obstructive coronary
Numerous studies have demonstrated that CAC scanning artery disease), the backlash from expert consensus panels and
provides incremental prognostic information compared to guidelines committees from its endorsement, and the persistent
CAD risk factors and global algorithms [26, 27]. For lack of coverage of the method, making it available only to
instance, within the MESA study, the CAC score potently patients who can afford to pay for their own scans.
3  The Long March into Clinical Practice: Cardiac CT and Its Competitors 25

Fig. 3.5  Total cardiac event

rates (per 1000 person-years) 33.1
30.4
according to the combined
consideration of coronary 35

Event Rate per 1000 person-years

artery calcium scores and 20.5
calculated Framingham Risk 30
Score (FRS) category within 15.3 22.1
the MESA study. (From 25
Silverman et al. [27], with 15.2
permission) 20 9.1 11.6
15.4
15
8.8
3.8
4.1
10
CAC > 300
5 1.1 CAC 101–300
1.6 3.5
0 2.5 CAC 1–100

FRS 0–6% CAC = 0

FRS 6–10%
FRS 10–20%
FRS > 20%

CAC = 0 CAC 1–100 CAC 101–300 CAC > 300

Table 3.3  Comparison of net reclassification improvement among p­rognosis. Importantly, this need was recognized in
1330 intermediate Framingham risk subjects in the MESA study ­centers around the world, such that multicenter registries
Variables % NRI became possible.
Brachial FMD 2.4% 2. The need for cardiologists and radiologists and imaging
Ankle-brachial index 3.6% scientists to be involved in order to have the source of
Hs-CRP 7.9% patient referrals and the physicians who generally control
Carotid IMT 10.2% the expensive scanners both being advocates of the
Family history 16.0% methods.
Coronary artery calcium 65.9%
3. The need for standardization of acquisition, processing,
and interpretation methods.
4. The need for a professional organization to be an ­advocate
 he Advent of Coronary CT
T for all aspects of the development of the field.
Angiography (CCTA)
Thus, in 2007, only a few years after the introduction of
Coronary CT angiography became possible only after mul- the 64-slice scanners, the Society of Cardiovascular CT
tislice scanners with rotation speeds became commercially (SCCT) was formed, to support research and education in
available in the early 2000s—with the general availability of cardiac CT and, from its beginning, dedicated to having the
64 slice scanners beginning around 1995. It was thus intro- organization be a collaborative effort between cardiologists
duced at a time in which its developers could learn from the and radiologists.
history of the other methods. Continual hardware and soft- Using the databases that had become part of imaging prac-
ware improvements have led to faster and higher quality tice, early studies focused on the accuracy of CCTA in detect-
scans (improved spatial and temporal resolution) and the ing stenosis compared to invasive coronary angiography, in
ability to perform these scans with very low radiation which it far surpassed the accuracy of the other noninvasive
exposure. imaging techniques—not surprisingly, since they were both
Lessons learned from the experience with the earlier anatomic imaging methods. The prospectively defined data-
methods were applied to coronary CTA. These included the bases provided the tool for rapid reporting of multiple studies
following: showing the prognostic value of CCTA above baseline CAD
risk factors and commonly used risk stratification schemes.
1. The need for systematic development of databases within Further, since centers around the world had learned the need
multiple centers that contained comprehensive clinical for prospective databases, it was possible to have multicenter
and coronary CTA data as well as follow-up information registries that quickly amassed large numbers of patients with
for cardiac events, which then could result in rapid comprehensive clinical and coronary CTA data. The desire of
published evidence for the field in diagnosis and
­ many leaders in the field to provide evidence that would lead
26 S. Uretsky et al.

to broad dissemination of the method and importantly to This limited coverage has its roots in two important develop-
payor coverage and inclusion in guidelines led multiple inves- ments over the prior decade. First, the development of a
tigators to pool their data in such a fashion that individual marked increase in the use of stress imaging modalities
centers might not be recognized for their contribution, but the between 2000 and 2006 [30] caught the attention of Congress
overall field could have the opportunity to rapidly gain in the United States and has led to demands to curb the
strength. Exemplifying this, in a very large registry of 23,854 growth of cardiac testing in general. Thus one aspect of the
patients without known CAD in the Coronary CT Angiography success of nuclear cardiology served to slow the growth of
Evaluation for Clinical Outcomes: An International coronary CTA, as coverage was not as easily gained. This
Multicenter (CONFIRM) registry, there was a significant demand has led to important methods for restraining medical
association between mortality and CCTA findings, according costs, such as the advent of appropriateness criteria, pre-­
to analysis of obstructive CAD by either a per-patient, per- certification, and reduced medical reimbursement. Moreover,
vessel, or per-segment basis [29]. Outcome studies have con- concerns for the rising costs of testing have led to the devel-
sistently demonstrated that the absence of CAD by CCTA is opment of a new standard for evaluating the efficacy of car-
associated with an excellent prognosis and a very long “war- diac tests. For decades, cardiac tests were evaluated solely
ranty period.” according to their diagnostic and prognostic efficacy.
The excellent diagnostic and prognostic accuracy of CCTA Demonstration of such efficacy was sufficient to gain third-­
makes it an attractive test in diagnostic populations. The high party insurance coverage. Instead, cardiac tests are now eval-
negative predictive accuracy of CCTA among patients with a uated according to a standard of “therapeutic efficacy,” as
low to intermediate risk of CAD—with the ability to define illustrated in Fig. 3.6 [31]. Whereas in the past, it was suffi-
absent or minimal CAD in a large proportion of patients cient to demonstrate that a test merely predicts clinical out-
undergoing testing—makes it superior to stress testing for the comes, third-party carriers now demand evidence that a new
diagnostic work-up of patients in this patient cohort. test will also improve clinical outcomes or reduce costs,
thereby providing value (Fig. 3.7). The optimal outcome is to
demonstrate that a test leads to reduction in hard cardiac
 he Advent of Comparative
T events, including cardiac death and myocardial infarction.
Effectiveness Research However, demonstration of improvement in intermediate
outcomes such as reduction in the number of invasive coro-
While each of the imaging modalities reviewed above have nary angiograms that reveal nonobstructive CAD, reduction
diagnostic and prognostic effectiveness, the newer imaging in time in the emergency department, and fewer admissions
modalities, such as the CAC scan and coronary CTA, have to for chest pain in patients who prove not to have acute coro-
date been limited in terms of third-party insurance coverage. nary syndromes. These might also include evidence that a

Fig. 3.6  Traditional versus TRADITIONAL PERSPECTIVE

new standard for assessing the PREDICTS adverse
efficacy of cardiac tests. clinical events?
Traditionally, the efficacy of
cardiac tests was assessed DEMONSTRATE
according to their ability to PROGNOSTIC
predict adverse clinical events EFFICACY
(prognostic efficacy). Now, Provides incremental
tests are also evaluated prognostic information?
according to their ability to
reduce adverse clinical events
(therapeutic efficacy). (From
Rozanski et al. [31], with
permission)
CURRENT PERSPECTIVE
REDUCES adverse
clinical events?

DEMONSTRATE
THERAPEUTIC
EFFICACY

Provides incremental
cost-effectiveness?
3  The Long March into Clinical Practice: Cardiac CT and Its Competitors 27

new test leads to a reduction in CAD risk factors. On the side or a new proposed imaging modality. Patients are then fol-
of costs a new test would show, reduced downstream medical lowed for a specified period of time and then assessed for the
testing, and interventions, such as percutaneous coronary relative frequency of clinical events or other specified
interventions and coronary bypass surgery, and unnecessary outcomes.
hospital admissions. The last decade has seen a proliferation of cardiac imag-
Because of today’s multiplicity of cardiac tests, third-­ ing trials designed to compare new imaging modalities ver-
party carriers also require demonstration of how new tests sus usual care or prior imaging modalities. Most of this
compare to existing tests or clinical management strategies research has centered upon the use of coronary CTA as the
with respect to their ability to improve clinical outcomes and experimental imaging modality. To date, coronary CTA has
reduce downstream medical costs and medical resource uti- been evaluated in three clinical settings: its use for the
lization. This has led to a new era of comparative effective- work-­up of patients presenting to emergency departments
ness research. A basic algorithm for performing such research (ED), patients admitted to the hospital for clinical work-up,
is schematized in Fig. 3.8. This research takes the form of and patients being evaluated for chest pain in outpatient
prospective randomized trials in which a cohort of patients is settings. Comparative effectiveness trials involving coro-
either randomized to usual medical care/conventional testing nary CTA that have been done in the ED setting are listed
in Table 3.4 [32–40].
In a meta-analysis of imaging trials, Hulten et al. found
Value-based Cardiac Imaging in CAD
that randomization to coronary CTA [41] resulted in
decreased length of stay and decreased costs in the emer-
Does the test gency department rather than usual care in the ED. However
result in coronary CTA was associated with an increase in referral for
improved
outcomes or
invasive coronary angiography.
reduce costs? Two small trials have assessed the use of CTA in the inpa-
tient setting. The cardiac CT in the treatment of acute chest
pain (CATCH) study randomized 600 patients to CCTA or
standard of care. The results showed a similarly low fre-
quency of events in both randomized arms, but the duration
of follow-up was short [39]. In the Prospective First
Does the test
PREDICT risk? Evaluation in Chest Pain Trial (PERFECT), 411 patients
admitted to the hospital for the work-up of chest pain were
randomized to either CCTA or stress testing, using either
Fig. 3.7  Concepts of value-based cardiac imaging in CAD stress echo or stress MPI [38]. The study showed no differ-

Fig. 3.8  Paradigm for

conducting comparative
RECRUITMENT OF
effective research for cardiac
CLINICAL COHORT
imaging tests

Management based on usual Management based on new

work-up imaging modality

CLINICAL OUTCOMES CLINICAL OUTCOMES

28 S. Uretsky et al.

Table 3.4  Randomized prospective comparative effectiveness studies for CCTA

Study n Setting Comparison
CCTA SOC
Goldstein et al. [32] 197 ED CCTA SN
CT-STAT [33] 699 ED CCTA SN
ACRIN-PA [34] 1370 ED CCTA ET, SI
ROMICAT II [35] 1000 ED CCTA SOC
PROMISE [36] 10,003 Outpatient CCTA ET, SE, SN
Levsky et al. [37] 400 Inpatient CCTA SN
PERFECT [38] 411 Inpatient CCTA SE, SN
CATCH [39] 600 Inpatient CCTA ET, SN
SCOT-HEART [40] 9849 Outpatient ET + CCTA ET
CCTA coronary CT angiography, ED emergency department, ET exercise testing, SE stress echocardiography, SI stress imaging, SN stress nuclear,
SOC standard of care

ence in downstream testing, hospitalization, or morbidity/ In part based on the SCOT-HEART findings, the NICE
mortality within the two arms of the study. As with some of criteria in Great Britain now have coronary CTA as the first-­
the ED studies, there was a higher frequency of invasive cor- line test of choice in patients with acute or stable chest pain
onary angiography and revascularization procedures within syndromes. While the growth of coronary CTA in the United
the coronary CTA arm. States has been slowed by lack of commercial coverage, its
Two large randomized trials have compared coronary growth worldwide has been dramatic.
CTA vs. either usual care or stress testing in the outpatient
setting. The Scottish Computed Tomography of the HEART
Trial (SCOT-HEART) trial randomized patients with chest Coronary CTA and Its Competitors
pain syndromes to coronary CTA or usual care [40]. The
coronary CTA arm demonstrated increased accuracy and There has increasingly been a recognition in each field that
certainty in the diagnosis of angina—the primary aim of the both anatomic and physiologic measures are needed for
trial. There was a trend toward a decrease in hard cardiac comprehensive use of noninvasive imaging in guiding patient
events. When a delay in implementation of changes in thera- management. For coronary CTA, this recognition has led to
peutic decision-making in the Scottish system was taken into the development of stress CT perfusion (CTP) and, more
account, a 50% reduction in myocardial infarction and car- recently, attempts to assess coronary physiology through CT
diac death was observed [42]. Those who underwent coro- fractional flow reserve (FFRCT). Both approaches have been
nary CTA also had a higher frequency of initiation of medical supported by a series of recent prospective multicenter trials
therapies and cancelation of therapies based on the results of [43–46]. Further, the ability of coronary CTA to evaluate
the coronary CTA. coronary plaque, with its potential to provide automated
The largest imaging trial to date has been the Prospective quantitative assessments of the global coronary plaque bur-
Multicenter Imaging Study for Evaluation of Chest Pain den of lipid-containing plaque, non-calcified plaque, and cal-
(PROMISE) [35]. In the PROMISE study, 10,003 symp- cified plaque, as well as adverse plaque characteristics, is
tomatic patients were randomized to CCTA or stress test- being developed. Quantitative plaque assessment may pro-
ing. At a median follow-up of 2 years, there was no vide important key to prediction of cardiac events and to
difference in clinical outcomes, but there was a higher fre- monitoring the effects of medical therapy. It is likely that the
quency of invasive angiography in the CCTA arm. However, profound ability of coronary CTA to detect CAD, evaluate its
among those referred for ICA, 72% of patients in the coro- extent and severity, assess its physiologic significance, and
nary CTA arm had obstructive CAD, compared to only provide quantitative measures of plaque and plaque vulner-
48% of patients in the stress testing arm. Of note, however, ability will revolutionize cardiac imaging. Furthermore, the
the prevalence of obstructive CAD was far lower in the previous high radiation doses initially associate with coro-
PROMISE trial than anticipated (10.7% observed versus nary CTA have been dramatically reduced with new instru-
53.3% predicted), and the overall event rate was very low. mentation and new modes of acquisition.
These crucial factors seriously limited the overall utility of These and other newer advances of CCTA are covered
this study. Nevertheless, the study demonstrated that coro- within the chapters of this book. Of note, however, many of
nary CTA had comparable efficacy compared to stress test- the “competitors” of CCTA are also “on the march,” benefit-
ing in the PROMISE cohort. ing from important improvements in their technology.
3  The Long March into Clinical Practice: Cardiac CT and Its Competitors 29

Advances in Nuclear Cardiology agent will allow use of PET MPI without the need for an
expensive on-site generator or cyclotron and can also be used
Nuclear cardiology has enjoyed substantial technological with exercise. Further, PET is routinely performed with car-
improvements in recent years. The recent development of diac CT, allowing the routine measurement of CAC—thus
solid-state detectors (e.g., cadmium zinc telluride) and adding competition to coronary CTA.  SPECT CT systems
improved collimation has resulted in improved energy reso- are also becoming more common. These developments could
lution and reduction in undesirable scatter events. Application allow for the improved effect of MPI testing on outcomes, by
of attenuation correction is used to improve the accuracy of assessment of coronary atherosclerosis as well as myocardial
SPECT-MPI by eliminating artificial defects, thus permitting perfusion.
a higher frequency of stress-only imaging studies. The com-
bination of new SPECT camera instrumentation with current
improvements in reconstruction software is not only Advances in Stress Echocardiography
improved image quality but also markedly reduced radiation
exposure to patients during the MPI scan. With new systems, As with the other imaging modalities, there have been impor-
stress-only SPECT-MPI can be performed with radiation tant advances in the technical capabilities of stress echocar-
doses as low as 1 mSv. The improvements in software have diography. Manufacturing improvements have improved the
also improved the automation of quantitative analysis of MPI post-processing of ultrasound studies and increased the tem-
images and now permit the incorporation of all observed poral resolution of the acquired images. Contrast echo has
variables for decision-making through the application of improved endocardial definition, further improved resolu-
machine learning. tion, and decreased the number of nondiagnostic studies. In
PET MPI is performed with very low radiation doses—in addition, newer ultrasound approaches for assessing periph-
the range of 1 mSv. Importantly, PET MPI provides a mea- eral arterial beds allow application of echocardiography for
surement of absolute stress myocardial blood flow and coro- the detection of subclinical atherosclerosis. This includes the
nary flow reserve. These features provide information that potential use of three-dimensional carotid ultrasound to
adds to the prognostic value of perfusion defect assessment characterize carotid plaque burden. In the BioImage Study,
and provides a better understanding of the cause of patients’ carotid plaque burden was comparable to CAC scanning for
symptoms [47] (Fig. 3.9). A new PET scanning agent with predicting adverse cardiac events [49]. Femoral ultrasound
ideal perfusion tracer characteristics—F-18 flurpiridaz— is can also be used for identifying atherosclerotic plaques in
undergoing phase III clinical trials [48] (Fig.  3.10). This this vascular bed. In the Aragon Workers’ Health Study,

Fig. 3.9 Unadjusted Coronary Flow Reserve (CFR) Predicts Mortality

annualized cardiac mortality Independent of Perfusion Defects
by tertiles of CFR and
categories of stress SPECT-­ N=2,783
MPI perfusion defect (PD).
CD = cardiac mortality. For CD=137
each category of perfusion
finding, there is increasing 12%
P<0.0001
Annualized Mortality

mortality associated with 10%

decreasing tertiles of CFR.
8%
(From Murthy et al. [47], with
permission) 6%
4%
Lower Tertile
2%
Middle Tertile
0%
Upper Tertile
≥10%
1-9%
0%

≥10% (n) 1-9% (n) 0% (n)

Upper Tertile 2.4% (119) 0.3% (195) 0.1% (614)
Middle Tertile 4.4% (217) 4.0% (202) 1.1% (509)
Lower Tertile 10.2% (416) 6.0% (190) 3.6% (321)

Murthy VL, et al. Circulation 2011

30 S. Uretsky et al.

Fig. 3.10  Adenosine stress

(adeno) and rest apical,
18F-Flurpiridaz PET
mid-ventricular, and basal
short-axis and vertical and
horizontal mid-ventricular BMS
7476158 APICAL MID BASAL VLA HLA
long-axis views acquired with Stress Extent(%)

flurpiridaz F-18 PET (top)

and Tc-99m sestamibi (MIBI) ADENO
SPECT bottom. Quantitative
polar maps reconstructed
from the same PET and
SPECT studies are shown on REST
the right. The extent and
severity of induced perfusion
defect are more apparent in
the PET images. The MIBI APICAL MID BASAL VLA HLA
Stress Extent(%)
quantitative TPD with the
99mTc-sestamibi SPECT ADENO
study was 17% and was 32%
with flurpiridaz F-18 PET.
(From Berman et al. [56],
with permission)
REST

assessment of the femoral arteries revealed greater plaque with CAC scanning has been placed almost exclusively on the
detection than assessment of atherosclerosis by carotid ultra- total Agatston CAC score. However, recent studies suggest
sound or CAC scanning [50]. A high sensitivity for detecting that additional useful information may be obtained from the
atherosclerotic plaque by femoral ultrasound was also dem- CAC scan, including assessment of the number of CAC
onstrated in the Progression of Early Atherosclerosis Study plaques, the density of the plaques, and pericardial fat [52, 53].
[51]. Ultrasound approaches could thus be used to assess
stress-induced ischemia as well as subclinical atherosclero-
sis, if routine combination of plaque imaging and cardiac Future Directions
imaging were performed.
As the technical capabilities of CCTA and its competitors
improve, the new field of cardiac comparative effectiveness
Advances in Cardiac MRI research will need to develop accordingly. As today’s imag-
ing technologies reach a relative technical maturation, there
Technological advances also continue to improve the capa- will be impetus to compare these imaging modalities with
bilities of cardiac MRI. Recent advances include the advent new prospective randomized imaging trials which will seek
of 3-Tesla scanners and the refinement of MRI sequences to to determine the relative effectiveness of these imaging com-
allow shorter breath holds. In addition, emerging modalities petitors for improving medical outcomes. These imaging tri-
such as T1 mapping and 4D flow sequences are refining the als will seek to determine which techniques provide the most
diagnostic capabilities of MRI.  Importantly, major strides effective reduction in adverse clinical events in the most eco-
are being made in the standardization of cardiac MRI acqui- nomical way. These trials may also consider combinations of
sition and processing methods, reducing the reliance on tech- technologies that can be used in a cost-effective manner. For
nical expertise. In the United States, it is anticipated that example, whereas exercise electrocardiography, a technique
approval of a gadolinium contrast agent for myocardial per- that has not changed since the 1960s, provides only limited
fusion may allow for a broader dissemination of information diagnostic and prognostic information, its utility can be
regarding the method. enhanced by combining it with a CAC scan into a hybrid test
that has been termed the “treadmill calcium test” [54].
Similarly, MPI scans can now be combined with atheroscle-
Advances in CAC Scanning rosis imaging using combined SPECT/CT and PET/CT cam-
era systems. These test combinations will deserve strong
The technological advances with CT have also benefited CAC consideration as trials are developed to determine the opti-
scanning by permitting studies that can now be obtained with mal method for the work- up patients for diagnostic and
less than 1 mSv of radiation exposure. Until recently, ­emphasis prognostic assessment.
3  The Long March into Clinical Practice: Cardiac CT and Its Competitors 31

Fig. 3.11  The temporal 45%

prevalence of abnormal n=39,515 pts
SPECT-MPI studies and 40%
ischemic SPECT-MPI studies
between 1991 and 2009. 35%
40.9%
(From Rozanski et al. [55],
with permission) 30%

25%

20% 29.6%

15%

10%

5% 8.7%

0%
1991 1993 1995 1997 1999 2001 2003 2005 2007 2009
5.0%
% Ischemia % Abnormal SPECT

An experience that has been garnered over four decades with growth in testing in areas of proven value and reduction
has taught us that there is no one ideal imaging test for all of testing in areas in which value is not provided or has not
clinical scenarios. Rather, the value of a test is always depen- been shown. Provision of value will be the metric upon which
dent upon the goal of testing and upon disease and/or isch- imaging tests will be judged. The long march into clinical
emic prevalence. Notably, the frequency of inducible practice of cardiac CT will be inherently linked to the evi-
ischemia is itself a now changing variable, as the frequency dence of the value of this powerful modality.
of inducible myocardial ischemia recedes due to enhanced
prevention and the earlier detection of CAD in modern clini-
cal practice (Fig. 3.11) [55]. In various settings, such as the References
ED, imaging technologies will also compete against more
effective serum markers of disease, such as high-sensitivity 1. Dalen JE, et al. The epidemic of the 20(th) century: coronary heart
disease. Am J Med. 2014;127(9):807–12.
troponin assays. Thus, the future assessment of what consti- 2. Master AM.  The two-step exercise electrocardiogram: a test for
tutes optimal imaging technologies is likely to be a dynamic coronary insufficiency. Ann Intern Med. 1950;32(5):842–63.
process. What is certain, however, is that as our imaging 3. Bruce RA, et  al. Exercising testing in adult normal subjects and
capabilities improve, they are likely to retain and even grow cardiac patients. Pediatrics. 1963;32:SUPPL 742–56.
4. Coplan NL. Evaluation of patients for coronary artery bypass sur-
in the central role that they have assumed for screening, diag- gery: the role of exercise testing. Am Heart J. 1991;122(6):1800–2.
nostic assessment, and/or risk stratification among patients 5. Mark DB, et al. Exercise treadmill score for predicting prognosis in
with suspected and known CAD. coronary artery disease. Ann Intern Med. 1987;106(6):793–800.
The rapid evolution of new medical devices and therapies 6. Zaret BL, et  al. Noninvasive regional myocardial perfusion with
radioactive potassium. Study of patients at rest, with exercise and
and the increasing numbers of patients being studied or during angina pectoris. N Engl J Med. 1973;288(16):809–12.
treated based on these developments are leading to an unsus- 7. Berman DS, et  al. Noninvasive detection of regional myocar-
tainable increase in health-care costs. In the United States, dial ischemia using rubidium-81 and the scintillation camera:
there will be a transition from a volume-based health-care comparison with stress electrocardiography in patients with
arteriographically documented coronary stenosis. Circulation.
reimbursement to a value-based reimbursement. Inherently, 1975;52(4):619–26.
as with any commodity, value is a function of quality and 8. Wackers FJ, et  al. Prognostic significance of normal quantitative
cost. In cardiology, quality ultimately implies not just risk planar thallium-201 stress scintigraphy in patients with chest pain.
stratification but an improvement in outcomes or reduction in J Am Coll Cardiol. 1985;6(1):27–30.
9. Germano G, et  al. Quantitative LVEF and qualitative regional
costs. In the patient with CAD, these outcomes might be function from gated thallium-201 perfusion SPECT. J Nucl Med.
reduced cardiac events such as death or myocardial infarction 1997;38(5):749–54.
or improvement in quality of life and other measures. In the 10. Garcia EV, Maddahi J, Berman DS, Waxman A. Space-time quan-
future, whether the payer is the government, insurance com- titation of thallium-201 myocardial scintigraphy. J Nucl Med.
1981;22(4):309–17.
panies, or the patients themselves, it is likely that only those 11. Garcia EV, Van Train K, Maddahi J, Prigent F, Friedman J, Areeda J,
approaches that provide value will be purchased. In imaging, Waxman A, Berman DS. Quantification of rotational thallium-201
this implies an increasing penetrance of value-­based imaging, myocardial tomography. J Nucl Med. 1985;26(1):17–26.
32 S. Uretsky et al.

12. Ladenheim ML, et al. Extent and severity of myocardial hypoperfu- 31. Rozanski A, Muhlestein JB, Berman DS.  Primary prevention

sion as predictors of prognosis in patients with suspected coronary of CVD: the role of imaging trials. JACC Cardiovasc Imaging.
artery disease. J Am Coll Cardiol. 1986;7(3):464–71. 2017;10(3):304–17.
13. Weiss AT, et al. Transient ischemic dilation of the left ventricle on 32.
Goldstein JA, et  al. The CT-STAT (coronary computed
stress thallium-201 scintigraphy: a marker of severe and extensive tomographic angiography for systematic triage of acute
coronary artery disease. J Am Coll Cardiol. 1987;9(4):752–9. chest pain patients to treatment) trial. J Am Coll Cardiol.
14. Berman DS, et  al. Separate acquisition rest thallium-201/stress 2011;58(14):1414–22.
technetium-99m sestamibi dual-isotope myocardial perfusion 33. Goldstein JA, et  al. A randomized controlled trial of multi-slice
single-­photon emission computed tomography: a clinical validation coronary computed tomography for evaluation of acute chest pain.
study. J Am Coll Cardiol. 1993;22(5):1455–64. J Am Coll Cardiol. 2007;49(8):863–71.
15. Hachamovitch R, et  al. Comparison of the short-term survival
34. Litt HI, et  al. CT angiography for safe discharge of patients

benefit associated with revascularization compared with medical with possible acute coronary syndromes. N Engl J Med.
therapy in patients with no prior coronary artery disease undergo- 2012;366(15):1393–403.
ing stress myocardial perfusion single photon emission computed 35. Douglas PS, et al. Outcomes of anatomical versus functional testing
tomography. Circulation. 2003;107(23):2900–7. for coronary artery disease. N Engl J Med. 2015;372:1291–300.
16. Rozanski A, et al. Comparison of long-term mortality risk follow- 36.
Hoffmann U, et  al. Coronary CT angiography versus
ing normal exercise vs adenosine myocardial perfusion SPECT. J standard evaluation in acute chest pain. N Engl J Med.
Nucl Cardiol. 2010;17(6):999–1008. 2012;367(4):299–308.
17. Rozanski A, et al. Long-term mortality following normal exercise 37. Levsky JM, Travin MI, Haramati LB. Coronary computed tomog-
myocardial perfusion SPECT according to coronary disease risk raphy angiography versus radionuclide myocardial perfusion imag-
factors. J Nucl Cardiol. 2014;21(2):341–50. ing in patients with chest pain admitted to telemetry: a randomized,
18. Supariwala A, et al. Influence of mode of stress and coronary risk controlled trial. Ann Intern Med. 2016;164(2):133–4.
factor burden upon long-term mortality following normal stress 38. Uretsky S, et al. Comparative effectiveness of coronary CT angi-
myocardial perfusion single-photon emission computed tomo- ography vs stress cardiac imaging in patients following hos-
graphic imaging. Am J Cardiol. 2013;111(6):846–50. pital admission for chest pain work-up: the Prospective First
19. Hachamovitch R, et al. Incremental prognostic value of myocardial Evaluation in Chest Pain (PERFECT) Trial. J Nucl Cardiol.
perfusion single photon emission computed tomography for the pre- 2017;24(4):1267–78.
diction of cardiac death: differential stratification for risk of cardiac 39. Linde JJ, et  al. Long-term clinical impact of coronary CT angi-
death and myocardial infarction. Circulation. 1998;97(6):535–43. ography in patients with recent acute-onset chest pain: the ran-
20. Kang X, et al. Incremental prognostic value of myocardial perfu- domized controlled CATCH trial. JACC Cardiovasc Imaging.
sion single photon emission computed tomography in patients with 2015;8(12):1404–13.
diabetes mellitus. Am Heart J. 1999;138(6 Pt 1):1025–32. 40.
Investigators, S.-H.  CT coronary angiography in patients
21. Yao SS, et al. Practical applications in stress echocardiography: risk with suspected angina due to coronary heart disease (SCOT-­
stratification and prognosis in patients with known or suspected HEART): an open-label, parallel-group, multicentre trial. Lancet.
ischemic heart disease. J Am Coll Cardiol. 2003;42(6):1084–90. 2015;385(9985):2383–91.
22. Rumberger JA, et  al. Coronary artery calcium area by electron- 41. Hulten E, et  al. Outcomes after coronary computed tomography
beam computed tomography and coronary atherosclerotic angiography in the emergency department: a systematic review and
plaque area. A histopathologic correlative study. Circulation. meta-analysis of randomized, controlled trials. J Am Coll Cardiol.
1995;92(8):2157–62. 2013;61(8):880–92.
23. Budoff MJ, et  al. Long-term prognosis associated with coronary 42. Williams MC, et al. Use of coronary computed tomographic angi-
calcification: observations from a registry of 25,253 patients. J Am ography to guide management of patients with coronary disease. J
Coll Cardiol. 2007;49(18):1860–70. Am Coll Cardiol. 2016;67(15):1759–68.
24. Detrano R, et  al. Coronary calcium as a predictor of coro-
43. Cury RC, et  al. Dipyridamole stress and rest myocardial per-
nary events in four racial or ethnic groups. N Engl J Med. fusion by 64-detector row computed tomography in patients
2008;358(13):1336–45. with suspected coronary artery disease. Am J Cardiol.
25. Polonsky TS, et  al. Coronary artery calcium score and risk
2010;106(3):310–5.
classification for coronary heart disease prediction. JAMA. 44. Min JK, et al. Diagnostic accuracy of fractional flow reserve from
2010;303(16):1610–6. anatomic CT angiography. JAMA. 2012;308(12):1237–45.
26. Nasir K, et al. Interplay of coronary artery calcification and tradi- 45. Norgaard BL, et  al. Diagnostic performance of noninvasive frac-
tional risk factors for the prediction of all-cause mortality in asymp- tional flow reserve derived from coronary computed tomography
tomatic individuals. Circ Cardiovasc Imaging. 2012;5(4):467–73. angiography in suspected coronary artery disease: the NXT trial
27. Silverman MG, et al. Impact of coronary artery calcium on coro- (analysis of coronary blood flow using CT angiography: next
nary heart disease events in individuals at the extremes of tradi- steps). J Am Coll Cardiol. 2014;63(12):1145–55.
tional risk factor burden: the Multi-Ethnic Study of Atherosclerosis. 46. Rocha-Filho JA, et  al. Incremental value of adenosine-induced

Eur Heart J. 2014;35(33):2232–41. stress myocardial perfusion imaging with dual-source CT at cardiac
28. Yeboah J, et al. Comparison of novel risk markers for improvement CT angiography. Radiology. 2010;254(2):410–9.
in cardiovascular risk assessment in intermediate-risk individuals. 47. Murthy VL, et  al. Improved cardiac risk assessment with non-
JAMA. 2012;308(8):788–95. invasive measures of coronary flow reserve. Circulation.
29. Min JK, et al. Age- and sex-related differences in all-cause mortality 2011;124(20):2215–24.
risk based on coronary computed tomography angiography findings 48. Berman DS, et  al. Phase II safety and clinical comparison with
results from the International Multicenter CONFIRM (Coronary CT single-photon emission computed tomography myocardial per-
Angiography Evaluation for Clinical Outcomes: An International fusion imaging for detection of coronary artery disease: flur-
Multicenter Registry) of 23,854 patients without known coronary piridaz F 18 positron emission tomography. J Am Coll Cardiol.
artery disease. J Am Coll Cardiol. 2011;58(8):849–60. 2013;61(4):469–77.
30. Shaw LJ, et  al. Why all the focus on cardiac imaging? JACC 49. Baber U, et  al. Prevalence, impact, and predictive value of

Cardiovasc Imaging. 2010;3(7):789–94. detecting subclinical coronary and carotid atherosclerosis in
3  The Long March into Clinical Practice: Cardiac CT and Its Competitors 33

asymptomatic adults: the BioImage study. J Am Coll Cardiol. 53. Rozanski A, Slomka P, Berman SD. Extending the use of coronary
2015;65(11):1065–74. calcium scanning to clinical rather than just screening populations:
50. Laclaustra M, et al. Femoral and carotid subclinical atherosclero- ready for prime time. Circ Cardiovasc Imaging. 2016;9(5):e004876.
sis association with risk factors and coronary calcium: the AWHS 54. Rozanski A, Cohen R, Uretsky S. The coronary calcium treadmill
study. J Am Coll Cardiol. 2016;67(11):1263–74. test: a new approach to the initial workup of patients with suspected
51. Fernandez-Friera L, et  al. Prevalence, vascular distribution, and coronary artery disease. J Nucl Cardiol. 2013;20(5):719–30.
multiterritorial extent of subclinical atherosclerosis in a middle-­ 55. Rozanski A, et  al. Temporal trends in the frequency of inducible
aged cohort: the PESA (progression of early subclinical atheroscle- myocardial ischemia during cardiac stress testing: 1991 to 2009. J
rosis) study. Circulation. 2015;131(24):2104–13. Am Coll Cardiol. 2013;61(10):1054–65.
52. Berman DS, Arnson Y, Rozanski A. Coronary artery calcium scan- 56. Berman DS, Germano G, Slomka PJ. Improvement in PET myo-
ning: the Agatston score and beyond. JACC Cardiovasc Imaging. cardial perfusion image quality and quantification with flurpiridaz
2016;9(12):1417–9. F 18. J Nucl Cardiol. 2012;19(Suppl 1):S38–45.
 Part II
Where We Are: Human Requisites
 Cardiac Computed Tomography:
A Team Sport 4
Sheldon E. Litwin

Nature abhors a vacuum. – Attributed to Aristotle

All politics are local. – Attributed to Tip O’Neill

Historical Perspective based coronary angiography and nuclear cardiology, it is not

altogether surprising that tensions also developed in this new
The concepts underlying computed tomography (CT) were landscape. The term “tribalism” refers to a strong cultural or
first developed by Godfrey Hounsfield and Allan Cormack in ethnic identity that separates one group from another. In
the early 1970s [1]. Although Hounsfield attempted to use medicine, tribes are most often defined by specialty or
gated CT for cardiac imaging in the mid-1970s [2], nearly departmental affiliations. The conflicts that exist regarding
30 years elapsed before this became a reality in clinical med- the modern practice of cardiac CT generally relate to debates
icine. In the mid-2000s, a number of hardware and software over which medical specialty is best suited to oversee the
advances including the use of multiple rows of detectors and equipment and technologists, the interpretation of the
helical imaging allowed us to overcome the technical limita- images, communications with referring providers, and the
tion imposed by cardiac and respiratory motion. The long financial aspects of the service being rendered. An underly-
slow road to the advent of modern cardiac CT was followed ing premise in the dispute is that cardiologists have tradition-
by a decade of accelerated development and refinement of ally “owned” the body part being studied (the heart) while
the hardware, imaging approaches, processing techniques, radiologists have been the keepers of the CT scanner. Before
and interpretation skills. This rapid expansion led to some- discussing the arguments put forth over these issues more
what of a vacuum in the proprietorship of this field. specifically, it is worth recalling that the pioneers in this field
were an engineer and a physicist, not cardiologist or radiolo-
gist. Arguably, neither medical specialty owns exclusive
Nature Abhors a Vacuum rights to the use of this imaging modality for evaluation of
the heart.
As one would expect, the initially empty space surrounding
the rapidly expanding use of cardiac CT strongly attracted
enterprising individuals with a variety of backgrounds and Imaging Turf Wars
training. The sudden realization that we could noninvasively
image coronary arteries at high spatial resolution was a pow- The term “turf war” has been used to describe a dispute
erful draw for cardiologists who have always been fascinated between criminals and gangs over the right to operate within
by coronary anatomy. The striking views of the heart and a particular area. More recently this term has been aptly
vasculature that could be created using newer approaches applied to conceptually similar struggles in the medical
such as curved multiplanar reconstructions and three-­ field. A recent Internet search of the terms “medical turf
dimensional volume rendering also proved irresistible to wars” yielded >240,000 hits. Obviously, this is a topic that
radiologists. Given a history of struggles over ownership of has commanded quite a bit of attention. Cardiac CT is not
other cardiac imaging modalities such as invasive or catheter-­ the only battleground, but it is one of the arenas where the
tug-­of-­war has been relatively public. The degree of conflict
over cardiac CT and the various solutions to the struggle
have varied rather widely in different parts of the country or
the world. From the “30,000 foot view,” it is relatively easy
S. E. Litwin (*) to see the benefits of a cooperative approach to running a
Division of Cardiology, Department of Medicine, Medical cardiac CT program with the opportunity for input and
University of South Carolina, Charleston, SC, USA
active participation from all interested and qualified parties.
e-mail: [email protected]

© Humana Press 2019 37

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_4
38 S. E. Litwin

However, many complexities such as institutional tracking services. Workflow for cardiologists and radiologists has tra-
of individual physician or departmental productivity can ditionally been different, and this impacts the ability to work
easily derail the ideal vision of how to optimally deliver a together in a seamless fashion. Radiology reading rooms are
clinical service. It is typically easier to endorse the broad often physically located near the CT scanners where cardiac
concept of truly integrated cardiac imaging services, than it imaging is performed, and hence, radiologists are more
is to bring that concept to fruition at one’s own institution. immediately available to troubleshoot issues at the time of
At a national or international level, organizations like the scans and to do immediate wet reads. Cardiologists on the
Society for Cardiac Computed Tomography bring together other hand have conventionally spent much of their time see-
many individuals with a variety of backgrounds. The shared ing patients in an office or hospital setting and doing invasive
enthusiasm for cardiac CT allows such a group to have com- or noninvasive procedures in areas that are distant from the
munal goals such as increasing availability, standardizing CT scanner. Thus, they are generally less available for imme-
reporting, improving insurance coverage, and other factors diate issues related to the scan procedures. Use of cardiac CT
relevant to this imaging modality. The big picture notwith- in the emergency department for evaluation of chest pain is
standing the primacy of local politics often wins out to the one of the most rapidly growing indications for the tech-
detriment of many. nique. To make the test most valuable to patients and to the
system, rapid interpretation is necessary. A shortened time to
diagnosis is one of the major advantages of cardiac CT com-
Competition Versus Collaboration pared to other approaches. Again, this favors the use of inter-
preting physicians who have proximate access to computer
The issue at hand is how to effectively bring together the workstations with post processing capability and the hard-
competing groups in a fashion that is mutually beneficial and ware and software solutions for immediate reporting. In
results in the best and most efficient patient care. Most would most institutions, the volume of cardiac CT scans is still rela-
agree that cardiologists and radiologists bring unique and tively low, and it is challenging for individual practitioners to
complementary skill sets and knowledge bases to the disci- maintain competency if this volume is spread among multi-
pline of cardiac CT. Cardiologists have substantial familiar- ple different readers. All of these issues are potentially solv-
ity with cardiac and coronary anatomy and physiology and able in an era where increasing numbers of CT technologists
are accustomed to looking at moving (cine) images of the are trained in performing ECG-gated acquisitions without
heart. Assessment of global function, regional wall motion, the need for physician supervision and thin client or cloud-­
and heart valves is second nature to most cardiologists. based image interpretation can be done anywhere.
However, they are less likely to be comfortable with CT Perhaps the most vexing challenge to shared ownership
technology, injection protocols, recognition of artifacts, use and management of cardiac CT programs relates to the eco-
of conventional CT imaging planes, image post processing, nomic issues resulting from the ever-increasing pressure to
and evaluation of extra-cardiac anatomy. In contrast, radiolo- raise productivity and lower costs for health systems. As a
gists tend to be more adept at performing mental three-­ means to achieve this goal, many systems closely track
dimensional reconstructions of anatomy while viewing axial individual physician revenues or relative value units (aka,
images and are highly familiar with most of the anatomy in RVUs). In general, only one physician can be the official
the chest outside of the heart. They are usually very comfort- interpreter of a test. That person is also the one who bills
able performing post processing of images and manipulating for the interpretation. This focus on individual productivity,
volumetric data sets. However, many radiologists are less at rather than team work, has an obvious chilling effect on any
ease with viewing and interpreting dynamic images, such as arrangement that results in splitting of reimbursement or on
those of the beating heart. Radiologists may also have less any process that may slow down the pace of interpreting
familiarity with the clinical implications of using certain studies (i.e., joint reading sessions). The technical charges
descriptive terms such as “coronary artery disease,” which associated with a procedure generally only go to one
can be used with reference to nonobstructive coronary ath- department – again, a strong disincentive to cross-depart-
erosclerosis, although this terminology is frequently assumed mental collaboration. Coming up with equitable ways to
to mean that obstructive plaques are present. share revenues, or to assign credit for communal work or
The advantages of a collaborative approach to the devel- program building, does not fit very with most current
opment and management of a cardiac CT program are abun- accounting systems.
dantly clear. Unfortunately, major barriers to cooperation Creativity and thinking outside the customary departmen-
still exist. Some of the more important obstacles include tra- tal structures are needed to make a program work. A commit-
dition, departmental silos, turf, time, billing, volume of stud- ment from hospital, departmental, and division leaders is a
ies, maintenance of interpretive competency, management of necessity in order to overcome the many obstacles. The cre-
equipment, supervision of technical staff, and location of ation of service lines is one approach to breaking down
4  Cardiac Computed Tomography: A Team Sport 39

departmental barriers. However, in most cases, the service invasive angiography suite. If this comes to pass, there will
line is not the fiscal accounting unit, and the solid and dotted be a need for many more physicians who are competent to
lines on the new organizational charts become blurred. interpret cardiac CT and perhaps competition will wane.
Frequently radiology is considered to be a service unit, like “Miniaturization” and reduced cost of CT hardware may
anesthesiology and pathology, and is not included as part of allow decentralization of cardiac CT imaging locations into
a cardiovascular service line. To be successful, there is a smaller hospitals and clinics. Recently an imaging think
need for real equality among the participants, with each tank was convened to discuss the future of cardiac imaging
member of the team playing an important role. Transparency [4]. The group made many recommendations including
about financial issues and clear delineation of responsibili- redefining a cardiac imager as someone with a high level of
ties and roles must exist. Excellent communication, shared expertise in at least one imaging modality, but not necessar-
values, and mutual respect are cornerstones of a successful ily more than one modality. In an accompanying editorial,
program, but these are not things that are taught in medical the authors made the case that such an approach may work
school nor nurtured in subsequent training. It is often easier well in academic institutions or large hospitals, but is not
to retreat to the tribal mentality, than to expend the effort to well suited to regional hospitals or smaller systems where
bridge political divides. most clinical care actually occurs. They opined that we
When cardiac CT is practiced as a team sport, it creates should instead direct our efforts toward developing imagers
many rich opportunities. As in any team sport, individuals with a broader skill and knowledge base [5]. Using this
play different positions, and the depth of talent on any team, approach, they believe that clinicians will be more likely to
not the skill of one or a few individuals, typically determines choose an optimal test for an individual patient based on
its success. The analogy to the varying contributions of car- their specific characteristics, rather than choosing on the
diologists and radiologists to the sport of cardiac CT is apt. basis of existing practice patterns or local islands of exper-
Jointly managed programs have clear symbiosis when it tise. How we train the next generation of cardiac imagers is
comes to education of medial trainees who can greatly ben- likely to affect what the imaging teams of the future will
efit from the collective wisdom of diverse teachers. Research look like and how they will function. Hopefully that future
programs, as well, are much more likely to thrive when the health-care system will be focused more on health and less
ideas of many are allowed to come together. Intriguingly, the on politics. In the meantime, efforts to forge a permanent
progressive push toward value-based health care, which partnership between the radiology and cardiology commu-
emphasizes the importance of population health and cost nities can yield benefits for all.
effectiveness, rather than volume of procedures or tests, may
be the thing that ultimately eliminates departmental silos and
reduces economic competition within institutions. If and References
when this enlightened approach to health care becomes
engrained, other benefits such as improved radiation stew- 1. Hounsfield GN.  Historical notes on computerized axial tomogra-
phy. J Can Assoc Radiol. 1976;27(3):135–42.
ardship [3], more flexible night and weekend coverage, and 2. Sagel SS, et al. Gated computed tomography of the human heart.
reduction of layered or redundant testing may become easier Investig Radiol. 1977;12(6):563–6.
to realize. As in any group practice, veneration among the 3. Douglas PS, et  al. Developing an action plan for patient radia-
involved individuals is the foundation upon which teamwork tion safety in adult cardiovascular medicine. Proceedings from the
Duke University Clinical Research Institute/American College of
is built. Cardiology Foundation/American Heart Association Think Tank
The field of cardiac CT is still relatively young and the Held on February 28, 2011. J Nucl Cardiol. 2012;19(3):534–50.
optimal organization of a program represents a moving tar- 4. Douglas PS, et al. The future of cardiac imaging: report of a think
get. A number of forces will affect how CT programs are tank convened by the American College of Cardiology. JACC
Cardiovasc Imaging. 2016;9(10):1211–23.
actualized as the field matures. It is likely that CT will pro- 5. Marwick TH, Chandrashekhar Y, Narula J.  Training in multimo-
gressively assume portions of the diagnostic imaging vol- dality CV imaging: is there an inclusive model? JACC Cardiovasc
ume previously performed in nuclear laboratories or in the Imaging. 2016;9(10):1235–7.
 Cardiac CT: Credentialing
and Accreditation 5
James M. Kofler, Heidi A. Edmonson, Shuai Leng,
and Eric E. Williamson

 ccreditation of Cardiac CT Scanners:

A pdf?la=en). For the application of the cardiac module, if one
Scanner and Technologist Requirements or more other modules are included in the same application,
at least one clinical adult coronary CTA case must be submit-
There are currently four different organizations recognized ted, unless the unit is used solely for patients of 18 years of
by the Centers for Medicare and Medicaid Services (CMS) age or younger, in which case a pediatric cardiac examina-
as accreditors for CT imaging, including the American tion is required. If the unit is used for both adult and pediatric
College of Radiology (ACR), the Joint Commission (TJC), patients, then at least one pediatric cardiac CT exam is
the Intersocietal Accreditation Commission (IAC, formerly required to be submitted. The non-cardiac modules require
Intersocietal Commission for the Accreditation of Vascular submission of additional cases specific to the module. For
Laboratories (ICAVL)), and RadSite [1]. In addition to gen- accreditation applications specific to a cardiac module, the
eral CT imaging accreditation, the ACR has a specific car- requirements are as follows: adult-only cardiac CT requires
diac module, and the IAC has accreditation programs for two coronary CT clinical cases; pediatric-only cardiac CT
coronary calcium scoring CT and coronary CTA exams. TJC requires two pediatric cases; and for both adult and pediatric
and RadSite have no specific cardiac requirements for CT cardiac CT, at least one (of three) clinical case must be a
scanners, although RadSite requires submission of a sample pediatric patient (http://www.acraccreditation.org/~/media/
cardiac case if cardiac CT is part of the clinical practice. The ACRAccreditation/Documents/CT/Requirements.
commonality of all four accreditors is that the CT equipment pdf?la=en). Complete case sets must be submitted for each
specifications and performance meet all state and federal CT unit application, including a copy of the clinic protocol
requirements. Beyond this requirement, the criterion for for each. Full details on the required cases for submission are
equipment accreditation and the technologist credentials dif- available from the ACR [2]. Minimum criteria for scan
fers among the accreditors. parameters for various exams, including cardiac exams, are
listed in the ACR accreditation program Testing Instructions
[3]. The clinical images are reviewed by multiple radiologist
The ACR CT Accreditation Program reviewers who use a range of criteria, such as image quality,
appropriateness of scan parameters, scan range, etc., to
Scanner Requirements determine if the submission meets the requirements for ACR
The ACR requires submission of clinical examples, which accreditation.
are assumed to represent the best work of the site, not just The ACR also requires phantom testing and radiation
typical cases. The type and number of clinical cases depend dosimetry measurements for accreditation. This testing must
on the number and type of modules that are being submitted be completed by a “qualified medical physicist” (QMP),
for accreditation. The module choices are adult abdomen, which is defined in the personnel requirements [4]. The
adult head, pediatric abdomen, pediatric head, and adult and/ imaging phantom required for the physics testing was devel-
or pediatric cardiac CT (http://www.acraccreditation.org/~/ oped specifically for the ACR accreditation program and is
media/ACRAccreditation/Documents/CT/Requirements. available through a separate vendor (Gammex 464, Sun
Nuclear, Middleton, WI). Sections of the phantom are
designed to measure CT number accuracy, low-contrast reso-
J. M. Kofler (*)
Department of Radiology, Mayo Clinic, Jacksonville, FL, USA lution, and image uniformity. Phantom measurements are
e-mail: [email protected] acquired using the site’s clinical scan parameters for the
H. A. Edmonson · S. Leng · E. E. Williamson ­relevant protocols, depending on the modules included for
Department of Radiology, Mayo Clinic, Rochester, MN, USA the submission, but none include cardiac-specific protocol

© Humana Press 2019 41

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_5
42 J. M. Kofler et al.

testing. Radiation dosimetry measurements and images must vant to CT imaging, radiological sciences, and patient care,
also be submitted. These are also acquired using the site’s is required. In these cases, a documentation of the continuing
clinical scan parameters, but, again, none are specific to car- education credits must be provided to the ACR.
diac CT imaging. Tolerances are provided for measured val-
ues, most of which, if exceeded, count as “minor” demerit
points for the submission. A unit may have up to six minors The IAC Accreditation Program
before failing the application. More severe issues can result
in a “major,” which is automatic failure. Some examples of Scanner Requirements
major issues are excessive dose, gross misuse of certain tech- The IAC has several exam-specific accreditation programs,
nical parameters, and excessive image artifacts. Sites may including coronary calcium scoring CT, coronary CTA, vas-
also fail if the submitted images were acquired using param- cular CTA, neurological CT, sinus and temporal bone CT,
eters that are substantially different from those listed in the and body CT. The general program includes basic guidelines
scan protocol. The detailed instructions for the physics tests and requirements for such items as image archive, timely
for accreditation submission are available from the ACR [3]. reporting, policies and procedures for patient safety, and a
The ACR also requires an ongoing CT quality assurance pathway to report incidents and adverse events, among oth-
program, which includes involvement from the technologist, ers. Details can be found in the IAC CT Standards [6]. The
radiologist, and medical physicist. The program requires roles of the personnel and supervisory staff, as well as facil-
acceptance testing at scanner installation and routine physics ity requirements, are also listed in this document.
testing done at least annually, completed by a QMP, and For coronary calcium scoring, a multi-detector system of
including reports and documentation. The technologist must four slices or greater must be used, with a recommended
perform daily measurements on the scanner prior to scanning rotation time of ≤0.5 s. The software and reconstruction sys-
patients. The measurements include recording of the CT tems must have the following capabilities: provide visual
number of water and image noise values, as well as a visual representation of coronary calcium exceeding protocol
inspection for artifacts. The technologist must also perform thresholds; quantify coronary calcification using Agatston
monthly visual checks of basic system operating and safety and mass and/or volume scoring methodologies; and provide
features. All measurements must be recorded and available user interaction with quantitative program to allow selecting
for review by the medical physicist or an ACR inspector. All or deselecting coronary calcification based on visual
required and suggested tests, including tolerances, are inspection.
described in the ACR CT QC Manual [5]. The ACR also has For coronary CTAs (used for coronary arteries and coro-
requirements for CT protocol maintenance, in which a team nary bypass grafts), the CT scanner must have ≥64 slices, a
consisting of at least a technologist, medical physicist, and rotation time of ≤0.5 s, and a dual auto-injector. The associ-
radiologist review protocol items such as appropriateness, ated software must meet the following requirements: (1) are
clarity of the instructions, and dose optimization opportuni- capable of displaying data as maximum intensity projection
ties. Although the ongoing quality assurance (QA) compo- (MIP), thick or thin slices, multi-planar reformats, and data
nent of the ACR accreditation program involves many in a curve plane reformat; (2) create 3D images and rotate
individuals, it is ultimately the responsibility of the radiolo- about all three axes; (3) provide measurements as described
gist to ensure that all aspects of the program are being met. in a facility-specific protocol; (4) simultaneously load mul-
tiple phases; and (5) perform quantification of coronary
Technologist Requirements calcium.
Technologists must meet all three of the following criteria General CT requirements are that all data be reviewed digi-
for ACR accreditation: (1) ARRT registered (RT) and radiog- tally on a monitor which is sufficient to prevent any loss of
raphy (R) and/or computed tomography (CT) certified and/ resolution and with standard window/level settings and which
or NMTCB registered (CNMT) and/or unrestricted state can also be adjusted manually. The datasets must be DICOM
license; (2) documented training and experience in CT; and compatible, and the systems must have the capability to opti-
(3) documented training and experience in operating CT mize the field of view based on patient size and protocol.
equipment, radiation physics, and protection. Passing an The IAC also requires a quality improvement (QI) pro-
advanced examination for CT certification is recommended. gram, which includes quality control testing, acceptance
For registered technologists, the continuing education testing after installation or major repair or upgrade, and room
requirements of the certifying organization, including credits shielding verification. QI program oversight should be done
pertinent to CT, must be met. Maintaining good standing sta- by a QI committee which consists of the technical director,
tus within the certifying organization is sufficient proof of medical director, service engineer, and/or site-appointed
meeting the continuing education credits. For state-licensed medical physicist. The facility must have at least one QI
and other technologists, 24 h of continuing education, rele- meeting per year to review the results of the QI program.
5  Cardiac CT: Credentialing and Accreditation 43

Acceptance testing includes measurements of image degree in another medical imaging profession (with focus on
quality, system performance characterization, and radiation radiation physics).
dose. The system parameters must be compared to the manu- The technologists must have at least 15 h of Category 1
facturer’s specifications. Daily and periodic QC tests are to AMA or RCEEN-approved CT-related continuing education
be performed as suggested by the manufacturer. The daily credits over a period of 3 years. Three of the 15 h must be
tests, which would be completed by a technologist, include related to radiation safety.
the measurement of the CT number of water and another ref-
erence material, measurement of image noise, and visual
artifact assessment. Additionally, the proper functioning of The Joint Commission Accreditation Program
the audible and visual patient safety equipment must be veri-
fied. Annual performance checks are also required to be Scanner Requirements
completed by a qualified medical physicist. The tests include The Joint Commission diagnostic imaging accreditation pro-
basic image quality and system checks and are listed in the gram includes CT but with no specific requirements for car-
standards manual [6]. A QI committee and/or the medical diac imaging. TJC has requirements for annual testing of the
director must evaluate the physicist’s recommendations for CT scanners, including measuring standard image quality
the type of testing, frequency, and designated personnel to metrics and radiation dose, as well as display device moni-
perform the test(s). Preventive maintenance service is also toring tests, although measurement methods and tolerances
required at least annually for each scanner at the site. All QC are not provided. A complete list of the required tests can be
testing, including preventative maintenance records, must be found in the Joint Commission Diagnostic Imaging
documented and maintained at the facility. Requirements [7]. The requirements also include shielding
As part of the QI program, facilities must have a process to design for new installations or remodels, including verifica-
assess the exam appropriateness and to evaluate aspects of the tion afterward.
technical quality, including review of clinical images for sub- TJC accreditation program mandates that patient age and
optimal image quality or artifacts, completeness of the study, recent imaging be considered when deciding the most appro-
adherence to protocol, and any patient or facility safety con- priate type of imaging. The CT protocols must be based on
cerns. The interpretive quality should be evaluated, such as current standards of practice, which includes clinical indica-
through peer review, including the final report completeness tion, contrast administration, etc. The protocols must be
and timeliness. Radiation safety items must also be recorded, reviewed and kept current with input from an interpreting
including documentation of the expected dose ranges based radiologist, medical physicist, and lead technologist.
on patient-specific characteristics, comparison of doses for Patient radiation doses must be documented in the patient
each protocol, and tracking of repeated CT exams. Proper record, in a retrievable format (if the scanner is capable of
usage of dose reduction must be documented, and staff occu- producing dose metrics). Some tracking or recording of
pational exposures must be monitored. patient doses is also necessary to meet the requirement that
The IAC accreditation program also has criteria and any incidents where the dose exceeds a pre-defined dose
guidelines for minimum procedure volumes, indication veri- range must be reviewed and analyzed. These doses then need
fication, and ensuring that the CT test is appropriately to be compared to external benchmarks.
ordered and scheduled. A list of elements for examination
performance is also included, as well as parameters that Technologist Requirements
should be included in the scan protocol. The site must have documentation that the technologists who
perform CT participate in ongoing education in radiation
Technologist Requirements dose optimization techniques and tools for pediatric and
All technologists performing CT scans must meet at least adult patients, as well as training in safe scanner operation.
one of the following criteria: (1) the American Registry of Specific hours of required training are not provided.
Radiologic Technologists (ARRT) or the Canadian
Association of Medical Radiation Technologists (CAMRT)
certification in CT imaging (i.e., ARRT(R) ARRT(CT)), (2) The RadSite Accreditation Program
an appropriate nationally recognized credential in another
medical imaging field (i.e., CNMT, ARRT(R), ARRT(R) Scanner Requirements
(MR)), or (3) a 12-month full-time clinical CT experience The requirements for accreditation are detailed in the
under direct supervision of a credentialed technologist plus RadSite Accreditation Standards [8]. Clinical patient
one of the following – completion of a formal 2-year p­ rogram images must be submitted for each CT unit. A selection of
(or equivalent) in another medical imaging profession (with choices is listed for each type of exam, including neuroim-
focus on radiation physics) or completion of a bachelor’s aging, musculoskeletal, and body [9]. A minimum of three
44 J. M. Kofler et al.

adult cases and protocols, as well as one pediatric case, if US hospitals employ credentialing committees to
applicable, must be included in the submission. For scan- review physician qualifications. It is the responsibility of
ners that perform cardiac imaging, at least one cardiac the committee to ensure individual physicians have the
exam must be submitted, as listed in the requirements for medical knowledge, technical skills, and clinical judgment
body scanners. Clinical exam submissions must also to perform their assigned duties. Once a physician has
include the clinical indication and any post-processed been credentialed at an individual hospital, he or she must
images, as well as the clinical report. A list of scan param- apply to be recredentialed every 3 years to maintain hospi-
eters for various exams, as well as pass/fail examples, is tal privileges [10].
included in the Standards document. Some examples for The credentialing committee is also responsible for deter-
failing the submission include incomplete coverage, exces- mining hospital privileges for each practitioner.
sive dose, or inappropriate scan parameters. The Standards Determination of criteria for individual staff privileges is at
document also includes samples of the image quality scor- the discretion of the committee. Examples of criteria for the
ing sheet, which contains all of the elements that are part of practice of cardiac CT could include specifying residency
the review process. An on-site audit checklist is also pro- and/or fellowship subspecialty training, minimum numbers
vided, such that sites can be prepared for an on-site of cases read and reported, or completion of a certifying
inspection. examination.
The RadSite accreditation program requires annual QA
testing that includes basic image quality and dose metrics
and is completed by a qualified medical physicist. There are
several phantom options available for testing, and tolerances
Recommendations for Physician Training
in Cardiac CT
are listed for each for the various required tests. Phantom
images must be supplied for adult and pediatric head and
Several different organizations have published training
abdomen exams that have been acquired using the site’s clin-
guidelines and recommendations for certification of special-
ical protocol, and the dose for each of these must be within
ists who are to interpret cardiac CT studies. Although none
the listed limits. The physicist’s report for each unit applying
of these recommendations has been universally adopted,
for accreditation must be included in the submission materi-
each represents an effort to improve patient safety through
als. Preventative maintenance, including documentation, is
standardization of training for physicians. Several of these
also required.
consensus documents are detailed below.
Technologist Requirements
For CT accreditation, the technologists must meet the fol-
lowing requirements: have registration or certification from  onjoint Committee of Australian and New
C
the American Registry of Radiologic Technologists Zealand Practitioners: Training Requirements
(ARRT); trained and/or hold current, unrestricted for CTCA Specialists
registration(s) or be licensed in each of the modalities per-
formed; have an associate’s or bachelor’s degree in radio- In 2009, a conjoint committee comprised of members of
logic science whenever required, with requisite job the Australasian Association of Nuclear Medicine
experience; and complete ­ongoing education as required by Specialists (AANMS), the Cardiac Society of Australia
the license, certification, and medical director’s directives. and New Zealand (CSANZ), and the Royal Australian and
Validated credentials by the technologist’s internal creden- New Zealand College of Radiologists (RANZCR) pub-
tialing program are sufficient evidence of meeting these lished a summary statement detailing their recommenda-
requirements. tions for training in coronary CT angiography. This
document has been revised several times, most recently in
2014 [11].
 redentialing of Physicians Interpreting
C The conjoint committee describes two distinct levels of
Cardiac CT specialist training. Level A training allows the practicing
physician to interpret coronary CT angiography indepen-
Unlike accreditation, which focuses on the technical aspects of dently and without supervision. Level B training allows the
cardiac CT and the equipment used to perform it, credentialing practitioner to operate independently and to supervise the
refers to the process used to ensure that the practitioners who training and recertification of other physicians. The training
interpret these studies are competent to do so. In the United consists of didactic work as well as a mix of live and “library”
States, credentialing is the responsibility of the hospital or imag- cases, each of which needs to have been supervised by a
ing center where these studies are performed or interpreted. Level B specialist.
5  Cardiac CT: Credentialing and Accreditation 45

Level A Training ered to be educated consumers but are not expected to be

A. 40  hours of didactic coursework, including 20  h of
able to interpret CT scans independently. This training con-
“hands-on” training sists of approximately 1 month of intensive exposure to car-
B. 150 coronary CTA cases, including: diac CT and should include mentored interpretation of at
(a) 50 live cases least 50 scans with correlation to other cardiac imaging. For
(b) 50 cases with correlation of findings to other
at least 15 of these studies, it is expected that the trainee will
modalities be present during the performance of the examination.
(c) 25 cases with non-coronary cardiac findings
(d) 25 cases with non-cardiac findings Level 2 Training
Level 2 training is the minimum level of training needed for
Level B Training independent interpretation of cardiac CT studies. Trainees
A. Level A training requirements should complete at least 1 additional month of training in
B. 150 coronary CTA cases beyond those provided for Level addition to the requirements for Level 1 and interpret at least
A (total of 300 cases), including: 200 additional cardiac CT studies, of which 65 require the
(a) 50 live cases (total of 100 live cases) trainee be present during performance of the examination.
(b) 30 cases with correlation of findings (total of 80 Additionally, the trainee is expected to review all CT cases
­correlated cases) for non-cardiac findings.

Upon completion of their training, physicians can apply for Level 3 Training
recognition of training, which functions as certification by the Level 3 training allows the practitioner to supervise an inde-
conjoint committee. Recertification is available for practitio- pendent cardiac CT facility or direct an academic cardiac CT
ners who document a minimum of 300 cases (for Level A) or practice. The ACC recommendations specify that this level
600 cases (for Level B) within a 3-year recertification period. of training requires additional time beyond the standard
3-year cardiovascular fellowship and is expected to include
training in at least one other cardiac imaging modality.
 merican College of Cardiology
A Additionally, Level 3 training should involve participation in
Recommendations for Training in research, teaching, and the administrative aspects of running
Cardiovascular CT: COCATS 4 Task Force 7 a cardiac CT service.

In 2015, the American College of Cardiology (ACC) pub-

lished its recommendations for fellows-in-training to become Certification in Cardiac CT
proficient in cardiac CT. This document was endorsed by the
American Society of Nuclear Cardiology (ASNC), the Although credentialing standards for individual physicians
Society for Cardiovascular Angiography and Interventions are determined by the hospitals where they work and do not
(SCAI), the Society of Atherosclerosis Imaging and necessarily include completion of a certification examina-
Prevention (SAIP), and the Society of Cardiovascular tion, two different organizations in the United States provide
Computed Tomography (SCCT). The recommendations are certification for specialists in cardiac CT.  Each of these
designed to meet the competency-based parameters endorsed ­certificates is designed to allow qualified practitioners from
by the American College of Graduate Medical Education varying medical specialties to demonstrate proficiency in the
(ACGME) and are intended to address the six ACGME core performance and interpretation of cardiac CT.
competencies for medical education [12].
As with the Australia and New Zealand guidelines, these  BCCT Certification in Cardiac CT
C
recommendations delineate distinct levels of proficiency that In 2008, the Certification Board of Cardiovascular Computed
can be attained with different amounts of training. Unlike the Tomography (CBCCT) began offering a certification exami-
conjoint committee’s document, these recommendations only nation in cardiac CT. The examination is a computer-based
cover fellows during their initial training. Recommendations test offered every 2 years and consists of up to 175 multiple
for established practitioners are not included. Details of these choice questions on a variety of topics relating to cardiac CT
training recommendations can be found in Table 5.1. (http://www.cccvi.org/cbcct) [13].
Qualified physicians who pass the examination are
Level 1 Training granted certification, which is valid for a period of 8 years.
Level 1 training is the minimum basic level of training Recertification is accomplished through maintenance of a
needed to gain basic familiarity with the principles of cardiac valid medical license and appropriate medical board certifi-
CT. Practitioners who have Level 1 training can be consid- cation, supplemented by continued medical education and
46 J. M. Kofler et al.

Table 5.1  Core competency components and curricular milestones for training in cardiovascular computed tomography
Milestones (months)
Competency component 12 24 36 Add
Medical knowledge
1 Know the principles of cardiovascular computed tomographic scanning and the scanning modes I
2 Know the risks and safety measures for cardiovascular computed tomographic scanning, I
including radiation reduction strategies
3 Know the appropriate indications for cardiovascular computed tomography for screening or I
evaluating symptoms in patients with suspected cardiac disease
4 Know the indications, potential adverse effects, prevention, and treatment of complications of I
iodinated contrast agent use in cardiovascular computed tomographic studies
5 Know the indications and protocols for beta-adrenergic blocking drugs and nitroglycerin during II
cardiovascular computed tomographic studies
6 Know the principles of cardiovascular computed tomographic scan collimation, temporal II
resolution, table speed, field of view, and window and level view settings
7 Know the principles of post-processing methods for cardiovascular computed tomographic II
scanning
8 Know the algorithms used for reconstruction, and recognize and isolate causes of artifacts II
9 Know the principles of quantitative coronary artery calcium scoring II
10 Know normal chest anatomy and common incidental extra cardiac findings II
11 Know the characteristic cardiovascular computed tomographic images of normal cardiac I
chambers and great vessels, normal coronary arteries and veins, and normal variants
12 Know the characteristic cardiovascular computed tomographic findings of coronary II
atherosclerosis including plaque morphology and assessment of stenosis severity
13 Know the characteristic cardiovascular computed tomographic findings of anomalous coronary II
arteries and other common congenital anomalies
14 Know the characteristic cardiovascular computed tomographic findings in postoperative cardiac II
surgical patients including internal mammary artery and saphenous vein bypass grafts
15 Know the characteristic cardiovascular computed tomographic findings of acquired and II
congenital valvular disease
16 Know the characteristic cardiovascular computed tomographic findings of left atrial and II
pulmonary and coronary venous abnormalities
17 Know the characteristic cardiovascular computed tomographic findings of pericardial disease II
18 Know the characteristic cardiovascular computed tomographic findings of cardiomyopathies and II
infiltrative myocardial diseases
19 Know the differential diagnosis of cardiac masses identified by cardiovascular computed II
tomography
20 Know the characteristic cardiovascular computed tomographic findings of common diseases of II
the aorta and great vessels
21 Know the characteristic cardiovascular computed tomographic findings of pulmonary embolism II
and primary and acquired pulmonary vascular diseases
22 Know when to request help with the interpretation of difficult studies, such as patients with I
complex congenital heart disease
Evaluation tools: conference presentation, direct observation, and in-training examination
Patient care and procedural skills
1 Skill to appropriately utilize cardiovascular computed tomography in the evaluation and I
management of patients with known or suspected cardiovascular disease
2 Skill to integrate cardiovascular computed tomographic findings with other clinical information in I
patient evaluation and management
3 Skill to recognize and treat contrast-related adverse reactions I
4 Skill to independently perform and interpret cardiovascular computed tomography II
5 Skill to perform and interpret hybrid CT/SPECT and CT/PET imaging III
Evaluation tools: conference presentation, direct observation, and logbook
Systems-based practice
1 Incorporate appropriate use criteria, risk/benefit, and cost considerations in the use of I
cardiovascular computed tomography and alternative imaging modalities
Evaluation tools: conference presentation, direct observation, and multisource evaluation
Practice-based learning and improvement
1 Identify knowledge and performance gaps and engage in opportunities to achieve focused I
education and performance improvement
5  Cardiac CT: Credentialing and Accreditation 47

Table 5.1 (continued)

Milestones (months)
Competency component 12 24 36 Add
2 Utilize point-of-care educational resources (e.g., guidelines, appropriate use criteria, and clinical I
trial results)
Evaluation tools: conference presentation, direct observation, and reflection and self-assessment
Professionalism
1 Work effectively in an interdisciplinary cardiovascular computed tomographic imaging I
environment
2 Reliably obtain patient-informed consent, ensuring that patients understand the risks and benefits I
of, and alternatives to, cardiovascular computed tomographic testing
3 Know and promote adherence to clinical practice guidelines I
Evaluation tools: conference presentation, direct observation, and multisource evaluation
Interpersonal and communication skills
1 Communicate testing results to physicians and patients in an effective and timely manner I
Evaluation tools: direct observation and multisource evaluation
From Garcia et al. [12], with permission
Add additional months beyond the 3-year cardiovascular fellowship, CT computed tomography, PET  positron emission tomography, SPECT
­single-photon emission computed tomography

ongoing interpretation of a minimum of 150 cardiac CT References

studies over a 36-month period. Details regarding eligibility
requirements for initial certification are available on the 1. https://www.cms.gov/Medicare/Provider-Enrollment-and-
CVCCT website [cccvi.org]. Certification/SurveyCertificationGenInfo/Accreditation-of-
Advanced-Diagnostic-Imaging-Suppliers.html. Accessed 27 Oct
2016.
 CR Certificate of Advanced Proficiency
A 2. CT Accreditation Program Requirements. http://www.acrac-
in Cardiac CT creditation.org/~/media/ACRAccreditation/Documents/CT/
In 2008, the American College of Radiology developed a Requirements.pdf?la=en. Accessed 27 Oct 2016.
3. American College of Radiology CT Accreditation Program
certification examination for practitioners to demonstrate Testing Instructions. http://www.acraccreditation.org/~/media/
advanced competency in cardiac CT. Initially, this examina- ACRAccreditation/Documents/CT/CT-Accreditation-Testing-
tion included a live workstation case review. The current Instructions.pdf?la=en. Accessed 27 Oct 2016.
examination is computer-based and consists of 100 multiple-­ 4. Personnel Requirements for Medical Physicists and MR Scientists.
http://www.acraccreditation.org/Revised-Program-Requirements.
choice questions administered over a 2-h period at more than Accessed 27 Oct 2016.
850 testing centers around the world. 5. 2012 Computed Tomography Quality Control Manual, available
The ACR Certificate of Advanced Proficiency (CoAP) from the American College of Radiology (http://www.acr.org/
examination covers a wide variety of topics related to Education/Education-Catalog/Products/8336734).
6. The IAC Standards and Guidelines for CT Accreditation. http://
cardiovascular CT (http://www.acr.org/Education/Exams-
­ www.intersocietal.org/ct/standards/IACCTStandards2015.pdf.
Certifications/CoAP). Successful completion of the examina- Accessed 27 Oct 2016.
tion conveys a certificate which is valid for a period of 7. Diagnostic Imaging Requirements. https://www.jointcommis-
10 years. Eligibility requirements to sit for the examination sion.org/assets/1/18/AHC_DiagImagingRpt_MK_20150806.pdf.
Accessed 27 Oct 2016.
and a list of designated testing centers are available on the 8. RadSite MAP Accreditation Standards, version 2 (MAP v 2.2),
ACR website [acr.org] [14]. available on request from http://www.radsitequality.com/cms-
accreditation-requirments. Accessed 27 Oct 2016.
9. http://www.radsitequality.com/cms-accreditation-requirments with
permission.
Conclusion 10. National Association Medical Staff Services. www.namss.org/por-
tals/0/education. Accessed 15 Jan 2017.
Although accreditation of equipment and credentialing of 11. Training Requirements for CTCA Specialists, Version 4 (Rep.).
practitioners are not the first topics that come to mind when 2014 Conjoint Committee for Recognition of Training in Computed
Tomography Coronary Angiography.
discussing cardiac CT, each is important for establishing and 12. Garcia MJ, Blankstein R, Budoff MJ, Dent JM, Drachman DE, Lesser
maintaining a successful cardiovascular imaging program. JR, et al. COCATS 4 task force 7: training in cardiovascular com-
As cardiovascular CT continues to develop as an imaging puted tomographic imaging. J Am Coll Cardiol. 2015;65(17):1810.
subspecialty, maintenance of certification and employment 13. Latest News-Certification Board of Cardiovascular Computed

Tomography. http://www.cccvi.org/cbcct. Accessed 28 Oct 2016.
of appropriate technology will become increasingly impor- 14. Cardiac CT Certificate of Advanced Proficiency (CoAP Exam). http://
tant as means of demonstrating expertise in this rapidly www.acr.org/Education/Exams-Certifications/CoAP.  Accessed 28
changing field. Oct 2016.
 Part III
Where We Are: Technical and Operational Requisites
 Cardiac CT Platforms: State of the Art
6
Bernhard Schmidt, Katharine Grant, Thomas G. Flohr,
and Thomas Allmendinger

Shorter scan times and the desire for higher resolution were data and thus due to acquisitions with reduced pitch result-
the driving forces behind the development of spiral scanning ing in substantially increased patient dose values [15].
in 1990 and the first multi-detector row computed tomogra- Irregular or arrhythmic heart rates posed a challenge for
phy (CT) systems in 1998 [1–6]. With the introduction of artifact-free combination of images that were reconstructed
ECG-gated scanning on four-slice CT scanners in 1999, the from multiple cardiac cycles into a single volume, often
first step toward cardiac imaging with multi-detector row yielding “stack artifacts” in these cases. Moreover, calcium
computed tomography (MDCT) had been made. blooming was still an issue – mainly due to operation in an
Available since 2004, 64-slice CT systems are currently X-ray tube limited, and therefore noise-dominated regime,
considered prerequisite for routine cardiac imaging in resulting in the routine application of medium-sharp recon-
clinical scenarios [7–10], and with its introduction, the structions in cardiac images exhibiting the undesired bloom-
task of imaging coronary arteries and calcium scoring ing of high-­ contrast calcium. Adaption of X-ray tube
replaced EBCT (electron beam CT) completely. 64-slice voltages (kVs) to lower kVs to reduce patient dose was used
CT scanners allowed for comprehensive diagnosis of mor- in particular for regular contrast-enhanced scans, to improve
phology and cardiac function within one integrated CT image quality in pediatric patients. These adaptations
examination, including high-resolution imaging of the occurred manually, often on weight-based selection criteria,
coronary arteries [11–14]. Improved temporal resolution and application was limited to a select group of the general
(< 200  ms) was enabled by faster gantry rotation times population due to the demand of high X-ray tube current
down to 0.33 s and leading to an increased clinical robust- capacities required for cardiac imaging [16]. Finally, scan
ness of ECG-gated scanning techniques at higher heart times for coronary CTAs were still in the range of 10–15 s
rates [9, 10]. ECG-gated 64-slice CT also started being given the limited detector coverage and the lack of alterna-
used for rapid triage of patients with acute chest pain in the tive scan technology.
emergency room and for diagnosis of pulmonary embo- In 2004, all major CT vendors offered 64-slice CT scan-
lism, aortic dissection or aneurysm, or significant coronary ners. The remaining challenges of cardiac CT imaging, the
artery disease in one scan. pursuit for even more clinical applications, the risk that MR
Despite the new clinical opportunities and the increasing imaging could further improve on and replace CT for not
use of CT for routine cardiac imaging, several challenges just functional but for morphological imaging, the tena-
still remained: For high heart rates, multi-segment recon- cious competition of the vendors, and the relentless
structions allowing for temporal resolutions beyond half the improvements in engineering and scientific innovations,
rotation time were still desired for robust imaging of coro- made cardiac imaging the main driving force for further CT
nary arteries. Multi-segment reconstructions unfortunately innovations, technical developments, and new CT plat-
do not guarantee images with a well-defined improved tem- forms. In contrary to previous innovations, technical solu-
poral resolution in combination with the need for redundant tions and directions of development differed substantially
between different vendors, ranging from wide detector sys-
tems with single-source systems over two X-ray source
B. Schmidt (*) · T. Allmendinger · T. G. Flohr
systems with a single detector to systems with two X-ray
Department of Computed Tomography, Siemens Healthcare
GmbH, Forchheim, Germany tubes and two detectors.
e-mail: [email protected] In 2005, a dual-source CT (DSCT) system, i.e., a CT sys-
K. Grant tem with two X-ray tubes and two corresponding detectors
Siemens Medical Solutions USA, Inc., Malvern, PA, USA offset by 90°, was introduced [17]. It provided improved

© Humana Press 2019 51

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_6
52 B. Schmidt et al.

temporal resolution of 83  ms independent of the patient’s  rom EBCT to Multi-detector CT for Cardiac
F
heart rate as compared to 165–190 ms with MDCT systems Imaging
at that time. DSCT scanners proved to be well suited for inte-
grated cardiothoracic examinations even in acutely ill EBCT
patients and for the triage of patients with acute chest pain
[18]. The introduction of dual-energy scanning with DSCT Generation of X-rays for medical imaging is typically
enabled tissue characterization and provided combined func- accomplished by using X-ray tubes. Electrons emitted from
tional and morphological information, i.e., to depict local a cathode are accelerated by a high-voltage power source and
perfusion deficits in the lung parenchyma in patients with collide with the anode, at which X-rays are generated by the
pulmonary embolism [19, 20]. interaction of electrons within the anode material. For CT
The second and third generations (introduced in 2009 and imaging, X-ray attenuation data from multiple angles around
2013, respectively) of DSCT systems offer high-pitch scan the patient are required. Thus, first CT systems were equipped
modes which enabled coverage of the heart in a single car- with X-ray tubes set up opposite a detector, which in con-
diac cycle [21, 22]. In addition, high-resolution CT scans of junction rotated around the patients at rotation times of about
the entire thorax in less than 1 s scan time with an acquisition 5  s or longer. Mechanical constraints and challenges with
time per image better than 100 ms were now possible [23, rotating components, mainly data and power transmission,
24]. These scan modes are potentially advantageous for eval- prevented continuous rotation at faster rotation times back
uating the lung parenchyma and vascular structures in then. Unfortunately, for cardiac imaging in particular, faster
patients who have difficulty complying with breath-holding rotation times were desired to minimize motion artifacts and
instructions [25]. High-pitch scan modes have also been used provide better temporal resolution. Already in the late 1970s,
for fast CTA scans of the aorta [26]. Combined with ECG electron beam scanners were proposed with the goal of pro-
triggering, they provide adequate visualization of the coro- viding a substantially higher temporal resolution for cardiac
nary arteries, the aortic valve, the aorta, and the iliac arteries imaging by avoiding rotating parts. In the early 1980s,
in one scan at low radiation dose, which is beneficial in the D.P. Boyd – a Stanford researcher and partner in the radiol-
planning of transcatheter aortic valve replacement (TAVR) ogy department at the University of California, San Francisco
procedures [27, 28]. (USA)  – invented the first clinically used EBCT system:
Another challenge for CT is the visualization of dynamic “Electron beam control assembly and method for a scanning
processes in extended anatomical ranges, i.e., to character- electron beam computed tomography scanner.” It related to
ize the inflow and outflow of contrast agent in the arterial “… a high speed multiple section computed-tomographic
and venous system in dynamic CT angiographies or to (CT) x-ray transmission scanner and more particularly to a
determine the enhancement characteristics of the contrast multiple target scanning-electron beam x-ray source provid-
agent in volume perfusion studies. One way to address this ing rapid scan …” (see also Figs. 6.1 and 6.2) [33]. Instead
problem is by utilizing wide-area detectors large enough to of rotating tube and detector around the patient, a stationary
cover organs such as the heart, the kidneys, or the brain in detector ring is used for data acquisition. To generate X-rays,
one axial scan. Besides temporal resolution, registration or an electron beam originating from an electron gun is deflected
mismatch artifacts from the combination of multiple cardiac and steered over an anode that was ringlike enclosing the
cycles or slabs in z-direction are another challenge, natu- patient. Temporal resolutions of 100  ms and shorter were
rally leading to the idea of increasing the detector coverage possible. The main application  – in particular in the early
to such a level that the entire heart is covered in a single days of EBCT – was the detection and quantification of coro-
prospectively gated sequential acquisition. Meanwhile, two nary calcium at dose levels of about 0.6 mSv [34]. Later on,
vendors have introduced CT scanners with 16 cm detector studies reported the use of EBCT for coronary CTA imaging
coverage at isocenter, providing 320 × 0.5 mm collimation [35], assessment of ventricular anatomy and function [36], as
at 0.28  s rotation time or 256  ×  0.625  mm collimation at well as myocardial perfusion [37].
0.28 s rotation time [29]. Additionally these scanners have Despite the unquestionable benefit of high temporal reso-
the potential to acquire dynamic volume data by repeatedly lution of EBCT, only a few systems were installed world-
scanning the same anatomical range without table move- wide, by the end of 1987 about 14 units. The high price, low
ment (e.g., [30–32]). production volume, system design issues (e.g., the inability
In the following sections, after a short historic review of to use scatter collimators), and limited clinical benefit beside
EBCT, general technical demands of MDCT for cardiac the heart were preventing cardiologists and radiologists to
image quality are discussed. Then a technical overview of invest into an EBCT system. Finally, the limited spatial reso-
the main currently available platforms and CT systems is lution was yet another factor making the already niche scan-
provided, and details about advantages and technical chal- ner even less attractive. Although with the latest generation
lenges of the different approaches are discussed. of EBCTs (GE’s e-Speed), slice thicknesses of 1.5 mm were
6  Cardiac CT Platforms: State of the Art 53

14

46

13
24 27
23 11
12

22
62
37 38 39 40

21

23 30
28 29 31
26

32 35
33 34

Fig. 6.1  Illustration of an EBCT (electron beam computerized tomogra- through the patient to the detector [14] on the opposite end of the scan
phy) system – taken from Boyd’s patent in 1982; US 4352021: Electrons tube tomography system. (From US patent US4352021 https://www.
emitted from the electron gun [22] are accelerated and steered around the google.com/patents/US4352021?dq=us+patent+4352021&hl=en&sa=X
patient. The electron beam hitting the target rings [28–31] created X-rays &ved=0ahUKEwitp6_N38vYAhXDY98KHT01DXwQ6AEIJzAA)
at the respective position on the ring. From there the X-rays passed

a b

Fig. 6.2  Axial image of a contrast-enhanced EBCT scan, showing the left main and left anterior descending coronary artery (arrow) (a). Surface-­
rendered view of the heart by visualizing only voxels above a certain threshold (b) (From Achenbach et al. [35], with permission)
54 B. Schmidt et al.

enabled [38], this relevant specification parameter for coro- tions depending on the rotation time, during a full gantry
nary CT imaging was already matched by 16-slice CT sys- rotation. All measurement values acquired at the same angu-
tems at the end of 1990s and then outperformed substantially lar position of the measurement system are called a “projec-
by the introduction of 64-slice MDCT systems providing tion” or “view.” Slip ring designs which pass the electrical
routinely submillimeter image resolution in the early 2000s. signals across sliding contacts allow for continuous rotation
In addition, volume scanning, the ability to apply modern of the measurement system.
dose reduction techniques such as low kV scanning, utiliza- State-of-the-art X-ray tubes are powered by onboard gen-
tion of shaped filters, or patient-specific tube current modu- erators and provide peak powers of 60–120 kW at different
lation never became available or were technically impossible user-selectable voltages ranging from 70 to 150 kV. Scanning
to implement on an EBCT system. at low tube voltage is favorable for dose-efficient scans, such
as pediatric CT [39, 40] or CT angiographic scanning [16,
41–43], because the X-ray attenuation of iodine significantly
Multi-detector Row CT Setup (MDCT) increases at lower kV [40].
All modern MDCT systems use solid-state scintillation
The basic CT system design of modern MDCT system is detectors. The incident X-rays interact with a radiation-­
shown in Fig. 6.3. Today, all manufacturers use the same fan-­ sensitive crystal or ceramic (such as gadolinium oxide, gado-
beam CT design, characterized by an X-ray tube and an linium oxysulfide, or garnets) with suitable doping. They are
opposing detector which are mounted on a rotating gantry absorbed, and their energy is converted into visible light
ring. The detector is a two-dimensional array, consisting of which is detected by a silicon photodiode attached to the
2–320 element rows aligned in the z-axis direction (the z-axis backside of the scintillation detector. The resulting electrical
is the length axis of the patient) and around 650–1000 detec- current is then converted into a digital signal. Key require-
tor elements in each row. The fan angle of the detector is ments for a detector material are good detection efficiency,
wide enough (approximately 45–55°) to cover a whole-body i.e., high atomic number, very good signal linearity, and very
scan field of view (SFOV) of typically 50 cm in diameter. In short afterglow time to enable fast readout at the high gantry
a CT scan, the detector array measures the X-ray attenuation rotation speeds that are essential for cardiothoracic CT.
profile of the patient at 1000–5000 different angular posi- The image noise in a CT image is caused by the quantum
noise of the X-ray photons and the electronic noise of the
detection system. In high-dose scanning situations, the image
noise is dominated by quantum noise, whereas when larger
x-ray tube
patients are scanned or in examinations at low radiation
dose, e.g., in low-dose thorax scans, electronic noise is more
dominant. In addition, electronic noise degrades image qual-
ity and the stability of CT values. Recently, detector systems
with integrated electronics were commercially introduced
Collimator (e.g., Stellar, Siemens Healthcare, Forchheim, Germany;
NanoPanel Elite, Philips Healthcare, Amsterdam,
High voltage generator
Netherlands), with the goal of reducing electronic noise and
Control
detector cross talk. In these designs, photodiodes and analog-­
systems to-­digital converters (ADC) are combined and directly
attached to the ceramic scintillators, without the need for
noise-sensitive analog connection cables (Fig.  6.4). In a
recent study, for a 30 cm phantom corresponding to an aver-
age abdomen, reduction of image noise by up to 40% was
demonstrated with the use of an integrated electronic detec-
Detector with electornics
tor at 80 kV [44]. According to the authors, this noise reduc-
tion translated into a dose reduction of up to 50% while
Patient table achieving equivalent image noise.
In early MDCT detectors, ADC electronics and data
transmission were a limiting factor and a strong contributor
Fig. 6.3  Main components of a modern MDCT system. The X-ray fan to the total cost of a system. An efficient way of enabling
beam is indicated in red; it covers a SFOV of typically 50 cm in diam-
eter. The data measurement system consists of detector and detector
higher scan speeds based on a larger z-axis collimation was
electronics. The patient is positioned within the gantry using a dedi- through the development of detector elements with a larger
cated patient table number of detector rows. The total beam width in the z-­
6  Cardiac CT Platforms: State of the Art 55

Fig. 6.4  Comparison of the Si-tile without scintillator ceramics – with (80 kV, 30 mA) are illustrated. The integrated design leads to substan-
and without integrated electronics (top) (Siemens Healthcare). At the tially lower noise values (right bottom image)
bottom CT images of a 30 cm water phantom acquired at same dose

direction is subsequently adjusted by pre-patient collimation total, 24 rows with a z-width of 19.2  mm at isocenter
at the expense of fusing two (or more) detectors along the (Fig. 6.5). By adjusting the X-ray beam width such that only
z-axis electronically into thicker slices. the central detector rows are illuminated, the system pro-
The detector of a 16-slice CT (Siemens SOMATOM vides 16 collimated 0.6 mm slices. By ­illuminating the entire
Emotion 16) as an example comprises 16 central rows, each detector, reading out all rows and electronically combining
with 0.6  mm collimated slice width, and 4 outer rows on the signals of every 2 central rows, the system provides 16
either side, each with 1.2  mm collimated slice width  – in collimated 1.2  mm slices. The 16-slice detectors of other
56 B. Schmidt et al.

anode
focal spot position 1
2

reading 2
reading 1

through-plane direction
(z-direction)
shift by half a collimated
slice-width at iso-center

detector

Fig. 6.6  Basic principle of a z-flying focal spot. Consecutive readings

are shifted periodically by half a slice width on the anode plate. Every
two readings are interleaved to one projection with double the number
of slices and half the z-sampling distance

Ingenuity DAS, Philips Healthcare, Amsterdam,

16 x 1.2 mm
Netherlands). The focal spot in the X-ray tube is periodically
moved between two z-positions on the anode plate by elec-
16 x 0.6 mm tromagnetic deflection. As a consequence, the measurement
rays of two readings are shifted by half a collimated slice
width at isocenter and can be interleaved to one projection
with double the number of slices but half the z-sampling dis-
4x1.2, 16x0.6, 4x1.2 mm tance (Fig.  6.6). Two 64-slice readings with 0.6  mm slice
width and 0.6 mm z-sampling distance, as an example, are
Fig. 6.5  Example of a 16-slice detector, which consists of 24 detector
combined to one projection with 128 overlapping 0.6  mm
rows and provides either 16 collimated 0.6 mm slices or – by combina-
tion of the signals of every 2 central rows – 16 collimated 1.2 mm slices slices at 0.3 mm z-sampling distance. The z-flying focal spot
provides improved data sampling in the z-direction for better
through-plane resolution, reconstruction of thinner slices,
manufacturers are similarly designed, with slightly different and reduced spiral windmill artifacts (Fig. 6.7).
collimated slice widths (0.5  mm, 0.6  mm, or 0.625  mm,
depending on the manufacturer).
With the introduction of MDCT detectors with 64 or more ECG-Based Imaging with MDCT
detector rows and a vendor-dependent minimal slice width of
0.5  mm, 0.6  mm, or 0.625  mm, the detector row fusing To acquire images of the heart, recording of the ECG signal
mechanisms described above was no longer necessary. As of is mandatory. On one hand one wants to be able to process
today a wide variety of detector row configurations exist and view a dedicated phase of the cardiac cycle which
(e.g., 96  ×  0.6  mm, z-width 6  cm at the isocenter; requires a time stamp-based link between the cardiac phase
128  ×  0.625  mm, z-width 8  cm at the isocenter) up to the of the patient and the acquired CT data. On the other hand, in
widest commercially available CT detectors covering 16 cm the case of limited detector coverage, data from multiple
at isocenter. The acquisition is either based on 320 colli- heart beats is acquired over time and then stitched together to
mated 0.5  mm slices (Aquilion ONE, Toshiba Medical, allow the reconstruction and display of the whole heart. For
Japan) or 256 collimated 0.625 mm slices (Revolution, GE MDCT with 64 or more detector rows, typically two differ-
Healthcare, USA). ent acquisition modes are used  – ECG-gated spiral/helical
All modern MDCT scanners enable reconstruction of and prospective ECG-triggered sequence acquisition. More
images with different image slice widths from the same raw details about data acquisition and reconstruction are pro-
data according to the clinical needs (e.g., 3 mm or 5 mm for vided in Chaps. 6 and 11. However, to illuminate recent tech-
initial viewing and additional submillimeter slices or 1 mm nical developments in CT technology, a short summary of
slices for post processing). the acquisition techniques and influencing parameters on
Some CT systems employ a hardware-based oversam- coverage and temporal resolution is provided.
pling scheme effectively doubling the number of simultane- During the first years of routine clinical integration of
ously acquired slices by means of a z-flying focal spot [45, coronary CT angiography, retrospectively ECG-gated spiral
46] (Z-FFS, Siemens Healthcare, Forchheim, Germany; (helical) scanning was the most widely used data acquisition
6  Cardiac CT Platforms: State of the Art 57

Fig. 6.7  Impact of slice

thickness on resolution.
Z-flying focal spot techniques
allow for the reconstruction of
thinner slices and this sharper
delineation of anatomical
structures

0.6 mm Slice 2.0 mm Slice

technique. A spiral CT scan with continuous data acquisition with cone-beam effects. The table is then moved to the next
and table movement at low table feed (low spiral pitch of z-position, and the next axial CT scan is performed at the
approximately 0.2) is performed to examine the patient’s corresponding temporal distance from the next R-wave. This
heart. Simultaneously, the patient’s ECG is recorded. way, the heart volume is sequentially covered by axial scans.
Following data acquisition, the ECG serves as guidance to ECG-triggered axial scanning is a very dose-efficient
select those data intervals in the spiral CT data set that scan technique. In a simple technical realization, a partial
describe the same user-selected cardiac phase, measured in scan data interval – the minimum amount of data necessary
different cardiac cycles. Those data are then used for phase-­ for image reconstruction  – is acquired in a user-selected
consistent image reconstruction of the coronary anatomy. phase of the patient’s cardiac cycle. Then, however, neither
The upper part of Fig. 6.8 shows the principle of retrospec- retrospective reconstruction in a slightly different phase of
tively ECG-gated spiral scanning. This scan mode is very the patient’s cardiac cycle to minimize motion artifacts is
versatile because it allows for retrospective optimization of possible nor are multiphase reconstructions for evaluation of
the reconstruction window and image reconstruction in dif- cardiac function. Furthermore, due to the fact that incom-
ferent cardiac phases. plete data is acquired, the temporal offset of scan data acqui-
As a downside, retrospectively ECG-gated spiral scan- sition to the next R-wave is based on a prospective estimate
ning of the heart requires strongly overlapping spiral data of the previous cycle’s length. This is a challenge in patients
acquisition at low table feed for phase-consistent imaging with irregular heart beat and arrhythmia, unless the acquisi-
and is therefore associated with relatively high radiation tion strategy is properly tailored and aimed at an end-systolic
dose to the patient, which can be somewhat reduced by using cardiac phase [47].
ECG-controlled dose modulation approaches (“ECG-­
pulsing”). Many modern CT scanners are equipped with
such versatile ECG-pulsing algorithms which react flexibly From 4 cm MDCT CT Systems to Area
to arrhythmia and ectopic beats and have the potential to Detector and Dual-Source CT
extend the clinical application spectrum of ECG-­
synchronized dose modulation to patients with irregular Area Detector CT
heart rates.
As an alternative scan mode, prospectively ECG-triggered ECG-synchronized CT volume imaging of the heart
sequential scanning in a “step-and-shoot” technique is more acquired by the abovementioned MDCT systems typically
commonly used with newer generations of CT systems. results in 3–4 sub-volumes reconstructed from data mea-
During the CT examination, the patient’s ECG is used to sured over multiple consecutive heart beats (Fig. 6.9) [48].
trigger the start time for the acquisition of axial (sequential) Variations of heart motion from one cardiac cycle to the
CT images; see lower part of Fig. 6.8. The patient table is next, as well as contrast media dynamics between the
moved to its defined start position: with a user-selectable cycles, can result in these image sub-volumes being shifted
temporal distance from an R-wave, an axial CT scan without relative to each other and, as a consequence, resulting in
table movement is performed. The CT system acquires scan stairstep or slab ­registration artifacts in multi-planar refor-
data covering a sub-volume of the patient’s cardiac anatomy mations (MPRs) or volume rendered images (VRTs).
which corresponds to around 90% of the total z-width for a Recently, 128-row, 256-slice CT systems with 80  mm
64-row detector and is reduced for CT systems with large z-axis coverage were commercially introduced (Philips
detectors in the z-axis direction which have to cope more ICT, Philips Healthcare, Best, the Netherlands). They facil-
58 B. Schmidt et al.

Recon

Recon

Recon
Continuous
table movement

Scan Detecor z-coverage

Scan

Table movement

Table movement

z-axis

Fig. 6.8  Established acquisition modes for cardiac scanning with CT covered. Similar scan times are accomplished in prospectively ECG-­
systems up to 4 cm coverage: In retrospective ECG-gated scan mode triggered axial (sequential) scanning of the heart (lower part). Based on
(upper part), a helical scan with a comparatively small, fixed heart rate-­ the user-selected phase position of ECG signal, an axial scan is per-
dependent table feed is performed. For higher heart rates, table feed is formed. After each axial scan, the table moves to the next z-position.
increased. In case of a 4 cm system, after 3–4 heart beats, the heart is This acquisition technique is also called “step-and-shoot”

itate examinations of the heart with two axial scans, thereby As of today, two CT vendors provide single-source CT
reducing the number of potential “steps” in the volume systems with large area detectors capable of imaging the
image to one [49]. In a study with 160 consecutive patients, entire heart in one beat, using one axial scan without table
prospectively triggered axial coronary CTA performed on movement. With 320 × 0.5 mm collimation at 0.28 s rotation
256-slice CT provided significantly improved and more time (Toshiba Aquilion ONE, Toshiba Medical Systems
stable image quality at an equivalent effective radiation Corporation, Tokyo, Japan) and 256 × 0.625 mm collimation
dose compared with 64-slice CT [50]. at 0.28  s rotation time (GE Revolution, GE Healthcare,
6  Cardiac CT Platforms: State of the Art 59

Waukesha, USA), these systems can cover 16  cm in the for reconstruction (Fig. 6.10). In addition, due to this effect
z-axis direction at isocenter, large enough to cover the entire also dose efficiency is compromised since not all acquired
heart in one beat and hence avoid stairstep and slab registra- data can be utilized fully for the reconstruction of the
tion artifacts, as well as problems related to inconsistent con- images. Another challenge of larger detector z-coverage is
trast enhancement in different sub-volumes [51]. increased X-ray scatter. Scattered radiation may cause
A possible consequence of the large cone-beam angle hypo-dense cupping or streaking artifacts, and the scatter-
that those systems have compared to systems with smaller induced noise may reduce the contrast-to-noise ratio (CNR)
z-­coverage can be limitations in the field of view available in the images [52]. To reduce scatter – in this case forward
scatter – some vendors have added scatter grids. While anti-
scatter grids are a well-known topic for dual-source CT sys-
tems, it is important to note that anti-scatter grids are also
recommended in the case of single-source system with wide
detectors. A study by Engel et al. showed that the contribu-
tion of scatter for a 16  cm single-source system is even
higher than in the case of an 8 cm dual-source system, which
again motivates the need for anti-scatter grids (ASG) even
for single-source systems in case of wide detectors [53].
The requirements were intensely evaluated by Vogtmeier
et al. [54]. Figure 6.11 shows the impact on scatter for 1D
and 2D scatter grids, for different collimations. In particular,
for collimations between 8 and 16  cm coverage, the 2D
grids performed substantially better.
Another challenge of area detectors are the often neces-
sary changes to the tube design. In order to compensate for
the so-called Heel effect and to accomplish a homogeneous
Fig. 6.9  Non-axial view of a cardiac examination performed with intensity on the detector along the scan direction (Fig. 6.12),
ECG-gated spiral mode. Due to the limited coverage, slab artifacts are
the anode angle is often flatted [55]. The intensity profile is
pronounced due to contrast dynamics and a slight phase mismatch
between different acquired heart beats combined with limited temporal improved, however, at the price of reduced tube power. This
resolution consequence can also be seen when looking at the maximum
16 cm

Z-FOV

XY-FOV
4 cm

16 cm 4 cm

Fig. 6.10  Comparison of systems with 16 cm and 4 cm detector cover- regions are exposed (highlighted in red), they cannot be reconstructed
age in axial scan mode. Left side: view from the side. The area of full with traditional reconstruction techniques. As a consequence distorted
FOV is substantially reduced in case of wide detector systems. Although regions are often cropped in the reconstructed images (see right side)
60 B. Schmidt et al.

Fig. 6.11  Contribution of Measured SPR for 1D-ASG and 2D-ASG

scatter to the primary signal
for different collimations and Water Phantom Diameter
designs of anti-scatter grid
(ASG). (From Engel et al. 0.25 16”
[53], with permission)

Scatter to Primary Ratio

0.2

1D
ASG
0.15
12”

0.1
8”
16”
2D
0.05
12” ASG
8”

0
0 20 40 60 80 100 120 140
Z-Coverage [mm]

a b

Anode Cathode

ANODE

Z 4.2˚
X

6.3˚

12.6˚

0˚ X 0˚
Z
CATHODE

Fig. 6.12  Due to the Heel effect, the footprint of the projected focal changed to a flatter design. The drawback might be a reduced power
spot becomes much narrower on the anode side. To compensate and output of the tube. (From Li et al. [55], with permission)
have a more homogeneous intensity, typically the anode angle is

tube currently available on high-end CTs. For non-16  cm Despite all the abovementioned challenges, successful
systems, tube currents up to 1000 mA (Philips) and 1300 mA use of CT systems with 16 cm detector coverage for coro-
(Siemens) are enabled, whereas the 16  cm systems have nary CTA has been demonstrated [53, 56, 57]. The applica-
maximum mA values of 735 mA (GE) and 600 mA (Toshiba). tion spectrum has been extended to scanning of patients with
6  Cardiac CT Platforms: State of the Art 61

atrial fibrillation [58] and coronary CTA combined with first Dual-Source CT
pass perfusion evaluation [59, 60]. The second generation of
320-row CT scanners has been shown to enable coronary A dual-source CT (DSCT) is a CT system with two X-ray
CTA at reduced radiation dose compared to the first genera- tubes and two detectors at an angle of approximately 90°
tion [52, 56]. (Fig. 6.14). Both measurement systems acquire CT scan data
As a second benefit, CT systems with wide-area detectors simultaneously at the same anatomical level of the patient
can acquire dynamic volume data by repeatedly scanning the (same z-position).
same anatomical range without table movement. This is use- The first generation of DSCT scanners with 2 × 64 slices
ful in dynamic CT angiographic examinations or for whole using z-flying focal spot (19.2  mm detector coverage) and
heart functional data acquisition. So et  al. published first 0.33  s gantry rotation time was introduced in 2006
results on phantoms and an animal model, demonstrating (SOMATOM Definition, Siemens Healthcare, Forchheim,
general feasibility for the quantitative evaluation of the myo- Germany), the second generation with 2  ×  128 slices
cardial blood flow [61].
As discussed above in detail, systems with 16 cm detec-
tors suffer multiple limitations. Some of the challenges are
related to the X-ray source. To compensate for some of the
effects, multiple sources in z-direction have been proposed
(Fig.  6.13). Fast alternating switching is realized between
two X-ray tubes during data acquisition. Due to the improved
geometry setup, this approach does not suffer, for example,
from cone-beam artifacts compared to traditional wide-area
16 cm CT systems. A commercially available system of this
kind was introduced in 2016 (SpotLight CT, Arineta Ltd.,
Israel). Optimized for a small footprint, the system has a
diagnostic field of view limited to 250  mm, 192 detector
rows with z-coverage of 140 mm and a fast gantry rotation
time of 0.24 s [62]. One can derive from these specifications
that the system comprises a highly specialized cardiac CT
system in contrast to the above-described multipurpose full-­
Fig. 6.14  DSCT with two independent measurement systems. The
body scanners. image shows the first-generation DSCT with an angle of 90° between
both measurement systems. To increase the SFOV of detector B, an
angle of 95° was chosen for the second- and third-generation
systems

Fig. 6.13  Setup of CT

system in so-called inverse
geometry
62 B. Schmidt et al.

(38.4 mm detector coverage) and 0.28 s gantry rotation time example, temporal resolution is about 0.4 times the rotation
in 2009 (SOMATOM Definition Flash, Siemens Healthcare, time – this is 100 ms with the second-generation DSCT [70].
Forchheim, Germany), and the third generation with 2 × 192 With the high-pitch scan mode, very high scan speed is
slices (57.6 mm detector coverage) and 0.25 s gantry rotation achieved – up to 450 mm/s with the second-generation DSCT
time in 2014 (SOMATOM Definition Force, Siemens (38.4 mm detector coverage, 0.28 s gantry rotation time) and
Healthcare, Forchheim, Germany). up to 737 mm/s with the third-generation DSCT (57.6 mm
DSCT systems provide significantly improved tempo- detector coverage, 0.25 s gantry rotation time). This is ben-
ral resolution for cardiothoracic imaging. The shortest eficial for the examination of larger anatomical ranges in
data acquisition time for an image corresponds to a quar- very short scan times, e.g., for chest CTA at high temporal
ter of the gantry rotation time. Close to the isocenter, 180° resolution [24], for the evaluation of pulmonary embolism
of scan data is the minimum needed for image reconstruc- and visualization of most cardiac structures and proximal
tion. Due to the 90° angle between both X-ray tubes, each coronary arteries [71], for fast CTA scans of the aorta at low
of the measurement systems needs to acquire only 90° of radiation and contrast dose [72], or when the patient has lim-
scan data. The two 90° segments at the same anatomical ited ability to cooperate, such as in pediatric radiology
level are put together to form the 180° scan. Using this [73–76].
technique, a temporal resolution of 83  ms, 75  ms, and The high-pitch scan mode can also be used in combina-
66 ms, respectively, is achieved for the three generations tion with ECG-triggering – the patient’s ECG triggers both
of DSCT systems. With the dual-source approach, tempo- table motion and data acquisition. The patient table is posi-
ral resolution is independent of the patient’s heart rate, tioned, and table acceleration is started in a way that the table
because data from only one cardiac cycle are used to arrives at the prescribed start z-position (e.g., the base or the
reconstruct an image. This is a major difference to single- apex of the heart) at the requested cardiac phase after full
source MDCT systems, which can provide similar tempo- table speed has been reached (Fig. 6.15). Then data acquisi-
ral resolution by combining data from several heart cycles tion begins. The scan data for images at adjacent z-positions
to an image in a multi-segment reconstruction. Then, are acquired at slightly different – typically diastolic – phases
however, temporal resolution strongly depends on the of the cardiac cycle. Meanwhile, several clinical studies have
relation of heart rate and gantry rotation time. Meanwhile, demonstrated the successful use of the high-pitch scan tech-
several clinical studies have demonstrated the potential of nique for coronary CT angiography in patients with suffi-
DSCT to reliably perform coronary CT angiographic ciently low and stable heart rate (<62  bpm with the
studies in patients with high and even irregular heart rates second-generation DSCT, <70 bpm with the third-generation
(e.g., [63–65]). DSCT is sufficiently accurate to diagnose DSCT), with the potential to scan the entire heart in one beat
clinically significant coronary artery disease in difficult to at very low radiation dose [23, 77–80].
image patients [66, 67]. The high temporal resolution is ECG-triggered high-pitch scans can be used also for com-
also beneficial for reducing motion artifacts in cardiotho- prehensive thorax examinations in the emergency room and
racic studies (e.g., [68]). in the planning of TAVR procedures, because they provide
With a DSCT system, both X-ray tubes can also be oper- adequate visualization of the coronary arteries, the aorta, and
ated at different kV settings, e.g., 80 and 140 kV, to acquire the iliac arteries in one scan at low radiation dose and high
dual-energy CT data. The advantages and disadvantages of temporal resolution. The very short total scan time may
different techniques to acquire dual-energy CT data as well potentially allow for a reduction of the amount of contrast
as clinically relevant applications will be discussed substan- agent; see, e.g., [27, 81]. Figure 6.15 shows an ECG-triggered
tially more in detail in Chap. 4. high-pitch CTA of the aorta as an example.
While the maximum spiral pitch in single-source CT Despite their clinical benefits, DSCT systems have to
images is limited to about 1.5 to ensure gapless volume cov- cope with some challenges. One challenge is the presence of
erage, DSCT systems can be operated at twice the pitch. cross-scattered radiation, i.e., scattered radiation from X-ray
Data acquired with the second measurement system a quarter tube B detected by detector A and vice versa. Cross-scattered
rotation after the first measurement system can be used to fill radiation – if not corrected for – can result in image artifacts
the sampling gaps up to a pitch of about 3.2  in a limited and degraded CNR of the images [69]. Another challenge is
SFOV that is covered by both detectors [69, 70]. At maxi- the limited SFOV of the second detector, which was increased
mum pitch, no redundant data are acquired, and a quarter from 25 cm in the first-generation DSCT to 35.5 cm in the
rotation of data per measurement system is used for image third-generation DSCT.
reconstruction. Temporal resolution is then a quarter of the The most straightforward correction approach is to
gantry rotation time. At decreasing pitch, temporal resolu- directly measure the cross-scattered radiation in detectors A
tion decreases because of the increasing angular data seg- and B and to subtract it from the measured signal. This tech-
ment that corresponds to an image. At a pitch of 2, for nique was first implemented in the second-generation
6  Cardiac CT Platforms: State of the Art 63

Aquisition time
interaval trot/4 per image

Requested
start phase
z-axis
60 60 60

Table speed
Scan
and z-position
Table speed reached at
correct start phase

Start table acceleration

Time

Fig. 6.15  ECG-triggered start of table movement and data acquisition for the high-pitch DSCT spiral

Fig. 6.16  Schematic diagram

of a detector module with
scatter sensors outside the
Collimator blades
direct beam. The center
region consists of the primary
detector pixels, positioned
below the collimator blades of
the anti-scatter collimator.
The scattered radiation
sensors, placed on both sides Scatter sensors
of the primary detector
Detector
module, are equipped with
elements
collimator blades as well. The
arrows indicate the z-direction
(row direction) and the
fan-angle direction (channel
direction)

z-direction
(detector rows q)

fan angle (β-) direction

(detector channels k)

DSCT.  It requires additional detector elements on each scattered radiation in each measured projection. In the
detector outside the direct beam (Fig. 6.16). An alternative to ­third-­generation DSCT, an iterative cross-scatter correction
direct measurement is a model-based cross-scatter correc- based on a simplified Monte Carlo simulation is applied.
tion. The primary source of cross-scattered radiation is While image artifacts caused by cross-scattered radiation
Compton scatter at the object surface. In the first-generation can be significantly reduced by either model-based or
DSCT, pre-stored cross-scatter tables for objects with similar measurement-­based correction approaches, these corrections
surface shape are used for an online correction of the cross-­ are often considered to come at the expense of increased
64 B. Schmidt et al.

image noise and reduced contrast-to-noise ratio (CNR) in the of about 90°. Motion artifacts in the corresponding A and B
images. It has been demonstrated, however, that careful low images can therefore be slightly different, which can affect
pass filtering of the scatter correction term can efficiently the material decomposition of the dual-energy images. In
mitigate the image noise increase without visually affecting practice, however, this problem is not relevant thanks to the
image detail resolution or image contrast. CNR can in fact be good temporal resolution of DSCT, and it can be further
increased beyond the values achieved without cross-scatter mitigated by nonrigid registration of A and B images.
correction, and it approaches or sometimes even surpasses
the CNR performance of a comparable single-source scan
using these techniques.
Another  – evitable  – challenge of dual-source dual-
References
energy CT (DE) is the 90° offset of projections acquired by 1. Kalender WA, Seissler W, Klotz E, Vock P. Spiral volumetric CT
both X-ray tubes at the same z-position. Because high- with single-breath-hold technique, continuous transport, and con-
energy and low-energy projections are not simultaneously tinuous scanner rotation. Radiology. 1990;176(1):181–3.
acquired at the same projection angle, raw data-based DE 2. Achenbach S, Ulzheimer S, Baum U, et  al. Noninvasive coro-
nary angiography by retrospectively ECG-gated multi-slice spiral
algorithms are difficult to realize. DE algorithms are there- CT. Circulation. 2000;102:2823–8.
fore image based – which has the huge advantage of avoid- 3. Becker C, Knez A, Ohnesorge B, Schöpf U, Reiser M. Imaging of
ing the limited and sensitive calibration of raw data-based non calcified coronary plaques using helical CT with retrospective
methods. It is often claimed that image-based methods are EKG gating. AJR. 2000;175:423–4.
4. Knez A, Becker C, Leber A, Ohnesorge B, Reiser M, Haberl R. Non-­
strongly limited by the problem of beam hardening. invasive assessment of coronary artery stenoses with multidetector
However, under certain conditions – which are typically ful- helical computed tomography. Circulation. 2000;101:e221–2.
filled in modern CT scanners  – image-based methods are 5. Ohnesorge B, Flohr T, Becker C, Kopp A, Schoepf U, Baum U,
practically equivalent for clinical tasks. A prerequisite for Knez A, Klingenbeck Regn K, Reiser M.  Cardiac imaging by
means of electro- cardiographically gated multisection spiral CT –
image-based material decomposition is the validity of the initial experience. Radiology. 2000;217:564–71.
thin absorber model. For example, if one uses water and 6. Nieman K, Oudkerk M, Rensing B, van Oijen P, Munne A, van
iodine as the basis materials for image-based dual-energy Geuns R, de Feyter P. Coronary angiography with multi-slice com-
decomposition, the maximum X-ray attenuation of the puted tomography. Lancet. 2001;357:599–603.
7. Hoffmann MHK, Shi H, Schmitz BL, Schmid FT, Lieberknecht M,
iodine along any measured ray path is expected to be so Schulze R, Ludwig B, Kroschel U, Jahnke N, Haerer W, Brambs
small that it is valid to assume a linear contribution to the H-J, Aschoff AJ.  Noninvasive coronary angiography with mul-
total attenuation. In clinical practice, this prerequisite is tislice computed tomography. JAMA. 2005;293:2471–8.
typically fulfilled. If necessary, still dedicated established 8. Leschka S, Wildermuth S, Boehm T, Desbiolles L, Husmann L,
Plass A, Koepfli P, Schepis T, Marincek B, Kaufmann PA, Alkadhi
single energy-based methods for beam-­hardening correc- H.  Noninvasive coronary angiography with 64-section CT: effect
tions can be applied to allow for respective data correction. of average heart rate and heart rate variability on image quality.
As a second prerequisite, both the CT value of water and the Radiology. 2006;241(2):378–85.
CT values of small iodine samples are expected to be inde- 9. Leber AW, Knez A, Ziegler F, Becker A, Nikolaou K, Paul S,
Wintersperger B, Reiser MF, Becker CR, Steinbeck G, Boekstegers
pendent from their position within the scanned object. P. Quantification of obstructive and nonobstructive coronary lesions
DSCT scanners are therefore equipped with an optimized by 64-slice computed tomography  – a comparative study with
bowtie filter of sufficient beam hardening, and the approxi- quantitative coronary angiography and intravascular ultrasound. J
mately cylindrical patient cross-section has to be centered Am Coll Cardiol. 2005;46(1):147–54.
10. Raff GL, Gallagher MJ, O’Neill WW, Goldstein JA. Diagnostic
within the SFOV – typically being fulfilled in case of clinical accuracy of non-invasive coronary angiography using
CT systems. Anyhow, in practice, electronic noise, scanner 64-slice spiral computed tomography. J Am Coll Cardiol.
calibration, stability of emitted spectra, cone-beam effects, 2005;46(3):552.
and scattered radiation can have a larger impact on the 11. Salem R, Remy-Jardin M, Delhaye D, Khalil C, Teisseire A,

Delannoy-Deken V, Duhamel A, Remy J. Integrated cardio-thoracic
obtained results than the raw data or image-based analysis imaging with ECG-gated 64-slice multidetector-row CT: initial
method itself. Further on, the ability for an optimized, inde- findings in 133 patients. Eur Radiol. 2006;16(9):1973–81.
pendent selection of X-ray voltage and spectral filtration – 12. Bruzzi JF, Rémy-Jardin M, Delhaye D, Teisseire A, Khalil

but still dose-­wise well and optimally balanced high and C, Rémy J.  When, why, and how to examine the heart dur-
ing thoracic CT: part 1, basic principles. AJR Am J Roentgenol.
low energy beam – and therefore an optimal spectral separa- 2006;186(2):324–32.
tion enables robust dual-energy imaging at a huge variety 13. Bruzzi JF, Rémy-Jardin M, Delhaye D, Teisseire A, Khalil

for applications and thus compensates for potential limita- C, Rémy J.  When, why, and how to examine the heart during
tions and challenges. Further details are provided in Chap. ­thoracic CT: part 2, clinical applications. AJR Am J Roentgenol.
2006;186(2):333–41.
4, dedicated to dual-energy imaging of the heart. 14. Delhaye D, Remy-Jardin M, Salem R, Teisseire A, Khalil C,

A third challenge is the fact that moving objects, such as Delannoy-Deken V, Duhamel A, Remy J. Coronary imaging quality
the heart, are seen by both detectors with an angular offset in routine ECG-gated multidetector CT examinations of the entire
6  Cardiac CT Platforms: State of the Art 65

thorax: preliminary experience with a 64-slice CT system in 133 30. Ohno Y, Koyama H, Matsumoto K, Onishi Y, Takenaka D,
patients. Eur Radiol. 2007;17(4):902–10. Fujisawa Y, Yoshikawa T, Konishi M, Maniwa Y, Nishimura
15. Flohr T, Ohnesorge B. Heart rate adaptive optimization of spatial Y, Ito T, Sugimura K. Differentiation of malignant and benign
and temporal resolution for electrocardiogram-gated multislice spi- pulmonary nodules with quantitative first-pass 320-detec-
ral CT of the heart. J Comput Assist Tomogr. 2001;25(6):907–23. tor row perfusion CT versus FDG PET/CT.  Radiology.
16. Hausleiter J, Martinoff S, Hadamitzky M, et  al. Image quality 2011;258(2):599–609.
and radiation exposure with a low tube voltage protocol for coro- 31. Willems PW, Taeshineetanakul P, Schenk B, Brouwer PA,

nary CT angiography: results of the PROTECTION II trial. JACC Terbrugge KG, Krings T.  The use of 4D-CTA in the diagnostic
Cardiovasc Imaging. 2010;3(11):1113–23. work-up of brain arteriovenous malformations. Neuroradiology.
17. Flohr TG, McCollough CH, Bruder H, Petersilka M, Gruber K, Süß 2012;54(2):123–31.
C, Grasruck M, Stierstorfer K, Krauss B, Raupach R, Primak AN, 32. Motosugi U, Ichikawa T, Sou H, Morisaka H, Sano K, Araki

Küttner A, Achenbach S, Becker C, Kopp A, Ohnesorge BM. First T. Multi-organ perfusion CT in the abdomen using a 320-­detector
performance evaluation of a dual-source CT (DSCT) system. Eur row CT scanner: preliminary results of perfusion changes in the
Radiol. 2006;16(2):256–68. liver, spleen, and pancreas of cirrhotic patients. Eur J Radiol.
18. Johnson TR, Nikolaou K, Busch S, Leber AW, Becker A,
2012;81(10):2533–7.
Wintersperger BJ, Rist C, Knez A, Reiser MF, Becker CR. Diagnostic 33. US Patent X-Ray transmission scanning system and method and
accuracy of dual-source computed tomography in the diagnosis of electron beam X-ray scan tube for use therewith. US 4352021.
coronary artery disease. Investig Radiol. 2007;42(10):684–91. 1982.
19. Pontana F, Faivre JB, Remy-Jardin M, Flohr T, Schmidt B,
34. Agatston AS, Janowitz WR, Hildner FJ, et al. Quantification of cor-
Tacelli N, Pansini V, Remy J.  Lung perfusion with dual-energy onary artery calcium using ultrafast computed tomography. J Am
multidetector-­row CT (MDCT): feasibility for the evaluation Coll Cardiol. 1990;15(4):827–32.
of acute pulmonary embolism in 117 consecutive patients. Acad 35. Achenbach S, Ropers D, Regenfus M, Muschiol G, Daniel WG,
Radiol. 2008;15(12):1494–504. Moshage W. Contrast enhanced electron beam computed tomogra-
20. Thieme SF, Becker CR, Hacker M, Nikolaou K, Reiser MF,
phy to analyse the coronary arteries in patients after acute myocar-
Johnson TR. Dual energy CT for the assessment of lung perfusion- dial infarction. Heart. 2000;84(5):489–93.
-correlation to scintigraphy. Eur J Radiol. 2008;68(3):369–74. 36. Rumberger JA, Sheedy PF, Breen JF. Use of ultrafast (cine) x-ray
21. Achenbach S, Marwan M, Ropers D, et  al. Coronary computed computed tomography in cardiac and cardiovascular imaging. In:
tomography angiography with a consistent dose below 1 mSv using Giuliani ER, Gersh BJ, McGoon MD, Hayes DL, Schaff HF, edi-
prospectively electrocardiogram-triggered high-pitch spiral acqui- tors. Mayo clinic practice of cardiology. 3rd ed. St. Louis: Mosby;
sition. Eur Heart J. 2010;31(3):340–6. 1996. p. 303–24.
22. Alkadhi H, Stolzmann P, Desbiolles L, Baumueller S, Goetti R, 37. Bell MR, Lerman LO, Rumberger JA.  Validation of minimally
Plass A, Scheffel H, Feuchtner G, Falk V, Marincek B, Leschka invasive measurement of myocardial perfusion using electron beam
S.  Low-dose, 128-slice, dual-source CT coronary angiography: computed tomography and application in human volunteers. Heart.
accuracy and radiation dose of the high-pitch and the step-and-­ 1999;81:628–35.
shoot mode. Heart. 2010;96(12):933–8. https://doi.org/10.1136/ 38. Lu B, Zhuang N, Mao S, Bakhsheshhi H, Liu S, Budoff M. Image
hrt.2009.189100. quality of three-dimensional electron beam coronary angiography.
23. Lell M, Hinkmann F, Anders K, Deak P, Kalender WA, Uder
J Comput Assist Tomogr. 2002;26:202–9.
M, Achenbach S.  High-pitch electrocardiogram-triggered com- 39. Niemann T, Henry S, Duhamel A, Faivre JB, Deschildre A,

puted tomography of the chest: initial results. Investig Radiol. Colas L, Santangelo T, Remy J, Remy-Jardin M.  Pediatric chest
2009;44(11):728–33. CT at 70 kVp: a feasibility study in 129 children. Pediatr Radiol.
24. Tacelli N, Remy-Jardin M, Flohr T, Faivre JB, Delannoy V, Duhamel 2014;44(11):1347–57.
A, Remy J. Dual-source chest CT angiography with high temporal 40. Durand S, Paul JF. Comparison of image quality between 70 kVp
resolution and high pitch modes: evaluation of image quality in 140 and 80 kVp: application to paediatric cardiac CT.  Eur Radiol.
patients. Eur Radiol. 2010;20(5):1188–96. 2014;24(12):3003–9.
25. Schulz B, Jacobi V, Beeres M, Bodelle B, Gruber T, Lee C, Bauer 41. LaBounty TM, Leipsic J, Poulter R, et al. Coronary CT angiogra-
R, Kerl M, Vogl T, Zangos S. Quantitative analysis of motion arti- phy of patients with a normal body mass index using 80 kVp versus
facts in high-pitch dual-source computed tomography of the thorax. 100 kVp: a prospective, multicenter, multivendor randomized trial.
J Thorac Imaging. 2012;27(6):382–6. AJR Am J Roentgenol. 2011;197(5):W860–7.
26. Beeres M, Schell B, Mastragelopoulos A, Herrmann E, Kerl JM, 42. Oda S, Utsunomiya D, Funama Y, et al. A low tube voltage tech-
Gruber-Rouh T, Lee C, Siebenhandl P, Bodelle B, Zangos S, Vogl nique reduces the radiation dose at retrospective ECG-gated car-
TJ, Jacobi V, Bauer RW. High-pitch dual-source CT angiography of diac computed tomography for anatomical and functional analyses.
the whole aorta without ECG synchronisation: initial experience. Acad Radiol. 2011;18(8):991–9.
Eur Radiol. 2012;22(1):129–37. 43. Nakagawa M, Ozawa Y, Sakurai K, Shimohira M, Ohashi K, Asano
27. Wuest W, Anders K, Schuhbaeck A, May MS, Gauss S, Marwan M, Yamaguchi S, Shibamoto Y. Image quality at low tube voltage
M, Arnold M, Ensminger S, Muschiol G, Daniel WG, Uder M, (70 kV) and sinogram-affirmed iterative reconstruction for com-
Achenbach S.  Dual source multidetector CT-angiography before puted tomography in infants with congenital heart disease. Pediatr
Transcatheter Aortic Valve Implantation (TAVI) using a high-pitch Radiol. 2015;45(10):1472–9.
spiral acquisition mode. Eur Radiol. 2012;22(1):51–8. 44. Duan X, Wang J, Leng S, Schmidt B, Allmendinger T, Grant

28. Plank F, Friedrich G, Bartel T, Mueller S, Bonaros N, Heinz A, K, Flohr T, McCollough CH.  Electronic noise in CT detec-
Klauser A, Cartes-Zumelzu F, Grimm M, Feuchtner G. Benefits of tors: impact on image noise and artifacts. AJR Am J Roentgenol.
high-pitch 128-slice dual-source computed tomography for plan- 2013;201(4):W626–32.
ning of transcatheter aortic valve implantation. Ann Thorac Surg. 45. Flohr T, Stierstorfer K, Raupach R, Ulzheimer S, Bruder

2012;94(6):1961–6. H.  Performance evaluation of a 64-slice CT-system with z-flying
29. Hsiao EM, Rybicki FJ, Steigner M.  CT coronary angiography: focal spot. Rofo. 2004;176:1803–10.
256-slice and 320-detector row scanners. Curr Cardiol Rep. 4 6. Flohr TG, Stierstorfer K, Ulzheimer S, Bruder H, Primak AN,
2010;12(1):68–75. https://doi.org/10.1007/s11886-009-0075-z. McCollough CH.  Image reconstruction and image quality
66 B. Schmidt et al.

evaluation for a 64-slice CT scanner with z-flying focal spot. coronary artery disease and severity of coronary artery calcifica-
Med Phys. 2005;32(8):2536–47. tion: the CORE320 study. Int J Cardiol. 2015;201:570–7.
47. Lee AM, Beaudoin J, Engel LC, Sidhu MS, Abbara S, Brady TJ, 61. So A, Imai Y, Nett B, Jackson J, Nett L, Hsieh J, Wisenberg G, Teefy
Hoffmann U, Ghoshhajra BB.  Assessment of image quality and P, Yadegari A, Islam A, Lee TY.  Technical Note: Evaluation of a
radiation dose of prospectively ECG-triggered adaptive dual-source 160-mm/256-row CT scanner for whole-heart quantitative myocar-
coronary computed tomography angiography (cCTA) with arrhyth- dial perfusion imaging. Med Phys. 2016;43(8):4821. https://doi.
mia rejection algorithm in systole versus diastole: a retrospective org/10.1118/1.4957389.
cohort study. Int J Cardiovasc Imaging. 2013;29(6):1361–70. 62. FDA, 510k Summary Document https://www.accessdata.fda.gov/
48. Flohr TG, Schoepf UJ, Ohnesorge BM.  Chasing the heart: new cdrh_docs/pdf16/K161066.pdf
developments for cardiac CT. J Thorac Imaging. 2007;22(1):4–16. 63. Sun ML, Lu B, Wu RZ, Johnson L, Han L, Liu G, Yu FF, Hou
49. Muenzel D, Noel PB, Dorn F, Dobritz M, Rummeny EJ, Huber ZH, Gao Y, Wang HY, Jiang S, Yang YJ, Qiao SB.  Diagnostic
A.  Coronary CT angiography in step-and-shoot technique with accuracy of dual-source CT coronary angiography with prospec-
256-­slice CT: impact of the field of view on image quality, cra- tive ECG-triggering on different heart rate patients. Eur Radiol.
niocaudal coverage, and radiation exposure. Eur J Radiol. 2011;21(8):1635–42.
2012;81(7):1562–8. 64. Lee AM, Engel LC, Shah B, Liew G, Sidhu MS, Kalra M, Abbara
50. Klass O, Walker M, Siebach A, Stuber T, Feuerlein S, Juchems S, Brady TJ, Hoffmann U, Ghoshhajra BB.  Coronary computed
M, Hoffmann MH.  Prospectively gated axial CT coronary angi- tomography angiography during arrhythmia: radiation dose reduc-
ography: comparison of image quality and effective radiation dose tion with prospectively ECG-triggered axial and retrospectively
between 64- and 256-slice CT. Eur Radiol. 2010;20(5):1124–31. ECG-gated helical 128-slice dual-source CT. J Cardiovasc Comput
51. Rybicki FJ, Otero HJ, Steigner ML, Vorobiof G, Nallamshetty Tomogr. 2012;6(3):172–183.e2.
L, Mitsouras D, Ersoy H, Mather RT, Judy PF, Cai T, Coyner K, 65. Paul JF, Amato A, Rohnean A.  Low-dose coronary-CT angiogra-
Schultz K, Whitmore AG, Di Carli MF. Initial evaluation of coro- phy using step and shoot at any heart rate: comparison of image
nary images from 320-detector row computed tomography. Int J quality at systole for high heart rate and diastole for low heart rate
Cardiovasc Imaging. 2008;24(5):535–46. with a 128-slice dual-source machine. Int J Cardiovasc Imaging.
52. Tomizawa N, Maeda E, Akahane M, Torigoe R, Kiryu S, Ohtomo 2013;29(3):651–7.
K.  Coronary CT angiography using the second-generation 66. Westwood ME, Raatz HD, Misso K, Burgers L, Redekop K,

320-detector row CT: assessment of image quality and radiation Lhachimi SK, Armstrong N, Kleijnen J. Systematic review of the
dose in various heart rates compared with the first-generation scan- accuracy of dual-source cardiac CT for detection of arterial stenosis
ner. Int J Cardiovasc Imaging. 2013;29(7):1613–8. in difficult to image patient groups. Radiology. 2013;267(2):387–95.
53. Engel KJ, Herrmann C, Zeitler G. X-ray scattering in single- and 67. Sidhu MS, Venkatesh V, Hoffmann U, et  al. Advanced adaptive
dual-source CT. Med Phys. 2008;35(1):318–32. axial-sequential prospectively electrocardiogram-triggered dual-­
54. Vogtmeier G, Dorscheida R, Engela KR, Luhtab K, Mattsonb R, source coronary computed tomographic angiography in a patient
Harwoodb B, Applebyc M, Randolphc B, KlingerVogtmeier J, with atrial fibrillation. J Comput Assist Tomogr. 2011;35(6):747–8.
et al. Two-dimensional anti-scatter grids for computed tomography 68. Hutt A, Tacelli N, Faivre JB, Flohr T, Duhamel A, Remy J, Remy-­
detectors. Medical imaging. International Society for Optics and Jardin M. Is bronchial wall imaging affected by temporal resolu-
Photonics. 2008. tion? Comparative evaluation at 140 and 75 ms in 90 patients. Eur
55. Li B, Toth TL, Hsieh J, et al. Simulation and analysis of image qual- Radiol. 2016;26(2):469–77.
ity impacts from single source, ultra-wide coverage CT scanner. J 69. Petersilka M, Stierstorfer K, Bruder H, Flohr T. Strategies for scat-
Xray Sci Technol. 2012;20(4):395–404. ter correction in dual source CT. Med Phys. 2010;37(11):5971–92.
56. Chen MY, Shanbhag SM, Arai AE. Submillisievert median radia- 70. T G F, Leng S, Yu L, Allmendinger T, Bruder H, Petersilka M,
tion dose for coronary angiography with a second-­ generation Eusemann CD, Stierstorfer K, Schmidt B, McCollough C.  Dual-­
320-detector row CT scanner in 107 consecutive patients. source spiral CT with pitch up to 3.2 and 75 ms temporal resolution:
Radiology. 2013;267(1):76–85. image reconstruction and assessment of image quality. Med Phys.
57. Nasis A, Leung MC, Antonis PR, Cameron JD, Lehman SJ,
2009;36(12):5641–53.
Hope SA, Crossett MP, Troupis JM, Meredith IT, Seneviratne 71. Hou DJ, Tso DK, Davison C, Inacio J, Louis LJ, Nicolaou S,
SK.  Diagnostic accuracy of noninvasive coronary angiography Reimann AJ, Hou DJ, Tso DK, Davison C, Inacio J, Louis LJ,
with 320-detector row computed tomography. Am J Cardiol. Nicolaou S, Reimann AJ. Clinical utility of ultra high pitch dual
2010;106(10):1429–35. source thoracic CT imaging of acute pulmonary embolism in the
58. Kondo T, Kumamaru KK, Fujimoto S, Matsutani H, Sano T,
emergency department: are we one step closer towards a non-gated
Takase S, Rybicki FJ.  Prospective ECG-gated coronary 320-­ triple rule out? Eur J Radiol. 2013;82(10):1793–8.
MDCT angiography with absolute acquisition delay strategy for 72. Apfaltrer P, Hanna EL, Schoepf UJ, Spears JR, Schoenberg SO,
patients with persistent atrial fibrillation. AJR Am J Roentgenol. Fink C, Vliegenthart R. Radiation dose and image quality at high-­
2013;201(6):1197–203. pitch CT angiography of the aorta: intraindividual and interindi-
59. George RT, Mehra VC, Chen MY, Kitagawa K, Arbab-Zadeh A, vidual comparisons with conventional CT angiography. AJR Am J
Miller JM, Matheson MB, Vavere AL, Kofoed KF, Rochitte CE, Roentgenol. 2012;199(6):1402–9.
Dewey M, Yaw TS, Niinuma H, Brenner W, Cox C, Clouse ME, 73. Han BK, Overman DM, Grant K, et al. Non-sedated, free breathing
Lima JA, Di Carli M.  Myocardial CT perfusion imaging and cardiac CT for evaluation of complex congenital heart disease in
SPECT for the diagnosis of coronary artery disease: a head-to-head neonates. J Cardiovasc Comput Tomogr. 2013;7(6):354–60.
comparison from the CORE320 multicenter diagnostic perfor- 74. Lee JH, Han D, Danad I, et  al. Multimodality imaging in coro-
mance study. Radiology. 2015;274(2):626. nary artery disease: focus on computed tomography. J Cardiovasc
60. Sharma RK, Arbab-Zadeh A, Kishi S, Chen MY, Magalhães TA, Ultrasound. 2016;24(1):7–17.
George RT, Dewey M, Rybicki FJ, Kofoed KF, de Roos A, Tan SY, 75. Lell MM, May M, Deak P, Alibek S, Kuefner M, Kuettner A, Köhler
Matheson M, Vavere A, Cox C, Clouse ME, Miller JM, Brinker JA, H, Achenbach S, Uder M, Radkow T. High-pitch spiral computed
Arai AE, Di Carli MF, Rochitte CE, Lima JA. Incremental diagnos- tomography: effect on image quality and radiation dose in pediatric
tic accuracy of computed tomography myocardial perfusion imag- chest computed tomography. Investig Radiol. 2011;46(2):116–23.
ing over coronary angiography stratified by pre-test probability of
6  Cardiac CT Platforms: State of the Art 67

76. Bridoux A, Hutt A, Faivre JB, Flohr T, Duhamel A, Pagniez J, Remy 79. Gordic S, Husarik DB, Desbiolles L, Leschka S, Frauenfelder T,
J, Remy-Jardin M.  Coronary artery visibility in free-­ breathing Alkadhi H. High-pitch coronary CT angiography with third genera-
young children on non-gated chest CT: impact of temporal resolu- tion dual-source CT: limits of heart rate. Int J Cardiovasc Imaging.
tion. Pediatr Radiol. 2015;45(12):1761–70. 2014;30(6):1173–9.
77. Achenbach S, Marwan M, Schepis T, Pflederer T, Bruder H, 80. Morsbach F, Gordic S, Desbiolles L, Husarik D, Frauenfelder T,
Allmendinger T, Petersilka M, Anders K, Lell M, Kuettner A, Schmidt B, Allmendinger T, Wildermuth S, Alkadhi H, Leschka
Ropers D, Daniel WG, Flohr T.  High-pitch spiral acquisition: S.  Performance of turbo high-pitch dual-source CT for coronary
a new scan mode for coronary CT angiography. J Cardiovasc CT angiography: first ex vivo and patient experience. Eur Radiol.
Comput Tomogr. 2009;3:117–21. 2014;24(8):1889–95.
78. Leschka S, Stolzmann P, Desbiolles L, Baumueller S, Goetti R, 81. Azzalini L, Abbara S, Ghoshhajra BB.  Ultra-low contrast com-
Schertler T, Scheffel H, Plass A, Falk V, Feuchtner G, Marincek B, puted tomographic angiography (CTA) with 20-mL total dose for
Alkadhi H. Diagnostic accuracy of high-pitch dual-source CT for transcatheter aortic valve implantation (TAVI) planning. J Comput
the assessment of coronary stenoses: first experience. Eur Radiol. Assist Tomogr. 2014;38(1):105–9.
2009;19(12):2896–903.
 Principles of Cardiac CT Image
Acquisition 7
Thomas Henzler, Patricia Carrascosa, Brian S. Ko,
and Ronen Rubinshtein

Ongoing technical developments in CT have established car- that allow whole-heart coverage. Moreover, the last years
diac CT angiography (cCTA) as a robust noninvasive imag- have also seen significant progress in novel iterative recon-
ing test of the heart and coronary arteries, with ever-improving struction techniques that allow further dose reduction for
image quality and diagnostic accuracy [1–3]. However, con- cCTA as well as improved image quality [4, 5]. However, all
cerns have been raised (with the first- and second-generation CT vendors have slightly different approaches within their
scanners) regarding radiation exposure from cCTA, which cCTA protocols that make it difficult to provide a universal
has traditionally been comparatively high relative to other overview on cCTA acquisition principles. Therefore, this
CT applications, although the dose levels compare well with chapter separately describes cCTA acquisition protocols for
other longer established cardiac imaging tests such as the major vendors from the perspective of experienced users
radionuclide-­ based myocardial perfusion imaging. in order to provide the readership with a practical approach
Considering the increasing number of cCTA investigations that can be directly transferred into their own clinical
around the globe, however, and expected sustained growth, routine.
the contribution of cCTA to the radiation exposure of the
general population is not negligible. While there is no evi-
dence that the level of radiation that is ordinarily applied at  eneral Principles for Cardiac CT Image
G
cCTA bestows any kind of risk, the current uncertainties Acquisition
regarding the biological effects of radiation, our commit-
ment to the ALARA (As Low As Reasonably Achievable) Scan Coverage Optimization
principles, and increasing public radiation awareness
behoove us to keep radiation dose at medical imaging to a The scan coverage is planned using the topogram or “scout”
minimum. view identical to other CT acquisitions. As the total radiation
Over the past years, several acquisition techniques for dose delivered is directly proportional to the scan coverage,
cCTA have been developed in order to minimize patients’ precise collimation is important to minimize the DLP (dose
radiation exposure including prospective ECG gating, high-­ length product = scan length × CTDIvol). With the improved
pitch cCTA, and single-heartbeat cCTA using large detectors capabilities of CT systems, there is a general tendency to
increase the area of coverage, due to faster acquisition and
T. Henzler (*) less limitation of tube power. If the z-axis of the scan is too
Institute of Clinical Radiology and Nuclear Medicine, University long, for example, including the upper part of the abdomen
Medical Center Mannheim, Medical Faculty Mannheim,
Heidelberg University, Heidelberg, Germany or the neck, there will be unnecessary radiation delivered to
e-mail: [email protected] abdominal organs or the thyroid gland. However, if the scan
P. Carrascosa length is too short, excluding a portion of the coronary artery
Department of Cardiovascular Imaging, Diagnóstico Maipú, tree, the examination would not be complete and repeat scan-
Buenos Aires, Argentina ning would be required, resulting in increased radiation and
B. S. Ko contrast media volume to the patient. For dedicated coronary
Department of Medicine Monash Medical Centre (MMC), Monash artery imaging (i.e., excluding coronary artery bypass and
Cardiovascular Research Centre, MonashHEART, Monash Health “triple-rule-out” acute chest pain studies to rule out aortic
and Monash University, Melbourne, Australia
dissection or pulmonary embolism in addition to coronary
R. Rubinshtein disease), scanning should begin at the level of the bifurcation
Department of Cardiovascular Medicine, Lady Davis Carmel
Medical Center, The Ruth and Bruce Rappaport School of of the trachea in most cases, but the exact starting position is
Medicine, Technion-Israel Institute of Technology, Haifa, Israel dependent on individual anatomy. For example, in thin

© Humana Press 2019 69

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_7
70 T. Henzler et al.

patients with a small, vertically oriented heart, scanning may radiation dose. Thus, the high temporal, spatial, and contrast
start lower than the carina without the risk of missing the resolution required for retrospectively ECG-gated cCTA is
proximal coronary artery tree. It is also important to be able obtained at the expense of an increase in radiation exposure.
to stop the scan during the acquisition once the entire cardiac Increased awareness of these relationships has prompted the
anatomy is covered, so as to avoid unnecessary radiation to development of a variety of strategies aimed at improving the
the upper abdomen. Thus, adjustment of scan length, and efficiency of ECG-synchronized acquisition techniques.
thus optimization of the radiation dose, requires the careful With appropriate application of these techniques, a more
attention of the CT technologist. Scan length optimization efficient utilization of X-ray photons does not necessarily
solely based on the topogram may be difficult, as the heart sacrifice the image quality of a cCTA study [12]. Choice of
position may vary with diaphragmatic motion, and some the ECG synchronization method at cCTA should be
patients may have difficulty keeping the same respiratory informed by various factors, such as patients’ heart rate or
state for successive acquisitions. If a noncontrast calcium rhythm, and the clinical question. Until a few years, the most
scoring scan is acquired preceding cCTA, as it is practice at frequently used scan technique for ECG synchronization at
many institutions, the landmarks obtained at calcium scoring cCTA has been retrospective ECG gating. The ECG signal is
can be advantageously used for more exact planning of the continuously recorded, while a low-pitch spiral/helical CT
scan range of the contrast-enhanced study. scan with continuous emission of X-rays is acquired. Because
the spiral/helical pitch used for this technique is very low
compared to a standard CT examination (usually around
 etrospective Versus Prospective ECG
R 0.2), the radiation exposure may be up to three times higher
Synchronization than that of a standard (non-ECG-gated) CT examination of
the same anatomic region (Fig. 7.1). This scan technique is
Retrospective ECG gating has been the traditional method of still commonly employed in patients with an irregular heart
ECG synchronization with mechanical CT scanners, which rhythm or in whom functional assessment of the myocar-
comprises simultaneous acquisition of the ECG signal com- dium or the valves is desired. However, the increased radia-
bined with constant X-ray emission during slow-pitch spiral/ tion burden with use of this technique has to be individually
helical scan acquisitions [6–10]. Because of this constant, justified by the additional diagnostic benefit of functional
slow-pitch application of radiation, which enables the acqui- assessment, particularly if other radiation-free tests such as
sition of oversampled data sets that allow interrogation of the echocardiography or MRI are available.
entire cardiac anatomy across the RR interval, cardiac CT Prospective ECG triggering, also in jargon referred to as
has traditionally been considered radiation dose-intensive, “step-and-shoot mode” [13], has developed to be the most
especially in light of the small volume covered (i.e., only frequently used technique for cCTA over the last years. This
approximately 12–13 cm in the z-axis for adult hearts) (Fig. scan mode eliminates redundant radiation that occurs from
6.1). The relatively high radiation dose is the result of the low-pitch cardiac spiral/helical scanning, since radiation is
particular requirements of this imaging test, where, more exclusively applied during a defined, predetermined phase of
than with other applications, diagnostic evaluation of the the cardiac cycle (typically during diastole with slower heart
small and rapidly moving coronary arteries depends on the rates and during systole with faster heart rates), while the
combination of high temporal resolution, spatial resolution, X-ray tube is completely shut off or turned down by approxi-
and signal-to-noise ratio [7]. Image noise is mainly influ- mately 80% during the remainder of the cardiac cycle
enced by the number of X-ray photons reaching the detector, (Fig. 7.1). With this technique, the effective radiation dose
which in turn is a factor of the object attenuation and tube equivalent can be an order of magnitude lower than with tra-
voltage and is proportional to the slice width, tube current, ditional retrospective ECG gating. However, prospective
and amount of time required to acquire the projection data ECG triggering may suffer from image degradation as more
needed for image reconstruction [11]. For cCTA, the best heartbeats are necessary for the acquisition of the entire scan
possible temporal resolution is required to minimize artifacts volume. Accordingly, prospective ECG triggering will only
resulting from cardiac motion. This goal is achieved by using provide adequate image quality if the detector of the CT sys-
only projections gathered in a defined, short time window for tem is wide enough to scan the entire heart within a limited
image reconstruction, which in turn requires relatively high number of cardiac cycles (around four steps with a detector
X-ray tube settings to ensure sufficient photon flux during of 4 cm), as is the case with 64-slice and higher CT systems.
this narrow acquisition window. High temporal resolution is More importantly, the use of prospective ECG triggering has
ordinarily achieved by fast gantry rotation times [8]; how- traditionally been restricted to patients with regular sinus
ever, in most available CT scanners, this necessitates the use rhythm, while this acquisition technique has been considered
of a slower pitch to avoid discontinuities in the anatomic unsuitable for patients with an arrhythmia, because the
coverage of the heart. A slow pitch in turn yields a higher occurrence of the desired cardiac phase for image acquisition
7  Principles of Cardiac CT Image Acquisition 71

a b
Retrospective ECG gating without ECG Prospective ECG triggering without ECG
dependent tube current modulation dependent tube current pulsing

e High-pitch spiral acquisition

c d
Retrospective ECG gating with ECG Prospective ECG triggering with ECG
dependent tube current modulation dependent tube current pulsing

Heart rate

Radiation dose

Fig. 7.1  General principles of cardiac CT angiography acquisition protocols. The light blue bars represent the time and intensity of the tube
­current during the different ECG phases

becomes increasingly unpredictable with varying length RR radiation compared with traditional prospective ECG
intervals. Accordingly, patients with an arrhythmia have tra- ­triggering albeit with dose levels that are still substantially
ditionally not benefited from this radiation protection strat- lower than with conventional retrospectively ECG-triggered
egy and had to be imaged with conventional retrospective techniques and ECG-based tube current modulation.
ECG gating, which leaves the operator with more flexibility
to choose the portion of the cardiac cycle with the least car-
diac motion for image reconstruction. High Pitch Versus Broad Detector Coverage
There are various technical attempts at improving the
robustness of prospective ECG triggering vis-à-vis faster and Dual-source CT (DSCT) was introduced in 2005 and pro-
more irregular heart rates. These include prolonging the vides particularly high temporal resolution for cCTA.  The
acquisition interval during the RR cycle (“ECG padding”) to system uses two X-ray tubes and two detectors arranged at
provide more flexibility in choosing the most suitable phase an angle of 90° (95° in the second and third generation).
of image reconstruction. While ECG padding provides Therefore, only one-quarter rotation of the gantry is neces-
greater flexibility and improves the robustness to account for sary to acquire the X-ray data for one cross-sectional image.
variations in heart rhythm, it also increases radiation expo- This effectively doubles the temporal resolution when com-
sure by widening the temporal window (“tolerance”) during pared with single-source CT at the same rotation speed. With
which the X-ray tube is turned on. A more sophisticated the introduction of second-generation DSCT, so-called high-­
method to account for heart rate variability is adaptive online pitch single-heartbeat acquisition became feasible. In early
monitoring of the ECG.  This technique detects the occur- reports, the ability to perform ECG-triggered spiral data
rence of extrasystoles and automatically rejects arrhythmic acquisition using very high-pitch values (around 3.2) has
beats in order to ensure that image acquisition occurs only been described for DSCT [14, 15]. The high pitch allows
during the desired cardiac phase. Last, recent developments image acquisition of the entire volumetric data set of the
have successfully addressed the inability of obtaining func- heart within a single cardiac cycle. Since with high-pitch
tional information with prospective ECG triggering. Hybrid acquisitions no redundant data is acquired, this scan mode is
acquisition techniques use a short, prospectively ECG-­ associated with approximately 1/10 the exposure of a retro-
triggered pulse of nominal tube output radiation to supply spectively gated spiral scan and 1/2 to 1/3 the dose of a pro-
morphological data flanked by phases of low-radiation image spectively ECG-triggered scan. In combination with low
acquisition covering end-systole and end-diastole, which tube voltage techniques, effective radiation doses of 1 mSv
enables functional evaluation (Fig.  7.1). However, again, or less can be routinely obtained [16]. When applied to the
since this involves exposure through a longer period of the entire chest for the purpose of “triple-rule-out” acute chest
cardiac cycle, this approach is associated with an increase in pain studies, the effective radiation dose equivalent still
72 T. Henzler et al.

remains at below 2 mSv, depending on the X-ray tube set- ECG-based dose modulation is activated prospectively,
ting. However, since the heart must remain motion-free for according to an automated analysis of the preceding heart-
the entire scan time, this mode has been predominantly beats. One drawback associated with the use of ECG-based
applied in patients with slow heart rates below 65–75 beats dose modulation is that, similar to prospective ECG trigger-
per minute [15, 16]. Similar to second- and third-generation ing, the exact occurrence of systolic and/or diastolic phases
DSCT, broad detector 320-row CT scanners allow image cannot be reliably predicted in patients with arrhythmia or
acquisition of the entire heart within a single heartbeat. With frequent premature ventricular contractions. In these cases,
this technology, the entire z-axis of the heart is covered by a there is a risk of incurring image reconstructions during a
broad (i.e., 16 cm) detector, allowing for the single-heartbeat low-radiation dose phase, with consequent increases in
acquisition. image noise and loss of image quality within the z-extent of
the data acquired during the arrhythmic beat. Thus, in cases
of cardiac arrhythmia, the use of ECG-based tube current
ECG-Based Tube Current Modulation modulation should be used with caution. Similar as with
automated online ECG monitoring techniques for prospec-
ECG-based tube current modulation is currently the most tive ECG triggering, however, this limitation is increasingly
widely used technical tool to reduce radiation dose without addressed by the recent availability of sophisticated algo-
sacrificing image quality at retrospectively ECG-gated rithms that automatically detect premature contraction of the
acquisitions as well as during prospective ECG triggering heart and suspend ECG-based tube current modulation mid-
with a second low-dose pulsing window. Radiation may be scan in a real-time fashion. Despite these current limitations,
reduced by up to 60% by the application of a simple princi- ECG-based tube current modulation has become one of the
ple: radiation dose is decreased during phases of the cardiac most successful and widely used techniques for radiation
cycle that are not anticipated to be useful for morphological protection in clinical routine [18].
evaluation, and the full dose is delivered only during the car-
diac phase which is expected to yield diagnostic results for
morphological evaluation [17]. Some CT manufacturers Tube Voltage
allow adjustment of the temporal window receiving full
dose, which can be beneficial for optimizing the performance Lowering the X-ray tube voltage is one of the most effective
of ECG-dependent tube current modulation. For example, if and simplest methods for radiation dose reduction, since the
the cardiac rhythm is regular, sinusoidal, and slow, there is a radiation dose roughly changes with the square of the tube
high probability that the most suitable phase for reconstruc- voltage [19–21]. However, reduced tube voltage causes an
tion will be found in diastole. In such cases it is advisable to increase in image noise [19, 22]; thus, use of low tube volt-
apply the full dose during end-diastole, representing a tem- age protocols requires an increase in the tube current and
poral window of only 20% of the full RR interval. The performs best with state-of-the-art iterative reconstruction
remaining 80% of the cardiac cycle receives reduced tube techniques in order to maintain a diagnostic contrast-to-­noise
current. For faster heart rates, the most suitable phase for ratio [21, 23]. With further development of X-ray tubes, a
reconstruction becomes more difficult to predict and may wider variety of tube voltage options has become available in
vary from 30% to 80%. In this scenario, it is possible to set cCTA with potential benefits considering luminal contrast
the full-dose period from 30% to 80% (or less), which and radiation exposure. X-ray photon absorption of a partic-
increases the likelihood of including the optimal reconstruc- ular object is dependent on the object’s material composi-
tion phase in the full-dose window, albeit with reduced radia- tion. Different materials show different mass attenuation
tion savings. Future developments should aim at providing coefficients, dependent on the photon energy. In addition, the
operators with the ability to adjust the temporal amplitude of photon energy itself is dependent on the particular X-ray
the modulation more flexibly across the RR cycle. For exam- tube and tube voltage. Iodine molecules effectively absorb
ple, with faster heart rates, it may be desirable to have avail- photons with photon energies ranging from 33.2 to 43.3 keV
able both mid-diastolic and end-systolic images at full dose. with the highest absorption at 33.2 keV (k-edge). However
It is also important to remember that the effectiveness of clinical CT systems’ X-ray tubes always generate a poly-
ECG-based tube current modulation varies with heart rate: chromatic X-ray beam with resulting mean photon energy.
the slower the heart rate, the more efficient the modulation, This limits the capability to generate photons in the range of
because the full-dose phase is proportionally shorter within iodine k-edge to some extent. In recent years the X-ray tubes
the RR interval. Consequently, the use of rate-controlling with 80 kV were the method of choice to get close to iodine
agents for lowering the heart rate will lengthen the diastolic k-edge. 80kV tube voltage has been successfully used in
phase and thus enhance the effectiveness of ECG-based tube chest CTA with beneficial effects on image quality and
current modulation with lower incident radiation doses. improved vascular attenuation. A further decrease in tube
7  Principles of Cardiac CT Image Acquisition 73

voltage to 70 kV, which is equivalent to mean photon energy improvements in image quality, IR has a great impact on
of approximately 53,5 keV, stresses these benefits. However, radiation exposure in cCTA by two different means. In gen-
until recently the applicability of 70 kV was limited by the eral IR enables reduction of tube current during primary
maximum tube current output of 500 mA in previous CT sys- image acquisition with preserved image quality compared to
tems. This issue has been overcome with the introduction of FBP. This is caused by IR inherently better denoising capa-
new-generation scanners allowing higher tube output (such bilities and results in dose savings of around 50% [4].
as the third-generation DSCT with a potential maximum Furthermore the lower image noise enables reduction of tube
tube current of up to 1300 mA) which enables routine cCTA voltage in the latest CT systems with potential for further
at 70  kV tube voltage in patients with BMI <26  kg/m2. reduction of radiation exposure [16].
Moreover, higher tube output (higher mA) and low tube volt-
age (lower kV) allow better tissue penetrance and may be
suitable for obese patients thus allowing better image quality  verview of Cardiac CT Acquisition
O
without significant increase in image noise. Platforms and Techniques Used by
the Various CT Vendors
Iterative Reconstruction
Iterative reconstruction (IR) techniques were developed Siemens
almost two decades ago, and their advantages over tradi-
tional filtered back projection (FBP) reconstruction have  T Systems for cCTA
C
long been recognized. However, only a few years ago, the Nowadays, robust cCTA imaging vendor independently
increasing performance of mainframe computers enabled requires CT systems with 64 slices or higher. In general,
their entry into clinical CT routine. IR methods substantially Siemens CT systems have to be divided into single-source CT
reduce image noise and thereby may also allow substantial systems (SSCT) and dual-source CT systems (DSCT). DSCT
reduction of radiation dose. Modern IR algorithms are all systems are equipped with two X-ray tubes and two corre-
vendor specific leading to slight differences in the image sponding detectors that are oriented in the gantry with an
appearance. By now every big vendor (e.g., General Electric, angular offset of 90° (Fig. 7.2). The two X-ray source/detector
Philips, Toshiba, and Siemens) offers and develops different systems rotate simultaneously capturing image data in half the
IR algorithms with increasing capabilities considering image time required with conventional technology which is of sig-
quality and noise reduction. While early algorithms mainly nificant importance in patients with high heart rates and or
relied on statistical IR, increasingly sophisticated and com- arrhythmia. Using this technology, DSCT systems have been
plex full or partially raw data domain-based methods are the leader in temporal resolution when compared to other
implemented and available with the latest scanner manufacturers over the last 10 years. Moreover, the temporal
technology. resolution continuously decreased from first-, second-, to
The use of IR algorithms results in reduced image noise third- generation DSCT from 83 ms, 75 ms, to 66 ms. DSCT
thus enabling primary acquisition with decreased tube volt- was also the first technology that allowed dual-­energy CT
age and/or tube current while maintaining or even improving (DECT) without a temporal offset. DECT means that the sys-
diagnostic image quality with cCTA. Besides lower radiation tem uses two X-ray sources simultaneously at different energy
exposure, a reduction in tube voltage has other beneficial levels. This makes it possible to differentiate not only between
effects on vessel-to-tissue contrast as explained in the tube fat, soft tissue, and bone but also between calcifications and
voltage section. In addition, IR is a possibility for noise iodinated contrast material on the basis of their unique energy-
reduction in obese patients where other measures of noise dependent attenuation profiles. For cCTA, DECT allows to
reduction, like increased tube voltage and tube current, might quantitatively measure the myocardial iodine concentration as
be limited by the specific CT system used. Since use of IR a surrogate of myocardial perfusion at a single time point
means a tremendous change to the image generation process, (“snapshot” perfusion during peak enhancement) [27, 28].
there have been concerns this might alter image characteris- Moreover, novel DECT approaches from Siemens also allow
tics such as degree of stenosis or plaque composition. At this subtraction of coronary calcifications using the spectral differ-
juncture it is uncontroversial that the use of IR techniques ences between iodine and calcium. This approach has the
results in fewer blooming artifacts on cCTA which induces potential to further improve the specificity of cCTA which was
less overestimation of degree of stenosis, especially in heav- traditionally hampered by blooming artifacts in patients with
ily calcified vessels and stents [24, 25]. Consecutively there severe calcifications leading to false-positive test results. The
is a tendency to lower percentage of high attenuation plaque selective visualization of iodinated contrast material does also
if compared to FBP. Studies evaluating the effect on plaque allow the calculation of so-called virtual noncontrast images
volume and plaque composition assessment have shown which could be used for coronary artery calcium scoring based
minimal effect in quantified plaque volume [26]. Besides on a single cCTA acquisition [29]. However, although DECT
74 T. Henzler et al.

a c d

Fig. 7.2  Retrospective ECG-gated cardiac CT angiography (cCTA) of resulting in a total radiation dose of 3.2 mSv. The case demonstrates the
a 52-year-old male patient with acute chest pain performed on a third-­ high potential of radiation dose reduction due to low tube voltage imag-
generation dual-source CT. The retrospective cCTA protocol was cho- ing even in adult normal weight patients in which retrospective ECG
sen due to the high irregular heart rate of the patient 95 bpm. In order to gating was necessary
save radiation dose, the protocol was performed at 70 kVp tube voltage

offers several advantages for cCTA, it is important to keep in Likewise, to other manufacturers Siemens continuously
mind that the inherent high temporal resolution of DSCT sys- developed and improved IR techniques over the past 6 years.
tems is reduced by 50% using DECT settings. However, novel Whereas the first-generation of IR (IRIS) was slightly ham-
advantages in image reconstruction that are currently under pered by a so-called plastic-like appearance of the CT images
investigation have the potential to virtually save the high tem- compared to traditional FBP, this issue was solved by the two
poral resolution during raw data reconstruction using iterative following generations of IR techniques (SAFIRE/ADMIRE).
reconstructions that require less than 180° of projections [30]. IR is of paramount importance for low kVp imaging since
State-of-the-art SSCT systems from Siemens also provide a image noise significantly increases with decreased tube volt-
high temporal resolution for cCTA up to 142  ms (e.g., age. Thus, it is recommended to use IR techniques in all
SOMATOM Edge) due to a high gantry rotation time of 0.28 s. cCTA studies acquired with a tube voltage below 100 kVp.
Lowering the tube voltage in combination with IR has the
Scan Acquisition advantage to simultaneously reduce radiation dose and the
Siemens recognized early the general need for low tube amount of contrast material. A recently published study
voltage imaging in cardiovascular CT.  Today, all cardiac demonstrated the total amount of contrast material can be
CT systems from Siemens provide sufficient tube current reduced from 80  cc for a cCTA performed at 100  kVp to
output to perform low tube voltage cCTA studies dependent 50  cc for a cCTA performed at 70  kVp, recommending a
on patients’ BMI. With the introduction of third-generation reduction of 10 cc per 10 kVp tube voltage reduction [31].
DSCT, routine cCTA at 70 kVp has become a reality even
in adult patients with a high BMI.  Technically, routine  cquisition Protocol Selection
A
70  kVp imaging was realized by the integration of two Selection of the scan acquisition protocols (retrospective gat-
Vectron tubes that both provide a tube current of up to ing, prospective gating, high-pitch acquisition) should be
1300 mA, utilizing power reserves from two 120 kW gen- based on patient’s heart rate and rhythm and the clinical sce-
erators. One of the main breakthroughs related to this tech- nario (high-risk elderly patient  =  robust imaging; low risk
nology is the possibility to have full tube voltage setting young patient = robust but as gently as possible; LV function
flexibilities since the system allows the tube voltage modi- required = no high-pitch acquisition). Generally, high-pitch
fication from 70 to 150 kVp in 10 kVp intervals. Siemens acquisitions deliver high diagnostic image quality in patients
also integrated automated tube voltage selection into all of with regular heart rates of up to 70 bpm (up to 75 with third-­
their novel CT systems. Automated tube voltage selection generation DSCT) (Fig.  7.3). With Siemens CT systems
interacts with the well-­known automated tube current mod- above 64 slices, prospective ECG triggering has become the
ulation and automatically selects the optimal tube voltage most frequently used technique for cCTA especially since
tube current combination based on patients’ anatomy the technique also enabled LV function analysis. Traditionally,
detected with the topogram. evaluation of cardiac function with cCTA was restricted to
7  Principles of Cardiac CT Image Acquisition 75

Fig. 7.3  High-pitch (3.2)

cardiac CT acquisition of a
59-year-old female patient
with chest pain. The study
was performed with 70 kVp
using automated tube voltage
selection in combination with
automated tube current
modulation resulting in a
radiation dose of 0.2 mSv.
The heart rate of the patient
was regular with 73 bpm

retrospective ECG gating with or without ECG-dependent ity and clinical skills through the junction of coverage, spa-
tube current modulation which resulted in high radiation tial and temporal resolution: all in one system [32]. This
dose levels. Using the concept of ECG-dependent tube cur- scanner has been built using the new Gemstone Clarity
rent modulation, Siemens introduced low-dose imaging of detector of 256 detector rows of 0.625 mm thickness. The z
the whole cardiac cycle also for prospective ECG gating. By coverage extends up to 160 mm width achieving up to 512
using this so-called adaptive cardiac sequence, the tube volt- slices with 280 ms of gantry rotation and temporal resolu-
age is reduced to 20% of the maximum tube current output tion of 140 ms (Fig. 7.4).
during the desired phase of the cardiac cycle before moving
to the next table position. However, this technique should Cardiac Protocols
only be used with CT systems that allow to cover the heart in Calcium Score  The technical parameters to perform this
only three single steps in order to minimize contrast inhomo- study in this scanner are 120 Kv, 300 mA, 0.28 s gantry rota-
geneities throughout the coronary arteries in the z-axis direc- tion, noise index 20, and ASIR 50%.
tion. Based on our experience, the full tube current should be
applied throughout a larger part of the cardiac cycle (30–
80%) in patients with heart rated above 80 bpm in order to Coronary CT Angiography  Revolution CT has a wide
obtain both systolic and diastolic images at full dose. coverage thus allowing to scan the heart in a single gantry
Table  7.1 provides an overview on basic characteristics of rotation.
three typical CT systems from Siemens.
It has a wide cone cardiac axial (WCCA) that provides an
approximately 35% dose reduction with constant mA profile
General Electric for single-phase acquisition, generating a great dose advan-
tage compared to Revolution HD and Revolution EVO.
GE Healthcare with the Revolution CT family offers three If the region to be covered is larger such as patients with
options to perform these studies: (1) Revolution CT, (2) bypass grafts, two axial volumes have to be selected for opti-
Revolution HD (with /without GSI), and (3) Revolution mal visualization.
EVO. It is recommended to scan the heart at least an additional
All of them operate with any of the two contrast injection 1 cm in all directions because there is excellent temporal and
protocols: (a) bolus timing scan (BTS) or SmartPrep (SM) spatial resolution in the center of the selected region but
also known as automatic bolus tracking system. reduction of both at the outer margins. The first and last
There are differences between each of the three scanners images will have reduced field of view and will be more sus-
regarding coverage, temporal resolution, and number of ceptible to motion artifacts.
slices among other issues that are shown in Table 7.2. Revolution CT has a tool called “auto-gating” that auto-
Revolution CT is GE’s newest CT scanner, also known as matically determines the heart rate-dependent settings for
volume high definition (VHD). triggered acquisition and gated reconstructions. In this
It has a wide coverage detector width of 160 mm allow- way the scanner automatically establishes the best scan
ing to scan a complete organ as the heart in a single gantry mode for that patient: snapshot pulse or cardiac helical.
rotation. Revolution CT offers uncompromised image qual- With regard to patients’ heart rate, it will also establish the
76 T. Henzler et al.

Table 7.1  Basic characteristics of Siemens CT scanners

Rotation
Model Source Coverage Slice Detection time Acquisition mode Recon mode Recon algorithm
Definition Single source, 38.4 mm 2 × 64 × 0.6 Single 0.28 s Sequential Standard, Sequential:
edge with dual focal layer (step-and-shoot), second gen. multi-cycle, single
spot, 100 kW helical Iterative cycle
(SAFIRE) Helical: multi-
cycle, single cycle
Definition Dual source, with 38.4 mm 2 × 64 × 0.6 Single 0.28 s High pitch (flash Standard, High pitch: single
flash dual focal spot, per source/ layer mode), sequential second gen. cycle
2 × 100 kW detector pair (step-and-shoot), Iterative Sequential:
helical (SAFIRE) multi-cycle, single
cycle
Helical: multi-
cycle, single cycle
Definition Dual source, with 52.2 mm 2 × 96 × 0.6 Single 0.25 s High pitch (flash Standard, High pitch: Single
force dual focal spot, per source/ layer mode), sequential third gen. cycle
2 × 120 kW detector pair (step-and-shoot), Iterative Sequential:
helical (ADMIRE) multi-cycle, single
cycle
Helical: multi-
cycle, single cycle

Table 7.2  Comparison between GE scanners for cardiac studies Parameter settings such as pitch and ECG modulation are
Technical parameters Revolution Revolution also determined automatically. Based on patient size
and tools Revolution CT HD EVO observed in a scout view, kV assist will provide with optimal
Detector row (N°) 128 64 64 mA and kVp. Cardiac modulation of mA is available allow-
Slices per rotation 512 128 128 ing to provide the adequate exposure to the patient in order
(N°)
to attain the best image quality with the lowest radiation dose
Gantry rotation (ms) 280 350 350
Temporal resolution 140 175 175
in keeping with the ALARA principle (as low as reasonably
(ms) achievable). This new tool admits simplification and robust-
Iterative ASIR V ASIR ASIR V ness pressing only one button and no longer in need of heart
reconstruction rate-specific protocols.
Snapshot freeze ✓(Intelligent) ✓ ✓ Revolution CT applies new iterative reconstruction algo-
Dual energy mode Work in ✓optional x rithm called ASIR-V that enables up to 80% radiation dose
progress
reduction compared to standard filtered back projection
Other cardiac tools Cardiac Cardiac Cardiac
auto-gating assist assist (FBP) with similar reconstruction speed. It also improves
low-contrast detectability up to 135% at the same dose, spa-
tial resolution, and image noise and can reduce low-signal
cardiac phase range to acquire, which phase needs to be
artifacts such as streak artifacts.
reconstructed, and when snapshot freeze should be applied.
Snapshot freeze (SSF) can be used to reduce coronary
Regarding patients’ beat-to-beat HR variation, it will
artery motion artifacts in high-definition or in standard car-
determine how much additional padding may be needed,
diac studies. A three-phase image series is generated for sub-
whether +ms-based trigger should be used instead of % of
sequent motion correction processing on the AW workstation
the RR interval or if retrigger double scan is appropriate in
or AW server [33].
patients with irregular rhythm such as arrhythmias or pre-
ventricular contraction. 1. Revolution HD (with/without GSI): consists in a high-­
For functional evaluation, regardless of the heart rate, definition scanner that works in single energy (SE) or dual
the tube current will be 100% power at 40% and 75% and energy (DE) using Gemstone Spectral Imaging (GSI).
only 20% at the other phases. In this acquisition mode,
additional phases every 10% will be reconstructed to show This scanner is 64-row detectors and has coverage of 4 cm.
the complete heartbeat and allow ejection fraction Other technical information: collimation width 0.625  mm,
calculation. gantry rotation 0.35 s, and temporal resolution 175 ms.
7  Principles of Cardiac CT Image Acquisition 77

Fig. 7.4  The newest CT

scanner with wide coverage
of 160 mm that allows
scanning the complete heart
in one gantry rotation. In
cases of patients with bypass
grafts, two volumes need to
be acquired. The fast gantry
rotation of 0.28 ms gives an
ideal temporal resolution to
scan patients even with high
heart rates and obtains
excellent image quality.
Cardiac auto-gating tool
automatically determines the
best scan mode for that
patient (snapshot pulse or
cardiac helical), the cardiac
phase range to acquire, which
phase needs to be
reconstructed, and when
snapshot freeze should be
applied making the scan very
simple. Also it determines if
retrigger double scan is
necessary in patients with
irregular rhythm such as
arrhythmias or pre-ventricular
contraction

It operates with Smart Dose technologies such as kV Dual-Energy GSI  It operates exclusively in prospective
assist, dosed organ modulation, 3D dose modulation, and mode for all cardiac studies. It acquires the information per-
dose check among others. It also operates with ASIR. forming rapid switching (0.3–0.5 ms) between low and high
tube potentials (80–140  kVp) using only one source [37].
Single Energy   Single energy works in standard or high-­ According to the patient’s heart rate and variability, different
definition mode allowing to improve spatial resolution phases and additional padding could be necessary. For patients
to 0.28 mm [34]. Snapshot assist (SSA) helps the user to with (a) <65 bpm, center phase at 75%. Additional padding of
determine the acquisition mode according to the patient’s 100  ms can be complemented if the rhythm is unstable: (b)
heart rate, beat-to-beat variation, and size. SSF can be 65–74 bpm center phase at 60%, padding 150–200 ms, and (c)
applied in prospective or retrospective techniques. For >75 bpm center phase at 40%, padding 100 ms. SSF can be
intermediate heart rates (65–74  bpm), especially with applied to improve temporal resolution with similar results
moderate to high variability, SSF has to be prescribed at with single energy [38]. ASIR is actually available in GSI in all
75% phase (mid-­ d iastole) and 45% RR interval energy levels (40–140 keV) admitting to reduce noise mainly at
­( end-systole), and in cases with high HR > 75 bpm, only 40–60 keV (levels with highest noise). Dual-energy CT signifi-
45% phase is needed [35]. Appropriate cardiac phase cantly reduces blooming in patients with severe calcification of
selection is mandatory for optimal performance. the coronary arteries and beam-hardening artifacts improving
Excellent contrast opacification is essential for ideal combined ­anatomo-­functional evaluation present with single-
SSF ­correction [36]. energy CT in determined population (Fig. 7.5) [39].
78 T. Henzler et al.

Stress Rest Stress Rest

a d g j

Monochromatic 85 keV

b e h l

c f i ll

m n

Fig. 7.5  A 75-year-old female with cardiovascular risk factors such as septal, anteroseptal, apical, and inferior walls (white arrows). (d–f) Rest
hypertension, dyslipidemia, and diabetes. She presents with angina and CT in the same views demonstrating normal perfusion stress (g, h, l)
dyspnea FC II–III. A combined anatomo-functional evaluation was per- and rest (j, l, ll) SPECT images show the same findings than dual-­
formed with dual-energy CT (GSI) to assess coronary stenosis and energy CT confirming an anteroseptal, septal, apical, and inferior isch-
myocardial perfusion with reduction of booming and beam-hardening emia. (m) CTCA of the left anterior descendent (LAD) artery exhibiting
artifacts. Monochromatic evaluation at medium energy levels (85 keV) a severe stenosis in the midportion (arrow head). (n) Invasive coronary
was used for the analysis. (a–c) Stress CT at short-axis, four-chamber, angiography showing the same lesion in the LAD
and long-axis views, respectively, showing severe hypoperfusion in

2. Revolution EVO: this scanner has a coverage of 4  cm, fore allowing for 128-row and 256-row acquisition (Brilliance
gantry rotation of 0.35 ms, collimation width of 0.625 mm, iCT-SP (128), and Brilliance iCT (256)) were introduced in
and clarity detector that improves dose efficiency and 2008 for clinical use and underwent several iterations until
signal-to-noise ratio. its current configuration. Philips takes special care toward its
detector technology and rapid gantry rotation time. Image
It operates with ASIR V and applies Smart Dose technol- analysis may be performed via dedicated software available
ogies, snapshot assist, snapshot pulse, ECG dose modula- on their Extended Brilliance Workstation (EBW) and also
tion, 3D mA modulation, and pre-scan dose estimates. via a PC-based portal that allows most analysis processes to
be performed on raw data accessed from a laptop or office-/
home-based personal computer.
Philips Recently, Philips introduced a new version of the 64-slice
scanner equipped with their new detector technology
Several CT scanners from Philips Medical Systems are cur- (NanoPanel Prism) allowing separation between low- and
rently available (Table 7.3). Those include the 64-slice ver- high-energy raw data resulting in spectral imaging. This new
sion (Brilliance 64) which was the first 64-slice scanner spectral CT scanner (IQon) technology will not be discussed
offered for clinical use in October 2004. The later 64- and here. The IQon gantry rotation speed is higher than the con-
128-slice scanners using a dual focal spot correction, there- ventional Brilliance 64 and similar to the Brilliance iCT-SP,
7  Principles of Cardiac CT Image Acquisition 79

Table 7.3  Basic characteristics of Philips CT scanners

Rotation Acquisition
Model Source Coverage Slice Detection time mode Recon mode Recon algorithm
Br64/ Single source, with 40 mm 64 × 0.625 Single 0.42 s/rot Axial Standard, Helical: multi-
Inginuity64 dual focal spot, layer (step-and-­ iDose4, IMR cycle, single cycle
80 kW shoot), helical Axial: single cycle
iCT-SP (128) Single source, with 40 mm 128 × 0.625 Single 0.27 s/rot Axial Standard, Helical: multi-
dual focal spot, layer (step-and-­ iDose4, IMR cycle, single cycle
120 kW shoot), helical Axial: single cycle
iCT-SP (256) Single source, with 80 mm 256 × 0.625 Single 0.27 s/rot Axial Standard, Helical: multi-
dual focal spot, layer (step-and-­ iDose4, IMR cycle, single cycle
120 kW shoot), helical Axial: single cycle
IQon Single source, with 40 mm 128 × 0.625 Dual 0.27 s/rot Axial Standard, Helical: multi-
dual focal spot, layer (step-and-­ iDose4, IMR cycle, single cycle
120 kW shoot), helical Axial: single cycle

and most acquisition parameters are similar too. The funda- rithm is used that identified the same physiologic phases of
mentals of dual energy and spectral imaging are presented the cardiac cycle while taking into account the nonlinear
elsewhere in this textbook. Additionally, all Philips platforms changes in the individual cardiac states with the heart rate
allow for the use of iterative reconstruction (either conven- variations during CT acquisition. As mentioned above car-
tional “iDose” or the model-based “iMR”). diac adaptive multi-cycle (or multi-segment) reconstruction
technique is typically used that combined data from consec-
 rilliance 64: Typical Image Acquisition
B utive cardiac cycles, thus improving temporal resolution.
and Commonly Used Protocols Dedicated software (Arrhythmia Editing, Philips Medical
Calcium Score Acquisition  The coronary artery calcium Systems, Cleveland, OH) allowed correction for R wave
score (Agatston units) may be measured on a non-enhanced irregularity after data acquisition in patients with marked
scan using prospective ECG triggering. Typical scan param- sinus arrhythmia or frequent premature beats. Local “phase
eters include a tube voltage of 120  kV, tube current of point” editing is also available post acquisition (not only
55  mAs, gantry rotation time of 0.42/s, and collimation of R-tag detection).
64 × 0.625 mm. Reconstruction is performed with a CB filter
with 2.5  mm or 3  mm slice thickness without overlap.  rilliance iCT (256)
B
Coronary CT angiography when using the 60 kW tube of the cCTA is performed on this 256-row scanner which has a lon-
scanner may be performed with retrospective electrocardio- gitudinal coverage of 8 cm, offers options for different rota-
graphic (ECG) gating or with prospective ECG triggering tion times which can be as fast as 0.27/s, and has a 120 kW
(step-and-shoot mode). The contrast-enhanced scan is per- generator. cCTA is performed either as prospectively trig-
formed using a bolus of approximately 70 (range 40–120) ml gered “step-and-shoot” mode or with helical retrospective
of contrast media injected into an antecubital vein at a flow ECG gating. ECG-based tube current modulation can be
rate of 5–7  ml/s, followed by a 50  ml saline chaser bolus. used when possible with retrospective gating (DoseRight
Scanning is typically performed at 120  kV, effective tube Cardiac).
current 600–1000 mAs, slice collimation of 64 × 0.625 mm The coronary artery calcium score (Agatston units) may
acquisition, 0.42  s gantry rotation time, and pitch 0.2 as a be measured on a non-enhanced scan using prospective ECG
standard. ECG-based tube current modulation for retrospec- triggering. Typical scan parameters include a tube voltage of
tive gating (“DoseRight Cardiac”) may be used to decrease 120 kV, tube current of 55 mAs, and gantry rotation time of
radiation when possible. Reconstruction is performed using 0.27/s, and autocollimation is frequently 128  ×  0.625  mm.
a window centered at 75% of the RR interval as default (to Reconstruction is performed with a XCB filter with 2.5 mm
coincide with ventricular diastasis). For heart rates above or 3 mm slice thickness without overlap.
70 beats/min, an earlier reconstruction phase (usually 45% Contrast-enhanced scan is performed on the iCT using a
or less) is frequently used (coinciding with isovolumic relax- bolus of 65 (range 50–100) ml contrast media injected into
ation). Typically a multi-cycle scan image reconstruction an antecubital vein at a flow rate of 5–7 ml/s, followed by a
with XCC filter (Philips) and slice thickness reconstruction mixed 50% contrast/saline injection and then a 20–30  ml
of 0.67 mm with 50% overlap is used (allowing spatial reso- saline chaser bolus. Contrast media dose with Philips scan-
lution of 10–24  lp/cm depending on scanner version). If ners is calculated according to the following formula: [(pre-
movement artifacts are apparent, additional reconstruction dicted scan time in seconds +5)  ×  IV contrast flow rate].
windows may be analyzed. A dedicated cardiac gating algo- Scans are performed at 120 kV [or 80/100/140 kV depending
80 T. Henzler et al.

on body mass index (BMI)] with a slice collimation of even in patients with atrial fibrillation [40] and reduces the
128  ×  0.625  mm, with dual z-focal spot positions (which time required for patient breath hold (1–2  seconds) hence
leads to a double number of simultaneous imaged slices per rendering the technique very convenient for the patient while
gantry rotation, therefore 256-row acquisition) and a rotation minimizing potential artifacts secondary to respiratory
time of 0.27 or 0.33 s. For low kV scans (80 or 100 kV), a motion [41].
slower injection rate (4 ml/s) is frequently used resulting in Image data is reconstructed using a half-scan reconstruc-
an overall lower volume of contrast thus maintaining a rela- tion algorithm, typically with a 512 × 512 matrix and 0.5 mm
tively stable contrast to noise ratio. The helical scans (retro- thick sections. Wide detector imaging eliminates “stair-step”
spective ECG gating) are performed with an effective tube or misalignment artifacts and permits uniform attenuation
current (rotation time product normalized by the pitch) in the when evaluating the coronary arteries, when analyzing
range of 900–1500 mAs (effective) depending on BMI and ­myocardial defects [42, 43], or when measuring contrast
body habitus and a pitch of 0.14–0.18. The step-and-shoot attenuation gradients across atherosclerotic lesions [44]
(prospective triggering) scans are performed in patients with (Figs. 7.7, 7.8, and 7.9). The ability to cover the entire heart
stable heart rhythm and heart rate < 65 beats per minute with in one shot permits dynamic volume assessment including of
a tube current X-ray ON time product of 160–300 mAs. The left ventricular function and assessment of absolute myocar-
scanner provides several predefined protocols which deter- dial blood flow [45]. Radiation doses are lower with wide
mines the pitch and rotation time in each protocol according detector scanners owing to shorter scan times and reduction
to scanning mode, heart rate, and whether a “bariatric of redundant radiation from either overlapping of sequential
mode”(high mA) is used (to allow optimal temporal resolu- axial scans or helical oversampling [46]. In addition, total
tion at all settings). Reconstruction is again performed using iodinated contrast volume can be decreased to as low as
a window centered at 75% of the RR interval as default. For 40 ml at 5 ml/s in patients with normal cardiac output [47].
heart rates above 70 BPM, an earlier reconstruction phase There have been three generations of 320 detectors CT
(usually 45% or less) is frequently used with retrospective (Aquilion ONE, Aquilion VISION, and Aquilion GENESIS).
gating. Typical filter used for reconstruction is XCC and Despite equivalent spatial resolution (0.5  mm), the gantry
same 0.67 mm slice thickness with 50% overlap (Fig. 7.6). speed for the VISION and GENESIS generation scanner is
faster than that of the first-generation scanner (275 vs.
350  ms), with a temporal resolution of 138 vs. 175  ms,
Toshiba respectively. In patients with elevated heart rates, in order to
improve the temporal resolution, data from two or three car-
Toshiba manufactures both narrow (80 detectors) and wide diac cycles can be used for image reconstruction, which is
detector (320 detectors) CT scanners (Table 7.4), the latter is called multi-segment reconstruction. Accordingly, in the
the focus of the chapter. Wide detector scanners with 320 first-generation scanner, the temporal resolution of multi-­
detector rows and up to 16  cm of craniocaudal coverage segment reconstruction can be theoretically improved to 88
uniquely enable whole-heart coverage in an axial (not heli- or 58 ms, using two- and three-beat reconstruction. With its
cal) scan mode with one gantry rotation typically within a improved temporal resolution, the newer-generation scan-
single heartbeat in as little as 0.138  s. The short duration ners have been demonstrated to permit single-beat imaging
required for scan acquisition makes the technique applicable in vast majority (93%) in those with heart rates of up to

Fig. 7.6 High-risk/ruptured
plaque (napkin-ring sign)
(right panel-long axis, left
panel-short axis) in the
proximal left anterior
descending coronary artery
demonstrating the high-­
resolution image that may be
acquired with Brilliance iCT
7  Principles of Cardiac CT Image Acquisition 81

Table 7.4  Basic characteristics of Toshiba cardiac CT scanners

Aquilion PRIME Aquilion ONE Aquilion ONE vision Aquilion ONE GENESIS
Basic No. of detectors 80 320 320 320
performance
Thickness of detectors 0.5 mm 0.5 mm 0.5 mm 0.5 mm
No. of slices acquirable 160 using double-slice 640 using double- 640 using double- 640 using double-slice
technology slice technology slice technology technology
Cranial caudal coverage 4 cm 16 cm 16 cm 16 cm
Gantry rotation speed 350 ms 350 ms 275 ms 275 ms
Opening size 78 78 cm 78 cm 78 cm
Tilt angle 30 22 30 30
Acquisition mode Prospective or Axial volume Axial volume Axial volume
retrospective helical
Single beat cardiac CT No, 4–5 beats Yes – up to 65BPM Yes up to 65 bpm Yes – up to 75BPM
acquisition
Tube and Generator capacity 72 kW 72 kW 100 kW 100 kW
generator
Tube voltage 80/100/120/135 kV 80/100/120/135 kV 80/100/120/135 kV 80/100/120/135 kV
Tube current 10–600 mA 10–600 mA 10-10-900 mA 10–900 mA
Computer Iterative reconstruction AIDR 3D hybrid AIDR 3D hybrid AIDR 3D hybrid/
system FIRST model based
Reconstruction speed 30 ips 50 ips 80 ips

a b c

d e f

Fig. 7.7  Wide detector CT examination of a patient with critical steno- arrows) observed during CT stress myocardial perfusion imaging (d–f),
sis in the proximal left anterior descending artery secondary to mixed which are also present but to a much less extent on the corresponding
calcified and noncalcified atherosclerotic plaque as demonstrated on rest images (g–i). This case highlights volumetric assessment providing
axial, curved planar reformat and volume-rendered images on CTCA both coronary anatomy and perfusion using single-beat imaging.
(a–c). There are corresponding perfusion defects in the basal, mid-, and (Adapted from Ko et al. [42], with permission.)
distal anteroseptal, anterior, and anterolateral myocardial segments (red
82 T. Henzler et al.

g h i

Fig. 7.7 (continued)

a b 75 bpm [48]. When compared with the first-generation scan-

ner in patients with elevated heart rates, it has been demon-
strated to provide superior image quality with 35% radiation
dose reduction [49, 50].
Until recently, body weight or body mass index (BMI)
has been used to manually select the tube voltage and tube
current for cardiac CT applications. Based on observations
that attenuation and body habitus over the thorax can be
c discrepant from BMI or weight, Toshiba scanners are cur-
rently equipped with automated exposure control
(SureExposure 3D) which determine a tube potential on the
basis of the X-ray attenuation on anterior-posterior and lat-
eral scout images, body profile of the individual patient and
the reconstruction kernel [51]. SureExposure 3D adopts the
standard deviation noise as a measure of image quality and
aims to match the image noise to the targeted standard
d e f g deviation [52]. This has demonstrated to reduce radiation
exposure by up to 42% [51], without significant impair-
ment of image quality compared to the results obtained
using a BMI-based protocol.
h Similar to other CT vendors, as part of a dose reduc-
800 TAG = -38.3, FFR = 0.76 tion strategy, Toshiba scanners are equipped with adap-
tive iterative dose reduction (AIDR 3D), a reconstruction
Luminal Attenuation (HU)

600 algorithm which is uniquely adapted for large cone-beam

Y = -3.83 X + 625.1
CT examinations and wide-volume acquisitions. The
400 AIDR algorithm can be applied to all acquisition modes
for routine clinical use including CTCA and calcium
200 scores [53] and efficiently eliminates noise and streak
artifact due to photon starvation with a statistical noise
0 and scanner model considering both photon and elec-
0 50 100 150
tronic noise in the raw data domain. The raw data are
Length From Ostium (mm)
reconstructed with filtered back projection (FBP), and
Fig. 7.8  Left anterior descending artery (LAD) with significant obstruc- then a sophisticated iterative technique which optimizes
tive plaque burden imaged by invasive angiography (a) and 320-detector reconstructions for the particular regions is applied in the
CTCA (b). Curved planar reformat and representative cross-sectional image domain, resulting in noise reduction while pre-
views with corresponding luminal attenuation on CCTA (c–g), which per- serving spatial resolution and image texture [54]. It
mitted the calculation of the intraluminal attenuation gradient (h) across a
single heartbeat. The gradient was considered significant (−38.3); the cor- enables the elimination of up to 50% of image noise [55],
responding invasive fractional flow reserve value was 0.76  in the distal compared with FBP reconstruction, and may reduce the
LAD. (Adapted from Wong et al. [44], with permission.) radiation dose up to 22–65% [56, 57].
7  Principles of Cardiac CT Image Acquisition 83

Fig. 7.9  The same patient

underwent CTCA with Aquilion One 2008 Aquilion Vision 2016
Toshiba Acquilion ONE
scanner in 2008 using
retrospective ECG gating and a b
standard filter back projection
reconstruction (a, b) and
again with the Toshiba Vision
scanner in 2016 using
prospective ECG gating and
AIDR3D iterative
reconstruction (c, d). The
representative right coronary
arteries are displayed,
respectively (b, d). The heart
rate acquired in both scans
was 58 bpm. DLP has
reduced by 60% from 440 to
176 mGy-cm

c d

Arineta Inc. SpotLight CT, Cesarea, Israel 2. Meijboom WB, Weustink AC, Pugliese F, et  al. Comparison of
diagnostic accuracy of 64-slice computed tomography coronary
angiography in women versus men with angina pectoris. Am J
A new type of “stereo” CT with a compact size and high Cardiol. 2007;100:1532–7.
gantry rotation speed intended specifically for cardiovascu- 3. Roberts WT, Bax JJ, Davies LC. Cardiac CT and CT coronary angi-
lar imaging was recently introduced and received FDA ography: technology and application. Heart. 2008;94:781–92.
4. Moscariello A, Takx RA, Schoepf UJ, et al. Coronary CT angiog-
approval. This new scanner has two overlapping cone beams
raphy: image quality, diagnostic accuracy, and potential for radia-
facing a single detector array rotating around the patient. It tion dose reduction using a novel iterative image reconstruction
has a coverage (14 cm) sufficient for whole heart scanning technique-comparison with traditional filtered back projection. Eur
in a single beat (acquisition time of 120 ms). This scanner Radiol. 2011;21:2130–8.
5. Den Harder AM, Willemink MJ, De Ruiter QM, et  al. Dose
has the ­fastest CT gantry rotation time currently available
reduction with iterative reconstruction for coronary CT angi-
(0.24 s). Default acquisition is with prospective ECG trig- ography: a systematic review and meta-analysis. Br J Radiol.
gering. This 280-slice CT scanner (192 detector rows) with 2016;89:20150068.
a temporal resolution of 120 ms is of small size (60 cm gan- 6. Brenner D, Elliston C, Hall E, Berdon W.  Estimated risks of
radiation-­induced fatal cancer from pediatric CT.  AJR Am J
try), compatible with small rooms, and was designed for Roentgenol. 2001;176:289–96.
optimal spatial and temporal resolution at a lower cost. The 7. Brenner DJ.  Radiation risks potentially associated with low-­
initial clinical assessment is promising with this scanner. dose CT screening of adult smokers for lung cancer. Radiology.
2004;231:440–5.
8. Brenner DJ, Hall EJ. Risk of cancer from diagnostic X-rays. Lancet.
2004;363:2192; author reply −3.
9. Brenner DJ, Hall EJ. Computed tomography – an increasing source
References of radiation exposure. N Engl J Med. 2007;357:2277–84.
10. Einstein AJ, Moser KW, Thompson RC, Cerqueira MD, Henzlova
1. Bastarrika G, Lee YS, Huda W, Ruzsics B, Costello P, Schoepf MJ.  Radiation dose to patients from cardiac diagnostic imaging.
UJ. CT of coronary artery disease. Radiology. 2009;253:317–38. Circulation. 2007;116:1290–305.
84 T. Henzler et al.

11. Primak AN, McCollough CH, Bruesewitz MR, Zhang J, Fletcher 30. Apfaltrer P, Schoendube H, Schoepf UJ, et  al. Enhanced tempo-
JG. Relationship between noise, dose, and pitch in cardiac multi-­ ral resolution at cardiac CT with a novel CT image reconstruction
detector row CT. Radiographics. 2006;26:1785–94. algorithm: initial patient experience. Eur J Radiol. 2013;82:270–4.
12. Hausleiter J, Meyer T, Hermann F, et  al. Estimated radia-
31. Mangold S, Wichmann JL, Schoepf UJ, et al. Automated tube volt-
tion dose associated with cardiac CT angiography. JAMA. age selection for radiation dose and contrast medium reduction
2009;301:500–7. at coronary CT angiography using 3(rd) generation dual-source
13. Efstathopoulos EP, Kelekis NL, Pantos I, et  al. Reduction of the CT. Eur Radiol. 2016;26:3608–16.
estimated radiation dose and associated patient risk with prospec- 32. Latif MA, Sanchez FW, Sayegh K, et  al. Volumetric single-beat
tive ECG-gated 256-slice CT coronary angiography. Phys Med coronary computed tomography angiography: relationship of
Biol. 2009;54:5209–22. image quality, heart rate, and body mass index. Initial patient expe-
14. Flohr TG, Leng S, Yu L, et  al. Dual-source spiral CT with pitch rience with a new computed tomography scanner. J Comput Assist
up to 3.2 and 75 ms temporal resolution: image reconstruction and Tomogr. 2016;40:763–72.
assessment of image quality. Med Phys. 2009;36:5641–53. 33. Pontone G, Andreini D, Bertella E, et al. Impact of an intra-cycle
15. Achenbach S, Marwan M, Ropers D, et  al. Coronary computed motion correction algorithm on overall evaluability and ­diagnostic
tomography angiography with a consistent dose below 1 mSv using accuracy of computed tomography coronary angiography. Eur
prospectively electrocardiogram-triggered high-pitch spiral acqui- Radiol. 2016;26:147–56.
sition. Eur Heart J. 2010;31:340–6. 34. Pontone G, Bertella E, Mushtaq S, et al. Coronary artery disease:
16. Meyer M, Haubenreisser H, Schoepf UJ, et al. Closing in on the diagnostic accuracy of CT coronary angiography  – a compari-
K edge: coronary CT angiography at 100, 80, and 70 kV-initial son of high and standard spatial resolution scanning. Radiology.
comparison of a second- versus a third-generation dual-source CT 2014;271:688–94.
system. Radiology. 2014;273:373–82. 35. Andreini D, Pontone G, Mushtaq S, et al. Low-dose CT coronary
17. Poll LW, Cohnen M, Brachten S, Ewen K, Modder U. Dose reduc- angiography with a novel IntraCycle motion-correction algorithm
tion in multi-slice CT of the heart by use of ECG-controlled tube in patients with high heart rate or heart rate variability. Eur Heart J
current modulation (“ECG pulsing”): phantom measurements. Cardiovasc Imaging. 2015;16:1093–100.
Rofo. 2002;174:1500–5. 36. Carrascosa P, Deviggiano A, Capunay C, De Zan MC, Goldsmit
18. Hausleiter J, Meyer T, Hadamitzky M, et al. Radiation dose esti- A, Rodriguez-Granillo GA.  Effect of intracycle motion correc-
mates from cardiac multislice computed tomography in daily tion algorithm on image quality and diagnostic performance of
practice: impact of different scanning protocols on effective dose ­computed tomography coronary angiography in patients with sus-
estimates. Circulation. 2006;113:1305–10. pected coronary artery disease. Acad Radiol. 2015;22:81–6.
19. Huda W, Scalzetti EM, Levin G.  Technique factors and image 37. So A, Lee TY, Imai Y, et  al. Quantitative myocardial perfusion
quality as functions of patient weight at abdominal CT. Radiology. imaging using rapid kVp switch dual-energy CT: preliminary expe-
2000;217:430–5. rience. J Cardiovasc Comput Tomogr. 2011;5:430–42.
20. Niemann T, Henry S, Duhamel A, et  al. Pediatric chest CT at 38. Carrascosa P, Deviggiano A, Leipsic JA, et al. Dual energy imag-
70 kVp: a feasibility study in 129 children. Pediatr Radiol. ing and intracycle motion correction for CT coronary angiography
2014;44:1347–57. in patients with intermediate to high likelihood of coronary artery
21. Leschka S, Stolzmann P, Schmid FT, et al. Low kilovoltage cardiac disease. Clin Imaging. 2015;39:1000–5.
dual-source CT: attenuation, noise, and radiation dose. Eur Radiol. 39. Rodriguez-Granillo GA, Carrascosa P, Cipriano S, et  al. Beam
2008;18:1809–17. hardening artifact reduction using dual energy computed tomog-
22. Stolzmann P, Leschka S, Scheffel H, et al. Dual-source CT in step-­ raphy: implications for myocardial perfusion studies. Cardiovasc
and-­shoot mode: noninvasive coronary angiography with low radia- Diagn Ther. 2015;5:79–85.
tion dose. Radiology. 2008;249:71–80. 40. Xu L, Yang L, Fan Z, Yu W, Lv B, Zhang Z.  Diagnostic perfor-
23. Hsieh J, Londt J, Vass M, Li J, Tang X, Okerlund D. Step-and-shoot mance of 320-detector CT coronary angiography in patients with
data acquisition and reconstruction for cardiac x-ray computed atrial fibrillation: preliminary results. Eur Radiol. 2011;21:936–43.
tomography. Med Phys. 2006;33:4236–48. 41. Kang EJ, Lee J, Lee KN, Kown H, Ha DH, Kim RB. An initial ran-
24. Renker M, Nance JW Jr, Schoepf UJ, et  al. Evaluation of heav- domised study assessing free-breathing CCTA using 320-detector
ily calcified vessels with coronary CT angiography: comparison CT. Eur Radiol. 2013;23:1199–209.
of iterative and filtered back projection image reconstruction. 42. Ko BS, Cameron JD, Leung M, et  al. Combined CT coronary
Radiology. 2011;260:390–9. angiography and stress myocardial perfusion imaging for hemody-
25. Renker M, Ramachandra A, Schoepf UJ, et  al. Iterative image namically significant stenoses in patients with suspected coronary
reconstruction techniques: applications for cardiac CT. J Cardiovasc artery disease: a comparison with fractional flow reserve. JACC
Comput Tomogr. 2011;5:225–30. Cardiovasc Imaging. 2012;5:1097–111.
26. Naoum C, Blanke P, Leipsic J. Iterative reconstruction in cardiac 43. Rochitte CE, George RT, Chen MY, et al. Computed tomography
CT. J Cardiovasc Comput Tomogr. 2015;9:255–63. angiography and perfusion to assess coronary artery stenosis caus-
27. Meinel FG, De Cecco CN, Schoepf UJ, et  al. First-arterial-pass ing perfusion defects by single photon emission computed tomog-
dual-energy CT for assessment of myocardial blood supply: do raphy: the CORE320 study. Eur Heart J. 2014;35(17):1120–30.
we need rest, stress, and delayed acquisition? Comparison with 44. Wong DT, Ko BS, Cameron JD, et  al. Transluminal attenuation
SPECT. Radiology. 2014;270:708–16. gradient in coronary computed tomography angiography is a novel
28. Weininger M, Schoepf UJ, Ramachandra A, et  al. Adenosine- noninvasive apporach to the identification of functionally signifi-
stress dynamic real-time myocardial perfusion CT and adenos- cant coronary artery stenosis: a comparison with fractional flow
ine-stress first-pass dual-energy myocardial perfusion CT for reserve. J Am Coll Cardiol. 2013;61:1271–9.
the assessment of acute chest pain: initial results. Eur J Radiol. 45. Kikuchi Y, Oyama-Manabe N, Naya M, et  al. Quantification of
2012;81:3703–10. myocardial blood flow using dynamic 320-row multi-detector CT as
29. Schwarz F, Nance JW Jr, Ruzsics B, Bastarrika G, Sterzik A, compared with (1)(5)O-H(2)O PET. Eur Radiol. 2014;24:1547–56.
Schoepf UJ.  Quantification of coronary artery calcium on the 46. Hsiao EM, Rybicki FJ, Steigner M.  CT coronary angiography:
basis of dual-­ energy coronary CT angiography. Radiology. 256-slice and 320-detector row scanners. Curr Cardiol Rep.
2012;264:700–7. 2010;12:68–75.
7  Principles of Cardiac CT Image Acquisition 85

47. Kim R, Park EA, Lee W, Chung JW. Feasibility of 320-row area 52. Lee CH, Goo JM, Ye HJ, et  al. Radiation dose modulation tech-
detector CT coronary angiography using 40 mL of contrast mate- niques in the multidetector CT era: from basics to practice.
rial: assessment of image quality and diagnostic accuracy. Eur Radiographics. 2008;28:1451–9.
Radiol. 2016;26(11):3802–10. 53. Choi AD, Leifer ES, Yu J, et al. Prospective evaluation of the influ-
48. Chen MY, Shanbhag SM, Arai AE. Submillisievert median radia- ence of iterative reconstruction on the reproducibility of coronary
tion dose for coronary angiography with a second-­ generation calcium quantification in reduced radiation dose 320 detector row
320-detector row CT scanner in 107 consecutive patients. CT. J Cardiovasc Comput Tomogr. 2016;10:359–63.
Radiology. 2013;267:76–85. 54. Tatsugami F, Higaki T, Fukumoto W, et al. Radiation dose reduc-
49. Wong DT, Soh SY, Ko BS, et al. Superior CT coronary angiography tion for coronary artery calcium scoring at 320-detector CT with
image quality at lower radiation exposure with second generation adaptive iterative dose reduction 3D.  Int J Cardiovasc Imaging.
320-detector row CT in patients with elevated heart rate: a compari- 2015;31:1045–52.
son with first generation 320-detector row CT.  Cardiovasc Diagn 55. Gervaise A, Osemont B, Lecocq S, et al. CT image quality improve-
Ther. 2014;4:299–306. ment using Adaptive Iterative Dose Reduction with wide-volume
50. Tomizawa N, Maeda E, Akahane M, Torigoe R, Kiryu S, Ohtomo acquisition on 320-detector CT. Eur Radiol. 2012;22:295–301.
K.  Coronary CT angiography using the second-generation 56. Yoo RE, Park EA, Lee W, et  al. Image quality of adaptive itera-
320-detector row CT: assessment of image quality and radiation tive dose reduction 3D of coronary CT angiography of 640-slice
dose in various heart rates compared with the first-generation scan- CT: comparison with filtered back-projection. Int J Cardiovasc
ner. Int J Cardiovasc Imaging. 2013;29:1613–8. Imaging. 2013;29:669–76.
51. Lim J, Park EA, Lee W, Shim H, Chung JW.  Image quality and 57. Chen MY, Steigner ML, Leung SW, et al. Simulated 50% radiation
radiation reduction of 320-row area detector CT coronary angiogra- dose reduction in coronary CT angiography using adaptive iterative
phy with optimal tube voltage selection and an automatic exposure dose reduction in three-dimensions (AIDR3D). Int J Cardiovasc
control system: comparison with body mass index-adapted proto- Imaging. 2013;29:1167–75.
col. Int J Cardiovasc Imaging. 2015;31(Suppl 1):23–30.
 Dual Energy and Spectral CT Techniques
in Cardiovascular Imaging 8
B. Krauss and C. H. McCollough

The terms “Dual Energy CT” and “spectral CT” or “spectral atomic numbers from materials with high atomic num-
imaging” [1, 2] refer to X-ray computed tomography scan- bers [7, 8]. In contrast to some other imaging modalities,
ners that image a patient using two or more X-ray spectra so CT is not sensitive to chemical binding but mainly
that differences in the chemical composition of body materi- depends on the total number of electrons in a given vol-
als can be analyzed. At the moment all commercially avail- ume (electron density) and the atomic number of the
able CT scanners of this type work with only two X-ray atoms (effective atomic number). The same applies also
spectra, and therefore the term Dual Energy CT will be used to Dual Energy CT, which also allows for explicitly mea-
for all of them in the following. Several technologies can be suring these two quantities [9, 10]. One of the earliest
used to either generate different spectra at the side of the clinical applications of Dual Energy was related to the
X-ray source or to measure two different spectra at the side detection of calcium. In the 1980s, the first commercially
of the detector [3]. Effectively, all of these technologies gen- available Dual Energy CT scanner was introduced, which
erate similar information, but they differ in terms of dose could measure bone mineral density in the spine with
efficiency, sensitivity to motion, and effectiveness for spe- improved accuracy [11]. However, at that time multi-
cific applications. One of the most important clinical appli- slice spiral CT scanning was not yet invented, limiting
cations of Dual Energy CT is the selective [4], quantitative the clinical capabilities of the technique. Dual Energy CT
[5], and possibly enhanced visualization [6] of iodinated disappeared again from the market, until it was revived in
contrast material, which is enabled by the special X-ray 2005, when the first clinical Dual-Source Dual Energy
attenuation properties of iodine. In contrast to soft body tis- CT scanner was introduced [12].
sues, the X-ray attenuation of iodine decreases strongly with Since then, there has been an increasing interest in this
energy within the spectral range that is commonly used for technology, indicated by the several hundred peer-
clinical CT scanners. This is a consequence of the high reviewed clinical publications on this topic. Each of the
atomic number of iodine (Z = 53), which results in a domi- large manufacturers of CT scanners is currently offering
nance of the photoelectric effect. For water and soft body Dual Energy CT scanners, and many new diagnostic
tissues, the decrease of the attenuation coefficient with applications have been identified [1]. Some of them are
energy is much less pronounced due to a dominance of related to oncological questions [3], but there are also
Compton scattering. Since the definition of the Hounsfield benefits for other fields like urology [13], rheumatology
unit uses water as reference material, the CT number of these [14], and pulmonology [15]. As with single energy CT,
tissues in Hounsfield units (HU) is almost independent of the most of the cardiovascular applications require high-
X-ray spectrum. speed arterial phase imaging and are therefore mainly fea-
Dual Energy CT has a long history and is almost as old sible with CT scanners that have at least 64 data channels
as single energy CT.  In the 1970s, it was demonstrated along the z-axis and rotation times no greater than 0.3 s. A
that Dual Energy CT could be used to make CT images power injector is needed, and for pulmonary CT angiogra-
more quantitative and to differentiate materials with low phy (CTA), it may be advantageous if the injector allows
for several injection phases with different saline admix-
B. Krauss (*) tures within the same bolus [16].
Siemens Healthcare GmbH, Forchheim, Germany
e-mail: [email protected]
C. H. McCollough
Department of Radiology, Mayo Clinic, Rochester, MN, USA

© Humana Press 2019 87

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_8
88 B. Krauss and C. H. McCollough

Dual Energy CT Technology both beams. Data acquisition at both spectra is simultaneous,
as well as continuous, so that motion artifacts can be sup-
Scanner Types pressed by conventional motion correction algorithms [21].
However, there is less freedom in the choice of spectra, and
Dual Energy CT scanning can be realized in two very differ- the delay between the acquisition of data for the same ana-
ent ways, as the measured X-ray spectrum depends on the tomic region with the low and high beam spectra is about one
X-ray spectrum emitted by the X-ray source as well as the rotation. Because of this one rotation delay, slow kV-­
energy-dependent measurement response of the detector. switching scanners and split-filter scanners are in principle
Hence, energy separation can take place either at the side of suitable for Dual Energy CTA of the head-neck area [22] or
the X-ray source or at the side of the detector. Under these of the abdomen. Even pulmonary angiography has been
two main categories exist several subcategories representing shown to be feasible [23], but cardiac scanning is likely not
different technical implementations. feasible. Both methods appear to be most useful for coopera-
Because of the requirement for high scan speed, not all of tive patients and relatively static situations (body parts that
the currently available Dual Energy CT scanners are suitable do not move quickly or iodine concentrations that do not
for cardiovascular imaging. This refers, for example, to the change quickly).
case of Dual Spiral (also called single-source helical) Dual The lowest sensitivity to patient motion is obtained with
Energy CT, where two independent spiral scans, each at a Fast kV-switching (also called single-source rapid-­switching)
different X-ray tube voltage, are performed one after the [24], Dual-Layer [25], and Dual-Source [26] Dual Energy
other [17]. In this case, the time delay between measuring CT technologies. With Fast kV-switching, the tube voltage is
the same anatomic region with the two different spectra is in switched after each acquisition of an individual “frame” of
the order of several seconds, so that arterial phase iodine the X-ray projection data. Switching speed is reported to be
concentrations in the vessels change considerably. For non-­ below one millisecond so that corresponding projection data
contrast Dual Spiral scans, effects from patient motion can frames obtained with the two spectra are well aligned. In
be reduced by appropriate motion correction algorithms. addition, gantry rotation times of 0.28 s have been achieved.
This is not possible, however, if CT numbers change due to Possible disadvantages of the technology are a reduced
changing concentrations of the iodinated contrast material. frame rate (number of projections acquired per unit time) at
Under these conditions, an approach that scans a non-­contrast each energy, finite switching time which reduces spectral
phase and a contrast-enhanced phase, and subtracts the two separation, and limitations concerning the combination of
resulting image sets, may be more promising for measuring this technique with fast and automatic tube current
or isolating iodine enhancement [18]. modulation.
Lower sensitivity to patient motion or contrast agent Dual-Layer detectors have the advantage that no special
dynamics is realized with slow kV-switching (also called technology is needed on the side of the X-ray tube.
single-source sequential) Dual Energy CT scanners [19] and Conventional tube voltages such as 120 kV or 140 kV can be
split-filter (also called single-source twin beam) Dual Energy used, and the fairly wide spectrum is decomposed into two
CT scanners [20]. For slow kV-switching, data are acquired effective energy spectra by the detector. Technically this is
during one gantry rotation using one tube voltage, and for the achieved by using a detector consisting of two layers of scin-
following gantry rotation, data are acquired at a different tillators. While low-energy photons are typically stopped in
tube voltage. Since the minimum amount of projection data the top layer, high-energy photons are typically stopped in
that are needed for image reconstruction of the scanned anat- the bottom layer. Due to the statistical nature of photon
omy is less than one rotation (i.e., 180° + fan angle), these absorption in the scintillator material, energy discrimination
data can be acquired during a portion of the first gantry rota- in such a setup is less than for other methods [27].
tion, and the remaining time in the rotation can be used to Dual-Source Dual Energy configurations have the advan-
switch the tube voltage and tube current before the next rota- tage that no modification is needed for the X-ray tube or for
tion. This can be done in a sequential or spiral acquisition the detector. In order to increase spectral separation, the
mode; in both cases there is a delay of approximately one available X-ray spectra are modified by applying additional
rotation between the projection data acquired with different filtration to the high kV spectra. Because the two tube/detec-
spectra for the same anatomic region. tor systems are separate, the tube potential and tube current
For split-filter scanners, two different filter materials sep- settings, and also tube current modulation, are independent.
arate the fan beam into two different spectra so that half of While spectral quality and the potential for low radiation
the detector (along the z-axis) is exposed to one spectrum dose [28] are the strengths of this technique, fast and sub-
and the other half to the other spectrum. This acquisition stantial motion can lead to false iodine enhancement at ­air/
mode is only possible with spiral CT scanning, when the tissue boundaries because of lower temporal coherence (as
same anatomic region of the patient is translated through will be discussed below).
8  Dual Energy and Spectral CT Techniques in Cardiovascular Imaging 89

All three of the discussed techniques are suitable for vas- tion. The effect of reduced resolution in combination with
cular imaging in all contrast phases and also pulmonary realistic noise is shown in Fig. 8.1c.
angiography. For cardiac imaging at higher heart rates, there Good temporal resolution and coherence are essential for
are limitations with all three methods because of potentially cardiac imaging and have already been mentioned in the con-
rapid motion during data acquisition. As a conclusion, not all text of different scanner technologies. Typically, temporal
Dual Energy CT scanners can be used for vascular imaging, resolution is defined by the time interval that is needed in
and even fewer can be used for cardiac imaging. order to collect all projection data for the reconstruction of at
least one image. On single-source systems and especially in
cardiac mode, it is assumed to be gantry rotation time divided
Special Requirements for Cardiovascular by two. If some structure moves faster than this, motion arti-
Imaging facts will be present in both projection data sets, so that ves-
sel blurring or false stenosis may be seen on the resulting
For cardiovascular Dual Energy CT, strong iodine enhance- images. For cardiac scanning the three fastest Dual Energy
ment is required in the CT images so that vessel anatomy can methods have similar temporal resolution, with Dual-Source
be assessed. This means that the average photon energy of Dual Energy being the fastest with its 0.25 s gantry rotation
the applied X-ray spectra should be rather low, which is in time. The effect of low temporal resolution on spectral data
conflict with the idea that a wide spectral separation is is schematically shown in Fig. 8.1d.
needed for best post-processing results. Typically, the noise-­ Temporal coherence denotes the time over which motion
optimized weighted average of the low- and high-energy leads to a displacement between object contours in the low-
images has similar iodine enhancement as a conventional and high-energy projection data set. Here the Dual-Source
120 kV image, but techniques like iterative reconstruction or Dual Energy scan mode has a temporal coherence that is
nonlinear image filter algorithms can be used to further about a factor of 2 better than its temporal resolution, while
increase the iodine contrast-to-noise ratio. Fast kV-switching and Dual-Layer detectors have almost
In spite of good iodine contrast, there are other effects that perfect temporal coherence. The occurrence of motion arti-
can impede the spectral evaluation of small vessels. In order facts depends on both time components, temporal resolution
to demonstrate this, the cross-sectional view of a partly calci- and coherence. As a consequence, the visual appearance of
fied vessel will be considered. Under ideal conditions it motion artifacts may differ between the three CT scanner
looks like as shown in Fig. 8.1a. types, but all of them are sensitive to motion that occurs on a
One important effect is spatial resolution of the images time scale of less than 150 ms. The combined effect of fast
(Fig. 8.1b), which must be sufficient to see the vascular anat- vessel motion and non-zero temporal coherence is schemati-
omy of interest [29]. Spatial resolution depends on the size cally shown in Fig. 8.1e.
of the detector pixels, the use of flying focal spot techniques Another criterion to be considered is spectral resolution
[30], and the reconstruction kernel. A typical clinical case is [33]: in order to separate, e.g., calcium from iodine with high
the question of in-stent restenosis in coronary arteries, where spatial resolution, the original spatial resolution of the low-
the typical diameter of the vessel is less than 5 mm. In this and high-energy images is not the only concern. For most
situation, improved spatial resolution helps to reduce the size body materials, low- and high-energy images only show
of blooming artifacts, which are caused by the high CT num- small differences. For Dual Energy CT, these small differ-
bers of metal [31]. If spatial resolution is not sufficient to ences have to be amplified, and this leads to an enormous
separate the plaque from the blooming, visualization of ves- increase in image noise if, e.g., iodine enhancement is to be
sel anatomy cannot really be improved by using spectral selectively visualized. There are different strategies to reduce
information. Similar problems arise when plaque composi- noise: radiation dose can be increased – which may not be
tion is assessed on contrast-enhanced images; for both spec- appropriate for each patient – or spectral resolution can be
tra the neighboring dense iodine in the vessel may increased. In both cases, dedicated denoising algorithms are
contaminate the measured attenuation values of the plaque also needed. However, for very small structures, the effec-
due to its close proximity [32]. In general, a quantitative tiveness of such algorithms may be limited. The effect of
interpretation of CT numbers often requires structures that increased spectral resolution in the absence of motion is
are considerably larger than the nominal spatial resolution of shown in Fig.  8.1f, which can be compared directly to
a CT scanner, which is usually assessed based on the visibil- Fig. 8.1c.
ity of high-contrast features. Moreover, depending on the In addition to these properties of Dual Energy technolo-
technique, the Dual Energy or spectral scan mode may offer gies, there are several effects in cardiac imaging that can
poorer spatial resolution than the corresponding single decrease image quality, although they may be difficult to
energy scan mode. In reality, one additional complication is avoid [34]. The first limitation arises as a consequence of
that there is image noise in addition to finite spatial resolu- scattered radiation: all multi-slice CT scanners that are fast
90 B. Krauss and C. H. McCollough

a b
800 800
CT-value (HU)

600 600

CT-value (HU)
400 400

200 200

0 0

c 800 d 800

600 600
CT-value (HU)

CT-value (HU)
400 400

200 200

0 0
e f
800 800

600 600
CT-value (HU)

CT-value (HU)

400 400

200 200

0 0

position (mm) position (mm)

Fig. 8.1  Schematic cross-section of a cardiac vessel that contains a resolution; (c) finite spatial resolution and image noise; (d) finite spatial
calcified plaque (high CT values on the left side) as well as highly con- resolution, image noise, and patient motion (high temporal coherence);
centrated iodine (plateau on the right side). The low-energy profile is (e) finite spatial resolution, image noise, and patient motion (lower tem-
shown in red; the high-energy profile is shown in blue. (a) Ultrahigh poral coherence); (f) finite spatial resolution, image noise, and improved
spatial resolution at infinite radiation dose; (b) effect of finite spatial spectral resolution

enough for cardiac imaging either use a wide collimation or cles. Beam-hardening artifacts may either be reduced by
a Dual-Source configuration. In both cases the amount of projection data Dual Energy processing [37] or by projection
scattered radiation relative to transmitted radiation increases data-based iterative beam-hardening correction [38]. Strong
and can lead to a potential distortion of CT numbers in both iodine enhancement of ventricles and vessels may also cause
spectral channels. This effect is more prominent if only half cone-beam artifacts related to these structures.
a rotation of raw data is collected instead of a full rotation, as Finally there may be step artifacts, if the detector is not
is commonly done for cardiac CT [35]. Strategies against wide enough to cover the heart in a single heartbeat. All
scattered radiation include improved anti-scatter grids as effects together imply that high-quality quantitative CT
well as the subtraction of a simulated or measured scatter imaging of the beating heart is a much more challenging task
signal [36]. than Dual Energy imaging of the abdomen.
In addition, beam-hardening artifacts may originate from
the large amount of dense contrast material inside the ven-
tricles, which may exceed an enhancement-length product of Data and Image Processing
5000 HU-cm along certain directions. Apart from the
enhancement-length product, the strength of the observed Dual Energy CT data contain a wealth of additional informa-
artifacts also depends on the tube voltage and prefiltration tion, and the analysis of this information can be performed in
and thus on the scanner type. The generated artifacts can be different ways. The traditional approach starts with a
especially problematic for the detection of iodine in the myo- ­projection data-based base material decomposition that con-
cardium and the assessment of blood clots inside the ventri- verts the measured X-ray attenuations at low and high energy
8  Dual Energy and Spectral CT Techniques in Cardiovascular Imaging 91

into base material column densities from which material time, the contrast material concentration in mg/ml or the
density images are reconstructed. Typical base materials for iodine enhancement in HU can be calculated.
this approach are aluminum and Lucite or iodine and water. One feature of base material decomposition is that the
Because of the dominance of the photoelectric effect and the user has to make assumptions about the chemical composi-
Compton effect, all naturally occurring body materials will tion of the body materials that are present in the image. For
receive meaningful equivalent water and iodine densities [7]. example, a three-material decomposition into air, water, and
However, contrast agents with high effective atomic num- iodine will show false-negative enhancement in body fat,
ber – like gadolinium – will not be handled consistently in while a three-material decomposition into fat, soft tissue, and
this approach. The generated base material images can be iodine will be corrected in fat and liver tissue but not in the
used to perform further material differentiation and quantifi- lungs. It is therefore important to choose the correct material
cation or to calculate beam-hardening free CT images decomposition or post-processing application for the clinical
(monoenergetic images). Alternatively, images can first be question that has to be answered. The image obtained after
corrected for beam-hardening using an iterative projection iodine subtraction is commonly known as “virtual non-­
data-based method and the following analysis performed in contrast” image, since it shows similar CT numbers as a true
CT number space. In both cases, image-based analysis can non-contrast scan. The deviation between the two images
be used once the beam-hardening effects are removed. depends on the calibration of the scanner and the post-­
In image space, the concept of base material decomposi- processing software and can be in the order of ±10 Hounsfield
tion can be understood by considering a CT number diagram units. This has to be compared with deviations between dif-
(Fig. 8.2) that plots the low-energy CT number against the ferent scanners and scan modes in single energy mode,
high-energy CT number. Different body materials are located which may also be in the order of several HU [39].
at different positions in this diagram. Most soft body tissues Base material decomposition is also the first step for the
are located on one of two lines: the air-water line or the fat-­ calculation of monoenergetic or monochromatic images.
soft tissue line. Adding iodine or calcium to soft body tissues These images show the same CT numbers as images that
simply shifts their CT values in a characteristic direction. If have been acquired with a real monochromatic X-ray source
this direction is known, it is mathematically possible to such as a synchrotron. After measuring the equivalent water
reverse this shift by means of simple geometrical projection. and iodine concentrations with Dual Energy CT, the t­ abulated
In this way the CT numbers of the underlying body material iodine attenuation for each X-ray energy can be used to
can be obtained from the measured CT numbers. At the same derive the monoenergetic CT number at the selected energy.

Fig. 8.2  Base material 400

decomposition (left) and soft tissues blood
material classification or
350 iodine (direction) iodine
labeling (right). In case of 250
base material decomposition, calcium
the measured attenuation is 300 separation line
Measured
projected to the soft tissue
Attenuation
line in order to obtain the
virtual non-contrast image 250
200
low energy CT-value (HU)

low energy CT-value (HU)

and the iodine concentration.

In case of material
200 Measured
classification, the measured
CT value is above the Attenuation
separation line (blue dot) so 150
that the voxel is classified as 150
iodine
100

50
100

0 After
Subtraction
-50
50

-100
-100 0 100 200 50 100 150
high energy CT-value (HU) high energy CT-value (HU)
92 B. Krauss and C. H. McCollough

Due to the dominance of photoelectric effect and Compton single energy CT images can be obtained with no loss in
effect, this works not only for mixtures of these base materi- information from Dual Energy CT examinations. At least in
als but also for all other body tissues. the absence of major beam-hardening effects by bone or
An alternative method for the analysis of Dual Energy iodine, this is indeed the case. Virtual monoenergetic images
results is material classification or material labeling. In this that are generated without additional denoising algorithms
case two or more body materials have to be distinguished, have been shown to have the same CT number values as sin-
which are embedded in the same body tissue. The in vivo dif- gle energy CT images for naturally occurring body materials
ferentiation of various kidney stone types is a typical applica- as long as the correct virtual monoenergetic energy level is
tion of Dual Energy CT, but also the differentiation between selected [45]. The same image can also be generated by cal-
calcified plaques and contrast agent belongs into this category. culating the appropriately weighted average of the original
Separation between the various materials is typically achieved low- and high-energy images.
by using separation lines in the CT value diagram (Fig. 8.2). In Thus, two CT data sets, acquired with different spectra
the right side of Fig. 8.2, calcifications are located below the and reduced radiation dose levels relative to a standard clinic
separation line, while iodine contrast material is located above CT exam, can be combined into one conventional CT image
the separation line. For low concentrations of iodine and cal- that uses all of the applied dose, which is typically set at the
cium, it is not possible to distinguish the two materials because level of a standard clinical exam. As noted as early as 1979
of the small distance between the two lines. [46], it is in general not true that radiation dose must increase
with the usage of Dual Energy techniques to obtain the same
diagnostic value. It is true, however, that the average dose
Denoising efficiency of the low- and the high-energy spectra can be
worse than the dose efficiency of some optimum spectrum in
Image noise is one of the known limitations of Dual Energy the middle. Hence, the dose efficiency of a Dual Energy scan
CT. Even for iodine, the difference between the low-energy mode depends on the selected spectra but also on patient
and the high-energy image is typically much smaller than the size, electronic noise of the detector, and specifications of the
iodine enhancement in the low kV image. So the spectral dif- Dual Energy scan mode, such as a different gantry rotation
ferences to be evaluated are relatively small, and image noise time or a different exposure time per reading.
in virtual non-contrast images is strongly increased relative For example, low noise readout electronics [47] and addi-
to the original images. This means that a straightforward tional filtration for the high-energy spectrum [48] can be
base material decomposition is usually too noisy to be clini- combined, so that similar image noise and iodine contrast
cally useful. Different strategies have been developed to can be generated at the same radiation dose as with a conven-
compensate for this effect. An example is shown in Fig. 8.3. tional 120  kV spectrum [49]. A similar behavior can be
Some methods exploit the anticorrelation [40] of iodine expected from split-filter scanners [19] and scanners with
and water images after base material decomposition to dis- energy-sensitive detectors that are operated at 120 kV [50].
tinguish between image noise and image features. Other For other techniques, differences in dose efficiency may also
researchers have proposed to tailor iterative reconstruction be small depending on protocol optimization and patient
techniques specifically for Dual Energy data [41] or to use diameter. One major consideration is the extent to which the
spectral denoising [42]. Another noise reduction method approach supports automatic tube current modulation and
uses frequency band decomposition to extract the high spa- whether the minimum and maximum dose levels for a given
tial frequencies from the low noise average of the original patient and protocol can actually be achieved. Software mod-
images [43]. For all techniques, it is important to note that ifications such as iterative reconstruction [51] may help to
nonlinear processing can modify image resolution, as well as improve dose efficiency, although the change in diagnostic
noise texture, depending on the image content. On clinical image quality is hard to assess in terms of simple metrics
products, denoising is usually applied as part of the base such as image noise or contrast-to-noise ratio [52].
material decomposition, which may be accomplished in one A very different criterion to measure dose efficiency is
step during image reconstruction or post-processing. the quality of the resultant Dual Energy images. For exam-
ple, the best dose efficiency of weighted average images
may be obtained with one scan mode, but the best quality of
Radiation Dose virtual non-contrast (VNC) images may be obtained with
another scan mode [33]. In general, the use of extra radia-
The question of radiation dose and radiation dose efficiency tion for a Dual Energy CT scan (relative to a single energy
is important for all new techniques in medical imaging, technique) is only justified if the additional information can-
including Dual Energy [44]. First it must be clarified whether, not otherwise be obtained or if competing methods have
for the same applied dose, images that emulate conventional similar risks and benefits.
8  Dual Energy and Spectral CT Techniques in Cardiovascular Imaging 93

a b

c d

Fig. 8.3  Image-based processing was used to generate virtual monoen- noise is substantially reduced by the denoising algorithm for the 40 keV
ergetic images (a and b). These are in comparison to virtual monoener- images (d). This makes, for example, structures in the contrast-­
getic images with an applied denoising algorithm (bottom row). enhanced pancreas clearer
Although there is negligible difference between the 70 keV images (c),

In summary, the dose efficiency of Dual Energy CT depends tion is the automatic removal of the bone from CTA images.
on the comparison that is made: typical approaches include the The iodine and bone can have similar high CT number val-
comparison of single energy and Dual Energy reference proto- ues, but because they are chemically different, they can be
cols on the same scanner, the comparison of protocols on differ- distinguished using Dual Energy CT and a material classi-
ent scanners that are designed to accomplish the same diagnostic fication approach. This method can be compared with sin-
task, or the comparison between Dual Energy and highly opti- gle energy techniques, which try to differentiate between
mized single energy protocols. Because results differ depending the bone and iodine based on anatomical and morphologi-
on the approach, published studies have shown mixed results cal criteria. Both approaches can achieve high-quality
[53–56]. The observed differences between Dual Energy and results, but especially on scanners with good spectral sepa-
single energy dose results also have to be compared with the ration, the Dual Energy method provides superior results
significant dose variations that have been observed between dif- for unusual vessel anatomy or aneurysms close to the bone
ferent single energy scanners for the same diagnostic task [57]. [59]. One feature of Dual Energy bone removal is that hard
plaques and bones are both removed, since both contain
large amounts of calcium. In order to display the calcified
 linical Applications for Cardiovascular
C plaques without displaying the bone, morphological algo-
Imaging rithms are used to ­identify plaques. An example for this is
shown in Fig. 8.4. Also, trabecular bone containing yellow
Vascular Imaging bone marrow requires special processing, as its Dual
Energy behavior is similar to iodinated contrast media in
Dual Energy CT has numerous applications in the field of the blood and Dual Energy techniques alone therefore can-
vascular imaging [58]. Probably the most obvious applica- not remove all of the bone.
94 B. Krauss and C. H. McCollough

a b

c d

Fig. 8.4  Four views of the same patient: original CTA image (maxi- on morphological criteria (MIP, c); combined visualization with differ-
mum intensity projection (MIP), a); after Dual Energy-based calcium ent coloring of bone and vessels (bottom right, d). No manual editing
subtraction (MIP, b); after adding plaques and other calcifications based was applied

However, material classification may not be the best Virtual monoenergetic imaging can be helpful for assess-
approach to remove calcified plaques in small vessels. Due ing the vessel lumen; the CT number values of iodinated
to the finite spatial resolution of CT images, there is a transi- contrast media change more quickly with energy than do the
tion range (one or more voxels) between the plaque and the CT number values of calcium. In this way, the relative con-
vessel lumen (e.g., see Fig.  8.1). Segmentation algorithms trast between the lumen and plaques can be optimized [61].
that try to preserve the vessel lumen will typically differenti- Monoenergetic images can also be used to reduce metal arti-
ate between plaque and lumen at the middle of this transition facts and blooming from stents [62], so that an optimum bal-
region, but this may leave a bright rim surrounding the ance can be found between artifact strength, which decreases
lumen. At the expense of increased image noise, this can be with energy, and iodine contrast-to-noise ratio, which also
avoided by explicitly performing a base material decomposi- decreases with energy, but not necessarily at the same rate.
tion [60]. That is, by decomposing the imaging into the At the lower photon energies, virtual monoenergetic imaging
blood, calcium hydroxyapatite, and iodine, the calcium can also be used to enhance the contrast-to-noise ratio of
hydroxyapatite signal can be subtracted from the image, iodinated vessels. This may be especially helpful for the
which also subtracts all blooming effects. However, since visualization of small vessels on maximum intensity projec-
this approach does not simultaneously increase image reso- tion (MIP) images. Energy levels of 40–55 keV are usually
lution, quantification of the remaining lumen may still be a recommended for this purpose, provided that state-of-the-art
challenge. denoising methods are implemented in the software [63].
8  Dual Energy and Spectral CT Techniques in Cardiovascular Imaging 95

a b

Fig. 8.5  Endoleak detection with Dual-Source Dual Energy. (a) The shown in red and orange colors; the measured iodine density is 7.3 mg/
leakage of contrast agent is seen as an orange area outside the stent ml. (b) At the same position, the VNC image alone does not show
(circled region of interest). In this fused image, the virtual non-contrast increased CT number values; hence, calcium and hemorrhage can be
(VNC) image is displayed in gray, and the iodine base material image is excluded

On scanners with wide spectral separation, similar effects single-phase Dual Energy CT scan achieves very high sensi-
can also be obtained from nonlinear image blending of the tivity and specificity for this diagnostic task [73, 74]. An
low- and high-energy images, where the low-energy image example for this application is shown in Fig. 8.5. In addition,
receives a higher weight in vessels, such that the iodine con- endoleaks may be more conspicuous on low-energy virtual
trast is increased [64–66]. Also, the optimization of linear monoenergetic images [75].
blending can lead to improved contrast-to-noise ratio of The lungs are especially interesting for Dual Energy tech-
iodinated vessels and differentiation of plaque from the niques, since they represent a highly movable and compress-
lumen [67]. ible organ with a strong air/soft tissue contrast and delicate
Analysis of the vessel wall [68] and plaques attached to it structures. One limitation of single energy CT is the diffi-
is a challenging task. Diagnostic tasks such as the detection culty to assess the iodine content of lung vessels below
of intramural hemorrhage [69] are very sensitive to the use of 1–2 mm in diameter because of partial volume effects with
different noise reduction algorithms and may also be difficult the neighboring low-density lung tissue. Dual Energy CT
to answer if the dense contrast agent in the vessel lumen is can be used to visualize the iodine content in such small ves-
not fully subtracted. In this context, it is important to note sels, as it is possible to calculate a ratio which is almost inde-
that the subtraction of iodine enhancement is usually easier pendent of the partial volume effects. This technique has
at low concentrations than at high concentrations, so that been used for the detection of peripheral emboli [76].
VNC images tend to be less reliable in the vessels during Another popular method for the visualization of pulmo-
arterial phase than during portal venous phase [70]. Attempts nary embolism with Dual Energy CT is the measurement of
have been made to look for chemical anomalies, such as the the iodine uptake in the lung parenchyma [77, 78]. After a
presence of iron in soft plaques [71] or iodine uptake in soft good arterial injection, iodine enhancement in the lung
plaques [72]. The results of these studies suggest a possible parenchyma can reach values of more than 30 HU. But with
clinical value, but more research is necessary before such single energy CT, a reduction of iodine uptake is not easily
methods can be used for routine imaging of human patients distinguished from variation in parenchymal density, which
under realistic conditions. can produce effects on a similar scale. With Dual Energy CT,
It has been convincingly demonstrated that patients who the iodine uptake can be selectively visualized so that perfu-
have undergone endovascular repair may benefit from the sion defects can be readily recognized. Depending on the
application of Dual Energy CT techniques. Acute leakage of analysis software, iodine enhancement in the lungs may be
contrast agent after such a procedure (referred to as an displayed in the parenchyma only or together with the
endoleak) is not always easily distinguishable from previous enhancement of the vessels. Results can be compared with
bleedings or calcifications in the same area. With single other methods to quantify lung perfusion such as scintigra-
energy CT, multiphase examinations are usually performed phy, SPECT [79], and MR [80]. Because of the different
including a non-contrast phase, so that false-positive find- contrast agent kinetics, scintigraphy and Dual Energy CT
ings can be avoided. Alternatively, it has been shown that a may not agree in all cases.
96 B. Krauss and C. H. McCollough

Cardiac Applications Dual Energy CT can be used to extract first-pass enhance-

ment of the myocardium [89], as shown in Fig. 8.6. Compared
Dual Energy CT techniques can also be applied to the mov- to other imaging techniques, this method has been
ing heart [81], provided that the scanner is fast enough. ­demonstrated to achieve high sensitivity and specificity for
Similarly to other vessels in the human body, virtual the detection of relevant reduction of myocardial perfusion.
monoenergetic imaging can be used for the coronary arter- In order to stabilize the measured CT number values, full
ies to either optimize iodine contrast-to-noise ratio [82] or rotation reconstructions, instead of half rotation reconstruc-
to reduce calcium blooming and metal artifacts [83–85]. tions, tend to be more suitable for this application.
Vessel segments that are subject to very rapid motion may Dual Energy CT has also been used in stress testing,
have to be excluded from the evaluation in the same way as where patients receive adenosine before the coronary CTA
for single energy CT. There are also sources of metal arti- exam [90]. One side effect of this approach is that – after the
facts, which are only found for cardiac CT scans, such as application of adenosine – the heart rate can increase; con-
valve implants, pacemakers, and sternal wires. Especially sequently it may not be possible to reconstruct motion-free
for the less massive metal objects such as clips and sternal images. Motion will have a bigger effect on the iodine-only
wires, Dual Energy-­based metal artifact reduction has been images than on the low- or high-energy images or weighted
reported to be helpful [86]. CT images. Thus, a failure of the iodine/tissue decomposi-
Other artifact patterns, however, are not completely tion process does not automatically imply a waste of radia-
reduced. For example, projection- as well as image-based tion dose.
Dual Energy CT approaches show limited improvement Due to the absence of dense contrast media in the ventri-
for artifacts associated with dense contrast media in the cle, image quality appears to be better later after the injection
superior vena cava [87]. These artifacts are related to a of iodinated contrast media. In particular, late enhancement
combination of effects, including motion, beam-harden- scans have been proposed, which can highlight contrast
ing, and scattered radiation. By removing only the beam- media in an infarcted area several minutes after application
hardening component, the artifact pattern may not improve of the contrast agent. Promising results have been obtained
substantially. for chronic myocardial infarction [91].
Beam-hardening artifacts are more prominent in other VNC imaging has also been studied to determine whether
situations. In coronary CTA, beam-hardening is often seen VNC images can be used for calcium scoring. This approach
along the axis of the left ventricle, the aorta, and the spine as has been confirmed by several studies [92, 93] and may be
a dark shadow in the soft tissue between these areas. The close to clinical routine.
dark shadow may also extend beyond the tip of the ventricle The analysis of soft cardiac plaque composition is a chal-
into the myocardium. Such a feature can cause difficulties in lenging task. Currently, single energy CT has limitations
the assessment of myocardial enhancement; Dual Energy CT concerning quantitative assessment and correlation with
[88] as well as single energy iterative methods [38] can be clinical symptoms [94], and it is not clear if Dual Energy can
used to improve image quality. Beam-hardening correction add any additional relevant information [95]. At very high
also appears to be relevant when measuring the attenuation spatial resolution, Dual Energy could potentially be useful to
properties of ventricular thrombi. detect diffuse calcium accumulation or iodine uptake inside

Fig. 8.6  Dual Energy CT for

a b
the detection of reduced
myocardial perfusion.
Occluded left anterior
descending coronary artery
(a) and resulting perfusion
defect in the anterior wall and
septum in a patient with acute
myocardial infarction (b). The
reduced iodine uptake (dark
area) is visualized by using
three-material decomposition
(Courtesy of R. Bauer,
University Hospital Frankfurt,
Frankfurt, Germany)
8  Dual Energy and Spectral CT Techniques in Cardiovascular Imaging 97

the plaque [96]. Neither technique has yet been verified Conclusions
in vivo with existing clinical CT scanners.
Finally, the use of Dual Energy CT has been suggested in More and more clinical CT scanners have Dual Energy CT
order to detect chemical anomalies in soft tissues. For exam- capabilities, and routine usage of this technique, as well as
ple, Dual Energy CT can be used for the detection of iron in the number of scientific publications on this topic, is increas-
the myocardium [97]. It may also reveal iodine deposition ing rapidly. At the moment, several technical approaches are
from amiodarone in the liver [98]. Both indications require commercially available that have different strengths and
non-contrast Dual Energy CT scans. weaknesses. Also, post-processing features differ consider-
ably between different software products. However, the need
for a certain degree of standardization has already been rec-
Future Directions: Multi-energy CT ognized by the community [2].
For cardiovascular imaging, the main focus of Dual
For true spectral CT imaging, referred to as multi-energy CT, Energy CT is on the visualization and quantification of iodin-
more than two X-ray spectra have to be measured simultane- ated contrast media in vessels. Virtual monoenergetic images
ously [99]. This is feasible with energy-sensitive detectors, allow for an interactive optimization of iodine contrast rela-
where the individual energy of each X-ray photon can be tive to soft tissue, calcium, or stent material. Virtual monoen-
measured. In this way, X-ray spectroscopy becomes possible, ergetic CT number values are independent of the patient or
which is potentially useful to separate K-edge contrast agents, the prefiltration of the used scanner; they are in this way
such as gadolinium or gold, from iodine and body tissues. more objective and reproducible. Additional diagnostic
Technically, multi-energy CT is realized by using a high information is available through the calculation of iodine
atomic number semiconductor like CdTe or CdZnTe, which maps for the heart and lungs. While some of the discussed
can stop X-ray photons within a short range due to a strong applications are at the very early stage of feasibility studies,
photoelectric effect. A rather large voltage is applied to the others, such as bone removal, virtual monoenergetic imag-
semiconductor so that the generated charges drift toward the ing, or the visualization of iodine uptake in the lung paren-
top and bottom electrodes. In this way, a voltage pulse is cre- chyma, are already now widely used.
ated at the input of the readout electronics, and its height There are also other opportunities for the future. As auto-
correlates with the energy of the original photon. Technically, matic segmentation algorithms will have access to chemical
large photon fluence rates of the order of 109 photons/mm2/ information as well as X-ray attenuation information,
second have to be accurately measured, which has been one increased automation based on Dual Energy data may be pos-
of the biggest obstacles in developing such CT scanners. For sible. In addition, further technological improvements such as
this reason, it is not possible to measure and store the energy better temporal resolution can be expected on the scanner
of each photon. Instead, a number of counters are imple- side, so that new clinical applications may become feasible.
mented in the readout electronics so that each counter counts
the number of photons above a preset energy threshold.
X-ray attenuation is measured as a fractional reduction of the References
photon fluence rate for each energy threshold.
So far, there is no commercially available clinical scanner 1. McCollough CH, Leng S, Yu L, Fletcher JG.  Dual- and multi-­
energy CT: principles, technical approaches, and clinical applica-
of this type. However, prototype installations have been
tions. Radiology. 2015;276(3):637–53.
reported in literature [100, 101], and these have shown prom- 2. Siegel MJ, Kaza RK, Bolus DN, Boll DT, Rofsky NM, De Cecco
ising results for phantoms, animals, and a limited number of CN, Foley WD, Morgan DE, Schoepf UJ, Sahani DV, Shuman
patients. The advantage of such photon-counting detectors is WP, Vrtiska TJ, Yeh BM, Berland LL. White paper of the society
of computed body tomography and magnetic resonance on dual-­
that very good spectral resolution can be combined with per-
energy CT, part 1: technology and terminology. J Comput Assist
fect temporal coherence, which has advantages for vascular Tomogr. 2016;40(6):841–5.
as well as cardiac imaging. 3. Johnson TR.  Dual-energy CT: general principles. AJR Am J
The analysis of multi-energy data with more than two spec- Roentgenol. 2012;199(5 Suppl):S3–8.
4. De Cecco CN, Darnell A, Rengo M, Muscogiuri G, Bellini D,
tral channels is much more involved [27]. If K-edge contrast
Ayuso C, Laghi A. Dual-energy CT: oncologic applications. AJR
agents such as gadolinium or gold are to be used, a corre- Am J Roentgenol. 2012;199(5Suppl):S98–S105.
sponding multi-material decomposition is needed to measure 5. Koonce JD, Vliegenthart R, Schoepf UJ, Schmidt B, Wahlquist
the concentrations of these contrast agents in addition to the AE, Nietert PJ, Bastarrika G, Flohr TG, Meinel FG. Accuracy of
dual-energy computed tomography for the measurement of iodine
CT number values of body tissues or iodine. The efficient cali-
concentration using cardiac CT protocols: validation in a phantom
bration of a projection-based multi-material decomposition is model. Eur Radiol. 2014;24(2):512–8.
not a trivial task. Additionally, noise propagation and system- 6. Caruso D, De Cecco CN, Schoepf UJ, Schaefer AR, Leland PW,
atic uncertainties have to be carefully evaluated and addressed. Johnson D, Laghi A, Hardie AD.  Can dual-energy computed
98 B. Krauss and C. H. McCollough

tomography improve visualization of hypoenhancing liver lesions pulmonary artery imaging with rapid kVp switching for the diag-
in portal venous phase? Assessment of advanced image-based vir- nostic analysis and quantitative evaluation of acute pulmonary
tual monoenergetic images. Clin Imaging. 2016;41:118–24. embolism. Acad Radiol. 2015;22(6):743–51.
7. Alvarez RE, Macovski A.  Energy-selective reconstruc- 23. Zhang D, Li X, Liu B. Objective characterization of GE discovery
tions in X-ray computerized tomography. Phys Med Biol. CT750 HD scanner: gemstone spectral imaging mode. Med Phys.
1976;21(5):733–44. 2011;38(3):1178–88.
8. Hünemohr N, Krauss B, Tremmel C, Ackermann B, Jäkel O, 24. Hidas G, Eliahou R, Duvdevani M, Coulon P, Lemaitre L, Gofrit
Greilich S. Experimental verification of ion stopping power pre- ON, Pode D, Sosna J. Determination of renal stone composition
diction from dual energy CT data in tissue surrogates. Phys Med with dual-energy CT: in vivo analysis and comparison with x-ray
Biol. 2014;59(1):83–96. diffraction. Radiology. 2010;257(2):394–401.
9. Goodsitt MM, Christodoulou EG, Larson SC. Accuracies of the 25. Johnson TR, Krauss B, Sedlmair M, Grasruck M, Bruder H,
synthesized monochromatic CT numbers and effective atomic Morhard D, Fink C, Weckbach S, Lenhard M, Schmidt B, Flohr
numbers obtained with a rapid kVp switching dual energy CT T, Reiser MF, Becker CR. Material differentiation by dual energy
scanner. Med Phys. 2011;38(4):2222–32. CT: initial experience. Eur Radiol. 2007;17(6):1510–7.
10. Kalender WA, Klotz E, Suess C. Vertebral bone mineral analysis: 26. Faby S, Kuchenbecker S, Sawall S, Simons D, Schlemmer HP,
an integrated approach with CT. Radiology. 1987;164(2):419–23. Lell M, Kachelrieß M.  Performance of today's dual energy
11. Flohr TG, McCollough CH, Bruder H, Petersilka M, Gruber K, CT and future multi energy CT in virtual non-contrast imag-
Süss C, Grasruck M, Stierstorfer K, Krauss B, Raupach R, Primak ing and in iodine quantification: a simulation study. Med Phys.
AN, Küttner A, Achenbach S, Becker C, Kopp A, Ohnesorge 2015;42(7):4349–66.
BM.  First performance evaluation of a dual-source CT (DSCT) 27. Siegel MJ, Curtis WA, Ramirez-Giraldo JC.  Effects of dual-­
system. Eur Radiol. 2006;16(2):256–68. energy technique on radiation exposure and image quality in pedi-
12. Stolzmann P, Kozomara M, Chuck N, Müntener M, Leschka S, atric body CT. AJR Am J Roentgenol. 2016;4:1–10. [Epub ahead
Scheffel H, Alkadhi H. In vivo identification of uric acid stones of print].
with dual-energy CT: diagnostic performance evaluation in 28. Ulrich A, Burg MC, Raupach R, Bunck A, Schuelke C, Maintz D,
patients. Abdom Imaging. 2010;35(5):629–35. Heindel W, Seifarth H. Coronary stent imaging with dual-source
13. Choi HK, Burns LC, Shojania K, Koenig N, Reid G, Abufayyah CT: assessment of lumen visibility using different convolution
M, Law G, Kydd AS, Ouellette H, Nicolaou S.  Dual energy kernels and postprocessing filters. Acta Radiol. 2015;56(1):42–50.
CT in gout: a prospective validation study. Ann Rheum Dis. 29. Flohr TG, Stierstorfer K, Ulzheimer S, Bruder H, Primak AN,
2012;71(9):1466–71. McCollough CH. Image reconstruction and image quality evalua-
14. Hoffman EA, Lynch DA, Barr RG, van Beek EJ, Parraga tion for a 64-slice CT scanner with z-flying focal spot. Med Phys.
G. IWPFI investigators. Pulmonary CT and MRI phenotypes that 2005;32(8):2536–47.
help explain chronic pulmonary obstruction disease pathophysiol- 30. Gassenmaier T, Petri N, Allmendinger T, Flohr T, Weng AM,
ogy and outcomes. J Magn Reson Imaging. 2016;43(3):544–57. Kunz AS, Petritsch B, Voelker W, Bley TA.  In vitro compari-
15. Kerl JM, Bauer RW, Renker M, Weber E, Weisser P, Korkusuz H, son of second- and third-generation dual-source CT for coro-
Schell B, Larson MC, Kromen W, Jacobi V, Vogl TJ. Triphasic con- nary stent visualization at different tube potentials. Acad Radiol.
trast injection improves evaluation of dual energy lung perfusion 2016;23(8):961–8.
in pulmonary CT angiography. Eur J Radiol. 2011;80(3):e483–7. 31. Horiguchi J, Fujioka C, Kiguchi M, Shen Y, Althoff CE, Yamamoto
16. Leng S, Shiung M, Ai S, Qu M, Vrtiska TJ, Grant KL, Krauss B, H, Ito K. Soft and intermediate plaques in coronary arteries: how
Schmidt B, Lieske JC, McCollough CH. Feasibility of discrimi- accurately can we measure CT attenuation using 64-MDCT? AJR
nating uric acid from non-uric acid renal stones using consecutive Am J Roentgenol. 2007;189(4):981–8.
spatially registered low- and high-energy scans obtained on a con- 32. Krauss B, Grant KL, Schmidt BT, Flohr TG. The importance of
ventional CT scanner. AJR Am J Roentgenol. 2015;204(1):92–7. spectral separation: an assessment of dual-energy spectral separa-
17. Tomandl BF, Hammen T, Klotz E, Ditt H, Stemper B, Lell tion for quantitative ability and dose efficiency. Investig Radiol.
M. Bone-subtraction CT angiography for the evaluation of intra- 2015;50(2):114–8.
cranial aneurysms. AJNR Am J Neuroradiol. 2006;27(1):55–9. 33. Kalisz K, Buethe J, Saboo SS, Abbara S, Halliburton S, Rajiah
18. Diekhoff T, Ziegeler K, Feist E, Kiefer T, Mews J, Hamm B, P. Artifacts at cardiac CT: physics and solutions. Radiographics.
Hermann KG.  First experience with single-source dual-energy 2016;36(7):2064–83.
computed tomography in six patients with acute arthralgia: a fea- 34. Stenner P, Schmidt B, Bruder H, Allmendinger T, Haberland U,
sibility experiment using joint aspiration as a reference. Skelet Flohr T, Kachelriess M. Partial scan artifact reduction (PSAR) for
Radiol. 2015;44(11):1573–7. the assessment of cardiac perfusion in dynamic phase-correlated
19. Euler A, Parakh A, Falkowski AL, Manneck S, Dashti D, Krauss CT. Med Phys. 2009;36(12):5683–94.
B, Szucs-Farkas Z, Schindera ST. Initial results of a single-source 35. Ohnesorge B, Flohr T, Klingenbeck-Regn K. Efficient object scat-
dual-energy computed tomography technique using a split-filter: ter correction algorithm for third and fourth generation CT scan-
assessment of image quality, radiation dose, and accuracy of ners. Eur Radiol. 1999;9(3):563–9.
dual-energy applications in an in vitro and in vivo study. Investig 36. Carrascosa PM, Cury RC, Deviggiano A, Capunay C, Campisi
Radiol. 2016;51(8):491–8. R, López de Munain M, Vallejos J, Tajer C, Rodriguez-Granillo
20. Sotiras A, Davatzikos C, Paragios N.  Deformable medi- GA. Comparison of myocardial perfusion evaluation with single
cal image registration: a survey. IEEE Trans Med Imaging. versus dual-energy CT and effect of beam-hardening artifacts.
2013;32(7):1153–90. Acad Radiol. 2015;22(5):591–9.
21. Kaemmerer N, Brand M, Hammon M, May M, Wuest W, Krauss 37. Bucher AM, Wichmann JL, Schoepf UJ, Wolla CD, Canstein C,
B, Uder M, Lell MM. Dual-energy computed tomography angiog- McQuiston AD, Krazinski AW, De Cecco CN, Meinel FG, Vogl
raphy of the head and neck with single-source computed tomog- TJ, Geyer LL. Quantitative evaluation of beam-hardening artefact
raphy: a new technical (split filter) approach for bone removal. correction in dual-energy CT myocardial perfusion imaging. Eur
Investig Radiol. 2016;51(10):618–23. Radiol. 2016;26(9):3215–22.
22. Cai XR, Feng YZ, Qiu L, Xian ZH, Yang WC, Mo XK, Wang 38. Birnbaum BA, Hindman N, Lee J, Babb JS.  Multi-detector
XB. Iodine distribution map in dual-energy computed tomography row CT attenuation measurements: assessment of intra- and
8  Dual Energy and Spectral CT Techniques in Cardiovascular Imaging 99

i­nterscanner variability with an anthropomorphic body CT phan- source dual-energy CT angiography with reduced iodine load in
tom. Radiology. 2007;242(1):109–19. patients referred for aortoiliofemoral evaluation before transcath-
39. Kalender WA, Klotz E, Kostaridou L. An algorithm for noise sup- eter aortic valve implantation: impact on image quality and radia-
pression in dual energy CT material density images. IEEE Trans tion dose. Eur Radiol. 2014;24(11):2659–68.
Med Imaging. 1988;7(3):218–24. 55. Takrouri HS, Alnassar MM, Amirabadi A, Babyn PS, Moineddin
40. Dong X, Niu T, Zhu L.  Combined iterative reconstruction and R, Padfield NL, BenDavid G, Doria AS. Metal artifact reduction:
image-domain decomposition for dual energy CT using total-­ added value of rapid-kilovoltage-switching dual-energy CT in
variation regularization. Med Phys. 2014;41(5):051909. relation to single-energy CT in a piglet animal model. AJR Am J
41. Leng S, Yu L, Fletcher JG, McCollough CH. Maximizing iodine Roentgenol. 2015;205(3):W352–9.
contrast-to-noise ratios in abdominal CT imaging through use of 56. Cody DD, Kim HJ, Cagnon CH, Larke FJ, McNitt-Gray MM,
energy domain noise reduction and virtual monoenergetic dual-­ Kruger RL, Flynn MJ, Seibert JA, Judy PF, Wu X.  Normalized
energy CT. Radiology. 2015;276(2):562–70. CT dose index of the CT scanners used in the National Lung
42. Grant KL, Flohr TG, Krauss B, Sedlmair M, Thomas C, Schmidt Screening Trial. AJR Am J Roentgenol. 2010;194(6):1539–46.
B.  Assessment of an advanced image-based technique to calcu- 57. Machida H, Tanaka I, Fukui R, Shen Y, Ishikawa T, Tate E, Ueno
late virtual monoenergetic computed tomographic images from E. Dual-energy spectral CT: various clinical vascular applications.
a dual-energy examination to improve contrast-to-noise ratio in Radiographics. 2016;36(4):1215–32.
examinations using iodinated contrast media. Investig Radiol. 58. Watanabe Y, Uotani K, Nakazawa T, Higashi M, Yamada N, Hori
2014;49(9):586–92. Y, Kanzaki S, Fukuda T, Itoh T, Naito H. Dual-energy direct bone
43. Foley WD, Shuman WP, Siegel MJ, Sahani DV, Boll DT, Bolus removal CT angiography for evaluation of intracranial aneurysm
DN, De Cecco CN, Kaza RK, Morgan DE, Schoepf UJ, Vrtiska or stenosis: comparison with conventional digital subtraction
TJ, Yeh BM, Berland LL.  White paper of the society of com- angiography. Eur Radiol. 2009;19(4):1019–24.
puted body tomography and magnetic resonance on dual-energy 59. Andreini D, Pontone G, Mushtaq S, Bertella E, Conte E,
CT, part 2: radiation dose and iodine sensitivity. J Comput Assist Segurini C, Giovannardi M, Baggiano A, Annoni A, Formenti
Tomogr. 2016;40(6):846–50. A, Petullà M, Beltrama V, Volpato V, Bartorelli AL, Trabattoni
44. Yu L, Leng S, McCollough CH.  Dual-energy CT-based mono- D, Fiorentini C, Pepi M.  Diagnostic accuracy of rapid kilo-
chromatic imaging. AJR Am J Roentgenol. 2012;199(5 volt peak-switching dual-energy CT coronary angiography in
Suppl):S9–S15. patients with a high calcium score. JACC Cardiovasc Imaging.
45. Alvarez R, Seppi E. A comparison of noise and dose in conven- 2015;8(6):746–8.
tional and energy selective computed tomography. IEEE Trans 60. Mannelli L, MacDonald L, Mancini M, Ferguson M, Shuman
Nucl Sci. 1979;NS-26:2853–6. WP, Ragucci M, Monti S, Xu D, Yuan C, Mitsumori LM.  Dual
46. Christe A, Heverhagen J, Ozdoba C, Weisstanner C, Ulzheimer energy computed tomography quantification of carotid plaques
S, Ebner L. CT dose and image quality in the last three scanner calcification: comparison between monochromatic and poly-
generations. World J Radiol. 2013;5(11):421–9. chromatic energies with pathology correlation. Eur Radiol.
47. Primak AN, Giraldo JC, Eusemann CD, Schmidt B, Kantor B, 2015;25(5):1238–46.
Fletcher JG, McCollough CH. Dual-source dual-energy CT with 61. Mangold S, De Cecco CN, Schoepf UJ, Yamada RT, Varga-­
additional tin filtration: dose and image quality evaluation in phan- Szemes A, Stubenrauch AC, Caruso D, Fuller SR, Vogl TJ,
toms and in vivo. AJR Am J Roentgenol. 2010;195(5):1164–74. Nikolaou K, Todoran TM, Wichmann JL.  A noise-optimized
48. Uhrig M, Simons D, Kachelrieß M, Pisana F, Kuchenbecker S, virtual monochromatic reconstruction algorithm improves stent
Schlemmer HP.  Advanced abdominal imaging with dual energy visualization and diagnostic accuracy for detection of in-stent re-­
CT is feasible without increasing radiation dose. Cancer Imaging. stenosis in lower extremity run-off CT angiography. Eur Radiol.
2016;16(1):15. 2016;26(12):4380–9.
49. Gutjahr R, Halaweish AF, Yu Z, Leng S, Yu L, Li Z, Jorgensen 62. Beeres M, Trommer J, Frellesen C, Nour-Eldin NE, Scholtz JE,
SM, Ritman EL, Kappler S, McCollough CH.  Human imaging Herrmann E, Vogl TJ, Wichmann JL.  Evaluation of different
with photon counting-based computed tomography at clinical keV-settings in dual-energy CT angiography of the aorta using
dose levels: contrast-to-noise ratio and cadaver studies. Investig advanced image-based virtual monoenergetic imaging. Int J
Radiol. 2016;51(7):421–9. Cardiovasc Imaging. 2016;32(1):137–44.
50. Ohana M, Labani A, Jeung MY, El Ghannudi S, Gaertner S, Roy 63. Apel A, Fletcher JG, Fidler JL, Hough DM, Yu L, Guimaraes LS,
C.  Iterative reconstruction in single source dual-energy CT pul- Bellemann ME, McCollough CH, Holmes DR 3rd, Eusemann
monary angiography: is it sufficient to achieve a radiation dose CD.  Pilot multi-reader study demonstrating potential for dose
as low as state-of-the-art single-energy CTPA? Eur J Radiol. reduction in dual energy hepatic CT using non-linear blending of
2015;84(11):2314–20. mixed kV image datasets. Eur Radiol. 2011;21(3):644–52.
51. Kofler JM, Yu L, Leng S, Zhang Y, Li Z, Carter RE, McCollough 64. Li S, Wang C, Jiang X, Xu G. Effects of dual-energy CT with non-­
CH.  Assessment of low-contrast resolution for the American linear blending on abdominal CT angiography. Korean J Radiol.
College of Radiology Computed Tomographic Accreditation 2014;15(4):430–8.
Program: what is the impact of iterative reconstruction? J Comput 65. Paul J, Tan MM, Farhang M, Beeres M, Vogl TJ. Dual-energy CT
Assist Tomogr. 2015;39(4):619–23. spectral and energy weighted data sets: carotid stenosis and plaque
52. Schick D, Pratap J. Radiation dose efficiency of dual-energy CT detection. Acad Radiol. 2013;20(9):1144–51.
benchmarked against single-source, kilovoltage-optimized scans. 66. Jorgensen SM, Korinek MJ, Vercnocke AJ, Anderson JL,
Br J Radiol. 2016;89(1058):20150486. Halaweish A, Leng S, McCollough CH, Ritman EL. Arterial wall
53. Purysko AS, Primak AN, Baker ME, Obuchowski NA, Remer perfusion measured with photon counting spectral X-ray CT. Proc
EM, John B, Herts BR. Comparison of radiation dose and image SPIE Int Soc Opt Eng. 2016;28:9967.
quality from single-energy and dual-energy CT examinations in 67. Lemos AA, Pezzullo JC, Fasani P, Gullo M, Giannitto C, Lo Gullo
the same patients screened for hepatocellular carcinoma. Clin R, Biondetti PR. Can the unenhanced phase be eliminated from
Radiol. 2014;69(12):e538–44. dual-phase CT angiography for chest pain? Implications for diag-
54. Dubourg B, Caudron J, Lestrat JP, Bubenheim M, Lefebvre V, nostic accuracy in acute aortic intramural hematoma. AJR Am J
Godin M, Tron C, Eltchaninoff H, Bauer F, Dacher JN.  Single-­ Roentgenol. 2014;203(6):1171–80.
100 B. Krauss and C. H. McCollough

68. Toepker M, Moritz T, Krauss B, Weber M, Euller G, Mang T, Wolf third generation dual-source dual-energy CT: a phantom study.
F, Herold CJ, Ringl H. Virtual non-contrast in second-generation, Eur Radiol. 2016;26(6):1871–8.
dual-energy computed tomography: reliability of attenuation val- 82. Halpern EJ, Halpern DJ, Yanof JH, Amin-Spector S, Fischman
ues. Eur J Radiol. 2012;81(3):e398–405. D, Aviram G, Sosna J.  Is coronary stent assessment improved
69. Wang J, Garg N, Duan X, Liu Y, Leng S, Yu L, Ritman EL, Kantor B, with spectral analysis of dual energy CT? Acad Radiol.
McCollough CH. Quantification of iron in the presence of calcium 2009;16(10):1241–50.
with dual-energy computed tomography (DECT) in an ex  vivo 83. Stehli J, Clerc OF, Fuchs TA, Possner M, Gräni C, Benz DC,
porcine plaque model. Phys Med Biol. 2011;56(22):7305–16. Buechel RR, Kaufmann PA. Impact of monochromatic coronary
70. Shinohara Y, Sakamoto M, Kuya K, Kishimoto J, Iwata N, Ohta Y, computed tomography angiography from single-source dual-­
Fujii S, Watanabe T, Ogawa T. Assessment of carotid plaque com- energy CT on coronary stenosis quantification. J Cardiovasc
position using fast-kV switching dual-energy CT with gemstone Comput Tomogr. 2016;10(2):135–40.
detector: comparison with extracorporeal and virtual histology-­ 84. Secchi F, De Cecco CN, Spearman JV, Silverman JR, Ebersberger
intravascular ultrasound. Neuroradiology. 2015;57(9):889–95. U, Sardanelli F, Schoepf UJ.  Monoenergetic extrapolation of
71. Stolzmann P, Frauenfelder T, Pfammatter T, Peter N, Scheffel cardiac dual energy CT for artifact reduction. Acta Radiol.
H, Lachat M, Schmidt B, Marincek B, Alkadhi H, Schertler 2015;56(4):413–8.
T.  Endoleaks after endovascular abdominal aortic aneurysm 85. Geyer LL, Scherr M, Körner M, Wirth S, Deak P, Reiser MF,
repair: detection with dual-energy dual-source CT.  Radiology. Linsenmaier U.  Imaging of acute pulmonary embolism using a
2008;249(2):682–91. dual energy CT system with rapid kVp switching: initial results.
72. Flors L, Leiva-Salinas C, Norton PT, Patrie JT, Hagspiel Eur J Radiol. 2012;81(12):3711–8.
KD.  Endoleak detection after endovascular repair of thoracic 86. So A, Hsieh J, Imai Y, Narayanan S, Kramer J, Procknow K,
aortic aneurysm using dual-source dual-energy CT: suitable scan- Dutta S, Leipsic J, Min JK, Labounty T, Lee TY.  Prospectively
ning protocols and potential radiation dose reduction. AJR Am J ECG-triggered rapid kV-switching dual-energy CT for quantita-
Roentgenol. 2013;200(2):451–60. tive imaging of myocardial perfusion. JACC Cardiovasc Imaging.
73. Maturen KE, Kaza RK, Liu PS, Quint LE, Khalatbari SH, Platt JF. 2012;5(8):829–36.
“Sweet spot” for endoleak detection: optimizing contrast to noise 87. Vliegenthart R, Pelgrim GJ, Ebersberger U, Rowe GW, Oudkerk
using low keV reconstructions from fast-switch kVp dual-energy M, Schoepf UJ.  Dual-energy CT of the heart. AJR Am J
CT. J Comput Assist Tomogr. 2012;36(1):83–7. Roentgenol. 2012;199(5 Suppl):S54–63.
74. Lee CW, Seo JB, Song JW, Kim MY, Lee HY, Park YS, Chae EJ, 88. Ko SM, Song MG, Chee HK, Hwang HK, Feuchtner GM, Min
Jang YM, Kim N, Krauss B. Evaluation of computer-aided detec- JK.  Diagnostic performance of dual-energy CT stress myocar-
tion and dual energy software in detection of peripheral pulmo- dial perfusion imaging: direct comparison with cardiovascular
nary embolism on dual-energy pulmonary CT angiography. Eur MRI. AJR Am J Roentgenol. 2014;203(6):W605–13.
Radiol. 2011;21(1):54–62. 89. Wichmann JL, Arbaciauskaite R, Kerl JM, Frellesen C, Bodelle B,
75. Pontana F, Faivre JB, Remy-Jardin M, Flohr T, Schmidt B, Lehnert T, Monsefi N, Vogl TJ, Bauer RW. Evaluation of monoen-
Tacelli N, Pansini V, Remy J.  Lung perfusion with dual-energy ergetic late iodine enhancement dual-energy computed tomog-
multidetector-­row CT (MDCT): feasibility for the evaluation of raphy for imaging of chronic myocardial infarction. Eur Radiol.
acute pulmonary embolism in 117 consecutive patients. Acad 2014;24(6):1211–8.
Radiol. 2008;15(12):1494–504. 90. Schwarz F, Nance JW Jr, Ruzsics B, Bastarrika G, Sterzik A,
76. Apfaltrer P, Bachmann V, Meyer M, Henzler T, Barraza JM, Schoepf UJ.  Quantification of coronary artery calcium on the
Gruettner J, Walter T, Schoepf UJ, Schoenberg SO, Fink basis of dual-energy coronary CT angiography. Radiology.
C.  Prognostic value of perfusion defect volume at dual energy 2012;264(3):700–7.
CTA in patients with pulmonary embolism: correlation with CTA 91. Yamada Y, Jinzaki M, Okamura T, Yamada M, Tanami Y, Abe T,
obstruction scores, CT parameters of right ventricular dysfunction Kuribayashi S. Feasibility of coronary artery calcium scoring on
and adverse clinical outcome. Eur J Radiol. 2012;81(11):3592–7. virtual unenhanced images derived from single-source fast kVp-­
77. Yang GF, Yang X, Zhang LJ, Zhu H, Chai X, Hu XB, Hu YX, switching dual-energy coronary CT angiography. J Cardiovasc
Lu GM.  Pulmonary enhancement imaging with dual energy CT Comput Tomogr. 2014;8(5):391–400.
for the detection of pulmonary embolism in a rabbit model: com- 92. Sudarski S, Fink C, Sueselbeck T, Kayed H, Schoenberg SO,
parison to perfusion planar scintigraphy, SPECT and SPECT-CT Borggrefe M, Vliegenthart R, Oudkerk M, Henzler T. Quantitative
modalities. Acad Radiol. 2011;18(5):605–14. analysis of coronary plaque composition by dual-source CT in
78. Zhang LJ, Lu L, Bi J, Jin LX, Chai X, Zhao YE, Chen B, Lu patients with acute non-ST-elevation myocardial infarction com-
GM. Detection of pulmonary embolism comparison between dual pared to patients with stable coronary artery disease correlated
energy CT and MR angiography in a rabbit model. Acad Radiol. with virtual histology intravascular ultrasound. Acad Radiol.
2010;17(12):1550–9. 2013;20(8):995–1003.
79. Danad I, Ó Hartaigh B, Min JK. Dual-energy computed tomog- 93. Henzler T, Porubsky S, Kayed H, Harder N, Krissak UR, Meyer
raphy for detection of coronary artery disease. Expert Rev M, Sueselbeck T, Marx A, Michaely H, Schoepf UJ, Schoenberg
Cardiovasc Ther. 2015;13(12):1345–56. SO, Fink C. Attenuation-based characterization of coronary ath-
80. Raju R, Thompson AG, Lee K, Precious B, Yang TH, Berger A, erosclerotic plaque: comparison of dual source and dual energy
Taylor C, Heilbron B, Nguyen G, Earls J, Min J, Carrascosa P, CT with single-source CT and histopathology. Eur J Radiol.
Murphy D, Hague C, Leipsic JA. Reduced iodine load with CT 2011;80(1):54–9.
coronary angiography using dual-energy imaging: a prospective 94. Barreto M, Schoenhagen P, Nair A, Amatangelo S, Milite M,
randomized trial compared with standard coronary CT angiogra- Obuchowski NA, Lieber ML, Halliburton SS.  Potential of dual-­
phy. J Cardiovasc Comput Tomogr. 2014;8(4):282–8. energy computed tomography to characterize atherosclerotic
81. Mangold S, Cannaó PM, Schoepf UJ, Wichmann JL, Canstein plaque: ex vivo assessment of human coronary arteries in compari-
C, Fuller SR, Muscogiuri G, Varga-Szemes A, Nikolaou K, De son to histology. J Cardiovasc Comput Tomogr. 2008;2(4):234–42.
Cecco CN.  Impact of an advanced image-based monoenergetic 95. Hazirolan T, Akpinar B, Unal S, Gümrük F, Haliloglu M, Alibek
reconstruction algorithm on coronary stent visualization using S.  Value of dual energy computed tomography for detection of
8  Dual Energy and Spectral CT Techniques in Cardiovascular Imaging 101

myocardial iron deposition in Thalassaemia patients: initial expe- of conventional imaging performance in a research whole-body
rience. Eur J Radiol. 2008;68(3):442–5. CT system with a photon-counting detector array. Phys Med Biol.
96. Suzuki H, Imafuji A, Kato M, Omiya H.  Dual energy 2016;61(4):1572–95.
CT assessment of amiodarone induced liver damage. 99. Pourmorteza A, Symons R, Sandfort V, Mallek M, Fuld MK,
SOMATOM Sessions Online. 2014. https://health.sie- Henderson G, Jones EC, Malayeri AA, Folio LR, Bluemke
mens.com/ct_applications/somatomsessions/index.php/ DA. Abdominal imaging with contrast-enhanced photon-counting
dual-energy-ct-assessment-of-amiodarone-induced-liver-damage/. CT: first human experience. Radiology. 2016;279(1):239–45.
97. Schlomka JP, Roessl E, Dorscheid R, Dill S, Martens G, Istel T, 100. Holmes DR 3rd, Fletcher JG, Apel A, Huprich JE, Siddiki H, Hough
Bäumer C, Herrmann C, Steadman R, Zeitler G, Livne A, Proksa DM, Schmidt B, Flohr TG, Robb R, McCollough C, Wittmer M,
R.  Experimental feasibility of multi-energy photon-counting Eusemann C.  Evaluation of non-linear blending in dual-energy
K-edge imaging in pre-clinical computed tomography. Phys Med computed tomography. Eur J Radiol. 2008;68(3):409–13.
Biol. 2008;53(15):4031–47. 101. Zatz LM. The effect of the kVp level on EMI values. Radiology.
98. Yu Z, Leng S, Jorgensen SM, Li Z, Gutjahr R, Chen B, Halaweish 1976;119:683–8.
AF, Kappler S, Yu L, Ritman EL, McCollough CH.  Evaluation
 Drugs in Cardiac CT
9
Sebastian Rogowski, Virginia W. Lesslie,
and Ullrich Ebersberger

In spite of the advances in scanner technology in coronary Role of β-Blockers in Controlling Heart Rate
CT angiography (CCTA), image quality is still influenced by
heart rate (HR) and the consistency of the cardiac cycle. At many centers, clinicians control HR by administering a
Additionally, HR is related to the diagnostic value and level cardio-selective β-adrenergic antagonist [7]. Metoprolol has
of radiation exposure of CCTA [1–3]. become the standard agent (Table  9.1). It is safe to use in
The reason for pharmaceutical intervention in patients’ patients with congestive heart failure as well as in patients
CCTA examinations for the assessment of coronary disease with significant chronic obstructive pulmonary disease. In
(CAD) and myocardial perfusion is the reduction of HR and addition, the agent is low cost and reliable [8]. The most
the instigation of coronary vasodilatation in order to achieve common oral approach involves a dose of 100 mg of meto-
maximal coronary flow. The current guidelines propose a HR prolol. Hence, one possible protocol is to give 100 mg orally
of <65 bpm before imaging [4]. 1 h before the scan (slow-release forms should not be used).
If HR remains ˃60 bpm, additional metoprolol may be given
intravenously to expedite HR reduction. Atenolol may be
 harmacological Interventions to Lower
P chosen for patients with significant hepatic dysfunction
the Heart Rate because of its renal route of clearance.
Beta-blockers have the potential to cause a variety of side
The use of medication to lower HR in cardiac CT is common effects related to their β1-receptor activity. At high one-off
in clinical practice. In many centers, the intravenous (IV) and doses, the risk of β2-receptor-mediated adverse effects
oral administration of β-blockers is used to reduce HR [5, 6]. increases. Potential adverse events include bradycardia, atrio-
However, there is no widely accepted protocol for this prac- ventricular block, hypotension, bronchospasm, Raynaud phe-
tice. Furthermore, the use of non-dihydropyridine calcium nomenon, and anaphylaxis [9]. The half-life of IV metoprolol
channel blockers (CCBs), nitrates, and ivabradine is is approximately 3–7  h; therefore, an adverse effect may
described. debilitate the patient for an extended period of time [10].
In such cases, esmolol may be used as an alternative.
Esmolol is an ultrashort-acting cardio-selective IV β-receptor
blocking agent with rapid onset (within 2–3 min) and ultra-
short duration of action (mean half-life 9 min) [10, 11]. In con-
S. Rogowski trast to the longer-acting β-receptor antagonists, the activity of
Department of Cardiology and Intensive Care Medicine, Heart
Center Munich-Bogenhausen, Munich, Germany
V. W. Lesslie Table 9.1  β-Blockers
Division of Cardiovascular Imaging, Department of Radiology and Administration Drug Dose Onset Half-life
Radiological Science, Medical University of South Carolina, Oral Metoprolol 50–100 mg 1 h 3–7 h
Charleston, SC, USA
Atenolol 50–100 mg 2 h 6–10 h
U. Ebersberger (*) Intravenous Metoprolol 2.5–5.0 mg 1 min 3–4 h
Department of Cardiology and Intensive Care Medicine, Heart (up to 15 mg)
Center Munich-Bogenhausen, Munich, Germany Atenolol 2.5–5.0 mg 1 min 6–10 h
Division of Cardiovascular Imaging, Department of Radiology and (up to 10 mg)
Radiological Science, Medical University of South Carolina, Esmolol 0.5 mg/kg 1–2 min 9 min
Charleston, SC, USA (25–50 mg)
e-mail: [email protected]

© Humana Press 2019 103

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_9
104 S. Rogowski et al.

Table 9.2  Side effects of B-blockers Role of Ivabradine in Controlling Heart Rate
Non-cardiovascular
Drug class Cardiac adverse effects side effects Ivabradine is another attractive option for reducing patient
Ivabradine Dizziness, weakness Eyesight problems: HR for CCTA. It selectively inhibits the If-current, which is
(symptomatic bradycardia) Increased brightness
Phosphenes (bright
a mixed K+-Na+-inward current activated by hyperpolariza-
spots) tion, and contributes to the activity of the sinoatrial node
Blurred vision [15, 16]. It slows the diastolic depolarization phase of the
Metoprolol, Hypotension Fatigue sinoatrial cells and slows the HR at rest and during exercise
esmolol, and Bradycardia Erectile dysfunction without affecting myocardial contractility [17, 18].
other Adrenergic-mediated Insomnia
beta-blockers coronary spasm on sudden Hyperlipidemia Therefore, it can be administered in patients in sinus rhythm
withdrawal Depression but not other rhythms such as atrial fibrillation. Guaricci
Airway disease et  al. showed that the administration of ivabradine over a
Raynaud’s 5-day period prior to CCTA is successful in reducing HR for
phenomenon
CCB Hypotension Constipation
CCTA [19].
Bradycardia Drowsiness A recently published trial with administration of an IV
bolus of ivabradine achieved rapid, safe, and sustained HR
reduction during CCTA, increased procedural convenience,
esmolol is short with a fast onset and short duration of activity,
and reduced radiation exposure when compared to the pla-
which constitutes a major safety factor [12]. Efficient HR con-
cebo [20, 21].
trol is crucial during CCTA; therefore, esmolol might be an
There is not yet a consensus about the protocol for
effective alternative to the standard of care metoprolol. At
ivabradine administration. In a recent study, a larger number
present, esmolol is routinely used in intensive care treatment
of patients achieved a HR under 65 bpm by oral administra-
of acute supraventricular arrhythmias; however, administra-
tion of a 7.5  mg dose of ivabradine in comparison with a
tion before CCTA for HR reduction is an “off-label” indica-
5 mg dose [21].
tion. In an observational study, Degertekin et al. [13] reported
that 65% of patients achieved a HR of <65 bpm using 50 mg
of oral β-blocker (atenolol) combined with 1–2 mg/kg of IV
esmolol (range, 50–300 mg). Pharmacological Interventions
More evidence is needed regarding the safety and effi- for Vasodilatation
ciency of IV esmolol administered in a stepwise bolus proto-
col. Furthermore, no direct comparison of esmolol and Reducing motion artifact alone may not be enough to achieve
metoprolol administration for HR control during CCTA is the best possible image results. The diameter of coronary
available (Table 9.2). arteries can be increased significantly by coronary vasodila-
Active bronchospastic disease is considered a contraindi- tation. Administration of sublingual nitrate can dilate the
cation to treatment with β-blockers. Another medical treat- coronary arteries by approximately 15%. Administration of
ment option for these patients is the use of short-acting CCBs nitrates also increases the visualization of side branches
or ivabradine. However, there is no robust data available in including septal branches.
the literature. Caution is advised in the use of β-blockers in Nitroglycerin leads to the release of nitric oxide, which in
patients with the following: known or suspected sick sinus turn initiates smooth muscle relaxation independent of endo-
syndrome, unexplained pre-syncope or collapse, current use thelial function [22] and reduces the likelihood of coronary
of other antiarrhythmic medications (including but not lim- artery vasospasm. The administration of nitrates was recom-
ited to CCBs, digoxin, or amiodarone), depressed left or mended by the American College of Cardiology/American
right ventricular function, or allergy to β-blockers. Heart Association (ACC/AHA) in 1999 for a reliable assess-
ment of the extent of coronary stenosis [23, 24]. Nitroglycerin
is also used for dilatation of coronary arteries in CCTA
 ole of Calcium Channel Blockers (CCBs)
R examinations, optimizing visualization of the vessel lumen
in Controlling Heart Rate and improving stenosis assessment [25].
Administration in the form of a sublingual spray
When β-blockers are contraindicated, CCBs can be an alter- (0.4–0.8  mg) is the most efficacious method with the
native method to control HR (Fig.  9.1). Phenylalkylamine least amount of side effects. The most common side
non-dihydropyridine, verapamil, and diltiazem are the only effect is dizziness caused by nitroglycerin-induced hypo-
CCBs used for this purpose; however, poor HR control with tension or headaches, which are of short duration and
CCBs and benzodiazepines was reported in a recent trial [14]. self-limiting [9].
9  Drugs in Cardiac CT 105

Fig. 9.1  Heart rate control in

CCTA Heart rate <65 bpm?

Yes No
Proceed with CCTA Contraindication to β-Blockers?

Yes No
Sinus Rhthym?
Oral β-Blocker
(Metoprolol 50-100mg one
hour before scan)
Intravenous β-Blocker
Yes No (Metoprolol 5mg with repetitive
Ivabradin 7.5 mg p.o. No Ivabradin administration or Esmolol
1-2mg/kg

Another point of concern with the administration of nitro- Adenosine has a short half-life (<10  s) and is rapidly
glycerin is the potential decrease in blood pressure or reflex removed from the body upon administration. Thus, it has a
tachycardia, which can result in more motion artifacts on rapid onset and short duration of action. Continuous infu-
CCTA [26]. Furthermore nitroglycerin should never be used sion of 140 mcg/kg/min is required for the purpose of CT
in patients taking erectile dysfunction medications like silde- myocardial perfusion (CT-MPI). At the 3-minute mark, the
nafil or patients with severe anemia. In patients with severe stress radiopharmaceutical is injected and the infusion is
aortic stenosis, nitroglycerin should also be administered continued for 3 additional minutes. Due to the short dura-
with extreme caution. tion of the treatment with adenosine, medication is not
expected to be associated with prolonged side effects
(Tables 9.3, 9.4, and 9.5).
 harmacological Interventions for Stress
P Absolute contraindications for adenosine stress testing
Perfusion Evaluation include asthma, second- or third-degree AV block, sick sinus
syndrome, or the recent use of dipyridamole.
The diagnostic accuracy for CAD is similar for all vasodila-
tors, which have a higher cardiac uptake than catecholamines
and improve the heart/background contrast ratios. In combi-
Dipyridamole
nation with exercise, dobutamine may double coronary per-
fusion as compared to coronary flow at rest. However,
Dipyridamole increases the level of endogenous adenos-
vasodilators may even quadruple the flow.
ine by decreasing the level of reuptake by endothelial
Pharmacological stress agents most extensively validated
cells [28]. Both adenosine and dipyridamole have good
by feasibility trials include adenosine and dipyridamole [27,
sensitivity and specificity for the detection of myocardial
28]. Dipyridamole blocks adenosine reuptake. Adenosine
perfusion defect with stress CT-MPI (Table  9.5). They
dilates coronary arteries by interaction with adenosine A2
have similar side effects, including reflex tachycardia, and
receptors.
are similarly contraindicated in patients with asthma,
COPD, and high-grade atrioventricular block in the
Adenosine absence of a pacemaker. The standard dose of dipyridam-
ole (0.142 mg/kg/min) is infused over 4 min. However, its
The main effect of adenosine consists of a coronary arterio- effects last longer than 4 min. If side effects occur, admin-
lar vasodilation by specific activation of the A2A receptor istering aminophylline  – an adenosine receptor antago-
and leads to a 3.5- to 4-fold increase in myocardial blood nist  – might be necessary to reverse the effects of
flow. If perfusion is limited by coronary stenosis, there is a dipyridamole. Both dipyridamole and adenosine require
diminished hyperemic response. weight-based dosing.
106 S. Rogowski et al.

Table 9.3  Adenosine receptor activation dose, half-life, and effects of medications as well as their
A1 receptor AV block indications, contraindications, and interaction with other
A2B receptor Peripheral vasodilation, bronchospasm medications.
A3 receptor Bronchospasm

Table 9.4  Adenosine side effects References

Symptoms Incidence (%)
1. Stolzmann P, Goetti RP, Maurovich-Horvat P, et  al. Predictors of
Flushing 35–40
image quality in high-pitch coronary CT angiography. AJR Am J
Chest pain* 25–30 Roentgenol. 2011;197:851e858.
Dyspnea 20 2. Matt D, Scheffel H, Leschka S, et  al. Dual-source CT coronary
Dizziness 7 angiography: image quality, mean heart rate, and heart rate vari-
Nausea 5 ability. AJR Am J Roentgenol. 2007;189:567e573.
Symptomatic hypotension 5 3. Dewey M, Zimmermann E, Deissenrieder F, et al. Noninvasive cor-
onary angiography by 320-row computed tomography with lower
Chest pain is nonspecific and is not necessarily indicative of CAD
*
radiation exposure and maintained diagnostic accuracy: compari-
son of results with cardiac catheterization in a head-to-head pilot
Table 9.5  Stress perfusion drug evaluation investigation. Circulation. 2009;120:867e875.
Adenosine Good sensitivity and specificity 4. Abbara S, Arbab-Zadeh A, Callister T, et  al. SCCT guidelines
Dipyridamole Good sensitivity and specificity, inexpensive for performance of coronary computed tomographic angiog-
Regadenoson Easier dosing (10-s bolus), fewer side effects raphy: a report of the Society of Cardiovascular Computed
Tomography Guidelines Committee. J Cardiovasc Comput Tomogr.
Dobutamine Physiological mechanism – increase oxygen
2009;3:190–204.
consumption in the myocardium
5. Groen JM, Greuter MJ, van Ooijen PM, Willems TP, Oudkerk
M. Initial results on visualization of coronary artery stents at mul-
tiple heart rates on a moving heart phantom using 64-MDCT.  J
Regadenoson Comput Assist Tomogr. 2006;30:812–7.
6. Pannu H, Alvarez W, Fishman E. Beta-blockers for cardiac CT: a
Another A2A adenosine receptor agonist is regadenoson. primer for the radiologist. Am J Roentgenol. 2006;186:S341–5.
Similar to adenosine, it induces selective coronary vasodila- 7. Mahabadi AA, Achenbach S, Burgstahler C, et  al. Safety, effi-
cacy, and indications of beta-adrenergic receptor blockade to
tion by binding to A2A adenosine receptors in the walls of
reduce heart rate prior to coronary CT angiography. Radiology.
the heart blood vessels. 2010;257:614e623.
The recommended dose of regadenoson is a single injec- 8. Le Jemtel TH, Padeletti M, Jelic S.  Diagnostic and therapeutic
tion of 400 micrograms (5  mL), without dose adjustment challenges in patients with coexistent chronic obstructive pul-
monary disease and chronic heart failure. J Am Coll Cardiol.
according to the body weight. Regadenoson is injected into a
2007;49:171–82.
peripheral vein for 10 s. The injection should be immediately 9. López-Sendón J, Swedberg K, McMurray J, et al. Expert consen-
followed by a 5 mL saline flush. The same contraindications sus document on beta-adrenergic receptor blockers. Eur Heart J.
must be observed as with adenosine. 2004;25:1341–62.
10. Westfall TC, Westfall DP. Beta adrenergic receptor antagonists. In:
Brunton LL, editor. Goodman & Gilman’s the pharmacological
basis of therapeutics. New York: The McGraw-Hill; 2011.
Dobutamine 11. Wiest D. Esmolol. A review of its therapeutic efficacy and pharma-
cokinetic characteristics. Clin Pharmacokinet. 1995;28:190e202.
12. Wiest DB, Haney JS.  Clinical pharmacokinetics and therapeutic
Another option for the medical stress test is the administra-
efficacy of esmolol. Clin Pharmacokinet. 2012;51:347e356.
tion of a synthetic catecholamine  - dobutamine. It raises 13. Degertekin M, Gemici G, Kaya Z, et  al. Safety and efficacy of
blood pressure, myocardial contractility, and HR by unselec- patient preparation with intravenous esmolol before 64-slice
tive stimulation of the β-receptor. computed tomography coronary angiography. Coron Artery Dis.
2008;19:33e36.
According to protocol, it should be given at a starting
14. Maffei E, Palumbo AA, Martini C, et al. “In-house” pharmacologi-
dose of 5 mcg/kg/min for 3 min with subsequent increases cal management for computed tomography coronary angiography:
(10, 20, 30, and 40 mcg/kg/min) until the stress endpoint is heart rate reduction, timing and safety of different drugs used dur-
achieved. The endpoints are reaching target HR or the expe- ing patient preparation. Eur Radiol. 2009;19:2931e2940.
15. Bois P, Bescond J, Renaudon B, Lenfant J. Mode of action of bra-
rience of chest pain, ECG changes, or hypotension. In case
dycardic agent, S 16257, on ionic currents of rabbit sinoatrial node
of adverse side effects, administration should be stopped cells. Br J Pharmacol. 1996;118:1051–7.
immediately and can be reversed using 0.5  mg/kg of IV 16. DiFrancesco D. Characterization of single pacemaker channels in
esmolol. cardiac sino-atrial node cells. Nature. 1986;324:470–3.
17. Gardiner SM, Kemp PA, March JE, Bennett T. Acute and chronic
The application of medications in CCTA enables demon-
cardiac and regional haemodynamic effects of the novel bra-
strably improved results by means of better image quality dycardic agent, S16257, in conscious rats. Br J Pharmacol.
and dose reduction. Practitioners should be aware of the 1995;115:579–86.
9  Drugs in Cardiac CT 107

18. Ragueneau I, Laveille C, Jochemsen R, Resplandy G, Funck-­

24. Scanlon PJ, Faxon DP, Audet AM, et  al. ACC/AHA guidelines
Brentano C, Jaillon P. Pharmacokinetic-pharmacodynamic model- for coronary angiography. A report of the American College of
ing of the effects of ivabradine, a direct sinus node inhibitor, on heart Cardiology/American Heart Association task force on practice
rate in healthy volunteers. Clin Pharmacol Ther. 1998;64:192–203. guidelines (committee on coronary angiography). Developed
19. Guaricci AI, Schuijf JD, Cademartiri F, et al. Incremental value and in collaboration with the Society for Cardiac Angiography and
safety of oral ivabradine for heart rate reduction in computed tomog- Interventions. J Am Coll Cardiol. 1999;33:1756–824.
raphy coronary angiography. Int J Cardiol. 2012;156(1):28–33. 25. Johnson PT, Eng J, Pannu HK, Fishman EK. 64-MDCT angiogra-
20. Cademartiri F, Garot J, Tendera M, et al. Intravenous ivabradine for phy of the coronary arteries: nationwide survey of patient prepara-
control of heart rate during coronary CT angiography: a random- tion practice. AJR Am J Roentgenol. 2008;190:743–7.
ized, double-blind, placebo-controlled trial. J Cardiovasc Comput 26. Khan M, Cummings KW, Gutierrez FR, Bhalla S, Woodard PK,
Tomogr. 2015;9(4):286–94. Saeed IM.  Contraindications and side effects of commonly used
21. Guaricci AI, et al. Heart rate control with oral ivabradine in com- medications in coronary CT angiography. Int J Cardiovasc Imaging.
puted tomography coronary angiography: a randomized compari- 2011;27:441–9.
son of 7.5 mg vs 5 mg regimen. Int J Cardiol. 2012;168(1):362–8. 27. Blankstein R, Shturman LD, Rogers IS, et al. Adenosine-induced
more. stress myocardial perfusion imaging using dual-source cardiac
22. Ignarro LJ, Napoli C, Loscalzo J. Nitric oxide donors and cardio- computed tomography. J Am Coll Cardiol. 2009;54:1072–84.
vascular agents modulating the bioactivity of nitric oxide: an over- 28. Cury RC, Magalhaes TA, Borges AC, et  al. Dipyridamole stress
view. Circ Res. 2002;90:21–8. and rest myocardial perfusion by 64-detector row computed tomog-
23. Jost S, Rafflenbeul W, Reil GH, et al. Reproducible uniform coro- raphy in patients with suspected coronary artery disease. Am J
nary vasomotor tone with nitrocompounds: prerequisite of quan- Cardiol. 2010;106:310–5.
titative coronary angiographic trials. Catheter Cardiovasc Diagn.
1990;20:168–73.
 Contrast Media Injection Protocols in CT
Coronary Angiography 10
Casper Mihl, Madeleine Kok, Joachim E. Wildberger,
and Marco Das

Optimal contrast enhancement of the coronary arteries is Concentration

of pivotal importance for the proper assessment of coro-
nary artery disease (CAD). Despite the degree of intravas- Different CM concentrations are used in CCTA and vary
cular enhancement, overall image quality in coronary between 240 and 400 mg iodine per millilitre (mgI/ml). In
computed tomography angiography (CCTA) will depend terms of the degree of intravascular enhancement, all con-
on patient- and scan-related parameters. Previous studies centrations can be comparably used in the clinical setting, as
have demonstrated that at a tube voltage of 120  kV long as the iodine delivery rate (IDR) or iodine flux (gI/s)
enhancement levels in the coronary arteries above 325 remains constant.
Hounsfield units (HU) are necessary for optimal diagnosis
[1, 2]. Attenuation values >500 HU may lead to underesti-
mation of the degree of stenosis [3]. This necessitates syn- Flow Rate
chronization of contrast material (CM) administration
protocols with the scanning protocol offering individual Contrast material will be intravenously delivered at a dedi-
timing based on the patient-­ specific situation [4, 5]. cated speed of injection. Flow rates <5 ml/s are most com-
Technological advances in CCTA including shortening of monly used in clinical practice. Higher flow rates (up to
the overall acquisition time and the broad availability of 8.8  ml/s), however, can also be used safely [6]. Injections
advanced reconstruction techniques (iterative reconstruc- with high flow rates lead to a shorter and steeper peak
tion, [IR]) allow utilization of individual low-dose proto- enhancement and thus a proportional increase in vascular
cols without compromising image quality parameters in a and parenchymal enhancement at the expense of a shorter
broad range of patients. time window in which these high enhancement levels are
available [5, 7].

Contrast Material Application

Volume
In order to achieve optimal opacification and contrast
between vessel lumen and surrounding tissue, respectively, The overall iodine volume is the key determinant for solid
high and constant iodine contrast must be maintained organ imaging. In general, shorter acquisition times may to
throughout acquisition of the entire dataset. A state-of-the-­ some extent reduce the total amount of CM required. For
art CM delivery in CCTA consists of several components, CCTA, this aspect is of minor importance. However, fixed
each of which will be addressed separately and linked to a CM volumes show large variation between different weight
concise recipe for robust clinical use. groups. Diagnostic attenuation of the coronary arteries can
be acquired with total injected CM volumes varying between
50  ml and 94  ml (tube voltage: 100  kV, concentration:
300 mgI/ml), depending on weight of the patient (39–109 kg)
C. Mihl (*) · M. Kok · J. E. Wildberger · M. Das and scan duration [8].
Department of Radiology and Nuclear Medicine, Maastricht
University Medical Center, Maastricht, The Netherlands
CARIM School for Cardiovascular Diseases, Maastricht University
Medical Center, Maastricht, The Netherlands

© Humana Press 2019 109

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_10
110 C. Mihl et al.

Saline Chaser Iodine Delivery Rate (IDR)

A saline chaser pushes the tail of the injected CM bolus The IDR is the decisive factor for attenuation of the coronary
into the central blood volume. Otherwise, approximately arteries and can easily be calculated by multiplying the iodine
20–30  ml of CM will be retained in the “dead space” concentration of the CM used (mg/ml) by the flow rate of the
between the brachial vein and the superior vena cava [9]. injector (ml/s). Injection of lower concentrated CM will pro-
Consequently, saline flushing improves arterial enhance- vide identical attenuation of the coronary arteries in compari-
ment and reduces the amount of contrast needed for a diag- son to higher concentrated CM when keeping IDR identical
nostic examination. Therefore, the use of dual-head power [16, 17]. As a rule of thumb, an IDR of 2 gI/s is reasonable for
injectors and a saline chaser is highly recommended. The CCTA applications at 120 kV. This can be done by either using
latter should be injected at the same flow rate as used for a higher flow rate or using a higher concentration of CM. This
the CM bolus. implies a flow rate of approximately 6.7 ml/s for a CM con-
centration of 300 mgI/ml, and a flow rate of 5 ml/s for a CM
concentration of 400 mgI/ml would produce identical attenua-
Preheating tion. Normalizing IDR is a straightforward means of making
different injection protocols comparable. At first glance, it
There is an exponential relationship between iodine seems obvious that the choice of high iodine-concentration
­concentration and viscosity. In general, the higher the contrast agents permits the use of lower flow rates. However,
iodine concentration, the higher the viscosity. High vis- as already outlined, it should be kept in mind that the relation-
cosity increases the friction resistance and the pressure ship between iodine concentration and viscosity is exponen-
needed for the same flow rate [10]. In addition, viscosity tial. It follows that pressure created by the power injector,
is directly influenced by temperature and can be lowered resulting pressures at the site of injection, and subsequent
by preheating [11–13]. Low viscosity can be advanta- pressures inside the patient’s veins can even be higher with
geous in several ways and has a direct influence on differ- high iodine-­concentration agents than with low iodine-con-
ent aspects in CM injection: accelerated CM distribution, centration agents. The benefits of preheated CM (to a mean
sustaining mixing of CM with blood as well as lower body temperature of 37 °C) in order to reduce viscosity are
injection pressure [14, 15]. Therefore, standardized pre- obvious. As technical advances are aimed at scanning with
heating of CM should be a prerequisite for clinical CM lower tube voltages, which leads to an increase in attenuation,
administration (Fig. 10.1). IDR will have to be modified (e.g. lowered) accordingly.

Fig. 10.1  Standardized preheating of CM should be a prerequisite for one CM bottle of up to 500  ml, to prevent temperature drop on site.
clinical CM administration. Left: Heating cabinet for storage of CM Note: the small warming pads regularly delivered with the power injec-
bottles up to 500  ml in the preparation room. CM will be heated to tor are usually not capable of maintaining a constant temperature of
38 °C (100°F), in order to ensure that the CM is injected at body tem- 37 °C (99 °F)
perature. Right: Small incubator in the scanner room, accommodating
10  Contrast Media Injection Protocols in CT Coronary Angiography 111

Fig. 10.2 Comparison Group RCA

between fixed CM injection Cx
(left) and individualized fixed individualized LAD
injection protocol (right)
based on CCTA protocols on
a third-generation dual-source 700
scanner. Two different
protocols are presented with
600
acquisition times of around
1 s and 7 s, respectively.
Error bars show the mean

flash
500
attenuation values with
standard errors for all
coronary arteries in both 400
injection protocols and all
weight classes. Reference
300
line for attenuation was set to
Attenuation (HU)

Scan Protocol
325 HU
200

700

600

adaptive
500

400

300

200

39–59 60–74 75–94 95–109 39–59 60–74 75–94 95–109

Weight Classes (kg)

Total Iodine Load (TIL) circulation of the elbow or the subclavian vein through the
cephalic vein. The clinical effect of different flow rates and
The TIL can easily be calculated by multiplying the iodine con- CM viscosities on the incidence of paravasation is not fully
centration of the CM used (mgI/ml) by the overall volume of understood due to the absence of prospective large popula-
CM to be delivered (ml) and is mostly relevant for organ stud- tion studies. Large-access lines (e.g. 18G needles) are rec-
ies and, as a minor factor, as a weight-related aspect in CCTA. ommended for CCTA.  When experiencing difficulties, due
In summary, all concentrations can be used in the clinical to impairment of peripheral veins (e.g. old patients, patients
setting, as long as the IDR remains constant. However, in terms under chemotherapy or who are receiving corticosteroids),
of injection pressure, the use of lower concentrations is advan- consider using 20G needles and bring down IDR accordingly
tageous as viscosity increases exponential with higher concen- (e.g. 1.5 gI/s max.).
trations of CM. The latter results in higher pressures created by
the power injector despite usage of lower flow rates.
Comfort

Patient Care Results from the EICAR trial nicely show that a broad range
of flow rates, up to 8.3 ml (at 240 mgI/ml) can be applied
Injection Site without an increased risk of extravasation and without the
presence of negative side effects such as discomfort, pain or
The preferred site for CM injection is the right antecubital stress [6]. Injection with higher flow rates should not be con-
vein. The administrated CM will drain into the basilic vein or sidered a drawback in clinical practice. Furthermore, this
a forearm vein that subsequently drains into the deep venous creates a final doorway towards applicability of a broad
112 C. Mihl et al.

a b c

d e f

Fig. 10.3  Example of individualized CM injection protocol tailored to weight, 95 kg; length, 182 cm; BMI, 28.7; total injected CM volume,
patient’s weight. Patient 1 (a–c): body weight, 59 kg; length, 162 cm; 100  ml. Note an equal homogeneous attenuation in RCA in both
BMI, 22.5; total injected CM volume, 71  ml. Patient 2 (d–f): body patients

v­ariety in flow rates and IDR’s and subsequently more tionship between patient weight and scan duration in order to
­individually tailored injection protocols. achieve diagnostic and comparable intravascular enhance-
ment in all patients (Fig. 10.2). For a CCTA acquisition at
100  kV and patients weighing between 39 and 109  kg, a
Weight and Body Mass Index graded IDR between 1.6 and 2.2  gI/s and TIL between 15
and 28 gI are recommended (Fig. 10.3) [8].
Body weight and body mass index (BMI) are known to have
a substantial impact on vascular attenuation and time to peak
in CCTA. A fixed injection protocol causes a large variation Cardiac Output
between different weight groups indicating suboptimal use
of CM in different patient weight groups, ranging from high Physical factors such as cardiac output may cause variability
contrast levels in patients with low body weight (as a result of vascular enhancement as well. The patient’s cardiovascu-
of too much contrast) to low contrast levels in heavy patients lar circulation critically affects CM bolus transit time which
(e.g. not enough contrast) [8]. Individually tailored CM needs to be considered when determining individualized
injections are beneficial in terms of individual adaptation of scan timing [7, 18]. Empiric scan delays can no longer be
the IDR and the overall volume based upon a non-linear rela- recommended. Test bolus technique and bolus tracking
10  Contrast Media Injection Protocols in CT Coronary Angiography 113

t­echnique are both proper techniques for calculation of the c­oncordantly increasing tube-current-time product (mAs),
CM bolus transit time in routine clinical practice and the image noise will increase and invariably the signal-to-noise
total amount of CM which is needed. The use of test bolus ratio (SNR) will decrease as a result of radiation dose reduc-
technique can overcome suboptimal attenuation due to tion. Therefore, low tube voltage settings should only be
timing-­related issues. Alternatively, the scan delay can be used in combination with an adaptation of the tube-current-
determined individually by acquisition of a single-level time product (mAs) to higher levels.
dynamic CT series (automated bolus tracking). In this mode, Practically, when the selection of lower tube voltage is pos-
a pre-­monitoring scan is performed within the target volume sible by the available scanner, the total amount of iodine can
to localize the target vessel for contrast timing; a low-dose be reduced accordingly. In general, 12% reduction of iodine
scanning technique (e.g. 120  kVp; 20  mAseff.) is usually dose is diagnostically acceptable for a 100 kV scan compared
applied for this purpose. A region of interest (ROI) is then to a 120 kV scan. For 80 kV and 70 kV, reduction percentages
placed in the target vessel (ascending aorta), and attenuation of 45% and 56% might be considered, ­respectively [21].
values (in HU) are measured every 2  s during the contrast
injection. When the predefined trigger threshold level is
reached (e.g. 140 HU; range 50–150 HU based on the cur- Reconstruction Technique
rent literature) [7], an automated start of the CT scan is ini-
tialized. In this mode the additional transition delay has to be Newly available reconstruction technologies such as IR
considered, defined as the delay between reaching the thresh- reduce image noise and significantly improve image
old and the start of the actual CT scan (e.g. 6–7 s; usually quality in terms of SNR and contrast-to-noise ratio
ranging between 0 and 18 s). (CNR) compared to routinely used ­filtered back projection

Table 10.1  Checklist for various parameters in CM injection protocol

Scanner in CCTA, based on a tube voltage of 100 kVp
Factors influencing
Bolus Geometry Parameters Range range
CM concentration (mgI/ml) 240–400 Supplier
Flow rate (ml/s) 2.5–8.8 Based on IDR
Bolus geometry is the pattern of enhancement plotted on a
CM concentration
time (s)/attenuation (Hounsfield units [HU]) curve measured Acquisition time
in the ROI [19]. A physiological time enhancement curve Needle size
shows a rapid rise in aortic contrast enhancement followed Needle position
by a very short steady state plateau and a steady decline in Volume (ml) 50–94 Based on TIL
CM concentration
enhancement due to recirculation and hemodynamic pertur- Acquisition time
bation [5, 19]. Bi- and multiphasic protocols aim to create a Saline chaser (ml) ~30–50a volume Flow rate
longer plateau of enhancement during scan time by means of Preheating Body
a higher injection rate at the beginning of the injection and a temperature
lower rate in the second part [20]. Sixty-four-slice systems (37 °C)
IDR (gI/s) 1.6–2.2 Tube voltage
provide a scan time of approximately 7–12  s for a spiral
Patient size/
CCTA. As the acquisition time is rather in the range of 1–5 s weight
for latest equipment, these (complex) injection protocols are TIL (gI) 15–28 Tube voltage
no longer needed and a monophasic injection protocol is Patient size/
appropriate for clinical routine. weight
Injection site Right Patient condition
antecubital vein
Needle size (G) 18 Needle position
Tube Voltage Vessel size
Timing Bolus tracking Cardiac output
Technical developments of CT technique have also made the Test bolus (Ir)regular heart
use of lower tube voltage (kVp) feasible. The use of lower rate
Reconstruction technique IR (preferred) Scanner technique
kVp settings is accompanied by higher contrast enhance-
Tube voltage (kVp) 70–120b Patient size/
ment; this is explained by the fact that lower tube voltage weight
translates into lower effective photon energy (effective pho- Scanner technique
ton energy being approximately half the kVp), bringing the a
Make sure the flow rate is identical to the flow rate used for main bolus
latter closer to the K-edge of iodine (33.2  keV) [21, 22]. b
All values estimated for this CM injection protocol are based on a tube
However, if lower tube voltages will be used without voltage of 100 kVp. For lower kVp settings, reductions of TIL and IDR
can be applied [21]. If so, please go through the checklist (again)
114 C. Mihl et al.

Table 10.2  CM injection parameters based upon body weight adapted 4. Budoff MJ, Dowe D, Jollis JG, Gitter M, Sutherland J, Halamert
CM injection software, currently used in our clinical routine, using E, et  al. Diagnostic performance of 64-multidetector row coro-
second-generation as well as third-generation dual-source CT for nary computed tomographic angiography for evaluation of coro-
­clinical CCTA nary artery stenosis in individuals without known coronary artery
disease: results from the prospective multicenter ACCURACY
Scan CM Saline Flow
(assessment by coronary computed tomographic angiography of
Weight protocol volume TIL flush rate IDR
individuals undergoing invasive coronary angiography) trial. J Am
classes (100 kV) (ml) (gI) (ml) (ml/s) (gI/s)
Coll Cardiol. 2008;52(21):1724–32.
39–59 kg Sequential/ ~60 ~18 30 5.3 1.6 5. Bae KT.  Optimization of contrast enhancement in thoracic
spiral MDCT. Radiol Clin N Am. 2010;48(1):9–29.
High pitch ~50 ~15 6. Kok M, Mihl C, Hendriks BM, Altintas S, Eijsvoogel NG,
60–74 kg Sequential/ ~70 ~21 30 5.6 1.7 Kietselaer BL, et al. Patient comfort during contrast media injection
spiral in coronary computed tomographic angiography using varying con-
High pitch ~60 ~18 trast media concentrations and flow rates: results from the EICAR
75–94 kg Sequential/ ~80 ~24 30 6.7 2.0 trial. Investig Radiol. 2016;51(12):810–5.
spiral 7. Bae KT.  Intravenous contrast medium administration and
High pitch ~70 ~21 scan timing at CT: considerations and approaches. Radiology.
2010;256(1):32–61.
95–109 kg Sequential/ ~90 ~27 30 7.3 2.2
8. Mihl C, Kok M, Altintas S, Kietselaer BL, Turek J, Wildberger
spiral
JE, et  al. Evaluation of individually body weight adapted con-
High pitch ~80 ~24 trast media injection in coronary CT-angiography. Eur J Radiol.
Data from references [8, 27] 2016;85(4):830–6.
9. Schoellnast H, Tillich M, Deutschmann HA, Deutschmann MJ,
Fritz GA, Stessel U, et al. Abdominal multidetector row computed
tomography: reduction of cost and contrast material dose using
(FBP) in contrast-enhanced CT [23–26]. When available, saline flush. J Comput Assist Tomogr. 2003;27(6):847–53.
10. Kok M, Mihl C, Mingels AA, Kietselaer BL, Muhlenbruch
the use of IR and dedicated cardiac reconstruction ker- G, Seehofnerova A, et  al. Influence of contrast media vis-
nels are recommended to ensure diagnostic image quality cosity and temperature on injection pressure in computed
whilst using reduced radiation dose. tomographic angiography: a phantom study. Investig Radiol.
In summary, optimal CM injection during CCTA is key. 2014;49(4):217–23.
11. Halsell RD. Heating contrast media: role in contemporary angiog-
Many parameters influence attenuation patterns and overall raphy. Radiology. 1987;164(1):276–8.
image quality of CCTA.  A thorough understanding of the 12. Halsell RD.  Heating contrast media in a microwave oven.

underlying principles and their interaction is necessary to Radiology. 1987;163(1):279–80.
provide the full picture on the interplay of the various CM 13. Schwab SA, Kuefner MA, Anders K, Adamietz B, Heinrich

MC, Baigger JF, et  al. Peripheral intravenous power injection
aspects, scanner-related factors as well as patient-related of iodinated contrast media: the impact of temperature on maxi-
aspects. The ultimate goal is to create a personalized CM mum injection pressures at different cannula sizes. Acad Radiol.
injection protocol, where all these parameters are tailored 2009;16(12):1502–8.
individually in order to provide adequate diagnostic attenua- 14. Bae KT, Tran HQ, Heiken JP. Uniform vascular contrast enhance-
ment and reduced contrast medium volume achieved by using
tion values and image quality, based on a standardized con- exponentially decelerated contrast material injection method.
cept. Table  10.1 is provided to serve as a checklist, where Radiology. 2004;231(3):732–6.
recommended values and ranges of all parameters stated 15. Muhlenbruch G, Behrendt FF, Eddahabi MA, Knackstedt C, Stanzel
above are briefly addressed. In addition, Table 10.2 states the S, Das M, et al. Which iodine concentration in chest CT? a prospec-
tive study in 300 patients. Eur Radiol. 2008;18(12):2826–32.
CM injection parameters based upon body weight adapted 16. Mihl C, Wildberger JE, Jurencak T, Yanniello MJ, Nijssen EC,
CM injection software, currently used in our clinical routine. Kalafut JF, et al. Intravascular enhancement with identical iodine
delivery rate using different iodine contrast media in a circulation
phantom. Investig Radiol. 2013;48(11):813–8.
17. Mihl C, Kok M, Wildberger JE, Altintas S, Labus D, Nijssen EC,
References et  al. Coronary CT angiography using low concentrated contrast
media injected with high flow rates: feasible in clinical practice.
1. Isogai T, Jinzaki M, Tanami Y, Kusuzaki H, Yamada M, Kuribayashi Eur J Radiol. 2015;84(11):2155–60.
S.  Body weight-tailored contrast material injection protocol for 18. Bae KT, Heiken JP, Brink JA. Aortic and hepatic contrast medium
64-detector row computed tomography coronary angiography. Jpn enhancement at CT.  Part I.  Prediction with a computer model.
J Radiol. 2011;29(1):33–8. Radiology. 1998;207(3):647–55.
2. Cademartiri F, Maffei E, Palumbo AA, Malago R, La Grutta L, 19. Cademartiri F, van der Lugt A, Luccichenti G, Pavone P, Krestin
Meiijboom WB, et al. Influence of intra-coronary enhancement on GP.  Parameters affecting bolus geometry in CTA: a review. J
diagnostic accuracy with 64-slice CT coronary angiography. Eur Comput Assist Tomogr. 2002;26(4):598–607.
Radiol. 2008;18(3):576–83. 20. Fleischmann D, Rubin GD, Bankier AA, Hittmair K.  Improved
3. Fei X, Du X, Yang Q, Shen Y, Li P, Liao J, et al. 64-MDCT coronary uniformity of aortic enhancement with customized contrast
angiography: phantom study of effects of vascular attenuation on medium injection protocols at CT angiography. Radiology.
detection of coronary stenosis. Am J Roentgenol. 2008;191(1):43–9. 2000;214(2):363–71.
10  Contrast Media Injection Protocols in CT Coronary Angiography 115

21. Mahesh M. MDCT physics: the basics: technology, image quality 25. Pontana F, Pagniez J, Duhamel A, Flohr T, Faivre JB, Murphy
and radiation dose. 1st ed. Philadelphia: Lippincott Williams and C, et  al. Reduced-dose low-voltage chest CT angiography with
Wilkins; 2009. sinogram-­ affirmed iterative reconstruction versus standard-dose
22. Brooks RA.  A quantitative theory of the Hounsfield unit and its filtered back projection. Radiology. 2013;267(2):609–18.
application to dual energy scanning. J Comput Assist Tomogr. 26. Wang R, Schoepf UJ, Wu R, Reddy RP, Zhang C, Yu W, et al. Image
1977;1(4):487–93. quality and radiation dose of low dose coronary CT angiography
23. Renker M, Nance JW Jr, Schoepf UJ, O'Brien TX, Zwerner PL, in obese patients: sinogram affirmed iterative reconstruction versus
Meyer M, et al. Evaluation of heavily calcified vessels with coro- filtered back projection. Eur J Radiol. 2012;81(11):3141–5.
nary CT angiography: comparison of iterative and filtered back pro- 27. Seifarth H, Puesken M, Kalafut JF, Wienbeck S, Wessling J, Maintz
jection image reconstruction. Radiology. 2011;260(2):390–9. D, et al. Introduction of an individually optimized protocol for the
24. Wang R, Schoepf UJ, Wu R, Nance JW Jr, Lv B, Yang H, et  al. injection of contrast medium for coronary CT angiography. Eur
Diagnostic accuracy of coronary CT angiography: comparison of Radiol. 2009;19(10):2373–82.
filtered back projection and iterative reconstruction with different
strengths. J Comput Assist Tomogr. 2014;38(2):179–84.
 Cardiovascular CT: Image
Reconstruction 11
Annemarie M. den Harder, Arnold M. R. Schilham,
and Martin J. Willemink

Image reconstruction is the formation of images based on CT Raw Data

projection data. This process has undergone tremendous
improvements in recent years. This chapter focuses on the X-ray photons travel from the CT tube to the detector. Once
different types of image reconstructions. First, the steps they hit the detector, x-ray photons are converted to light
before the actual reconstruction can take place will be photons which are converted to an electrical signal that
explained, including calibration of the CT system, formation scales with the number of light photons and thus the intensity
of raw data, z-interpolation for spiral acquisitions, and ECG of the detected x-rays. Those electrical signals have certain
editing. Subsequently, the most common reconstruction intensities. The detector will receive multiple signals with
techniques filtered back projection (FBP) and iterative recon- different intensities. This can be used to create a signal inten-
struction and their applications in cardiac CT will be sity profile, which displays the number of signals with a spe-
discussed. cific intensity. All signals received from a certain beam angle
form a projection. The signal intensity profiles of each pro-
jection are transformed into x-ray attenuation values. If an
Calibration object is placed between the CT tube and the detector, a
decreased electrical signal is measured, which reflects an
Low-energy x-ray photons attenuate more easily than high-­ increase of the total x-ray attenuation along the line between
energy photons. Therefore, the fraction of low-energy pho- the detector and the tube.
tons decreases more than the fraction of high-energy photons Projection data can be stored in a two-dimensional image,
when passing through different tissues. This results in beam which is called a sinogram because of its sinusoid shape. A
hardening: an increased mean energy of the beam, which can sinogram is a set of projections for certain orientations using
cause artifacts. This is especially the case when dense mate- line integrals (Fig. 11.1). The x-axis represents the distance
rials are imaged (e.g., bone or metal implants). The CT sys- along the projection direction, and the y-axis represents the
tem is calibrated to correct for beam hardening and angle of the projection. The intensity in the sinogram repre-
inhomogeneities based on assumptions of tissue attenuation sents the attenuation values; more attenuation results in
and x-ray photon energy [1, 2]. Inhomogeneities can, for lower signal and vice versa. The collection of projection data
example, be caused by a defective detector element or low in a sinogram is called the raw data set, which can be used to
attenuation in air. Calibration phantoms with known attenua- reconstruct an image.
tion profiles are used to compare known attenuation proper-
ties to acquired attenuations.
Z-Interpolation

In spiral CT acquisitions, the x-ray beam rotates continu-

ously with simultaneous movement of the table. The table
movement per rotation divided by collimator width is defined
as the pitch. A pitch below 1 results in beam overlap, while a
pitch above 1 results in gaps (Fig. 11.2). In cases of the latter,
A. M. den Harder · A. M. R. Schilham · M. J. Willemink (*)
Department of Radiology, University Medical Center Utrecht, data from different projections have to be interpolated to
Utrecht, The Netherlands form transverse slices of projection data. Interpolation is the

© Humana Press 2019 117

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_11
118 A. M. den Harder et al.

computation of an unknown value by using known values on plane. It also offers the possibility to increase the pitch to 2
either side. The distance between the values on either side is without sampling gaps. However, images are noisier com-
the z-spacing. A higher pitch results in an increased distance, pared to 360° interpolation [4].
which can result in interpolation artifacts [3]. Values can be
separated by a full rotation (360° interpolation) or a half
rotation (180° interpolation). 180 degrees interpolation can ECG Editing
be performed by synthesizing data using projections from
opposite directions. Linear interpolation of 180° is often pre- With retrospective electrocardiography (ECG)-gated acquisi-
ferred, since this will result in improved resolution and thin- tions, images can be reconstructed at different time intervals.
ner slices because the values used are closer to the desired R-waves are automatically detected, and images can be
reconstructed throughout the entire cardiac cycle, usually at
an increment of 5–10%. The most optimal reconstruction
phase can be selected in the post-processing phase. Although
usually the end-diastolic phase is used for reconstruction
(approximately 75% of the R-R interval), occasionally a dif-
ferent phase might result in better image quality, for example,
in patients with atrial fibrillation [5]. The most optimal phase
can be selected manually by evaluating reconstructions of all
R-R intervals; however this is a very time-­consuming pro-
cess. To overcome this drawback, algorithms are available
which automatically reconstruct the entire cardiac cycle and
Angle of the projection

select the best reconstruction based on motion [6].

Due to irregular or high heart rates, the detection of R
waves might be incorrect, leading to artifacts, or the tempo-
ral window can be too short leading to a lack of data. As
illustrated in Fig. 11.3, the affected temporal window can be
deleted, and additional windows can be added in the next
heartbeat. If there are long pauses, additional windows can
also be added [7].

Back Projection

Images are created from projection data using a technique

called back projection. Every projection contains the sum
Distance of the single attenuation values of each pixel along the line
of projection. The total sum of attenuation values is known;
Fig. 11.1  Sinogram of a cardiac CT acquisition
however the exact positions of those values along the

Fig. 11.2 Schematic
a b
illustration of z-interpolation.
An acquisition with a pitch of
1 (a) and a pitch of 2 (b) is
shown. Using 360°
interpolation, a pitch of 2 will
result in sampling gaps. With
180° interpolation it is still Z Z
possible to use a pitch of 2
without sampling gaps,
because the spaces between
the rotations are filled by
synthesizing data using
projections from opposite
directions
Pitch = 1 Pitch = 2
11  Cardiovascular CT: Image Reconstruction 119

Fig. 11.3 Schematic a Correct phases

illustration of ECG editing.
The first ECG signal (a)
represents a normal sinus
rhythm resulting in correct
phase detection. The second
ECG signal (b) is arrhythmic
resulting in incorrect phase
detections. With ECG editing
the wrongly detected phases
can be deleted (c), and images
are based on correct phases
b Correct phases

Incorrect phases

c Correct phases

p­ rojection are unknown. To compute an image from those Convolution Kernels

values, the total sum of attenuations is averaged over the
different pixels on the projection line. This is explained in Because back projection alone results in blurred images, a
Fig. 11.4. In Fig. 11.4a, an image is shown with four pixels convolution kernel is applied. The kernel is also called a fil-
(a). The number in the left upper matrix represents the ter. Convolution takes the values of neighboring pixels into
attenuation value of a pixel. P1 to P4 show the total mea- account with certain weights based on the distance.
sured attenuation values from different projections acquired Depending on the kernel, this can result in a smooth or a
from the true object. It is possible to get an approximation sharp image and everything in between (Fig. 11.5). All man-
of the numbers based on the total sum of attenuation values. ufacturers developed their own kernels with different names.
To this aim, the values are smeared out across the projec- After the convolution kernel is applied to the projection data,
tion, and the different projections are summed and renor- back projection is performed to create an image.
malized by dividing the total amount by the number of The choice of convolution kernel will influence the
projections (in this example 4). Increasing the number of amount of noise and the spatial resolution. A smooth kernel
projections will lead to a better approximation of the true will result in reduced noise at the cost of spatial resolution,
object. However, even with hundreds of projections, the while a sharper kernel results in improved spatial resolution
image will still remain an approximation, and with simple at the cost of increased noise. In cardiac CT this can, for
back projection, the final image will be blurred (Fig. 11.4b). example, influence quantification of coronary plaques [8].
A convolution kernel is necessary for mathematical filter- Sharper kernels will reduce blooming and are therefore pre-
ing to improve the approximation. ferred when stents are present or for heavily calcified ­arteries.
120 A. M. den Harder et al.

True object
a b
2 8 10 5 5
P1 P1 P1 P1
6 3 9 4.5 4.5

P4 P2
2 8 P3
2.5 8
14 5
P2 P2
3 6 P2
P4 6 2.5
8 11

P3

3.4 6.4 13.5 25.5 2 7 4 5.5

P3 P3
5.4 3.9 21.5 15.5 7 3 4 5.5

Divide by 4 P1+P2+P3+P4 P4
Σ Pn P4

Fig. 11.4  Back projection schematically explained (a). The upper left because these are projections of a single pixel. Subsequently the projec-
matrix indicates the measured attenuation values of the true object. In tions are summed. Finally the numbers are divided by the number of
this example there are four pixels, and four projections are acquired at projections, which is 4 in this case. The final reconstruction is shown in
different angles (P1 to P4). The first projection (P1) shows the measured the left lower corner. Increasing numbers of projections improve the
attenuation from left to right, and the other projections (P2 to P4) are quality of the approximated final reconstruction. However, even with a
rotated by 45°. The values are smeared out across the projection. For large number of projections, the reconstructed image will be blurred
example, the 10 in P1 is divided by two (number of pixels); therefore a with simple back projection (b)
5 is allocated to each pixel. While in P2 the 8 and 6 are not divided

Fig. 11.5  Example of Original profile Convolution kernel Filtered profile

different convolution kernels.
A convolution kernel uses the
value of neighboring pixels to
smoothen the image or to
enhance the edges. The use of
× =
a convolution kernel will
result in a filtered profile,
which can be used for back
Standard kernel
projection

× =

Smooth kernel

× =

Sharp kernel
11  Cardiovascular CT: Image Reconstruction 121

However, smooth kernels will result in less noise, which will The first step of iterative reconstruction is a FBP. From the
improve lumen visualization. Usually medium-smooth ker- images derived from the FBP, new projection data are gener-
nels are used for cardiac CT angiography, while sharper ker- ated based on prior information concerning the scanner geom-
nels are applied for coronary calcium scoring. etry, the x-ray spectrum, the detector characteristics, and a
noise model. Subsequently the new projection data are com-
Iterative Reconstruction pared to the original projection data, and corrections in noise
modeling are performed. This process is repeated (iterated)
A commonly used reconstruction method is FBP, in which a several times. Each time the newly generated projection data
convolution kernel is applied to the projection data, followed will represent a closer match to the original projection data.
by a back projection as explained above. This method works This iteration process can be repeated until the differences are
well in most situations; however at low-radiation doses and very small or until a certain number of iterations are reached.
in large patients, FBP produces noisy images. To reduce the This process is also called model-­based iterative reconstruc-
amount of noise, iterative reconstruction algorithms have tion. A simplified, less time-­consuming method is hybrid iter-
been developed [9]. A form of iterative reconstruction was ative reconstruction. In hybrid iterative reconstruction, noise is
first described in the 1970s under the name algebraic recon- iteratively filtered in the projection domain and in the image
struction [10]. Due to the limited computational power at domain without multiple iterations in forward and backward
that time, it was not suitable for clinical practice. projection steps (Fig.  11.6). In image-based iterative recon-

a
Projection data Image data

Measured projection data

True object CT Projection 1 Back Projection 2

3 Filter 4 IR 5

Noise identification

b
Measured projection data

True object CT Projection 1

Noise identification
5
n
tio
ec

3 6 8
IR Filter 7 IR
oj

4
Pr
ck
Ba

Noise identification

Model projection data

Fig. 11.6  Image-based (a), hybrid (b), and model-based (c) iterative Model-based iterative reconstruction (a) is more advanced and com-
reconstruction schematically explained. The numbers represent the sub- pares and corrects the modeled projection data to the true projection
sequent steps undertaken. In image-based iterative reconstruction (b), a data (step 4) after which a new back projection is performed. This pro-
FBP is performed after which noise is reduced in the image domain cess is iterated several times to acquire model projection data that are a
(steps 5–6). Hybrid iterative reconstruction (c) reduces noise both in the close match to the measured projection data
projection domain (steps 3–4) and in the image domain (steps 7–8).
122 A. M. den Harder et al.

c
Measured projection data

True object CT Projection 1 Back Projection 2

6
Compare and correct

n
tio
ec
oj
Pr
5

ck
Ba
4
IR

3 Forward projection
Model projection data

Fig. 11.6 (continued)

struction, noise is only reduced in the image domain. In both Table 11.1  Overview of current clinically used iterative reconstruc-
hybrid and image-­based iterative reconstruction, no forward tion algorithms
projection steps are performed based on the acquired image Algorithm Vendor Type
resulting in less computational demanding algorithms. Iterative Reconstruction in Image Space Siemens Image-­
(IRIS) based
The current clinically available iterative reconstruction Adaptive Iterative Dose Reduction 3D Toshiba Hybrid
algorithms are provided in Table  11.1. Most current clini- (AIDR 3D)
cally available algorithms have several levels or percentages Adaptive Statistical Iterative General Hybrid
indicating the amount of noise reduction. A larger level or Reconstruction (ASiR) Electric
percentage implies more noise reduction, which can, for Adaptive Statistical Iterative General Hybrid
Reconstruction (ASiR – V) Electric
example, be achieved by a higher number of iterations.
Advanced Modeled Iterative Siemens Hybrid
Iterative reconstruction is nowadays widely implemented Reconstruction (ADMIRE)
and used. Some of the latest CT systems do not offer FBP iDose4 Philips Hybrid
anymore but only use iterative reconstruction. Iterative recon- Sinogram-Affirmed Iterative Siemens Hybrid
struction has several advantages and disadvantages. The main Reconstruction (SAFIRE)
advantage is noise reduction, which results in two major Model-Based Iterative Reconstruction GE Model-­
(MBIR-Veo) based
applications. First, image quality is improved compared to
Iterative Model Reconstruction (IMR) Philips Model-­
FBP at a similar dose level. Second, it also offers the oppor- based
tunity to acquire images at lower radiation dose levels with Forward projected model-based Iterative Toshiba Model-­
comparable image quality compared to FBP at routine dose Reconstruction SoluTion (FIRST) based
levels. The reported achievable dose reduction for cardiac CT
acquisitions using iterative reconstruction varies widely and
is dependent on several factors including the reconstruction power, and therefore reconstruction times can be prolonged.
algorithm, CT system, and study population. Recent studies The delay with image-based and hybrid iterative algorithms
have shown that iterative reconstruction allowed for coronary is limited and is not likely to result in clinically relevant
CT angiography at radiation dose levels of 2–3 mSv, while delays, not even in the emergency setting [14]. Model-based
with FBP doses up to 10 mSv are used [11]. Using iterative algorithms initially required several hours for image recon-
reconstruction, non-contrast-enhanced acquisitions for coro- struction, but advancements in reconstruction algorithms and
nary calcium scoring can be reduced to submillisievert dose computational power have reduced this to less than half an
levels [12, 13]. These results are mainly based on hybrid iter- hour and for some algorithms to only a few minutes [15]. It is
ative reconstruction algorithms. Further dose reduction is likely that the reconstruction time will be reduced even fur-
expected using model-based iterative reconstruction. ther in the future and therefore will not be a barrier anymore.
Reported disadvantages of iterative reconstruction are longer The second commonly mentioned disadvantage is the differ-
reconstruction times and a different, smoother image appear- ent image appearance, also described as “plastic,” “blotchy,”
ance. Iterative reconstruction requires more computational or “artificial” (Fig. 11.7). This is caused by low noise levels
11  Cardiovascular CT: Image Reconstruction 123

a b c

FBP Hybrid IR Model-based IR

Fig. 11.7  Contrast-enhanced CT acquisition of the heart in a 56-year-­ iDose level 4, Philips), and model-based iterative reconstruction (c,
old male with a heart rate of 59 beats/minute and a weight of 80 kilo- IMR level 2, Philips). Note the smoother image appearance with model-­
grams. Reconstructed with FBP (a), hybrid iterative reconstruction (b, based iterative reconstruction, due to low noise levels

7. Kruger BL, Lee F, Bruesewitz MR, Primak A, Williamson EE,

and is more pronounced in higher noise reduction settings for
McCollough CH. Quality recovery in cardiac CT angiography: how
hybrid iterative reconstruction algorithms and for all model- to edit an ECG trace to improve image quality. Abstract Archives of
based iterative reconstruction algorithms. There are concerns the RSNA 2007(LL-PH5183).
that this will result in decreased diagnostic accuracy; however 8. Cademartiri F, La Grutta L, Runza G, et al. Influence of convolu-
tion filtering on coronary plaque attenuation values: observations
this has not been investigated yet.
in an ex vivo model of multislice computed tomography coronary
angiography. Eur Radiol. 2007;17(7):1842–9.
9. Willemink MJ, de Jong PA, Leiner T, et al. Iterative reconstruction
References techniques for computed tomography part 1: technical principles.
Eur Radiol. 2013;23(6):1623–31.
10. Fleischmann D, Boas FE.  Computed tomography–old ideas and
1. Ketcham RA, Hannah RD.  Beam hardening correction for X-ray
new technology. Eur Radiol. 2011;21(3):510–7.
computed tomography of heterogeneous materials. Comput Geosci.
11. Den Harder AM, Willemink MJ, De Ruiter QM, et al. Dose reduction
2014;67:49–61.
with iterative reconstruction for coronary CT angiography: a sys-
2. Boas FE, Fleischmann D. CT artifacts: causes and reduction tech-
tematic review and meta-analysis. Br J Radiol. 2015;89:20150068.
niques. Imaging Med. 2012;4:229–40.
12. Hecht HS, de Siqueira ME, Cham M, et al. Low- vs. standard-dose
3. Goldman LW. Principles of CT: multislice CT. J Nucl Med Technol.
coronary artery calcium scanning. Eur Heart J Cardiovasc Imaging.
2008;36(2):57,68; quiz 75–6.
2015;16(4):358–63.
4. Brink JA, Heiken JP, Wang G, McEnery KW, Schlueter FJ,
13. van Osch JA, Mouden M, van Dalen JA, et al. Influence of iterative
Vannier MW. Helical CT: principles and technical considerations.
image reconstruction on CT-based calcium score measurements. Int
Radiographics. 1994;14(4):887–93.
J Cardiovasc Imaging. 2014;30(5):961–7.
5. Matsumoto H, Kondo T, Watanabe S, et  al. ECG-edited middia-
14. Willemink MJ, Schilham AM, Leiner T, Mali WP, de Jong PA, Budde
stolic phase reconstruction improves image quality at 64-MDCT
RP. Iterative reconstruction does not substantially delay CT imag-
coronary angiography of patients with atrial fibrillation. AJR Am J
ing in an emergency setting. Insights Imaging. 2013;4(3):391–7.
Roentgenol. 2008;191(6):1659–66.
15. Oda S, Weissman G, Vembar M, Weigold WG. Cardiac CT for plan-
6. Cademartiri F, Mollet NR, Runza G, et  al. Improving diagnos-
ning redo cardiac surgery: effect of knowledge-based iterative model
tic accuracy of MDCT coronary angiography in patients with
reconstruction on image quality. Eur Radiol. 2015;25(1):58–64.
mild heart rhythm irregularities using ECG editing. AJR Am J
Roentgenol. 2006;186(3):634–8.
 The Challenging Patient
12
Damiano Caruso, Domenico De Santis, Taylor M. Duguay,
Sheldon E. Litwin, and Carlo N. De Cecco

In the past decade, coronary CT angiography (CCTA) has prehensive evaluation of CAD and direct assessment of myo-
established itself as a robust imaging technique for the exclu- cardial ischemia in a single examination [12].
sion of coronary artery disease (CAD) in patients presenting Technically speaking, CCTA presents a greater challenge
with acute chest pain [1, 2] and moderate cardiovascular risk and is more demanding than other CT applications due to the
[3–5] due to its high sensitivity and negative predictive value nature of its target, the continuously moving heart [13, 14].
in detecting significant coronary artery stenosis [6]. Thus, higher spatial and temporal resolutions are mandatory
Noninvasive CCTA has become a reliable alternative to inva- for a proper visualization of small-caliber vessels such as cor-
sive angiography [7] and, thanks to recent technical innova- onary arteries. In addition, image acquisition should be syn-
tions, has raised its power from just a morphological evaluation chronized with the patient’s electrocardiogram (ECG) in order
to a more comprehensive morphological and functional to achieve optimal image quality. Due to its technical com-
assessment [2, 8]. Specifically, novel CT technological plexity, the accurate selection and preparation of the patient
advancements such as CT myocardial perfusion imaging are mandatory for obtaining a high-level examination [15].
(CT-MPI) provide a reliable functional analysis of the heart Despite the continuous technological advances that have
capable of accurately presenting hemodynamically significant unleashed the diagnostic power of CCTA and allowed accu-
coronary artery stenosis [9–11]. CCTA coupled with CT-MPI rate diagnoses even in difficult clinical scenarios, some
could raise CT as the stand-alone image modality for a com- minor limitations should still be taken into account when
performing CCTA examinations. The comprehensive defini-
tion of a challenging patient embraces some clinical condi-
tions that should be investigated in order to always guarantee
D. Caruso
Department of Radiological Sciences, Oncological and a safe and accurate CT examination.
Pathological Sciences, University of Rome “Sapienza”,
Latina, Italy
D. De Santis The Ideal Patient
Department of Radiological Sciences, Oncological and
Pathological Sciences, University of Rome “Sapienza”, Due to the technical limitations of CT in its early years,
Latina, Italy
CCTA had to be reserved to a select population in order to
Division of Cardiovascular Imaging, Department of Radiology achieve diagnostic results [16–18]. On one hand, the accu-
and Radiological Science, Medical University of South Carolina,
Charleston, SC, USA rate selection of patients who can benefit from a CCTA is
important in order to avoid unnecessary invasive procedures
T. M. Duguay (*)
Division of Cardiovascular Imaging, Department of Radiology such as coronary angiography, while on the other hand, it is
and Radiological Science, Medical University of South Carolina, necessary to ensure that all CCTA examinations provide
Charleston, SC, USA diagnostic input. The most important factors that allow suc-
e-mail: [email protected] cessful CCTA examinations are a heart rate less than 60
S. E. Litwin beats per minute (bpm) and a sinus rhythm [19, 20].
Division of Cardiology, Department of Medicine, Medical Additionally, sufficient renal function is also mandatory for
University of South Carolina, Charleston, SC, USA
patients undergoing CCTA due to the need for IV iodinated
C. N. De Cecco (*) contrast media.
Department of Radiology and Radiological Science, Medical
University of South Carolina, Charleston, SC, USA The main goal of CCTA is to obtain high-quality images
e-mail: [email protected] while delivering the lowest reasonably achievable radiation

© Humana Press 2019 125

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_12
126 D. Caruso et al.

dose to the patient [21, 22]. Generally speaking, image qual- The most commonly seen heart rate irregularity is atrial
ity and radiation dose are intrinsically linked  – high-dose fibrillation. In order to overcome the irregular heart rhythm,
examinations return higher perceived image quality. CCTA image acquisition is usually performed using a retro-
However, radiation dose is not the only contributing factor to spective acquisition protocol. CCTA has been described to
the quality of images. Patient weight has been shown to have a high diagnostic accuracy in patient affected by atrial
affect image quality [23, 24]; and in fact the diagnostic qual- fibrillation, as reported in a recent meta-analysis [30]; how-
ity of CCTA images of obese patients is often severely lim- ever, the retrospective acquisition protocol often results in an
ited. In addition, the ideal patient must be hemodynamically increased radiation dose. A recent study evaluated the feasi-
stable, able to perform a sufficient breath-hold of 15 s, and bility of prospective ECG-triggered sequential acquisition in
able to lay in a supine position with their arms extended patients affected by atrial fibrillation and reported an excel-
behind their head. The administration of beta-blockers and/ lent diagnostic accuracy (around 90%) [20].
or sublingual nitroglycerine is common in clinical practice in Ivabradine, a selective inhibitor of If channels located in
order to achieve an optimal heart rate less than 60 bpm and the sinus node, is an alternative choice to metoprolol tartrate
sufficient vasodilation, both of which have been shown to for heart rate reduction. Unfortunately, it is recommended
provide superior image quality [25, 26]. However, patients that ivabradine only be administered to patients with a sinus
contraindicated for beta-blockers and/or sublingual nitro- rhythm [31]. On the other hand, ivabradine has a good safety
glycerine (discussed further in subsequent sections) also profile and does not lower left ventricle contractility.
present a challenge in terms of producing optimal image Moreover, no rebound effects were reported after drug cessa-
quality [27]. tion nor was there an onset of tolerance reported after pro-
longed use. Ivabradine is administered orally and generally
allows for a heart rate reduction of around 10 bpm at the rec-
The Challenging Patient ommended dosage (no more than 15 mg/day). Several trials
have demonstrated the efficacy of ivabradine in lowering the
Elevated Heart Rates and Arrhythmias heart rate [27, 32–36] and established it as a single-dose pre-
medication proven to be effective and safe in heart rate
The accuracy of CCTA is highly dependent on the image reduction prior to CCTA [27, 35].
quality of the dataset, and robust heart rate control, ideally The most recent generation of CT scanners have provided
less than 60 bpm, is essential for optimizing image quality superior image quality and improved diagnostic performance
and minimizing radiation dose [28]. Beta-blockers are rou- that is less dependent on heart rate. These scanners allow for
tinely administered in clinical practice for heart rate reduc- the acquisition of the entire heart in a single heartbeat, pro-
tion and cardiac rhythm stabilization. Drugs classified as viding great image quality with a lower radiation dose and
beta-blockers are grouped based on the β1 selectivity, lipid administered contrast media amount [37].
solubility, and partial agonistic activity [28]. Dedicated
guidelines have been established in order to avoid complica-
tions with the administration of beta-blockers. These contra- Reduced Renal Function
indications are as follows: severe chronic obstructive
pulmonary disease, asthma, sensitivity to beta-blockers, sec- Patients with a reduced renal function present a unique chal-
ond- or third-degree heart block, and sinus bradycardia with lenge as they are limited to the amount of intravenous con-
systolic blood pressure less than 100 mmHg [29]. trast media they can be administered. Contrast-induced acute
Patients who have a heart rate < 60 bpm during pre-­scan kidney injury (CI-AKI) is an iatrogenic disease that occurs
monitoring do not require beta-blockers, whereas patients within 3 days of the intravascular injection of iodinated con-
with irregular rhythms or a heart rate > 60 bpm require beta- trast media in absence of an alternative etiology. It is defined
blockers to avoid any motion artifacts. Despite having sev- as an absolute increase in serum creatinine concentration of
eral cardio-selective beta-blockers, metoprolol tartrate ≥0.5 mg/dL or a relative increase of ≥25% above baseline.
(5 mg) is typically chosen due to its availability and cardio- CI-AKI has received significant public attention which has
selectivity. Metoprolol tartrate (5  mg) is typically intrave- resulted in the administration of iodinated contrast media
nously administered in order for the drug to take effect being considered a general risk factor for worsening renal
quickly (5–10 min), whereas metoprolol oral tablets need to function [38].
be administered 60–90 min before the scan. In the absence of Estimated glomerular filtration rate (eGFR), calculated
contraindications, an initial bolus of 5 mg is usually injected from the serum creatinine, is the preeminent method of esti-
once the patient is lying on the scanner table. In cases of mating renal function prior to contrast media administration.
unsatisfactory heart rate response, a maximum 15  mg of The Chronic Kidney Disease Epidemiology Collaboration
metoprolol tartrate can be administered. (CKD-EPI) formula gives the most accurate eGFR.  An
12  The Challenging Patient 127

eGFR value below 45  ml/min/1.73m2 is considered a risk The introduction of the latest generation of dual-source
factor for contrast-induced nephropathy (CIN) as reported CT scanner overcame the need to increase the tube voltage to
from dedicated guidelines [39], and preventive measures are 140  kV.  These new CT scanners, equipped with the latest
required in such patients. technical developments, are able to improve the image qual-
Hydration is considered the common strategy for CIN ity even in obese patients thanks to the increased power of
prevention. In patients with eGFR between 30 and 45 mL/ the tubes, a higher spatial resolution related to a smaller focal
min/1.73m2, intravenous hydration should be arranged spot, and an improved temporal resolution due to shorter
before and after the administration of contrast media. ESUR gantry times, along with a fully integrated circuit detector
guidelines suggest intravenous protocol of 1.0–1.5 mL/kg/h system for noise reduction. Mangold et  al. have demon-
for at least 6 h before and after contrast media administration strated that even severe obese patients with a BMI greater
[38, 40]. If intravenous hydration cannot be appropriately than 40 kg/m2 can be imaged at 120 kVp if high amperage
performed, oral hydration should be prescribed. and dedicated iterative reconstructions are available. This
Contrast media volume has also been shown to be associ- technical approach led to satisfactory objective and subjec-
ated with CIN [41]. As such, reducing contrast media vol- tive CCTA image quality and adequate coronary intravascu-
ume used during CCTA is a reliable option in reducing the lar attenuation with an approximate contrast media dose of
risk of CIN. The ability to lower contrast media volume is 80  mL [24]. In addition, automated tube voltage selection
directly related to reducing the tube current. Low tube ener- (ATVS) technology and the most recent generation of itera-
gies are closer to the iodine k-edge (33.2  keV), and thus, tive reconstructions can be reliably used to assess coronary
lower energy levels maximize iodine attenuation allowing a artery disease in obese patients returning an excellent diag-
substantial reduction contrast media [42]. Recent studies nostic accuracy at a reduced radiation dose compared to
confirm the possibility to use low volumes of contrast media standard 120  kV CCTA examinations. ATVS individually
(30  mL) in CCTA by using prospectively ECG-triggered selects the tube voltage based on the patient’s habitus, sys-
high-pitch gated acquisition at 70 kVp without any lack of tem specifics, and the clinical task. With the implementation
image quality in selected patients with BMIs <23 kg/m2 and of this technology, even obese patients can undergo CCTA
heart rates <65  bpm [43]. Unfortunately, high BMI values examination with tube voltages less than 120 kV and main-
significantly limit the image quality when low tube voltages tain adequate diagnostic accuracy [48].
are applied.

The Anxious Patient

The Obese Patient
Beta-blockers are administered in CCTA in order to achieve
Obese patients are considerable challenges for CCTA, as low and stable heart rates that are mandatory for obtaining
the higher image noise generated by the photon starvation diagnostic image quality at the lowest radiation dose.
phenomenon [44] degrades both contrast and spatial reso- However, not all patients respond to beta-blocker treatment,
lution causing a poor evaluation of small vessels and non- which is sometimes due to patient anxiety [49]. Different
calcified atherosclerotic plaques. With previous generations strategies have been considered to reduce patient anxiety
of CT scanners, increasing the tube voltage to 140 kVp was such as the use of benzodiazepines. Even if benzodiazepines
considered one of the most effective solutions for improv- are not significantly involved in the mechanism for lowering
ing the image quality in obese patients [45]. Although this the heart rate, they can be useful in reducing the physiologi-
increases photon flux, higher tube voltage returns lower cal burden of the examination [50]. According to one study,
attenuation and consequently lowers the signal of iodinated benzodiazepines are usually used when beta-blockers are
contrast media, all while bearing the cost of an increased required but contraindicated. A comprehensive overlook of
radiation dose. different challenging scenarios and related countermeasures
In such patients, slower gantry rotation speeds may be are summarized in Table 12.1.
advisable to enhance photon flux and improve the signal-to-­
noise ratio. In addition, the use of a high iodine concentra-
tion contrast medium at a faster flow rate could be helpful in  ow to Manage Angina Onset During
H
compensating for the increased levels of image noise that Stress Cardiac CT
limit contrast resolution. Moreover, dedicated iterative
reconstructions, by using different strategies and mathemati- With the use of pharmacological stress agent, patients can
cal algorithms, allow for a reduced image noise and an experience common side effects such as flushing, chest
improved image quality compared to standard filtered back pain, dyspnea, headache, and gastrointestinal discomfort.
projection [46, 47]. The most common pharmacological stress agents used
128 D. Caruso et al.

Table 12.1  Challenging scenarios and related strategies

Challenge Strategy Drug Dose Peak Side effects Contraindications
Elevated Βeta-blockers Metoprolol 5 mg (up to 15 mg) 5–10 min LV contraction Sinus bradycardia, heart
heart rates (IV) depression, block greater than first
and Metoprolol 100 mg 60– arteriolar and degree, systolic blood
arrhythmias (oral) 90 min bronchiolar pressure < 100 mmHg
Atenolol (IV) 5 mg 5 min constriction
Selective If Ivabradine 15 mg 60 min Luminous Acute decompensated heart
channel inhibitor (oral) phenomenon failure, blood
(phosphenes); pressure < 90/50 mmHg,
enhanced sick sinus syndrome
brightness in
visual field
eGFR Hydration Saline or Saline: IV – – –
<45 mL/ bicarbonate injection1.0–1.5 mL/
min/1.73 m2 kg/h for at least 6 h
before and after CM
Bicarbonate: 154 mEq/L
in dextrose (5% water):
3 mL/kg/h for 1 h before
CM and 1 mL/kg/h for
6 h after CM
Lower the CM
amount
Obese ATVS – – – – Increased radiation
patient Iterative exposure
reconstruction
Anxious Benzodiazepines Lorazepam 0.5–1 mg (20–40 drops) 60 min Sedation, Acute narrow-angle
patient (oral) tremor, glaucoma
Diazepam 2 mg (10 drops) 30– agitation,
(oral) 120 min dizziness,
hypersomnia
Heart Methylxanthines Aminophylline 50–250 mg at slow rate 1–2 min Abdominal Not to be used in case of
ischemia (IV) (50–100 mg over pain, blurred seizures
30–60 s) vision, fainting
Nitrates Nitroglycerine 0.4 mg 5 min Headache, Severe anemia, increased
(SL) blurred vision, intracranial pressure,
fainting hypersensitivity to
nitroglycerin
Cardiopulmonary
resuscitation
maneuvers
Transfer the
patient to the
emergency
department
Abbreviations: LV Left ventricle, eGFR estimated glomerular filtration rate, CM contrast media, ATVS automatic tube voltage selection

during CT-MPI are regadenoson and adenosine. Patients affected by angina during a stress cardiac CT are
Regadenoson has been shown to have a lower incidence of somewhat fortunate in the sense that they are already hospi-
flushing and chest pain and higher incidence of abdominal talized and can receive treatment almost immediately, avoid-
discomfort and headache compared to adenosine [51]. ing the common delay of 1.5–2 h reported for patients that
Both stress agents have reported extremely rare but serious develop stress-induced angina outside of a hospital setting
side effects such as seizures and myocardial infarction [55, 56]. Since the patient is ECG-monitored during the
[52]. In cases where a patient experiences pharmacological exam, any suspicious change in the baseline ECG, even if
stressor-induced seizures, the use of methylxanthines non-specific for myocardial ischemia, is immediately identi-
(aminophylline, theophylline, and caffeine) is not recom- fiable. Intravenous injection of aminophylline (50–250 mg)
mended [53, 54]. Moreover, it should be pointed out that at a slow rate (50–100 mg over 30–60 s) is the first choice in
some seizures can be prolonged and may require anticon- cases of adverse side effects [57] and usually leads to resolu-
vulsive therapy. tion of the patient’s symptoms.
12  The Challenging Patient 129

Cases of asystole following regadenoson administration 10. Caruso D, et al. Dynamic CT myocardial perfusion imaging. Eur J
Radiol. 2016;85(10):1893–9.
have been reported [58]. In such rare scenarios, CT person- 11. Jin KN, et  al. Myocardial perfusion imaging with dual energy
nel must be prepared to perform immediate cardiopulmonary CT. Eur J Radiol. 2016;85(10):1914–21.
resuscitation and medical stabilization. Rare cases of acute 12. Pelgrim GJ, et al. The dream of a one-stop-shop: meta-analysis on
myocardial infarct up to 15 min after pharmacological stress myocardial perfusion CT. Eur J Radiol. 2015;84(12):2411–20.
13. Flohr TG, et  al. Computed tomographic assessment of coronary
test were also reported [59, 60] and are typically treated with artery disease: state-of-the-art imaging techniques. Radiol Clin N
sublingual nitroglycerine and intravenous administration of Am. 2015;53(2):271–85.
heparin and saline. IV atropine administration (0.5–1  mg) 14. Flohr TG, et  al. Pushing the envelope: new computed tomogra-
should also be considered. It is highly recommended that in phy techniques for cardiothoracic imaging. J Thorac Imaging.
2010;25(2):100–11.
the CT section, there is an availability of transcutaneous pac- 15. Taylor CM, Blum A, Abbara S.  Patient preparation and scanning
ing for bradyarrhythmias or heart block associated with techniques. Radiol Clin N Am. 2010;48(4):675–86.
regadenoson or adenosine. Most importantly, patients suffer- 16. Schussler JM, Grayburn PA.  Non-invasive coronary angiography
ing from rare but serious side effects must be transferred to using multislice computed tomography. Heart. 2007;93(3):290–7.
17. Kopp AF, et al. Coronary arteries: retrospectively ECG-gated multi-­
the emergency department for comprehensive monitoring detector row CT angiography with selective optimization of the
and further medical management. image reconstruction window. Radiology. 2001;221(3):683–8.
18. Nieman K, et al. Coronary angiography with multi-slice computed
tomography. Lancet. 2001;357(9256):599–603.
19. Fuchs TA, et  al. Impact of a new motion-correction algo-

Summary rithm on image quality of low-dose coronary CT angiography
in patients with insufficient heart rate control. Acad Radiol.
The aforementioned scenarios represent just some of the 2014;21(3):312–7.
challenges that can be observed around the CT scanner in 20. Yang L, et al. Prospectively ECG-triggered sequential dual-source
coronary CT angiography in patients with atrial fibrillation: influ-
everyday practice. Despite being coined a “challenge,” the ence of heart rate on image quality and evaluation of diagnostic
challenging patient represents an objective issue in clinical accuracy. PLoS One. 2015;10(7):e0134194.
management. However, with the continuous technological 21. Yin WH, et  al. Detection of coronary artery stenosis with sub-­
and pharmacological advances in addition to an effective milliSievert radiation dose by prospectively ECG-triggered high-­
pitch spiral CT angiography and iterative reconstruction. Eur
training of CT personnel, we are nowadays able to effec- Radiol. 2013;23(11):2927–33.
tively face and overcome many of these issues, all in an effort 22. Zhang LJ, et al. Feasibility of prospectively ECG-triggered high-­
to guarantee safe and adequate examinations for all patients. pitch coronary CT angiography with 30 mL iodinated contrast agent
at 70 kVp: initial experience. Eur Radiol. 2014;24(7):1537–46.
23. Cecco CN, Schoepf UJ. Future of cardiac computed tomography.
World J Radiol. 2015;7(12):421–3.
References 24. Mangold S, et al. Coronary CT angiography in obese patients using
3(rd) generation dual-source CT: effect of body mass index on
1. Meinel FG, et  al. Coronary computed tomographic angiog- image quality. Eur Radiol. 2016;26(9):2937–46.
raphy in clinical practice: state of the art. Radiol Clin N Am. 25. Takx RA, et  al. Sublingual nitroglycerin administration in coro-
2015;53(2):287–96. nary computed tomography angiography: a systematic review. Eur
2. Takx RA, et al. Computed tomography-derived parameters of myo- Radiol. 2015;25(12):3536–42.
cardial morphology and function in black and white patients with 26. Shim SS, Kim Y, Lim SM.  Improvement of image quality with
acute chest pain. Am J Cardiol. 2016;117(3):333–9. beta-blocker premedication on ECG-gated 16-MDCT coronary
3. Hoffmann U, et al. ACR appropriateness criteria acute nonspecific angiography. AJR Am J Roentgenol. 2005;184(2):649–54.
chest pain-low probability of coronary artery disease. J Am Coll 27. Celik O, et al. Single dose ivabradine versus intravenous metopro-
Radiol. 2015;12(12 Pt A):1266–71. lol for heart rate reduction before coronary computed tomogra-
4. Bastarrika G, et  al. CT of coronary artery disease. Radiology. phy angiography (CCTA) in patients receiving long-term calcium
2009;253(2):317–38. channel-­blocker therapy. Acta Radiol. 2014;55(6):676–81.
5. Bastarrika G, Schoepf UJ.  Evolving CT applications in ischemic 28. Sabarudin A, Sun Z.  Beta-blocker administration protocol for

heart disease. Semin Thorac Cardiovasc Surg. 2008;20(4):380–92. prospectively ECG-triggered coronary CT angiography. World J
6. von Ballmoos MW, et al. Meta-analysis: diagnostic performance of Cardiol. 2013;5(12):453–8.
low-radiation-dose coronary computed tomography angiography. 29. Pannu HK, Alvarez W Jr, Fishman EK. Beta-blockers for cardiac
Ann Intern Med. 2011;154(6):413–20. CT: a primer for the radiologist. AJR Am J Roentgenol. 2006;186(6
7. De Cecco CN, et al. Beyond stenosis detection: computed tomog- Suppl 2):S341–5.
raphy approaches for determining the functional relevance of coro- 30. Vorre MM, Abdulla J.  Diagnostic accuracy and radiation dose of
nary artery disease. Radiol Clin N Am. 2015;53(2):317–34. CT coronary angiography in atrial fibrillation: systematic review
8. Ruzsics B, et al. Comparison of dual-energy computed tomography and meta-analysis. Radiology. 2013;267(2):376–86.
of the heart with single photon emission computed tomography for 31. National Institute for Health and Care Excellence (NICE).

assessment of coronary artery stenosis and of the myocardial blood Ivabradine for treating chronic heart failure (NICE website).
supply. Am J Cardiol. 2009;104(3):318–26. 2012. Published online at https://www.nice.org.uk/guidance/
9. Varga-Szemes A, et al. CT myocardial perfusion imaging. AJR Am ta267/resources/ivabradine-for-treating-chronicheart-failure-
J Roentgenol. 2015;204(3):487–97. pdf-82600557030853
130 D. Caruso et al.

32. Adile KK, et al. Safety and efficacy of oral ivabradine as a heart 47. Geyer LL, et al. State of the art: iterative CT reconstruction tech-
rate-reducing agent in patients undergoing CT coronary angiogra- niques. Radiology. 2015;276(2):339–57.
phy. Br J Radiol. 2012;85(1016):e424–8. 48. Mangold S, et  al. Diagnostic accuracy of coronary CT angiogra-
33. Guaricci AI, et al. Incremental value and safety of oral ivabradine phy using 3rd-generation dual-source CT and automated tube volt-
for heart rate reduction in computed tomography coronary angiog- age selection: clinical application in a non-obese and obese patient
raphy. Int J Cardiol. 2012;156(1):28–33. population. Eur Radiol. 2017 Jun;27(6):2298–308.
34. Lambrechtsen J, Egstrup K. Pre-treatment with a sinus node block- 49. Jimenez-Juan L, et al. Failed heart rate control with oral metoprolol
ade, ivabradine, before coronary CT angiography: a retrospective prior to coronary CT angiography: effect of additional intravenous
audit. Clin Radiol. 2013;68(10):1054–8. metoprolol on heart rate, image quality and radiation dose. Int J
35. Pichler P, et al. Ivabradine versus metoprolol for heart rate reduc- Cardiovasc Imaging. 2013;29(1):199–206.
tion before coronary computed tomography angiography. Am J 50. Maffei E, et al. “In-house” pharmacological management for com-
Cardiol. 2012;109(2):169–73. puted tomography coronary angiography: heart rate reduction, tim-
36. Bayraktutan U, et al. Efficacy of ivabradin to reduce heart rate prior ing and safety of different drugs used during patient preparation.
to coronary CT angiography: comparison with beta-blocker. Diagn Eur Radiol. 2009;19(12):2931–40.
Interv Radiol. 2012;18(6):537–41. 51. Cerqueira MD, et  al. Effects of age, gender, obesity, and dia-
37. Morsbach F, et al. Performance of turbo high-pitch dual-source CT betes on the efficacy and safety of the selective A2A agonist
for coronary CT angiography: first ex vivo and patient experience. regadenoson versus adenosine in myocardial perfusion imaging
Eur Radiol. 2014;24(8):1889–95. integrated ADVANCE-MPI trial results. JACC Cardiovasc Imaging.
38. Tao SM, et  al. Contrast-induced nephropathy in CT: inci-
2008;1(3):307–16.
dence, risk factors and strategies for prevention. Eur Radiol. 52. Kitt T, Jiang J. Serious complications associated with regadenoson
2016;26(9):3310–8. administration for myocardial perfusion imaging. J Nucl Cardiol.
39. Stacul F, et  al. Contrast induced nephropathy: updated ESUR
2015;22(2):395.
Contrast Media Safety Committee guidelines. Eur Radiol. 53. Agarwal V, DePuey EG. Regadenoson and seizures: a real clinical
2011;21(12):2527–41. concern. J Nucl Cardiol. 2014;21(5):869–70.
40. Ohno I, et al. Guidelines on the use of iodinated contrast media in 54. Adenoscan [package insert]. Northbrook: I.A., Inc.; 1995. Updated
patients with kidney disease 2012: digest version: JSN, JRS, and in 2014.
JCS Joint Working Group. Clin Exp Nephrol. 2013;17(4):441–79. 55. Goldberg RJ, et al. Prehospital delay in patients with acute coro-
41. Nozue T, et al. Contrast medium volume to estimated glomerular nary syndromes (from the global registry of acute coronary events
filtration rate ratio as a predictor of contrast-induced nephropathy [GRACE]). Am J Cardiol. 2009;103(5):598–603.
developing after elective percutaneous coronary intervention. J 56. Spencer FA, et al. Delay to reperfusion in patients with acute myo-
Cardiol. 2009;54(2):214–20. cardial infarction presenting to acute care hospitals: an interna-
42. Yuan R, et al. Reduced iodine load at CT pulmonary angiography tional perspective. Eur Heart J. 2010;31(11):1328–36.
with dual-energy monochromatic imaging: comparison with stan- 57. Grady EC, Barron JT, Wagner RH. Development of asystole requir-
dard CT pulmonary angiography–a prospective randomized trial. ing cardiac resuscitation after the administration of regadenoson in
Radiology. 2012;262(1):290–7. a patient with pulmonary fibrosis receiving n-acetylcysteine. J Nucl
43. Zhang LJ, et  al. High-pitch coronary CT angiography at 70 kVp Cardiol. 2011;18(3):521–5.
with low contrast medium volume: comparison of 80 and 100 kVp 58. AlJaroudi WA, et al. Safety and tolerability of regadenoson in 514
high-pitch protocols. Medicine (Baltimore). 2014;93(22):e92. SPECT MPI patients with and without coronary artery disease and
44. Barrett JF, Keat N.  Artifacts in CT: recognition and avoidance. submaximal exercise heart rate response. Eur J Nucl Med Mol
Radiographics. 2004;24(6):1679–91. Imaging. 2013;40(3):341–8.
45. Lee AM, et al. Coronary computed tomography angiography at 140 59. Hsi DH, et  al. Regadenoson induced acute ST-segment elevation
kV versus 120 kV: assessment of image quality and radiation expo- myocardial infarction and multivessel coronary thrombosis. J Nucl
sure in overweight and moderately obese patients. Acta Radiol. Cardiol. 2013;20(3):481–4.
2014;55(5):554–62. 60. Shah S, Parra D, Rosenstein RS. Acute myocardial infarction dur-
46. Gebhard C, et  al. Image quality of low-dose CCTA in obese
ing regadenoson myocardial perfusion imaging. Pharmacotherapy.
patients: impact of high-definition computed tomography and adap- 2013;33(6):e90–5.
tive statistical iterative reconstruction. Int J Cardiovasc Imaging.
2013;29(7):1565–74.
 Cardiac CT: Contemporary Clinical
Image Data Display, Analysis, 13
and Quantification

Moritz H. Albrecht, Marwen Eid, and Pal Spruill Suranyi

 hy Do We Need “Advanced Visualization”

W clinical use to measure myocardial mass and function, to
and Quantification? evaluate cardiac chamber sizes, and to assess coronary
plaques and stenoses.
One of the most significant breakthroughs in cardiac CT Lastly, as we are striving to understand the correlation
imaging was being able to acquire isotropic resolution data- between pathophysiology, qualitative and quantitative image
sets with sufficient temporal resolution to take “snapshots” findings, and patient outcomes, multiple investigational meth-
of this dynamic, ever-moving organ. Isotropic three-­ ods are still being developed, holding the promise of future
dimensional (3D) datasets acquired with either prospective clinical utility for myocardial perfusion imaging and analysis
triggering or retrospective gating, and further complicated of coronary hemodynamics. In this chapter, we will briefly
by the dimension of time within the cardiac cycle, effectively touch on each of these methods and will try to elucidate their
yielding four-dimensional (4D) datasets, allow for a variety diagnostic value and potential future clinical implementation.
of post-processing techniques useful for advanced visualiza-
tion of cardiovascular anatomy and pathology. Some of these
techniques are fancy, colorful, and eye-pleasing or just serve Clinical Image Data Display and Analysis
as an aid to student or patient education, but others have
become a crucial part of our clinical routine when evaluating 3D Volume Rendering Techniques (VRT)
cardiac CT images and discussing findings with non-­
radiology physicians. Although initially developed by the animation movie indus-
Secondly, in the era of “precision medicine” and “imag- try, 3D VRT reconstructions have been rapidly adopted in
ing biomarkers,” it is paramount to strive for more quantita- clinical medical imaging [1]. While the exact technical
tive analysis of our findings. Thus, teasing out objective details are beyond the scope of this chapter, the general idea
numbers and values from our large datasets is becoming is that interactive probability-based algorithms and worksta-
increasingly important. Initially calcium scoring was the pio- tions build virtual 3D models of an organ using the 3D matrix
neer in quantitative cardiac evaluation. Since then various of enhancement values by deciding whether a given volume
workstations and algorithms have been developed for routine element, i.e., voxel, belongs to an adjacent structure or not.
To stay on focus, this chapter will include all of these
techniques under the umbrella of 3D VRT.  However, it
M. H. Albrecht
Department of Diagnostic and Interventional Radiology, University should be noted that there are fine nuances differentiating the
Hospital Frankfurt, Frankfurt am Main, Germany various surface rendering techniques, e.g., shaded realistic
M. Eid surface displays (Fig.  13.1), and volume rendering
Division of Cardiovascular Imaging, Department of Radiology and approaches which can be further manipulated to generate fly-­
Radiological Science, Medical University of South Carolina, through or fly-around views of various organs [2]. Using cre-
Charleston, SC, USA
ative display colors for different tissue types as well as
P. S. Suranyi (*) various lighting methods and depth shading, these techniques
Division of Cardiovascular Imaging, Department of Radiology
enable the analysis of a dataset from any desired angle,
and Radiological Science, Medical University of South Carolina,
Charleston, SC, USA including from within the vessel lumen. By changing the
attenuation level-based visualization thresholds, one can
Division of Cardiology, Department of Medicine, Medical
University of South Carolina, Charleston, SC, USA hide or accentuate certain organs and tissue types or even
e-mail: [email protected] make them semitransparent. In order to simplify the s­ election

© Humana Press 2019 131

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_13
132 M. H. Albrecht et al.

Fig. 13.1  Several approaches

a b
exist for 3D visualization of
the heart and coronaries.
While traditional 3D volume
rendering technique (VRT)
images (a) facilitate the
orientation and anatomical
relations to adjacent organs,
novel cinematic rendering
(b) allows furthermore for a
photo-realistic appearance
and analysis of the organ
surfaces

a b c

Fig. 13.2  3D reconstruction images displaying the utility of 3D VRT pressing on the left main bronchus (arrowhead). The various tissue
images in a patient with Marfan syndrome with prior history of ascend- types and metallic sternal wires can be clearly delineated by assigning
ing aorta repair (note sternal wires highlighted in purple, a), presenting different colors based on CT attenuation enabling us to demonstrate the
with acute chest pain and shortness of breath. The sagittal image (a) relationship of airways to great vessels and the proximity of anterior
shows a type-B aortic dissection (arrow). The oblique axial (b) and structures to sternal wires
sagittal (c) images show the aneurysmal descending aorta (arrow) com-

of the best technique for a given indication, workstations When emphasizing vessels, however, one must under-
have presets stored for various purposes which often require stand that the goal of contrast media administration in car-
further interactive manual adjustment by the reader. diac CT is to improve the visualization of the vascular lumen
Volume rendering of the heart and adjacent anatomic rather than its surface. Also, when analyzing vessels with
structures is helpful to obtain an overall understanding of stents, 3D VRT shows the presence of the stent but does not
the spatial relationship between organs, airways, the chest show its content. Thus, 3D VRT cannot determine whether a
wall, and vascular pathologies (Fig. 13.2). 3D VRT is also stented vessel lumen is patent, occluded, or if it has intimal
beneficial in providing an overview of the quality of our hyperplasia.
datasets, especially in the case of arrhythmias or misregis- Further shortcomings of 3D VRT include locating the
trations, also called “step-offs,” which can be easily spotted presence of calcifications, which may appear to be located
in a 3D view. on the surface of the vessel when in fact they are in the vessel
wall. It is also important to note that motion artifacts and
13  Cardiac CT: Contemporary Clinical Image Data Display, Analysis, and Quantification 133

streak artifacts may cause artifactual vessel stenosis on 3D that improve the signal-to-noise ratio in compromised datas-
VRT, which may be misleading for beginners in cardiac ets and help visualize structures that may be moving in and
CT. Therefore, experts do not recommend making diagnoses out of the imaging plane during the cardiac cycle (e.g., aortic
solely based on VRT reconstructions but advise using 3D valve). Furthermore, thick-MPRs are also useful if structures
VRT as an auxiliary technique to supplement standard cross-­ have both high- and low-density components, e.g., a sclerotic
sectional images when making a diagnosis. aortic valve with superimposed vegetation or thrombus
(Fig. 13.5).

 utomated, Semiautomated, or Manual

A
Segmentation of the Heart Curved MPR, Stretched MPR, and MAR

Sophisticated workstations offer the option for automatic The basic principle of curved MPR (cMPR) is similar to that
segmentation of the heart, which virtually “removes” the rib- of linear MPR in that it uses the reconstructed dataset with-
cage and the lungs to allow for 360-degree visualization of out additional manipulation of the Hounsfield units (HU)
the heart’s chambers and vessels. Such automated segmenta- but still allows for the close-up analysis of vessels. The key
tion simplifies the visualization of routine cases, but semiau- feature of cMPR is the “centerline,” which is either auto-
tomated or manual segmentation is preferred when trying to matically or semiautomatically predetermined.
analyze more challenging postsurgical cases or patients with Subsequently, the dataset is reformatted with resulting
unusual anatomies. images showing data that is perpendicular to this centerline.
Unfortunately, automated segmentation algorithms may It is paramount that the centerline should go through the
“remove” venous bypass grafts, the internal mammary arter- vessel lumen, because otherwise the vessel may appear arti-
ies, or parts thereof, which may lead to the erroneous diagno- factually stenosed. Thus, before actually examining the
sis of occluded bypass grafts. Similarly, human observers cMPRs, one must carefully check – and if needed, adjust –
may inadvertently “remove” vessel segments or create the centerline. Normally there are two cMPR images dis-
“nicks” in vessels during post-processing when generating played next to each other, where one is perpendicular to the
the volume-rendered segmented heart, thus causing potential other and both are perpendicular to the centerline. This data-
problems for interpretation. This is why most experts set can then be manually “swiveled” about the centerline to
strongly emphasize looking at the “source images” whenever look at all possible aspects of a vessel and take a closer look
a 3D VRT image suggests abnormal pathology. at any particular vessel segment of interest (Fig. 13.6).
The alternative to segmenting out the heart is to look at In this context, it is important to emphasize the relevance
the entire 3D dataset using a cut plane or a slab, which allows of the window/level settings when studying vessels with ste-
the user to scroll through the organs in the chest in 3D while nosis, especially with the presence of calcifications or
also making sure that no structures remain hidden due to implanted stents. The apparent size of the calcified lesions
erroneous segmentation (Fig. 13.3). greatly depends on the window/level settings, as shown in
Fig. 13.7. Therefore, when stents or calcifications are present
and cause blooming artifacts, it is important to take a second
Dedicated Cardiac Planes, MPR, look at the vessel after “darkening” the image prior to mak-
and Thick-MPR ing a diagnosis. This can be achieved by widening the win-
dow settings, which reduces the influence of blooming
Multi-planar reformatting (MPR) images represent the stan- artifacts that can potentially lead to an overestimation of
dard cross-sectional reconstructions in cardiac CT. A 3D or luminal narrowing.
4D dataset with essentially isotropic resolution allows one to A second caveat to remember is that stents and calcified
arbitrarily choose the ideal plane when studying the heart. vessel segments are better evaluated using a sharper recon-
Whether it is one of the standard long- and short-axis planes struction kernel and iterative reconstruction techniques.
(Fig. 13.4) or a more dedicated view such as one showing the Unfortunately, these need to be applied at the time of recon-
aortic valve leaflets or any other structure of interest, this can structing the axial datasets from the raw imaging data [3]
be performed at a workstation in a matter of seconds using (Figs. 13.8 and 13.9).
MPR of the original source imaging data. Thus, a “simple” Another capability of the above technique is that the
MPR image shows the reconstructed data in arbitrary non-­ centerline’s natural curves can be virtually stretched,
axial planes with a thickness of one voxel. These views are resulting in sMPR images (“s” for stretched or straight).
useful in performing anatomic measurements in vessels or However, some experts consider this method too unnatu-
valves or within the chambers. ral and vulnerable to over manipulation. Additionally,
Most workstations also offer the option of generating multiple aligned reformats can be generated by combin-
“thick-MPRs,” which consist of virtually thickened CT slices ing the cMPRs of multiple vessels. This technique uses
134 M. H. Albrecht et al.

a b

c d

Fig. 13.3  3D reconstruction of the heart in a patient after coronary Unfortunately, the heart and grafts cannot be seen in their entirety on
arterial bypass graft (CABG) surgery. The heart can be automatically the full volume either, due to overlying ribs and sternum. (b) The solu-
segmented (a) and isolated from the remaining structures; however, this tion is a semitransparent (c) rib cage or a clip plane/slab, which both can
technique sometimes artificially removes important parts of the anat- visualize structures obscured by the chest wall (d), without losing infor-
omy, like the proximal aspect of the LIMA graft in this case. mation, thus reducing the likelihood of misinterpretation

the same principles as cMPR but aims to depict several perfusion (Fig. 13.10), which has the added benefit of being
vessel paths simultaneously by taking into account the able to “pair” vessel-specific stenosis with corresponding
multiple centerlines of multiple vessels. This technique regional myocardial dysfunction or perfusion abnormality.
provides an overview of the coronary anatomy and may
prove useful in planning the placement of multiple stents
near bifurcations. Projection Techniques
The efficient segmentation of the coronary tree can pro-
vide us with a quick overall understanding of the blood Maximum and minimum intensity projections are obtained
supply routes of the myocardium. The coronary tree may when a 3D slab of the original reconstructed dataset is taken
then be projected onto polar maps of regional function or (more than one voxel in thickness) and a single two-­
13  Cardiac CT: Contemporary Clinical Image Data Display, Analysis, and Quantification 135

a d e

f g

Fig. 13.4  This figure shows the standard planes for cardiac CT evalu- necessary to define true long and short axes of the heart. The spatial
ation. As the cardiac axis is highly variable in every individual, conven- correlation among the short-axis (a, apical; b, mid; c, basal), 2-chamber
tional radiological axial, sagittal, and coronal images are not optimal (d), 3-chamber (e), 4-chamber (f), and 5-chamber views (g) is demon-
for assessment of the heart. Manual adjustment on a 3D workstation is strated here by white reference lines

dimensional image is generated. This can be performed by taken when using it to assess the severity of vessel stenosis
either taking only the lowest (minimum intensity projec- due to its potential to mislead the observer and cause over-
tion  (MINIP)) or only the highest attenuation (maximum or underestimation of stenosis. Thus, for the quantification
intensity projection  (MIP)) across the slab. An additional of stenosis, it is most appropriate to use either the source
technique called average intensity projection (AIP), also images or cMPR images. An additional shortcoming of
called “thick-MPR” as previously described, uses the aver- only using MIPs is the potential to “miss” low-density
age attenuation values of all voxels across the slab. lesions such as small emboli, intimal flaps, abnormal valve
apparatus (e.g., ruptured chordae), or even myocardial
hypoattenuation.
MIP

Thin (<10 mm) and thick (>10 mm) MIPs are most useful MINIP
for the assessment of vessels or bypass grafts when the
contrast medium-filled lumen is of high interest. MIPs When the structures of interest are of low attenuation, it is help-
facilitate the tracking of the vessel paths, bifurcation anal- ful to use MINIPs, which may reveal pathologies that would be
ysis, and the evaluation of the length of stenotic segments. too subtle, if not invisible when using MPRs or MIPs.
It is worth mentioning that some workstations have the Examples include ischemic regions of the myocardium,
capability to generate curved MIP images (cMIP), which infarcted areas, fatty infiltration of the myocardium, valves,
uses the same principles as cMPR except that the end chordae tendineae, papillary muscles, vegetations, thrombi or
result is a MIP that is perpendicular to the centerline dissection membranes, and even airways. Figure  13.11 pro-
throughout the length of a given vessel. Although MIP is vides an example where MINIP, MPR, and MIP are compared
very useful for the aforementioned purposes, care must be side by side.
136 M. H. Albrecht et al.

a b c

d e

Fig. 13.5  Varying cross-sectional visualization techniques lead to dif- with high spatial resolution, does not show the entirety of the valve and
ferent depictions of a calcified and stenotic aortic valve. (a) Minimum the true extent of disease. (d) Maximum intensity projection (MIP)
intensity projection (MINIP), showing only low-density portions of the highlighting calcifications but hiding noncalcified portions of the valve.
valve. (b) Thick multi-planar reformat (thick-MPR) shows all portions (e) Valve orifice measurement on the thick-MPR
of the valve but is blurry due to volume averaging. (c) Standard MPR,

Quantification (MESA) study (https://www.mesa-nhlbi.org/Calcium/input.

aspx), patients can be assigned to standardized risk categories
In addition to the relatively well-established contemporary with a low user-dependent variability compared to purely sub-
clinical image data display and analyzing techniques, the field jective or qualitative CT analysis. More advanced quantitative
of objective image quantification is rapidly growing and repre- methods include calcium volume and density scores, and the
sents the most investigated topic in state-of-the-art cardiac CT. hope is that fully automated Agatston scoring may be available
in the near future [4]. Further details regarding Agatston scor-
ing are discussed in another chapter of this book.
Calcium Scoring

Coronary artery calcium scoring, also known as Agatston scor-  oronary Stenosis Quantification/Fractional
C
ing, is the most validated method for the prediction of patient Flow Reserve
outcome and risk stratification and has substantially enhanced
the utility of cardiac CT imaging. In a standardized fashion, the Workstations have evolved to the point where algorithms
Agatston score is semiautomatically calculated based on two based on attenuation levels within the coronary lumen allow
parameters: the area of calcified coronary plaques with an us to quantitatively evaluate vessel stenosis [5, 6]. Currently,
attenuation greater than 130 HU on an ECG-­synchronized non- this is performed in a semiautomated fashion in which the
contrast cardiac CT and a density factor which is based on the observer defines the segment of interest by placing a mark
maximal attenuation of the same plaque. Using an online cal- just proximal to the starting point of the plaque and another
culator based on the Multi-Ethnic Study of Atherosclerosis just beyond the distal end of the plaque using cMPR images
13  Cardiac CT: Contemporary Clinical Image Data Display, Analysis, and Quantification 137

a b c

d e f

Fig. 13.6  The assessment of curved multi-planar reformat (cMPR) head). The full extent of this mixed calcified-noncalcified plaque, and
images from multiple views is crucial for accurate coronary artery eval- its effect on the lumen, is best understood when all different projections
uation. Varying angulations, achieved by manual swiveling around the are considered. Figure (d) accentuates the noncalcified portion (arrow-
coronary centerline, reveal different details of a coronary stenosis head) and makes the stenosis appear more severe than it really is. (a),
(arrow) with mixed calcified and soft-plaque of the LAD proximal to an (b), (c), (e), and (f) demonstrate the mixed nature of the plaque and less
implanted stent. Note that the severity of luminal narrowing could be severe stenosis
over- or underestimated by only examining one cMPR image (d, arrow-

a b c

Fig. 13.7  cMPR images of a diseased LAD show the effect of adjusted (c, arrow). The same effect can be observed for an implanted stent
window/level settings on the apparent size of calcifications within a (arrowhead). This highlights the necessity of widening preset settings
significant stenotic lesion. Both window width decreases and level for improved luminal visualization in the presence of calcified plaque
increases from image (a) to (c). The calcifications are overestimated or stents
using narrower and brighter window settings due to blooming artifacts
138 M. H. Albrecht et al.

Fig. 13.8  In this patient,

a b
post-processing was
performed using filtered back
projection (a and c), as well
as with iterative
reconstruction (b and d). A
mixed stenosing plaque of the
LAD (arrow) and a coronary
stent in the RCA (arrowhead)
are more sharply depicted by
means of iterative
reconstruction, whereas
filtered back projection
images show a grainier image
appearance and blurry
coronary lumen

c d

Fig. 13.9  In this patient, an

a b
LAD stent is visualized on a
cMPR (a) and a sMPR (b),
both of which are based on
soft tissue kernel
reconstructions. The same
coronary is then depicted
using a dataset reconstructed
with stent (sharp) kernel on
cMPR (c) and sMPR (d). The
coronary stent is more sharply
depicted by means of the stent
kernel reducing blooming c d
artifacts and increasing our
ability to detect subtle intimal
hyperplasia in the midportion
of the stent
13  Cardiac CT: Contemporary Clinical Image Data Display, Analysis, and Quantification 139

Fig. 13.10  Segmentation and

evaluation of the coronary
a b
tree (a) of this patient
revealed subtotal occlusion of
the LAD (cMPR, b) and
severely diseased, stented
RCA (cMPR, c). By
projecting the coronary tree
onto polar maps of regional
function obtained from the c
functional CT dataset (wall
motion, d; wall thickening, e),
the significance of coronary
disease and consequent
functional impairment (shown
in cold colors) becomes
apparent

d e

a b c

Fig. 13.11 (a–c) Three-chamber views applying MPR (0.75  mm), such as valves or dissection membranes (arrowheads) are best visual-
MINIP (10 mm), and MIP (10 mm) algorithms of the same location of ized with MINIP while they are almost invisible on MIP images
a patient with aortic dissection. Note that hypoattenuating structures
140 M. H. Albrecht et al.

(see above). Some workstations will also ask for the user to to decide between discharge, observation, functional stress
mark the point of maximal stenosis in the vessel of interest. testing, and invasive catheterization.
However, defining the point of maximal stenosis is often dif- An emerging technique of CT-derived fractional flow
ficult due to the presence of bifurcations, calcifications, and reserve (CT-FFR) is gaining increasing interest by using
potentially motion artifacts that may cause erroneous computational fluid dynamic models to quantify stenosis
measurements. severity and to “simulate” how a coronary would respond to
Therefore, automated stenosis quantification using car- vasodilator stress testing. In general, FFR describes the
diac multi-detector CT (MDCT) data is a controversial issue, decrease of blood pressure over a coronary stenosis and thus
despite the fact that some investigators have shown that spe- indicates the hemodynamic relevance of a lesion. Notably,
cialized software may be just as efficient as visual grading or current CT-FFR algorithms are based on normal static CCTA
invasive coronary angiography with regard to quantitative datasets and can be retrospectively employed in addition to
analysis [6, 7]. Many studies have demonstrated the out- the standard examination when more information is required
standing negative predictive value of MDCT [8, 9]. However, [11]. To date, CT-FFR is performed mainly by outsourcing
the accurate determination of whether a stenotic lesion CCTA images to off-site, specialized core labs. This method
causes myocardial ischemia remains a challenge. Even if our takes the purely anatomic imaging a step further to virtual
images and anatomic stenosis quantification methods were physiologic testing but is still highly dependent on the qual-
impeccable, simulating the hemodynamic characteristics ity of anatomic data acquired (Fig. 13.12).
over a narrowed coronary is incredibly complex and is
dependent on numerous additional factors beyond the extent
of stenosis. For example, stenosis length, lesion geometry, Plaque Composition Analysis
blood viscosity, blood pressure, heart rate, myocardial mass,
and peripheral resistance in the capillary bed are all Current approaches to plaque composition analysis are
­contributing factors to determining the hemodynamic sig- still mainly investigational tools used in research projects;
nificance of a lesion. In an effort to standardize stenosis- however, these techniques hold great promise for future
severity reporting, a new reporting system (CAD-RADS) has CCTA evaluation and subsequent risk stratification [12].
just been published [10]. This system stratifies patients into Using threshold limits for various plaque components
risk categories and provides the probability that the source of allows for the differentiation and volumetric determina-
observed chest pain is visible in the epicardial coronary tion of calcified, dominantly fibrous, and lipid-rich plaque
arteries using CTA. The current consensus is that a radiolo- components and provides the cardiac imager with quanti-
gist – whose decision is based solely on anatomic imaging – tative measurements (Fig. 13.13). Analogous to the quan-
is unable to definitively assess the hemodynamic significance tification of total ­calcified plaque burden in the context of
of a lesion and should defer their diagnosis to the treating calcium/Agatston scoring described above, these tech-
clinicians who should ultimately exercise clinical judgment niques may allow for more comprehensive and robust

a b c

Fig. 13.12  While a stenotic calcified lesion of the LAD is clearly tation can be overcome. A CT-FFR of 0.76 was shown, with almost
depicted using only cMPR images (a) and invasive catheterization perfect agreement to invasive FFR measurement of 0.75, an indication
(arrow in b), the assessment of its hemodynamic relevance remains of a flow-­restrictive coronary stenosis. (Courtesy of Christian Tesche,
uncertain. By means of retrospective CT-FFR calculation (c), this limi- MD, Medical University of South Carolina, Charleston, SC, USA)
13  Cardiac CT: Contemporary Clinical Image Data Display, Analysis, and Quantification 141

Fig. 13.13  Entire vessel

plaque burden measurement
using user-defined thresholds
for lumen (blue), lipid-rich
plaques (yellow), fibrous
plaques (green), and
calcifications (white).
Epicardial fat and severely
fatty plaques appear red

evaluation of CCTA examinations by providing quantita- tions. Although the temporal resolution of cardiac CTA is
tive results on the total lipid-rich versus fibrous plaque sub-par relative to MRI or echocardiography, in patients
burden. Figure 13.13 demonstrates the processing steps of where the first-line modalities fail or are contraindicated,
VRT, MIP, cMPR, and plaque analysis in a patient with CTA becomes an important surrogate to assess biventricular
calcified and noncalcified plaque components. Other than size and function. Various post-processing software which
quantitative approaches, qualitative criteria for plaque offer semiautomated and automated applications for volu-
composition and vulnerability have also been described in metric analysis of cardiac chambers have become widely
the literature [13]. adopted in clinical workflows (Fig.  13.14). Most software
use a combination of thresholding and contour-detection
algorithms to delineate the actual complex shape of cardiac
Chamber Volumes and Ejection Fraction chambers, offering a tremendous advantage over conven-
tional 2D echocardiography which mainly uses geometrical
The assessment of cardiac chamber dimensions and ejection assumptions.
fraction is an important diagnostic tool for managing patients Quantification of ventricular chamber volumes is not only
with heart disease and making clinical decisions regarding useful in assessing global systolic function and ventricular
device implantation and catheter-based or surgical interven- dilation but can also be utilized in assessing regurgitant frac-
142 M. H. Albrecht et al.

a b

c Blood Vol. Mode LV RV Normal Values

Ejection Fraction % 64 47 47 - 80
Myocardial Mass ED g 119.57 - 24 - 55
Stroke Volume ml 60.03 60.05 33 - 98
ED Volume ml 93.92 127.64 58 - 154
ES Volume ml 33.89 67.6 23 - 103
Cursor Volume ml 93.92 127.64
Cardiac Output (l/min) 5.7 5.7 2.82 - 8.82

Height: 165 cm 5 ft 5 in Sex: Female

Weight: 75 kg 165 lbs BSA / m2 1.85

Fig. 13.14  Myocardial function analysis. Long- (a) and short-axis (b) Subsequently, a table with the different volumetric values is generated
views of the heart showing the segmentation of the right ventricle (yel- (c), providing the cardiac imager with quantitative data to assess global
low line) and left ventricle (red and green lines) for function analysis. cardiac function, myocardial mass, and ventricular volumes

tion or pulmonic to systemic flow (Qp/Qs) in patients with  ssessment of the Myocardium, Regional
A
valve disease or shunts. Function, and Epicardial Fat
As with any other post-processing technique, there are
certain caveats one should bear in mind before “blindly” Hypertrophic cardiomyopathy, hypertension, and aortic ste-
accepting the automatically generated results. The data are nosis are the most common causes of myocardial
only as good as the source images, and the results will be ­“overgrowth.” By delineating endo- and epicardial contours,
influenced by overall image quality, i.e., temporal resolu- the total left ventricular myocardial mass can be determined,
tion, partial coverage of the cardiac cycle, signal-to-noise which can be helpful to assess disease severity and progno-
and contrast-­to-noise relationships, streak artifacts from sis. If the contouring is performed for both end-diastole and
devices, motion artifacts, misregistrations due to arrhyth- end-systole, objective and quantitative parameters of regional
mia, or patient motion. Moreover, erroneous segmentation function can be obtained (such as wall thickening, wall
of the chamber vs. myocardium, inclusion of atria or out- motion, segmental ejection fraction), which may help in
flow tracts, and the faulty delineation of valve levels can all identifying vessel territories with impaired blood supply
lead to an incorrect estimation of cardiac volumes (Fig. 13.16). However, in routine clinical practice, the tradi-
(Fig. 13.15). tional subjective, visual inspection of regional wall motion
13  Cardiac CT: Contemporary Clinical Image Data Display, Analysis, and Quantification 143

Fig. 13.15  It is crucial to

a b
visually assess the
segmentation provided by
automated software
algorithms to avoid erroneous
functional measurements.
Automatically segmented
images are demonstrated in
systole (a and c) and diastole
(b and d). In this particular
case, the initial segmentation
(a and b) included large parts
of the left atrium (black
arrows), and this leads to
marked underestimation of
the ejection fraction (38%).
After manual adjustment of
the mitral valve level (white
arrows), the ejection fraction c d
was calculated to be 70%,
within normal range

Fig. 13.16  Dual-energy CT

a b
myocardial iodine maps. This
figure shows a short-axis
MPR image of the heart (a)
where the myocardium
appears normal. Post-­
processing of the dual-energy
CT dataset allowed the
generation of an iodine map
of the myocardium (b) which
helped detect areas of subtle
hypoperfusion (arrows)

remains the standard method of choice and is performed [14, 15]. Extracellular volume fraction (ECVF) calculations
using dedicated cardiac planes in the long and short axis, as can also be performed to quantify and assess myocardial
discussed above. fibrosis [16]. ECVF is calculated based on attenuation
While contrast-enhanced MRI remains the gold standard ­c haracteristics on unenhanced and contrast-enhanced CT
for myocardial viability assessment, the phenomenon of images, taking the patient’s hematocrit into account. These
delayed hyperenhancement has also been described as useful techniques are discussed in more detail elsewhere in this
in delineating subacute infarcts and fibrosis with cardiac CT book, but are not currently part of routine clinical evaluation.
144 M. H. Albrecht et al.

Several studies have quantified epicardial fat in both semi- codes can be used to visualize hypoperfused regions of the
automated and automated fashions and suggested an added myocardium (Fig. 13.17).
value for predicting patient outcome and risk stratification CT myocardial perfusion during vasodilator stress, based
[17]. The most sophisticated, cutting-edge post-­processing to on principles similar to MRI-based first-pass perfusion, is
date is using CTA-derived data to depict infiltrating myocar- yet another exciting field emerging and receiving growing
dial fat and delayed hyperenhancing infarcts/fibrosis fused acceptance from the radiology community. One approach
with electrophysiological mapping data [18]. This method provides a “subtraction” dataset, depicting areas of differen-
holds the promise to potentially guide radiofrequency abla- tial perfusion between rest and stress datasets (Fig. 13.17).
tion for ventricular or atrial arrhythmias. While other chapters in this book will discuss the various
ways of acquiring these images, we wanted to include an
example of a post-processed dataset where myocardial blood
Dual-Energy CT/Perfusion Imaging flow is quantitatively mapped over time with dynamic, also
known as time-resolved, perfusion CT technique (Fig. 13.18).
Another rapidly growing field in cardiac CT is the develop-
ment of gated dual-energy cardiac scans, which is further
detailed in other chapters of this book. One method to gener- Fusion with Other Modalities
ate these images is to use two different tube currents (e.g.,
80 kV and 140 kV) in two tubes of a dual-source scanner, The volumetric data from cardiac CTA is also being used to
which acquires two accurately co-registered image sets with enhance visualization and provide guidance for several
a temporal resolution of a single-source 64-slice CT scanner. additional modalities. While a comprehensive overview of
In many cases, this is sufficient for simultaneous CCTA. The this topic is beyond the scope of our current chapter, CT has
mAs setting is usually adjusted so that the noise levels are been fused with nuclear medicine SPECT or PET myocar-
similar in the two image sets. Thus, the trade-off is temporal dial perfusion imaging [19], where the power of anatomic
resolution (for instance, 165 ms vs. 83 ms), but the benefit is detail with CT is paired with the physiologic information
the ability to identify, localize, and quantify the amount of from SPECT or PET. CT has also been instrumental in the
iodine present in varying tissues. This value is correlated developments of electrophysiology, where surface mapping
with tissue perfusion, which can be quantified in mg/ml, and fused with CT data helps identify the substrate for ventricu-
may be beneficial in cases of suspected perfusion deficits and lar arrhythmia and 3D visualization of pulmonary venous
myocardial tissue characterization. Using dedicated filtering anatomy can aid in radiofrequency ablations for atrial fibril-
(D kernels) for reconstruction is essential, and various color lation [20].

Fig. 13.17  Myocardial stress

perfusion images can be
a b
subtracted from the rest
dataset. The resulting 3D VRT
(a) and superimposed polar
map (b) show green areas of
stress-induced ischemia.
(Courtesy of FUJIFILM
Medical Systems USA, Inc.)
13  Cardiac CT: Contemporary Clinical Image Data Display, Analysis, and Quantification 145

Outsourcing, Web Clients, and the “Cloud” entire vessel when using cMPR, and adjust the thickness of a
MIP or a MINIP to fully assess and understand the true
Outsourcing of the post-processing steps  – whether to the extent of the pathology. Thus, interpreters of cardiac CT
well-trained CT technologist in the department or to need to train themselves and become comfortable with
­dedicated 3D-workstation experts overseas  – has become applying these techniques to allow for a comprehensive anal-
somewhat of a routine in some centers. This approach is ysis of CT images of the heart (Fig. 13.19). A great aid for
meant to speed up interpretation by providing the interpret- this are “web-client”-based systems, allowing simultaneous
ing physician with snapshots of MIPs, 3D VRTs, and cMPRs use of a 3D workstation by multiple users from anywhere, at
of the heart and allowing them to “flip through” the pictures, any time, provided they have a computer and a reasonable
which clearly has advantages in terms of workflow and internet connection. In addition, ongoing and future research
throughput. Still, we firmly believe that in complex or chal- is needed to investigate the clinical value and cost-­
lenging cases, the ability to interact with the post-processed effectiveness of various existing visualization and quantifica-
dataset firsthand is crucial for avoiding errors in interpreta- tion techniques. Workflow optimization remains a major
tion. The clinician needs to be able to adjust the window/ concern as we strive to adopt these novel approaches into the
level settings, rotate the 3D VRT heart, swivel around the routine clinical workflow of cardiac imaging.

150.00 15.00
[1] VOI
[1] VOI

[3] ROI

[3] ROI (MPR)

Mean: 217.18 mL/100mL/min

a 0.00
b 0.00

HU
370 5 ROI/VOI Legend AVG MBF MBV FE PCBV
AIF

295 4
[1] VOI 85.92 130.47 16.08 94.25 8.58
220 3
[3] ROI 99.99 217.18 26.29 95.36 19.50
145 2
[3] ROI
70 1
-5
c0 5 10 15 20 25 30 35
sec d

Fig. 13.18  Quantitative myocardial perfusion analysis. Color-coded tion during the dynamic sequential image acquisition. As a result, a
maps of myocardial blood flow (a) and myocardial blood volume (b) table of objective calculated values (d), such as myocardial blood flow
illustrating myocardial perfusion. Note the apical area of hypoperfusion (MBF) and myocardial blood volume (MBV), can be generated for
(blue area). ROIs can be placed to generate arterial input functions quantitative myocardial perfusion analysis
(AIF) and perfusion curves (c), where each point indicates the attenua-
146 M. H. Albrecht et al.

Fig. 13.19  Working up a

a b
case: 3D VRT showed a
suspicious bifurcation of the
LAD and the first diagonal
(a). MIP showed the presence
of mixed, calcified and
noncalcified components (b).
cMPRs of the LAD (c) and
color-coded plaque analysis
(d) showed the lipid-rich
nature of the noncalcified
portions of the lesion

c d

6. Arbab-Zadeh A, Hoe J. Quantification of coronary arterial stenoses

by multidetector CT angiography in comparison with conventional
References angiography. JACC Cardiovasc Imaging. 2011;4:191–202.
7. Boogers MJ, Schuijf JD, Kitslaar PH, et al. Automated quantifica-
1. Fishman EK, Ney DR, Heath DG, et al. Volume rendering versus tion of stenosis severity on 64-slice CT. JACC Cardiovasc Imaging.
maximum intensity projection in CT angiography: what works best, 2010;3:699–709.
when, and why. Radiographics. 2006;26:905–22. 8. Stein PD, Yaekoub AY, Matta F, Sostman HD. 64-slice CT for diag-
2. van Ooijen PMA, Ho KY, Dorgelo J, Oudkerk M. Coronary artery nosis of coronary artery disease: a systematic review. Am J Med.
imaging with multidetector CT: visualization issues. Radiographics. 2008;121:715–25.
2003;23:e16. 9. Janne d’Othée B, Siebert U, Cury R, et al. A systematic review on
3. Geyer LL, Glenn GR, De Cecco CN, et al. CT evaluation of small-­ diagnostic accuracy of CT-based detection of significant coronary
diameter coronary artery stents: effect of an integrated circuit detec- artery disease. Eur J Radiol. 2008;65:449–61.
tor with iterative reconstruction. Radiology. 2015;276:706–14. 10. Cury RC, Abbara S, Achenbach S, et  al. CAD-RADS™: coro-
4. Hecht HS.  Coronary artery calcium scanning. JACC Cardiovasc nary artery disease – reporting and data system. J Am Coll Radiol.
Imaging. 2015;8:579–96. 2016;13(12 Pt A):1458–1466.e9.
5. Busch S, Johnson TRC, Nikolaou K, et  al. Visual and auto- 11. Gaur S, Taylor CA, Jensen JM, et  al. FFR derived from coro-
matic grading of coronary artery stenoses with 64-slice CT nary CT angiography in nonculprit lesions of patients with recent
angiography in reference to invasive angiography. Eur Radiol. STEMI. JACC Cardiovasc Imaging. 2016;10:424–33.
2007;17:1445–51.
13  Cardiac CT: Contemporary Clinical Image Data Display, Analysis, and Quantification 147

12. Gitsioudis G, Schüssler A, Nagy E, et  al. Combined assessment contrast-­enhanced CT: validation against histologic findings.
of high-sensitivity troponin T and noninvasive coronary plaque Radiology. 2013;269:396–403.
composition for the prediction of cardiac outcomes. Radiology. 17. Spearman JV, Meinel FG, Schoepf UJ, et al. Automated quantifica-
2015;276:73–81. tion of epicardial adipose tissue using CT angiography: evaluation
13. Maurovich-Horvat P, Schlett CL, Alkadhi H, et al. The napkin-ring of a prototype software. Eur Radiol. 2014;24:519–26.
sign indicates advanced atherosclerotic lesions in coronary CT 18. Truong QA, Thai W, Wai B, et  al. Myocardial scar imaging by
angiography. JACC Cardiovasc Imaging. 2012;5:1243–52. standard single-energy and dual-energy late enhancement CT:
14. Lardo AC.  Contrast-enhanced multidetector computed tomogra- comparison with pathology and electroanatomic map in an experi-
phy viability imaging after myocardial infarction: characterization mental chronic infarct porcine model. J Cardiovasc Comput
of myocyte death, microvascular obstruction, and chronic scar. Tomogr. 2015;9:313–20.
Circulation. 2006;113:394–404. 19. Thilo C, Schoepf UJ, Gordon L, et  al. Integrated assessment of
15. Gerber BL.  Characterization of acute and chronic myocar-
coronary anatomy and myocardial perfusion using a retractable
dial infarcts by multidetector computed tomography: compari- SPECT camera combined with 64-slice CT: initial experience. Eur
son with contrast-enhanced magnetic resonance. Circulation. Radiol. 2009;19:845–56.
2006;113:823–33. 20. Cochet H, Dubois R, Sacher F, et al. Cardiac arrhythmias: multi-
16. Bandula S, White SK, Flett AS, et  al. Measurement of myo- modal assessment integrating body surface ECG mapping into car-
cardial extracellular volume fraction by using equilibrium diac imaging. Radiology. 2014;271:239–47.
 Workflow Optimization
14
Thomas Allmendinger, Andrew N. Primak,
and Christian D. Eusemann

Cardiac CT imaging is considered to be complicated due to a  Brief Recap of the Available Tools

A
wide variety of rules, recommendations, and different guide- for Cardiac CT Data Acquisition
lines in combination with a rapid technological development
in the last 15 years. However, the primary underlying acqui- A detailed technical description of the state-of-the-art car-
sition challenge stems from the fact that cardiac imaging is a diac CT technology and the different acquisition modes was
patient-dependent optimization problem with diagnostic provided in Chap. 6. Therefore, only a brief summary and
image quality being the first and X-ray radiation dose the refresher of the technical nomenclature is provided here.
second important degree of freedom which need to be bal- From a cardiac CT point of view, the oldest most widely
anced against each other. If there were absolutely no concern used scanning technique is retrospective ECG-gated spiral
for radiation dose, cardiac CT acquisition would most likely scanning also called helical scanning. The scan is character-
be identical to the approach 15 years back – a retrospective ized by a continuous data acquisition in combination with a
ECG-gated spiral scanning without any tube current modula- continuous table motion at relatively low table feed (low spi-
tion enabling maximum freedom for image reconstruction ral pitch) with simultaneous recording of the patient’s
based on the acquired data. However, as this approach con- ECG. After the data acquisition, the recorded ECG is utilized
flicts with the ALARA principle, today’s approaches are as guidance to select those data intervals in the CT data set
more finely tuned. that were measured in different heartbeats in the same user-­
This chapter has two parts. The first part focuses on acqui- selected cardiac phase, subsequently building a consistent
sition strategy recommendations applicable to a broad spec- image volume. Over the years the dose efficiency and com-
trum of state-of-the-art CT systems. In the second part, fort of this “straight-forward” approach were improved by
specific key elements of the patient preparation, important technological features like the tube current modulation dur-
for a successful cardiac workflow, are discussed  – patient ing the individual cardiac cycles [1, 2], automatic calculation
positioning, ECG, and breath-hold training. of the optimal spiral pitch based on the patient’s heart rate
[3], and algorithms for optimized reconstruction with mini-
mal cardiac motion [4, 5].
In contrast to the spiral acquisition, the sequence tech-
nique, also called a step-and-shoot acquisition, was devel-
oped from the opposite direction. In its earliest form, an axial
CT scan without table movement was performed in a single
prospectively ECG-triggered cardiac phase. The CT system
acquired scan data covering a sub-volume of the patient’s car-
diac anatomy corresponding roughly to the total z-width of
T. Allmendinger (*)
Department of Computed Tomography, Siemens Healthcare the CT detector. The duration of such a sub-volume acquisi-
GmbH, Forchheim, Germany tion was limited to a quick-scan range to enable the recon-
e-mail: [email protected] struction of a single set of images at this position corresponding
A. N. Primak to around three quarters of a rotation. Afterwards the table is
Department of Diagnostic Imaging, Siemens Healthineers, moved to the next z-position, and the next axial CT scan is
Malvern, PA, USA
performed at the same cardiac phase in a later cardiac cycle.
C. D. Eusemann This way, the heart volume is sequentially covered by multi-
Department of Collaborations, Siemens Medical Solutions USA,
ple axial scans. By design the table feed of such a sequential
Inc., Malvern, PA, USA

© Humana Press 2019 149

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_14
150 T. Allmendinger et al.

scan is only determined by geometrical factors of the CT sys-  ustomizing Cardiac Data Acquisition
C
tem and independent of the patient’s heart, which contributes to the Individual Patient
already significantly to the improved dose efficiency of this
scan acquisition. By further limiting the data acquisition to a Whether to scan in a retrospective gated spiral mode or a
single cardiac phase as it was done in the early years, the prospectively triggered sequence mode is considered the
sequence technique was considered optimized for “low-dose basic first decision step for a cardiac CT acquisition. As
scanning” however at the cost of a limited robustness. described above, the very steep, near steplike, transition
Technological developments of later years vastly increased from a retrospective spiral without tube current modulation
the flexibility and robustness of sequential techniques by to a prospective sequence limited to a single cardiac phase
introducing features like arrhythmia rejection and recovery should be a thing of the past. Basically all modern CT sys-
scanning [6, 7], a flexible definition of the desired cardiac tems enhanced the capabilities of their respective sequence
phase acquisition window and an automatic adjustment of acquisition modes with an automatic heart rate rhythm-­
this acquisition based on the cardiac rhythm [8]. dependent acquisition window adjustment, the possibility to
A common remaining challenge for both acquisition tech- manually select wide acquisition windows, which allow cov-
niques described above is the proper alignment and registra- erage of end-systole and end-diastole phases, combined with
tion of the images acquired and reconstructed in stacks over mechanisms for handling extra-systolic events (PVCs, SVC).
multiple cardiac cycles. A hardware-based system solution The reasons for sticking to retrospective spiral acquisitions
to this “last” challenge of cardiac imaging, which is cur- are a possible lack of some of those features in sequence
rently available in the form of several high-end CT systems mode on some scanner platforms and an improved shorter
from different manufacturers, is provided in two rather fun- total scan time for high heart rates due to the commonly
damentally different ways – single-source volume scanning applied heart rate-dependent automatic pitch adaptation.
with wide-area detector systems [9, 10] and ECG-triggered This shortening of the total scan time is important for higher
high-pitch scanning with dual-source systems [11]. heart rates as it counters the effect of increased cardiac
As of today, cardiac volume scanning is offered by two throughput, which would otherwise trigger a necessary
vendors. With a detector collimation of 320  ×  0.5  mm at increase of the contrast injection time as it is the case in high
0.28 s rotation time (Toshiba Aquilion One, Toshiba Medical heart rate sequence acquisitions.
Systems Corporation, Tokyo, Japan) and 256  ×  0.625  mm Besides the scan mode, the second decision to be made is
detector collimation at 0.28 s rotation time (GE Revolution, the targeted cardiac phase, typically choosing between end-­
GE Healthcare, Waukesha, USA), these systems are capable diastole, end-systole, or both and whether the phase target is
of a 16-cm volume coverage in the z-axis direction at the limited to a single time point or if a slightly wider acquisition
isocenter, large enough to scan the entire heart in one beat in window is configured to enable the possibility to search
most of the clinical cases, therefore avoiding stair-step and locally for the best image reconstruction time. In addition,
banding artifacts and problems with inconsistent contrast due to the nature of the contraction and filling of the cardiac
enhancement in different sub-volumes. Even if these wide-­ chambers throughout the cardiac cycle, the question whether
area detector CT systems do not need to answer the basic to make the phase definition in a relative fashion (in units of
scan mode acquisition technique question, the planning of percent relative to the 100% of a cardiac cycle RR interval)
the suitable cardiac phases and the width of the acquisition or in an absolute millisecond-based parameter setting needs
window remain. Therefore, these systems are also discussed to be answered. A more detailed discussion on the acquisi-
as a part of the single-source discussion below. tion and reconstruction consequences of the phase unit selec-
As an alternative, single-heartbeat volume scanning can tion is done further below.
also be achieved with dual-source CT systems by means of a Finally, similar to the desired cardiac acquisition phase
spiral acquisition at high pitch values of around 3.2–3.4. and unit, the target for the image reconstruction needs to be
Depending on the generation of the CT system, this approach defined as well. However, this decision is secondary to the
yields table feeds of 450 mm/s (SOMATOM Definition Flash, acquisition as one can add multiple reconstructions within
Siemens Healthineers) or 736  mm/s (SOMATOM Force, the available acquired data set or utilize motion-optimized
Siemens Healthineers) enabling the coverage of an entire car- image reconstructions [4, 5]. Nevertheless, a manual selec-
diac volume in the diastolic rest phase by controlling the tion of the target phase unit is, in most cases, still required,
acquisition in a prospective ECG-triggered fashion. The pro- which could potentially have significant effects on the final
spective nature of the scan mode limits the application of the image quality.
scan mode in patients with irregular heart rate, as discussed A “correct” level of tuning and separation into individual
below; however the possibility to cover an extended range scan parameter sets or classes cannot be easily specified due
beyond the heart anatomy in a single acquisition makes it a to the fact that a very detailed description into many indi-
clinical valuable tool for CABG or TAVI scans [12, 13]. vidual classes based on clinical question, heart rate, cardiac
14  Workflow Optimization 151

rhythm, patient size, age, sex, and possible other additional with a necessary focus on the requirements of cardiac CT
factors might be desirable from an academic point of view. imaging. The established clinical classifications of cardiac
However, such a detailed approach is not feasible in a daily arrhythmia are too detailed and therefore not really suitable
routine, if the protocol adjustments are made by manual for this kind of application. A simple, robust, and mostly
parameter modifications. As a consequence, the approach descriptive classification scheme which can be applied is the
discussed below is, to some extent, a compromise with a lim- utilization of the three different cardiac rhythm groups – reg-
ited amount of individual classes still maintaining a certain ular, unstable, and arrhythmic. A regular heart rate is classi-
high level of detail. fied as a normal sinus rhythm with minimal variation in
Of note, the subgroups described below can be combined correlation with the breathing rate of the patient. An unstable
if further simplification is desired. rhythm is characterized by continuous beat-to-beat variation
The first selection step is indication-driven, based on the of up to 8–10 BPM or multiple extra beats within a few
required diagnostic image quality level. Even if there is a cycles. An arrhythmic cardiac rhythm, as it is used in this
very wide range of clinical questions, the two most common context, is characterized by the fact that the determination of
applications on opposite sides of the image quality spectrum a valid average heart rate is already difficult if looking at
are coronary calcium scoring (CaSc) and coronary CT angi- 10–15 cardiac cycles. Typical examples of ECG yielding a
ography (CCTA). CaSc, applied since the early 1990s and classification like that would be a bigeminy rhythm or atrial
originally driving cardiac CT system development with dedi- fibrillation.
cated designs (e.g., EBCT [14]), has nowadays become a In addition to the clinical and physiological grouping
“screening-like test” focusing primarily on limiting X-ray detailed above, the CT hardware itself needs to be clustered,
radiation dose. The image quality requirements for accurate to some extent, to allow the development of a practical deci-
calculation of a calcium score are rather relaxed even allow- sion guide. The solution chosen here splits the guidance into
ing the evaluation in non-ECG-gated scans [15]. In contrast four different blocks illustrated by Figs. 14.1, 14.2, 14.3, and
to this, the image quality requirements for CCTA were rela- 14.4. These blocks are separated according to the clinical
tively high to begin with, in terms of spatial resolution, visi- indication, CCTA (Figs. 14.1 and 14.2) and CaSc (Figs. 14.3
ble contrast, and absence of cardiac motion. These and 14.4), and one key technical differentiator, the temporal
requirements tightened in recent years, due to the fact that resolution for cardiac imaging. Based on this temporal reso-
the reporting of additional coronary plaque features, beyond lution difference, the guidance is also separated into single-­
a “simple” stenosis quantification, became clinically rele- source (Figs.  14.1 and 14.3) and dual-source technologies
vant, driving the image quality needs even further [16]. (Figs. 14.2 and 14.4).
Besides this obvious clinical choice, a characterization of Illustrated in Fig.  14.1 is a CCTA classification scheme
the individual patient’s cardiac heart rate and rhythm needs for single-source CT systems. The classification assumes
to take place in order to adjust the scan parameters that the prospective sequence has modern features, which
appropriately. can handle extra beats properly and allow a separate defini-
The determination of an average heart rate is rather tion of the start and end of a cardiac phase. To provide some
straightforward and often provided automatically by the guidance despite the wide range of available rotation times,
attached ECG system. One should however be mindful that ranging from 0.27 s to 0.35 s [17], and detector collimations,
these values are typically derived during a free-breathing ranging from 2 cm to 8 cm, the cutoff values for the classifi-
patient state, while the CT exam is normally done with the cation into low, moderate, and high heart rates are made sys-
patient holding his/her breath. Patient heart rates typically tem dependent as shown on the left side of Fig. 14.1. As a
drop by an average of 4–6 BPM during a breath-hold and general pattern, one can explain the approach as follows,
often in combination with a reduction of the overall heart starting in the bottom left with a regular low rhythm, which
rate variability. In general, this behavior is beneficial for the enables the utilization of an end-diastolic single-phase car-
CT acquisition. One still needs to pay attention in cases of an diac sequence acquisition. With increasing heart rate or vari-
automatic pitch selection with retrospectively gated spiral ability, the uncertainty of where to image optimally is
scans, as the heart rate determining the pitch value is mea- mitigated by increasing the acquisition window, slightly in
sured during the free-breathing state of the patient and is set the unstable direction or considerably in the elevated heart
prior to the scan, with a safety reserve of around 8–10 rate region, as it is undetermined if an end-diastolic or an
BPM. In rare cases, the heart rate of patients can drop below end-systolic image reconstruction provides the more
these values during breath-hold, requiring a manual pitch ­diagnostic images. For high heart rates, the full transition
selection in order to avoid image artifacts from data gaps. into a solely end-systolic acquisition is natural as diastole
A standardized method of calculation or classification of simply becomes too short for imaging. The explanation for
cardiac rhythm, comparable to a mean or median value for the strategy of end-systolic imaging, based on an absolute
the heart rate, is not established and agreed on a wide basis millisecond phase definition, can be found in the next section
152 T. Allmendinger et al.

Coronary CTA – Single Source Systems

Regular (±3-4 BPM) Unstable (±10 BPM) Arrhythmic (> ±10 BPM)

End-Systolic Scan End-Systolic Scan Not recommended,

Phase: 30-40% RR Phase: 200ms-400ms RR If necessary end-systolic
> 80 BPM
> 75 BPM
High

Mode: Spiral (BiSegment) Mode: Spiral (BiSegment) millisecond strategy.

Recon: % Syst. Recon: ms Syst.

Unspecific RR Target Unspecific RR Target Questionable !

Systolic + Diastolic Scan End-Systolic Scan
Moderate

< 80 BPM

Systolic + Diastolic Scan

< 75 BPM

Phase: 30-80% RR Phase: 30-80% RR Phase: 200ms-400ms RR

Mode: Spiral Mode: Spiral Mode: Spiral
Recon: % Diast. Recon: % Diast. + ms Syst. Recon: ms Syst.

End-Diastolic Scan End-Diastolic, wider window Unspecific RR Target

< 65 BPM

< 70 BPM

Phase: 70% RR Phase: 60-80% RR Systolic + Diastolic Scan

Low

Mode: Sequence Mode: Sequence Phase: 30-80% RR

Recon: % Diast.e Recon: % Diast. Mode: Spiral
Recon: % Diast.

System 1 : Single Source, 64×0.6mm, rot. time ≥ 0.33s, e.g. Siemens Definition AS64
System 2 : Single Source, 128×0.6mm, rot. time ≤ 0.3s, e.g. Siemens Definition Edge
System 1

System 2

Fig. 14.1  Practical decision guide – coronary CT angiography using a trated for a mid-range and high-end single-source scanner system with
single-source imaging system. The cutoff values for the different heart different temporal resolution and detector coverage
rate ranges depend on the capabilities of the CT scanner system, illus-

as it is a clinically proven concept established for dual-source second and third dual-source generations. The most notice-
CT. Even though this concept can be applied to single-source able difference compared to the single-source approach is
systems, it does (as all end-systolic approaches) depend on a the lack of an undetermined region, which would require
very good temporal resolution and, therefore, is more suit- both end-systolic and end-diastolic combined acquisitions,
able for dual-source systems. but a rather sharp transition from a diastolic to a systolic
Handling the heart rate rhythm in dedicated separate approach at a certain heart rate. The change of the cardiac
groups with specific measures, switching between percent- phase unit toward absolute milliseconds is an effective mea-
age diastolic or millisecond systolic acquisitions, is to a sure to limit the radiation dose despite an increasing irregular
large extent attributed to the previously described stack heart rate. The reason for this behavior is simply the fact that
alignment and the registration challenge of systems acquir- all percent-based definitions are relative with respect to the
ing only a part of the heart anatomy in single detector z-cov- best guess time point into the future where the R-peak clos-
erage. With wide-area detector systems, the dedicated ing the current cardiac cycle should appear. If the quality or
handling of the cardiac rhythm can be neglected to some precision of this estimate deteriorates due to an irregular
extent and the scan parametrization can be solely based on rhythm, nearly all modern CT systems react with an auto-
the regular heart rate in combination with the average heart matic widening of the acquisition window, which can lead to
rate determined by the ECG. a greatly increased radiation dose. If the cardiac target phase
Figure 14.2 illustrates the CCTA selection scheme for is configured by an absolute millisecond approach, aiming at
dual-source systems. The grouping into the three different the end-systole, no prospective estimate is required, and the
heart rate levels (low, moderate, and high) is also system necessary X-ray control time points are derived in a stop-
dependent reflecting the technical advancement between the watch-like manner, without any effect on the width of the
14  Workflow Optimization 153

Coronary CTA – Dual Source Systems

Regular (±3-4 BPM) Unstable (±10 BPM) Arrhythmic (> ±10 BPM)

End-Systolic Scan End-Systolic Scan End-Systolic Scan

> 75 BPM

> 80 BPM

Phase: 30-40% RR Phase: 150ms-350ms RR Phase: 150ms-350ms RR

High

Mode: Spiral Mode: Spiral Mode: Spiral

Recon: % Syst. Recon: ms Syst. Recon: ms Syst.

End-Diastolic End-Diastolic End-Systolic Scan

Moderate

Phase: 60-80% RR Phase: 60-80% RR Phase: 200ms-400ms RR

< 75 BPM

< 80 BPM

Mode: Sequence Mode: Sequence Mode: Sequence

Recon: % Diast. Recon: % Diast. Recon: ms Syst.

End-Diastolic Scan End-Diastolic End-Systolic Scan

Phase: n.a. Phase: 70% RR Phase: 200ms-400ms RR
< 65 BPM

< 70 BPM

Mode: High-Pitch
Low

Mode: Sequence Mode: Sequence

Recon: n.a. Recon: % Diast. Recon: ms Syst.

System 1 : Dual Source, 128×0.6mm, rot. time 0.28s, e.g. Siemens Definition Flash
System 1

System 2

System 2 : Dual Source, 196×0.6mm, rot. time 0.25s, e.g. Siemens Force

Fig. 14.2  Practical decision guide – coronary CT angiography using a eration. This is especially important for the definition of the “low heart
dual-source imaging system. The cutoff values for the different heart rate” category, where second- and third-generation systems differ in
rate ranges depend on the capabilities of the dual-source scanner gen- their upper limits for dual-source high-pitch scanning

acquisition window. A detailed description of this approach high-pitch scanning, the FAST Phase feature on Siemens
including its clinical results in terms of dose and image qual- systems with syngo CT VB10 [22] or the Auto Gating fea-
ity can be found here [18–20]. ture on GE Revolution CT [23].
The approaches for CaSc illustrated in Figs. 14.3 and 14.4 However, a truly intelligent system would need to include
can be summarized as single-phase limited acquisitions way more patient- and indication-specific input factors in
focusing primarily on radiation dose in combination with a order to reach a dedicated tailored scan and reconstruction
simplified grouping and classification scheme for easy solution, which would lead to a dramatic increase in the level
application. of details in the decision tree. While such a system can easily
handle a finely tuned classification scheme from a technical
point of view, the user operating the system still needs to be
 uture: Fully Automated Cardiac
F empowered to understand the underlying reasoning of the sys-
CT Systems? tem (i.e., “why will this patient be scanned and reconstructed
in this specific way?”) beyond the actual displaying of all scan
In the advent of self-driving cars and deep learning tech- and reconstruction parameters (“what will be done?”), as he/
nologies which are rapidly evolving, one can raise the ques- she still bears the full responsibility for the CT scan and there-
tion if such a manual set of recommendations as presented fore for patient image quality and radiation dose. This “trust in
above are still state of the art and will be relevant in a few automation” challenge is well known [24, 25], and a satisfying
years. So far, initial steps have been undertaken by different answer needs to be provided in the context of CT systems by
manufacturers to support the user in their decision and the manufacturers to enable the widespread acceptance of par-
selection process. Examples are the FlashCheck [21] for tially or fully automated CT scanners.
154 T. Allmendinger et al.

Calcium Scoring – Single Source Systems

Regular (±3-4 BPM) Unstable (±10 BPM) Arrhythmic (> ±10 BPM)

End-Systolic Scan
> 75 BPM

> 80 BPM

Phase: 250ms RR
High

Mode: Sequence
Recon: ms Syst.

End-Diastolic Scan
Moderate

< 75 BPM

< 80 BPM

Phase: 70%
Mode: Sequence
Recon: 70%
< 65 BPM

< 70 BPM
Low

System 1 : Single Source, 64×0.6mm, rot. time ≥ 0.33s, e.g. Siemens Definition AS64
System 2 : Single Source, 128×0.6mm, rot. time ≤ 0.3s, e.g. Siemens Definition Edge
System 1

System 2

Fig. 14.3  Practical decision guide – coronary calcium scoring using a what reduced overall image quality. This approach is reflected by a
single-source imaging system. The primary focus of calcium scoring guide, which is drastically simplified compared to the CTA approach
exams is the minimization of X-ray dose therefore accepting a some- limiting it to a single decision point

 atient Positioning, ECG, and Breath-Hold

P s­ ystem or if the entire chest should be placed in the system
Training ­isocenter. Depending on the individual patient anatomy, the
two approaches can yield different table height settings,
Preparing the patient prior to the CT scan includes multiple between 5 cm and 10 cm, and can significantly influence the
steps, typically performed by the technologist staff, prior to results of the exam. Unfortunately, a decisive answer to this
the actual scan, ranging from communicating to the patient question is difficult as both approaches have their merits and
about the upcoming exam, checking laboratory values up to limitations. A suitable approach needs to be picked individu-
the management of contrast media administration. From this ally based on the available CT hardware, requested clinical
wide range of responsibilities, three are discussed in greater procedure, and body habitus.
detail here as they are an integral part of a successful cardiac Some CT systems, e.g., GE VCT scanners [26], enable
CT examination – patient positioning, ECG acquisition, and the manual selection of a dedicated cardiac bowtie filter,
breath-hold training. which limits the X-ray radiation to a smaller scan field of
view (SFOV). This selection needs to be done prior to the
scan and therefore, if selected, mandates the positioning of
Patient Positioning the heart into the isocenter of the system as it leads to a limi-
tation of the reconstruction field of view to values ranging
When positioning the patient on the CT table, one question from 32 cm to 40 cm. On the other hand, the benefit of such
arising in cardiac exams is whether to put the heart itself, as an aggressive bowtie filter approach is a possible lowering of
anatomic organ of interest, into the isocenter of the CT the radiation dose of up to 40% [26]. The routine use of small
14  Workflow Optimization 155

Calcium Scoring – Dual Source Systems

Regular (±3-4 BPM) Unstable (±10 BPM) Arrhythmic (> ±10 BPM)

End-Systolic Scan
> 75 BPM

> 80 BPM
Phase: 250ms RR
High

Mode: Sequence
Recon: ms Syst.

End-Diastolic Scan End-Systolic Scan

< 75 BPM

< 80 BPM
Moderate

Phase: n.a. Phase: 300ms RR

Mode: High-Pitch Mode: Sequence
Recon: n.a. Recon: ms Syst.

or

End-Diastolic Scan
Phase: 70%
< 65 BPM

< 70 BPM

Mode: Sequence
Low

Recon: 70%

System 1 : Dual Source, 128×0.6mm, rot. time 0.28s, e.g. Siemens Definition Flash
System 1

System 2

System 2 : Dual Source, 196×0.6mm, rot. time 0.25s, e.g. Siemens Force

Fig. 14.4  Practical decision guide – coronary calcium scoring using a at the same time optimizing the acquisition in the direction of an overall
dual-source imaging system. The guide is drastically simplified reflect- low X-ray dose
ing the lowered image quality expectations for this kind of exam while

bowtie filters and small FOV will both reduce radiation dose ence relying on a single AP topogram (or scout) for the esti-
and limit incidental findings related to non-cardiac disease mation of the necessary scan exposure settings. If in this case
(primarily non-cancerous lung nodules). Whether these inci- the positioning of the patient is chosen systematically off-­
dental findings improve outcomes is still debated in the sci- center, the image aberration introduced in the topogram
entific literature [27–29]. image due to the fan beam nature of the CT system could
Other CT systems, e.g., Siemens scanners, do utilize car- yield an unwanted bias in the AEC estimation. In case of a
diac bowtie filters focusing the radiation dose more strongly personal preference for the heart isocenter placement
to the isocenter of the scanner yet still enable the reconstruc- approach, it is recommended to add a second lateral topo-
tion images in the full field of view (on most scanner systems gram to the exam, which provides the full attenuation infor-
50  cm) at the price of slightly elevated image noise in the mation to the AEC system, therefore removing all possible
periphery. The achievable dose reduction comparing to a bias [31].
standard body bowtie setting with the corresponding cardiac Another quite interesting, more specialized technique is
settings is approximately 20–30% in this approach [3, 30]. the less known so-called breast displacement technique for
This freedom naturally leads to the recommendation to posi- female patients. The idea is simply to push the breasts up so
tion the chest into the isocenter as it enables the reconstruc- they are covering a smaller portion of the scan range. A recent
tion of the heart and lung anatomy in separate reconstruction publication describes this approach in a cardiac CT practice,
jobs in its entirety. An important benefit of this approach can which is accomplished by using the Velcro patient positioning
be derived from the fact that, on some systems, the automatic strap attached to the scanner table. This “reshaping” of the
exposure control (AEC) of a scanner is due to user prefer- patient anatomy prior to the acquisition of the topogram (or
156 T. Allmendinger et al.

scout) images supports overall dose reduction by keeping the removed, along with any natural oils and dirt that impede
more radiation-sensitive breast tissue partially outside of the electrical flow and thus create a resistance to signal quality.
cardiac scan range in combination with the reduced patient Prepare the skin by cleaning the area for application of elec-
chest attenuation, which leads to an improved image quality trodes. This can be done with soap and water. Alcohol may
or alternatively a lowered exposure [32]. only be used on particularly oily skin as alcohol dries out the
skin while removing the oil. Wipe the skin area multiple
times with gentle pressure and an abrasive skin prep pad
ECG (e.g., 3M 2236 One Step Skin Prep).
The use of suitable quality electrodes is an important part
The acquisition of quality ECG data is critical to provide a of the overall monitoring process. When selecting electrodes
reliable and resilient signal for controlling the X-ray exposure the following simple guidelines apply:
during a cardiac CT examination. This statement is valid for
all acquisition strategies, a retrospective spiral with ECG puls- 1. All electrodes selected should be of the same brand and
ing, which modulates the tube current, based on the phase in type to help minimize noise.
the cardiac cycle and all prospectively triggered acquisitions, 2. Adhere to all instructions for use found on the electrode
which determine the X-ray on and off time points from the packaging (e.g., ‘Use by’ date).
ECG information. Poor ECG quality can result in the CT sys- 3. Electrodes that are packaged together in large quantities
tem’s inability to recognize the R-wave, thus affecting X-ray should be used shortly after the packaging is open or the
exposure with a high likelihood of increasing the patient dose packaging should be tightly resealed since extended
substantially combined with a potentially detrimental effect on exposure to air will dry out electrodes prematurely and
the final image quality of the exam. This section only provides reduce their adhesive and conductive properties.
a brief summary of essential elements; more detailed guidance 4. Using wet gel electrodes (e.g., 2700 Cleartrace 2) is ben-
can be found in these white papers [33, 34]. eficial for the needs of a short-term ECG in combination
Citing a relatively old article from 1984, in Critical Care with a CT exam.
Nursing, this statement discussing ECG artifacts still holds
true despite advancements in ECG devices: “Improvement in Recent technological developments support the efforts of
electrode preparation techniques and a better understanding the clinicians at this critical patient-machine interface in the
of the sources of artifact can enhance equipment performance, form of an automated ECG electrode impedance measure-
resulting in improved patient assessment, more effective uti- ment combined with reporting the result in real-time on the
lization of nursing time, and reduced operating costs” [35]. gantry display of the CT system (ECG Monitor, Siemens
Improving ECG quality is an important factor built on the Drive, Siemens Force). The example of Fig. 14.5 illustrates
following key elements: correct placement, good skin prepa- such a support mechanism yielding, in case of a poor critical
ration, use of quality electrodes, proper electrode application, left arm electrode connection, a very high electrical imped-
good electrode-to-patient contact, and proper lead selection. ance and a red warning triangle (Fig. 14.5a), and an accept-
When positioning ECG electrodes, it is important to avoid able, but not very good, “yellow” warning (Fig.  14.5b). In
bony protuberances and scar tissue and keep enough distance case of good electrode connections, the patient view is
between the individual electrodes. Bones and scar tissue can replaced by a display of the current heart rate, the overall
interfere with the electrical impulse transmission. Larger dis- ECG signal strength, and the selected ECG lead channel
tances between the electrodes typically provide signals with (Fig. 14.5c).
larger amplitude. One should avoid, if possible, positioning
electrodes in the scan field, i.e., near or at the level of the
heart, because ECG trace artifacts may be generated when  reath-Hold: Training Prior to the Scan
B
electrodes are hit by the X-ray beam. and Image Analysis After the Scan
The skin is a poor conductor of electricity, and artifacts
from the outer layer of the skin (epidermis) are more trouble- It is well known that the suspension of the respiration during
some than other types of artifacts because it is difficult to scan acquisition is of great importance for a successful and
filter electronically, and its amplitude is often larger than the high-quality CT of the upper chest area such as a cardiac CT
ECG signal [36]. Unfortunately, clinical studies have shown exam. On the other hand, one could get the impression that this
that the electrode-skin interface is frequently overlooked as a topic is treated as the “annoying poor cousin” of cardiac
major source of artifacts affecting many electrophysiological CT. For example, an otherwise very good book on technologist
recordings [37]. Although skin preparation can add a little training devotes only few lines of text to this crucial topic [38].
time, the extra effort may reduce the time spent dealing with The first part of the section provides a comprehensive sum-
a poor CT examination. To prepare the skin for electrode mary of respiratory instructions based on this excellent detailed
placement, dry, dead epidermal layers of the skin must be publication [39]. In the second part, illustrations of breathing
14  Workflow Optimization 157

Fig. 14.5  Example of a poor

a b
and acceptable ECG electrode
connection indicated on the
gantry display either by (a)
red (“bad impedance”) or (b)
yellow (“weak impedance”)
warning signs. A proper
high-quality electrode
connection is indicated by (c)
displaying average heart rate,
overall ECG signal strength,
and ECG lead channel

artifacts in cardiac exams are shown. They provide some image repeatability. However, it needs to be checked initially if the
analysis strategies answering the question: “How to distinguish available built-in breath-hold commands are suitable for the
breathing artifacts from cardiac motion artifacts?” specific needs of a cardiac exam as they tend to be quite short
in order to minimize effects on the inter-scan delay of auto-
matically coupled scan ranges. A possible solution, which is
Breath-Hold Training offered by most modern CT systems, is the possibility to
record and save an individual breath-hold command of the
All explanations to the patient should be as simple and figu- technologist during the setup of the scan protocol, which can
rative as possible. A cardiac CT exam might be a simple rou- be applied if the need for such a dedicated command arises.
tine task for a radiologist or technologist; however, it can be Finally, there are often additional CT scans occurring prior
an unusual and potentially alienating experience for most to the more critical contrast-enhanced cardiac exam (e.g.,
patients. Using brief, simple, and figurative sentences topogram, a test bolus series scan, CaSc scan). Briefly review-
explaining the procedure and what is expected from the ing these images for the presence of breathing motion arti-
patient serves as framework for the following steps. facts can prevent a failed exam by reinstructing the patient.
The breathing instructions should be performed in the
same body position as the CT exam that is typically supine
and with the arms extended behind the head. Ideally, training Breathing Artifacts in Cardiac Images
should be performed with the patient positioned on the CT
table simulating the experience of the later CT scan. The In Fig. 14.6 one can see examples of curved planar reformat
technologist or radiologist should proactively demonstrate (CPR) images of the main coronaries with clear indications
the breathing maneuver as many patients are not properly of stack registration and misalignment artifact of the under-
instructed or have never seen a demonstration on how maneu- lying prospectively triggered cardiac sequence. In case of
vers are supposed to be performed. One should carefully this kind of artifact, a question arises, whether these align-
check if the training is successful and the patient adheres to ment issues are due to cardiac motion, which could poten-
the instructions by closely monitoring motion of the chest tially be fixed by alternative reconstructions at different
during the breath-hold or if there is a considerable delay cardiac phase points, or if the misalignment is due to an
between the breath-hold command and the actual breath-­ incomplete breath-hold of the patient.
hold state of the patient. Figure 14.7 shows coronal and sagittal reconstructions of
Almost all CT systems provide digitally recorded voice the lung. The original thin-slice axial reconstructions were
commands, a convenient solution with a consistent level of chosen with an option that limits each individual image
158 T. Allmendinger et al.

Fig. 14.6  Examples of curved planar reformat (CPR) images of the RCA (left), LCA (middle), and CX (right) coronaries. The reconstructions
were derived from a prospectively triggered cardiac sequence and stack misalignment artifacts are clearly visible

Fig. 14.7  Coronal (left) and sagittal (right) MPR images of the characteristic signature of breathing through the diaphragm can be
lung. The misalignment of the different stacks is clearly visible in seen as the misalignment is limited to the lung vessels in combina-
the lung vessels with a considerable gap between the structures tion with the flat liver dome and no misalignment is visible for the
shown in these images and any fast-moving cardiac structures. A thorax surface or ribs

source data to a single cardiac cycle without allowing any slice MIP images with a slice thickness of 40–50 mm illus-
data stack overlap (TrueStack, Siemens Healthineers). The trated in Fig.  14.8. Breathing motion artifacts through the
location of the coronal and sagittal image planes were set at diaphragm can also be easily demonstrated if visualized
a deliberate distance of a few centimeters away from the using thick-slice MIP images. Two additional cases of insuf-
heart minimizing the influence of residual cardiac motion on ficient breath-hold and subsequent breathing artifacts are
these images. The misalignment of the individual stacks is illustrated in Fig. 14.9, both visualized as thick 40-mm MIP
clearly visible in these image views, indicating breathing images in the coronal view making the root cause of the
motion as the root cause of the image artifacts. In addition, as problem easily identifiable.
a characteristic sign of breathing through the diaphragm, one
can see that the misalignment is limited to the lung vessels
including the flat liver dome and no signs of misalignment or Conclusion
steps are present in the ribs or thorax surface of the patient.
Due to the good alignment of the outer thorax and ribs, this In this chapter key workflow elements that are crucial for a suc-
kind of diaphragm breathing-induced artifact is often wrong- cessful cardiac CT examination were described. Ignoring any of
fully attributed to cardiac motion. them can easily result in a suboptimal or non-­diagnostic exam.
A viable alternative for the identification of these kinds of Since the described workflow is quite complex, a dedicated car-
artifacts to the MPR reconstructions of Fig. 14.8 are thick-­ diac CT training of the technologists is highly desirable. Keeping
14  Workflow Optimization 159

Fig. 14.8  Coronal (left) and sagittal (right) 40-mm thick MIP images of the lung. These are from the same case as above illustrating the improved
visibility of the breathing artifacts by switching from a MPR to a MIP display

Fig. 14.9 (a, b) Additional a b

case examples of breathing
artifacts visualized as
thick-slice MIP images in a
coronal image view

in mind that cardiac CT technology is quickly evolving, such 4. Ruzsics B, Gebregziabher M, Lee H, Brothers RL, Allmendinger
T, Vogt S, Costello P, Schoepf UJ. Coronary CT angiography: auto-
training should occur on a regular basis. Clinical practices with matic cardiac-phase selection for image reconstruction. Eur Radiol.
well-trained dedicated cardiac CT technologists perform excep- 2009;19:1906–13.
tionally well. For example, a recent publication from one of 5. Hoffmann MH, Lessick J, Manzke R, Schmid FT, Gershin E, Boll
such practices reports only one non-diagnostic CCTA in 1022 DT, Rispler S, Aschoff AJ, Grass M. Automatic determination of
minimal cardiac motion phases for computed tomography imaging:
consecutive examinations [40]. This example clearly proves that initial experience. Eur Radiol. 2006;16:365–73.
cardiac CT can (and should) be successfully performed when all 6. Lee AM, Engel LC, Shah B, Liew G, Sidhu MS, Kalra M, Abbara
the key workflow instructions are followed. This chapter at hand S, Brady TJ, Hoffmann U, Ghoshhajra BB.  Coronary computed
only focused on two major aspects of the cardiac CT workflow: tomography angiography during arrhythmia: radiation dose reduc-
tion with prospectively ECG-triggered axial and retrospectively
patient preparation and selection of the most appropriate acqui- ECG-gated helical 128-slice dual-source CT. J Cardiovasc Comput
sition mode. Other important aspects such as use of beta-block- Tomogr. 2012;6(3):172–183.e2.
ers, contrast injection, dealing with difficult-to-­image patients 7. Lee AM, Beaudoin J, Engel LC, Sidhu MS, Abbara S, Brady TJ,
Hoffmann U, Ghoshhajra BB.  Assessment of image quality and
(e.g., morbidly obese or with poor renal function), and cardiac
radiation dose of prospectively ECG-triggered adaptive dual-source
CT post-processing were deliberately omitted here because they coronary computed tomography angiography (cCTA) with arrhyth-
are described in other chapters of this book. mia rejection algorithm in systole versus diastole: a retrospective
cohort study. Int J Cardiovasc Imaging. 2013;29(6):1361–70.
8. Siemens data sheet: Flex: http://www.neocoil.com/docs/16CH_
Flex_Siemens_MRI/NC00S007%20Rev%20A%2016ch%20
References Siemens%20Flex%20Data%20Sheet.pdf.
9. Williams MC, Weir NW, Mirsadraee S, Scott AE, Uren NG,
1. Jakobs TF, Becker CR, Ohnesorge B, et al. Multislice helical CT McKillop G, van Beek EJ, Reid JH, Newby DE. Image quality with
of the heart with retrospective ECG gating: reduction of radiation single-heartbeat 320-multidetector computed tomographic coro-
exposure by ECG-controlled tube current modulation. Eur Radiol. nary angiography. J Comput Assist Tomogr. 2014;38(3):444–50.
2002;12(5):1081–6. 10. Latif MA, Sanchez FW, Sayegh K, Veledar E, Aziz M, Malik R,
2. Weustink AC, Mollet NR, Pugliese F, et al. Optimal electrocardio- Haider I, Agatston AS, Batlle JC, Janowitz W, Peña C, Ziffer JA,
graphic pulsing windows and heart rate: effect on image quality Nasir K, Cury RC.  Volumetric single-beat coronary computed
and radiation exposure at dual-source coronary CT angiography. tomography angiography: relationship of image quality, heart rate,
Radiology. 2008;248(3):792–8. and body mass index. Initial patient experience with a new computed
3. McCollough CH, Primak AN, Saba O, et  al. Dose perfor- tomography scanner. J Comput Assist Tomogr. 2016;40(5):763–72.
mance of a 64-channel dual-source CT scanner. Radiology. 11. Meyer M, Haubenreisser H, Schoepf UJ, Vliegenthart R, Leidecker
2007;243(3):775–84. C, Allmendinger T, Lehmann R, Sudarski S, Borggrefe M,
160 T. Allmendinger et al.

Schoenberg SO, Henzler T. Closing in on the K edge: coronary CT 24. Lee JD, See KA.  Trust in automation: designing for appropriate
angiography at 100, 80, and 70 kV-initial comparison of a second-­ reliance. Hum factors. Spring. 2004;46(1):50–80.
versus a third-generation dual-source CT system. Radiology. 2014 25. Parasuraman R, Sheridan TB, Wickens CD. A model for types and
Nov;273(2):373–82. levels of human interaction with automation. IEEE Trans Syst Man
12. Felmly LM, De Cecco CN, Schoepf UJ, Varga-Szemes A, Mangold Cybern Syst Hum. 2000;30(3):286–97.
S, McQuiston AD, Litwin SE, Bayer RR 2nd, Vogl TJ, Wichmann 26. Budoff MJ.  Maximizing dose reductions with cardiac CT.  Int J
JL.  Low contrast medium-volume third-generation dual-source Cardiovasc Imaging. 2009;25(Suppl 2):279–87.
computed tomography angiography for transcatheter aortic valve 27. Budoff MJ.  Ethical issues related to lung nodules on cardiac

replacement planning. Eur Radiol. 2017;27(5):1944–53. CT. AJR Am J Roentgenol. 2009;192(3):W146.
13. Koplay M, Guneyli S, Akbayrak H, Demir K, Sivri M, Avci A, 28. Budoff MJ, Gopal A.  Incidental findings on cardiac computed
Erdogan H, Paksoy Y. Diagnostic accuracy and effective radiation tomography. Should we look? J Cardiovasc Comput Tomogr.
dose of high pitch dual source multidetector computed tomography 2007;1(2):97–105.
in evaluation of coronary artery bypass graft patency. Wien Klin 29. Goehler A, McMahon PM, Lumish HS, Wu CC, Munshi V, Gilmore
Wochenschr. 2016;128(13–14):488–94. M, Chung JH, Ghoshhajra BB, Mark D, Truong QA, Gazelle GS,
14. Budoff MJ, Georgiou D, Brody A, Agatston AS, Kennedy J,
Hoffmann U. Cost-effectiveness of follow-up of pulmonary nodules
Wolfkiel C, Stanford W, Shields P, Lewis RJ, Janowitz WR, Rich incidentally detected on cardiac computed tomographic angiogra-
S, Brundage BH. Ultrafast computed tomography as a diagnostic phy in patients with suspected coronary artery disease. Circulation.
modality in the detection of coronary artery disease: a multicenter 2014;130(8):668–75.
study. Circulation. 1996;93:898–904. 30. Co SJ, Mayo J, Liang T, Krzymyk K, Yousefi M, Nicolaou

15. Hecht HS, Cronin P, Blaha MJ, Budoff MJ, Kazerooni EA, Narula J, S. Iterative reconstructed ultra high pitch CT pulmonary angiogra-
Yankelevitz D, Abbara S. 2016 SCCT/STR guidelines for coronary phy with cardiac bowtie-shaped filter in the acute setting: effect on
artery calcium scoring of noncontrast noncardiac chest CT scans: dose and image quality. Eur J Radiol. 2013;82(9):1571–6.
a report of the Society of Cardiovascular Computed Tomography 31. Franck C, Bacher K. Influence of localizer and scan direction on
and Society of Thoracic Radiology. J Cardiovasc Comput Tomogr. the dose-reducing effect of automatic tube current modulation in
2017;32(5):W54–66. computed tomography. Radiat Prot Dosim. 2016;169(1–4):136–42.
16. Cury RC, Abbara S, Achenbach S, Agatston A, Berman DS, 32. Vadvala H, Kim P, Mayrhofer T, Pianykh O, Kalra M, Hoffmann
Budoff MJ, Dill KE, Jacobs JE, Maroules CD, Rubin GD, Rybicki U, Ghoshhajra B. Coronary CTA using scout-based automated tube
FJ, Schoepf UJ, Shaw LJ, Stillman AE, White CS, Woodard potential and current selection algorithm, with breast displacement
PK, Leipsic JA.  CAD-RADS(TM) Coronary Artery Disease  – results in lower radiation exposure in females compared to males.
Reporting and Data System. An expert consensus document of the Cardiovasc Diagn Ther. 2014 Dec;4(6):470–9.
Society of Cardiovascular Computed Tomography (SCCT), the 33. http://www.mysupplies.philips.com/Documents/lit/453564119681.
American College of Radiology (ACR) and the North American pdf.
Society for Cardiovascular Imaging (NASCI). Endorsed by the 34. http://multimedia.3m.com/mws/media/358372O/proper-skin-prep-
American College of Cardiology. J Cardiovasc Comput Tomogr. ecg-trace-quality-white-paper.pdf.
2016;10(4):269–81. 35. Smith M.  Rx for ECG monitoring artifact. Crit Care Nurse.

17. Halliburton S, Arbab-Zadeh A, De D, Einstein AJ, Gentry R, et al. 1984;4:64–6.
State-of-the-art in CT hardware and scan modes for cardiovascular 36. Odma S, Oberg P.  Movement-induced potentials in surface elec-
CT. J Cardiovasc Comput Tomogr. 2012;6:154–63. trodes. Med Biol Eng Comput. 1982;20:159–66.
18. Celeng C, Vadvala H, Puchner S, Pursnani A, Sharma U, Kovacs 37. Adams-Hamoda MG, Caldwell MA, Stotts NA, Drew BJ. Factors to
A, Maurovich-Horvat P, Hoffmann U, Ghoshhajra B.  Defining consider when analyzing 12-lead electrocardiograms for evidence
the optimal systolic phase targets using absolute delay time for of acute myocardial ischemia. Am J Crit Care. 2003;12:9–18.
reconstructions in dual-source coronary CT angiography. Int J 38. Romans LE. Computed tomography for technologists: exam

Cardiovasc Imaging. 2016;32(1):91–100. review. Lippincott Williams & Wilkins. 2010. https://www.amazon.
19. Srichai MB, Barreto M, Lim RP, Donnino R, Babb JS, Jacobs com/Computed-Tomography-Technologists-Lippincott-Williams/
JE. Prospective-triggered sequential dual-source end-systolic coro- dp/0781777968.
nary CT angiography for patients with atrial fibrillation: a feasibil- 39. Bankier AA, O’Donnell CR, Boiselle PM.  Quality initiatives.

ity study. J Cardiovasc Comput Tomogr. 2013;7(2):102–9. Respiratory instructions for CT examinations of the lungs: a hands-
20. Kondo T, Kumamaru KK, Fujimoto S, Matsutani H, Sano T, Takase ­on guide. Radiographics. 2008;28:919–31.
S, Rybicki FJ. Prospective ECG-gated coronary 320-­MDCT angiog- 40. Ghoshhajra BB, Takx RA, Staziaki PV, Vadvala H, Kim P, Neilan TG,
raphy with absolute acquisition delay strategy for patients with persis- Meyersohn NM, Bittner D, Janjua SA, Mayrhofer T, Greenwald JL,
tent atrial fibrillation. AJR Am J Roentgenol. 2013;201(6):1197–203. Truong QA, Abbara S, Brown DF, Januzzi JL, Francis S, Nagurney JT,
21. Siemens data sheet: FlashCheck. Hoffmann U, MGH Emergency Cardiac CTA Program Contributors.
22. Siemens syngo CT VB10 data sheet. Clinical implementation of an emergency department coronary com-

23. GE lookup tables. http://www3.gehealthcare.com/en/ puted tomographic angiography protocol for triage of patients with
education/product_education_clinical/tip_applications/ suspected acute coronary syndrome. Eur Radiol. 2017;27(7):2784–93.
computed_tomography_hq_class/revolution_ct.
 Defining the Role and Benefits of a 3D
Laboratory for Cardiovascular CT 15
Laura J. Pierce, Daniel T. Boll, and Geoffrey D. Rubin

In addition to understanding methods for successfully and on some of the operational aspects to consider to
acquiring and interpreting computed tomographic angio- assure consistently high-quality output and service.
gram (CTA) images of the heart and cardiovascular struc-
tures, cardiovascular (CV) radiologists and cardiologists
must also be facile with the creation and evaluation of  eveloping a 3D/CV Service to Meet Needs
D
post-processed three- and four-dimensional images and of the Medical Community
analyses of these structures. Particularly for assessing car-
diovascular structure and function across modalities and When performing 3D analyses of CT data, there are many
over time for longitudinal comparisons, CT post-process- priorities that imaging practices and hospitals may have.
ing is essential (Fig. 15.1). The primary interpreter of the These include the creation of consistently high-quality visu-
CTA should be capable of exploring the CT data using alizations and measurements, operations and results that
three- and four-dimensional analysis tools and performing comply with regulations for payment and communication of
a full spectrum of assessments including the creation of the results, efficient operations to off-load tasks from physi-
volume renderings, curved planar reformations (CPR), cians, freeing them to engage in higher value activities,
maximum intensity projections (MIP), and the measure- enhanced and intuitive communication with patients and
ment of the cardiac and vascular dimensions, quantitation referring providers, marketing of clinical programs, facilita-
of coronary calcium, and assessing myocardial function tion of clinical research protocols, and education and train-
through qualitative review of regional wall motion and ing of medical staff (Fig. 15.2).
quantitation of ventricular ejection fraction [1]. This work
is most often accomplished separately from the CT scan-
ner, using independent 3D workstations or client applica- Assessing Community Needs and Priorities
tions to remotely interact with server-based analytical
tools over a network. While physicians bear primary Cardiovascular 3D image post-processing can be very time-­
responsibility for fully interpreting the CT data, there are consuming for physicians, especially when the CTA dataset
compelling reasons for physicians to partner with technol- contains several thousand images from multiple time points
ogists within a formal “3D laboratory” to facilitate the cre- throughout the cardiac cycle. If the local medical community
ation, communication, and curation of post-­ processed has a large population with cardiovascular imaging needs, it
images and measurements and, in collaboration with radi- may be prudent and more efficient for physicians to central-
ologists and referring clinicians, facilitate structured ize the CV image post-processing in a 3D Laboratory, staffed
reporting. This chapter focuses on the rationale for estab- by 3D radiologic technologists who have received advanced
lishing a 3D laboratory for managing CV CTA workflows training in cardiovascular image post-processing and super-
vised by radiologists to grow and evolve offerings according
to the needs of referring clinicians and their patients.
Because the source datasets are very large and present the
L. J. Pierce · G. D. Rubin (*) anatomy from perspectives that may not be intuitive to refer-
Department of Radiology, Duke University School of Medicine,
rers and their patients, cardiologists, cardiovascular sur-
Durham, NC, USA
e-mail: [email protected] geons, and other medical specialists rely on 3D images to
understand a patient’s anatomy and pathologies, to plan
D. T. Boll
Department of Radiology, University of Basel, Basel, Switzerland interventions, and to communicate these details to their

© Humana Press 2019 161

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_15
162 L. J. Pierce et al.

Fig. 15.1  Schematic emphasizing the multidimensionality of imaging (pictogram second from left) or only milliseconds apart (middle picto-
data. High-resolution isometric-voxel acquisitions (left pictogram) are gram) and then compared in an intermodal approach (right pictogram)
acquired either days, weeks, or years apart as follow-up examinations

• Vessel Tacking
• Volumetric Analysis
• (Curved) MPR
• MIP
• Volume Rendering
• Colonography
Thin Client • DECT Post-Processing
• Lung Nodule Assessment
Workstations • Cardiac Analysis
• Organ Segmentation
Workflow
• Multiparametric Analysis PACS Server
• First in / First out Archive • Organ Perfusion Analysis
• Most Recent Exams Quality
• ECG Editing
• Prefetched Exams Assurance

Thin Client • Gigabit Network Standalone

Dicom Server Workstations
Finalized
Image & Image &
Text Result Text Result
Reporting Reporting
• Multi Archive
• Exam Prefetching
• Dicom Conversion
• pdf Conversion

Imaging Laboratory Finalized

pdf Result
Information:
Core Lab Dicom & Database Server Reporting Reporting
Text Reporting DICOM / pdf
Time-resolved Single Energy Dual Energy CT/PET Reporting
MRA Source Data Source Data Source Data
0.6 & 1.2 mm 0.6 & 1.2 mm 0.6 & 1.2 mm

eBrowser
Information:
Prefetch &
DICOM
MRI MDCT DECT PET/CT Transfer Image Reporting
All
Imaging
Data

Fig. 15.2  Flowchart highlighting the complex workflows of a post-­ structured reporting are performed on different types of post-processing
processing laboratory. Following the source data acquisition (lower solutions. Every workflow step is logged and thereby lends itself to
left-hand corner), various forms of post-processing and creation of continuous workflow optimization procedures
15  Defining the Role and Benefits of a 3D Laboratory for Cardiovascular CT 163

patients. When interpreting CTA scans, radiologists and car- specific needs of referring clinicians and their patients. Some
diologists are principally focused upon the interpretation and referring physicians may express an interest in manipulating
its narrative report and less on the creation of images for the and constructing the post-processed images themselves, thus
referrers. Even when considering the needs of referring pro- requiring remote access to the 3D server software in the clin-
viders, physicians interpreting the scans rarely take an orga- ical or surgical environment. Remote access would require
nized approach to image creation, preferring to capture a few additional technical and training support that may be pro-
views analogous to “spot views” captured during other real-­ vided by the 3D laboratory staff. Workflow prioritization,
time imaging explorations such as with gastrointestinal fluo- reporting of quantitative analysis, and coordination of image
roscopy or sonography. While a couple of well-positioned review between cardiothoracic surgery and radiology ser-
and annotated images can be valuable for referring physi- vices are also services that may be facilitated by a 3D
cians and their patients, most prefer a standard set of views laboratory.
similar to those that are routinely acquired by sonographers In a large medical center where many surgical specialists
and radiologic technologists following upper or lower gas- may practice, protocol standardization is imperative for effi-
trointestinal tract contrast studies. The task of creating reli- cient workflows and reliable results. The development of five
ably formatted and labeled standard views is an excellent different 3D coronary CTA protocols for a group of five dif-
task for a 3D technologist. Labeled standard views of previ- ferent cardiothoracic surgeons is a recipe for errors and
ous imaging can be easily incorporated to document longitu- unhappy referrers, particularly when cross-covering one
dinal assessments of cardiovascular atherosclerosis and another. A physician champion in the role of 3D laboratory
cardiac function, for example. director may need to intervene with the group of referrers to
Establishing these standard views is best done in partner- help them organize themselves and create a single consensus-­
ship with the physicians who need to use them to manage driven solution. Regardless of the nuances and ­idiosyncrasies
their patients. Understanding and assessing referring physi- encountered, a fundamental approach to a successful system-
cians’ needs and expectations for CV imaging is an impor- wide 3D lab resource is outreach to referrers and regular
tant first step in the development of a 3D laboratory service. engagement to maximize the value of the lab’s output and
The optimal approach for extracting this information is to enhance dependencies between the referrers and the lab.
identify and interview key referring cardiovascular physi-
cians in the local community. Significant insights may be
extracted from these interviews, which will assist a radiology Quality, Efficiency, and Compliance
group in establishing a service that reflects the voice of its
customers [2] (Fig. 15.3). Some physicians may be princi- According to the American College of Radiology [3]
pally focused on standard image creation, while other may practice parameters for cardiac CT, 3D post-processing
need detailed measurements recorded on standardized forms. should be performed by a physician, a registered radio-
Structured reporting facilitates reports that are targeted to the logic ­technologist, or other experienced personnel who

Fig. 15.3  Example of an affinity chart analysis performed to learn wants and needs from referring clinicians. Which type of post-processing and
reporting is desired as part of a “voice of the customer” service improvement initiative
164 L. J. Pierce et al.

have ­sufficient knowledge of cardiovascular anatomy and Beyond the creation and curation of 3D images and mea-
pathophysiology [3]. Because 3D post-processing for surements, a 3D laboratory can act as a central point of com-
large cardiac CT datasets can be very time-consuming, it munication with the medical center to address any 3D
may be cost-efficient to off-load most of the 3D tasks to a post-processing requests or questions. A consistent point of
trained 3D technologist, rather than an MD. Assigning the contact built into the imaging workflow facilitates follow-
bulk of the creation and quantitation of 3D images to a through and accountability to assure that the required pro-
technologist allows the radiologist or cardiologist to focus cessing is completed and that exams do not slip through the
upon interpretation and reporting of the examination. cracks.
Because 3D technologists devote all their time to post-
processing and quantitative data analysis, they can become
very efficient at these tasks. 3D lab staffing models ideally Compliance
should be flexible and responsive to variations in demand
to assure that turnaround time goals are met. In the event Beyond assuring that images are created consistently, the 3D
of after-hours emergencies, 3D technologists may be lab can assure that necessary requirements are met to comply
given access to the 3D processing server to process urgent with requirements for billing. When scheduling a CTA study
CTA studies remotely. of the heart in the United States, the correct Current
Quantitative analyses of cardiovascular structures and Procedural Terminology (CPT) code must be assigned to
pathology are important measures of disease progress. Key ensure that the correct bill is released to the payer. Radiology
management decisions may be based on subtle changes in CPT codes are maintained by the American Medical
vascular dimensions. However, 3D cardiovascular structure Association (AMA) and uniformly describe diagnostic ser-
measurement results vary with operator experience, intra- vices. CPT codes for cardiac CT range from 75571 to 75574,
and interobserver inconsistencies, CT acquisition, and dif- varying in complexity from a non-contrast CT of the coro-
ferences among 3D software tools and algorithms [4]. nary arteries to a postoperative pre-/post-intravenous con-
3D technologists can provide baseline presurgical and trast CTA of the heart that includes arterial and/or venous
follow-up postsurgical aortic diameters, areas, volumes, structures. Reimbursement is not only dependent on valid
angles, and lengths, as well as coronary artery stenosis and medical indication but also correct documentation of the pro-
cardiac chamber function measurements (Fig. 15.4). A qual- cedure performed. All cardiac CT CPT codes include some
ified 3D Laboratory will provide oversight and established type of 3D post-processing. Because the 3D service is bun-
consistency in cardiac and aortic measurements, perma- dled with the CTA charges, there is not a separate charge for
nently store measurement images on a radiology imaging 3D imaging. To compliantly bill for the cardiac CT examina-
archive, and record quantitative findings into the electronic tion, 3D imaging must be performed, and the 3D results must
CTA examination report for review and signature by the be included in the final report. The American College of
interpreting radiologist or cardiologist. Preoperative mea- Radiology [3] also recommends that the post-processed
surements may be useful for screening for patient candidacy images be labeled with patient and facility identification,
for aortic stent graft or aortic valve replacement, as well as examination date, and anatomic location depicted. The 3D
for stent/valve sizing selection. When follow-up exams images must be saved and archived with the CTA sectional
come through, 3D technologists can recall the prior mea- images [3].
surements and assure that measurements are made in the The 3D lab is excellently positioned to serve as a single
same location and with the same techniques, providing point of contact to confirm billing compliance by reviewing
greater reliability of serial measurements when compared to exam charges, proofreading reports for proper technique ver-
standard clinical reading. Graphical displays of variations in biage, and validating that the 3D images are labeled correctly
key measurements over time can be useful adjuncts to the and stored in the proper CT examination folder for viewing
standard images and measurement tables (Fig.  15.5). To in the enterprise PACS.  The Centers for Medicare and
reduce the potential for transposition errors in transcription, Medicaid Services (CMS) state that if a technologist per-
quantitative data may be exported directly from the 3D forms the CTA 3D examination, the procedure must be per-
interface into the CTA text report or stored in a permanent formed under the direct supervision of a physician (level 2).
database where it may be accessed for comparison with Direct supervision means the physician must be present on
postoperative or follow-up CTA studies. The measurements premises and immediately available to furnish assistance and
may also be programmed to be automatically exported from direction throughout the performance of the procedure. The
the database into the CTA report for confirmation and final physician need not be present in the room when a 3D proce-
signature by the attending radiologist or cardiologist. When dure is performed [5]. For audit purposes, a 3D laboratory
quantitative results are outside normal value, they may be administrator can confirm documentation of physician super-
flagged automatically for review. vision for 3D examinations.
15  Defining the Role and Benefits of a 3D Laboratory for Cardiovascular CT 165

Fig. 15.4  Longitudinal report describing the evolution of a treated aor- row of thumbnail images and corresponding graphical chart) and con-
tic aneurysm over time following endovascular treatment as a function figuration of the implanted endograft over time with a volumetric
of cross-sectional diameter (upper row of thumbnail images and corre- assessment of aneurysm size (lower role of thumbnail images and cor-
sponding graphical chart), aneurysms’ craniocaudal extension (middle responding graphical chart)
166 L. J. Pierce et al.

Category Measurement Name Major Minor

Orthogonal Diameter in Left ventricular outflow tract 28.8 18.2
mm(wall to wall,
Major Aortic Annulus diameter (systole) 27.9
including calcium)
Perpendicular Minor Aortic annulus diameter (systole) 26.0
Aortic annulus in diastole 27.7 25.1
Sinus of Valsalva diameter, right coronary 32.0
Sinus of Valsalva diameter, left coronary 33.7
Sinus of Valsalva diameter, non-coronary 32.3
Sinotubular junction 30.6 60.4
Aorta 4 cm distal to annulus 32.4 30.6
Descending thoracic aorta distal to the LSA 27.6 25.8
Descending thoracic aorta at the level of diaphragm 23.4 22.5
Orthogonal Diameter in Rightsubclavian artery minimum lumen diameter (MLD) 5.2
mm(flowlumen, within Rightsubclavian artery diameter perpendicular to MLD
7.0
calcium)
Left subclavian artery minimum lumen diameter 5.8
Left subclavian artery diameter perpendicular to MLD 5.2
Abdominal aorta proximal to renal arteries 20.9 18.1
Abdominal aorta distal to renal arteries 16.2 14.9
Minimum abdominal aortic diameter 12.5
Perpendicular abdominal aortic diameter at minimum 14.2
Distal abdominal aortic proximal to aortic bifurcation 13.6 12.8
Right common iliac artery proximal 10.0 8.2
Right common iliac artery minimum lumen diameter 7.7
Right common iliac artery perpendicular to MLD 9.0
Right common iliac artery distal 9.0 7.7
Right external iliac artery minimum lumen diameter 6.5
Right external iliac artery perpendicular to MLD 7.0
Right common femoral artery minimum lumen diameter 6.5
Right common femoral artery perpendicular to MLD 7.4
Left common iliac artery proximal 10.3 10.0
Left common iliac artery minimum lumen diameter 8.6
Left common iliac artery perpendicular to MLD 9.6
Left common iliac artery distal 9.5 8.8
Left external iliac artery minimum lumen diameter 7.8
Left external iliac artery perpendicular to MLD 7.8
Left common femoral artery minimum lumen diameter 7.8
Left common femoral artery perpendicular to MLD 7.9
Angle in degrees Aortic Root Angulation (diastole) 41.6
Angle of SOV leaflets, RAO 7.0
Angle of SOV leaflets, CAU 3.0
Length in mm Annulus to RCA origin 14.2
Annulus to Lt main origin 11.8
Sinus of Valsalva height, Right coronary (diastole) 22.0
Sinus of Valsalva height, Left coronary (diastole) 19.7
Sinus of Valsalva height, Non coronary (diastole) 21.9
Average Sinuses of Valsalva height 21.2
Circumference in mm Aortic annulus perimeter systole including mural calcium 76.0
Ejection Fraction in LV Ejection Fraction 58.0
Percentage

Fig. 15.5  Tabulated measurements (a) and corresponding measurement locations (b) as desired by referring clinicians for planning prior to TAVR
procedure
15  Defining the Role and Benefits of a 3D Laboratory for Cardiovascular CT 167

Fig. 15.5 (continued)

Another critical compliance need in the United States Communications in Medicine (DICOM) image header. 3D
centers around confidentiality of medical information and laboratory policies and procedures should contain language
personal records. All staff in the 3D lab must follow protocol about the need to access only needed patient information, not
to maintain compliance with the Health Information sharing passwords, not emailing confidential images or
Portability and Accountability Act (HIPAA). Training for information outside the institution, and not sharing patient
HIPAA compliance is mandatory for all 3D staff. To ensure information with unauthorized personnel. Lab policies must
confidentiality, when 3D images are exported for education also include HIPAA protections for the enterprise-wide 3D
or research purposes, the 3D technologist can anonymize or software that resides on central servers. Rules must limit
de-identify the images electronically, stripping the 18 HIPAA access to the 3D software to currently trained staff, with
Privacy Rule identifiers from the Digital Imaging and strong passwords and time-out limits to assure that 3D
168 L. J. Pierce et al.

c­ omputers are not left unattended when patient accounts are performance, policies, and procedures, a 3D laboratory can
open and active. This becomes even more important as users assure compliance to the Food and Drug Administration
can access the 3D server remotely from off-site locations. (FDA) and research sponsors.
Analogously and particularly for patients who are enrolled in Within a large medical center, the 3D laboratory can act
clinical trials, the ability to de-identify or anonymize com- as a centralized 3D imaging training center for technologists,
plex source data and processed imaging series may represent radiologists, fellows, residents, clinicians, researchers, and
another important type of “post-processing” if imaging data students. The completion of formalized training may be
needs to be sent to an external facility. required prior to qualifying users for secure access to
enterprise-­wide 3D servers. If a remote user has any diffi-
culty with any 3D software, a 3D laboratory technologist
Outreach Facilitated by 3D Imaging may be made available for on-the-spot instruction.
Training protocols may be tailor-made for each group’s
Patient Centricity clinical 3D post-processing needs. This is important for con-
sistency and quality in imaging throughout an institution, as
The formal creation of 3D images provides benefits beyond there is the potential for variation in 3D imaging depending
those derived by referring physicians and interpreting physi- on a user’s abilities and training. After completing and being
cians directly. When surgeons are discussing options for car- signed off for training in the 3D laboratory, cardiovascular or
diothoracic surgery with their patients, 3D images that display radiology residents and fellows may be utilized for any real-­
aberrant anatomy or abnormal conditions may be very helpful time post-processing needs during emergency or clinical
for demonstrating the gravity of a newly diagnosed condition readout of CTA cardiac/thoracic studies. In addition,
or to explain a surgical approach. Patients are better able to researchers from other medical departments who are trained
understand the benefits of a pending intervention after seeing in the 3D laboratory may in turn develop collaborations with
a lifelike 3D model of their anatomy, compared with viewing radiologists and exchange valuable information regarding
the hundreds of stacked axial images that are produced in a the latest advances in surgical technology. This increased
CTA. Engaging patients in their treatment and diagnosis ben- communication between departments has been shown to be
efits the patient and the referring physicians. Many referring advantageous to diminishing organizational silos [2].
physicians relish the opportunity to show beautiful 3D refor-
mations to their patients who provide word-of-mouth endorse-
ments that may enhance future referrals.  perational Considerations for an Effective
O
3D Laboratory

Enhancing Referrals  stablishment of Proper Communication

E
Channels
3D images can also be used by the imaging practice or the
hospital to support a strategy that promotes referrals for spe- One of the most challenging and important functions of a 3D
cialized cardiovascular procedures, such as an atrial fibrilla- laboratory is the establishment of proper communication
tion treatment program or a percutaneous trans-aortic valve channels between all stakeholders in the cardiovascular
replacement for aortic stenosis. 3D images in brochures, imaging service line. A key component of 3D laboratory pro-
websites, and advertisements can highlight the technical cedures is the establishment of two-sided communication
expertise of an institution, which the public may find attrac- channels. Examples include communication of initial imag-
tive when compared with other regional or international ing orders between the radiology information system and the
competitors. A 3D laboratory can prepare and anonymize 3D laboratory, exchange of relevant clinical information
high-resolution 3D static images and 4D temporal videos between the referring MD and the responsible radiologist,
that highlight institution offerings for marketing or educa- assignment of proper scanning and post-processing proto-
tional purposes. cols from the radiologist to the CT and 3D technologists,
requesting retrospective image reconstructions between the
3D laboratory and CT technologists, and communication of
Benefits for Research and Education CT and 3D results to the referring MD.
Because 3D post-processing is often performed on equip-
When an institution performs clinical research, a 3D labora- ment that is located apart from the CT scanners, communica-
tory with trained technologists can ensure that consistency, tion between the scanners and the 3D laboratory is
quality, and integrity are built into the data collection for the particularly critical to the acquisition and processing of clini-
research protocol. With proper documentation of training, cally effective imaging exams. Procedures should be
15  Defining the Role and Benefits of a 3D Laboratory for Cardiovascular CT 169

d­eveloped within the 3D laboratory to collect and then errors in imaging. This is an important quality control func-
review problematic CT acquisitions with the CT technolo- tion that should be built into 3D laboratory workflow.
gists. Mutual respect and regular communication between
these two key groups is critical to assuring effective cardiac
CT performance. Imaging workflow map should be devel-  D Laboratory Technologist Creates
3
oped and monitored starting from the generation of the CT Protocolled Images/Reports
examination request in the clinical arena to the finalized
report in the medical record. This will assist in identifying As discussed previously, input from requesting physicians is
and addressing key points of communication failure which is critical to the creation of clinically valuable post-processing
inherent in a system with several healthcare personnel protocols. 3D laboratory personnel must also work with the
addressing a patient’s clinical data in separate geographical radiologists and cardiologists to formalize and harmonize
locations. standardized 3D imaging protocols and reports in a manner
A solution to assure effective communication may utilize that attempts to accommodate the preference of requesting
customized workflow software that may be integrated with physicians as much as possible. All 3D technologists should
the institution’s electronic medical records. Continuous HL7 be trained to produce the specific 3D images and standard-
data streams containing timestamped updates of clinical ized measurements contained within each protocol in a con-
information may be transmitted from the electronic health sistent manner to reduce output variability. 3D lab report
record to the 3D laboratory workflow software, assisting formatting is another key target for formalization to assure
with workflow prioritizations. The real-time 3D laboratory consistency. One important source of flexibility to be built
worklist may be accessed remotely by authorized 3D tech- into the protocols is an opportunity for 3D technologists to
nologists, radiologists, and referring clinicians. Users may include additional 3D images that demonstrate unexpected
electronically convey any updated quantitative imaging pathologies.
needs or additional clinical information. Radiologists and
clinicians may electronically request priority processing of
urgent CTA examinations and may view the post-processing  egmentation of CT Data for Surgical Planning
S
status of a CTA examination in real time [2]. and Guidance Programs

With their advanced knowledge of anatomy and pathology, 3D

 anaging Source Image Flow Across
M technologists have the capability to segment anatomical struc-
the Enterprise tures from a CTA dataset by using a combination of automatic
and manual segmentation tools (Fig. 15.6). These segmented
Following the imaging workflow map, 3D laboratory person- CT images can be exported as a new series, either in a standard
nel can instruct enterprise information technology services to radiology image format (Digital Imaging and Communications
automatically route all thin-section CTA images from the CT in Medicine (DICOM)) or in a format that is specific to an
scanners to a server serving as a central repository, where the external application that might be used by the referring physi-
images may then be accessed throughout the enterprise. If cians to plan or guide procedures. For example, a segmented
desired, the images may be routed to the 3D server before CT dataset of the aorta may be fused with live fluoroscopic
being routed to PACS to allow preliminary 3D viewing and images in surgery for catheter and device implantation guid-
post-processing to begin immediately. After reviewing the ance (Vessel Navigator, Philips Healthcare, Cleveland, Ohio).
initial imaging, the radiologist or 3D technologist may con- The segmented images may also be exported as stereolithog-
tact the CT technologist directly for any additional image raphy files for rapid prototyping of 3D models or to create
reconstructions to be performed retrospectively. For example, life-sized models of anatomical structures using a 3D printer.
another series may be needed from a cardiac phase with less These models may be used for presurgical planning or cre-
motion to improve the visualization of the coronary arteries ation of tailor-made anatomical implants [6].
or the cardiac valves on the 3D post-processed images.
In addition to 3D laboratory output, 3D images may be
produced throughout the enterprise by physicians and their Software Upgrades
trainees. It is important to monitor which 3D images become
a permanent part of the patient’s medical record in the 3D software companies release frequent patches (bug fixes)
PACS.  It is easy for someone with insufficient training to and updates with new features and functionalities. Although
create a 3D image that falsely represents the patient’s anat- these software upgrades are tested by the manufacturers in
omy. The 3D laboratory can validate all 3D images that their development environment, they cannot duplicate the
arrive on PACS and remove any 3D images that contain typical clinical environment where the software will actually
170 L. J. Pierce et al.

a b

Fig. 15.6  Volume renderings showing (a) the proximal and mid-right cal depiction shown in both views is the result of detailed segmentation
coronary artery and (b) ascending aortic repair with distal anastomotic by 3D laboratory technologists
pseudoaneurysm and residual aortic arch aneurysm. The clear anatomi-

be performing. Therefore, it is advisable to thoroughly assess should include peer review of 3D images and measurements
and validate new releases in a test environment at each clini- before they are submitted for interpretation. Offering a
cal site, using anonymized clinical datasets and running pre- fellowship-­like training program for external technologists
determined 3D imaging and quantitative functions on each may also increase the reputation of the local post-processing
dataset. This will ensure that each 3D software application laboratory and provide a source for external referrals.
accurately, reliably, and consistently fulfills its intended pur-
pose. This time-consuming validation task may be performed
by 3D personnel who will in turn feedback to the manufac- Quality Assurance
turer any performance issues before the software upgrade is
released into the clinical environment. This validation pro- A quality assurance (QA) program is critical for optimizing
cess can help prevent the consequences of software failures reliability and credibility in 3D imaging and measurements.
in the clinical environment, which could potentially have a Key components of a QA program include outlining proce-
negative impact on patient care and clinical workflow [7]. dures for communication, following standardized protocols,
quality control, obtaining and documenting measurements,
reviewing reports, rectifying exam billing or scheduling
Training and Quality Control of Technologists errors, validating 3D software and upgrades, monitoring per-
formance of personnel, and assuring that all 3D operators
The success of a 3D laboratory depends on the selection and receive consistent and updated training and mentoring [5, 8].
training of qualified radiologic technologists. Candidates
should have several years of experience in CT or MRI scan-
ning and have completed a formal in-house training program Documentation and Archive of Results
in a radiology 3D laboratory. A valuable component of train-
ing includes mentorship by a cardiovascular radiologist, Numerical results from 3D measurements should be included
including case review of cardiovascular anatomy and pathol- in the body of the formal CT report, but an institution’s
ogy, recognition of image artifacts, and preferred techniques ­electronic medical records (EMR) system may also have the
for acquiring quantitative analysis. Technologists in training capability to store screen captures of key 3D images and vid-
should participate in a quality control process that is speci- eos associated with the examination. The 3D laboratory can
fied within a 3D laboratory’s quality assurance policy and validate and insert these images along with text into the final
15  Defining the Role and Benefits of a 3D Laboratory for Cardiovascular CT 171

report within the RIS prior to sign-off by the attending radi- • Transarterial aortic valve implantation: preprocedural
ologist or cardiologist. This mixed media presentation can planning for aortic valve replacement. Results are used to
enhance the clinical value of the CT reports. plan delivery route, device type, and device size.
• Thoracic aortic CT angiography: depiction and measure-
ment of aortic aneurysm, traumatic injury, acute aortic
syndrome, vasculitis, or congenital abnormality [9].
Clinical Protocols • Central vein mapping: assessment of stenoses, occlu-
sions, or thrombi within the superior vena cava and its
The creation of specific post-processed images and mea- primary supplying veins.
surements are guided by standardized protocols, which as • Myocardial function: measurement of right or left ven-
discussed previously will be institution specific and tricular ejection fraction. Assessment of reginal wall
informed by the preferences of referring and interpreting motion abnormalities.
physicians. Broad classes of cardiac CT applications that • Pulmonary artery CT angiography: while 3D assessment is
are commonly implemented in 3D laboratories are as not routine in the setting of suspected acute thromboembo-
follows: lism, 3D assessment can be very helpful in mapping con-
genital lesions of the central pulmonary arteries and assessing
• Coronary CT angiography: depiction of native vessel dis- anatomical relationships and graft status in the pre- or post-
ease, bypass grafts, and stents. operative assessment of congenital heart disease.
• Coronary calcium measurement: risk assessment of coro- • Cardiac CT in congenital heart disease: depiction of the
nary heart disease. heart before or after repair of congenital heart disease.
• Pulmonary vein mapping: planning and guiding pulmonary
vein ablation in the management of atrial fibrillation. An example of a pulmonary vein mapping protocol is
illustrated in Table 15.1.

Table 15.1  Sample pulmonary vein mapping protocol

Task Number of images Notes
Curved planar reformations (two planes) Create curved images in two planes through each major coronary
artery. Also, create curved planar reformations through the entire
aorta and any takeoff vessels in the neck and abdomen, if included
in the scan
All coronary arteries (gated or not per LH) 6–8
Aorta and takeoff vessels ~11
Multiplanar reformations
Three transverse images of the left atrium to 3
demonstrate the esophagus
Volume rendering Create a volume rendered image of the left atrium, left atrial
appendage, and several cm of pulmonary vein branches. Save as
horizontal and vertical movies, ~36 images each. Also save
horizontal and vertical movies as avi files, 72 images each, for PDFs
Horizontal and vertical left atrium movie 36 × 2
including the left atrial appendage
Horizontal and vertical left atrium movie 72 × 2
including the left atrial appendage, avi files for pdf
Measurements
Pulmonary mapping measurement protocol Perform measurements per PV measurement template:
 Orthogonal diameters and areas of the origins of each pulmonary
vein. Measure only the contrast-­filled flow lumen. Screen capture
corresponding 2D image in two planes to demonstrate the
location of your measurement. In 3D, screen capture a posterior
VR reference image of the left atrium showing location of
diameter measurements
 Measure the left atrium volume including the left atrial
appendage
 Include Sup/inf and Ant/post diameters of the left atrium,
measured on true sagittal image
 Measure centerline distance between LSPV and RSPV
 If patient has a prior study, look at prior measurements and try to
duplicate the cardiac phase and positioning for comparison
172 L. J. Pierce et al.

Conclusion of Cardiology Foundation/American Heart Association/American

College of Physicians Task Force on Clinical Competence and
Training. J Am Coll Cardiol. 2007;50:1097–114.
Post-processing CT images is mandatory for many cardiac 2. Boll DT. Affinity chart analysis: a method for structured collection,
CT acquisitions. While the interpreting physician should aggregation, and response to customer needs in radiology. Am J
always assess the images using an interface that provides Roentgenol. 2017 Apr;208(4):W134–45.
3. ACR–NASCI–SPR practice parameter for the performance and
ready access to multiplanar reformations, maximum inten- interpretation of cardiac computed tomography (CT). Amended
sity projections, curved planar reformations, volume render- 2016 (Resolution 21). Accessed 21 Nov 2016.
ings, and measurement tools, a 3D laboratory is a key adjunct 4. Cayne NS, Veith FJ, Lipsitz EC, et al. Variability of maximal aor-
for an effective workflow. Staffed by technologists with tic aneurysm diameter measurements on CT scan: significance and
methods to minimize. J Vasc Surg. 2004;39:811–5.
administrative support, 3D laboratories offer consistent, 5. Physician supervision of diagnostic tests. In: cms.gov; 2001.
high-quality image and measurement preparations and Accessed 15 Jul 2016.
archive. They provide a reliable interface with referring phy- 6. Rengier F, Mehndiratta A, Tengg-Kobligk von H, Zechmann CM,
sicians and are better positioned to address special process- Unterhinninghofen R, Kauczor H-U, Giesel FL. 3D printing based
on imaging data: review of medical applications. Int J Comput
ing needs as they might arise. By serving as the primary Assist Radiol Surg. 2010;5:335–41.
resource for creating archival images for cardiac CT exams, 7. Center for Devices, Radiological Health, Center for Biologics
they off-load work from interpreting radiologists and cardi- Evaluation, Research. Guidance documents (medical devices and
ologists, allowing them to shift their focus away from record radiation-emitting products) – general principles of software valida-
tion; Final guidance for industry and FDA staff; 2002.
keeping and increase their focus on the interpretation of 8. Pierce L, Raman K, Rosenberg J, Rubin GD. Quality Improvement
imaging studies. in 3D Imaging. AJR Am J Roentgenol. 2012;198:150–5.
9. Budoff MJ, Cohen MC, Garcia MJ, et  al. ACCF/AHA clinical
competence statement on cardiac imaging with computed tomog-
raphy and magnetic resonance: a report of the American College
References of Cardiology Foundation/American Heart Association/American
College of Physicians Task Force on Clinical Competence and
1. Kramer CM, Budoff MJ, Fayad ZA, et al. ACCF/AHA 2007 clinical Training. J Am Coll Cardiol. 2005;46:383–402.
competence statement on vascular imaging with computed tomog-
raphy and magnetic resonance. A report of the American College
 Structured Reporting for Cardiac CT
16
Anil Attili and Ella A. Kazerooni

The cardiac CT report is an important document that first and reporting ensures the elements required for billing and bill-
foremost communicates the findings and their clinical impli- ing compliance are documented, including adequate infor-
cations to the consumers of the report, including the refer- mation to support clinical necessity, and a sufficient
ring physician, the patient and their families, and the description of exam technique for accurate first time Current
cardiovascular specialists who use the information to make Procedural Terminology (CPT) coding. While the final out-
decisions about procedural intervention appropriateness, eli- put and format do not need to be the same from practice to
gibility, and approach. The report also serves as the docu- practice, inclusion of standard elements, terminology, and
mentation of the imaging episode of care and services definitions are essential to meet the needs of those who con-
rendered, used for medical record documentation, billing, sume the reports. We will share our report formats with you.
and compliance. The structure of a report can also be used to Details of how to obtain, analyze, and interpret examinations
enable quality improvement, outcomes tracking, and are covered in chapters throughout this text.
research. To these ends, a comprehensive, consistent easily
readable cardiac CT report using standardized descriptive
terminology and quantitative metrics where relevant is rec- Components of a Structured Report
ommended [1–4].
Structured reports assure quality and consistency within A structured report begins with an organization of the key
and across practices, providing information in a consistent, components. Each of the following key components will be
predictable manner to those who consume it. Important ele- discussed in detail.
ments are less likely to be omitted using a structured report
template assembled using systematically defined and • Clinical indications
arranged elements. Standardized reports reduce the variabil- • Review of prior cardiac imaging examination images and/
ity of individual communication styles that may result in the or reports
same findings being perceived differently by the consumers. • Study technique
Specifically, the use of structured reporting has been shown • Image quality
to improve clinician’s comprehension of coronary CT angi- • Findings
ography (CCTA) reports [5]. This structure also facilitates • Impression and management recommendations when
research through the use of uniform data fields for data appropriate
extraction and use of standard definitions. Educators will
find structured reports aid trainees in learning the important
elements of study description, clinical elements, definitions, Clinical Indications
and terminology. As cardiac CT has matured and now rec-
ommended in numerous clinical guidelines, ranging from The specific reason(s) for the examination should be
stable and acute chest pain syndromes to structural heart dis- reviewed and documented, including the relevant clinical
ease, reporting templates can be tailored to specific study signs and symptoms, physical examination findings, and
protocols and to disease states [6–11]. Finally, structured medical and surgical history, and should be sufficient for
applicable ICD-10 coding. Cardiovascular risk factors and
A. Attili · E. A. Kazerooni (*) the inclusion of demographic information such as patient
Department of Radiology, Division of Cardiothoracic Radiology, age, gender, body mass index, procedure date, and referring
University of Michigan, Ann Arbor, MI, USA
physician are recommended. Clinical history should include
e-mail: [email protected]

© Humana Press 2019 173

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_16
174 A. Attili and E. A. Kazerooni

coronary risk factors, nature and status of cardiac symptoms, • Radiation exposure (such as dose length product) and use
prior tests and procedures (such as the location and extent of of any dose-reduction strategies
ischemia on stress testing), and the nature of any future • Clinical parameters if relevant, such as heart rate and
planned procedures following the cardiac CT. rhythm
The major indications for cardiac CT include [11]: • Any complications, such as a contrast reaction or
extravasation
1. Coronary artery calcium scoring in asymptomatic patients
with a low to intermediate risk of coronary artery disease Elements of image reconstruction that should be docu-
(CAD) mented include:
2. Evaluation of suspected or known coronary artery

anomalies • Slice thickness
3. Evaluation of low to intermediate risk stable patients with • Slice increment
chest pain or suspected acute coronary syndrome • Phase of cardiac cycle reconstructed
4. Discordant or inconclusive stress tests • Display field of view (DFOV)
5. Evaluation of non-coronary pathology, including the
• Additional elements to consider include:
great vessels, veins, cardiac chambers, myocardium, –– Reconstruction filter
valves, and pericardium –– Reconstructed imaging planes
6. Evaluation of cardiac chamber function –– Post-processing methods used, such as multiplanar
7. Evaluation of cardiac structure and morphology in the reformats (MPR), maximum intensity projections
setting of congenital heart disease (MIP), and 3D volume-rendered reformations

Study Technique Image Quality

The procedure section of the CT report describes both image Overall image quality, contributing factors to reduced image
acquisition and image reconstruction. quality, and the presence of artifacts should be reported.
These variables impact both the reader as they interpret the
Elements of image acquisition that should be documented study and the confidence in their findings and are used by
include: the consumer report in their assessment of their confidence
in the interpretation. Image quality may be degraded by
• Examination type or protocol, such as: technical factors and patient-related factors, and these
–– Coronary calcium score impact the ability to obtain a better-quality examination in
–– Coronary CTA the future. While there are no standard statements to describe
–– Cardiac structural exam, such as for cardiac mass or overall study quality, terms such as excellent, good, or diag-
congenital heart disease nostic when there is a mild reduction in quality or minor
–– Left atrium and pulmonary vein assessment artifacts that do not limit the interpretation, acceptable when
–– Left atrial appendage occluder planning there are further artifacts or reductions in quality that may
–– Transcatheter aortic valve planning (TAVR) or mitral limit a part of the interpretation, and nondiagnostic repre-
valve planning sent a scale of quality. The factors that contribute to any
• Technical acquisition parameters, including collimation, reduction in quality, particularly when it reduces diagnostic
mA and kVp quality or confidence, should be specifically mentioned. For
• Scanner type, such as detector configuration and single or example, when segments of the coronary tree are not inter-
dual source pretable, which segments and the reason(s) why should be
• ECG gating used: prospective and at what phases of the clearly stated.
cardiac cycle or retrospective Patient-related factors to reduced image quality include:
• Anatomic region scanned (e.g., carina to the base of the
heart for routine CCTA) • Large patient size which may reduce quality by an
• Contrast administration: increase in image noise
–– Volume, type, and rate • Metal implants or extensive calcification that may create
–– Timing method—bolus tracking or test bolus beam hardening artifact
–– Uni-, bi-, or triphasic • Motion of the body in general, respiratory motion due to
• Additional medications administered, such as nitroglyc- breathing, and cardiac motion due to a rapid and/or irreg-
erin and heart rate-reducing agents ular heart rate
16  Structured Reporting for Cardiac CT 175

• Poor contrast enhancement, due to poor cardiac output, Table 16.1  Components of a coronary calcium non-contrast CT report
hemodilution with increased circulating blood volume, Section Specific component(s)
intracardiac shunts including opening of a patent foramen Clinical data
ovale, and large patient size General Indication or reason for test, procedure date
Demographics Name, date of birth, sex, referring clinician,
height, weight
Technical factors that may contribute to reduced image
History Symptoms, risk factors, relevant diagnostic
quality include: tests
Procedure data
• Insufficient signal to noise due to low photon number (mA) Equipment Scanner type: Number of detectors, rotation
• Insufficient image contrast due to inappropriate photon time
energy selection (kVp) Acquisition Gating method
• Insufficient contrast volume or rate Tube voltage
Estimated radiation dose
• Inappropriate gating technique
Reconstruction Slice thickness
• Insufficient or lack of heart rate-lowering medications or Slice increment, reconstruction filter, phases
use of nitroglycerin of cardiac cycle
Results
Technical quality Overall quality
Findings Presence and type of artifact and effect on
interpretation
Coronary Agatston score for individual vessels and
The finding section should be tailored to the specific type of
total Agatston score
cardiac examination type; within a practice, consistency Agatston score percentile for age and sex
across templates is recommended where possible. A coro- Volume score and mass score
nary calcium CT report will have a different findings section Non-coronary Presence of calcium in aortic wall,
than a coronary CTA or a TAVR planning examination. cardiac pericardium, myocardium, and valves
Sample reporting templates and reporting checklists are Dilated chambers or total heart enlargement
Presence of pericardial effusion or thickening
provided.
Non-cardiac Abnormalities in the lungs, mediastinum,
esophagus, chest wall, and upper abdomen if
 oronary Artery Calcium Scoring
C present
Table 16.1 summarizes the reporting elements for a coro- Impressions and Total Agatston score and percentile
nary calcium CT.  This may be an independent exam or conclusions Correlation to other or prior cardiac studies
included with as part of a coronary CT in patients who have Documentation of communication to
referring physician for urgent finding(s)
not previously undergone coronary artery stent(s) place-
Clinical recommendations
ment or bypass surgery. The most frequently used measure
for coronary calcium scoring remains the Agatston score
(AJ-130 score) which has been validated with outcome
studies, histological studies, and available nomograms [12, ever if abnormally dilated should be reported. Low attenu-
13]. The AJ-130 score measures the amount of calcium ation of blood due to anemia reduces the density of blood
based on voxels which exceed 130 HU and reach a size in the cardiac chambers relative to myocardium and may
threshold. Software programs measure each calcification serve as internal contrast by which to separate myocar-
and sum the score for each major vessel and for all vessels dium from the chamber cavities. Coronary anomalies may
together, the latter representing the total calcium score. be evident and should be described (see section on CCTA).
Each of these values should be reported, as well as the per- Finally, all non-­cardiac structures in the field of view
centile rank by the patients’ age and gender as established should be reviewed and reported on (see section on non-
on nomograms [14]. The newer scores such as the volume cardiac findings).
score and mass score for each vessel and total score by these A standardized reporting and data classification system
methods may also be reported, as they have the advantage of for coronary artery calcification scoring known as CAC-­
greater reproducibility but have less data in aggregate RADS for the Coronary Artery Calcification Reporting and
behind them [15–18]. Data System was developed by the Society of Cardiovascular
The presence of calcium in other portions of the heart Computed Tomography (SCCT) and the Society of Thoracic
such as the aortic valve, aorta, myocardium, mitral valve, Radiology and is applicable for the reporting of coronary
and pericardium should be noted and reported semiquan- artery calcification on non-contrast, non-gated CT examina-
titatively as mild, moderate, or severe. Evaluation of the tions [19] and may be used when non-gated exams are
cardiac chambers is limited on non-contrast exams, how- obtained as part of a cardiac CT.
176 A. Attili and E. A. Kazerooni

 oronary CT Angiography (CCTA)

C the American Heart Association (AHA) segmentation devel-
Coronary findings: There are several sections that are impor- oped for catheter-based coronary angiography; initially pro-
tant to a structured report with respect to the coronary arter- posed in 1975, this schema has stood the test of time and
ies, including: been used in many long-term outcome studies relating the
location of stenosis to major adverse cardiac events (MACE)
• Coronary artery dominance [20]. An axially based version of this standard model has
• Coronary artery anatomy, specifically been adapted for CCTA by the Society of Cardiac Computed
–– Course: Tomography (SCCT) and more closely emulates CCTA
• Normal views than those obtained during invasive angiography [4],
• Normal variants as illustrated in Fig. 16.1 and described in Table 16.2. Each
• Anomalies, including origin; course; relationship to major epicardial coronary arteries is divided into proximal,
cardiac relevant anatomy including cardiac cham- mid, and distal segments. The proximal right coronary artery
bers, the ascending aorta, pulmonary artery, and (RCA) extends from the ostium to one-half the distance to
ventricular septum; anomalous artery lumen cali- the acute margin of the heart, mid RCA from the end of prox-
ber; shape and evidence of dynamic narrowing or imal RCA to the acute margin of the heart, and distal seg-
occlusion during the cardiac cycle; and any intra- ment from the end of mid RCA to PDA. The proximal left
mural course anterior descending (LAD) artery extends from the end of
–– Myocardial bridging the left main to the origin of the first large diagonal or septal
• Coronary artery atherosclerotic stenosis: by location, perforator, the mid LAD extends from the end of the proxi-
severity, plaque composition, and morphology mal LAD to one-half the distance to the apex, and the distal
• Coronary artery aneurysms LAD extends from the end of the mid LAD to the end of the
• Coronary artery bypass grafts LAD. The proximal left circumflex coronary artery (LCx) is
• Coronary artery stents the segment between the end of the left main and the origin
of the first obtuse marginal branch (OM1), and the mid and
The location of stenosis should be described for each distal LCx are the segments distal to the OM1 to the end of
major epicardial coronary artery using a standardized scheme the vessel or the origin of a left PDA.

Fig. 16.1  SCCT coronary

segmentation diagram. The
abbreviation and description
for each segment can be LAD
found in Table 16.2. (From 8
Leipsic et al. [4], with
permission)
RCA
1
2

7 10
Left main
Diagonal
(D2)
5 6
3
R-PDA 4
9 Diagonal (D1)
11

15 L-PDA 17 Ramus

Circumflex
12 Obtuse
marginal 1
18 L-PLB
(OM1)
13
R-PLB 16
14 Obtuse marginal 2 (OM2)
16  Structured Reporting for Cardiac CT 177

Table 16.2  Axial coronary anatomy legend when there is no evidence of any coronary artery disease (i.e.,
Segment Abbreviation Description normal lumen and no calcified and non-calcified plaque).
1. Proximal right pRCA Ostium of the RCA to one-half Segments containing nonobstructive atherosclerotic plaque
coronary artery the distance to the acute margin should not be described as normal. The following stenosis
(RCA) of the heart
grading is recommended based on a visual semiquantitative
2. Mid RCA mRCA End of proximal RCA to the
acute margin of heart maximal diameter stenosis assessment [4, 21]:
3. Distal RCA dRCA End of mid RCA to origin of
the PDA (posterior descending • Normal: Absence of plaque; no luminal stenosis
artery) • Minimal: Plaque with <25% stenosis
4. PDA-R R-PDA PDA from RCA
• Mild: Plaque with 25–49% stenosis
5. Left main (LM) LM Ostium of LM to bifurcation of
LAD (left anterior descending
• Moderate: Plaque with 50–69% stenosis
artery) and LCx (left • Severe: 7 plaque with 0–99% stenosis
circumflex artery) • Occlusion
6. Proximal LAD pLAD End of LM to the first large
septal or D1 (first diagonal; Plaque composition should be described as calcified when
>1.5 mm in size) whichever is
most proximal the entire plaque appears as calcium by attenuation, non-­
7. Mid LAD mLAD End of proximal LAD to calcified when the entire plaque is devoid of calcium, or par-
one-half the distance to the tially calcified when both elements are present [26]. Very low
apex attenuation within plaque, indicating lipid-rich plaque, should
8. Distal LAD dLAD End of mid LAD to end of also be described. Plaque morphology can be used to identify
LAD
9. D1 D1 First diagonal branch D1
vulnerable plaque, with some features independently associ-
10. D2 D2 Second diagonal branch D2 ated with future acute coronary events [27, 28]. These include
11. Proxiaml LCx pLAD End of LM to the origin of the positive remodeling (the ratio of outer vessel diameter at the
OM1 (first obtuse marginal) site of plaque divided by the average outer diameter of the
12. OM1 OM1 First OM1 traversing the lateral proximal and distal vessel greater than 1.1), low-attenuation
wall of the left ventricle plaque (less than 30 HU), spotty calcium defined as punctate
13. Mid and distal LCx Traveling in the atrioventricular
calcium within a plaque, and the napkin-­ring sign, defined as
LCx groove, distal to the OM1
branch to the end of the vessel or central low attenuation with a peripheral rim of higher CT
origin of the L-PDA (left PDA) attenuation. The presence of any of these high-risk features
14. OM2 OM2 Second marginal OM2 should be evaluated and reported. Other morphological
15. PDA-L L-PDA PDA from LCx descriptors including bifurcation or ostial involvement,
16. PLB-R R-PLB PLB from RCA plaque ulceration, dissection, or fissuring should be reported.
17. Ramus RI Vessel originating from the left
Arterial occlusion should be described by location and length.
intermedius main between the LAD and
LCx in case of a trifurcation Several features indicative of lesions that are less amenable to
18. PLB-L L-PLB PLD from LCx successful percutaneous coronary intervention should be
From Leipsic et al. [4], with permission reported if present, including stenosis length >3 cm, ostial or
Additional nomenclature may be added, for example, D3, R-PDA2, bifurcation location, calcification particularly at the proximal
saphenous vein graft, and mLAD cap, and the degree of collateralization [29–32]. CAD-RADS
PLB posterior-lateral branch
is a structured reporting tool for CCTA that uses stenosis
severity and plaque composition to grade the severity of dis-
Quantification of luminal stenosis, area stenosis, and ease and is discussed in detail later on in this chapter.
plaque content may be done using analysis software; however If bypass grafts are present, the type and number of grafts
current technology has not demonstrated sufficient reproduc- and their origin and insertion should be reported. There is
ibility or accuracy in predicting invasive coronary angiogra- extensive evidence validating the accuracy of CCTA for
phy findings to make such measurements a routine clinical evaluation of graft patency [33]. Patency of the proximal and
practice [21]. Studies have reported that visual CCTA quanti- distal anastomosis of each graft as well as the distal runoff
fication of lesion severity by percent maximal diameter steno- vessels should be documented. Graft stenosis should be reus-
sis has good correlation with quantitative invasive angiography ing the same schema use for native coronary arteries. Review
and intravascular ultrasound, but with a relatively large stan- of the surgical report from bypass surgery is critical to identi-
dard deviation [21–25]. These comparative studies suggest fying grafts. With chronic graft occlusion, grafts become
that at a 95% confidence limit, CCTA currently predicts inva- markedly atretic and may be identified as occlusion by the
sive quantitative angiography to within ±25%. The term “nor- inability to identify them. In the event of planned redo cardiac
mal” in reference to the coronary arteries should be used only surgery, a detailed description of graft location relative to the
178 A. Attili and E. A. Kazerooni

chest wall is important for operative planning; CT has been –– Pulmonary arteries
shown to reduce the rate of reentry injury and improved peri- –– Pulmonary veins
operative outcomes in redo cardiac surgery [34–36]. The • Cardiac chambers:
relationship of the brachiocephalic vessels, aorta, right ven- –– Size and morphology, including aneurysms,
tricle, and pericardium to the midline sternum should be diverticula
assessed in axial or reconstructed sagittal and volume-­ –– Myocardium: hypertrophy, thinning, fatty replace-
rendered CT images. High-risk CT findings include <10 mm ment, lack of enhancement
distance between the right ventricle and aorta to the chest –– Masses
wall or a graft crossing the midline within 10 mm of the ster- –– Congenital anomalies
num in the anteroposterior direction. In addition, extensive –– Wall motion and function if full cardiac cycle data is
atherosclerotic calcification of the ascending aorta, known as obtained
a “porcelain” aorta, is associated with a high risk of plaque • Cardiac valves:
dislodgement during cannulation for cardiopulmonary –– Atrioventricular and semilunar valve thickening, calci-
bypass that may result in cerebrovascular accidents. fication, or structural abnormalities such as a bicuspid
Calcification of the thoracic aorta should be described with aortic valve
respect to location and extent (e.g., mild, moderate, severe, –– Valvular stenosis or regurgitation if full cardiac cycle
or porcelain) and may be demonstrated on maximum inten- data is obtained, including planimetry of the open aor-
sity 3D reconstructions as a calcium map. tic valve at early/mid-systole
Coronary stents should be described with respect to loca- • Pericardium: thickening, calcification, or effusion
tion, quality of interpretability, and the presence of obstruction.
Evaluation of lumen patency inside stents is possible in most Shunting between the atria and ventricles may be visible
cases, though evaluation of in-stent stenosis is dependent on depending on the contrast injection protocol and should be
stent size and composition [37–39]. In general, 64-slice and mentioned. The interatrial and interventricular septum
newer-generation CT scanners exclude in-stent restenosis with should be anatomically evaluated for integrity. Measurement
a high negative predictive value; however, precise quantification and reporting of chamber diameter and wall thickness are
of the in-stent restenosis is not accurate. CT is considered considered optional. Most contemporary scan protocols do
appropriate for risk assessment after revascularization in not include the true end-diastolic phase and mid-diastolic or
asymptomatic patients with a history of left main coronary end-systolic measures of ventricular cavity size, and wall
artery stenting and a stent diameter of 3  mm or larger [11]. thickness may not be representative.
When interpretable, stents should generally be reported as Multiphase reconstruction of these structures may be
patent, with mild in-stent stenosis (<50%), significant in-stent available with retrospective ECG-gated protocols permitting
stenosis (50–99%), or occlusion. dynamic display of segmental wall motion and quantitative
evaluation of size and function. In such cases, left and right
ventricular end-diastolic volumes, end-systolic volumes,
Non-coronary Cardiac Findings stroke volumes, and ejection fractions referenced to body
surface area can be measured and reported providing addi-
All structures within the field of view should be reviewed and tional important information for patient management [40,
reported on. For CCTA, this includes the pericardium, cardiac 41]. Regional wall motion abnormalities should be reported
chambers, atrial and ventricular septum, atrioventricular and using the AHA 17-segment format and assigned to each of
ventriculoarterial valves, pulmonary arteries and veins, sys- the major coronary artery distributions [42]. Wall motion
temic veins, and portions of the thoracic aorta. Depending on may be reported as normal, hypokinetic, akinetic or dyski-
the contrast infusion protocol, typically, the left-sided cham- netic, and aneurysmal.
bers and sometimes the right-sided are suitable for interpreta- Myocardial CT enhancement patterns should be assessed.
tion. Reconstruction of the axial source data in the standard During the arterial, hypodense areas of myocardium may rep-
cardiac planes (short axis, four-chamber, two-chamber, and resent decreased myocardial perfusion as a result of prior
three-chamber/LVOT view) should be performed. A descrip- myocardial infarction or severe obstructive CAD. Old myo-
tion of the non-coronary cardiac structures should be pro- cardial infarcts are characterized by hypoenhancement with
vided in the report to include the following: wall thinning and/or fat replacement, ventricular remodeling,
and/or the presence of calcifications and thrombi. 5–8-mm-
• Thoracic vasculature: thick averaged MPR reconstructions, a narrow width and
–– Aorta include diameters of the aortic root, ascending level, and the application of minimum intensity projection
and descending aorta techniques improve detection of low-attenuation myocardial
–– Superior and inferior vena cava areas due to hypoperfusion and infarction. Perfusion defects
16  Structured Reporting for Cardiac CT 179

and myocardial abnormalities should be described using the Cardiovascular Computed Tomography (SCCT), the
AHA 17-segment model used for wall motion abnormalities. American College of Radiology (ACR), and the North
Aortic valve planimetry measurement has been shown to cor- American Society for Cardiovascular Imaging (NASCI) [45].
relate with the severity of aortic stenosis on transesophageal The intent of CAD-RADS is to create a standardized method
and transthoracic echocardiography [43, 44]. to communicate CCTA findings in order to facilitate decision-
making regarding patient management. The CAD-­ RADS
classification is applied at the patient level, based on the high-
 AD-RADS: The Coronary Artery Disease
C est-grade coronary lesion. CAD-RADS codes range from 0 in
Reporting and Data System the absence of stenosis and plaque to CAD-RADS 5 in the
presence of at least one totally occluded coronary artery. The
A standardized reporting and data classification system for CAD-RADS reporting and data system for patients with sta-
CCTA, called CAD-RADS, has been recently developed and ble and acute chest pain is shown in Tables 16.3 and 16.4,
reported as an expert consensus document of the Society of respectively. The grading scale for stenosis severity suggested

Table 16.3  CAD-RADS reporting and data system for patients presenting with stable chest pain
Degree of maximal Further cardiac
coronary stenosis Interpretation investigation Management
CAD-­RADS 0 0% (no plaque or Documented None Reassurance. Consider non- atherosclerotic
stenosis) absence of CAD causes of chest pain
CAD-­RADS 1 1–24%—minimal Minimal None Consider non-atherosclerotic causes of chest
stenosis or plaque with nonobstructive pain
no stenosisa CAD Consider preventive therapy and risk factor
modification
CAD-­RADS 2 25–49% Mild non-­ None Consider non-atherosclerotic causes of chest
mild stenosis obstructive CAD pain
Consider preventive therapy and risk factor
modification, particularly for patients with
non-obstructive plaque in multiple segments
CAD-­RADS 3 50–69% stenosis Moderate stenosis Consider functional Consider symptom-guided anti-ischemic and
assessment preventative pharmacotherapy as well as risk
factor modification per guideline-directed careb
Other treatments should be considered per
guideline-directed careb
CAD-­RADS 4 70–99% stenosis Severe stenosis A: Consider ICA or Consider symptom-guided anti-ischemic and
or functional assessment preventative pharmacotherapy as well as risk
B—left main >50% or B: ICA is factor modification per guideline-directed careb
three-vessel obstructive recommended Other treatments (including options of
(≥70%) disease revascularization) should be considered per
guideline-directed careb
CAD-­RADS 5 100% (total occlusion) Total coronary Consider ICA and/or Consider symptom-guided anti-ischemic and
occlusion viability assessment preventative pharmacotherapy as well as risk
factors modification per guideline-directed careb
Other treatments (including options of
revascularization) should be considered per
guideline-directed careb
CAD-­RADS N Nondiagnostic study Obstructive CAD Additional or
cannot be excluded alternative evaluation
may be needed
From Cury et al. [45], with permission
Note: The CAD-RADS classification should be applied on a per-patient basis for the clinically most relevant (usually highest-grade) stenosis. All
vessels greater than 1.5  mm in diameter should be graded for stenosis severity. CAD-RADS will not apply for smaller vessels (<1.5  mm in
diameter)
Modifiers: If more than one modifier is present, the symbol “/” (slash) should follow each modifier in the following order:
  First: modifier N (non-diagnostic)
  Second: modifier S (stent)
  Third: modifier G (graft)
  Fourth: modifier V (vulnerability)
CAD coronary artery disease, ICA invasive coronary angiography
a
CAD-RADS 1—This category should also include the presence of plaque with positive remodeling and no evidence of stenosis
b
Guideline-directed care per ACC Stable Ischemic Heart Disease Guidelines [46]
180 A. Attili and E. A. Kazerooni

Table 16.4  CAD-RADS reporting and data system for patients presenting with acute chest pain, negative first troponin, negative or nondiagnostic
electrocardiogram, and low to intermediate risk (TIMI risk score < 4) (emergency department or hospital setting)
Degree of maximal
coronary stenosis Interpretation Management
CAD-­RADS 0 0% ACS highly No further evaluation of ACS is required
unlikely Consider other etiologies
CAD-­RADS 1 1–24%a ACS highly Consider evaluation of non-ACS etiology, if normal troponin and no
unlikely ECG changes
Consider referral for outpatient follow-up for preventive therapy and risk
factor modification
CAD-­RADS 2 25–49%b ACS unlikely Consider evaluation of non-ACS etiology, if normal troponin and no
ECG changes
Consider referral for outpatient follow-up for preventive therapy and risk
factor modification
If clinical suspicion of ACS is high or if high-risk plague features are
noted, consider hospital admission with cardiology consultation
CAD-­RADS 3 50–69% ACS possible Consider hospital admission with cardiology consultation, functional
testing, and/or ICA for evaluation and management
Recommendation for anti-ischemic and preventive management should
be considered as well as risk factor modification. Other treatments should
be considered if there is presence of a hemodynamically significant
lesion
CAD-­RADS 4 A: 70–99% ACS likely Consider hospital admission with cardiology consultation. Further
B: Left main >50% or evaluation with ICA and revascularization as appropriate
three-vessel obstructive Recommendation for anti-ischemic and preventive management should
disease be considered as well as risk factor modification
CAD-­RADS 5 100% (total occlusion) ACS very likely Consider expedited ICA on a timely basis and revascularization if
appropriate if acute occlusion.c
Recommendation for anti-ischemic and preventive management should
be considered as well as risk factor modifications
CAD-­RADS N Nondiagnostic study ACS cannot be Additional or alternative evaluation for ACS is needed
excluded
From Cury et al. [45], with permission
Note: The CAD-RADS classification should be applied on a per-patient basis for the clinically most relevant (usually highest-grade) stenosis. All
vessels greater than 1.5  mm in diameter should be graded for stenosis severity. CAD-RADS will not apply for smaller vessels (<1.5  mm in
diameter)
Modifiers: If more than one modifier is present, the symbol “/” (slash) should follow each modifier in the following order:
  First: modifier N (non-diagnostic)
  Second: modifier S (stent)
  Third: modifier G (graft)
  Fourth: modifier V (vulnerability)
ACS acute coronary syndrome, ICA invasive coronary angiography
a
CAD-RADS 1—This category should also include the presence of plaque with positive remodeling and no evidence of stenosis
b
CAD-RADS 2—Modifier 2/V can be used to indicate vulnerable/high-­risk plaque
c
Unless the total coronary occlusion can be identified as chronic (through CT and clinical characteristics or patient history)

by the SCCT is used. All vessels greater than 1.5 mm in diam- incorporating CAD-RADS is as described in this chapter and
eter should be graded for stenosis severity; CAD-RADS does in accordance with professional society guidelines. A CAD-
not apply for vessels less than 1.5  mm in diameter. CAD- RADS code ranging from 0 to 5 is assigned in the impression
RADS categories can be complimented by modifiers to indi- section based on the highest-grade coronary stenosis along
cate that a study is not fully evaluable or nondiagnostic (N) or with corresponding management recommendations. CAD-
to indicate the presence of stents (S), grafts (G), and vulner- RADS reinforces the objective of structured reporting in
able plaques (V). Specific recommendations are provided for communicating findings in a clear consistent manner, facili-
further management of patients with stable or acute chest tating education, research, peer review, and quality
pain based on the CAD-RADS classification and code. These assurance.
suggestions for further patient management should always be Examples of CAD lesions categorized using the CAD-­
considered in light of the full clinical information available to RADS system are shown in Figs.  16.2, 16.3, 16.4, 16.5,
the treating physician. The basic structure of a CCTA report and 16.6.
16  Structured Reporting for Cardiac CT 181

a b

Fig. 16.2 (a–c) CAD-RADS 0. Normal left main and LAD, LCX, and RCA without plaque or stenosis

Cardiac CT for Electrophysiology Applications  re-procedural CT Imaging for Left Atrial

P
Ablation
Pre-procedural cardiac CT imaging provides information Pre-procedural left atrial (LA) imaging by CT focuses on
about cardiac structure and anatomy that may be crucial for determining atrial and pulmonary venous anatomy, ruling
successful atrial and ventricular ablation for arrhythmias and out atrial thrombi, and identifying adjacent structures.
structural interventions in electrophysiology. Pre-procedural Knowledge of patient-specific anatomy of the pulmonary
CT combined with imaging obtained during EP interven- veins facilitates guiding catheter manipulation and delivery
tions enhances procedural safety, reduces exposure to ioniz- of radiofrequency energy at antral or ostial locations, thereby
ing radiation from fluoroscopy, reduces exposure time, and reducing the risk of pulmonary vein stenosis [47, 48]. In
may improve outcomes. patients with paroxysmal AF, circumferential PV isolation
182 A. Attili and E. A. Kazerooni

a b

Fig. 16.3 (a, b) CAD-RADS 1. Calcified plaque in the mid LAD with minimal luminal narrowing (less than 25% diameter stenosis)

a b

Fig. 16.4 (a, b) CAD-RADS 3. Predominantly non-calcified plaque in the proximal LAD with 50–69% diameter stenosis

guided by image integration significantly improves clinical • Left atrial dimensions and size
outcome in comparison with both circumferential PV • Presence of absence of left atrial appendage thrombus
­isolation guided by 3D mapping alone and with segmental • Adjacent structures: esophagus and phrenic nerve
electrophysiologically guided PV isolation [49]. The pre-
procedural CT report should include: Conventional pulmonary venous anatomy is defined as
the presence of single right and left superior and inferior pul-
• Pulmonary veins: monary veins that drain into the LA without any accessory
–– Anatomy: number of veins and any anatomic variants veins. The most common variants are a common or con-
and accessory or anomalous veins joined vein when the superior and inferior veins combine
–– Ostial diameter of each vein in long and short axis proximal to the LA, resulting in only one atriopulmonary
–– Distance to first-order branch of each vein venous junction on the involved side, as most commonly
16  Structured Reporting for Cardiac CT 183

a b

Fig. 16.5 (a–c) CAD-RADS 4A. Non-calcified and calcified plaque in the mid LAD causing 70–99% diameter stenosis

a b

Fig. 16.6 (a, b) CAD-RADS 5. Complete occlusion of the LAD

184 A. Attili and E. A. Kazerooni

seen on the left. Supernumerary or accessory pulmonary pulmonary veins can be delineated on CT [56]; however it
veins are extra veins with independent atriopulmonary must be noted that the esophagus is a mobile structure and
venous junctions separate from the superior and inferior pul- can shift by ≥2 cm in patients undergoing AF ablation under
monary veins and are named for the pulmonary lobe or seg- conscious sedation. Therefore real-time fluoroscopic imag-
ment that they drain, sometimes crossing pulmonary lobar ing during the procedure is preferable [57].
fissures before emptying into the left atrium. Common
examples include an additional right middle lobe vein.  re-procedural CT for Ventricular Ablation
P
Anomalous pulmonary venous anatomy occurs when pulmo- Late gadolinium enhancement cardiac MR (CMR) is the ref-
nary veins drain into a structure other than the left atrium. erence standard for detection of myocardial scar and is used
A variety of post-processing methods are used to display to guide and target arrhythmogenic tissue with catheter abla-
left atrial and pulmonary venous anatomy including 3D vol- tion in ischemic and nonischemic cardiomyopathies.
ume rendering (extra-atrial surface views and intra-atrial However, many patients with ventricular tachycardia (VT)
endoscopic views) MPR, CPR, and MIP views. To obtain the have a cardiac implanted electronic device (CIED), with car-
correct measurement of ostial diameter of each PV, both diac MR services not widely available to these patients, or
MPR and CPR images should be selected to show the en face artifacts from the CIED may preclude assessment of scar tis-
ostial image. Then, ostial diameter should be measured as sue on the CMR exam. Regional wall thinning and hypoper-
minor and major axis diameter by orthogonal angles and fusion on cardiac CT have been found to correlate with the
reported [23, 50]. arrhythmogenic substrate in postinfarction VT [58]. Both epi-
LA size is related to the chronicity of AF and the likeli- cardial fat and scar tissue result in low epicardial voltage on
hood of success for achieving sinus rhythm following abla- EP mapping procedures. Cardiac CT can differentiate epicar-
tion procedures. Reference normal and indexed values for dial fat from scar or muscle on the basis of their distinct atten-
LA size, volumes, and function have been published for uation, with epicardial ablations at sites with >10 mm of fat
ECG-gated MDCT [51]. The LA diameter can be measured found to be ineffective [59]. The thickness of the epicardial
and reported as a single anterior-posterior measurement on a fat over areas of scar and low voltage on electroanatomic
three-chamber reconstruction in end-systole. However due mapping is therefore relevant to patient selection and the suc-
to the complex shape of the LA and asymmetrical remodel- cess of epicardial ablation procedures. Multiplanar and 3D
ing in AF, 2D measurements may not reflect the true size of images of epicardial fat generated from the pre-procedural
the structure. LA function and volume can be calculated and CT examination are co-registered with the electroanatomic
reported using a modified method of Simpson with manual voltage map. Other relevant imaging features on pre-proce-
tracing of the endocardial borders on successive slices at dural CT include the relationship and proximity of the site of
both end-systole and end-diastole [52, 53]. An incremental the arrhythmic focus to the coronary venous system, coronary
predictive value has been shown for LA volumes by CT over arteries, and phrenic nerve [60]. Evaluation and reporting of
echocardiographic dimensions in predicting the outcome of these measurements require close collaboration with refer-
AF ablations. ring electrophysiologists and knowledge of the abnormalities
The presence of LAA thrombus should be identified and on electroanatomic voltage maps.
reported. Poor contrast opacification of the LAA is often
caused by slow flow and a delayed acquisition 1–2 min after
the initial scan increases the accuracy for thrombus detection Cardiac CT for Structural Heart Disease
[54]. Phrenic nerve (PN) injury is a known complication of
RF catheter ablation of AF.  The right cardiophrenic artery CT offers high spatial resolution 3D anatomic assessment,
can be visualized by CT and serves as a landmark to identify with the temporal resolution enabled by gating permitting
the adjacent phrenic nerve [55]. The anatomic path of the focused assessment at specific phases of the cardiac cycle, all
right cardiophrenic artery consistently runs anterior to the important for patient selection, device eligibility, and pre-­
right superior pulmonary vein within the pericardial tissue procedural planning for interventions in structural heart dis-
anterior and adjacent to the right superior pulmonary vein ease. This includes planning for transaortic valve replacement
and the SVC.  Although the CT tissue density of the right (TAVR), trans-mitral valve implantation, and left atrial
phrenic nerve remains the same as tissue densities of the sur- appendage occlusion.
rounding structures, the right cardiophrenic artery enhanced
with contrast can be identified on the pre-procedural  ranscatheter Aortic Valve Replacement (TAVR)
T
CT. During LA catheter ablation, an atrioesophageal fistula TAVR has become the treatment of choice for symptomatic
can develop as a result of thermal injury of the esophagus severe aortic stenosis in both high-risk and intermediate-risk
during ablation along the posterior LA. The esophagus and patients. CT plays an important role in the workup of patients
its relationship to the left atrium, pulmonary vein ostia, and who are candidates for implantation of a transcatheter aortic
16  Structured Reporting for Cardiac CT 185

valve [61]. Essential elements of a TAVR CT report include anatomy, and papillary muscle structure. The geometry of
detailed information about the aortic outflow, annulus, root, the mitral valve and supporting structures is more complex
and proximal ascending aorta for valve sizing and assess- and variable than the aortic valve and annulus, making real-­
ment of common access vessels for appropriateness, the lat- time assessment with echocardiography more critical in
ter usually being the iliofemoral arteries but may also include decision-making. CT plays a complementary role to 3D
the upper extremity arteries as well. Details of each structure echocardiography in determining patient suitability by 3D
are provided below. For a detailed description of the use of anatomical quantification and landing zone characterization
CT in pre-procedural planning of a TAVR procedure, please [62]. Specific items for a structured report should include the
refer to Chap. 40. details listed below. Currently, TMVI devices in use and
Elements of a structured report for TAVR planning development are typically placed by a transapical route, less
include: often transeptal or transatrial route; therefore aortic and arte-
rial measurements are not required [63]. Leaflet length is
• Aortic valve: generally measured using echocardiography and not CT.
–– Number of leaflets
–– Presence and severity of aortic valve leaflet calcifica- • Mitral annulus calcification (none, mild, moderate, or
tion (none, mild, moderate, or severe severe)
• Aortic root: • Mitral annulus diameter across each commissure
–– Aortic annulus diameter (maximum and minimum) • Mitral annulus circumference
–– Aortic annulus perimeter • Mitral annulus area
–– Aortic annulus area • Angle between coronary sinus and the mitral annulus
–– Sinus of Valsalva height for the right, left, and non-­ • Presence of leaflet prolapse if cine imaging obtained
coronary sinuses • Subvalvular apparatus evaluation for unfavorable anat-
–– Sinus of Valsalva diameter for the right, left, and non-­ omy, including:
coronary sinuses –– Irregular papillary muscle head branching or fused
–– Height from annulus to left main and right coronary heads
artery origins –– Chordae tendinae extending to the subannular groove
–– Ascending aorta maximum diameter, including 4  cm of the mitral valve
above the aortic annulus –– Large strut chordae
–– Aortic root angulation in degrees • Left ventricle:
• Access vessel assessment with minimum diameter (maxi- –– End-diastolic diameter
mum and minimum diameter at this location), tortuosity –– Presence of basal septal bulging, hypertrophy, or
and calcification assessment of each segment, and if pres- thinning
ent any aneurysm (including maximum diameter and
location), stent, endograft or anatomic variant:  eft Atrial Appendage Occlusion
L
–– Thoracic aorta Percutaneous closure of the left atrial appendage (LAA) is an
–– Abdominal aorta effective and safe alternative for treating patients with atrial
–– Left common iliac artery, including high-riding fibrillation and contraindication for oral anticoagulants [64].
bifurcation The Watchman LAA occlusion device is FDA approved in
–– Left external iliac artery the United States for patients with nonvalvular atrial fibrilla-
–– Left femoral artery tion with acceptable anatomy who are at increased risk for
–– Right common artery, including high-riding stroke and would be candidates for anticoagulation in whom
bifurcation there is concern about the risk/benefit ratio of chronic antico-
–– Right external artery agulation [65]. Preimplantation cardiac CT can provide cru-
–– Right femoral artery cial information for safe device planning and selection and
–– Left subclavian artery has been shown to be more accurate than echocardiography
–– Right subclavian artery for LAA measurements and overall geometry of this complex
–– Right innominate artery eccentric structure [66]. Important preimplantation consider-
ations to be mentioned in the structured report include the
Mitral Valve Interventions following:
Transmitral valve implantation (TMVI) is an evolving treat-
ment strategy for patients with mitral regurgitation, requiring • Left atrial appendage:
detailed pre-procedural imaging for evaluation of annulus –– Morphology and shape: windsock, cactus, chicken
shape and dimensions, the valve leaflets, coronary sinus wing, or broccoli [66, 67]
186 A. Attili and E. A. Kazerooni

–– Length and diameter stress perfusion CT imaging is performed during the admin-
–– Volume istration of a pharmacological vasodilator such as adenosine,
–– Presence or absence of thrombus regadenoson, or dipyridamole. Several single-center and
–– Relationship to the left superior pulmonary vein multicenter trials indicate that stress CT perfusion is nonin-
–– Relationship to the adjacent pulmonary artery ferior to SPECT and FFR to detect myocardial ischemia and
• Atrial appendage ostium hemodynamically relevant stenosis [69–71]. CT assessment
–– Ostium shape: oval, foot-like, triangular, water drop-­ of myocardial perfusion is based on the distribution of iodin-
like, or round ated contrast material during its first pass through the myo-
–– Maximum and minimum diameter cardium. Contrast material distribution is dependent on the
–– Perimeter arterial supply; hence myocardial perfusion deficits are iden-
• Relationship of LAA to adjacent left upper lobe pulmo- tified as hypoattenuating areas. Details of exam acquisition
nary vein and pulmonary artery technique are covered in Chap. 62.
A stepwise approach to combined visual interpretation and
There are four types of LAA shape. The windsock type reporting of CCTA and CT perfusion imaging is recommended
has a dominant central lobe allowing straightforward inser- [72] and should include (1) coronary CTA interpretation as
tion of the occlusion device. The cactus type has a dominant covered earlier in this chapter, (2) reconstruction and process-
central lobe, with secondary lobes arising from the central ing of myocardial CT perfusion images, (3) image quality
lobe, however does not contraindicate insertion. The chicken assessment and the identification of potentially confounding
wing type has a dominant central lobe with a sharp kink. A artifacts, (4) rest and stress image interpretation for enhance-
kink very proximal to the ostium creates a small usable ment patterns and areas of hypoattenuation, and (5) correla-
length that may complicate insertion. The broccoli or cauli- tion of coronary anatomy and myocardial perfusion deficits.
flower type is the least common with multiple lobes and a CT perfusion images should be reconstructed in the three
short length, generally considered unsuitable for occlusion. traditional cardiac views: short axis, horizontal long axis,
The LAA ostium has a variable shape and may be classi- and vertical long axis that allow for evaluation of each myo-
fied into five types including oval, foot-like, triangular, water cardial perfusion territory in more than two planes and
drop-like, and round. The shape has implications for device viewed at a narrow window width and level (300/150 or
placement as the ostium has to adapt to the shape of the round 200/100). The data set should be examined for image qual-
device. The measured maximal diameter of a very flat ostium ity and artifacts such as poor signal to noise, misalignment,
will not be reflective of the diameter at device insertion. The and beam hardening. Typical locations for artifacts due to
perimeter and a perimeter-based ostial diameter have been beam hardening are the basal inferior lateral wall and the
reported to be the best parameters for sizing the LAA occlu- left ventricular apex. Multiphase data sets should be used if
sion device due to the variable shape and morphology of the available to differentiate true perfusion defects from arti-
ostium [68]. The approximate LAA length is measured as the facts, the former persisting over several cycles and are in the
distance from the LAA ostial plane to the LAA apex in the distribution of a coronary artery. Rest and stress perfusion
primary lobe. LAA thrombus is a contraindication to implan- images are read serially and compared, to assess for isch-
tation, and its presence or absence should be reported. The emia and infarction. The content of a structured report
LAA ridge is a barrier between the LAA orifice and the adja- should include:
cent left upper lobe pulmonary vein (LUPV). The width of
the LAA ridge and proximity of the LAA to the LUPV could • Quality of myocardial perfusion images and presence of
also affect a safe occlusion. For example, the expansion of an any artifacts
occluder may markedly compress the pulmonary vein when a • Visual assessment of the perfusion to each myocardial
LAA orifice is extremely flat and in close proximity to the segment at rest and stress:
LUPV. Perforation of the pulmonary artery and consequent –– 0 = normal myocardial perfusion
hemopericardium is a recognized complication of LAA –– 1  =  mild, perfusion deficit is less than one-third
occluder insertion. If the ostium is very close to the main pul- transmurality
monary artery, this should be described. –– 2 = moderate, ≤50% transmurality
–– 3 = severe, >50% transmurality
–– 4 = thinned chronic infarct
 yocardial Perfusion Imaging and Delayed
M • Reversibility of each defect
Enhancement CT • Transmyocardial perfusion ratio when relevant
• Whether each defect corresponds to an anatomic lesion
Technologic improvements allowing for faster lower radia- on CCTA
tion exposure imaging facilitate the ability to perform CT • Absolute myocardial blood flow if dynamic CT perfusion
myocardial perfusion imaging at rest and stress. Myocardial technique is used
16  Structured Reporting for Cardiac CT 187

A reversible perfusion deficit consistent with inducible segmental approach with a systematic description as below
myocardial ischemia will show normal resting perfusion and and include a detailed description of lung parenchyma, air-
abnormal stress perfusion. A predominately fixed perfusion ways, and chest wall given the association of other congeni-
deficit may have two different appearances: (1) rest perfu- tal abnormalities.
sion deficit may persist on stress imaging and remain the
same size, or (2) a rest perfusion deficit may partially hyper- • Visceral and atrial situs
enhance to differing degrees due to relative hyperenhance- • Position and location of the apex
ment of an infarct. The transmyocardial perfusion ratio • Systemic and pulmonary venous connections
(TPR) is a semiquantitative perfusion metric representing the • Atrioventricular connections
ratio of subendocardial attenuation and the entire subepicar- • Ventriculoarterial connections
dial attenuation within a myocardial segment. TPR reflects • Atrial size and interatrial septum
the normal relation between subendocardial and subepicar- • Ventricular size, function, and interventricular septum
dial blood flow, which is normally higher in the subendocar- • Relationship of the great vessels
dium and is also the area compromised first during coronary • Outflow tracts and semilunar valves
stenosis [72, 73]. In subjects with no coronary disease, the • Patency, size, and branching of the aorta and pulmonary
mean TPR is 1.12  ±  0.13, and it is considered potentially arteries
abnormal when it is <0.99 or more than 1 SD below the mean • Origins and course of the coronary arteries
normal TPR. In case of dynamic CT perfusion, a calculation • Functional and volumetric data relative to body surface
of absolute blood flow in the myocardium can be performed area and available normal values
and has been shown to have a good correlation to invasive
fractional flow reserve [69, 74].
Delayed enhancement CT allows accurate visualization Extracardiac Structures
of myocardial infarction and viability and can be performed
with a second CT scan performed 5–10 min after coronary Dedicated cardiac CT data sets include portions of noncar-
CTA to demonstrate hyperenhancement [75–79]. The diovascular thoracic and upper abdominal anatomy, which
regional extent of delayed enhancement is assessed using a should be reviewed and reported for two important reasons:
17-segment model. MinIP and thick MPR detect infarcted recognizing associated primary and secondary comorbid
myocardium with greater visibility and definition than MIP pathology and identification of abnormalities that may lead
and thin MPR [80]. Each segment is described and reported to alternative cause for symptoms (e.g., pulmonary embo-
as involved or healthy and the percentage of transmural lism or pneumonia in the acute setting). A full field of view
extent of enhancement graded using a 4-point scale (1–4): reconstruction should be performed, and all the extracardiac
0–25%, 26–50%, 51–75%, or 76–100%. structures in the field of view should be examined and abnor-
malities reported [4].

Congenital Heart Disease

Impressions and Management
Improved CT scanner technology and reconstruction meth- Recommendations
ods allowing rapid, high spatial and temporal resolution
imaging of the heart with low radiation doses make CT an The concluding section should summarize the major findings
attractive and increasingly used imaging modality for con- in a clear concise manner, answer the specific clinical ques-
genital heart disease (CHD) in both pediatric and adult tions posed, and guide further clinical management. In the
patients. The use of cardiac CT for CHD is a guideline rec- context of CCTA for ischemic symptoms, the major findings
ommended in specific circumstances, such as inability to regarding coronary artery and cardiac structure and function
undergo MRI due to claustrophobia, nondiagnostic MRI due should be concisely stated. Additional tests may be recom-
to metallic artifacts, and when higher spatial resolution ana- mended if the significance of scan findings is uncertain, e.g.,
tomical imaging is required [81–83]. nuclear medicine or other functional ischemia tests to analyze
Imaging targets and reporting elements of CT for CHD borderline stenosis. The results of the CCTA can be broadly
depend on the nature of the lesion and prior surgical inter- described in the conclusion as normal coronary arteries (no
ventions, e.g., right ventricular size, function, right ventricu- plaque or stenosis), nonobstructive CAD (1–49% stenosis),
lar outflow tract anatomy and branch pulmonary arteries or significant CAD (>/=50% or >/=70% stenosis) with further
after tetralogy of Fallot repair, recoarctation or pseudoaneu- qualification on the location and extent of disease. If the
rysm formation after coarctation repair, and patency of baf- CAD-RADS reporting and data system is used, the specific
fles and conduits after blood flow redirection surgery. CAD-RADS code for the findings and recommendations for
Structured reporting for CHD should follow a sequential further workup should be mentioned in the impressions.
188 A. Attili and E. A. Kazerooni

Table 16.5  Components of comprehensive cardiac CT report

Specific component(s) Necessity
Clinical data
General Indication or reason for test, procedure date Required
Demographics Name, date of birth, sex, referring clinician height, weight, BMI Required
History Symptoms, risk factors, relevant diagnostic tests Recommended
Procedure data
Description Test type (e.g., coronary CT angiography, calcium scoring, ventricular function, pulmonary Required
vein, others)
Equipment Scanner type: Number of detectors, rotation time, number of X-ray sources Recommended
Acquisition Gating method: Retrospective vs prospective ECG synchronization Required
Tube voltage, tube current, dose modulation (if used) Recommended
Estimated radiation dose, e.g., dose length product Recommended
Reconstruction Scanned or reconstructed phase of cardiac cycle Recommended
Slice thickness, increment, reconstruction filter Required
Medications Contrast type, volume, β-blockers, nitroglycerin, or any other, if given Required
Contrast rate Recommended
Patient parameters Complication(s), if present Required
Heart rate, arrhythmia, if present Recommended
Results
Technical quality Overall quality Required
Presence and type of artifact and effect on interpretation Required
Coronary Calcium score (if calcium scan performed) Required
Coronary anatomy: anomalies (origins and course), dominance, dilatation/aneurysms, Required
anatomical variance, bridging
Stenosis location and severity Required
Uninterpretable segments, arteries, or overall study Required
Stenosis plaque type: Calcified, non-calcified, or partially calcified Required
Stenosis extent and vulnerable features: length, ostial or branch involvement, positive Recommended
remodeling, low-attenuation plaque, napkin-ring sign, spotty calcifications, tortuosity
Use of SCCT stenosis severity classification Required
Use of SCCT axial coronary segmentation model Recommended
Calcium score percentile (if calcium scan performed) Required
Prior cardiac procedures Stents: Location, interpretability, patency Required
Prior CABG: Type, location, course and anastomosis, interpretability, patency, stenosis
Non-coronary cardiac
structures
Vessels Abnormalities of the aorta, vena cava, pulmonary arteries, and pulmonary veins, if present Required
Pulmonary vein morphology and ostia sizes (required for pre-ablation studies) Required
Cardiac chambers Abnormal chamber dilation, hypertrophy, masses, thrombus, shunts, and other structural Required
diseases, if present
Left ventricular size, EF, and volume (if function data is obtained) Required
Left atrial volume (for pre-ablation studies) Recommended
Right ventricular size and volume (if functional data obtained) Required
Left ventricular wall motion (17-segment model) and EF (if functional data obtained) Required
Myocardium Evidence of prior MI (hypoperfusion, wall thinning, fat, calcification, aneurysms) Required
End-diastolic left ventricular wall thickness Recommended
Perfusion defects and abnormal enhancement using a 17-segment model (if performed) Required
Pericardium Abnormal thickness, calcification, effusion if present Required
Valves Abnormal aortic and mitral valve calcification, thickness, stenosis, morphology Required
Prosthetic valves: type, location, pannus, thrombus, mobility
Non-cardiac Abnormalities in the lungs, mediastinum, esophagus, bony structures, chest wall, etc., if present Required
Impressions and Coronary interpretation Required
conclusions Abnormal non-coronary cardiac findings Required
Abnormal non-cardiac findings Required
Correlation to other or prior cardiac studies Recommended
Documentation of communication to referring physician for urgent finding(s) Recommended
Clinical recommendations Recommended
CAD-RADS code Recommended
16  Structured Reporting for Cardiac CT 189

In case of extracardiac findings, their relevance and any summarized in Table 16.5. A sample standardized reporting
recommendations for further workup should be stated. template for CCTA with calcium scoring (normal examina-
The components of a comprehensive cardiac CT report are tion) is provided below in Table 16.6.

Table 16.6  A sample standardized reporting template for CCTA (normal examination)
Procedure: CT cardiac (without and with contrast, with 3D reformatting)—coronary CTA protocol including coronary calcium CT
Clinical history: [ ]
Comparison: [ ]
Technique: Using a [scanner type], a preliminary scout study was obtained, followed by coronary artery calcium protocol. Following
administration of [x] ml iodinated intravenous contrast [contrast name], [0.5] mm collimated images were obtained from the carina to the base
of the heart. [Prospective; Retrospective>] ECG triggering was used. Heart rate at the time of acquisition was approximately [ ] bpm.
Multiplanar and 3D reformatted images were rendered and reviewed on an advanced processing workstation to further define anatomy and
possible pathology (DFOV = 25/cm)
Medications: [100 mg of oral metoprolol and x mg IV metoprolol was administered prior to scanning]. [0.4 mg sublingual nitroglycerine was
administered immediately prior to scanning]
Technical quality: [excellent, with no artifacts; good, with minor artifact but good diagnostic quality; acceptable, with moderate artifacts;
poor/suboptimal, with severe artifacts]
Findings
CT coronary calcium scoring
 LMA =
 LAD =
 LCX =
 RCA =
 PDA =
Total calcium score = [ ] using the AJ-130 method. This score is in the percentile rank for age and gender, meaning that [ ] % of patients of the
same age and gender will have a higher score. The total volume score is [ ]
Coronary CT angiography
The coronary artery origins and proximal course are normal. There is ____ (right/left/co) coronary artery dominance
Note: Stenoses are reported as maximum percentage diameter stenosis
Stenosis grading is reported using the following scheme:
Normal: no plaque
Minimal: negligible impact on the lumen (<25% stenosis)
Mild: no flow-limiting stenosis (25–49% stenosis)
Moderate: possible flow-limiting disease (50–69% stenosis)
Severe: probable flow-limiting disease (70–99% stenosis)
Occluded
Left main: The left main coronary artery is a _____ (short/medium/large) size vessel and (bifurcates in LAD and LCX/or trifurcates in LAD,
LCX, and a ramus branch). It is patent with no evidence of plaque or stenosis
LAD: The left anterior descending artery is patent with no evidence of plaque or stenosis. It gives off ____ patent diagonal branches
LCX: The left circumflex artery is patent with no evidence of plaque or stenosis. It gives off ____ patent obtuse marginal branches
RCA: The right coronary artery is patent with no evidence of plaque or stenosis. It gives off a patent posterior descending artery and a patent
posterior left ventricular branch
Cardiac structure and morphology
Cardiac chambers: Non-enlarged. No myocardial hypertrophy
Cardiac valves: There is no thickening or calcifications in the aortic and mitral valves
Pericardium: No effusion, thickening, or calcifications
Chest
Within the imaged field of view, evaluation of the visualized thoracic and upper abdominal structures demonstrates no additional
abnormalities; specifically, there are no pulmonary nodules or enlarged lymph nodes. There is a three-leaflet aortic valve with no aortic leaflet
calcification. The aortic root is normal in diameter, measuring – cm. The central pulmonary arteries are unremarkable
Impression
1. Total coronary calcium score = [ ] and is at the [ ] percentile for age and gender
2. Normal coronary arteries
3. Normal cardiac chamber size
4. Unremarkable aortic valve and aortic root
5. Unremarkable pericardium
190 A. Attili and E. A. Kazerooni

Time Line for Report Distribution clinicians’ comprehension of coronary CT angiography reports. J

Am Coll Radiol JACR. 2013;10(6):432–8.
6. Litt HI, Gatsonis C, Snyder B, Singh H, Miller CD, Entrikin DW,
To expedite patient flow in the emergency department (ED), Leaming JM, Gavin LJ, Pacella CB, Hollander JE.  CT angiogra-
cardiac CT studies for acute chest pain should be interpreted phy for safe discharge of patients with possible acute coronary syn-
and reported within 60 min of reconstruction. It is also rec- dromes. N Engl J Med. 2012;366(15):1393–403.
7. Goldstein JA, Chinnaiyan KM, Abidov A, Achenbach S, Berman
ommended that all emergency scans including normal stud- DS, Hayes SW, Hoffmann U, Lesser JR, Mikati IA, O'Neil BJ,
ies be reported verbally to the ED providers and that all Shaw LJ, Shen MY, Valeti US, Raff GL, Investigators C-S.  The
exams with moderate- or high-grade stenosis be communi- CT-STAT (coronary computed tomographic angiography for sys-
cated in all clinical settings, with documentation of the ver- tematic triage of acute chest pain patients to treatment) trial. J Am
Coll Cardiol. 2011;58(14):1414–22.
bal communication in the report. Physicians should be 8. Investigators S-H. CT coronary angiography in patients with sus-
notified immediately of life-threatening critical findings, pected angina due to coronary heart disease (SCOT-HEART):
such as pulmonary embolism or aortic dissection. Similarly, an open-label, parallel-group, multicentre trial. Lancet.
incidental unexpected findings requiring follow-up that may 2015;385(9985):2383–91.
9. Douglas PS, Hoffmann U, Patel MR, Mark DB, Al-Khalidi HR,
lead to later downstream medical problems, such as lung Cavanaugh B, Cole J, Dolor RJ, Fordyce CB, Huang M, Khan
nodules requiring follow-up CT exams to ensure their lack of MA, Kosinski AS, Krucoff MW, Malhotra V, Picard MH, Udelson
growth, should be communicated and documented. Elective JE, Velazquez EJ, Yow E, Cooper LS, Lee KL, Investigators
studies should be reported within 1–2 working days of the P. Outcomes of anatomical versus functional testing for coronary
artery disease. N Engl J Med. 2015;372(14):1291–300.
procedure. 10. Raff GL, Chinnaiyan KM, Cury RC, Garcia MT, Hecht HS,

Hollander JE, O’Neil B, Taylor AJ, Hoffmann U, Society of
Cardiovascular Computed Tomography Guidelines C.  SCCT
Conclusion guidelines on the use of coronary computed tomographic angiogra-
phy for patients presenting with acute chest pain to the emergency
department: a report of the Society of Cardiovascular Computed
The use of structured reporting in cardiac CT is recom- Tomography Guidelines Committee. J Cardiovasc Comput Tomogr.
mended to improve communication of results to referring 2014;8(4):254–71.
physicians. Furthermore structured reporting benefits 11. Taylor AJ, Cerqueira M, Hodgson JM, Mark D, Min J, O'Gara
P, Rubin GD.  ACCF/SCCT/ACR/AHA/ASE/ASNC/NASCI/
research, education, peer review, and quality assurance and SCAI/SCMR 2010 appropriate use criteria for cardiac computed
results in improved quality of care. The structured report tomography. A report of the American College of Cardiology
should contain adequate information to support the clinical Foundation Appropriate Use Criteria Task Force, the Society of
necessity for the procedure and a comprehensive description Cardiovascular Computed Tomography, the American College of
Radiology, the American Heart Association, the American Society
of the findings and their significance using standardized ter- of Echocardiography, the American Society of Nuclear Cardiology,
minology to enable proper clinical management. the North American Society for Cardiovascular Imaging, the
Society for Cardiovascular Angiography and Interventions, and
the Society for Cardiovascular Magnetic Resonance. J Cardiovasc
Comput Tomogr. 2010;4(6):407 e1–33.
References 12. Greenland P, Bonow RO, Brundage BH, Budoff MJ, Eisenberg MJ,
Grundy SM, Lauer MS, Post WS, Raggi P, Redberg RF, Rodgers
1. Sundaram B, Patel S, Bogot N, Kazerooni EA. Anatomy and ter- GP, Shaw LJ, Taylor AJ, Weintraub WS, Harrington RA, Abrams
minology for the interpretation and reporting of cardiac MDCT: J, Anderson JL, Bates ER, Grines CL, Hlatky MA, Lichtenberg
part 1, Structured report, coronary calcium screening, and coronary RC, Lindner JR, Pohost GM, Schofield RS, Shubrooks SJ Jr,
artery anatomy. AJR Am J Roentgenol. 2009;192(3):574–83. Stein JH, Tracy CM, Vogel RA, Wesley DJ, American College
2. Sundaram B, Patel S, Agarwal P, Kazerooni EA. Anatomy and ter- of Cardiology Foundation Clinical Expert Consensus Task F,
minology for the interpretation and reporting of cardiac MDCT: Society of Atherosclerosis I, Prevention, Society of Cardiovascular
part 2, CT angiography, cardiac function assessment, and non- Computed T.  ACCF/AHA 2007 clinical expert consensus docu-
coronary and extracardiac findings. AJR Am J Roentgenol. ment on coronary artery calcium scoring by computed tomog-
2009;192(3):584–98. raphy in global cardiovascular risk assessment and in evaluation
3. Stillman AE, Rubin GD, Teague SD, White RD, Woodard PK, of patients with chest pain: a report of the American College of
Larson PA.  Structured reporting: coronary CT angiography: a Cardiology Foundation Clinical Expert Consensus Task Force
white paper from the American College of Radiology and the North (ACCF/AHA Writing Committee to Update the 2000 Expert
American Society for Cardiovascular Imaging. J Am Coll Radiol Consensus Document on Electron Beam Computed Tomography).
JACR. 2008;5(7):796–800. Circulation. 2007;115(3):402–26.
4. Leipsic J, Abbara S, Achenbach S, Cury R, Earls JP, Mancini GJ, 13. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M
Nieman K, Pontone G, Raff GL.  SCCT guidelines for the inter- Jr, Detrano R. Quantification of coronary artery calcium using ultra-
pretation and reporting of coronary CT angiography: a report of fast computed tomography. J Am Coll Cardiol. 1990;15(4):827–32.
the Society of Cardiovascular Computed Tomography Guidelines 14. Detrano R, Guerci AD, Carr JJ, Bild DE, Burke G, Folsom AR, Liu
Committee. J Cardiovasc Comput Tomogr. 2014;8(5):342–58. K, Shea S, Szklo M, Bluemke DA, O'Leary DH, Tracy R, Watson
5. Ghoshhajra BB, Lee AM, Ferencik M, Elmariah S, Margey RJ, K, Wong ND, Kronmal RA.  Coronary calcium as a predictor of
Onuma O, Panagia M, Abbara S, Hoffmann U.  Interpreting the coronary events in four racial or ethnic groups. N Engl J Med.
interpretations: the use of structured reporting improves referring 2008;358(13):1336–45.
16  Structured Reporting for Cardiac CT 191

15. Rutten A, Isgum I, Prokop M.  Coronary calcification: effect of 28. Motoyama S, Sarai M, Harigaya H, Anno H, Inoue K, Hara T,
small variation of scan starting position on Agatston, volume, and Naruse H, Ishii J, Hishida H, Wong ND, Virmani R, Kondo T, Ozaki
mass scores. Radiology. 2008;246(1):90–8. Y, Narula J. Computed tomographic angiography characteristics of
16. Hong C, Becker CR, Schoepf UJ, Ohnesorge B, Bruening R, Reiser atherosclerotic plaques subsequently resulting in acute coronary
MF.  Coronary artery calcium: absolute quantification in nonen- syndrome. J Am Coll Cardiol. 2009;54(1):49–57.
hanced and contrast-enhanced multi-detector row CT studies. 29. Hsu JT, Kyo E, Chu CM, Tsuji T, Watanabe S. Impact of calcifica-
Radiology. 2002;223(2):474–80. tion length ratio on the intervention for chronic total occlusions. Int
17. Rumberger JA, Kaufman L. A rosetta stone for coronary calcium J Cardiol. 2011;150(2):135–41.
risk stratification: agatston, volume, and mass scores in 11,490 indi- 30. Martin-Yuste V, Barros A, Leta R, Ferreira I, Brugaletta S, Pujadas
viduals. AJR Am J Roentgenol. 2003;181(3):743–8. S, Carreras F, Pons G, Cinca J, Sabate M. Factors determining suc-
18. Callister TQ, Cooil B, Raya SP, Lippolis NJ, Russo DJ, Raggi cess in percutaneous revascularization of chronic total coronary
P.  Coronary artery disease: improved reproducibility of calcium occlusion: multidetector computed tomography analysis. Rev Esp
scoring with an electron-beam CT volumetric method. Radiology. Cardiol. 2012;65(4):334–40.
1998;208(3):807–14. 31. Yokoyama N, Yamamoto Y, Suzuki S, Suzuki M, Konno K, Kozuma
19. Hecht HS, Cronin P, Blaha MJ, Budoff MJ, Kazerooni EA, Narula J, K, Kaminaga T, Isshiki T. Impact of 16-slice computed tomography
Yankelevitz D, Abbara S. 2016 SCCT/STR guidelines for coronary in percutaneous coronary intervention of chronic total occlusions.
artery calcium scoring of noncontrast noncardiac chest CT scans Catheter Cardiovasc Interv. 2006;68(1):1–7.
a report of the Society of Cardiovascular Computed Tomography 32. Luo C, Huang M, Li J, Liang C, Zhang Q, Liu H, Liu Z, Qu Y, Jiang
and Society of Thoracic Radiology. J Cardiovasc Comput Tomogr. J, Zhuang J. Predictors of interventional success of antegrade PCI
2017;11(1):74–84. for CTO. J Am Coll Cardiol Img. 2015;8(7):804–813.8.
20. Austen WG, Edwards JE, Frye RL, Gensini GG, Gott VL, Griffith 33. Chan M, Ridley L, Dunn DJ, Tian DH, Liou K, Ozdirik J, Cheruvu
LS, McGoon DC, Murphy ML, Roe BB.  A reporting system on C, Cao C. A systematic review and meta-analysis of multidetector
patients evaluated for coronary artery disease. Report of the Ad Hoc computed tomography in the assessment of coronary artery bypass
Committee for Grading of Coronary Artery Disease, Council on grafts. Int J Cardiol. 2016;221:898–905.
Cardiovascular Surgery, American Heart Association. Circulation. 34. Kamdar AR, Meadows TA, Roselli EE, Gorodeski EZ, Curtin

1975;51(4 Suppl):5–40. RJ, Sabik JF, Schoenhagen P, White RD, Lytle BW, Flamm SD,
21. Cheng V, Gutstein A, Wolak A, Suzuki Y, Dey D, Gransar H, Desai MY.  Multidetector computed tomographic angiography in
Thomson LE, Hayes SW, Friedman JD, Berman DS.  Moving planning of reoperative cardiothoracic surgery. Ann Thorac Surg.
beyond binary grading of coronary arterial stenoses on coronary 2008;85(4):1239–45.
computed tomographic angiography: insights for the imager and 35. Aviram G, Sharony R, Kramer A, Nesher N, Loberman D, Ben-Gal
referring clinician. JACC Cardiovasc Imaging. 2008;1(4):460–71. Y, Graif M, Uretzky G, Mohr R. Modification of surgical planning
22. Achenbach S, Moselewski F, Ropers D, Ferencik M, Hoffmann U, based on cardiac multidetector computed tomography in reopera-
MacNeill B, Pohle K, Baum U, Anders K, Jang IK, Daniel WG, tive heart surgery. Ann Thorac Surg. 2005;79(2):589–95.
Brady TJ. Detection of calcified and noncalcified coronary athero- 36. Imran Hamid U, Digney R, Soo L, Leung S, Graham AN. Incidence
sclerotic plaque by contrast-enhanced, submillimeter multidetector and outcome of re-entry injury in redo cardiac surgery: benefits of
spiral computed tomography: a segment-based comparison with preoperative planning. Eur J Cardiothorac Surg. 2015;47(5):819–23.
intravascular ultrasound. Circulation. 2004;109(1):14–7. 37. de Graaf FR, Schuijf JD, van Velzen JE, Boogers MJ, Kroft LJ, de
23. Miller JM, Rochitte CE, Dewey M, Arbab-Zadeh A, Niinuma
Roos A, Reiber JH, Sieders A, Spano F, Jukema JW, Schalij MJ, van
H, Gottlieb I, Paul N, Clouse ME, Shapiro EP, Hoe J, Lardo AC, der Wall EE, Bax JJ. Diagnostic accuracy of 320-row multidetec-
Bush DE, de Roos A, Cox C, Brinker J, Lima JA. Diagnostic per- tor computed tomography coronary angiography to noninvasively
formance of coronary angiography by 64-row CT. N Engl J Med. assess in-stent restenosis. Investig Radiol. 2010;45(6):331–40.
2008;359(22):2324–36. 38. Kumbhani DJ, Ingelmo CP, Schoenhagen P, Curtin RJ, Flamm SD,
24. Raff GL, Gallagher MJ, O'Neill WW, Goldstein JA.  Diagnostic Desai MY.  Meta-analysis of diagnostic efficacy of 64-slice com-
accuracy of noninvasive coronary angiography using 64-slice spiral puted tomography in the evaluation of coronary in-stent restenosis.
computed tomography. J Am Coll Cardiol. 2005;46(3):552–7. Am J Cardiol. 2009;103(12):1675–81.
25. Budoff MJ, Dowe D, Jollis JG, Gitter M, Sutherland J, Halamert 39. Carrabba N, Schuijf JD, de Graaf FR, Parodi G, Maffei E, Valenti
E, Scherer M, Bellinger R, Martin A, Benton R, Delago A, Min R, Palumbo A, Weustink AC, Mollet NR, Accetta G, Cademartiri F,
JK.  Diagnostic performance of 64-multidetector row coronary Antoniucci D, Bax JJ.  Diagnostic accuracy of 64-slice computed
computed tomographic angiography for evaluation of coronary tomography coronary angiography for the detection of in-stent
artery stenosis in individuals without known coronary artery restenosis: a meta-analysis. J Nucl Cardiol. 2010;17(3):470–8.
disease: results from the prospective multicenter ACCURACY 40. Stojanovska J, Prasitdumrong H, Patel S, Sundaram B, Gross BH,
(Assessment by Coronary Computed Tomographic Angiography Yilmaz ZN, Kazerooni EA.  Reference absolute and indexed val-
of Individuals Undergoing Invasive Coronary Angiography) trial. J ues for left and right ventricular volume, function and mass from
Am Coll Cardiol. 2008;52(21):1724–32. cardiac computed tomography. J Med Imaging Radiat Oncol.
26. Weigold WG, Abbara S, Achenbach S, Arbab-Zadeh A, Berman 2014;58(5):547–58.
D, Carr JJ, Cury RC, Halliburton SS, McCollough CH, Taylor AJ, 41. Lin FY, Devereux RB, Roman MJ, Meng J, Jow VM, Jacobs A,
Society of Cardiovascular Computed T. Standardized medical ter- Weinsaft JW, Shaw LJ, Berman DS, Callister TQ, Min JK. Cardiac
minology for cardiac computed tomography: a report of the Society chamber volumes, function, and mass as determined by
of Cardiovascular Computed Tomography. J Cardiovasc Comput 64-­multidetector row computed tomography: mean values among
Tomogr. 2011;5(3):136–44. healthy adults free of hypertension and obesity. JACC Cardiovasc
27. Puchner SB, Liu T, Mayrhofer T, Truong QA, Lee H, Fleg JL, Imaging. 2008;1(6):782–6.
Nagurney JT, Udelson JE, Hoffmann U, Ferencik M.  High-risk 42. Cerqueira MD, Weissman NJ, Dilsizian V, Jacobs AK, Kaul

plaque detected on coronary CT angiography predicts acute coro- S, Laskey WK, Pennell DJ, Rumberger JA, Ryan T, Verani MS,
nary syndromes independent of significant stenosis in acute chest American Heart Association Writing Group on Myocardial S,
pain: results from the ROMICAT-II trial. J Am Coll Cardiol. Registration for Cardiac I.  Standardized myocardial segmenta-
2014;64(7):684–92. tion and nomenclature for tomographic imaging of the heart. A
192 A. Attili and E. A. Kazerooni

statement for healthcare professionals from the Cardiac Imaging 54. Romero J, Husain SA, Kelesidis I, Sanz J, Medina HM, Garcia
Committee of the Council on Clinical Cardiology of the American MJ. Detection of left atrial appendage thrombus by cardiac com-
Heart Association. Circulation. 2002;105(4):539–42. puted tomography in patients with atrial fibrillation: a meta-­
43. Feuchtner GM, Dichtl W, Friedrich GJ, Frick M, Alber H, Schachner analysis. Circ Cardiovasc Imaging. 2013;6(2):185–94.
T, Bonatti J, Mallouhi A, Frede T, Pachinger O, zur Nedden D, 55. Horton R, Di Biase L, Reddy V, Neuzil P, Mohanty P, Sanchez J,
Muller S. Multislice computed tomography for detection of patients Nguyen T, Mohanty S, Gallinghouse GJ, Bailey SM, Zagrodzky
with aortic valve stenosis and quantification of severity. J Am Coll JD, Burkhardt JD, Natale A.  Locating the right phrenic nerve by
Cardiol. 2006;47(7):1410–7. imaging the right pericardiophrenic artery with computerized
44. LaBounty TM, Sundaram B, Agarwal P, Armstrong WA, Kazerooni tomographic angiography: implications for balloon-based proce-
EA, Yamada E.  Aortic valve area on 64-MDCT correlates with dures. Heart Rhythm. 2010;7(7):937–41.
transesophageal echocardiography in aortic stenosis. AJR Am J 56. Lemola K, Sneider M, Desjardins B, Case I, Han J, Good E,
Roentgenol. 2008;191(6):1652–8. Tamirisa K, Tsemo A, Chugh A, Bogun F, Pelosi F Jr, Kazerooni E,
45. Cury RC, Abbara S, Achenbach S, Agatston A, Berman DS,
Morady F, Oral H. Computed tomographic analysis of the anatomy
Budoff MJ, Dill KE, Jacobs JE, Maroules CD, Rubin GD, Rybicki of the left atrium and the esophagus: implications for left atrial
FJ, Schoepf UJ, Shaw LJ, Stillman AE, White CS, Woodard PK, catheter ablation. Circulation. 2004;110(24):3655–60.
Leipsic JA.  CAD-RADS: Coronary Artery Disease  – Reporting 57. Good E, Oral H, Lemola K, Han J, Tamirisa K, Igic P, Elmouchi
and Data System: an expert consensus document of the Society D, Tschopp D, Reich S, Chugh A, Bogun F, Pelosi F Jr, Morady
of Cardiovascular Computed Tomography (SCCT), the American F. Movement of the esophagus during left atrial catheter ablation
College of Radiology (ACR) and the North American Society for atrial fibrillation. J Am Coll Cardiol. 2005;46(11):2107–10.
for Cardiovascular Imaging (NASCI). Endorsed by the American 58. Komatsu Y, Cochet H, Jadidi A, Sacher F, Shah A, Derval N, Scherr
College of Cardiology. J Am Coll Radiol JACR. 2016;13(12 Pt D, Pascale P, Roten L, Denis A, Ramoul K, Miyazaki S, Daly M,
A):1458–1466 e1459. Riffaud M, Sermesant M, Relan J, Ayache N, Kim S, Montaudon
46. Fihn SD, Gardin JM, Abrams J, Berra K, Blankenship JC, Dallas AP, M, Laurent F, Hocini M, Haissaguerre M, Jais P. Regional myocar-
Douglas PS, Foody JM, Gerber TC, Hinderliter AL, King SB III, dial wall thinning at multidetector computed tomography correlates
Kligfield PD, Krumholz HM, Kwong RYK, Lim MJ, Linderbaum to arrhythmogenic substrate in postinfarction ventricular tachycar-
JA, Mack MJ, Munger MA, Prager RL, Sabik JF, Shaw LJ, Sikkema dia: assessment of structural and electrical substrate. Circ Arrhythm
JD, Smith CR Jr, Smith SC Jr, Spertus JA, Williams SV. ACCF/ Electrophysiol. 2013;6(2):342–50.
AHA/ACP/AATS/PCNA/SCAI/STS guideline for the diagnosis and 59. Desjardins B, Morady F, Bogun F. Effect of epicardial fat on elec-
management of patients with stable ischemic heart disease: a report troanatomical mapping and epicardial catheter ablation. J Am Coll
of the American College of Cardiology Foundation/American Heart Cardiol. 2010;56(16):1320–7.
Association Task Force on, American Association for Thoracic 60. Njeim M, Desjardins B, Bogun F. Multimodality imaging for guid-
Surgery, Preventive Cardiovascular Nurses Association, Society ing EP ablation procedures. J Am Coll Cardiol Img. 2016;9(7):873–
for Cardiovascular Angiography and Interventions, and Society of 86. 9.
Thoracic Surgeons. J Am Coll Cardiol. 2012;60:e44–164. 61. Achenbach S, Delgado V, Hausleiter J, Schoenhagen P, Min JK,
47. Ang R, Hunter RJ, Baker V, Richmond L, Dhinoja M, Sporton S, Leipsic JA.  SCCT expert consensus document on computed
Schilling RJ, Pugliese F, Davies C, Earley M. Pulmonary vein mea- tomography imaging before transcatheter aortic valve implanta-
surements on pre-procedural CT/MR imaging can predict difficult tion (TAVI)/transcatheter aortic valve replacement (TAVR). J
pulmonary vein isolation and phrenic nerve injury during cryo- Cardiovasc Comput Tomogr. 2012;6(6):366–80.
balloon ablation for paroxysmal atrial fibrillation. Int J Cardiol. 62. Blanke P, Dvir D, Cheung A, Levine RA, Thompson C, Webb
2015;195:253–8. JG, Leipsic J. Mitral annular evaluation with CT in the context of
48. Cronin P, Kelly AM, Desjardins B, Patel S, Gross BH, Kazerooni transcatheter mitral valve replacement. J Am Coll Cardiol Img.
EA, Morady F, Oral H, Carlos RC. Normative analysis of pulmo- 2015;8(5):612–5.
nary vein drainage patterns on multidetector CT with measurements 63. Natarajan N, Patel P, Bartel T, Kapadia S, Navia J, Stewart W,
of pulmonary vein ostial diameter and distance to first bifurcation. Tuzcu EM, Schoenhagen P.  Peri-procedural imaging for trans-
Acad Radiol. 2007;14(2):178–88. catheter mitral valve replacement. Cardiovasc Diagn Ther.
49. Bertaglia E, Bella PD, Tondo C, Proclemer A, Bottoni N, De Ponti 2016;6(2):144–59.
R, Landolina M, Bongiorni MG, Coro L, Stabile G, Dello Russo A, 64. Xu H, Xie X, Wang B, Ma S, Wang F. Efficacy and safety of per-
Verlato R, Mantica M, Zoppo F. Image integration increases effi- cutaneous left atrial appendage occlusion for stroke prevention in
cacy of paroxysmal atrial fibrillation catheter ablation: results from nonvalvular atrial fibrillation: a meta-analysis of contemporary
the CartoMerge Italian Registry. Europace. 2009;11(8):1004–10. studies. Heart Lung Circ. 2016;25(11):1107–17.
50. Cronin P, Sneider MB, Kazerooni EA, Kelly AM, Scharf C, Oral H, 65. Holmes DR Jr, Reddy VY. Left atrial appendage and closure: who,
Morady F. MDCT of the left atrium and pulmonary veins in plan- when, and how. Circ Cardiovasc Interv. 2016;9(5):e002942.
ning radiofrequency ablation for atrial fibrillation: a how-to guide. 66. Kwong Y, Troupis J.  Cardiac CT imaging in the context of

AJR Am J Roentgenol. 2004;183(3):767–78. left atrial appendage occlusion. J Cardiovasc Comput Tomogr.
51. Stojanovska J, Cronin P, Patel S, Gross BH, Oral H, Chughtai K, 2015;9(1):13–8.
Kazerooni EA. Reference normal absolute and indexed values from 67. Di Biase L, Santangeli P, Anselmino M, Mohanty P, Salvetti I, Gili
ECG-gated MDCT: left atrial volume, function, and diameter. AJR S, Horton R, Sanchez JE, Bai R, Mohanty S, Pump A, Cereceda
Am J Roentgenol. 2011;197(3):631–7. Brantes M, Gallinghouse GJ, Burkhardt JD, Cesarani F, Scaglione
52. Parikh SS, Jons C, McNitt S, Daubert JP, Schwarz KQ, Hall M, Natale A, Gaita F.  Does the left atrial appendage morphol-
B. Predictive capability of left atrial size measured by CT, TEE, and ogy correlate with the risk of stroke in patients with atrial fibril-
TTE for recurrence of atrial fibrillation following radiofrequency cath- lation? Results from a multicenter study. J Am Coll Cardiol.
eter ablation. Pacing Clin Electrophysiol PACE. 2010;33(5):532–40. 2012;60(6):531–8.
53. Stojanovska J, Cronin P, Gross BH, Kazerooni EA, Tsodikov A, 68. Wang Y, Di Biase L, Horton RP, Nguyen T, Morhanty P, Natale
Frank L, Oral H.  Left atrial function and maximum volume as A. Left atrial appendage studied by computed tomography to help
determined by MDCT are independently associated with atrial planning for appendage closure device placement. J Cardiovasc
fibrillation. Acad Radiol. 2014;21(9):1162–71. Electrophysiol. 2010;21(9):973–82.
16  Structured Reporting for Cardiac CT 193

69. Bamberg F, Becker A, Schwarz F, Marcus RP, Greif M, von

77. Habis M, Capderou A, Ghostine S, Daoud B, Caussin C, Riou
Ziegler F, Blankstein R, Hoffmann U, Sommer WH, Hoffmann VS, JY, Brenot P, Angel CY, Lancelin B, Paul JF.  Acute myocardial
Johnson TR, Becker HC, Wintersperger BJ, Reiser MF, Nikolaou infarction early viability assessment by 64-slice computed tomog-
K. Detection of hemodynamically significant coronary artery ste- raphy immediately after coronary angiography: comparison with
nosis: incremental diagnostic value of dynamic CT-based myocar- low-dose dobutamine echocardiography. J Am Coll Cardiol.
dial perfusion imaging. Radiology. 2011;260(3):689–98. 2007;49(11):1178–85.
70. Blankstein R, Shturman LD, Rogers IS, Rocha-Filho JA, Okada 78. Jacquier A, Boussel L, Amabile N, Bartoli JM, Douek P, Moulin G,
DR, Sarwar A, Soni AV, Bezerra H, Ghoshhajra BB, Petranovic M, Paganelli F, Saeed M, Revel D, Croisille P. Multidetector computed
Loureiro R, Feuchtner G, Gewirtz H, Hoffmann U, Mamuya WS, tomography in reperfused acute myocardial infarction. Assessment
Brady TJ, Cury RC.  Adenosine-induced stress myocardial perfu- of infarct size and no-reflow in comparison with cardiac magnetic
sion imaging using dual-source cardiac computed tomography. J resonance imaging. Investig Radiol. 2008;43(11):773–81.
Am Coll Cardiol. 2009;54(12):1072–84. 79. Paul JF, Wartski M, Caussin C, Sigal-Cinqualbre A, Lancelin B,
71. Rochitte CE, George RT, Chen MY, Arbab-Zadeh A, Dewey M, Angel C, Dambrin G.  Late defect on delayed contrast-enhanced
Miller JM, Niinuma H, Yoshioka K, Kitagawa K, Nakamori S, multi-detector row CT scans in the prediction of SPECT infarct
Laham R, Vavere AL, Cerci RJ, Mehra VC, Nomura C, Kofoed KF, size after reperfused acute myocardial infarction: initial experience.
Jinzaki M, Kuribayashi S, de Roos A, Laule M, Tan SY, Hoe J, Paul Radiology. 2005;236(2):485–9.
N, Rybicki FJ, Brinker JA, Arai AE, Cox C, Clouse ME, Di Carli 80. Rogers IS, Cury RC, Blankstein R, Shapiro MD, Nieman K,

MF, Lima JA. Computed tomography angiography and perfusion to Hoffmann U, Brady TJ, Abbara S. Comparison of postprocessing
assess coronary artery stenosis causing perfusion defects by single techniques for the detection of perfusion defects by cardiac com-
photon emission computed tomography: the CORE320 study. Eur puted tomography in patients presenting with acute ST-segment
Heart J. 2014;35(17):1120–30. elevation myocardial infarction. J Cardiovasc Comput Tomogr.
72. Mehra VC, Valdiviezo C, Arbab-Zadeh A, Ko BS, Seneviratne SK, 2010;4(4):258–66.
Cerci R, Lima JA, George RT. A stepwise approach to the visual 81. Han BK, Rigsby CK, Leipsic J, Bardo D, Abbara S, Ghoshhajra B,
interpretation of CT-based myocardial perfusion. J Cardiovasc Lesser JR, Raman SV, Crean AM, Nicol ED, Siegel MJ, Hlavacek
Comput Tomogr. 2011;5(6):357–69. A. Computed tomography imaging in patients with congenital heart
73. George RT, Silva C, Cordeiro MA, DiPaula A, Thompson DR, disease, part 2: technical recommendations. An expert consensus
McCarthy WF, Ichihara T, Lima JA, Lardo AC.  Multidetector document of the Society of Cardiovascular Computed Tomography
­computed tomography myocardial perfusion imaging during ade- (SCCT): endorsed by the Society of Pediatric Radiology (SPR)
nosine stress. J Am Coll Cardiol. 2006;48(1):153–60. and the North American Society of Cardiac Imaging (NASCI). J
74. Rossi A, Dharampal A, Wragg A, Davies LC, van Geuns RJ,
Cardiovasc Comput Tomogr. 2015;9(6):493–513.
Anagnostopoulos C, Klotz E, Kitslaar P, Broersen A, Mathur 82. Han BK, Rigsby CK, Hlavacek A, Leipsic J, Nicol ED, Siegel MJ,
A, Nieman K, Hunink MG, de Feyter PJ, Petersen SE, Pugliese Bardo D, Abbara S, Ghoshhajra B, Lesser JR, Raman S, Crean
F.  Diagnostic performance of hyperaemic myocardial blood flow AM.  Computed tomography imaging in patients with congeni-
index obtained by dynamic computed tomography: does it predict tal heart disease part I: rationale and utility. An expert consensus
functionally significant coronary lesions? Eur Heart J Cardiovasc document of the Society of Cardiovascular Computed Tomography
Imaging. 2014;15(1):85–94. (SCCT): endorsed by the Society of Pediatric Radiology (SPR)
75. Mahnken AH, Koos R, Katoh M, Wildberger JE, Spuentrup E, and the North American Society of Cardiac Imaging (NASCI). J
Buecker A, Gunther RW, Kuhl HP.  Assessment of myocardial Cardiovasc Comput Tomogr. 2015;9(6):475–92.
viability in reperfused acute myocardial infarction using 16-slice 83. Bhatt AB, Foster E, Kuehl K, Alpert J, Brabeck S, Crumb S,
computed tomography in comparison to magnetic resonance imag- Davidson WR Jr, Earing MG, Ghoshhajra BB, Karamlou T, Mital
ing. J Am Coll Cardiol. 2005;45(12):2042–7. S, Ting J, Tseng ZH, American Heart Association Council on
76. Boussel L, Ribagnac M, Bonnefoy E, Staat P, Elicker BM, Revel Clinical C.  Congenital heart disease in the older adult: a scien-
D, Douek P.  Assessment of acute myocardial infarction using tific statement from the American Heart Association. Circulation.
MDCT after percutaneous coronary intervention: comparison with 2015;131(21):1884–931.
MRI. AJR Am J Roentgenol. 2008;191(2):441–7.
 Integration of CT Data into Clinical
Workflows: Role of Modern IT 17
Infrastructure Including Cloud
Technology

Paul Schoenhagen and Mathis Zimmermann

Since its introduction to medical imaging in the 1970s [1, 2], Imaging and image review/reconstruction are increasingly
CT has developed into a central imaging modality in a wide part of a stepwise decision-making process, transforming tra-
array of clinical conditions. In cardiovascular medicine, this ditional single-observer reading and reporting to a process
includes emergent indication (e.g., acute aortic syndromes involving a team of interdisciplinary clinical specialists. This
[AAS], pulmonary embolism [PE]), many elective indica- trend is observed in several subspecialties including oncology
tions, and novel indications (e.g., structural and valvular (e.g., “tumor board” for cancer treatment planning) and car-
heart disease). Its use has increased exponentially, supported diovascular medicine (e.g., “heart team” in the context of
by rapid developments of scanner technology and software transcatheter aortic valve replacement, TAVR). These devel-
applications [3, 4]. opments require a new level of data accessibility and perfor-
Initial developments of cardiovascular CT technology mance of imaging systems, including ability to share data
were focused on scanner hardware and image generation. A beyond the “reading room” in large healthcare systems. It is
critical step was ECG-synchronized imaging acquisition, supported by novel developments of IT architecture, allowing
which allowed to minimize cardiac motion artifacts. sharing of a centrally stored dataset between multiple periph-
Subsequent, consecutive generations of scanners with eral workstations (client-server). Connection of scanners and
increased cranio-caudal coverage and speed of acquisition workstations into a network or “cloud” with integration into
expanded cardiovascular applications. Initially the acquired the entire electronic health record (EHR) allows exchange of
series of CT images were printed on film sheets and reviewed information across healthcare systems and supports multidis-
on film boxes (Figs. 17.1 and 17.2). This was subsequently ciplinary teams working on defined clinical workflows. It has
replaced by digital review on expensive workstations in been adopted rapidly for CT imaging. In this chapter we will
“reading rooms” close to the CT scanner. The true potential review these developments and discuss the impact on clinical
of CT lies in the ability to combine multiple thin-slice axial workflows involving CT imaging.
images into a 3-D dataset and digital manipulation. This was
realized with development of increasingly powerful worksta-
tions and software allowing post-processing resulting in  omputed Tomography: CT Scanner
C
unlimited image planes oblique to the acquisition plane and Review Software/Workstation
(Fig. 17.3). The ability of 3-D reconstruction has been a key
element for the modern use of cardiovascular CT. Importantly, The diagnostic use of computed tomography (CT) is based
complex image reconstruction is performed by several users, on developments in the field of physics during the 1970s [1,
including imaging specialist/radiologist and clinical inter- 2]. Since then CT has developed into an established diagnos-
ventionalist/surgeon. Review by these users is typically sep- tic modality in cardiovascular medicine. The diagnostic
arated by time and location. spectrum includes routine indications such as the assessment
of aortic, pulmonary, and coronary vascular disease, as well
as novel applications, including the evaluation in the context
of minimally invasive cardiothoracic surgery and transcath-
P. Schoenhagen (*) eter interventions.
Imaging Institute, Cleveland Clinic, Lerner College of Medicine, Initially used electron-beam technology (EBCT) has
Cleveland, OH, USA
almost completely been replaced by multi-detector CT
e-mail: [email protected]
(MDCT) systems. In MDCT systems, the gantry (X-ray tube
M. Zimmermann
and detector) rotates rapidly around the patient [5, 6].
Digital Health Services, Siemens Healthineers, Malvern, PA, USA

© Humana Press 2019 195

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_17
196 P. Schoenhagen and M. Zimmermann

Fig. 17.1  This figure shows

an old-fashioned viewing
station. The X-ray films show
images obtained from an
ancient mummy

Subsequent generations of multi-slice scanners with 8, 16,

64, 256, and 320 slice acquisition per rotation were intro-
duced in rapid progression over the last two decades [7–10].
Modern systems with fast gantry rotation and thin collimated
detector rows with ECG-synchronized acquisition of multi-
ple slices per gantry rotation are optimized for cardiovascu-
lar imaging. Dual-source scanners, with two X-ray tubes/
detector systems, allow further reduction in temporal resolu-
tion [11, 12]. Today, high-end scanners permit rotation times
as low as 250  ms, with resulting temporal resolution of
135 ms (single source) and 66 ms (dual source). Modern sys-
tems acquire data with a minimum collimated detector row
width of 0.5 mm resulting in isotropic spatial resolution of
around 0.5 × 0.5 × 0.5 mm. The acquired raw data is recon-
structed at the scanner and then sent to workstations and/or
storage systems (PACS).
Early generations of workstations were local systems with
powerful local graphics hardware for 3-D reconstruction. The
complete data from individual studies was transferred to the
workstations as DICOM data, either directly from the scanner
or via the PACS system. Because of the high cost of such
workstations, only a few would be available, typically in the
“reading room” close to the scanner. Sharing of data was lim-
ited to a few reconstructed images (printed or saved on CD or
sent by e-mail). If a clinician wanted to directly review
images, this was only possible in the reading room.
Subsequent generations of local client-server systems
Fig. 17.2  This figure shows a traditional film library where X-ray combine a central server with local access stations at several
images are archived and retrieved if necessary locations. In these systems, storage of the complete data and
17  Integration of CT Data into Clinical Workflows: Role of Modern IT Infrastructure Including Cloud Technology 197

Fig. 17.3  This figure shows

a modern workstation with
multiple monitors. The
workstation is connected to
the central data center, where
the imaging data is archived
and processed

data reconstruction is performed at the central server level allows to consider the CT data in the context of the clinical
(“remote rendering”). The reconstructed images are dis- pathway of the patient. Enterprise-wide access has the poten-
played at the local client workstation, allowing use of less tial to streamline the ordering process (e.g., avoiding unnec-
expensive computer systems, such as those used for access essary repeat studies, tracking of cumulative radiation
of the EHR. The central storage of images in a clinical data exposure) and allows collaborative reading and reporting.
center facilitates the creation of a network of connected Data transfer of discrete values enables structured reporting
workstations but requires sufficient bandwidth for enterprise-­ and data mining/ analytics for clinical research and adminis-
wide data transfer (Fig. 17.4). This approach protects clinical trative purposes.
data because data is maintained within the hospital IT net- In large healthcare systems, this requires a complex infra-
work. It enables shared access to the same imaging study for structure/network spanning across multiple locations. A cen-
different clinical groups, e.g., surgeons or referring physi- tral data center connects a large number of servers (“server
cians at multiple locations contributing diagnostic or inter- farm”), with integration between EHR, PACS systems, and
ventional information. additional 3-D/4-D analysis tools (Figs. 17.4 and 17.5). The
The most recent approach is the use of “cloud servers.” core storage provides access to all data generated in the
Advantages are use of readily available cheaper cloud stor- healthcare system, which can be visualized in a patient-­
age and almost unlimited processing power. It allows new centric fashion through the electronic health record (EHR),
applications, e.g., retrospective and even prospective (pre- PACS, or more dedicated systems. These systems already are
dictive) data analytics, specifically in the context of integra- a form of cloud computing, described as a “private medical-­
tion into the EHR.  However, it introduces new challenges grade cloud.” Hospital systems may be able to maintain a
and requirements, e.g., data security aspects of storing “private cloud” or alternatively use third-party data service,
patient health information in the cloud on third-party servers but issues of reliable patient identification and data security
(Fig. 17.5). need to be clarified. While technically complex, the benefit
of “online” sharing of data files between professionals at
multiple locations within a treatment network appears to be
Integration of CT/Imaging Data and EHR obvious [13, 14].
“Smart” computer systems are not limited to data storage
The digital availability of data and ability to share as part of but also contribute to its collection and analysis. Examples
the EHR support enhanced integration of imaging data into are computer-aided detection (CAD) systems in diagnostic
clinical pathways. Integration of CT data and EHR servers imaging/radiology and automatic data analysis. A recent
198 P. Schoenhagen and M. Zimmermann

care in order to build predictive models, both for individual

patients and larger populations [19–21].
While these approaches are exciting, it is important to
consider potential limitations. The tremendous amount and
complexity of patient-related data accumulating in large
healthcare institutions require advanced software and hard-
ware and a dedicated group of IT professionals for mainte-
nance [22]. Data has to be available 24/7, and there are
mandated requirements for long-term data storage. A par-
ticular challenge is unequivocal patient identification, in
particular if data is to be shared across large healthcare sys-
tems and with other institutions. Complex issues related to
data security and ownership (hospital system or patient)
have to be addressed. Because of the complexity and asso-
ciated cost of maintaining such “data clouds”, healthcare
systems may consider external providers for data manage-
ment, further complicating oversight and regulation
(Fig. 17.5).

Impact on Clinical Management

The potential clinical value of systems combining imaging

data and the EHR can be studied using clinical scenarios,
where CT imaging is central, e.g., transcatheter aortic valve
replacement (TAVR) and acute aortic syndromes (AAS).
In the case of AAS, regional treatment networks have
been established, which coordinate diagnosis, triage, and
treatment between local emergency rooms as the initial
point of contact, and central specialized centers, experi-
enced in definitive pharmacological, interventional, or sur-
gical treatment [23–27]. In contrast to acute myocardial
infarction (MI), AAS lacks simple diagnostic biomarkers.
Triage of patients with suspected AAS requires definitive
Fig. 17.4  This figure shows a server farm in a data center imaging. Computed tomography is typically performed
and interpreted at the initial point of access in the local
ED, where subspecialized expertise is often lacking. As
paper examined the usefulness of CAD for the diagnosis of such, false-­positive transfer activation is not uncommon,
lung nodules on CT scans [15]. CAD systems detected lung especially in the setting of complex prior endovascular
cancers that were initially missed by a radiologist but failed repair [28–30]. The large image data files have tradition-
to detect other lung lesions identified by the radiologist, sug- ally been transferred with the patient either in film or CD
gesting a complementary role of CAD systems. Automatic format. At the tertiary referral center, the images were then
analysis (“data mining”) of large datasets can be performed downloaded and subsequently reinterpreted to form the
with minimal human input. This process is called “machine basis for definitive treatment. The availability of imaging
learning” (ML) and is used for various applications ­including data within a shared EHR would allow proactive triage
weather forecasting or recommending items of interest to prior or during patient transfer by specialists available
consumers while using online search engines [16–18]. within the network at different locations [31, 32]. Modern
Machine learning uses algorithms to identify expected and IT network structures and specifically “cloud” technology
unexpected patterns and can consider a greater number and allow upload of data from multiple sites onto central
complexity of variables than traditional methods of predic- “cloud servers” and subsequent access and review from
tive analysis. ML techniques are increasingly applied to anywhere within network [13, 14]. Mobile online access to
large amounts of imaging and other discrete data in health- the central electronic health record including imaging and
17  Integration of CT Data into Clinical Workflows: Role of Modern IT Infrastructure Including Cloud Technology 199

Fig. 17.5 Schematic Public and Commercial Clinical Cloud Services

state-of-the-art cloud services
in context of reading and • Anonymized and protected medical data
reporting • Long term archiving
• Retrospective and prospective data analytics
• Research exchange
• Ontologies

Authorized Users

Mobile
App

Private Clinical Cloud

(Protected) Firewall Client
application

Data Center

• Electronic Health Record

Physician Clients • Image data Physician Clients
• Performance data

non-imaging data from tablet computers and smartphones dependent. Increasingly, semiautomated software systems
allows review without limitations to place and time and pre-analyze the data, providing the evaluating physician with
enables surveillance and online communication within a a structured template for final analysis and documentation/
network of specialists [33]. reporting [44, 45].
The process of CT image acquisition and analysis in the In most clinical centers, selection of patients suitable for
context of TAVR has evolved and is increasingly standard- the TAVR procedure involves an interdisciplinary team of
ized [34–37]. After initial general review including non-­ physicians, including cardiologists, cardiothoracic surgeons,
cardiac structures, the cardiovascular analysis is focused on and imaging specialists/radiologists. After an initial clinical
the access site/route and aortic root deployment zone. The appointment, subsequent testing including imaging is per-
peripheral access arteries (iliac and axillary arteries) are formed. The imaging specialist analyzes the data and pre-
evaluated for luminal size, calcified and non-calcified pares a report including saved key images and presentation
­atherosclerotic plaque of the vessel wall, and vessel tortuos- states. The interventional cardiologist and surgeon review
ity [38–43]. Because TAVR stent/valve prostheses are avail- the accumulated data, which is used for decision-making and
able in limited size ranges, precise measurement of the aortic planning of the procedure. All these single steps do not nec-
annulus is critical. Specifically, the annulus is measured as essarily happen at the same place and time, they may be per-
the plane at the lowest insertion point of the aortic valve leaf- formed independently at different places in the hospital at
lets and minimal/maximal diameter, circumference, and area different time points. Therefore, organization of the huge
are described. In addition, the spatial relation of the coronary amount of accumulated data in a centralized archive and cus-
ostia to the annulus and the angiographic plane orthogonal to tomized reading software models optimized for the work-
the valve plane is defined to ensure optimal positioning/ flow of individual groups of clinicians support the clinical
alignment of the prosthesis with the vessel axis. The analysis pathway.
process is best organized into a multistep workflow, which The examples of TAVR and AAS demonstrate the impact
involves manual or semiautomated aortic root segmentation, of an integrated EHR including imaging data on clinical
centerline reconstruction of the aorta and iliac arteries, and workflows in acute and elective clinical scenarios involving
additional 4-D functional analysis of ventricular and valvular CT imaging. Other clinical examples including those involv-
function. While manual analysis with standard, basic 3-D ing other imaging modalities, e.g., echocardiography or
software is possible, it is time-intensive and operator-­ MRI, can be readily identified [46, 47].
200 P. Schoenhagen and M. Zimmermann

Conclusion 11. Flohr TG, CH MC, Bruder H, Petersilka M, Gruber K, Suss C,

Grasruck M, Stierstorfer K, Krauss B, Raupach R, Primak AN,
Kuttner A, Achenbach S, Becker C, Kopp A, Ohnesorge BM. First
Following hard- and software development to optimize performance evaluation of a dual-source CT (DSCT) system. Eur
image acquisition and analysis of individual scanner/work- Radiol. 2006;16:256–68.
station systems, recent and future developments increasingly 12. Achenbach S, Ropers D, Kuettner A, Flohr T, Ohnesorge B, Bruder
H, Theessen H, Karakaya M, Daniel WG, Bautz W, Kalender
create networks of systems integrated into the EHR. These WA, Anders K.  Contrast-enhanced coronary artery visualization
“cloud” systems are transforming clinical workflows, exem- by dual-source computed tomography  – initial experience. Eur J
plified by the examples of acute aortic syndromes (AAS) and Radiol. 2006;57:331–5.
transcatheter aortic valve replacement (TAVR), where CT 13. Schoenhagen P, Zimmermann M, Falkner J. Advanced 3-D analy-
sis, client-server systems, and cloud computing-Integration of car-
imaging has a central role. However, experience is limited, diovascular imaging data into clinical workflows of transcatheter
and further evaluation of the appropriate infrastructure aortic valve replacement. Cardiovasc Diagn Ther. 2013;3:80–92.
including requirement for reliable patient identification 14. Schoenhagen P, Falkner J, Piraino D.  Transcatheter aortic valve
between provider organizations and data safety is critical. repair, imaging, and electronic imaging health record. Curr Cardiol
Rep. 2013;15:319.
Further the potential clinical impact needs to be evaluated in 15. Liang M, Tang W, Xu DM, Jirapatnakul AC, Reeves AP, Henschke
clinical trials. The widespread use of CT and its standardized CI, Yankelevitz D.  Low-dose CT screening for lung cancer:
data structure makes CT a perfect example, but other imag- computer-aided detection of missed lung cancers. Radiology.
ing modalities including echocardiography and MRI are also 2016;281:279–88.
16. Waljee AK, Higgins PDR. Machine learning in medicine: a primer
evaluated extensively. for physicians. Am J Gastroenterol. 2010;105:1224–6.
17.
Deo RC.  Machine learning in medicine. Circulation.
2015;132:1920–30.
18. Dietterich TG. Ensemble methods in machine learning. Lect Notes
Comput Sci. 2000;1857:1–15.
References 19. Waljee AK, Joyce JC, Wang S, Saxena A, Hart M, Zhu J, Higgins
PD. Algorithms outperform metabolite tests in predicting response
1. Ambrose J, Hounsfield G. Computerized transverse axial tomogra- of patients with inflammatory bowel disease to thiopurines. Clin
phy. Br J Radiol. 1973;46:148–9. Gastroenterol Hepatol. 2010;8:143–50.
2. Hounsfield GN.  Computed medical imaging. Science. 20. Singal AG, Mukherjee A, Elmunzer BJ, Higgins PD, Lok AS, Zhu
1980;210:22–8. J, Marrero JA, Waljee AK. Machine learning algorithms outperform
3. Mettler FA Jr, Bhargavan M, Faulkner K, Gilley DB, Gray JE, conventional regression models in identifying risk factors for hepa-
Ibbott GS, Lipoti JA, Mahesh M, McCrohan JL, Stabin MG, tocellular carcinoma in patients with cirrhosis. Am J Gastroenterol.
Thomadsen BR, Yoshizumi TT.  Radiologic and nuclear medicine 2013;108:1723–30.
studies in the United States and worldwide: frequency, radiation 21. Motwani M, Dey D, Berman DS, Germano G, Achenbach S,

dose, and comparison with other radiation sources  – 1950-2007. Al-Mallah M. Machine learning for prediction of all-cause mortal-
Radiology. 2009;253:520–31. ity in patient with suspected coronary artery disease: a 5-year multi-
4. Brenner DJ, Hall EJ. Computed tomography – an increasing source center prospective registry analysis. Eur Heart J. 2017;38(7):500–7.
of radiation exposure. N Engl J Med. 2007;357:2277–84. 22. Drowning in big data? Reducing information technology complexi-
5. Kalendar WA, Seissler W, Klotz E, Vock P. Spiral volumetric CT ties and costs for healthcare organizations. http://www.emc.com/
with single-breath-hold technique, continuous transport, and con- collateral/analyst-reports/frost-sullivan-reducing-information-tech-
tinuous scanner rotation. Radiology. 1990;176:181–3. nology-complexities-ar.pdf.
6. Klingenbeck-Regn K, Schaller S, Flohr T, Ohnesorge B, Kopp A, 23. Boyd DR, Dunea MM, Flashner BA. The Illinois plan for a state-
Baum U. Subsecond multi-slice computed tomography: basics and wide system of trauma centers. J Trauma. 1973;13:24–31.
applications. Eur J Radiol. 1999;31:110–24. 24. Cowley RA, Hudson F, Scanlan E, Gill W, Lally RJ, Long W, et al.
7. Nieman K, Cademartiri F, Lemos PA, Raaijmakers R, Pattynama An economical and proved helicopter program for transporting the
PMT, de Feyter PJ.  Reliable noninvasive coronary angiography emergency critically ill and injured patient in Maryland. J Trauma.
with fast submillimeter multislice spiral computed tomography. 1973;13:1029–38.
Circulation. 2002;106:2051–4. 25. Harris KM, Strauss CE, Duval S, Unger BT, Kroshus TJ, Inampudi
8. Chao SP, Law WY, Kuo CJ, Hung HF, Cheng JJ, Lo HM, Shyu S, et  al. Multidisciplinary standardized care for acute aortic dis-
KG.  The diagnostic accuracy of 256-row computed tomographic section: design and initial outcomes of a regional care model. Circ
angiography compared with invasive coronary angiography in Cardiovasc Qual Outcomes. 2010;3:424–30.
patients with suspected coronary artery disease. Eur Heart J. 26. Henry TD, Sharkey SW, Burke MN, Chavez IJ, Graham KJ, Henry
2010;31:1916–23. CR, et al. A regional system to provide timely access to percutane-
9. Einstein AJ, Elliston CD, Arai AE, Chen MY, Mather R, Pearson ous coronary intervention for ST-elevation myocardial infarction.
GD, Delapaz RL, Nickoloff E, Dutta A, Brenner DJ.  Radiation Circulation. 2007;116:721–8.
dose from single-heartbeat coronary CT angiography per- 27. Jollis JG, Roettig ML, Aluko AO, Anstrom KJ, Applegate RJ,
formed with a 320-detector row volume scanner. Radiology. Babb JD, et  al., Reperfusion of Acute Myocardial Infarction in
2010;254:698–706. North Carolina Emergency Departments (RACE) Investigators.
10. George RT, Arbab-Zadeh A, Miller JM, Vavere AL, Bengel FM, Implementation of a statewide system for coronary reperfu-
Lardo AC, Lima JA. Computed tomography myocardial perfusion sion for ST-segment elevation myocardial infarction. JAMA.
imaging with 320-row detector computed tomography accurately 2007;298:2371–80.
detects myocardial ischemia in patients with obstructive coronary 28. Aggarwal B, Raymond CE, Randhawa MS, Roselli E, Jacob

artery disease. Circ Cardiovasc Imaging. 2012;5:333–40. J, Eagleton M, Kralovic DM, Kormos K, Holloway D, Menon
17  Integration of CT Data into Clinical Workflows: Role of Modern IT Infrastructure Including Cloud Technology 201

V.  Transfer metrics in patients with suspected acute aortic syn- angiographic planar imaging and 3-dimensional multidetector row
drome. Circ Cardiovasc Qual Outcomes. 2014;7:780–2. computed tomography. JACC Cardiovasc Interv. 2010;3:105–13.
29. Raymond CE, Aggarwal B, Schoenhagen P, Kralovic DM, Kormos 39. Gurvitch R, Wood DA, Leipsic J, et  al. Multislice computed

K, Holloway D, Menon V. Prevalence and factors associated with tomography for prediction of optimal angiographic deployment
false positive suspicion of acute aortic syndrome: experience in a projections during transcatheter aortic valve implantation. JACC
patient population transferred to a specialized aortic treatment cen- Cardiovasc Interv. 2010;3:1157–65.
ter. Cardiovasc Diagn Ther. 2013;3:196–204. 40. da Gama RV, Vouga L, Markowitz A, et  al. Vascular access in
30. Aggarwal B, Raymond C, Jacob J, Kralovic D, Kormos K,
transcatheter aortic valve implantation. Int J Cardiovasc Imaging.
Holloway D, Menon V.  Transfer of patients with suspected acute 2011;27:1235–43.
aortic syndrome. Am J Cardiol. 2013;112:430–5. 41. Petronio AS, De Carlo M, Bedogni F, et al. Safety and efficacy of
31. Schoenhagen P, Mehta N. Big data, smart computer systems, and the subclavian approach for transcatheter aortic valve implanta-
doctor-patient relationship. Eur Heart J. 2017;38(7):508–10. tion with the CoreValve revalving system. Circ Cardiovasc Interv.
32. Cowie MR, Chronaki CE, Vardas P.  E-health innovation: time
2010;3:359–66.
for engagement with the cardiology community. Eur Heart J. 42. Svensson LG, Dewey T, Kapadia S, et  al. United States feasibil-
2013;34:1864–8. ity study of transcatheter insertion of a stented aortic valve by the
33. Matar R, Renapurkar R, Obuchowski N, Menon V, Piraino D, left ventricular apex. Ann Thorac Surg. 2008;86:46–54; discussion
Schoenhagen P. Utility of hand-held devices in diagnosis and triage 54–5.
of cardiovascular emergencies. Observations during implementa- 43. Walther T, Simon P, Dewey T, et  al. Transapical minimally inva-
tion of a PACS-based system in an acute aortic syndrome (AAS) sive aortic valve implantation: multicenter experience. Circulation.
network. J Cardiovasc Comput Tomogr. 2015;9:524–33. 2007;116:I240–5.
34. Leon MB, Piazza N, Nikolsky E, et al. Standardized endpoint defi- 44. Schoenhagen P, Tuzcu EM, Kapadia SR, Desai MY, Svensson
nitions for transcatheter aortic valve implantation clinical trials: a LG. Three-dimensional imaging of the aortic valve and aortic root
consensus report from the valve academic research consortium. Eur with computed tomography: new standards in an era of transcath-
Heart J. 2011;32:205–17. eter valve repair/implantation. Eur Heart J. 2009;30:2079–86.
35. Vahanian A, Alfieri OR, Al-Attar N, et  al. Transcatheter valve 45. Schoenhagen P, Numburi U, Halliburton SS, Aulbach P, von

implantation for patients with aortic stenosis: a position state- Roden M, Desai MY, Rodriguez LL, Kapadia SR, Tuzcu EM,
ment from the European Association of Cardio-Thoracic Surgery Lytle BW. Three-dimensional imaging in the context of minimally
(EACTS) and the European Society of Cardiology (ESC), in invasive and transcatheter cardiovascular interventions using multi-­
collaboration with the European Association of Percutaneous detector computed tomography: from pre-operative planning to
Cardiovascular Interventions (EAPCI). Eur J Cardiothorac Surg. intra-operative guidance. Eur Heart J. 2010;31:2727–40.
2008;34:1–8. 46.
Cavalcante JL, Lalude OO, Schoenhagen P, Lerakis
36. Achenbach S, Delgado V, Hausleiter J, et al. SCCT expert consensus S.  Cardiovascular magnetic resonance imaging for structural and
document on computed tomography imaging before transcatheter valvular heart disease interventions. JACC Cardiovasc Interv.
aortic valve implantation (TAVI)/transcatheter aortic valve replace- 2016;9(5):399–425.
ment (TAVR). J Cardiovasc Comput Tomogr. 2012;6:366–80. 47. Natarajan N, Patel P, Bartel T, Kapadia S, Navia J, Stewart W,
37. Schoenhagen P, Hausleiter J, Achenbach S, et al. Computed tomog- Tuzcu EM, Schoenhagen P.  Peri-procedural imaging for trans-
raphy in the evaluation for transcatheter aortic valve implantation catheter mitral valve replacement. Cardiovasc Diagn Ther.
(TAVI). Cardiovasc Diagn Ther. 2011;1:44–56. 2016;6(2):144–59.
38. Kurra V, Kapadia SR, Tuzcu EM, et al. Pre-procedural imaging of
aortic root orientation and dimensions comparison between X-ray
 Thoughts on Coding
and Reimbursement 18
Adefolakemi Babatunde and Pamela K. Woodard

Accurate Current Procedural Terminology (CPT) [1] and CPT Codes

International Classification of Diseases (ICD) coding [2]
along with documentation within the report are critical parts In late 2009 four cardiac CPT codes were released by the
of the reporting and billing processes. Errors in coding ulti- American Medical Association (AMA) as category I codes
mately lead to denial of claims from Medicaid and Medicare to be used for all cardiac CT billing purposes. CPT codes are
and insurance payers. CPT codes provide third-party payers standardized codes published by the AMA for medical pro-
with information about the procedure that was performed, cedures [1]. There are four CPT codes for cardiac CT:
while ICD codes provide them with the indication for the
procedure. • 75571: computed tomography, heart, without contrast
In reporting cardiac computed tomography (CT) exami- material, with quantitative evaluation of coronary
nations, as with any imaging study, the report should clearly calcium
document the examination performed and the volume of • 75572: computed tomography, heart, with contrast mate-
contrast, and the title of the study performed in the report rial, for evaluation of cardiac structure and morphology
should match the CPT code. In addition, information on any (this includes 3D image post-processing, assessment of
3D processing that was performed on a dedicated worksta- cardiac function, and evaluation of venous structures, if
tion (apart from standard 3D reconstructions at the scanner) performed)
should be reported. Documenting processing performed is • 75573: computed tomography, heart, with contrast mate-
important in reporting cardiac CT scans as the processing rial, for evaluation of cardiac structure and morphology in
component of the examination is included within the CPT the setting of congenital heart disease (this includes 3D
code description. It is important to note that no additional 3D image post-processing, assessment of LV cardiac func-
processing code should be added to the cardiac CT codes. tion, RV structure and function and evaluation of venous
Other information important to include is the indication for structures, if performed)
the examination, for the purposes of ICD-10 coding. • 75574: computed tomographic angiography, heart, coro-
It is also important to note whether or not the insurance nary arteries, and bypass grafts (when present), with con-
provider requires precertification or preauthorization for the trast material, including 3D image post-processing
study. For cardiac CT studies, most insurance providers do (including evaluation of cardiac structure and morphol-
require preauthorization. Ensuring that the referring physi- ogy, assessment of cardiac function, and evaluation of
cian’s office has obtained insurance company preauthoriza- venous structures, if performed) [1]
tion prior to the examination and not afterwards significantly
increases the likelihood of reimbursement for the These codes are used with ICD-9 codes for indications for
procedure. scan use.

A. Babatunde Indications for Calcium Scoring

Cardiovascular Division, Department of Medicine, Washington
University School of Medicine, St. Louis, MO, USA The coronary artery calcium (CAC) scan is a non-contrast
P. K. Woodard (*) prospectively gated cardiac CT scan used to define coronary
Mallinckrodt Institute of Radiology, Washington University School calcium in the epicardial coronary arteries. Its indications are
of Medicine, St. Louis, MO, USA
coronary artery disease screening and further patient risk
e-mail: [email protected]

© Humana Press 2019 203

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_18
204 A. Babatunde and P. K. Woodard

stratification in asymptomatic patients of intermediate risk perfusion examination and the adenosine and/or regadeno-
[3]. It is rarely if ever appropriate for asymptomatic patients son stress agent separately using a J code. The dose of stress
with low global risk for CAD and for symptomatic patients. agent, along with the dose of iodinated contrast agent, should
The code 75571 should be billed alone and should not be be reported.
billed with either codes 75572 or 75574.
The first coronary calcium score was developed by Arthur  cute Chest Pain
A
Agatston and is defined as the product of calcified plaque Because of its high negative predictive value, use of coronary
area and weighted density score given to the highest attenua- CT angiography (CTA) (CPT code 75574) has shown to be
tion value (HU) in the area of calcium [4]. The threshold for useful in the emergency department. Several studies have
selection of calcium for Agatston scoring is any coronary supported the early use of cardiac CT in the emergency
calcium that is greater than 130 Hounsfield units (HU) [4]. department in patients that present with acute chest pain con-
Calcium score ranges from 0 for absence of calcium, cerning for acute coronary syndrome [7–9]. The Rule Out
1–10 for minimal plaque, 11–100 for mild plaque, 101–400 Myocardial Infarction using Computer-Assisted Tomography
for moderate plaque, and >400 for severe plaque [5]. The (ROMICAT) trial was an observational cohort study of
calcium volume score, defined as volume of calcium greater patients with chest pain that demonstrated low-to-­
than 130 HU in the epicardial coronary system, may also be intermediate risk patients with a negative CCTA were
reported and is thought to be more reproducible than the unlikely to be diagnosed with ACS [9]. These findings
Agatston score and thus a better indicator to be followed in resulted in a follow-­up trial, ROMICAT II, which random-
patients to determine disease progression [6]. ized patients who presented with symptoms of ACS to the
CAC score is currently a class of recommendation (COR) emergency department to either early coronary CTA or stan-
IIb with the principal indication for CAC scoring further risk dard of care. This study further supported the use of early
stratification of individuals that are at intermediate risk after coronary CTA in chest pain evaluation. Early coronary CTA
formal risk assessment [2]. led to decreased length of stay with no missed ACS events
The CAC scan (CPT code 75571), as a screening tool, [7]. Also supporting the use of coronary CTA in the emer-
initially was not covered by either the Center for Medicaid gency department for early acute chest pain triage in low-to-­
and Medicare Services (CMS) or private insurance compa- intermediate risk patients is the ACRIN 4005 trial (CT
nies, and most patients who underwent the examination paid Angiography for Safe Discharge of Patients with Possible
out of pocket. Currently some CMS carriers and insurance Acute Coronary Syndromes) [8]. This study was a similar
companies cover the cost of this examination. study of patients presenting to the emergency department
with possible ACS but with the primary end point of safety.
In ACRIN 4005, there was a higher rate of detection of CAD
Indications for Coronary CTA than in ROMICAT II; however, the high negative predictive
value of CCTA still allowed patients to be safely discharged
Indications for coronary CT angiography include both acute from the ED [8].
and chronic angina as well as newly diagnosed heart failure
in order to determine whether the etiology is ischemic or  hronic Stable Chest Pain
C
nonischemic. It is important to note that coronary CT angi- While there is less evidence for use of coronary CT angiog-
ography is not a screening examination. The CPT code raphy in patients with chronic stable angina, there are studies
75574, at present, is used whether or not the study is per- that suggest its utility [10].
formed alone as a dedicated coronary CT angiogram, or with The Prospective Multicenter Imaging Study for Evaluation
a stress agent such as adenosine or regadenoson for perfu- of Chest Pain (PROMISE) trial evaluated ambulatory patients
sion, and whether or not the CT data is sent for fractional with stable chest pain using coronary CTA compared to
flow reserve (FFR) calculation. The single code of 75574 functional testing. Patients who received coronary CTA had
should also be billed if the study is performed with calcium a better correlation with cardiac catheterization results.
scoring (the code 75571 should not also be billed). The code Coronary CTA again demonstrated excellent negative pre-
75574 should also be billed alone in the performance of a dictive value and, unlike functional testing, led to fewer inva-
“triple rule-out” for coronary CTA, aortic dissection, and sive coronary angiograms (ICAs) that showed no obstructive
pulmonary embolism assessment, although the entirety of CAD. This led to the finding of less radiation exposure with
the chest is covered. If billing for coronary CTA rest/stress coronary CTA, if the comparative functional testing included
perfusion, the 75574 code should be billed only once (not radiation because there would be an elimination of many
twice  – once for coronary assessment/rest and once for unnecessary cardiac catheterizations. Coronary CTA has the
stress). It is possible, depending upon the insurance com- advantage over other functional testing, in that it provides
pany, to bill for the stress supervision component of a CT anatomy. Although coronary CTA does not change clinical
18  Thoughts on Coding and Reimbursement 205

outcomes, it appears to reduce the number of unnecessary with age. In the United States, an estimated 5% of patients
invasive studies [10]. over the age of 65 are affected by atrial fibrillation [13].
In addition, investigators in the Randomized Evaluation One important mechanism described in the initiation and
of patients with Stable angina Comparing Utilization of non- maintenance of atrial fibrillation is the rapidly discharging
invasive Examinations (RESCUE) trial sought to use coro- triggers that arise in the smooth muscle of the pulmonary
nary CTA to specifically determine which patients had left veins. Electrophysiologists try to isolate the pulmonary
main disease and which did not, directing patients without veins to prevent propagation of the rapid discharge from the
left main disease to optimal medical therapy (OMT) alone musculature lining the pulmonary veins to the endocardium
and those with left main disease to OMT plus ICA and inter- [14]. An indication for cardiac CT is the mapping of pul-
vention [11]. This trial design was based on the results of the monary vein anatomy prior to pulmonary vein isolation
Clinical Outcomes Utilizing Revascularization and during atrial fibrillation RF ablation as it can be time con-
Aggressive Drug Evaluation (COURAGE) trial which suming when done with traditional angiography. In addi-
showed that in stable coronary artery disease percutaneous tion to mapping, cardiac CT can be used to assess for the
coronary intervention does not reduce the risk of death, MI, presence of pulmonary stenosis as a complication of RF
or other major cardiovascular events when compared to ablation [15].
OMT [12]. Whether CCTA will ultimately be used to triage
patients to ICA and intervention, or OMT alone, will depend Assessment of Prosthetic Valves
upon the results of larger CCTA trials designed in the man- The two main types of prosthetic heart valves (PHVs) are
ner of the COURAGE and RESCUE trials. biologic and mechanical. Biological valves have the advan-
tage of not requiring anticoagulation but over time deterio-
Other Indications rate. Multiple imaging modalities exist to monitor the
Coronary CTA is appropriate for assessing newly diagnosed integrity of PHVs. ACC/AHA guidelines recommend TTE
systolic heart failure in determining whether its etiology is for the initial evaluation of bioprosthetic valve hemodynam-
ischemic or nonischemic and may be appropriate in newly ics and then an annual TTE after the first 10 years, even in
diagnosed diastolic heart failure and syncope with interme- the absence of a change in clinical status [16].
diate for high global CAD risk. Coronary CTA can be used Mechanical PHV dysfunction is rare but should be sus-
in patients with arrhythmia, especially in diagnosing the eti- pected if there is a change in a patient’s clinical status. The
ology of sustained ventricular tachycardia (VT), ventricular estimated prevalence of mechanical valve dysfunction is
fibrillation, exercised induced VT, or nonsustained VT, and 0.01%–6% [17–20]. Because there are limitations in ana-
prior to initiation of anti-arrhythmia therapy in high global tomical visualization with TTE, cardiac CT provides an
CAD risk patients [3]. attractive alternative as it is quick and can allow for multiple
Coronary CTA can be used as follow-up imaging proce- cardiac phase reconstruction to assess for valve leaflet mobil-
dure to further assess a prior abnormal test or uncertain prior ity. Mechanical valves have reference guides that describe
results. This might include abnormal imaging studies such as opening and closing angles and the interpreting physician
an abnormal echocardiogram or nuclear medicine stress test can consult these to determine if there is a change in the
but may also be used to further assess abnormal rest ECG opening or closing leaflet angle. Moreover, contrast-­
findings in patients with low, intermediate, or high global enhanced cardiac CT can help to diagnose pannus formation,
CAD risk. Lastly, coronary CTA may be appropriate for thrombus, paravalvular leak, endocarditis, or mycotic aneu-
assessing symptomatic patients post revascularization. rysm [21–23] as well as provide information about adjacent
Coronary CTA is rarely appropriate for follow-up testing in structures including coronary arteries, distance from the ster-
asymptomatic patients [3]. num, and involvement of the aorta, all of which may be
important for surgical planning [24].

Indications for Anatomic Assessment  ongenital Heart Disease

C
CT can play a role in the assessment of congenital heart
The code 75572 should be used for routine anatomic assess- disease. When used for this purpose, cardiac CT has its own
ment, such as in pulmonary vein anatomy assessment for CPT code, 75563. This study can be performed with pro-
radiofrequency (RF) ablation, whereas 75573 should be used spective gating to define anatomy or retrospective gating in
in patients being assessed for congenital heart disease. order to diagnose functionality (ventricular function and
valve mobility), realizing retrospective gating techniques
 ulmonary Vein Assessment Pre-RF Ablation
P will provide the patient with greater radiation dose expo-
Atrial fibrillation is the most common arrhythmia and is sure. Regardless of the gating technique, the same CPT
associated with significant morbidity. Prevalence increases code is used.
206 A. Babatunde and P. K. Woodard

Some centers provide 3D printing services as part of pre- 3. Wolk MJ, Bailey SR, Doherty JU, Douglas PS, Hendel RC,
Kramer CM, et  al. ACCF/AHA/ASE/ASNC/HFSA/HRS/SCAI/
surgical planning for congenital heart patients. At present SCCT/SCMR/STS 2013 multimodality appropriate use criteria
there is no code to cover the added costs associated with 3D for the detection and risk assessment of stable ischemic heart dis-
printing. ease: a report of the American College of Cardiology Foundation
Appropriate Use Criteria Task Force, American Heart Association,
American Society of Echocardiography, American Society of
 ranscatheter Aortic Valve Replacement:
T Nuclear Cardiology, Heart Failure Society of America, Heart
A New Indication Rhythm Society, Society for Cardiovascular Angiography and
Aortic stenosis leads to progressive left ventricular outflow Interventions, Society of Cardiovascular Computed Tomography,
obstruction. Symptoms include angina, dyspnea, syncope, Society for Cardiovascular Magnetic Resonance, and Society of
Thoracic Surgeons. J Am Coll Cardiol. 2014;63(4):380–406.
and heart failure. Transcatheter aortic valve replacement 4. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M
(TAVR) is an alternative to surgical aortic valve replacement Jr, Detrano R. Quantification of coronary artery calcium using ultra-
for patients with severe aortic stenosis and high surgical risk fast computed tomography. J Am Coll Cardiol. 1990;15(4):827–32.
because of age and comorbidities. Significant pre-procedural 5. Hecht HS.  Coronary artery calcium scanning: past, present, and
future. J Am Coll Cardiol Img. 2015;8(5):579–96.
planning must occur, and computed tomographic angiogra- 6. Kopp AF, Ohnesorge B, Becker C, Schroder S, Heuschmid M,
phy is becoming a more central piece in planning. Information Kuttner A, et  al. Reproducibility and accuracy of coronary cal-
is needed regarding aortic root anatomy, aortic valve annulus cium measurements with multi-detector row versus electron-beam
(measured in systole), coronary arteries, deployment angle, CT. Radiology. 2002;225(1):113–9.
7. Hoffmann U, Truong QA, Schoenfeld DA, Chou ET, Woodard PK,
abdominal aorta and iliofemoral artery diameters, tortuosity, Nagurney JT, et al. Coronary CT angiography versus standard eval-
and extent of calcification [25, 26]. Patient eligibility for uation in acute chest pain. N Engl J Med. 2012;367(4):299–308.
TAVR depends upon annular size, location of coronary arter- 8. Litt HI, Gatsonis C, Snyder B, Singh H, Miller CD, Entrikin DW,
ies, and peripheral access. Precise and accurate deployment et al. CT angiography for safe discharge of patients with possible
acute coronary syndromes. N Engl J Med. 2012;366(15):1393–403.
of the valve is crucial for safety [25]. If the valve is too low, 9. Hoffmann U, Bamberg F, Chae CU, Nichols JH, Rogers IS,
there is increased risk of heart block, paravalvular regurgita- Seneviratne SK, et al. Coronary computed tomography angiogra-
tion, and mitral valve regurgitation. If deployment is too phy for early triage of patients with acute chest pain: the ROMICAT
high, risks of valve embolization, aortic root injury, and para- (Rule Out Myocardial Infarction using Computer Assisted
Tomography) trial. J Am Coll Cardiol. 2009;53(18):1642–50.
valvular regurgitation are increased. CT when used with 2D 10. Douglas PS, Hoffmann U, Patel MR, Mark DB, Al-Khalidi

echo reduces paravalvular regurgitation, as more information HR, Cavanaugh B, et  al. Outcomes of anatomical versus func-
can be gathered regarding aortic annulus size, geometry, and tional testing for coronary artery disease. N Engl J Med.
anatomy. Specific protocols have been designed for TAVR 2015;372(14):1291–300.
11. Stillman AE, Gatsonis C, Lima JA, Black WC, Cormack J, Gareen
assessment. Heart rate is marginally less important in data I, et  al. Rationale and design of the Randomized Evaluation of
acquisition for TAVR and beta blockers can be detrimental in patients with Stable angina Comparing Utilization of noninvasive
severe AS given the risk of low cardiac output in these Examinations (RESCUE) trial. Am Heart J. 2016;179:19–28.
patients with a fixed left ventricular outflow tract obstruction 12. Boden WE, O’Rourke RA, Teo KK, Hartigan PM, Maron DJ,
Kostuk WJ, et al. Optimal medical therapy with or without PCI for
[26]. The cardiac CT CPT code usually used for this purpose stable coronary disease. N Engl J Med. 2007;356(15):1503–16.
is 75572 rather than 75574 since the goal of this study is not 13. Sellers MB, Newby LK. Atrial fibrillation, anticoagulation, fall risk,
coronary atherosclerosis assessment but rather anatomic and outcomes in elderly patients. Am Heart J. 2011;161(2):241–6.
evaluation. An additional CPT code of 74174 is required 14. Friedman PA, Asirvatham SJ, Grice S, Glikson M, Munger TM, Rea
RF, et al. Noncontact mapping to guide ablation of right ventricular
when this study is performed along with CTA of the abdo- outflow tract tachycardia. J Am Coll Cardiol. 2002;39(11):1808–12.
men and pelvis for aorta and iliofemoral artery assessment. 15. Cronin P, Sneider MB, Kazerooni EA, Kelly AM, Scharf C, Oral
H, et al. MDCT of the left atrium and pulmonary veins in planning
radiofrequency ablation for atrial fibrillation: a how-to guide. AJR
Summary Am J Roentgenol. 2004;183(3):767–78.
16. Nishimura RA, Otto C. 2014 ACC/AHA valve guidelines:

earlier intervention for chronic mitral regurgitation. Heart.
In short, following the above rules when coding for cardiac 2014;100(12):905–7.
CT examinations will increase accuracy in coding and 17. Khan S. Long-term outcomes with mechanical and tissue valves. J
increase likelihood of reimbursement. Heart Valve Dis. 2002;11(Suppl 1):S8–14.
18. Hammermeister K, Sethi GK, Henderson WG, Grover FL, Oprian
C, Rahimtoola SH. Outcomes 15 years after valve replacement with
a mechanical versus a bioprosthetic valve: final report of the veter-
References ans affairs randomized trial. J Am Coll Cardiol. 2000;36(4):1152–8.
19. Grunkemeier GL, Li HH, Naftel DC, Starr A, Rahimtoola

1. American Medical Association Current Procedural Technology SH. Long-term performance of heart valve prostheses. Curr Probl
Professional Edition; 2013. Cardiol. 2000;25(2):73–154.
2. International Classification of Diseases (ICD), 11th Revision (ICD- 20. Cannegieter SC, Rosendaal FR, Briet E.  Thromboembolic and

11) [Cited 2018 October 30]. Available from: http://www.who.int/ bleeding complications in patients with mechanical heart valve
classifications/icd/en/ prostheses. Circulation. 1994;89(2):635–41.
18  Thoughts on Coding and Reimbursement 207

21. Symersky P, Budde RP, Prokop M, de Mol BA. Multidetector-row 24. Habets J, Mali WP, Budde RP.  Multidetector CT angiography in
computed tomography imaging characteristics of mechanical pros- evaluation of prosthetic heart valve dysfunction. Radiographics
thetic valves. J Heart Valve Dis. 2011;20(2):216–22. Rev Publ Radiological Soc North Am Inc. 2012;32(7):1893–905.
22. Konen E, Goitein O, Feinberg MS, Eshet Y, Raanani E, Rimon U, 25. Litmanovich DE, Ghersin E, Burke DA, Popma J, Shahrzad M,
et al. The role of ECG-gated MDCT in the evaluation of aortic and Bankier AA.  Imaging in Transcatheter Aortic Valve Replacement
mitral mechanical valves: initial experience. AJR Am J Roentgenol. (TAVR): role of the radiologist. Insights Imaging. 2014;5(1):123–45.
2008;191(1):26–31. 26. Blanke P, Schoepf UJ, Leipsic JA. CT in transcatheter aortic valve
23. Habets J, Symersky P, van Herwerden LA, de Mol BA, Spijkerboer replacement. Radiology. 2013;269(3):650–69.
AM, Mali WP, et  al. Prosthetic heart valve assessment with
multidetector-­row CT: imaging characteristics of 91 valves in 83
patients. Eur Radiol. 2011;21(7):1390–6.
 Part IV
Where We Are: Cardiac CT Fundamentals
 Pathology and Pathophysiology
of Coronary Atherosclerotic Plaques 19
Hiroyoshi Mori, Frank D. Kolodgie, Aloke V. Finn,
and Renu Virmani

The death rate from coronary artery disease has declined in the last century, have made enormous contributions to the
the past few decades through greater understanding of risk better understanding of early lesions, the influence of risk
factors of coronary heart disease as well as through better factors on plaque progression, and the role of plaque rupture
treatment, including the creation of coronary care units. in the occurrence of luminal thrombosis.
However, because of the lack of an animal model of unstable The cellular participants of atherosclerosis include smooth
plaque, our understanding of atherosclerotic plaque mor- muscle cells, endothelial cells, macrophages, T- and
phology comes only from static histology of lesion morphol- B-lymphocytes, red cells, platelets, neutrophils, and basophils.
ogy in patients dying of acute coronary syndromes [1]. The noncellular components include lipid, proteoglycans, col-
Although transgenic models of atherosclerosis have mark- lagen, elastic fibers, calcium, iron (hemosiderin), and blood
edly enhanced our understanding of certain aspects of plaque components, including fibrin and factor VIII.  These various
progression and regression, they have failed thus far to components help to form the various lesions that are recog-
explain the relationship of the coagulation parameters and nized as part of the atherosclerotic process. These include
plaque morphology that precipitate coronary thrombosis [1]. adaptive intimal thickening (AIT), intimal xanthomas, patho-
Until we are able to create a better model or study plaque logical intimal thickening (PIT), fibroatheroma, thin-cap fibro-
morphology prospectively and determine the mechanisms atheroma, plaque rupture, plaque erosion, calcified nodule,
and the anatomic markers of progression, we will make fibrocalcific plaque, healed plaque rupture, and fibrous plaque
progress very slowly. This review is based on the examina- (either from healed plaque erosion or propagated thrombus).
tion of human coronary artery pathology in patients dying a The best known of the classifications is the American
sudden coronary death, in order to ascertain the pathologic Heart Association (AHA) reported by Stary et  al. [2].
lesion morphologies that are linked to plaque progression However, this classification missed two important clinical eti-
and thrombosis, which will be necessary for us to be able to ologies of coronary thrombus that are distinct from rupture.
recognize by invasive or noninvasive means the prospective One is surface erosion, which accounts for 25–30% of throm-
lesions that are likely to produce symptoms. bosis, and the other is calcified nodule, which occurs in <5%
of patients. Furthermore, plaque fissure which is another form
of communication between the lumen and underlying necrotic
Classification of Atherosclerosis core was not included to the AHA classification [3]. Another
important concept missing in the AHA classification was the
Our understanding of atherosclerosis has been enhanced by recognition of precursor lesions to plaque rupture, known as
the development of the various classifications that have come “thin-cap fibroatheroma” or “vulnerable plaque.” Table 19.1
from the insights of scientists like Virchow, who was a pio- shows our modified AHA classification [4]. AHA lesion types
neer in pathology. Systematic studies of lesion development, I–IV were replaced by descriptive terms of adaptive intimal
described by the giants of atherosclerosis including Robert thickening, intimal xanthoma (fatty streak), pathological inti-
Wissler, Herbert Stary, Henry McGill, and Michael Davies in mal thickening, and fibroatheroma. AHA categories V and VI
were discarded because it doesn’t explain three etiologies of
coronary thrombosis (rupture, erosion, and calcified nodule).
H. Mori · F. D. Kolodgie · R. Virmani (*) The concept of healed plaque rupture and healed erosion was
CVPath Institute, Gaithersburg, MD, USA
also added, which can lead to increases in plaque burden,
e-mail: [email protected]
luminal ­narrowing, or even silent or symptomatic chronic
A. V. Finn
total occlusion.
CVPath Institute, University of Maryland, Gaithersburg, MD, USA

© Humana Press 2019 211

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_19
212 H. Mori et al.

Table 19.1  Updated classification of atherosclerotic lesions based on morphology

Type of lesion Subtype of lesion Morphological description
Nonatherosclerotic intimal Intimal thickening Natural accumulation of neointimal smooth muscle cells in the absence
lesions of lipid, macrophage foam cells, and thrombosis
Intimal xanthoma Superficial accumulation of foam cells without a necrotic core, fibrous
cap, or thrombosis
Progressive atherosclerotic Pathological intimal thickening Plaque rich in smooth muscle cells, with hyaluronan and proteoglycan
lesions matrix and focal accumulation of extracellular lipid. Absence of
thrombosis
Fibroatheroma During early necrosis: focal macrophage infiltration into areas of lipid
pools with an overlying fibrous cap. During late necrosis: loss of matrix
and extensive cellular debris with an overlying fibrous cap. With or
without calcification. Absence of thrombosis
Intraplaque hemorrhage or Large necrotic core (size >10% to plaque area) with hemorrhage, and
plaque fissure plaque area shows the presence of angiogenesis. Necrotic core
communicates with the lumen through a fissure. Minimal tear without
obvious thrombus
Thin-cap fibroatheroma A thin, fibrous cap (<65 μm) infiltrated by macrophages and
lymphocytes, with rare or no smooth muscle cells and relatively large
underlying necrotic core (size >10% to plaque area). Intraplaque
hemorrhage and/or fibrin might be present. Absence of thrombosis
Lesions with acute thrombi Plaque rupture Thin-cap fibroatheroma with cap disruption. Thrombosis is present and
may or may not be occlusive. The luminal thrombus communicates with
the underlying necrotic core
Plaque erosion Can occur on pathological intimal thickening or on a fibroatheroma.
Thrombosis is present and may or may not be occlusive. No
communication of the thrombus with the necrotic core.
Calcified nodule Eruptive (shedding) of calcified nodule with an underlying fibrocalcific
plaque with minimal or no necrosis. Thrombosis is usually not occlusive
Healed lesions Healed plaque rupture, erosion, Healed lesion composed of smooth muscle cells, proteoglycans, and type
or calcified nodule III collagen with or without underlying disrupted fibrous cap, necrotic
core, or nodular calcification. Lesions can contain large areas of
calcification with few inflammatory cells and have a small or no necrotic
core. The fibrotic or fibrocalcific collagen-­rich plaque is associated with
significant luminal stenosis. Absence of thrombosis
From Yahagi et al. [4], with permission
An updated version of the modified AHA classification [1, 4], which was based on the original AHA classification published in the mid-­1990s [2]

Adaptive Intimal Thickening (AIT) a lesion of atherosclerosis, as it eventually regresses in

humans. The Pathobiological Determinants of Atherosclerosis
AIT occurs in most arteries once flow is established in utero in Youth (PDAY) study showed that fatty streaks are promi-
or soon after birth, consisting primarily of smooth muscle nent in the third decade in the thoracic aorta but regress later
cells, which are strongly α-actin positive and surrounded by in life [8]. The site of advanced lesions is generally the
a proteoglycan-rich matrix (Fig.  19.1a). Macrophages are abdominal aorta, where fatty streaks are relatively uncom-
rarely detected. These lesions are most prominent at branch mon. Fatty streaks are rich in macrophages [9, 10] (hence the
points and are considered by many to be precursor lesions of term xanthoma) but lack necrotic cores or lipid pools
atherosclerosis [5]. Studies of neonates, adolescents, and (Fig. 19.1b).
young adults have indicated that intimal masses forming
near branch points enlarge with advancing age and might be
precursors to high-risk plaques [6, 7]. Pathological intimal thickening (PIT)

PIT lesion is also rich in smooth muscle cells in a proteogly-

Intimal Xanthomas (Fatty Streak) can matrix (Fig. 19.2a). However, the area close to the media
shows loss of smooth muscle cells and an accumulation of
We have used the microscopic term intimal xanthoma for fat as a lipid pool, which stains positive with oil red O.  A
fatty streak, the corresponding designation for the lesion as significant number of smooth muscle cells close to the media
grossly seen in the aorta. We believe intimal xanthoma is not show accumulation of intracellular fat and may appear
19  Pathology and Pathophysiology of Coronary Atherosclerotic Plaques 213

a b

Fig. 19.1  Adaptive intimal thickening (a) and intimal xanthoma (b). streak) is known to regress. Intimal thickening consists mainly of
Lesions uniformly present in all populations, although intimal xantho- smooth muscle cells in a proteoglycan-rich matrix, while intimal xan-
mas are more prevalent with exposure to a Western diet. Intimal xantho- thomas primarily contain macrophage-derived foam cells,
mas are commonly produced in animal models; however, they usually T-lymphocytes, and varying degrees of smooth muscle cells. (From
do not develop into progressive atherosclerotic lesions. Both lesions Virmani et al. [1], with permission)
occur soon after birth; the intimal xanthoma (otherwise known as a fatty

foamy, best appreciated by transmission electron micros- Fibroatheroma

copy. The smooth muscle cells may appear as ghosts with
surrounding thickened basement membrane, which stains Fibroatheromas are lesions with a necrotic core and a thick
strongly periodic acid-Schiff (PAS) positive [11] (3a). The fibrous cap (Fig. 19.2b). The necrotic core is rich in acellular
surrounding matrix also stains positive for oil red O and debris and varies in size. The fibrous cap covers the necrotic
often shows the presence of monohydrate cholesterol, which core completely, is rich in smooth muscle cells within a col-
may appear as cholesterol crystals (lipid pools). There may lagen/proteoglycan matrix, and may contain macrophages
be macrophage infiltration in the superficial regions of the and lymphocytes. The thickness of the fibrous cap varies; a
plaque, but these macrophages are not in proximity to the thin cap is believed to impart instability and therefore war-
lipid pools. It is believed that smooth muscle cells are under- rants a separate designation of thin-cap fibroatheroma. We
going apoptosis, degeneration, and calcification [12]. We recently subcategorized fibroatheromas into “early” and
believe that pathological intimal thickening is the precursor “late” phase [4]. Early fibroatheroma is associated with a
lesion of the fibroatheroma. decrease in the expression of biglycan, hyaluronan, and
214 H. Mori et al.

a b

Fig. 19.2 Pathological intimal thickening (a) vs. atheroma (b). phocytes may also be present but are usually sparse. The more defini-
Pathological intimal thickening is a poorly defined entity sometimes tive lesion or fibrous cap atheroma classically shows a “true” necrotic
referred to in the literature as an “intermediate lesion.” True necrosis is core (NC), containing cholesterol esters, free cholesterol, phospholip-
not apparent, and there is no evidence of cellular debris; some lipid may ids, and triglycerides. The fibrous cap consists of smooth muscle cells
be present deep in the lesion near the elastic lamina (EL), but it is dis- in a proteoglycan-collagenous matrix, with a variable number of mac-
persed. The fibrous cap overlying the areas of lipid is rich in smooth rophages and lymphocytes. The media underneath the plaque is often
muscle cells and proteoglycans. Some scattered macrophages and lym- thin. (From Virmani et al. [1], with permission)

v­ ersian within the lipid pool together with infiltrating macro- In support of this concept is the frequent finding of fibrin, by
phages, which are undergoing necrosis or apoptosis. Late immunohistochemical methods, within the necrotic core.
fibroatheroma is deficient in extracellular matrix and has
more cholesterol clefts and calcification [13, 14]. How the
necrotic core enlarges and becomes rich in free cholesterol is Thin-Cap Fibroatheroma
poorly understood. Hemorrhage into a plaque, either through
the leaky vasa vasorum or possibly through plaque fissuring, The original description of AHA classification did not men-
plays an important role in the enlargement of the necrotic tion the thin-cap fibroatheroma (TCFA) as a precursor lesion
core. Because the vasa vasorum become leaky, allowing of plaque rupture. Partly based on their morphologic resem-
fibrinogen to seep into the plaque, especially if pressure blance to acute ruptures, we believe that they are the precur-
within them rises, as they have a poorly developed basement sor lesions. The TCFA is defined as a lesion with a necrotic
membrane and often lack pericytes and smooth muscle cells. core and an overlying fibrous cap that is <65 μm and infil-
19  Pathology and Pathophysiology of Coronary Atherosclerotic Plaques 215

a Insignificant b Large eccentric c Large concentric d Healed

Plaque burden necrotic core necrotic core Rupture(s)

Fig. 19.3  Morphologic variants of the thin-cap atheroma. Thin caps s­ uggestive of a prior rupture and multiple areas of thin caps (arrows).
may emerge in fibroatheromas with insignificant plaque burden and (b) The necrotic core (NC) is large, with an area of thinned cap (arrow).
insignificant luminal narrowing, in fibroatheromas with large cores that (c) The necrotic core (NC) is concentric with an extensive area of cap
are eccentric or eccentric, and frequently in plaques with evidence of thinning (arrow). (d) There may be a healed rupture (arrows) with a
prior rupture. (a) In this plaque with insignificant plaque burden, there proteoglycan-­rich smooth muscle cell reparative layer (arrowhead)
is an area of proteoglycan-rich smooth muscle cells (arrowheads)

trated by macrophages (>25 per high-power magnification Table 19.2  Approximate sizes of necrotic core in fibroatheroma, thin-­
[0.03 mm diameter field]) [15] (Fig. 19.3). The thickness cri- cap atheroma, and acute rupture
terion of TCFA was chosen because 95% of fibrous caps Plaque type
adjacent to acute plaque rupture measured <65 μm, with a Fibroatheroma Thin-cap Acute plaque
mean cap thickness of 23 ± 19 μm. We have also compared Dimension atheroma rupture
other morphologic characteristics between plaque rupture Length (mm) 6 mm (range 8 mm (range 9 mm (range
mean/range 1–18 mm) 2–17 mm) 2.5 ± 22)
and TCFA. The necrotic core is smaller than that seen in rup-
Cross-­ 1.2 ± 2.2 1.7 ± 1.1 4.1 ± 5.5
ture (Table 19.2). In addition, the percent area occupied by sectional area,
the necrotic core is greater in rupture than TCFA, although mm2
we were unable to determine any significant differences in % Cross-­ 15 ± 20% 23 ± 17% 34 ± 17%
length (Table  19.2). The cross-­sectional luminal narrowing sectional area
of TCFA is less than that of acute rupture; nearly 80% of From Ref. [45], with permission
TCFAs had luminal narrowing <75% in contrast to rupture
(Fig. 19.4). ological intimal thickening, and a calcified plaque with or
without a necrotic core (Fig. 19.1).

Lesions with Thrombi
Plaque Rupture
From studies carried out in patients dying sudden coronary
death, we have observed three main causes of luminal Plaque rupture is defined as a disruption, discontinuous
thrombi: plaque rupture (the precursor of which is presum- fibrous cap, underlying necrotic core, and superimposed
ably the TCFA), erosion, and calcified nodule (Fig.  19.5). luminal thrombus. Ruptured plaques that result in occlusive
The most frequent cause of thrombosis is plaque rupture luminal thrombi almost invariably have a large necrotic
(over 60%), plaque erosion is the second most frequent core, which occupies =25% of the plaque area in 80% of
(about 35%), whereas calcified nodule is a rare cause of cases. Plaque rupture is the major underlying mechanism of
thrombosis (<5%). These three types of thrombosis occur, luminal thrombosis in sudden death (60%), acute myocar-
respectively, in the setting of a TCFA, fibroatheroma or path- dial infarction (75%), and unstable angina (70%) and is
 216 H. Mori et al.

73 ± 13%*

40 77 ± 16%*

% of plaque ruptures (n=90)

40 77 ± 15%*
35
% of plaque ruptures (n=90)

35
79 ± 13%* 30
30
71 ± 13%*
25 25

20 81 ± 9%* 78 ± 12%*
20
82 ± 9%* 15
15
10 10

5 5

0 0
% <10 % 10-25 % 25-50 >50% area<1 area 1-3 area 3-5 area > 5

Lipid core as % plaque area Cross sectional lipid core area, mm2

45 71 ± 17%* 65 ± 15%*
45
% of thin capatheromas (n=106)

% of thin capatheromas (n=106)

40 40
35 74 ± 15%*
35
30 30
65 ± 17%*
25 25
77 ± 9%*
20 76 ± 14%* 20
78 ± 12%*
15 15 77 ± 13%*

10 10
5 5
0 0
% <10 % 10-25 % 25-50 >50% area<1 area 1-3 area 3-5 area > 5

Fig. 19.4  The distribution frequency of plaque ruptures (a, b) and area. In the case of thin-cap atheromas, the degree of cross-sectional
thin-cap atheromas (c, d) by size of lipid core or lipid core as a percent area luminal narrowing and area of necrotic core is shifted to the left
of plaque area (x-axis). The majority of plaque ruptures occur when (lesser or smaller) as compared to plaque ruptures. (From Ref. [45],
lipid core area forms 25% to 50% of plaque area or 1–3 mm2 lipid core with permission)

rarely present in stable angina [16–18]. The mechanism of ture. Davies has shown that 11% of hearts have plaque fis-
rupture is poorly understood; however, it is generally sure with intra-intimal thrombus, 2% have intimal thrombus
believed that macrophages infiltrate the cap and release as the only mechanism of sudden death, while rupture with
matrix metalloproteinases, which are responsible for the thrombosis was identified in 74%, and only 5% had no acute
breakdown of the collagens [19, 20]. There is also evidence coronary lesion [3]. We have observed vasa vasorum to be
that macrophage myeloperoxidase may be responsible for more prominent in patients dying suddenly during exercise
the disruption of the fibrous cap by producing the prooxi- with plaque rupture, compared to patients dying at rest.
dant species, hypochlorous acid. We have shown that up to Also, hemorrhage into a plaque was more frequent in
15% of macrophages within ruptured fibrous caps are rich in patients dying during exercise than in patients dying at rest
myeloperoxidase. We have observed that the numbers of [22]. The site of rupture was located at the shoulder more
plaque hemorrhages in the coronary arteries of patients frequently in the cases dying at rest (65%) versus dying dur-
dying suddenly with acute plaque rupture are greater than in ing exercise (25%) (P  =  0.02) [22]. Plaque rupture corre-
patients with severe coronary disease dying with other lated with high total cholesterol, low HDL cholesterol, and
forms of thrombosis or without coronary thrombosis [21]. an elevated ratio of total/HDL cholesterol while smoking
These data suggest that hemorrhage, due to plaque fissure or was found to be a predictor of acute thrombosis regardless
rupture of vasa vasorum, may be precursors to plaque rup- of plaque etiology [15].
19  Pathology and Pathophysiology of Coronary Atherosclerotic Plaques 217

Fig. 19.5  Lesions with thrombi. Ruptured plaques are thin-cap fibro- cellular lipid pools, but necrotic cores are uncommon; when present,
atheroma with luminal thrombi (Th). These lesions usually have an the necrotic core does not communicate with the luminal thrombus.
extensive necrotic core containing large numbers of cholesterol crystals Inflammatory infiltrate is usually absent, but if present, it is sparse and
and a thin fibrous cap (<65 μm) infiltrated by foamy macrophages and consists of macrophages and lymphocytes. Calcified nodules are
a paucity of T-lymphocytes. The fibrous cap is thinnest at the site of plaques with luminal thrombi showing calcific nodules protruding into
rupture and consists of a few collagen bundles and rare smooth muscle the lumen through a disrupted thin fibrous cap (FC). There is an absence
cells. The luminal thrombus is in communication with the lipid-rich of endothelium at the site of the thrombus, and inflammatory cells
necrotic core. Erosions occur over lesions rich in smooth muscle cells (macrophages, T lymphocytes) are absent. (From Virmani et  al. [1],
and proteoglycans. Luminal thrombi overlay areas lacking surface with permission)
endothelium. The deep intima of the eroded plaque often shows extra-

Plaque Erosion the pathological intimal thickening or fibrous cap atheroma

(pathological intimal thickening 16%, early fibroatheroma
Plaque erosion is defined as an acute thrombus in direct con- 50%, late fibroatheroma 34%) [24]. Erosion lesions show
tact with intimal plaque rich in smooth muscle cells within a less calcification with more than half of erosions (56%)
proteoglycan matrix and an absence of endothelium [1, 17]. showing no calcification, while 40% show microcalcifica-
There are relatively few macrophages and lymphocytes adja- tion, while fragmented and sheet of calcification are rarely
cent to the thrombus in the majority of plaque erosions. observed [24]. The most frequent location for both erosion
Plaque erosion accounts for 20% of all sudden coronary and rupture is the proximal left anterior descending coronary
deaths and 35% of coronary thrombi in patients dying sud- artery (nearly two-thirds of patients) followed by the right
denly with coronary thrombosis [1, 17]. The mean age of coronary artery and the left circumflex. However, patients
patients dying with plaque erosion is less than that of patients dying with plaque erosion frequently do not have extensive
dying with acute rupture, and there is less severe narrowing disease, unlike patients dying with stable plaque or plaque
at the sites of thrombosis in plaque erosion. Plaque erosion rupture. Over one-half of deaths are seen in patients with
accounts for over 80% of thrombi occurring in women less single-vessel disease and one-quarter with double-vessel dis-
than 50 years of age. The lesions tend to be eccentric and are ease. The etiology of plaque erosion is poorly understood;
infrequently calcified [17]. Plaque erosions tend to embolize however, it is believed that coronary vasospasm may be
more frequently than plaque rupture (71%, vs. 42%, respec- involved.
tively) [23]. The thrombi of eroded plaques often show at
later stage of organization than that of ruptured plaques (88%
vs. 54%, respectively, P < 0.0001). The risk factors for ero- Calcified Nodule
sion are poorly understood; similar to acute rupture, there is
an increased risk among smokers, but there appears to be a The least frequent lesion that results in coronary luminal
lesser association with dyslipidemia, as compared to plaque thrombus is the calcified nodule, a plaque that contains calci-
rupture [15]. The underlying plaque in erosions is generally fied plates at the site of luminal disruption. The incidence of
218 H. Mori et al.

calcified nodules was only 5% [25]. There may or may not be Healed Plaque Ruptures
necrotic core within the plaque, but the characteristic fea-
tures are the breakage of calcified plates, often with a fibrotic Not all plaque ruptures result in symptoms; therefore, there
reaction, an interspersed fibrin, a disrupted surface, and an should be morphologic hallmarks that could be recognized
overlying thrombus. Occasionally, bone formation with within plaques that are representative of a previous site of
osteoblasts and osteoclasts is present [1]. Although the pre- thrombus [26]. Healed ruptures in the coronary vascular bed
cise mechanism is unknown, one hypothesis is that mechani- are readily detected microscopically by the identification of
cal stress might lead to breakage of the calcified sheets. In the breaks in the fibrous cap, which is rich in type I collagen,
between the calcified spicules, fibrin is commonly observed, and an overlying repaired lesion, which is richer in type III
potentially arising from the surrounding break in the capil- collagen. The healed lesion can be more easily recognized by
laries. Eruptive calcified nodules typically happen in eccen- Movat’s stain (healed site identified by the brilliant blue-­
tric lesions where protrusion causes disruption of the green color of the proteoglycan-rich matrix) and confirmed
overlaying cap and endothelium. The lesion is more com- by picrosirius red staining and polarized microscopy. When
monly located in the mid-right coronary artery or left ante- viewed under polarized light, this stain highlights the breaks
rior descending artery, generally in a heavily calcified in the fibrous cap, which is rich in type I collagen, as yellow-­
segment [1]. It is more common in older male individuals red birefringence with an underlying necrotic core (Fig. 19.6).
than women. We have observed these lesions in the carotid The plaque overlying the fibrous cap consists of smooth
arteries as well as the coronaries. muscle cells in a proteoglycan matrix, which is rich in type

a b

c
d

Fig. 19.6  Healed plaque rupture. (a) demonstrates areas of intra-­ orrhagic cores seen in corresponding view in panel a. (d) demonstrates
intimal lipid-rich core with hemorrhage and cholesterol clefts. (b) an image taken with polarized light. The dense collagen (type 1) which
shows a higher magnification of the looser smooth muscle cell forma- forms the fibrous cap is lighter reddish-yellow and is disrupted (arrow),
tion within a collagenous proteoglycan-rich neointima showing a clear with the newer greenish type III collagen on the right and above the
demarcation with the more fibrous regions of the old plaque to the right. rupture site (a and b, Movat’s pentachrome). (From Burke et al. [46],
(c) and (d) demonstrate the layers of collagen by Sirius red staining. In with permission)
(c), note the area of dense dark red collagen surrounding the lipid hem-
19  Pathology and Pathophysiology of Coronary Atherosclerotic Plaques 219

III collagen and has green birefringence under polarized

light (sirius red stained) [22].
We and others believe that healing of a disrupted plaque is
the main stimulus for plaque progression beyond 40% cross-­
sectional area narrowing and is one of the major factors lead-
ing to high-grade coronary stenosis [26]. Incidence of healed
plaque rupture was 16% in arteries with ≦20% diameter ste-
nosis, 19% with 21–50% diameter stenosis, and 73% for
plaque with >50% diameter stenosis [26]. This mechanism
would explain the phasic rather than linear progression of
coronary disease observed in angiograms carried out annu-
ally in patients with chronic ischemic heart disease [26].
However, these are speculations that need to be proven when
we can prospectively recognize the sites of vulnerability and
follow up the lesional morphologies with advanced imaging
techniques. We have also shown that these lesions are the
most heavily calcified as compared to acute rupture or thin-­
cap fibroatheroma, suggesting that healing of a ruptured
plaque also results in greater extent of calcification [27].

Total Occlusions

Among patients dying sudden coronary death without prior

symptoms, total occlusion is a frequent finding at autopsy
(Fig. 19.7). The frequency is as high as 40% when there are no
other sites with acute plaque rupture. Healed myocardial
infarcts are seen in 90% of patients with a total occlusion [1].
The lumen is composed of dense collagen and/or proteogly-
cans with interspersed capillaries, arterioles, smooth muscle
cells, and macrophages [28]. In the case of total occlusions
secondary to thrombi, the proximal and distal ends organize
initially, followed by the middle segment, which may demon-
strate entrapped red cells and fibrin without cellular ingrowth
for long periods. Sites of total occlusion are often associated
Fig. 19.7  Total occlusion. Plaques with total occlusion from prior
with negative remodeling of the vessel, probably related to thrombi contain mostly smooth muscle cells in a collagen-­proteoglycan-­
collagen replacement of the thrombus with eventual cross- rich matrix with capillaries and inflammatory cell infiltrate. This sec-
linking, resulting in artery shrinkage. There is often little cal- tion shows a necrotic core (NC), although this is not always present,
and the lumen is filled with an organized thrombus (orgTh) with mul-
cification within total occlusions, probably because the plaque
tiple capillary channels
has formed via organized thrombus, as opposed to successive
ruptures of necrotic cores with preserved blood flow.
cells in the absence of a significant influx of macrophages
and other inflammatory cells. The term fibrous plaque has
Fibrous and Fibrocalcific Plaques not been typically used in formal classifications of coronary
artery atherosclerosis, partly because it is a general category
A subset of coronary artery plaques have little evidence of that may represent an etiologically diverse set of lesions.
lipid pool or necrotic core formation and are designated as
fibrous or fibrocalcific plaques, depending on the presence of
calcification. The mechanism of plaque enlargement of such Intraplaque Hemorrhage
plaques is unknown but may be in some cases related to
propagated thrombus or healed plaque erosions (Fig. 19.8). As mentioned above, the interrelationship between intra-
These lesions are rich in type I collagen, but type III collagen plaque hemorrhage, fibrin, necrotic core, and vasa vasorum
and proteoglycans may also be present. The calcium deposi- has not been fully explained. Most vasa vasorum originate
tion may be related to calcified apoptotic smooth muscle from the parent artery and ramify plexogenically in the
220 H. Mori et al.

deposition in the coronary plaques is higher in patients with

acute coronary syndrome than in those dying of noncardiac
causes [32]. Since plaque hemorrhage is a frequent phenom-
enon, especially in advanced plaques, it is not surprising that
some of the free cholesterol may have its origin from the red
cell membrane. The increase in iron content correlated with
intraplaque hemorrhage is identified by glycophorin A stain-
ing, a red-cell specific anion-exchange protein [13].
Rupture of vasa vasorum is not the sole etiology of intra-
plaque hemorrhage. As Constantinides originally suggested,
plaque hemorrhage may occur from cracks or fissures origi-
nating from the lumen [33]. Intraplaque hemorrhage, whether
from vasa vasorum or the lumen, may be critical to the lesion
progression, as it may provide a significant source of non-
metabolic cholesterol [13, 34].

Coronary Calcification

In early coronary atherosclerotic plaques, coronary calcifica-

tion is identified using sensitive histochemical stains, such as
the von Kossa stain. When such tests are applied to fatty
streaks, there is little if any calcium is present. However,
pathological intimal thickening and fibrous plaques have a
high propensity for microscopic calcification, especially
when there is evidence of smooth muscle cell apoptosis. The
biology of intimal calcification within atherosclerotic regions
is complex and involves smooth muscle cell matrix vesicles
(related to apoptotic bodies), bone-associated proteins, lip-
ids, and inflammation in the formation of extracellular inti-
mal calcifications [12]. Almost all necrotic cores with
apoptotic smooth muscle cells contain calcification (at least
90%), and those cores with predominantly macrophage infil-
trates are less likely to initially calcify. The nature of the cal-
Fig. 19.8  Healed erosion shows deep multilayering of collagen sepa- cifications differs by cell type, in that fibrous cores contain
rated by elastin layers (arrowheads) and a paucity of smooth muscle finely granular calcifications, occasionally coalescing into
cells. The superficial plaque is rich in smooth muscle cells, collagen, masses, whereas macrophage “necrotic” cores often contain
and proteoglycans
larger crystalline deposits. Calcification of fibrous plaques in
which there is smooth muscle cell death cores progresses to
adventitia [29]. There is a clear relationship between num- plate-like sheets of calcification and often pipestem-like
bers of vasa vasorum (Fig.  19.9) and blood flow through arteries. However, calcification of necrotic, inflamed,
them and mass of coronary plaque [30]. In addition, increased macrophage-­rich cores tends to be more irregular, resulting
vasa vasorum are associated with atherosclerotic plaque in a radiographic appearance of irregular calcification.
expansion [29], and plaque regression is paralleled by a Ultrastructural study of coronary plaques will demonstrate
decrease in blood flow through the vasa vasorum [31]. calcifications in close contact with dying smooth muscle
Further evidence for a dynamic relationship between vasa cells, macrophages, and surrounding cholesterol clefts
vasorum and the atherosclerotic process are the fact that they (Fig. 19.10). In less than 5% of plaques, actual ossification
respond to vasoconstrictor stimuli and that their density is may occur, with osteoblasts, osteoclasts, and even marrow
associated with exertional plaque rupture [22]. formation (Fig. 19.11).
The mechanisms by which vasa vasorum contribute to Calcification does not in itself appear to play a role in the
plaque growth are poorly understood. Plaque iron content is thrombotic process, except under the unusual circumstances
a surrogate marker of preceding hemorrhage and correlates of nodular calcification. Successive ruptures or fissures of
with plaque neovascularization [29]. The frequency of iron the fibrous cap result in layering of the plaque, and the deeper
19  Pathology and Pathophysiology of Coronary Atherosclerotic Plaques 221

a b

c d

Fig. 19.9  The vasa vasorum within plaque. (a) At low magnification, area in B shows a muscular artery feeder vessel traversing the media.
there is thinning of the media with dilated spaces within the plaque. (b) (d) A different segment of near-total occlusion stained with Ulex euro-
A higher magnification of the boxed area in (a) shows dilated channels paeus lectin highlighting neovessels brown within the adventitia and
to be blood vessels (arrows). (c) A higher magnification of the boxed intima

layers close to the internal elastic lamina typically demon- erosion does not expose the contents of the necrotic core to
strate more severe calcification than the surface layers. the lumen. Eroded plaques, perhaps partly as a function of
We have performed quantitative analysis of calcified their pathogenesis, do not typically occur in calcified arteries
matrix in intermediate and later plaque stages, including and demonstrate far less calcified matrix than acute or healed
fibroatheromas, acute and healed plaque ruptures, total ruptures. The degree of calcification by plaque type is illus-
occlusions, and plaque erosion. In contrast to rupture, plaque trated in Fig. 19.12.
222 H. Mori et al.

a b c

Fig. 19.10  Ultrastructural images of coronary atherosclerotic plaques smooth muscle cell organelles (b), and small calcifications surrounding
show extracellular calcifications adjacent to an apoptotic smooth mus- cholesterol clefts in area of macrophage degeneration (c)
cle cell (SMC) (a), spherical calcification associated with degenerating

a b c

Fig. 19.11  Ossification in coronary plaque. (a) demonstrates a low-­ (a) is a low magnification of the cross section of the artery; contrast
magnification image of a coronary artery section (proximal left anterior material (black) was injected into the artery postmortem. (b) demon-
descending artery) in the left anterior descending coronary artery of an strates the bony trabeculae and (c) a higher magnification of lacunae
obese elderly woman with hypertensive atherosclerotic heart disease. containing osteoblasts

Fig. 19.12  Amount of a b

calcification by plaque
1.2 22.5
morphology. Morphometric
Calcified area, mm2, mean

measurements were made of 20

1
% cal area, mean

calcified matrix in several 17.5

hundred coronary lesions .8 15
classified by plaque type. (a) 12.5
Healed ruptures demonstrated the .6
10
greatest mean area of .4 7.5
calcification; plaque erosions 5
demonstrate almost no .2
2.5
calcification. (b) When expressed
0 0
as a percentage of the plaque
fibrous plaque
fibroatheroma

thin cap atheroma

hemorrhage into plaque

plaque rupture
healed rupture

plaque erosion
total occlusion

fibrous plaque
fibroatheroma

thin cap atheroma

hemorrhage into plaque

plaque rupture
healed rupture

plaque erosion
total occlusion

area, total occlusion and plaque

ruptures have relatively little
calcium and plaque erosions the
least. Fibroatheroma and healed
ruptures demonstrate the greatest
amount of calcified matrix. (From
Burke et al. [27], with
permission)
19  Pathology and Pathophysiology of Coronary Atherosclerotic Plaques 223

Table 19.3  Relationship between calcification and percent luminal narrowing and residual lumen, by arterial section
Calcified area vs. residual lumen
Calcified area vs. percent luminal narrowing (simple regression) area (simple regression)
Men Women Men and women (combined)
Artery ra P DF ra P DF rb P
LM .20 .20 30 .15 .59 14 .14 .58
PLAD .56 <.0001 122 .15 .27 53 .07 .53
MLAD .45 <.0001 91 .50 <.0001 66 .10 .29
DLAD .64 .0003 27 .31 .11 24 .18 28
LD .59 .001 26 .75 .0004 17 .13 .71
PLC .30 .01 78 .25 .07 51 .11 .32
MLC .02 .90 35 .31 .29 13 .21 .40
LOM .60 .0001 34 72 0.01 11 .13 .30
PRC .12 .15 143 .09 .54 54 .09 .38
MRC .29 .001 127 .63 <.0001 68 .09 .37
DRC .32 .0004 116 .36 .06 27 .17 .29
From Burke et al. [27], with permission
Segments are bolded showing r values >0.4 and a significant correlation P < 0.05
Abbreviations: LM left main, PLAD proximal left anterior descending, MLAD mid-left anterior descending, DLAD distal left anterior descending,
LD left diagonal, PLC proximal left circumflex, MLC mid-left circumflex, LOM left obtuse marginal, PRC proximal right coronary, MRC mid-­
right coronary, DRC distal right coronary, df degrees of freedom
a
T values are all positive (positive correlation)
b
T values are negative (PLAD, MRC, DRC, LD) or positive (LM, LOM, MLAD, PLC, MLC, PRC)

Fig. 19.13  Relationship between a erosions b acute ruptures

plaque morphology and radiographic
calcification. Plaque erosions (a)
were exclusively present in areas with
stippled or no calcification. Plaque
ruptures (b) were most frequently speckled
none
seen in areas of speckled calcification fragmented
speckled
but were also present in blocked or diffuse
diffuse calcification. Curiously, there
were no ruptures in segments devoid
of any calcification. Thin-capped
atheromas were most frequently
present in areas of speckled
calcification (c) but were also seen in c Vulnerable plaques d healed ruptures
heavily calcified or uncalcified areas,
suggesting that calcification pattern is
not helpful in diagnosing these
lesions. Healed ruptures are almost
always seen in areas of calcification none none
and most frequently in diffusely speckled speckled
calcified areas (d). (From Burke et al. fragmented fragmented
[27], with permission) diffuse diffuse

The relationship between coronary plaque calcification “process marker,” unlike markers of inflammation [36].
and plaque instability has been debated. Biomechanical stud- These findings are corroborated by autopsy studies that dem-
ies have calculated stress at different regions of the plaque. onstrate a good correlation between plaque size and morpho-
Mathematical models using large-strain finite element analy- metric analysis of calcification, but no correlation between
sis have shown that increased lipids are associated with areas residual lumen and calcification (Table  19.3). Few studies
of weakness of the fibrous cap, but not calcification [35]. have correlated the radiologic pattern of calcification with
Although calcification is a good marker for plaque burden, plaque instability [37], but there is some suggestion that
absolute calcium scores do not indicate plaques that are speckled or fragmented calcification patterns as determined
unstable or prone to result in clinical events. It has been radiologically are most likely associated with unstable
stated that calcification is a “disease marker” as opposed to a plaque types (Fig. 19.13).
224 H. Mori et al.

 ulnerable Plaque by Computed

V rotic plaques on coronary CT images characterized by a
Tomography necrotic core with low CT attenuation area surrounded by
a rim-like area of higher CT attenuation as “napkin ring”
Cardiac computed tomography (CT) imaging is now able [39–41] (Fig.  19.14). This “napkin-ring” finding has a
to visualize the coronary lumen and atherosclerotic high specificity and high positive predictive value for the
plaques. Motoyama et  al. reported that the features of presence of advanced lesions [42, 43]. The noninvasive
high-risk plaque for acute coronary syndrome were posi- identification of vulnerable plaque is one of the ultimate
tive remodeling, spotty calcification, and low plaque atten- goals of coronary imaging because it would improve risk
uation (<30 Hounsfield units [HU]) [38]. Other studies stratification of both symptomatic and asymptomatic
have described specific attenuation pattern of atheroscle- patients [44].

Longitudinal Image of CT

Fig. 19.14  The longitudinal CT image of the RCA demonstrates ath- fibrous plaque tissue. In addition, the presence of spotty calcification
erosclerotic plaques with spotty calcification in the proximal segment correlates to the CT findings. The CT images from PRCA2 show a
of the RCA.  The bars (PRCA 1 and PRCA2) indicate cross sections pronounced napkin-ring-like attenuation pattern. The circumferential
within the coronary plaques that were compared with histopathology. outer rim (red dashed line) of the plaque has a higher CT attenuation
The cross-sectional CT images from PRCA1 show a coronary athero- in both the noncontrast- (A) and contrast-enhanced (B) images
sclerotic plaque cross section with napkin-ring-like attenuation pattern (57.2 ± 8.8 HU, range 40.0–81.0 HU vs. 57.9 ± 8.7 HU, range 35.0–
and spotty calcification. The circumferential outer rim (red dashed 76.0 HU, respectively) as compared to the attenuation within the cen-
line) of the plaque has a higher CT attenuation in both the noncontrast- tral part of the plaque (21.8  ±  4.3 HU, range 13.5–31.6 HU and
(A) and contrast-enhanced (B) images (61.8 ± 9.3 HU, range 42.4–74.9 26.0  ±  2.0 HU, range 22.0–31.0 HU on noncontrast- and contrast-
HU vs. 67.3 ± 11.6 HU, range 43.4–87.0 HU, respectively) as com- enhanced images, respectively). The average plaque attenuation on
pared to the attenuation within the central part of the plaque (43.4 ± 7.5 nonenhanced CT was 48.1 ± 14.2 HU versus 52.2 ± 14.0 HU on the
HU, range 35.2–62.9 HU and 43.2 ± 14.0 HU, range 21.2–72.5 HU, on contrast-enhanced CT.  The corresponding histopathological section
noncontrast- and contrast-enhanced images, respectively). The aver- (C) demonstrates a late fibroatheroma. Again, the plaque contains a
age noncalcified plaque attenuation on nonenhanced CT was necrotic core (stars), which correlates to the low-attenuation plaque
52.8  ±  10.9 HU versus 58.5  ±  17.1 HU attenuation on the contrast- core on the CT images. The outer portion of the plaque (red-dashed
enhanced image. Histopathology revealed a thin-cap fibroatheroma (C, line) contains a significant amount of fibrous plaque tissue correlating
D) with spotty calcification (E). Again, the necrotic core (stars) cor- to the high-attenuation CT rim. Moreover, the histopathological analy-
relates with the low-attenuation plaque core on the CT images. Similar sis revealed significant vasa vasorum (C, arrowheads) accompanied by
to the previous figures, the outer-rim attenuation (red dashed line) on macrophage infiltration in the basal plaque area. (From Maurovich-
the noncontrast- and contrast-enhanced CT images corresponds to the Horvat et al. [42], with permission)
19  Pathology and Pathophysiology of Coronary Atherosclerotic Plaques 225

PRCA 1

a b c d

PRCA 2

a b c

Fig. 19.14 (continued)

rinone among patients with severe heart failure. Vesnarinone Trial

Investigators. N Engl J Med. 1998;339:1810–6.
References 8. McGill HC, McMahan CA, Herderick EE, Tracy RE, Malcom
GT, Zieske AW, et al. Effects of coronary heart disease risk factors
1. Virmani R, Kolodgie FD, Burke AP, Farb A, Schwartz SM. Lessons on atherosclerosis of selected regions of the aorta and right coro-
from sudden coronary death: a comprehensive morphological clas- nary artery. PDAY Research Group. Pathobiological Determinants
sification scheme for atherosclerotic lesions. Arterioscler Thromb of Atherosclerosis in Youth. Arterioscler Thromb Vasc Biol.
Vasc Biol. 2000;20:1262–75. 2000;20:836–45.
2. Stary HC, Chandler AB, Dinsmore RE, Fuster V, Glagov S, Insull 9. Fan J, Watanabe T. Inflammatory reactions in the pathogenesis of
W, et al. A definition of advanced types of atherosclerotic lesions atherosclerosis. J Atheroscler Thromb. 2003;10:63–71.
and a histological classification of atherosclerosis. A report from the 10. Aikawa M, Rabkin E, Okada Y, Voglic SJ, Clinton SK, Brinckerhoff
Committee on Vascular Lesions of the Council on Arteriosclerosis, CE, et  al. Lipid lowering by diet reduces matrix metallopro-
American Heart Association. Arterioscler Thromb Vasc Biol. teinase activity and increases collagen content of rabbit ather-
1995;15:1512–31. oma: a potential mechanism of lesion stabilization. Circulation.
3. Davies MJ, Thomas A.  Thrombosis and acute coronary-artery 1998;97:2433–44.
lesions in sudden cardiac ischemic death. N Engl J Med. 11. Kockx MM, De Meyer GR, Bortier H, de Meyere N, Muhring J,
1984;310:1137–40. Bakker A, et al. Luminal foam cell accumulation is associated with
4. Yahagi K, Kolodgie FD, Otsuka F, Finn AV, Davis HR, Joner smooth muscle cell death in the intimal thickening of human saphe-
M, et  al. Pathophysiology of native coronary, vein graft, and in-­ nous vein grafts. Circulation. 1996;94:1255–62.
stent atherosclerosis. Nat Rev Cardiol. 2016;13:79–98. Nature 12. Proudfoot D, Shanahan CM.  Biology of calcification in vascular
Publishing Group. cells: intima versus media. Herz. 2001;26:245–51.
5. Nakashima Y, Chen Y-X, Kinukawa N, Sueishi K. Distributions of 13. Kolodgie FD, Gold HK, Burke AP, Fowler DR, Kruth HS, Weber
diffuse intimal thickening in human arteries: preferential expres- DK, et al. Intraplaque hemorrhage and progression of coronary ath-
sion in atherosclerosis-prone arteries from an early age. Virchows eroma. N Engl J Med. 2003;349:2316–25.
Arch. 2002;441:279–88. 14. Sluimer JC, Kolodgie FD, Bijnens AP, Maxfield K, Pacheco E,
6. Ikari Y, McManus BM, Kenyon J, Schwartz SM. Neonatal intima Kutys B, et al. Thin-walled microvessels in human coronary athero-
formation in the human coronary artery. Arterioscler Thromb Vasc sclerotic plaques show incomplete endothelial junctions. Relevance
Biol. 1999;19:2036–40. of compromised structural integrity for intraplaque microvascular
7. Cohn JN, Goldstein SO, Greenberg BH, Lorell BH, Bourge RC, leakage. J Am Coll Cardiol. 2009;53:1517–27. American College
Jaski BE, et al. A dose-dependent increase in mortality with vesna- of Cardiology Foundation.
226 H. Mori et al.

15. Burke AP, Farb A, Malcom GT, Liang YH, Smialek J, Virmani 32. Virmani R, Roberts WC. Extravasated erythrocytes, iron, and fibrin
R.  Coronary risk factors and plaque morphology in men in atherosclerotic plaques of coronary arteries in fatal coronary
with coronary disease who died suddenly. N Engl J Med. heart disease and their relation to luminal thrombus: frequency
1997;336:1276–82. and significance in 57 necropsy patients and in 2958 five mm
16. Michel JB, Virmani R, Arbustini E, Pasterkamp G.  Intraplaque segments of 224 major epicardial coronary arteries. Am Heart J.
haemorrhages as the trigger of plaque vulnerability. Eur Heart J. 1983;105:788–97.
2011;32:1977–85. 33. Constantinides P. Coronary thrombosis linked to fissure in athero-
17. Farb A, Burke AP, Tang AL, Liang TY, Mannan P, Smialek J, et al. sclerotic vessel wall. JAMA. 1964;188 Suppl:35–7.
Coronary plaque erosion without rupture into a lipid core. A fre- 34. Pasterkamp G, Virmani R. The erythrocyte: a new player in athero-
quent cause of coronary thrombosis in sudden coronary death. matous core formation. Heart. 2002;88:115–6.
Circulation. 1996;93:1354–63. 35. Huang H, Virmani R, Younis H, Burke AP, Kamm RD, Lee RT. The
18. Roberts WC, Kragel AH, Gertz SD, Roberts CS, Kalan JM. The heart impact of calcification on the biomechanical stability of atheroscle-
in fatal unstable angina pectoris. Am J Cardiol. 1991;68:22B–7B. rotic plaques. Circulation. 2001;103:1051–6.
19. Libby P.  Molecular bases of the acute coronary syndromes.
36. Hunt ME, O’Malley PG, Vernalis MN, Feuerstein IM, Taylor

Circulation. 1995;91:2844–50. AJ.  C-reactive protein is not associated with the presence or
20. Libby P, Sukhova G, Lee RT, Galis ZS.  Cytokines regulate vas- extent of calcified subclinical atherosclerosis. Am Heart J.
cular functions related to stability of the atherosclerotic plaque. J 2001;141:206–10.
Cardiovasc Pharmacol. 1995;25 Suppl 2:S9–12. 37. Burke AP, Taylor A, Farb A, Malcom GT, Virmani R. Coronary cal-
21. Virmani R, Kolodgie FD, Burke AP, Finn AV, Gold HK, Tulenko cification: insights from sudden coronary death victims. Z Kardiol.
TN, et  al. Atherosclerotic plaque progression and vulnerability 2000;89 Suppl 2:49–53.
to rupture: angiogenesis as a source of intraplaque hemorrhage. 38. Motoyama S, Kondo T, Sarai M, Sugiura A, Harigaya H, Sato T,
Arterioscler Thromb Vasc Biol. 2005;25:2054–61. et  al. Multislice computed tomographic characteristics of coro-
22. Burke AP, Farb A, Malcom GT, Liang Y, Smialek JE, Virmani nary lesions in acute coronary syndromes. J Am Coll Cardiol.
R. Plaque rupture and sudden death related to exertion in men with 2007;50:319–26.
coronary artery disease. JAMA. 1999;281:921–6. 39. Narula J, Achenbach S.  Napkin-ring necrotic cores: defining cir-
23. Schwartz RS, Burke A, Farb A, Kaye D, Lesser JR, Henry TD, cumferential extent of necrotic cores in unstable plaques. JACC
et al. Microemboli and microvascular obstruction in acute coronary Cardiovasc Imaging. 2009;2:1436–8.
thrombosis and sudden coronary death. relation to epicardial plaque 40. Kashiwagi M, Tanaka A, Kitabata H, Tsujioka H, Kataiwa H,
histopathology. J Am Coll Cardiol. 2009;54:2167–73. Elsevier Inc. Komukai K, et  al. Feasibility of noninvasive assessment of thin-­
24. Yahagi K, Zarpak R, Sakakura K, Otsuka F, Kutys R, Ladich E, cap fibroatheroma by multidetector computed tomography. JACC
et al. Multiple simultaneous plaque erosion in 3 coronary arteries. Cardiovasc Imaging. 2009;2:1412–9.
JACC Cardiovasc Imaging. 2014;7:1172–4. 41. Tanaka A, Shimada K, Yoshida K, Jissyo S, Tanaka H, Sakamoto
25. Yahagi K, Davis HR, Arbustini E, Virmani R.  Sex differences in M, et  al. Non-invasive assessment of plaque rupture by 64-slice
coronary artery disease: pathological observations. Atherosclerosis. multidetector computed tomography – comparison with intravascu-
2015;239:260–7. lar ultrasound. Circ J. 2008;72:1276–81.
26. Mann J, Davies MJ.  Mechanisms of progression in native coro- 42. Maurovich-Horvat P, Hoffmann U, Vorpahl M, Nakano M, Virmani
nary artery disease: role of healed plaque disruption. Heart. R, Alkadhi H. The napkin-ring sign: CT signature of high-risk coro-
1999;82:265–8. nary plaques? JACC Cardiovasc Imaging. 2010;3:440–4.
27. Burke AP, Weber DK, Kolodgie FD, Farb A, Taylor AJ, Virmani 43. Maurovich-Horvat P, Schlett CL, Alkadhi H, Nakano M, Otsuka
R.  Pathophysiology of calcium deposition in coronary arteries. F, Stolzmann P, et al. The napkin-ring sign indicates advanced ath-
Herz. 2001;26:239–44. erosclerotic lesions in coronary CT angiography. JACC Cardiovasc
28. Sakakura K, Nakano M, Otsuka F, Yahagi K, Kutys R, Ladich E, Imaging. 2012;5:1243–52.
et al. Comparison of pathology of chronic total occlusion with and 44. Narula J, Garg P, Achenbach S, Motoyama S, Virmani R, Strauss
without coronary artery bypass graft. Eur Heart J. 2014;35:1683–93. HW. Arithmetic of vulnerable plaques for noninvasive imaging. Nat
29. Kumamoto M, Nakashima Y, Sueishi K. Intimal neovascularization Clin Pract Cardiovasc Med. 2008;5 Suppl 2:S2–10.
in human coronary atherosclerosis: its origin and pathophysiologi- 45. Can atherosclerosis imaging techniques improve the detection of
cal significance. Hum Pathol. 1995;26:450–6. patients at risk for ischemic heart disease? Proceedings of the 34th
30. Heistad DD, Armstrong ML.  Blood flow through vasa vasorum Bethesda Conference. Bethesda, Maryland, USA. October 7, 2002.
of coronary arteries in atherosclerotic monkeys. Arteriosclerosis. J Am Coll Cardiol. 2003;41:1856–917.
1986;6:326–31. 46. Burke AP, Kolodgie FD, Farb A, Weber DK, Malcom GT, Smialek
31. Williams JK, Armstrong ML, Heistad DD.  Vasa vasorum in ath- J, et  al. Healed plaque ruptures and sudden coronary death: evi-
erosclerotic coronary arteries: responses to vasoactive stimuli and dence that subclinical rupture has a role in plaque progression.
regression of atherosclerosis. Circ Res. 1988;62:515–23. Circulation. 2001;103:934–40.
 CT Cardiac Anatomy
20
Michael A. Kadoch and Hans-Christoph R. F. Becker

The heart is a mediastinal structure that rests on the dia- The LAD courses in the anterior interventricular groove
phragm and is separated from the lung parenchyma on either and gives off septal perforator braches that supply the inter-
side by the pericardium, pleura, and fat. The innermost layer ventricular septum in addition to diagonal branches that sup-
of the ventricle is the endocardium, the bulk of the ventricu- ply the anterior and anterolateral walls of the left ventricle.
lar wall is comprised of the myocardium, and the outermost Myocardial bridging is a common congenital anomaly that is
layer is the epicardium. The fat layer just beyond the epicar- seen in approximately 25% of the population and is most
dium represents epicardial adipose tissue. The visceral and commonly identified within the midportion of the LAD. This
parietal layers of the pericardium are identified just beyond bridged segment is classically described as being protected
the epicardial fat. A small amount of pericardial fluid is from the development of atherosclerotic disease but can
physiologically present between these layers. The normal occasionally be a cause of anginal symptoms and may pre-
thickness of the pericardium is usually 1–2 mm and is con- dispose to increased plaque development within more proxi-
sidered abnormal when greater than 4 mm. The fat layer just mal portions of the LAD with an associated increased risk of
beyond the parietal pericardium represents paracardial acute coronary events due to rupture of this plaque forma-
adipose tissue. The base of the heart refers to its posterior tion. The LAD typically gives off two diagonal branches (D1
surface and is formed by the atria, mainly the left atrium, and and D2), and its distal end can be seen wrapping around the
is separated from the vertebral bodies by the esophagus and left ventricular apex in the majority of patients.
aorta. The apex of the heart refers to its inferior tip and is The LCx courses in the left atrioventricular groove and
usually formed by the left ventricle. contributes obtuse marginal branches (typically two, OM1
and OM2) that supply the lateral and inferolateral walls of
the left ventricle. The most common coronary artery anom-
Coronary Arteries aly involves the LCx arising from either the right coronary
sinus independently or as a branch of the right coronary
The coronary arteries (Fig. 20.1) originate normally from the artery (RCA) before taking a retroaortic course to the left
aortic sinuses of Valsalva that face the pulmonary root. The atrioventricular groove. Knowledge of this anomaly is
left main coronary artery (LM) arises from the left coronary important in patients undergoing aortic valve surgery since
sinus and courses between the left atrial appendage and the its presence can introduce complications during the
right ventricular outflow tract. The LM usually bifurcates procedure.
into the left anterior descending (LAD) and left circumflex The RCA arises from the right coronary sinus and courses
(LCx) coronary arteries. In approximately 20% of patients, within the right atrioventricular groove. The first branch of
the LM trifurcates with a ramus intermedius (RI) branch the RCA is often the conus artery, which typically supplies
identified between the LAD and LCx. the anterior interventricular septum in addition to the conal
tissue of the right ventricular outflow tract. Independent ori-
gin of the conus artery from the right coronary sinus is a
common variant seen in up to 50% of patients. The second
M. A. Kadoch (*) branch of the RCA is typically the sinoatrial (SA) nodal
Department of Radiology, University of California, Davis, CA, artery, which courses posteriorly to supply the SA node. The
USA SA nodal artery is a branch of the LCx as a variant in approx-
e-mail: [email protected]
imately 40% of patients. The acute marginal branches arise
H.-C. R. F. Becker from the mid to distal RCA and are its largest branches,
Department of Radiology, Stanford University, Stanford, CA, USA

© Humana Press 2019 227

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_20
228 M. A. Kadoch and H.-C. R. F. Becker

a b c

d e f

g h i

Fig. 20.1 (a–q) Axial CT cardiac anatomy. 1. Main pulmonary artery. Aortic valve. 26. First obtuse marginal branch. 27. Acute marginal
2. Ascending thoracic aorta. 3. Descending thoracic aorta. 4. Right pul- branch. 28. Pericardium. 29. Anterior leaflet of the mitral valve. 30.
monary artery. 5. Superior vena cava. 6. Left superior pulmonary vein. Posterior leaflet of the mitral valve. 31. Right atrium. 32. Interatrial
7. Left main coronary artery. 8. Left atrial appendage. 9. Right atrial septum. 33. Right inferior pulmonary vein. 34. Left ventricle. 35. Fossa
appendage. 10. Right superior pulmonary vein. 11. Left atrium. 12. Left ovalis. 36. Crista terminalis. 37. Moderator band. 38. Interventricular
anterior descending coronary artery. 13. First septal perforator. 14. First septum. 39. Anterior papillary muscle of the left ventricle. 40. Right
diagonal branch. 15. Left inferior pulmonary vein. 16. Conus artery. 17. ventricle. 41. Tricuspid valve. 42. Posterior papillary muscle of the left
Right coronary artery. 18. Left circumflex coronary artery. 19. Conus. ventricle. 43. Coronary sinus. 44. Second obtuse marginal branch. 45.
20. Great cardiac vein. 21. Coumadin ridge. 22. Right middle pulmo- Posterolateral branch. 46. Inferior vena cava. 47. Posterior descending
nary vein. 23. Second diagonal branch. 24. Sinoatrial nodal artery. 25. coronary artery. 48. Middle cardiac vein
20  CT Cardiac Anatomy 229

j k l

m n o

p q

Fig. 20.1 (continued)

which supply the right ventricular free wall. The atrioven- right-dominant system (approximately 80% of patients),
tricular nodal artery is usually a branch of the RCA at the both branches are supplied by the RCA. In a left-dominant
crux of the heart where it can be seen coursing anteriorly to system (approximately 15% of patients), both branches are
supply the atrioventricular node. The posterior descending supplied by the LCx. In a codominant system (approximately
artery (PDA) and the posterolateral branch (PLB) are the ter- 5% of patients), the PDA and/or PLB are supplied by
minal branches of the RCA in a right-dominant system sup- branches of both the RCA and LCx.
plying the inferior and inferolateral walls of the left ventricle. In 2009, the Society for Cardiovascular Computed
The PDA courses in the posterior interventricular groove. Tomography (SCCT) released guidelines for the interpreta-
Coronary artery dominance is most accurately defined by tion and reporting of coronary computed tomographic
the coronary artery that supplies the PDA and PLB.  In a angiography (CCTA) (Fig. 20.2; Table 20.1) [1]. As part of
230 M. A. Kadoch and H.-C. R. F. Becker

Fig. 20.2  SCCT coronary

segmentation diagram. See
Table 20.1. (From Raff et al.
[1], with permission) 1
RCA
2 LAD

3 LEFT MAIN 7
10
6 DIAGONAL
4 5 (D2)
R-PDA
9
11 DIAGONAL
15 17
L-PDA (D1)

L-PLB CIRCUMFLEX
14 12

16 13
R-PLB 14 OBTUSE
MARGINAL 1
(OM1) RAMUS

OBTUSE MARGINAL 2
(OM2)

Table 20.1  SCCT segmentation of axial coronary anatomy

Segment Abbreviation Description
Proximal RCA pRCA Ostium of the RCA (right coronary artery) to one-half the distance to the acute margin of
heart
Mid-RCA mRCA End of proximal RCA to the acute margin of heart
Distal RCA dRCA End of mid-RCA to origin of the PDA (posterior descending artery)
PDA-RCA R-PDA PDA from RCA
PLB-RCA R-PLB PLB (posterior-lateral branch) from RCA
LM LM Ostium of LM (left main) to bifurcation of LAD (left anterior descending artery) and LCx
(left circumflex artery)
Proximal LAD pLAD End of LM to the first large septal or D1 (first diagonal), whichever is most proximal
Mid LAD mLAD End of proximal LAD to one-half the distance to the apex
Distal LAD dLAD End of mid-LAD to end of LAD
Diagonal 1 D1 First diagonal branch D1
Diagonal 2 D2 Second diagonal branch D2
Proximal LCx pCx End of LM to the origin of the OM1 (first obtuse marginal)
OM1 OM1 First OM1 traversing the lateral wall of the left ventricle
Mid and distal LCx LCx Traveling in the AV groove, distal to the first obtuse marginal branch to the end of the
vessel or origin of the L-PDA (left posterior descending artery)
OM2 OM2 Second marginal OM2
PDA-LCx L-PDA PDA from LCx
Ramus intermedius RI Vessel originating from the left main between the LAD and LCx in case of a trifurcation
PLB-L L-PLB PLB from LCx
Dashed lines represent division between RCA, LAD, and LCx and the end of the
LMPLB = PLV (posterior left ventricular branch). Additional nomenclature may be added,
for example, D3, R-PDA2, SVF (saphenous vein graft), mLAD
See Fig. 20.2 for accompanying diagram. From Raff et al. [1], with permission

that publication, the SCCT advocates for the use of an angiography. Adopting this standardized approach to coro-
18-­segment model of the axial coronary anatomy, which nary segmentation can improve the description and com-
more closely emulates the CCTA views of the coronary munication of findings among radiologists and with
arteries than the standard views obtained from invasive referring clinicians.
20  CT Cardiac Anatomy 231

Coronary Veins outflow tract (RVOT), which is in distinction to the fibrous

continuity of the mitral and aortic valves of the left heart.
The anatomy of the coronary veins (Fig.  20.1) is not as Blood returns to the left atrium (LA) via the pulmonary
straightforward as the arterial anatomy since variations are veins. The LAA is variable in shape but can be distinguished
common [2]. However, knowledge of the coronary venous from the RAA by its smaller size, finger-like configuration,
system is increasingly important as more CT imaging is and corrugated internal surface owing to the presence of par-
being performed prior to cardiac pacing and transvenous allel pectinate muscles. The atria are separated by the inter-
ablation procedures. atrial septum. The fossa ovalis (FO) is an oval-shaped
The great cardiac vein accompanies the LAD along the depression on the right atrial side of the interatrial septum,
anterior interventricular groove and the LCx along the left which may be traversed during interventional procedures
atrioventricular groove before terminating in the left end of requiring transseptal puncture. The FO is an important land-
the coronary sinus (CS). The middle cardiac vein accompa- mark to identify because it may be spared in cases of lipoma-
nies the PDA along the posterior interventricular groove tous septal hypertrophy, is the location where a patent
before terminating in the right end of the CS. The small car- foramen ovale may be seen, and is the most common region
diac vein accompanies the RCA along the right atrioventric- of attachment for cardiac myxomas. Maximum anteroposte-
ular groove before terminating in the right end of the CS. The rior diameter measurements of the LA greater than 4.5 cm at
right marginal vein runs along the inferior margin of the the level of the aortic root on axial images are 91% specific
heart before terminating in the small cardiac vein. The for LA enlargement [3]. LA enlargement is a common find-
oblique vein of the left atrium runs along the posterior sur- ing in patients with atrial fibrillation.
face of the left atrium before terminating in the left end of the Blood enters the left ventricle (LV) via the mitral valve
CS.  The posterior vein of the left ventricle runs along the (MV). The LV lies posterior and to the left of the RV and is
diaphragmatic surface of the left ventricle before terminating distinguished by its smoother wall and two large papillary
in the middle of the CS. muscles (anterior and posterior). The left ventricular apical
The CS is identified within the posterior atrioventricular thin point (LVATP) is an important landmark to identify in
groove and is of variable length. A prominent Thebesian order to generate the standard cardiac planes. Maximum
valve is occasionally seen at the ostium of the CS, which diameter measurements of the LV greater than 5.5 cm at the
may pose an obstacle to interventionalists trying to engage level of the papillary muscle tips on axial images are 93%
the CS. specific for LV enlargement [3].
The anterior cardiac veins of the right ventricle and the The American Heart Association (AHA) recommends
smallest cardiac veins (Thebesian veins) terminate directly dividing the heart into 17 segments for assessment of the
into the right atrium. The small cardiac vein may also termi- myocardium and the left ventricular cavity [4]. According to
nate directly into the right atrium. this segmentation system, the heart is divided into equal
thirds perpendicular to its long-axis with the basal portion
identified at the tips of the mitral valve leaflets, the mid-­
cavity identified at the level of the papillary muscles, and the
Cardiac Chambers apical portion identified beyond the papillary muscles but
before the cavity ends (Fig.  20.3). Individual myocardial
The right atrium (RA) receives blood from the superior vena segments are then assigned to the three major coronary arter-
cava (SVC), inferior vena cava (IVC), and the CS. The crista ies recognizing that there is anatomic variability (Fig. 20.4).
terminalis is a vertically oriented smooth muscle ridge within
the superior portion of the right atrium that might occasion-
ally be mistaken for a mass when increased in prominence. Cardiac Valves
The Eustachian valve is occasionally seen at the junction of
the IVC and the RA.  The right atrial appendage (RAA) is The TV is comprised of anterior, septal, and posterior leaf-
pyramidal in shape and larger in size than the left atrial lets. The anterior leaflet is the largest and separates the inflow
appendage (LAA). and outflow tracts of the RV.  The septal leaflet has many
Blood enters the right ventricle (RV) from the RA via the direct chordal attachments to the interventricular septum,
tricuspid valve (TV). The RV is more coarsely trabeculated and the posterior leaflet is usually the smallest. The normal
than the left ventricle (LV) and can also be distinguished by TV is slightly displaced apically as compared with the
the presence of the moderator band, which is seen extending MV.  In Ebstein’s anomaly, this displacement will exceed
from the base of the anterior papillary muscle to the interven- 8 mm/m2 of the body surface area.
tricular septum. The TV and pulmonic valve (PV) are sepa- The MV is comprised of anterior and posterior leaflets
rated by the muscular conal tissue of the right ventricular (Fig. 20.1). The large semicircular anterior leaflet separates
232 M. A. Kadoch and H.-C. R. F. Becker

Fig. 20.3 (a–d) Short axis

view of the heart with AHA
a b
17-segment model of the left
ventricle. Segment 1: basal
anterior wall. Segment 2:
basal anteroseptal wall.
Segment 3: basal inferoseptal
wall. Segment 4: basal
inferior wall. Segment 5:
basal inferolateral wall.
Segment 6: basal anterolateral
wall. Segment 7: mid-anterior
wall. Segment 8: mid-­
anteroseptal wall. Segment 9:
mid-inferoseptal wall.
Segment 10: mid-inferior
wall. Segment 11: mid-­
inferolateral wall. Segment c d
12: mid-anterolateral wall.
Segment 13: apical anterior
wall. Segment 14: apical
septal wall. Segment 15:
apical inferior wall. Segment
16: apical lateral wall.
Segment 17: apex

The aortic valve (AV) is normally tricuspid and is com-

Coronary Artery Territories
prised of semilunar-shaped left, right, and noncoronary
Vertical cusps and associated commissures extending from the aortic
Short Axis Long Axis annulus to the sinotubular junction (Fig. 20.5). The noncoro-
Apical Mid Basal Mid
nary cusp typically faces the interatrial septum and the right
1 coronary cusp typically faces the sternum. The PV is also
7
13 2 6 normally tricuspid and defines the point of transition from
8 12
14 16 17
the RVOT to the main pulmonary artery (MPA). The MV and
9 5
AV are better evaluated with CT than the TV and PV.
15 3
10
4

LAD RCA LCX

Pulmonary Veins

Fig. 20.4  Assignment of the 17 AHA myocardial segments to the cor- The muscular sleeves of the distal pulmonary veins are a fre-
onary artery territories. (LAD left anterior descending coronary artery, quent source of ectopic foci in patients with atrial fibrilla-
RCA right coronary artery, LCX left circumflex coronary artery). (From tion. CT is routinely performed for anatomic mapping of the
Cerqueira et al. [4], with permission)
pulmonary veins prior to ablation therapy (Fig.  20.6). The
the inflow and outflow tracts of the LV. The anterior and pos- most common configuration of the pulmonary veins includes
terior papillary muscles arise from the left ventricular wall four separate branches for the right superior, right inferior,
and attach to the MV leaflets via fan-like chordae tendineae. left superior, and left inferior pulmonary veins. Common
Mitral valve prolapse (MVP) is defined by systolic atrial dis- variants include independent drainage of the right middle
placement of one or both of the MV leaflets by greater than lobe vein into the left atrium and a single left-sided venous
2 mm above the mitral annulus in a long-axis view. ostium [5, 6]. The “Coumadin ridge” is a ridge of smooth
20  CT Cardiac Anatomy 233

Fig. 20.5 (a–d) The aortic root. 1. Aortic a b

annulus. 2. Sinuses of Valsalva. 3. Sinotubular
junction. 4. Right coronary sinus. 5. Left-right
commissure. 6. Left coronary sinus.
7. Left-noncoronary commissure.
8. Noncoronary sinus. 9. Right-noncoronary
commissure. 10. Left main coronary artery.
11. Right coronary artery. 12. Sinoatrial nodal
artery

c d

Fig. 20.6 (a–c) Pulmonary veins. 1. Left a

superior pulmonary vein. 2. Left inferior
pulmonary vein. 3. Right superior pulmonary
vein. 4. Right middle pulmonary vein. 5.
Right inferior pulmonary vein. 6. Coumadin
ridge. 7. Left atrial appendage. 8. Left main
coronary artery

b c
234 M. A. Kadoch and H.-C. R. F. Becker

a b c

d e

Fig. 20.7  Cardiac planes. (a) Horizontal long-axis (HLA) view. (b) tricular outflow tract (RVOT) view. 1. Left ventricular apical thin point.
Left ventricular outflow tract (LVOT) view. (c) Vertical long-axis 2. Mitral valve. 3. Tricuspid valve. 4. Aortic valve. 5. Left atrial append-
(VLA) view. (d) Right ventricular three-chamber view. (e) Right ven- age. 6. Pulmonic valve. 7. Right ventricular apex

muscle found along the wall of the LA at the junction of the center of the TV, the center of the PV, and the RV apex; and
LAA and the entrance of the left superior pulmonary vein. Its the RVOT view using the center of the TV, the center of the
name is derived from the fact that it is frequently mistaken PV, and the bifurcation line of the MPA.  The short-axis
for a thrombus (or mass) with subsequent initiation of anti- (SAX) view can then be generated as a stack of perpendicu-
coagulation therapy. Partial anomalous pulmonary venous lar sections through either of the long-axis views of the heart.
return is occasionally encountered as an incidental finding in
asymptomatic adults with the left upper lobe being the most
common site for this abnormality. References
1.  Raff GL, Abidov A, Achenbach S, et  al. SCCT guidelines for the
interpretation and reporting of coronary computed tomographic
Cardiac Planes angiography. J Cardiovasc Comput Tomogr. 2009;3(2):122–36.
2.  Saremi F, Muresian H, Sánchez-Quintana D. Coronary veins: com-
In addition to routine axial, sagittal, and coronal planes, stan- prehensive CT-anatomic classification and review of variants and
dard cardiac planes analogous to those obtained with echo- clinical implications. Radiographics. 2012;32(1):E1–32.
3. Huckleberry J, Haltom S, Issac T, et  al. Accuracy of non-ECG-­
cardiography and MRI are often generated from CT data sets gated computed tomography angiography of the chest in assess-
for better evaluation of the valves, ventricular walls, and out- ment of left-sided cardiac chamber enlargement. J Thorac Imaging.
flow tracts (Fig. 20.7). These planes may be obtained using 2012;27(6):354–8.
three points as follows (Fig. 20.7): the horizontal long-axis 4.  Cerqueira MD, Weissman NJ, Dilsizian V, et al. Standardized myo-
cardial segmentation and nomenclature for tomographic imaging of
(HLA, or four-chamber) view using the center of the MV, the the heart. A statement for healthcare professionals from the Cardiac
center of the TV, and the LVATP; the left ventricular outflow Imaging Committee of the Council on Clinical Cardiology of the
tract (LVOT, or LV three-chamber) view using the center of American Heart Association. Circulation. 2002;105(4):539–42.
the MV, the center of the AV, and the LVATP; the vertical 5.  Marom EM, Herndon JE, Kim YH, et al. Variations in pulmonary
venous drainage of the left atrium: implications for radiofrequency
long-axis (VLA, or LV two-chamber) view using the center ablation. Radiology. 2004;230(3):824–9.
of the MV, the LVATP, and the bifurcation line between the 6.  Ghaye B, Szapiro D, Dacher JN, et al. Percutaneous ablation for atrial
HLA and LVOT views; the RV three-chamber view using the fibrillation: the role of cross-sectional imaging. Radiographics.
2003;23(suppl_1):S19–33.
 Patient Selection: When to Use Cardiac
CT Versus Other Imaging or 21
Non-­imaging Tests

Pal Spruill Suranyi, Akos Varga-Szemes,
Marques L. Bradshaw, Richard R. Bayer II,
Salvatore A. Chiaramida, Peter L. Zwerner,
and David Gregg

Preamble: The Importance cally revise guidelines and appropriateness criteria [1–10];

of Appropriateness however, our aim here is to give a practical overview of how
cardiac CT can complement  – rather than compete with  –
Technological advances in imaging may be contributing to the other imaging and non-imaging tests (Table 21.1).
increasing healthcare expenditures as our ability to vividly
visualize pathology has skyrocketed the utilization of these
imaging techniques. These trends collide with our fears of  verview of the Disadvantages of CT When
O
potentially limited resources in the near future as baby boom- Compared to Other Modalities
ers reach old age and healthcare expenditures are cut back.
This has led to the crescendoing outcry for cost-­effective, pre- The critics of cardiac CT voice concerns over the inherent
cision medicine. Responsible care dictates that we do not per- radiation exposure, the need for iodinated contrast media, the
form novel imaging tests just because we can but rather relatively low temporal resolution, the inability to measure
because we are convinced that there is significant benefit in flow, the limits on myocardial tissue characterization, and
guiding patient management. Heart disease is the leading suboptimal viability imaging with CT.  Several factors can
cause of death in both women and men and consumes a sig- degrade image quality including arrhythmias, inability to do
nificant amount of advanced imaging. Appropriate utilization breath-­holds or raise arms, large body habitus, and metallic
of cardiac CT remains a topic of discussion in the evolving hardware in the chest. Cardiac CTA is felt to be a more “ana-
healthcare delivery system, and we will need to be mindful in tomic” than a “physiologic” test, which is certainly true in
our quest to improve outcomes in all cardiovascular diseases. the way CT is used clinically today; however, newer
Numerous societies have joined forces to create and periodi- advances, such as CT fractional flow reserve (FFR), iodine
content mapping, and quantitative CT perfusion for myocar-
dial blood flow during vasodilator stress, are making strides
P. S. Suranyi (*) · R. R. Bayer II to mitigate those claims. We refer the reader to the respective
Division of Cardiovascular Imaging, Department of Radiology chapters for further detail on these new approaches. Lastly,
and Radiological Science, Medical University of South Carolina,
Charleston, SC, USA the relatively high cost of equipment is a disadvantage when
compared to gamma cameras or echo machines.
Division of Cardiology, Department of Medicine, Medical
University of South Carolina, Charleston, SC, USA
e-mail: [email protected]
A. Varga-Szemes  verview of the Advantages of CT Relative
O
Division of Cardiovascular Imaging, Department of Radiology and to Other Modalities
Radiological Science, Medical University of South Carolina,
Charleston, SC, USA To date, cardiac CT is the most easily accessible, fastest, non-
M. L. Bradshaw invasive, high spatial resolution (submillimeter) tomographic
Department of Radiology and Radiological Sciences, Vanderbilt imaging modality (Table  21.2). Rapid advances in radiation
University School of Medicine, Nashville, TN, USA
dose reduction are minimizing exposure to a point where, in
S. A. Chiaramida · P. L. Zwerner some patients, sub-milliSievert whole-chest coronary CTAs
D. Gregg
Division of Cardiology, Department of Medicine, Medical have become a reality [11]. These low-dose scans are far
University of South Carolina, Charleston, SC, USA below the nuclear exposure by SPECT nuclear imaging [12].

© Humana Press 2019 235

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_21
236 P. S. Suranyi et al.

Table 21.1  Simplified summary for common “appropriate” (ACR appropriateness criteria) indications for cardiac CT
Indication Patient population Notes
CAD risk stratification (Ca scoring) Asymptomatic patient with low to intermediate
risk with strong family history
Preoperative assessment of CAD risk Asymptomatic patients with intermediate
(coronary CTA) probability for CAD, undergoing major
noncardiac or valve surgery
Acute chest pain Low to intermediate probability with negative or In patients with low clinical suspicion of acute
equivocal initial cardiac enzymes and aortic syndrome or pulmonary embolism, a
nondiagnostic/equivocal ECG or serial troponins cardiac CTA is preferred over an entire chest
and ECG still negative or borderline for triple-rule-out
NSTEMI/ACS
Chronic chest pain Low-, intermediate-, or high-probability patients
(especially following a negative or inconclusive
ECG exercise stress or discordant stress ECG
and stress imaging results)
Post-­revascularization Symptomatic (ischemic equivalent) patients Whole chest should be covered to include
after CABG to assess graft patency or location if origins of internal mammary arteries
graft markers not present
Asymptomatic patients with prior left main stent
or stents ≥3 mm
Planning of open heart surgery, Minimally invasive valve surgeries
(coronary, valve, or congenital) Any redo surgery (retrosternal anatomy)
Native or prosthetic valve disease Clinically suspected valve dysfunction with CT occasionally may act as an arbitrator if CT
inadequate images from other noninvasive and ECHO yield conflicting results
methods
Pre-procedural evaluation for Most commonly in patients with severe aortic The indication may soon be expanded even to
transcatheter valve placement stenosis who are poor surgical candidates patients who would be good surgical candidates
Mitral, tricuspid, and pulmonic valve disease Note CT is a one-stop shop also for assessing
(native or bioprosthetic valves) may also be vascular access, which could be imaged using
good candidates the same contrast bolus
Known or suspected congenital heart Patients where echocardiography was Scanning of the entire chest is likely warranted
disease inconclusive or raised additional questions to image all great vessels for potential anomalies
New heart failure (systolic or Patients where echocardiography was Adjust injection protocol (longer bolus) when
diastolic), assessment of left or right inconclusive and CAD risk is low to right heart function assessment is needed
heart function intermediate
Suspected cardiac mass/thrombus Inadequate images from other methods or Precontrast and delayed scans should also be
contraindications for MRI obtained
Pericardial abnormality Congenital or acquired diseases of the
pericardium
Arrhythmia/EP procedure, PM/ICD Suspected ARVD/ARVC
placement Planning for RFA – pulmonary vein anatomy
Planning RFA to treat potential sources for VF/
VTACH
Coronary venous mapping

Table 21.2  Advantages and challenges of cardiac CT

Advantage Challenge Solution on the horizon?
Spatial resolution High spatial resolution Spatial resolution still inferior Decreasing detector size
to invasive angiography
Contrast resolution Depending on body habitus, Relatively low soft tissue Soft tissue contrast improved by iodine
low-Kv scanning provides contrast resolution mapping with dual energy, ECVF, spectral CT
improved soft tissue contrast
even at low iodine doses
Temporal resolution Low effective temporal Dual source scanners offer highest temporal
resolution, dependent on resolution to date
scanner type and scanning mode
Anatomic and Excellent depiction of anatomy Less information gained on CT-FFR, CT myocardial perfusion
physiological physiology and function
information gained
Function Can assess gross systolic Relatively low temporal Improvements in gantry rotation time and other
function, wall motion, chamber resolution, cannot assess flow, means of increasing temporal resolution (e.g.,
volumes, and ejection fraction regurgitation, velocity, or dual-source techniques) may make CT
diastolic function competitive for function assessment in the future
21  Patient Selection: When to Use Cardiac CT Versus Other Imaging or Non-imaging Tests 237

Table 21.2 (continued)

Advantage Challenge Solution on the horizon?

Speed Faster and more easily Difficulties with IV access and Scanners with improved temporal and spatial
accessible than MRI, SPECT, need for medications (beta resolution can scan without medications
PET, and even echo (especially blockers, nitrates) may delay
transesophageal) scanning
Radiation Some techniques allow Relatively large breast dose; Low kV scanning
sub-mSv scanning arrhythmic or large patients will
receive higher dose
Contrast Peripheral IV access is Ideally a power-injectable Low kV scanning can help reduce the contrast
sufficient access is needed; severe volume needed
allergies or acute renal
insufficiency can be
contraindications
Equipment cost Typically cheaper than MRI More expensive than gamma
camera or echo machines
Implanted medical Hardware in patient is never a Metallic streak artifact may Streak-artifact reduction algorithms,
hardware, electronics, contraindication somewhat degrade image dual-energy scanning
prior intervention/ quality
surgeries
Patient comfort Claustrophobia is rarely a Need for power injection of Advanced low kV and dual-energy techniques
problem; scan itself lasts only contrast, may lead to can reduce contrast amount needed
seconds, noninvasive extravasation

For a successful cardiac CTA in its current iteration, all pected. Calcium scoring, discussed in detail elsewhere in
one needs is an imaging unit and power-injectable intrave- this book, is most appropriate for intermediate-risk
nous access. With advances in scanner technology, diagnos- asymptomatic patients, in whom it may provide reassur-
tic images can be obtained even in patients with high heart ance of low cardiac risk or, on the other hand, may help
rate and arrhythmias. Although prior surgeries, sternal wires, guide preventive measures and lipid-lowering therapies
stents, prosthetic valves, and implanted devices may cause [13]. Newer, low-voltage tin filtration protocols allow
some artifacts, it is rare that the entire study is nondiagnostic. 75% radiation dose reduction with accurate cardiac risk
Even many bariatric patients can be imaged successfully, as stratification [14].
well as those unable to do breath-holds. An additional indication where CTA may precede or sub-
Cardiac CT can be useful when contrast cannot be admin- stitute other imaging tests is preoperative assessment of car-
istered. ECG-synchronized non-contrast CT is invaluable for diac risk, which can be appropriate in certain patient groups
risk stratification of patients for coronary disease and is also undergoing evaluation for major surgeries (noncardiac sur-
able to identify some acute aortic pathologies (e.g., acute geries or cardiac valve surgeries) [15, 16].
intramural hematoma and sometimes dissection).
Additionally, surgeons and most cardiologists  – even if
imaging is not their primary interest – have become familiar Acute Chest Pain
and comfortable with CT images. This ability of clinicians to
“look for themselves,” especially with web-based intuitive Acute chest pain is one indication, where, following a nor-
and interactive 3D workstations, has further contributed to mal CXR and equivocal initial workup in the emergency
the appeal of this modality. department (vitals, cardiac enzymes, and ECG), a cardiac
CTA (or a whole-chest triple-rule-out) could prove useful for
swift and cost-effective patient management [17, 18]. This
 hen Should Cardiac CT Be the First
W can be referred to as the “early assessment pathway” [5]. The
Advanced Imaging Test? appropriateness of CTA depends on the pretest probability,
which ideally should be low to intermediate, with a TIMI
Risk Stratification risk score of 0 [19]. CTA is also useful in cases where
­ischemic symptoms have resolved hours before presenting to
In very low-risk patients, imaging is likely unwarranted, the emergency room. Even in patients who undergo the
but calcium scoring is an inexpensive, quick, and effective “observational pathway”, and serial ECG and troponins
risk stratification tool. This study can be considered in remain negative or borderline, CTA is considered an appro-
asymptomatic and low-risk patients, in whom – based on priate choice.
lifestyle, family history of premature CAD, or lab results – There is extensive evidence in the literature that CTA
advanced or premature coronary atherosclerosis is sus- offers an outstanding negative predictive value and low-
238 P. S. Suranyi et al.

to intermediate-risk patients can be safely discharged if suspected anginal chest pain [26, 27]. In the United Kingdom,
no or only mild coronary stenosis is found [20]. Moreover, the updated National Institute for Health and Care Excellence
a chest or cardiac CTA will not only visualize coronaries (NICE) guidelines have included coronary CT as a first-line
but is very efficient in pointing the clinician to other test for coronary artery disease, citing its relatively low cost and
sources of chest pain, such as acute aortic pathology, pul- high sensitivity [28]. CTA has been shown to be at least equiva-
monary embolism, lung pathology, chest wall, or gastro- lent in the workup of patients with stable chest pain with good
intestinal (e.g., hiatal hernia, reflux esophagitis) problems outcomes. CTA’s ability to detect nonobstructive coronary dis-
[21]. Admittedly, these findings may lead to increased ease by way of visualizing not only the lumen but also plaques
downstream testing that detracts from the appeal of coro- is very helpful in our efforts trying to understand microvascular
nary CTA use in the ER as shown in the ROMICAT II disease, plaque erosion, and myocardial infarction with no
study [17]. These will be further discussed in a separate obstructive coronary artery (MINOCA) [29].
chapter in this book. A particularly interesting group is competitive athletes, in
When acute aortic syndrome is included in the differen- whom cardiac CT may prove useful to visualize anomalous
tial, ECG-synchronized CTA is preferred over non-gated coronaries or other cardiac pathologies. Indications in this
chest CTA, as ECG gating is crucial when visualizing the patient group would include family history of sudden cardiac
aortic root and ascending aorta, with the added benefit of death, uncertain syncope, post-resuscitation, positive stress
visualizing coronary, aortic valve, and pericardial involve- ECG without symptoms, and equivocal ischemic symptoms
ment, when present [22]. Of note, CTA is the only appropri- with normal stress ECG [30].
ate imaging modality in hemodynamically unstable patients,
regardless of whether or not they had prior aorta interven-
tions, since it is “less invasive” than transesophageal echo. How Can CT Complement Echocardiography?
Some have advocated using a negative Ca score scan
(NPV 97%) as a means to rule out acute coronary syndrome Echocardiography is likely the first choice for cardiologists
in the emergency room (ER) setting [23, 24]. However, this to “screen” a patient for suspected congenital heart disease,
approach is certainly not universally accepted, especially valvular disease, new heart failure (systolic or diastolic), car-
because it is possible to have ruptured noncalcified plaques, diac embolic source, and pericardial abnormalities [31–34].
coronary emboli, thrombosed coronary, or dissected coro- Albeit echocardiography is easily accessible and effective in
naries even in the absence of any calcium. many patients, body habitus and suboptimal acoustic win-
The last thing to consider in the acute setting is what to do dows may prevent us from reaching a conclusive diagnosis
when there are extensive coronary calcifications. The 2010 solely with echo in others. Further, echo has clear shortcom-
Appropriateness Criteria warn us that the use of CTA in ings when it comes to visualization of great arteries and
acutely symptomatic patients with Ca score above 400 is not veins. The right heart is challenging due to its retrosternal
definitively appropriate. The assessment of “uncertain appro- position, the left atrial appendage is faintly seen with trans-
priateness” was rendered due to concerns over poor visual- thoracic approach, and the contents of the pericardium can
ization of the coronaries from beam hardening and look rather ominous, when there is anything in there, aside
“blooming” effects. However, in everyday practice, this is from clear serous fluid. The assessment of regional myocar-
rarely a major problem, especially considering rapidly dial function and the quantification of chamber volumes and
improving scanner technology. Additionally, performing Ca systolic function are known to be operator dependent, sub-
scoring “on the fly” while the patient is in the scanner is not jective, and poorly reproducible when using two-­dimensional
very practical, as this would require the radiologist to be methods. Lastly, echocardiography is rarely able to visualize
physically present for every scan. While extensive coronary the coronaries, and its ability to determine myocardial perfu-
calcium is certainly a drawback in assessing the coronaries, sion and viability is also limited, unless an aggressive stress
it does not completely negate the usefulness of a triple-rule-­ agent like dobutamine is used. Below we will discuss what
out or coronary CTA in finding the source for chest pain. cardiac CT can offer when an initial echo is unable to answer
the clinical question or when echo raises new questions
based on its findings.
Chronic Chest Pain

In the past, guidelines recommended stress ECG to be among  ongenital Heart Disease (CHD) and Great
C
the first to evaluate patients with persistent or chronic chest Vessel Anomalies/Abnormalities
pain, despite the fact that sensitivity and specificity are only a
little better than a coin toss, and its accuracy has been shown to In newborns and adults alike, CT can prove useful in visual-
be poor in women [25]. Growing evidence suggest that coro- izing anatomic abnormalities of the coronaries, the airways,
nary CTA should be the first test when evaluating patients with and the great arteries and veins (anomalous pulmonary
21  Patient Selection: When to Use Cardiac CT Versus Other Imaging or Non-imaging Tests 239

veins, left SVC, etc.), aortopulmonary or other systemic to duit, as well as calcifications, stenoses, aneurysms, and
pulmonary collaterals, and pulmonary arteriovenous mal- nearby structures that may be compromised (e.g., coronar-
formations [35, 36]. Cardiology guidelines emphasize that ies). In addition, the same contrast bolus can be used in some
cardiac CT might be useful in detailed assessment of com- scanners to interrogate also the entire aorta, iliofemoral, and
plex cardiac anatomy, e.g., in unusual VSDs, such as inlet or axillary-subclavian access routes to elucidate where the
apical defects, which may be poorly seen by echocardiogra- cleanest and safest pathway can be found. For further detail
phy [37]. on these, we refer to other chapters of this book.
When undetected in childhood, the suspicion of congeni-
tal heart disease may be raised in adults due to chamber
enlargement or a murmur in the absence of significant valve  ew Systolic or Diastolic Heart Failure
N
disease or visible intracardiac shunt, prompting the cardiolo- and Regional Wall Motion Abnormalities
gist to look for an extracardiac shunt or anomaly.
Although due to its low temporal resolution CT-derived sys- Ischemic heart disease is one of the most common reasons for
tolic function is not as accurate as cine MRI, functional images impaired global or regional cardiac function, which can be
can be used to assess regional and global systolic function to efficiently ruled out or detected with cardiac CT, if the overall
provide reasonably accurate volumes of chamber sizes (both clinician impression about the pretest probability is low to
the left and right heart) in patients in whom MRI is contraindi- moderate. In patients where ARVD/ARVC is suspected and
cated (e.g., unapproved devices or severe claustrophobia). MRI cannot be done, CT can be performed to evaluate right
Chest CTA can be useful in assessment of aortic dimen- ventricular function and areas of fatty infiltration.
sions (degenerative aortic aneurysms, bicuspid valve, or
Marfan-/Ehlers-Danlos-related aneurysms), for which  –
truthfully  – only ECG-synchronized or superfast spiral Cardiac Embolic Source
acquisitions with excellent temporal resolution should be
used as only these techniques provide motion-free, accurate In patients with embolic events (e.g., strokes, splenic infarct,
measurements of the aortic root and ascending aorta [22]. etc.) where transthoracic echocardiography is unrevealing
Cardiac CT is a great tool for postsurgical evaluation and (e.g., poor acoustic windows) or equivocal (e.g., questionable
follow-up (anastomoses, baffles, conduits, homografts, mass) and transesophageal echo is unavailable (e.g., at night in
shunts, patches, and plasties) and also for preoperative an understaffed hospital), contraindicated (e.g., unstable/unco-
assessment prior to repeat or multistage operations. The operative patient), or unsuccessful, cardiac CT is a reasonable
ability to visualize retrosternal anatomy in detail is an next step to visualize intracardiac masses. Of note, it is crucial
important benefit before redo sternotomy to plan a safe that the imager reviews the images immediately after the scan
reentry into the chest. to decide whether a delayed scan is warranted (i.e., question-
able clot in the left atrial appendage). When this is logistically
not possible, a low-dose ECG-­ synchronized delayed scan
Valve Disease should be obtained routinely. If a cardiac neoplasm is sus-
pected, a precontrast scan should also be obtained to assess
When echocardiography is unable to visualize the valves enhancement of the mass, when present. Of the common “nor-
reliably and endocarditis or other acute valve pathology is mal variants” that can look ominous on echo, CT can usually
suspected (e.g., papillary muscle rupture), CTA can be more identify the majority with certainty providing reassurance
efficacious. This is especially useful in prosthetic valves, (lipomatous hypertrophy of the interatrial septum, crista termi-
where artifact from the metallic frames often renders echo nalis, prominent Eustachian valve, accessory appendages).
studies suboptimal. Vegetations, thrombi, pannus, leaflet
malfunction, paravalvular leaks, and aneurysms can be beau-
tifully visualized with CT. Pericardium
Rarely, as a third-line evaluation, when echocardiography
and catheter-based pressure measurements disagree and an CT can visualize the size of a pericardial effusion but is lim-
MRI is contraindicated, CT can act as an arbitrator to mea- ited in the ability to detect hemodynamic significance. CT is
sure valve orifice area [38]. unable to conclusively demonstrate whether constriction is
Preceding minimally invasive mitral and aortic valve sur- physiologically present, since ventricular interdependence
geries, CTA can help in planning the best surgical approach cannot be demonstrated during a breath-hold. Pericardial
and visualizing the complexities of the mitral valve appara- enhancement and thickening can only rarely be seen on early
tus. For planning percutaneous valve placement, CT is arterial phase; therefore it is difficult to decide if active
extremely helpful in visualizing the anatomy of the valve inflammation is present on CTAs; however a delayed pro-
annulus, or the potential landing site in a homograft or con- spectively gated scan may highlight pericardial inflammation.
240 P. S. Suranyi et al.

CT is certainly useful in visualizing the extent and distribu- next step after ECHO, the following questions may remain
tion of calcification in chronic constrictive pericarditis and unanswered even after the CT scan, to which MRI is likely to
may aid in surgical planning for pericardial stripping. When a be able to provide answers:
small pericardial effusion is present, prominent epicardial fat
may give the impression of an intrapericardial mass on echo- –– What is the Qp/Qs if a shunt is present?
cardiograms, which can be easily visualized on CT. In meta- –– What is the peak velocity when valvular, subvalvular, or
static malignancies, pericardial metastases are also usually supravalvular stenosis is noted on CT?
well visualized with cardiac CT. –– What is the regurgitant fraction in valvular
insufficiency?
–– What is the pulmonary blood flow to each of the lungs
Following Cardiac Devices and Cardiac Surgery (balanced or unbalanced)?
–– Is there a clot in poorly opacified parts of the cardiovascu-
Pacemaker/ICD leads, ventricular assist devices, or other lar system, when delayed CT was not obtained to clarify
embolized devices (e.g., IVC filters) can best be visualized areas of mixing or suboptimal contrast enhancement (e.g.,
with CTA. Visualizing retrosternal and great vessel anatomy Fontan shunt, when lower extremity is not injected and
is crucial prior to LVAD implantation. Following LVAD delays were not obtained)?
implantation, CT will be the best modality to visualize –– Is there focal myocardial fibrosis/scarring from prior sur-
thrombus formation in or around the inflow and outflow can- gery or ischemic event that may lead to arrhythmia?
nulae or kinking/compression of the non-radiopaque tubing –– Is there stress-inducible ischemia in the myocardium
connecting to the ascending aorta. when coronary anomalies/stenoses are detected on CT?
Following cardiac or aortic surgery, CT can be useful to –– Depending on the contrast injection protocol, it may be
identify early complications such as anastomotic leaks or challenging to determine right ventricular function
pseudoaneurysms. When sternal osteomyelitis and mediasti- with CT, not to mention that due to its low temporal
nitis are suspected, a gated cardiac scan will provide better resolution, CT-derived ventricular volumes are not as
visualization of paracardiac structures to accurately assess accurate as MRI anyways. Thus, MRI might be needed
the extent of infection, and as a bonus, it also gives superb for accurate assessment of biventricular size and sys-
view of the sternum itself. tolic function.
–– If based on clinical symptoms and CT findings an active
vasculitis is suspected, MRI (T2-weighted imaging,
Should We Get a Cardiac CT or an MRI Next? dynamic contrast enhancement, and diffusion-weighted
imaging) might prove helpful after CTA.
In the United States, between the two major advanced tomo- –– When an aneurysm is found that needs follow-up in
graphic methods, cardiac CT seems to be preferred, more so younger patients, it is prudent to schedule routine follow-
than in other parts of the world. When MRI is unavailable or ­up with MRI rather than CT to avoid radiation. Of note, if
contraindicated (unapproved electronic devices, metallic for- simple sizing is the question, a non-contrast MRA may be
eign bodies, severe claustrophobia), CT is obviously a better sufficient.
choice. For some MRI applications, gadolinium-based con-
trast is crucial, which is contraindicated in pregnancy, end-­ When MRI/MRA is ordered first in CHD patients, there
stage renal failure, or severe allergies to MRI contrast. may be a number of dilemmas raised that cannot be com-
Nonetheless, for many questions detailed in the above para- pletely answered without a CT:
graph, cardiovascular MRI is just as efficacious and offers
ionizing radiation-free assessment. –– Questionable coronary anomaly, coronary fistula, or
In the below paragraphs, we briefly discuss when MRI ­stenosis due to postsurgical changes
should be the next step after CT and, vice versa, when a CT –– Questionable coronary atherosclerosis in older patients
might be indicated following cardiac MRI. –– Calcifications in baffles, patches, shunts, conduits, and
homografts
–– In-stent intimal hyperplasia
 ongenital Heart Disease, Great Vessel
C –– Stent-graft evaluation where MRI was suboptimal
Abnormalities, Valve Disease, –– Incomplete closure with a closure device
and Biventricular Function –– Suspected artificial valve complications (mechanical fail-
ure, clot, pannus, vegetation, small paravalvular aneurysm,
When CHD is suspected based on clinical workup and echo or leak)
findings, or patients have history of CHD and valve disease –– Poorly visualized retrosternal anatomy due to artifact
of great vessel anomalies, and cardiac CT is chosen as the from sternal wires
21  Patient Selection: When to Use Cardiac CT Versus Other Imaging or Non-imaging Tests 241

Cardiac Masses rently only practiced at some institutions (detailed in other

chapters), and the widespread use of these CT techniques has
When evaluating a patient with CT for cardiac mass or not yet been implemented. Ultimately it seems that the best
embolic source, the following issues may arise: diagnostic accuracy will be achieved when anatomic and
functional imaging is combined in some form, providing
–– Visualization of lesions in the right heart (particularly the “hybrid cardiac imaging”  – as shown in a recent meta-­
right atrium) might be compromised due to swirling and analysis by Rizvi et al. [43].
mixing of contrasted and non-opacified blood. A currently hot topic of recanalizing chronic total occlu-
–– If no precontrast scan is available, it may be difficult to sions (CTO) may also benefit from a CTA after MRI estab-
decide if a mass is enhancing. lishes viability of the vessel territory: a CTA might be the
–– Some masses, e.g., fibroma, will only show delayed enhance- next best step to serve as a 3D “map” prior to the recanaliza-
ment and may mimic a thrombus on an arterial phase CT. tion procedure.
–– If a thrombus is suspected on CT but delayed images are Transplant-related coronary disease is another realm
not available, MRI might be helpful in clarification. where CTA may be helpful to rule out epicardial disease, but
MRI may be better at picking up on microvascular disease in
On the other hand, sometimes a mass is detected on the peripheral capillary bed. However, to date, neither MRI
MRI, for which further clarification is needed using car- nor CT is fully reliable in assessing rejection.
diac CTA:

–– Pericardial/paracardiac/mediastinal/pleural/lung mass with  ardiac CT and Nuclear Cardiac Imaging:

C
suspected coronary or myocardial invasion. Friends or Adversaries?
–– Suspected calcifications.
–– Suspected perivalvular lesion/thrombus/vegetation/pan- Nuclear cardiac imaging boasts with extensive historical
nus/aneurysm adjacent to a prosthetic valve. data, tradition, and a vast following. However, SPECT
–– Central line or ICD/pacemaker lead-associated thrombi nuclear imaging has been shown to yield particularly low
are better characterized with CT. sensitivity and specificity in women, compared to men,
seems to struggle demonstrating three-vessel disease and
small subendocardial infarcts, and has no ability to visualize
Pericardial Disease epicardial coronary lumens or plaques. In a recent meta-­
analysis, where invasive angiography and fractional flow
As described above, the strength of CT is its ability to visual- reserve were used as the ground truth, SPECT myocardial
ize calcifications, but differentiating of pericardial thicken- perfusion was outperformed by cardiac CTA, myocardial
ing from dense fluid is often fraught with difficulty as perfusion MRI, and even PET perfusion [44]. In another
pericardial enhancement is only rarely visible with CT. CT is recent meta-analysis, coronary CTA showed higher diagnos-
also unable to generate real-time deep inspiratory imaging, tic performance than exercise ECG or SPECT testing in sta-
which is necessary to detect ventricular interdependence, a ble angina [45].
hallmark of constriction. Further, MRI is the best modality to Since attenuation artifacts have been identified as one of
show pericardial tethering (myocardial tagging) and enhance- the important causes for false positives, many centers have
ment (early and delayed). now implemented an ECG-gated non-contrast CT scan for
attenuation correction and used this scan for concurrent cal-
cium scoring for additional risk stratification, which cer-
Ischemic Heart Disease tainly seems like a viable solution for enhancing the accuracy
of nuclear imaging.
If based on CT findings and clinical workup the thought of Cardiac CT has other measurable advantages, namely,
catheterization and revascularization is entertained, it may be that it is able to expedite discharge and has lower radiation
prudent to obtain a cardiac MRI stress test, where stress-­ exposure and lower cost in the emergency room setting when
induced ischemia and viability can be detected with great compared to the standard workup, which – interestingly – is
accuracy. Growing evidence suggests that MRI is superior to even more pronounced in women vs. men [46]. Additionally,
nuclear cardiac SPECT myocardial perfusion imaging [39, in some institutions, cardiac CT is available 24/7 [47],
40]. It has also been shown that MRI stress is much superior whereas nuclear imaging is offered typically only during
to SPECT in women [41]. The anatomic findings of coronary regular work hours. While a complete stress-rest SPECT
CTA may be complemented by dual-energy iodine mapping, may take hours and even PET perfusion (where available)
CT myocardial perfusion assessment, and CT-FFR for a will take at least an hour, CT can be done in a matter of min-
more functional assessment [42]. However, these are cur- utes. The in-plane spatial resolution of nuclear imaging tests
242 P. S. Suranyi et al.

even with the latest improvements is an order of magnitude best next step in patient management. The question should
lower, around 10–5  mm (SPECT vs. PET), as opposed to not be about one modality lording over another but, rather,
CT’s half a millimeter resolution. how we can integrate cardiac CT best into the comprehen-
As detailed in other chapters, the radiation dose with sive multidisciplinary workup and management of cardio-
cardiac CT has been a downward moving target, and vascular diseases. It is also important to mention that
SPECT certainly is unable to keep up with the race. PET professional society-driven appropriateness criteria cannot
perfusion, however, offers relatively lower radiation doses keep up with the incredible speed of scanner technology
that compete with advanced CT scanners (when function is development and rapid advances in novel cardiac device uti-
acquired with tube current modulation) in the 2–5  mSv lization. Therefore, imagers have to maintain flexibility and
effective dose range. Thus, depending on the scanners used be ready to take cardiac CTA even to previously uncharted
at one’s institution, PET perfusion may be a reasonable territories for the benefit of our patients.
option particularly for women in whom quantitative myo- It seems that cardiac CT is best utilized in an environment
cardial perfusion measurement may reveal ischemic heart where the stakeholders of various modalities communicate
disease where CTA or even conventional angiography can- and collaborate even before a test is ordered, to ensure that a
not find luminal obstruction. Of note, as detailed in another diversity of opinions is considered and optimal image acqui-
chapter, myocardial perfusion CT has also been getting sition is performed, tailored to the specific clinical questions
increasing coverage in the “press.” PET imaging has fur- and suspicions, while also keeping individual patient-­specific
ther advantages over CT, especially if MRI is not available factors in sight. Then, following image acquisition and post-­
in imaging active inflammation, which is particularly use- processing, the team should again reconvene to allow for
ful in sarcoidosis, where CT would be unlikely to detect thorough interpretation involving several subspecialties.
any changes in the heart. This is the only way that cardiac CT will find its rightful
It is not unusual after an equivocal nuclear cardiac scan – place in heart disease management and will provide true
or after a scan when the results are not what the clinician value while reducing cost and minimizing risk.
hoped for – that a cardiologist will order a cardiac CTA, to
conclusively demonstrate the absence or presence of epicar-
dial disease, before the patient is subjected to an invasive
References
procedure. And indeed, CT has been most helpful in exclud-
ing coronary disease when a false-positive nuclear result is 1. Taylor AJ, et  al. ACCF/SCCT/ACR/AHA/ASE/ASNC/NASCI/
suspected, given its high negative predictive value [48]. CT SCAI/SCMR 2010 appropriate use criteria for cardiac computed
has also proven to be helpful in visualizing epicardial coro- tomography. A report of the American College of Cardiology
Foundation Appropriate Use Criteria Task Force, the Society of
nary disease, positive remodeling, and noncalcified plaques Cardiovascular Computed Tomography, the American College of
in patients where perfusion defects were absent, but the per- Radiology, the American Heart Association, the American Society
sistent clinical signs and symptoms prompted the clinician to of Echocardiography, the American Society of Nuclear Cardiology,
pursue the disease with further imaging (i.e., clinically the North American Society for Cardiovascular Imaging, the
Society for Cardiovascular Angiography and Interventions, and
deemed false negatives). In the recent PROMISE trial, com- the Society for Cardiovascular Magnetic Resonance. J Cardiovasc
paring coronary CTA to functional testing, CTA has proven Comput Tomogr. 2010;4(6):407 e1–33.
to be a significantly better prognostic indicator than the alter- 2. Expert Panel on Cardiac, I, et al. ACR appropriateness criteria(R)
natives, including exercise electrocardiography, exercise or known or suspected congenital heart disease in the adult. J Am Coll
Radiol. 2017;14(5S):S166–76.
pharmacological nuclear myocardial perfusion imaging, and 3. Expert Panel on Cardiac, I, et al. ACR appropriateness criteria(R)
exercise or pharmacological stress echocardiography [49]. chronic chest pain-high probability of coronary artery disease. J
This, again, is particularly important in women, in whom Am Coll Radiol. 2017;14(5S):S71–80.
nonobstructive coronary disease (or as some would prefer 4. Expert Panel on Cardiac, I, et  al. ACR appropriateness
criteria(R) dyspnea-suspected cardiac origin. J Am Coll Radiol.
ischemic heart disease) is more prevalent than in men. 2017;14(5S):S127–37.
5. Emergency Department Patients With Chest Pain Writing, P,
et  al. 2015 ACR/ACC/AHA/AATS/ACEP/ASNC/NASCI/SAEM/
SCCT/SCMR/SCPC/SNMMI/STR/STS appropriate utilization
 onclusion: Comprehensive Care,
C of cardiovascular imaging in emergency department patients with
Collaboration, Communication, chest pain: a joint document of the American College of Radiology
and Coexistence Appropriateness Criteria Committee and the American College of
Cardiology Appropriate Use Criteria Task Force. J Am Coll Radiol.
2016;13(2):e1–e29.
There is obviously no simple algorithm for deciding when 6. Hoffmann U, et al. ACR appropriateness criteria acute nonspecific
and how cardiac CT should be used. One has to consider risk chest pain-low probability of coronary artery disease. J Am Coll
factors, pretest probability, clinical presentation, and other Radiol. 2015;12(12 Pt A):1266–71.
7. White RD, et  al. 2013 ACCF/ACR/ASE/ASNC/SCCT/SCMR
noninvasive test results to decide whether cardiac CTA is the appropriate utilization of cardiovascular imaging in heart failure:
21  Patient Selection: When to Use Cardiac CT Versus Other Imaging or Non-imaging Tests 243

an executive summary: a joint report of the ACR Appropriateness artery disease: systematic review and meta-analysis. Ann Emerg
Criteria (R) Committee and the ACCF Appropriate Use Criteria Med. 2016;68(6):659–70.
Task Force. J Am Coll Radiol. 2013;10(7):493–500. 25. Kwok Y, et al. Meta-analysis of exercise testing to detect coronary
8. Dill KE, et al. ACR appropriateness criteria imaging for transcatheter artery disease in women. Am J Cardiol. 1999;83(5):660–6.
aortic valve replacement. J Am Coll Radiol. 2013;10(12):957–65. 26. Moss AJ, Newby DE. CT coronary angiographic evaluation of sus-
9. Earls JP, et al. ACR appropriateness criteria asymptomatic patient at pected anginal chest pain. Heart. 2016;102(4):263–8.
risk for coronary artery disease. J Am Coll Radiol. 2014;11(1):12–9. 27.
Investigators, S.-H.  CT coronary angiography in patients
10. Mammen L, et  al. ACR appropriateness criteria(R) nonischemic with suspected angina due to coronary heart disease (SCOT-­
myocardial disease with clinical manifestations (ischemic cardio- HEART): an open-label, parallel-group, multicentre trial. Lancet.
myopathy already excluded). J Thorac Imaging. 2014;29(4):W44–7. 2015;385(9985):2383–91.
11. Meinel FG, et al. ECG-synchronized CT angiography in 324 con- 28. Moss AJ, et  al. The updated NICE guidelines: cardiac CT as the
secutive pediatric patients: spectrum of indications and trends in first-line test for coronary artery disease. Curr Cardiovasc Imaging
radiation dose. Pediatr Cardiol. 2015;36(3):569–78. Rep. 2017;10(5):15.
12. Meinel FG, et al. Radiation risks from cardiovascular imaging tests. 29. Baldassarre LA, et  al. Noninvasive imaging to evaluate women
Circulation. 2014;130(5):442–5. with stable ischemic heart disease. JACC Cardiovasc Imaging.
13. Vliegenthart R, Morris PB.  Computed tomography coronary
2016;9(4):421–35.
artery calcium scoring: review of evidence base and cost-­ 30. Grani C, et al. Multimodality imaging in individuals with anomalous
effectiveness in cardiovascular risk prediction. J Thorac Imaging. coronary arteries. JACC Cardiovasc Imaging. 2017;10(4):471–81.
2012;27(5):296–303. 31. Baumgartner H, et al. Echocardiographic assessment of valve ste-
14. Tesche C, et al. Accuracy and radiation dose reduction using low-­ nosis: EAE/ASE recommendations for clinical practice. J Am Soc
voltage computed tomography coronary artery calcium scoring Echocardiogr. 2009;22(1):1–23. quiz 101-2
with tin filtration. Am J Cardiol. 2017;119(4):675–80. 32. Lang RM, et  al. Recommendations for chamber quantification: a
15. Hwang JW, et al. Assessment of perioperative cardiac risk of patients report from the American Society of Echocardiography’s Guidelines
undergoing noncardiac surgery using coronary computed tomo- and Standards Committee and the Chamber Quantification Writing
graphic angiography. Circ Cardiovasc Imaging. 2015;8(3):e002582. Group, developed in conjunction with the European Association of
16. Morris MF, et al. Computed tomography as an alternative to cath- Echocardiography, a branch of the European Society of Cardiology.
eter angiography prior to robotic mitral valve repair. Ann Thorac J Am Soc Echocardiogr. 2005;18(12):1440–63.
Surg. 2013;95(4):1354–9. 33. Group, T.T.A.C.W, et  al. ACCF/ASE/ACEP/ASNC/SCAI/
17. Hoffmann U, et al. Coronary CT angiography versus standard eval- SCCT/SCMR 2007 appropriateness criteria for transthoracic
uation in acute chest pain. N Engl J Med. 2012;367(4):299–308. and transesophageal echocardiography: a report of the American
18. Truong QA, et al. Coronary CT angiography versus standard emer- College of Cardiology Foundation Quality Strategic Directions
gency department evaluation for acute chest pain and diabetic Committee Appropriateness Criteria Working Group, American
patients: is there benefit with early coronary CT angiography? Society of Echocardiography, American College of Emergency
Results of the Randomized Comparative Effectiveness ROMICAT Physicians, American Society of Nuclear Cardiology, Society
II Trial. J Am Heart Assoc. 2016;5(3):e003137. for Cardiovascular Angiography and Interventions, Society of
19. Rybicki FJ, et  al. 2015 ACR/ACC/AHA/AATS/ACEP/ASNC/
Cardiovascular Computed Tomography, and the Society for
NASCI/SAEM/SCCT/SCMR/SCPC/SNMMI/STR/ STS appro- Cardiovascular Magnetic Resonance. Endorsed by the American
priate utilization of cardiovascular imaging in emergency depart- College of Chest Physicians and the Society of Critical Care
ment patients with chest pain: a joint document of the American Medicine. J Am Soc Echocardiogr. 2007;20(7):787–805.
College of Radiology Appropriateness Criteria Committee and the 34. American College of Cardiology Foundation Appropriate Use

American College of Cardiology Appropriate Use Criteria Task Criteria Task, F, et  al. ­ ACCF/ASE/AHA/ASNC/HFSA/HRS/
Force. J Am Coll Cardiol. 2016;67(7):853–79. SCAI/SCCM/SCCT/SCMR 2011 appropriate use criteria for echo-
20. El-Hayek G, et  al. Meta-analysis of coronary computed tomog- cardiography. A report of the American College of Cardiology
raphy angiography versus standard of care strategy for the evalu- Foundation Appropriate Use Criteria Task Force, American Society
ation of low risk chest pain: are randomized controlled trials of Echocardiography, American Heart Association, American
and cohort studies showing the same evidence? Int J Cardiol. Society of Nuclear Cardiology, Heart Failure Society of America,
2014;177(1):238–45. Heart Rhythm Society, Society for Cardiovascular Angiography
21. Lee HY, Yoo SM, White CS. Coronary CT angiography in emer- and Interventions, Society of Critical Care Medicine, Society of
gency department patients with acute chest pain: triple rule-out pro- Cardiovascular Computed Tomography, Society for Cardiovascular
tocol versus dedicated coronary CT angiography. Int J Cardiovasc Magnetic Resonance American College of chest physicians. J Am
Imaging. 2009;25(3):319–26. Soc Echocardiogr. 2011;24(3):229–67.
22. Hiratzka LF, et  al. 2010 ACCF/AHA/AATS/ACR/ASA/SCA/
35. Han BK, et al. Computed tomography imaging in patients with con-
SCAI/SIR/STS/SVM Guidelines for the diagnosis and management genital heart disease part I: rationale and utility. An expert consensus
of patients with thoracic aortic disease. A Report of the American document of the Society of Cardiovascular Computed Tomography
College of Cardiology Foundation/American Heart Association (SCCT): Endorsed by the Society of Pediatric Radiology (SPR)
Task Force on Practice Guidelines, American Association for and the North American Society of Cardiac Imaging (NASCI). J
Thoracic Surgery, American College of Radiology, American Cardiovasc Comput Tomogr. 2015;9(6):475–92.
Stroke Association, Society of Cardiovascular Anesthesiologists, 36.
Hlavacek AM.  Imaging of congenital cardiovascular dis-
Society for Cardiovascular Angiography and Interventions, ease: the case for computed tomography. J Thorac Imaging.
Society of Interventional Radiology, Society of Thoracic 2010;25(3):247–55.
Surgeons, and Society for Vascular Medicine. J Am Coll Cardiol. 37. Warnes CA, et  al. ACC/AHA 2008 guidelines for the manage-
2010;55(14):e27–e129. ment of adults with congenital heart disease: a report of the
23. Hinzpeter R, Higashigaito K, Morsbach F, et  al. Coronary artery American College of Cardiology/American Heart Association
calcium scoring for ruling-out acute coronary syndrome in chest Task Force on Practice Guidelines (writing committee to develop
pain CT. Am J Emerg Med. 2017;35(10):1565–7. guidelines on the management of adults with congenital heart dis-
24. Chaikriangkrai K, et  al. Prognostic value of coronary artery cal- ease). Developed in collaboration with the American Society of
cium score in acute chest pain patients without known coronary Echocardiography, Heart Rhythm Society, International Society
244 P. S. Suranyi et al.

for Adult Congenital Heart Disease, Society for Cardiovascular artery disease compared with stand-alone coronary computed
Angiography and Interventions, and Society of Thoracic Surgeons. tomography angiography: a Meta-analysis. JACC Cardiovasc
J Am Coll Cardiol. 2008;52(23):e143–263. Imaging. 2018;11(4):589–99.
38. Li X, et  al. Aortic valves stenosis and regurgitation: assessment 44. Takx RA, et  al. Diagnostic accuracy of stress myocardial perfu-
with dual source computed tomography. Int J Cardiovasc Imaging. sion imaging compared to invasive coronary angiography with
2009;25(6):591–600. fractional flow reserve meta-analysis. Circ Cardiovasc Imaging.
39. Schwitter J, et  al. Superior diagnostic performance of perfusion-­ 2015;8(1):e002666.
cardiovascular magnetic resonance versus SPECT to detect coro- 45. Nielsen LH, et  al. The diagnostic accuracy and outcomes after
nary artery disease: the secondary endpoints of the multicenter coronary computed tomography angiography vs. conventional
multivendor MR-IMPACT II (Magnetic Resonance Imaging for functional testing in patients with stable angina pectoris: a system-
Myocardial Perfusion Assessment in Coronary Artery Disease atic review and meta-analysis. Eur Heart J Cardiovasc Imaging.
Trial). J Cardiovasc Magn Reson. 2012;14:61. 2014;15(9):961–71.
40. de Jong MC, et  al. Diagnostic performance of stress myocardial 46. Truong QA, et al. Sex differences in the effectiveness of early coro-
perfusion imaging for coronary artery disease: a systematic review nary computed tomographic angiography compared with standard
and meta-analysis. Eur Radiol. 2012;22(9):1881–95. emergency department evaluation for acute chest pain: the rule-­
41. Greenwood JP, et  al. Comparison of cardiovascular magnetic
out myocardial infarction with Computer-Assisted Tomography
resonance and single-photon emission computed tomography in (ROMICAT)-II Trial. Circulation. 2013;127(25):2494–502.
women with suspected coronary artery disease from the Clinical 47. Garrett KG, et al. Residents’ performance in the interpretation of
Evaluation of Magnetic Resonance Imaging in Coronary Heart on-call “triple-rule-out” CT studies in patients with acute chest
Disease (CE-MARC) Trial. Circulation. 2014;129(10):1129–38. pain. Acad Radiol. 2014;21(7):938–44.
42. Gonzalez JA, et  al. Meta-analysis of diagnostic performance of 48. Chong FY, et al. Negative coronary CT angiography for chest pain
coronary computed tomography angiography, computed tomogra- assessment predicts low event rate in 5 years. J Med Imaging Radiat
phy perfusion, and computed tomography-fractional flow reserve Oncol. 2012;56(1):55–7.
in functional myocardial ischemia assessment versus invasive frac- 49. Hoffmann U, et  al. Prognostic value of noninvasive cardiovas-
tional flow reserve. Am J Cardiol. 2015;116(9):1469–78. cular testing in patients with stable chest pain: insights from the
43. Rizvi A, Han D, Danad I, et al. Diagnostic performance of hybrid PROMISE trial (prospective multicenter imaging study for evalua-
cardiac imaging methods for assessment of obstructive coronary tion of chest pain). Circulation. 2017;135(24):2320–32.
 Current Guidelines
22
Kweku Appau and Arthur E. Stillman

In an effort to improve quality, various societies across the document (CAD-RADS) that not only provides the severity
globe have developed appropriate use criteria and guidelines of coronary artery disease (CAD) as detected by cardiac CT
for the use of cardiac CT. Imaging guidelines are generally but also attempts to guide both patients and ordering physi-
incorporated into various disease conditions or clinical sce- cians on the appropriate interpretation and management of
narios [1–8] and include qualifications of interpreting physi- patients based on the degree of CAD stenosis identified on
cians, equipment standards, and drugs used for heart rate cardiac CT (Tables 22.3 and 22.4) [15].
control and vasodilation [9, 10]. There are two major sources The ACR and the ACCF have critically evaluated cardiac
of guidelines in the USA, the American College of Radiology imaging modalities to guide patients and healthcare provid-
(ACR) and the American College of Cardiology Foundation ers on the utility of the available cardiac imaging procedures
(ACCF), that discuss the training expectations of physicians in terms of which are appropriate (A), maybe appropriate
for interpretation of cardiac CT. The ACR focuses on radiol- (M), or rarely appropriate (R) depending on various cardiac
ogy training requirements [9], while the ACCF is more conditions. There are guidelines which discuss the use of
focused on cardiologists [11]. The Society of Cardiovascular imaging for stable coronary heart disease and heart failure
Computed Tomography (SCCT) has incorporated the train- [2, 5, 7, 8]. These recommendations are summarized in
ing requirements of the ACCF and encourages board certifi- Tables 22.4, 22.5, 22.6, 22.7, 22.8, 22.9, 22.10, 22.11, 22.12,
cation through the Certification Board of Cardiovascular 22.13, 22.14, 22.15, 22.16, 22.17, 22.18, 22.19, 22.20, 22.21,
Computed Tomography (CBCCT) [10]. The ACR offers a 22.22, 22.23, 22.24, 22.25, 22.26, 22.27, 22.28, 22.29, 22.30,
Certificate of Advanced Proficiency (CoAP) after passing a 22.31, 22.32, 22.33, and 22.34.
rigorous examination [12]. This has greater requirements for There is a strong consensus agreement across the globe
experience in interpreting cardiac CT than the minimum with respect to class II and class III indications for the use
ACR requirements. The European Board of Cardiac of cardiac CT [2, 4, 5, 7, 8, 17–19]. Most of the guidelines
Radiology and the British Society of Cardiovascular Imaging stress that there are a class II indication for the use of car-
(BSCI) also provide a diploma confirming competence in diac CT in patients at low to intermediate risk for CAD and
performing and interpreting cardiac CT [13, 14]. Tables 22.1 a class III indication for patients at high risk. Class I indica-
and 22.2 show the differences and similarities between the tions for the use of cardiac CT are very limited except in
ACR and the ACCF training requirement guidelines. The Brazil where there a class I indication for the use of cardiac
ACR and SCCT have helped immensely by providing com- CT to detect coronary calcium in asymptomatic patients
prehensive guidelines on optimized image acquisition, image with intermediate risk for CAD [1]. There is also strong
processing, as well as image interpretation and reporting [9, consensus agreement among the various cardiac imaging
15, 16]. Recently, both the ACR and SCCT leaders in con- and cardiovascular societies when it comes to appropriate
junction with the North American Society for Cardiovascular and rarely use criteria for cardiac CT [2, 3, 5, 17, 19–21].
Imaging (NASCI) have developed a standardized reporting The gray area lies in the maybe appropriate use criteria
where clinical judgment (just as in any patient encounter) is
K. Appau strongly advised. In South Korea and most Asian countries,
Department of Cardiology, Emory University School of Medicine, the cardiovascular societies tend to rely heavily on appropri-
Atlanta, GA, USA ate use criteria alone [19, 20], whereas Western counterparts
A. E. Stillman (*) utilize both appropriate use criteria and guidelines on the
Department of Radiology and Imaging Science, Emory University use of cardiac CT [2, 3, 5, 7, 8, 22]. Currently, most other
School of Medicine, Atlanta, GA, USA
countries across the world tend to utilize the ACCF and the
e-mail: [email protected]

© Humana Press 2019 245

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_22
246 K. Appau and A. E. Stillman

Table 22.1  Qualifications and responsibilities of personnel: ACR–NASCI–SPR practice parameter for the performance and interpretation of
cardiac computed tomography
Cardiac CT procedures must be supervised and The supervising physician must also Additional qualifications include:
interpreted by a physician with one of the have the following:
following qualifications below:
Certification in radiology or diagnostic radiology  (a) The physician should have Cardiac CT Category I CME or training in
by the: documented training in the cardiac CT in a training program approved by:
 (a) American Board of Radiology physics of diagnostic radiology  (a) The Accreditation Council for Graduate
 (b) American Osteopathic Board of Radiology  (b) The physician must be familiar Medical Education (ACGME)
 (c) Royal College of Physicians and Surgeons of with:  (b) The Royal College of Physicians and
Canada    (a) The principles of radiation Surgeons of Canada (RCPSC)
 (d) Collège des Médecins du Québec protection  (c) The Collège des Médecins du Québec
 (e) Involvement with the supervision,    (b) The hazards of radiation  (d) The American Osteopathic Association
interpretation, and reporting of 300 CT    (c) Radiation monitoring (AOA)
examinations in the past 36 months requirements as they apply to
both patients and personnel
Completion of a diagnostic radiology residency The physician should be thoroughly Education in cardiac anatomy, physiology,
program approved by: acquainted with the many morphologic pathology, and cardiac CT imaging for a time
 (a) The Accreditation Council for Graduate and pathophysiologic manifestations equivalent to at least 30 h of CME
Medical Education (ACGME) of artifacts demonstrated on CT.
 (b) The Royal College of Physicians and Additionally, supervising physicians
Surgeons of Canada (RCPSC) should have appropriate knowledge of
 (c) The Collège des Médecins du Québec alternative imaging methods including
 (d) The American Osteopathic Association the use of and indications for general
(AOA) radiography and specialized studies
 (e) Involvement with the supervision, such as:
interpretation, and reporting of 500 CT  (a) Angiography
examinations in the past 36 months  (b) Ultrasonography
 (c) Magnetic resonance imaging
 (d) Nuclear medicine studies
Physicians not board certified in radiology or not The physician should be familiar with Supervision, interpretation, or reporting of at
trained in a diagnostic radiology residency patient preparation for the examination least 50 cardiac CT examinations in the last
program who assume the responsibilities for CT The physician must have training in 36 months. Coronary artery calcium scoring
imaging exclusively in a specific anatomical area the recognition and treatment of does not qualify as meeting these requirements
should meet the following criteria: adverse effects of contrast materials
 (a) Completion of an ACGME-­approved used for these studies (see the ACR
residency program in the specialty practiced Manual on Contrast Media and the
plus 200 h of Category I CME in the ACR–SPR Practice Parameter for the
performance and interpretation of CT in the Use of Intravascular Contrast Media)
subspecialty where CT reading occurs
 (b) Supervision, interpretation, and reporting of
500 cases in that subspecialty area during the
past 36 months in a supervised situation
Completion of an accredited radiology residency The physician must have the For any physician or any other physician who
in the past 24 months will be presumed to be responsibility for: assumes responsibilities for cardiac CT
satisfactory experience for the reporting and  (a) Reviewing all indications for the imaging, additional qualifications should
interpreting requirement examination include:
 (b) Specifying the use, dosage, and  (a) Completion of an ACGME-approved
rate of administration of contrast training program in the specialty
agents practice plus 200 h of Category I CME
 (c) Specifying the imaging in the performance and interpretation of
technique, including available CT in the subspecialty where CT
techniques to reduce radiation reading occurs
dose
 (d) Interpreting images
 (e) Generating official/final
interpretation reports
 (f) Maintaining the quality of the
images and the interpretations
22  Current Guidelines 247

Table 22.1 (continued)

The physician’s continuing medical Supervision, interpretation, or reporting of

education should be in accordance 500 cases in cardiothoracic imaging that
with the ACR Practice Parameter for include at least 50 cardiac CT examinations
Continuing Medical Education (CME) during the past 36 months in a supervised
of 150 h of approved education every situation and at least 450 additional thoracic
3 years and should include CME in CT or thoracic CT angiography cases
cardiac CT as is appropriate to the Coronary artery calcium scoring does not
physician’s practice needs qualify as meeting these requirements
Completion of at least 30 h of Category I
CME in cardiac imaging, including cardiac
CT, anatomy, physiology, and/or pathology, or
documented equivalent supervised experience
in a facility actively performing cardiac CT
Data from ACR–NASCI–SPR practice parameter for the performance and interpretation of cardiac computed tomography (CT) [9]

Table 22.2  Qualifications and responsibilities of personnel (ACCF: COCATS 4 Task Force 7: training in cardiovascular computed tomographic
imaging)
It is required that at least 1 physician should have a It is recommended that interpretation physicians must have:
minimum of 2 years of clinical experience and/or 300  (a) CBCCT certification
scan interpretation of coronary CTA  (b) ACR board certification
 (c) Dedicated fellowship training in advanced cardiac imaging
It is required that all other physicians must maintain level Cardiac CT fellows must be familiar with cardiovascular:
2 coronary CTA certification or equivalent  (a) Anatomy
 (b) Physiology
 (c) Pathophysiology
Interpreting physicians must be trained in the best-practice Fellows must also understand the:
protocol selection of the scanner(s) in use  (a) Clinical application of cardiac CT
 (b) Principles of cardiac CT physics
 (c) Radiation generation/exposure
A qualified physician must interpret the non-cardiac To be eligible to sit for the CBCCT examination:
anatomy on all scans either as the primary reader or as a  (a) US-trained cardiovascular fellows must have undergone training in a
consulting physician program accredited by the ACGME
 (b) Meet Level II training requirements
Cardiac CT fellows must be familiar with the protocols: Every cardiovascular fellow should develop familiarity with CCT:
 (a) For contrast administration and contrast kinetics  (a) Technical performance
 (b) The potential adverse events resulting from contrast  (b) Strengths
exposure and appropriate treatment  (c) Limitations of CCT
 (d) Gain an understanding of how to effectively use the information provided
by CCT
Level II training requirements: Level III training requirements (in addition to Level II):
 (a) Be present in the scanning suite control room and  (a) Additional training and experience beyond the cardiovascular fellowship
actively participate in the acquisition of 50 cases  (b) Acquire specialized knowledge and experience in performing, interpreting,
 (b) View a maximum of 50 cases from an educational and training others to perform specific procedures or render advanced
CD or presentation granting continuing medical specialized care at a high level of skill
education credit that contains CCT data review  (c) Be involved in the acquisition and interpretation of imaging examinations
 (c) 150 cases must involve interactive manipulation of  (d) Participation in research, teaching, and the administrative aspects of
reconstructed datasets using a three-dimensional laboratory operations
imaging workstation  (e) Data management, report generation and distribution, quality
 (d) 20 cases must include evaluation of cardiac function improvement, and accreditation as well as development of an
 (e) 20 cases should involve evaluation of structural and/ understanding of evolving multimodality imaging technologies
or congenital heart disease
 (f) 15 cases must involve evaluation of bypass graft
vessels
 (g) 40 cases should be correlated with invasive
angiography and/or myocardial perfusion imaging
 (h) Be actively involved in acquisition, interpretation,
and reporting of 50 cases CCT images
Data from Garcia et al. [11]
248 K. Appau and A. E. Stillman

Table 22.3  Equipment specifications (ACR–NASCI–SPR practice parameter for the performance and interpretation of cardiac computed
­tomography [9])
For diagnostic-quality cardiac CT, the CT scanner should meet or exceed the following specifications:
ECG synchronization for all scans, with the ability to perform Non-contrast-enhanced MDCT for coronary artery calcium scoring
prospective triggering and retrospective gating may be adequately performed on a scanner with a temporal
resolution of 0.50 s using prospectively ECG-­triggered “step-and-
shoot” sequential acquisition
Setup for bolus tracking of the administered contrast material for Minimum section thickness should be:
appropriate timing of contrast-enhanced cardiac CT exams  (a) ≤5 mm
 (b) ≤3 mm for coronary calcium scoring
 (c) ≤1.5 mm for CT coronary arteriography
Automated tube modulation during image acquisition for dose Volumetric computed tomography dose index (CTDI vol) and dose
reduction length product (DLP) must be available after each scan for transfer
to individual PACS workstations
Contrast-enhanced cardiac CT by MDCT (should meet or exceed a
64-detector scanner), including:
 (a) CT coronary arteriography
 (b) A scanner capable of achieving in-plane spatial resolution
≤0.5 × 0.5 mm axial
 (c) z-axis spatial resolution ≤1 mm longitudinal
 (d) Temporal resolution ≤0.25 s
Data from ACR–NASCI–SPR practice parameter for the performance and interpretation of cardiac computed tomography (CT) [9]

Table 22.4  CAD-RADS reporting and data system for patients presenting with stable chest pain
Corresponding Recommendation on
Degree of maximal CAD-RAD further cardiac
coronary stenosis classification Interpretation investigation Recommendation on management
0% (no plaque or CAD-RADS 0 Documented None Reassurance. Consider non-atherosclerotic
stenosis) absence of CAD causes of chest pain
1–24% minimal stenosis CAD-RADS 1 Minimal None Consider non-atherosclerotic causes of chest
or plaque with no non-obstructive pain
stenosis CAD Consider preventive therapy and risk factor
modification
25–49% mild stenosis CAD-RADS 2 Mild non-­ None Consider non- atherosclerotic causes of chest
obstructive CAD pain
Consider preventive therapy and risk factor
modification particularly for patients with
non-obstructive plaque in multiple segments
50–69% stenosis CAD-RADS 3 Moderate stenosis Consider functional Consider symptom-guided anti-ischemic and
assessment preventive pharmacotherapy as well as risk
factor modification per guideline-directed
care
Other treatments should be considered per
guideline-directed care
A: 70–99% stenosis CAD-RADS 4 Severe stenosis A: Consider ICA or Consider symptom-guided anti-ischemic and
or functional preventive pharmacotherapy as well as risk
B: Left main >50% or assessment factor modification per guideline-directed
3- vessel obstructive B: ICA is care
(≥70%) disease recommended Other treatments (including options of
revascularization) should be considered per
guideline-directed care
100% (total occlusion) CAD-RADS 5 Total coronary Consider ICA and/or Consider symptom-guided anti-ischemic and
occlusion viability preventive pharmacotherapy as well as risk
assessment factors modification per guideline-directed
care
Other treatments (including options of
revascularization) should be considered per
guideline-directed care
Nondiagnostic study CAD-RADS N Obstructive CAD Additional or
cannot be alternative evaluation
excluded may be needed
Data from Cury et al. [15]
22  Current Guidelines 249

Table 22.5  CAD-RADS reporting and data system for patients presenting with acute chest pain, negative first troponin, negative or nondiagnostic
electrocardiogram, and low to intermediate risk (TIMI risk score <4) (emergency department or hospital setting)
Corresponding
Degree of maximal CAD-RAD
coronary stenosis classification Interpretation Recommendation on management
0% CAD-RADS 0 ACS highly No further evaluation of ACS is required
unlikely Consider other etiologies
1–24% CAD-RADS 1 ACS highly Consider evaluation of non-ACS etiology, if normal troponin and no
unlikely ECG changes
Consider referral for outpatient follow-up for preventive therapy and
risk factor modification
25–49% CAD-RADS 2 ACS unlikely Consider evaluation of non-ACS etiology, if normal troponin and no
ECG changes
Consider referral for outpatient follow-up for preventive therapy and
risk factor modification
If clinical suspicion of ACS is high or if high-risk plaque features are
noted, consider hospital admission with cardiology consultation
50–69% CAD-RADS 3 ACS possible Consider hospital admission with cardiology consultation, functional
testing, and/or ICA for evaluation and management
Recommendation for anti-ischemic and preventive management should
be considered as well as risk factor modification. Other treatments
should be considered if presence of hemodynamically significant lesion
A:70–99% CAD-RADS 4 ACS likely Consider hospital admission with cardiology consultation. Further
B: Left main >50% or evaluation with ICA and revascularization as appropriate
three-vessel obstructive Recommendation for anti-ischemic and preventive management should
disease be considered as well as risk factor modification
100% (total occlusion) CAD-RADS 5 ACS very Consider expedited ICA on a timely basis and revascularization if
likely appropriate if acute occlusion
Recommendation for anti-ischemic and preventive management should
be considered as well as risk factor modifications
Nondiagnostic study CAD-RADS N ACS cannot be Additional or alternative evaluation for ACS is needed
excluded
Data from Cury et al. [15]

Table 22.6  Recommendation on advance imaging in patients with suspected SIHD who require noninvasive testing (ACCF/AHA/ACP/AATS/
PCNA/SCAI/STS)
Indications Class Level of evidence
CCTA may be reasonable for patients with intermediate pretest probability of IHD IIb B
CCTA is reasonable for patients with low to intermediate pretest probability of IHD who are unable to IIa B
perform moderate physical functioning or have disabling comorbidity
CCTA is reasonable for patients with an intermediate pretest probability of IHD who: IIa C
 (a) Continue to have symptoms despite normal prior test results
 (b) Have inconclusive results from prior ETT/pharm stress
 (c) Are unable to undergo stress nuclear MPI or stress echo
For patients with low to intermediate pretest probability of obstructive IHD, non-contrast cardiac CT to IIb C
determine CAC score may be considered
In asymptomatic adults with diabetes, 40 years of age and older, measurement of CAC is reasonable for IIa B
cardiovascular risk assessment
Measurement of CAC may be reasonable for cardiovascular risk assessment in persons at low to IIb B
intermediate risk (6–10% 10-year risk)
Persons at low risk (<6% 10-year risk) should not undergo CAC measurement for cardiovascular risk III B
assessment
Cardiac CT is not recommended for routine evaluation of LV function in patients who have normal ECG, III C
have no history of MI, have no symptoms, have no signs indicative of heart failure, and have no complex
ventricular arrhythmias
Repeat cardiac CT for LV function evaluation in patients who have had prior LV functional evaluation less III C
than 1 year ago and in whom there has been no change in clinical status or anticipation of a change in
therapy is not recommended
Data from Fihn et al. [2]
250 K. Appau and A. E. Stillman

Table 22.7  Risk assessment in patients able and unable to exercise (ACCF/AHA/ACP/AATS/PCNA/SCAI/STS)
Indications Class Level of evidence
In patients with SIHD who have uninterpretable ECG, CCTA may be reasonable for risk assessment even IIb B
if they can exercise
In patients who are able to exercise and have interpretable ECG, CCTA is not recommended III C
CCTA can be used as a first-line test for cardiovascular risk evaluation in patients with SIHD who are IIa C
unable to exercise even if they have interpretable ECG
CCTA can be used for cardiovascular risk evaluation in SIHD patients with indeterminate functional test IIa C
results
CCTA may be used for cardiovascular risk evaluation in SIHD patients who cannot undergo stress imaging IIb C
tests; or it may be used as an alternative to invasive coronary angiogram in patients with unknown coronary
anatomy in whom functional test show moderate- to high-risk result
Performance of stress imaging and CCTA at the same time is not recommended III C
Data from Fihn et al. [2]

Table 22.8  Follow-up noninvasive testing in patients with known SIHD who have new, recurrent, or worsening symptoms that is not consistent
with unstable angina (ACCF/AHA/ACP/AATS/PCNA/SCAI/STS) [2]
Indications Class Level of evidence
In patients with new or worsening symptoms that are not indicative of UA and who have prior stents (with IIb B
stent diameter of ≥3 mm) or CABG, CCTA can be useful to evaluate for patency of coronary stents and
CABG even if they can exercise
In patients with new or worsening symptoms that is not indicative of UA and who have prior stents (with IIb B
stent diameter of ≥3 mm), CCTA may be useful if there is no evidence of moderate or severe coronary
calcifications even if they can exercise
CCTA is not recommended in patients who have evidence of moderate to severe coronary calcifications or III B
prior stents (with stents diameter of less than 3 mm) whether or not they can or cannot exercise
CCTA is not recommended in asymptomatic patients with prior CABG less than 5 years or prior PCI less III C
than 2 years ago
Data from Fihn et al. [2]

Table 22.9  Heart failure imaging guidelines (ACCF/ACR/ASE/ASNC/SCCT/SCMR [5])

Recommendations Class type Level of evidence
In patients with angina/ischemic equivalent syndrome/angina, noninvasive imaging may be considered to IIb C
define the likelihood of CAD in patients with HF and LV dysfunction
In patient without angina/ischemic equivalent syndrome/angina, noninvasive imaging may be considered to IIb C
define the likelihood of CAD in patients with HF and LV dysfunction
Data from Patel et al. [5]

Table 22.10  Stable ischemic heart disease guidelines (ACCF/AHA/ASE/ASNC/HFSA/HRS/SCAI/SCCT/SCMR/STS [8])

Test indications Calcium scoring CCTA
Low pretest probability of CAD, ECG interpretable, and able to exercise R R
Low pretest probability of CAD, ECG uninterpretable, or unable to exercise R M
Intermediate pretest probability of CAD, ECG interpretable, and able to exercise R M
Intermediate pretest probability of CAD, ECG uninterpretable, or unable to exercise R A
High pretest probability of CAD, ECG interpretable, and able to exercise R M
High pretest probability of CAD, ECG uninterpretable, and unable to exercise R M
Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate; R – Rarely appropriate; NA – No determination of appropriateness

Table 22.11  Asymptomatic (without symptoms or ischemic equivalent) (ACCF/AHA/ASE/ASNC/HFSA/HRS/SCAI/SCCT/SCMR/STS [8])

Test indications Calcium scoring CCTA
Low global CHD risk, regardless of ECG interpretability and ability to exercise R R
Intermediate global CHD risk and ECG interpretable and able to exercise M R
Intermediate global CHD risk ECG uninterpretable or unable to exercise M R
High global CAD risk and ECG interpretable and able to exercise M M
High global CAD risk and ECG uninterpretable or unable to exercise M M
Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate; R – Rarely appropriate; NA – No determination of appropriateness
22  Current Guidelines 251

Table 22.12  Newly diagnosed heart failure (resting LV function pre- Table 22.14 Syncope without ischemic equivalent (ACCF/AHA/
viously assessed but no prior CAD evaluation) (ACCF/AHA/ASE/ ASE/ASNC/HFSA/HRS/SCAI/SCCT/SCMR/STS [8])
ASNC/HFSA/HRS/SCAI/SCCT/SCMR/STS [8])
Test indications Calcium scoring CCTA
Test indications Calcium scoring CCTA Low global CAD risk R R
Newly diagnosed systolic heart failure R A Intermediate or high global CAD risk R M
Newly diagnosed diastolic heart failure R M Data from Wolk et al. [8] A – Appropriate; M – May be appropriate;
Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate; R – Rarely appropriate; NA – No determination of appropriateness
R – Rarely appropriate; NA – No determination of appropriateness

Table 22.13  Evaluation of arrhythmias without ischemic equivalent (no prior cardiac evaluation) (ACCF/AHA/ASE/ASNC/HFSA/HRS/SCAI/
SCCT/SCMR/STS [8])
Test indications Calcium scoring CCTA
Sustained VT R M
Ventricular fibrillation R M
Exercise-induced VT or nonsustained VT R M
Frequent PVCs R M
Infrequent PVCs R R
New-onset atrial fibrillation R R
Prior to initiation of anti-arrhythmia therapy in high global CAD risk patients R M
Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate; R – Rarely appropriate; NA – No determination of appropriateness

Table 22.15  Prior testing without intervening revascularization (sequential testing (≤90 days): abnormal prior test/study) (ACCF/AHA/ASE/
ASNC/HFSA/HRS/SCAI/SCCT/SCMR/STS)
Test indications Calcium scoring CCTA
Abnormal rest ECG findings (potentially ischemic in nature such as LBBB, T-wave inversions) and low R M
global CAD risk
Abnormal rest ECG findings (potentially ischemic in nature such as LBBB, T-wave inversions) and R M
intermediate to high global CAD risk
Abnormal prior exercise ECG test R A
Abnormal prior stress imaging study (assumes not repeat of same type of stress imaging) R A
Obstructive CAD on prior CCTA study NA NA
Obstructive CAD on prior invasive coronary angiography R R
Abnormal prior CCT calcium (Agatston score >100) NA M
Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate; R – Rarely appropriate; NA – No determination of appropriateness

Table 22.16 Sequential or follow-up testing (≤90  days): uncertain prior results (ACCF/AHA/ASE/ASNC/HFSA/HRS/SCAI/SCCT/
SCMR/STS)
Test indication Calcium scoring CCTA
Abnormal rest ECG findings (potentially ischemic in nature such as LBBB, T-wave inversions) and low R M
global CAD risk
Abnormal rest ECG findings (potentially ischemic in nature such as LBBB, T-wave inversions) and R M
intermediate to high global CAD risk
Abnormal prior exercise ECG test R A
Abnormal prior stress imaging study (assumes not repeat of the same type of stress imaging) R A
Obstructive CAD on prior CCTA study NA NA
Obstructive CAD on prior invasive coronary angiography R R
Abnormal prior CCT calcium (Agatston score >100) NA M
Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate; R – Rarely appropriate; NA – No determination of appropriateness
252 K. Appau and A. E. Stillman

Table 22.17  Sequential or follow-up testing (≤90  days): uncertain Table 22.22  Prior coronary calcium Agatston score asymptomatic
prior results  – equivocal, borderline, or discordant prior noninvasive (without ischemic equivalent) or stable symptoms (ACCF/AHA/ASE/
evaluation where obstructive CAD remains a concern (ACCF/AHA/ ASNC/HFSA/HRS/SCAI/SCCT/SCMR/STS [8])
ASE/ASNC/HFSA/HRS/SCAI/SCCT/SCMR/STS)
Calcium
Calcium Test indications scoring CCTA
Test indications scoring CCTA Agatston score <100 R R
Prior exercise ECG test R A Low to intermediate global CAD risk Agatston R R
Prior stress imaging study (assumes not repeat R A score between 100 and 400
of same type of stress imaging) High global CAD risk Agatston score between R R
Prior CCTA NA NA 100 and 400
Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate; Agatston score >400 R R
R – Rarely appropriate; NA – No determination of appropriateness Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate;
R – Rarely appropriate; NA – No determination of appropriateness
Table 22.18  Prior coronary angiography (invasive or noninvasive)
(ACCF/AHA/ASE/ASNC/HFSA/HRS/SCAI/SCCT/SCMR/STS [8])
Table 22.23  Normal prior exercise ECG test asymptomatic (without
Calcium
ischemic equivalent) (ACCF/AHA/ASE/ASNC/HFSA/HRS/SCAI/
Test indication scoring CCTA
SCCT/SCMR/STS)
Coronary stenosis or anatomic abnormality of NA NA
unclear significance found on cardiac CCTA Calcium
Coronary stenosis or anatomic abnormality of R R Test indications scoring CCTA
unclear significance on previous coronary Low global CAD risk R R
angiography Intermediate to high global CAD risk and test R R
<2 years ago
Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate;
R – Rarely appropriate; NA – No determination of appropriateness Intermediate to high global CAD risk and test R R
≥2 years ago
Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate;
Table 22.19  Follow-up testing (>90  days): asymptomatic or stable R – Rarely appropriate; NA – No determination of appropriateness
symptoms – abnormal prior exercise ECG test asymptomatic or stable
symptoms (ACCF/AHA/ASE/ASNC/HFSA/HRS/SCAI/SCCT/
SCMR/STS [8])
Table 22.24  Normal prior stress imaging study or non-obstructive
Test indication Calcium scoring CCTA CAD on angiogram (invasive or noninvasive) asymptomatic (without
Last test <2 years ago R R ischemic equivalent) (ACCF/AHA/ASE/ASNC/HFSA/HRS/SCAI/
Last test ≥2 years ago R R SCCT/SCMR/STS)
Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate; Calcium
R – Rarely appropriate; NA – No determination of appropriateness Test indications scoring CCTA
Low global CAD risk R R
Table 22.20  Abnormal prior stress imaging study asymptomatic or Intermediate to high global CAD risk and R R
stable symptoms (ACCF/AHA/ASE/ASNC/HFSA/HRS/SCAI/SCCT/ study <2 years ago
SCMR/STS [8]) Intermediate to high global CAD risk and R R
study ≥2 years ago
Test indication Calcium scoring CCTA
Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate;
Last study <2 years ago R R
R – Rarely appropriate; NA – No determination of appropriateness
Last study ≥2 years ago R R
Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate;
R – Rarely appropriate; NA – No determination of appropriateness
Table 22.25  Normal prior exercise ECG test stable symptoms (ACCF/
AHA/ASE/ASNC/HFSA/HRS/SCAI/SCCT/SCMR/STS)
Table 22.21  Obstructive CAD on prior coronary angiography (inva- Calcium
sive or noninvasive) asymptomatic (without ischemic equivalent) or Test indications scoring CCTA
stable symptoms (ACCF/AHA/ASE/ASNC/HFSA/HRS/SCAI/SCCT/
Low global CAD risk R R
SCMR/STS [8])
Intermediate to high global CAD risk and test R R
Test indications Calcium scoring CCTA <2 years ago
Last study <2 years ago R R Intermediate to high global CAD risk and test R R
Last study ≥2 years ago R R ≥2 years ago
Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate; Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate;
R – Rarely appropriate; NA – No determination of appropriateness R – Rarely appropriate; NA – No determination of appropriateness
22  Current Guidelines 253

Table 22.26  Normal prior stress imaging study or non-obstructive Table 22.30 Preoperative evaluation for non-cardiac surgery  –
CAD on angiogram (invasive or noninvasive) stable symptoms (ACCF/ Moderate-to-good functional capacity (≥4 METs) or no clinical risk
AHA/ASE/ASNC/HFSA/HRS/SCAI/SCCT/SCMR/STS) factors (ACCF/AHA/ASE/ASNC/HFSA/HRS/SCAI/SCCT/SCMR/
STS)
Calcium
Test indications scoring CCTA Test indications Calcium scoring CCTA
Low global CAD risk R R Any surgery R R
Intermediate to high global CAD risk and R R Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate;
prior study <2 years ago R – Rarely appropriate; NA – No determination of appropriateness
Intermediate to high global CAD risk and R R
prior study ≥2 years ago
From Wolk et  al. [8]. A – Appropriate; M – May be appropriate;
R – Rarely appropriate; NA – No determination of appropriateness Table 22.31  Asymptomatic and  <1  year post any of the following:
normal CT or invasive angiogram, normal stress test for CAD, or revas-
cularization (ACCF/AHA/ASE/ASNC/HFSA/HRS/SCAI/SCCT/
Table 22.27  Follow-up testing: new or worsening symptoms (ACCF/ SCMR/STS)
AHA/ASE/ASNC/HFSA/HRS/SCAI/SCCT/SCMR/STS) Test indication Calcium scoring CCTA
Calcium Any surgery R R
Test indications scoring CCTA Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate;
Normal exercise ECG test R A R – Rarely appropriate; NA – No determination of appropriateness
Non-obstructive CAD on coronary R R
angiography
(invasive or noninvasive) or normal prior
stress imaging study Table 22.32  Poor or unknown functional capacity (<4 METs) (ACCF/
Abnormal exercise ECG test R A AHA/ASE/ASNC/HFSA/HRS/SCAI/SCCT/SCMR/STS)
Abnormal prior stress imaging study R A Test indications Calcium scoring CCTA
Obstructive CAD on CCTA study R R Low-risk surgery R R
Obstructive CAD on invasive coronary R R ≥1 clinical risk factor
angiography Intermediate-risk surgery R R
Abnormal CCTA calcium (Agatston R M ≥1 clinical risk factor
Score > 100) Vascular surgery R R
Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate; ≥1 clinical risk factor
R – Rarely appropriate; NA – No determination of appropriateness Kidney transplant R R
Liver transplant R R
Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate;
Table 22.28 Post-revascularization (PCI or CABG) symptomatic R – Rarely appropriate; NA – No determination of appropriateness
(ischemic equivalent) (ACCF/AHA/ASE/ASNC/HFSA/HRS/SCAI/
SCCT/SCMR/STS)
Test indication Calcium scoring CCTA
Evaluation of ischemic equivalent R M Table 22.33  Determine exercise level prior to initiation of exercise
prescription or cardiac rehabilitation exercise prescription (ACCF/
Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate; AHA/ASE/ASNC/HFSA/HRS/SCAI/SCCT/SCMR/STS)
R – Rarely appropriate; NA – No determination of appropriateness
Test indication Calcium scoring CCTA
No prior revascularization R R
Table 22.29  Asymptomatic (without ischemic equivalent) (ACCF/ Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate;
AHA/ASE/ASNC/HFSA/HRS/SCAI/SCCT/SCMR/STS) R – Rarely appropriate; NA – No determination of appropriateness
Calcium
Test indications scoring CCTA
Incomplete revascularization/additional R R
revascularization feasible Table 22.34  Prior to the initiation of cardiac rehabilitation (as a stand-­
Prior left main coronary stent R M alone indication): able to exercise (ACCF/AHA/ASE/ASNC/HFSA/
<5 years after CABG R R HRS/SCAI/SCCT/SCMR/STS)
≥5 years after CABG R R Test indications Calcium scoring CCTA
<2 years after PCI R R Post-revascularization (PCI or CABG) R R
≥2 years after PCI R R Heart failure R R
Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate; Data from Wolk et al. [8]. A – Appropriate; M – May be appropriate;
R – Rarely appropriate; NA – No determination of appropriateness R – Rarely appropriate; NA – No determination of appropriateness
254 K. Appau and A. E. Stillman

World Health Organization [23, 24] guidelines on the use of Committee and the ACCF Appropriate Use Criteria Task Force. J
Am Coll Radiol. 2013;10(7):493–500.
cardiac CT. Recommendations from various appropriate use 8. Wolk MJ, Bailey SR, Doherty JU, Douglas PS, Hendel RC,
criteria and guidelines are shown in Tables 1-34. Kramer CM, et  al. ACCF/AHA/ASE/ASNC/HFSA/HRS/SCAI/
By far, the USA has the most comprehensive set of guide- SCCT/SCMR/STS 2013 multimodality appropriate use criteria
lines and appropriate use criteria, and therefore many other for the detection and risk assessment of stable ischemic heart dis-
ease: a report of the American College of Cardiology Foundation
countries tend to utilize the guidelines and the appropriate Appropriate Use Criteria Task Force, American Heart Association,
use criteria found in the USA. Furthermore, the USA is cur- American Society of Echocardiography, American Society of
rently shifting from single modality appropriate use criteria Nuclear Cardiology, Heart Failure Society of America, Heart
to multimodality imaging for various disease scenarios. Rhythm Society, Society for Cardiovascular Angiography and
Interventions, Society of Cardiovascular Computed Tomography,
There is a strong agreement among the guidelines found Society for Cardiovascular Magnetic Resonance, and Society of
around the globe with respect to class II and III indications Thoracic Surgeons. J Am Coll Cardiol. 2014;63(4):380–406.
for the use of cardiac CT. There is also a strong agreement 9. ACR–NASCI–SPR practice parameter for the performance
among the guidelines across the globe regarding the use of and interpretation of cardiac computed tomography (CT).
2016. Available from: http://www.acr.org/~/media/f4720a18f-
cardiac CT in low- to intermediate- and intermediate-risk 03b4a26a0c9c3cc18637d87.pdf.
patients. There is currently no class I indications for the use 10. Abbara S, Arbab-Zadeh A, Callister TQ, Desai MY, Mamuya W,
of cardiac CT except in Brazil. Thomson L, et  al. SCCT guidelines for performance of coronary
computed tomographic angiography: a report of the Society of
Cardiovascular Computed Tomography Guidelines Committee. J
Cardiovasc Comput Tomogr. 2009;3(3):190–204.
References 11. Garcia MJ, Blankstein R, Budoff MJ, Dent JM, Drachman

DE, Lesser JR, et  al. COCATS 4 Task Force 7: Training in
1. Cesar LA, Ferreira JF, Armaganijan D, Gowdak LH, Mansur AP, Cardiovascular Computed Tomographic Imaging. J Am Coll
Bodanese LC, et  al. Guideline for stable coronary artery disease. Cardiol. 2015;65(17):1810–21.
Arq Bras Cardiol. 2014;103(2 Suppl 2):1–56. 12. ACR Cardiac CT Certificate of Advanced Proficiency Cardiac

2. Fihn SD, Gardin JM, Abrams J, Berra K, Blankenship JC, Dallas AP, CT: American College of Radiology. Available from: http://www.
et  al. 2012 ACCF/AHA/ACP/AATS/PCNA/SCAI/STS Guideline acr.org/~/media/ACR/Documents/PDF/Education/Exams/CoAP/
for the diagnosis and management of patients with stable isch- CoAPBrochure.pdf.
emic heart disease: a report of the American College of Cardiology 13. European Diploma Cardiac Imaging: European Society of Cardiac
Foundation/American Heart Association Task Force on Practice Radiology. Available from: http://escr.org/cms/website.php?id=/en/
Guidelines, and the American College of Physicians, American education/diploma.htm.
Association for Thoracic Surgery, Preventive Cardiovascular 14. CT Coronary Angiography Service Specification: British Society of
Nurses Association, Society for Cardiovascular Angiography Cardiovascular Imaging. 2012. Available from: https://bsci.org.uk/
and Interventions, and Society of Thoracic Surgeons. J Am Coll standards-guidelines/ct-coronary-angiography-service-specification.
Cardiol. 2012;60(24):e44–e164. 15. Cury RC, Abbara S, Achenbach S, Agatston A, Berman DS, Budoff
3. Mancini GB, Gosselin G, Chow B, Kostuk W, Stone J, Yvorchuk MJ, et al. CAD-RADS: Coronary Artery Disease - Reporting and
KJ, et al. Canadian Cardiovascular Society guidelines for the diag- Data System.: An Expert Consensus Document of the Society of
nosis and management of stable ischemic heart disease. Can J Cardiovascular Computed Tomography (SCCT), the American
Cardiol. 2014;30(8):837–49. College of Radiology (ACR) and the North American Society
4. Montalescot G, Sechtem U, Achenbach S, Andreotti F, Arden C, for Cardiovascular Imaging (NASCI). Endorsed by the American
Budaj A, et al. 2013 ESC guidelines on the management of stable College of Cardiology. J Am Coll Radiol. 2016;13(12 Pt A):1458–
coronary artery disease: the Task Force on the management of sta- 66.e9
ble coronary artery disease of the European Society of Cardiology. 16. Leipsic J, Abbara S, Achenbach S, Cury R, Earls JP, Mancini GJ,
Eur Heart J. 2013;34(38):2949–3003. et al. SCCT guidelines for the interpretation and reporting of cor-
5. Patel MR, White RD, Abbara S, Bluemke DA, Herfkens RJ, Picard onary CT angiography: a report of the Society of Cardiovascular
M, et al. 2013 ACCF/ACR/ASE/ASNC/SCCT/SCMR appropriate Computed Tomography Guidelines Committee. J Cardiovasc
utilization of cardiovascular imaging in heart failure: a joint report Comput Tomogr. 2014;8(5):342–58.
of the American College of Radiology Appropriateness Criteria 17. Beanlands RS, Chow BJ, Dick A, Friedrich MG, Gulenchyn KY,
Committee and the American College of Cardiology Foundation Kiess M, et al. CCS/CAR/CANM/CNCS/CanSCMR joint position
Appropriate Use Criteria Task Force. J Am Coll Cardiol. statement on advanced noninvasive cardiac imaging using positron
2013;61(21):2207–31. emission tomography, magnetic resonance imaging and multidetec-
6. Rybicki FJ, Udelson JE, Peacock WF, Goldhaber SZ, Isselbacher tor computed tomographic angiography in the diagnosis and evalu-
EM, Kazerooni E, et  al. 2015 ACR/ACC/AHA/AATS/ACEP/ ation of ischemic heart disease--executive summary. Can J Cardiol.
ASNC/NASCI/SAEM/SCCT/SCMR/SCPC/SNMMI/STR/STS 2007;23(2):107–19.
Appropriate Utilization of Cardiovascular Imaging in Emergency 18. Kitagawa K, Choi BW, Chan C, Jinzaki M, Tsai IC, Yong HS, et al.
Department Patients With Chest Pain: A Joint Document of ASCI 2010 appropriateness criteria for cardiac magnetic resonance
the American College of Radiology Appropriateness Criteria imaging: a report of the Asian Society of Cardiovascular Imaging
Committee and the American College of Cardiology Appropriate cardiac computed tomography and cardiac magnetic resonance
Use Criteria Task Force. J Am Coll Cardiol. 2016;67(7):853–79. imaging guideline working group. Int J Cardiovasc Imaging.
7. White RD, Patel MR, Abbara S, Bluemke DA, Herfkens RJ, Picard 2010;26(Suppl 2):173–86.
M, et al. 2013 ACCF/ACR/ASE/ASNC/SCCT/SCMR appropriate 19. Tsai IC, Choi BW, Chan C, Jinzaki M, Kitagawa K, Yong HS, et al.
utilization of cardiovascular imaging in heart failure: an executive ASCI 2010 appropriateness criteria for cardiac computed tomog-
summary: a joint report of the ACR Appropriateness Criteria (R) raphy: a report of the Asian Society of Cardiovascular Imaging
22  Current Guidelines 255

Cardiac Computed Tomography and Cardiac Magnetic Resonance Society for Cardiovascular Angiography and Interventions, and
Imaging Guideline Working Group. Int J Cardiovasc Imaging. the Society for Cardiovascular Magnetic Resonance. J Am Coll
2010;26(Suppl 1):1–15. Cardiol. 2010;56(22):1864–94.
20. Kim YJ, Yong HS, Kim SM, Kim JA, Yang DH, Hong YJ, et  al. 22. Greenland P, Alpert JS, Beller GA, Benjamin EJ, Budoff MJ, Fayad
Korean guidelines for the appropriate use of cardiac CT. Korean J ZA, et al. 2010 ACCF/AHA guideline for assessment of cardiovas-
Radiol. 2015;16(2):251–85. cular risk in asymptomatic adults: a report of the American College
21. Taylor AJ, Cerqueira M, Hodgson JM, Mark D, Min J, O'Gara of Cardiology Foundation/American Heart Association Task Force
P, et  al. ACCF/SCCT/ACR/AHA/ASE/ASNC/NASCI/SCAI/ on Practice Guidelines. J Am Coll Cardiol. 2010;56(25):e50–103.
SCMR 2010 appropriate use criteria for cardiac computed 23. Global health risks: mortality and burden of disease attributable to
tomography. A report of the American College of Cardiology selected major risks: World Health Organization. 2009. Available
Foundation Appropriate Use Criteria Task Force, the Society of from: http://www.who.int/healthinfo/global_burden_disease/
Cardiovascular Computed Tomography, the American College of GlobalHealthRisks_report_full.pdf.
Radiology, the American Heart Association, the American Society 24. Global atlas on cardiovascular disease prevention and control:

of Echocardiography, the American Society of Nuclear Cardiology, World Health Organization. 2011. Available from: http://www.who.
the North American Society for Cardiovascular Imaging, the int/cardiovascular_diseases/publications/atlas_cvd/en/.
 Part V
Where We Are:
Risk Stratification and Management
 Clinical Application of the Coronary
Artery Calcium Score and Implications 23
for Cardiovascular Disease Prevention

Pamela B. Morris and Michael D. Shapiro

Introduction: The Burden of Atherosclerosis ages biomarkers that most reliably identify individuals at
greatest risk and most likely to benefit from intervention.
Despite advances in medical technologies and pharmaceuti- More recently, there has been a paradigm shift in the
cal treatments, the burden of atherosclerotic cardiovascular identification of at-risk individuals to include strategies for
disease (ASCVD) remains formidable [1]. Applications of early detection of ASCVD (the presence of subclinical ath-
evidence-based therapies, including the use of thrombolysis, erosclerosis), in addition to the use of traditional ASCVD
coronary artery bypass grafting, percutaneous transluminal risk factors. Quantification of coronary artery calcification
coronary angioplasty and stents, beta-blockers, angiotensin-­ (CAC) with electrocardiographic (ECG)-gated non-contrast
converting enzyme inhibitors, spironolactone, and HMG-­ cardiac computed tomography (CCT) has emerged as the
CoA reductase inhibitors (statins), have contributed to a most effect means to refine ASCVD risk estimation. Studies
substantial decline in ASCVD morbidity and mortality in the consistently demonstrate that CAC scores (CACS) provide
United States over the past two decades [2]. However, the incremental predictive value above and beyond traditional
direct and indirect costs of these treatments for symptomatic risk factors and improve the identification of at-risk patients
heart disease continue to rise, making primary prevention of as discussed below.
new onset ASCVD a global priority [1].
To reduce morbidity and mortality from ASCVD, two
approaches have been widely implemented: a population Global Risk Assessment Tools for ASCVD:
(public health) approach and a high-risk (clinical strategy) Matching Intensity of Intervention
approach. The population approach focuses on community-­ with Severity of Risk
wide education encouraging individuals to maintain an opti-
mal ASCVD risk factor profile, including adoption of a heart Matching intensity of preventive treatments to the severity of
healthy dietary pattern, maintenance of optimal weight, reg- risk is based on the concept that high-risk patients receive the
ular physical activity, avoidance of tobacco products, and greatest benefit from therapeutic intervention, thereby opti-
achievement of optimal lipid and blood pressure levels. This mizing efficacy, safety, and cost-effectiveness [3]. Risk
approach can be expected to have the greatest long-term assessment algorithms have been developed from epidemio-
impact on reducing the burden of ASCVD across the globe. logic studies for a variety of populations using conventional
However, in individuals with established risk factors for ath- ASCVD risk factors to guide selection of candidates for
erosclerosis, the high-risk approach intensifies management evidence-based preventive drug therapies. The Framingham
and specifies evidence-based interventions proven to reduce Risk Score (FRS) algorithms are derived from a predomi-
ASCVD death and disability. The high-risk approach lever- nantly Caucasian population in the USA [4], whereas the
Prospective Cardiovascular Munster (PROCAM—Germany)
[5] and the Systemic Coronary Risk Evaluation (SCORE) [6]
P. B. Morris (*) models are derived from European population studies. The
Seinsheimer Cardiovascular Health Program, Department of 2013 American College of Cardiology (ACC)/American
Medicine and Cardiology, Medical University of South Carolina,
Heart Association (AHA) Cardiovascular (CV) Risk
Charleston, SC, USA
e-mail: [email protected] Calculator (based on the Pooled Cohort Equations) is derived
from a US population that is composed of primarily
M. D. Shapiro
Department of Medicine, Division of Cardiovascular Medicine, Caucasians and blacks [7]. Each algorithm predicts either
Oregon Health & Science University, Portland, OR, USA 10-year risk of coronary heart disease (CHD) events

© Humana Press 2019 259

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_23
260 P. B. Morris and M. D. Shapiro

(­coronary death or myocardial infarction [MI]) or global Ethnic Study of Atherosclerosis), four of the tools, including
ASCVD events (fatal or nonfatal MI, fatal or nonfatal stroke, the 2013 AHA/ACC CV Risk Calculator, demonstrated over-
CHD death). The ACC/AHA CV Risk Calculator also pro- estimation of risk (25% to 115%) in this cohort without base-
vides an estimate of lifetime or 30-year risk of ASCVD line clinical ASCVD [10]. The Reynolds Risk Score
events for individuals between the ages of 20 and 59 years. overestimated risk by 9% in men but underestimated risk by
As cardiovascular risk is most strongly related to age, risk 21% in women.
prediction algorithms may underestimate the relative or Nearly 660,000 previously asymptomatic people present
long-term absolute risk of younger individuals, particularly annually in the USA with a first MI, and 610,000 people
younger men and middle-aged men and women. Recognizing present with a first stroke [1]. This would suggest that
these limitations, the Reynolds Risk Score was developed to although our current risk assessment algorithms are useful
more accurately predict ASCVD risk in men and women for population-based prediction, clinicians evaluating the
before age 60 years [8, 9]. This algorithm includes not only individual patient may need additional information to accu-
conventional ASCVD risk factors (age, systolic blood pres- rately identify those at higher risk. While most of the excess
sure, total- and high-density lipoprotein-cholesterol, and risk for ASCVD events in a population may be explained
tobacco use) but also incorporates hemoglobin A1C levels, by conventional ASCVD risk factors, there is considerable
high-sensitivity C-reactive protein (hs-CRP), and family his- variability in how the disease is expressed in individuals
tory of premature MI. In a recent analysis of calibration and within the population. The majority of ASCVD events con-
discrimination among five cardiovascular risk scoring algo- tinue to occur in patients considered either low- or interme-
rithms in the more contemporary MESA population (Multi-­ diate-risk at baseline assessments, and most individuals are

a Traditional 10-Year Risk Model b “Lifetime” Risk Model

Unaccounted for Unaccounted for

Risk Factor Exposure Risk Factor Exposure
RISK ®

RISK ®

ATHEROSCLEROSIS ATHEROSCLEROSIS

0 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 70 80 90

AGE ® AGE ®

c Subclinical Disease Detection Model

RISK ®

Integrated Risk
Exposure

0 10 20 30 40 50 60 70 80 90
AGE ®

Fig. 23.1  Strategies for risk prediction. Traditional risk models (a) use Subclinical atherosclerosis (c), by directly visualizing the accumulated
one-time measurements of risk factors taken later in life—in this exam- burden of atherosclerosis, integrates lifetime risk exposure and may
ple at age 50—and project coronary heart disease (CHD) risk during the provide more accurate estimates of near-term and long-term risk. (From
coming 10 years. Lifetime risk models (b) similarly use one-time mea- Blaha et al. [15], with permission)
surements of risk factors but instead project long-term CHD risk.
23  Clinical Application of the Coronary Artery Calcium Score and Implications for Cardiovascular Disease Prevention 261

not recognized as high-risk prior to first CHD event [11– a 90th percentile of CAC, Women
14]. Previous risk assessment algorithms have focused only
on a single measure of traditional risk factors taken at a 1200
relatively later stage in adult life and may not accurately White
Black
capture the risk of cumulative exposure to all risk determi- 1000 Hispanic
nants over a lifetime (Fig.  23.1). This can contribute to Chinese
over- and underestimation of risk and resultant over- or
800
undertreatment with preventive therapies. To enhance the
predictive power in individuals, current interest is focused
on the identification of emerging risk factors that influence 600

CAC
how the major risk factors affect absolute ASCVD risk, as
well as early disease detection strategies such as preventive 400
imaging techniques to identify subclinical atherosclerotic
disease. The 2013 ACC/AHA cardiovascular risk assess-
ment guideline suggests consideration of these risk markers 200
if, after quantitative risk assessment, a risk-based treatment
decision is uncertain [7]. Recommended risk markers 0
include family history of premature ASCVD, hs-CRP, and
50 60 70 80
measures of subclinical atherosclerotic disease (ankle bra-
Age
chial index [ABI] and CACS).
The presence of subclinical atherosclerosis can inform b 90th percentile of CAC, Men
the treatment decisions (e.g., initiation, intensification) for
4000
preventive therapies in carefully selected asymptomatic
individuals. CACS directly quantifies the accumulated White
Black
burden of coronary atherosclerosis and provides a more Hispanic
personalized assessment of risk. As noted by Blaha et al., 3000 Chinese
by integrating the cumulative interaction of early risk
determinants (genetic and epigenetic factors) with lifetime
exposure to traditional risk factors and other influences
CAC

2000
(e.g., passive tobacco smoke exposure, air quality), CACS
provides a unique measure of the resultant effect on the
vasculature [15, 16].
1000

 oronary Artery Calcium: Age, Gender,

C
and Ethnic Differences 0

The extent of CAC varies with ethnicity, gender, and age 50 60 70 80

as demonstrated in the MESA study [17, 18]. One analysis Age
from MESA included 6110 subjects of which 53% were
female. Ethnicity distribution was balanced across gender Fig. 23.2  Estimated 90th percentile of the CACS distribution by gen-
with 41% white, 11.8% Chinese American, 26.4% black, der, age, and race/ethnicity. Note that the y-axis scales for these two
plots are very different, with the scale for the men (b) >3 times as large
and 20.9% Hispanic individuals. Men had higher CACS as that for women (a). (From McClelland et al. [18], with permission)
than women in every age group for each ethnicity.
Moreover, the extent and severity of CAC increased with
age in both men and women. Approximately 62% of  rognostic Implications of CACS
P
women demonstrated no evidence of CAC, whereas only and Prognostic Implications of Coronary
40% of men had a CACS = 0. Overall, for men, blacks had Artery Calcium in ASCVD Risk Evaluation
the lowest, and whites had the highest prevalence of
CAC.  White women had the highest prevalence of Discrimination is a measure of a marker’s ability to differen-
CACS > 0, whereas Hispanic women had the lowest. The tiate between individuals who do and do not have events. The
prevalence of subclinical atherosclerosis was intermediate most popular measure of discrimination is the receiver oper-
in Chinese and black women (Fig. 23.2). ating characteristic (ROC) curve, a plot of sensitivity vs
262 P. B. Morris and M. D. Shapiro

1-specificity [19]. The area under the curve (AUC) is also Cholesterol, patients with 10-year ASCVD risk of ≥7.5
known as the c-statistic and can range from 0.5 (no predic- to <20% are identified as one of four statin benefit groups in
tive ability) to 1 (perfect discrimination). CACS has been primary prevention. Patients with risk from 5% to <7.5 may
repeatedly demonstrated to be a robust predictor of ASCVD also be considered for statin therapy in the setting of shared
events, including CHD, stroke, and fatal ASCVD and is decision making [41, 42]. Thus, the 10-year risk of ASCVD
associated with a significantly better c-statistic (improved events as estimated by CACS may help to inform the patient
discrimination) as compared to traditional risk factors, par- clinician discussion regarding preventive therapies.
ticularly in patients who are deemed to be in the intermediate Historically, duration of follow-up after baseline CACS
risk category (Table  23.1) [20–35]. CACS has emerged as was only approximately 4–5  years. This is a significant
the most powerful predictor of risk in asymptomatic indi- shortcoming given the fact that contemporary risk assess-
viduals as demonstrated in every cohort in which it has been ment algorithms, such as the ACC/AHA Pooled Cohort
studied. One stunningly consistent observation is the approx- Equations, predict 10-year and lifetime risk of ASCVD
imate tenfold increased hazard of hard events associated events [7]. However, two recent studies have demon-
with high CACS compared to CACS = 0 [21, 36]. In a head-­ strated that CACS accurately predicts 15-year mortality
to-­head comparison of novel risk markers (CACS, ABI, hs-­ and support effective long-term risk prediction and reclas-
CRP, and family history) for improvement in ASCVD risk sification [43, 44].
assessment in intermediate-risk individuals, all were inde- The net reclassification index (NRI) is used for calibra-
pendent risk markers, but CACS provided markedly superior tion of risk markers to measure the improvement in risk
discrimination and risk reclassification compared to other prediction or risk reclassification when comparing new
markers [37]. In MESA, findings suggested that CACS was biomarkers or measures of subclinical atherosclerosis to
associated more strongly than carotid intima-media thick- traditional risk prediction strategies [19]. Several large
ness (cIMT) with the risk of incident ASCVD with a prospective, population-based studies have demonstrated
c-­statistic of 0.81 (95% CI 0.78–0.83) for CACS versus 0.78 that CACS improves not only risk stratification and dis-
for cIMT (95% CI 0.75–0.81) [38]. crimination but also improves accurate reclassification of
The data from five large prospective, randomized studies ASCVD risk beyond traditional risk factors and risk
provide approximate 10-year event rates associated with assessment algorithms [37, 40, 45–48]. Though CACS has
increasing levels of CACS (ASCVD events defined in three primarily been recommended for use in patients at inter-
studies as CHD death, MI, and revascularization and in two mediate ASCVD risk as determined by conventional risk
studies as CHD death and MI) (Table 23.2) [20, 25, 30, 34, assessment algorithms, there has also been considerable
35, 39, 40]. Asymptomatic patients with CACS 101–400 interest in defining its potential role in individuals at low
Agatston had a 10-year event rate of 12.8–16.4%, and risk. An early analysis of the MESA study included 2684
patients with CACS 1–100 Agatston had an approximate women followed for a mean of 3.75 years who were clas-
10-year event rate between 2.3% and 5.9%. According to the sified as low-risk based on the FRS [33]. The prevalence of
2018 ACC/AHA Guideline on the Management of Blood any CAC in these low-risk women was 32%. There was a

Table 23.1  Prognostic power of coronary artery calcium in asymptomatic patients

First author Mean age Follow-up Calcium score cutoff Comparator group for relative risk Relative risk
(Reference #) N (years) (years) (Agatston) calculation ratio
Arad et al. [22] 1172 53 3.6 >160 <160 20.2
Wong et al. [24] 926 54 3.3 Top quartile First quartile 8.8
(>270)
Greenland et al. [29] 1312 66 7.0 >300 CACS = 0 3.9
Shaw et al. [28] 10,377 53 5 ≥400 <10 8.4
Arad et al. [30] 4903 59 4.3 ≥100 <100 10.7
Taylor et al. [31] 2000 43 3.0 >44 0 11.8
Vliegenthart et al. 2013 71 3.3 >1000 <100 8.3
[25] 400–1000 <100 4.6
Budoff et al. [32] 25,503 56 6.8 >400 0 9.2
Lakoski et al. [33] 3601 60 3.75 >0 0 6.5
Becker et al. [34] 1726 57.7 3.4 >400 0 6.8 (men)
7.9 (women)
Detrano et al. [20] 6722 62.2 3.8 >300 0 9.67
Erbel et al. [35] 4129 45–75 5 >75th percentile <25th percentile 11.1 (men)
3.2 (women)
Adapted from Weber et al. [21], with permission
23  Clinical Application of the Coronary Artery Calcium Score and Implications for Cardiovascular Disease Prevention 263

Table 23.2  Summary of absolute event rates by coronary artery CACS = 0, the presence of any CAC (CACS > 0) was associ-
­calcium score (CACS) from 14,856 patients in five prospective studies ated with increased risk of ASCVD event (multivariable-­
FRS 2013 ACC/AHA statin 10-year event adjusted hazard ratio, 2.04 [95% CI, 1.44–2.90]) and modest
CACS equivalenta benefit groupb rate (%) improvement in prognostic accuracy compared to traditional
0 Very low No 1.1–1.7 risk factor-based algorithm (c-statistic 0.77 [95% CI, 0.74–
1–100 Low May consider 2.3–5.9
0.81] versus 0.73 [95% CI, 0.69–0.77]).
101– Intermediate Yes 12.8–16.4
400 The role of CACS in patients with a low lifetime risk of
>400 High Yes 22.5–28.6 ASCVD as measured by optimal risk factor status (total cho-
>1000 Very high Yes 37.0 lesterol <180  mg/dL, untreated blood pressure
Adapted from Hecht [39], with permission <120/80  mmHg, non-smoking and nondiabetic status) was
a
FRS indicates Framingham Risk Score (Ref. [4]) evaluated in the MESA cohort [51]. At 10.4 years of follow-
b
2013 ACC/AHA statin benefit group indicates eligibility for statin ­up, the event rates for CACS = 0, CACS > 0, and CACS > 100
therapy based on 10-year ASCVD risk as calculated by the Pooled
Cohort Equations
were 0.9/1000, 5.7/1000, and 11.0/1000 person-­ years,
respectively. The presence of CAC was the strongest predic-
tor of incident CHD in this low-risk cohort.
sixfold greater risk for a CHD event in women with CAC has been found to be predictive of CHD events
CACS > 0 compared to women with no evidence of CAC across all age groups [52]. The MESA cohort ranged in age
(HR 6.5; 95% CI 2/6–16.4, p < 0.001). from 45 to 84  years. With increasing age, there was an
Among another cohort of 9715 patients referred for increased proportion of the cohort with CACS 1–100 and
CACS, a total of 2363 asymptomatic men and women were CACS  >  100 and a decreased proportion of patients with
classified as low-intermediate risk using the FRS [47]. After CACS  =  0 (Fig.  23.3). Compared with CACS  =  0, CACS
a mean follow-up of 14.6  years, the total mortality ranged 1–100 and CACS > 100 were associated with a 2-fold and
from 5.0% for women with a CACS = 0 to 23.5% for those 12-fold increased risk of all CHD events for younger indi-
with CACS > 400 (p < 0.001). For men, the 15-year mortal- viduals (ages 45–54), respectively. Among patients
ity ranged from 3.5% among those with CACS = 0 to 18% 75–84  years of age, CACS 1–100 and CACS  >  100 were
for those with CACS > 400 (p < 0.001). Thus, even among associated with 11-fold and 20-fold increased risk of events
men and women with low-intermediate risk factor burden, compared to those with no evidence of CAC. CACS = 0 in
the presence of CAC effectively identifies those who are at the 75–84 year-old group had an 8.5-year survival of approx-
significantly higher risk than predicted by the FRS alone. imately 98% which was comparable to younger patients with
Women with CACS > 10 were noted to have higher mortality no CACS.
risk compared to men with similar CACS. The fact that risk In 2015, a novel risk score was derived in the MESA pop-
assessment based on the traditional risk factor-based algo- ulation to estimate 10-year CHD risk using CACS in con-
rithm performed poorly in women, suggests that CACS may junction with traditional ASCVD risk factors and validated
play an important role to improve risk prediction in women. in the Heinz Nixdorf Recall Study (HNR) and the Dallas
More recently, data from MESA was used to calibrate the Heart Study (DHS) [53]. The addition of CACS to the tradi-
Pooled Cohort Equations used in the 2013 ACC/AHA cho- tional risk factors improved risk prediction significantly
lesterol guideline and to determine whether use of nontradi- (c-statistic 0.80 vs. 0.75; p  <  0.0001) (Fig.  23.4). External
tional risk markers can improve assessment in low-risk validation in both the HNR and DHS populations provided
individuals who fall below the threshold for statin therapy very good discrimination and validation, with Harrell’s
[49]. During the 10 years of follow-up, 57% of events in the c-statistic of 0.779 in HNR and 0.816 in DHS. The MESA
MESA cohort occurred among participants with an esti- risk score is available online at https://www.mesa-nhlbi.org/
mated 10-year ASCVD risk <7.5%. The presence of signifi- MESACHDRisk/MesaRiskScore/RiskScore.aspx and may
cant subclinical atherosclerosis with CACS ≥ 300 reclassified be used to aid clinicians when communicating risk to patients
6.8% as being at higher risk with an event rate of 13.3% and and as part of the determination of risk-based treatment
relative risk of 4.0 (95% CI 2.8–5.7, p < 0.0001). strategies.
A recent meta-analysis assessed the utility of CACS for The dominant effect of age on eligibility for statin therapy
ASCVD risk estimation among 6739 women with 10-year according to the ACC/AHA cholesterol guideline means that
risk <7.5% by the Pooled Cohort Equations in five large all healthy people, even with optimal risk factors, will auto-
population-­ based cohorts (the Dallas Heart Study, the matically qualify for statin therapy between the ages of 63
Framingham Heart Study, the Heinz Nixdorf Recall study, and 71 years, depending upon sex and ethnicity. A disease-­
MESA, and the Rotterdam Study) [50]. The median follow- guided approach based upon the presence of subclinical ath-
­up period ranged from 7.0 to 11.6 years. CAC was present in erosclerosis as assessed by CACS or carotid plaque burden
36.1% of these low-risk women. Compared to women with was evaluated in elderly individuals in the BioImage Study
264 P. B. Morris and M. D. Shapiro

Fig. 23.3  Proportion of 100

patients in each coronary
artery calcium (CAC) group 90
with increasing age. (From
Tota-Maharaj et al. [52], with 80
permission)
70

Patients (%)
60

50

40

30

20

10

0
45 50 55 60 65 70 75 80

Age (y)

CAC score 0 CAC score 1-100 CAC score >100

to more accurately assign statin therapy (Fig.  23.5) [54].

1.0

Participants (N = 5805) who were mean age 69 years (range

55–80 years) underwent 10-year ASCVD risk assessment by
the ACC/AHA Pooled Cohort Equations, as well as CACS
0.8

and carotid ultrasound imaging to detect and quantify carotid

plaque. In the study population, 86% of participants quali-
fied for risk-based statin therapy, but imaging demonstrated
0.6

that up to one third had no evidence of CAC. The investiga-

Sensitivity

tors tested the disease-guided reclassification approach to

improve specificity (less overtreatment) and sensitivity (less
0.4

undertreatment) among elderly people. Those individuals

with 10-year ASCVD risk ≥7.5% were down-classified from
Risk with CAC: AUC=0.8136 statin eligible to ineligible if there was no evidence of sub-
0.2

Risk without CAC: AUC=0.7600 clinical atherosclerosis. Individuals with intermediate

10-year ASCVD risk (5% to <7.5%) were up-classified from
statin optional to statin eligible if CACS ≥ 100 or there was
0.0

evidence of significant carotid plaque. During the median

follow-up of 2.7 years, there were 91 CHD events and 138
0.0 0.2 0.4 0.6 0.8 1.0
CVD events. The specificity for CHD and CVD events
(1-Specificity) improved significantly with CACS-guided reclassification
without loss in sensitivity, leading to a binary NRI of 0.20 for
Fig. 23.4  Discrimination of the MESA CHD Risk Score Within the
Development Cohort. The area under the survival ROC curve within
CHD and 0.14 for CVD. The positive NRIs were driven pri-
MESA was 0.81 for the model with CAC, indicating excellent discrimi- marily by down-classifying the patients with CACS = 0 or no
nation between events and nonevents. Receiver-operator characteristic evidence of carotid plaque. The investigators proposed that
(ROC) curves for the risk scores with and without coronary artery cal- limiting statin therapy to those with CACS > 0 or evidence of
cium (CAC) applied within the MESA (Multi-Ethnic Study of
Atherosclerosis) cohort. (Adapted from McClelland et  al. [53], with
carotid plaque could spare a significant proportion of asymp-
permission) tomatic elderly people from lifelong statin therapy.
23  Clinical Application of the Coronary Artery Calcium Score and Implications for Cardiovascular Disease Prevention 265

Fig. 23.5 Disease-Guided
CENTRAL ILLUSTRATION: Disease-Guided Primary Prevention With Statins
Primary Prevention With
Statins. (a) Following
guideline-recommended A. Simple Disease-guided Reclassification Approach
formal risk assessment and
Individuals Free ACC/AHA Subclinical
atherosclerosis imaging, Statin Eligibility
Disease-guided
of ASCVD Atherosclerosis? Reclassification
statin-eligible individuals are
down-classified to ineligible in Step 1 Step 2
the absence of atherosclerosis Yes=Up-classify
and other high-risk features, Not
and statinineligible patients are eligible
Not
up-classified to eligible if eligible
significant atherosclerosis is Statin
Statin
present. This principle eligible eligible
facilitates an informed
clinician-patient discussion,
leading to an individualized No=Down-classify
treatment decision. (b) In the Improved statin allocation
after clinician-patient discussion
BioImage cohort, beneficial
reclassification was achieved B. Clinical Implications in the Biolmage Study
using coronary artery calcium
(CAC) and carotid plaque Net Reclasification
burden (cPB). ACC/AHA CAC-guided 64% Improvement*
American College of Statin Eligible CHD: 20%
Cardiology/American Heart 2013 CVD: 14%
5,805 Biolmage
Association, ASCVD participants free of
ACC/AHA
*NRI = ∆Sensitivity + ∆Specificity
atherosclerotic cardiovascular ASCVD at baseline 86%
disease, CHD coronary heart Statin Eligible Net Reclasification
Improvement*
disease, CVD cardiovascular 70%
Statin Eligible CHD: 9%
disease, NRI net cPB-guided
CVD: 6%
reclassification index. (From
Mortensen et al. [54], with
permission)
Improved Statin Allocation

Diabetes CACS < 10 [57]. The relative risk for CV events comparing

CACS ≥ 10 to CACS < 10 was 9.22 (95% CI, 2.73–31.07,
Asymptomatic individuals with diabetes exhibit more p = 0.005) (Fig. 23.6). The risk of CV events increased with
extensive subclinical coronary atherosclerosis than do non- increasing CACS levels of ≥100, ≥400, and ≥1000 com-
diabetic patients. Moreover, younger patients with diabetes pared to patients with diabetes and CACS < 10.
manifest CACS comparable to older individuals without Multiple large prospective studies have demonstrated that
diabetes [55]. Among traditional ASCVD risk factors, dia- baseline CACS is an independent predictor of ASCVD
betes is the strongest correlate for CACS > 0. In a study of events in patients with diabetes. The prognostic value of
young individuals with diabetes (age  <  40  years), 3723 CACS in patients with and without diabetes was assessed in
asymptomatic patients (4% with diabetes) underwent 10,377 asymptomatic individuals (903 with diabetes) over an
CACS [56]. Overall, the prevalence of CAC was 43% in average follow-up of 5.0 years [58]. Overall, mortality was
young patients with diabetes and was >50% in those aged significantly greater in patients with diabetes than in those
>35 years. The presence of diabetes was associated with a without diabetes. Mortality increased with increasing CACS
fourfold higher odds of CACS ≥ 100 when compared to for both diabetic and nondiabetic individuals. However, for
nondiabetic individuals. every increase in CACS, there was a greater increase in mor-
In a systematic review and meta-analysis of eight obser- tality for diabetics compared to nondiabetic (Fig. 23.7).
vational studies with a mean follow-up of 5.18 years, includ- In the MESA population, there were 881 participants
ing 6521 patients with diabetes, the relative risk for all-cause (13%) with diabetes [59]. Compared to individuals without
mortality or CV events, or both, was 5.47 (95% CI, 2.59– diabetes, the adjusted hazard ratios for CHD events in dia-
11.53, p < 0.001) when comparing CACS ≥ 10 compared to betics with CACS 1–99, CACS 100–399, and CACS ≥ 400
266 P. B. Morris and M. D. Shapiro

Fig. 23.6  Meta-analyses of Study Sample size Relative risk Weight Relative risk
association between coronary All cause mortality or cardiovascular (95% Cl) (%) (95% Cl)
artery calcium score ≥10 and events (fatal and non-fatal), or both
outcome in people with type 2
Raggi 200430 903 12.49 6.84 (2.48 to 18.88)
diabetes. Weights are from
random effects analysis. Anand 200631 181 7.55 24.15 (3.31 to 176.10)
(From Kramer et al. [57], with Elkeles 200814 589 10.36 9.79 (2.43 to 39.49)
permission) Chiu 201012 225 14.62 1.46 (0.79 to 2.70)
Agarwal 201132 1048 14.03 3.86 (1.85 to 8.05)
Malik 201133 871 14.26 18.43 (9.23 to 36.78)
Shemesh 201234 164 12.30 2.13 (0.75 to 6.07)
Silverman 201213 2384 14.39 4.92 (2.53 to 9.57)
Overall: P<0.001, I2=82.4% 100.00 5.47 (2.59 to 11.53)

Cardiovascular events (fatal and non-fatal)

Anand 200631 181 17.93 24.15 (3.31 to 176.10)
Elkeles 200814 589 23.69 9.79 (2.43 to 39.49)
Malik 201133 871 30.97 18.43 (9.23 to 36.78)
Shemesh 201234 164 27.41 2.13 (0.75 to 6.07)
Overall: P<0.005, I2=76.7% 100.00 9.22 (2.73 to 31.07)
0.00568 1 176

Fig. 23.7  Cox proportional No Diabetes (n=9,474) Diabetes (n=903)

hazards survival (n = 10,377) 1.00 1.00
by electron beam tomography 0-10 (n=5,600)
.98 .98 0-10 (n=351)
coronary calcium 11-100 (n=1,854)
measurements in subjects .96 101-400 (n=1,251) .96 11-100 (n=189)
with and without diabetes .94 .94
Cumulative Survival

(chi-square = 204, 401-1000 (n=508) 101-400 (n=182)

.92 .92
p < 0.0001). The number of
subjects in each calcium score .90 .90
401-1000 (n=110)
category is in parentheses. .88 .88
>1000 (n=261)
(From Raggi et al. [58], with .86 .86
permission)
.84 .84
.82 .82
>1000 (n=71)
.80 .80
0 1 2 3 4 5 0 1 2 3 4 5
Follow-up (Years) Follow-up (Years)

were 2.9, 3.3, and 6.2, respectively. Thus, although patients more intensive preventive medical therapies. However, mul-
with diabetes are at higher ASCVD risk overall, recent evi- tiple population-based studies have clearly demonstrated the
dence does suggest a wide range of risk and supports a role favorable prognosis associated with the absence of CAC,
for CACS in selected individuals (particularly in those with leading many experts to suggest that a CACS = 0 may iden-
diabetes who are younger and with duration of disease of tify patients who could avoid lifelong pharmacotherapy [60,
less than 10 years). 61]. This is particularly relevant as the 2013 ACC/AHA cho-
lesterol guideline increased the number of US adults eligible
for statin therapy from 43 million to 56 million, with the
The Power of Zero majority of these newly eligible candidates being individuals
without clinical ASCVD [62]. Therefore, the value of CACS
The focus for integration of CACS into clinical practice for may now be finding CACS = 0 in a patient with an intermedi-
ASCVD risk assessment has largely been on the intermediate-­ ate 10-year ASCVD risk or uncertainty after consideration of
risk patient with evidence of subclinical atherosclerosis to risk estimation to inform clinical decision making when con-
identify higher-risk individuals who may be candidates for sidering the initiation of statin therapy.
23  Clinical Application of the Coronary Artery Calcium Score and Implications for Cardiovascular Disease Prevention 267

Asymptomatic Patients The implications of CACS  =  0 among statin candidates

according to the ACC/AHA cholesterol guideline were also
In a systematic review of 49 studies of the diagnostic and evaluated in the MESA population [65]. Approximately 50%
prognostic utility of CACS (N = 64,873 individuals), 29,312 of the 4758 MESA participants in this analysis qualified for
asymptomatic patients demonstrated a CACS = 0 [45]. Over statin therapy as per the guidelines, the majority because of
a mean follow-up period of 50 months, there was a very low an estimated 10-year ASCVD risk ≥7.5% by the Pooled
event rate (154 ASCVD events; 0.47%) in this cohort. Cohort Equations. Of all participants who were statin eligi-
Compared to patients with any CAC, the cumulative relative ble (recommend or consider statin therapy per guideline),
risk rate for ASCVD events with a CACS = 0 was 0.15 (95% 44% had CACS = 0 and an event rate of 4.2 per 1000 person-­
CI: 0.11–0.21, p < 0.001). years, which is a lower observed ASCVD risk than the
The absence of CAC and all-cause mortality was threshold recommended for treatment. Thus, it is suggested
assessed in a study of 44,052 consecutive patients referred that there is significant heterogeneity of risk among patients
for CACS [63]. Among the 19,898 patients (45%) with no who are eligible for statin therapy according to the guideline,
evidence of CAC, there were 104 deaths (0.52%) over a and the absence of CAC may reclassify approximately half
mean follow-up period of 5.6  years. The annualized all- of these individuals to a lower-risk group.
cause mortality rate for CACS = 0 was 0.87 per 1000 per- The landmark JUPITER trial demonstrated that healthy
son-years (Table 23.3). patients without clinical ASCVD, normal LDL-C, and ele-
In the MESA population, there were a total of 3415 vated hs-CRP benefit from high-intensity statin therapy with
asymptomatic participants with no detectable CAC [64]. a significant reduction in major cardiovascular events [66].
During a median follow-up period of 4.1 years, there were 17 In a subset of MESA participants who met criteria for entry
all CHD events (MI, cardiac arrest, fatal CHD, angina, revas- into the JUPITER study, nearly half of the population had no
cularization) and 10 (0.3%) hard CHD events (nonfatal MI evidence of CAC (47%) [67]. During the mean follow-up
and CHD death). Compared to participants with CACS = 0, period of 5.8 years, the rate of CHD events was 0.8 per 1000
there was a more than threefold increase in all CHD events person-years, and there was an unfavorable 5-year number
(HR, 3.66; p65% CI 1.71–7.85) when even only minimal needed to treat (NNT) of 549 (Table 23.4). Most of the events
CAC (CACS 1–10) was present. occurred in individuals with CACS > 100 which was associ-

Table 23.3  All-cause mortality rates by CAC scores in overall population (N = 44,052) of consecutive asymptomatic individuals
Rate/1000 person-years
No. of patients No. of events at risk 95% CI for rate
CAC = 0 19,898 (45%) 104 (0.52%) 0.87 0.72–1.05
CAC 1–10 5388 (12%) 58 (1.06%) 1.92 1.48–2.48
CAC > 10 18,766 (43%) 739 (3.96%) 7.48 6.95–8.04
Total 44,052 (100%) 901 (2.05%) 3.62 3.39–3.89
From Blaha et al. [63], with permission

Table 23.4  CHD and CVD events by CAC status in the MESA population eligible for JUPITER
N (%) CHD events CVD events
CHD events Event rate (per 1000 Hazard ratio CVD events Event rate (per 1000 Hazard ratio
(%) person-years) (95% CI) (%) person-years) (95% CI)
CAC 0 444 2 (0.5%) 0.8 1* (ref) 9 (2.0%) 3.7 1* (ref)
(47%)
CAC 1–100 267 7 (2.6%) 4.8 4.91 12 (4.5%) 8.4 1.86
(28%) (0.97–24.9) (0.73–4.76)
CAC >100 239 25 (10.6%) 20.2 27.8 32 (13.4%) 26.4 6.16
(25%) (5.97–129.8) (2.51–15.1)
Any CAC 506 32 (6.3%) 11.0 11.0 44 (8.7%) 16.6 3.20
present (53%) (2.51–48.5) (1.41–7.24)
From Blaha et al. [67], with permission
CHD coronary heart disease, CVD cardiovascular disease, CAC coronary artery calcium, MESA Multi-ethnic Study of Atherosclerosis, JUPITER
Justification for the Use of statins in Prevention: an Intervention Trial Evaluating Rosuvastatin
*
Hazard ratios of 1 were used as a reference
268 P. B. Morris and M. D. Shapiro

ated with a NNT of only 24. These findings led the authors to of the MESA population [48]. The study cohort consisted of
suggest that CACS may be used to identify patients most 6814 participants who were followed for a mean of
likely to benefit from statin therapy, focusing treatment on 10.3 years. The lowest proportion of total CVD, CHD, and
those patients with evidence of subclinical atherosclerosis. hard CHD events occurred in participants with CACS = 0.
One of the most important considerations when using Among all negative risk factors, CACS = 0 was the strongest
CACS = 0 to reclassify an individual to a lower ASCVD risk and resulted in the largest, most accurate downward risk
category is whether the absence of subclinical atherosclerosis reclassification (Fig. 23.9). CACS = 0 was particularly infor-
confers long-term protection that is incremental to and inde- mative in older adults with lower diagnostic likelihood ratios
pendent of traditional risk assessment tools. The “warranty with increasing age, making the test more useful for down-
period” (defined as <1% annual mortality) associated with the grading disease risk compared to younger adults.
absence of CAC was evaluated in 9715 consecutive asymp- In the BioImage Study of elderly people, the improved
tomatic individuals referred for CACS at a single site [44]. sensitivity, specificity, and NRI of a disease-guided approach
Participants were stratified by age and CVD risk (FRS and that incorporates CACS were largely driven by down-­
National Cholesterol Education Program Adult Treatment classifying, or “de-risking,” statin-eligible patients who had
Panel III [NCEP ATP III] categories) and were followed for a a CACS  =  0 [54]. CHD and CVD event rates were low in
mean of 14.6 years. Among the 4864 patients with CACS = 0, individuals with CACS = 0 and the absence of carotid plaque,
the annual observed rate of mortality remained <1% during even among participants with diabetes. Investigators suggest
the entire follow-up period with a warranty period of 15 years that withholding statin therapy in older adults without
(Fig.  23.8). Compared with the FRS or NCEP ATP III risk ­subclinical atherosclerosis could spare patients from lifelong
assessment tools alone, CACS significantly improved dis- statin therapy that may benefit only a few.
crimination when added to the models. At all levels of calcu-
lated risk, CACS improved risk classification (Table  23.5).
Interestingly, a similar study was performed in asymptomatic Low- and High-Risk Patients
diabetic subjects who were followed for a mean of 14.7 years.
A CACS = 0 in diabetic patients was associated with a favor- CACS  =  0 has also been demonstrated to provide important
able 5-year prognosis. However, after 5 years, the risk of mor- prognostic information among asymptomatic individuals with-
tality increased significantly among diabetics even in the out traditional risk factors who are considered to be at low- or
presence of a baseline CACS = 0 [68]. high-risk of ASCVD events. In a study cohort of 44,052 indi-
The role of negative risk markers for ASCVD (including viduals referred for assessment of CACS, patients were fol-
CACS = 0, cIMT <25th percentile, absence of carotid plaque, lowed for a mean of 5.6 years [69]. Risk factors were assessed
and other risk markers) was also evaluated in recent analysis by self-report, and patients were characterized as having 0, 1, 2,

Fig. 23.8  The 15-year 12.0

cumulative mortality rate for
low-risk individuals. (Adapted
from Valenti et al. [44], with 10.0
Cumulative Mortality Rate (%)

permission)

8.0

6.0

4.0

2.0

0.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Year

Overall (N = 9,715) FRS < 10% (N = 3,954) NCEP ATP III < 10% (N = 5,185)
No CV risk factors (N = 545) CAC = 0 (N = 4,864)
23  Clinical Application of the Coronary Artery Calcium Score and Implications for Cardiovascular Disease Prevention 269

Table 23.5  Net reclassification index with the addition of coronary artery calcium (CAC) Scoring to a Model Including the FRS or NCEP ATP
III Score
% of events correctly % of nonevents Noneventp
NRI 95% CI p Value reclassified Event p value correctly reclassified value
Overall cohort (N = 9715)
FRS + CAC 0.58966 0.5251– <0.0001 29 <0.0001 30 <0.0001
0.6542
NCEP ATP 0.57966 0.5149– <0.0001 28 <0.0001 30 <0.0001
III + CAC 0.6444
Low cardiovascular risk
FRS + CAC 0.52561 0.3961– <0.0001 12 0.07 41 <0.0001
0.6551
NCEP ATP 0.49967 0.3878– <0.0001 13 0.02 37 <0.0001
III + CAC 0.6116
Intermediate cardiovascular risk
FRS + CAC 0.46417 0.3611– <0.0001 23 <0.0001 23 <0.0001
0.5672
NCEP ATP 0.51544 0.3955– <0.0001 28 <0.0001 23 <0.0001
III + CAC 0.6354
High cardiovascular risk
FRS + CAC 0.58347 0.4709– <0.0001 40 <0.0001 18 <0.0001
0.6961
NCEP ATP 0.55956 0.4494– <0.0001 37 <0.0001 19 <0.0001
III + CAC 0.6698
From Valenti et al. [44], with permission
CI confidence interval, NRI net reclassification improvement
30%
10% 15% 20% 25%
5%
0%

0% 5% 10% 15% 20% 25% 30%

Pre-Test Risk

Zero CAC BNP <100 pg/mL

cIMT <25th percentile No Microalbuminuria
No Carotid Plaque No Family History
Flow-Mediated Dilation >5% No Family History of Premature CHD
Normal ABI No Metabolic Syndrome
hsCRP <2 mg/dL Healthy Lifestyle
Homocysteine <10 umol/L

Fig. 23.9 Negative risk markers for cardiovascular disease. were obtained by plotting the pretest and posttest risk on the basis of the
Relationship between pretest and posttest cardiovascular disease (CVD) diagnostic likelihood ratio of each MESA participant and then applying
risk after the knowledge of the negative result of each risk marker. The a linear fit. MESA indicates Multi-Ethnic Study of Atherosclerosis,
regression lines display the relationship between the pretest predicted ABI ankle-brachial index, BNP brain natriuretic peptide, CAC coro-
10-year CVD risk (x axis) and the posttest risk (y axis) after the knowl- nary artery calcium, CHD coronary heart disease, CIMT carotid intima-­
edge of the negative result of each risk marker. A broken back line is media thickness, FMD flow-mediated dilation, hsCRP high-sensitivity
displayed as reference (risk shift with no additional testing). Results C-reactive protein. (From Blaha et al. [48], with permission)
270 P. B. Morris and M. D. Shapiro

or  >  2 risk factors. A total of 19,898 patients (45%) had CAC. However, 65% of events occurred in individuals with
CACS = 0. As shown in Fig. 23.10, more than half of individu- only 0–1 lipid abnormality. CACS stratified ASCVD risk
als with no ASCVD risk factors (53%) had evidence of sub- across varying levels of dyslipidemia, and when CACS = 0,
clinical atherosclerosis with CACS > 0. The lowest all-cause the absolute rates of ASCVD events were relatively low
mortality occurred in those with no CAC and no ASCVD risk regardless of the dyslipidemia burden (Fig. 23.12). Even in
factors. Of interest, individuals with no risk factors but evi- the presence of three lipid abnormalities, when there was no
dence of significant disease with CACS > 400 had a substan- evidence of subclinical atherosclerosis, the absolute event
tially higher mortality compared to individuals with multiple rate over the 7.6-year follow-up period was <6%. In such
risk factors but CACS = 0 (16.89 versus 2.72 per 1000 person- individuals, the NNT with statin therapy was 154. The inves-
years). The authors suggested that the exclusive use of risk tigators suggested that CACS may help to personalize pre-
assessment tools which rely only upon traditional risk factors ventive strategies, especially in individuals with a history of
may be inadequate in all low- and high-risk patients for guiding statin-related side effects.
the intensity of primary prevention therapies.
The role of CACS in patients at the extremes of calculated
ASCVD risk was also evaluated in the MESA cohort [70]. A Diabetes
total of 6698 individuals were followed for a mean of
7.1 years. Fifty percent of the study population had CACS = 0, Diabetes was previously considered to be a CHD risk equiv-
and only 13% of all CHD events occurred in this group with alent by the NCEP ATP III guidelines due to the high inci-
a total event rate of 1.8 per 1000 person-years (Table 23.6). dence of ASCVD events among patients with this diagnosis
The annualized event rate increased slightly with increasing [72]. Patients with diabetes have been shown to have more
risk factors; however, the absolute event rate remained very extensive CAC and an increase in risk of events when com-
low at all levels of traditional risk factors. CACS was also pared to nondiabetic as discussed above. However, CACS = 0
able to further risk stratify individuals at the extremes of risk is associated with a low short-term risk of death, even in the
as estimated by the FRS (Fig. 23.11). These findings indicate presence of diabetes [58]. In a study of 10,377 asymptomatic
that assessment of subclinical coronary atherosclerosis may patients (903 diabetics), there are 4800 patients without dia-
play a role even in patients who are considered to be at low- betes and 267 patients with diabetes who had CACS  =  0.
or high-risk by traditional risk factor assessments. Over a mean follow-up period of 5.0 years, participants with
The MESA cohort was also used to examine two para- diabetes and CACS  =  0 demonstrated a survival similar to
digms of risk assessment, in this case dyslipidemia as a risk that of individuals without diabetes and CACS  =  0
factor and CACS as a measure of disease burden [71]. (Fig. 23.13).
Participants (N = 5534) were not on baseline lipid-lowering Participants in the Diabetes Heart Study (N = 1051) were
medications and were classified as having 0–3 lipid abnor- followed for a mean of 7.4 years [73]. The estimated annual
malities (including elevated LDL-C, low HDL-C, and/or risk of all-cause mortality was lowest in individuals with
high triglycerides). Over a median follow-up of 7.6  years, CACS = 0–9 (0.9%), compared to annual risk of 20.0% in
82% of events occurred in participants with evidence of those with CACS ≥ 1000.

Table 23.6  Coronary heart disease events by risk factor burden (A) and coronary artery calcium score (B)
All CHD events Total CHD event rate (per Hard CHD events Hard CHD event rate
n (%) (% of total events) 1000 person-­years) (% of total events) (per 1000 person-­years)
A
 0 RF 1067 (16%) 16 (5%) 2.1 11 (5%) 1.4
 1 RF 2310 (34%) 82 (24%) 5.0 50 (25%) 3.0
 2 RF 2116 (32%) 126 (37%) 8.7 83 (41%) 5.7
 ≥3 RF 1205 (18%) 115 (34%) 14.6 60 (29%) 7.4
B
 CAC 0 3349 (50%) 44 (13%) 1.8 31 (15%) 1.3
 CAC 1–100 1774 (27%) 85 (25%) 6.9 54 (26%) 4.3
 CAC 101–300 747 (11%) 73 (22%) 14.4 46 (23%) 8.8
 CAC >300 828 (12%) 137 (40%) 26.3 73 (36%) 13.3
From Silverman et al. [70], with permission
RFs considered: high LDL-C, low HDL-C, hypertension, diabetes, current cigarette smoking. Coronary heart disease event rates were significantly
different by RF burden (P < 0.001). Coronary heart disease event rates were significantly different by CAC score (P < 0.001)
23  Clinical Application of the Coronary Artery Calcium Score and Implications for Cardiovascular Disease Prevention 271

Fig. 23.10  The prevalence P<0.0001

and extent of coronary artery 100
calcium (CAC) in
90
asymptomatic patients
according to burden 80
traditional risk factors (RFs).
70
(From Nasir et al. [69], with
permission) 60
50
40
30
20
10
0
0 RF 1 RF 2 RF >2 RF
N=18,819 N=10,093 N=8,754 N=6,386
CAC=0 CAC 1-100 CAC 101-400 CAC>400

Fig. 23.11  Total (a) and hard a

(b) coronary heart disease 34.1
event rates (per 1000
35 26.8
person-years) with increasing
Event Rate per 1000 person-years

24.3
coronary artery calcium
scores according to 30
Framingham risk score 22.4
categories. (From Silverman 25
10.9 15.3
et al. [70], with permission)
20
8.0 10.0
15 13.2
7.9
3.4
10 3.5 CAC > 300

5 0.6 CAC 101–300

1.6 2.5 3.1
0 CAC 1–100
0 RF
1 RF CAC = 0
2 RF
3 RF
CAC = 0 CAC 1-100 CAC 101-300 CAC > 300

15.7
b 14.1
11.9 14.5
Event Rate per 1000 person-years

16

14 8.1

12 8.6
10 5.9 6.1

8 6.8
5.6
6
1.3 2.7
4 CAC > 300

2 0.6 1.1 2.1 CAC 101–300

0 1.4 CAC 1–100

0 RF CAC = 0
1 RF
2 RF
3 RF
CAC = 0 CAC 1-100 CAC 101-300 CAC > 300
272 P. B. Morris and M. D. Shapiro

Fig. 23.12 Atherosclerotic
cardiovascular disease events
per 1000 person-years by 29.3 29.5
strata of coronary artery 30 28.7
calcium score (CAC) and
lipid abnormality (LA). (From 22.7
25

Rate of CVD per 1,000

Martin et al. [71], with
permission) 20

person-years
15
11.7 15.3
10 7.4 10.5

5
≥100
2.7
0 2.8 4.1 5.9 1 to 99

CAC
0
1 0
LA 2
3

Fig. 23.13  Cox proportional 1.000

hazards cumulative survival
for subjects with and without No diabetes (n=4,800)
diabetes mellitus with a .990
calcium score of 0. The Diabetes (n=267)
number of subjects in each .980
calcium score category is in
parentheses. (Raggi et al.
[58], with permission) .970
Cumulative Survival

.960

.950

.940

.930

.920

.910
χ2=0.5, p=0.49
.900
0 1 2 3 4 5
Follow-up (Years)

In the MESA population, 38% of patients with diabetes  ymptomatic Patients: Stable and Acute
S
(n  =  881) had CACS  =  0 [58]. The annualized CHD and Chest Pain
CVD event rates in this group were 0.4% and 0.8%, respec-
tively, compared to CHD and CVD event rates of 4% and Assessment of CAC is primarily recommended for ASCVD
5.1%, respectively, in diabetics with CACS > 400. The ten- risk assessment in asymptomatic individuals, but the prog-
fold increase in CHD event rates according to the extent of nostic implication of CACS = 0 has also been evaluated in
CAC suggests that CACS may play an important role in patients presenting with symptoms of suspected CHD [60,
improvement in risk stratification in patients with diabetes. 74]. The CONFIRM (Coronary CT Angiography Evaluation
23  Clinical Application of the Coronary Artery Calcium Score and Implications for Cardiovascular Disease Prevention 273

for Clinical Outcomes: An International Multicenter) myocardial perfusion imaging, stress echocardiography,
Registry evaluated the prevalence and severity of CAD in CCTA, or invasive coronary angiography based on presenta-
relation to prognosis in symptomatic patients who under- tion and risk assessment.
went coronary angiography and concomitant CCTA [75].
Among 10,037 individuals, 51% had CACS  =  0. Among
these patients, CCTA showed no evidence of CAD in 84% of Summary
patients. Approximately 13% of patients had nonobstructive
CAD, 3.5% had ≥50% stenosis, and 1.4% had ≥70% steno- Though there is overwhelming and consistent evidence that
sis. CACS = 0 had a sensitivity and specificity for stenosis the absence of CAC is associated with a low risk of mortality
≥50% of 89% and 59%, respectively. The negative and posi- (warranty up to 15  years) and hard CHD events (≥7  years
tive predictive values were 96% and 29%, respectively. based on MESA) in multiple patient populations, there is not
During the 2.1-year period of follow-up, there was no differ- yet clear randomized clinical trial (RCT) evidence that
ence in mortality among patients with CACS = 0 regardless CACS = 0 justifies avoidance of prolonged use of preventive
of the presence of obstructive CAD. However, CACS did not therapies such as statins [79]. Concern remains that ASCVD
add incremental prognostic information compared with risk factors have a long latency period of effect and that even
extent of CAD at CCTA for the composite endpoint of mor- though patients with no evidence of CAC are at low near-­term
tality, MI, or coronary revascularization. Among patients risk, they may have significant lifetime risk of events [80]. In
with CACS = 0, the presence of ≥1 coronary stenosis ≥50% addition, in high-risk symptomatic patients, the absence of
was predictive of nonfatal MI and late revascularization. CAC is not sufficient to rule out the presence of obstructive
The CRESCENT (Computed Tomography vs. Exercise atherosclerotic coronary artery disease. The 2018 ACC/AHA
Testing in Suspected Coronary Artery Disease) trial evalu- blood cholesterol guidelines recommend that in individuals at
ated the effectiveness and safety of a CCTA algorithm com- intermediate risk with 10-year ASCVD risk of >7.5% to
pared with functional testing with a symptom-limited <20%, if the risk decision is uncertain, calcium scoring may
treadmill stress test in patients with stable angina referred to be considered in selected adults [42]. Authors note that a
an outpatient clinic [76]. Myocardial perfusion imaging or CACS = 0 is useful to reclassify patients to a lower-risk
stress echocardiography was performed when there was a group, often allowing statin therapy to be delayed or witheld
contraindication to treadmill testing or in cases of non-­ unless additional high-risk conditions are present. For patients
interpretable or equivocal results. The tiered cardiac CT pro- with a CAC score of 0, it is currently uncertain when follow-
tocol included CACS followed by CCTA if CACS > 1 to up studies should be performed, though many experts con-
400. A total of 350 patients with stable angina were prospec- sider repeating evaluation at 5–10 years.
tively randomized 2:1 to cardiac CT (N = 242) or functional
testing (N  =  108). In the CT group, 100 patients had
CACS = 0 and 2 went on to CCTA due to pretest probability  ACS-Guided Prevention and Adherence
C
of CAD > 70%. During the mean 1.2-year follow-up period, to Therapies
none of the 98 patients with CACS  =  0 underwent further
testing, and no adverse events occurred. The authors In view of the numerous studies demonstrating that CACS
­suggested that a tiered cardiac CT protocol offers an effec- can improve risk assessment in primary prevention, there has
tive and safe alternative to functional testing in patients with been strong interest in understanding whether noninvasive
stable chest pain. imaging can improve cardiovascular outcomes by guiding
Among 2088 patients presenting to the hospital with a the implementation and intensity of preventive strategies and
chest pain syndrome (considered atypical or non-anginal whether knowledge of CACS can increase adherence to pre-
chest pain by clinicians) who underwent CCTA, CACS = 0 ventive therapies [81].
was identified in 1114 patients (53%) [77]. Obstructive CAD The EISNER (Early Identification of Subclinical
was found in 48 patients (4.3%), and during the mean 2.8-­ Atherosclerosis by Noninvasive Imaging Research) study
year follow-up, there were 14 CHD events (1.3%) among the was a prospective RCT to evaluate the effects of CACS on
CACS = 0 patients. CAD severity was a strong predictor of risk factor management and downstream diagnostic testing
events in this group. [82]. Trial participants (N = 2137) were middle-aged (mean
The 2013 ACCF/AHA/ASE/ASNC/HFSA/SCAI/SCCT/ 58.5 years) individuals with established ASCVD risk factors
SCMR/STS Multimodality Appropriate Use Criteria for the who were randomized 2:1 to undergo CACS or no imaging
Detection and Risk Assessment of Stable Ischemic Heart and followed for a mean of 4  years. All subjects received
Disease indicate that CACS is rarely appropriate for evalua- standardized risk factor counseling. Compared with the no-­
tion of acute chest pain among patients at any level of base- CACS group, those who underwent CACS had greater
line ASCVD risk [78]. Preferred diagnostic strategies include improvement in systolic blood pressure, LDL-C, and waist
274 P. B. Morris and M. D. Shapiro

circumference (for those with elevated baseline waist girth). RCTs of CACS-guided ASCVD prevention strategies [7].
There was no difference in net utilization of downstream However, there also have not been RCTs demonstrating
diagnostic testing, such as stress tests, carotid ultrasound, reduction in ASCVD events based upon risk assessment
invasive coronary angiogram, or revascularization proce- tools or other biomarkers, such as hs-CRP [87, 88]. In view
dures. Approximately two-thirds of subjects with CACS of the limitations of current risk assessment tools which are
≥400 underwent some form of noninvasive stress testing, but strongly influenced by age and the increasing numbers of
the frequency of cardiac catheterization and revasculariza- statin-eligible individuals, there is interest among many
tion was substantially lower among patients who underwent stakeholders to conduct such a trial. Recent processes for
CACS compared to those who did not. In patients with guideline development have restricted consideration to only
CACS  =  0, there was a 37% reduction in procedure costs. large RCTs and have limited the use of observational data,
These results were consistent with the hypothesis that CACS thereby possibly necessitating such a CACS-guided RCT to
can improve ASCVD risk factor management without sig- strengthen recommendations for use of imaging to personal-
nificant increases in downstream health-care costs.
A recent systematic review examined the effects of CACS a 60
on behavioral or lifestyle modification, perception of risk for Calcium=0
CHD, and medication adherence [83]. Of the 15 studies 50 Calcium>0
reviewed, 3 were RCTs and 12 were observational. The
strongest impact of CACS was found to be improved medi- 40
cation utilization and adherence. The results on lifestyle

Percent
modification and risk perception were either nonsignificant 30
or inconsistent. In an observational study of 980 asymptom-
atic patients who underwent assessment of CAC, initiation of 20
aspirin, improvement in dietary habits, and increase in exer-
cise were strongly associated with greater CACS [84]. The 10
prospective incidence of aspirin and statin use was assessed 1640 1546 1152 1058 978 716
in the PACC (Prospective Army Coronary Calcium) Project 0
li ne r1 r2 3 4
[85]. Men ages 40–50 years were screened for CHD risk fac- se a a ar ar a r5 ar6
Ba Ye Ye Ye Ye Ye Ye
tors and CACS. Preventive medications were not prescribed Follow-up Year
by the study investigators. Over the subsequent 6-year fol- b 60
low-­up period, patients with evidence of CAC were three Calcium=0
times more likely to be prescribed a statin and more likely to 50 Calcium>0
receive aspirin than those with CACS = 0. CAC was strongly
and independently associated with the use of statin, aspirin, 40
Percent

or both (Fig. 23.14). In the largest RCT, the EISNER trial,

30
the CACS group were noted to have high adherence rates for
lipid-lowering and antihypertensive medications [83].
20
The motivational effects of CACS on adherence to statin
therapy and behavioral modification for weight loss were ret- 10
rospectively analyzed in 2608 patients who had underwent
1640 1546 1152 1058 978 716
baseline study and returned for follow-up at a mean of 0
e
4.1 years [86]. Statin adherence was lowest (27.4%) among el
in
a r1 a r2 ar
3
ar
4
r5 r6
s a a
those with CACS = 0 and increased with increasing levels of Ba Ye Ye Ye Ye Ye Ye
Follow-up Year
CAC to 58.8% adherence among those with CACS  >  400.
Similarly, behavioral modification resulting in weight loss Fig. 23.14  Incidence of statin (a) and aspirin (b) use during 6-year
was lowest (19.8%) among those with CACS = 0 and gradu- actuarial follow-up in the PACC Project Cohort. (a) Men only; n = 1640.
ally increased to 33.6% among those with CACS > 400. This Ever use of a statin was noted in 23% of participants, including 48.5%
study is the largest study to date demonstrating that the of those with coronary artery calcium and 15.5% of those without coro-
nary artery calcium (p < 0.001), which remained significant after con-
CACS results are associated with improved adherence with trolling for NCEP risk variables (odds ratio 3.53; 95% CI 2.66–4.69).
ASCVD risk reduction behaviors. (b) Men only; n-1640. Ever-use of aspirin was noted in 31.2% of par-
Current guidelines have generally only provided a level ticipants, including 51.5% of those with coronary artery calcium versus
IIB (“may be considered”) endorsement for addition of 25.3% of those without coronary artery calcium (p  <  0.001), which
remained significant after controlling for NCEP risk variables (odds
CACS into risk assessment algorithms. This tepid recom- ratio 3.05; 95% CI 2.30–4.05). (Adapted from Taylor et al. [85], with
mendation is largely due to the lack of large, prospective permission)
23  Clinical Application of the Coronary Artery Calcium Score and Implications for Cardiovascular Disease Prevention 275

ize preventive strategies [89]. The numerous considerations  rogression of CACS and Role of Serial
P
in conducting such a large RCT for CACS are discussed in a Scanning
recent review and include cost associated with such a trial
and previous evidence for cost-effectiveness of CACS, expo- Multiple retrospective and prospective observational studies
sure to ionizing radiation, trial feasibility and adequate sta- have suggested that progression of CAC is associated with
tistical power, ethical considerations in randomizing patients increased risk of CHD events and predicts all-cause mortal-
with extensive CAC to placebo, selection of patients based ity [91, 92]. In the MESA cohort, the absolute change in
on the level of risk-based factor risk assessment, and chal- CACS was assessed from baseline study to follow-up scan
lenges of blinding [82]. Authors also propose consideration obtained approximately 2.5  years later [93]. Among those
of incorporating such a trial into current guideline-based individuals with CACS > 0 at baseline, greater change in
­recommendations for low-dose CT screening for lung cancer absolute CACS was associated with higher cumulative
for persons 55–74 years of age with >30-pack year history of ­proportion of incident CHD events (Fig. 23.15). Those with
smoking. annual progression of ≥300 had adjusted HRs of 3.8 and 6.3
The ROBINSCA (Risk Or Benefit IN Screening for for total and hard CHD events, respectively, compared to
Cardiovascular Disease) trial is currently being conducted individuals without CAC progression. For patients with
in Europe and will include 33,000 individuals from the CACS = 0 at baseline, compared to those with no increase in
general population [90]. Participants will be randomized CAC, any increase was associated with 1.4- and 1.5-fold
(non-­blinded) to systematic screening by traditional greater risks for total and hard CHD events, respectively.
ASCVD risk factors (SCORE algorithm), systematic A cohort of 4609 asymptomatic individuals underwent
screening by CACS, or no intervention (control). repeat CACS with a mean interscan time of 3.1 years [94].
Participants in the risk factor-­based intervention group with CAC progression was evaluated as the absolute change in
intermediate or high risk for ASCVD are referred to a pri- CACS, percent annualized difference between scans, and the
mary care provider (PCP) for guideline-based preventive difference between square root of the baseline and square
management. Participants in the CACS-based intervention root of follow-up CACS > 2.5 (the “SQRT” method). Among
group with a high or very high risk for ASCVD based on patients with baseline CACS  >  0, CAC progression added
their CACS are also referred to a PCP and for treatment incremental value in predicting all-cause mortality compared
according to the study protocol. All participants will be fol- to baseline score and traditional risk factors. Progression in
lowed for 5 years, and the primary study outcome is 5-year the setting of increasing baseline CACS (1–10, 11–100,
fatal and nonfatal CHD.  This study is currently enrolling 101–400, and >400) was associated with increasing risk of
with estimated completion in 2022. all-cause mortality. At all levels of CACS, the risk for all-

Fig. 23.15 Kaplan-Meier
0.40

plot of cumulative incidence CAC Change per Year

of total CHD among persons
300+ (2%)
with CAC > 0 at baseline.
Numbers in parentheses 200-300 (3%)
indicate proportion of subjects
0.30
Cumulative incidence of total CHD

100-200 (12%)
in each group. Imputed and
nonimputed subjects are .001-100 (68%)
included. For log-rank test (-) or None (15%)
across CAC change groups,
p < 0.001. CHD indicates
0.20

coronary heart disease, CAC

coronary artery calcium score.
(From Budoff et al. [91], with
permission)
0.10 0.00

0 1 2 3 4 5 6 7 8
Years from baseline exam
 276 P. B. Morris and M. D. Shapiro

a a 2500 Men
1.00 b: Median b1 b: 75. PCTL b1 b: 90. PCTL b1
CAC = 0 b: Median
5y: Median b1
b: 76 PCTL
5y: 75. PCTL b1
b: 90 PCTL
5y: 90. PCTL b1
CAC > 1 to 100 2250 5y: Median 5y: 75. PCTL 5y: 90. PCTL

CAC > 100 to 400

2000
Survival Probability

CAC > 400

1750
1500

CAC score
0.75
1250
1000
750

0.50 500
250
0
40 45 50 55 60 65 70 75 80 85
0 5 10 15 20 Age (Years)
Follow-up (Years)
In Non-Progressors b 1000 Women
b 1.00 b: Median b1
b: Median
b: 75. PCTL b1
b: 76 PCTL
b: 90. PCTL b1
b: 90 PCTL
5y: Median b1 5y: 75. PCTL b1 5y: 90. PCTL b1
900 5y: Median 5y: 75. PCTL 5y: 90. PCTL
CAC = 0
800
Survival Probability

700
CAC > 1 to 100
600
CAC score
0.75
500

CAC > 100 to 400 400

CAC > 400 300

0.50 200
100
0
40 45 50 55 60 65 70 75 80 85
0 5 10 15 20 Age (Years)
Follow-up (Years)
In Progressors Fig. 23.17  Age- and gender-related percentiles of CAC distribution
follow an exponential curvature. Observed and fitted 50th, 75th, and
Fig. 23.16  Survival curves demonstrating time to all-cause mortal- 90th percentile of the CAC distribution for men (n  =  1633) (a) and
ity for nonprogressors and progressors. Cox proportional hazards women (n = 1848) (b) by age categories. In dark colors for the baseline
survival curves demonstrating time to all-cause mortality for patients values, when the participants (1633 men) were aged between 45 and
with a yearly change using survival curve of progression based on 74  years, and in light colors for the 5-year follow-up data, when the
our best-­fitting model of square root >2.5 according to baseline cal- cohort was aged 50–79 years. Note the exponential shape of the increase
cium score: (a) nonprogressors and (b) outcomes in progressors. of CAC. Dots represent observed percentile values for each 5-year age
CAC indicates coronary artery calcium score. (From Budoff et  al. category; lines show linear quantile regression on a log scale after
[91], with permission) retransformation. The y-axis range in (a) and (b) differs by a factor of
2.5  in men compared with women. (From Erbel et  al. [94], with
permission)

cause mortality was higher among those with significant pro- over 5  years of follow-up (Fig.  23.17) [94]. Based on the
gression compared to those with less progression (Fig. 23.16). observed CACS distribution of the cohort, investigators
In the MESA cohort, individuals with the metabolic syn- developed a mathematical tool to predict CAC progression
drome, diabetes, or both were demonstrated to have greater for individuals in the study. The model indicated that pro-
progression of CAC compared to individuals without these gression of CAC is nearly inevitable and is only modestly
conditions. Furthermore, progression predicted CHD events influenced by adjustment for risk factors and lipid lowering
among those in the highest tertile of CAC increase versus no and antihypertensive medications.
increase [93]. Continued progression of CAC in patients currently being
Data from HNR demonstrated that age- and gender-­ treated with statin therapy has been shown to be associated
related percentiles of CAC distribution follow an exponential with increased risk of MI in patients with and without
curve, with a nearly indistinguishable shift along the ­baseline ­diabetes [95].
23  Clinical Application of the Coronary Artery Calcium Score and Implications for Cardiovascular Disease Prevention 277

Preventive Therapies and CAC Progression age and gender [103]. Participants were followed for a mean
of 4.3  years. Though treatment reduced LDL-C by 43%,
In a systematic review of trials of preventive therapies and there was no effect on progression of CAC in the treatment
progression of CAC, patients with clinical ASCVD received group compared to individuals receiving placebo. The mean
treatments including statins, antihypertensive agents, or pla- increase in CACS was 331  ±  421  in the treatment group
cebo [96]. Patients with chronic kidney disease received compared with 323 ± 385 in the control group. Though there
interventions including low phosphorous diet, sevelamer were fewer ASCVD events in the overall treatment group,
hydrochloride, and calcium-based phosphate binders. There this difference did not achieve statistical significance
was no significant effect of any of these therapies on mean (p  =  0.08). However, among those individuals with
weighted annualized CAC increase in any of these patient CACS > 400 (47% of study population), treatment reduced
populations. the risk of events by 42% (p = 0.046).
The effects of statin therapy on plaque composition, Recent serial IVUS studies have provided new insights
plaque progression, and the extent and progression of CAC into the role of statins and plaque composition, progression
are of specific interest. The dose-dependent beneficial effects of calcification, and plaque stability. An analysis of IVUS
of statins on slowing progression of coronary atheroma vol- data from eight large multicenter clinical trials assessing the
ume and plaque regression have been well established with impact of medical therapies on serial changes in coronary
intravascular ultrasound (IVUS) [97, 98]. Given the docu-
mented correlation between the extent of CAC and total a p=0.007
coronary plaque burden, one could hypothesize that statin p=0.03
therapy would be associated with a decrease in progression 1.5 0.12
of CAC [99]. However, despite promising early data, large
0.08
Change in PAV (%)

Change in Cal
prospective trials have failed to demonstrate a reduction in 0.75
CAC with low-, moderate-, or high-intensity statin [100]. 0.04
Investigators for SALTIRE (Scottish Aortic Stenosis and 0
0
Lipid Lowering Trial: Impact on Regression) randomized
102 patients with calcific aortic stenosis and CAC to atorvas- -0.04
-0.75
tatin 80  mg or placebo [101]. After a mean follow-up of -0.08
2 years, despite significant reductions in LDL-C and hs-CRP, p<0.001
-1.5 -0.12
there was a nonsignificant increase in CAC progression in p<0.001
the atorvastatin group (26% per year) compared to the
HIST LIST No-Statin
placebo-­treated group (18% per year).
The BELLES (Beyond Endorsed Lipid Lowering with b p=0.004

EBT Scanning) trial was a randomized, double-blind, multi- p=0.01

8 0.12
center study of atorvastatin 80 mg or pravastatin 40 mg daily
Change in TAV (mm3)

in 615 hyperlipidemia women without clinical ASCVD 0.08

Change in Cal
[102]. Women underwent baseline and 12-month CACS to 4
0.04
measure percent change in total and single artery calcium
volume score from baseline, as well as absolute changes in 0 0
CACS.  Among the 475 women who completed the study, -0.04
LDL-C was reduced by approximately 47% in those ran- -4
-0.08
domized to atorvastatin and 25% in those randomized to
pravastatin. Although, high-intensity statin therapy caused a p<0.001 -0.12
-8
greater reduction in atherogenic lipoproteins, there was no p<0.001
effect upon any measure of CAC progression. In the absence HIST LIST No-Statin
of a placebo group, it was not possible to determine if statin
therapy similarly reduced progression of CAC in both treat- Fig. 23.18  Statins and Coronary Plaque Calcification: Changes in
Coronary Atheroma Volume and Calcium Indices According to Therapy.
ment groups. In addition, the short duration of the follow-up
(a) Percent atheroma volume (PAV)-adjusted model, depicting corre-
may have precluded observation of longer-term effects of sponding changes in PAV and calcium index (CaI). (b) Total atheroma
statin therapy. volume (TAV)-adjusted model depicting corresponding changes in TAV
The St. Francis Heart Study was a randomized, double-­ and CaI. Changes in PAV and TAV (blue boxes) are reported as least
squares mean ± standard error of the mean, whereas the changes in CaI
blind, placebo-controlled trial of atorvastatin 20  mg daily,
(salmon boxes) are reported as median (interquartile range). CI confi-
vitamin C, and vitamin E in 1005 asymptomatic individuals dence interval, HIST high-intensity statin therapy, LIST low-intensity
on baseline aspirin 81 mg with CACS ≥80th percentile for statin therapy. (From Puri et al. [104], with permission)
278 P. B. Morris and M. D. Shapiro

atheroma burden included two trials of intensive lipid low- cular smooth muscle cell calcification, possibly promoting
ering with statins (REVERSAL [Reversal of Atherosclerosis plaque stability (Fig. 23.19).
with Aggressive Lipid Lowering] and SATURN [the Study The effect of high-intensity statin on plaque composition
of Coronary Atheroma by Intravascular Ultrasound: Effect and plaque burden by serial IVUS was also assessed in 82
of Rosuvastatin Versus Atorvastatin]) [104]. High-intensity patients following non-ST segment elevation MI in IBIS-4
statin therapy was associated with mean 36.2% reduction in (Integrated Biomarker Imaging Study) [105]. Patients were
LDL-C and significant regression in percent atheroma vol- treated with rosuvastatin 40 mg daily, and IVUS with analy-
ume (PAV) from baseline during 18- to 24-month follow-up, sis of matched segments was performed at baseline and
whereas low-intensity statin and no statin therapy were 13  months. Statin therapy was associated with significant
associated with significant PAV progression (Fig.  23.18a). reductions in percent atheroma volume, percent fibrous
All treatment groups had significant progression of CAC plaque volume, and necrotic core volume. The proportion of
from baseline, and the change in CAC was greatest in the calcified plaque components was, however, significantly
statin-­treated groups versus the no-statin group (Fig. 23.18b). increased (+1.28%; CI 0.66–1.90%; p < 0.0001).
The change in CAC was not statistically greater in the high-­
intensity statin versus low-intensity statin comparison. The
greatest increases in CAC were noted in patients on high-­ Summary
intensity statin with significant plaque regression, whereas
in patients on no-statin therapy, there was profound progres- Progression of CAC is clearly associated with an increased
sion of coronary atheroma with smaller increases in risk of future ASCVD events. Despite significant event
CAC.  Authors suggest that evidence supports the role of reduction in patients treated with statin therapy and evidence
statin therapy in de-lipidation of atheroma and probable vas- of reduction in atheroma volume by IVUS, CAC progression

No Statin Therapy
• Plaque Progression
• Micro-calcification within Lipid Pools

High-Intensity Statin Therapy

• Plaque Regression
• Delipidation
• ? Vascular Smooth Muscle Cell Calcification

Fig. 23.19  Plaque Calcification in the Setting of No-Statin Therapy or Following long-term high-intensity statin therapy, plaque regression mani-
High-Intensity Statin Therapy. Natural plaque progression likely involves fests as delipidation and probable vascular smooth muscle cell calcifica-
lipid-pool expansion coupled with microcalcifications within lipid pools. tion, promoting plaque stability. (From Puri et al. [104], with permission)
23  Clinical Application of the Coronary Artery Calcium Score and Implications for Cardiovascular Disease Prevention 279

is not retarded and may even be accelerated on treatment. incorporation of CACS into models of risk assessment and
Greater progression of CAC on statin therapy has been inter- risk reduction were offered in this summary.
preted by many as benign conversion of pre-existing non-­ The NCEP ATP III released guidelines for detection, eval-
calcified to calcified plaque. However, greater progression of uation, and treatment of high blood cholesterol levels in
CAC on statin therapy has been associated with higher major 2002 [72]. Authors supported the position of the Prevention
adverse cardiovascular event rates and does not support this V conference writers, recommending against routine CACS
prevailing view. Instead, progression of CAC, irrespective of in asymptomatic persons for risk assessment. However, it
statin therapy, supports the predominant formation of new was felt that data supported a higher risk of major coronary
atherosclerotic plaque that then becomes calcified. At the events in persons with advanced subclinical atherosclerosis.
current time, there are no medical therapies which have been The ATP III guidelines supported the “measurement of coro-
demonstrated to achieve reductions [98]. nary calcium as an option for advanced risk assessment in
The 2013 ACCF/AHA/ASE/ASNC/HFSA/SCAI/SCCT/ appropriately selected persons, provided the test is ordered
SCMR/STS Multimodality Appropriate Use Criteria for the by a physician who is familiar with the strengths and weak-
Detection and Risk Assessment of Stable Ischemic Heart nesses of noninvasive testing.” The guideline writers sug-
Disease indicate that repeat CACS is rarely appropriate for gested that the presence of subclinical atherosclerosis could
patients at any level of ASCVD risk or any baseline level of be used to justify intensification of cholesterol therapy in
CAC [79]. intermediate-risk patients.
The first gender-specific guidelines for the use of noninva-
sive testing in the evaluation of women with suspected CHD
Evolution of Guidelines, Recommendations, were published in 2005 [108]. Women develop coronary ath-
and Appropriate Use Criteria for Calcium erosclerosis approximately a decade later than men, and the
Scoring in ASCVD Risk Assessment occurrence of CAC tracks this incidence of clinical events.
These differences make it important to consider gender-­
In 2000, the ACC Task Force on Clinical Expert Consensus specific criteria when evaluating the clinical significance of
Documents published the first recommendations for CACS CACS.  In the noninvasive testing guidelines for women,
in the diagnosis and prognosis of coronary artery disease CACS was characterized as an “emerging imaging” modality,
[106]. The Writing Group reviewed literature published along with CCTA, magnetic resonance imaging, and carotid
between 1988 and 1999, which at that time primarily utilized intima-media thickness. By the time of this report, a number
electron beam computed tomography (EBCT). Findings of of studies were available confirming that the presence and
this meta-analysis demonstrated a high sensitivity of EBCT severity of CAC have independent and incremental value
for the presence of obstructive CHD, a much lower specific- when added to major ASCVD risk factors in estimation of risk
ity, and an overall accuracy of approximately 70%. Available of MI or CHD death. Despite limited data in women, this con-
data at that time did not clarify whether the presence of CAC sensus statement proposed that CACS is useful in risk stratifi-
was additive to the FRS for risk assessment in asymptomatic cation in women, particularly in women at intermediate CHD
patients. Additional information was felt to be needed on the risk. Since the publication of these guidelines, MESA has pro-
value of a positive or negative calcium score in intermediate-­ vided a large prospective database for evaluation of age-, gen-
risk individuals for risk adjustment. The test was felt to be der-, and ethnic-related percentiles for CACS [17, 18].
equivalent but not superior to alternative noninvasive imag- Given technological advancements and the rapidly evolv-
ing methods such as single- photon emission computed ing literature, the AHA reissued a scientific statement on
tomography (SPECT) imaging. CCT [109] in 2006 and an expert consensus document on
The Writing Group III of the American Heart Association CACS in 2007 [110]. The AHA scientific statement reviewed
(AHA) Prevention Conference V: Noninvasive Tests of the most current evidence for the use of CCT for noninvasive
Atherosclerotic Burden also considered the role of CACS in imaging of coronary atherosclerosis. Though the authors
assessment of CHD risk in 2000 [107]. While acknowledg- addressed CCT with and without contrast enhancement, the
ing CAC as a definitive marker of atherosclerosis, the authors current discussion focuses on recommendations provided for
found relatively few prospective studies directly linking non-contrast-enhanced CCT for assessment of CAC.
CACS to subsequent CHD outcomes. The extent to which For reproducible and valid measurement of CAC, the 2006
CACS predicted events independently of traditional risk fac- AHA Writing Group established minimum criteria for scan-
tors was also not evident from the data at that time. It was ning. Studies comparing measurement of CACS with EBCT
proposed that CAC had the greatest potential for risk adjust- and MDCT demonstrate good correlation over a large range
ment in patients who had an intermediate risk score as calcu- of values in large, multiethnic population-based studies [111].
lated by one of the above risk scoring algorithms, but further However, to minimize motion artifacts, reduce radiation
research was suggested. No definitive recommendations for exposure, and improve intra- and inter-scan reproducibility, it
280 P. B. Morris and M. D. Shapiro

was recommended to use an EBCT or at least four-slice treatment decision is uncertain, CACS may be considered to
MDCT scanner and to maintain consistency in gating, trig- inform treatment decision making (Class IIb, LOE B) [7].
gering, and slice thickness parameters. The Agatston score According to the 2013 ACCF/AHA/ASNC/HFSA/HRS/
remains the primary tool for risk assessment because of the SCAI/SCCT/SCMR/STS Multimodality Appropriate Use
availability of large prognostic databases that allow clinicians Criteria for the Detection and Risk Assessment of Stable
and patients to understand the implication of any given score. Ischemic Heart Disease, CACS “may be appropriate” in
The 2007 ACCF/AHA Clinical Expert Consensus patients with intermediate or high global CHD risk regard-
Document on Coronary Artery Calcium Scoring by less of baseline ECG or ability of patient to exercise [79].
Computed Tomography in Global Cardiovascular Risk However, CACS is considered “rarely appropriate” for
Assessment and in Evaluation of Patients with Chest Pain asymptomatic patients with low global CHD risk or in symp-
represented an updated review of scientific data published tomatic patients.
subsequent to the original ACC/AHA Expert Consensus doc- The American College of Radiology Appropriateness
ument on EBCT published in 2000 [112]. The focus of the Criteria for the Asymptomatic Patient at Risk for Coronary
Writing Committee was on clinical application of CAC Artery Disease provides a rating scale for CACS as follows:
quantification by cardiac CT in two areas: (1) CHD risk 1,2,3, usually not appropriate; 4,5,6, may be appropriate; and
assessment in the asymptomatic patient to improve selection 7,8,9, usually appropriate [115]. For the intermediate-risk
of candidates for preventive interventions and (2) to improve patient, CACS is rated 8 or usually appropriate. However, for
triage of symptomatic patients, refining selection of candi- patients at high and low ASCVD risk, CACS is rated 3 or
dates for subsequent hospitalization or further diagnostic usually not appropriate.
testing. At the time of this document, significant scientific The 2016 European Society of Cardiology Guidelines on
data was available supporting a directly proportional rela- Cardiovascular Disease Prevention in Clinical Practices rec-
tionship between risk of CHD death or nonfatal MI, overall ommend that “coronary artery calcium scoring may be con-
CHD events, all-cause mortality, and the extent of CAC, as sidered as a risk modifier in CV risk assessment” (Class IIb,
previously described above. Thus, the measurement of CAC LOE B) [116]. However, routine screening with imaging
for risk assessment was considered “reasonable” in asymp- modalities to predict future ASCVD events is generally not
tomatic patients with an intermediate (10–20% absolute recommended in clinical practice.
10-year) risk of MI or CHD death by the FRS (Class IIb, The 2018 ACC/AHA blood cholesterole guidelines rec-
Level of Evidence B). The presence of significant CAC in an ommend calcium scoring for patients at intermediate risk
intermediate-risk individual could appropriately reassign the (>7.5% to <20%) when the risk decision for statin therapy is
patient to a higher-risk category and justify more aggressive uncertain (Class IIb, LOE B-NR). For patients with CAC = 0
reduction of CHD risk factors [113]. If an intermediate-risk and no additional high-risk clinical features, consideration
patient had CACS = 0 or a very low score, the risk stratifica- may be given to withholding or delaying initiation of statin
tion may be reduced. The probability of a CHD event in a therapy [42].
patient without CAC present was approximately 0.1–0.6% Nearly 7.1 million individuals undergo diagnostic non-­
per year (10-year CHD risk of 1.0–6.3%, consistent with contrast, non-ECG-gated, chest CT examinations each year
FRS low-risk <10%) based on data available at that time. in the USA [117]. Recommendations by the 2014 US
Imaging of CAC was not recommended for patients at low Preventive Services Task Force for routine lung cancer
risk of CHD (<10% 10-year risk) or for high-risk (>20% screening in high-risk individuals (based on age and smok-
10-year risk) patients (Class III, Level of Evidence B). ing history) may increase the number of low-dose screening
The 2010 ACCF/AHA Guideline for Assessment of CT scans by another 7–10 million annually [118]. In 2017,
Cardiovascular Risk in Asymptomatic Adults evaluated the Society of Computed Tomography and the Society of
CACS among other emerging risk markers and provided the Thoracic Radiology published consensus guidelines for the
highest endorsement for its clinical application thus far assessment of CAC from non-contrast chest CT scans [119].
[114]. CACS was considered reasonable for CVD risk CACS on non-gated CT examinations has been shown to
assessment in asymptomatic adults at intermediate risk (10– correlate well with scores obtained from conventional ECG-­
20% 10-year risk) as defined by the global FRS (Class IIa, gated scans and provide similar prognostic value in predict-
Level of Evidence [LOE] B). CACS was also considered rea- ing ASCVD events [120–123]. The guidelines recommend
sonable in persons at low to intermediate risk (6–10% incorporation of CACS into all non-contrast chest examina-
10-year risk) (Class IIb recommendation, LOE B). CACS tions as the appropriate standard of care. This is an extension
was not recommended for persons at low risk (<6% 10-year of the already existing requirement of the American College
risk) (Class III: No Benefit, LOE B). of Radiology to report “coronary arterial calcification mod-
As previously noted, the 2013 ACC/AHA Guideline on the erate or severe” when present in all lung cancer screening
Assessment of Cardiovascular Risk indicates that if a risk-based registry patients [124].
23  Clinical Application of the Coronary Artery Calcium Score and Implications for Cardiovascular Disease Prevention 281

Conclusion of fatal cardiovascular disease in Europe: the SCORE project. Eur

Heart J. 2003;24:987–1003.
7. Goff DF, Lloyd-Jones DM, Bennett G, et  al. 2013 ACC/AHA
Cardiovascular disease continues to be the leading cause of guideline on the assessment of cardiovascular risk. J Am Coll
morbidity and mortality around the globe, and CACS is an Cardiol. 2014;63:2935–57.
important tool to readily identify individuals at high risk of 8. Ridker PM, Paynter NP, Rifai N, et  al. C-reactive protein and
parental history improve global cardiovascular risk prediction: the
events to initiate early lifesaving preventive interventions. Reynolds risk score for men. Circulation. 2008;118:2243–51.
The assessment of subclinical coronary atherosclerosis by 9. Ridker PM, Buring JE, Firai N, Cook NR.  Development and
CACS has been shown to be a robust predictor of both near- validation of improved algorithms for the assessment of global
and long-term clinical ASCVD events and provides infor- cardiovascular risk in women: the Reynolds Risk score. JAMA.
2007;297:611–9.
mation that outperforms traditional risk factor-based 10. DeFilippis AP, Young R, Carrubba CJ, et  al. An analysis of
assessments. Current guidelines focus on the consideration calibration and discrimination among multiple cardiovascu-
of CACS in intermediate-risk patients to inform clinical lar risk scores in a modern multiethnic cohort. Ann Intern Med.
decision making for initiation of preventive therapies. 2015;162:266–75.
11. Akosah KO, Schaper A, Cogbill C, Schoenfeld P. Preventing myo-
Knowledge of the subclinical disease can improve adher- cardial infarction in the young adult in the first place: how do the
ence to preventive medications and lifestyle interventions. National Cholesterol Education Panel III guidelines perform? J
There is expanding evidence that imaging for subclinical Am Coll Cardiol. 2003;41:1475–9.
atherosclerosis also provides incremental prognostic infor- 12. Cooney MT, Dudina A, Whincup P, et al., SCORE Investigators.
Re-evaluating the Rose approach: comparative benefits of the pop-
mation in low- and high-risk individuals. This information ulation and high-risk preventive strategies. Eur J Cardiovasc Prev
may be used to “down-classify” (less overtreatment) or “up-­ Rehabil. 2009;16:541–9.
classify” (less undertreatment) patient risk to more accu- 13. Michos ED, Nasir K, Braunstein JB, et al. Framingham risk equa-
rately evaluate the role of lifelong pharmacotherapy for tion underestimates subclinical atherosclerosis risk in asymptom-
atic women. Atherosclerosis. 2006;184:2010–6.
prevention of ASCVD events. Though there is compelling 14. Schlendorf KH, Nasir K, Blumenthal RS.  Limitations of the
evidence that the absence of subclinical atherosclerosis by Framingham risk score are now much clearer. Prev Med.
CACS may provide a rationale to de-escalate statins and 2009;48:115–6.
other preventive therapies, a prospective randomized con- 15. Blaha MJ, Silverman MG, Budoff MJ. Is there a role for coronary
artery calcium scoring for management of asymptomatic patients
trolled trial is needed to determine if imaging identifies a at risk for coronary artery disease? Circ Cardiovasc Imaging.
lower-risk population and results in accurate assignment of 2014;7:398–408.
treatments for optimal ASCVD risk reduction. With the 16. Kianoush S, Al Rifai M, Cainzos-Achirica M, et  al. An update
expansion of guideline-based low-dose lung cancer screen- on the utility of coronary artery calcium scoring for coronary
heart disease and cardiovascular disease risk prediction. Curr
ing in high-risk patients and recommendations for simulta- Atheroscler Rep. 2016;18:13.
neous reporting of CACS, it will be possible to accurately 17. Bild DE, Detrano R, Peterson D, et al. Ethnic differences in coro-
characterize ASCVD risk for millions of individuals. nary calcification: the Multi-Ethnic Study of Atherosclerosis.
Circulation. 2005;111:1313–20.
18. McClelland RL, Chung H, Detrano R, et al. Distribution of coro-
nary artery calcium by race, gender, and age: results from the Multi-­
References Ethnic Study of Atherosclerosis. Circulation. 2006;113:30–7.
19. Cook NR.  Statistical evaluation of prognostic versus diagnostic
1. Mozaffarian D, Benjamin EJ, Go AS, et  al. Heart disease and models: beyond the ROC curve. Clin Chem. 2008;54:17–23.
stroke statistics—2016 update. A report from the American Heart 20. Detrano R, Guerci AD, Carr JJ, et al. Coronary calcium as a pre-
Association. Circulation. 2016;133:e38–e360. dictor of coronary events in four racial or ethnic groups. N Engl J
2. Ford ES, Ajani UA, Croft JB, et  al. Explaining the decrease in Med. 2008;358:1336–45.
U.S. deaths from coronary disease, 1980-2000. N Engl J Med. 21. Weber LA, Cheezum MK, Reese JM, et al. Cardiovascular imag-
2007;356:2388–98. ing for the primary prevention of atherosclerotic cardiovascular
3. Greenland P, Bonow RO, Brundage BH, et al. ACCF/AHA 2007 disease events. Curr Cardiovasc Imaging Rep. 2015;8:36.
Clinical expert consensus document on coronary artery cal- 22. Arad Y, Spadaro LA, Goodman K, et  al. Prediction of coronary
cium scoring by computed tomography in global cardiovascular events with electron beam computed tomography. J Am Coll
risk assessment and in evaluation of patients with chest pain. Cardiol. 2000;36:1253–60.
Circulation. 2007;115:402–26. 23. Raggi P, Callister TQ, Cooil B, et  al. Identification of patients
4. Wilson PWF, D’Agostino RB, Levy D, et  al. Prediction of cor- at increased risk of first unheralded acute myocardial infarc-
onary heart disease using risk factor categories. Circulation. tion by electron-beam computed tomography. Circulation.
1998;97:1837–47. 2000;101:850–5.
5. Assmann G, Cullen P, Schulte H.  Simple scoring scheme for cal- 24. Wong ND, Hsu JC, Detrano RC, et al. Coronary artery calcium
culating the risk of acute coronary events based on the 10-year evaluation by electron beam tomography and its relation to new
follow-up of the prospective cardiovascular Munster (PROCAM) cardiovascular events. Am J Cardiol. 2000;86:495–8.
study. [Erratum: Circulation. 2002;105:900]. Circulation. 2002;105: 25. Vliegenthart R, Oudkerk M, Song B, et al. Coronary calcification
310–5. detected by electron-beam computed tomography and myocardial
6. Conroy R, Pylorala K, Fitzgerald AP, et al. The SCORE (Systemic infraction: the Rotterdam Coronary Calcification Study. Eur Heart
Coronary Risk Evaluation) manuscript. Prediction of ten-year risk J. 2002;23:1596–603.
282 P. B. Morris and M. D. Shapiro

26. Park R, Detrano RC, Xiang M, et al. Combined use of computed 45. Sarwar A, Shaw LJ, Shapiro MD, et al. Diagnostic and prognos-
tomography coronary calcium scores and C-reactive protein lev- tic value of absence of coronary artery calcification. J Am Coll
els in predicting cardiovascular events in nondiabetic individuals. Cardiol Imaging. 2009;2:678–88.
Circulation. 2002;106:2073–7. 46. Polonsky TS, McClelland RL, Jorgensen NW, et  al. Coronary
27. Kondos GT, Hoff JA, Sevrukov A, et  al. Electron-beam tomog- artery calcium score and risk classification for coronary heart dis-
raphy coronary artery calcium and cardiac events. Circulation. ease prediction. JAMA. 2010;303:1610–6.
2003;107:2571–6. 47. Kelkar AA, Schultz WM, Khosa F, et  al. Long-term prognosis
28. Shaw LJ, Raggi P, Schisterman E, et al. Prognostic value of car- after coronary artery calcium scoring among low-intermediate
diac risk factors and coronary artery calcium screening for all-­ risk men and women. Circ Cardiovasc Imaging. 2016;9:e003742.
cause mortality. Radiology. 2003;228:826–33. https://doi.org/10.1161/circimaging.115.003742.
29. Greenland P, LaBree L, Azen SP, et al. Coronary artery calcium 48. Blaha MJ, Cainzos-Achirica M, Greenland P, et  al. Role of
score combined with Framingham score for risk prediction in coronary artery calcium score of zero and other negative risk
asymptomatic individuals. JAMA. 2004;291:210–5. markers for cardiovascular disease: the Multi-Ethnic Study of
30. Arad Y, Goodman KJ, Roth M, et al. Coronary calcification, coro- Atherosclerosis (MESA). Circulation. 2016;133:849–58.
nary disease risk factors, C-reactive protein, and atherosclerotic 49. Yeboah J, Polonsky TS, Young R, et al. Utility of nontraditional
cardiovascular disease events. J Am Coll Cardiol. 2005;46:158–65. risk markers in individuals ineligible for statin therapy accord-
31. Taylor AJ, Bindeman J, Feuerstein I, et al. Coronary calcium inde- ing to the 2013 American College of Cardiology/American
pendently predicts incident premature coronary heart disease over Heart Association cholesterol guidelines. Circulation. 2015;132:
measure cardiovascular disease risk factors. J Am Coll Cardiol. 916–22.
2005;46:807–14. 50. Kavousi M, Desai CS, Ayers C, et al. Prevalence and prognostic
32. Budoff MJ, Shaw LJ, Liu ST, et al. Long-term prognosis associated significance of coronary artery calcification in low-risk women: a
with coronary calcification. J Am Coll Cardiol. 2007;49:1860–70. meta-analysis. JAMA. 2016;316:2126–34.
33. Lakoski SG, Greenland P, Wong ND, et al. Coronary artery cal- 51. Joshi PH, Patel B, Blaha MJ, et  al. Coronary artery calcium
cium scores and risk for cardiovascular events in women classified predicts cardiovascular events in participants with a low life-
as “low risk” based on Framingham risk score. Arch Intern Med. time risk of cardiovascular disease: the Multi-Ethnic Study of
2007;167:2437–42. Atherosclerosis (MESA). Atherosclerosis. 2016;246:367–73.
34. Becker A, Leber A, Becker C, Knez A. Predictive value of coro- 52. Tota-Maharaj R, Blaha MJ, Blankstein R, et  al. Association
nary calcifications for future cardiac events in asymptomatic indi- of coronary artery calcium and coronary heart disease events
viduals. Am Heart J. 2008;155:154–60. in young and elderly participants in the Multi-Ethnic Study of
35. Erbel R, Mohlenkamp S, Moebus S, et al. Coronary risk stratifica- Atherosclerosis: a secondary analysis of a prospective, population-­
tion, discrimination, and reclassification improvement based on based cohort. Mayo Clin Proc. 2014;89:1350–9.
quantification of subclinical coronary atherosclerosis. J Am Coll 53. McClelland RL, Jorgensen NW, Budoff M, et al. 10-year coronary
Cardiol. 2010;56:1397–406. heart disease risk prediction using coronary artery calcium and
36. Hecht HS, Narula J. Coronary artery calcium scanning in asymp- traditional risk factors. J Am Coll Cardiol. 2015;66:1643–53.
tomatic patients with diabetes. J Diabetes. 2012;4:342–50. 54. Mortensen MB, Fuster V, Muntendam P, et al. A simple disease-­
37. Yeboah J, McClelland RL, Polonsky TS, et  al. Comparison of guided approach to personalize ACC/AHA-recommended statin
novel risk markers for improvement in cardiovascular risk assess- allocation in elderly people: the BioImage Study. J Am Coll
ment in intermediate-risk individuals. JAMA. 2012;308:788–95. Cardiol. 2016;68:881–91.
38. Folsom AR, Kronmal RA, Detrano RC, et al. Coronary artery 55. Hoff JA, Quinn L, Sevrukov A, et al. The prevalence of coronary
calcification compared with carotid intima-media thick- artery calcium among diabetic individuals without known coro-
ness in the prediction of cardiovascular disease incidence: nary artery disease. J Am Coll Cardiol. 2003;41:1008–12.
the Multi-Ethnic Study of Atherosclerosis. Arch Intern Med. 56. Daga N, Nasir K, Hamirani Y, et al. Prevalence and severity of cor-
2008;168:1333–9. onary artery calcium in young persons with diabetes. J Cardiovasc
39. Hecht HS. Coronary artery calcium scanning: past, present, and Comput Tomogr. 2013;7:241–7.
future. JACC Cardiovasc Img. 2015;8:580–96. 57. Kramer CK, Ziman B, Gross JL, et al. Coronary artery calcium
40. Elias-Smale SE, Vliegenthart R, Koller MT, et al. Coronary cal- score prediction of all cause mortality and cardiovascular events in
cium score improves classification of coronary heart disease in the people with type 2 diabetes: systematic review and meta-analysis.
elderly. J Am Coll Cardiol. 2010;56:1407–14. BMJ. 2013;346:f1654.
41. Stone NJ, Robinson JG, Lichtenstein AH, et al. The 2013 ACC/ 58. Raggi P, Shaw LJ, Berman DS, Callister TQ. Prognostic value of
AHA guideline on the treatment of blood cholesterol to reduce coronary artery calcium screening in subjects with and without
atherosclerotic cardiovascular risk in adults. J Am Coll Cardiol. diabetes. J Am Coll Cardiol. 2004;43:1663–9.
2014;63:2889–934. 59. Malik S, Budoff MJ, Katz R, et  al. Impact of subclinical ath-
42. Grundy SM, Stone NJ, Bailey AL, et al. 2018 ACC/AHA/ erosclerosis on cardiovascular disease events in individu-
AACVPR/AAPA/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/ als with metabolic syndrome and diabetes. Diabetes Care.
PCNA guideline on the management of blood cholesterol: a 2011;34:2285–90.
report of the American College of Cardiology Foundation and 60. Hecht HS. A zero coronary artery calcium score: priceless. J Am
the American Heart Association Task Force on Clinical Practice Coll Cardiol. 2010;55:1118–20.
Guidelines. J Am Coll Cardiol. 2018; https://doi.org/10.1016/j. 61. Shareghi S, Ahmadi N, Young E, et al. Prognostic significance of
jacc.2018.11.003. zero coronary calcium scores on cardiac computed tomography. J
43. Shaw LJ, Giambrone AE, Blaha MJ, et  al. Long-term progno- Cardiovasc Comput Tomogr. 2007;1:155–9.
sis after coronary artery calcification testing in asymptomatic 62. Pencina MJ, Navar-Boggan AM, D’Agostino RB Sr, et  al.
patients: a cohort study. Ann Intern Med. 2015;163:14–21. Application of new cholesterol guidelines to a population-based
44. Valenti V, Ó Hartaigh B, Heo R, et al. A 15-year warranty period sample. N Engl J Med. 2014;370:1422–31.
for asymptomatic individuals without coronary artery calcium: 63. Blaha M, Budoff MJ, Shaw LJ, et al. Absence of coronary artery
a prospective follow-up of 9,715 individuals. JACC Cardiovasc calcification and all-cause mortality. J Am Coll Cardiol Img.
Imaging. 2015;8:900–9. 2009;2:692–700.
23  Clinical Application of the Coronary Artery Calcium Score and Implications for Cardiovascular Disease Prevention 283

64. Budoff MJ, McClelland RL, Nasir K, et al. Cardiovascular events 81. McEvoy JW, Martin SS, Blaha MJ, et al. The case for and against
with absent or minimal coronary calcification: the Multi-Ethnic a coronary artery calcium trial. JACC Cardiovasc Imaging.
Study of Atherosclerosis (MESA). Am Heart J. 2009;158:554–61. 2016;9:994–1002.
65. Nasir K, Bittencourt MS, Blaha MJ, et al. Implications of coro- 82. Rozanski A, Gransar H, Shaw LJ, et al. Impact of coronary artery
nary artery calcium testing among statin candidates according calcium scanning on coronary risk factors and downstream testing.
to American College of Cardiology/American Heart Association The EISNER (Early Identification of Subclinical Atherosclerosis
cholesterol management guidelines. J Am Coll Cardiol. by Noninvasive Imaging Research) prospective randomized trial.
2015;66:1657–68. J Am Coll Cardiol. 2011;57:1622–32.
66. Ridker PM, Danielson E, Fonseca FA, et  al., for the JUPITER 83. Mamudu HM, Paul TK, Veeranki SP, Budoff M.  The effects of
Study Group. Rosuvastatin to prevent vascular events in men coronary artery calcium screening on behavioral modification,
and women with elevated C-reactive protein. N Engl J Med. risk perception, and medication adherence among asymptomatic
2008;359:2195–202. adults. Atherosclerosis. 2014;236:338–50.
67. Blaha MJ, Budoff MJ, DeFilippis AP, et al. Associations between 84. Orakzai RH, Nasir K, Orakzai SH, et al. Effect of patient visual-
C-reactive protein, coronary artery calcium, and cardiovascular ization of coronary calcium by electron beam computed tomog-
events: implications for the JUPITER population from MESA, a raphy on changes in beneficial lifestyle changes. Am J Cardiol.
population-based cohort study. Lancet. 2001;378:684–92. 2008;101:999–1002.
68. Valenti V, Hartaigh BO, Cho I, et al. Absence of coronary artery 85. Taylor AJ, Bindeman J, Feuerstein I, et al. Community-based pro-
calcium identifies asymptomatic diabetic individuals at low near-­ vision of statin and aspirin after the detection of coronary artery
term but not long-term risk of mortality: a 15-year follow-up study calcium within a community-based screening cohort. J Am Coll
of 9715 patients. Circ Cardiovasc Imaging. 2016;9:e003528. Cardiol. 2008;51:1337–41.
https://doi.org/10.1161/CIRCIMAGING.115.003528. 86. Kalia NK, Cespedes L, Youssef G, et al. Motivational effects of
69. Nasir K, Rubin J, Blaha MJ, et al. Interplay of coronary artery cal- coronary artery calcium scores on statin adherence and weight
cification and traditional risk factors for the prediction of all-cause loss. Coron Artery Dis. 2015;26:226–30.
mortality in asymptomatic individuals. Circ Cardiovasc Imaging. 87. McEvoy JW, Blaha MJ. Coronary artery calcium testing: explor-
2012;5:467–73. ing the need for a randomized trial. Circ Cardiovasc Imaging.
70. Silverman MG, Blaha MJ, Krumholz HM, et al. Impact of coro- 2014;7:578–80.
nary artery calcium on coronary heart disease events in individuals 88. Kim J, McEvoy JW, Nasir K, et  al. Critical review of high-­
at the extremes of traditional risk factor burden: the Multi-Ethnic sensitivity C-reactive protein and coronary artery calcium for
Study of Atherosclerosis. Eur Heart J. 2014;35:2232–41. the guidance of statin allocation. Circ Cardiovasc Imaging.
71. Martin SS, Blaha MJ, Blankstein R, et al. Dyslipidemia, coronary 2014;7:315–22.
artery calcium, and incident atherosclerotic cardiovascular dis- 89. Institute of Medicine. Clinical practice guidelines we can trust.
ease: implications for statin therapy from the Multi-Ethnic Study Available at https://www.nap.edu/read/13058/chapter/1. Accessed
of Atherosclerosis. Circulation. 2014;129:77–86. 4 Feb 2017.
72. National Cholesterol Education Program. Third report of the 90. ROBINSCA Study Design. Available at http://www.robinsca.nl/
National Cholesterol Education Program (NCEP) Expert Panel on for-researchers/study-design/. Accessed 4 Feb 2017.
Detection, Evaluation, and Treatment of High Blood Cholesterol 91. Budoff MJ, Young R, Lopez VA, et  al. Progression of coronary
in Adults (Adult Treatment Panel III): final report. Circulation. calcium and incident coronary heart disease events: MESA
2002;106:3149–421. (Multi-Ethnic Study of Atherosclerosis). J Am Coll Cardiol.
73. Agarwal S, Morgan T, Herrington DM, et  al. Coronary calcium 2013;61:1231–9.
score and prediction of all-cause mortality in diabetes. Diabetes 92. Budoff MJ, Hokanson JE, Nasir K, et  al. Progression of coro-
Care. 2001;34:1219–24. nary artery calcium predicts all-cause mortality. JACC Cardiovas
74. Budoff MJ, Diamond GA, Raggi P, et  al. Continuous probabi- Imaging. 2010;3:1229–36.
listic prediction of angiographically significant coronary artery 93. Wong ND, Nelson JC, Granston T, et al. Metabolic syndrome, dia-
disease using electron beam tomography. Circulation. 2002;105: betes, and incidence and progression of coronary calcium. JACC
1791–6. Cardiovasc Imaging. 2012;5:358–66.
75. Villines TC, Julten EA, Shaw LJ, et al., for the CONFIRM Registry 94. Erbel R, Lehmann N, Churzidse S, et al., on behalf of the Heinz
Investigators. Prevalence and severity of coronary artery disease Nixdorf Recall Study Investigators. Progression of coronary
and adverse events among symptomatic patients with coronary artery calcification seems to be inevitable, but predictable—
artery calcification scores of zero undergoing coronary computed results of the Heinz Nixdorf Recall (HNR) study. Eur Heart J.
tomography angiography. J Am Coll Cardiol. 2011;58:2533–40. 2014;35:2960–2971.
76. Lubbers M, Dedic A, Coenen A, et  al. Calcium imaging and 95. Raggi P, Callister TQ, Shaw LJ.  Progression of coronary artery
selective computed tomography angiography in comparison to calcium and risk of first myocardial infarction in patients receiv-
functional testing for suspected coronary artery disease: the multi- ing cholesterol-lowering therapy. Arterioscler Thromb Vasc Biol.
center, randomized CRESCENT trial. Eur Heart J. 2016;37:1232. 2004;24:1272–7.
https://doi.org/10.1093/eurheartj/ehv700. 96. McCullough PA, Chinnaiyan KM.  Annual progression of coro-
77. Kim YJ, Hur J, Lee HJ, et al. Meaning of zero calcium score in nary calcification in trials of preventive therapies: a systematic
symptomatic patients referred for coronary computed tomogra- review. Arch Intern Med. 2009;169:2064–70.
phy. Eur Heart J Cardiovasc Imaging. 2012;13:776–85. 97. Nissen SE, Nicholls SJ, Sipahi I, et  al. Effects of very high-­
78. Wolk MJ, Bailey SR, Doherty JU, et al. ACCF/AHA/ASE/ASNC/ intensity statin therapy on regression of coronary atherosclerosis:
HFSA/HRS/SCAI/SCCT/SCMR/STS 2103 Multimodality appro- the ASTEROID trial. JAMA. 2006;295:1556–65.
priate use criteria for the detection and risk assessment of stable 98. Nicholls SJ, Ballantyne CM, Barter PJ, et al. Effect of two inten-
ischemic heart disease. J Am Coll Cardiol. 2014;63:381–406. sive statin regimens on progression of coronary disease. N Engl J
79. Greenland P. When nothing is really something. JACC Cardiovasc Med. 2001;365:2078–87.
Imaging. 2015;8:910–2. 99. Ferencik M, Chatzizisis YS. Statins and the coronary plaque cal-
80. Taylor AJ, Raggi J, Raggi P. The power of nothing: the zero cal- cium “paradox”: insights from non-invasive and invasive imaging.
cium score. J Cardiovasc Comput Tomogr. 2007;1:160–1. Atherosclerosis. 2015;241:783–5.
284 P. B. Morris and M. D. Shapiro

100. McElvoy JW, Blaha MJ, DeFilippis AP, et al. Coronary artery cal- 112. Greenland P, Bonow RO, Brundage BH, et al. ACCF/AHA 2007
cium progression: an important clinical measurement? J Am Coll clinical expert consensus document on coronary artery calcium
Cardiol. 2010;56:1613–22. scoring by computed tomography in global cardiovascular risk
101. Houslay ED, Cowell SJ, Prescott RJ, et al. Progressive coronary assessment and in evaluation of patients with chest pain: a report
calcification despite intensive lipid-lowering treatment: a random- of the American College of Cardiology Foundation Clinical
ized controlled trial. Heart. 2006;92:1207–12. Expert Consensus Task Force (ACCF/AHA Writing Committee
102. Raggi P, Davidson M, Callister TQ, et  al. Aggressive versus to Update the 2000 Expert Consensus Document on Electron
moderate lipid-lowering therapy in hypercholesterolemic post- Beam Computed Tomography) developed in collaboration with
menopausal women: beyond endorsed lipid lowering with EBT the Society of Atherosclerosis Imaging and Prevention and the
scanning (BELLES). Circulation. 2005;112:563–71. Society of Cardiovascular Computed Tomography. J Am Coll
103. Arad Y, Spadaro LA, Roth M, et al. Treatment of asymptomatic Cardiol. 2007;49:378–402.
adults with elevated coronary calcium scores with atorvastatin, 113. Greenland P, Smith SC Jr, Grundy SM. Improving coronary heart
vitamin C, and vitamin E: the St. Francis Heart Study Randomized disease risk assessment in asymptomatic people: role of traditional
Clinical Trial. J Am Coll Cardiol. 2005;46:166–72. risk factors and noninvasive cardiovascular tests. Circulation.
104. Puri R, Nicholls SJ, Shao M, et al. Impact of statin serial coronary 2001;104:1863–7.
calcification during atheroma progression and regression. J Am 114. Greenland P, Alpert JS, Beller GA, et al. 2010 ACCF/AHA guide-
Coll Cardiol. 2015;65:1273–82. line for assessment of cardiovascular risk in asymptomatic adults.
105. Raber L, Taniwaki M, Zaugg S, et  al. Effect of high-intensity J Am Coll of Cardiol. 2010;56:e50–103.
statin therapy on atherosclerosis in non-infarct-related coronary 115. Earls JP, Woodard PK, Abbara S, et al. ACR appropriateness crite-
arteries (IBIS-4): a serial intravascular ultrasonography study. Eur ria asymptomatic patient at risk for coronary artery disease. J Am
Heart J. 2015;36:490–500. Coll Radiol. 2014;11:12–9.
106. O’Rourke RA, Brundage BH, Froelicher VF, et  al. American 116. Piepoli MF, Hoes AW, Agewall S, et  al. 2016 European guide-
College of Cardiology/American Heart Association Expert lines on cardiovascular disease prevention in clinical practice. The
Consensus Document on electron-beam computed tomogra- Sixth Joint Task Force of the European Society of Cardiology and
phy for the diagnosis and prognosis of coronary artery disease. Other Societies on cardiovascular disease prevention in clinical
Circulation. 2000;102:126–40. practice. Eur Heart J. 2016;37:2315–81.
107. Greenland P, Abrams J, Aurigemma GP, et  al. Prevention 117. Berrington de Gonzalex A, Mahesh M, Kim KP, et al. Projected
Conference V. Beyond secondary prevention: identifying the high-­ cancer risks from computed tomographic scans performed in the
risk patient for primary prevention. Noninvasive tests of athero- United States in 2007. Arch Intern Med. 2009;169:2071–7.
sclerotic burden. Circulation. 2000;101:e16–322. 118. Moyer VA, on behalf of the U.S. Preventive Services Task Force.
108. Mieres JH, Shaw LJ, Arai A, et al. Role of noninvasive testing in Screening for lun cancer: U.S. Preventive Services Task Force rec-
the clinical evaluation women with suspected coronary artery dis- ommendation statement. Ann Intern Med. 2014;160:330–8.
ease. Consensus Statement from the Cardiac Imaging Committee, 119. Hecht HS, Cronin P, Blaha MJ, et  al. 2016 SCCT/STR guide-
Council on Clinic Cardiology, and the Cardiovascular Imaging and lines for coronary artery calcium scoring of noncontrast noncar-
Intervention Committee, Council on Cardiovascular Radiology diac chest CT scans: a report of the Society of Cardiovascular
and Intervention, American Heart Association. Circulation. Computed Tomography and Society of Thoracic Radiology. J
2005;111:682–96. Cardiovasc Comput Tomogr. 2017; https://doi.org/10.1016/j.
109. Budoff MJ, Achenbach S, Blumenthal RS, et  al. Assessment jcct.2016.11.003.
of coronary artery disease by cardiac computed tomography. 120. Budoff MJ, Nasir K, Kinney GL, et al. Coronary artery and tho-
AHA scientific statement from the American Heart Association racic calcium on non-contrast thoracic CT scans: comparison of
Committee on Cardiovascular Imaging and Intervention, Council ungated and gated examinations in patients from the COPD gene
on Cardiovascular Radiology and Intervention, and Committee cohort. J Cardiovasc Comput Tomogr. 2011;5:113–8.
on Cardiac Imaging, Council on Clinical Cardiology. Circulation. 121. Hughes-Austin JM, Dominquez A, Allison MA, et al. Relationship
2006;114:1761–91. of coronary artery calcium on standard chest computed tomogra-
110. Greenland P, Bonow RO, Brundage BH, et al. ACCF/AHA 2007 phy scans with mortality. J Am Coll Cardiol Img. 2016;9:152–9.
clinical expert consensus document on coronary artery calcium 122. Jacob PC, Prokop M, van der Graaf Y, et al. Comparing coronary
scoring by computed tomography in global cardiovascular artery calcium and thoracic aorta calcium for prediction of all-­
risk assessment and in evaluation of patients with chest pain. cause mortality and cardiovascular events on low-dose non-gated
A report of the American College of Cardiology Foundation computed tomography in high-risk population of heavy smokers.
Clinical Expert Consensus Task Force (ACCF/AHA Writing Atherosclerosis. 2010;209:455–62.
Committee to update the 2000 Expert Consensus Document 123. Mmets OM, Vliegenthart R, Gondrie MJ, et  al. Lung cancer
on electron beam computed tomography). Circulation. screening CT-based prediction of cardiovascular events. J Am
2007;115:402–26. Coll Cardiol Img. 2013;6:899–907.
111. Detrano RC, Anderson M, Nelson J, Wong ND, et al. Coronary 124. Lung Cancer Screening Registry. American College of Radiology.
calcium measurements: effect of CT scanner type and calcium Available at https://www.acr.org/~/media/ACR/Documents/PDF/
measure on rescan reproducibility—MESA study. Radiology. QualitySafety/NRDR/Lung-Cancer-Screening-Practice-Registry/
2005;236:477–84. ACR-LCSR-User-Guide.pdf?la=en. Accessed 1 Apr 2017.
 The Many Uses of Epicardial Fat
Measurements 24
Mohamed Marwan

The term “epicardial adipose tissue” (EAT)  – often called tissue enclosed within the pericardial sac, “epicardial adi-
“epicardial fat”  – refers to the visceral adipose tissue sur- pose tissue” is more accurate. It better describes the location
rounding the heart and coronary arteries. Along with the below the epicardium and on the immediate surface of the
pericardial fluid, epicardial adipose tissue creates a smooth heart (whereas “pericardial fat” could be on the outside of
surface that facilitates motion of the heart within the pericar- the pericardium), and “adipose tissue” is more accurate than
dium. Notably, both the volume and the distribution of epi- simply “fat” since the tissue is metabolically active. Adipose
cardial adipose tissue vary widely between individuals and tissue on the outer surface of the pericardial is termed “para-
do not show a strict linear relationship to body weight, body cardiac fat”(see figure. 24.1). Epicardial fat is a visceral fat
mass index, obesity, or the extent of abdominal visceral fat. depot surrounding the heart similar to intra-abdominal fat,
In the last years, there has been an increasing interest in and both evolve embryologically from brown adipose tissue
imaging and quantification of epicardial adipose tissue. [16]. Vascular supply to epicardial fat is secured through the
Evidence has so far accumulated showing a significant asso- coronary arteries. Importantly, epicardial fat lies in direct
ciation between epicardial adipose tissue, risk factors for contact with the myocardium with no muscle fascia in
atherosclerosis, coronary artery disease, and cardiovascular between which explains the potential interaction between
outcome [1–6]. In addition, significant associations have both the myocardium and epicardial fat both supplied by
been found between epicardial adipose tissue and non-­ coronary microcirculation [16].
atherosclerotic diseases, such as atrial fibrillation or left ven-
tricular diastolic dysfunction [7–9]. This chapter summarizes
the current data concerning epicardial adipose tissue, its Imaging and Quantification of Epicardial Fat
quantification, and its potential use as a novel risk factor for
cardiovascular morbidity and mortality. Over the past decade, several imaging modalities have been
used to detect and quantify epicardial adipose tissue. They
include echocardiography, magnetic resonance, and com-
 at Stores Surrounding the Heart
F puted tomography (CT). Surrogate imaging measures to esti-
and Their Development mate the true volume of epicardial fat include epicardial fat
thickness, pericoronary fat thickness or volume, and epicar-
The nomenclature used in the current literature to define adi- dial fat areas in single cross sections. However, given the con-
pose tissue surrounding the heart is heterogeneous. Terms siderable inter-individual differences in epicardial fat spatial
such as “epicardial,” “pericardial,” “paracardiac,” and “intra- distribution, it remains questionable whether such surrogate
thoracic” fat have been used [2, 3, 5, 10–15]. “Epicardial measures correlate well with the absolute amount of epicar-
adipose tissue” (EAT) refers to the fat between the myocar- dial fat. In fact, data indicate that full volume quantification is
dium and the visceral layer of the pericardial sac, the serous preferable to surrogate measurements; Bastarrika et  al.
epicardium. Although the simpler term “pericardial fat” has directly compared volumetric and distance measurements of
been often used in the imaging literature to refer to adipose epicardial fat in relation to coronary artery stenosis in 45
patients who underwent CT imaging and invasive coronary
angiography [17]. Patients with significant coronary artery
M. Marwan (*)
Department of Medicine, Cardiology and Angiology,
stenosis had greater epicardial fat volumes than those without
Universitätsklinikum Erlangen, Erlangen, Germany significant arterial stenosis, while epicardial fat distance
e-mail: [email protected] failed to show a difference between the two patient groups.

© Humana Press 2019 285

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_24
286 M. Marwan

a b

Fig. 24.1  Fifty-two-year-old female patient. (a) Three-dimensional reconstruction of the heart within the epicardial adipose tissue depicted in
yellow (light green arrow). (b) The pericardium is marked by the white arrow. Fat outside the pericardium is termed “paracardiac fat”

Of the noninvasive imaging tools, CT and magnetic reso- diaphragm, the left ventricular apex, and a single slice beyond
nance tomography offer complete volumetric coverage of the the posterior descending artery have been proposed [2, 7, 11,
heart. Quantification of epicardial fat from CT is appealing 15, 18, 20]. Processing time for epicardial fat measurements
given the high spatial resolution of CT, the full volume cov- ranges from 5 to 11 min [11, 20].) Recently, automated EAT
erage of the entire heart in all routinely acquired cardiac volume quantification tools particularly for contrast-enhanced
scans, and the fact that fat has distinctly lower attenuation CT are available on some commercial workstations, requiring
values than other tissue types. only minor user interaction, which further decreases the time
Semiautomated software are readily available which for EAT volume quantification. EAT volume measurements
allow quantification of epicardial fat volume based on CT based on CT are highly reproducible, with reported interob-
attenuation thresholds, after setting the upper and lower server variabilities of 7%, 8%, and 15% in three studies
boundaries of the heart and tracing the pericardial sac. Fat [3, 11, 21], with lower variability for volume measurements
attenuation thresholds in CT typically range from a lower compared to surrogate markers such as fat thickness [17]. For
value of −250 to −190 Hounsfield Units to an upper value of semiautomated measurements of epicardial and thoracic fat
−50 and −30 HU [1, 10, 11, 15, 18]. For non-contrast CT, a from non-contrast CT, the interscan reproducibility has also
typical fat attenuation range of −190 HU to −30 HU has been reported to be high, with correlation coefficients ≥0.98
been used. Contrast scans have regularly been used for quan- between two separate measurements [22].
tification of epicardial adipose tissue volume in published
literature. However, care should be taken that the threshold
for detection of epicardial fat needs to be adjusted for con- Reference Values for Epicardial Fat
trast scans. In fact, La Grutta et al. recently compared epicar-
dial fat volumes quantified in non-contrast and contrast scans A major limitation for widespread clinical use of epicardial
in 76 patients systematically [19]. They observed an under- fat measurements is the absence of standardized reference
estimation of epicardial adipose tissue volume even after values for what is to be considered as “normal” epicardial fat
adjustment of post-contrast attenuation. volumes. There is limited data to determine normal limits
Anatomical landmarks for measurement of total epicardial beyond which epicardial fat volume can be considered
fat volumes include the pulmonary artery bifurcation, the mid abnormally high. In a community-based sample of more than
left atrium, and the aortic root. For the lower boundary, the 3000 middle-aged individuals, Thanassouslis et al. quantified
24  The Many Uses of Epicardial Fat Measurements 287

Table 24.1  Summarizes different thresholds and binary cut-off used in literature for epicardial fat quantification
Cohort Threshold Outcome/reference
Thanassoulis et al. [10] (healthy reference 139.4 cm3 for men, 119 cm3 for women Abnormally high epicardial fat volumes on the basis
sample, n = 3312) of age and gender-specific 90th percentiles
Iwasaki et al. [23] (n = 197) 100 cm3 Higher incidence of CAD and higher calcium score
Sarin et al. [18] (n = 151)
Cheng et al. [20] (n = 232) 125 cm3 MACE on follow-up
Tamarappoo et al. [24] (n = 219) 125 cm3 Ischemia on SPECT
Shmilovich et al. [25] (n = 458) 68 cm3/m2 95th percentile threshold for abnormally high
epicardial fat volume, MACE on follow-up
Lee et al. [26] (n = 846) %EAT >41% (from total body fat) Obstructive CAD

epicardial and paracardiac fat volumes in participants from >50%) and %EAT >47% for severe CAD (stenosis >70%)
the Framingham Heart Study [10]. For the overall sample, Table 24.1.
the median values for epicardial fat volume were 117.5 cm3
and 93.9 cm3, respectively, in men and women and 117.7 cm3
and 58.4  cm3 for paracardiac fat, respectively. The authors Function of Epicardial Fat
defined abnormally high epicardial and paracardiac fat vol-
umes on the basis of the age and gender-specific 90th percen- Epicardial fat not only serves as a mechanical buffer for car-
tiles for these fat depots in a healthy reference sample diac motion and coronary artery distension but also has
(139.4 cm3 in men and 119 cm3 in women for epicardial fat direct metabolic and biochemical effects on the myocar-
and 150.5 cm3 and 73.3 cm3 for paracardiac fat). Using these dium and coronary vessels. The anatomical, embryological,
thresholds, 29.3% of men and 26.3% of women with coro- and biochemical diversity of adipose tissue enclosed within
nary artery disease had elevated epicardial fat volumes. In the pericardial sac compared to other adipose tissues, such
the two studies, a simple threshold of 100 mm3 for epicardial as paracardiac fat, subcutaneous fat, or visceral adipose tis-
fat volume was used, and higher epicardial fat volumes were sue, has triggered numerous hypotheses concerning specific
associated with a significantly higher coronary artery cal- metabolic properties of epicardial adipose tissue [13, 14].
cium score and higher incidence of CAD [18, 23]. Cheng Epicardial adipose tissue has greater capacity for release
et al. used a threshold of 125 cm3 and reported that epicardial and uptake of free fatty acids than other visceral fat depots.
fat volumes above that threshold were found more frequently Since free fatty acid oxidation is responsible for about
in baseline CT scans of asymptomatic patients experiencing 50–70% of the energy production of the heart, epicardial fat
MACE on follow-up [20]. Tamarappoo et al. used the same has been proposed to function as a buffer to protect the heart
threshold when comparing patients with myocardial isch- against exposure to excessively high levels of free fatty
emia by SPECT to controls [24]. acids and to provide energy for the myocardium [27].
Beyond absolute epicardial fat measurements, several Furthermore, convincing evidence has emerged confirming
authors suggested to index the volume of epicardial adipose the fact that epicardial adipose tissue is metabolically active
tissue to the body surface area. In a healthy asymptomatic and serves as a source of several adipokines. This has led
population of 226 subjects, Shmilovich et  al. identified a numerous authors to assume a direct local influence of epi-
value of 68.1  cm3/m2 as the 95th percentile threshold for cardial adipose tissue on the neighboring coronary arteries
abnormally high epicardial fat volume [25]. They could through paracrine and vasocrine mechanisms [13, 14]. This
show that values exceeding this threshold predicted major inherent endocrine potential includes the secretion of sev-
adverse cardiovascular events, with a trend to add to standard eral pro- and anti-­inflammatory mediators and cytokines
coronary calcium scoring and Framingham risk score in pre- such as adiponectin, interleukin 6 (Il-6), and tumor necrosis
dicting cardiovascular events. Furthermore, Lee et al. could factor (TNF) alpha [1, 28–35]. With increasing amounts of
show recently that percentage epicardial fat and total body visceral fat, plasma levels of adiponectin decrease [36, 37],
fat showed the maximum area under the curve for predicting which in turn causes an increase in TNF alpha levels and
presence of coronary artery disease and superior discrimina- hence to a local increase of inflammation [32]. The resulting
tive performance compared to epicedial adipose tissue vol- imbalance of pro- and anti-inflammatory mediators as well
ume and other indexed parameters of EAT [26]. After as cytokines secreted by epicardial adipose tissue is thought
adjustment for Framingham risk score and calcium score, to contribute to a local influence on the embedded coronary
%EAT >41% was a predictor of obstructive CAD (stenosis arteries [38].
288 M. Marwan

Epicardial Fat and Cardiovascular Disease In a community-based sample of the Framingham Heart

Study (n = 1155), Rosito et al. observed a significant asso-
Coronary Atherosclerosis ciation between epicardial fat, but not intrathoracic fat,
with coronary artery calcification quantified by computed
Two lines of evidence support the “local effect” of epicardial tomography (CT) after adjustment for visceral adipose tis-
adipose tissue on the coronary arteries and atherosclerosis sue and traditional cardiovascular risk factors [5].
affecting these vessels. First, the volume of epicardial adi- Furthermore, Konishi et al. could show that epicardial fat
pose tissue appears to be relatively independent from overall volume measured in CT, but not waist circumference, was
body adipose tissue but, like intra-abdominal visceral fat, significantly associated with the presence of any coronary
shows an association to the presence and severity of athero- plaques, non-­stenotic plaques confirmed by coronary angi-
sclerosis. Second, histopathological and immunohistochemi- ography, and non-calcified plaques [44]. Moreover, in a
cal studies indicate specific patterns of mediator secretion by study of 214 consecutive patients referred for coronary CT
epicardial adipose tissue – but not by fatty tissue in remote angiography, Alexopoulos et  al. observed a significantly
locations – which correlate to atherosclerosis. In a landmark larger epicardial fat volume in patients with mixed or non-
study, Mazurek et al. analyzed epicardial fat as well as sub- calcified plaques compared to patients with calcified
cutaneous fat of the lower extremity in obese patients referred plaques or no plaques [15]. This relationship remained sig-
for coronary artery bypass grafting. They observed an nificant after adjustment for traditional risk factors of coro-
increased expression and secretion of several inflammatory nary artery disease as well as after indexing the epicardial
mediators as well as chronic inflammatory cell infiltration fat volume to body surface area. In an analysis of 402
with macrophages, lymphocytes, and basophils in epicardial patients with no prior known CAD, Rajani et al. found that
adipose tissue as compared to subcutaneous fat [30]. Along epicardial adipose tissue volume was elevated in patients
the same line, an increased expression of CD45 mRNA in the with calcified plaque, non-­calcified plaque, and partially
epicardial fat of subjects with coronary artery disease has calcified plaque [45]. Furthermore in patients with non-
been observed in other studies, with elevated macrophage calcified plaque or partially calcified plaque, epicardial fat
infiltration [28] and an increase of mast cells in the adventitia volume was found to be an independent predictor of coro-
of coronary lesions [1, 29]. Moreover, in an immunohisto- nary stenosis ≥70% even after adjustment of conventional
chemical study by Konishi et  al., more leucocyte common cardiovascular risk factors [45]. Interestingly, in a cross-
antigen (LCA)-positive cells were observed in pericoronary sectional study of 128 patients with angina undergoing
fat of autopsy cases with coronary artery disease compared invasive coronary angiography, Gorter et al. observed that
to patients without coronary disease, suggesting inflamma- the significant association between the severity of coronary
tion of pericoronary fat in the former group [39]. Hirata et al. atherosclerosis and epicardial as well as pericoronary fat
analyzed patterns of macrophage polarization in epicardial volumes as determined by CT was limited to patients with
adipose tissue of patients with and without coronary artery a low body mass index [12]. The authors assume a potential
disease (CAD) [40]. They observed a change of the ratio of role of insulin resistance in patients with high body mass
M1/M2 macrophages with an abundance of the proinflam- index where higher systemic plasma levels of inflammatory
matory M1-polarized state in patients with CAD [40]. cytokines accelerate atherogenesis, so that – in that case –
Interestingly, recently Uchida et  al. could show in a study there is little addition to the atherosclerotic process through
using immunohistochemical techniques that pericoronary fat the local production of adipocytokines in fat surrounding
store oxidized low-density lipoproteins (oxLDL) as well as the coronary arteries. In a study by Wang et al., epicardial
high-density lipoproteins (HDL), both being supplied either fat volume and the thickness of epicardial fat in the left
by CD 68 (+)- macrophages or through vasa vasorum with atrioventricular groove were found to be associated with
the balance between oxLDL and HDL influencing either the presence and extent of coronary atherosclerosis as
plaque growth (more oxLDL) or plaque maturation (more defined by four CT-based scores (severity score, extent
HDL) [41]. score, calcium volume score, and number of coronary arter-
Clinical observations support a potential direct and local ies with ≥50% luminal stenosis) [46]. Interestingly, only
effect of epicardial adipose tissue on the coronary arteries. In epicardial fat thickness in the left atrioventricular groove,
an intravascular ultrasound (IVUS) study, Prati et  al. sug- but not total epicardial fat volume, was significantly associ-
gested a permissive role of epicardial adipose tissue on the ated with all four parameters of coronary atherosclerosis –
positive remodeling of atherosclerotic plaques [42]. Some in a dose-­ dependent manner  – after adjustment for
authors attribute the fact that coronary arteries within myo- conventional risk factors, body mass index, waist circum-
cardial bridges are typically spared from atherosclerosis to ference, C-reactive protein, and intra-abdominal visceral
the lack of contact to epicardial fat, which may protect the fat area [46]. The authors concluded that measurement of
arteries from developing atherosclerosis [43]. epicardial fat thickness in the left atrioventricular groove
24  The Many Uses of Epicardial Fat Measurements 289

may be more predictive concerning atherogenic risk than referred for 64-slice CT [49]. They found that a high epicar-
measurements of the total epicardial fat volume. These data dial fat volume (above 100 cm3) was associated with vulner-
again support the concept that the influence of epicardial able plaque components (non-calcified plaque, low-density
adipose tissue on coronary atherosclerosis may be spatially plaque, and positive remodeling) independent of obesity
heterogeneous and potentially due to local influence in measurements (BMI and visceral adipose tissue) as well as
addition to or instead of a systemic effect. Along the same coronary calcium score. Furthermore, Ito et  al. measured
line, Mahabadi et al. showed in 78 patients and 311 coro- epicardial adipose tissue in 1308 consecutive symptomatic
nary segments that pericoronary fat as quantified by cardiac patients with a zero calcium score [50]. Significant coronary
CT was associated with the presence of calcified or non- stenoses were present in 7% of the patients, and CT features
calcified coronary lesions in coronary CT angiography. of vulnerable plaques, specifically positively remodeled
They could demonstrate a 2.5-­fold increase of the likeli- plaques with a mean density <30 HU, were found in 5% of
hood to detect plaque in the underlying coronary segment the patients. Volume of epicardial adipose tissue was signifi-
per each doubling of local pericoronary fat volume [1]. As cantly larger both in patients with significant coronary steno-
a possible confounder of their observation, it is worth men- ses (mean 124.3 ± 43.2 cm3) and in patients with CT features
tioning that, based on anatomical reasons, epicardial fat of vulnerable plaques (mean 133.0 ± 40.2 cm3) as compared
accumulations are typically more pronounced in some to controls (mean 95.1 ± 40.3 cm3). In multivariate analysis,
regions of the coronary artery tree, such as the proximal left epicardial adipose tissue was found to be a predictor for both
anterior descending coronary artery. The same regions typi- significant coronary stenoses and CT features of vulnerable
cally show the first signs of atherosclerosis, so that the plaque. In an earlier study of patients who underwent both
association between atherosclerosis and local fat volume CT and intracoronary optical coherence tomography (OCT),
may be a coincidence as well as a causal relationship. the same authors showed that patients in the highest tertile of
epicardial adipose tissue (>130.7 cm3) were more likely to
display thin-capped fibroatheroma in OCT [51]. Harada
Coronary Spasm et al. measured epicardial adipose tissue volume using CT in
80 patients hospitalized for acute coronary syndromes
Owing to the metabolic activity of epicardial adipose tissue (ST-segment elevation and non-ST-segment elevation infarc-
that potentially influences coronary artery morphology by tions) in comparison to 90 outpatient controls with suspected
playing a role in the initiation and aggravation of the process acute coronary syndrome but in whom CT results were
of atherosclerosis, evidence is also emerging supporting a ­normal [52]. Epicardial fat volume was significantly higher
possible effect of epicardial adipose tissue on coronary artery in patients with acute coronary syndromes, and a cutoff value
function. Ohyama et  al. could show recently that coronary of 100  ml was found to be independently associated with
perivascular adipose tissue volume measured using CT was ACS in the multivariate analysis. Recently, in 467 patients
significantly higher in patients with vasospastic angina com- enrolled in the ROMICAT II trial, indexed epicardial adipose
pared to controls [47]. Furthermore, in 97 patients with sus- tissue volume was significantly associated with high-risk
pected coronary spastic angina who underwent ergonovine plaque features even after adjustment for cardiovascular risk
provocation testing, Ito et al. measured epicardial fat volume factors, coronary artery calcium score, and obstructive coro-
in CT data sets [48]. In 27 patients with coronary spasm nary artery disease [53].
(28%), epicardial fat volume and male gender were signifi-
cant predictors of coronary spasm. A threshold of 149.4 cm3
epicardial fat volume was found to be the optimal threshold Myocardial Ischemia
for identifying patient with coronary spasms (70% sensitiv-
ity, 76% specificity, AUC 0.76). These data support a poten- Earlier studies have investigated the relationship between
tial role of epicardial fat in inducing endothelial epicardial fat and inducible myocardial ischemia in SPECT
dysfunction. and uniformly report epicardial fat to be associated with an
increased likelihood of myocardial perfusion defects [24,
54]. Interestingly in a recent study of 396 consecutive
Acute Coronary Syndromes patients undergoing SPECT perfusion imaging, regional
thickness, cross-sectional areas, and total epicardial fat vol-
The association between epicardial fat and “vulnerable” ume were quantified using CT.  In this study the authors
plaque features potentially leading to acute coronary syn- could demonstrate increased regional epicardial fat in
dromes (ACS) and future events have been examined in sev- patients with reversible perfusion defect compared to
eral patient cohorts. Oka et  al. assessed coronary plaque patients with normal perfusion or myocardial scars (fixed
components and epicardial fat volume in 357 patients defects) [55].
290 M. Marwan

Coronary Artery Disease Progression follow-­up [59]. In unadjusted analysis, and even after adjust-
ment for cardiovascular risk factors, intrathoracic fat vol-
Regarding the influence of epicardial adipose tissue on coro- ume, but not BMI, was associated with the risk of coronary
nary artery progression, Nakanishi et  al. could show that heart disease. More recently, Cheng et  al. confirmed these
increase in epicardial fat volume is associated with greater results in a case-­control study based on a registry of 2751
progression of coronary artery calcification in intermediate-­ asymptomatic individuals and a 4-year follow-up for major
risk subjects [56]. Moreover, Yerramasu et  al. performed a adverse cardiovascular events (MACE), comparing 58
follow-up study of 333 asymptomatic diabetics referred for a patients with MACE to 174 event-free controls matched for
repeat coronary calcium scan at a mean interval of 2.7 years. age, risk factors, and coronary calcium score [20]. In this
They observed that epicardial fat volume was an independent analysis, significantly higher mean epicardial and intratho-
predictor for the presence and severity of coronary artery cal- racic fat volumes were observed in patients experiencing
cification in the baseline scan as well as coronary calcium MACE compared to controls. In agreement with the findings
progression during follow-up [57]. In a long-term CT fol- of the Framingham investigators, adding epicardial fat vol-
low-­up study of asymptomatic individual without prior his- ume and not intrathoracic fat to the coronary calcium score
tory of coronary artery disease (median 65 months), Hwang and Framingham risk score improved specificity and overall
et  al. observed new development of non-calcified plaque accuracy in predicting MACE, which supports the hypothe-
with higher indexed epicardial fat volume at baseline CT sis of a local influence of epicardial adipose tissue on disease
(79.9  ±  30.3  cm3/m2) compared to patients with no plaque development.
(62.5  ±  24.7  cm3/m2) or newly developed calcified plaque In 2013, investigators of the Heinz Nixdorf Recall Study
(63.7 ± 22.7 cm3/m2) [58]. reported the association of epicardial fat volume, quantified
from cardiac CT, with incident myocardial infarction in a
European, population-based random cohort [60]. The asso-
Cardiovascular Outcome ciation of epicardial adipose tissue with coronary events was
independent of traditional cardiovascular risk factors and
Patient-based cross-sectional analyses provided the first remained statistically significant even after further adjust-
reports of an association between epicardial adipose tissue ment for the coronary calcium score. This finding was partly
and prevalent cardiovascular disease. In 2001, Taguchi et al. explained by a stronger correlation between epicardial adi-
described an association of epicardial fat determined by car- pose tissue and incident events in subjects with low or no
diac computed tomography, with the presence of CAD in coronary calcium. This led the authors to conclude that
251 Japanese men [6]. Consecutively, investigators of the ­quantification of epicardial adipose tissue from cardiac CT
Framingham Heart Study described increased epicardial adi- may complement prognostic information above coronary
pose tissue volumes in subjects with prevalent CAD in a calcium scoring. In a population of subjects with acute chest
cross-sectional analysis in 1267 subjects from the Offspring pain, however, Forouzandeh et al. found an additional prog-
cohort [10]. Intrathoracic fat, which was defined as fat inside nostic value of epicardial adipose tissue over coronary cal-
the thorax but outside the pericardial sac, showed no associa- cium only in subjects with calcium scores >400 [61]. In a
tion with cardiac events. While associations of epicardial recent review of 9 studies assessing the prognostic value of
adipose tissue with CAD were independent of age and gen- epicardial fat volume including 10,252 patients, the majority
der as well as BMI and waist circumference, no significant of the studies suggested that epicardial fat volume is associ-
association was found after adjusting for traditional cardio- ated with clinical outcomes and provides incremental prog-
vascular risk factors. nostic value to coronary artery calcium scoring [62]. In this
For risk stratification purposes, prospective follow-up review, the authors observed different thresholds and aggre-
studies provide more accurate evidence compared to cross-­ gates of epicardial fat volume used in different studies which
sectional studies. Several prospective analyses have become renders comparisons between different cohorts difficult.
available for epicardial adipose tissue. In a case-cohort study More recently, Oikonomou et al. [63] assessd a novel imag-
based on the Multi-Ethnic Study of Atherosclerosis (MESA), ing biomarker  – the so-called perivascular fat attenuation
intrathoracic fat volume was quantified in the baseline CT of index – in two large clincal cohorts from Europe and North
972 randomly selected individuals with no incident coronary America [63]. This index traces spacial changes of pericoro-
artery disease and compared to the intrathoracic fat volume nary fat attenuation on coronary CT angiography secondary
of 147 MESA participants who developed incident coronary to coronary inflammation. In 3912 patients from the deriva-
heart disease (myocardial infarction, resuscitated cardiac tion and validation cohort, the perivascular fat attenuation
arrest, angina or fatal coronary heart disease) during index around the right coronary artery and the left anterior
24  The Many Uses of Epicardial Fat Measurements 291

descending coronary artery were predictive of all-cause and anisms for the link of epicardial adipose tissue with mea-
cardiac mortality in a median follow-up of 72 months and 54 sures of left ventricular size and function.
months for both cohorts, respectively. These data may poten-
tially help tailoring and intensifying preventive measures in
patients deemed to be at high risk of cardiovascular events. Conflicting Evidence

Besides the large body of evidence supporting an association

 picardial Fat and Non-atherosclerotic
E between epicardial fat and coronary artery disease, several
Cardiovascular Disease studies observed conflicting evidence concerning a causal
association between epicardial fat and CAD. In 380 patients
Interestingly the association of epicardial adipose tissue to enrolled in the CORE320 trial who underwent coronary CT
cardiac disease is not limited to coronary atherosclerosis. A angiography, myocardial perfusion imaging, and clinically
number of studies have shown associations of epicardial fat driven invasive angiography, no association was found
with non-coronary heart disease. Several studies indicated a between epicardial fat volume and presence, severity of cor-
link between epicardial adipose tissue and atrial fibrillation onary artery disease, and myocardial perfusion abnormali-
which was suggested to be mediated by the inflammatory ties [66]. Similarly, in 557 subjects enrolled in the
effects of epicardial fat. Al Chekakie et al. reported higher Cardiovascular Risk in Young Finns Study, associations of
epicardial adipose tissue volume in subjects with atrial coronary heart disease risk markers with epicardial fat vol-
fibrillation compared to subjects with sinus rhythm, and ume were attenuated after multivariable adjustment with no
subjects with persistent atrial fibrillation had even higher fat evidence of increased EFV being independently associated
volume than subjects with paroxysmal atrial fibrillation [7]. with pre-clinical atherosclerosis [67]. On the same line,
Wong et  al. quantified periatrial, periventricular, and total Romijn et al., in a study of 122 patients who underwent non-­
epicardial fat volume in 110 patients undergoing first-time contrast-­enhanced CT for quantification of coronary artery
ablation of atrial fibrillation compared to 20 controls in calcium and epicardial fat volume as well as invasive angiog-
sinus rhythm. They found a significant association between raphy with fractional flow reserve measurements, could not
epicardial fat volume and left atrial volumes, the presence demonstrate an incremental value of epicardial fat volume
and severity of atrial fibrillation, and poorer outcome after beyond coronary artery calcium score and traditional risk
ablation [64]. These data support a potential local effect of factors for detection of hemodynamically significant coro-
adipose tissue as a trigger for atrial fibrillation. Framingham nary artery disease [68].
investigators also described a cross-sectional association of
epicardial adipose tissue with atrial fibrillation, independent
of age, gender, hypertension, and BMI in over 3000 subjects Summary
[65]. However, it has to be acknowledged that longitudinal
analyses of the role of epicardial adipose tissue regarding In summary, the accumulating evidence over the last years
the prediction of future atrial fibrillation are lacking. strongly suggests a potential causal relationship between the
Furthermore, Fox et  al. described a significant correlation metabolic and paracrine characteristics of epicardial adipose
between epicardial fat – quantified by CT – and left atrial tissue and cardiovascular disease. Computed tomography is
dimension, which by itself is a strong predictor of atrial the most accurate imaging method to quantify epicardial adi-
fibrillation [4]. pose tissue. Low-dose, non-contrast acquisitions could
A potential lipotoxic effect of epicardial adipose tissue on potentially be performed specifically to quantify epicardial
left ventricular function has been suggested in the literature. fat. In addition, the information on individual volume and
However, an analysis in participants from the Framingham distribution of epicardial adipose tissue comes for free with
Heart Study failed to confirm a correlation of epicardial adi- the increasing numbers of clinically driven cardiac and coro-
pose tissue volume and magnetic resonance measures of left nary CT examinations. Several limitations must be consid-
ventricular structure or function, independent of general adi- ered: First of all, there may a coincidental co-localization of
posity [4]. In a study in 110 patients from a health screening higher amounts of perivascular adipose tissue and the typical
program without prevalent CAD, Cavalcante et al. described predilection sites of coronary atherosclerotic lesions or vul-
a significant association of EAT with e′ and E/e′ additionally nerable plaque (such as the proximal left anterior descending
to LA size, suggesting that epicardial adipose tissue may be coronary artery). Second, the best “threshold” to identify
associated with diastolic dysfunction [9]. Further studies are abnormally elevated epicardial fat volumes, probably some-
needed to confirm these results and establish potential mech- where around 100–125 cm3, has not yet been identified, and
292 M. Marwan

third, no data is available that would permit to base treatment coronary calcium and Metabolic Syndrome. Atherosclerosis.
decisions – specifically, intensification or withholding of risk 2010;209(1):136–41.
12. Gorter PM, de Vos AM, van der Graaf Y, Stella PR, Doevendans
modifying measures  – on the individually determined vol- PA, Meijs MF, et  al. Relation of epicardial and pericoronary
ume of epicardial adipose tissue. All the same, epicardial fat fat to coronary atherosclerosis and coronary artery calcium
clearly seems to be more than just a substance to fill empty in patients undergoing coronary angiography. Am J Cardiol.
space, and further prospective investigations will serve to 2008;102(4):380–5.
13. Sacks HS, Fain JN. Human epicardial adipose tissue: a review. Am
clarify its intriguing association with coronary artery disease Heart J. 2007;153(6):907–17.
and other cardiac disorders. 14. Iacobellis G, Corradi D, Sharma AM.  Epicardial adipose tissue:
Further prospective studies are needed to determine age- anatomic, biomolecular and clinical relationships with the heart.
and gender-specific reference values to help clinicians inter- Nat Clin Pract Cardiovasc Med. 2005;2(10):536–43.
15. Alexopoulos N, McLean DS, Janik M, Arepalli CD, Stillman AE,
pret individual epicardial fat volumes regarding their Raggi P. Epicardial adipose tissue and coronary artery plaque char-
association with cardiovascular risk and to potentially iden- acteristics. Atherosclerosis. 2010;210(1):150–4.
tify symptomatic and asymptomatic subjects at increased 16. Iacobellis G. Epicardial fat: a new cardiovascular therapeutic target.
risk who require more intensive risk factor modification. Curr Opin Pharmacol. 2016;27:13–8.
17. Bastarrika G, Broncano J, Schoepf UJ, Schwarz F, Lee YS, Abro
JA, et  al. Relationship between coronary artery disease and epi-
cardial adipose tissue quantification at cardiac CT: comparison
between automatic volumetric measurement and manual bidimen-
References sional estimation. Acad Radiol. 2010;17(6):727–34.
18. Sarin S, Wenger C, Marwaha A, Qureshi A, Go BD, Woomert
1. Mahabadi AA, Reinsch N, Lehmann N, Altenbernd J, Kalsch H, CA, et  al. Clinical significance of epicardial fat measured
Seibel RM, et al. Association of pericoronary fat volume with ath- using cardiac multislice computed tomography. Am J Cardiol.
erosclerotic plaque burden in the underlying coronary artery: a seg- 2008;102(6):767–71.
ment analysis. Atherosclerosis. 2010;211(1):195–9. 19. La Grutta L, Toia P, Farruggia A, Albano D, Grassedonio E, Palmeri
2. Mahabadi AA, Massaro JM, Rosito GA, Levy D, Murabito JM, A, et al. Quantification of epicardial adipose tissue in coronary cal-
Wolf PA, et  al. Association of pericardial fat, intrathoracic fat, cium score and CT coronary angiography image data sets: com-
and visceral abdominal fat with cardiovascular disease burden: the parison of attenuation values, thickness and volumes. Br J Radiol.
Framingham heart study. Eur Heart J. 2009;30(7):850–6. 2016;89(1062):20150773.
3. Greif M, Becker A, von Ziegler F, Lebherz C, Lehrke M, Broedl 20. Cheng VY, Dey D, Tamarappoo B, Nakazato R, Gransar H,

UC, et al. Pericardial adipose tissue determined by dual source CT Miranda-Peats R, et al. Pericardial fat burden on ECG-gated non-
is a risk factor for coronary atherosclerosis. Arterioscler Thromb contrast CT in asymptomatic patients who subsequently experi-
Vasc Biol. 2009;29(5):781–6. ence adverse cardiovascular events. JACC Cardiovasc Imaging.
4. Fox CS, Gona P, Hoffmann U, Porter SA, Salton CJ, Massaro 2010;3(4):352–60.
JM, et  al. Pericardial fat, intrathoracic fat, and measures of left 21. Nichols JH, Samy B, Nasir K, Fox CS, Schulze PC, Bamberg
ventricular structure and function: the Framingham Heart Study. F, et  al. Volumetric measurement of pericardial adipose tissue
Circulation. 2009;119(12):1586–91. from contrast-enhanced coronary computed tomography angi-
5. Rosito GA, Massaro JM, Hoffmann U, Ruberg FL, Mahabadi AA, ography: a reproducibility study. J Cardiovasc Comput Tomogr.
Vasan RS, et al. Pericardial fat, visceral abdominal fat, cardiovascu- 2008;2(5):288–95.
lar disease risk factors, and vascular calcification in a community-­ 22. Nakazato R, Shmilovich H, Tamarappoo BK, Cheng VY, Slomka
based sample: the Framingham Heart Study. Circulation. PJ, Berman DS, et al. Interscan reproducibility of computer-aided
2008;117(5):605–13. epicardial and thoracic fat measurement from noncontrast cardiac
6. Taguchi R, Takasu J, Itani Y, Yamamoto R, Yokoyama K, Watanabe CT. J Cardiovasc Comput Tomogr. 2011;5(3):172–9.
S, et  al. Pericardial fat accumulation in men as a risk factor for 23. Iwasaki K, Matsumoto T, Aono H, Furukawa H, Samukawa M.
coronary artery disease. Atherosclerosis. 2001;157(1):203–9. Relationship between epicardial fat measured by 64-­multidetector
7. Al Chekakie MO, Welles CC, Metoyer R, Ibrahim A, Shapira AR, computed tomography and coronary artery disease. Clin Cardiol.
Cytron J, et  al. Pericardial fat is independently associated with 2011;34(3):166–71.
human atrial fibrillation. J Am Coll Cardiol. 2010;56(10):784–8. 24. Tamarappoo B, Dey D, Shmilovich H, Nakazato R, Gransar H,
8. Zhu W, Zhang H, Guo L, Hong K. Relationship between epicardial Cheng VY, et al. Increased pericardial fat volume measured from
adipose tissue volume and atrial fibrillation : a systematic review noncontrast CT predicts myocardial ischemia by SPECT.  JACC
and meta-analysis. Herz. 2016;41(5):421–7. Cardiovasc Imaging. 2010;3(11):1104–12.
9. Cavalcante JL, Tamarappoo BK, Hachamovitch R, Kwon DH, 25. Shmilovich H, Dey D, Cheng VY, Rajani R, Nakazato R, Otaki Y,
Alraies MC, Halliburton S, et  al. Association of epicardial fat, et  al. Threshold for the upper normal limit of indexed epicardial
hypertension, subclinical coronary artery disease, and metabolic fat volume: derivation in a healthy population and validation in an
syndrome with left ventricular diastolic dysfunction. Am J Cardiol. outcome-based study. Am J Cardiol. 2011;108(11):1680–5.
2012;110(12):1793–8. 26. Lee BC, Lee WJ, Lo SC, Hsu HC, Chien KL, Chang YC, et al. The
10. Thanassoulis G, Massaro JM, Hoffmann U, Mahabadi AA,
ratio of epicardial to body fat improves the prediction of coronary
Vasan RS, O'Donnell CJ, et  al. Prevalence, distribution, and artery disease beyond calcium and Framingham risk scores. Int J
risk factor correlates of high pericardial and intrathoracic fat Cardiovasc Imaging. 2016;32(Suppl 1):117–27.
depots in the Framingham heart study. Circ Cardiovasc Imaging. 27. Iacobellis G, Bianco AC.  Epicardial adipose tissue: emerging

2010;3(5):559–66. physiological, pathophysiological and clinical features. Trends
11. Dey D, Wong ND, Tamarappoo B, Nakazato R, Gransar H, Cheng Endocrinol Metab. 2011;22(11):450–7.
VY, et  al. Computer-aided non-contrast CT-based quantifica- 28. Baker AR, Silva NF, Quinn DW, Harte AL, Pagano D, Bonser
tion of pericardial and thoracic fat and their associations with RS, et al. Human epicardial adipose tissue expresses a pathogenic
24  The Many Uses of Epicardial Fat Measurements 293

profile of adipocytokines in patients with cardiovascular disease. severe coronary stenosis, and high-risk coronary plaque features
Cardiovasc Diabetol. 2006;5:1. assessed by coronary CT angiography. J Cardiovasc Comput
29. Laine P, Kaartinen M, Penttila A, Panula P, Paavonen T, Kovanen Tomogr. 2013;7(2):125–32.
PT. Association between myocardial infarction and the mast cells 46. Wang TD, Lee WJ, Shih FY, Huang CH, Chen WJ, Lee YT,

in the adventitia of the infarct-related coronary artery. Circulation. et  al. Association of epicardial adipose tissue with coronary ath-
1999;99(3):361–9. erosclerosis is region-specific and independent of conventional
30. Mazurek T, Zhang L, Zalewski A, Mannion JD, Diehl JT, Arafat H, risk factors and intra-abdominal adiposity. Atherosclerosis.
et al. Human epicardial adipose tissue is a source of inflammatory 2010;213(1):279–87.
mediators. Circulation. 2003;108(20):2460–6. 47. Ohyama K, Matsumoto Y, Shimokawa H. Impact of epicardial adi-
31. Ouchi N, Kihara S, Arita Y, Okamoto Y, Maeda K, Kuriyama H, pose tissue volume quantified by non-contrast electrocardiogram-­
et  al. Adiponectin, an adipocyte-derived plasma protein, inhibits gated computed tomography on ergonovine-induced epicardial
endothelial NF-kappaB signaling through a cAMP-dependent path- coronary artery spasm. Int J Cardiol. 2017;229:40.
way. Circulation. 2000;102(11):1296–301. 48. Ito T, Fujita H, Ichihashi T, Ohte N. Impact of epicardial adipose
32. Ouchi N, Kihara S, Funahashi T, Nakamura T, Nishida M,
tissue volume quantified by non-contrast electrocardiogram-gated
Kumada M, et  al. Reciprocal association of C-reactive protein computed tomography on ergonovine-induced epicardial coronary
with adiponectin in blood stream and adipose tissue. Circulation. artery spasm. Int J Cardiol. 2016;221:877–80.
2003;107(5):671–4. 49. Oka T, Yamamoto H, Ohashi N, Kitagawa T, Kunita E, Utsunomiya
33. Saijo Y, Kiyota N, Kawasaki Y, Miyazaki Y, Kashimura J, Fukuda H, et al. Association between epicardial adipose tissue volume and
M, et al. Relationship between C-reactive protein and visceral adi- characteristics of non-calcified plaques assessed by coronary com-
pose tissue in healthy Japanese subjects. Diabetes Obes Metab. puted tomographic angiography. Int J Cardiol. 2012;161(1):45–9.
2004;6(4):249–58. 50. Ito T, Suzuki Y, Ehara M, Matsuo H, Teramoto T, Terashima M,
34.
Wajchenberg BL.  Subcutaneous and visceral adipose tis- et  al. Impact of epicardial fat volume on coronary artery disease
sue: their relation to the metabolic syndrome. Endocr Rev. in symptomatic patients with a zero calcium score. Int J Cardiol.
2000;21(6):697–738. 2013;167(6):2852–8.
35. Yatagai T, Nagasaka S, Taniguchi A, Fukushima M, Nakamura T, 51. Ito T, Nasu K, Terashima M, Ehara M, Kinoshita Y, Ito T, et al. The
Kuroe A, et al. Hypoadiponectinemia is associated with visceral fat impact of epicardial fat volume on coronary plaque vulnerability:
accumulation and insulin resistance in Japanese men with type 2 insight from optical coherence tomography analysis. Eur Heart J
diabetes mellitus. Metabolism. 2003;52(10):1274–8. Cardiovasc Imaging. 2012;13(5):408–15.
36. Arita Y, Kihara S, Ouchi N, Maeda K, Kuriyama H, Okamoto 52. Harada K, Amano T, Uetani T, Tokuda Y, Kitagawa K, Shimbo Y,
Y, et  al. Adipocyte-derived plasma protein adiponectin acts as a et al. Cardiac 64-multislice computed tomography reveals increased
platelet-­derived growth factor-BB-binding protein and regulates epicardial fat volume in patients with acute coronary syndrome.
growth factor-induced common postreceptor signal in vascular Am J Cardiol. 2011;108(8):1119–23.
smooth muscle cell. Circulation. 2002;105(24):2893–8. 53. Lu MT, Park J, Ghemigian K, Mayrhofer T, Puchner SB, Liu T,
37. Hajer GR, van Haeften TW, Visseren FL.  Adipose tissue dys- et al. Epicardial and paracardial adipose tissue volume and attenu-
function in obesity, diabetes, and vascular diseases. Eur Heart J. ation  – association with high-risk coronary plaque on computed
2008;29(24):2959–71. tomographic angiography in the ROMICAT II trial. Atherosclerosis.
38. Tavora F, Kutys R, Li L, Ripple M, Fowler D, Burke A. Adventitial 2016;251:47–54.
lymphocytic inflammation in human coronary arteries with intimal 54. Janik M, Hartlage G, Alexopoulos N, Mirzoyev Z, McLean DS,
atherosclerosis. Cardiovasc Pathol. 2010;19(3):e61–8. Arepalli CD, et al. Epicardial adipose tissue volume and coronary
39. Konishi M, Sugiyama S, Sato Y, Oshima S, Sugamura K, Nozaki T, artery calcium to predict myocardial ischemia on positron emis-
et al. Pericardial fat inflammation correlates with coronary artery sion tomography-computed tomography studies. J Nucl Cardiol.
disease. Atherosclerosis. 2010;213(2):649–55. 2010;17(5):841–7.
40. Hirata Y, Tabata M, Kurobe H, Motoki T, Akaike M, Nishio C, 55. Khawaja T, Greer C, Thadani SR, Kato TS, Bhatia K, Shimbo
et  al. Coronary atherosclerosis is associated with macrophage D, et al. Increased regional epicardial fat volume associated with
polarization in epicardial adipose tissue. J Am Coll Cardiol. reversible myocardial ischemia in patients with suspected coronary
2011;58(3):248–55. artery disease. J Nucl Cardiol. 2015;22(2):325–33.
41. Uchida Y, Uchida Y, Shimoyama E, Hiruta N, Kishimoto T, Watanabe 56. Nakanishi R, Rajani R, Cheng VY, Gransar H, Nakazato R,

S. Human pericoronary adipose tissue as storage and possible sup- Shmilovich H, et al. Increase in epicardial fat volume is associated
ply site for oxidized low-density lipoprotein and high-­density lipo- with greater coronary artery calcification progression in subjects at
protein in coronary artery. J Cardiol. 2017;69(1):236–44. intermediate risk by coronary calcium score: a serial study using
42. Prati F, Arbustini E, Labellarte A, Sommariva L, Pawlowski T, non-contrast cardiac CT. Atherosclerosis. 2011;218(2):363–8.
Manzoli A, et  al. Eccentric atherosclerotic plaques with posi- 57. Yerramasu A, Dey D, Venuraju S, Anand DV, Atwal S, Corder R,
tive remodelling have a pericardial distribution: a permissive role et al. Increased volume of epicardial fat is an independent risk fac-
of epicardial fat? A three-dimensional intravascular ultrasound tor for accelerated progression of sub-clinical coronary atheroscle-
study of left anterior descending artery lesions. Eur Heart J. rosis. Atherosclerosis. 2012;220(1):223–30.
2003;24(4):329–36. 58. Hwang IC, Park HE, Choi SY.  Epicardial adipose tissue contrib-
43. Ishii T, Asuwa N, Masuda S, Ishikawa Y.  The effects of a myo- utes to the development of non-calcified coronary plaque: a 5-year
cardial bridge on coronary atherosclerosis and ischaemia. J Pathol. computed tomography follow-up study. J Atheroscler Thromb.
1998;185(1):4–9. 2017;24(3):262–74.
44. Konishi M, Sugiyama S, Sugamura K, Nozaki T, Ohba K, Matsubara 59. Ding J, Hsu FC, Harris TB, Liu Y, Kritchevsky SB, Szklo M, et al.
J, et  al. Association of pericardial fat accumulation rather than The association of pericardial fat with incident coronary heart
abdominal obesity with coronary atherosclerotic plaque formation ­disease: the Multi-Ethnic Study of Atherosclerosis (MESA). Am J
in patients with suspected coronary artery disease. Atherosclerosis. Clin Nutr. 2009;90(3):499–504.
2010;209(2):573–8. 60. Mahabadi AA, Berg MH, Lehmann N, Kalsch H, Bauer M,

45. Rajani R, Shmilovich H, Nakazato R, Nakanishi R, Otaki Y, Cheng Kara K, et  al. Association of epicardial fat with cardiovascu-
VY, et al. Relationship of epicardial fat volume to coronary plaque, lar risk factors and incident myocardial infarction in the general
294 M. Marwan

population: the Heinz Nixdorf Recall Study. J Am Coll Cardiol. 65. Thanassoulis G, Massaro JM, O'Donnell CJ, Hoffmann U, Levy
2013;61(13):1388–95. D, Ellinor PT, et  al. Pericardial fat is associated with prevalent
61. Forouzandeh F, Chang SM, Muhyieddeen K, Zaid RR, Trevino atrial fibrillation: the Framingham Heart Study. Circ Arrhythm
AR, Xu J, et al. Does quantifying epicardial and intrathoracic fat Electrophysiol. 2010;3(4):345–50.
with noncontrast computed tomography improve risk stratifica- 66. Tanami Y, Jinzaki M, Kishi S, Matheson M, Vavere AL, Rochitte
tion beyond calcium scoring alone? Circ Cardiovasc Imaging. CE, et  al. Lack of association between epicardial fat volume
2013;6(1):58–66. and extent of coronary artery calcification, severity of coronary
62. Spearman JV, Renker M, Schoepf UJ, Krazinski AW, Herbert TL, artery disease, or presence of myocardial perfusion abnormali-
De Cecco CN, et al. Prognostic value of epicardial fat volume mea- ties in a diverse, symptomatic patient population: results from
surements by computed tomography: a systematic review of the the CORE320 multicenter study. Circ Cardiovasc Imaging.
literature. Eur Radiol. 2015;25(11):3372–81. 2015;8(3):e002676.
63. Oikonomou E, Marwan M, et al. Non-invasive detection of coro- 67. Hartiala O, Magnussen CG, Bucci M, Kajander S, Knuuti J,

nary inflammation using computed tomography and prediction of Ukkonen H, et  al. Coronary heart disease risk factors, coronary
residual cardiovascular risk (the CRISP CT study): a post-hoc anal- artery calcification and epicardial fat volume in the Young Finns
ysis of prospective outcome data. Lancet. 2018;392(10151):929– Study. Eur Heart J Cardiovasc Imaging. 2015;16(11):1256–63.
39. https://doi.org/10.1016/S0140-6736(18)31114-0. 68. Romijn MA, Danad I, Bakkum MJ, Stuijfzand WJ, Tulevski II,
64. Wong CX, Abed HS, Molaee P, Nelson AJ, Brooks AG, Sharma G, Somsen GA, et al. Incremental diagnostic value of epicardial adi-
et al. Pericardial fat is associated with atrial fibrillation severity and pose tissue for the detection of functionally relevant coronary artery
ablation outcome. J Am Coll Cardiol. 2011;57(17):1745–51. disease. Atherosclerosis. 2015;242(1):161–6.
 Part VI
Where We Are: Non-invasive Coronary Artery
Imaging
 Nonatherosclerotic Coronary Artery
Disease 25
Toru Sakuma, Kotaro Ouchi, and Kunihiko Fukuda

Atherosclerotic coronary artery disease (CAD) is not appar- early 1970s, coronary angiography demonstrated spasms of
ent in coronary angiography or necropsy in approximately the epicardial coronary artery during such an attack [3]. We
4–7% of all patients with acute myocardial infarction (AMI) now know that coronary artery spasm can lead to arrhyth-
[1]. Showing only the lumina of arteries, conventional coro- mias and various myocardial ischemic diseases, including
nary angiography offers very low specificity for diagnosing the variant angina, which is now described as vasospastic
CAD etiology. However, the temporal and spatial resolution angina, as well as acute coronary syndrome (ACS), acute
of coronary angiography using computed tomography (CT) myocardial infarction, and sudden death [4, 5]. Coronary
allows depiction of the vessel wall and adjacent tissue as artery spasm is considered a disease of middle- to old-aged
well as the arterial lumen that permits the safe and noninva- men and postmenopausal women [6]; premenopausal women
sive detection of various disease manifestations of both non- benefit from the protective effects of estrogen on dilator
atherosclerotic and atherosclerotic CAD. response in coronary arteries [7].
Nonatherosclerotic CAD is rare but can result in severe A recent survey in Japan showed coronary artery spasm in
and life-threatening complications, including coronary artery 40.9% of consecutive patients with angina pectoris who
aneurysm, spasm, dissection, stenosis, and intraluminal underwent coronary angiography [8]. The prominent risk
thrombosis, that may occur in very young patients and are factors for such spasm are age, smoking, and presence in the
often silent in the early stages of disease. Accurate diagnosis plasma of high sensitivity C-reactive protein (hsCRP) and
and proper patient management require that clinicians recog- remnant lipoproteins, and the rate of smoking is significantly
nize CT findings associated with the various manifestations higher among patients with coronary spastic angina than
of nonatherosclerotic CAD. those with stable effort angina [9]. The exact mechanisms of
We present an overview that includes the pathophysiol- coronary artery spasm remain unknown and may be related
ogy, etiology, and findings on coronary CT angiography of to endothelial dysfunction, polymorphisms of the eNOS
coronary artery spasm, dissection, and vasculitis as well as gene, oxidative stress, chronic low-grade inflammation,
immune-mediated conditions and lymphoproliferative disor- hypercontractility of the coronary smooth muscle, and mag-
der – information that is clinically useful and that can facili- nesium deficiency [10].
tate accurate diagnosis and treatment. Coronary artery spasm can be reliably diagnosed by prov-
ocation test using ergonovine or acetylcholine during coro-
nary angiography, but the test is invasive and can induce
Coronary Artery Spasm severe myocardial ischemia or arrhythmia [11]. As a nonin-
vasive alternative, coronary CT angiography can depict
In 1959, Prinzmetal et al. [2] described what they termed a abnormalities of the coronary arteries, but its use does not
variant form of classic angina pectoris that occurred at rest, guarantee visualization of the spasm, which is temporary and
was associated with elevation of the ST segment on electro- rarely occurs during the daytime, and imaging follows sub-
cardiography (ECG), and was not induced by exertion. In the lingual administration of nitroglycerine to dilate the coro-
nary arteries. Nevertheless, Ito’s group [12] reported the
utility of coronary CT angiography to predict coronary artery
T. Sakuma (*) · K. Ouchi · K. Fukuda spasm, observing noncalcified plaques with intermediate
Department of Radiology, The Jikei University School of attenuation and negative remodeling of the vessel diameter at
Medicine, Tokyo, Japan
e-mail: [email protected]
spasm sites (Fig. 25.1). Male patients with these c­ haracteristic

© Humana Press 2019 297

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_25
298 T. Sakuma et al.

c d

Fig. 25.1  Coronary computed tomography angiography in a 53-year-­ at the middle of the LCX. Cross-sectional images (b) at the middle of
old man with chest pain at rest in the morning and suspected vaso- the LCX show noncalcified plaque with negative remodeling (white
spastic angina. Stretched multiplanar reconstruction (MPR) image arrow). This patient was diagnosed with coronary artery spasm using
(a), curved MPR image (c) of the left circumflex artery (LCX), and the provocation test
angiographic view of the left coronary artery (d) demonstrate ­stenosis

plaques should be given medication to prevent coronary Coronary Artery Dissection

artery spasm and/or undergo provocation testing to prevent
fatal cardiac events. A pilot study by Kang and colleagues Coronary artery dissection has been categorized as either pri-
[13] demonstrated the safety and feasibility of a protocol mary, that is, occurring spontaneously, or secondary, occur-
using double-acquisition coronary CT angiography with and ring after coronary angiography, coronary intervention,
without intravenous infusion of nitrate as a noninvasive cardiac surgery, trauma, or dissection of the aortic root.
means of depicting vasospastic angina, yielding moderate Spontaneous coronary artery dissection (SCAD) is defined
sensitivity (73%) and high specificity (100%). Nonetheless, as an idiopathic separation of the coronary arterial wall by
sensitivity remained insufficient to rule out vasospastic intramural hemorrhage that creates a false lumen with or
angina without the provocation test. without intimal tear. If the separation occurs in the media,
25  Nonatherosclerotic Coronary Artery Disease 299

Fig. 25.2 Coronary
computed tomography
angiography in a 70-year-old
man with chest pain during
walking and suspected effort
angina. Electrocardiography-­
gated reconstruction images
of the right coronary artery
(RCA) show dual channels
and the intimal flap in the
coronary lumen (arrowheads),
which represent spontaneous
coronary artery dissection
(SCAD). The middle part of
the SCAD shows aneurysmal
dilatation with the
thrombosed false lumen and
narrowed true lumen (white
arrow)

the resulting intramural hematoma or enlarged false lumen diagnosis and assessment of SCAD because it allows direct
compresses the true arterial lumen, thereby compromising visualization of the coronary wall as well as the arterial
blood flow of the coronary artery, and may cause chest pain, lumen without the use of invasive methods like IVUS or
subsequent myocardial ischemia, infarction, acute coronary OCT (Fig. 25.2).
syndrome, ventricular fibrillation, or sudden death. SCAD is A large prospective study by Tweet and colleagues [21]
atherosclerotic or nonatherosclerotic. Predisposing factors of suggests initial conservative management as a reasonable
the nonatherosclerotic type include peripartum state, con- strategy for patients with SCAD. Those authors observed a
nective tissue disorders, coronary artery spasm, fibromuscu- 10-year rate of recurrence after a primary SCAD event of
lar dysplasia, and systemic inflammatory diseases [14]. 29.4%, with the median time to a second episode of 2.8 years
In 1931, Pretty [15] first reported the discovery of SCAD (range, 3 days–12 years), and a 10-year rate of major adverse
at autopsy in a 42-year-old woman who had died suddenly, cardiac events (death, recurrent SCAD, myocardial infarc-
and Forker’s team [16] reported the first angiographic diag- tion, and congestive heart failure) of 47.4% [21]. Although
nosis in 1973. SCAD has been reported in an estimated SCAD demonstrates favorable long-term survival, the rate of
0.07–1.1% of patients undergoing coronary angiography associated major adverse events is high, and its recurrence is
[17], with higher prevalence (3–4%) among patients with unpredictable. Its management therefore requires both close
acute coronary syndrome than those with stable symptoms and long-term follow-up [22]. Noninvasive coronary CT
(0.3%) and even higher prevalence (8.7%) among women angiography may be the optimal imaging method for the
younger than 50 years with ACS [18]. follow-up of patients with SCAD who are initially managed
Though coronary angiography is most widely used to conservatively.
diagnose SCAD, the inability of conventional coronary angi-
ography to depict abnormalities of the arterial wall can pre-
vent definitive diagnosis. Thus, the appearance of multiple Takayasu Arteritis
lumina on conventional coronary angiography demonstrates
medial dissection with intimal tear, but stenosis or occlusion Takayasu arteritis (TAK) is an idiopathic systemic granulo-
of the lumina caused by the dissecting medial hematoma fol- matous vasculitis that involves medium and large arteries
lowing rupture of the vasa vasorum can be mistaken for ste- and is especially prominent in the aortic arch and subclavian
nosis or occlusion due to atherosclerosis. In this setting, and carotid arteries. Mikito Takayasu first described the dis-
intravascular ultrasound (IVUS) or optical coherence tomog- ease in 1905, when he presented the case of a 21-year-old
raphy (OCT) provides more detailed information about the woman with arteriovenous anastomosis surrounding the
coronary lumen and vessel wall of the lesion of SCAD [19, papilla of the eyeground. The 2012 revision of the
20]. Coronary CT angiography may also be useful in the Nomenclature of Vasculitides of the International Chapel
300 T. Sakuma et al.

Fig. 25.3 Coronary
a b
computed tomography
angiography in a 29-year-old
woman with active-phase
Takayasu arteritis. (a) Axial
image depicts marked wall
thickening of the ascending
aorta (arrowheads). (b)
Cross-sectional multiplanar
reconstruction image of the
root of the ascending aorta
demonstrates constriction of
the left main trunk (white
arrow) passing through the
thickened wall of the sinus of
Valsalva

Hill Consensus Conference (CHCC) now categorizes TAK coronary aneurysm [33]. Coronary ostial stenosis in patients
along with giant cell arteritis (GCA) as a vasculitis of large with TAK is attributed to the extension of the active inflam-
vessels (LVV), defining it as arteritis that is often granuloma- matory process from the ascending aorta with thickening and
tous and that predominantly affects the aorta and/or its major enhancement of the arterial wall in the early and active dis-
branches [23]. Onset usually occurs in patients younger than ease phases (Fig. 25.3). In the later and chronic phases, pro-
age 50 but has been observed in older patients. Reported all longed inflammation of the coronary vessels may produce
over the world, TAK is seen more often in Japan, Southeast diffuse or focal stenosis and the formation of aneurysms
Asia, Korea, India, China, and Mexico than in the United (Fig. 25.4). Kang and colleagues [24] observed a lower inci-
States and Europe [24]. Its annual incidence is relatively dence of non-ostial coronary arterial stenosis in patients with
similar in different countries and ranges from 0.4 to 2.6 cases active disease.
per million in the population [25]. The highest rate is 40
cases per million in the Japanese population. Patients with
TAK are predominantly female (82.9–97.0%) and most com- Kawasaki Disease
monly young women.
Cardiac manifestations of TAK include myocarditis, val- Kawasaki disease (KD) is an acute febrile illness that causes
vular regurgitation, and coronary involvement that may lead mucosal inflammation, skin rash, and cervical lymphade-
to myocardial infarction, heart failure, or sudden death [26– nopathy, is most often recognized in children younger than
28]. As minor diagnostic criteria of TAK, Ishikawa included 4  years of age, and is of unknown etiology. Dr. Tomisaku
annuloaortic ectasia with aortic regurgitation in 1988 [29], Kawasaki described the disease first in the Japanese litera-
and Sharma’s group added the presence of a coronary artery ture in 1967 [34] and then in English in 1974 [35]. KD pro-
lesion in the absence of risk factors in patients under age duces systemic vasculitis that is most severe in medium-sized
30 in 1995 [30]. arteries and especially prominent in the coronary arteries.
Conventional coronary angiography reveals coronary The 2012 CHCC nomenclature now categorizes KD along
artery disease in approximately 10–30% of cases [27, 31, with polyarteritis nodosa (PAN) as a vasculitis of medium
32]. In contrast, CT angiography can show abnormalities of vessels (MVV), defining it as arteritis associated with muco-
the vascular wall as well as luminal information and has been cutaneous lymph node syndrome that predominantly affects
reported to depict coronary involvement in 53.2% of patients medium and small arteries. The disease often involves the
with TAK regardless of symptoms and disease activity [24]. coronary arteries, may involve the aorta and large arteries,
An autopsy study of patients with TAK highlighted three and usually occurs in infants and young children [23].
abnormal conditions of the coronary artery  – stenosis or There is no specific test to identify the presence of KD, so
occlusion of the ostia, diffuse or focal coronary arteritis, and it is diagnosed according to the guidelines of the American
25  Nonatherosclerotic Coronary Artery Disease 301

a b

Fig. 25.4  Computed tomography angiography of the thoracic and Sagittal image of the thoracic and abdominal aorta also shows a large
abdominal aorta in a 72-year-old woman with chronic-phase Takayasu aneurysm (arrowhead) in the distal right coronary artery. Severe calcifi-
arteritis. (a) Coronal image of the chest shows a large coronary artery cation of the thoracic and abdominal aorta, aneurysmal dilatation of the
aneurysm (white arrow) in the left main trunk to the proximal left ante- ascending aorta, and narrowing of the descending and abdominal aorta
rior descending artery. Severe calcification is noted in the aortic arch, are also noted with occlusion of the left common carotid and subclavian
and dilated collateral arteries are seen in the right lateral chest wall. (b) arteries

Heart Association/American Academy of Pediatrics [36] and Orenstein’s team [40] reported three linked vasculopathic
JCS Joint Working Group [37]. When coronary aneurysm is processes in the arterial wall – necrotizing arteritis, subacute/
recognized on transthoracic echocardiography or coronary chronic arteritis, and luminal myofibroblastic proliferation
angiography (CAG) in patients with fever of at least 5 days’ (LMP). Necrosis of the vessel wall in response to the infiltra-
duration, KD can be diagnosed by the observation of four of tion of neutrophils contributes to aneurysmal formation in
its five principal clinical features – changes in the extremi- the coronary arteries in the acute phase of KD, and in the late
ties, polymorphous exanthem, bilateral bulbar conjunctival phase, LMP narrows the lumina of the arteries [41].
injection without exudate, changes in the lips and oral cavity, Echocardiography is employed for the initial evaluation and
and cervical lymphadenopathy [36]. Associated morbidity follow-up of lesions of the coronary arteries in KD because
and mortality may be considerable, mostly resulting from it is considered to be the most useful and essential method to
myocardial involvement and such coronary artery complica- evaluate coronary aneurysms. However, as the child grows,
tions as aneurysm, calcification, and stenosis. KD is a lead- visualization of the coronary arteries becomes progressively
ing cause of acquired heart disease in children in the United more difficult by transthoracic echocardiography. Evaluation
States as well as Japan [38]. of aneurysms also becomes more difficult as wall calcifica-
Prevention of these cardiac abnormalities is the primary tion and luminal narrowing progress. For these reasons, cor-
goal of its treatment. Though high-dose intravenous immu- onary CT angiography is the primary modality used in the
noglobulin (IVIG) has been administered as the standard lifelong surveillance of the coronary arteries.
therapy in the acute phase of KD, failure of this treatment has Medium and giant aneurysms are often associated with
been shown in 20.3% of patients, and these patients are at thrombotic occlusion in the relatively early stage of KD, but
high risk to develop coronary artery complications, particu- neovascularization can occur. Such development of new ves-
larly aneurysm [39]. A second infusion of IVIG and inflix- sels has been reported in 15% of patients with coronary
imab is currently used as a follow-up treatment in this group artery lesions and on the right coronary artery in 90% of
of initial nonresponders. these cases [42] (Fig. 25.5). In addition, localized stenosis,
302 T. Sakuma et al.

c d

Fig. 25.5  Coronary computed tomography angiography in a 32-year-­ part. Cross-sectional images (b) show multiple channels (white arrow)
old man with Kawasaki disease. Stretched multiplanar reconstruction in the coronary lumen, which represent recanalization of the throm-
(MPR) image (a), curved MPR image (c), and volume-rendering image bosed aneurysm
(d) of the right coronary artery show a giant aneurysm in its proximal

particularly at the entrance or exit of an aneurysm, developed I mmunoglobulin G4-Related Disease

in 4.7–12% of patients with coronary artery lesions between (IgG4-­Related Disease)
10 and 20  years after onset (Fig.  25.6), and findings of a
Japanese multicenter survey reported by Suzuki’s group [43] IgG4-related disease is a recently recognized multi-organ,
showed that half of patients with giant aneurysms underwent immune-mediated condition characterized by a high concen-
coronary artery bypass grafting (CABG). Coronary CT angi- tration of IgG4-positive plasma cells in the site of disease
ography is also thought to be a promising noninvasive alter- with or without elevation of the serum level of IgG4.
native for the assessment of the patency of bypass grafts or Kamisawa and associates [44, 45] first recognized IgG4-­
residual aneurysms and stenosis. related disease as a systemic disorder in 2003, noting
25  Nonatherosclerotic Coronary Artery Disease 303

e­ xtrapancreatic features in patients with autoimmune pan-

creatitis. In Japan, the overall prevalence of IgG4-related
autoimmune pancreatitis has been estimated as 2.2 cases per
100,000 in the population [46]. Because it affects multiple
organs in addition to the pancreas, its true prevalence is
unclear and may be underestimated. A national study of
autoimmune pancreatitis in Japan [47] showed a greater
prevalence of IgG4-related disease among middle- to old-
aged men and greater prevalence in male than female patients
in a ratio of 2.8 to 1. However, no male predilection is noted
in the head and neck lesions. IgG4-related disease has been
described in multiple organs  – the biliary tree, salivary
glands, periorbital tissues, kidneys, lungs, lymph nodes,
meninges, aorta, breast, prostate, thyroid, pericardium, and
skin. It is also now known to comprise many and various
conditions, including autoimmune pancreatitis, eosinophilic
angiocentric fibrosis, fibrosing mediastinitis, hypertrophic
pachymeningitis, idiopathic hypocomplementemic tubuloin-
terstitial nephritis with extensive tubulointerstitial deposits,
Fig. 25.6 Coronary computed tomography angiography in a inflammatory pseudotumor, Küttner’s tumor (chronic scle-
36-year-­old man with Kawasaki disease. Short-axis multiplanar rosing sialadenitis), Mikulicz syndrome, multifocal fibro-
reconstruction image shows a giant aneurysm (25 mm in diameter)
in the proximal right coronary artery. The entrance of the aneurysm sclerosis, periaortitis and periarteritis, inflammatory aortic
(white arrow) is occluded, and the aneurysm is thrombosed com- aneurysm, retroperitoneal fibrosis, Riedel’s thyroiditis, and
pletely. Giant aneurysms are also detected in the proximal left ante- sclerosing mesenteritis [48]. In the late phases after the
rior descending artery, which is patent, and the proximal left administration of contrast material on CT, arterial lesions,
circumflex artery, which also remains patent despite the large
thrombosis including inflammatory aneurysm, retroperitoneal fibrosis,

Fig. 25.7 Coronary
computed tomography
a b
angiography in a 70-year-old
man with suspected
immunoglobulin G4-related
disease. Long-axis
multiplanar reconstruction
image (a) and axial image (b)
of the proximal right coronary
artery (RCA) show
periarterial disease (white
arrow) of the soft tissue
density that causes narrowing
of the RCA
304 T. Sakuma et al.

periaortitis, and periarteritis, have shown homogeneous wall been associated with malignant tumor [53]. We must be
thickening and enhancement [49]. In contrast to the vasculi- aware that the pseudotumor of IgG4-related disease resem-
tis of large vessels – giant cell and Takayasu arteritis – that bles the invasion of malignant lymphoma around the coro-
affects the aortic branches, especially the subclavian arteries, nary artery [54] (Fig. 25.8).
IgG4-related disease often avoids these vessels. Several
recent studies have suggested its association with such coro-
nary artery diseases as coronary aneurysm, periarterial pseu- Chronic Active Epstein-Barr Virus Infection
dotumor (Fig. 25.7), wall calcification, and intimal thickening
[50–52]. However, the coronary artery abnormalities of Epstein, Achong, and Barr first discovered the Epstein-Barr
IgG4-related disease have not been fully revealed, and their virus (EBV) in 1964 within cells cultured from Burkitt’s
incidence remains unclear. IgG4-related disease has also lymphoma tissue [55]. EBV is a herpes virus that infects
over 90% of humans and persists throughout life. Primary
EBV infection usually occurs in infants and children and is
generally asymptomatic or shows nonspecific symptoms.
Infections in adolescents and adults frequently result in
infectious mononucleosis [56]. EBV directly infects resting
B cells of the oropharynx via oral secretions [57]. The infec-
tion is controlled by cellular and humoral immune response,
but the virus persists in the resting memory B cells [58].
Patients with congenital or acquired immunodeficiency have
impaired cellular immunity and are unable to regulate the
proliferation of EBV-infected B cells. Their EBV infections
result in infectious mononucleosis and localized or dissemi-
nated lymphoproliferative disease involving the lymph
nodes, liver, lung, kidney, bone marrow, central nervous sys-
tem, or small intestine [59]. In response to unknown signals,
memory B cells may express uncontrolled growth that leads
to the development of such tumors as nasopharyngeal carci-
Fig. 25.8  Computed tomography of the chest in a 77-year-old woman noma, Burkitt’s lymphoma, and Hodgkin’s lymphoma, as
with diffuse large B-cell lymphoma reveals a recurrent tumor (white well as other cancers [57].
arrow) of the soft tissue density around the calcified right coronary CAEBV demonstrates the clonal expansion of EBV-­
artery. It is sometimes difficult to distinguish immunoglobulin
G4-related disease of the coronary artery from malignant lymphoma of infected T or natural killer cells. Clinically, patients with
the heart CAEBV show persistent and recurrent symptoms that

Fig. 25.9 Coronary
a b
computed tomography
angiography in a 37-year-old
man with chronic active
Epstein-Barr virus infection.
Volume-rendering image (a)
and short-axis multiplanar
reconstruction image (b)
show diffuse and
inhomogeneous dilatation of
the coronary arteries. The
anomalous origin of the right
coronary artery is also
depicted
25  Nonatherosclerotic Coronary Artery Disease 305

able to incidentally detect the findings of the disease before

the development of life-threatening complications.

References
1. Waller BF.  Nonatherosclerotic coronary heart disease. In: Fuster
V, Walsh RA, Harrington RA, editors. Hurst’s the heart. 13th ed.
New York: McGraw-Hill Education; 2011. p. 1257–86.
2. Prinzmetal M, Kennamer R, Merliss R, Wada T, Bor N.  Angina
pectoris. I. A variant form of angina pectoris; preliminary report.
Am J Med. 1959;27:375–88.
3. Oliva PB, Potts DE, Pluss RG.  Coronary arterial spasm in
Prinzmetal angina. Documentation by coronary arteriography. N
Engl J Med. 1973;288:745–51.
4. Nakamura M, Takeshita A, Nose Y. Clinical characteristics associ-
ated with myocardial infarction, arrhythmias, and sudden death in
Fig. 25.10  Contrast-enhanced computed tomography of the abdomen
patients with vasospastic angina. Circulation. 1987;75:1110–6.
in the same patient shown in Fig. 25.9 shows diffuse wall thickening of
5. Ong P, Athanasiadis A, Hill S, Vogelsberg H, Voehringer M,
the abdominal aorta, which represents chronic active arteritis
Sechtem U.  Coronary artery spasm as a frequent cause of acute
coronary syndrome: the CASPAR (Coronary Artery Spasm in
r­ esemble those of infectious mononucleosis, including fever, Patients with Acute Coronary Syndrome) study. J Am Coll Cardiol.
lymphadenopathy, and hepatosplenomegaly, and they expe- 2008;52:523–7.
6. Yasue H, Kugiyama K.  Coronary spasm: clinical features and
rience many other complications, such as hematological, pathogenesis. Intern Med. 1997;36:760–5.
digestive tract, neurological, pulmonary, ocular, dermal, and/ 7. Williams JK, Adams MR, Klopfenstein HS.  Estrogen modu-
or cardiovascular disorders (including aneurysm and valvu- lates responses of atherosclerotic coronary arteries. Circulation.
lar disease) [60]. 1990;81:1680–7.
8. Yasue H, Sasayama S, Kikuchi K, Okumura K, Matsubara T, Miwa
First described by Virelizier’s group [61] in 1978, CAEBV K, et  al. The role of coronary spasm in ischemic heart disease
is now recognized as not only an infectious disease but also among Japanese. In: Annual report of the research on cardiovascu-
as one of the EBV-associated T- and NK-cell lar diseases. Osaka: National Cardiovascular Center; 2000. p. 96–7.
­lymphoproliferative disorders. Okano and associates [62] [in Japanese].
9. Takaoka K, Yoshimura M, Ogawa H, Kugiyama K, Nakayama M,
proposed a guideline for diagnosing CAEBV in 2005. Shimasaki Y, et  al. Comparison of the risk factors for coronary
Patients with CAEBV commonly experience cardiovascular artery spasm with those for organic stenosis in a Japanese popula-
complications because the EBV-infected cells infiltrate the tion: role of cigarette smoking. Int J Cardiol. 2000;72:121–6.
vascular wall and induce vasculitis, and aneurysm of the 10. Yasue H, Nakagawa H, Itoh T, Harada E, Mizuno Y.  Coronary
artery spasm--clinical features, diagnosis, pathogenesis, and treat-
coronary artery or sinus of Valsalva has been reported [63, ment. J Cardiol. 2008;51:2–17.
64]. Muneuchi’s group [65] reported cardiovascular disease 11. Szlachcic J, Waters DD, Miller D, Théroux P. Ventricular arrhyth-
in 60% of patients with CAEBV, which included coronary mias during ergonovine-induced episodes of variant angina. Am
artery aneurysm (Fig. 25.9), myocarditis, large-vessel arteri- Heart J. 1984;107:20–4.
12. Ito T, Terashima M, Kaneda H, Nasu K, Ehara M, Kinoshita Y, et al.
tis (Fig. 25.10), pulmonary artery hypertension, pericardial In vivo assessment of ergonovine-induced coronary artery spasm
effusion, decreased cardiac function, and complete atrioven- by 64-slice multislice computed tomography. Circ Cardiovasc
tricular block, and Kimura and colleagues [66] reported cor- Imaging. 2012;5:226–32.
onary artery aneurysm and myocarditis as the complications 13. Kang EJ, Kim MH, De Jin C, Seo J, Kim DW, Yoon SK, et  al.
Noninvasive detection of coronary vasospastic angina using a
with the poorest prognoses. double-­acquisition coronary CT angiography protocol in the
presence and absence of an intravenous nitrate: a pilot study. Eur
Radiol. 2016. https://doi.org/10.1007/s00330-016-4476-2.
Summary 14. Saw J, Ricci D, Starovoytov A, Fox R, Buller CE.  Spontaneous
coronary artery dissection: prevalence of predisposing conditions
including fibromuscular dysplasia in a tertiary center cohort. JACC
Nonatherosclerotic CAD is associated with fatal myocardial Cardiovasc Interv. 2013;6:44–52.
infarction or sudden cardiac death in all age groups including 15. Pretty HC.  Dissecting aneurysm of coronary artery in a woman
young age. Cardiologists and radiologists need to be aware aged 42: rupture. Br Med J. 1931;1:667.
16. Forker AD, Rosenlof RC, Weaver WF, Carveth SW, Reese

of the differential diagnosis of nonatherosclerotic CAD in HE. Primary dissecting aneurysm of the right coronary artery with
patients with cardiac symptoms. Coronary CT angiography survival. Chest. 1973;64:656–8.
is a very useful and noninvasive tool to investigate the vari- 17.
Vrints CJ.  Spontaneous coronary artery dissection. Heart.
ous manifestations of the nonatherosclerotic CAD and is 2010;96:801–8.
306 T. Sakuma et al.

18. Vanzetto G, Berger-Coz E, Barone-Rochette G, Chavanon N,

37. JCS Joint Working Group. Guidelines for diagnosis and manage-
Bouvaist H, Hacini R, et  al. Prevalence, therapeutic management ment of cardiovascular sequelae in Kawasaki disease (JCS 2008)—
and medium-term prognosis of spontaneous coronary artery dissec- digest version. Circ J. 2010;74:1989–2020.
tion: results from a database of 11,605 patients. Eur J Cardiothorac 38.
Taubert KA, Rowley AH, Shulman ST.  Nationwide sur-
Surg. 2009;35:250–4. vey of Kawasaki disease and acute rheumatic fever. J Pediatr.
19. Maehara A, Mintz GS, Castagna MT, Pichard AD, Satler
1991;119:279–82.
LF, Waksman R, et  al. Intravascular ultrasound assessment 39. Dajani AS, Taubert KA, Takahashi M, Bierman FZ, Freed

of spontaneous coronary artery dissection. Am J Cardiol. M, Ferrieri P, et  al. Guidelines for long-term management of
2002;89:466–8. patients with Kawasaki disease. Report from the Committee on
20. Alfonso F, Paulo M, Gonzalo N, Dutary J, Jimenez-Quevedo P, Rheumatic Fever, Endocarditis, and Kawasaki Disease, Council on
Lennie V, et  al. Diagnosis of spontaneous coronary artery dis- Cardiovascular Disease in the Young, American Heart Association.
section by optical coherence tomography. J Am Coll Cardiol. Circulation. 1994;89:916–22.
2012;59:1073–9. 40. Uehara R, Belay ED, Maddox RA, Holman RC, Nakamura Y,
21. Tweet MS, Hayes SN, Pitta SR, Simari RD, Lerman A, Lennon RJ, Yashiro M, et al. Analysis of potential risk factors associated with
et al. Clinical features, management, and prognosis of spontaneous nonresponse to initial intravenous immunoglobulin treatment
coronary artery dissection. Circulation. 2012;126:579–88. among Kawasaki disease patients in Japan. Pediatr Infect Dis J.
22. Alfonso F, Paulo M, Lennie V, Dutary J, Bernardo E, Jiménez-­ 2008;27:155–60.
Quevedo P, et  al. Spontaneous coronary artery dissection: long-­ 41. Orenstein JM, Shulman ST, Fox LM, Baker SC, Takahashi M,
term follow-up of a large series of patients prospectively managed Bhatti TR, et al. Three linked vasculopathic processes characterize
with a “conservative” therapeutic strategy. JACC Cardiovasc lnterv. Kawasaki disease: a light and transmission electron microscopic
2012;5:1062–70. study. PLoS One. 2012;7(6):e38998.
23. Jennette JC, Falk RJ, Bacon PA, Basu N, Cid MC, Ferrario F, et al. 42. Newburger JW, Takahashi M, Burns JC. Kawasaki disease. J Am
2012 revised International Chapel Hill Consensus Conference Coll Cardiol. 2016;67:1738–49.
Nomenclature of Vasculitides. Arthritis Rheum. 2013;65:1–11. 43. Suzuki A, Kamiya T, Tsuda E, Tsukano S. Natural history of coro-
24. Kang EJ, Kim SM, Choe YH, Lee GY, Lee KN, Kim DK. Takayasu nary artery lesions in Kawasaki disease. Prog Pediatr Cardiol.
arteritis: assessment of coronary arterial abnormalities with 1997;6:211–8.
128-section dual-source CT angiography of the coronary arteries 44. Kamisawa T, Egawa N, Nakajima H.  Autoimmune pancre-

and aorta. Radiology. 2014;270:74–81. atitis is a systemic autoimmune disease. Am J Gastroenterol.
25. de Souza AW, de Carvalho JF. Diagnostic and classification criteria 2003;98:2811–2.
of Takayasu arteritis. J Autoimmun. 2014;48-49:79–83. 45. Kamisawa T, Funata N, Hayashi Y, Eishi Y, Koike M, Tsuruta K,
26. Subramanyan R, Joy J, Balakrishnan KG. Natural history of aorto- et al. A new clinicopathological entity of IgG4-related autoimmune
arteritis (Takayasu’s disease). Circulation. 1989;80:429–37. disease. J Gastroenterol. 2003;38:982–4.
27. Park YB, Hong SK, Choi KJ, Sohn DW, Oh BH, Lee MM, et al. 46. Kanno A, Nishimori I, Masamune A, Kikuta K, Hirota M, Kuriyama
Takayasu arteritis in Korea: clinical and angiographic features. S, et al. Research Committee on Intractable Diseases of Pancreas.
Heart Vessels Suppl. 1992;7:55–9. Nationwide epidemiological survey of autoimmune pancreatitis in
28. Lee GY, Jang SY, Ko SM, Kim EK, Lee SH, Han H, et  al.
Japan. Pancreas. 2012;41:835–9.
Cardiovascular manifestations of Takayasu arteritis and their rela- 47. Nishimori I, Tamakoshi A, Otsuki M, Research Committee on
tionship to the disease activity: analysis of 204 Korean patients at a Intractable Diseases of the Pancreas, Ministry of Health, Labour,
single center. Int J Cardiol. 2012;159:14–20. and Welfare of Japan. Prevalence of autoimmune pancreati-
29. Ishikawa K.  Diagnostic approach and proposed criteria for the tis in Japan from a nationwide survey in 2002. J Gastroenterol.
clinical diagnosis of Takayasu’s arteriopathy. J Am Coll Cardiol. 2007;42(Suppl 18):6–8.
1988;12:964–72. 48. Kamisawa T, Zen Y, Pillai S, Stone JH.  IgG4-related disease.

30. Sharma BK, Jain S, Suri S, Numano F.  Diagnostic criteria for Lancet. 2015;385:1460–71.
Takayasu arteritis. Int J Cardiol. 1996;54(Suppl):141–7. 49. Inoue D, Zen Y, Abo H, Gabata T, Demachi H, Yoshikawa J, et al.
31. Amano J, Suzuki A.  Coronary artery involvement in Takayasu’s Immunoglobulin G4-related periaortitis and periarteritis: CT find-
arteritis. Collective review and guideline for surgical treatment. J ings in 17 patients. Radiology. 2011;261:625–33.
Thorac Cardiovasc Surg. 1991;102:554–60. 50. Matsumoto Y, Kasashima S, Kawashima A, Sasaki H, Endo M,
32. Endo M, Tomizawa Y, Nishida H, Aomi S, Nakazawa M, Tsurumi Kawakami K, et al. A case of multiple immunoglobulin G4-related
Y, et al. Angiographic findings and surgical treatments of coronary periarteritis: a tumorous lesion of the coronary artery and abdomi-
artery involvement in Takayasu arteritis. J Thorac Cardiovasc Surg. nal aortic aneurysm. Hum Pathol. 2008;39:975–80.
2003;125:570–7. 51. Ishizaka N, Sakamoto A, Imai Y, Terasaki F, Nagai R. Multifocal
33. Matsubara O, Kuwata T, Nemoto T, Kasuga T, Numano F. Coronary fibrosclerosis and IgG4-related disease involving the cardiovascu-
artery lesions in Takayasu arteritis: pathological considerations. lar system. J Cardiol. 2012;59:132–8.
Heart Vessels Suppl. 1992;7:26–31. 52. Hourai R, Miyamura M, Tasaki R, Iwata A, Takeda Y, Morita
34. Kawasaki T.  Acute febrile mucocutaneous syndrome with lym- H, et  al. A case of IgG4-related lymphadenopathy, pericarditis,
phoid involvement with specific desquamation of the fingers and coronary artery periarteritis and luminal stenosis. Heart Vessel.
toes in children. Allergy. 1967;16:178–222. [in Japanese]. 2016;31:1709–13.
35. Kawasaki T, Kosaki F, Okawa S, Shigematsu I, Yanagawa H.  A 53. Yamamoto M, Takahashi H, Tabeya T, Suzuki C, Naishiro Y,

new infantile acute febrile mucocutaneous lymph node syndrome Ishigami K, et  al. Risk of malignancies in IgG4-related disease.
(MLNS) prevailing in Japan. Pediatrics. 1974;54:271–6. Mod Rheumatol. 2012;22:414–8.
36. Newburger JW, Takahashi M, Gerber MA, Gewitz MH, Tani LY, 54. Ishizaka N.  IgG4-related disease underlying the pathogenesis of
Burns JC, et al. Diagnosis, treatment, and long-term management coronary artery disease. Clin Chim Acta. 2013;415:220–5.
of Kawasaki disease: a statement for health professionals from 55. Epstein MA, Achong BG, Barr YM. Virus particles in cultured lym-
the Committee on Rheumatic Fever, Endocarditis and Kawasaki phoblasts from Burkitt’s lymphoma. Lancet. 1964;1:702–3.
Disease, Council on Cardiovascular Disease in the Young, American 56.
Cohen JI.  Epstein-Barr virus infection. N Engl J Med.
Heart Association. Circulation. 2004;110:2747–71. 2000;343:481–92.
25  Nonatherosclerotic Coronary Artery Disease 307

57. Thorley-Lawson DA, Gross A.  Persistence of the Epstein-Barr

62. Okano M, Kawa K, Kimura H, Yachie A, Wakiguchi H, Maeda A,
virus and the origins of associated lymphomas. N Engl J Med. et  al. Proposed guidelines for diagnosing chronic active Epstein-­
2004;350:1328–37. Barr virus infection. Am J Hematol. 2005;80:64–9.
5 8.
Babcock GJ, Decker LL, Volk M, Thorley-Lawson 63. Kikuta H, Taguchi Y, Tomizawa K, Kojima K, Kawamura N,

DA.  EBV persistence in memory B cells in  vivo. Immunity. Ishizaka A, et  al. Epstein-Barr virus genome-positive T lympho-
1998;9:395–404. cytes in a boy with chronic active EBV infection associated with
59. Cohen JI.  Epstein-Barr virus lymphoproliferative disease asso-
Kawasaki-like disease. Nature. 1988;333:455–7.
ciated with acquired immunodeficiency. Medicine (Baltimore). 64. Sato Y, Tsuboi T, Mikami T, Kurosawa H, Kanou K, Sugita K, et al.
1991;70:137–60. Chronic active Epstein-Barr virus infection with dilatation of the
60. Kimura H. Pathogenesis of chronic active Epstein-Barr virus infec- Valsalva sinus. Pediatr Int. 2006;48:643–5.
tion: is this an infectious disease, lymphoproliferative disorder, or 65. Muneuchi J, Ohga S, Ishimura M, Ikeda K, Yamaguchi K, Nomura
immunodeficiency? Rev Med Virol. 2006;16:251–61. A, et al. Cardiovascular complications associated with chronic active
61. Virelizier JL, Lenoir G, Griscelli C. Persistent Epstein-Barr virus Epstein-Barr virus infection. Pediatr Cardiol. 2009;30:274–81.
infection in a child with hypergammaglobulinaemia and immuno- 66. Kimura H, Morishima T, Kanegane H, Ohga S, Hoshino Y, Maeda
blastic proliferation associated with a selective defect in immune A, et  al. Prognostic factors for chronic active Epstein-Barr virus
interferon secretion. Lancet. 1978;2:231–4. infection. J Infect Dis. 2003;187:527–33.
 The Many Faces of Atherosclerosis
26
Nandini (Nina) M. Meyersohn, Jan-Erik Scholtz,
and Brian B. Ghoshhajra

Coronary CT angiography (CTA) has been well established degree of calcification of a plaque is not linked directly to the
as a noninvasive imaging modality for the assessment of degree of stenosis.
coronary artery disease (CAD) and the evaluation of athero- Heavily calcified plaque can result in minimal or no ste-
sclerotic plaque within the coronary arteries, as reviewed nosis, while noncalcified plaque can be obstructive or
earlier in this volume. While the pathology and pathophysi- occlusive. Coronary artery calcium (CAC) is an indepen-
ology of atherosclerosis leading to the development of dent predictor of adverse events, however it remains
plaques within the coronary arteries have been covered in unclear whether progression of CAC represents true pro-
earlier chapters, familiarity with the various imaging appear- gression of coronary artery disease or rather stabilization
ances of coronary artery plaque and their clinical signifi- and healing of existing plaque [3]. From a technical per-
cance is essential for the cardiac imager. In this chapter, the spective, the presence of calcification complicates inter-
CT imaging features of the range of presentations of athero- pretation of CTA due to what is known as “blooming
sclerotic plaque within the coronary arteries will be reviewed artifact,” which is the tendency of calcified objects to
and illustrated. appear larger than their true size on CT images because of
a complex interaction of volume averaging and beam hard-
ening [4]. Due to the contribution of beam hardening, this
Plaque Calcification artifact is accentuated at lower tube potentials, meaning
that dose reduction strategies can often increase the degree
As described in earlier chapters, coronary artery calcification of blooming artifact. Care must be taken by the imager to
is part of the continuum of development of atherosclerosis account for blooming artifact when evaluating the degree
[1]. Atherosclerotic plaques may be noncalcified, partially of luminal stenosis secondary to calcified plaques; exam-
calcified, or predominantly/fully calcified. Of note, noncalci- ining the degree of remaining contrast-­opacified lumen
fied and partially calcified plaques have been previously that is visible can help resolve this issue, as can newer
referred to as “soft” and “mixed” plaque, respectively; how- technologies such as dual-energy scanning and volume
ever these terms have fallen out of favor and are no longer calcium subtraction.
used in major society guidelines [2]. Calcification is never Figure 26.1a depicts a noncalcified plaque in the distal
normal within the wall of a coronary artery, however the right coronary artery (RCA), which not only demonstrates
lack of calcification but also notably low attenuation values
as measured by Hounsfield units (HU), a high-risk feature
which will be discussed below. Figure 26.1b depicts a par-
N. (N). M. Meyersohn (*) · B. B. Ghoshhajra tially calcified plaque in the left anterior descending coro-
Division of Cardiovascular Imaging, Department of Radiology, nary artery (LAD). Partially calcified plaques can lead to
Massachusetts General Hospital/Harvard Medical School,
diagnostic uncertainty when the density of calcification has
Boston, MA, USA
a similar attenuation of contrast within the coronary artery
Cardiac MR PET CT Program, Massachusetts General Hospital,
lumen. Figure 26.1c demonstrates a fully calcified plaque
Boston, MA, USA
e-mail: [email protected] in the left main coronary artery (LM). Note that in this case,
the density of the calcium in this plaque is of significantly
J.-E. Scholtz
Cardiac MR PET CT Program, Massachusetts General Hospital, higher attenuation than the contrast within the vessel
Boston, MA, USA lumen.

© Humana Press 2019 309

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_26
310 N. (N). M. Meyersohn et al.

Fig. 26.1  Coronary CT

angiography images
demonstrating long-axis and
short-axis (inset) views of
noncalcified plaque with low
attenuation (<30 HU)
resulting in mild stenosis of
the distal RCA (left, white
arrow), partially calcified
plaque and positive
remodeling resulting in mild
stenosis of the distal LAD
(middle, light gray arrow),
and calcified plaque in the
wall of the LM resulting in
mild stenosis (right, dark gray
arrow)

Coronary Artery Stenosis 5 for complete occlusion. The CAD-RADS system also
includes management recommendations to facilitate clinical
As coronary CTA evolved from an area of research into a decision-making authored in collaboration with the American
growing clinical subspecialty, the reporting of severity of College of Radiology and the North American Society of
stenosis resulting from atherosclerotic plaque has also Cardiovascular Imaging and endorsed by the American
evolved. Reports early on tended to be qualitative in nature College of Cardiology. Separate recommendations exist for
and were not necessarily consistent between readers regard- patients with acute chest pain and for stable outpatients with
ing overall burden of atherosclerosis as well as hemody- chest pain and are meant to integrate evolving clinical expe-
namic significance of each individual plaque. In 2009, Raff rience with the natural history of stenosis found on coronary
et  al. from the Society of Cardiovascular Computed CTA.  Figures  26.2, 26.3, and 26.4 demonstrate atheroscle-
Tomography (SCCT) issued society guidelines on the rotic plaques on coronary CTA resulting in mild [CAD-­
reporting of coronary CTA that were adopted in clinical RADS 2), moderate (CAD-RADS 3), and severe stenosis
practice at many institutions seeking to standardize interpre- (CAD-RADS 4), respectively, according to these
tation [5]. Under this system, if atherosclerotic plaque of guidelines.
any composition was identified in a coronary artery, the ste- Atherosclerosis in the coronary arteries is also the most
nosis could be graded as minimal (negligible impact on the common cause of coronary artery aneurysm in adults,
lumen), mild (less than 49% stenosis), moderate (50–69% defined as abnormal focal dilatation of a vessel with a diam-
stenosis), severe (70–99% stenosis), or occluded. There has eter exceeding the diameter of adjacent unaffected coronary
been some overlap in the literature with use of the terms segment by 1.5 times or greater [7]. As in other vessels, an
“significant” plaque and “obstructive” plaque; one method aneurysm is defined as saccular if the transverse diameter
of resolving this overlap is to reserve the use of “obstruc- exceeds the length of the aneurysm or fusiform if the con-
tive” for stenoses graded as severe and to use the more gen- verse is true. Figure 26.5 demonstrates the CT appearance of
eral term “significant” for stenosis graded as moderate or a saccular atherosclerotic coronary artery aneurysm.
higher, as these are of potential but not certain hemody- A specific entity that should be carefully evaluated for on
namic significance. coronary CTA is stenosis of the left main coronary artery.
A more recent multisociety guideline released in 2016 is Left main stenosis is important to identify as it is associated
the CAD-RADS™, or Coronary Artery Disease Reporting with a high rate of complications during catheterization, and
and Data System [6]. This system addresses the standardized patients are often treated instead with coronary artery bypass
reporting of degree of coronary artery stenosis in a fashion grafting [8]. Evaluating left main stenosis on coronary CTA
similar to that described in the 2009 guidelines but with the mandates the use of multiplanar reformatted images, and it is
addition of a numerical score ranging from 0 for no plaque to important for the imager to note that in both classification
26  The Many Faces of Atherosclerosis 311

Fig. 26.2  Coronary CT

angiography images
demonstrate partially calcified
plaque in the wall of the
proximal RCA resulting in
mild (CAD-RADS 2) stenosis
(left, yellow arrow). Invasive
coronary angiography
correlation demonstrating
mild stenosis (right, yellow
arrow). Comparison of the
diameters at the level of mild
stenosis (left inset, left panel)
and distal to the stenosis
(left inset, right panel)

Fig. 26.3  Coronary CT

angiography demonstrates
partially calcified plaque
resulting in moderate luminal
narrowing (CAD-RADS 3) at
its proximal and mid segment
(left and left inset, orange
arrows). Coronary
angiography confirms diffuse
proximal-to-mid vessel
calcification in the LAD
resulting in focal moderate
stenosis (right, orange arrow).
Angiography shows
additional severe stenosis in
the mid LAD (red arrows)

systems described above, stenosis of >50% in the left main as opposed to for a single plaque, two potential commonly
coronary artery constitutes a severe (CAD-RADS 4B) lesion. used clinical scores are the segment stenosis score and the
Figure 26.6 demonstrates the CT appearance of a severe left ­segment involvement score [9]. The segment involvement
main stenosis. score can be calculated as the total number of coronary
When describing the overall extent of coronary artery artery segments demonstrating any degree of plaque, with
disease burden throughout the entire coronary artery tree a minimum of 0 and a maximum of 16 when considering
312 N. (N). M. Meyersohn et al.

Fig. 26.4  Coronary CT

angiography demonstrates
partially calcified plaque in
the proximal LAD with severe
(CAD-RADS 4) luminal
narrowing in the long (left
top, red arrow) and short axis
(left bottom). Coronary
angiography confirms 80%
stenosis in the proximal LAD
(left, red arrow)

Fig. 26.5 Atherosclerotic
aneurysm within the distal
RCA in a region of partially
calcified plaque (top, red
arrow). The diameter within
the aneurysm (bottom right)
measures >1.5 times that of
the vessel proximal to the
aneurysm (bottom left, white
arrow in top image)

the entire ­coronary tree. Alternatively, the segment steno- RCA (each 1 point) as well as moderate stenosis in the
sis score grades each coronary segment as demonstrating proximal LAD, proximal LCx, and distal RCA (each 2
no to severe plaque (i.e., scores from 0 to 3), which are points), the segment stenosis score would be
then summed to yield a total score ranging from 0 to 48. 1+1+1+1+2+2+2 for a total of 10. The segment involve-
Figure  26.7 demonstrates multi-segment plaque involve- ment score would be 7, as 7 total coronary artery segments
ment of all three coronary arteries. Given that there is mild demonstrate plaque.
stenosis in the left main, mid LAD, and proximal and mid
26  The Many Faces of Atherosclerosis 313

Chronic Total Occlusion

On the spectrum of coronary artery stenosis secondary to

atherosclerotic plaque, the most extreme manifestation is
chronic total occlusion (CTO). This entity is defined as
complete occlusion of a coronary artery for greater than
3  months [10] and in some situations can be clinically
silent for an extended period of time before coming to
attention after an acute event in another vessel. In recent
years, techniques have been developed to attempt percuta-
neous coronary intervention (PCI) for revascularization of
CTOs, although they remain quite technically challenging.
Factors for which the imager should evaluate that are cor-
related with a lower chance of technical success during
revascularization include a blunt stump, lesion c­ alcification,
a CTO that is in a vessel other than the left anterior
descending coronary artery, and lesion length ≥20  mm
[11]. Detailed reporting of these factors may assist the
referring interventional physician in understanding the
likelihood of technical success and determining appropri-
ate candidates. Figure 26.8 demonstrates the CTA appear-
Fig. 26.6  Severe (>50%) stenosis of the LM due to a noncalcified
plaque viewed in long axis (left, red arrow). Short-axis images demon- ance of a chronic total occlusion of the left anterior
strate the vessel appearance at the level of the stenosis (top right) and descending coronary artery.
distal to the stenosis (bottom right, white arrow in long-axis image)

Fig. 26.7 Multi-segment
plaque involvement of all
three coronary arteries (LM
and LAD, left; LCx, middle;
RCA, right). Curved planar
reformats demonstrate mild
luminal narrowing (yellow
arrows) in the LM, mid LAD,
and proximal and mid
RCA. Moderate luminal
stenoses (orange arrows) are
in the proximal LAD,
proximal LCx, and distal
RCA
314 N. (N). M. Meyersohn et al.

High-Risk Plaque Features or greater [17, 18]. The “­ napkin-­ring” sign is defined as a non-
calcified portion of coronary artery plaque with a higher atten-
Multiple coronary artery plaque features have been identified uation peripheral ring and a lower attenuation core [19–21].
that are independently linked with adverse outcomes. These The “napkin-ring” sign is thought to correspond by histology
have been termed high-risk plaque (HRP) features and are to a thin-cap fibroatheroma which may be vulnerable to rup-
important to recognize clinically. The four most commonly ture. Figures 26.9, 26.10, 26.11, and 26.12 demonstrate the CT
referenced are low-attenuation plaque, spotty calcium, posi- appearance of these four high-risk plaque features.
tive remodeling, and the “napkin-ring” sign. Low-attenuation
plaque has been defined as attenuation less than 30 Hounsfield
units (HU) in three regions of interest (ROIs) within the non-
calcified portion of a plaque [12, 13]. Low-attenuation plaque
is thought to reflect lipid-rich plaque which is histologically
vulnerable to rupture. Spotty calcium has been defined as the
presence of calcified plaque with a diameter of less than 3 mm
in direction, length less than 1.5 times the vessel diameter, and
width less than two-thirds of the vessel diameter [14–16].
Positive remodeling is generally defined as a ratio of outer ves-
sel diameter in the region of plaque to uninvolved vessel of 1.1

Fig. 26.8  Curved planar reformat demonstrates chronic total occlu- Fig. 26.9  Coronary CT angiography images demonstrating mild ste-
sion of approximately 30 mm length (red arrows demarcate full length) nosis with positive remodeling (ratio of plaque to uninvolved vessel
of the mid RCA with following additional characteristics: tapered >1.1) due to a partially calcified plaque in the wall of the proximal LAD
stump and calcification greater than 50% (white arrow) with corresponding short-axis view (inset)

Fig. 26.10  Coronary CT

angiography images in long- and
short-axis views demonstrating
focal noncalcified plaque causing
moderate luminal narrowing in
the proximal LAD (left, white
arrow). This plaque exhibits
high-risk plaque features
including low attenuation, positive
remodeling, and napkin-ring sign
(right, light gray arrows)
26  The Many Faces of Atherosclerosis 315

Fig. 26.11  Curved planar reformat demonstrates spotty calcification

in the proximal LAD (white arrows)

Fig. 26.13  Coronary CT angiography curved planar reformat demon-

strating short-segment atherothrombotic occlusion of the RCA at its
ostium (red arrows)

High-risk plaque features increase the rate of plaque rup-

ture, which can be responsible for acute coronary syndrome
(ACS) secondary to atherothrombosis, which is the develop-
ment of coronary thrombosis superimposed on a ruptured
plaque. Atherothrombosis is a main cause of unstable angina
and acute myocardial infarction [22]. Figure 26.13 demon-
strates the CT appearance of an atherothrombotic occlusion
of a coronary artery.
Perivascular fat stranding has been identified as a sign of
plaque rupture, which is an important entity to identify in the
emergency setting. Figure  26.14 demonstrates the CT
appearance of perivascular stranding adjacent to an acute
plaque rupture [23].
After a percutaneous intervention has been performed and
a stent has been placed across a stenosis felt to be hemody-
namically significant, subsequent in-stent stenosis can
develop over time due to neointimal hyperplasia. Challenges
in imaging in-stent restenosis include beam hardening and
partial volume averaging artifacts, which can be mitigated in
part by the use of scanners with high spatial resolution and
iterative reconstruction [24]. Figure 26.15 demonstrates the
CT appearance of in-stent restenosis in a diagonal branch
coronary artery.
A final entity which should not be confused for athero-
sclerotic plaque is a myocardial bridge, where a segment of
coronary artery travels through the myocardium (Fig. 26.16).
Plaque within a bridge is unusual, although plaque can form
proximally, perhaps due to unusual shear stress at the entry
point to the myocardium.
Fig. 26.12  Coronary CT angiography images in long-axis (left)
and short-axis (right) views demonstrating noncalcified plaque with
low attenuation (<30 HU) in the wall of the distal RCA (white
arrows and ROI)
316 N. (N). M. Meyersohn et al.

Fig. 26.14  Coronary CT

angiography demonstrating
perivascular fat stranding
(white arrows) adjacent to an
acute plaque rupture within the
proximal RCA in axial view
(left) and curved planar
reformat (right)

Fig. 26.15  Curved planar reformat of the LAD with a partially calci-
fied plaque (white arrow) and a stent that spans approximately 40 mm
from the distal LM to the mid LAD (light gray arrows) with segments
of in-stent restenosis (red arrow)

Fig. 26.16  Coronary CT

angiography demonstrating
myocardial bridge in the mid
LAD with myocardium
overlying the vessel (top left,
white arrows). Evidence of
systolic compression (middle
left, gray arrow; right, white
arrow) and diastolic
decompression (bottom left,
black arrow)
26  The Many Faces of Atherosclerosis 317

Conclusion coronary arteries. A consensus document: part I.  Circulation.

2005;112(15):2364–72.
11. Alessandrino G, Chevalier B, Lefevre T, Sanguineti F, Garot P,
Atherosclerosis in the coronary arteries can have many faces Unterseeh T, et al. A clinical and angiographic scoring system to
and take on a wide variety of appearances on CT. Familiarity predict the probability of successful first-attempt percutaneous
with the characteristic types of plaque composition as well as coronary intervention in patients with total chronic coronary occlu-
sion. JACC Cardiovasc Interv. 2015;8(12):1540–8.
the degrees of stenosis in which they result is at the core of 12. Motoyama S, Sarai M, Harigaya H, Anno H, Inoue K, Hara T, et al.
cardiac CT imaging. Knowledge of the appearance of chronic Computed tomographic angiography characteristics of atheroscle-
total occlusions and the relationship with PCI, as well as of the rotic plaques subsequently resulting in acute coronary syndrome. J
role of specific plaque features in conferring an elevated risk Am Coll Cardiol. 2009;54(1):49–57.
13. Motoyama S, Kondo T, Sarai M, Sugiura A, Harigaya H, Sato T,
of adverse events, is also an essential component of a high- et  al. Multislice computed tomographic characteristics of coro-
quality practice of acquiring and interpreting CT of the heart. nary lesions in acute coronary syndromes. J Am Coll Cardiol.
2007;50(4):319–26.
14. van Velzen JE, de Graaf FR, de Graaf MA, Schuijf JD, Kroft LJ,
de Roos A, et  al. Comprehensive assessment of spotty calcifica-
References tions on computed tomography angiography: comparison to plaque
characteristics on intravascular ultrasound with radiofrequency
1. O’Rourke RA, Brundage BH, Froelicher VF, Greenland P, backscatter analysis. J Nucl Cardiol Off Publ Am Soc Nucl Cardiol.
Grundy SM, Hachamovitch R, et  al. American College of 2011;18(5):893–903.
Cardiology/American Heart Association Expert Consensus 15. Kitagawa T, Yamamoto H, Horiguchi J, Ohhashi N, Tadehara F,
document on electron-­beam computed tomography for the diag- Shokawa T, et al. Characterization of noncalcified coronary plaques
nosis and prognosis of coronary artery disease. Circulation. and identification of culprit lesions in patients with acute coronary
2000;102(1):126–40. syndrome by 64-slice computed tomography. J Am Coll Cardiol
2. Weigold WG, Abbara S, Achenbach S, Arbab-Zadeh A, Berman D, Img. 2009;2(2):153–60.
Carr JJ, et al. Standardized medical terminology for cardiac computed 16. Kashiwagi M, Tanaka A, Kitabata H, Tsujioka H, Kataiwa H,
tomography: a report of the Society of Cardiovascular Computed Komukai K, et al. Feasibility of noninvasive assessment of thin-cap
Tomography. J Cardiovasc Comput Tomogr. 2011;5(3):136–44. fibroatheroma by multidetector computed tomography. J Am Coll
3. Rodriguez-Granillo GA, Carrascosa P, Bruining N.  Progression Cardiol Img. 2009;2(12):1412–9.
of coronary artery calcification at the crossroads: sign of progres- 17. Gauss S, Achenbach S, Pflederer T, Schuhback A, Daniel WG,
sion or stabilization of coronary atherosclerosis? Cardiovasc Diagn Marwan M.  Assessment of coronary artery remodelling by dual-­
Therapy. 2016;6(3):250–8. source CT: a head-to-head comparison with intravascular ultra-
4. Andrew M, John H.  The challenge of coronary calcium on coro- sound. Heart. 2011;97(12):991–7.
nary computed tomographic angiography (CCTA) scans: effect on 18. Hoffmann U, Moselewski F, Nieman K, Jang IK, Ferencik
interpretation and possible solutions. Int J Cardiovasc Imaging. M, Rahman AM, et  al. Noninvasive assessment of plaque
2015;31(Suppl 2):145–57. morphology and composition in culprit and stable lesions in
5. Raff GL, Abidov A, Achenbach S, Berman DS, Boxt LM, Budoff acute coronary syndrome and stable lesions in stable angina
MJ, et  al. SCCT guidelines for the interpretation and reporting by multidetector computed tomography. J Am Coll Cardiol.
of coronary computed tomographic angiography. J Cardiovasc 2006;47(8):1655–62.
Comput Tomogr. 2009;3(2):122–36. 19. Ito T, Terashima M, Kaneda H, Nasu K, Matsuo H, Ehara M, et al.
6. Abbara S, Blanke P, Maroules CD, Cheezum M, Choi AD, Han BK, Comparison of in vivo assessment of vulnerable plaque by 64-slice
et al. SCCT guidelines for the performance and acquisition of coro- multislice computed tomography versus optical coherence tomog-
nary computed tomographic angiography: a report of the society raphy. Am J Cardiol. 2011;107(9):1270–7.
of Cardiovascular Computed Tomography Guidelines Committee: 20. Otsuka K, Fukuda S, Tanaka A, Nakanishi K, Taguchi H, Yoshikawa
endorsed by the North American Society for Cardiovascular Imaging J, et  al. Napkin-ring sign on coronary CT angiography for the
(NASCI). J Cardiovasc Comput Tomogr. 2016;10(6):435–49. prediction of acute coronary syndrome. J Am Coll Cardiol Img.
7. Diaz-Zamudio M, Bacilio-Perez U, Herrera-Zarza MC, Meave-­ 2013;6(4):448–57.
Gonzalez A, Alexanderson-Rosas E, Zambrana-Balta GF, et  al. 21. Maurovich-Horvat P, Hoffmann U, Vorpahl M, Nakano M, Virmani
Coronary artery aneurysms and ectasia: role of coronary CT R, Alkadhi H. The napkin-ring sign: CT signature of high-risk coro-
angiography. Radiographics Rev Publ Radiol Soc North Am Inc. nary plaques? J Am Coll Cardiol Img. 2010;3(4):440–4.
2009;29(7):1939–54. 22.
Fuster V, Moreno PR, Fayad ZA, Corti R, Badimon
8. Virani SS, Mendoza CE, Ferreira AC, de Marchena E. Left main JJ.  Atherothrombosis and high-risk plaque: part I: evolving con-
coronary artery stenosis: factors predicting cardiac events in patients cepts. J Am Coll Cardiol. 2005;46(6):937–54.
awaiting coronary surgery. Tex Heart Inst J. 2006;33(1):23–6. 23. Laine P, Kaartinen M, Penttila A, Panula P, Paavonen T, Kovanen
9. Min JK, Shaw LJ, Devereux RB, Okin PM, Weinsaft JW, Russo DJ, PT. Association between myocardial infarction and the mast cells
et al. Prognostic value of multidetector coronary computed tomo- in the adventitia of the infarct-related coronary artery. Circulation.
graphic angiography for prediction of all-cause mortality. J Am 1999;99(3):361–9.
Coll Cardiol. 2007;50(12):1161–70. 24. Andreini D, Pontone G, Mushtaq S, Bartorelli AL, Bertella E,
10. Stone GW, Kandzari DE, Mehran R, Colombo A, Schwartz RS, Trabattoni D, et al. Coronary in-stent restenosis: assessment with
Bailey S, et al. Percutaneous recanalization of chronically occluded CT coronary angiography. Radiology. 2012;265(2):410–7.
 Chronic Chest Pain
27
Richard A. P. Takx and Csilla Celeng

Chronic chest pain is a multifactorial clinical problem, and population is less frequent due to a high threshold for pain.
the diagnosis is based on patients’ history. It is characterized Diabetes and hypertension might be accompanied by renal
by recurrent episodes of chest pain occurring in a relatively dysfunction, which has a negative impact on the prognosis of
stable pattern. About four million cardiac stress tests (in 87% patients with stable angina. In patients with chronic kidney
of cases combined with imaging) are performed annually in disease (CKD), symptoms of chest pain are often misleading
the United States [1] for chest pain evaluation. There is con- due to diabetic and uremic neuropathy [8]. Besides CAD,
cern regarding the rising healthcare costs, inappropriate uti- coronary artery anomalies can also lead to the development
lization, and patient safety of diagnostic testing, with of chronic chest pain. Myocardial bridging is one of the most
approximately one-third of cardiac stress tests being likely common coronary anomalies. The reported prevalence of
inappropriately requested [1]. In addition, around 25% of myocardial bridging depends on the diagnostic method used.
invasive coronary angiography is performed among asymp- Autopsy studies report a wide prevalence of 5–86% [9], cor-
tomatic patients [2]. onary angiography studies 0.4–15.8% [10], and in CT stud-
The primary diagnostic consideration in patients with ies numbers range between 4% and 58% [11].
chronic chest pain is coronary artery disease (CAD). The
prevalence of stable angina increases with age for both sexes,
from 5–7% at 45–64  years to 10–12% at 65–84  years in Pathophysiology of Stable Chest Pain
females and from 4–7% at 45–64  years to 12–14% at
65–84 in males [3]. Of note is that angina is more prevalent Stable CAD is typically characterized by recurrent episodes
in middle-aged females than in males, most likely due to of myocardial ischemia due to a mismatch between oxygen
microvascular disease, while at older age angina is more demand and supply, which can be induced by exercise, emo-
prevalent in males [4, 5]. There is no accurate epidemiologi- tion, or other forms of stress, but it may also occur sponta-
cal data on the frequency of microvascular angina and vaso- neously. These ischemic/hypoxic episodes are commonly
spastic angina. However, recent clinical data suggest that associated with transient chest discomfort (i.e. angina pec-
coronary microvascular spasm is present in two-third of toris). The occurrence of angina can be attributable to vari-
patients with normal angiograms [6]. Conventional cardio- ous pathological conditions including (1) plaque-related
vascular risk factors such hypertension, hyperlipidemia, dia- coronary artery stenosis, (2) focal or diffuse spasm of the
betes, physical inactivity, obesity, smoking and family coronary arteries, (3) microvascular disease, (4) ischemic
history facilitate the progression of CAD. Patients with dia- cardiomyopathy/left ventricular dysfunction, and (5) coro-
betes mellitus have a similar risk for myocardial infarction as nary anomalies. These mechanisms may act separate but can
patients without diabetes with prior myocardial infarction also overlap in the same patient, and some may change over
[7]. Nevertheless, the occurrence of chest pain in this patient time [4]. The clinical manifestation of angina may be spe-
cific for the underlying pathology. Effort-induced angina is
associated with coronary artery stenosis and/or local vaso-
constriction at the site of the stenosis and/or microvascular
R. A. P. Takx (*)
Department of Radiology, University Medical Center Utrecht, dysfunction, while rest angina is related to focal or diffuse
Utrecht, The Netherlands vasospasm of the epicardial vessels or the microvasculature.
e-mail: [email protected] Hypoxia-induced metabolic changes initiate the ischemic
C. Celeng cascade, and the consequences occur in a predefined
Department of Radiology, Heart and Vascular Center, Semmelweis ­temporal manner (Fig. 27.1):
University, Budapest, Hungary

© Humana Press 2019 319

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_27
320 R. A. P. Takx and C. Celeng

Fig. 27.1  The ischemic

cascade; perfusion
abnormality caused by Normal function
reduced coronary blood flow
initiates the ischemic cascade,
and the consequences occur in
a predefined temporal manner

Perfusion abnormality
Progression of myocardial ischemia
Metabolic alterations

Diastolic dysfunction

Regional wall motion

abnormalities

Global wall motion

abnormalities

Systolic dysfunction

ST-T changes

Exposure time

1. During ischemia, the anaerobic conditions prevail, which 3. These mechanisms result in the development of ST-T
results in a decrease of cell pH. To balance the accumula- ­segment changes on the ECG.
tion of hydrogen ions, the Na+/H+ exchanger excretes 4. Finally myocardial ischemia manifests in chest pain [13].
excess hydrogen ions, which entails a large influx of Other “angina equivalent” symptoms such as dyspnea,
sodium ions. Hypoxia also depletes the cellular concen- fatigue, palpitations, or syncope can be also present in
tration of ATP, which leads to the decreased function of addition to or instead of angina.
Na+/K+-ATPase pump and K-ATP channels. These mech-
anisms lead to accumulation of extracellular K+ and
membrane depolarization, which inactivates the Na+ Histological Characteristics
channels. Decreased action potential upstroke leads to of Atherosclerotic Plaques
reduced conduction velocity, which can result in the
development of arrhythmias. Furthermore, ATP pump Previous investigations revealed that atherosclerotic plaques
dysfunction reduces the active efflux of Ca2+ and its reup- in patients with stable chest pain show different histological
take to the endoplasmic reticulum, thus causing intracel- characteristics than those with acute coronary syndrome:
lular Ca2+ overload, which induces Ca2+-dependent stable plaques are associated with smaller necrotic core and
apoptotic cascade and tissue damage [12]. a thicker cap (>84 μm) [14] made up of smooth muscle cells
2. Perfusion abnormalities lead to metabolic changes and and contain less macrophages [15–17]. In contrast plaques
abnormal myocardial performance that involves diastolic that are prone to rupture usually have a large necrotic core
dysfunction (slowed ventricular relaxation) and systolic with an overlying thin fibrous cap (between 55  μm and
dysfunction with regional and later global wall motion 84 μm), which is infiltrated by macrophages [14]. The pres-
abnormalities (reduced ejection fraction). ence of calcification as a marker of plaque stability is still
27  Chronic Chest Pain 321

unclear. Unstable plaques detected in patients in their fourth have higher plaque volume and area, higher remodeling indi-
decade tend to contain more calcification than stable plaques; ces and contain a greater proportion of non-calcified compo-
however, with increasing age this difference disappears and nent representing the necrotic core (Fig. 27.2) [19, 20].
more calcification is observed in stable plaques compared to
unstable plaques [18]. Coronary computed tomography
angiography (CTA) provides a noninvasive approach for the  ocal or Diffuse Spasm
F
detection of coronary atherosclerotic lesions. Beyond the of the Coronary Arteries
visualization of the coronary artery lumen, the submillimeter
spatial resolution of CTA allows for detailed characterization Vasospastic angina (often called as variant or Prinzmetal’s
as well as quantification of coronary plaques. Similar to his- angina) is determined as a focal and/or diffuse narrowing of
tology, stable and unstable plaques on CTA show different normal or diseased epicardial artery that can result in the
characteristics: stable plaques have smaller plaque volume reduction of coronary blood flow [21]. Focal spasm is typi-
and area, lower remodeling indices, and lower proportion of cally characterized by ST-segment elevation, while diffuse
non-calcified material [19, 20]. In contrast unstable lesions vasoconstriction is usually followed by ST depression.

a b c

d e

Fig. 27.2  Curved multiplanar reconstruction (a) of the LM-LAD smaller amount of fibrous tissue (green). Cross-sectional images of
with the presence of non-calcified plaques. Horizontal lines repre- the second plaque without (d) and with (e) color overlay show a
sent two different plaque types. Cross-sectional images of the first stable plaque with a smaller necrotic core and more pronounced
plaque without (b) and with color overlay (c) represent an unstable fibrous tissue
plaque, with positive remodeling, large necrotic core (red), and
322 R. A. P. Takx and C. Celeng

Coronary spasm can result in myocardial ischemia, arrhyth- in the absence of provocation tests, for example, by per-
mias, and heart failure and may even lead to myocardial forming double-acquisition coronary CTA in the absence
infarction [22]. Spastic motion can be frequently found in and presence of intravenous infusion of nitrate [26].
the coronary microvasculature alone (microvascular
angina) and might suggest the presence of the underlying
microvascular disease [6]. The pathogenesis of abnormal Microvascular Dysfunction
coronary vasomotion is heterogeneous and involves several
contributing factors such as increased kinase activity, which Stable microvascular angina (often referred to as cardiac syn-
induces the phosphorylation of myosin light chain and con- drome X) is described as effort-induced chest pain that is
sequently the contraction of smooth muscle cells [23]; attributable to abnormal circulation in small coronary ­arteries
endothelial dysfunction that results in an altered paracrine (<500 μm) [27]. Similar to epicardial stenosis, microvascular
signaling as well as in enhanced release of vasoactive sub- dysfunction is also characterized by the reduction of coronary
stances such as endothelin [24]; hyperreactivity of the auto- blood flow; however, the pattern of myocardial ischemia is
nomic nervous system, in which imbalances in both substantially different. Flow-limiting stenosis in epicardial
sympathetic and parasympathetic neurotransmitters can arteries results in a homogenously distributed perfusion
induce coronary artery spasm; smoking that results in the defect and is accompanied by impairment of segmental ven-
exposure of toxic agents, which can cause damage of the tricular contraction. In contrast, microvascular dysfunction is
vascular system, thus increasing the risk of coronary spasm; characterized by a more scattered manner of perfusion defects
and a high prevalence of coronary vasospasm that was and preserved segmental ventricular contraction, due to the
observed in Asian people [25]. Also, in a recent European presence of normal myocardial cells. This phenomenon might
cohort, microvascular spasm was found in two-third of explain the arising difficulties when trying to identify the
patients with normal coronary arteriograms and was shown presence of microvascular ischemia by standard imaging
to occur more frequently in females [6]. Currently, vaso- methods (Fig. 27.3). Microvascular dysfunction can be asso-
spasm is demonstrated using invasive provocation tests ciated with various pathological mechanisms. Impaired bio-
(ergonovine or acetylcholine). In the future coronary CTA activity of the endothelium-dependent nitric oxide is the most
might become a useful tool for the diagnosis of vasospasm common cause of reduced vascular relaxation. Abnormal

a Microvasculature

Epicardial vessel Periarterioles Arterioles

Endo
Epi Myo
b

Fig. 27.3  Epicardial stenosis (a) and attributable gradual decreasing of leads to the development of more localized perfusion defects. Patchy
coronary blood flow within myocardial layers covering the epicardial-­ distribution of ischemic territories hampers the detection of abnormal
endocardial axis. Homogenous distribution of the perfusion defect contractile function due to the presence of normal contractile myocar-
results in detectable segmental impairment of contractile function. dial cells within the same myocardial segment. Epi: epicardium, myo:
Microvascular dysfunction (b) involves small myocardial areas and myocardium, endo: endocardium
27  Chronic Chest Pain 323

endothelial function in patients with higher cholesterol level hibernation, and scarring) in which the “smart heart” reduces
leads to reduced release of nitric oxide [28]. Similarly, the myocardial demand by a loss in contractile apparatus [34].
decreased level of estradiol in postmenopausal females is Myocardial stunning evolves after 15–20 min of reduced per-
associated with increased breakdown of nitric oxide [29]. In fusion and is most probably caused by the abnormal function
the Women’s Ischemia Syndrome Evaluation (WISE) study of the sarcoplasmic reticulum due to oxygen radicals [35].
during a 5.4-year follow-up period, a significant association After restoration of blood flow, impaired contractility can per-
was found between lower coronary flow reserve (due to sist up to several hours, but eventually it leads to full recovery.
microvascular dysfunction) and major adverse outcomes in More prolonged periods of ischemia lead to the development
females of whom the majority (81%) had either no CAD or of myocardial hibernation where besides metabolic alterations
non-obstructive CAD on invasive angiography [30]. Other several structural changes in the contractile material occur
conditions including hypertension, diabetes, and ventricular (e.g., loss of sarcoplasmic reticulum and t-tubules). The hiber-
hypertrophy can also contribute to the development of micro- nated myocardium represents the hallmarks of left ventricular
vascular dysfunction. Reduced vasodilator response of the remodeling; however, at early stages no scarring mechanism is
diseased small vessels supports the development of coronary present. In case of chronic hibernation, structural changes
steal phenomenon by normal microvessels, thus leading to such as diffuse fibrosis and expansion of extracellular matrix
myocardial ischemia. The spatial resolution of current CT stay fixed and eventually result in myocardial scarring. The
scanners (≈400 μm) limits the precise detection of microvas- ability of delayed contrast-enhanced CTA to differentiate via-
cular disease; however, future technical advances in stress-­ ble and nonviable myocardium has been already established in
induced CT myocardial perfusion might serve as a noninvasive animal studies, where both acute and chronic infarction
imaging method for detecting microvascular disease [31]. appeared as hyperenhanced areas on CTA ≈5  min after the
contrast injection [36]. Since the method is accompanied by
only a slight increase in radiation dose in the future, its appli-
Ischemic Cardiomyopathy cation might be relevant in the comprehensive evaluation of
patients with advanced cardiovascular disease.
Chronic chest pain can be a clinical manifestation of ischemic
cardiomyopathy, which has been defined as left ventricular
systolic dysfunction with a history of prior myocardial infarc- Coronary Anomaly
tion or revascularization, or ≥75% stenosis of left main or
proximal LAD, or ≥75% stenosis of two or more epicardial Myocardial bridging may also be associated with exertional
vessels [32]. The development of chronic dysfunction is of angina. It is characterized by the presence of cardiac muscle
particular interest due to the observation that chronically overlying part of a coronary artery [37] and usually involves
hypoperfused myocardium (i.e., hibernating myocardium) the left descending artery [38] (Fig. 27.4). Ultrasound stud-
after revascularization can recover [33]. The pathophysiologi- ies showed that during early diastole, when coronary blood
cal background of this phenomenon is characterized by a spec- flow is highest, myocardial bridging results in delayed relax-
trum of adaptive mechanism (including myocardial stunning, ation, thus limiting coronary flow [39]. Coronary lesions

a b

Fig. 27.4  Myocardial bridge in a 53-year-old male with chronic chest Short-axis reconstruction (b) and cross section of the LAD (arrow) with
pain. Curved multiplanar reconstruction of the LM-LAD (a) with the an overlying myocardial tissue
presence of a myocardial bridge in the middle part of the LAD (arrows).
324 R. A. P. Takx and C. Celeng

a  iagnosing CAD in Patients

D
with Chronic Chest Pain

All patients suspected of chronic chest pain without an obvi-

ous noncardiac cause should undergo 12-lead resting ECG
[4, 41]. Though a normal resting ECG does not exclude the
presence of myocardial ischemia, it may demonstrate signs
of prior myocardial infarction and is useful as a baseline ref-
erence. Prior to diagnostic imaging, estimation of pre-test
probability may assist in the decision on appropriate imaging
technique and help to avoid unnecessary invasive procedures
[42]. The pre-test probability can be determined using risk
calculators such as ASCVD risk (https://tools.acc.org/
ASCVD-Risk-Estimator/) or SCORE (https://www.escardio.
org/Guidelines-&-Education/Practice-tools/CVD-
prevention-toolbox/SCORE-Risk-Charts). Patients with a
pre-test probability between 15 and 85% are recommended
to undergo noninvasive testing (Table  27.1) [4]. Exercise
ECG can be used as an initial test for diagnosing stable coro-
nary artery disease, with a good specificity of 85–90% and a
limited sensitivity of 45–50% [43, 44]. Prior to exercise ECG
patients should temporarily stop taking anti-ischemic medi-
cation and have a relative normal ECG at baseline.
b Nonetheless the test may be inconclusive in some patients
(usually when 85% of maximum heart rate is not reached)
[4]. Therefore if available (i.e. equipment and expertise),
myocardial perfusion imaging (MPI) or coronary CTA is
recommended [4]. MPI can be performed using single-­
photon emission computed tomography (SPECT),
­echocardiography (i.e., wall motion abnormality), magnetic
resonance imaging (MRI), positron emission tomography
(PET), and computed tomography (CT). In a meta-analysis
[45], the diagnostic accuracy of the various MPI techniques
was compared to invasive coronary angiography (ICA) with
fractional flow reserve (FFR) for the diagnosis of hemody-
namically significant CAD. PET, CT, and MRI demonstrated
superior diagnostic accuracy compared to SPECT and echo-
cardiography (Fig.  27.6). Considering that MRI does not
expose the patient to ionizing radiation and has good diag-
nostic performance [46], MRI could be regarded as the tech-
Fig. 27.5  Axial (a) and coronal (b) reconstructions of a patient with- nique of choice, with CT being an appropriate alternative in
out CAD but with an abnormal origin of the right coronary artery
those who have contraindications for undergoing MRI.
(arrow) with a malignant course (*) between the aorta (Ao) and the
pulmonary trunk (P) above the pulmonary valve (arrowheads) Adenosine stress-rest cardiac CT is able to identify stress-­
induced myocardial perfusion defects and as such is capable
usually evolve proximal to the bridge [40]. Anomalous ori- of reducing the number of false-positive CTA findings [47,
gin of a coronary artery arising from the opposite sinus can 48]. Using dual-source CT ventricular function and wall
be also a cause of chronic chest pain. A course between the motion abnormalities can be obtained by prospectively ECG-­
pulmonary trunk and aorta is considered malignant or inter- triggered (80% reduced mAs) dual-step pulsing [49] or full
arterial course (Fig. 27.5). The 3D nature of CTA allows for radiation using retrospective ECG-gated CTA.
precise characterization of the myocardial bridge and enables Coronary CTA is an increasingly popular alternative for
accurate visualization of the anomalous origin and course of ruling out CAD.  The examination traditionally consists of
the coronary artery. two diagnostic parts: first a non-contrast scan is made for
27  Chronic Chest Pain 325

Table 27.1  Pre-test probability of CAD in patients with chest pain split on sex

Non-cardiac chest pain Atypical chest pain Typical chest pain

Age (y) Female Male Female Male Female Male

30 - 39 Intermediate Low Intermediate Intermediate Intermediate

40 - 49 Intermediate Low Intermediate Intermediate High

50 - 59 Low Intermediate Intermediate Intermediate Intermediate High

60 - 69 Intermediate Intermediate Intermediate Intermediate Intermediate High

70 - 79 Intermediate Intermediate Intermediate High High Very high

>80 Intermediate Intermediate Intermediate High High Very high

Low: <15% pre-test probability no further testing needed

Intermediate: 15–65% pre-test probability additional testing is warranted, consider MPI
High: 66-85% pre-test probability additional testing is warranted, consider coronary CTA or MPI
Very high: >85% pre-test probability CAD is assumed to be present.
Pre-test probability based on data from Genders et al. [80]

Fig. 27.6  Diagram relating 1

pre- and posttest probability
SPECT
of hemodynamically
significant coronary artery 0.9 Echocardiography
disease for the various
techniques, with the upper left
MRI
half representing a positive
test and the lower right half a 0.8
PET
negative test. (From Takx
et al. [45], with permission) CT
0.7

0.6
Post-test probability

0.5

0.4

0.3

0.2

0.1
326 R. A. P. Takx and C. Celeng

measuring coronary artery calcification (CAC); second, such the role of CTA is expected to increase in therapy
iodine-containing contrast agent is administered intrave- guidance [61].
nously to visualize the coronary lumen, in combination with
nitroglycerin [50] and depending on heart rate and site policy
short-acting β-blockers. Examinations should be performed Prognostic Value of CTA
using modern hardware (≥64 slice).
CAC is defined as pixels above a threshold of 130 A meta-analysis on the prognostic value of coronary CTA
Hounsfield units (HU) and is traditionally quantified using demonstrated that 50% stenosis degree is a strong indepen-
the Agatston score [51], which is calculated by multiplying dent predictor of cardiovascular events [62]. In detail, the
the lesion area (mm2) by a density factor (between 1 and 4 presence of significant coronary stenosis was associated
based on the voxel with the highest density). However, with a sixfold increased hazard ratio (HR) for death, myo-
recent research suggests that density might be inversely cardial infarction, and unstable angina pectoris independent
related to cardiovascular events [52]. In addition, a CAC of CAC score. Also, in patients with diabetes, no CAD on
score of zero does not exclude the presence of a significant coronary CTA allows for safely ruling out future events, and
non-calcified stenosis [53]. Novel automated and semiauto- the detection of CAD is associated with a higher risk for
mated software algorithms enable the quantification of events [63]. In patients with CKD, a clear increase is seen in
CAC score using contrast-enhanced CTA series. Several annualized event rates (myocardial infarction, cerebrovas-
studies demonstrated good correlations (above 0.95) and cular accident, and death) when patients have both moderate
kappa values (above 0.87) for the quantification of CAC CKD and obstructive CAD on CTA (3.7% for both vs. 2.5%
score using ECG-gated contrast-enhanced CTA series com- for only moderate CKD or obstructive CAD) [64].
pared to standard ECG-gated unenhanced CT, suggesting Multivariate Cox models demonstrated that both moderate
that the use of quantification tools might allow for a lower CKD (HR 2.4, 95%CI 1.1–5.2, p  =  0.03) and obstructive
radiation dose in the future by omitting the non-contrast CAD (HR 2.8, 95%CI 1.4–5.4, p  <  0.01) are independent
CAC scan from the scan protocol [54, 55]. predictors of events [64]. Recently two large randomized
Some centers use high CAC scores (≥400 Agatston controlled trials investigated the effect of coronary CTA on
score) as a threshold for not performing coronary CTA; survival compared to standard-of-care treatment. The
however, modern scanners demonstrate good diagnostic PROspective Multicenter Imaging Study for Evaluation of
performance even in cases with high CAC scores [56]. A Chest Pain (PROMISE) trial randomized 10,003 symptom-
64-row or higher coronary CTA has an excellent sensitiv- atic patients with suspected CAD to an initial strategy of
ity of 95.6% and good specificity of 81.5%; nondiagnostic either coronary CTA or to functional testing during a median
scans occur in 2.5% of patients with CAD and 7.5% with- follow-up period of 25  months [65]. The PROMISE trial
out CAD [57]. A significant stenosis on coronary CTA showed no significant difference between the CTA strategy
requires verification to confirm the hemodynamic severity and functional testing in respect to cardiovascular events
of CAD (angiography optionally with FFR), nondiagnos- (HR 1.04, 95%CI 0.83–1.29, p  =  0.75). These results are
tic coronary CTA requires alternative imaging (e.g., MPI), partly influenced by the fact that CTA images were read at
and a negative CTA reliably excludes significant CAD. Of nonexpert centers and the limited statistical power (actual
note is that majority of CTA studies define obstructive event rate 3.0% vs. 8.0% estimated). However the trial
CAD as ≥50% stenosis degree, while ICA studies usually showed a reduction in the number of invasive angiograms
consider ≥70% stenosis degree. Lesion characterization with non-­obstructive disease in the CTA group vs. func-
by stenosis degree alone on ICA might not be sufficient tional testing group (3.4% vs. 4.3%, respectively, p = 0.02).
for intermediate stenosis (40–70%); therefore FFR mea- The Scottish COmputed Tomography of the HEART (SCOT-
surement is recommendable [58]. In addition, FFR values HEART) trial randomly assigned 4146 patients with stable
have a continuous and independent relationship with car- chest pain due to suspected coronary heart disease to stan-
diac events (i.e., the lower the FFR value, the more car- dard of care plus CTA and standard of care alone group [66].
diac events occur) [59]. Even so, there is an ongoing During a median follow-up of 1.7 years, the SCOT-HEART
debate on the true reference standard for the quantifica- trial showed a considerable, though nonsignificant reduc-
tion of myocardial ischemia, and this question might be tion of 38% in fatal and nonfatal myocardial infarction in
answered by results as obtained in the International Study standard of care plus CTA vs. standard of care alone group
of Comparative Health Effectiveness with Medical and (HR 0.62, 95%CI 0.38–1.01, p = 0.053). Combined, these
Invasive Approaches (ISCHEMIA) trial (ClinicalTrials. data demonstrate that coronary CTA is at least equivalent to
gov Identifier: NCT01471522). Furthermore advances conventional functional testing, and with growing experi-
in computational fluid dynamics allow for the modeling ence, it is expected that it will be applied more widely in
of FFR from coronary CTA datasets (FFRCT) [60]; as clinical practice.
27  Chronic Chest Pain 327

Sex Differences nary CTA [78]. Guidelines on secondary prevention could

be interpreted as supporting aggressive statin treatment in
Despite improvements in healthcare, hospitalization rates patients with obstructive CAD.  However, in case of non-­
for acute myocardial infarction have not declined in people obstructive CAD and high-risk plaque features, guidelines
younger than 55 years old, besides young females have are unclear, despite literature showing that extensive non-­
more comorbidity and higher in-hospital mortality than obstructive CAD (≥4 involved coronary segments) yields
young males [67]. Females commonly present with atypi- similar event rates as obstructive CAD [79]. Also, high-risk
cal chest pain (i.e. less exertional symptoms) [68]. In addi- plaque features (low-attenuation plaque and positive
tion, risk scores poorly categorize females for pre-test CAD remodeling) result in a significant higher risk over >50%
probability [69]. Traditionally, the goal of coronary CTA is stenosis degree [80]. The available data suggest that coro-
focused on diagnosing obstructive CAD, while in females nary CTA-guided therapy is expected to be incorporated in
non-­obstructive CAD is relatively common [30, 70]. Also, future guidelines.
positive remodeling, plaque erosion, and microvascular
embolization are more prevalent in females [71]. The
importance of non-obstructive plaque in females is stressed Conclusions
by the finding that the number of non-obstructive plaques is
predictive of mortality in females even after adjusting for In the past years, coronary CTA has become an integral part
the presence of obstructive CAD [72]. Results from the of strategies for evaluation of chronic chest pain. It allows
PROMISE trial found evidence that CTA yields more prog- for quick and reliable ruling out of the presence of CAD;
nostic information in females than stress testing compared furthermore it is capable to perform perfusion imaging and
to males [73], while the Coronary CT Angiography has the potential to guide and monitor therapy. In patients
Evaluation for Clinical Outcomes: An international multi- with diabetes mellitus and CKD, CAD on CTA remains an
center (CONFIRM) registry observed no difference in independent predictor of events. In females coronary CTA
prognostic value of CTA between both sexes with a similar also adds prognostic information, though the exact differ-
disease extent [74]. As such there is still equipoise consid- ences between CAD findings on CTA between the sexes still
ering the role of non-­obstructive CAD on coronary CTA in have to be elucidated. Advances in quantification algorithms
females. as well as modeling of FFR from coronary CTA datasets
have the potential to increase the role of CTA in decision-­
making and therapy guidance.
Therapy of Patients with Stable CAD

The Clinical Outcomes Utilizing Revascularization and References

Aggressive Drug Evaluation (COURAGE) trial [75] random-
ized 1149 patients with stable chest pain to percutaneous 1. Ladapo JA, Blecker S, Douglas PS.  Physician decision making
and trends in the use of cardiac stress testing in the United States:
coronary intervention (PCI) plus optimal medical therapy an analysis of repeated cross-sectional data. Ann Intern Med.
(OMT) and 1138 for OMT alone. The trial demonstrated that 2014;161:482–90.
during a median follow-up of 4.6 years, the addition of PCI 2. Bradley SM, Spertus JA, Kennedy KF, et  al. Patient selection
to OMT reduced the occurrence of angina, but it did not for diagnostic coronary angiography and hospital-level percu-
taneous coronary intervention appropriateness: insights from
reduce the risk of mortality, myocardial infarction, and hos- the National Cardiovascular Data Registry. JAMA Intern Med.
pitalization for acute coronary syndromes compared to OMT 2014;174:1630–9.
alone [75]. Coronary artery bypass graft (CABG) surgery 3. National Institutes of Health NHL, and Blood Institute. Morbidity
should be considered in patients with two- or three-vessel & mortality: 2012 chart book on cardiovascular, lung, and blood
diseases. Bethesda: National Heart, Lung, and Blood Institute;
disease (complex CAD), while in less complex cases and in 2012.
patients with high operative risk, PCI is recommended [76]. 4. Task Force M, Montalescot G, Sechtem U, et al. 2013 ESC guide-
In patients with diabetes and significant coronary disease lines on the management of stable coronary artery disease: the Task
(≥50% stenosis on coronary angiography with a positive Force on the management of stable coronary artery disease of the
European Society of Cardiology. Eur Heart J. 2013;34:2949–3003.
stress test or ≥70% stenosis with classic angina pectoris), 5. Crea F, Camici PG, Bairey Merz CN. Coronary microvascular dys-
CABG is preferred over OMT [77]. function: an update. Eur Heart J. 2014;35:1101–11.
CTA has the potential to improve the identification and 6. Ong P, Athanasiadis A, Borgulya G, Mahrholdt H, Kaski JC,
monitoring of high-risk patients who would benefit from Sechtem U. High prevalence of a pathological response to acetyl-
choline testing in patients with stable angina pectoris and unob-
lipid-lowering therapy. A recent study demonstrated that structed coronary arteries. The ACOVA Study (Abnormal COronary
the use of statins results in the reduction of non-calcified VAsomotion in patients with stable angina and unobstructed coro-
plaque volume and low-attenuation plaque volume on coro- nary arteries). J Am Coll Cardiol. 2012;59:655–62.
328 R. A. P. Takx and C. Celeng

7. Haffner SM, Lehto S, Ronnemaa T, Pyorala K, Laakso M. Mortality 29. Collins P, Rosano GM, Sarrel PM, et al. 17 beta-estradiol attenuates
from coronary heart disease in subjects with type 2 diabetes and in acetylcholine-induced coronary arterial constriction in women but
nondiabetic subjects with and without prior myocardial infarction. not men with coronary heart disease. Circulation. 1995;92:24–30.
N Engl J Med. 1998;339:229–34. 30. Pepine CJ, Anderson RD, Sharaf BL, et  al. Coronary microvas-
8. Cai Q, Mukku VK, Ahmad M. Coronary artery disease in patients cular reactivity to adenosine predicts adverse outcome in women
with chronic kidney disease: a clinical update. Curr Cardiol Rev. evaluated for suspected ischemia results from the National Heart,
2013;9:331–9. Lung and Blood Institute WISE (Women’s Ischemia Syndrome
9. Mohlenkamp S, Hort W, Ge J, Erbel R.  Update on myocardial Evaluation) study. J Am Coll Cardiol. 2010;55:2825–32.
bridging. Circulation. 2002;106:2616–22. 31. de Roos A. Myocardial perfusion imaging with multidetector CT:
10. Ishikawa Y, Kawawa Y, Kohda E, Shimada K, Ishii T. Significance beyond lumenography. Radiology. 2010;254:321–3.
of the anatomical properties of a myocardial bridge in coronary 32. Felker GM, Shaw LK, O’Connor CM.  A standardized definition
heart disease. Circ J. 2011;75:1559–66. of ischemic cardiomyopathy for use in clinical research. J Am Coll
11. Wirianta J, Mouden M, Ottervanger JP, et al. Prevalence and pre- Cardiol. 2002;39:210–8.
dictors of bridging of coronary arteries in a large Indonesian popu- 33. Rahimtoola SH.  A perspective on the three large multicenter

lation, as detected by 64-slice computed tomography scan. Neth ­randomized clinical trials of coronary bypass surgery for chronic
Heart J. 2012;20:396–401. stable angina. Circulation. 1985;72:V123–35.
12. Sanada S, Komuro I, Kitakaze M. Pathophysiology of myocardial 34. Rahimtoola SH.  The hibernating myocardium. Am Heart J.

reperfusion injury: preconditioning, postconditioning, and trans- 1989;117:211–21.
lational aspects of protective measures. Am J Physiol Heart Circ 35. Poole-Wilson PA, Holmberg SR, Williams AJ. A possible molec-
Physiol. 2011;301:H1723–41. ular mechanism for ‘stunning’ of the myocardium. Eur Heart J.
13. Detry JM.  The pathophysiology of myocardial ischaemia. Eur
1991;12(Suppl F):25–9.
Heart J. 1996;17(Suppl G):48–52. 36. Lardo AC, Cordeiro MA, Silva C, et  al. Contrast-enhanced mul-
14. Narula J, Nakano M, Virmani R, et al. Histopathologic character- tidetector computed tomography viability imaging after myocar-
istics of atherosclerotic coronary disease and implications of the dial infarction: characterization of myocyte death, microvascular
findings for the invasive and noninvasive detection of vulnerable obstruction, and chronic scar. Circulation. 2006;113:394–404.
plaques. J Am Coll Cardiol. 2013;61:1041–51. 37. De Giorgio F, Grassi VM, Polacco M, Pascali VL, d’Aloja E, Arena
15. Virmani R, Burke AP, Farb A, Kolodgie FD. Pathology of the vul- V. Myocardial bridging and sudden cardiac death: is the actual clas-
nerable plaque. J Am Coll Cardiol. 2006;47:C13–8. sification exhaustive? Int J Cardiol. 2014;172:e383–4.
16. Celeng C, Takx RA, Ferencik M, Maurovich-Horvat P.  Non-­
38. Loukas M, Curry B, Bowers M, et al. The relationship of myocar-
invasive and invasive imaging of vulnerable coronary plaque. dial bridges to coronary artery dominance in the adult human heart.
Trends Cardiovasc Med. 2016;26:538–47. J Anat. 2006;209:43–50.
17. Tarkin JM, Dweck MR, Evans NR, et al. Imaging atherosclerosis. 39. Yamada R, Tremmel JA, Tanaka S, et al. Functional versus anatomic
Circ Res. 2016;118:750–69. assessment of myocardial bridging by intravascular u­ltrasound:
18. Otsuka F, Finn AV, Virmani R. Do vulnerable and ruptured plaques impact of arterial compression on proximal atherosclerotic plaque.
hide in heavily calcified arteries? Atherosclerosis. 2013;229:34–7. J Am Heart Assoc. 2016;5:e001735.
19. Hoffmann U, Moselewski F, Nieman K, et al. Noninvasive assessment 40. Corban MT, Hung OY, Eshtehardi P, et  al. Myocardial bridging:
of plaque morphology and composition in culprit and stable lesions in contemporary understanding of pathophysiology with implica-
acute coronary syndrome and stable lesions in stable angina by multi- tions for diagnostic and therapeutic strategies. J Am Coll Cardiol.
detector computed tomography. J Am Coll Cardiol. 2006;47:1655–62. 2014;63:2346–55.
20. Pflederer T, Marwan M, Schepis T, et al. Characterization of culprit 41. Fihn SD, Gardin JM, Abrams J, et al. 2012 ACCF/AHA/ACP/AATS/
lesions in acute coronary syndromes using coronary dual-source PCNA/SCAI/STS Guideline for the diagnosis and management of
CT angiography. Atherosclerosis. 2010;211:437–44. patients with stable ischemic heart disease: a report of the American
21. Maseri A, Chierchia S.  Coronary artery spasm: demonstration, College of Cardiology Foundation/American Heart Association
definition, diagnosis, and consequences. Prog Cardiovasc Dis. Task Force on Practice Guidelines, and the American College of
1982;25:169–92. Physicians, American Association for Thoracic Surgery, Preventive
22. Hao PP, Shang R, Liu YP, et  al. Cardiogenic shock from acute Cardiovascular Nurses Association, Society for Cardiovascular
ST-segment elevation myocardial infarction induced by severe mul- Angiography and Interventions, and Society of Thoracic Surgeons.
tivessel coronary vasospasm. Eur Heart J. 2014;35:146. J Am Coll Cardiol. 2012;60:e44–e164.
23. Amano M, Ito M, Kimura K, et al. Phosphorylation and activation 42. Diamond GA, Forrester JS.  Analysis of probability as an aid in
of myosin by Rho-associated kinase (Rho-kinase). J Biol Chem. the clinical diagnosis of coronary-artery disease. N Engl J Med.
1996;271:20246–9. 1979;300:1350–8.
24. Toyo-oka T, Aizawa T, Suzuki N, et al. Increased plasma level of 43. Morise AP, Diamond GA. Comparison of the sensitivity and speci-
endothelin-1 and coronary spasm induction in patients with vaso- ficity of exercise electrocardiography in biased and unbiased popu-
spastic angina pectoris. Circulation. 1991;83:476–83. lations of men and women. Am Heart J. 1995;130:741–7.
25. Shimokawa H, Nagasawa K, Irie T, et  al. Clinical characteristics 44. Froelicher VF, Lehmann KG, Thomas R, et al. The electrocardio-
and long-term prognosis of patients with variant angina. A com- graphic exercise test in a population with reduced workup bias:
parative study between western and Japanese populations. Int J diagnostic performance, computerized interpretation, and multi-
Cardiol. 1988;18:331–49. variable prediction. Veterans Affairs Cooperative Study in Health
26. Kang EJ, Kim MH, De Jin C, et al. Noninvasive detection of coro- Services #016 (QUEXTA) Study Group. Quantitative exercise test-
nary vasospastic angina using a double-acquisition coronary CT ing and angiography. Ann Intern Med. 1998;128:965–74.
angiography protocol in the presence and absence of an intravenous 45. Takx RA, Blomberg BA, El Aidi H, et al. Diagnostic accuracy of
nitrate: a pilot study. Eur Radiol. 2016. https://doi.org/10.1007/ stress myocardial perfusion imaging compared to invasive coro-
s00330-016-4476-2. nary angiography with fractional flow reserve meta-analysis. Circ
27. Lanza GA. Cardiac syndrome X: a critical overview and future per- Cardiovasc Imaging. 2015;8(1). pii: e002666.
spectives. Heart. 2007;93:159–66. 46. Morton G, Chiribiri A, Ishida M, et al. Quantification of absolute
28. Casino PR, Kilcoyne CM, Quyyumi AA, Hoeg JM, Panza JA. The myocardial perfusion in patients with coronary artery disease: com-
role of nitric oxide in endothelium-dependent vasodilation of parison between cardiovascular magnetic resonance and positron
hypercholesterolemic patients. Circulation. 1993;88:2541–7. emission tomography. J Am Coll Cardiol. 2012;60:1546–55.
27  Chronic Chest Pain 329

47. Feuchtner G, Goetti R, Plass A, et al. Adenosine stress high-pitch 65. Douglas PS, Hoffmann U, Patel MR, et al. Outcomes of anatomical
128-slice dual-source myocardial computed tomography perfu- versus functional testing for coronary artery disease. N Engl J Med.
sion for imaging of reversible myocardial ischemia: compari- 2015;372:1291–300.
son with magnetic resonance imaging. Circ Cardiovasc Imaging. 66. investigators S-H. CT coronary angiography in patients with sus-
2011;4:540–9. pected angina due to coronary heart disease (SCOT-HEART):
48. De Cecco CN, Harris BS, Schoepf UJ, et al. Incremental value of an open-label, parallel-group, multicentre trial. Lancet.
pharmacological stress cardiac dual-energy CT over coronary CT 2015;385:2383–91.
angiography alone for the assessment of coronary artery disease in 67. Gupta A, Wang Y, Spertus JA, et  al. Trends in acute myocardial
a high-risk population. AJR Am J Roentgenol. 2014;203:W70–7. infarction in young patients and differences by sex and race, 2001
49. Takx RA, Moscariello A, Schoepf UJ, et al. Quantification of left to 2010. J Am Coll Cardiol. 2014;64:337–45.
and right ventricular function and myocardial mass: comparison 68. Bugiardini R.  Women, ‘non-specific’ chest pain, and normal or
of low-radiation dose 2nd generation dual-source CT and cardiac near-normal coronary angiograms are not synonymous with favour-
MRI. Eur J Radiol. 2012;81:e598–604. able outcome. Eur Heart J. 2006;27:1387–9.
50. Takx RA, Sucha D, Park J, Leiner T, Hoffmann U. Sublingual nitro- 69. Cheng VY, Berman DS, Rozanski A, et al. Performance of the tradi-
glycerin administration in coronary computed tomography angiog- tional age, sex, and angina typicality-based approach for estimating
raphy: a systematic review. Eur Radiol. 2015;25:3536–42. pretest probability of angiographically significant coronary artery
51. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M disease in patients undergoing coronary computed tomographic
Jr, Detrano R. Quantification of coronary artery calcium using ultra- angiography: results from the multinational coronary CT angiog-
fast computed tomography. J Am Coll Cardiol. 1990;15:827–32. raphy evaluation for clinical outcomes: an international multicenter
52. Criqui MH, Denenberg JO, Ix JH, et  al. Calcium density of cor- registry (CONFIRM). Circulation. 2011;124:2423–32. 2421-2428
onary artery plaque and risk of incident cardiovascular events. 70. Pepine CJ, Ferdinand KC, Shaw LJ, et  al. Emergence of nonob-
JAMA. 2014;311:271–8. structive coronary artery disease: a woman’s problem and need
53. Akram K, O’Donnell RE, King S, Superko HR, Agatston A, Voros for change in definition on angiography. J Am Coll Cardiol.
S.  Influence of symptomatic status on the prevalence of obstruc- 2015;66:1918–33.
tive coronary artery disease in patients with zero calcium score. 71. Bairey Merz CN, Shaw LJ, Reis SE, et al. Insights from the NHLBI-­
Atherosclerosis. 2009;203:533–7. sponsored Women’s Ischemia Syndrome Evaluation (WISE) study:
54. Ebersberger U, Eilot D, Goldenberg R, et al. Fully automated deri- Part II: gender differences in presentation, diagnosis, and outcome
vation of coronary artery calcium scores and cardiovascular risk with regard to gender-based pathophysiology of atherosclerosis
assessment from contrast medium-enhanced coronary CT angiog- and macrovascular and microvascular coronary disease. J Am Coll
raphy studies. Eur Radiol. 2013;23:650–7. Cardiol. 2006;47:S21–9.
55. Schuhbaeck A, Otaki Y, Achenbach S, et al. Coronary calcium scor- 72. Shaw LJ, Min JK, Narula J, et al. Sex differences in mortality asso-
ing from contrast coronary CT angiography using a semiautomated ciated with computed tomographic angiographic measurements of
standardized method. J Cardiovasc Comput Tomogr. 2015;9:446–53. obstructive and nonobstructive coronary artery disease: an explor-
56. Westwood ME, Raatz HD, Misso K, et al. Systematic review of the atory analysis. Circ Cardiovasc Imaging. 2010;3:473–81.
accuracy of dual-source cardiac CT for detection of arterial stenosis 73. Pagidipati NJ, Hemal K, Coles A, et  al. Sex differences in func-
in difficult to image patient groups. Radiology. 2013;267:387–95. tional and CT angiography testing in patients with suspected coro-
57. Menke J, Kowalski J. Diagnostic accuracy and utility of coronary nary artery disease. J Am Coll Cardiol. 2016;67:2607–16.
CT angiography with consideration of unevaluable results: a sys- 74. Schulman-Marcus J, ó Hartaigh B, Gransar H, et  al. Sex-specific
tematic review and multivariate Bayesian random-effects meta-­ associations between coronary artery plaque extent and risk of
analysis with intention to diagnose. Eur Radiol. 2016;26:451–8. major adverse cardiovascular events: the CONFIRM long-term
58. Tobis J, Azarbal B, Slavin L.  Assessment of intermediate sever- registry. JACC Cardiovasc Imaging. 2016;9:364–72.
ity coronary lesions in the catheterization laboratory. J Am Coll 75. Boden WE, O’Rourke RA, Teo KK, et al. Optimal medical therapy
Cardiol. 2007;49:839–48. with or without PCI for stable coronary disease. N Engl J Med.
59. Johnson NP, Toth GG, Lai D, et al. Prognostic value of fractional 2007;356:1503–16.
flow reserve: linking physiologic severity to clinical outcomes. J 76. Deb S, Wijeysundera HC, Ko DT, Tsubota H, Hill S, Fremes
Am Coll Cardiol. 2014;64:1641–54. SE. Coronary artery bypass graft surgery vs percutaneous interven-
60. Celeng C, Leiner T, Maurovich-Horvat P, et al. Anatomical and tions in coronary revascularization: a systematic review. JAMA.
functional computed tomography for diagnosing hemodynami- 2013;310:2086–95.
cally significant coronary artery disease. J Am Coll Cardiol Img. 77. Group BDS, Frye RL, August P, et al. A randomized trial of thera-
2018:S1936-878X(18)30681–8. pies for type 2 diabetes and coronary artery disease. N Engl J Med.
61. Taylor CA, Fonte TA, Min JK.  Computational fluid dynamics
2009;360:2503–15.
applied to cardiac computed tomography for noninvasive quantifi- 78. Zeb I, Li D, Nasir K, et al. Effect of statin treatment on coronary
cation of fractional flow reserve: scientific basis. J Am Coll Cardiol. plaque progression  – a serial coronary CT angiography study.
2013;61:2233–41. Atherosclerosis. 2013;231:198–204.
62. Bamberg F, Sommer WH, Hoffmann V, et al. Meta-analysis and sys- 79. Bittencourt MS, Hulten E, Ghoshhajra B, et al. Prognostic value of
tematic review of the long-term predictive value of assessment of nonobstructive and obstructive coronary artery disease detected by
coronary atherosclerosis by contrast-enhanced coronary computed coronary computed tomography angiography to identify cardiovas-
tomography angiography. J Am Coll Cardiol. 2011;57:2426–36. cular events. Circ Cardiovasc Imaging. 2014;7:282–91.
63. Celeng C, Maurovich-Horvat P, Ghoshhajra BB, Merkely B, Leiner 80. Yamamoto H, Kitagawa T, Ohashi N, et al. Noncalcified atheroscle-
T, Takx RA. Prognostic value of coronary computed tomography rotic lesions with vulnerable characteristics detected by coronary
angiography in patients with diabetes: a meta-analysis. Diabetes CT angiography and future coronary events. J Cardiovasc Comput
Care. 2016;39:1274–80. Tomogr. 2013;7:192–9.
64. Yiu KH, de Graaf FR, Schuijf JD, et  al. Prognostic value of
81. Genders TS, Steyerberg EW, Alkadhi H, et al. A clinical prediction
renal dysfunction for the prediction of outcome versus results rule for the diagnosis of coronary artery disease: validation, updat-
of computed tomographic coronary angiography. Am J Cardiol. ing, and extension. Eur Heart J. 2011;32:1316–30.
2011;108:968–72.
 Coronary CT Angiography
for Evaluation of Acute Coronary 28
Syndrome in the Emergency
Department

Nam Ju Lee and Harold Litt

There is nearly one death from heart disease every 38 s in the discharged directly from the ED [7]. The presence or absence
United States [1]. Acute chest pain is the single most com- of specific symptoms is not reliable when used to rule out
mon complaint of patients older than 15 years of age present- myocardial ischemia [8], and risk factors (hypercholesterol-
ing to the emergency department (ED) [2] and accounts for emia, hypertension, family history, and tobacco use) remain
about 4% of ED visits in the United States [3]. Origins of poor predictors of ACS [9, 10]. The standard 12-lead ECG
chest pain include diseases of the heart, aorta, pulmonary cannot exclude ACS conclusively because diagnostic ECG
system, esophagus, upper abdomen, and chest wall and even findings are present only in a minority of patients, although
psychiatric disorders. Determination of the etiology of the it is the single best test to identify ST-segment elevation MI
chest pain is often difficult, although different types of chest (STEMI) [11, 12]. Despite improvements in cardiac bio-
pain are classically ascribed to different corresponding dis- markers (creatine kinase, CK-MB, myoglobin, troponin I,
eases. Acute coronary syndrome (ACS) is estimated to be troponin T) for early diagnosis and risk stratification of acute
responsible for 20% of all clinical encounters for acute chest MI [13–16], they are not universally elevated in patients with
pain [4]. Patients with ACS present with unstable angina, unstable angina or transient myocardial ischemia [17].
acute myocardial infarction, or sudden cardiac death [5]; Therefore, negative markers cannot exclude ACS entirely,
therefore, timely triage of ACS is important as it affects treat- and evaluation is still needed according to their clinical
ment and prognosis. Also timely triage may save significant presentations.
costs. Using coronary CTA to evaluate patients instead of Patients with low–intermediate TIMI score without initial
admitting patients for a rule-out approach with serial tropo- enzyme level elevation or ECG changes are typically admitted
nin has the potential to save significant costs to the health-­ for complete evaluation with further enzyme analysis and
care system. often for myocardial stress imaging, whereas patients with a
high-risk score are referred for intravenous (IV) heparin and
further investigation with catheterization for intervention [18].
Risk Assessment The HEART score in another risk classification system is
based on history, ECG, age, risk factors, and troponin. In the
The initial step in evaluation of potential ACS in the ED is initial derivation cohort, MACE occurred in 1.7% of those
assessment of patient risk; this may be performed using with low HEART scores (values 0–3), which was 36.4% of
scales such as the thrombosis in myocardial infarction patients; therefore low HEART scores (0–3) exclude short-­
(TIMI) score, including clinical and medical history, coro- term MACE with >98% certainty. Conservative evaluation
nary artery disease risk factors, electrocardiogram (ECG) policies might be considered for those patients compared to
results, and serum cardiac enzyme levels [6]. In approxi- more aggressive evaluation in patients with high HEART
mately 10% of all ED chest pain patients, with a TIMI risk scores (7–10) who have a higher risk of MACE [19]. The
score of zero combined with negative serial cardiac bio- c-statistic of the HEART score (0.83) is significantly higher
marker testing, the patient may be considered at low risk and than the c-statistic of TIMI (0.75) and GRACE (0.70),
respectively (p < 0.0001) [19].
Weinstock et al. showed in adult patients with chest pain
N. J. Lee admitted with two negative serial biomarkers, no concerning
Department of Radiology, Mayo Clinic, Rochester, MN, USA vital signs, and nonischemic ECG findings, short-term
H. Litt (*) ­clinically relevant adverse cardiac events were rare and com-
Department of Radiology, Perelman School of Medicine of the monly iatrogenic and suggested that routine inpatient admis-
University of Pennsylvania, Philadelphia, PA, USA
e-mail: [email protected] sion may not be a beneficial strategy for this group [20].

© Humana Press 2019 331

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_28
332 N. J. Lee and H. Litt

However, the reported 2–4% rate of missed diagnosis of ACS presenting with a possible acute coronary syndrome. Of 640
[21–23] may expand the number of tests and hospitalizations patients with a negative CCTA examination, none died or
in relatively low-risk populations with chest pain without had a myocardial infarction within 30 days, and patients in
definite evidence of ACS [24]. the CCTA group had a higher rate of direct discharge from
the ED (49.6% vs. 22.7%), a shorter length of stay (LOS)
(median, 18.0 h vs. 24.8 h), and a higher rate of detection of
CCTA for ACS Patients in ER coronary disease (9.0% vs. 3.5%) compared with patients
receiving traditional care [39].
Patients with a low to intermediate risk of ACS are usually CCTA is comparable to myocardial stress perfusion imag-
judged clinically according to initial risk assessment data ing in terms of safety and accuracy for excluding or diagnos-
including clinical history, ECG, biomarkers, and TIMI score ing ACS [33].
of 0–2. CCTA may be performed for patients with prior neg- The Computed Tomographic Angiography for the
ative stress test or even for those with TIMI score 3–4, which Systematic Triage of Acute Chest Pain Patients to Treatment
would allow cardiac CTA to be ordered for approximately (CT-STAT) trial showed that CCTA and stress SPECT myo-
60–70% of patients [25–27]. Recently, appropriate use of cardial perfusion imaging led to a similar number of patients
criteria for the imaging of ED patients presenting with chest referred to invasive coronary angiography, 6.9% and 6.2%,
pain has been published [28, 29]. respectively, similar to findings in ACRIN PA 4005. Time to
Most international guidelines suggest that the least expen- diagnosis and hospital costs were significantly reduced with
sive and most widely available stress test, treadmill ECG, CCTA, to an average of 3 h compared with 7 h for those who
should be the first-line test for patients at intermediate risk received stress SPECT MPI; direct costs were reduced by
for acute coronary syndrome [30]. However, Hermann et al. 38%, from roughly $3500 to $2000. However, Salerno et al.
demonstrated that routine provocative cardiac testing gener- pointed out that there was lack of stress ECG as a comparator
ated a small therapeutic yield with new diagnoses of coro- [44].
nary artery disease being uncommon, but many false-positive CCTA also reduces cost in addition to reduction in diag-
results in an emergency department-based chest pain unit nostic time in the intermediate-risk population [32, 45].
[31]. In their study, routine provocative testing was positive Recent randomized controlled trials in acute ED patients that
for coronary ischemia in 470/4181 (11.2%) patients, of compared CCTA to nuclear SPECT [39] or to a mixed stan-
whom 123 underwent coronary angiography. Obstructive dard of care showed lower length of stay at lower cost [38,
disease was confirmed in 63 of 123 (51.2% true positive), 40].
and 28 (0.7% overall) had findings consistent with the poten- In the ROMICAT II trial, 1000 patients with symptoms
tial benefit from revascularization (American Heart suggestive of ACS were randomized to an early CCTA or
Association class I or IIa) [31]. standard ED evaluation. Similar to the other studies,
The application of CCTA in the ED is based on trial data ROMICAT II demonstrated a shorter length of stay (LOS)
and observation from a limited number of centers [32–37], for patients who underwent CCTA than standard evaluation,
and CCTA is now seen as a viable alternative to functional and ED costs were lower. However, catheterization and
testing in the evaluation of patients with acute chest pain in revascularization rates were higher in the CCTA arm, leading
the ED [38–40]. to higher inpatient costs and overall cost neutrality. Women
Early utilization of CCTA in the ED for appropriately in the CCTA arm had greater reduction in LOS, lower hospi-
selected patients with a low to intermediate risk of ACS iden- tal admission rates, and a smaller increase in cumulative
tifies low-risk patients quickly without the added delay of radiation dose than men when comparing ED strategies
serial biomarkers or prolonged observation. It facilitates (p-interactions ≤ 0.02) [46]. While women had lower ACS
more rapid discharge from the ED compared to those under- rates than men, sex differences in LOS persisted after adjust-
going non-CCTA standard care including serial ECGs, car- ment for baseline differences including ACS rate (p-­
diac biomarkers, and subsequent cardiac testing such as interaction < 0.03). LOS was similar between sexes with
exercise testing, stress perfusion imaging, or cardiac cathe- normal CCTA findings (p = 0.11). However, there was limi-
terization without a statistically significant difference in tation of the lack of stress ECG as a comparator.
major adverse cardiac events (MACE) [34, 38–43]. In a pro- The recent CT Coronary Angiography Compared to
spective cohort trial of 568 patients in the ED with low TIMI Exercise ECG (CT-COMPARE) study represents the largest
score, 84% of patients were discharged based on a negative prospective, randomized trial of CCTA comparing to
CCTA without adverse outcomes in 30 days with only one ­treadmill exercise ECG as part of the standard of care in
potential cardiac death at 1-year follow-up [36]. low–intermediate risk possible ACS patients presenting to
ACRIN PA 4005 was a randomized controlled study of the ED to evaluate diagnostic performance measures and the
CCTA versus usual care for low-to-intermediate-risk patients hospital-­based costs of CCTA-based care as compared to
28  Coronary CT Angiography for Evaluation of Acute Coronary Syndrome in the Emergency Department 333

ECG-based care [30]. The data showed that CCTA-based Contraindications to nitroglycerin include recent phosphodi-
evaluation is 35% faster and 20% less expensive than ExECG esterase inhibitor therapy, commonly used for erectile dys-
in patients at low–intermediate risk of acute coronary syn- function or pulmonary hypertension, hypotension, and
drome. These results suggest that coronary CT angiography critical aortic stenosis.
is faster and less expensive, with improved diagnostic perfor- Cardiovascular computed tomography angiography
mance compared to exercise treadmill ECG-based care in (CCTA) had been associated with considerable radiation
symptomatic patients at low–intermediate risk for acute cor- doses, with previous studies reporting high average radiation
onary syndrome presenting to the emergency department doses of >10  mSv [55]. Retrospective electrocardiographi-
[30]. These data add additional proof that protocols using cally (ECG)-gated CCTA was used in which radiation is
CCTA rather than stress testing-based care in emergency applied during the entire cardiac cycle with 100% tube cur-
departments should be considered [30]. rent. A substantial reduction in radiation dose was achieved
CCTA has shown its safety and efficiency in excluding with the introduction of prospective ECG-triggered CCTA in
coronary artery disease or relevant stenosis for advanced risk which radiation is only administered during a predefined
stratification over the past decade in the setting of ambiguous phase of the cardiac cycle and the tube current is turned off
acute chest pain of a low to intermediate risk of ACS with or reduced outside that phase [56]. Although retrospective
86–100% sensitivity, 92–98% specificity, 93–100% negative ECG-triggering yields functional information during the car-
predictive value, and 50–90% positive predictive value [32, diac cycle, its use exposes patients to more radiation, whereas
47–50]. However, patients with indeterminate lesions or prospective ECG-triggering provides fewer cardiac phases
nondiagnostic CCTA will be referred for stress testing to for interpretation but reduces the radiation dose. A meta-
evaluate the physiologic relevance of intermediate lesions. analysis comparing retrospective and prospective ECG-
There are minority of uninterpretable cases due to patient’s triggering found that an almost fourfold reduction in radiation
motion or unexpected variability of heart rate, with a goal to dose (11.3–3.5 mSv) was possible using prospective ECG-
be below 5% of total studies. triggering, while image quality remained comparable [57].
Therefore, prospective ECG-­triggering is now widely used
for the majority of CCTA acquisitions. Despite these
Performing CCTA advances, the ability to lower dose in CCTA by prospective
ECG-triggering alone remains problematic in patients with
Patients with a serum creatinine level of greater than 1.5 mg/ high heart rates or arrhythmias [58].
dL or a severe allergy to iodinated contrast material and Conventionally, a low pitch is used for retrospectively
those who are pregnant are not eligible for CCTA. Conditions gated CCTA acquisitions to ensure adequate data sampling;
that may be not suitable for CCTA are active asthma or other however, with second- and third-generation dual-source
contraindications to β-blocker administration, irregular heart scanners, high-pitch acquisitions are possible in which the
rhythms such as atrial fibrillation, and weight more than entire chest including the heart is covered within a single
150  kg, depending on the specific CT technology in use. heartbeat. This is possible because the second detector fills in
Heart rate control to less than 70 bpm is required for high-­ the gaps in data sampling from the first detector. It is possible
quality CCTA, depending upon the technology in use [51]. A to maintain image quality using high-pitch spiral prospective
fast-acting β-blocker with a short half-life, typically oral ECG-triggering in patients with a regular heart rhythm up to
metoprolol, is given at least 1 h before imaging for patients 70 beats/min [56]. Wide area or volume detector scanners are
with fast heart rates when there is no contraindications, e.g., not constrained by pitch because the whole heart can be
active or unstable small airways disease, hypotension, sinus imaged within a single heartbeat [57].
bradycardia, or recent cocaine use. The initial dose is Blooming artifact from densely calcified plaque limits the
50–100  mg with additional doses up to 200  mg if needed accurate evaluation of stenosis, and the degree of stenosis is
[52]. IV metoprolol can be administered in the CT suite, and often overestimated [59]. Despite concerns about image
the effect occurs typically in 5–10 min; 5 mg IV can be given quality in patients with calcium score over 400, a study
initially, with additional doses up to 20 mg if required. For showed that only a small percentage of these scans in ED
patients with contraindications to β-blockers, calcium chan- patients are uninterpretable [60]. Newer dual-energy tech-
nel blockers, preferably diltiazem with the least negative ino- nology using simultaneous or alternating imaging at two dif-
tropic effect, can be considered as an alternative (initial IV ferent X-ray tube potentials allows improved quantification
bolus of up to 0.25 mg/kg). Careful monitoring of vital signs of calcified plaque by reducing tissue blooming and beam
is required with IV medication. Utilizing sublingual nitro- hardening beyond single-energy MDCT [61]. Imagers may
glycerin before contrast-enhanced images is valuable for decide to proceed with scanning based on the likelihood of a
improving the contrast-to-noise ratio and vessel visualiza- diagnostic study even in patients with high calcium score.
tion by vasodilation of the coronary arteries [53, 54]. Evaluation of coronary stent patency is limited due to
334 N. J. Lee and H. Litt

b­ looming artifacts similar to dense calcium [62]. However, alone cannot be used to determine risk in patients presenting
stents are unlikely to be present in a patient with low to inter- with potential ACS in the ED and does not add prognostic
mediate risk of ACS presenting to the ED. value for acute events although it may be useful for long-­
Triple-rule-out (TRO) CT examines the coronary arteries, term management of cardiac risk in outpatients. One study
the thoracic aorta, and the pulmonary arteries simultaneously showed that a significant number of patients with a zero
[63]. For the diagnosis of coronary artery disease, TRO CT CAC had CTA findings [74]; therefore, CTA is better than
has a sensitivity of 94.3%, a specificity of 97.4%, and a nega- CAC scoring in determining the atheroma burden, especially
tive predictive value of 99%, findings that are similar to those in patients with risk factors.
obtained with the use of dedicated coronary CTA [64]. However, higher Agatston score is correlated with
Compared with dedicated CCTA, TRO CT has greater ana- increased risk of cardiac events and worse overall prognosis
tomic coverage, including the structures above the carina, [75]. Framingham risk score combined with coronary cal-
leading to a higher radiation dose [65]. Therefore, TRO CT cium score can improve 10-year risk stratification of asymp-
requires a slightly higher IV contrast load to opacify both the tomatic people, especially those with intermediate risk
right- and left-side circulations [64, 65]. TRO CT is thought (Framingham risk score 10–20%) [76]. Calcium scoring may
to be beneficial in the evaluation of patients with chest pain be recommended in this group to support clinical decision-­
for whom additional diagnoses other than ACS are consid- making, particularly whether to start aggressive medical
ered in addition to a low to intermediate risk for ACS at base- therapy. The risk is highest when the Agatston score is above
line, such as pulmonary embolism or aortic dissection [66]. 400. Cardiovascular risk increases proportionally to the
Because of the relatively low prevalence of pulmonary amount of calcium, and an annual progression of more than
embolism and aortic dissection in this patient population 15% enhances the risk of myocardial infarctions although
[64], appropriate patient selection is important for the effec- positive predictive value of CAC progression is low as a
tive application of TRO CT [63]. marker of risk [77–79]. After myocardial infarction, patients
have higher CAC progression than event-free subjects [80].
CAC scoring does not necessarily add relevant information
Coronary Calcium Evaluation in high- or low-risk populations [47]. While an unenhanced
series may be useful to customize the CCTA scan field and
Coronary artery calcification (CAC) can be seen in most other parameters to reduce doses, the effective dose of a cal-
patients with ACS or sudden cardiac death, and it has been cium scoring acquisition is now similar to CCTA doses
established as a quantitative marker and indicator of athero- obtained with lower tube voltage technique, and therefore its
sclerosis using the Agatston score [67–70]. The Agatston value in the ED is uncertain [81].
scoring scale calculates an area for all pixels above a thresh-
old of 130 HU, in contiguous 3 mm slices, (the slice thick-
ness and spacing used by Agatston et al), and multiplies it by Coronary Stenosis
a density factor. The score is the summed or total “score” for
the entire epicardial coronary system. Prospective ECG-­ There is a very low MACE rate (0.5%) in the 30 days after
triggering is the current mode for measuring coronary cal- ED presentation in patients with a stenosis of 25–50% on
cium at most centers. cardiac CTA [82], although one study reported that 2.6% of
A study of 1031 patients with a median CACS of 0 (61% ED patients with visible atherosclerosis but less than 50%
with CACS of 0) showed a majority of patients evaluated for stenosis at CCTA experienced ACS as identified by serial
chest pain of uncertain cardiac cause have a CACS of 0, cardiac markers but not by stress testing [37]. All patients
which predicts both a normal SPECT result and an excellent had small vessel disease, which is associated with a low inci-
short-term outcome [71]. Only 2 events occurred in 625 dence of adverse events and mortality, and is not generally
patients with a CACS of 0 (0.3%, 95% confidence interval amenable to revascularization [83–87]. Nonetheless, small
0.04–1.1%). Both of these patients developed increased tro- vessels are a limitation of CCTA evaluation [88], and there-
ponin levels during their index visit but had normal serial fore, serial cardiac makers may be beneficial in patients with
ECG and SPECT study results and no cardiac events at less than 50% stenosis on CCTA, but further evaluation such
6-month follow-up [71]. as stress testing is not indicated during the index visit [82].
However, a lack of coronary calcium does not definitely Short-interval outpatient follow-up is important to establish
exclude coronary stenosis, and a high amount of calcium strategies for primary prevention.
does not necessarily correlate with angiographic luminal ste- The ROMICAT I trial demonstrated a 77% sensitivity of a
nosis or vulnerability of plaques [72, 73]. The role of cal- stenosis ≥50% for detection of ACS [37]. In the ROMICAT
cium scoring in ED patients with suspected ACS continues to II study, a stenosis ≥50% was detected in 78% of patients
be debated, because it has been shown that calcium scores with ACS, which is similar to invasive angiography studies
28  Coronary CT Angiography for Evaluation of Acute Coronary Syndrome in the Emergency Department 335

that observed an absence of significant stenosis in 12–14% Coronary Atherosclerotic Plaques

of patients with ACS [89]. A clinical registry by Velzen et al.
showed sensitivity, specificity, and positive and negative pre- Atherosclerotic lesions can be characterized as calcified,
dictive values to detect significant CAD on CTA were 100, non-calcified, and partially calcified (mixed plaques) on CT
87, 93, and 100%, respectively [90]. During mean follow-up (Fig. 28.1). On a cellular basis, atherosclerotic changes can
of 13.7 months, no cardiovascular events occurred in patients be classified in to six types (I–VI). Type III is the border
with a normal CTA examination. In patients with nonsignifi- where atherosclerotic changes become visible to the eyes
cant CAD on CTA, no cardiac deaths or myocardial infarc- [93]. Types IV, V, and VI are classified as advanced plaques
tions occurred and only one patient underwent [94]. The entire vessel grows with increasing plaque volume
revascularization due to unstable angina. In patients present- so that the lumen diameter is maintained; therefore, type IV
ing with acute chest pain, an excellent clinical performance lipid core atheroma and type V fibroatheroma (atheroma
for the noninvasive assessment of significant CAD was dem- with fibrous tissue) can be asymptomatic with adaptive
onstrated using CTA.  Normal or nonsignificant CAD on change of the arteries [95]. When thrombus or hemorrhage
CTA predicted a low rate of adverse cardiovascular events develops in type IV–V lesions, patients can be symptomatic
and favorable outcome during follow-up [90]. ROMICAT I (type VI, complicated). Most ACS are thought to be the
demonstrated limited positive value of a stenosis >50% at result of sudden luminal thrombosis, which occurs from
CCTA for ACS because matching perfusion abnormalities plaque rupture, plaque erosion, or calcified nodules [5, 96–
on SPECT MPI testing were seen in only 46% of these 100]. Plaque rupture refers to luminal thrombosis occurring
patients [91]. The optimal management of patients with when the highly necrotic core of a ruptured thin-cap fibroath-
intermediate lesions at CCTA remains uncertain and may eroma contracts with platelets. Hemodynamically nonsig-
result in an increased frequency of downstream testing and nificant plaque may be prone to rupture depending on the
interventions [38–40, 92]. However, patients with significant morphology and biochemical structure [89, 101–105]. Thin-­
stenosis on CTA cannot be discharged from the ED after ini- cap fibroatheroma is often seen pathologically in plaque rup-
tial troponin and ECG [38–40, 92]. ture with myocardial infarctions, where speckled calcification

Fig. 28.1  Characterization of

atherosclerotic plaques by
CT. Curved planar
reformatted images from
coronary CT studies
performed on ED chest pain
patients demonstrate
non-calcified (left), mixed
(center), and calcified (right)
plaques
336 N. J. Lee and H. Litt

can be visualized [5]. Luminal thrombus from plaque ero- niques may assess plaque composition [61], and myocardial
sion occurs when there is minimal inflammation of an under- perfusion provides prognostic information [120]. Dual-­
lying base rich in proteoglycans and smooth muscle cells, in energy CT may assist in distinguishing densely calcified
which plaque is mostly devoid of necrotic core or the necrotic plaques from other plaque types although further plaque dis-
core dose not communicate with the lumen due to a thick crimination such as fibrous vs. lipid-rich plaque is not cur-
fibrous cap. Calcified nodules are the least common in lumi- rently possible [121].
nal thrombosis, and their effect on instability is less evident High-risk plaque features have been associated with an
although coronary calcification correlates highly with plaque increased risk for future cardiovascular events in patients
burden [5]. with stable chest pain syndromes [110, 122, 123]. These
ACS develops frequently from relatively mild or moder- findings are more often seen with culprit lesions in patients
ate lesions angiographically in the early stage of atheroscle- with ACS compared with similarly stenotic plaques in
rosis [104, 106, 107]. Coronary artery plaque features may patients with stable angina [63]. Studies reported 5–15%
assist in further stratifying risk beyond that obtained using prevalence of high-risk plaque features in the acute chest
percent stenosis [108–110]. Reported features of high-risk pain population, in patients undergoing invasive coronary
plaque (Fig. 28.2) are large plaque volume, positive remod- angiography, and in non-culprit vessels of patients with
eling, spotty calcium, greater proportion of non-calcified ACS (Providing Regional Observations to Study Predictors
plaque, low mean and minimal CT attenuation value of Events in the Coronary Tree, PROSPECT trial) [92,
(Hounsfield units, HU), and a contrast-enhanced rim seen 108–110, 114, 115]. The value of high-risk plaque features
around the central filling defect (napkin-ring sign) [109, for the diagnosis of ACS in patients with significant steno-
111–117], which are similar to high-risk findings on intra- sis was demonstrated in the ROMICAT I trial [124]. In the
vascular imaging (positive remodeling, larger plaque area, ROMICAT II trial, high-risk plaques on coronary CTA
spotty calcium, and large necrotic core) [118, 119] as well as increased the likelihood of ACS independent of stenotic
histology. These changes can directly be visualized on CCTA CAD and clinical risk assessment (age, sex, and number of
but will not appear on invasive angiography because the ves- cardiovascular risk factors) in patients presenting to the
sel wall is not visualized [47]. Multi-energy imaging tech- ED with acute chest pain but negative initial electrocardio-

a b c d

Fig. 28.2  High-risk plaque features. Curved planar and MPR images and positive remodeling, (b) spotty calcification, (c) low attenuation
from coronary CT studies performed on ED chest pain patients show plaque, positive remodeling, and spotty calcification, and (d, e) the
various high-risk plaque features including (a) low attenuation plaque napkin-ring sign
28  Coronary CT Angiography for Evaluation of Acute Coronary Syndrome in the Emergency Department 337

gram and troponin [92]. The ROMICAT score derived results on the diagnostic performance of an alternative
from semiautomated quantitative measurements of high- on-site algorithm (Siemens Healthineers, Forchheim,
­
risk plaque features (volume of <60HU plaque, remodel- Germany); these results were subsequently corroborated
ing index, spotty calcium, plaque length) was an by a larger study [133, 134]. These investigations have
independent predictor of ACS during the index hospital- demonstrated that algorithm-based noninvasive FFR deri-
ization and was incremental to gender and presence of vation from CCTA compares favorably to the diagnostic
≥50% stenosis [125]. When applying the ROMICAT score gold standard of invasive FFR.
derived from the ROMICAT I trial to the patient popula- CT myocardial perfusion is growing as a new tech-
tion of the ROMICAT II trial, the ROMICAT score (OR nique that provides a functional assessment of the myo-
2.9, 95% CI 1.4–6.0, p  =  0.003) was a predictor of ACS cardium along with a comprehensive evaluation of
after adjusting for gender and  ≥50% stenosis [125]. The coronary artery disease within a single modality [135].
AUC of the model containing ROMICAT score, gender, The addition of dynamic CT myocardial perfusion to stan-
and ≥50% stenosis was 0.91 (95% CI 0.86–0.96) and was dard coronary CTA can provide insightful information on
better than with a model that included only gender significant coronary stenosis, particularly for hemody-
and  ≥  50% stenosis (AUC 0.85, 95% CI 0.77–0.92, namically relevant lesions, which may be helpful for
p = 0.002) [125]. patient management [135]. Recently, CT myocardial per-
Interestingly, nonalcoholic fatty liver disease (NAFLD) fusion studies have been performed with promising results
was reported to be associated with advanced high-risk cor- [136–138]. Although results are promising, this technique
onary plaque, independent of traditional cardiovascular comes at the price of additional radiation exposure and
risk factors and the extent and severity of coronary artery contrast media along with the need for pharmacologic
disease [126]. stress agents [139].
Patients with remaining obstructive non-culprit plaques
and higher plaque burden index had a higher risk of
MACE. In multivariate analysis, with diabetes, dyslipidemia, CCTA Reporting
and plaque burden index, obstructive non-culprit plaques
remained an independent predictor of MACE [127]. Almost CAD-RADS categories depend on stenosis severity, and a
a quarter of the study population experienced a new event classification system suggested by the Society of
arising from a non-culprit plaque during a follow-up of Cardiovascular Computed Tomography is used to grade
almost 5  years. ACS patients with remaining obstructive stenosis severity [140–142]. They range from CAD-
non-culprit plaques or high plaque burden have an increased RADS 0 (absence of atherosclerosis) to CAD-RADS 5
risk of future MACE [127]. (presence of at least one total occlusion) (Table  28.1)
(Fig.  28.3). Categories should reflect the clinically most
relevant finding per patient and be complementary to the
Advanced Application of CCTA final impression of the report to provide specific informa-
tion regarding the location and extent of coronary plaque
In the presence of coronary stenosis, fractional flow and stenosis.
reserve (FFR) can be measured angiographically to deter-
mine the coronary blood volume. Typically FFR over
0.75–0.8 has been associated with more positive outcomes
and decreased ischemia after revascularization [128]. Table 28.1  Grading scale for stenosis severity and CAD-RADS
CT-FFR adds functional aspects to stenosis assessment CAD-­ Degree of luminal
and reduces false-­ positive test results, which leads to RADS diameter stenosis (%) Terminology interpretation
N Nondiagnostic ACS cannot be
improved specificity [129]. A recent meta-analysis [129]
excluded
suggests that noninvasive CT-FFR offers improved speci- 1 0 No visible ACS highly
ficity without noticeably altering the sensitivity of CCTA stenosis unlikely
in detecting hemodynamically relevant lesions in patients 2 1–24 Minimal ACS highly
with suspected or known coronary artery disease over ana- stenosis unlikely
tomic interpretation by CCTA alone [129]. At this time, 3 25–49 Mild stenosis ACS unlikely
4 50–69 Moderate ACS possible
three large prospective clinical trials have demonstrated
stenosis
the diagnostic value of a noninvasive FFR (fractional flow 5 70–99 Severe ACS likely
reserve from coronary computed tomographic angiogra- stenosis
phy [CT-FFR]) algorithm (HeartFlow, Inc., Redwood City, 6 100 Occluded ACS very
CA) [130–132]. Furthermore, an initial study presented likely
338 N. J. Lee and H. Litt

a b c d

e f

Fig. 28.3  CAD-RADS reporting categories. Curved planar and 50–69% stenosis of the LAD; (e) CAD-RADS5, non-calcified plaque
MPR images from coronary CT studies performed on ED chest pain causing 70–99% stenosis of the LAD (arrow) and a corresponding
patients show (a) CAD-RADS1, no luminal stenosis in the RCA; (b) subendocardial perfusion defect in the anterior wall (arrowhead); (f)
CAD-­RADS2, calcified and non-calcified plaque causing 1–24% ste- CAD-RADS6, occlusion of the circumflex artery (arrow) and a cor-
nosis of the LAD; (c) CAD-RADS3, mixed plaque causing 25–49% responding subendocardial perfusion defect in the lateral wall
stenosis of the LAD; (d) CAD-RADS4, calcified plaque causing (arrowhead)
28  Coronary CT Angiography for Evaluation of Acute Coronary Syndrome in the Emergency Department 339

Fig. 28.4  Myocardial bridging. MPR image from a coronary CT study

performed on an ED patient with chest pain demonstrates long intra-
myocardial course of the mid-LAD, without narrowing of the bridged
segment
Fig. 28.5  Coronary anomalies. Thin-slab MIP image from a coronary
CT study performed on an ED patient with chest pain demonstrates
Non-atherosclerotic Causes of Chest Pain anomalous origin of the right coronary artery from the left cusp with
interarterial course. This course has been associated with arrhythmia
and sudden cardiac death
Myocardial bridging (Fig. 28.4), a segment of an epicardial
coronary artery coursing within the myocardium, is present
in about one third of adults [143]. In the majority of cases, artery anomalies are at risk of dying, especially with exercise
systolic compression of the tunneled segment remains clin- [152]. Sudden cardiac arrest (SCA) secondary to congenital
ically silent. With provocative testing in patients with angi- anomalous coronary artery disease occurs due to insufficient
ographically normal coronary arteries, myocardial bridges coronary flow by the anomalous LCA to meet elevated left
are revealed in ≤40% of cases by increasing systolic com- ventricular myocardial metabolic demand, usually during
pression [144]. Rare complications of myocardial bridging exertion or exercise. In a majority of previously reported
include angina, myocardial ischemia, myocardial infarc- cases, SCA was triggered by exertion, and most of these
tion, left ventricular dysfunction, myocardial stunning, par- patients have a positive exercise stress test [153]. Contributing
oxysmal AV blockade, exercise-induced ventricular factors to an increased resistance in the LCA include com-
tachycardia, and sudden cardiac death [145–148]. The like- pression between the great vessels, a slit-like ostium, myo-
lihood of ischemia also increases with the intramyocardial cardial bridging, or unfavorable geometry [154]. High-risk
depth of the tunneled segment; however, there is no consis- defects include those involved with the proximal coronary
tent association between the patient’s symptom and the artery or coursing of the anomalous artery between the aorta
length and depth of the tunneled segment or the degree of and pulmonary trunk [152]. Left anterior descending coro-
systolic ­compression [149, 150]. nary arteries arising from the right cusp with an interarterial
CCTA is ideal to demonstrate the anatomic course of cor- course have been considered malignant (Fig.  28.6), and
onaries and caliber change throughout the cardiac cycle anomalous coronary origins can be associated with acute
when retrospectively ECG-gated CCTA is employed. About chest pain [152].
80% of coronary artery anomalies are benign and incidental
findings at the time of catheterization [151]. Although the
significance of coronary anomalies is mostly unclear, poten- High-sensitivity Troponin
tially serious anomalies (Fig.  28.5) which include ectopic Assays (Hs–Troponins)
coronary origin from the pulmonary artery or opposite aortic
sinus, single coronary artery, and large coronary fistulae can The introduction of high-sensitivity troponin (hs-troponin)
result in angina pectoris, myocardial infarction, heart failure, assays has changed standard optimal care (SOC) consider-
arrhythmias, and sudden cardiac arrest [151]. Younger ably, and measurement has become standard practice in
patients in their first three decades with isolated coronary many institutions, allowing for more accurate and faster
340 N. J. Lee and H. Litt

a b

Fig. 28.6  Malignant coronary anomaly. (a) Thin-slab MIP image from arterial course. The LAD is markedly narrowed in the interarterial
a coronary CT performed on an ED patient with chest pain demon- segment, leading to a large subendocardial perfusion defect seen on a
strates anomalous origin of the LAD from the right cusp, with an inter- four-chamber MPR image (b)

­rule-­out of ACS [15, 16]. These new assays are more s­ ensitive strategy of avoiding further testing in subjects with unde-
and reach negative predictive values of >97% for myocardial tectable initial hsTnI, performing CACS on subjects with
infarction within 3 h [15, 16]. Early observations indicated detectable initial hsTnI but nonincreased hsTnI (less than
that hs-troponins would allow fast and accurate exclusion of 99th percentile), and obtaining CTA in subjects with
ACS in a substantial proportion of low- to intermediate-risk Agatston greater than 0 will have a negative predictive
patients, obviating the need for prolonged observation and value of 100.0% (95% confidence interval, 98.2%–
in-hospital diagnostic testing in the absence of elevated high-­ 100.0%). The addition of CACS to hsTnI improves the
sensitivity cardiac biomarkers or precarious ECG identification of low-risk subjects in whom CTA might be
­abnormalities [155–157]. avoided [159].
The BEACON (Better Evaluation of Acute Chest Pain A cohort study in patients with suspected ACS enrolled in
with Coronary Computed Tomography Angiography) trial is the ROMICAT II trial and randomized to coronary CTA who
a European randomized trial comparing a diagnostic strategy also had hsTnI measurement at the time of the emergency
supplemented by early CCTA with SOC for patients sus- department presentation assessed coronary CTA for tradi-
pected of ACS in the era of hs-troponins [158]. In their study, tional (no CAD, nonobstructive CAD, ≥50% stenosis) and
CCTA applied early in the work-up of suspected ACS is safe advanced features of CAD (≥50% stenosis, high-risk plaque
and associated with less outpatient testing and lower costs. features: positive remodeling, low <30-Hounsfield units
However, CCTA does not identify more patients with signifi- plaque, napkin-ring sign, spotty calcium) [160]. hsTnI at the
cant CAD requiring coronary revascularization, shorten hos- time of presentation followed by early advanced coronary
pital stay, or allow for more direct discharge from the ED in CTA assessment improved the risk stratification and diag-
the era of hs-troponins [158]. Discharge from the ED was not nostic accuracy for ACS as compared to conventional tropo-
more frequent after CCTA (65% vs. 59%, p  =  0.16), and nin and traditional coronary CTA assessment [160]. The
length of stay was similar (6.3 h in both groups, p = 0.80). study showed absence of ≥50% stenosis and high-risk plaque
The CCTA group had lower direct medical costs (€337 vs. ruled out ACS in patients with intermediate hsTnI (n = 87,
€511, p < 0.01) and less outpatient testing after the index ED 54.4%; ACS rate 0%), whereas patients with both ≥50% ste-
visit (10 [4%] vs. 26 [10%], p < 0.01). There was no differ- nosis and high-risk plaque were at high risk (n = 13, 8.1%;
ence in incidence of undetected ACS [158]. ACS rate 69.2%) and patients with either ≥50% stenosis or
Korley FK et al. showed that no subjects with undetect- high-risk plaque were at intermediate risk for ACS (n = 39,
able hsTnI or zero calcium score (CACS) had a MACE. A 24.4%; ACS rate 7.7%) [160].
28  Coronary CT Angiography for Evaluation of Acute Coronary Syndrome in the Emergency Department 341

Alternative Tests diagnostic interventions and resource use for low- to

intermediate-­ risk populations [177, 178]. Applying risk
In patients with low to intermediate pretest probability, other stratification tools (such as the HEART score), combined
tests with high negative predictive value to rule in or rule out with normal ECG findings and negative cardiac biomarker
ACS or myocardial infarct (MI) to enable early discharge level results, is sufficient to identify a low- to intermediate-­
include rest CMR, echocardiography, and SPECT MPI, risk patient population who will not benefit from the addition
though these must be performed while the patient is experi- of coronary CT angiography or other means of objective
encing chest pain. Resting SPECT MPI with technetium-­ assessment [178, 179].
99m sestamibi is useful in the setting of suspected ACS Also, at least one study has shown that patients who did
[161–164], and its use in the ED setting resulted in earlier not undergo initial noninvasive testing were no more likely
discharge, lower cost, and fewer unnecessary admissions to experience an MI than were those who did receive testing
[165–167]. Observational studies demonstrated its high [177]. Compared with no testing, exercise electrocardiogra-
­negative predictive value to rule out MI or short-term cardiac phy, myocardial perfusion scintigraphy, and coronary com-
events. However, a considerable limitation of resting SPECT puted tomography angiography were associated with
MPI in ED patients is the limited availability of tracers and significantly higher odds of cardiac catheterization and
interpreting personnel as well as decreased accuracy when revascularization procedures without a concomitant improve-
not performed during an episode of chest pain. Resting two-­ ment in the odds of experiencing an MI [177]. Patients with
dimensional (2D) echocardiography provides information by chest pain evaluated in the ED who do not have an MI are at
evaluating wall motion and ejection fraction rapidly and non- very low risk of experiencing an MI during short- and longer-­
invasively [168, 169]. However, the positive predictive value term follow-up in a cohort of privately insured patients. This
is only 0–44% [168, 170, 171] with difficulty distinguishing low risk does not appear to be affected by the initial testing
acute from chronic ischemia. In addition, coronary stenosis strategy. Deferral of early noninvasive testing appears to be
cannot be evaluated unless the patient has a wall motion reasonable [177].
abnormality; as with multiphase CCTA, rest echo may be
useful for evaluation of nonischemic causes of chest pain.
Cardiac MR (CMR) can image coronary anatomy, ventricu- References
lar function, myocardial perfusion, and myocardial fibrosis/
scar. In observational studies, CMR has a sensitivity of 1. Lloyd-Jones D, Adams RJ, Brown TM, et  al. Heart disease and
stroke statistics--2010 update: a report from the American Heart
70–85% to detect ischemic heart disease [172–174], and Association. Circulation. 2010;121(7):e46–e215. https://doi.
coronary MR angiography showed 72% accuracy detecting org/10.1161/CIRCULATIONAHA.109.192667.
stenosis. Although a normal CMR is associated with good 2. Pitts SR, Niska RW, Xu J, Burt CW. National Hospital Ambulatory
Medical Care Survey: 2006 emergency department summary. Natl
prognosis and a very low risk, lower spatial resolution and
Health Stat Report. 2008;7:1–38.
limited availability compared to CCTA limit its clinical 3. McCaig LF, Burt CW.  National Hospital Ambulatory Medical
application in the ED setting [175]. Conventional catheter Care Survey: 2002 emergency department summary. Adv Data.
angiography, considered a gold standard for diagnosis of 2004;340:1–34.
4. Pozen MW, D'Agostino RB, Selker HP, Sytkowski PA, Hood WB
coronary artery disease, should be reserved for use as a con-
Jr. A predictive instrument to improve coronary-care-unit admis-
firmative test and for intervention for patients with signifi- sion practices in acute ischemic heart disease. A prospective mul-
cant stenosis or occlusion who will undergo intervention ticenter clinical trial. N Engl J Med. 1984;310(20):1273–8.
with certain likelihood rather than for initial evaluation of 5. Virmani R, Burke AP, Farb A, Kolodgie FD. Pathology of the vul-
nerable plaque. J Am Coll Cardiol. 2006;47(8 Suppl):C13–8.
patients at less than high risk for ACS. Hence, invasive cath-
6. Antman EM, Cohen M, Bernink PJ, McCabe CH, Horacek T,
eterization should be a purely therapeutic option. Papuchis G, Mautner B, Corbalan R, Radley D, Braunwald
E. The TIMI risk score for unstable angina/non-ST elevation MI:
a method for prognostication and therapeutic decision making.
JAMA. 2000;284(7):835–42.
Future Direction: More or Less 7. Than M, Cullen L, Reid CM, et  al. A 2-h diagnostic protocol
to assess patients with chest pain symptoms in the Asia-Pacific
Data suggest that the management of low- to intermediate-­ region (ASPECT): a prospective observational validation study.
Lancet. 2011;377(9771):1077–84.
risk chest pain patients is changing to enable safe discharge 8. Hickam DH, Sox HC Jr, Sox CH.  Systematic bias in record-
and reduced testing at the index visit [19, 20]. With mounting ing the history in patients with chest pain. J Chronic Dis.
evidence that early objective testing rarely changes manage- 1985;38(1):91–100.
ment [31] and that hospitalization provides no benefit in 9. Jayes RL Jr, Beshansky JR, D’Agostino RB, Selker HP.  Do
patients’ coronary risk factor reports predict acute cardiac isch-
patients with timely and appropriate follow-up [20] and may emia in the emergency department? A multicenter study. J Clin
incur harm [176], care should be taken to avoid unnecessary Epidemiol. 1992;45(6):621–6.
342 N. J. Lee and H. Litt

10. Han JH, Lindsell CJ, Storrow AB, Luber S, Hoekstra JW, of emergency department patients with potential acute coronary
Hollander JE, Peacock WF, Pollack CV, Gibler WB.  The role syndrome. Acad Emerg Med. 2009;16(8):693–8.
of cardiac risk factor burden in diagnosing acute coronary syn- 27. Lee NJ, Litt H. Cardiac CT angiography for evaluation of acute
dromes in the emergency department setting. Ann Emerg Med. chest pain. Int J Cardiovasc Imaging. 2016;32(1):101–12.
2007;49(2):145–52. 152 e141. 28. Rybicki FJ, Udelson JE, Peacock WF, Goldhaber SZ, Isselbacher
11. Lloyd-Jones DM, Camargo CA Jr, Lapuerta P, Giugliano RP, EM, Kazerooni E, Kontos MC, Litt H, Woodard PK. 2015 ACR/
O'Donnell CJ.  Electrocardiographic and clinical predictors of ACC/AHA/AATS/ACEP/ASNC/NASCI/SAEM/SCCT/SCMR/
acute myocardial infarction in patients with unstable angina pec- SCPC/SNMMI/STR/STS appropriate utilization of cardiovas-
toris. Am J Cardiol. 1998;81(10):1182–6. cular imaging in emergency department patients with chest
12. Chase M, Brown AM, Robey JL, Pollack CV Jr, Shofer FS, pain: a joint document of the American College of Radiology
Hollander JE. Prognostic value of symptoms during a normal or Appropriateness Criteria Committee and the American College
nonspecific electrocardiogram in emergency department patients of Cardiology Appropriate use Criteria Task Force. J Am Coll
with potential acute coronary syndrome. Acad Emerg Med. Radiol. 2016;13(2):e1–e29.
2006;13(10):1034–9. 29. Rybicki FJ, Udelson JE, Peacock WF, Goldhaber SZ, Isselbacher
13. Green GB, Beaudreau RW, Chan DW, DeLong D, Kelley CA, EM, Kazerooni E, Kontos MC, Litt H, Woodard PK. 2015 ACR/
Kelen GD. Use of troponin T and creatine kinase-MB subunit levels ACC/AHA/AATS/ACEP/ASNC/NASCI/SAEM/SCCT/SCMR/
for risk stratification of emergency department patients with pos- SCPC/SNMMI/STR/STS appropriate utilization of cardiovas-
sible myocardial ischemia. Ann Emerg Med. 1998;31(1):19–29. cular imaging in emergency department patients with chest
14. Newby LK. The emerging role of myoglobin for risk stratification. pain: a joint document of the American College of Radiology
Am Heart J. 2001;142(1):4–6. Appropriateness Criteria Committee and the American College
15. Reichlin T, Hochholzer W, Bassetti S, Steuer S, Stelzig C, of Cardiology Appropriate Use Criteria Task Force. J Am Coll
Hartwiger S, Biedert S, Schaub N, Buerge C, Potocki M, Noveanu Cardiol. 2016;67(7):853–79.
M, Breidthardt T, Twerenbold R, Winkler K, Bingisser R, Mueller 30. Hamilton-Craig C, Fifoot A, Hansen M, Pincus M, Chan J, Walters
C. Early diagnosis of myocardial infarction with sensitive cardiac DL, Branch KR.  Diagnostic performance and cost of CT angi-
troponin assays. N Engl J Med. 2009;361(9):858–67. ography versus stress ECG--a randomized prospective study of
16. Keller T, Zeller T, Peetz D, et  al. Sensitive troponin I assay in suspected acute coronary syndrome chest pain in the emergency
early diagnosis of acute myocardial infarction. N Engl J Med. department (CT-COMPARE). Int J Cardiol. 2014;177(3):867–73.
2009;361(9):868–77. 31. Hermann LK, Newman DH, Pleasant WA, Rojanasarntikul D,
17. Januzzi JL Jr, Bamberg F, Lee H, Truong QA, Nichols JH, Lakoff D, Goldberg SA, Duvall WL, Henzlova MJ.  Yield of
Karakas M, Mohammed AA, Schlett CL, Nagurney JT, Hoffmann routine provocative cardiac testing among patients in an emer-
U, Koenig W. High-sensitivity troponin T concentrations in acute gency department-based chest pain unit. JAMA Intern Med.
chest pain patients evaluated with cardiac computed tomography. 2013;173(12):1128–33.
Circulation. 2010;121(10):1227–34. 32. Rubinshtein R, Halon DA, Gaspar T, Jaffe R, Karkabi B, Flugelman
18. Braunwald E, Antman EM, Beasley JW, et al. ACC/AHA guide- MY, Kogan A, Shapira R, Peled N, Lewis BS.  Usefulness of
line update for the management of patients with unstable angina 64-slice cardiac computed tomographic angiography for diagnos-
and non-ST-segment elevation myocardial infarction--2002: sum- ing acute coronary syndromes and predicting clinical outcome in
mary article: a report of the American College of Cardiology/ emergency department patients with chest pain of uncertain ori-
American Heart Association Task Force on Practice Guidelines gin. Circulation. 2007;115(13):1762–8.
(Committee on the Management of Patients With Unstable 33. Goldstein JA, Gallagher MJ, O’Neill WW, Ross MA, O’Neil BJ,
Angina). Circulation. 2002;106(14):1893–900. Raff GL.  A randomized controlled trial of multi-slice coronary
19. Backus BE, Six AJ, Kelder JC, et al. A prospective validation of computed tomography for evaluation of acute chest pain. J Am
the HEART score for chest pain patients at the emergency depart- Coll Cardiol. 2007;49(8):863–71.
ment. Int J Cardiol. 2013;168(3):2153–8. 34. Hoffmann U, Nagurney JT, Moselewski F, Pena A, Ferencik M,
20. Weinstock MB, Weingart S, Orth F, VanFossen D, Kaide C, Chae CU, Cury RC, Butler J, Abbara S, Brown DF, Manini A,
Anderson J, Newman DH.  Risk for clinically relevant adverse Nichols JH, Achenbach S, Brady TJ. Coronary multidetector com-
cardiac events in patients with chest pain at hospital admission. puted tomography in the assessment of patients with acute chest
JAMA Intern Med. 2015;175(7):1207–12. pain. Circulation. 2006;114(21):2251–60.
21. Pope JH, Aufderheide TP, Ruthazer R, et al. Missed diagnoses of 35. Gallagher MJ, Ross MA, Raff GL, Goldstein JA, O’Neill WW,
acute cardiac ischemia in the emergency department. N Engl J O’Neil B. The diagnostic accuracy of 64-slice computed tomogra-
Med. 2000;342(16):1163–70. phy coronary angiography compared with stress nuclear imaging
22. Goldberg A, Litt HI.  Evaluation of the patient with acute chest in emergency department low-risk chest pain patients. Ann Emerg
pain. Radiol Clin N Am. 2010;48(4):745–55. Med. 2007;49(2):125–36.
23. Lee TH, Rouan GW, Weisberg MC, et  al. Clinical character- 36. Hollander JE, Chang AM, Shofer FS, McCusker CM, Baxt WG,
istics and natural history of patients with acute myocardial Litt HI.  Coronary computed tomographic angiography for rapid
infarction sent home from the emergency room. Am J Cardiol. discharge of low-risk patients with potential acute coronary syn-
1987;60(4):219–24. dromes. Ann Emerg Med. 2009;53(3):295–304.
24. Fineberg HV, Scadden D, Goldman L.  Care of patients with a 37. Hoffmann U, Bamberg F, Chae CU, Nichols JH, Rogers IS,
low probability of acute myocardial infarction. Cost effectiveness Seneviratne SK, Truong QA, Cury RC, Abbara S, Shapiro MD,
of alternatives to coronary-care-unit admission. N Engl J Med. Moloo J, Butler J, Ferencik M, Lee H, Jang IK, Parry BA, Brown
1984;310(20):1301–7. DF, Udelson JE, Achenbach S, Brady TJ, Nagurney JT. Coronary
25. Galperin-Aizenberg M, Cook TS, Hollander JE, Litt HI. Cardiac computed tomography angiography for early triage of patients
CT angiography in the emergency department. AJR Am J with acute chest pain: the ROMICAT (Rule Out Myocardial
Roentgenol. 2015;204(3):463–74. Infarction using Computer Assisted Tomography) trial. J Am Coll
26. Hollander JE, Chang AM, Shofer FS, Collin MJ, Walsh KM, Cardiol. 2009;53(18):1642–50.
McCusker CM, Baxt WG, Litt HI. One-year outcomes following 38. Hoffmann U, Truong QA, Schoenfeld DA, Chou ET, Woodard
coronary computerized tomographic angiography for evaluation PK, Nagurney JT, Pope JH, Hauser TH, White CS, Weiner SG,
28  Coronary CT Angiography for Evaluation of Acute Coronary Syndrome in the Emergency Department 343

Kalanjian S, Mullins ME, Mikati I, Peacock WF, Zakroysky P, Society of Cardiovascular Computed Tomography Guidelines
Hayden D, Goehler A, Lee H, Gazelle GS, Wiviott SD, Fleg JL, Committee. J Cardiovasc Comput Tomogr. 2009;3(3):190–204.
Udelson JE. Coronary CT angiography versus standard evaluation 53. Dewey M, Hoffmann H, Hamm B. Multislice CT coronary angi-
in acute chest pain. N Engl J Med. 2012;367(4):299–308. ography: effect of sublingual nitroglycerine on the diameter of
39. Litt HI, Gatsonis C, Snyder B, Singh H, Miller CD, Entrikin DW, coronary arteries. Rofo. 2006;178(6):600–4.
Leaming JM, Gavin LJ, Pacella CB, Hollander JE. CT angiogra- 54. Decramer I, Vanhoenacker PK, Sarno G, Van Hoe L, Bladt
phy for safe discharge of patients with possible acute coronary O, Wijns W, Parizel PM.  Effects of sublingual nitroglycerin
syndromes. N Engl J Med. 2012;366(15):1393–403. on coronary lumen diameter and number of visualized septal
40. Goldstein JA, Chinnaiyan KM, Abidov A, Achenbach S, Berman branches on 64-MDCT angiography. AJR Am J Roentgenol.
DS, Hayes SW, Hoffmann U, Lesser JR, Mikati IA, O'Neil BJ, 2008;190(1):219–25.
Shaw LJ, Shen MY, Valeti US, Raff GL.  The CT-STAT (coro- 55. Hausleiter J, Meyer T, Hermann F, Hadamitzky M, Krebs M,
nary computed tomographic angiography for systematic triage of Gerber TC, McCollough C, Martinoff S, Kastrati A, Schomig A,
acute chest pain patients to treatment) trial. J Am Coll Cardiol. Achenbach S.  Estimated radiation dose associated with cardiac
2011;58(14):1414–22. CT angiography. JAMA. 2009;301(5):500–7.
41. Cook TS, Galperin-Aizenberg M, Litt HI. Coronary and cardiac 56. Earls JP, Berman EL, Urban BA, Curry CA, Lane JL, Jennings
computed tomography in the emergency room: current status and RS, McCulloch CC, Hsieh J, Londt JH. Prospectively gated trans-
future directions. J Thorac Imaging. 2013;28(4):204–16. verse coronary CT angiography versus retrospectively gated heli-
42. Cury RC, Feuchtner GM, Batlle JC, Pena CS, Janowitz W, Katzen cal technique: improved image quality and reduced radiation dose.
BT, Ziffer JA. Triage of patients presenting with chest pain to the Radiology. 2008;246(3):742–53.
emergency department: implementation of coronary CT angiog- 57. Menke J, Unterberg-Buchwald C, Staab W, Sohns JM, Seif
raphy in a large urban health care system. AJR Am J Roentgenol. Amir Hosseini A, Schwarz A. Head-to-head comparison of pro-
2013;200(1):57–65. spectively triggered vs retrospectively gated coronary computed
43. Hoffmann U, Pena AJ, Moselewski F, Ferencik M, Abbara S, Cury tomography angiography: meta-analysis of diagnostic accuracy,
RC, Chae CU, Nagurney JT. MDCT in early triage of patients with image quality, and radiation dose. Am Heart J. 2013;165(2):154–
acute chest pain. AJR Am J Roentgenol. 2006;187(5):1240–7. 63. e153.
44. Salerno M, Bourque JM, Beller GA.  Coronary angiographic 58. den Harder AM, Willemink MJ, de Jong PA, Schilham AM, Rajiah
evaluation of low-risk chest pain in the emergency department P, Takx RA, Leiner T. New horizons in cardiac CT. Clin Radiol.
CT-STAT, or maybe not quite that fast? J Am Coll Cardiol. 2016;71(8):758–67.
2011;58(14):1423–5. 59. Lau GT, Ridley LJ, Schieb MC, Brieger DB, Freedman SB, Wong
45. Rubinshtein R, Halon DA, Gaspar T, Jaffe R, Goldstein J, LA, Lo SK, Kritharides L.  Coronary artery stenoses: detection
Karkabi B, Flugelman MY, Kogan A, Shapira R, Peled N, Lewis with calcium scoring, CT angiography, and both methods com-
BS.  Impact of 64-slice cardiac computed tomographic angiog- bined. Radiology. 2005;235(2):415–22.
raphy on clinical decision-making in emergency department 60. Chang AM, Le J, Matsuura AC, Litt HI, Hollander JE.  Does
patients with chest pain of possible myocardial ischemic origin. coronary artery calcium scoring add to the predictive value of
Am J Cardiol. 2007;100(10):1522–6. coronary computed tomography angiography for adverse cardio-
46. Truong QA, Hayden D, Woodard PK, Kirby R, Chou ET, vascular events in low-risk chest pain patients? Acad Emerg Med.
Nagurney JT, Wiviott SD, Fleg JL, Schoenfeld DA, Udelson JE, 2011;18(10):1065–71.
Hoffmann U.  Sex differences in the effectiveness of early coro- 61. Boll DT, Merkle EM, Paulson EK, Mirza RA, Fleiter TR. Calcified
nary computed tomographic angiography compared with standard vascular plaque specimens: assessment with cardiac dual-energy
emergency department evaluation for acute chest pain: the rule-­ multidetector CT in anthropomorphically moving heart phan-
out myocardial infarction with Computer-Assisted Tomography tom. Radiology. 2008;249(1):119–26. https://doi.org/10.1148/
(ROMICAT)-II Trial. Circulation. 2013;127(25):2494–502. radiol.2483071576.
47. Eckert J, Schmidt M, Magedanz A, Voigtlander T, Schmermund 62. Boll DT, Merkle EM, Paulson EK, Fleiter TR.  Coronary stent
A. Coronary CT angiography in managing atherosclerosis. Int J patency: dual-energy multidetector CT assessment in a pilot study
Mol Sci. 2015;16(2):3740–56. with anthropomorphic phantom. Radiology. 2008;247(3):687–95.
48. Fazel P, Peterman MA, Schussler JM.  Three-year outcomes 63. Wnorowski AM, Halpern EJ.  Diagnostic yield of triple-­
and cost analysis in patients receiving 64-slice computed tomo- rule-­out CT in an emergency setting. AJR Am J Roentgenol.
graphic coronary angiography for chest pain. Am J Cardiol. 2016;207(2):295–301.
2009;104(4):498–500. 64. Ayaram D, Bellolio MF, Murad MH, Laack TA, Sadosty AT,
49. Romero J, Husain SA, Holmes AA, Kelesidis I, Chavez P, Erwin PJ, Hollander JE, Montori VM, Stiell IG, Hess EP. Triple
Mojadidi MK, Levsky JM, Wever-Pinzon O, Taub C, Makani rule-out computed tomographic angiography for chest pain: a
H, Travin MI, Pina IL, Garcia MJ. Non-invasive assessment of diagnostic systematic review and meta-analysis. Acad Emerg
low risk acute chest pain in the emergency department: a com- Med. 2013;20(9):861–71.
parative meta-­ analysis of prospective studies. Int J Cardiol. 65. Burris AC 2nd, Boura JA, Raff GL, Chinnaiyan KM.  Triple
2015;187:565–80. rule out versus coronary CT angiography in patients with acute
50. Beigel R, Oieru D, Goitein O, Chouraqui P, Konen E, Shamiss chest pain: results from the ACIC consortium. JACC Cardiovasc
A, Hod H, Or J, Matetzky S. Usefulness of routine use of mul- Imaging. 2015;8(7):817–25.
tidetector coronary computed tomography in the “fast track” 66. Takakuwa KM, Halpern EJ. Evaluation of a “triple rule-out” coro-
evaluation of patients with acute chest pain. Am J Cardiol. nary CT angiography protocol: use of 64-Section CT in low-to-­
2009;103(11):1481–6. moderate risk emergency department patients suspected of having
51. Shim SS, Kim Y, Lim SM.  Improvement of image quality with acute coronary syndrome. Radiology. 2008;248(2):438–46.
beta-blocker premedication on ECG-gated 16-MDCT coronary 67. Pohle K, Ropers D, Maffert R, Geitner P, Moshage W, Regenfus
angiography. AJR Am J Roentgenol. 2005;184(2):649–54. M, Kusus M, Daniel WG, Achenbach S. Coronary calcifications
52. Abbara S, Arbab-Zadeh A, Callister TQ, Desai MY, Mamuya W, in young patients with first, unheralded myocardial infarction: a
Thomson L, Weigold WG. SCCT guidelines for performance of risk factor matched analysis by electron beam tomography. Heart.
coronary computed tomographic angiography: a report of the 2003;89(6):625–8.
344 N. J. Lee and H. Litt

68. Schmermund A, Schwartz RS, Adamzik M, Sangiorgi G, Pfeifer cholesterol-lowering therapy. Arterioscler Thromb Vasc Biol.
EA, Rumberger JA, Burke AP, Farb A, Virmani R. Coronary ath- 2004;24(7):1272–7.
erosclerosis in unheralded sudden coronary death under age 50: 81. Herzog C, Britten M, Balzer JO, Mack MG, Zangos S, Ackermann
histo-pathologic comparison with ‘healthy’ subjects dying out of H, Schaechinger V, Schaller S, Flohr T, Vogl TJ.  Multidetector-­
hospital. Atherosclerosis. 2001;155(2):499–508. row cardiac CT: diagnostic value of calcium scoring and CT coro-
69. Budoff MJ, Achenbach S, Blumenthal RS, Carr JJ, Goldin JG, nary angiography in patients with symptomatic, but atypical, chest
Greenland P, Guerci AD, Lima JA, Rader DJ, Rubin GD, Shaw pain. Eur Radiol. 2004;14(2):169–77.
LJ, Wiegers SE. Assessment of coronary artery disease by cardiac 82. Miller CD, Litt HI, Askew K, Entrikin D, Carr JJ, Chang AM,
computed tomography: a scientific statement from the American Kilkenny J, Weisenthal B, Hollander JE. Implications of 25% to
Heart Association Committee on cardiovascular imaging and 50% coronary stenosis with cardiac computed tomographic angi-
intervention, council on cardiovascular radiology and interven- ography in ED patients. Am J Emerg Med. 2012;30(4):597–605.
tion, and committee on cardiac imaging, council on clinical cardi- 83. Wu E, Ortiz JT, Tejedor P, Lee DC, Bucciarelli-Ducci C, Kansal P,
ology. Circulation. 2006;114(16):1761–91. Carr JC, Holly TA, Lloyd-Jones D, Klocke FJ, Bonow RO. Infarct
70. O’Rourke RA, Brundage BH, Froelicher VF, Greenland P, Grundy size by contrast enhanced cardiac magnetic resonance is a stron-
SM, Hachamovitch R, Pohost GM, Shaw LJ, Weintraub WS, ger predictor of outcomes than left ventricular ejection fraction
Winters WL Jr. American College of Cardiology/American Heart or end-systolic volume index: prospective cohort study. Heart.
Association expert consensus document on electron-beam com- 2008;94(6):730–6.
puted tomography for the diagnosis and prognosis of coronary 84. Miller TD, Christian TF, Hopfenspirger MR, Hodge DO, Gersh
artery disease. J Am Coll Cardiol. 2000;36(1):326–40. BJ, Gibbons RJ.  Infarct size after acute myocardial infarction
71. Nabi F, Chang SM, Pratt CM, Paranilam J, Peterson LE, measured by quantitative tomographic 99mTc sestamibi imaging
Frias ME, Mahmarian JJ.  Coronary artery calcium scoring in predicts subsequent mortality. Circulation. 1995;92(3):334–41.
the emergency department: identifying which patients with 85. Elezi S, Kastrati A, Neumann FJ, Hadamitzky M, Dirschinger J,
chest pain can be safely discharged home. Ann Emerg Med. Schomig A.  Vessel size and long-term outcome after coronary
2010;56(3):220–9. stent placement. Circulation. 1998;98(18):1875–80.
72. Marwan M, Ropers D, Pflederer T, Daniel WG, Achenbach 86. Togni M, Eber S, Widmer J, Billinger M, Wenaweser P, Cook S,
S.  Clinical characteristics of patients with obstructive coronary Vogel R, Seiler C, Eberli FR, Maier W, Corti R, Roffi M, Luscher
lesions in the absence of coronary calcification: an evaluation by TF, Garachemani A, Hess OM, Wandel S, Meier B, Juni P,
coronary CT angiography. Heart. 2009;95(13):1056–60. https:// Windecker S. Impact of vessel size on outcome after implantation
doi.org/10.1136/hrt.2008.153353. of sirolimus-eluting and paclitaxel-eluting stents: a subgroup anal-
73. Davies MJ. The composition of coronary-artery plaques. N Engl J ysis of the SIRTAX trial. J Am Coll Cardiol. 2007;50(12):1123–31.
Med. 1997;336(18):1312–4. 87. Schunkert H, Harrell L, Palacios IF. Implications of small refer-
74. Ergun E, Kosar P, Ozturk C, Basbay E, Koc F, Kosar U. Prevalence ence vessel diameter in patients undergoing percutaneous coro-
and extent of coronary artery disease determined by 64-slice CTA nary revascularization. J Am Coll Cardiol. 1999;34(1):40–8.
in patients with zero coronary calcium score. Int J Cardiovasc 88. Herzog C, Zwerner PL, Doll JR, Nielsen CD, Nguyen SA,
Imaging. 2011;27(3):451–8. Savino G, Vogl TJ, Costello P, Schoepf UJ.  Significant coro-
75. Sarwar A, Shaw LJ, Shapiro MD, Blankstein R, Hoffmann U, nary artery stenosis: comparison on per-patient and per-vessel
Cury RC, Abbara S, Brady TJ, Budoff MJ, Blumenthal RS, Nasir or per-segment basis at 64-section CT angiography. Radiology.
K. Diagnostic and prognostic value of absence of coronary artery 2007;244(1):112–20.
calcification. JACC Cardiovasc Imaging. 2009;2(6):675–88. 89. Roe MT, Harrington RA, Prosper DM, Pieper KS, Bhatt DL,
76. Greenland P, LaBree L, Azen SP, Doherty TM, Detrano Lincoff AM, Simoons ML, Akkerhuis M, Ohman EM, Kitt MM,
RC. Coronary artery calcium score combined with Framingham Vahanian A, Ruzyllo W, Karsch K, Califf RM, Topol EJ. Clinical
score for risk prediction in asymptomatic individuals. JAMA. and therapeutic profile of patients presenting with acute coronary
2004;291(2):210–5. syndromes who do not have significant coronary artery disease.
77. Raggi P, Cooil B, Ratti C, Callister TQ, Budoff M.  Progression The platelet glycoprotein IIb/IIIa in unstable angina: receptor sup-
of coronary artery calcium and occurrence of myocardial infarc- pression using integrilin therapy (PURSUIT) trial investigators.
tion in patients with and without diabetes mellitus. Hypertension. Circulation. 2000;102(10):1101–6.
2005;46(1):238–43. 90. van Velzen JE, de Graaf FR, Kroft LJ, de Roos A, Reiber
78. Greenland P, Bonow RO, Brundage BH, Budoff MJ, Eisenberg JH, Bax JJ, Jukema JW, Schuijf JD, Schalij MJ, van der Wall
MJ, Grundy SM, Lauer MS, Post WS, Raggi P, Redberg RF, EE.  Performance and efficacy of 320-row computed tomogra-
Rodgers GP, Shaw LJ, Taylor AJ, Weintraub WS.  ACCF/AHA phy coronary angiography in patients presenting with acute chest
2007 clinical expert consensus document on coronary artery cal- pain: results from a clinical registry. Int J Cardiovasc Imaging.
cium scoring by computed tomography in global cardiovascular 2012;28(4):865–76.
risk assessment and in evaluation of patients with chest pain: a 91. Ahmed W, Schlett CL, Uthamalingam S, Truong QA, Koenig W,
report of the American College of Cardiology Foundation Clinical Rogers IS, Blankstein R, Nagurney JT, Tawakol A, Januzzi JL,
Expert Consensus Task Force (ACCF/AHA Writing Committee Hoffmann U. Single resting hsTnT level predicts abnormal myo-
to Update the 2000 Expert Consensus Document on Electron cardial stress test in acute chest pain patients with normal initial
Beam Computed Tomography) developed in collaboration with standard troponin. JACC Cardiovasc Imaging. 2013;6(1):72–82.
the Society of Atherosclerosis Imaging and Prevention and the 92. Puchner SB, Liu T, Mayrhofer T, Truong QA, Lee H, Fleg JL,
Society of Cardiovascular Computed Tomography. J Am Coll Nagurney JT, Udelson JE, Hoffmann U, Ferencik M.  High-risk
Cardiol. 2007;49(3):378–402. plaque detected on coronary CT angiography predicts acute
79. Budoff MJ, Hokanson JE, Nasir K, Shaw LJ, Kinney GL, Chow coronary syndromes independent of significant stenosis in acute
D, Demoss D, Nuguri V, Nabavi V, Ratakonda R, Berman DS, chest pain: results from the ROMICAT-II trial. J Am Coll Cardiol.
Raggi P. Progression of coronary artery calcium predicts all-cause 2014;64(7):684–92.
mortality. JACC Cardiovasc Imaging. 2010;3(12):1229–36. 93. Stary HC, Chandler AB, Glagov S, Guyton JR, Insull W Jr,
80. Raggi P, Callister TQ, Shaw LJ.  Progression of coronary artery Rosenfeld ME, Schaffer SA, Schwartz CJ, Wagner WD, Wissler
calcium and risk of first myocardial infarction in patients ­receiving RW. A definition of initial, fatty streak, and intermediate lesions of
28  Coronary CT Angiography for Evaluation of Acute Coronary Syndrome in the Emergency Department 345

atherosclerosis. A report from the committee on vascular lesions lesions in acute coronary syndromes using coronary dual-source
of the council on arteriosclerosis, American Heart Association. CT angiography. Atherosclerosis. 2010;211(2):437–44.
Arterioscler Thromb. 1994;14(5):840–56. 110. Motoyama S, Sarai M, Harigaya H, Anno H, Inoue K, Hara T,
94. Stary HC, Chandler AB, Glagov S, Guyton JR, Insull W Jr, Naruse H, Ishii J, Hishida H, Wong ND, Virmani R, Kondo T,
Rosenfeld ME, Schaffer SA, Schwartz CJ, Wagner WD, Wissler Ozaki Y, Narula J.  Computed tomographic angiography charac-
RW. A definition of initial, fatty streak, and intermediate lesions of teristics of atherosclerotic plaques subsequently resulting in acute
atherosclerosis. A report from the Committee on Vascular Lesions coronary syndrome. J Am Coll Cardiol. 2009;54(1):49–57.
of the Council on Arteriosclerosis, American Heart Association. 111. Hoffmann U, Moselewski F, Nieman K, Jang IK, Ferencik M,
Circulation. 1994;89(5):2462–78. Rahman AM, Cury RC, Abbara S, Joneidi-Jafari H, Achenbach
95. Glagov S, Weisenberg E, Zarins CK, Stankunavicius R, Kolettis S, Brady TJ. Noninvasive assessment of plaque morphology and
GJ.  Compensatory enlargement of human atherosclerotic coro- composition in culprit and stable lesions in acute coronary syn-
nary arteries. N Engl J Med. 1987;316(22):1371–5. drome and stable lesions in stable angina by multidetector com-
96. Burke AP, Farb A, Malcom GT, Liang YH, Smialek J, Virmani puted tomography. J Am Coll Cardiol. 2006;47(8):1655–62.
R.  Coronary risk factors and plaque morphology in men 112. Gauss S, Achenbach S, Pflederer T, Schuhback A, Daniel WG,
with coronary disease who died suddenly. N Engl J Med. Marwan M. Assessment of coronary artery remodelling by dual-­
1997;336(18):1276–82. source CT: a head-to-head comparison with intravascular ultra-
97. Burke AP, Farb A, Malcom GT, Liang Y, Smialek J, Virmani sound. Heart. 2011;97(12):991–7.
R.  Effect of risk factors on the mechanism of acute throm- 113. Motoyama S, Kondo T, Sarai M, Sugiura A, Harigaya H, Sato
bosis and sudden coronary death in women. Circulation. T, Inoue K, Okumura M, Ishii J, Anno H, Virmani R, Ozaki Y,
1998;97(21):2110–6. Hishida H, Narula J. Multislice computed tomographic character-
98. Farb A, Tang AL, Burke AP, Sessums L, Liang Y, Virmani istics of coronary lesions in acute coronary syndromes. J Am Coll
R. Sudden coronary death. Frequency of active coronary lesions, Cardiol. 2007;50(4):319–26.
inactive coronary lesions, and myocardial infarction. Circulation. 114. Kashiwagi M, Tanaka A, Kitabata H, Tsujioka H, Kataiwa H,
1995;92(7):1701–9. Komukai K, Tanimoto T, Takemoto K, Takarada S, Kubo T, Hirata
99. Virmani R, Kolodgie FD, Burke AP, Farb A, Schwartz SM. Lessons K, Nakamura N, Mizukoshi M, Imanishi T, Akasaka T. Feasibility
from sudden coronary death: a comprehensive morphological of noninvasive assessment of thin-cap fibroatheroma by mul-
classification scheme for atherosclerotic lesions. Arterioscler tidetector computed tomography. JACC Cardiovasc Imaging.
Thromb Vasc Biol. 2000;20(5):1262–75. 2009;2(12):1412–9.
100. Kolodgie FD, Burke AP, Farb A, Gold HK, Yuan J, Narula J, Finn 115. Kitagawa T, Yamamoto H, Horiguchi J, Ohhashi N, Tadehara F,
AV, Virmani R. The thin-cap fibroatheroma: a type of vulnerable Shokawa T, Dohi Y, Kunita E, Utsunomiya H, Kohno N, Kihara
plaque: the major precursor lesion to acute coronary syndromes. Y. Characterization of noncalcified coronary plaques and identifi-
Curr Opin Cardiol. 2001;16(5):285–92. cation of culprit lesions in patients with acute coronary syndrome
101. Burke AP, Taylor A, Farb A, Malcom GT, Virmani R. Coronary by 64-slice computed tomography. JACC Cardiovasc Imaging.
calcification: insights from sudden coronary death victims. Z 2009;2(2):153–60.
Kardiol. 2000;89:Suppl 2:49–53. 116. Obaid DR, Calvert PA, Gopalan D, Parker RA, Hoole SP, West
102. Wallentin L. Non-ST-elevation acute coronary syndrome: fuel for NE, Goddard M, Rudd JH, Bennett MR. Atherosclerotic plaque
the invasive strategy. Lancet. 2002;360(9335):738–9. composition and classification identified by coronary computed
103. Burke AP, Kolodgie FD, Farb A, Weber DK, Malcom GT, Smialek tomography: assessment of computed tomography-generated
J, Virmani R. Healed plaque ruptures and sudden coronary death: plaque maps compared with virtual histology intravascular ultra-
evidence that subclinical rupture has a role in plaque progression. sound and histology. Circ Cardiovasc Imaging. 2013;6(5):655–64.
Circulation. 2001;103(7):934–40. https://doi.org/10.1161/CIRCIMAGING.112.000250.
104. Nakamura M, Nishikawa H, Mukai S, Setsuda M, Nakajima K, 117. Pundziute G, Schuijf JD, Jukema JW, Decramer I, Sarno G,
Tamada H, Suzuki H, Ohnishi T, Kakuta Y, Nakano T, Yeung Vanhoenacker PK, Reiber JH, Schalij MJ, Wijns W, Bax
AC. Impact of coronary artery remodeling on clinical presentation JJ.  Head-to-head comparison of coronary plaque evaluation
of coronary artery disease: an intravascular ultrasound study. J Am between multislice computed tomography and intravascular ultra-
Coll Cardiol. 2001;37(1):63–9. sound radiofrequency data analysis. JACC Cardiovasc Interv.
105. Maehara A, Mintz GS, Bui AB, Walter OR, Castagna MT, Canos 2008;1(2):176–82.
D, Pichard AD, Satler LF, Waksman R, Suddath WO, Laird JR Jr, 118. Hong MK, Mintz GS, Lee CW, Suh J, Kim JH, Park DW, Lee SW,
Kent KM, Weissman NJ. Morphologic and angiographic features Kim YH, Cheong SS, Kim JJ, Park SW, Park SJ. Comparison of
of coronary plaque rupture detected by intravascular ultrasound. J virtual histology to intravascular ultrasound of culprit coronary
Am Coll Cardiol. 2002;40(5):904–10. lesions in acute coronary syndrome and target coronary lesions in
106. Ambrose JA, Tannenbaum MA, Alexopoulos D, Hjemdahl-­Monsen stable angina pectoris. Am J Cardiol. 2007;100(6):953–9.
CE, Leavy J, Weiss M, Borrico S, Gorlin R, Fuster V. Angiographic 119. Pundziute G, Schuijf JD, Jukema JW, Decramer I, Sarno G,
progression of coronary artery disease and the development of Vanhoenacker PK, Boersma E, Reiber JH, Schalij MJ, Wijns
myocardial infarction. J Am Coll Cardiol. 1988;12(1):56–62. W, Bax JJ.  Evaluation of plaque characteristics in acute coro-
107. Giroud D, Li JM, Urban P, Meier B, Rutishauer W.  Relation nary syndromes: non-invasive assessment with multi-slice
of the site of acute myocardial infarction to the most severe computed tomography and invasive evaluation with intravas-
coronary arterial stenosis at prior angiography. Am J Cardiol. cular ultrasound radiofrequency data analysis. Eur Heart J.
1992;69(8):729–32. 2008;29(19):2373–81.
108. Stone GW, Maehara A, Lansky AJ, de Bruyne B, Cristea E, 120. Ruzsics B, Lee H, Zwerner PL, Gebregziabher M, Costello P,
Mintz GS, Mehran R, McPherson J, Farhat N, Marso SP, Parise Schoepf UJ. Dual-energy CT of the heart for diagnosing coronary
H, Templin B, White R, Zhang Z, Serruys PW.  A prospective artery stenosis and myocardial ischemia-initial experience. Eur
natural-history study of coronary atherosclerosis. N Engl J Med. Radiol. 2008;18(11):2414–24.
2011;364(3):226–35. 121. Barreto M, Schoenhagen P, Nair A, Amatangelo S, Milite M,
109. Pflederer T, Marwan M, Schepis T, Ropers D, Seltmann M, Obuchowski NA, Lieber ML, Halliburton SS.  Potential of dual-­
Muschiol G, Daniel WG, Achenbach S. Characterization of culprit energy computed tomography to characterize atherosclerotic
346 N. J. Lee and H. Litt

plaque: ex vivo assessment of human coronary arteries in compari- tomography angiography in suspected coronary artery disease: the
son to histology. J Cardiovasc Comput Tomogr. 2008;2(4):234–42. NXT trial (analysis of coronary blood flow using CT angiography:
122. Yamamoto H, Kitagawa T, Ohashi N, Utsunomiya H, Kunita E, next steps). J Am Coll Cardiol. 2014;63(12):1145–55.
Oka T, Urabe Y, Tsushima H, Awai K, Kihara Y.  Noncalcified 133. Renker M, Schoepf UJ, Wang R, Meinel FG, Rier JD, Bayer
atherosclerotic lesions with vulnerable characteristics detected RR 2nd, Mollmann H, Hamm CW, Steinberg DH, Baumann
by coronary CT angiography and future coronary events. J S. Comparison of diagnostic value of a novel noninvasive coro-
Cardiovasc Comput Tomogr. 2013;7(3):192–9. nary computed tomography angiography method versus standard
123. Versteylen MO, Kietselaer BL, Dagnelie PC, Joosen IA, Dedic A, coronary angiography for assessing fractional flow reserve. Am J
Raaijmakers RH, Wildberger JE, Nieman K, Crijns HJ, Niessen Cardiol. 2014;114(9):1303–8.
WJ, Daemen MJ, Hofstra L.  Additive value of semiautomated 134. Coenen A, Lubbers MM, Kurata A, Kono A, Dedic A, Chelu RG,
quantification of coronary artery disease using cardiac computed Dijkshoorn ML, Gijsen FJ, Ouhlous M, van Geuns RJ, Nieman
tomographic angiography to predict future acute coronary syn- K. Fractional flow reserve computed from noninvasive CT angi-
drome. J Am Coll Cardiol. 2013;61(22):2296–305. ography data: diagnostic performance of an on-site clinician-­
124. Ferencik M, Schlett CL, Ghoshhajra BB, Kriegel MF, Joshi operated computational fluid dynamics algorithm. Radiology.
SB, Maurovich-Horvat P, Rogers IS, Banerji D, Bamberg F, 2015;274(3):674–83.
Truong QA, Brady TJ, Nagurney JT, Hoffmann U. A computed 135. Caruso D, Eid M, Schoepf UJ, Jin KN, Varga-Szemes A, Tesche C,
tomography-­based coronary lesion score to predict acute coronary Mangold S, Spandorfer A, Laghi A, De Cecco CN. Dynamic CT
syndrome among patients with acute chest pain and significant myocardial perfusion imaging. Eur J Radiol. 2016;85(10):1893–9.
coronary stenosis on coronary computed tomographic angiogram. 136. De Cecco CN, Varga-Szemes A, Meinel FG, Renker M, Schoepf
Am J Cardiol. 2012;110(2):183–9. UJ. Beyond stenosis detection: computed tomography approaches
125. Ferencik M, Mayrhofer T, Puchner SB, Lu MT, Maurovich-Horvat for determining the functional relevance of coronary artery dis-
P, Liu T, Ghemigian K, Kitslaar P, Broersen A, Bamberg F, Truong ease. Radiol Clin North Am. 2015;53(2):317–34.
QA, Schlett CL, Hoffmann U. Computed tomography-based high-­ 137. Varga-Szemes A, Meinel FG, De Cecco CN, Fuller SR, Bayer RR
risk coronary plaque score to predict acute coronary syndrome 2nd, Schoepf UJ.  CT myocardial perfusion imaging. AJR Am J
among patients with acute chest pain--results from the ROMICAT Roentgenol. 2015;204(3):487–97.
II trial. J Cardiovasc Comput Tomogr. 2015;9(6):538–45. 138. Cury RC, Magalhaes TA, Borges AC, Shiozaki AA, Lemos PA,
126. Puchner SB, Lu MT, Mayrhofer T, Liu T, Pursnani A, Ghoshhajra Junior JS, Meneghetti JC, Rochitte CE. Dipyridamole stress and
BB, Truong QA, Wiviott SD, Fleg JL, Hoffmann U, Ferencik rest myocardial perfusion by 64-detector row computed tomog-
M.  High-risk coronary plaque at coronary CT angiography is raphy in patients with suspected coronary artery disease. Am J
associated with nonalcoholic fatty liver disease, independent of Cardiol. 2010;106(3):310–5.
coronary plaque and stenosis burden: results from the ROMICAT 139. Techasith T, Cury RC.  Stress myocardial CT perfusion: an
II trial. Radiology. 2015;274(3):693–701. update and future perspective. JACC Cardiovasc Imaging.
127. Dedic A, Kurata A, Lubbers M, Meijboom WB, van Dalen BM, 2011;4(8):905–16.
Snelder S, Korbee R, Moelker A, Ouhlous M, van Domburg R, 140. Cury RC, Abbara S, Achenbach S, Agatston A, Berman DS,
de Feijter PJ, Nieman K.  Prognostic implications of non-culprit Budoff MJ, Dill KE, Jacobs JE, Maroules CD, Rubin GD, Rybicki
plaques in acute coronary syndrome: non-invasive assessment FJ, Schoepf UJ, Shaw LJ, Stillman AE, White CS, Woodard PK,
with coronary CT angiography. Eur Heart J Cardiovasc Imaging. Leipsic JA.  CAD-RADS(TM) coronary artery disease  - report-
2014;15(11):1231–7. ing and data system. An expert consensus document of the
128. Pijls NH, Sels JW. Functional measurement of coronary stenosis. Society of Cardiovascular Computed Tomography (SCCT), the
J Am Coll Cardiol. 2012;59(12):1045–57. American College of Radiology (ACR) and the North American
129. Baumann S, Renker M, Hetjens S, Fuller SR, Becher T, Lossnitzer Society for Cardiovascular Imaging (NASCI). Endorsed by the
D, Lehmann R, Akin I, Borggrefe M, Lang S, Wichmann JL, American College of Cardiology. J Cardiovasc Comput Tomogr.
Schoepf UJ.  Comparison of coronary computed tomography 2016;10(4):269–81.
angiography-­derived vs invasive fractional flow reserve assess- 141. Cury RC, Abbara S, Achenbach S, Agatston A, Berman DS,
ment: meta-analysis with subgroup evaluation of intermedi- Budoff MJ, Dill KE, Jacobs JE, Maroules CD, Rubin GD,
ate stenosis. Acad Radiol. 2016; https://doi.org/10.1016/j. Rybicki FJ, Schoepf UJ, Shaw LJ, Stillman AE, White CS,
acra.2016.07.007. Woodard PK, Leipsic JA. CAD-RADS: coronary artery disease -
130. Min JK, Leipsic J, Pencina MJ, Berman DS, Koo BK, van reporting and data system.: an expert consensus document of the
Mieghem C, Erglis A, Lin FY, Dunning AM, Apruzzese P, Budoff Society of Cardiovascular Computed Tomography (SCCT), the
MJ, Cole JH, Jaffer FA, Leon MB, Malpeso J, Mancini GB, Park American College of Radiology (ACR) and the North American
SJ, Schwartz RS, Shaw LJ, Mauri L.  Diagnostic accuracy of Society for Cardiovascular Imaging (NASCI). Endorsed by the
fractional flow reserve from anatomic CT angiography. JAMA. American College of Cardiology. J Am Coll Radiol. 2016;13(12
2012;308(12):1237–45. Pt A):1458–66.
131. Koo BK, Erglis A, Doh JH, Daniels DV, Jegere S, Kim HS, 142. Cury RC, Abbara S, Achenbach S, Agatston A, Berman DS,
Dunning A, DeFrance T, Lansky A, Leipsic J, Min JK. Diagnosis Budoff MJ, Dill KE, Jacobs JE, Maroules CD, Rubin GD, Rybicki
of ischemia-causing coronary stenoses by noninvasive frac- FJ, Schoepf UJ, Shaw LJ, Stillman AE, White CS, Woodard PK,
tional flow reserve computed from coronary computed tomo- Leipsic JA. Coronary artery disease - reporting and data system
graphic angiograms. Results from the prospective multicenter (CAD-RADS): an expert consensus document of SCCT, ACR
DISCOVER-FLOW (diagnosis of ischemia-causing stenoses and NASCI: endorsed by the ACC.  JACC Cardiovasc Imaging.
obtained via noninvasive fractional flow reserve) study. J Am Coll 2016;9(9):1099–113.
Cardiol. 2011;58(19):1989–97. 143. Mohlenkamp S, Hort W, Ge J, Erbel R.  Update on myocardial
132. Norgaard BL, Leipsic J, Gaur S, Seneviratne S, Ko BS, Ito H, bridging. Circulation. 2002;106(20):2616–22.
Jensen JM, Mauri L, De Bruyne B, Bezerra H, Osawa K, Marwan 144. Iversen S, Hake U, Mayer E, Erbel R, Diefenbach C, Oelert
M, Naber C, Erglis A, Park SJ, Christiansen EH, Kaltoft A, Lassen H.  Surgical treatment of myocardial bridging causing coro-
JF, Botker HE, Achenbach S. Diagnostic performance of nonin- nary artery obstruction. Scand J Thorac Cardiovasc Surg.
vasive fractional flow reserve derived from coronary computed 1992;26(2):107–11.
28  Coronary CT Angiography for Evaluation of Acute Coronary Syndrome in the Emergency Department 347

145. Noble J, Bourassa MG, Petitclerc R, Dyrda I.  Myocardial

160. Ferencik M, Liu T, Mayrhofer T, Puchner SB, Lu MT, Maurovich-­
bridging and milking effect of the left anterior descending Horvat P, Pope JH, Truong QA, Udelson JE, Peacock WF, White
coronary artery: normal variant or obstruction? Am J Cardiol. CS, Woodard PK, Fleg JL, Nagurney JT, Januzzi JL, Hoffmann
1976;37(7):993–9. U. hs-Troponin I followed by CT angiography improves acute
146. Lee SS, Wu TL. The role of the mural coronary artery in preven- coronary syndrome risk stratification accuracy and work-up in
tion of coronary atherosclerosis. Arch Pathol. 1972;93(1):32–5. acute chest pain patients: Results from ROMICAT II trial. JACC
147. Endo M, Lee YW, Hayashi H, Wada J. Angiographic evidence of Cardiovasc Imaging. 2015;8(11):1272–81.
myocardial squeezing accompanying tachyarrhythmia as a pos- 161. Hilton TC, Fulmer H, Abuan T, Thompson RC, Stowers
sible cause of myocardial infarction. Chest. 1978;73(3):431–3. SA.  Ninety-day follow-up of patients in the emergency depart-
148. Tauth J, Sullebarger T.  Myocardial infarction associated with ment with chest pain who undergo initial single-photon emission
myocardial bridging: case history and review of the literature. computed tomographic perfusion scintigraphy with technetium
Catheter Cardiovasc Diagn. 1997;40(4):364–7. 99m-labeled sestamibi. J Nucl Cardiol. 1996;3(4):308–11.
149. Ferreira AG Jr, Trotter SE, Konig B Jr, Decourt LV, Fox K, Olsen 162. Conti A, Gallini C, Costanzo E, Ferri P, Matteini M, Paladini B,
EG.  Myocardial bridges: morphological and functional aspects. Francois C, Grifoni S, Migliorini A, Antoniucci D, Pieroni C,
Br Heart J. 1991;66(5):364–7. Berni G.  Early detection of myocardial ischaemia in the emer-
150. Roberts WC, Dicicco BS, Waller BF, Kishel JC, McManus BM, gency department by rest or exercise (99m)Tc tracer myocardial
Dawson SL, Hunsaker JC 3rd, Luke JL. Origin of the left main SPET in patients with chest pain and non-diagnostic ECG. Eur J
from the right coronary artery or from the right aortic sinus with Nucl Med. 2001;28(12):1806–10.
intramyocardial tunneling to the left side of the heart via the ven- 163. Schaeffer MW, Brennan TD, Hughes JA, Gibler WB, Gerson
tricular septum. The case against clinical significance of myo- MC.  Resting radionuclide myocardial perfusion imaging in a
cardial bridge or coronary tunnel. Am Heart J. 1982;104(2 Pt chest pain center including an overnight delayed image acquisi-
1):303–5. tion protocol. J Nucl Med Technol. 2007;35(4):242–5.
151. Yamanaka O, Hobbs RE. Coronary artery anomalies in 126,595 164. Kontos MC, Haney A, Ornato JP, Jesse RL, Tatum JL. Value of
patients undergoing coronary arteriography. Catheter Cardiovasc simultaneous functional assessment in association with acute rest
Diagn. 1990;21(1):28–40. perfusion imaging for predicting short- and long-term outcomes
152. Taylor AJ, Rogan KM, Virmani R. Sudden cardiac death associ- in emergency department patients with chest pain. J Nucl Cardiol.
ated with isolated congenital coronary artery anomalies. J Am 2008;15(6):774–82.
Coll Cardiol. 1992;20(3):640–7. 165. Radensky PW, Hilton TC, Fulmer H, McLaughlin BA, Stowers
153. Basso C, Maron BJ, Corrado D, Thiene G. Clinical profile of con- SA.  Potential cost effectiveness of initial myocardial perfusion
genital coronary artery anomalies with origin from the wrong aor- imaging for assessment of emergency department patients with
tic sinus leading to sudden death in young competitive athletes. J chest pain. Am J Cardiol. 1997;79(5):595–9.
Am Coll Cardiol. 2000;35(6):1493–501. 166. Udelson JE, Beshansky JR, Ballin DS, Feldman JA, Griffith
154. Bartoli CR, Wead WB, Giridharan GA, Prabhu SD, Koenig SC, JL, Handler J, Heller GV, Hendel RC, Pope JH, Ruthazer R,
Dowling RD. Mechanism of myocardial ischemia with an anoma- Spiegler EJ, Woolard RH, Selker HP.  Myocardial perfusion
lous left coronary artery from the right sinus of Valsalva. J Thorac imaging for evaluation and triage of patients with suspected
Cardiovasc Surg. 2012;144(2):402–8. acute cardiac ischemia: a randomized controlled trial. JAMA.
155. Reichlin T, Schindler C, Drexler B, Twerenbold R, Reiter M, 2002;288(21):2693–700.
Zellweger C, Moehring B, Ziller R, Hoeller R, Rubini Gimenez 167. Kontos MC, Schmidt KL, McCue M, Rossiter LF, Jurgensen M,
M, Haaf P, Potocki M, Wildi K, Balmelli C, Freese M, Stelzig C, Nicholson CS, Jesse RL, Ornato JP, Tatum JL. A comprehensive
Freidank H, Osswald S, Mueller C. One-hour rule-out and rule-in strategy for the evaluation and triage of the chest pain patient: a
of acute myocardial infarction using high-sensitivity cardiac tro- cost comparison study. J Nucl Cardiol. 2003;10(3):284–90.
ponin T. Arch Intern Med. 2012;172(16):1211–8. 168. Kontos MC, Arrowood JA, Paulsen WH, Nixon JV.  Early

156. Haaf P, Reichlin T, Twerenbold R, Hoeller R, Rubini Gimenez M, echocardiography can predict cardiac events in emergency
Zellweger C, Moehring B, Fischer C, Meller B, Wildi K, Freese department patients with chest pain. Ann Emerg Med.
M, Stelzig C, Mosimann T, Reiter M, Mueller M, Hochgruber 1998;31(5):550–7.
T, Sou SM, Murray K, Minners J, Freidank H, Osswald S, 169. Muscholl MW, Oswald M, Mayer C, von Scheidt W. Prognostic
Mueller C.  Risk stratification in patients with acute chest pain value of 2D echocardiography in patients presenting with acute
using three high-sensitivity cardiac troponin assays. Eur Heart J. chest pain and non-diagnostic ECG for ST-elevation myocardial
2014;35(6):365–75. infarction. Int J Cardiol. 2002;84(2–3):217–25.
157. Body R, Carley S, McDowell G, Jaffe AS, France M, Cruickshank 170. Lim SH, Sayre MR, Gibler WB. 2-D echocardiography prediction
K, Wibberley C, Nuttall M, Mackway-Jones K.  Rapid exclu- of adverse events in ED patients with chest pain. Am J Emerg
sion of acute myocardial infarction in patients with undetect- Med. 2003;21(2):106–10.
able troponin using a high-sensitivity assay. J Am Coll Cardiol. 171. Weston P, Alexander JH, Patel MR, Maynard C, Crawford L,
2011;58(13):1332–9. Wagner GS. Hand-held echocardiographic examination of patients
158. Dedic A, Lubbers MM, Schaap J, Lammers J, Lamfers EJ, with symptoms of acute coronary syndromes in the emergency
Rensing BJ, Braam RL, Nathoe HM, Post JC, Nielen T, Beelen department: the 30-day outcome associated with normal left ven-
D, le Cocq d’Armandville MC, Rood PP, Schultz CJ, Moelker A, tricular wall motion. Am Heart J. 2004;148(6):1096–101.
Ouhlous M, Boersma E, Nieman K. Coronary CT angiography for 172. Kwong RY, Schussheim AE, Rekhraj S, Aletras AH, Geller N,
suspected ACS in the era of high-sensitivity troponins: random- Davis J, Christian TF, Balaban RS, Arai AE. Detecting acute coro-
ized multicenter study. J Am Coll Cardiol. 2016;67(1):16–26. nary syndrome in the emergency department with cardiac mag-
159. Korley FK, George RT, Jaffe AS, Rothman RE, Sokoll LJ, netic resonance imaging. Circulation. 2003;107(4):531–7.
Fernandez C, Falk H, Post WS, Saheed MO, Gerstenblith G, 173. Ingkanisorn WP, Kwong RY, Bohme NS, Geller NL, Rhoads KL,
Berkowitz SA, Hill PM. Low high-sensitivity troponin I and zero Dyke CK, Paterson DI, Syed MA, Aletras AH, Arai AE. Prognosis
coronary artery calcium score identifies coronary CT angiography of negative adenosine stress magnetic resonance in patients pre-
candidates in whom further testing could be avoided. Acad Radiol. senting to an emergency department with chest pain. J Am Coll
2015;22(8):1060–7. Cardiol. 2006;47(7):1427–32.
348 N. J. Lee and H. Litt

174. Cury RC, Shash K, Nagurney JT, Rosito G, Shapiro MD, Nomura 176. James JT. A new, evidence-based estimate of patient harms associ-
CH, Abbara S, Bamberg F, Ferencik M, Schmidt EJ, Brown ated with hospital care. J Patient Saf. 2013;9(3):122–8.
DF, Hoffmann U, Brady TJ.  Cardiac magnetic resonance with 177. Foy AJ, Liu G, Davidson WR Jr, Sciamanna C, Leslie
T2-weighted imaging improves detection of patients with acute DL.  Comparative effectiveness of diagnostic testing strategies
coronary syndrome in the emergency department. Circulation. in emergency department patients with chest pain: an analysis of
2008;118(8):837–44. downstream testing, interventions, and outcomes. JAMA Intern
175. Kim WY, Danias PG, Stuber M, Flamm SD, Plein S, Nagel E, Med. 2015;175(3):428–36.
Langerak SE, Weber OM, Pedersen EM, Schmidt M, Botnar 178. Finnerty NM, Weinstock MB. Coronary computed tomography angi-
RM, Manning WJ.  Coronary magnetic resonance angiogra- ography for low-risk chest pain. Ann Emerg Med. 2016;68(5):645.
phy for the detection of coronary stenoses. N Engl J Med. 179. Redberg RF.  Coronary CT angiography for acute chest pain. N
2001;345(26):1863–9. Engl J Med. 2012;367(4):375–6.
 The Role of Cardiac CT in Patients
with Metabolic Disorders 29
Gianluca Pontone, Giuseppe Muscogiuri,
and Mark Rabbat

Diabetes inducible ischemia [7]. In the CACTI study, Dabelea et al.

showed that CAC is more extensive and prevalent in patients
Diabetes mellitus (DM) is a complex and heterogeneous dis- with type 1 DM compared to sex-matched controls [8].
order that can be classified in two subtypes [1]. Type 1 has a Therefore, if the threshold of calcium scores associated with
juvenile onset and is caused by autoimmune destruction of a high prevalence of ischemia in nondiabetic patients is
the pancreatic beta cell and subsequent insulin deficiency. >400, in diabetics the threshold maybe lower [9, 10].
Type 2 is typically characterized by adult onset and associ- Despite the CACTI study demonstrating an increased
ated with being overweight, a sedentary lifestyle, and a diet CAC value in patients with diabetes, a large portion of
high in fat and/or calories [1]. Oftentimes, diabetes is associ- patients with DM show a very low or CAC score of 0 [5].
ated with the presence of cardiovascular disease. Therefore, This finding supports the evidence that diabetes by itself is
optimal cardiovascular risk stratification is needed for an not a CAD equivalent.
optimal therapeutic approach [2]. Additionally, diabetics with a low CAC score have a low
Before the institution of contemporary coronary care, the short-term risk of death similar to nondiabetic patients and
mortality for myocardial infarction (MI) in diabetic patients do not appear to benefit from aspirin and statin use [11–13].
was higher than 40% and at least double compared with Extending the follow-up to 5 years, Valenti et al. support the
patients without diabetes [3]. In a large cohort of patients, theory that patients with a calcium score of 0 maintain low
Haffner et  al. demonstrated that comparing patients with cardiovascular risk [14].
prior myocardial infarction, patients with DM had poor out- Although a large portion of diabetics have very low or
comes with a higher probability of developing a myocardial CAC scores of 0, those with elevated CAC scores appear to
infarction compared to nondiabetics [4]. have poor mid- and long-term outcome compared to nondia-
Although diabetes has been associated with coronary betics [11, 14]. Therefore the evaluation of CAC scores in
artery disease (CAD), recently, some data suggests that it patients with DM is extremely important to help determine
should not be considered a CAD equivalent. Indeed, associ- the best therapeutic approach and risk stratification. In fact,
ating imaging findings, functional tests and severity of diabe- CAC scoring demonstrated a specificity and sensitivity of
tes have reclassified the therapeutic approach and risk diagnosing CAD in patients with asymptomatic type 2 dia-
stratification to better identify and predict ischemic event in betes mellitus of 100% and 88%, respectively [15].
patients with DM [5]. Taking into account the importance of CAC for risk stratifica-
A coronary artery calcium (CAC) score is easily acquired tion of patients, McClelland et al. suggested the use of a new
using cardiac CT and provides a wealth of diagnostic and algorithm that combines the CAC score with clinical information
prognostic information. The association of CAC scores with such as presence of diabetes, total cholesterol, HDL cholesterol,
atherosclerotic plaque burden has been well described [6]. smoking, and systolic blood pressure in order to better predict
Hence, patients with high CAC are more likely to develop the 10-year risk of developing coronary heart disease [16].
In addition to CAC, many papers in the literature have
G. Pontone (*) · G. Muscogiuri studied the role of coronary computed angiography (cCTA)
Centro Cardiologico Monzino, IRCCS, Milan, Italy for the assessment of CAD in patients with DM. In diabetics,
e-mail: [email protected]
although cCTA has high sensitivity for early identification of
M. Rabbat significant CAD, its diagnostic performance is reduced com-
Department of Medicine, Division of Cardiology,
Loyola University Chicago, Chicago, IL, USA
pared to nondiabetics [17, 18]. In several studies, risk strati-
fication in patients with diabetes using cCTA has
Department of Medicine, Division of Cardiology, Edward Hines Jr.
VA Hospital, Hines, IL, USA
demonstrated greater extent, severity, and prevalence of

© Humana Press 2019 349

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_29
350 G. Pontone et al.

CAD, leading to an increased risk of mortality [19]. Kim Table 29.1  Current guidelines in patients with diabetes mellitus [5]
et  al. demonstrated that a longer duration of diabetes in Risk factor assessment Class I
asymptomatic patients is associated with increased major Coronary artery calcium scanning Class IIa
adverse cardiac and cerebrovascular events in 10  years of Hemoglobin A1c for risk assessment Class IIb
follow-up. Another study by Min et al. demonstrated similar Not supported
results despite shorter-term follow-up of 3 years [20, 21]. In Carotid intima media thickness Class III
addition, other studies have demonstrated that diabetics had Routine functional testing Class III
worse prognosis at both 5 and 7 years [22, 23]. Data from Budoff et al. [5]
Class III
Despite the strong evidence of a real impact in progno-
sis and risk stratification of cCTA, currently, its use for
screening of CAD is not recommended. In the FACTOR-64
Table 29.2  Computed tomography in metabolic disorders
trial, at mean follow-up of 4 years, the primary composite
end point of all-cause mortality, nonfatal myocardial Metabolic disorders Role of CT
infarction and hospitalization for unstable angina did not Diabetes CAC
Extent and severity of CAD
differ for DM patients compared to the control group. Possible functional evaluation of CAD
Furthermore, the secondary end point of ischemic major  1. FFR-CT
adverse cardiovascular events was similar [24]. Thus, the  2. CTP
findings from FACTOR-64 do not support cCTA for Gaucher disease Calcifications of valves and thoracic
aorta
screening of CAD.
Mucopolysaccharidoses Evaluation of valve morphology
However the lack of significant improvement for both Narrowing of coronary arteries
primary and secondary outcomes observed in the Alkaptonuria CAC
FACTOR-64 trial may be explained by the low-risk study Myocardial, valve, vascular calcification
population and by the lack of benefit from an invasive thera- Cystinosis CAC
peutic approach based solely on the anatomical evaluation Obesity CAC
EAT
of coronaries [24].
Studies comparing invasive fractional flow reserve (FFR) CT computed tomography, CAC coronary artery calcium, CAD coro-
nary artery disease, FFR-CT fractional flow reserve computed tomogra-
to anatomical imaging, including quantitative coronary angi- phy, CTP computed tomography perfusion, EAT epicardial adipose
ography (QCA), cCTA, or intravascular ultrasound (IVUS), tissue
have consistently demonstrated an unreliable relationship
between anatomic measures of stenosis and lesion-specific
ischemia [25, 26]. In addition, overestimation of stenosis in
calcified plaque is a well-known limitation, and some ven- Gaucher Disease
dors have adopted technical strategies to reduce blooming
artifact [27]. Gaucher disease (GD) is the most common lysosomal stor-
To overcome these limitations, CT perfusion (CTP) and age disorder. It is characterized by the loss of function of the
fractional flow reserve CT (FFR-CT) have developed to catabolic lysosomal enzyme acid beta-glucocerebrosidase
bridge the anatomic evaluation of CAD to the functional causing accumulation of sphingolipid glycosylceramide in
relevance of stenosis [28, 29]. While the current guidelines lysosomes. Different types of genetic mutations involving
state a level IIa recommendation for CAC scoring and the GBA gene are responsible for the diminished function of
level III recommendation for routine cCTA for risk strati- the enzyme [33, 34].
fication, the novel noninvasive approach based on CTP, Clinical manifestation of GD can be very heterogeneous
FFR-CT, and plaque analysis will likely expand and but can be divided into three types [34].
broaden the indication for cCTA in order to accurately risk
stratify patients with DM (Table  29.1) [30, 31]. In fact, • Type 1: Characterized by hepatosplenomegaly, anemia,
Vliegenthart et al. recently identified early abnormal CTP thrombocytopenia, and bone disease.
in patients with diabetes which may improve CAD risk • Type 2: Typical onset in infancy with neurologic and vis-
stratification [32]. ceral manifestations associated with arrest of growth and
In conclusion, both CAC scoring and cCTA may be death in early childhood.
helpful in risk stratification and management of asymp- • Type 3: Characterized by neurologic involvement usu-
tomatic patients with DM.  Moreover, a combined ally with onset in early childhood. Based upon the
approach utilizing FFR-CT, CTP, and plaque characteris- involvement of visceral disease and neurological signs,
tics may play a key role in future CAD risk stratification type 3 GD can be further divided into three subgroups:
(Table 29.2). 3a, 3b, and 3c.
29  The Role of Cardiac CT in Patients with Metabolic Disorders 351

Cardiac involvement in GD is manifested as calcification Alkaptonuria

of the myocardium and mitral and aortic valve [35–37]. Few
case reports describe the role of CT for the evaluation of Alkaptonuria is an autosomal recessive metabolic disorder
patients with GD. The papers do not use ECG-gated acquisi- caused by a deficient activity of homogentisate 1,
tion and describe the appearance of calcified valves and tho- 2-­dioxygenase (HGO), an enzyme involved in the metabo-
racic aorta in patients with GD [38, 39]. lism of tyrosine. The decreased function of HGO leads to an
increase of homogentisic acid (HGA). Excessive production
of HGA is eliminated in the urine or accumulates in connec-
Mucopolysaccharidoses tive tissue causing ochronosis and subsequent darkening of
bones and cartilage, arthritis, joint destruction, and degener-
Mucopolysaccharidoses (MPS) is a genetic disease due ation of cardiac valves [47].
to a lysosomal storage disorder that causes impaired deg- In alkaptonuria, ochronosis seems to be upstream cause
radation of glycosaminoglycans (GAG) [40, 41]. There of cardiac valvular deterioration. The deposits of HGA mani-
are nine different subtypes of MPS depending on the fest a classic pigmentation to the cardiac valves that subse-
enzyme deficit with heterogeneous clinical manifesta- quently undergo a gradual progression to valvular
tions [40]. calcification [48]. The association of degenerative aortic ste-
MPS is a multivisceral disease and involves the musculo- nosis and alkaptonuria has been described [49–51]. As in
skeletal system, heart, lung, eyes, liver, spleen, and occasion- GD, cCTA can be useful for the evaluation of myocardial,
ally the cerebral nervous system [41]. valve, or vascular calcification [52]. CAC is commonly
Cardiac involvement in MPS is caused by the accumula- found in patients with alkaptonuria and can be quantified
tion of GAG. Thickening of the cardiac valves, intramyocar- using the Agatston score [47, 52].
dial deposition of GAG, patchy interstitial fibrosis, coronary
artery narrowing from myointimal proliferation, and sys-
temic hypertension may all contribute to the development of Cystinosis
heart failure [42].
Left-sided cardiac valves are commonly affected in MPS Cystinosis is a rare, autosomal recessive disease in which
and these individuals typically have MPS subtypes I, II, and cystine is accumulated in the lysosomes. The biallelic muta-
IV [43]. The mitral valve leaflets are usually thickened, the tion of the cystinosin, lysosomal cystine transporter (CTNS)
chordae tendineae are short, and the papillary muscles are gene, leads to deficient cystinosin (which is a cystine-proton
dysmorphic. The abnormal morphologic changes of the cotransporter) and consequently cystine is accumulated in
valves result in pathologic valvular function and the develop- the lysosomes [53].
ment of stenosis and insufficiency. Cystinosis is the most common cause of Fanconi syn-
Transthoracic two-dimensional echocardiography is the drome. Typically onset of symptoms occurs in early child-
first-line noninvasive technique that allows reliable evalua- hood. Polyuria, polydipsia, vomiting, constipation, growth
tion of valvular function and morphology. However, in retardation, rickets, dehydration, and acidosis have been
patients with poor acoustic windows or in cases where trans- associated with the disease [54]. Photophobia and blepharo-
esophageal echocardiography cannot be performed, cCTA spasm are common in patients with cystinosis as result of
represents an alternative diagnostic technique. While cCTA corneal cystine crystals accumulation. Loss of thyroid func-
can provide key information regarding valvular morphology, tion is caused by the gradual cystine accumulation and crys-
it is important and good practice to consider the radiation tal formation in thyroid follicular cells [53, 55]. Other
exposure to patients [43]. extrarenal manifestations include pancreatic and pulmonary
Narrowing of coronaries is described in all subtypes of dysfunction, muscle atrophy or weakness, and rarely impair-
MPS; however, it is more common in subtypes I and II. A ment of the central nervous system [54].
decrease in coronary lumen diameter is caused by the depo- The progressive renal impairment in patients with cysti-
sition of GAG that leads to a progressive diffuse intimal pro- nosis may lead to kidney failure and eventual kidney
liferation [43–45]. transplantation.
It is debatable whether cCTA is useful and maintains high Ueda et al. observed the increase of coronary artery and
diagnostic performance in patients with MPS mainly from vascular calcification in patients with cystinosis after kid-
the diffuse involvement of the vessels [43]. However, Felice ney transplantation [56]. They report that increased values
et al. describe a case of a 30-year-old patient with MPS, aor- of CAC were associated with to the onset of diabetes and/
tic stenosis, and left ventricular hypertrophy who underwent or abnormalities of calcium and phosphate homeostasis
cCTA and successfully ruled out involvement of coronary [56]. Furthermore, in one cystinosis patient with significant
narrowing [46]. CAC noted on CT, coronary angiography was performed to
352 G. Pontone et al.

evaluate the severity of CAD [56]. However, it is yet to be report that there may be incremental prognostic value of
determined whether patients with cystinosis may benefit EAT over CAC alone [72].
from the evaluation of CAC scoring and severity of CAD Although cardiac CT can be useful in obese patients for
using cCTA. the evaluation of CAC, EAT, and CAD extent and severity,
image quality may be limited if tube potential and current
are not adequately adjusted according to the individuals
Obesity BMI [73].
However, one must be cognizant that the increase of tube
Obesity is characterized by a body mass index (BMI) higher voltage and current leads to increased radiation dose to the
than 30 Kg/m2 in adults and above the 95th percentile in the patient. Hence, several strategies have been developed by the
pediatric population. The development of obesity results vendors in order to overcome this issue. Mangold et al. report
from the complex interaction between genetic, socioeco- high diagnostic image quality in obese patients with third-­
nomic, and cultural influences [57]. generation dual-source CT using a tube voltage of 120-kV. In
Obesity represents an important and serious health fact, diagnostic accuracy between obese and nonobese
problem in developed and developing countries which patients showed no significant difference in image quality
greatly influence the cost to healthcare system [58]. even when 44% of the obese patients underwent cCTA with
Furthermore, obesity is associated with several comorbidi- tube voltage <120 kV [74, 75].
ties that may adversely affect the clinical condition of
patients, resulting in poor outcomes. The combination of
being overweight and obese in 2010 was estimated to References
cause 3.4 million deaths [59, 60].
Comorbidities often associated with an obese body habi- 1. Polonsky KS.  The past 200 years in diabetes. N Engl J Med.
2012;367(14):1332–40.
tus include but not limited to obstructive sleep apnea, female 2. Inzucchi SE. Diagnosis of diabetes. N Engl J Med. 2013;368(2):193.
infertility, diabetes, and coronary heart disease [59, 61]. The 3. Nesto RW, Zarich S.  Acute myocardial infarction in diabe-
relationship between obesity and development of CAC is tes mellitus: lessons learned from ACE inhibition. Circulation.
debatable. In 2005, Cassidy et al. showed a positive correla- 1998;97(1):12–5.
4. Haffner SM, Lehto S, Ronnemaa T, Pyorala K, Laakso M. Mortality
tion between CAC and waist circumference, waist-to-hip from coronary heart disease in subjects with type 2 diabetes and in
ratio, and BMI [62]. In 2010, the CARDIA study demon- nondiabetic subjects with and without prior myocardial infarction.
strated that baseline BMI was associated positively and lin- N Engl J Med. 1998;339(4):229–34.
early with CAC at year 20 of follow-up [63]. In contrast to 5. Budoff MJ, Raggi P, Beller GA, Berman DS, Druz RS, Malik
S, et  al. Noninvasive cardiovascular risk assessment of the
the above results, Kovacic et al. found an inverse correlation asymptomatic diabetic patient: the imaging council of the
between BMI and CAC, while other studies report the lack American College of Cardiology. JACC Cardiovasc Imaging.
of significant association [64–66]. 2016;9(2):176–92.
The conflicting results regarding the association of 6. Rumberger JA, Simons DB, Fitzpatrick LA, Sheedy PF, Schwartz
RS.  Coronary artery calcium area by electron-beam computed
CAC and BMI were further investigated by two large stud- tomography and coronary atherosclerotic plaque area. A histopath-
ies. Analysis performed on the CONFIRM registry found ologic correlative study. Circulation. 1995;92(8):2157–62.
that obese patients have a statistically significant higher 7. Bavishi C, Argulian E, Chatterjee S, Rozanski A.  CACS and the
CAC compared to normal weight patients [67]. Similar frequency of stress-induced myocardial ischemia during MPI: a
meta-analysis. JACC Cardiovasc Imaging. 2016;9(5):580–9.
results were observed by Fujiyoshi et  al. where after 8. Dabelea D, Kinney G, Snell-Bergeon JK, Hokanson JE, Eckel RH,
adjusting for age, smoking, alcohol, hypertension, lipids, Ehrlich J, et al. Effect of type 1 diabetes on the gender difference
and diabetes mellitus, CAC was independently and posi- in coronary artery calcification: a role for insulin resistance? The
tively associated with prevalent CAC in a large population Coronary Artery Calcification in Type 1 Diabetes (CACTI) Study.
Diabetes. 2003;52(11):2833–9.
of 1212 patients [68]. 9. He ZX, Hedrick TD, Pratt CM, Verani MS, Aquino V, Roberts
Another finding that can be evaluated in obese and non- R, et  al. Severity of coronary artery calcification by electron
obese patients is the amount of epicardial adipose tissue beam computed tomography predicts silent myocardial ischemia.
(EAT). EAT, the visceral adipose tissue located between the Circulation. 2000;101(3):244–51.
10. Wong ND, Rozanski A, Gransar H, Miranda-Peats R, Kang X,
myo-epicardium and visceral pericardium, is an independent Hayes S, et  al. Metabolic syndrome and diabetes are associated
risk factor for CAD. EAT plays a local homeostatic role in with an increased likelihood of inducible myocardial ischemia
management of lipid and energy. In particular, it is respon- among patients with subclinical atherosclerosis. Diabetes Care.
sible for free fatty acid release and uptake [69]. The amount 2005;28(6):1445–50.
11. Raggi P, Shaw LJ, Berman DS, Callister TQ.  Prognostic value
of EAT is correlated with components of metabolic syn- of coronary artery calcium screening in subjects with and with-
drome, and the increase in EAT has been associated with the out diabetes. J Am Coll Cardiol. 2004;43(9):1663–9. https://doi.
progression of CAC [70, 71]. In addition, Spearman et  al. org/10.1016/j.jacc.2003.09.068.
29  The Role of Cardiac CT in Patients with Metabolic Disorders 353

12. Silverman MG, Blaha MJ, Budoff MJ, Rivera JJ, Raggi P, Shaw LJ, with diabetes: the FACTOR-64 randomized clinical trial. JAMA.
et al. Potential implications of coronary artery calcium testing for 2014;312(21):2234–43.
guiding aspirin use among asymptomatic individuals with diabetes. 25. Pontone G, Andreini D, Baggiano A, Bertella E, Mushtaq S, Conte
Diabetes Care. 2012;35(3):624–6. E, et  al. Functional relevance of coronary artery disease by car-
13. Nasir K, Bittencourt MS, Blaha MJ, Blankstein R, Agatson AS, diac magnetic resonance and cardiac computed tomography: myo-
Rivera JJ, et  al. Implications of coronary artery calcium test- cardial perfusion and fractional flow reserve. Biomed Res Int.
ing among statin candidates according to American College of 2015;2015:297696.
Cardiology/American Heart Association cholesterol management 26. Pontone G, Muscogiuri G, Andreini D, Guaricci AI, Guglielmo
guidelines: MESA (Multi-Ethnic Study of Atherosclerosis). J Am M, Mushtaq S, et  al. The new frontier of cardiac computed
Coll Cardiol. 2015;66(15):1657–68. tomography angiography: fractional flow reserve and stress
14. Valenti V, Hartaigh BO, Cho I, Schulman-Marcus J, Gransar H, myocardial perfusion. Curr Treat Options Cardiovasc Med.
Heo R, et al. Absence of coronary artery calcium identifies asymp- 2016;18(12):74.
tomatic diabetic individuals at low near-term but not long-term 27. Pontone G, Bertella E, Mushtaq S, Loguercio M, Cortinovis S,
risk of mortality: a 15-year follow-up study of 9715 patients. Circ Baggiano A, et al. Coronary artery disease: diagnostic accuracy of
Cardiovasc Imaging. 2016;9(2):e003528. CT coronary angiography--a comparison of high and standard spa-
15. Rassi CH, Churchill TW, Tavares CA, Fahel MG, Rassi FP, Uchida tial resolution scanning. Radiology. 2014;271(3):688–94.
AH, et  al. Use of imaging and clinical data to screen for cardio- 28. Wang R, Renker M, Schoepf UJ, Wichmann JL, Fuller SR, Rier JD,
vascular disease in asymptomatic diabetics. Cardiovasc Diabetol. et al. Diagnostic value of quantitative stenosis predictors with coro-
2016;15:28. nary CT angiography compared to invasive fractional flow reserve.
16. McClelland RL, Jorgensen NW, Budoff M, Blaha MJ, Post WS, Eur J Radiol. 2015;84(8):1509–15.
Kronmal RA, et al. 10-year coronary heart disease risk prediction 29. Ko BS, Cameron JD, Leung M, Meredith IT, Leong DP, Antonis
using coronary artery calcium and traditional risk factors: deriva- PR, et  al. Combined CT coronary angiography and stress myo-
tion in the MESA (Multi-Ethnic Study of Atherosclerosis) with cardial perfusion imaging for hemodynamically significant
validation in the HNR (Heinz Nixdorf Recall) study and the DHS stenoses in patients with suspected coronary artery disease: a com-
(Dallas Heart Study). J Am Coll Cardiol. 2015;66(15):1643–53. parison with fractional flow reserve. JACC Cardiovasc Imaging.
17. Andreini D, Pontone G, Bartorelli AL, Agostoni P, Mushtaq S, 2012;5(11):1097–111.
Antonioli L, et  al. Comparison of the diagnostic performance of 30. Nakazato R, Park HB, Gransar H, Leipsic JA, Budoff MJ, Mancini
64-slice computed tomography coronary angiography in diabetic GB, et al. Additive diagnostic value of atherosclerotic plaque char-
and non-diabetic patients with suspected coronary artery disease. acteristics to non-invasive FFR for identification of lesions causing
Cardiovasc Diabetol. 2010;9:80. ischaemia: results from a prospective international multicentre trial.
18. Ulimoen GR, Ofstad AP, Endresen K, Gullestad L, Johansen OE, EuroIntervention. 2016;12(4):473–81.
Borthne A. Low-dose CT coronary angiography for assessment of 31. Gaur S, Ovrehus KA, Dey D, Leipsic J, Botker HE, Jensen JM,
coronary artery disease in patients with type 2 diabetes--a cross-­ et al. Coronary plaque quantification and fractional flow reserve by
sectional study. BMC Cardiovasc Disord. 2015;15:147. coronary computed tomography angiography identify ischaemia-­
19. Rana JS, Dunning A, Achenbach S, Al-Mallah M, Budoff MJ, causing lesions. Eur Heart J. 2016;37(15):1220–7.
Cademartiri F, et  al. Differences in prevalence, extent, severity, 32. Vliegenthart R, De Cecco CN, Wichmann JL, Meinel FG, Pelgrim
and prognosis of coronary artery disease among patients with GJ, Tesche C, et  al. Dynamic CT myocardial perfusion imaging
and without diabetes undergoing coronary computed tomog- identifies early perfusion abnormalities in diabetes and hyperten-
raphy angiography: results from 10,110 individuals from the sion: insights from a multicenter registry. J Cardiovasc Comput
CONFIRM (COronary CT Angiography EvaluatioN For Clinical Tomogr. 2016;10(4):301–8.
Outcomes): an InteRnational Multicenter Registry. Diabetes Care. 33. Horowitz M, Elstein D, Zimran A, Goker-Alpan O. New directions
2012;35(8):1787–94. in Gaucher disease. Hum Mutat. 2016;37:1121.
20. Kim JJ, Hwang BH, Choi IJ, Choo EH, Lim S, Kim JK, et  al. 34. Bendikov-Bar I, Horowitz M.  Gaucher disease paradigm: from
Impact of diabetes duration on the extent and severity of coronary ERAD to comorbidity. Hum Mutat. 2012;33(10):1398–407.
atheroma burden and long-term clinical outcome in asymptomatic 35. Shah S, Misri A, Bhat M, Maheshwari S. Gaucher’s disease type III
type 2 diabetic patients: evaluation by Coronary CT angiography. C: unusual cause of intracardiac calcification. Ann Pediatr Cardiol.
Eur Heart J Cardiovasc Imaging. 2015;16(10):1065–73. 2008;1(2):144–6.
21. Min JK, Labounty TM, Gomez MJ, Achenbach S, Al-Mallah M, 36. Chabas A, Cormand B, Grinberg D, Burguera JM, Balcells S,
Budoff MJ, et al. Incremental prognostic value of coronary com- Merino JL, et  al. Unusual expression of Gaucher’s disease: car-
puted tomographic angiography over coronary artery calcium score diovascular calcifications in three sibs homozygous for the D409H
for risk prediction of major adverse cardiac events in asymptomatic mutation. J Med Genet. 1995;32(9):740–2.
diabetic individuals. Atherosclerosis. 2014;232(2):298–304. 37. Abrahamov A, Elstein D, Gross-Tsur V, Farber B, Glaser Y, Hadas-­
22. Andreini D, Pontone G, Mushtaq S, Bertella E, Conte E, Baggiano Halpern I, et  al. Gaucher’s disease variant characterised by pro-
A, et  al. Prognostic value of multidetector computed tomogra- gressive calcification of heart valves and unique genotype. Lancet.
phy coronary angiography in diabetes: excellent long-term prog- 1995;346(8981):1000–3.
nosis in patients with normal coronary arteries. Diabetes Care. 38. Ceyhan M, Celik FK, Elmali M, Gurmen N. An unusual form of
2013;36(7):1834–41. Gaucher’s disease: pulmonary and cardiovascular involvement and
23. Halon DA, Azencot M, Rubinshtein R, Zafrir B, Flugelman MY, cholelitiasis. Cent Eur J Med. 2010;5(4):495–8.
Lewis BS. Coronary computed tomography (CT) angiography as a 39. Altunbas G, Ercan S, Inanc IH, Ozer O, Kervancioglu S, Davutoglu
predictor of cardiac and noncardiac vascular events in asymptom- V. Extensive vascular and valvular involvement in Gaucher disease.
atic type 2 diabetics: a 7-year population-based cohort study. J Am Asian Cardiovasc Thorac Ann. 2015;23(4):446–8.
Heart Assoc. 2016;5(6):e003226. 40. Coutinho MF, Lacerda L, Alves S. Glycosaminoglycan storage dis-
24. Muhlestein JB, Lappe DL, Lima JA, Rosen BD, May HT, Knight orders: a review. Biochem Res Int. 2012;2012:471325.
S, et al. Effect of screening for coronary artery disease using CT 41. Valayannopoulos V, Wijburg FA.  Therapy for the mucopolysac-
angiography on mortality and cardiac events in high-risk patients charidoses. Rheumatology (Oxford). 2011;50(Suppl 5):v49–59.
354 G. Pontone et al.

42. Braunlin E, Wang R. Cardiac issues in adults with the mucopoly- 61. Mitchell A, Fantasia HC.  Understanding the effect of obesity on
saccharidoses: current knowledge and emerging needs. Heart. fertility among reproductive-age women. Nurs Womens Health.
2016;102(16):1257–62. 2016;20(4):368–76.
43. Braunlin EA, Harmatz PR, Scarpa M, Furlanetto B, Kampmann C, 62. Cassidy AE, Bielak LF, Zhou Y, Sheedy PF 2nd, Turner ST, Breen
Loehr JP, et al. Cardiac disease in patients with mucopolysacchari- JF, et al. Progression of subclinical coronary atherosclerosis: does
dosis: presentation, diagnosis and management. J Inherit Metab obesity make a difference? Circulation. 2005;111(15):1877–82.
Dis. 2011;34(6):1183–97. 63. Lee DH, Steffes MW, Gross M, Park K, Holvoet P, Kiefe CI, et al.
44. Brosius FC 3rd, Roberts WC. Coronary artery disease in the hurler Differential associations of weight dynamics with coronary artery
syndrome. Qualitative and quantitative analysis of the extent of calcium versus common carotid artery intima-media thickness: The
coronary narrowing at necropsy in six children. Am J Cardiol. CARDIA Study. Am J Epidemiol. 2010;172(2):180–9.
1981;47(3):649–53. 64. Kovacic JC, Lee P, Baber U, Karajgikar R, Evrard SM, Moreno P,
45. Kettles DI, Sheppard M, Liebmann RD, Davidson C. Left ventricular et al. Inverse relationship between body mass index and coronary
aneurysm, aortic valve disease and coronary narrowing in a patient artery calcification in patients with clinically significant coronary
with Hunter’s syndrome. Cardiovasc Pathol. 2002;11(2):94–6. lesions. Atherosclerosis. 2012;221(1):176–82.
46. Felice T, Murphy E, Mullen MJ, Elliott PM.  Management of
65. Bild DE, Detrano R, Peterson D, Guerci A, Liu K, Shahar E, et al.
aortic stenosis in mucopolysaccharidosis type I.  Int J Cardiol. Ethnic differences in coronary calcification: the Multi-Ethnic Study
2014;172(3):e430–1. of Atherosclerosis (MESA). Circulation. 2005;111(10):1313–20.
47. Phornphutkul C, Introne WJ, Perry MB, Bernardini I, Murphey 66. Taylor AJ, Feuerstein I, Wong H, Barko W, Brazaitis M, O'Malley
MD, Fitzpatrick DL, et al. Natural history of alkaptonuria. N Engl PG. Do conventional risk factors predict subclinical coronary artery
J Med. 2002;347(26):2111–21. disease? Results from the Prospective Army Coronary Calcium
48. Butany JW, Naseemuddin A, Moshkowitz Y, Nair V.  Ochronosis Project. Am Heart J. 2001;141(3):463–8.
and aortic valve stenosis. J Card Surg. 2006;21(2):182–4. 67. Labounty TM, Gomez MJ, Achenbach S, Al-Mallah M, Berman
49. Ffolkes LV, Brull D, Krywawych S, Hayward M, Hughes
DS, Budoff MJ, et al. Body mass index and the prevalence, sever-
SE.  Aortic stenosis in cardiovascular ochronosis. J Clin Pathol. ity, and risk of coronary artery disease: an international multi-
2007;60(1):92–3. centre study of 13,874 patients. Eur Heart J Cardiovasc Imaging.
50. Hiroyoshi J, Saito A, Panthee N, Imai Y, Kawashima D, Motomura 2013;14(5):456–63.
N, et  al. Aortic valve replacement for aortic stenosis caused by 68. Fujiyoshi A, Sekikawa A, Shin C, Masaki K, David Curb J, Ohkubo
alkaptonuria. Ann Thorac Surg. 2013;95(3):1076–9. T, et al. A cross-sectional association of obesity with coronary cal-
51. Fisher AA, Davis MW. Alkaptonuric ochronosis with aortic valve cium among Japanese, Koreans, Japanese Americans, and U.S.
and joint replacements and femoral fracture: a case report and lit- whites. Eur Heart J Cardiovasc Imaging. 2013;14(9):921–7.
erature review. Clin Med Res. 2004;2(4):209–15. 69. Iacobellis G, Malavazos AE, Corsi MM.  Epicardial fat: from the
52. Hannoush H, Introne WJ, Chen MY, Lee SJ, O’Brien K, Suwannarat biomolecular aspects to the clinical practice. Int J Biochem Cell
P, et al. Aortic stenosis and vascular calcifications in alkaptonuria. Biol. 2011;43(12):1651–4.
Mol Genet Metab. 2012;105(2):198–202. 70. Yorgun H, Canpolat U, Hazirolan T, Ates AH, Sunman H, Dural
53. Elmonem MA, Veys KR, Soliman NA, van Dyck M, van den M, et  al. Increased epicardial fat tissue is a marker of metabolic
Heuvel LP, Levtchenko E.  Cystinosis: a review. Orphanet J Rare syndrome in adult patients. Int J Cardiol. 2013;165(2):308–13.
Dis. 2016;11:47. 71. Nakanishi R, Rajani R, Cheng VY, Gransar H, Nakazato R,

54. Emma F, Nesterova G, Langman C, Labbe A, Cherqui S, Goodyer Shmilovich H, et al. Increase in epicardial fat volume is associated
P, et al. Nephropathic cystinosis: an international consensus docu- with greater coronary artery calcification progression in subjects at
ment. Nephrol Dial Transplant. 2014;29(Suppl 4):iv87–94. intermediate risk by coronary calcium score: a serial study using
55. Nesterova G, Gahl WA. In: Pagon RA, Adam MP, Ardinger HH, non-contrast cardiac CT. Atherosclerosis. 2011;218(2):363–8.
Wallace SE, Amemiya A, LJH B, et al., editors. Cystinosis. Seattle: 72. Spearman JV, Renker M, Schoepf UJ, Krazinski AW, Herbert TL,
GeneReviews; 1993. De Cecco CN, et al. Prognostic value of epicardial fat volume mea-
56. Ueda M, O’Brien K, Rosing DR, Ling A, Kleta R, McAreavey D, surements by computed tomography: a systematic review of the
et al. Coronary artery and other vascular calcifications in patients literature. Eur Radiol. 2015;25(11):3372–81.
with cystinosis after kidney transplantation. Clin J Am Soc Nephrol. 73. Halliburton SS, Abbara S, Chen MY, Gentry R, Mahesh M, Raff
2006;1(3):555–62. GL, et al. SCCT guidelines on radiation dose and dose-­optimization
57. Apovian CM. Obesity: definition, comorbidities, causes, and bur- strategies in cardiovascular CT.  J Cardiovasc Comput Tomogr.
den. Am J Manag Care. 2016;22(7 Suppl):s176–85. 2011;5(4):198–224.
58. Finkelstein EA, Trogdon JG, Cohen JW, Dietz W. Annual medical 74. Mangold S, Wichmann JL, Schoepf UJ, Litwin SE, Canstein C,
spending attributable to obesity: payer-and service-specific esti- Varga-Szemes A, et al. Coronary CT angiography in obese patients
mates. Health Aff (Millwood). 2009;28(5):w822–31. using 3(rd) generation dual-source CT: effect of body mass index
59. Haslam D.  Obesity: a medical history. Obes Rev. 2007;8(Suppl on image quality. Eur Radiol. 2016;26(9):2937–46.
1):31–6. 75. Mangold S, Wichmann JL, Schoepf UJ, Caruso D, Tesche C,

60. Ng M, Fleming T, Robinson M, Thomson B, Graetz N, Margono Steinberg DH, et al. Diagnostic accuracy of coronary CT angiogra-
C, et  al. Global, regional, and national prevalence of overweight phy using 3rd-generation dual-source CT and automated tube volt-
and obesity in children and adults during 1980-2013: a systematic age selection: clinical application in a non-obese and obese patient
analysis for the Global Burden of Disease Study 2013. Lancet. population. Eur Radiol. 2017;27(6):2298–308.
2014;384(9945):766–81.
 Use of Coronary Computed Tomography
Angiography in Cardiac Risk 30
Assessment for Non-cardiac Surgery

Gregory Jackson and Richard R. Bayer II

It has been estimated that approximately 900,000 patients Preoperative Risk Assessment
each year will suffer a major adverse cardiac event in the
postoperative period [1]. As such preoperative cardiac risk The Revised Cardiac Risk Index (RCRI) is a multifactorial
stratification is an important aspect in the clincal practice of risk index developed in 1999 by Lee et  al. which stratifies
internal medicine physicians, cardiologists, anesthesiolo- patients based upon specific medical comorbidities [3]. Both
gists, and surgeons alike. Traditional risk assessment models the American Heart Association (AHA) and the American
utilize patient-specific comorbidities to allow estimation of College of Cardiology (ACC) have emphasized the use of
perioperative cardiac risk. Such models have historically the RCRI for preoperative cardiac risk evaluation [9]. This
included the Multifactorial Cardiac Risk Index, Revised risk model utilizes patient-specific clinical risk factors
Cardiac Risk Index (RCRI), and Detsky’s Modified Cardiac including chronic kidney disease with a creatinine greater
Risk Index [2–4]. Once a patient-specific risk has been esti- than 2.0, insulin dependent diabetes mellitus, history of cere-
mated, the practitioner must not only determine the appropri- brovascular accident (CVA), history of ischemic heart dis-
ateness of further testing but if the results of subsequent ease, and history of congestive heart failure (Table 30.1) [9].
testing would alter treatment plans [5]. Traditionally, if fur- In combination with assessing the risk of surgery itself, a
ther testing was deemed appropriate, this was accomplished patient’s risk of major adverse cardiac events can and should
utilizing stress treadmill electrocardiography (ECG), single- also be estimated. This can range from a very low risk of
photon emission computed tomography (SPECT), or stress 0.4% to high risk with an 11% risk of perioperative cardiac
echocardiography [6]. In recent years, coronary computed complications (Table 30.2) [10]. In addition to clinical risk
tomography angiography (CCTA) has become an alternative factors, any assessment of preoperative risk also includes a
method for the evaluation of coronary artery disease. CCTA functional assessment (Table  30.3) [5]. Utilizing both the
has repeatedly demonstrated the ability to exclude signifi- RCRI risk score and patient’s functional capacity, the ACC
cant coronary artery disease with an extremely high sensitiv- and AHA have issued a stepwise algorithm for preoperative
ity and negative predictive value [7, 8]. However, to date its cardiac risk assessment (Fig.  30.1) [9]. The highlighted
use in the evaluation of preoperative cardiac risk stratifica-
tion has not been well established [6]. Table 30.1  Clinical risk factors utilized in preoperative cardiac risk
assessment via the Revised Cardiac Risk Index (RCRI)
Revised Cardiac Risk Index (RCRI)
History of ischemic heart disease
History of congestive heart failure
History of cerebrovascular accident
Chronic kidney disease with creatinine >2.0
G. Jackson History of insulin dependent diabetes mellitus
Department of Cardiology, Medical University of South Carolina,
Charleston, SC, USA
Table 30.2  Risk of perioperative complications based on number of
R. R. Bayer II (*) RCRI risk factors
Division of Cardiovascular Imaging, Department of Radiology
and Radiological Science, Medical University of South Carolina, Risk class # of risk factors Risk of complications
Charleston, SC, USA Very low risk 0 0.4%
Low risk 1 0.9%
Division of Cardiology, Department of Medicine,
Medical University of South Carolina, Charleston, SC, USA Moderate risk 2 7.0%
e-mail: [email protected] High risk ≥3 11.0%

© Humana Press 2019 355

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_30
356 G. Jackson and R. R. Bayer II

Table 30.3  Physical activity and corresponding metabolic equivalents sections represent situations where further testing may be
(METs) appropriate.
Physical activity METs With patients unable to complete ≥4 METs or who have
Sleeping 0.9 an unknown functional capacity and are intermediate to high
Watching television 1.0 risk, further noninvasive testing may be appropriate. This of
Walking slowly on level ground (1.7 mph) 2.3 course is with the caveat that it will alter management prior
Walking 3.0 miles per hour 3.3
to proceeding with surgery [5]. In accordance with the
Vacuuming, sweeping floor 3.3
Bicycling <10 miles per hour 4.0
appropriate use criteria, only SPECT and stress echocardiog-
Climb two flights of stairs 4.0 raphy receive scores of “appropriate” [9]. Additionally exer-
Washing car 4.5 cise stress ECG and stress cardiac magnetic resonance
Golf (carrying clubs) 4.5 (CMR) imaging earn scores of “may be appropriate” [9].
Walking 4.0 miles per hour 5.0 Conversely, in all situations CCTA is deemed “rarely appro-
Weightlifting 6.0 priate” [9]. These recommendations stem largely from the
Jogging 7.0 lack of current evidence for the use of CCTA in the setting of
Basketball 8.0 preoperative risk stratification [11].
Jump roping 10.0
Running (9 min/mile) 11.0

Fig. 30.1  Stepwise approach to

Yes
preoperative risk stratification prior Emergent Proceed with
Surgery? Surgery
to undergoing non-cardiac surgery

No

Yes Evaluation and

Acute Cardiac Management Prior to
Issues Proceeding
with Surgery

No

Yes
Proceed with
Low Risk Surgery?
Surgery

No

Yes
Proceed with
Functional
Surgery
Capacity
≥4 METs

No/Unknown

No Clinical Risk
≥3 Risk Factors 1-2 Risk Factors
Factors

Intermediate Risk Intermediate Risk Proceed with

Vascular Surgery Vascular Surgery
Surgery Surgery Surgery

Proceed with Proceed with Proceed with

Consider Non- Surgery OR Surgery OR Surgery OR
invasive Testing Consider Non- Consider Non- Consider Non-
invasive Testing invasive Testing invasive Testing
30  Use of Coronary Computed Tomography Angiography in Cardiac Risk Assessment for Non-cardiac Surgery 357

CCTA for Preoperative Risk Assessment myocardial infarction, pulmonary edema, ventricular fibrilla-
tion/tachycardia, and complete heart block [6]. MACE was
Although the indication for CCTA as part of preoperative seen in 3.5% of patients with minimal coronary artery disease
evaluation has not been well defined, the excellent sensitivity and 7.1% of patients with nonsignificant disease [6]. However,
and negative predictive value provides the rational for its use in patients with significant coronary artery disease, the rate of
as a screening test to exclude significant coronary disease adverse events increased dramatically to 17.2% [6]. The pres-
prior to non-cardiac surgery. While the volume of data is ence of multivessel disease increased rates of adverse events
insufficient to allow recommendation for the use of CCTA as above that of single-vessel disease, and higher CACS also
a noninvasive method in existing preoperative risk assess- resulted in higher rates of adverse cardiac events [6]. In
ment models, there does exist several small studies investi- patients with a CACS ≥113, the odds ratio (OR) for adverse
gating its utility. cardiac events was 5.84 [6]. This difference persisted when
The first study included 100 patients whom underwent stratified based on RCRI scores. In patients with CACS <113
CCTA prior to undergoing scheduled non-cardiac surgery compared to CACS ≥113, event rates in the RCRI score of 0
[12]. Patients whom were found to have significant coronary were 1.4% vs. 15.2%, 3.6% vs. 4.7% with a risk score of 1,
artery stenoses or in whom motion artifact or significant cal- 0% vs. 21.1% with a risk score of 2, and 50% vs. 66.7% with
cification prevented luminal assessment were referred for a risk score of 3 or more [6].
invasive angiography and excluded from analysis [12].
Significant stenosis was defined as luminal narrowing >50%
[12]. Of the 100 patients enrolled, CCTA was able to exclude
significant coronary artery disease in 81 patients and subse-
quently proceeded to surgery without further cardiac testing
[12]. These patients were followed for the onset of any major
adverse cardiac events (MACE) during the perioperative
period and for 3  months postoperatively [12]. MACE was
defined as death due to cardiac causes, acute myocardial
infarction, cardiac arrest, or nonelective coronary revascular-
ization [12]. In these 81 patients, no MACE were noted dur-
ing either the observed perioperative nor the postoperative
follow-up period [12].
A second study evaluated consecutive patients undergoing
intermediate risk non-cardiac surgery [6]. RCRI scores were
calculated in all patients [6]. Additionally, patients were not
only evaluated with CCTA but also underwent coronary
artery calcium scoring (CACS) [6]. While CACS alone is
unable assess the degree of coronary artery stenosis or the
precise location of potentially stenotic lesions, elevated
CACS portends an increased risk of major cardiovascular
events [13]. In one study of over 6500 asymptomatic patients
who underwent calcium scoring, the baseline calcium score
and the number of vessels involved were each independent
predictors of future coronary revascularizations during an
8.5-year follow-up period [13]. In another large observational
registry of more than 25,000 patients, unadjusted survival
over the follow-up period was 97.4% for patients with cal-
cium scores ranging from 11 to 100 and 73% for patients with
calcium scores ≥1000 [14]. Conversely, the absence of coro-
nary artery calcification was associated with extremely low
rates adverse cardiac events [15]. In the aforementioned
study, 239 patients underwent evaluation with CACS and 234
with CCTA [6]. Coronary stenoses were graded as minimal
(<30%) (Fig.  30.2), nonsignificant (>30% but <50%), and
significant (>50%) (Fig. 30.3) [6]. Patients were then subse-
quently followed for 30 days postoperatively for the presence Fig. 30.2  Calcified lesion in the left anterior descending (LAD) artery
of MACE, defined as cardiac death, unstable angina, nonfatal with resultant mild stenosis
358 G. Jackson and R. R. Bayer II

Additional Considerations

At present when additional noninvasive testing is indicated prior

to non-cardiac surgery, guidelines recommend the use of func-
tional testing (myocardial perfusion imaging, stress echocar-
diography, stress CMR). However, these modalities do not
provide anatomic assessment of the coronary arteries nor iden-
tify potential non-flow limiting stenosis, the latter of which
could pose risk for plaque rupture and subsequent adverse clini-
cal outcomes. Indeed it has been demonstrated that many post-
operative MACE are a result of nonobstructive coronary lesions
that may be more prone to plaque rupture in the increased
inflammatory setting of the operative period [19–21]. The abil-
ity to diagnose the presence of any coronary artery disease may
allow for modification of medical management in the preopera-
tive period resulting in mitigation of risk postoperatively. It has
been observed that the addition of statins in patients undergoing
surgery reduces rates of atrial fibrillation, myocardial infarction,
as well as length of hospital stay [22]. Additionally, while data
regarding beta-blockade in the perioperative period is conflict-
ing, there is data to support improved patient outcomes if beta-­
blockade is initiated early enough in the preoperative period
[23–25]. Apart from the anatomical assessment of the coronar-
ies themselves, CCTA has the ability to provide additional clini-
Fig. 30.3  Calcified and non-calcified plaque in the LAD with m
­ oderate cally relevant information pertaining to the lung parenchyma,
stenosis
pleura, mediastinum, and great vessels [26, 27].
Conversely, while myocardial perfusion imaging, stress
ECG, stress echocardiography, and stress CMR do not pro-
Finally, the third and largest study included 844 patients vide anatomical assessment of the coronary arteries, CCTA
undergoing CCTA prior to elective non-cardiac surgery [16]. does not provide functional assessment. However, with the
Similar to prior studies, patients were followed for 30 days advent of computed tomography-derived fractional flow
postoperatively for the development of MACE, defined as reserve measurements (FFRCT), the ability to assess the
pulmonary edema, myocardial infarction, or cardiac death [16]. hemodynamic significance of lesions identified on CCTA
As expected patients with a higher RCRI risk index had noninvasively exists [28–30]. Coupling CCTA with FFRCT
increasing postoperative MACE, an event rate of 1.2% with would allow not only for anatomical assessment of the coro-
an RCRI score of 1 compared to a 5.8% event rate with a nary arteries as well as surrounding cardiac structures but
RCRI score of ≥3 [16]. With respect to CCTA findings, also functional assessment as to the presence of ischemia. In
regardless of RCRI score patients with nonsignificant coro- a recent meta-analysis, the sensitivity for FFRCT was on par
nary stenosis (<50%) had fewer events that those with sig- with that of stress CMR (90% and 90%, respectively) and
nificant stenosis (≥50%), 1.7% vs. 8%, respectively [16]. surpassed that of both SPECT and stress echocardiography
When the two measures were combined, CCTA demon- (70% and 77%, respectively) [31]. Specificity was similar
strated the ability to reclassify patients within the same RCRI between FFRCT, SPECT, and stress echocardiography at
score group. In patients with an RCRI score of 2 or 3, the 71%, 78%, and 75%, respectively [31].
presence of any significant coronary artery stenosis increased
postoperative risk of MACE 3–17 times [16].
The previous studies would suggest that in patients with-
out significant coronary artery stenosis on preoperative Clinical Challenges
CCTA, there is a low rate of postoperative MACE.
Furthermore, positive and negative predictive values with A significant challenge in the decision-making process of
CCTA were on par with those of myocardial perfusion imag- preoperative evaluation is the timing and appropriateness of
ing and dobutamine stress echocardiography published pursuing intervention if significant coronary artery disease is
within the European Society of Cardiology (ESC), AHA, discovered. One must balance these findings against addi-
and ACC guidelines [16–18]. tional aspects of surgical planning, such as delay of non-­
30  Use of Coronary Computed Tomography Angiography in Cardiac Risk Assessment for Non-cardiac Surgery 359

emergent surgery to pursue revascularization, duration of 7. Achenbach S, Schuhbaeck A, Marwan M, et al. Multicenter evalu-
dual antiplatelet therapy, and further evaluation of extracar- ation of dual source ct coronary angiography in patients with inter-
mediate likelihood of coronary artery stenoses (medic): accuracy
diac findings (such as previously undiagnosed malignancies for the detection of individuals with significant coronary artery ste-
or incidental findings warranting further workup). Regardless noses. J Am Coll Cardiol. 2012;59:E1337.
of modality, preoperative noninvasive cardiac testing should 8. Meijboom WB, Meijs MF, Schuijf JD, et  al. Diagnostic accu-
be reserved for appropriate patients as determined by risk racy of 64-slice computed tomography coronary angiography: a
prospective, multicenter, multivendor study. J Am Coll Cardiol.
assessment and in whom results of testing would alter preop- 2008;52:2135–44.
erative management [5]. 9. Ronan G, Wolk MJ, Bailey SR, et al. ACCF/AHA/ASE/ASNC/
HFSA/HRS/SCAI/SCCT/SCMR/STS 2013 multimodal-
ity appropriate use criteria for the detection and risk assess-
ment of stable ischemic heart disease: a report of the American
Conclusions College of Cardiology Foundation Appropriate Use Criteria
Task Force, American Heart Association, American Society of
CCTA has proven to be a highly sensitive and reliable modal- Echocardiography, American Society of Nuclear Cardiology,
ity for the exclusion of significant coronary artery disease. Heart Failure Society of America, Heart Rhythm Society,
Society for Cardiovascular Angiography and Interventions,
When utilized in conjunction with techniques such as nonin- Society of Cardiovascular Computed Tomography, Society for
vasive fractional flow reserve, it yields excellent sensitivity – Cardiovascular Magnetic Resonance, and Society of Thoracic
on par with that of stress CMR – as well as good specificity, Surgeons. J Nucl Cardiol. 2014;21:192–220.
similar to that of SPECT and stress echocardiography. 10. Lee SP, Jang EJ, Kim YJ, et al. Cost-effectiveness of coronary CT
angiography in patients with chest pain: Comparison with myo-
Despite these data, a paucity of data exists for the use of cardial single photon emission tomography. J Cardiovasc Comput
CCTA in preoperative cardiac risk assessment. As such cur- Tomogr. 2015;9:428–37.
rent guidelines and appropriate use criteria classify CCTA as 11. Taylor AJ, Cerqueira M, Hodgson JM, et  al. ACCF/SCCT/ACR/
“rarely appropriate” as it pertains to preoperative testing. AHA/ASE/ASNC/NASCI/SCAI/SCMR 2010 appropriate use cri-
teria for cardiac computed tomography. A report of the American
While CCTA holds promise to be a valuable modality for College of Cardiology Foundation appropriate use criteria task
preoperative risk assessment, further investigation and larger force, the society of cardiovascular computed tomography, the
studies will be needed prior to its adoption into this arena. American College of Radiology, the American Heart Association,
the American Society of Echocardiography, the American
Society of Nuclear Cardiology, the North American Society for
Cardiovascular Imaging, the society for cardiovascular angiogra-
References phy and interventions, and the society for cardiovascular magnetic
resonance. Circulation. 2010;122:e525–55.
1. Devereaux PJ, Goldman L, Cook DJ, Gilbert K, Leslie K, Guyatt 12. Buffa V, De Cecco CN, Cossu L, et al. Preoperative coronary risk
GH. Perioperative cardiac events in patients undergoing noncardiac assessment with dual-source CT in patients undergoing noncoro-
surgery: a review of the magnitude of the problem, the pathophysi- nary cardiac surgery. Radiol Med. 2010;115:1028–37.
ology of the events and methods to estimate and communicate risk. 13. Silverman MG, Harkness JR, Blankstein R, et al. Baseline subclini-
CMAJ. 2005;173:627–34. cal atherosclerosis burden and distribution are associated with fre-
2. Goldman L, Caldera DL, Nussbaum SR, et al. Multifactorial index quency and mode of future coronary revascularization: multi-ethnic
of cardiac risk in noncardiac surgical procedures. N Engl J Med. study of atherosclerosis. JACC Cardiovasc Imaging. 2014;7:476–86.
1977;297:845–50. 14. Budoff MJ, Shaw LJ, Liu ST, et al. Long-term prognosis associated
3. Lee TH, Marcantonio ER, Mangione CM, et al. Derivation and pro- with coronary calcification: observations from a registry of 25,253
spective validation of a simple index for prediction of cardiac risk patients. J Am Coll Cardiol. 2007;49:1860–70.
of major noncardiac surgery. Circulation. 1999;100:1043–9. 15. Kelkar AA, Schultz WM, Khosa F, et  al. Long-term prognosis
4. Detsky AS, Abrams HB, McLaughlin JR, et al. Predicting cardiac after coronary artery calcium scoring among low-intermediate risk
complications in patients undergoing non-cardiac surgery. J Gen women and men. Circ Cardiovasc Imaging. 2016;9:e003742.
Intern Med. 1986;1:211–9. 16. Hwang JW, Kim EK, Yang JH, et  al. Assessment of periopera-
5. Fleisher LA, Beckman JA, Brown KA, et  al. ACC/AHA 2007 tive cardiac risk of patients undergoing noncardiac surgery using
guidelines on perioperative cardiovascular evaluation and care coronary computed tomographic angiography. Circ Cardiovasc
for noncardiac surgery: executive summary: a report of the Imaging. 2015;8:e002582.
American College of Cardiology/American Heart Association 17. Wijeysundera DN, Duncan D, Nkonde-Price C, et al. Perioperative
Task Force on Practice Guidelines (Writing Committee to beta blockade in noncardiac surgery: a systematic review for
Revise the 2002 Guidelines on Perioperative Cardiovascular the 2014 ACC/AHA guideline on perioperative cardiovascular
Evaluation for Noncardiac Surgery) Developed in Collaboration evaluation and management of patients undergoing noncardiac
­
With the American Society of Echocardiography, American surgery: a report of the American College of Cardiology/American
Society of Nuclear Cardiology, Heart Rhythm Society, Society Heart Association Task Force on Practice Guidelines. Circulation.
of Cardiovascular Anesthesiologists, Society for Cardiovascular 2014;130:2246–64.
Angiography and Interventions, Society for Vascular Medicine 18. Kristensen SD, Knuuti J, Saraste A, et  al. 2014 ESC/ESA

and Biology, and Society for Vascular Surgery. J Am Coll Cardiol. Guidelines on non-cardiac surgery: cardiovascular assessment and
2007;50:1707–32. management: The Joint Task Force on non-cardiac surgery: cardio-
6. Ahn JH, Park JR, Min JH, et  al. Risk stratification using computed vascular assessment and management of the European Society of
tomography coronary angiography in patients undergoing intermediate-­ Cardiology (ESC) and the European Society of Anaesthesiology
risk noncardiac surgery. J Am Coll Cardiol. 2013;61:661–8. (ESA). Eur Heart J. 2014;35:2383–431.
360 G. Jackson and R. R. Bayer II

19. Cohen MC, Aretz TH.  Histological analysis of coronary artery 26. Akhtar NJ, Markowitz AH, Gilkeson RC. Multidetector computed
lesions in fatal postoperative myocardial infarction. Cardiovasc tomography in the preoperative assessment of cardiac surgery
Pathol. 1999;8:133–9. patients. Radiol Clin N Am. 2010;48:117–39.
20. Dawood MM, Gutpa DK, Southern J, Walia A, Atkinson JB, Eagle 27. Bastarrika G, Thilo C, Headden GF, Zwerner PL, Costello P,

KA. Pathology of fatal perioperative myocardial infarction: impli- Schoepf UJ. Cardiac CT in the assessment of acute chest pain in the
cations regarding pathophysiology and prevention. Int J Cardiol. emergency department. AJR Am J Roentgenol. 2009;193:397–409.
1996;57:37–44. 28. Koo BK, Erglis A, Doh JH, et al. Diagnosis of ischemia-caus-
21. Duvall WL, Sealove B, Pungoti C, Katz D, Moreno P, Kim
ing coronary stenoses by noninvasive fractional flow reserve
M.  Angiographic investigation of the pathophysiology of peri- computed from coronary computed tomographic angio-
operative myocardial infarction. Catheter Cardiovasc Interv. grams. Results from the prospective multicenter DISCOVER-
2012;80:768–76. FLOW (Diagnosis of Ischemia-Causing Stenoses Obtained
22. Chopra V, Wesorick DH, Sussman JB, et  al. Effect of periopera- Via Noninvasive Fractional Flow Reserve) study. J Am Coll
tive statins on death, myocardial infarction, atrial fibrillation, and Cardiol. 2011;58:1989–97.
length of stay: a systematic review and meta-analysis. Arch Surg. 29. Min JK, Leipsic J, Pencina MJ, et  al. Diagnostic accuracy of

2012;147:181–9. fractional flow reserve from anatomic CT angiography. JAMA.
23. Andersson C, Merie C, Jorgensen M, et  al. Association of beta-­ 2012;308:1237–45.
blocker therapy with risks of adverse cardiovascular events and 30. Norgaard BL, Leipsic J, Gaur S, et al. Diagnostic performance
deaths in patients with ischemic heart disease undergoing noncar- of noninvasive fractional flow reserve derived from coronary
diac surgery: a Danish nationwide cohort study. JAMA Intern Med. computed tomography angiography in suspected coronary
2014;174:336–44. artery disease: the NXT trial (Analysis of Coronary Blood
24. Group PS, Devereaux PJ, Yang H, et  al. Effects of extended-­
Flow Using CT Angiography: Next Steps). J Am Coll Cardiol.
release metoprolol succinate in patients undergoing non-cardiac 2014;63:1145–55.
surgery (POISE trial): a randomised controlled trial. Lancet. 31. Danad I, Szymonifka J, Twisk JW, et al. Diagnostic performance of
2008;371:1839–47. cardiac imaging methods to diagnose ischaemia-causing coronary
25. Chopra V, Plaisance B, Cavusoglu E, Flanders SA, Eagle
artery disease when directly compared with fractional flow reserve as
KA.  Perioperative beta-blockers for major noncardiac surgery: a reference standard: a meta-analysis. Eur Heart J. 2017;38(13):991–8.
Primum Non Nocere. Am J Med. 2009;122:222–9.
 CT for Guiding Successful
Revascularization 31
Maksymilian P. Opolski

Both percutaneous coronary intervention (PCI) and coronary rial wall [2]. Consequently, it often underestimates lesion
computed tomography angiography (coronary CTA) belong length and vessel size compared with intravascular ultra-
to the most evolving fields in modern cardiology. In parallel sound (IVUS) and coronary CTA (specifically when diffuse
to the expanding indications for coronary CTA, a continually disease or negative remodeling are present) [7]. Whereas
increasing number of patients undergo PCI following coro- IVUS is still the current gold standard to optimize stent
nary CTA examination [1, 2]. In this regard, the abundance implantation [1], there is compelling evidence that measure-
of anatomic, morphological, and physiological data incorpo- ments of lumen area and diameter performed with coronary
rated from a one-stop-shop computed tomographic study can CTA and IVUS are well correlated [8], and noninvasive CTA
aid to optimize PCI strategy [3, 4]. The purpose of this chap- is the only imaging technique to quantify lesion length and
ter is to highlight contemporary and future advancements of vessel size before the procedure has even started. The con-
coronary CTA that might help guide PCI. cept of CTA-guided stent sizing was first reported in a simu-
lation study by LaBounty et al. who reevaluated 18 patients
with 24 coronary lesions and pre-procedural CTA datasets
Interventional Applications [9]. Interestingly, by using CTA operators selected longer
stents with larger stent diameters than with ICA. In another
Once a patient has been identified as having a flow-limiting small pilot study of 17 patients with 26 native coronary
plaque in need of stenting (Fig.  31.1), coronary CTA data lesions, Kass et  al. indicated relatively good agreement
can be used to better plan percutaneous strategy, particularly between CTA- and IVUS-derived stent sizes [10]. These
in angiographically demanding lesions such as coronary findings were subsequently replicated in 22 saphenous vein
chronic total occlusions (CTO), bifurcation lesions, or coro- graft lesions, wherein both coronary CTA and IVUS demon-
nary anomalies. Below a number of potential applications of strated similar measurements of the proximal and distal ref-
coronary CTA for PCI have been described. erence lumen diameters as well as lesion length [11]
Additionally, the assessment of stent length by coronary
CTA but not invasive coronary angiography (ICA) dictates
Selection of Stent Size and Length the exact actual length of the implanted stent as reported by
the stent’s manufacturer [12]. Of particular interest, in the
Numerous studies in the drug-eluting stent era have repeat- prospective randomized single-center study of 71 patients,
edly identified greater residual reference segment disease Pregowski et al. investigated whether the addition of coro-
and stent underexpansion as independent predictors of sub- nary CTA to PCI planning could affect immediate procedural
sequent stent edge restenosis and thrombosis [5, 6]. Hence, results as defined by IVUS endpoints [13]. To this end, angi-
complete lesion coverage and well-matched vessel diameters ography plus CTA-guided PCI versus angiography-guided
are paramount to decrease the risk of stent failure and PCI was compared, wherein computed tomographic criteria
improve clinical outcomes. Invasive coronary angiography for lumen area  >4.5  mm2 in the reference segments were
(ICA) represents a two-dimensional technique that cannot used to determine the optimal landing zone for stent implan-
reliably visualize the atherosclerotic involvement of the arte- tation. Notably, coronary stents selected under CTA guid-
ance were significantly longer with a trend toward larger
nominal stent diameter compared with angiography-­guided
M. P. Opolski (*) strategy alone (Fig. 31.2). This further translated into more
Department of Interventional Cardiology and Angiology,
complete lesion coverage and larger stent expansion as con-
Institute of Cardiology, Warsaw, Poland

© Humana Press 2019 361

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_31
362 M. P. Opolski

a b

Fig. 31.1  High-grade stenosis in coronary computed tomography angiography. (a) Curved multiplanar reconstruction showing high-grade lesion
in the mid-segment of the right coronary artery (arrow). (b) Corresponding invasive angiographic view

a b

Fig. 31.2  Computed tomography angiography-assisted planning of with proximal and distal stent diameters of 4.0 mm (cross section 1) and
percutaneous coronary intervention. (a) Curved multiplanar reconstruc- 3.0 mm (cross section 4), respectively. (b) Corresponding invasive angi-
tions of the high-grade stenosis in the mid-segment of the left anterior ographic view. In the angiography-guided strategy, the significant refer-
descending artery with corresponding cross-sectional areas (1 to 4). ence plaque burden (cross section 3) would likely be overlooked and
Based on computed tomographic data the planned stent length is 33 mm stent diameters underestimated

firmed by smaller peri-stent plaque burden and larger mini- [14]. Whereas CTA-assisted b­ifurcation PCI was equally
mal stent area in the CTA group. Lately Wolny et al. evaluated safe (with similar rate of periprocedural myocardial infarc-
the value of CTA in planning PCI strategy in bifurcation tion, amount of contrast and radiation dose during PCI) com-
lesions in a prospective randomized study of 93 patients pared with angiography-­ guided revascularization, it
31  CT for Guiding Successful Revascularization 363

prompted PCI operators to choose simpler PCI technique  laque Assessment for Planning PCI
P
with higher use of single-stent procedures and proximal opti- Interventions
mization technique along with less frequent side-branch
stenting. Possible explanation of this observation is that Several CT studies have consistently confirmed the close
angiography-defined lesion length is often underestimated, relationship between high-risk features of coronary plaques
leading to a selection of too short a stent – a phenomenon and invasive FFR [18–21]. In the study by Nakazato et al. in
that may be occasionally associated with suboptimal imme- 58 patients with 58 intermediate coronary lesions, percent
diate angiographic result (i.e., stent edge dissection or plaque aggregate plaque volume (cumulative plaque volume as a
protrusion caused by stent placement in a region with high function of total vessel volume) showed the highest predic-
plaque burden) – and thus requires additional stent implanta- tive value for identification, discrimination, and reclassifica-
tion. Whether the anatomic approach for CTA-guided stent tion of ischemic lesions compared with anatomic
implantation may also improve lesion-­ specific ischemia measurements of coronary stenosis [18]. In the larger multi-
compared to the standard-of-care angiography-­guided PCI is center evaluation of 407 stenoses in 252 patients from the
yet to be determined. Until these trials are completed, selec- DeFACTO study, positive remodeling was an independent
tion of stent size and length based on CTA data will likely be predictor of ischemia across all coronary lesions, whereas
slowly incorporated into clinical practice, notwithstanding percent aggregate plaque volume and low-attenuation plaque
the absence of clear guideline recommendation. were associated with ischemia in obstructive lesions (≥50%)
only [19]. Further, in the substudy of the NXT trial of 254
patients with 484 vessels employing semiautomated CT soft-
I schemia-Guided PCI by Computerized ware, low-density noncalcified plaque volume ≥30 mm3 was
Tomography Fractional Flow Reserve the only independent determinant of ischemia with incre-
mental predictive value over CTA stenosis >50% [20].
Given the ascendancy of ischemia-guided intervention over The presence, location, and extent of calcification on CTA
anatomic-guided intervention [15], the ability of computer- might potentially affect the results of catheter-based inter-
ized tomography fractional flow reserve (FFR-CT) to detect ventions. For example, for lesions with severe calcium
lesion-specific ischemia opens a door for FFR-CT-guided (mostly defined as calcification >50% of vessel cross sec-
PCI. In the RIPCORD study, 200 consecutive patients from tion), plaque modification with rotablation, intravascular
the NXT trial underwent sequential decision-making for the lithotripsy or high-pressure balloons may be required to
most optimal treatment strategy (optimal medical therapy achieve complete stent expansion [22]. Whereas coronary
alone, PCI, coronary artery bypass grafting, or more infor- dissections are often encountered at the junction of calcified
mation required) based on coronary CTA alone or the hybrid and noncalcified plaque, the optimal landing zone for stent
anatomic-physiological approach of coronary CTA plus implantation in calcium-free vessel segments should be con-
FFR-CT [16]. The decision was made by consensus of three sidered. The potential difficulties in calcified lesions have
experienced interventional cardiologists reporting on the been emphasized in a study showing that hard plaques with
changes in their management plan after disclosure of CT density  >120 Hounsfield units were associated with
FFR-CT. Overall, there was a change in management in 36% occurrence of dissection or perforation during PCI [23].
of patients, with a 23% increase in the use of optimal medi- Moreover, plaques with a higher calcium score had larger
cal therapy alone, a 5% decrease in PCI, and a 0.5% increase reference plaque burden following stent implantation and
in CABG. Among patients assigned to PCI, in 18% the target were more likely to require balloon post-dilation under IVUS
vessel for PCI was changed, whereas 30% were reallocated guidance [24].
to optimal medical therapy. Not surprisingly, FFR-CT can Assessment of plaque composition by coronary CTA can
also optimize PCI by dictating the landing zone for stent also predict the occurrence of distal emboli during PCI that
implantation necessary to resolve ischemia (Fig. 31.3). In a may ultimately result in periprocedural myocardial infarc-
pilot study of 48 ischemia-causing stenoses with post-PCI tion [25–29]. In the study by Harigaya et  al. including 78
invasive FFR data, the ability of FFR-CT to predict the func- patients with acute coronary syndrome or stable angina, the
tional benefit of coronary revascularization by “virtual stent- presence of CTA-derived low-attenuation plaque with a
ing” was tested [17]. Specifically, “virtual stenting” was length  >4.7  mm was an independent predictor of the no-­
performed by modification of the computational model in reflow phenomenon [27]. Consequently, Kodama et al. found
the region of the virtual stent according to the proximal and that low-attenuation plaque surrounded by circumferential
distal reference areas, and the diagnostic accuracy of FFR-CT plaque calcification conferred a higher risk of slow-flow phe-
to predict ischemia after stenting was 96%. Whether FFR-CT nomenon after PCI in a series of 40 patients [28]. Also, a
could determine which method of coronary revascularization larger prospective study of 189 stable patients corroborated
maximally reduces ischemia (“virtual stenting” versus “vir- the decisive role of low-attenuation plaque in predicting peri-
tual bypass surgery”) based on patient-specific anatomy is to procedural myocardial injury as confirmed by positive tropo-
be elucidated. nin-­T at 18 h post-PCI [26]. Ultimately, large analysis of 180
364 M. P. Opolski

a b

1 1

0.8 0.8

0.6 0.6

Fig. 31.3 Noninvasive treatment planning using fractional flow the left anterior descending artery (arrow). (b) Computerized tomogra-
reserve derived from computed tomographic data with virtual stenting phy fractional flow reserve showing resolution of ischemia in the left
algorithm. (a) Computerized tomography fractional flow reserve dem- anterior descending coronary artery after virtual stenting (arrow)
onstrating functionally significant stenosis in the proximal segment of

patients with non-ST-segment elevation acute coronary volume, and positive remodeling measured by CTA por-
­syndrome and pre-procedural CTA yielded low-attenuation tended the occurrence of in-stent restenosis with incremental
plaque and napkin-ring sign as independent predictors of predictive value (sensitivity of 90% and specificity of 84%).
slow-flow phenomenon in culprit lesions [29]. All of these Logical explanation for this observation may reside in the
studies reinforce the notion that coronary plaques with large inherent properties of a diffusely distributed low-attenuation
lipid area may be prone to being broken down into small plaque that could promote inflammatory endothelial response
atherosclerotic particles following balloon dilatation, with along with exaggerated neointimal proliferation as a reaction
subsequent embolic complications of the distal coronary cir- to stent strut penetration. In this regard, visualization of a
culation (Fig. 31.4). In this regard, identification of high-risk low-attenuation plaque with positive remodeling on CTA
plaques on CTA may help in shifting PCI strategy toward the should be considered as a “bad omen” (increasing the risk of
use of direct stenting, distal protection devices, or stronger both slow-flow phenomenon and in-stent restenosis) for PCI
antithrombotic therapy (e.g., glycoprotein IIb/IIIa receptor operators. Potential solution for mitigating the risk for in-­
inhibitors). Another option might include deferring PCI and stent restenosis should involve selection of well-expanded
starting statin therapy for plaque stabilization. Finally, in drug-eluting stents with minimal late lumen loss that cover
patients with multiple high-risk plaques (particularly involv- the entire length of coronary plaque.
ing left main or proximal left anterior descending coronary
artery), coronary bypass grafting rather than PCI could be
considered. Vulnerable Plaque
Coronary CTA can also predict mechanical complications
of stent implantation. Recently Tesche et  al. proposed an The concept of vulnerable plaque is based on the assumption
elegant method derived from pre-procedural CTA for pre- that certain coronary plaques are more prone to rupture than
dicting in-stent restenosis as defined by quantitative ICA others (Fig.  31.5) [31]. Given the compelling evidence for
[30]. In a retrospective analysis of 74 coronary lesions, their multifocal nature of vulnerable plaques as well as the strenu-
main observation was that lesion length, noncalcified plaque ous effort for invasive intracoronary imaging of all three
31  CT for Guiding Successful Revascularization 365

a b

c d

Fig. 31.4  Coronary computed tomography angiography for character- cification in the obtuse marginal branch (arrow). (b) Corresponding
ization of a high-risk plaque resulting in a no-reflow phenomenon fol- invasive angiographic view confirming severe stenosis in the obtuse
lowing percutaneous coronary intervention. (a) Curved multiplanar marginal branch (arrow). (c) Stent positioning. (d) No-reflow in the left
reconstructions along with corresponding cross-sectional areas (1 to 7) circumflex artery and slow-flow in the obtuse marginal branch after
show severely stenotic low-attenuation plaque with circumferential cal- stent implantation

coronary vessels, noninvasive CTA has become one of the n­ on-ST-­segment elevation myocardial infarction undergoing
most versatile techniques for prediction of future coronary coronary CTA, the total amount of noncalcified plaque in
events [32]. In a landmark study by Motoyama et al. includ- nonobstructive coronary lesions was an independent predic-
ing 1059 patients with stable chest pain, the presence of low-­ tor of subsequent coronary events [34]. Moreover, the focus
attenuation plaque and/or positive remodeling conferred a on napkin-ring sign morphology (defined as the presence of
23% risk for future coronary event during a mean follow-up a ring of high attenuation around coronary plaque) was
of 27  months [33]. In a study of 312 patients with undertaken in a prospective study of 895 patients [35]. After
366 M. P. Opolski

a b

Fig. 31.5  Coronary computed tomography angiography for character- with positive remodeling and a napkin-ring sign (cross sections 1 to 3)
ization of a vulnerable plaque causing future ischemic event. (a) Curved along with isolated circumferential contrast accumulation that might
multiplanar reconstructions of the left anterior descending artery with correspond to plaque ulceration (cross section 4). (b) Invasive angio-
corresponding cross-sectional areas (1 to 4) at the site distal to the ori- gram showing acute occlusion of the proximal left anterior descending
gin of the first diagonal branch demonstrate low-attenuation plaque artery at follow-up (arrowhead). SB side-branch

a mean of 2.3  years, both the napkin-ring sign along with duration, CTA reproducibly demonstrates a complete lack of
positive remodeling and low-attenuation plaque showed an contrast opacification within the occlusion site with subse-
independent relationship with future episodes of acute coro- quent contrast enhancement in the distal vessel segment
nary syndrome. Finally, in the latest meta-analysis across six (Fig. 31.6) [22].
CT studies, the risk of future coronary event was signifi- Unlike ICA, coronary CTA allows for complete visualiza-
cantly higher in high-risk plaques (defined by at least one of tion of the occlusion site and distal vessel segment [22, 38].
the following criterions: low attenuation, increased remodel- This is particularly apparent in long and tortuous CTO
ing index, or the presence of spotty calcification) versus low-­ lesions or when the distal CTO segment cannot be entirely
risk plaques with odds ratio of 12.14 [36]. On a practical visualized in ICA (e.g., patients after coronary artery bypass
level, the potential clinical use of CTA to detect vulnerable grafting) [22, 39]. Furthermore, it has been consistently
plaques would require the development of effective preven- established in several trials that CTA reliably quantifies mor-
tive measures mitigating the risk of subsequent plaque rup- phological (i.e., calcification, number of occlusion sites)
ture [31]. Futuristic approaches may involve vulnerable [39–42] and anatomical (e.g., tortuosity, length) [40, 41, 43]
plaque sealing with coronary stents or biodegradable vascu- features of CTO. Three-dimensional reconstructions enable
lar scaffolds. the visualization of the exact vessel trajectory with accurate
measurement of the occlusion length avoiding calibration
limitations or lesion foreshortening [7]. Also, CTA can be
Coronary Chronic Total Occlusions used for providing guidance for the location of ostial CTO
and multiple occlusion sites, characteristics usually missed
A CTO may be incidentally identified on a diagnostic coro- in ICA [39, 42].
nary CTA, or CTA may be specifically performed for a better All features described above have been used in a number
characterization of CTO (usually in cases with a previously of CT studies predicting the procedural outcome with CTO
failed recanalization attempt) [22]. In the former case, the PCI (Table 31.1) [38, 40, 42–52]. Mollet et al. were the first
prevalence of CTO is non-negligible and may reach up to to show that calcifications involving >50% of the vessel
6.2% in patients with obstructive coronary artery disease cross-sectional area along with an occlusion length >15 mm
[37]. Although CTO are defined based on the angiographic were independent predictors of guidewire crossing failure in
criterion of antegrade blood flow interruption of ≥3 months’ a series of 47 CTO lesions [43]. In a study by Ehara et al.
31  CT for Guiding Successful Revascularization 367

a b

Fig. 31.6  Chronic total occlusion in coronary computed tomography right coronary artery (arrow). The vessel distal to the occluded segment
angiography. (a) Multiplanar reformation demonstrating noncalcified is opacified with contrast. (b) Corresponding invasive angiographic
occlusion that lacks contrast opacification in the mid-segment of the view. (From Opolski and Achenbach [22], with permission)

Table 31.1  Studies investigating the ability of computed tomography angiography to predict successful percutaneous coronary intervention in
chronic total occlusion
No of Success
Author, year Type of CT CTO Study endpoint rate Independent predictors of PCI failure
Mollet et al. 2005 16-slice 47 Guidewire crossing 55% Calcium >50% CSA
[43] Occlusion length >15 mm
Soon et al. 2007 16-slice 43 Guidewire crossing with stenting 56% Calcium ≥50% CSA
[40] (TIMI 3 flow and <25% stenosis)
Ehara et al. 2009 64-slice 110 Guidewire crossing 85% Bending (>45°)
[44] Shrinkage (<1 mm in vessel diameter)
Severe calcium (~360° CSA)
García-García et al. 16/64-­slice 142 Guidewire crossing (TIMI 2–3 flow 63% Calcium >50% CSA
2009 [38] and <50% stenosis)
Cho et al. 2010 64-slice 72 Guidewire crossing with stenting 76% %Calcium area/CSA
[45] (TIMI 3 flow)
Li et al. 2010 [46] 64-slice 74 Guidewire crossing (TIMI 2–3 flow 77% Occlusion length
dual source and <20% stenosis) Severity of calcification (noncalcified,
spot-­calcified, arc-calcified,
circular-calcified)
Hsu et al. 2011 64-slice 82 Guidewire crossing (TIMI 2–3 flow 89% Calcification length ratio (calcification
[47] and <30% stenosis) length/occlusion length) >0.5
Choi et al. 2011 64-slice 186 Guidewire crossing (TIMI 3 flow 77% Occlusion length >18 mm
[48] and <30% stenosis) Segmental density >139HU
CTO duration >12 months or unknown
Martín-Yuste et al. 64-slice 77 Guidewire crossing 62% Calcium arc/CSA
2012 [49]
Li et al. 2013 [50] 64-slice 88 Guidewire crossing (TIMI 3 flow 58% Linear intrathrombus enhancement
dual source and <25% stenosis)
Opolski et al. 2015 64/128-­slice 240 Guidewire crossing within 30 min 62% Calcium >50% CSA
[42] dual source Bending (>45°)
Multiple occlusion sites
Blunt stump
CTO duration ≥12 months or unknown
Previously failed PCI at CTO
(continued)
368 M. P. Opolski

Table 31.1 (continued)
No of Success
Author, year Type of CT CTO Study endpoint rate Independent predictors of PCI failure
Luo et al. 2015 256-­slice 108 Guidewire crossing (TIMI 3 flow 74% Negative remodeling
[51] and <20% stenosis without MACE) Occlusion length ≥32 mm
Chen et al. 2015 64-slice 281 Guidewire crossing (TIMI 3 flow 85% Attenuation of proximal CTO segment
[52] and <25% stenosis) Occlusion length
Total coronary calcium score
CSA cross-sectional area, CT computed tomography, CTO chronic total occlusion, MACE major adverse cardiac events, PCI percutaneous coro-
nary intervention, TIMI thrombolysis in myocardial infarction

including 110 CTO lesions, bending (defined as an angle with a good retrograde collateral flow, makes intuitive sense for
>45°) was the most powerful predictor of wire failure (fol- choosing the retrograde approach. Furthermore, by showing a
lowed by vessel shrinkage and severe calcium) [44]. highly tortuous, calcified and/or long CTO, the “knuckle-wire”
Subsequently, García-García et  al. found that calcification technique can be recommended for easier tracking of the curved
>50% of vessel cross section was the only independent cor- path of occluded vessel segments. Finally, CT-derived scores
relator of failed PCI among 142 CTO, conferring an almost may be used to assess CTO difficulty level, both to allocate
sixfold higher risk for procedural failure [38]. Consistent resources and improve training programs (e.g. by assigning less
findings were consequently replicated by other authors, experienced operators to simple CTO lesions) (Fig. 31.8).
highlighting the deleterious role of severe calcium in ham-
pering successful guidewire passage [40, 45, 46–49]. There
are conflicting data regarding the impact of CTA-derived Coronary Anomalies
occlusion length on PCI result. Whereas some studies repeat-
edly indicated that long CTO (defined as either >15  mm Although coronary anomalies are uncommon, they consti-
or  >18  mm or  ≥32  mm or continuous variable) confer a tute a major diagnostic puzzle for most of the cardiologists
higher risk for procedural failure [43, 46, 48, 50–52], others interpreting ICA results and may thus mislead clinicians into
failed to demonstrate any relationship between occlusion employing inappropriate therapeutic decisions. This issue is
length and PCI outcome [38, 40, 42, 44, 45, 47, 49]. In lieu further complicated because anomalous arteries significantly
of these findings, the relevance of CT-derived severe calcifi- increase the difficulty level of coronary revascularization
cation and occlusion length cannot be overemphasized as (not to mention high contrast and radiation loads) – a phe-
both of these features (along with minimal vessel area) are nomenon that keeps awake most of the interventional cardi-
associated with adverse clinical outcomes after PCI of CTO ologists and cardiac surgeons. Fortunately, CTA comes as a
[53]. More recently, the CT-RECTOR score was introduced “release” for accurate characterization of the origin, course,
as a noninvasive robust prediction tool to predict successful and termination of coronary anomalies in a one-stop-shop
guidewire crossing through CTO within 30 min [42]. It was examination [55]. In a prospective study by Pregowski et al.,
derived from 240 consecutive CTO lesions from a multi- ostial anomalies comprised 24% of all nondiagnostic cathe-
center European registry, and the outcome variable was spe- terizations, all of which were readily seen on coronary CTA
cifically set to represent the difficulty level intrinsic to CTO [56]. Specifically, CTA can be used to locate coronary ostia
rather than operator skills and perseverance. Of note, the (e.g., anomalous origin from the opposite aortic sinus, high
CT-RECTOR score is calculated by assigning one point for origin of the coronary artery) and overcome hurdles related
each of the independent predictors (multiple occlusions, to failed catheter placement [55]. Also, CTA can depict the
blunt stump, severe calcification, bending, CTO exact position of the coronary artery in relation to the aorta
age ≥12 months, previously failed PCI) and adding them per and pulmonary artery for accurate differentiation between
CTO lesion (Fig. 31.7). With an increasing score, the likeli- the potentially malignant and benign courses of the anoma-
hood of guidewire crossing ≤30 min decreased, ranging from lous artery originating from the opposite aortic sinus [55].
95% for a CT-RECTOR score of 0 to 22% for a CT-RECTOR To some extent, multiplanar and three-dimensional volume-
score of ≥3. As an alternative, the noninvasive counterpart of rendered CT reconstructions can identify the high-­risk fea-
the commonly used angiographic J-CTO score (the so called tures of the interarterial course (e.g., slitlike ostium, acute
J-CTOCT score) has been shown high predictive value for angle of takeoff, or stretch of the intramural segment within
predicting guidewire crossing ≤30 min [54]. the aortic wall) and thus help guide revascularization strate-
In summary, the entirety of information proffered by CTA gies [57]. To date, only single case reports have reported the
can affect the technique of PCI for CTO. For example, identifi- modification of interventional treatment strategies in coro-
cation of severe and diffuse calcium at the entry site, together nary anomalies based on CTA datasets [58, 59].
31  CT for Guiding Successful Revascularization 369

Fig. 31.7  Calculation sheet CT-RECTOR Score Calculator

for the CT-RECTOR score
with illustrated definitions of Predictors Definitions
each variable and listing of
Multiple Occlusion Multiple Occlusion
the difficulty groups. (From
Opolski et al. [42], with
permission) Presence of ≥2 complete interruptions Presence (1)
of the contrast opacification separated
by contrast-enhanced segment of ≥5 mm. Absence (0)

Blunt Stump Blunt Stump

Absence of any tapered stump Presence (1)

at the entry or exit site.
Absence (0)

Severe Calcification Severe Calcification

Presence of any calcium involving Presence (1)

≥50% of the vessel cross-sectional
Absence (0)
area at the entry or exit site or within
the occlusion route.

Bending ≥45° Bending ≥45°

Presence of any bending ≥45º Presence (1)

at the entry or exit site or within
Absence (0)
the occlusion route.

Second Attempt Second Attempt

Previously failed PCI at CTO Yes (1)

No (0)

Duration of CTO Duration of CTO

Duration of CTO ≥12 months or unknown Yes (1)

No (0)

Difficulty Group Total Score

Easy (0) Difficult (2)

Intermediate (1) Very Difficult (≥3)

 oronary CTA for Periprocedural

C laboratory for periprocedural guidance of PCI. To date, one
Guidance of PCI approach has been to project three-dimensional CTA data
onto separate monitors in the catheterization laboratory. This
Coronary CTA can be directly used in the catheterization application allows operators to simultaneously display CTA
370 M. P. Opolski

a b

c d

Fig. 31.8  Very difficult lesion from the guidewire-failure group. (a) ≥50% vessel area. (c) Three-dimensional reconstruction displaying
Curved multiplanar reconstruction displaying two occlusion sites (yel- bending of 55° within the occlusion route. (d) Coronary angiography
low asterisks) and severe calcification in the mid-segment of the right after failed attempt of guidewire crossing within 55 min. (From Opolski
coronary artery. (b) Cross-sectional view of the calcification occupying et al. [42], with permission)

projections and fluoroscopic angulations for identification of devices superimposing augmented reality data onto the user’s
the most optimal working angles without foreshortening and field of view has garnered particular attention in the context
can be particularly used during guidewire advancement in of periprocedural planning for PCI. The advantage of wear-
CTO with suboptimal visualization in ICA [22]. Indeed, this able computers relies on limitless display and intuitive
has been affirmed in a study by Rolf et al. who showed sig- manipulation of multiple imaging data using simple hand
nificantly higher success rates for CTO PCI with periproce- gestures, voice commands, or motion tracking [60]. Indeed,
dural CTA monitors compared with PCI without CTA overlay recently Google Glass worn by cardiologists in the catheter-
[41]. Alternatively, the concept of wearable h­ ead-­mounted ization laboratory has been used for successful CTO PCI by
31  CT for Guiding Successful Revascularization 371

a b

Fig. 31.9  Computed tomography data displayed in a wearable com- View, CA) (arrows). (b) Three-dimensional computed tomographic
puter during recanalization attempt of a chronic total occlusion. (a) reconstruction projected directly onto the Google Glass showing the
Interventional cardiologists viewing computed tomographic images in exact trajectory of the right coronary artery (arrow). (From Opolski
the upper right visual field on Google Glass (Google Inc., Mountain et al. [61], with permission)

providing access to a three-dimensional CTA “roadmap” of also associated with a numerically higher recanalization
the occluded vessel (Fig. 31.9) [61]. rates compared with cases in which the position of the guide-
Alternatively, a more computationally intense approach wire was indeterminate (83% vs 63%). However, not every
of real-time fusion of three-dimensional coronary CTA with CTA result provided adequate information about guidewire
X-ray fluoroscopy has been introduced (Fig.  31.10). In a location (not to mention potentially higher labor input, time,
small series of 43 CTO patients who underwent attempted and radiation dose of intraprocedural CTA scanning).
PCI based on CTA data in combination with magnetic navi-
gation, only a moderate success rate of 44% was achieved
[62]. This may reflect the essential shortcomings of the Future Perspectives
fusion techniques (such as patient, cardiac, and respiratory
motion artifacts) resulting in a relatively wide margin of reg- Most likely, future advancements for guiding revasculariza-
istration errors that consequently compromise the submilli- tion procedures by coronary CTA would evolve into further
metric precision required for CTO PCI (e.g., differentiating validation and wider adoption of FFR-CT, steady improve-
intraluminal versus subintimal guidewire location). ments in CT technology for better characterization of coro-
Nevertheless, recently Ghoshhajra et  al. raised hope for a nary plaque, and ongoing technical developments for mobile
wider adoption of coupling CTA and angiographic data for display of CTA datasets in the catheterization laboratory.
CTO PCI by showing that CTA-derived extent and localiza- Specifically, coupling the anatomical and physiological
tion of coronary calcification along with vessel tortuosity information by coronary CTA may be the best approach for
may influence the strategy of antegrade dissection reentry accurate stent selection. Also, the accuracy of CTA is a
technique [63]. This however did not translate into higher “moving target” and could be further enhanced by wider
success rates compared with PCI without CTA overlay. adoption of automated plaque assessment tools, machine
Most likely, the best “remedy” for real-time integration of learning algorithms, and ­computational fluid dynamics to
CTA datasets with fluoroscopic environment is to bring CT provide time-saving multidimensional information on indi-
scanner directly to the catheterization laboratory. Although it vidual plaques prior to PCI.  Finally, the value of wearable
might appear to be a herculean task, Kim et al. installed an computers for projection of CTA data to the catheterization
advanced CTA system in the catheterization room (allowing laboratory has yet to receive widespread attention.
for interchangeable CT scanning and invasive fluoroscopy)
to investigate its role for the identification of guidewire loca-
tion during CTO PCI in a series of 61 patients [64]. By on-­ Conclusions
site review of the CT images in a 360-degree view, the exact
position of the guidewire tip was classified as intraluminal, The ability of coronary CTA to show and characterize ath-
subintimal, or outside the vessel wall. Interestingly, this erosclerotic plaque can assist interventional cardiologists
study demonstrated that successful prediction of the guide- with planning percutaneous coronary intervention strate-
wire location by intraprocedural CTA is not only feasible but gies. This is particularly salient in cases where noninvasive
372 M. P. Opolski

a b

Fig. 31.10  Image fusion of coronary computed tomography angiogra- fluoroscopic data are combined to show coronary segments in green (no
phy and fluoroscopic images. (a) Three-dimensional anatomy of the foreshortening) or red (substantial foreshortening). Prototype, Siemens
coronary arteries derived from computed tomography angiography Healthcare, Erlangen, Germany. (From Opolski and Achenbach [22],
facilitates identification of projections without foreshortening for each with permission)
respective vessel segment. (b) Computed tomography angiography and

CTA exceeds the characterization of coronary lesions com- References

pared to invasive angiogram (e.g., coronary chronic total
occlusions, bifurcation lesions, or coronary anomalies) [22]. 1. Windecker S, Kolh P, Alfonso F, et al. 2014 ESC/EACTS Guidelines
on myocardial revascularization. Eur Heart J. 2014;35:2541–619.
Moreover, the importance of noninvasive CTA detection of 2. Achenbach S. Coronary CTA and percutaneous coronary interven-
“adverse plaque characteristics” (mainly low-attenuation tion - a symbiosis waiting to happen. J Cardiovasc Comput Tomogr.
plaque) that confer a higher risk for periprocedural compli- 2016;10:384–5.
cations cannot be overemphasized in the context of tailored 3. Min JK, Chandrashekhar Y, Narula J. Diagnosis of coronary disease
and icing on the cake. JACC Cardiovasc Imaging. 2015;8:1117–20.
catheter interventions [25–29]. Of particular interest, the 4. Wink O, Hecht HS, Ruijters D. Coronary computed tomographic
ability of FFR-CT to pinpoint lesion-specific ischemia goes angiography in the cardiac catheterization laboratory: current appli-
far beyond enhancing diagnostic performance of coronary cations and future developments. Cardiol Clin. 2009;27:513–29.
CTA and may aid in ischemia-­guided “virtual stenting” pro- 5. Sonoda S, Morino Y, Ako J, et al. Impact of final stent dimensions
on long-term results following sirolimus-eluting stent implantation:
cedures – a phenomenon not evaluable by any other single serial intravascular ultrasound analysis from the sirius trial. J Am
imaging technique [17]. In this context, if coronary CTA is Coll Cardiol. 2004;43:1959–63.
available prior to percutaneous intervention, it would be 6. Sakurai R, Ako J, Morino Y, et al. Predictors of edge stenosis fol-
desirable that the operator takes good note of all the infor- lowing sirolimus-eluting stent deployment (a quantitative intravas-
cular ultrasound analysis from the SIRIUS trial). Am J Cardiol.
mation provided starting from lesion-specific ischemia and 2005;96:1251–3.
ending up with optimal viewing angles, true lesion length 7. van Velzen JE, de Graaf MA, Ciarka A, et al. Non-invasive assess-
(most often underestimated in ICA), reference vessel diam- ment of atherosclerotic coronary lesion length using multidetector
eters, and morphological plaque characteristics. Noteworthy, computed tomography angiography: comparison to quantitative
coronary angiography. Int J Cardiovasc Imaging. 2012;28:2065–71.
this task may be substantially simplified by projection of 8. Leber AW, Knez A, von Ziegler F, et al. Quantification of obstructive
CTA datasets directly to the catheterization laboratory. and nonobstructive coronary lesions by 64-slice computed tomog-
Although more data are clearly needed to uncover the full raphy: a comparative study with quantitative coronary angiography
potential of CTA for planning PCI, coronary CTA is already and intravascular ultrasound. J Am Coll Cardiol. 2005;46:147–54.
9. LaBounty T, Sundaram B, Chetcuti S, et  al. Stent size selection
gaining recognition as the only imaging tool for crossing the using 64-detector coronary computed tomography angiography:
borders between noninvasive and invasive management of a comparison to invasive coronary angiography. Acad Radiol.
coronary artery disease. 2008;15:820–6.
31  CT for Guiding Successful Revascularization 373

10. Kass M, Glover CA, Labinaz M, So DY, Chen L, Yam YCB. Lesion 26. Uetani T, Amano T, Kunimura A, et  al. The association between
characteristics and coronary stent selection with computed tomo- plaque characterization by CT angiography and post-procedural
graphic coronary angiography: a pilot investigation comparing myocardial infarction in patients with elective stent implantation.
CTA, QCA and IVUS. J Invasive Cardiol. 2010;22:328–34. JACC Cardiovasc Imaging. 2010;3:19–28.
11. Pregowski J, Kepka C, Kalinczuk L, et  al. Comparison of intra- 27. Harigaya H, Motoyama S, Sarai M, et al. Prediction of the no-
vascular ultrasound, quantitative coronary angiography, and reflow phenomenon during percutaneous coronary intervention
dual-source 64-slice computed tomography in the preprocedural using coronary computed tomography angiography. Heart Vessel.
assessment of significant saphenous vein graft lesions. Am J 2011;26:363–9.
Cardiol. 2016;107:1453–9. 28. Kodama T, Kondo T, Oida A, et al. Computed tomographic angi-
12. Ciszewski M, Zalewska J, Pregowski J, et al. Comparison of stent ography–verified plaque characteristics and slow-flow phenom-
length reported by the stent’s manufacturer to that determined by enon during percutaneous coronary intervention. JACC Cardiovasc
quantitative coronary angiography at the time of implantation ver- Interv. 2012;5:636–43.
sus that determined by coronary computed tomographic angiogra- 29. Miura K, Kato M, Dote K, et al. Association of nonculprit plaque
phy at a later time. Am J Cardiol. 2013;111:1111–6. characteristics with transient slow flow phenomenon during percu-
13. Pregowski J, Kepka C, Kruk M, et  al. Comparison of usefulness taneous coronary intervention. Int J Cardiol. 2015;181:108–13.
of percutaneous coronary intervention guided by angiography plus 30. Tesche C, De Cecco CN, Vliegenthart R, et  al. Coronary CT

computed tomography versus angiography alone using intravascu- angiography-­derived quantitative markers for predicting in-stent
lar ultrasound end points. Am J Cardiol. 2011;108:1728–34. restenosis. J Cardiovasc Comput Tomogr. 2016;10:377–83.
14. Wolny R, Pregowski J, Kruk M, et al. Computed tomography angi- 31. Opolski MP, Kepka C, Ruzyłło W.  Computed tomography for

ography versus angiography for guiding percutaneous coronary detection of vulnerable coronary plaque  - a Cassandra’s dream?
interventions in bifurcation lesions – A prospective randomized Postepy Kardiol Interwencyjnej. 2014;10:147–52.
pilot study. J Cardiovasc Comput Tomogr. 2017;11:119–28. 32. Maurovich-Horvat P, Ferencik M, Voros S, et  al. Comprehensive
15. Tonino PAL, De Bruyne B, Pijls NHJ, et al. Fractional flow reserve plaque assessment by coronary CT angiography. Nat Rev Cardiol.
versus angiography for guiding percutaneous coronary interven- 2014;11:390–402.
tion. N Engl J Med. 2009;360:213–24. 33. Motoyama S, Sarai M, Harigaya H, et al. Computed tomographic
16. Curzen NP, Nolan J, Zaman AG, et al. Does the routine availabil- angiography characteristics of atherosclerotic plaques subse-
ity of CT–derived FFR influence management of patients with quently resulting in acute coronary syndrome. J Am Coll Cardiol.
stable chest pain compared to CT angiography alone?: the FFRCT 2009;54:49–57.
RIPCORD study. JACC Cardiovasc Imaging. 2016;9:1188–94. 34. Kristensen TS, Kofoed KF, Kühl JT, et al. Prognostic implications of
17. Kim K-H, Doh J-H, Koo B-K, et al. A novel noninvasive technol- nonobstructive coronary plaques in patients with non–ST-segment
ogy for treatment planning using virtual coronary stenting and com- elevation myocardial infarction: a multidetector computed tomog-
puted tomography-derived computed fractional flow reserve. JACC raphy study. J Am Coll Cardiol. 2011;58:502–9.
Cardiovasc Interv. 2014;7:72–8. 35. Otsuka K, Fukuda S, Tanaka A, et al. Napkin-ring sign on coronary
18. Nakazato R, Shalev A, Doh J-H, et al. Aggregate plaque volume by CT angiography for the prediction of acute coronary syndrome.
coronary computed tomography angiography is superior and incre- JACC Cardiovasc Imaging. 2013;6:448–57.
mental to luminal narrowing for diagnosis of ischemic lesions of 36. Thomsen C, Abdulla J. Characteristics of high-risk coronary

intermediate stenosis severity. J Am Coll Cardiol. 2013;62:460–7. plaques identified by computed tomographic angiography and asso-
19. Park H-B, Heo R, ó Hartaigh B, et al. Atherosclerotic plaque char- ciated prognosis: a systematic review and meta-analysis. Eur Heart
acteristics by CT angiography identify coronary lesions that cause J Cardiovasc Imaging. 2016;17:120–9.
ischemia: a direct comparison to fractional flow reserve. JACC 37. Opolski MP, ó Hartaigh B, Berman DS, et  al. Current trends in
Cardiovasc Imaging. 2015;8:1–10. patients with chronic total occlusions undergoing coronary CT
20. Gaur S, Øvrehus KA, Dey D, et  al. Coronary plaque quantifica- angiography. Heart. 2015;101:1212–8.
tion and fractional flow reserve by coronary computed tomogra- 38. García-García HM, Van Mieghem H, Gonzalo N, et al. Computed
phy angiography identify ischaemia-causing lesions. Eur Heart J. tomography in total coronary occlusions (CTTO registry): radiation
2016;37:1220–7. exposure and predictors of successful percutaneous intervention.
21. Tesche C, De Cecco CN, Caruso D, et al. Coronary CT angiography EuroIntervention. 2009;4:607–16.
derived morphological and functional quantitative plaque markers 39. Opolski MP, Kepka C, Achenbach S, et  al. Coronary computed
correlated with invasive fractional flow reserve for detecting hemo- tomographic angiography for prediction of procedural and interme-
dynamically significant stenosis. J Cardiovasc Comput Tomogr. diate outcome of bypass grafting to left anterior descending artery
2016;10:199–206. occlusion with failed visualization on conventional angiography.
22. Opolski MP, Achenbach S.  CT angiography for revascularization Am J Cardiol. 2012;109:1722–8.
of CTO: crossing the borders of diagnosis and treatment. JACC 40. Soon KH, Cox N, Wong A, et al. CT coronary angiography predicts
Cardiovasc Imaging. 2015;8:846–58. the outcome of percutaneous coronary intervention of chronic total
23. Kinohira Y, Akutsu Y, Li H-L, et  al. Coronary arterial plaque occlusion. J Interv Cardiol. 2007;20:359–66.
characterized by multislice computed tomography predicts 41. Rolf A, Werner GS, Schuhbäck A, et al. Preprocedural coronary CT
complications following coronary intervention. Int Heart J. angiography significantly improves success rates of PCI for chronic
2007;48:25–33. total occlusion. Int J Cardiovasc Imaging. 2013;29:1819–27.
24. Pregowski J, Jastrzebski J, Kępka C, et al. Relation between coro- 42. Opolski MP, Achenbach S, Schuhbäck A, et al. Coronary computed
nary plaque calcium deposits as described by computed tomogra- tomographic prediction rule for time-efficient guidewire crossing
phy coronary angiography and acute results of stent deployment through chronic total occlusion: insights from the CT-RECTOR multi-
as assessed by intravascular ultrasound. Postepy Kardiol center registry (computed tomography registry of chronic total occlu-
Interwencyjnej. 2013;9:115–20. sion revascularization). JACC Cardiovasc Interv. 2015;8:257–67.
25. Nakazawa G, Tanabe K, Onuma Y, et al. Efficacy of culprit plaque 43. Mollet NR, Hoye A, Lemos PA, et al. Value of preprocedure mul-
assessment by 64-slice multidetector computed tomography to pre- tislice computed tomographic coronary angiography to predict the
dict transient no-reflow phenomenon during percutaneous coronary outcome of percutaneous recanalization of chronic total occlusions.
intervention. Am Heart J. 2008;155:1150–7. Am J Cardiol. 2005;95:240–3.
374 M. P. Opolski

44. Ehara M, Terashima M, Kawai M, et al. Impact of multislice com- CT angiography to invasive angiography. J Cardiovasc Comput
puted tomography to estimate difficulty in wire crossing in percuta- Tomogr. 2015;9:578–84.
neous coronary intervention for chronic total occlusion. J Invasive 55. Earls JP.  Coronary artery anomalies. Tech Vasc Interv Radiol.

Cardiol. 2009;21:575–82. 2006;9:210–7.
45. Cho JR, Kim YJ, Ahn C-M, et  al. Quantification of regional cal- 56. Pregowski J, Kepka C, Kruk M, et al. The clinical significance and
cium burden in chronic total occlusion by 64-slice multi-detector management of patients with incomplete coronary angiography and
computed tomography and procedural outcomes of percutaneous the value of additional computed tomography coronary angiogra-
coronary intervention. Int J Cardiol. 2010;145:9–14. phy. Int J Cardiovasc Imaging. 2014;30:825–32.
46. Li P, Gai L, Yang X, et  al. Computed tomography angiography-­ 57. Opolski MP, Pregowski J, Kruk M, et al. Prevalence and character-
guided percutaneous coronary intervention in chronic total occlu- istics of coronary anomalies originating from the opposite sinus of
sion. J Zhejiang Univ Sci B. 2010;11:568–74. valsalva in 8,522 patients referred for coronary computed tomogra-
47. Te HJ, Kyo E, Chu CM, et al. Impact of calcification length ratio phy angiography. Am J Cardiol. 2013;111:1361–7.
on the intervention for chronic total occlusions. Int J Cardiol. 58. Jo Y, Uranaka Y, Iwaki H, et al. Sudden cardiac arrest associated
2011;150:135–41. with anomalous origin of the right coronary artery from the left
48. Choi J-H, Bin SY, Hahn J-Y, et al. Three-dimensional quantitative main coronary artery. Tex Hear Inst J. 2011;38:539–43.
volumetry of chronic total occlusion plaque using coronary multi- 59. Kiefer TL, Vavalle J, Halim S, et  al. Anterograde percutaneous
detector computed tomography. Circ J. 2011;75:366–75. coronary–cameral fistula closure employing a guide-in-guide tech-
49. Martín-Yuste V, Barros A, Leta R, et  al. Determinantes del éxito nique. JACC Cardiovasc Interv. 2013;6:1105–7.
de la revascularización de las oclusiones coronarias crónicas: estu- 60. Dugas CM, Schussler JM. Advanced technology in interventional
dio mediante tomografía computarizada con multidetectores. Rev cardiology: a roadmap for the future of precision coronary interven-
Española Cardiol. 2012;65:334–40. tions. Trends Cardiovasc Med. 2016;26:466–73.
50. Li M, Zhang J, Pan J, Lu Z. Coronary total occlusion lesions: linear 61. Opolski MP, Debski A, Borucki BA, et al. First-in-man computed
intrathrombus enhancement at CT predicts better outcome of per- tomography-guided percutaneous revascularization of coronary
cutaneous coronary intervention. Radiology. 2013;266:443–51. chronic total occlusion using a wearable computer: proof of con-
51. Luo C, Huang M, Li J, et al. Predictors of interventional success of cept. Can J Cardiol. 2016;32:829.e11–3.
antegrade PCI for CTO. JACC Cardiovasc Imaging. 2015;8:804–13. 62. Ramcharitar S, van der Giessen W, van der Ent M, et al. The fea-
52. Chen Y, Lu B, Hou Z, et  al. Predicting successful percutaneous sibility and safety of applying the magnetic navigation system to
coronary intervention in patients with chronic total occlusion: the manage chronically occluded vessels: a single Centre experience.
incremental value of a novel morphological parameter assessed by EuroIntervention. 2011;6:711–6.
computed tomography. Int J Cardiovasc Imaging. 2015;31:1263–9. 63. Ghoshhajra BB, Takx RAP, Stone LL, et al. Real-time fusion of
53. Ito T, Tsuchikane E, Nasu K, et al. Impact of lesion morphology on coronary CT angiography with x-ray fluoroscopy during chronic
angiographic and clinical outcomes in patients with chronic total total occlusion PCI. Eur Radiol. 2017;27:2464–73.
occlusion after recanalization with drug-eluting stents: a multislice 64. Kim B-K, Cho I, Hong M-K, et al. Usefulness of intraprocedural cor-
computed tomography study. Eur Radiol. 2015;25:3084–92. onary computed tomographic angiography during intervention for
54. Li Y, Xu N, Zhang J, et  al. Procedural success of CTO recanali- chronic total coronary occlusion. Am J Cardiol. 2016;117:1868–76.
zation: comparison of the J-CTO score determined by coronary
 Stent Assessment
32
Junjie Yang, Christian Tesche, Taylor M. Duguay,
Lucas L. Geyer, and Yundai Chen

Percutaneous coronary intervention (PCI) with stent place- for exclusion of in-stent restenosis (ISR) [4–6]. However,
ment is a standard therapy for myocardial revascularization blooming and beam-hardening artifacts caused by stent
of hemodynamically significant coronary artery disease struts and stents with diameters <3  mm are existing chal-
(CAD). Although PCI procedures are being performed at lenges that limit the diagnostic accuracy of CCTA in the
increasing rates, in-stent restenosis (ISR) after stent place- evaluation on stent patency [7, 8]. Improvements in image
ment remains a frequent complication [1, 2]. Three genera- acquisition and post-processing, including several new
tions of stents have been introduced, which include approaches to enhance the evaluation of stent patency, ISR,
bare-metal stents, durable polymer drug-eluting stents, and and stent structure, have been developed. Iterative recon-
biodegradable-polymer drug-eluting stents. An appropriate struction techniques, use of high convolution kernels, and
noninvasive imaging technique to serve as a post-procedural dual-energy CT (DECT) acquisitions have shown promising
follow-up strategy and for the evaluation of stent patency has results [9–12].
been desired for some time. Coronary CT angiography
(CCTA) has been studied extensively for this purpose and
has been incorporated into current accepted guidelines for In-Stent Restenosis and Stent Fracture
the evaluation of larger coronary stents (≥3 mm) [3]. Recent
technical refinements have improved the accuracy of CCTA The introduction of bare-metal stents in the mid-1980s was a
stent evaluation, showing negative predictive values >90% milestone in the evolution of PCI. However, the appearance
of ISR with related increased morbidity and mortality was a
drawback of this technique. Incidences of ISR up to 20–30%
J. Yang · Y. Chen after bare-metal stenting leading to repeated target vessel
Division of Cardiovascular Imaging, Department of Radiology and revascularization have been reported [13]. Newer-generation
Radiological Science, Medical University of South Carolina, drug-eluting stents have decreased the incidence of ISR sig-
Charleston, SC, USA
nificantly as compared to bare-metal stents; however ISR
Department of Cardiology, People’s Liberation Army General rates between 5% and 30% are still observed [14, 15]. Several
Hospital, Beijing, China
procedural factors, lesion-related characteristics, and patient
C. Tesche (*) attributes are related to the development of ISR [16, 17].
Division of Cardiovascular Imaging, Department of Radiology and
Angiographic factors such as vessel size, lesion length, loca-
Radiological Science, Medical University of South Carolina,
Charleston, SC, USA tion, and plaque burden as well as clinical factors such as
smoking status, diabetes mellitus, and inflammatory status
Department of Cardiology and Intensive Care Medicine, Heart
Center Munich-Bogenhausen, Munich, Germany attribute to ISR. Furthermore, stent characteristics like strut
thickness and mechanical problems associated with stent
T. M. Duguay
Division of Cardiovascular Imaging, Department of Radiology and deployment contribute to ISR. Stent fracture is a well-­
Radiological Science, Medical University of South Carolina, established cause of ISR typically observed after drug-­
Charleston, SC, USA eluting stent implantation due to the reduced stent strut
L. L. Geyer thickness and fewer stent cell connections. These character-
Division of Cardiovascular Imaging, Department of Radiology and istics have led to weaker longitudinal strength in the new-­
Radiological Science, Medical University of South Carolina,
generation DES [18]. A representative example of in-stent
Charleston, SC, USA
restenosis is illustrated in Fig. 32.1.
Institute for Clinical Radiology, Ludwig-Maximilians-University
Hospital Munich, Munich, Germany

© Humana Press 2019 375

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_32
376 J. Yang et al.

a tation of ECG-synchronized acquisition protocols and the

use of CT scanners with high temporal resolution, e.g., dual-
source CT (DSCT) systems providing up to 75 ms, are the
b foundation of current cardiac CT imaging. In addition to
these features, optimal patient preparation is required,
including the appropriate use of beta-blockers for heart rate
control and nitroglycerin for optimal coronary artery visual-
ization and reduction of motion artifacts. However, residual
coronary motion remains a major problem as it causes blur-
ring and, as stents are high-contrast objects, it intensifies the
adverse effect of blooming on the images.

Blooming Artifacts

Beam-hardening and partial-volume artifacts are the main

causes of the artificial thickening of stent struts with
CCTA.  This phenomenon is more commonly known as a
“blooming artifact.” This blooming effect results in signifi-
cant luminal narrowing and an underestimation of the stent
lumen. The impact and degree of blooming artifacts are
influenced by image acquisition and reconstruction parame-
ters as well as stent type and diameter. Blooming artifacts
increase with strut thickness and the corresponding density
Fig. 32.1  A 55-year-old man presenting with chest pain. (a) CCTA of the strut material. Furthermore, overlapping stents and
using curved multi-planar reconstruction with sinogram-affirmed itera- additional calcifications in the stent segment aggravate
tive reconstruction demonstrates presence of in-stent restenosis in the
3.5 mm diameter drug-eluting stent of the right coronary artery (arrow).
beam-hardening effects resulting in further impairment of
(b) Boxed area clarifies the cross-section view CT stent assessment. With current state-of-the-art CT sys-
tems, visualization of stent lumen diameter ranges between
50% and 80% [19, 20]. Newly biodegradable-polymer drug-­
CT Assessment of Coronary Stents eluting stents may overcome the limitation of blooming arti-
facts caused by stent struts and may allow for sufficient stent
Contrast Enhancement assessment. A representative example of a biodegradable-­
polymer drug-eluting stent is shown in Fig. 32.2.
Sufficient intravascular contrast enhancement is a prerequi-
site for adequate delineation of lumen and stent visibility.
High contrast is crucial as factors like image noise (due to Tube Potential
sharp convolution kernels in the presence of stents) reduce
the contrast-to-noise ratio. Furthermore, appropriate window The appropriate use of a high kV acquisition is another
setting requires optimal intravascular attenuation to allow approach to reduce blooming artifacts and thus improve stent
correct identification of vessel lumen, intimal hyperplasia visualization [21]. However, the increase in kV level results
within the stent, and stent struts. To achieve high-contrast in a significantly higher radiation exposure to the patient and
concentration in the coronary vessel, appropriate contrast is therefore not recommended. In an in vitro study, Sirineni
agents and injection protocols with bolus tracking or test et al. investigated the impact of different kV settings on the
bolus application are essential. visualization of stent luminal diameter and showed no sig-
nificant improvement in stent lumen visibility when increas-
ing from 120 to 140 kV [22].
Motion Artifacts

Coronary artery and respiratory motions and irregular or Reconstruction Kernels

high heart rates deteriorate CCTA image quality. As stent
assessment is challenging, per se, the importance of follow- The applied convolution kernel for image reconstruction has
ing acquisition practice according to current guidelines for the greatest impact of stent lumen visualization and the appro-
the best possible image quality is imperative. The implemen- priate ability for stent patency evaluation. For the ­visualization
32  Stent Assessment 377

Fig. 32.2  A 43-year-old male b

a
patient with a biodegradable-­
polymer drug-eluting stent in
the right coronary artery
15 months after stent
implantation. Curved planar
reconstruction for both B26
kernel (a) and B46 kernel (b)
illuminated the presence of
remaining biodegradable stent
markers. The long arrow
indicates the proximal end of
the biodegradable-polymer
drug-eluting stent, and short
arrow points to the distal end

and assessment of coronary artery plaques, a medium smooth is vital due to the fact that increased image noise is the main
convolution kernel is recommended while an edge-enhancing drawback of technical solutions such as high-­resolution ker-
kernel is necessary for the delineation of the stent lumen. nels or submillimeter data acquisition. In recent years, itera-
These vendor-specific dedicated sharp convolution kernels tive reconstruction algorithms have been introduced and have
(high-spatial-frequency reconstruction algorithms) for stent successively replaced filtered back projection as the standard
evaluation lead to an enhancement of edges in high-attenua- method of data reconstruction in routine CT imaging [25].
tion structures like stents. This results in a significant reduc- Concerning image quality, iterative reconstruction techniques
tion in blooming artifact severity; however adversely focus on reducing artifacts, such as streak artifacts caused by
increased image noise is observed hindering the evaluation of dense objects and image noise. As a result, iterative recon-
coronary plaques in non-stented coronary segments [23]. struction has shown significant improvements in CT stent
Recently, a combined approach using both kernels for image imaging. Besides a potential dose reduction, the key feature is
reconstruction that were post-processed with a corresponding a more accurate assessment of the stent lumen resulting in a
filter algorithm investigated lumen visibility, intraluminal more precise detection of in-stent stenosis [4]. Furthermore,
attenuation, and image noise. However, the use of edge- iterative reconstruction facilitates the implementation of inte-
enhanced post-processing filter failed to compensate the pro- grated circuit detectors offering a minimal section thickness of
nounced blooming artifacts caused by stent struts and was 0.5 mm without increasing image noise. Therefore, the visual-
proven inferior to the dedicated edge-­enhancing kernel [24]. ization of small-diameter stents (< 3 mm) can be improved [5].

Spatial Resolution Monoenergetic Reconstruction with DECT

Technical improvements in CT technology leading to ongo- DECT offers another approach to reduce beam-hardening arti-
ing improved spatial resolution are the main key for sufficient facts which occur if polychromatic X-ray energy spectra are
stent assessment. Third-generation dual-source CT (DSCT) used for image reconstruction. In contrast, image generation
provides isotropic submillimeter resolution to visualize the from monochromatic (or monoenergetic) X-rays will not lead
coronary arteries. These improvements of spatial resolution to such artifacts. Mathematical methods allow the calculation
also result in an optimized in-stent assessment with changes of virtual monoenergetic images at different levels of kiloelec-
of in-stent lumen visibility from 30% to 60% with the use of tron volt (keV) from DECT data. As a consequence, artifacts
16-slice and 64-slice CT systems, respectively, and up to 80% caused by metallic objects, e.g., implants, can be significantly
with third-generation DSCT systems [19]. reduced [10]. CT stent assessment can also benefit from vir-
tual monoenergetic imaging. Although luminal CT numbers
decrease at high keV levels, the overall in-stent visualization
Iterative Reconstruction and accuracy of diameter measurements increase remarkably
due to the reduction of blooming artifacts. Even the lumen
Although blooming artifacts are mainly caused by beam-­ visualization of small stents (<3  mm) can be improved by
hardening and partial-volume effects, the reduction of image using advanced methods of virtual monoenergetic imaging
noise allows for significantly improved stent visualization and when compared to standard, polychromatic CT imaging [12].
378 J. Yang et al.

Fig. 32.3  A 46-year-old male

patient with new onset of
a b
dyspnea and angina
22 months after placement of
overlapping stents in the right
coronary artery. (a) Maximum
intensity projection and (b)
curved planar reconstruction
reveal stent fracture (white
arrows) with stent
malposition. Hypodensity
indicates in-stent restenosis
mainly observed at the stent
gaps

Clinical Evidence a

With the introduction of four-slice CT systems by all major

vendors in 1998 and gating techniques for multi-slice CT in
2000, CT became more important as a noninvasive imaging
technique for stent evaluation (Figs. 32.3 and 32.4). However,
the limited spatial and temporal resolution did not allow for
appropriate visualization of stent lumen visibility; thus stent
patency was assessed indirectly by evaluating the visual
assessment of the distal runoff [26]. Stent lumen visibility
ranged between 20% and 40% with an unacceptable high b
number of excluded stents (up to 80–90% due to insufficient
lumen visibility) [27]. When 16-slice CT systems enabled
direct visualization of the stent lumen with improved tempo-
ral resolution and submillimeter spatial resolution, CT
became more accepted for the assessment of ISR.  Studies
using 16-slice CT reported sensitivities and specificities of
54–100% and 88–100%, respectively [28]. However, a major
drawback was the high number of unevaluable stents result-
ing in an exclusion rate up to 14%. With the introduction
64-slice CT systems, noninvasive stent assessment with CT
was widely accepted. In a recent meta-analysis, the diagnos-
tic value of 64-slice CT for the detection of ISR was signifi-
cantly higher than that of 16-slice CT angiography (91 vs
81%) as a result of the increased spatial and temporal resolu-
tion [6]. Pooled sensitivity and specificity of 87% and 95%
were reported. The diagnostic accuracy of ISR detection is
highly dependent on stent diameter. For stents with a diam- Fig. 32.4  A 65-year-old male with a drug-eluting stent in the left main
eter of ≤3 mm, sensitivity was 54%, whereas for stent diam- to proximal left anterior descending artery. (a) Curved multi-planar
eter with ≥3  mm, sensitivity increased up to 86%. Thus, reconstruction shows uneven shortening of the stent strut in in the prox-
based on the results of 64-slice CT systems, current guide- imal portion of the stent (arrow) due to longitudinal compression. (b)
Cross-sectional image demonstrates diffuse in-stent restenosis
lines recommend CCTA for the assessment of stents ≥3 mm.
32  Stent Assessment 379

References 14. Palmerini T, Biondi-Zoccai G, Della Riva D, et  al. Clinical out-
comes with bioabsorbable polymer- versus durable polymer-based
drug-eluting and bare-metal stents: evidence from a comprehensive
1. Smith SC Jr, Dove JT, Jacobs AK, et  al. ACC/AHA guidelines
network meta-analysis. J Am Coll Cardiol. 2014;63:299–307.
for percutaneous coronary intervention (revision of the 1993
15. Cohen DJ, Van Hout B, Serruys PW, et al. Quality of life after PCI
PTCA guidelines)-executive summary: a report of the American
with drug-eluting stents or coronary-artery bypass surgery. N Engl
College of Cardiology/American Heart Association task force on
J Med. 2011;364:1016–26.
practice guidelines (Committee to revise the 1993 guidelines for
16. Holmes DR Jr, Kereiakes DJ, Garg S, et al. Stent thrombosis. J Am
percutaneous transluminal coronary angioplasty) endorsed by the
Coll Cardiol. 2010;56:1357–65.
Society for Cardiac Angiography and Interventions. Circulation.
17. Mehran R, Dangas G, Abizaid AS, et al. Angiographic patterns of
2001;103:3019–41.
in-stent restenosis: classification and implications for long-term
2. Kim LK, Feldman DN, Swaminathan RV, et al. Rate of percutane-
outcome. Circulation. 1999;100:1872–8.
ous coronary intervention for the management of acute coronary
18. Park KW, Park JJ, Chae IH, et al. Clinical characteristics of coro-
syndromes and stable coronary artery disease in the United States
nary drug-eluting stent fracture: insights from a two-center des reg-
(2007 to 2011). Am J Cardiol. 2014;114:1003–10.
istry. J Korean Med Sci. 2011;26:53–8.
3. Raff GL, Abidov A, Achenbach S, et al. SCCT guidelines for the
19. Gassenmaier T, Petri N, Allmendinger T, et  al. In vitro compari-
interpretation and reporting of coronary computed tomographic
son of second- and third-generation dual-source CT for coro-
angiography. J Cardiovasc Comput Tomogr. 2009;3:122–36.
nary stent visualization at different tube potentials. Acad Radiol.
4. Ebersberger U, Tricarico F, Schoepf UJ, et  al. CT evaluation of
2016;23:961–8.
coronary artery stents with iterative image reconstruction: improve-
20. Gebhard C, Fiechter M, Fuchs TA, et  al. Coronary artery stents:
ments in image quality and potential for radiation dose reduction.
influence of adaptive statistical iterative reconstruction on
Eur Radiol. 2013;23:125–32.
image quality using 64-HDCT.  Eur Heart J Cardiovasc Imaging.
5. Geyer LL, Glenn GR, De Cecco CN, et al. CT evaluation of small-­
2013;14:969–77.
diameter coronary artery stents: effect of an integrated circuit detec-
21. Suzuki S, Furui S, Kuwahara S, et  al. Assessment of coronary
tor with iterative reconstruction. Radiology. 2015;276:706–14.
stent in  vitro on multislice computed tomography angiography:
6. Sun Z, Almutairi AM.  Diagnostic accuracy of 64 multislice CT
improved in-stent visibility by the use of 140-kV tube voltage. J
angiography in the assessment of coronary in-stent restenosis: a
Comput Assist Tomogr. 2007;31:414–21.
meta-analysis. Eur J Radiol. 2010;73:266–73.
22. Sirineni GK, Kalra MK, Pottala K, Waldrop S, Syed M, Tigges
7. Sheth T, Dodd JD, Hoffmann U, et al. Coronary stent assessabil-
S. Effect of contrast concentration, tube potential and reconstruc-
ity by 64 slice multi-detector computed tomography. Catheter
tion kernels on MDCT evaluation of coronary stents: an in  vitro
Cardiovasc Interv. 2007;69:933–8.
study. Int J Cardiovasc Imaging. 2007;23:253–63.
8. Chung SH, Kim YJ, Hur J, et  al. Evaluation of coronary artery
23. Seifarth H, Raupach R, Schaller S, et al. Assessment of coronary
in-stent restenosis by 64-section computed tomography: factors
artery stents using 16-slice MDCT angiography: evaluation of a
affecting assessment and accurate diagnosis. J Thorac Imaging.
dedicated reconstruction kernel and a noise reduction filter. Eur
2010;25:57–63.
Radiol. 2005;15:721–6.
9. Eisentopf J, Achenbach S, Ulzheimer S, et  al. Low-dose dual-­
24. Ulrich A, Burg MC, Raupach R, et  al. Coronary stent imaging
source CT angiography with iterative reconstruction for coronary
with dual-source CT: assessment of lumen visibility using differ-
artery stent evaluation. JACC Cardiovasc Imaging. 2013;6:458–65.
ent convolution kernels and postprocessing filters. Acta Radiol.
10. Bamberg F, Dierks A, Nikolaou K, Reiser MF, Becker CR, Johnson
2015;56:42–50.
TR. Metal artifact reduction by dual energy computed tomography
25. Geyer LL, Schoepf UJ, Meinel FG, et al. State of the art: iterative
using monoenergetic extrapolation. Eur Radiol. 2011;21:1424–9.
CT reconstruction techniques. Radiology. 2015;276:339–57.
11. Meinel FG, Bischoff B, Zhang Q, Bamberg F, Reiser MF, Johnson
26. Kruger S, Mahnken AH, Sinha AM, et  al. Multislice spiral com-
TR. Metal artifact reduction by dual-energy computed tomography
puted tomography for the detection of coronary stent restenosis and
using energetic extrapolation: a systematically optimized protocol.
patency. Int J Cardiol. 2003;89:167–72.
Investig Radiol. 2012;47:406–14.
27. Maintz D, Grude M, Fallenberg EM, Heindel W, Fischbach

12. Mangold S, Cannao PM, Schoepf UJ, et al. Impact of an advanced
R. Assessment of coronary arterial stents by multislice-CT angiog-
image-based monoenergetic reconstruction algorithm on coronary
raphy. Acta Radiol. 2003;44:597–603.
stent visualization using third generation dual-source dual-energy
28. Sun Z, Davidson R, Lin CH. Multi-detector row CT angiography in
CT: a phantom study. Eur Radiol. 2016;26:1871–8.
the assessment of coronary in-stent restenosis: a systematic review.
13. Chen MS, John JM, Chew DP, Lee DS, Ellis SG, Bhatt DL. Bare
Eur J Radiol. 2009;69:489–95.
metal stent restenosis is not a benign clinical entity. Am Heart J.
2006;151:1260–4.
 Multidetector CT Angiography
for Coronary Bypass Graft Assessment 33
and Reoperative Cardiac Surgery

Lloyd M. Felmly

Coronary Bypass Grafting coagulopathy, and temporary myocardial dysfunction asso-

ciated with CPB and may decrease the risk of stroke associ-
Coronary artery bypass grafting (CABG) is a surgical ated with aortic cross-clamping. However it is important to
method of treating coronary artery disease that involves the note that distal anastomotic patency may be compromised
harvest of conduits from the patient’s own vasculature, fol- due to the increased technical difficulty of these anastomoses
lowed by myocardial revascularization by the implantation [3], especially in the hands of a surgeon who does not rou-
of these autologous grafts into the coronary arteries distal to tinely perform off-pump surgery. Minimally invasive tech-
known obstructions. While the number of CABGs has niques have been employed in an on- or off-pump fashion to
steadily declined in recent years, it remains the most com- avoid sternotomy while bypassing certain coronary territo-
mon cardiac surgery and one of the most common surgeries ries. More recently, studies on bypass grafting using all-­
worldwide, with approximately 800,000 CABGs performed arterial bypass grafts have shown improved patency rates
yearly [1]. The decline in CABG volume can be correlated and outcomes [4]. Reoperative CABG is undertaken in a
with improved medical management and the proliferation of similar fashion to primary CABG but with important differ-
percutaneous coronary intervention (PCI), but it continues to ences. Entry to the chest is typically via redo sternotomy,
be a relevant method of treating coronary disease. which can be time-consuming and hazardous. In addition,
Radiographic assessment of the coronary arteries is integral choice of grafts may be limited due to prior graft harvesting.
to the ability to perform bypass surgery. However its conduct, CABG as an operation is highly depen-
Proximally, graft blood flow may originate from the ves- dent on imaging to define the course and stenoses of the
sel’s native takeoff as for in situ right (RIMA) and left inter- coronary vasculature and in reoperative surgery, to define the
nal mammary arteries (LIMA), and less commonly the mediastinal anatomy and course of prior bypass grafts.
gastroepiploic artery, or from a proximal aortic anastomosis
in the case of saphenous vein, radial artery, or pedicled IMA
grafts. The LIMA was originally used as the primary graft  istorical Comment on CT for the Evaluation
H
early in the history of CABG, especially for bypassing the of Coronary Bypass Grafts
left anterior descending artery (LAD); however it fell from
favor due to the larger lumens of saphenous vein graft and Computed tomography (CT) has been used to radiographi-
greater technical ease of implanting these grafts. However cally characterize cardiac disease since the 1970s. The excel-
the LIMA re-emerged as a critical aspect of any CABG oper- lent temporal resolution of electron beam computed
ation after a landmark paper from Loop et al. at the Cleveland tomography (EBCT), approximately 100  ms, allowed the
Clinic demonstrated improved patency and long-term sur- accurate detection of coronary artery calcium, a useful tool
vival when this graft was used to bypass the LAD [2]. CABG to assess disease and predict the risk of adverse events such
can be conducted with the use of cardiopulmonary bypass as heart attack or stroke. Coronary calcium scanning is still a
(CPB) and cardioplegic arrest or on a beating heart with the useful risk-predictive tool; however the data it provides is
aid of stabilizing devices. This “off-pump” approach allows limited to assessment of calcific plaque load. CT imaging of
the surgeon to avoid the systemic inflammatory response, the coronaries combined with concurrent administration of
iodinated contrast media in the arterial phase, known as cor-
L. M. Felmly (*) onary computed tomographic angiography (CCTA), allows
Department of Surgery, Division of Cardiothoracic Surgery, for a more detailed analysis of vessel lumens as well as non-
Medical University of South Carolina, Charleston, SC, USA
calcific plaque characteristics. Conventional CT with con-
e-mail: [email protected]

© Humana Press 2019 381

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_33
382 L. M. Felmly

Fig. 33.1  Occluded bypass

grafts imaged on a 3rd
a b
generation DSCT scanner
with high fidelity. The takeoff
of the occluded vein graft to
the left circumflex artery and
distally occluded LIMA graft
can be visualized (a), as can
the proximally occluded vein
graft to the right coronary
artery (b)

trast attenuation was able to image venous grafts; however of 96.9%, 95.5%, 93.9%, 97.7%, and 96.1%, respectively
this was limited to basic binary assessment of patency versus [12] (Fig. 33.1).
occlusion [5–7]. The introduction of spiral CT, which These hardware advances in MDCT have contributed to the
reduced motion artifact, allowed adequate assessment of reduction in motion artifact. Changes in technique have allowed
venous bypass grafts, but visualizing distal anastomotic sites reductions in iodinated contrast medium volume and radiation
as well as arterial grafts remained a challenge [8]. While the exposure, allowing for safer scanning. Software innovations
temporal resolution of EBCT was far superior to that of con- such as new reconstructive techniques have also improved
ventional and spiral CT scanners at the time of its introduc- image quality. Taken together these advances have allowed
tion, it failed to achieve widespread use, mostly due to high MDCT to overtake EBCT as the dominant cross-­sectional and
cost and lack of indications outside of cardiac scanning. 3D imaging modality for coronary arteries and bypass grafts.
Technological advances in the more commonplace spiral CT, Future directions will likely focus on improved reconstructive
specifically subsecond rotation time, the addition of detector algorithms to further decrease artifact from motion, ECG dys-
rows to acquire multiple slices simultaneously, and the synchrony, and objects with high attenuation, as well as contin-
development of electrocardiogram (ECG) synchronization, ued development of detector technology.
allowed for spiral CT to take a larger role in the evaluation of
cardiac pathology. Decreases in acquisition time in single-
slice CT improved visualization of graft patency for both Preoperative Applications
arterial and venous grafts [9], but the progression from sin-
gle-slice spiral scanners to multidetector computed tomogra- CCTA is a useful methodology for the diagnosis and assess-
phy (MDCT) represented a critical jump in CT scanner ment of CAD. The validity of CCTA as an initial diagnostic
technology. Early 4-slice scanners had improved temporal test for CAD has been studied extensively and is covered in
resolution compared to single-slice scanners, between 125 previous chapters. Instead we will focus on the use of CCTA
and 250  ms at a 0.5  s gantry rotation time, and brought in the context of preoperative CABG planning. In the last
MDCT of the coronaries to a broad audience. There have several years, there have been several studies evaluating
been great advances in multi-­row detector technology since CCTA as a sole diagnostic test with which to proceed with
this time, related to increasing detector rows and therefore CABG [13, 14], showing sensitivities and specificities as
greater simultaneous slice acquisition, dual-source acquisi- high as 98.5% and 99.1% in detecting significant stenoses
tion which reduces gantry rotation time, and improved [15]. Despite advances in MDCT technology allowing CT
reconstructive algorithms. Coronary arteries can now be reli- to approach the spatial and temporal resolutions of typical
ably followed out to third-order segmental coronary arteries fluoroscopy systems, at 0.16mm and 30 ms [16], respec-
with MDCT [10], and temporal resolution now compares tively, invasive coronary angiography (ICA) remains the test
favorably to EBCT.  Recent generations of dual-source CT of choice among interventionalists and cardiac surgeons in
(DSCT) scanners allow acquisition of 384 slices with a gan- proceeding CABG, likely with some of this preference due
try rotation time of 0.25 s, producing a temporal resolution to familiarity [17]. However, as the accuracy of CCTA
of 66  ms and a spatial resolution of 0.24  mm [11]. Third- improves with newer generations of scanners, functional
generation DSCT has been shown to diagnose significant flow data becomes more available, and surgeon familiarity
coronary stenoses with sensitivity, specificity, positive pre- with CCTA increases, it is possible that CT may become the
dictive value, negative predictive value (NPV), and accuracy “gatekeeper” to not only the catheterization lab but the oper-
ating room as well.
33  Multidetector CT Angiography for Coronary Bypass Graft Assessment and Reoperative Cardiac Surgery 383

Computed tomography fractional flow reserve (CT-FFR) is raphy, and identification of stenosis proximal to the LIMA
a new technique that determines the hemodynamic significance origin, which carries a prevalence of 2–7% [26, 27], allows
of coronary artery stenoses via flow dynamics and computa- for adaptive strategies such as stenting of the subclavian
tional algorithms. This is in comparison to invasive catheter arteries or use of alternative bypass grafts [28]. While CCTA
FFR, the current gold standard, which relies on proximal and has not been extensively studied in this respect, it has the
distal stenosis pressure measurements to determine whether a benefit over catheter evaluation of being noninvasive. In
lesion is flow-limiting and therefore a likely source of isch- addition, routine preoperative catheter evaluation of the
emia. CT-FFR is discussed in detail elsewhere in this text, and IMAs and/or subclavian arteries may not be common in
therefore we will only briefly make note of its applications to some practices, and therefore abnormalities of this vascular
CABG. CT-FFR is currently not widely available due to the tree may be missed. As CCTA offers comprehensive evalua-
complexity of the employed algorithms, which require transfer tion of the thoracic vasculature, it is reasonable to assume
of data to an off-site supercomputer; however multiple studies that this modality would have a higher sensitivity for detect-
have shown that it compares favorably to invasive ing abnormalities due to a more consistent technique. The
FFR. Invasive FFR has been studied in the context of CABG author’s institution recently performed a reoperative CABG
and has revealed that less significant lesions are associated in which the initial operation had been an all-saphenous vein
with decreased graft patency [18], likely due to competitive graft CABG at another hospital, with the LIMA still in its
flow, and that certain lesion characteristics, such as diffuse native position. Prior operative reports and notes were not
instead of focal disease, are also associated with decreased available; however as we perform preoperative CCTA before
graft patency [19], likely due to impaired runoff. CT-FFR all redo cases, we identified a proximal left subclavian artery
has not been extensively studied in the context of planning stenosis that was not identified on preoperative catheteriza-
for CABG; however Curzen et al. demonstrated that provid- tion, allowing us to change our operative strategy and use
ing CT-FFR data changed management plans in more than alternative grafts (Fig. 33.2).
one third of patients with significant CAD when compared to While the majority of CABG procedures in the USA are
CCTA data alone [20]. While decisions to proceed to CABG performed in an open fashion through a median sternotomy,
are not routinely based on CCTA data, as CCTA for bypass
planning evolves and CT-FFR becomes more widespread,
CT-FFR may become more important in these management
decisions.
While CCTA has not been widely adopted as the sole
imaging test before coronary surgery, it offers more informa-
tion on cardiac and mediastinal anatomy than ICA, which is
limited in this regard due to its application being confined
almost exclusively to evaluation of the vessel lumen. CCTA
of the chest further delineates coronary artery course, as well
as the anatomy of the subclavian and mammary arteries.
While calcium can be seen on fluoroscopy and ICA, accurate
and extensive assessment of calcium load, which may influ-
ence choice of bypass target sites, is limited [21]. CCTA
adequately visualizes calcium and clearly defines coronary
artery course, revealing vessels that take an intramyocardial
course or run deep within the epicardial fat. This can predict
success in minimally invasive CABG as well as bypass distal
to a total coronary occlusion which may not be well visual-
ized on ICA [22]. In addition to anatomic information, CT
can also provide information regarding cardiac function [23]
and myocardial viability [24].
The LIMA is often used as a graft without preoperative
evaluation; however it has been shown that angiographic
visualization may change surgical strategy. In a series of 262
patients, anomalies in the origin, course, branching, or ath-
erosclerotic lesions resulted in a more meticulous dissection
Fig. 33.2  3D reconstruction of a patient with three-vessel disease pre-
of the artery in 26% of patients and a complete modification
senting for repeat CABG. Preoperative CT scan identified a subclavian
in surgical strategy in 4% of patients [25]. Subclavian artery artery stenosis proximal to the LIMA origin, and therefore the LIMA
anatomy has also been studied preoperatively using angiog- was not used as a bypass grafting
384 L. M. Felmly

it is possible to perform bypass grafting via non-sternotomy imally invasive approaches if it is deep in the epicardial fat,
approaches. Minimally invasive direct coronary artery especially if an off-pump strategy is employed. For this rea-
bypass (MIDCAB) utilizes an approach via a left anterior son preoperative CCTA is an important aspect of preopera-
minithoracotomy, whereas total endoscopic coronary artery tive planning for minimally invasive CABG, particularly in
bypass (TECAB) is a robotic procedure that allows a com- TECAB, as a robotic approach precludes tactile feedback
plete closed-chest approach to revascularization [29]. These and makes extensive dissection hazardous.
approaches offer the benefits of avoiding a sternotomy and in Off-pump CABG via sternotomy suffers from some of the
many cases the need for cardiopulmonary bypass, providing same limitations of minimally invasive CABG, in that coro-
an “off-pump” procedure. These approaches may offer a less nary arterial courses deep into the cardiac muscle or fat may
morbid procedure and a more appealing cosmetic result; make bypass challenging or impossible. While off-pump sur-
however they have several limitations, leading to less wide- gery through a traditional sternotomy retains the option of
spread adoption than might be expected. Firstly, access to converting to an on-pump surgery for better visualization,
these procedures may be limited to specific centers and sur- planning with CCTA delineates coronary anatomy and may
geons with the technology and expertise to perform them. lead to a change in bypass target site or a wholesale change
Secondly, the ability to achieve a full revascularization may in surgical strategy to on-pump surgery. This is significant,
be limited by coronary anatomy, especially when adopting as intraoperative conversion from off-pump to on-pump sur-
an off-pump approach, as the cardiac manipulation required gery has been associated with worsened outcomes [33]. In
to reach specific targets in the right coronary or left circum- addition, CT allows the identification of patients with a “por-
flex distributions may cause intraoperative hemodynamic celain aorta,” or aorta with extensive circumferential calcifi-
instability. MIDCAB offers only visualization of the left side cation. Under these circumstances the application of an
of the heart, limiting its application to two-vessel coronary aortic cross-clamp may represent a prohibitive risk of stroke
disease on the left side of the heart. TECAB allows better [7], and appropriate locations for the placement of proximal
visualization of the posterior aspect of the heart; however a aortocoronary anastomoses may not be present.
full revascularization then requires the application of cardio- As mentioned before, CCTA allows evaluation of both
pulmonary bypass via femoral cannulation with aortic mammary arteries to determine their suitability as bypass
“clamping” via endo-aortic balloon occlusion. Either one of grafts. This is significant, as there has been a trend toward
these techniques may be applied in concert with percutane- bypass grafting using all-arterial grafts due to higher demon-
ous intervention to achieve full revascularization via a hybrid strated patencies, which may lead to better long-term out-
approach. comes [4]. An all-arterial strategy is also attractive in that it
In cases in which minimally invasive CABG is pursued, is conducive to an off-pump approach due to the reduced
preoperative planning is paramount in order to determine need for aortocoronary anastomoses (Fig. 33.3). While mini-
appropriate vessel anatomy and relationships. EBCT has mally invasive CABG is useful in select populations, a mini-
been successfully used to determine appropriate incision mally invasive approach may preclude use of the right
placement based on coronary anatomy and desired targets internal mammary artery (RIMA) as a graft, which has at
[30]; however MDCT may also be employed successfully. least equal patency as compared to the radial artery and may
Distances between the LIMA and LAD as assessed by be higher in some circumstances [34]. This may be another
16-slice MDCT have been shown to correlate with operative reason that there has not been widespread adoption of mini-
time in TECAB [31]. This has been confirmed in other stud- mally invasive techniques. Regardless, preoperative CT is
ies, which have in addition shown that assessing soft-tissue useful in planning minimally invasive CABG, off-pump
thickness around the mammary as well as the pericardial fat CABG, and all-arterial bypasses.
pad around the LAD assists in operative planning [32].
Importantly, an intramyocardial vessel course or a course
deep within the epicardial fat is important to delineate preop- Postoperative Applications
eratively. In open CABG, appropriate anastomotic targets,
free of calcific plaque, are often determined by digital palpa- Invasive catheter angiography (ICA) has been the traditional
tion. If it is necessary to graft to an intramyocardial segment, method of evaluating bypass graft patency postoperatively;
myocardial dissection can be carried out safely, especially however contrast-enhanced coronary CT offers an appealing,
with the assistance of CPB and cardioplegic arrest. Dissection noninvasive method of evaluation, with the additional ability
and palpation of intramyocardial vessels may be difficult but to offer plaque characterization in late postoperative
still possible in MIDCAB. Vessel palpation is impossible in ­assessments. Initial studies were ungated, and while this
TECAB, and therefore intramyocardial vessel course is a allows evaluation of proximal and mid-vessel aspects of the
contraindication to TECAB.  In addition, accurate vessel graft, which tend to demonstrate less motion in the cardiac
course may be difficult to appreciate or dissect out from min- cycle than the coronary vasculature, an ungated approach
33  Multidetector CT Angiography for Coronary Bypass Graft Assessment and Reoperative Cardiac Surgery 385

multiple studies reported high sensitivities for assessment of

significant stenosis in evaluable grafts; however the percent-
age varies between studies, with reports of 78–92% of grafts
being completely evaluable [41–43]. An interesting study of
64-slice CT for bypass assessment found that CCTA demon-
strated perfect sensitivity in determining graft occlusion or
distal anastomotic stenosis when compared to ICA; however,
CCTA tended to overcall both [44]. The authors attributed the
false positives for occlusion to “string sign” visualized on
ICA, suggesting a lack of adequate spatial resolution, and to
competitive flow at distal anastomotic sites. While this study
did identify some shortcomings with CCTA, it is important to
note that all grafts were evaluable by 64-slice CCTA and that
performance overall was excellent. A recent meta-­analysis
comparing 64-slice CCTA to ICA in the evaluation of graft
occlusion and stenosis also yielded excellent results for
CCTA [45]. This meta-analysis by Chan et al. comprised 31
studies with 1975 total patients and demonstrated a sensitiv-
ity of 96.3% for identifying stenosis or occlusion compared to
ICA.  They did note, however, that CCTA was significantly
better (p = 0.004) for identifying abnormalities of saphenous
vein grafts compared to arterial grafts, indicating once again
Fig. 33.3  3D reconstruction of a patient with two-vessel disease now spatial resolution as a potential limiting factor.
status post a two-vessel CABG with bilateral IMA grafts. This approach CCTA has not been specifically studied for the evaluation
can obviate the need for aortocoronary anastomoses and is a useful
technique in patients with porcelain aorta in which avoiding aortic
of all-arterial bypass grafting; however there are plenty of
manipulation is desirable data regarding the evaluation of arterial grafts in combined
arterial and venous bypass. Arterial grafts have traditionally
been more challenging to image than venous grafts – due to
does not allow assessment of distal anastomoses [35]. surgical technique, they tend to have more clip artifacts than
Prospective ECG-triggering was also found to be inadequate venous grafts, and their smaller lumens require high spatial
early on [36]; however retrospective ECG-gating has been resolution to visualize. Early studies gave conflicting results
found to be highly successful in evaluating grafts, including on the accuracy of CT evaluation of arterial grafts as com-
their distal anastomoses, albeit at the cost of a higher radia- pared to venous grafts; however as CT technology has
tion dose [37, 38]. More recent studies have shown prospec- improved, this difference has become less significant. One
tive ECG synchronization to be equal to retrospective gating, study of single-slice CT gave accuracies of 96.4%, 65.7%,
however, allowing either of these methods to be employed and 45.5% for evaluating patency in venous, IMA, and radial
for diagnostic image quality [39]. artery grafts, respectively [46]. However a contemporaneous
Assessment of bypass grafts has improved as newer gen- study comparing IMA grafts to venous grafts resulted in
erations of CT scanners have added additional detector rows. accuracies of 88% and 96%, respectively, with no significant
Stein et al. published a systematic review in 2005 comparing difference [9]. The evaluation of arterial grafts on a 4-slice
results of single-slice, 4-slice, and 16-slice scanning in assess- system rendered similar results, with a sensitivity, specific-
ing bypass grafts in a pooled review comprising 1570 total ity, and overall accuracy of 100%, 98.7%, and 98.9%, respec-
patients [40]. They found that single-slice CT was not sensi- tively, for evaluating patency versus obstruction in arterial
tive for detecting occlusion, with a sensitivity of 81%, com- grafts, as compared to 81.8%, 93.7%, and 88.9% in venous
pared to 93% with 4-slice CT and 99% with 16-slice grafts [47]. Data from 64-slice CT showed a sensitivity,
CT. While the differences in occlusion detection were nonsig- specificity, and accuracy of 100% for evaluating venous graft
nificant between 4- and 16-slice CT, the ability to detect sig- occlusion versus 83.3%, 100%, and 98.9% for evaluating
nificant stenoses was significantly different, with a sensitivity arterial graft occlusion, respectively [48]. This study also
of 74% for 4-slice CT and 88% for 16-slice CT.  Multiple evaluated severity of stenosis in patent grafts and found a
studies have furthered the validation of 16-slice CT as a use- sensitivity, specificity, and accuracy of 94.4%, 98.4%, and
ful tool to assess bypass graft patency, with sensitivities as 96.9% in evaluating significant stenoses in venous grafts ver-
high as 100% for detecting occlusion [41]. One caveat that sus 100%, 97.7%, and 98% in arterial grafts, respectively.
remained was with regard to the number of evaluable grafts; More recently, 320-slice CT evaluation for the detection of
386 L. M. Felmly

Fig. 33.4  Artifact can be

introduced by metallic objects a b c
in and around coronaries and
bypass grafts such as surgical
clips adjacent to the LIMA
(a), extensive calcification in
a native coronary artery (b),
and coronary stents (c)

significant stenoses gave sensitivity, specificity, and diagnos- tion include the performance of CCTA in the presence of
tic accuracy of 95%, 93%, and 93% for venous grafts versus metallic artifact such as clips, stents, or heavy calcification
100%, 91%, and 93% for arterial grafts [49]. Taken with the (Fig. 33.4), which may also cause overestimation of steno-
results of the preceding study by Chan et al., it is important ses. However newer generations of CT scanners may over-
to note that there is still conflict in the literature with regard come these limitations, especially with software advances in
to the evaluation of arterial grafts. However, the overall body reconstructive techniques. With continued improvement in
of evidence indicates that CCTA is a useful modality for MDCT design, it is likely that CCTA may serve as a first step
evaluating arterial grafts and with current results would be in evaluating CABG grafts, ideally allowing ICA to be
acceptable as a first-line evaluation of patients with an all- reserved for those requiring procedural intervention.
arterial bypass to avoid the complications of ICA in patients
who do not require further investigation or intervention.
While invasive catheter FFR has applications in the Reoperative Surgery
assessment of bypass graft stenoses, CT-FFR has yet to be
studied in this context but may eventually find utility in cer- Reoperative surgery is an important aspect of cardiac surgi-
tain types of grafts. Saphenous vein grafts in particular tend cal practice, and one in which MDCT plays an increasingly
to develop atherosclerosis at an aggressive rate compared to important role. Due to the progressive nature of coronary
arterial grafts, and the resultant plaques are very friable and artery disease, as well as the effect of atherosclerosis on pre-
subject to rupture with distal embolization of contents when viously placed bypass grafts, a certain percentage of CABG
manipulated. Therefore, CT-FFR may provide added value patients will require repeat surgical revascularization in their
to the analysis of these lesions, as it avoids the risks of lifetime. In addition to revascularization, patients who have
manipulation associated with invasive assessment, namely, undergone valvular or other cardiac surgeries, with or with-
distal embolization and resultant periprocedural myocardial out concomitant bypass grafting, may require repeat cardiac
infarction. surgical intervention.
Currently, CCTA is a useful modality to assess previous Decreased graft patency with time, leading to symptoms
bypass grafts. Continued limitations besides spatial resolu- and cardiovascular events, is the major determinant for pro-
33  Multidetector CT Angiography for Coronary Bypass Graft Assessment and Reoperative Cardiac Surgery 387

ceeding to PCI or reoperative CABG. It is important to note may be increasing, the number of redo surgeries will likely
that different grafts demonstrate different patencies. A large see a decrease as transcatheter aortic valve implantation
series published in 1996 demonstrated SVG patency of 81% at (TAVI) is used more commonly in a “valve-in-valve” fashion
1 year, 75% at 5 years, and 50% at 15 years [50]. This trans- to treat bioprosthetic aortic valve stenosis. CT is indispens-
lates into a need for reoperative CABG of 2% at 5 years, 12% able in the planning of TAVI; however that topic is covered
at 10 years, and 20% at 12 years, in addition to patients under- in detail elsewhere in this text and will not be discussed here.
going percutaneous interventions [51]. Arterial grafts demon- Conversely, reoperation for mitral valve pathology is increas-
strate much higher patency, however, with the 10-year LIMA ing, with more than 10% of mitral operations being redo
graft patency of 94.6% when grafted to the LAD [52]. Long- operations, 35% of whom had undergone prior CABG [57].
term RIMA and radial artery patencies are less than that of the Due to a current lack of robust transcatheter mitral valve
LIMA but still superior to SVG [53]. Therefore adoption of an therapies, this number is likely to increase in the near future.
“all-arterial” strategy as noted before may lead to improved Reoperative cardiac surgery has traditionally been associ-
long-term graft patency with decreased need for reinterven- ated with significantly higher morbidity and mortality than
tion, and has been linked to improved ­outcomes [4]. primary cardiac surgery, with perioperative mortality tradi-
The frequency of reoperative cardiac surgery is evolving, tionally being reported in the 4–9% range, although mortality
and the incidence of reoperation depends on the indication of can exceed 20% in selected populations [55, 58]. This can be
the initial operation. There has been a steady decrease in attributed to several reasons, such as older age, increased
reoperative CABG in the past several decades, likely due to medical comorbidities, and damage to substernal structures
improved medical management and the proliferation of on reentry. Redo sternotomy can be a hazardous procedure, as
PCI. Conversely, there has been an increased need for valvu- the inflammatory and scar responses from primary sternot-
lar reintervention, as this patient population have experi- omy result in dense mediastinal adhesions. The aorta, right
enced greater longevity, along with increasing usage of coronary artery, right ventricle, and anteriorly located bypass
bioprosthetic valves [54], which have limited long-term grafts may be in close proximity, if not adhered, to the poste-
durability. Reports on regional or institutional trends regard- rior table of the sternum, making them susceptible to injury.
ing reoperative volume are subject to local practice patterns, The location of patent bypass grafts is of particular impor-
but national databases are useful in determining general tance, as myocardial dependence on these sources of blood
trends. Ghanta et  al. reported in a Society of Thoracic flow may be increased from the primary operation due to pro-
Surgeons (STS) database analysis that between 2000 and gression of native coronary disease, and injury sustained dur-
2009, redo CABG decreased from 6.0% to 3.4% of overall ing reentry may be difficult if not impossible to repair.
national CABG volume [55]. 5.9% of aortic valve proce- Specific difficult situations include a RIMA used to bypass a
dures nationally were redo operations in the period between left coronary territory or SVGs that come anteriorly off the
2011 and 2013 [56]. While need for aortic valve intervention aorta to bypass left-sided territories, as both of these graft pat-

Fig. 33.5  Preoperative CT is

critical prior to redo
a c
sternotomy, as bypass grafts
may be vulnerable to injury
on reentry. In this particular
instance, there are two venous
grafts that arise anteriorly
from the aorta and cross the
midline to supply the left
circumflex artery and a
diagonal branch of the
LAD. Grafts that cross
midline can be at risk for
damage; however in this b
instance they are a safe
distance away from the
sternum
388 L. M. Felmly

terns cross midline and may be in close proximity to the site native coronary arteries and bypass grafts. While CCTA is not
of re-sternotomy (Fig. 33.5). Regardless of sternal proximity, routinely used as a sole test to proceeding with CABG, it pro-
it is always important to localize the LIMA, as control of this vides more information on mediastinal anatomy than ICA and
graft is critical for myocardial protection. SVGs are placed on is indispensable for minimally invasive CABG and off-pump
the proximal aorta and therefore can carry cold, high-potas- CABG and for evaluating potential arterial grafts for all-arte-
sium cardioplegia to their supplied territories after cross- rial bypass. In addition, CCTA has been shown to accurately
clamping distally on the ascending aorta. However an in situ assess coronary artery bypass grafts for stenosis and overall
LIMA to LAD graft obligatorily originates past the aortic patency, in both venous and arterial grafts. CCTA is indispens-
cross-clamp and if not controlled and occluded during car- able to reoperative surgery in determining the location of
dioplegic arrest will carry warmer, lower-potassium blood bypass grafts and other mediastinal structures in order to avoid
from the bypass circuit to the LAD and any collateralized ter- damage during sternal reentry. The current status of MDCT
ritories, compromising myocardial protection. for use with CABG is promising, and advances in newer-gen-
Despite the difficulties associated with reoperative CABG eration CT scanners may lead to it becoming the “gatekeeper”
and other cardiac surgeries, outcomes have improved in to the catheterization lab and the operating room.
recent years and approach if not match outcomes of primary
surgery in some contemporary reports [55, 57]. The reason
for this is multifactorial, with improvements in anesthesia,
References
medical management of comorbidities, and postoperative
care contribution. However, the importance of preoperative 1. Roach GW, et al. Adverse cerebral outcomes after coronary bypass
MDCT cannot be understated. Preoperative CT is a prerequi- surgery. Multicenter Study of Perioperative Ischemia Research
site for redo surgery in the context of previous bypass grafts Group and the Ischemia Research and Education Foundation
in allowing appropriate planning with which many injuries Investigators. N Engl J Med. 1996;335(25):1857–63.
2. Loop FD, et  al. Influence of the internal-mammary-artery graft
upon reentry can be avoided. Three-dimensional reconstruc- on 10-year survival and other cardiac events. N Engl J Med.
tions are particularly useful in visualizing the mediastinal 1986;314(1):1–6.
anatomy. It is standard practice to acquire preoperative CT 3. Zhang B, et  al. Comparison of graft patency between off-pump
for all reoperative adult cardiac surgeries at the author’s and on-pump coronary artery bypass grafting: an updated meta-­
analysis. Ann Thorac Surg. 2014;97(4):1335–41.
institution. Multiple studies have shown that preoperative 4. Locker C, et  al. Multiple arterial grafts improve late survival
CT often changes management, with a high rate of adoption of patients undergoing coronary artery bypass graft surgery:
of protective strategies such as alternate incision sites, expo- analysis of 8622 patients with multivessel disease. Circulation.
sure of peripheral vascular access sites, and preemptive ini- 2012;126(9):1023–30.
5. Arad Y, et al. Predictive value of electron beam computed tomog-
tiation of cardiopulmonary bypass prior to redo sternotomy raphy of the coronary arteries. 19-month follow-up of 1173 asymp-
[59–61]. It has been demonstrated that preoperative CCTA is tomatic subjects. Circulation. 1996;93(11):1951–3.
associated with at least a trend toward decreased complica- 6. Godwin JD, et al. Clinical value of coronary bypass graft evaluation
tions after redo surgery [62, 63], which would likely be sig- with CT. AJR Am J Roentgenol. 1983;140(4):649–55.
7. Daniel WG, et  al. Value and limitations of computed tomogra-
nificant in larger series. In addition, preoperative CCTA phy in assessing aortocoronary bypass graft patency. Circulation.
appears cost-effective in reoperative cardiac surgery, as it 1983;67(5):983–7.
may prevent the cost of treating complications, although this 8. Tello R, et al. Spiral CT evaluation of coronary artery bypass graft
must be balanced with the cost of the additional OR time and patency. J Comput Assist Tomogr. 1993;17(2):253–9.
9. Engelmann MG, et  al. Accuracy of spiral computed tomography
equipment necessary for protective maneuvers [64]. As CT for identifying arterial and venous coronary graft patency. Am J
technology improves, so does this ability to plan for reopera- Cardiol. 1997;80(5):569–74.
tive cardiac surgery. Improved spatial and temporal resolu- 10. Becker CR, et  al. Current development of cardiac imaging with
tion affords accurate localization of substernal structures, as multidetector-row CT. Eur J Radiol. 2000;36(2):97–103.
11. Flohr T, Schmidt B. Technical aspects of dual energy CT with
do improved reconstructive models [65]. dual source CT systems. In: Carrascosa PM, Cury RC, Garcia MJ,
Leipseic JA, editors. Dual-energy CT in cardiovascular imaging.
New York: Springer International Publishing; 2015. p. 11–32.
Conclusion 12. Mangold S, et  al. Diagnostic accuracy of coronary CT angiogra-
phy using 3rd-generation dual-source CT and automated tube volt-
age selection: clinical application in a non-obese and obese patient
Multidetector CCTA is a useful tool for diagnosing and evalu- population. Eur Radiol. 2017;27:2298.
ating coronary artery disease, as well as elucidating mediasti- 13. Treede H, et al. Multidetector computed tomography (MDCT) in
nal anatomy pertinent to coronary artery bypass grafting. coronary surgery: first experiences with a new tool for diagnosis of
coronary artery disease. Ann Thorac Surg. 2002;74(4):S1398–402.
Multiple developments in hardware, software, and technique 14. Plass A, et al. Sixteen-channel multidetector row computed tomog-
have contributed to improved spatial and temporal resolution raphy versus coronary angiography in a surgical view. Heart Surg
while imaging the heart, giving high-quality information on Forum. 2006;9(2):E572–8.
33  Multidetector CT Angiography for Coronary Bypass Graft Assessment and Reoperative Cardiac Surgery 389

15. Bedi HS, Gill JA, Bakshi SS.  Can we perform coronary artery 36. von Smekal A.  The potential of cardio-computed tomography.

bypass grafting on the basis of computed tomographic angiography Multislice CT: a practical guide. Proceedings of the 5th International
alone? A comparison with conventional coronary angiography. Eur Somatom CT User Conference, Zurich, June 2000.
J Cardiothorac Surg. 2008;33(4):633–8. 37. Nieman K, et al. Coronary angiography with multi-slice computed
16. Otero HJ, Steigner ML, Rybicki FJ. The “post-64” era of coronary tomography. Lancet. 2001;357(9256):599–603.
CT angiography: understanding new technology from physical 38. Ropers D, et al. Investigation of aortocoronary artery bypass grafts by
principles. Radiol Clin N Am. 2009;47(1):79–90. multislice spiral computed tomography with electrocardiographic-­
17. Sadigh G, et  al. Impact of coronary CT angiography on surgical gated image reconstruction. Am J Cardiol. 2001;88(7):792–5.
decision-making for coronary artery bypass graft surgery. Acad 39. Lee JH, et  al. Prospective versus retrospective ECG-gated

Radiol. 2013;20(9):1083–90. 64-detector coronary CT angiography for evaluation of coro-
18. Botman CJ, et al. Does stenosis severity of native vessels influence nary artery bypass graft patency: comparison of image quality,
bypass graft patency? A prospective fractional flow reserve-guided radiation dose and diagnostic accuracy. Int J Cardiovasc Imaging.
study. Ann Thorac Surg. 2007;83(6):2093–7. 2011;27(5):657–67.
19. Shiono Y, et  al. Impact of functional focal versus diffuse coro- 40. Stein PD, et al. Usefulness of 4-, 8-, and 16-slice computed tomog-
nary artery disease on bypass graft patency. Int J Cardiol. raphy for detection of graft occlusion or patency after coronary
2016;222:16–21. artery bypass grafting. Am J Cardiol. 2005;96(12):1669–73.
20. Curzen NP, et al. Does the routine availability of CT-derived FFR 41. Anders K, et  al. Coronary artery bypass graft (CABG) patency:
influence management of patients with stable chest pain compared assessment with high-resolution submillimeter 16-slice
to CT angiography alone?: The FFRCT RIPCORD Study. JACC multidetector-­row computed tomography (MDCT) versus coronary
Cardiovasc Imaging. 2016;9(10):1188–94. angiography. Eur J Radiol. 2006;57(3):336–44.
21. Alexopoulos D, Moulias A. In the search of coronary calcium. Int J 42. Yamamoto M, et al. Noninvasive assessment of off-pump coronary
Cardiol. 2013;167(2):310–7. artery bypass surgery by 16-channel multidetector-row computed
22. Kepka C, et  al. Computed tomography to predict surgical revas- tomography. Ann Thorac Surg. 2006;81(3):820–7.
cularization of a left anterior descending artery occlusion incom- 43. Houslay ES, et  al. Non-invasive assessment of coronary artery
pletely visualized by conventional angiography. J Thorac Imaging. bypass graft patency using 16-slice computed tomography angiog-
2012;27(3):184–93. raphy. J Cardiothorac Surg. 2007;2:27.
23. Jones CM, et  al. Multi-slice computed tomography in coronary 44. Tochii M, et  al. Accuracy of 64-slice multidetector computed

artery disease. Eur J Cardiothorac Surg. 2006;30(3):443–50. tomography for diseased coronary artery graft detection. Ann
24. Hsiao EM, Rybicki FJ, Steigner M.  CT coronary angiography: Thorac Surg. 2010;89(6):1906–11.
256-slice and 320-detector row scanners. Curr Cardiol Rep. 45. Chan M, et al. A systematic review and meta-analysis of multide-
2010;12(1):68–75. tector computed tomography in the assessment of coronary artery
25. Bauer EP, et  al. Internal mammary artery anomalies. Thorac
bypass grafts. Int J Cardiol. 2016;221:898–905.
Cardiovasc Surg. 1990;38(5):312–5. 46. Ueyama K, et al. Evaluation of coronary artery bypass grafts using
26. Prasad A, et al. Prevalence and treatment of proximal left subcla- helical scan computed tomography. Catheter Cardiovasc Interv.
vian artery stenosis in patients referred for coronary artery bypass 1999;46(3):322–6.
surgery. Int J Cardiol. 2009;133(1):109–11. 47. Marano R, et al. Non-invasive assessment of coronary artery bypass
27. Byrne C, et  al. Ten-year experience in subclavian revasculari-
graft with retrospectively ECG-gated four-row multi-detector spiral
sation. A parallel comparative observational study. Vascular. computed tomography. Eur Radiol. 2004;14(8):1353–62.
2016;24(4):378–82. 48. Romagnoli A, et al. Diagnostic accuracy of 64-slice CT in evaluat-
28. Rigatelli G, Rigatelli G.  Simultaneous preoperative brachioce-
ing coronary artery bypass grafts and of the native coronary arter-
phalic angiography and coronary angiography to prevent coronary-­ ies. Radiol Med. 2010;115(8):1167–78.
subclavian steal syndrome in coronary surgery candidates. Heart 49. de Graaf FR, et al. Diagnostic performance of 320-slice multide-
Surg Forum. 2005;8(3):E175–7. tector computed tomography coronary angiography in patients
29. Holzhey DM, Rastan AJ, Falk V, Mohr FW. Minimally invasive myo- after coronary artery bypass grafting. Eur Radiol. 2011;21(11):
cardial revascularization. In: Cohn LH, editor. Cardiac surgery in the 2285–96.
adult. 4th ed. New York: McGraw-Hill Education; 2012. p. 553–66. 50. Fitzgibbon GM, et al. Coronary bypass graft fate and patient out-
30. Gulbins H, et  al. Preoperative 3D-reconstructions of ultrafast- come: angiographic follow-up of 5,065 grafts related to survival
CT images for the planning of minimally invasive direct coro- and reoperation in 1,388 patients during 25 years. J Am Coll
nary artery bypass operation (MIDCAB). Heart Surg Forum. Cardiol. 1996;28(3):616–26.
1998;1(2):111–5. 51. Weintraub WS, et al. Frequency of repeat coronary bypass or coro-
31. Schachner T, et al. Does preoperative multislice computed tomog- nary angioplasty after coronary artery bypass surgery using saphe-
raphy predict operative times in total endoscopic coronary artery nous venous grafts. Am J Cardiol. 1994;73(2):103–12.
bypass grafting? Heart Surg Forum. 2005;8(5):E314–8. 52. Zeff RH, et  al. Internal mammary artery versus saphenous vein
32. Wehman B, et  al. Patient anatomy predicts operative time in
graft to the left anterior descending coronary artery: prospec-
robotic totally endoscopic coronary artery bypass surgery. Interact tive randomized study with 10-year follow-up. Ann Thorac Surg.
Cardiovasc Thorac Surg. 2014;19(4):572–6. 1988;45(5):533–6.
33. Stevens LM, et al. Conversion after off-pump coronary artery bypass 53. Achouh P, et al. Reappraisal of a 20-year experience with the radial
grafting: the CORONARY trial experience. Eur J Cardiothorac artery as a conduit for coronary bypass grafting. Eur J Cardiothorac
Surg. 2017 Mar 1;51(3):539–46. Surg. 2012;41(1):87–92.
34. Benedetto U, et al. Searching for the second best graft for coronary 54. Javadzadegan H, Javadzadegan A, Mehdizadeh Baghbani J. Factors
artery bypass surgery: a network meta-analysis of randomized con- influencing mortality after bioprosthetic valve replacement; a mid-
trolled trialsdagger. Eur J Cardiothorac Surg. 2015;47(1):59–65; term outcome. J Cardiovasc Thorac Res. 2013;5(4):163–5.
discussion 65. 55. Ghanta RK, et  al. Evolving trends of reoperative coronary artery
35. Yamaguchi A, et  al. Three-dimensional computed tomographic
bypass grafting: an analysis of the Society of Thoracic Surgeons
angiography as preoperative evaluation of a patent internal thoracic Adult Cardiac Surgery Database. J Thorac Cardiovasc Surg.
artery graft. J Thorac Cardiovasc Surg. 2000;120(4):811–2. 2013;145(2):364–72.
390 L. M. Felmly

56. Kaneko T, et  al. Contemporary outcomes of repeat aortic valve 61. Kamdar AR, et al. Multidetector computed tomographic angiogra-
replacement: a benchmark for transcatheter valve-in-valve proce- phy in planning of reoperative cardiothoracic surgery. Ann Thorac
dures. Ann Thorac Surg. 2015;100(4):1298–304; discussion 1304. Surg. 2008;85(4):1239–45.
57. Ghoreishi M, et  al. Repeat sternotomy: no longer a risk fac-
62. Aviram G, et  al. Open heart reoperations after coronary artery
tor in mitral valve surgical procedures. Ann Thorac Surg. bypass grafting: the role of preoperative imaging with multidetec-
2013;96(4):1358–65. tor computed tomography. Isr Med Assoc J. 2009;11(8):465–9.
58. Christenson JT, Schmuziger M, Simonet F.  Reoperative coronary 63. den Harder AM, et al. Effect of computed tomography before car-
artery bypass procedures: risk factors for early mortality and late diac surgery on surgical strategy, mortality and stroke. Eur J Radiol.
survival. Eur J Cardiothorac Surg. 1997;11(1):129–33. 2016;85(4):744–50.
59. Gasparovic H, et  al. Three dimensional computed tomographic 64. Gada H, Desai MY, Marwick TH.  Cost-effectiveness of com-

imaging in planning the surgical approach for redo cardiac sur- puted tomographic angiography before reoperative coronary artery
gery after coronary revascularization. Eur J Cardiothorac Surg. bypass grafting: a decision-analytic model. Circ Cardiovasc Qual
2005;28(2):244–9. Outcomes. 2012;5(5):705–10.
60. Imran Hamid U, et  al. Incidence and outcome of re-entry injury 65. Oda S, et al. Cardiac CT for planning redo cardiac surgery: effect of
in redo cardiac surgery: benefits of preoperative planning. Eur J knowledge-based iterative model reconstruction on image quality.
Cardiothorac Surg. 2015;47(5):819–23. Eur Radiol. 2015;25(1):58–64.
 Cardiac CT in the Setting of Heart
Transplantation 34
Gorka Bastarrika and Gregorio Rábago

Heart transplantation is a definitive treatment option for in spatial and temporal resolution also allow visualization of
patients with end-stage heart failure. According to the 28th the coronary artery lumen and wall with unprecedented ana-
report of the International Society for Heart and Lung tomical detail. Further, ECG-triggered acquisition enables to
Transplantation (ISHLT) transplant registry, in the last evaluate cardiac function using the same CT datasets, without
decade, between 3600 and 3850 heart transplants were regis- the need of additional contrast media or radiation dose.
tered every year [1]. Moreover, CT is helpful in detecting transplant-related ancil-
Prognosis and long-term survival after heart transplanta- lary findings that may influence patient outcome. In this chap-
tion has significantly improved in the last decades particu- ter the usefulness of CT to provide an integrative approach to
larly due to effective antirejection therapy and infection the transplanted heart is reviewed.
control. The first year after transplant represents the period
with the highest risk of death, mainly due to graft failure,
infection, multiple organ failure, and rejection. Conversely,  ardiovascular Anatomy After Heart
C
long-term mortality is related to cardiac allograft vasculopa- Transplantation
thy (CAV), renal failure, and malignancy [1].
In fact, CAV is a well-known factor limiting the success- Orthotopic Heart Transplantation
ful long-term outcome of heart transplantation. Because of
the allograft denervation occurring after transplantation, this The first surgical technique for orthotopic heart transplanta-
entity is frequently asymptomatic. Clinical manifestations of tion was described by Shumway and Lower in 1966 [2]. The
CAV vary from arrhythmia to congestive heart failure or sud- so-called biatrial approach consists in leaving posterior right
den cardiac death. and left atrial cuffs to facilitate the reimplantation of the
Nowadays, conventional coronary angiography is the donor heart. The donor’s superior vena cava cuff is fixed
standard of reference to rule out CAV, although intravascular above the right atrium, and the donor’s inferior vena cava is
ultrasound is the most sensitive tool for early detection of the included in the right atrial anastomosis.
disease. Due to the fact that these two techniques carry a More recently, the standard biatrial technique was modi-
relatively high cost and risk, there has been vigorous research fied to launch the so-called bicaval approach [3, 4]. With this
on imaging modalities allowing noninvasive detection, char- surgical approach, the donor’s left atrium is sutured to the
acterization, and follow-up of CAV. In the last decade, car- posterior cuff of the recipient’s native left atrium containing
diac computed tomography (CT) has emerged as a very pulmonary vein insertion, the donor’s inferior and superior
useful imaging test in this clinical setting. vena cava are sutured to the corresponding recipient cuffs,
In the field of cardiac transplantation, CT allows to assess and the intervention is finalized with anastomosis of the
the typical appearance of the transplanted heart and to detect ascending aorta and main pulmonary artery. The bicaval
complications. Advances in CT technology and improvement technique allows better preservation of the right atrial mor-
phology and function, thus reducing the incidence of arrhyth-
G. Bastarrika (*) mias, right atrial dilatation, and tricuspid regurgitation [5].
Cardiothoracic Imaging Division, Department of Radiology, CT imaging shows characteristic findings that will allow
Clínica Universidad de Navarra, Pamplona, Spain to differentiate between the two surgical techniques. In
e-mail: [email protected]
patients undergoing heart transplantation with the biatrial
G. Rábago surgical technique, CT will show enlargement of both atrial
Department of Cardiac Surgery, Clínica Universidad de Navarra,
cavities, particularly of the left atrium (Fig. 34.1), whereas
Pamplona, Spain

© Humana Press 2019 391

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_34
392 G. Bastarrika and G. Rábago

patients undergoing heart transplantation with the bicaval surveillance (Fig. 34.4), emphasizing its role in the compre-
approach show a clockwise-rotated heart with normal size hensive evaluation of the transplanted heart.
atria (Fig.  34.2). The anastomotic sutures of the aorta and
main pulmonary artery are easily depicted, particularly when
there is size discrepancy between recipient and donor arter- Heterotopic Heart Transplantation
ies. In these cases, postsurgical changes may show the
appearance of perivascular soft tissue, redundancy, stenosis, Heterotopic heart transplantation is reserved for patients
or aneurysmal dilatation [6] (Fig. 34.3). with irreversible high pulmonary vascular resistance and
Graft outcome will depend on variables such as enlarge- body weight mismatch (over 20%) between recipient and
ment of the atria, asynchronous contraction of the donor and donor or for patients with potentially reversible myocardial
recipient atria, tendency for thrombus formation, conduction dysfunction. The donor heart is positioned side by side with
disturbances, and residual atrioventricular valve regurgita- the recipient heart. Surgeons perform anastomosis of right
tion [7–9]. Cardiac CT will also detect potential complica- atria, left atria, aorta, and main pulmonary arteries [10]. In
tions of the procedure early after surgery or during the heterotopic heart transplantation, the native right ventricle

a b

c d

Fig. 34.1  Orthotopic heart transplantation. (a, b) Biatrial technique. In clearly depicted (arrow). (c, d) Bicaval technique. CT exam of a
this 63-year-old woman who underwent heart transplantation with the 61-year-old man who underwent heart transplantation with the bicaval
biatrial surgical technique, CT imaging demonstrated enlargement of approach. In this technique the donor’s left atrium is sutured to the pos-
both atrial cavities, particularly of the left atrium. Note that most of the terior cuff of the recipient’s native left atrium containing pulmonary
left atrial cavity is provided by the receptor (R), whereas only a small vein insertion. As a result, there is no significant enlargement of atrial
portion is provided by the donor (D). The site of the anastomosis is cavities
34  Cardiac CT in the Setting of Heart Transplantation 393

a b c

Fig. 34.2  CT appearance of the graft in a patient with orthotopic heart wise rotation of the origin of the coronary vessels (b, left main, LM;
transplantation operated with the bicaval technique. (a) Axial image. right coronary artery, RCA). Despite this, the volumetric image is simi-
(b) Maximum intensity projection (MIP) image. (c) Volume-rendered lar to that of a non-transplanted heart (c). (LAD left anterior descending
image. The heart is clockwise-rotated (curved arrow in a). Note clock- coronary artery, LCx left circumflex, RCA right coronary artery)

Fig. 34.3  Postsurgical CT

appearance of vascular
anastomoses in a 68-year-old
male after heart
transplantation. (a) Axial
image. (b) Coronal image. (c)
Volume-rendered image. The
anastomotic sutures of the
aorta (Ao) and main
pulmonary artery (PA) after
transplantation are clearly
depicted (arrows)
a

b c

a b c

Fig. 34.4  Cardiac CT in a 65-year-old woman after heart transplantation with the biatrial technique. (a) Axial image. (b) Two-chamber view. (c)
Short-axis view. The exam showed a large intracavitary thrombus within the left atrium (LA, arrows)
394 G. Bastarrika and G. Rábago

provides most of right-side cardiac output, and the donor left mal hyperplasia causing diffuse luminal narrowing of the
ventricle delivers the majority of left-side cardiac output [6]. coronary vasculature. This will ultimately lead to myocardial
Complications of this surgical technique include native heart ischemia and ventricular dysfunction.
angina, ventricular fibrillation, and thromboembolic events Pathophysiological similarities and differences between
[11]. Graft stenosis is a frequent complication when grafts CAV and atherosclerosis are reviewed by Rahmani et  al.
are interposed between the recipient and donor pulmonary [13]. As opposed to conventional atherosclerotic disease,
arteries and may lead to pulmonary hypertension [10]. CAV typically affects distal segments and progresses to
proximal segments. Vessel involvement is concentric, longi-
tudinal, and diffuse. Collateral vasculature is generally
Cardiac Allograft Vasculopathy absent. In contrast, conventional atherosclerosis usually
involves the proximal coronary segments and progresses dis-
Cardiac allograft vasculopathy (CAV) is a progressive fibro- tally. Atherosclerotic plaques are commonly focal and non-­
proliferative process that extensively affects the coronary circumferential and show calcification more frequently than
arteries of a transplanted heart and is likely the consequence CAV. Thus, early after heart transplantation, focal and non-­
of cumulative endothelial injuries [12]. As said, this entity is circumferential thickening may be observed, usually trans-
responsible for a significant proportion of deaths after trans- mitted by the donor and representing conventional
plant. According to recent ISHLT data, the prevalence of atherosclerotic disease (Fig.  34.5). Late after transplant,
CAV remains high: 20% at 3 years, 30% at 5 years, and 45% however, focal atherosclerotic plaques and/or diffuse intimal
at 8  years after transplant. The disease is present in thickening may be seen representing CAV (Fig. 34.6).
­approximately 54% of survivors 10 years after heart trans- As mentioned, CAV is one of the main factors that will
plantation [1]. influence the prognosis and long-term survival after heart
Risk factors of CAV development depend on donor and transplantation. Thus, early detection of the disease becomes
recipient characteristics. Donor characteristics include old mandatory [14]. This is generally performed using diverse
donor age, male sex, high body surface area, history of imaging modalities.
hypertension, history of infection, and cause of death,
whereas recipient characteristics related with CAV comprise
history of ischemic heart disease, ventricular assist device Diagnosis of Cardiac Allograft Vasculopathy
implant before transplant, and history of infection [1]. Other
factors, such as the number of HLA-DR mismatches and the According to the ISHLT guidelines for the care of heart
use of certain type of immunosuppression before discharge, transplant recipients, annual or biannual coronary angiogra-
also increase the risk of CAV. phy should be considered to assess the development of
The pathophysiology of CAV entails endothelial dysfunc- CAV. Patients free of CAV at 3–5 years after heart transplan-
tion and vascular injury. Adaptive and innate immune tation, especially those with renal insufficiency, may undergo
responses, alloantigen-independent factors, and endothelial less frequent invasive evaluation (recommendation Class I,
replacement in the transplanted heart have been suggested to Level of Evidence: C) [15]. Recommendations for the use of
play a role in the pathophysiology of the disease [12]. imaging modalities for the diagnosis and management of
Persistent inflammatory response results in concentric inti- CAV are shown in Table 34.1.

Fig. 34.5  Coronary CT

angiography in a 64-year-old
man 2 years after heart
transplantation. (a) Axial
image. (b) Perpendicular
cross-sectional image. The
study demonstrated a
non-calcified plaque in the
left main coronary vessel
(arrow). Note that the plaque
is focal, eccentric, and
non-circumferential (arrows
in b), likely representing
atherosclerotic disease
transmitted by the donor
a b
34  Cardiac CT in the Setting of Heart Transplantation 395

Fig. 34.6  Coronary CT

angiography in a 75-year-old
asymptomatic man 12 years
after heart transplantation.
(a) Axial image.
(b) Perpendicular cross-
sectional image. Note the
almost concentric,
longitudinal, and diffuse
involvement of the left
anterior descending coronary
artery caused by cardiac
allograft vasculopathy
(arrows)
a b

Table 34.1  Recommendations for the diagnosis and management of ment and possess prognostic significance for graft survival,
cardiac allograft vasculopathy (CAV) according to the International patient survival, and adverse cardiac events [16]. However,
Society for Heart and Lung Transplantation (ISHLT) guidelines for the there is concern about the limited sensitivity of conventional
care of heart transplant (HT) recipients
coronary angiography for the detection of early stage CAV
Class I
and about its accuracy to establish the presence and extent of
Annual or biannual coronary angiography should be considered to
assess the development of CAV. Patients free of CAV at 3 to 5 years
the disease, particularly when compared with intravascular
after HT, especially those with renal insufficiency, may undergo less ultrasound (IVUS) [17, 18], optical coherence tomography
frequent invasive evaluation (Level of Evidence: C) (OCT) [19, 20], and histopathologic studies [21]. As a major
Selective coronary angiography is the investigation of choice for the limitation, conventional coronary angiography does not
diagnosis of CAV in pediatric HT recipients. It should be performed
allow to evaluate the arterial wall and to detect early intimal
at yearly or biannual intervals (Level of Evidence: C)
Class IIa
thickening. Thus, this technique can underdiagnose the prev-
A baseline coronary angiogram at 4 to 6 weeks after HT may be alence as well as the extent of CAV.
considered to exclude donor coronary artery disease (Level of
Evidence: C) IVUS
Intravascular ultrasound (IVUS) in conjunction with coronary Due to its ability to define arterial wall layers and to provide
angiography with a baseline study at 4 to 6 weeks and at 1 year
after HT is an option to exclude donor coronary artery disease, to
accurate quantitative assessment of lumen size, intimal
detect rapidly progressive CAV, and to provide prognostic thickening, and vessel wall morphology and composition,
information (Level of Evidence: B) IVUS is the most sensitive tool to detect CAV [22, 23]. There
IVUS can be safely used in older pediatric HT recipients to assess are, however, several limitations of this technique that have
CAV (Level of Evidence: C) impeded routine use of IVUS for the surveillance of CAV
Evaluation of coronary flow reserve in conjunction with coronary
angiography may be useful for the detection of small-vessel coronary
during regular follow-up of heart transplant recipients,
disease, which is a manifestation of CAV (Level of Evidence: C) including its high cost, lack of widespread expertise, the
Treadmill or dobutamine stress echocardiography and myocardial relatively large size of currently available IVUS catheters,
perfusion imaging may all be useful for the detection of CAV in HT and the complication rate of multivessel imaging. Similarly,
recipients unable to undergo invasive evaluation. Noninvasive testing the potential role of OCT in heart transplant recipients, an
for CAV is technically possible in children (Level of Evidence: B)
Class IIb
intravascular imaging technique able to provide detailed
Ultrafast computed tomography (CT) for the detection of coronary information of vessel wall composition, is under active
calcium has been used mostly as an investigational tool for research [24].
assessing CAV in HT recipients but is being superseded by
advances in CT angiography (Level of Evidence: C) Noninvasive Stress Testing
CT coronary angiography shows promise in the evaluation of CAV
Because of the reduced exercise capacity of heart transplant
in HT recipients, although higher resting heart rates in these patients
limit the technical quality of this study (Level of Evidence: C) recipients and the blunted heart rate response to exercise, the
Modified from Costanzo et al. [15], with permission accuracy of exercise stress tests for detection of CAV is
­limited [25].
Imaging modalities performed under pharmacological
 onventional Coronary Angiography
C stress show improved performance. Overall, sensitivity and
Conventional coronary angiography remains the standard of negative predictive values of noninvasive stress tests are
care for the screening and surveillance of CAV. Recent data high, whereas their specificity and positive predictive values
show that angiographic findings are able to guide manage- are low. Therefore, these tests can be employed to rule out
396 G. Bastarrika and G. Rábago

significant CAV, and, consequently, some authors recom- Technical Aspects

mend their use to avoid the need of annual conventional
coronary angiography for surveillance. Despite advances in CT technology and spatial and temporal
In the case of dobutamine stress echocardiography, its resolution, heart transplant recipients still represent a chal-
sensitivity and specificity to detect CAV range from 65% to lenge for noninvasive coronary artery imaging with CT [41].
95% and from 55% to 95% when compared to conventional The anatomical detail achieved with state-of-the-art technol-
coronary angiography [26, 27] and from 72% to 79% and ogy allows to reliably exclude significant coronary disease
from 83% to 88% when compared to IVUS [28, 29]. involving proximal and mid-coronary segments, but the spa-
Mastrobuoni et  al. showed that coronary CT angiography tial resolution still remains limited to accurately detect CAV
(CCTA) was able to demonstrate evidence of CAV despite involving more distal and smaller caliber coronary segments.
normal dobutamine stress echocardiography [30]. Another concerning aspect is the temporal resolution of
Regarding stress radionuclide myocardial perfusion CT systems and the high resting heart rates and the insuffi-
imaging, Ciliberto et al. showed sensitivity of 92% and spec- cient response to beta-blocker premedication of heart trans-
ificity of 86% of dipyridamole technetium-99m sestamibi plant recipients. After heart transplantation there is
SPECT to detect CAV manifested as angiographic focal ste- impairment of the sympathetic and parasympathetic innerva-
nosis greater than 50% or diffuse disease. The sensitivity tions of the graft. As a consequence, patient resting heart rate
dropped to 56% when the technique was used to detect angi- is high (mean resting heart rates up to 89 beats per minute)
ographic abnormalities of any grade [31]. Several studies [42], and they do not fully respond to different stimuli, such
have shown similar sensitivity and specificity values with as chronotropic and inotropic medication, including beta-­
dobutamine and dipyridamole, despite the fact that accord- blockers [43–45]. Therefore, the temporal resolution of even
ing to some authors, the effect of dipyridamole may be lim- 64-multidetector CT scanners with a 0.33-second rotation
ited under microvascular CAV [32]. time remains insufficient to accurately evaluate the coronary
Less is known about the usefulness of cardiac magnetic vasculature in this patient population, and more advanced
resonance (CMR) in the setting of CAV.  CMR provides CT systems with faster gantry rotation times are required.
accurate measurements of left and right ventricular volumes, Dual-source CT systems have shown to provide good image
ejection fraction, and mass. Further, late gadolinium-­ quality of the coronary vasculature in heart transplant recipi-
enhanced images [33, 34] and recently developed multipara- ents [42, 46–48] (Fig. 34.7).
metric T1 and T2 mapping techniques allow to establish and One interesting feature of heart transplant recipients is
quantify changes in the myocardial tissue architecture [35– their low heart rate variability. This allows to optimize car-
37]. In the field of stress perfusion CMR imaging, studies diac CT acquisition protocols to fully adjust the radiation
performed under adenosine demonstrated reduced myocar- dose. In this sense, prospectively ECG-triggered protocols
dial blood flow in patients with CAV [38]. In a recent study, may represent a valid alternative to conventional retrospec-
Miller et al. demonstrated that CAV, including epicardial and tively ECG-gated acquisition methods [49]. According to
microvascular components, could be detected more accu- recent publications, the mean radiation dose reported for
rately using noninvasive CMR-based absolute myocardial prospectively ECG-triggered CCTA in heart transplant
blood flow assessment than with conventional coronary recipients is significantly lower (mean of 4.5 mSv) [46] than
angiography [39]. In another study, Chih et al. showed that the radiation dose reported for the conventional retrospec-
CMR-derived myocardial perfusion reserve (MPR) index tively ECG-gated acquisition mode (mean of 10.2  mSv)
≤1.68 had sensitivity, specificity, negative, and positive pre- [50]. Of note, triggering the acquisition during the systolic
dictive values of 100%, 63%, 100%, and 86%, respectively, phase of the cardiac cycle has been advocated for the pro-
to detect CAV defined as maximal intimal thickness > 0.5 mm spective acquisition protocols (Fig.  34.8). Its rationale is
by IVUS of the left anterior descending artery [40]. based on the observation that the systolic phase has demon-
strated to be the optimal reconstruction phase for ­conventional
retrospectively ECG-gated coronary DSCT examinations
performed in subjects with high resting heart rates [51] and
 oronary CT Angiography and Cardiac
C also in heart transplant recipients [42].
Allograft Vasculopathy Finally, regarding the use of iodinated intravenous con-
trast during CCTA acquisition, it is important to note that
Coronary CT angiography (CCTA) has been investigated as chronic kidney disease is a common comorbidity in heart
a noninvasive alternative to conventional coronary angiogra- transplant recipients. Thus, the amount of contrast should be
phy to evaluate coronary artery vasculature and to detect adjusted to patient weight and scan duration. In most cases,
CAV in heart transplant recipients. Being noninvasive, one of the use of 60 or 70 mL of iodinated contrast should be suf-
the key advantages of this technique is that it provides infor- ficient to achieve adequate filling of the coronary
mation about the coronary artery lumen and wall. vasculature.
34  Cardiac CT in the Setting of Heart Transplantation 397

a b c d

Fig. 34.7  Retrospectively ECG-gated coronary CT angiography in a coronary artery (RCA), left anterior descending (LAD), and left cir-
63-year-old patient who underwent heart transplantation 5 years earlier. cumflex (LCx) coronary arteries. The study showed normal coronary
(a) Volume-rendered image. (b–d) Curved planar reformats of the right arteries, excluding cardiac allograft vasculopathy

a b c d

Fig. 34.8  Prospectively ECG-triggered coronary CT angiography in a ies. Despite a high resting heart rate (mean of 81 beats per minute) in
56-year-old heart transplant recipient. (a) Volume-rendered image. (b– this patient, the acquisition was prospectively ECG-triggered in systole
d) Curved planar reformats of the right coronary artery (RCA), left (35–45% of the R-R´ interval). The exam allowed to reliably exclude
anterior descending (LAD), and left circumflex (LCx) coronary arter- cardiac allograft vasculopathy

shows promise in the evaluation of CAV in heart transplant

 etection of Cardiac Allograft Vasculopathy
D recipients, although higher resting heart rates in these
with CCTA patients limit the technical quality of this study (Class IIb
recommendation, Level of Evidence: C).” Most recent appro-
According to the most recent ISHLT guidelines for the care priate use criteria for cardiac computed tomography endorsed
of heart transplant recipients, “CT coronary angiography by the American College of Cardiology Foundation along
398 G. Bastarrika and G. Rábago

with key specialty and subspecialty societies rate as uncer- predictive value of 71%, negative predictive value of 95%,
tain (score of 6) the routine evaluation of coronary arteries and accuracy of 93% for detection of coronary stenosis
by CCTA following heart transplantation [52]. greater than 50%. They concluded that 16-slice multidetec-
tor CT can replace the invasive procedure in de novo heart
Coronary Artery Calcification transplant patients and in patients with strictly normal CCTA
There is controversy regarding the clinical value of at follow-up [44].
CT-quantified coronary artery calcification as a marker of In their study, Mittal et  al. evaluated the largest heart
CAV. In a past publication, Ludman et al. reported that cal- transplant population published to date with CT [48]. They
cium is detected more frequently with CT than would be included 138 heart transplant recipients scheduled for rou-
suggested by studies using intravascular ultrasound [53]. tine conventional coronary angiography who prospectively
These authors observed that coronary calcium was associ- underwent CT to evaluate coronary artery calcification and
ated with the presence of angiographic disease and con- retrospectively ECG-gated CCTA with a 64-section scanner
cluded that at follow-up the presence of coronary calcium to assess vessel status. At the patient level, authors observed
was associated with an adverse clinical outcome [53]. More that CCTA had an area under the receiver operating charac-
recently, after evaluating 161 heart transplant recipients, von teristic curve of 0.942. Sensitivity, specificity, and positive
Ziegler et  al. determined that calcium scoring with dual-­ and negative predictive values of CCTA to detect CAV with
source CT is not a valuable noninvasive modality for CAV 50% or more stenosis were of 96%, 93%, 72%, and 99%,
detection. In fact, these authors suggested that calcium may respectively. Interestingly, authors found that none of the 61
represent preexisting or de novo traditional coronary athero- patients with normal CCTA angiographic results had CAV
sclerosis rather than CAV [54]. on the basis of the invasive procedure, concluding that CCTA
compares favorably with conventional coronary angiography
Coronary CTA Angiography Versus Conventional in detecting CAV in this patient population and that it may be
Coronary Angiography used as a screening technique [48].
The potential of CCTA for detecting CAV has been investi- Using a more recent generation dual-source CT scanner,
gated with different generation CT systems. Results of repre- von Ziegler et al. found similar results [59]. In 46 heart trans-
sentative studies comparing the diagnostic accuracy of plant recipients undergoing CCTA, authors observed sensi-
CCTA for the detection of CAV with respect to conventional tivity of 100%, specificity of 86%, and diagnostic accuracy
coronary angiography are shown in Table 34.2 [44, 45, 48, of 93% on a patient-based evaluation for detection of signifi-
50, 55–60]. cant stenosis. Authors concluded that improved image acqui-
In 2005, Romeo et al. sought to determine if CCTA per- sition and robustness of this CT technology allowed to
formed with a 16-row multidetector CT scanner could repre- exclude a low number of coronary segments, and, therefore,
sent a valid noninvasive alternative to conventional coronary it could be employed to reliably rule out CAV in heart trans-
angiography for serial detection and follow-up of coronary plant recipients, avoiding unnecessary conventional coro-
stenosis due to CAV [44]. For their study purpose, authors nary angiography for annual follow-up [59].
included 50 consecutive heart transplant patients who In a systematic review of the literature performed to deter-
­underwent CCTA within 24  h before or after their annual mine the accuracy of CCTA in CAV assessment, Khan et al.
routine conventional coronary angiography examination. included three studies performed using 16-row multidetector
Authors found sensitivity of 83%, specificity of 95%, p­ ositive CT and four studies using 64-row multidetector CT.  Per-

Table 34.2  Patient-based diagnostic accuracy of coronary CT angiography for the detection of hemodynamically significant coronary stenosis
compared with conventional coronary angiography
Number of Positive predictive Negative predictive
CT-system Author patients Sensitivity (%) Specificity (%) value (%) value (%)
MDCT-­16 Romeo et al. [44] 50 83 95 71 95
Sigurdsson et al. [45] 54 94 79 65 97
Moro et al. [55] 22 50 81 50 81
Pichler et al. [56] 60 88 97 88 97
MDCT-­64 Iyengar et al. [57] 19 100 94 50 100
von Ziegler et al. [58] 26 100 81 56 100
Mittal et al. [48] 138 96 78 77 98
DSCT Schepis et al. [50] 30 93 93 72 99
von Ziegler et al. [59] 46 100 86 33 100
Kepka et al. [60] 20 100 94 67 100
Note: MDCT multidetector computed tomography, DSCT dual-source computed tomography
34  Cardiac CT in the Setting of Heart Transplantation 399

segment analysis revealed pooled estimates for sensitivity, In their study, Gregory et al. reported a sensitivity of 70%,
specificity, and negative predictive value for CT ranging from specificity of 92%, positive predictive value of 89%, and
82% to 89%, 89% to 99%, and 99%, respectively [61]. negative predictive value of 77% for the detection of CAV
A recent meta-analysis evaluated the diagnostic accuracy with CCTA with respect to IVUS, concluding that CCTA has
of CCTA for detecting CAV in comparison with ­conventional moderate to excellent test characteristics for the detection of
coronary angiography, with similar conclusions [62]. After CAV [43].
analyzing 13 studies including 615 patients, the patient- In a third study, Schepis et al. found sensitivity of 85%,
based analysis showed mean weighted sensitivity of 97%, specificity of 84%, positive predictive value of 76%, and
specificity of 81%, positive predictive value of 78%, and negative predictive value of 91% for the detection of CAV by
negative predictive value of 97% for detection of any CAV DSCT with respect to IVUS [50].
and mean weighted sensitivity of 94%, specificity of 92%, Finally, in a recent meta-analysis, Wever-Pinzon et  al.
positive predictive value of 67%, and negative predictive found a mean weighted sensitivity of 81% and specificity of
value of 99% for detection of significant CAV (stenosis 75% to detect CAV by CCTA compared to IVUS [62].
≥50%) [62]. Authors concluded that CCTA is a reliable Therefore, according to most recent literature, CCTA
­noninvasive imaging alternative to conventional coronary seems to be a robust, reliable, and accurate technique for the
angiography to detect CAV [62]. diagnosis of CAV (Fig. 34.9). In particular, as in the field of
The outcome of heart transplant recipients using CCTA non-transplanted population, CCTA possesses a very high
for serial follow-up was investigated by Rohnean et al. [63]. negative predictive value. This allows to reliably exclude
These authors prospectively enrolled 62 individuals under- CAV in heart transplant patients. As a result, CCTA could be
going yearly CCTA for coronary allograft vasculopathy considered as an alternative to conventional coronary angi-
detection. According to their observations, time to stenosis ography in the routine surveillance of heart transplant
was consistently greater than 3 years, and the mean interval recipients.
without de novo significant stenosis was 10.31+/−4  years.
With these results, authors concluded that CCTA seems to be
a safe technique for monitoring heart transplant patients and Cardiac Function
suggested a follow-up every 2 years with CCTA in patients
with normal baseline exams [63]. The management and follow-up of heart transplant recipients
require accurate quantification of ventricular volumes, func-
 oronary CTA Angiography Versus
C tion, and mass [64]. Evaluation of these parameters is rou-
Intravascular Ultrasound tinely performed with transthoracic echocardiography,
As said, IVUS is the standard of reference to evaluate the although limited acoustic window, postsurgical changes, and
vessel wall. As CCTA also allows to assess vessel wall com- elevated body mass indices of heart transplant recipients
position, this technique may play a role in the early diagnosis may represent a limitation to this imaging modality. Further,
of CAV.  Three representative studies have compared the the limited interobserver reproducibility of echocardiogra-
diagnostic accuracy of the latter with respect to IVUS [43, phy is well known [65–67]. The reference standard to assess
45, 50] (Table 34.3). ventricular morphology, volumes, and function is magnetic
A subset of 54 heart transplant recipients enrolled by resonance imaging (MRI) [68–71].
Sigurdsson et al. were evaluated with CCTA, conventional As stated, cardiac CT may be used as an alternative to con-
coronary angiography, and IVUS [45]. Authors were able ventional coronary angiography to detect CAV. Interestingly,
to assess a total of 154 coronary segments in 13 individuals. when patients are examined with the retrospectively ECG-
The sensitivity, specificity, and positive and negative pre- gated technique, additional reconstructions including all
dictive values to identify proliferative changes by CCTA phases of the cardiac cycle provide the ability to evaluate ven-
with respect to IVUS were 96%, 88%, 80%, and 98%, tricular morphology and function without the need of further
respectively [45]. acquisitions. Its value in heart transplant patients has been

Table 34.3  Diagnostic accuracy of coronary CT angiography for the detection of hemodynamically significant coronary stenosis compared with
intravascular ultrasound (IVUS)
Number of Positive predictive Negative predictive
CT-system Author patients Sensitivity (%) Specificity (%) value (%) value (%)
MDCT-­16 Sigurdsson et al. [45] 13 96 88 80 98
MDCT-­64 Gregory et al. [43] 20 70 92 89 77
DSCT Schepis et al. [50] 30 85 84 76 91
Note: MDCT multidetector computed tomography, DSCT dual-source computed tomography
400 G. Bastarrika and G. Rábago

Fig. 34.9  Coronary CT

angiography in a 59-year-old
asymptomatic patient 15 years
after heart transplantation. (a)
Axial image. (b)
Perpendicular cross-sectional
image. (c) Curved planar
reformat. The exam
demonstrated longitudinal and
diffuse involvement of the left
anterior descending coronary
artery along its course by
cardiac allograft vasculopathy
(arrows in a and c). Note the
concentric appearance of the
disease (arrows in b)

b c

shown with 64-row multidetector CT scanners [72], as well as patients, skin cancer, solid-organ cancers including gastroin-
with dual-source CT technology [73, 74] (Fig.  34.10). testinal, lung cancer, oral cavity/oropharyngeal cancer, and
Usefulness of CT to assess left atrial function in heart trans- urologic cancers, as well as hematologic malignancies, such
plant recipients has also been demonstrated [75]. According as non-Hodgkin lymphoma, have been described [77].
to the results, agreement between echocardiography, CT, and In general practice, small field-of-view reconstructions
MRI-derived calculations for the assessment of cardiac struc- limited to the heart are employed to evaluate the coronary
ture and function seems to be between acceptable limits, sug- vasculature and the cardiac structures. There is discrep-
gesting the CT imaging may be used to assess atrial and ancy about the benefit of increasing the field of view to
ventricular parameters in heart transplant recipients. encompass the entire chest and the upper abdomen so as to
search for extracardiac findings [78–83]. Of course, the
prevalence of extracardiac findings in cardiac CT is
 ransplant-Related Ancillary Findings
T directly related to patient’s age, clinical symptoms, and
on Cardiac CT risk factors. In non-­ transplanted patients undergoing
CCTA, a variable prevalence of extracardiac findings rang-
Cardiac allograft vasculopathy is one of the four major ing from 15% to 67% has been reported [84, 85]. Most of
causes of death after heart transplantation. In the first year these findings do not require further action, whereas occa-
after transplant, most deaths are related to acute rejection or sionally clinically relevant entities are found [86]. Given
infection, whereas malignancy is the second most common the relatively high incidence of transplant-related compli-
long-term cause of mortality in heart transplant recipients, cations that may affect patient survival and the role of CT
likely related to immunosuppression. The carcinogenic con- in their detection and management, full field-of-view
sequences of long-term immunosuppression in transplant reconstructions may be particularly beneficial in this
recipients are well known [76]. In the case of heart transplant patient population (Fig. 34.11).
34  Cardiac CT in the Setting of Heart Transplantation 401

a b c

d e f

Fig. 34.10  Cardiac CT in a 62-year-old heart transplant recipient. (a, data acquisition. Software solutions automatically generate endocardial
d) Four-chamber view. (b, e) Two-chamber view. (c, f) Short-axis view. (red) and epicardial (green) left ventricular contours to provide left ven-
(a–c) Diastole. (d–f) Systole. Retrospectively ECG-gated cardiac CT tricular volumes, ejection fraction, and mass
protocols allow to assess cardiac function parameters without a­ dditional

Fig. 34.11  Cardiac CT in

68-year-old patient with heart
transplantation 21 years earlier. (a)
Small field-of-view reconstruction.
(b) Full field-of-view reconstruction.
Small field-of-view reconstructions
limited to the heart showed nodular
hepatic contours as well as
hypertrophy of the left lobe of the
liver. These findings were confirmed
in full field-of-view reconstructions,
which also allowed to depict an
arterially hyperenhancing lesion in
the right lobe consistent with
hepatocellular carcinoma in the
setting of cirrhosis
a b

Conclusion imaging modality that allows to assess posttransplant anat-

omy and to detect early and late complications related to the
Cardiac transplantation is the treatment of choice for patients procedure. Coronary artery lumen and wall are now viewed
with end-stage heart failure. Long-term prognosis and sur- and analyzed with unprecedented anatomical detail, whereas
vival after heart transplantation is related to cardiac allograft specific CT acquisition modes also enable to evaluate ­cardiac
vasculopathy, renal failure, and malignancy. Cardiac CT, morphologic and functional parameters in good agreement
particularly using latest generation equipment, is a unique with established imaging modalities. Although conventional
402 G. Bastarrika and G. Rábago

coronary angiography remains the reference standard for the Heart and Lung Transplantation working formulation of a standard-
ized nomenclature for cardiac allograft vasculopathy-2010. J Heart
diagnosis of CAV, the results of most recent studies and Lung Transplant. 2010;29(7):717–27.
meta-analysis suggest that CCTA could be considered as an 17. St Goar FG, Pinto FJ, Alderman EL, Valantine HA, Schroeder JS,
alternative to conventional coronary angiography in the rou- Gao SZ, et al. Intracoronary ultrasound in cardiac transplant recipi-
tine surveillance of heart transplant recipients. Still, further ents. In vivo evidence of “angiographically silent” intimal thicken-
ing. Circulation. 1992;85(3):979–87.
research is warranted to determine the impact of this imaging 18. Spes CH, Klauss V, Rieber J, Schnaack SD, Tammen AR, Uberfuhr
technology in the management and clinical outcomes of this P, et al. Functional and morphological findings in heart transplant
patient population. recipients with a normal coronary angiogram: an analysis by dobu-
tamine stress echocardiography, intracoronary Doppler and intra-
vascular ultrasound. J Heart Lung Transplant. 1999;18(5):391–8.
19. Khandhar SJ, Yamamoto H, Teuteberg JJ, Shullo MA, Bezerra HG,
References Costa MA, et al. Optical coherence tomography for characteriza-
tion of cardiac allograft vasculopathy after heart transplantation
1. Stehlik J, Edwards LB, Kucheryavaya AY, Benden C, Christie JD, (OCTCAV study). J Heart Lung Transplant. 2013;32(6):596–602.
Dobbels F, et al. The registry of the International Society for Heart 20. Hou J, Lv H, Jia H, Zhang S, Xing L, Liu H, et al. OCT assess-
and Lung Transplantation: twenty-eighth adult heart transplant ment of allograft vasculopathy in heart transplant recipients. JACC
report – 2011. J Heart Lung Transplant. 2011;30(10):1078–94. Cardiovasc Imaging. 2012;5(6):662–3.
2. Shumway NE, Lower RR, Stofer RC. Transplantation of the heart. 21. Johnson DE, Alderman EL, Schroeder JS, Gao SZ, Hunt S,

Adv Surg. 1966;2:265–84. DeCampli WM, et  al. Transplant coronary artery disease: histo-
3. Dreyfus G, Jebara V, Mihaileanu S, Carpentier AF. Total orthotopic pathologic correlations with angiographic morphology. J Am Coll
heart transplantation: an alternative to the standard technique. Ann Cardiol. 1991;17(2):449–57.
Thorac Surg. 1991;52(5):1181–4. 22. Rickenbacher PR, Pinto FJ, Chenzbraun A, Botas J, Lewis NP,
4. Sievers HH, Weyand M, Kraatz EG, Bernhard A.  An alternative Alderman EL, et  al. Incidence and severity of transplant coro-
technique for orthotopic cardiac transplantation, with preservation nary artery disease early and up to 15 years after transplanta-
of the normal anatomy of the right atrium. Thorac Cardiovasc Surg. tion as detected by intravascular ultrasound. J Am Coll Cardiol.
1991;39(2):70–2. 1995;25(1):171–7.
5. Weiss ES, Nwakanma LU, Russell SB, Conte JV, Shah 23. Tsutsui H, Ziada KM, Schoenhagen P, Iyisoy A, Magyar WA,
AS.  Outcomes in bicaval versus biatrial techniques in heart Crowe TD, et al. Lumen loss in transplant coronary artery disease
transplantation: an analysis of the UNOS database. J Heart Lung is a biphasic process involving early intimal thickening and late
Transplant. 2008;27(2):178–83. constrictive remodeling: results from a 5-year serial intravascular
6. Knisely BL, Mastey LA, Collins J, Kuhlman JE. Imaging of cardiac ultrasound study. Circulation. 2001;104(6):653–7.
transplantation complications. Radiographics. 1999;19(2):321–39; 24. Shan P, Dong L, Maehara A, Nazif TM, Ali ZA, Rabbani LE, et al.
discussion 40–1 Comparison between cardiac allograft vasculopathy and native
7. Triposkiadis F, Starling RC, Haas GJ, Sparks E, Myerowitz PD, coronary atherosclerosis by optical coherence tomography. Am J
Boudoulas H. Timing of recipient atrial contraction: a major deter- Cardiol. 2016;117(8):1361–8.
minant of transmitral diastolic flow in orthotopic cardiac transplan- 25. Cohn JM, Wilensky RL, O'Donnell JA, Bourdillon PD, Dillon JC,
tation. Am Heart J. 1993;126(5):1175–81. Feigenbaum H. Exercise echocardiography, angiography, and intra-
8. Stevenson LW, Dadourian BJ, Kobashigawa J, Child JS, Clark SH, coronary ultrasound after cardiac transplantation. Am J Cardiol.
Laks H.  Mitral regurgitation after cardiac transplantation. Am J 1996;77(14):1216–9.
Cardiol. 1987;60(1):119–22. 26. Akosah KO, McDaniel S, Hanrahan JS, Mohanty PK. Dobutamine
9. Gorcsan J 3rd, Snow FR, Paulsen W, Arrowood JA, Thompson stress echocardiography early after heart transplantation predicts
JA, Nixon JV.  Echocardiographic profile of the transplanted development of allograft coronary artery disease and outcome. J
human heart in clinically well recipients. J Heart Lung Transplant. Am Coll Cardiol. 1998;31(7):1607–14.
1992;11(1 Pt 1):80–9. 27.
Derumeaux G, Redonnet M, Soyer R, Cribier A, Letac
10. Lai HY, Chen JH, Chiu KM, Wang KL, Cheung WK, Li AH, et al. B. Assessment of the progression of cardiac allograft vasculopathy
CT of two hearts beating in one chest. AJR Am J Roentgenol. by dobutamine stress echocardiography. J Heart Lung Transplant.
2008;191(6):1711–6. 1998;17(3):259–67.
11. Roux C, Varnous S, Leprince P, Cluzel P.  Two hearts, one soul: 28. Spes CH, Klauss V, Mudra H, Schnaack SD, Tammen AR, Rieber
heterotopic heart transplantation follow-up with cardiac computed J, et al. Diagnostic and prognostic value of serial dobutamine stress
tomography. Eur Heart J. 2016;37:3356. echocardiography for noninvasive assessment of cardiac allograft
12. Schmauss D, Weis M. Cardiac allograft vasculopathy: recent devel- vasculopathy: a comparison with coronary angiography and intra-
opments. Circulation. 2008;117(16):2131–41. vascular ultrasound. Circulation. 1999;100(5):509–15.
13. Rahmani M, Cruz RP, Granville DJ, McManus BM. Allograft vas- 29. Spes CH, Mudra H, Schnaack SD, Klauss V, Reichle FM,

culopathy versus atherosclerosis. Circ Res. 2006;99(8):801–15. Uberfuhr P, et al. Dobutamine stress echocardiography for nonin-
14. Lamich R, Ballester M, Marti V, Brossa V, Aymat R, Carrio I, vasive diagnosis of cardiac allograft vasculopathy: a comparison
et  al. Efficacy of augmented immunosuppressive therapy for with angiography and intravascular ultrasound. Am J Cardiol.
early vasculopathy in heart transplantation. J Am Coll Cardiol. 1996;78(2):168–74.
1998;32(2):413–9. 30. Mastrobuoni S, Bastarrika G, Ubilla M, Castano S, Azcarate P,
15. Costanzo MR, Dipchand A, Starling R, Anderson A, Chan M, Desai Barrero EA, et  al. Dual-source CT coronary angiogram in heart
S, et al. The International Society of Heart and Lung Transplantation transplant recipients in comparison with dobutamine stress echo-
Guidelines for the care of heart transplant recipients. J Heart Lung cardiography for detection of cardiac allograft vasculopathy.
Transplant. 2010;29(8):914–56. Transplantation. 2009;87(4):587–90.
16. Mehra MR, Crespo-Leiro MG, Dipchand A, Ensminger SM,
31. Ciliberto GR, Ruffini L, Mangiavacchi M, Parolini M, Sara R, Massa
Hiemann NE, Kobashigawa JA, et  al. International Society for D, et al. Resting echocardiography and quantitative dipyridamole
34  Cardiac CT in the Setting of Heart Transplantation 403

technetium-99m sestamibi tomography in the identification of car- 47. Bastarrika G, De Cecco CN, Arraiza M, Mastrobuoni S, Pueyo JC,
diac allograft vasculopathy and the prediction of long-term progno- Ubilla M, et al. Dual-source CT for visualization of the coronary
sis after heart transplantation. Eur Heart J. 2001;22(11):964–71. arteries in heart transplant patients with high heart rates. AJR Am J
32. Elhendy A, Sozzi FB, van Domburg RT, Vantrimpont P, Valkema R, Roentgenol. 2008;191(2):448–54.
Krenning EP, et al. Accuracy of dobutamine tetrofosmin myocardial 48. Mittal TK, Panicker MG, Mitchell AG, Banner NR.  Cardiac

perfusion imaging for the noninvasive diagnosis of transplant coro- allograft vasculopathy after heart transplantation: electrocar-
nary artery stenosis. J Heart Lung Transplant. 2000;19(4):360–6. diographically gated cardiac CT angiography for assessment.
33. Braggion-Santos MF, Lossnitzer D, Buss S, Lehrke S, Doesch Radiology. 2013;268(2):374–81.
A, Giannitsis E, et  al. Late gadolinium enhancement assessed by 49. Shuman WP, Branch KR, May JM, Mitsumori LM, Lockhart DW,
cardiac magnetic resonance imaging in heart transplant recipients Dubinsky TJ, et al. Prospective versus retrospective ECG gating for
with different stages of cardiac allograft vasculopathy. Eur Heart J 64-detector CT of the coronary arteries: comparison of image qual-
Cardiovasc Imaging. 2014;15(10):1125–32. ity and patient radiation dose. Radiology. 2008;248(2):431–7.
34. Pedrotti P, Bonacina E, Vittori C, Frigerio M, Roghi A. Pathologic 50. Schepis T, Achenbach S, Weyand M, Raum P, Marwan M, Pflederer T,
correlates of late gadolinium enhancement cardiovascular mag- et al. Comparison of dual source computed tomography versus intra-
netic resonance in a heart transplant patient. Cardiovasc Pathol. vascular ultrasound for evaluation of coronary arteries at least one year
2015;24(4):247–9. after cardiac transplantation. Am J Cardiol. 2009;104(10):1351–6.
35. Riesenkampff E, Chen CK, Kantor PF, Greenway S, Chaturvedi 51. Adler G, Meille L, Rohnean A, Sigal-Cinqualbre A, Capderou
RR, Yoo SJ, et  al. Diffuse myocardial fibrosis in children after A, Paul JF.  Robustness of end-systolic reconstructions in coro-
heart transplantations: a magnetic resonance T1 mapping study. nary dual-source CT angiography for high heart rate patients. Eur
Transplantation. 2015;99(12):2656–62. Radiol. 2010;20(5):1118–23.
36. Ellims AH, Shaw JA, Stub D, Iles LM, Hare JL, Slavin GS, et al. 52. Taylor AJ, Cerqueira M, Hodgson JM, Mark D, Min J, O'Gara P, et al.
Diffuse myocardial fibrosis evaluated by post-contrast t1 map- ACCF/SCCT/ACR/AHA/ASE/ASNC/NASCI/SCAI/SCMR 2010
ping correlates with left ventricular stiffness. J Am Coll Cardiol. appropriate use criteria for cardiac computed tomography. A report
2014;63(11):1112–8. of the American College of Cardiology Foundation Appropriate
37. Usman AA, Taimen K, Wasielewski M, McDonald J, Shah S, Giri Use Criteria Task Force, the Society of Cardiovascular Computed
S, et al. Cardiac magnetic resonance T2 mapping in the monitoring Tomography, the American College of Radiology, the American
and follow-up of acute cardiac transplant rejection: a pilot study. Heart Association, the American Society of Echocardiography,
Circ Cardiovasc Imaging. 2012;5(6):782–90. the American Society of Nuclear Cardiology, the North American
38. Muehling OM, Wilke NM, Panse P, Jerosch-Herold M, Wilson Society for Cardiovascular Imaging, the Society for Cardiovascular
BV, Wilson RF, et  al. Reduced myocardial perfusion reserve Angiography and Interventions, and the Society for Cardiovascular
and transmural perfusion gradient in heart transplant arteriopa- Magnetic Resonance. J Am Coll Cardiol. 2010;56(22):1864–94.
thy assessed by magnetic resonance imaging. J Am Coll Cardiol. 53. Ludman PF, Lazem F, Barbir M, Yacoub M. Incidence and clini-
2003;42(6):1054–60. cal relevance of coronary calcification detected by electron beam
39. Miller CA, Sarma J, Naish JH, Yonan N, Williams SG, Shaw computed tomography in heart transplant recipients. Eur Heart J.
SM, et  al. Multiparametric cardiovascular magnetic resonance 1999;20(4):303–8.
assessment of cardiac allograft vasculopathy. J Am Coll Cardiol. 54. von Ziegler F, Kaczmarek I, Knez A, Greif M, Rummler J, Meiser
2014;63(8):799–808. B, et al. Coronary calcifications detected by computed tomography
40. Chih S, Ross HJ, Alba AC, Fan CS, Manlhiot C, Crean AM. Perfusion are not markers of cardiac allograft vasculopathy. Transplantation.
cardiac magnetic resonance imaging as a rule-out test for cardiac 2011;92(4):493–8.
allograft vasculopathy. Am J Transplant. 2016;16(10):3007–15. 55. Moro J, Almenar L, Igual B, Martinez-Dolz L, Sanchez E, Martin
41. Ferencik M, Brady TJ, Hoffmann U. Computed tomography imag- J, et  al. Multislice CT in graft vascular disease. A pilot study.
ing of cardiac allograft vasculopathy. J Cardiovasc Comput Tomogr. Transplant Proc. 2006;38(8):2563–5.
2012;6(4):223–31. 56. Pichler P, Loewe C, Roedler S, Syeda B, Stadler A, Aliabadi A,
42. Bastarrika G, De Cecco CN, Arraiza M, Ubilla M, Mastrobuoni S, et  al. Detection of high-grade stenoses with multislice computed
Pueyo JC, et al. Dual-source CT coronary imaging in heart trans- tomography in heart transplant patients. J Heart Lung Transplant.
plant recipients: image quality and optimal reconstruction interval. 2008;27(3):310–6.
Eur Radiol. 2008;18(9):1791–9. 57. Iyengar S, Feldman DS, Cooke GE, Leier CV, Raman SV. Detection
43. Gregory SA, Ferencik M, Achenbach S, Yeh RW, Hoffmann U, of coronary artery disease in orthotopic heart transplant recipients
Inglessis I, et al. Comparison of sixty-four-slice multidetector com- with 64-detector row computed tomography angiography. J Heart
puted tomographic coronary angiography to coronary angiography Lung Transplant. 2006;25(11):1363–6.
with intravascular ultrasound for the detection of transplant vascu- 58. von Ziegler F, Leber AW, Becker A, Kaczmarek I, Schonermarck
lopathy. Am J Cardiol. 2006;98(7):877–84. U, Raps C, et al. Detection of significant coronary artery stenosis
44. Romeo G, Houyel L, Angel CY, Brenot P, Riou JY, Paul
with 64-slice computed tomography in heart transplant recipients:
JF. Coronary stenosis detection by 16-slice computed tomography a comparative study with conventional coronary angiography. Int J
in heart transplant patients: comparison with conventional angi- Cardiovasc Imaging. 2009;25(1):91–100.
ography and impact on clinical management. J Am Coll Cardiol. 59. von Ziegler F, Rummler J, Kaczmarek I, Greif M, Schenzle J,
2005;45(11):1826–31. Helbig S, et  al. Detection of significant coronary artery stenosis
45. Sigurdsson G, Carrascosa P, Yamani MH, Greenberg NL, Perrone with cardiac dual-source computed tomography angiography in
S, Lev G, et  al. Detection of transplant coronary artery disease heart transplant recipients. Transpl Int. 2012;25(10):1065–71.
using multidetector computed tomography with adaptative multi- 60. Kepka C, Sobieszczansk-Malek M, Pregowski J, Kruk M, Bekta P,
segment reconstruction. J Am Coll Cardiol. 2006;48(4):772–8. Opolski M, et al. Usefulness of dual-source computed tomography
46. Bastarrika G, Broncano J, Arraiza M, Azcarate PM, Simon-
for the evaluation of coronary arteries in heart transplant recipients.
Yarza I, Levy Praschker BG, et  al. Systolic prospectively Kardiol Pol. 2012;70(11):1111–9.
ECG-triggered dual-source CT angiography for evaluation of 61. Khan R, Jang IK.  Evaluation of coronary allograft vasculopathy
the coronary arteries in heart transplant recipients. Eur Radiol. using multi-detector row computed tomography: a systematic
2011;21(9):1887–94. review. Eur J Cardiothorac Surg. 2012;41(2):415–22.
404 G. Bastarrika and G. Rábago

62. Wever-Pinzon O, Romero J, Kelesidis I, Wever-Pinzon J,

73. Bastarrika G, Arraiza M, De Cecco CN, Broncano J, Mastrobuoni
Manrique C, Budge D, et  al. Coronary computed tomography S, Ubilla M, et  al. Dual-source CT in heart transplant recipients:
angiography for the detection of cardiac allograft vasculopa- quantification of global left ventricular function and mass. J Thorac
thy: a meta-analysis of prospective trials. J Am Coll Cardiol. Imaging. 2009;24(2):103–9.
2014;63(19):1992–2004. 74. Bastarrika G, Arraiza M, De Cecco CN, Mastrobuoni S, Ubilla M,
63. Rohnean A, Houyel L, Sigal-Cinqualbre A, To NT, Elfassy E, Paul Rabago G. Quantification of left ventricular function and mass in
JF.  Heart transplant patient outcomes: 5-year mean follow-up by heart transplant recipients using dual-source CT and MRI: initial
coronary computed tomography angiography. Transplantation. clinical experience. Eur Radiol. 2008;18(9):1784–90.
2011;91(5):583–8. 75. Bastarrika G, Zudaire B, Ferreira M, Arraiza M, Saiz-Mendiguren
64. Globits S, De Marco T, Schwitter J, Sakuma H, O'Sullivan M, R, Rabago G.  Assessment of left atrial volumes and function in
Rifkin C, et  al. Assessment of early left ventricular remodeling orthotopic heart transplant recipients by dual-source CT: compari-
in orthotopic heart transplant recipients with cine magnetic reso- son with MRI. Investig Radiol. 2010;45(2):72–6.
nance imaging: potential mechanisms. J Heart Lung Transplant. 76. Cole WH. The increase in immunosuppression and its role in the
1997;16(5):504–10. development of malignant lesions. J Surg Oncol. 1985;30(3):139–44.
65. Bellenger NG, Marcus NJ, Davies C, Yacoub M, Banner NR,
77. Roussel JC, Baron O, Perigaud C, Bizouarn P, Pattier S, Habash
Pennell DJ.  Left ventricular function and mass after orthotopic O, et  al. Outcome of heart transplants 15 to 20 years ago: graft
heart transplantation: a comparison of cardiovascular magnetic survival, post-transplant morbidity, and risk factors for mortality. J
resonance with echocardiography. J Heart Lung Transplant. Heart Lung Transplant. 2008;27(5):486–93.
2000;19(5):444–52. 78. Budoff MJ, Gopal A.  Incidental findings on cardiac computed
66. Kronik G, Slany J, Mosslacher H.  Comparative value of eight tomography. Should we look? J Cardiovasc Comput Tomogr.
M-mode echocardiographic formulas for determining left ven- 2007;1(2):97–105.
tricular stroke volume. A correlative study with thermodilution 79. Earls JP. The pros and cons of searching for extracardiac findings
and left ventricular single-plane cineangiography. Circulation. at cardiac CT: studies should be reconstructed in the maximum
1979;60(6):1308–16. field of view and adequately reviewed to detect pathologic findings.
67. Allison JD, Flickinger FW, Wright JC, Falls DG 3rd, Prisant LM, Radiology. 2011;261(2):342–6.
VonDohlen TW, et  al. Measurement of left ventricular mass in 80. Hlatky MA, Iribarren C.  The dilemma of incidental find-

hypertrophic cardiomyopathy using MRI: comparison with echo- ings on cardiac computed tomography. J Am Coll Cardiol.
cardiography. Magn Reson Imaging. 1993;11(3):329–34. 2009;54(16):1542–3.
68. Pennell DJ, Sechtem UP, Higgins CB, Manning WJ, Pohost GM, 81. Killeen RP, Dodd JD, Cury RC.  Noncardiac findings on car-

Rademakers FE, et  al. Clinical indications for cardiovascular diac CT part I: pros and cons. J Cardiovasc Comput Tomogr.
magnetic resonance (CMR): Consensus Panel report. Eur Heart J. 2009;3(5):293–9.
2004;25(21):1940–65. 82. White CS.  The pros and cons of searching for extracardiac find-
69. Sandstede J, Lipke C, Beer M, Hofmann S, Pabst T, Kenn W, et al. ings at cardiac CT: use of a restricted field of view is acceptable.
Age- and gender-specific differences in left and right ventricular Radiology. 2011;261(2):338–41.
cardiac function and mass determined by cine magnetic resonance 83. Budoff MJ, Fischer H, Gopal A. Incidental findings with cardiac CT
imaging. Eur Radiol. 2000;10(3):438–42. evaluation: should we read beyond the heart? Catheter Cardiovasc
70. Natori S, Lai S, Finn JP, Gomes AS, Hundley WG, Jerosch-Herold Interv. 2006;68(6):965–73.
M, et  al. Cardiovascular function in multi-ethnic study of ath- 84. Dewey M, Schnapauff D, Teige F, Hamm B. Non-cardiac findings
erosclerosis: normal values by age, sex, and ethnicity. AJR Am J on coronary computed tomography and magnetic resonance imag-
Roentgenol. 2006;186(6 Suppl 2):S357–65. ing. Eur Radiol. 2007;17(8):2038–43.
71. Myerson SG, Bellenger NG, Pennell DJ. Assessment of left ven- 85. Kirsch J, Araoz PA, Steinberg FB, Fletcher JG, McCollough CH,
tricular mass by cardiovascular magnetic resonance. Hypertension. Williamson EE. Prevalence and significance of incidental extracar-
2002;39(3):750–5. diac findings at 64-multidetector coronary CTA. J Thorac Imaging.
72. Ferencik M, Gregory SA, Butler J, Achenbach S, Yeh RW,
2007;22(4):330–4.
Hoffmann U, et al. Analysis of cardiac dimensions, mass and func- 86.
Johnson KM.  Extracardiac findings on cardiac computed
tion in heart transplant recipients using 64-slice multi-detector com- tomography: a radiologist's perspective. J Am Coll Cardiol.
puted tomography. J Heart Lung Transplant. 2007;26(5):478–84. 2010;55(15):1566–8.
 Part VII
Where We Are: CT Assessment of Ventricular
Function
 CT of Cardiac Function and Wall Motion
35
Prabhakar Rajiah and Suhny Abbara

Functional evaluation is an important component of cardio- may be limited by acoustic windows, especially in patients
vascular imaging, contributing to diagnosis, risk stratifica- with obesity, COPD, chest wall deformities, and surgeries,
tion, prognosis, determining optimal therapy, and follow-up particularly in the evaluation of RV.  It also relies on geo-
after treatment. For example, left ventricular (LV) function is metrical assumptions, which make evaluation of remodeled
an important predictor of survival in patients following myo- and complex hearts challenging. There is also poor contrast
cardial infarction [1] and CABG surgery [2] as well as other between blood and myocardium, as a result of which it has
nonischemic cardiomyopathies. Right ventricular (RV) func- lower accuracy and reproducibility [3]. While LV volume is
tion is a prognostic determinant in conditions such as acute obtained by Simpson’s rule, mass is evaluated in M-mode.
pulmonary embolism (PE), congenital heart disease (CHD), Transesophageal echocardiography (TEE) has improved
and arrhythmogenic right ventricular dysplasia (ARVD). acoustic window, but there are associated complications of
Left atrial function is a predictor of recurrence of atrial fibril- an invasive procedure and general anesthesia. New tech-
lation after pulmonary vein ablation. Although cardiac CT is niques such as 3D result in improved quantification, with
primarily utilized in the evaluation of coronary arteries, it is higher accuracy and reproducibility [3], but add to the time
also a valuable tool in the evaluation of cardiac function as and are prone to artifacts.
well, particularly when echocardiography is not adequate CMR has several advantages including good spatial and
and cardiovascular magnetic resonance (CMR) is contraindi- temporal resolutions, high contrast between blood and myo-
cated or associated with significant artifacts. cardium which allows accurate delineation of the endocar-
In this chapter, we review the role and technique of CT in dial contour, no radiation, and no geometrical assumptions.
the evaluation of ventricular and atrial function. It can perform both qualitative and quantitative evaluation of
global and regional systolic function, with possibilities of
diastolic functional evaluation as well. CMR has been shown
Imaging of Cardiac Function to be the gold standard in quantification of ventricular func-
tion, due to high accuracy and reproducibility [4].
There are several imaging modalities available for the evalu- Quantification is usually performed in short-axis view utiliz-
ation of ventricular and atrial function, including echocar- ing modified Simpson’s method. Cine images in other planes
diography, magnetic resonance imaging (MRI), nuclear including two, three, and four chamber are used to evaluate
medicine, and angiography. Echocardiography is the most regional wall motion abnormalities. CMR also allows for
commonly used modality for the evaluation of cardiac func- anatomical evaluation and tissue characterization in the same
tion. It is widely available, inexpensive, portable, can be per- study. However, CMR is contraindicated in patients with
formed in hemodynamically unstable patients at bedside, several intracardiac devices as well as claustrophobic
does not have radiation, and there are no major contraindica- patients. CMR is also not widely available, is expensive,
tions. However, echocardiography is operator dependent and requires higher technical expertise, takes longer time, and is
challenging in patients with arrhythmia or those unable to
lay flat. In children, intubation and anesthesia are often
required.
Nuclear medicine techniques for quantification of LV
P. Rajiah · S. Abbara (*) function include gated radionuclide ventriculography (RVG),
Department of Radiology, Cardiothoracic Imaging Division, UT
ECG-gated single-photon emission computed tomography
Southwestern Medical Center, Dallas, TX, USA
[email protected] (SPECT), and positron emission tomography (PET). With

© Humana Press 2019 407

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_35
408 P. Rajiah and S. Abbara

RVG or MUGA (multiple-gated acquisition), patient’s RBCs available, and has rapid turnaround time, making it suitable
are labelled, and ECG-gated cardiac scintigraphy is obtained for patients who cannot lay still for prolonged periods of
over several heartbeats to generate a single composite car- time. For functional evaluation, CT is typically utilized when
diac cycle. This can quantify global and regional function MRI is contraindicated or suboptimal due to artifacts or if
and volumes, both at rest and stress. The advantage of this echocardiography does not provide adequate information
technique is that it does not rely on geometrical assumptions due to poor acoustic window, especially for right ventricle.
[5]. Disadvantages include lower temporal and spatial reso- Common contraindications for MRI include metallic objects
lution and limited accuracy in arrhythmias. There are long in the eye, older versions of intracardiac devices (pacemak-
preparation and examination times. Function can also be ers, defibrillators), claustrophobia, and inability to lay flat for
obtained from a gated SPECT utilized for myocardial perfu- prolonged period. CT for functional evaluation has been con-
sion imaging, by analyzing the ventricular contours [6], sidered an appropriate indication in the 2010 ACCF guide-
which correlates well with CMR [7]. Regional EF with hand- lines [10]. Disadvantages of CT include the use of ionizing
grip has been shown to be highly sensitive and specific for radiation which is associated with several theoretical risks
coronary artery disease [8]. Disadvantages of SPECT include and the use of contrast media in patients with renal dysfunc-
lower spatial resolution, inaccuracy in ischemia/infarct due tion which is associated with nephrotoxicity. Premedication
to lower uptake in subendocardial region [9], and decreased is required for patients with history of contrast allergy. Image
accuracy in attenuation artifacts as well as patients with quality is compromised in patients with severe arrhythmias,
small or large hearts [3]. Variability has been noted with with poor definition of borders, and with potential increases
Tl-201 scans and between different software algorithms [3]. of radiation dose in some scanners due to the opening of
Repeated radionuclide injections and radiation dose are image acquisition window.
additional concerns. Similarly, function can be obtained
from ECG-gated PET performed for myocardial perfusion
and metabolism imaging, which correlates well with CMR Scan Mode
[6]. PET has higher spatial resolution than SPECT but has
lower temporal resolution than CMR, which results in under- Cardiac cycle starts with isovolumetric contraction at the end
estimation of function. Due to inclusion of papillary muscles of ventricular filling, followed sequentially by ventricular
with myocardial wall, the volumes are lower than CMR [6]. systole, isovolumic ventricular relaxation, and then diastole.
Radiation is a concern with all the nuclear medicine To capture the cardiac volumes and function, ECG gating
techniques. and high temporal resolution are required to essentially
Cine ventriculography with catheter angiography is not freeze the cardiac motion without any motion artifacts. Since
routinely used for quantification of ventricular function. It is multiple cardiac phases are required in the R-R interval, par-
an invasive procedure, requires contrast, uses radiation, has ticularly end-systole (ES) and end-diastole (ED), retrospec-
geometric assumptions and restrictions from projection tive ECG gating is required. The heart is scanned in the spiral
images. It is limited in evaluation of some regions, especially mode, with the table continuously moving through the gan-
in complex, irregular shapes, and has significant variability try, and images are acquired throughout the cardiac cycle,
in regional wall motion assessment due to variable cardiac over multiple heartbeats. ECG tracing is also simultaneously
rotation and lack of reliable landmarks [3]. acquired, and images are retrospectively reconstructed. A
low pitch (<1.0), dependent on the heart rate, is utilized to
image the heart throughout R-R interval with areas of overs-
Computed Tomography ampling. Due to this, there is an overlap of the same anatomy
at different R-R intervals, which can be used for multi-­
Advantages and Disadvantages segment reconstruction. High spatial and temporal resolu-
tions are required for accurate quantification. The highest
CT scan has several advantages in the evaluation of ventricu- possible spatial resolution should be selected to accurately
lar function such as high isotropic spatial resolution, good delineate the endocardial borders and minimize partial vol-
temporal resolution, high contrast between ventricular lumen ume averaging at the blood pool-myocardium interface
and myocardium, and multi-planar reconstruction capabili- which can be seen at the apex [3].
ties. CT quantification does not rely on geometric assump- High temporal resolution is required to achieve artifact-­
tions. Additional information can be obtained on the free images of the myocardial contraction. A temporal reso-
morphology of the heart as well as coronary arteries and any lution of 100–250 milliseconds (ms) is recommended for
adjacent structure. CT is not operator dependent, is widely the evaluation of systole and diastole and higher if resolu-
35  CT of Cardiac Function and Wall Motion 409

tion of rapid ventricular filling and ejection phase is recom- RV, i.e., >176 HU, the volume measurements are more com-
mended. 80–90 milliseconds has been suggested for parable with CMR [17]. In the first phase, a full dose of con-
adequate evaluation of wall motion in CMR [3, 11], but trast is administered at a rate of 4–7 ml/s to opacify the left
19.1 ms or less is necessary to avoid motion artifacts [12], ventricle. In the second phase, either a slower injection of
especially in the posterior LV wall [11]. The temporal reso- contrast at 2–4  ml/s or a contrast saline admixture (50:50,
lution of a scanner is determined primarily by the gantry 60:40, or 70:30) at the same rate of 4–7 ml/s is injected to
rotation time. Since only a partial scan reconstruction algo- opacify the right ventricle. In the final phase, saline flush is
rithm is used, effective temporal resolution is the time taken injected at 5–7 ml/s to eliminate the streak artifacts in SVC
to rotate the tube 180 plus fan angle. In the earlier scanners and the right atrium. Scan acquisition is triggered either by
with few detectors, the gantry rotation time was long, with using a bolus tracking technique, with the ROI placed in the
limited temporal resolution, as a result of which, the EDV descending aorta, or a timing bolus technique following
can be underestimated and ESV can be overestimated since injection of a small bolus of contrast first, although this takes
the position of ventricles at EDV and ESV is averaged over additional time and also wastes some amount of contrast
multiple heartbeats. However, improvements in scanner [18]. Bolus tracking has been shown to provide more homo-
technology over the last decade have resulted in shorter gan- geneous enhancement [19]. Scanning is typically 4–6 s after
try rotation times and improved temporal resolution. This is peak aortic opacification [20].
further improved in dual-source scanners, where the tempo-
ral resolution is one fourth of the gantry rotation time,
resulting in effective temporal resolution of up to 75 milli- Radiation Dose Reduction Strategies
seconds. Another option to improve temporal resolution in a
spiral acquisition is to use multi-segment reconstruction, With retrospective ECG gating, the radiation dose is sig-
i.e., reconstruct data from same phase of cardiac cycle in nificantly high since data is obtained throughout the cardiac
consecutive heartbeats. The temporal resolution is equal to cycle (Fig. 35.1), with values from 14.8 to 21.1 mSv [21].
gantry rotation time divided by 2n, with n being the number However, the radiation dose can be reduced by using ECG-­
of segments. The higher the number of segments, the higher based tube current modulation, where the peak tube current
will be the temporal resolution. However, this requires a is delivered only in specific cardiac phase(s), i.e., ED and
steady heart rate; otherwise there will be beat to beat varia- ES, and ramped down to 4–20% of peak tube current for the
tion of the position of the heart resulting in artifacts. Even rest of the R-R interval (Fig. 35.2). This technique has been
with steady heartbeat, the heart may not be in the same loca- shown to allow radiation dose savings of up to 60% [22,
tion due to complex cardiac contraction. Multi-segment 23]. A prospective ECG-triggering mode, where data is
reconstruction may also require reduction of spiral pitch, acquired only in one portion of the cardiac cycle, is not use-
which will increase scan time and thus radiation dose. ful for estimation of function. However, prospective ECG
Motion artifacts can be eliminated by using wide-array or gating with wider padding can be utilized to evaluate car-
volume scanners with z-coverage up to 16 cm or dual-source diac function, although this does not provide any signifi-
scanners with high pitch (up to 3.4), as a result of which the cant radiation dose savings compared to retrospective ECG
entire heart can be scanned in one heartbeat. gating [24]. Another technique is a prospectively ECG-
triggered dual-­step pulsing protocol in a dual-source scan-
ner, in which a prospectively ECG-triggered window is
Medications and Contrast obtained at low radiation dose (20% tube current) across a
selected portion of cardiac cycle (10–90%), with a full tube
Beta-blockers are routinely used in coronary CTA to reduce current only in a selective portion of cardiac cycle (65–75%
the heart rate and minimize coronary artery motion, but this RR). X-ray tube is also turned off in between heartbeats.
is not required for evaluation of cardiac function. In addition, This provides both functional data at lower radiation dose
studies have shown there is a small but not negligible reduc- as well as normal radiation dose data for detailed evalua-
tion in LV function after administration of propranolol in tion of coronary arteries. This provides accurate quantifica-
healthy subjects [13, 14] as well as those with heart disease tion compared to CMR with lower radiation dose
[15, 16], due to negative inotropic action and increased ESV. (6.2 + 1.8 mSv) [21]. Other radiation dose reduction strate-
Contrast injection should be optimized depending on the gies can also be applied for functional evaluation, including
chamber of interest. If both ventricles are being evaluated, a lowering tube voltage, lowering tube current, automatic
triphasic injection protocol is used to adequately opacify tube current modulation techniques, and iterative recon-
both the ventricles. With adequate contrast opacification of struction algorithms.
410 P. Rajiah and S. Abbara

Fig. 35.1  Retrospective ECG gating without ECG-based tube current systole (bottom row). Since full tube current is delivered throughout the
modulation. Short-axis (left), two-chamber (middle), and four-chamber cardiac cycle, there is no significant difference in image quality between
(right) reconstructions of the heart at end-diastole (top row) and end-­ the two sets of images, albeit a high radiation dose

Fig. 35.2  Retrospective ECG gating with ECG-based tube current images since these images are obtained at 20% of maximum tube cur-
modulation. Short-axis (left), two-chamber (middle), and four-chamber rent. In spite of the higher noise at end-systole, these images are ade-
(right) reconstructions of the heart at end-diastole (top row) and end-­ quate for delineation of endocardial and epicardial contours for
systole (bottom row). Note the higher noise seen in the end-systolic functional, volume, and mass measurements
35  CT of Cardiac Function and Wall Motion 411

Image Reconstruction cross-sectional area and S is the section thickness.

Multiple parallel short-axis slices are utilized.
After acquisition of data, images are constructed at 5–10% Endocardial contours are drawn, either manually, auto-
increments throughout the R-R interval, generating either matically, or semiautomatically, both at end-diastole
10 or 20 cardiac phases. In a single-source scanner, no and end-systole in all the slices along the entire length
advantage is gained by using 20 instead of 10 phases, but of ventricle (Fig. 35.3). Although the LV limits can be
there is an increase in reconstruction time and space taken automatically determined using the software, determi-
for storing these images. In dual-source scanners with good nation of the most basal and apical slices can be chal-
temporal ­resolution, the full benefit of good temporal reso- lenging. The most basal slice is just forward to the
lution may be gained by using 20 phases [20]. No significant atrioventricular ring and has myocardium in at least
difference in EF and volumes was noted between 10 and 20 50% of its perimeter. The most apical slice is the last
phases, both for single-source [25] and dual-source scan- image with contrast-­opacified lumen. Determination of
ners [26]. Images can be displayed as 2D cine loops, and the correct ED and ES phase can also be done manually,
most workstations allow manipulation of the planes in any semiautomatically, or automatically. End-diastole is
direction. Typically short-axis, two-chamber, three-cham- defined as QRS onset, which corresponds to the phase
ber, and four-chamber planes are used for qualitative global with the largest cardiac dimension and following mitral
and regional functional evaluation. Thin slices are not man- valve closure. End-­systole is the phase where the car-
datory for evaluation of function, with 1.5, 2.0, or 2.5 mm diac dimension is the smallest and precedes mitral valve
slices with no overlap of intervals of 1 or 1.5 mm commonly opening. Papillary muscles are included in the ventricu-
used. Noise reduction strategies, including iterative recon- lar lumen. For evaluation of myocardial volume, epicar-
struction, increase the CNR and hence improve the correla- dial contours are drawn in end-diastolic phase, and the
tion of ventricular volumes, mass, and function with CMR mass is calculated by using the formula ∑AN × S x ῤ,
[27]. where ῤ is the myocardial density (1.05  g/cm3). This
three-dimensional technique is considered accurate
since it does not rely on geometric assumptions. It has
Post Processing been shown superior to the area-length technique, which
overestimates EDV, ESV, and SV [28] and agrees more
Quantitation of the volumes and function can be performed with 2D echo, MRI, CVG, and gated SPECT. Sources
using either manual, semiautomatic, or fully automatic of errors include inappropriate plane, selecting basal
techniques. and apical slices, selecting ED and ES phases, and
drawing endocardial and epicardial borders. Manual
(a) Area-length method. In this technique, volume of the left contouring, even with semiautomatic software, is time-
ventricle (V) is calculated from the area (A) and length consuming, subjective, and observer dependent.
Semiautomatic and automatic techniques significantly
8 A2
(L) of the left ventricle, using the formula V = ´ shorten the processing time due to advanced software,
3 p L.
automatic generation of cardiac views, automatic selec-
This is measured in one of the long-axis views (vertical tion of ED and ES phases, and automatic segmentation.
or horizontal). The length is measured from mitral Automatic techniques also decrease interobserver vari-
annulus to the apex, and the area is measured in the ability without significant changes in volumes and func-
same plane. This technique uses geometrical assump- tion [29–31], although some studies showed differences
tions to measure the volume, i.e., assuming an ellipsoid in LVEDV [29].
shape.
(b) Biplane area-length method. In this variation of the area-­
length product, both the vertical and horizontal long axes Threshold-Based Segmentation
are used to calculate the volume. This method also relies
on geometrical assumptions. Threshold-based segmentation is an automatic technique
(c) Simpson’s method. In Simpson’s method, the volumes that utilizes the differences in attenuation between the
are determined by summing up areas of LV cavity at blood pool and myocardium. The mitral annular plane
each short-axis slice and multiplying it with slice thick- may have to be verified or defined manually since the con-
ness, using the formula volume ∑AN × S, where A is the trast attenuation is similar in the atrium and ventricle,
412 P. Rajiah and S. Abbara

f­ollowing which the contrast-filled LV is automatically ation material such as dense contrast in veins or right
segmented in all the phases. Ventricular volume is calcu- heart or pacemaker leads, which may lead to inclusion of
lated by summing all the contiguous voxels whose attenu- RA or RV in LV blood pool. Similar errors can also ema-
ation is above that of a predefined threshold (Fig.  35.4). nate from unbroken column of contrast in VSD [30, 32].
Although volumes are typically calculated from short- In this technique, the papillary muscles are not included
axis view in 2D, recent software also allow 3D volume in the blood pool due to their different attenuation and
quantification. Segmented pixels are displayed in color. If hence do not contribute to the volume, as a result of which
the segmentation does not look accurate, the attenuation the EDV and ESV are smaller, with no change in EF [20].
threshold can be altered to make the color fit the cavity. In Threshold-­based technique has been shown to be accurate
most softwares, the LV function automation works read- and reproducible with lower interobserver variability with
ily, while RV function requires some kind of manual cor- shorter-processing time [3]. With manual contours, the
rection. Time-volume curves can be generated since all papillary muscles are within the blood pool, and hence the
the phases are segmented. This technique requires good volumes are larger, which correspond to MRI numbers
opacification of the blood pool with contrast, with at least due to similar measurement [33]. A study found differ-
150–200 HU attenuation difference between myocardium ences of 5.6–30.1% for LV volumes, 5.8–9.4% for LV
and blood pool. Errors will be seen if there is high attenu- mass, and 4.3–6.0% for LV EF [34].

Fig. 35.3  Quantification of the left ventricle. (a) Values obtained from (left), two-­chamber (middle), and four-chamber (right) reconstructions
the above contours include ejection fraction, end-diastolic volume, end-­ of the heart at end-diastole (top row) and end-systole (bottom row).
systolic volume, stroke volume, cardiac output, and myocardial mass. Endocardial (red) and epicardial (green) contours are drawn in the left
All these can be indexed to the body surface area (BSA) (b) Short-axis ventricle, both in end-diastolic and end-systolic images.
35  CT of Cardiac Function and Wall Motion 413

b
Standard Mode
ml
ES (LV) ED (LV)
LV 36.88ml 99.65ml

0% 10% 20% 30% 40% 50% 60% 70% 80% 90%

0ms 65ms 131ms 196ms 261ms 326ms 392ms 457ms 522ms 588ms

Standard Values Indexed Values

Standard Mode LV Normal Values

Ejection Fraction % 63 56 - 78
Myocardial Mass ED g 96.82 75 - 175
Stroke Volume ml 62.77 33 - 97
ED Volume ml 99.55 58 - 154
ES Volume ml 36.88 13 - 51
Cursor Volume ml 36.88
Cardiac output (I/min) 5.71 2.65 - 5.98

Fig. 35. 3 (continued)

Fig. 35.4  Quantification using threshold segmentation. Short-axis attenuation between contrast in the blood pool and the myocardium. All
(left), two-chamber (middle), and four-chamber (right) reconstructions the contiguous voxels whose attenuation is above that of a predefined
of the heart at end-diastole (top row) and end-systole (bottom row). The threshold are identified in pink. Ventricular volume is calculated by
threshold-based segmentation technique utilizes the differences in summing all these voxels
414 P. Rajiah and S. Abbara

Interpretation of Results myocardial volume × density. Usually, a density of 1.05 g/

cm3 is used. All these measurements can be indexed to the
Qualitative body surface area (BSA) to derive indexed values. If images
are reconstructed in all cardiac phases, a time-volume curve
Global function can be qualitatively evaluated in cine images can also be generated, which provides additional time-­
at multiple planes. All the currently available softwares auto- dependent variables such as peak ejection rate (PER), peak
matically generate the standard cardiac planes, typically a filling rate (PFR), time to PER, and time to PFR.
short-axis view and two long-axis views. Any other plane can Regional wall motion can also be quantified using several
also be generated by manual interaction. Global systolic func- indices if at least 8–10 cardiac phases are available. Normal
tion can be qualitatively classified as normal, reduced (mild, LV ejection is a combination of radial wall thickening, cir-
moderate, and severe), or hyperdynamic. Regional wall cumferential shortening, and longitudinal shortening [11].
motion abnormalities can be evaluated using the 17-­segment Parameters evaluated include myocardial wall thickness,
AHA model and multi-planar CT reconstructions, including which is calculated in orthogonal plane between endocardial
short-axis, two-chamber, three-chamber, and four-chamber and epicardial contour. Myocardium normally measures
views. There are six basal, six mid, and four apical segments, 6–8 mm in end-diastole and 10–14 mm in end-systole [35].
accounting for 35%, 35%, and 30% of myocardial volume, It is focally decreased in chronic ischemia or thickened in
respectively along with an apical cap. Regional wall motion cardiomyopathy [36, 37]; wall attenuation, mean attenuation
can be classified as normal (inward motion in systole), hypoki- of the myocardial segment, which is decreased in infarction;
nesis (decreased wall motion), akinesis (no wall motion), and and systolic percentage wall thickening, which represents
dyskinesis (paradoxical wall motion). Abnormalities have to be regional percentage of wall thickening during systole (Ws-­
seen in two different views to be considered true positive. Wd/Wd  ×  100%). Normal wall thickening is 5  mm. Wall
thickening is decreased in coronary artery stenosis, particu-
larly with stress imaging due to regional hypoperfusion. In
Quantitative ischemic heart disease, wall motion may be difficult to eval-
uate, and wall thickening percentage is more accurate [11]:
Several quantitative parameters are derived from the post-­ wall motion or shortening, displacement of points on the
processed images as above. Parameters of global systolic endocardial contour from ED to ES phase using displace-
function include the stroke volume, which is the volume of ment lines called chords, either in the longitudinal or circum-
blood ejected from the ventricle with each heartbeat, i.e., ferential plane. The length of each chord represents the
SV = EDV-ESV. This represents the blood ejected both for- amount of endocardial displacement; and regional/segmen-
ward and backward and hence includes regurgitant flow. The tal EF, which represents the EF relative to a specific myocar-
stroke volume of the right and left ventricles is the same, dial region/segment, is calculated by dividing ESV and EDV
assuming there is no regurgitation or shunting. Cardiac out- for a specific area beneath a myocardial segment or region
put (CO), the volume of blood pumped in 1 min, is SV × HR [11]. Due to high variability of quantitative regional indices,
(ml/min). Ejection fraction is a percentage, which is SV/ qualitative or semiquantitative metrics are used, with graph-
EDV × 100%. Myocardial mass is quantified as end-diastolic ics, color polar maps, and diagrams (Fig. 35.5).

WALL THICKNESS WALL THICKENING WALL MOTION

10.00
1
1 1

7
7 7
2 6
2 6 13 2 6
8 12
8 13 12 8 13 12

14 17 16
14 17 16 14 17 16

9 11
9 15 9
15 11 15 11
3 5
3 5 3 5
10
10 10

4
4 4
0.00

Fig. 35.5  Quantification of regional function. Polar maps showing values of myocardial wall thickness, myocardial wall thickening, and myocar-
dial wall motion across the 17 myocardial segments as per AHA classification scheme
35  CT of Cardiac Function and Wall Motion 415

Diastolic Function Ischemic Heart Disease

Diastolic dysfunction is an important factor providing diag- LV dysfunction indicates CAD in symptomatic patients,
nostic, therapeutic, and prognostic information in several although it is not highly sensitive. Contraction remains nor-
disorders including CAD. This is typically performed with mal till a particular threshold for coronary blood flow is
echocardiography. CT-derived transmitral velocity, mitral reached, below which there is an exponential fall in LV func-
septal tissue velocity, and estimation of LV filling pressures tion. Dysfunction is also not specific and can happen in non-
have shown good correlation with echocardiogram [38] with ischemic cardiomyopathy, valvular heart disease, and other
comparable assessment of diastolic dysfunction; but this is disorders. Wall motion abnormalities can be seen due to
not widely used clinically. myocardial ischemia, hibernating myocardium, stunned
myocardium, or infarct, with the last entity being irrevers-
ible. In myocardial ischemia, regional wall motion abnor-
Left Ventricle malities are seen in the early stages of the disease. In acute
myocardial infarction, akinesia is seen due to stunned myo-
Normal functional values for the left ventricle are shown cardium, which may improve with revascularization [47].
in Table 35.1 [36, 37, 39, 40]. Earlier studies showed over- Hibernating myocardium also shows functional abnormality,
estimation of ESV and underestimation of EF in CT com- which improves with revascularization. CT has 89% sensi-
pared to CMR due to poorer temporal resolution of the tivity and 92% specificity in diagnosing abnormal segments
earlier scanners [41, 42]. Several newer studies have shown using CMR as gold standard [48]. In patients with acute
excellent correlation between CT and CMR values, with coronary syndrome (ACS) with negative troponin and EKG,
single- and dual-source scanners of several detector coronary CTA alone has a 77% accuracy and 35% positive
widths, both at single- and multi-segment reconstruction predictive value for detection of coronary stenosis, whereas
[33, 43, 44]. Other studies have shown good correlation adding functional CT improves accuracy to 87% and PPV to
with and even superior accuracy than cineventriculogra- 80% [49] and also providing incremental value [50]. Wall
phy and transthoracic echocardiography [45, 46]. motion abnormalities also highlight subtle coronary artery
Interobserver variability of CT quantification ranges from abnormalities that are not always evident on coronary
2% to 11% for LV EDV, 6% to 9% for ESV [11], and 2% CTA.  In chronic infarct, regional wall motion abnormality
to 8.5% for EF [3]. Most studies have also shown good such as akinesis or dyskinesis is seen with thinned myocar-
sensitivity and accuracy of regional LV function assess- dium in a vascular territory. Fat metaplasia, calcification,
ment with CT compared to CMR, echocardiography, aneurysm, or pseudoaneurysm may be seen. Low EF pre-
SPECT, and CVG [3]. dicts heart failure in patients with uncomplicated MI [51].
Global systolic dysfunction is seen in several cardiomy- ESV is an additional independent prognostic factor in
opathies, both ischemic and nonischemic. Heart failure can patients with previous MI and EF <40% [52]. Worsening LV
present with either reduced or preserved ejection fractions. EF and larger LV volumes predicted mortality in patients
Hyperdynamic systolic function is seen in hypertrophic car- with CAD in the CONFIRM registry [53]. In the CASS reg-
diomyopathies. LV function is important in the diagnosis, istry, low EF was a better predictor of mortality than the
management, and prognosis of patients with ischemic heart number of vessels involved [2].
disease, heart failure, malignant ventricular arrhythmia,
chronic valvular regurgitation, type 2 diabetes mellitus, and
congenital heart diseases [3]. Heart Failure

Heart failure can be with preserved (HFpEF) or reduced

ejection fraction (HFrEF). In restrictive cardiomyopathies
Table 35.1  Normal regional LV functional parameters adapted from such as amyloidosis, sarcoidosis, hemochromatosis, and gly-
CMR data cogen storage disorders, the global systolic function is nor-
Men Women mal, but there is impaired diastolic relaxation. There is a
EDWT (mm) 7.6 ± 1.4 6.3 ± 1.0 direct correlation between LV EF and adverse outcomes
ESWT (mm) 13.2 ± 1.8 12.2 ± 1.6 including mortality. This relationship is not linear, but there
SWT (mm) 5.5 ± 0.8 5.8 ± 1.2
is a cutoff, such as 45%, below which there is a linear rela-
SWTH (%) 75 ± 16 96 ± 24
tionship [2]. Nonischemic cardiomyopathies can be dilated,
Data from Refs. [36, 37]
EDWT end-diastolic wall thickness, ESWT end-systolic wall thickness,
restrictive, hypertrophic cardiomyopathies, arrhythmogenic
SWT segmental wall thickening, SWTH segmental wall thickening right ventricular dysplasia (ARVD), and unclassified dis-
percentage ease. Most of these disorders have global systolic dysfunc-
416 P. Rajiah and S. Abbara

tion, with regional wall motion abnormalities rarely seen, In patients with left ventricular assist device (LVAD) for
such as in myocarditis, sarcoidosis, and stress-induced car- heart failure, CT provides accurate biventricular functional
diomyopathy. In dilated cardiomyopathy, the LV is dilated values, as well as comprehensive assessment of the device
with global systolic dysfunction. Hypertrophic cardiomyop- and its complications, such as location, angulation, throm-
athy is characterized by LV thickening of several types and bus, infection, fluid collections, pericardial effusion, and dia-
hyperdynamic global systolic function, with global systolic phragmatic hernia [57].
dysfunction seen in late phases of the disease. In
stress-­
­ induced cardiomyopathy (Takotsubo cardiomyopa-
thy), there is reversible global systolic dysfunction and hypo- Right Ventricle
kinesis/akinesis of the apical segments associated with
normal contraction of the basal segments in a nonvascular Right ventricle volumes and function can be qualitatively and
distribution, which resembles a Japanese octopus pot. In quantitatively evaluated using retrospective ECG-gated data
reverse Takotsubo cardiomyopathy, there is hypokinesis/aki- obtained from triphasic contrast injection protocol that opaci-
nesis of the basal segments and normal contraction of apical fies both the ventricles simultaneously. Manual, automated, or
segment. A common scenario is that in a patient with new semiautomated techniques can be used for quantifying vol-
onset heart failure and low to intermediate probability, CTA umes and function (Fig. 35.6), after determining the tricuspid
is used to exclude CAD, which can also provide functional annular plane. Normal values for the right ventricle in CT are
information in the same study [10]. Similarly, in patients shown in Table 35.2 [39]. Echocardiography is limited in the
being evaluated for CRT, CT can provide functional informa- evaluation of RV due to its location behind the sternum which
tion as well as venous anatomy in one study [20]. limits the acoustic window, complex shape which is exagger-
Intraventricular dysynchrony can also be evaluated to predict ated in abnormalities, heavy trabeculations, and thin wall [20].
response to CRT [20]. Ventricular function is also accurately CMR is considered the gold standard in the evaluation and
measured in CT in patients with cardiac transplants [54, 55] quantification of RV [58]. However, CT is the ideal modality
who are surviving longer, with 30% survival at 15 years [56]. in the evaluation of patients who have contraindications to

Fig. 35.6  Quantification of the right ventricle. (a) Short-axis (left), both in end-diastolic and end-systolic images. (b) Values obtained from
two-chamber (middle), and four-chamber (right) reconstructions of the the above contours include ejection fraction, end-­diastolic volume, end-
heart at end-diastole (top row) and end-systole (bottom row). Endocardial systolic volume, stroke volume, cardiac output, and myocardial mass.
(red) and epicardial (green) contours are drawn in the right ventricle All these can be indexed to the body surface area (BSA)
35  CT of Cardiac Function and Wall Motion 417

b
Standard Mode
ml
ED (RV) ES (RV)
RV 187.09ml 129.83ml

0% 10% 20% 30% 40% 50% 60% 70% 80% 90%

0ms 65ms 131ms 196ms 261ms 326ms 392ms 457ms 522ms 588ms

Standard Values Indexed Values

Standard Mode RV Normal Values

Ejection Fraction % 31 47 - 80
Myocardial Mass ED g - 24 - 55
Stroke Volume ml 57.27 33 - 98
ED Volume ml 187.09 58 - 154
ES Volume ml 129.83 23 - 103
Cursor Volume ml 129.83
Cardiac Output (l/min) 5.21 2.82 - 8.82

Fig. 35. 6 (continued)

Table 35.2  Normal right ventricular values

ESD (mm) 29.6 ± 5.3
and RV strain [62] (Fig. 35.7a). Other secondary signs of RV
EDD (mm) 37.0 ± 5.7 strain include septal straightening or bowing (Fig.  35.7b),
EDV (ml) 174.9 ± 48.0 which is seen throughout cardiac cycle due to pressure over-
ESV (ml) 82.1 ± 29.2 load and only in diastole for volume overload; dilated SVC
EF (%) 57.9 ± 8.0 (>25  mm), coronary sinus (>16  mm), and azygos vein
Adapted from Lin et al. [39], with permission (>25 mm); and reflux of contrast into IVC and hepatic veins
[63] (Fig. 35.7c). CT can be utilized for serial changes in RV
CMR or when artifacts are expected. Several studies have function and volumes, particularly with therapy and in those
shown the high accuracy and reproducibility of CT compared patients with contraindications for CMR (Fig. 35.8). RV dys-
to MRI in evaluation of RV function and volumes [59, 60]. function is an adverse prognostic determinant in other pulmo-
RV function is an important parameter in the diagnosis, nary disorders including cor pulmonale, interstitial lung
management, and prognosis of several disorders such as diseases, obstructive sleep disorders, and pulmonary hyper-
­pulmonary embolism (PE), pulmonary hypertension, ARVD, tension. For example, increased RVEDV and decreased EF
and congenital heart disorders. In acute PE, RV dysfunction is have been associated with lower survival in cor pulmonale
seen when there is >30–40% of pulmonary arterial tree [64]. Higher RV volumes have been observed in liver diseases
obstruction that has been shown to be an independent and due to hepatopulmonary syndrome, portopulmonary hyper-
powerful determinant of adverse prognosis [61], with higher tension, and TIPS therapy [65].
ICU admission and death within 30 days [11]. Early identifi- Congenital disorders including tetralogy of Fallot, trans-
cation of these high-risk patients will help in guiding them to position of great arteries, double outlet right ventricle, single
catheter-directed thrombolysis or surgical embolectomy to ventricle physiologies, and shunt lesions (ASD, VSD, PDA,
reduce recurrence and death. Although ECG-gated CT is ideal anomalous pulmonary venous return) may be associated
for quantification of RV volumes, in the setting of acute PE, with RV dilation and dysfunction [11]. For example, RV
ECG gating is not always available and also not necessary. hypertrophy/dilation and dysfunction are seen in D-TGA
The regular axial plane or a four-chamber reconstruction can following atrial switch procedure or in L-TGA due to inabil-
be used, and a RV/LV ratio >0.9 indicates elevated RV volume ity of the morphologic right ventricle (“systemic ventricle”)
418 P. Rajiah and S. Abbara

a b c

Fig. 35.7  Features of RV dysfunction in CT. (a) Axial CT scan shows RV/LV diameter ratio of 2.4. (b) Short-axis reconstruction of CT shows
dilated RV and bowing of the interventricular septum (arrow) toward the LV. (c) Reflux of contrast in hepatic veins and IVC (arrow)

a b c

Fig. 35.8  Serial changes of RV volumes. 3D volume-rendered images in a patient with pulmonary hypertension show changes in RV volumes
with time, with RVV/LVV values of 3.8 (a), 4.2 (b), and 2.6 (c) at 3 monthly intervals

to cope up with the systemic circulation. RV dysfunction and tricular dysplasia (ARVD). Recent task force criteria specify
pulmonic regurgitation are common complications seen after major criteria as a major wall motion abnormality (aneu-
surgical treatment of tetralogy of Fallot. RV function and rysm, dyskinesis, dysynchrony, akinesia) along with RV
volumes are important variables for repeat surgeries or inter- dilation (EDVi >110 ml/m2 in males, 100 ml/m2 in females)
ventional procedures such as percutaneous valve implanta- or RV dysfunction (<40%). EDVi>100  ml/m2 in males,
tion. For example, pulmonary valve replacement is indicated 90 ml/m2 in females, or RV dysfunction (40–45%) is consid-
in these patients when RVEDVI <170 ml/m2 or RVDESVI ered as a minor criteria, although these criteria are derived
<85  ml/m2 [66] with lower cutoffs for chronic pulmonic from MRI [68]. ARVD is an appropriate indication for car-
regurgitation of 163 ml/m2 and 80 ml/m2, respectively [67]. diac CT in 2010 ACCF guidelines [10]. Fatty tissue, bulging,
Abnormal RV volumes and function are seen in several and RV dilation help in distinguishing ARVD from tachyar-
cardiomyopathies, particularly arrhythmogenic right ven- rhythmia [69].
35  CT of Cardiac Function and Wall Motion 419

Fig. 35.9  Left atrial function. Short-axis (left), two-chamber (middle), the left atrium both in end-diastole and end-systole, the left atrial vol-
and 3D volume-rendered (right) reconstructions of the heart at end-­ umes and function can be quantified. All these can be indexed to the
diastole (top row) and end-systole (bottom row). By segmenting (green) body surface area (BSA)

Left Atrium Conclusion

Left atrial volumes and function can be obtained from the CT provides accurate and reliable quantification of ventricu-
same image data that is utilized for LV and RV quantifica- lar and atrial volumes and function, comparable to MRI and
tion. Manual, semiautomated, or automated techniques are echocardiography. CT is used in specific circumstances
used to draw endocardial contours in the ED and ES phases where echocardiography is inadequate and MRI cannot be
with Simpson’s method or used in threshold-based segmen- performed due to contraindications or artifacts. Retrospective
tation (Fig. 35.9). ECG-gated spiral acquisitions are required for functional
Normal atrial filling has three phases, with an early rapid evaluation in CT, with the radiation dose minimized by using
filling accounting for 75–80%, followed by diastasis ECG-based tube current modulation.
accounting for 5%, and then an atrial systole accounting for
15–25% of LV blood volume [70]. Impaired relaxation of
LV, with increased LA volume and lower LA EF, is seen in References
patients with restrictive cardiomyopathies (heart failure
with preserved ejection fraction), and this can be used to 1. Hammermeister KE, DeRouen TA, Dodge HT.  Variables predic-
tive of survival in patients with coronary disease. Selection by
distinguish it from other causes of LV hypertrophy includ- univariate and multivariate analyses from the clinical, electrocar-
ing hypertension [71]. Although echocardiography is the diographic, exercise, arteriographic, and quantitative angiographic
most commonly used modality for evaluation of LA func- evaluations. Circulation. 1979;59(3):421–30.
tion, it has several limitations as discussed above. Body 2. Mock MB, Ringquist I, Fisher LD, et  al. Survival of medically
treated patients in the coronary artery surgery study (CASS) regis-
habitus and poor lateral resolution make it difficult to delin- try. Circulation. 1982;66:562–8.
eate the endocardium accurately in the posteriorly located 3. Sayyed SH, Cassidy MM, Hadi MA.  Use of multidetector com-
left atrium, which results in lower atrial volumes and low puted tomography for evaluation of global and regional left ventric-
accuracy and reproducibility [72–75]. CT values correlate ular function. J Cardiovasc Comput Tomogr. 2009;3(S1):S23–34.
4. Setser RM, Eischer SE, Lorenz CH. Quantification of left ventricu-
with those of CMR, but usually, the LA volume is overesti- lar function with magnetic resonance images acquired in real time.
mated, and LA EF is underestimated [76]. J Magn Reson Imaging. 2000;12:430–8.
420 P. Rajiah and S. Abbara

5. de Geus-Oei LF, Mavinkurve-Groothuis AM, Bellerson L, et  al. 24. Feuchtner G, Goetti R, Plass A, et al. Dual-step prospective ECG-­
Scintigraphic techniques for early detection of cancer treatment-­ triggered 128-slice dual source CT for evaluation of coronary arter-
induced cardiotoxicity. J Nucl Med. 2011;52(4):560–71. ies and cardiac function without heart rate control a technical note.
6. Slart RH, Bax JJ, de Jong RM, et  al. Comparison of gated PET Eur Radiol. 2010;20:2092–9.
with MRI for evaluation of left ventricular function in patients with 25. Ko YJ, Kim SS, Park WJ, et  al. Comparison of global left ven-
coronary artery disease. J Nucl Med. 2004;45(2):176–82. tricular function using 20 phases with 10-phase reconstructions
7. Bavelaar-Croon CD, Kayser HW, van der Wall EE, et  al. Left in multidetector-­ row computed tomography. Int J Cardiovasc
ventricular function: correlation of quantitative gated SPECT Imaging. 2012;28(3):603–11.
and MR imaging over a wide range of values. Radiology. 26. Puesken M, Fischbach R, Wenker M, et al. Global left-ventricular
2000;217(2):572–5. function assessment using dual-source multidetector CT: effect of
8. Bodenheimer MMBV, Fooshee CM, Hermann GA, et  al. improved temporal resolution on ventricular volume measurement.
Comparison of wall motion and regional ejection fraction at rest Eur Radiol. 2008;18(10):2087–94.
and during isometric exercise: concise communication. J Nucl 27. Wai B, Thai WE, Brown H, et al. Novel phase-based noise reduc-
Med. 1979;20:724–32. tion strategy for quantification of left ventricular function and mass
9. Stollfuss JC, Haas F, Matsunari I, et al. Regional myocardial wall assessment by cardiac CT: comparison with cardiac magnetic reso-
thickening and global ejection fraction in patients with low angi- nance. Eur J Radiol. 2013;82:e337–41.
ographic left ventricular ejection fraction assessed by visual and 28. Lessick J, Ghersin E, Abadi S, Yalonetsky S. Accuracy of the long-­
quantitative resting ECG-gated 99mTc-tetrofosmin single-photon axis area-length method for the measurement of left ventricular
emission tomography and magnetic resonance imaging. Eur J Nucl volumes and ejection fraction using multidetector computed tomo-
Med. 1998;25(5):522–30. graph. Can J Cardiol. 2008;24(9):685–9.
10. Taylor AJ, Cerqueira M, Hodgson JM, et  al. ACCF/SCCT/
29. Greupner J, Zimmermann E, Hamm B, Dewey M. Automatic ver-
ACR/AHA/ASE/ASNC/SCAI/SCMR 2010 appropriate use sus semiautomatic global cardiac function assessment using 64-row
criteria for cardiac computed tomography. J Am Coll Cardiol. computed tomography. Br J Radiol. 2012;85(1015):e243–53.
2010;56(22):1864–94. 30. Juergens KU, Seifarth H, Range F, et  al. Automated threshold-­
11. Savino G, Zwerner P, Herzog C, et  al. CT of cardiac function. J based 3D segmentation versus short-axis planimetry for assessment
Thorac Imaging. 2007;22:86–100. of global left ventricular function with dual-source MDCT. Am J
12. Ritchie C, Godwin J, Crawford C, et al. Minimum scan speeds for Roentgenol. 2008;190(2):308–14.
suppression of motion artifacts in CT. Radiology. 1992;185:37–42. 31. Plumhans C, Keil S, Ocklenburg C, et al. Comparison of manual,
13. Port S, Cobb FR, Jones RH. Effects of propranolol on left ventricu- semi- and fully automated heart segmentation for assessing global
lar function in normal men. Circulation. 1989;61(12):358–66. left ventricular function in multidetector computed tomography.
14. Mo YH, Jaw FS, Wang YC, et al. Effects of propranolol on the left Investig Radiol. 2009;44(8):476–82.
ventricular volume of normal subjects during CT coronary angiog- 32. van Ooijen PM, de Jonge GJ, Oudkerk M.  Informatics in radi-
raphy. Korean J Radiol. 2011;12(3):319. ology: Postprocessing pitfalls in using CT for automatic and
15. Dell'Italia LJ, Walsh RA. Effect of intravenous metoprolol on left semiautomatic determination of global left ventricular function.
ventricular performance in Q-wave acute myocardial infarction. Radiographics. 2012;32(2):589–99.
Am J Cardiol. 1989;63(3):166–71. 33. de Jonge GJ, can der Vleuten PA, Overbosch J, et al. Semiautomatic
16. Jensen CJ, Jochims M, Hunold P, et al. Assessment of left ventricu- measurement of left ventricular function on dual source computed
lar function and mass in dual-source computed tomography coro- tomography using five different software tools in comparison with
nary angiography: influence of beta-blockers on left ventricular magnetic resonance imaging. Eur J Radiol. 2011;80(3):755–66.
function: comparison to magnetic resonance imaging. Eur J Radiol. 34. Mao SS, Li D, Rosenthal DG, et al. Dual-standard reference values
2010;74(3):484–91. of left ventricular volumetric parameters by multidetector CT angi-
17. Lee H, Kim SY, Gebregziabher M, et al. Impact of ventricular con- ography. J Cardiovasc Comput Tomogr. 2013;7:234–40.
trast medium attenuation on the accuracy of left and right ventricu- 35. Juergens UK, Fischbach R. Left ventricular function studied with
lar function analysis at cardiac multi detector-row CT compared MDCT. Eur Radiol. 2006;16:342–57.
with cardiac MRI. Acad Radiol. 2012;19(4):395–405. 36. Alfakih K, Plein S, Thiele H, et  al. Normal human left and right
18. Gao Y, Du X, Liang L, et al. Evaluation of right ventricular func- ventricular dimensions for MRI as assessed by turbo gradient echo
tion by 64-row CT in patients with chronic obstructive pulmonary and steady state free precession imaging sequences. J Magn Reson
disease and cor pulmonale. Eur J Radiol. 2012;81(2):345–53. Imaging. 2003;17:323–9.
19. Cademartiri F, Nieman K, van der Lugt A, et al. Intravenous con- 37. Sandstede J, Lipke C, Beer M, et al. Age and gender specific dif-
trast material administration at 16-detector row helical CT coronary ferences in left and right ventricular cardiac function and mass
angiography: test bolus versus bolus-tracking technique. Radiology. determined by cine magnetic resonance imaging. Eur Radiol.
2004;233(3):817–23. 2000;10:438–42.
20. Rizvi A, Deano RC, Bachman DP, et  al. Analysis of ventricular 38. Boogers MJ, van Werkhoven JM, Shuijf JD, et  al. Feasibility of
function by computed tomography. J Cardiovasc Comput Tomogr. diastolic functional assessment with cardiac CT: feasibility study
2015;9(1):1–12. in comparison with tissue Doppler imaging. JACC Cardiovasc
21. Takx RAP, Moscariello A, Schoepf UJ, et al. Quantification of left Imaging. 2011;4:246–56.
and right ventricular function and myocardial mass: comparison 39. Lin FY, et  al. Cardiac chamber volumes, function, and mass as
of low-radiation dose 2nd generation dual-source CT and cardiac determined by 64-multidetector row computed tomography: mean
MRI. Eur J Radiol. 2012;81(4):e598–604. values among healthy adults free of hypertension and obesity.
22. Hausleiter J, Meyer T, Hadamitzky M, et al. Radiation dose esti- JACC Cardiovasc Imaging. 2008;1(6):782–6.
mates from cardiac multislice computed tomography in daily prac- 40. Nevsky G, Jacobs JE, Lim RP, et al. Sex-specific normalized ref-
tice. Circulation. 2006;113:1305–10. erence values of heart and great vessel dimensions in cardiac CT
23. Hausleiter J, et  al. A new algorithm for ECG-based tube current angiography. Am J Roentgenol. 2011;196:788–94.
modulation (“MinDose”) reduces radiation dose estimates in car- 41. Juergens KU, Grude M, Maintz D, et  al. Multi-detector row CT
diac dual source CT angiography. Circulation. 2007;116(Suppl of left ventricular function with dedicated analysis software versus
16):II-575. MR imaging: initial experience. Radiology. 2004;230:403–10.
35  CT of Cardiac Function and Wall Motion 421

42. Mahnken AH, Speuntrup E, Neithammer M, et al. Quantitative and 58. Geva T. Is MRI the preferred method for evaluating right ventricu-
qualitative assessment of left ventricular volume with ECG-gated lar size and function in patients with congenital heart disease? Circ
multislice spiral CT: value of different image reconstruction algo- Cardiovasc Imaging. 2014;7:190–7.
rithms in comparison to MRI. Acta Radiol. 2003;44(6):604–11. 59. Raman SV, Shah M, McCarthy B, et al. Multi-detector row cardiac
43. van der Vleuten PA, Willems TP, Gotte MJ, et al. Quantification of computed tomography accurately quantifies right and left ventricu-
global left ventricular function: comparison of multidetector computed lar size and function compared with cardiac magnetic resonance.
tomography and magnetic resonance imaging. A meta-analysis and Am Heart J. 2006;151:736–44.
review of the current literature. Acta Radiol. 2006;47(10):1049–57. 60. Guo YK, Gao HL, Zhang XC, et  al. Accuracy and reproducibil-
44. Grude M, Juegens KU, Wichter T, et al. Evaluation of global left ity of assessing right ventricular function with 64-section multi-­
ventricular myocardial function with electrocardiogram gated mul- detector row CT: comparison with magnetic resonance imaging. Int
tidetector computed tomography. Comparison with magnetic reso- J Cardiol. 2010;139:254–62.
nance imaging. Investig Radiol. 2003;38:653–61. 61. Kang DK, Thilo C, Schoepf UJ, et al. CT signs of right ventricular
45. Asferg C, Usinger L, Kristensen TS, et al. Accuracy of multi-slice dysfunction. J Am Coll Cardiol Img. 2011;4(8):841–9.
computed tomography for measurement of left ventricular ejec- 62. Quiroz R, Kucher N, Schoepf UJ, et al. Right ventricular enlarge-
tion fraction compared with cardiac magnetic resonance imaging ment on chest computed tomography: prognostic role in acute pul-
and two-dimensional transthoracic echocardiography: a systematic monary embolism. Circulation. 2004;109(20):2401–4.
review and metaanalysis. Eur J Radiol. 2011;81:e756–62. 63. Staskiewicz G, Czekajska-Chehab E, Przegalinski J, et al. Widening
46. Dewey M, Muller M, Eddicks S, et  al. Evaluation of global and of coronary sinus in CT pulmonary angiography indicates right
regional left ventricular function with 16-slice computed tomog- ventricular dysfunction in patients with acute pulmonary embolism.
raphy, biplane cineventriculography and two-dimensional trans- Eur Radiol. 2010;20:1615–20.
thoracic echocardiography: comparison with magnetic resonance 64. Burgess MI, Mogulkoc N, Bright-Thomas RJ, et  al. Comparison
imaging. J Am Coll Cardiol. 2006;48:2034. of echocardiographic markers of right ventricular function in
47. Mahias-Narvarte AHKF, Willis PW. Evolution of regional left ven- determining prognosis in chronic pulmonary disease. J Am Soc
tricular wall motion abnormalities in acute Q and non-Q wave myo- Echocardiogr. 2002;15:633–9.
cardial infarction. American Heart Journal. 1987;113:1369–75. 65. Dupont MVM, Dragean CA, Coche EE.  Right ventricle function
48. Sarwar A, Shapiro MD, Nasir K, et al. Evaluating global and regional assessment by MDCT. Am J Roentgenol. 2011;196(1):77–86.
left ventricular function in patients with reperfused acute myocardial 66. Therrien J, Provost Y, Merchant N, et al. Optimal timing for pulmo-
infarction by 64-slice multidetector CT: a comparison to magnetic nary valve replacement in adults after tetralogy of Fallot repair. Am
resonance imaging. J Cardiovasc Comput Tomogr. 2009;3(3):170–7. J Cardiol. 2005;95:779–82.
49. Seneviratne SK, Troung QA, Bamberg F, et al. Incremental diag- 67. Lee C, Kim YM, Lee CH, et  al. Outcomes of pulmonary valve
nostic value of regional left ventricular function over coronary replacement in 170 patients with chronic pulmonary regurgitation
assessment by cardiac computed tomography for the detection of after relief of right ventricular outflow tract obstruction: implica-
acute coronary syndrome in patients with acute chest pain: from the tions for optimal timing of pulmonary valve replacement. J Am
ROMICAT trial. Circ Cardiovasc Imaging. 2010;3(8):375–83. Coll Cardiol. 2012;60:1005–14.
50. Bezerra HG, Loureiro R, Irlbeck T, et  al. Incremental value of 68. Marcus F, McKenna WJ, Sherrill D, et al. Diagnosis of arrhyth-
myocardial perfusion over regional left ventricular function and mogenic right ventricular cardiomyopathy/dysplasia: pro-
coronary stenosis by cardiac CT for detection of acute coro- posed modification of the task force criteria. Eur Heart J.
nary syndromes in high risk patients: a subgroup analysis of the 2010;31(7):806–14.
ROMICAT trial. J Cardiovasc Comput Tomogr. 2011;5:382–91. 69. Nakajima T, Kimura F, Kajimoto K, et  al. Utility of ECG-gated
51. Candell-Riera JLJ, Santana C, Castell J, et  al. Prognostic assess- MDCT to differentiate patients with ARVC/D from patients with
ment of uncomplicated first myocardial infarction by exercise ventricular tachyarrhythmias. J Cardiovasc Comput Tomogr.
echocardiography and Tc-99m tetrofosmin gated SPECT.  J Nucl 2013;7:223–33.
Cardiol. 2001;81:122–8. 70. Stojanovska J, Cronin P, Patel S, et al. Reference normal absolute
52. Antonini-Canterin FNG. Valutazione eoardiografica dei volume e and indexed values from ECG-gated MDCT: left atrial volume,
della funzione sistolica globale del ventricolo sinistro. Ital Heart J. function, and diameter. Am J Roentgenol. 2011;197(3):631–7.
2000;1:1261–72. 71. Melenovsky V, Borlaug B, Rosen, et  al. Cardiovascular features
53. Arsanjani R, Berman DS, Gransar H, et al. Left ventricular function of heart failure with preserved ejection fraction versus non-failing
and volume with coronary CT angiography improves risk stratifica- hypertensive left ventricular hypertrophy in the urban Baltimore
tion and identification of patients at risk for incident mortality: results community. J Am Coll Cardiol. 2007;49:198.
from 7758 patients in the prospective multinational CONFIRM 72. Avelar E, Durst R, Rosito GA, et al. Comparison of the accuracy
observational cohort study. Radiology. 2014;273(1):70–7 of multidetector computed tomography versus two-dimensional
54. Mastrobuoni S, Dell’aquila AM, Arraiza M, et  al. Allograft mor- echocardiography to measure left atrial volume. Am J Cardiol.
phology and function in heart transplant recipients surviving more 2010;106(1):104–9.
than 15 years by magnetic resonance imaging and dual-source com- 73. Rodevan O, Bjornerheim R, Ljosland M, et  al. Left atrial

puted tomography. Eur J Cardiothorac Surg. 2011;40(1):e62–6. volumes assessed by three- and two-dimensional echocar-
55. Bastarrika G, Arraiza M, DeCecco CN, et al. Quantification of left diography compared to MRI estimates. Int J Card Imaging.
ventricular function and mass in heart transplant recipients using 1999;15:397–410.
dual-source CT and MRI: initial clinical experience. Eur Radiol. 74. Christiaens L, Lequeux B, Ardilouze P, et  al. A new method for
2008;18(9):1784–90. measurement of LA volumes using 64-slice spiral CT: comparison
56. Taylor DO, Edwards LB, Aurora P, et al. Registry of the International with 2DE techniques. Int J Cardiol. 2009;131:217–24.
Society for Heart and Lung Transplantation: twenty-fifth official 75. Kircher B, Abbott JA, Paul S, et al. Left atrial volume determination
adult heart transplant report—2008. J Heart Lung Transplant. by biplane two-dimensional echocardiography: validation by cine
2008;27(9):943–56. CT. Am Heart J. 1991;121:864–71.
57. Acharya D, Singh S, Tallaj JA, et al. Use of gated cardiac computed 76. Wen Z, Zhang Z, Yu W, et al. Assessing the left atrial phase volume
tomography angiography in the assessment of left ventricular assist and function with dual-source CT: comparison with 3T MRI. Int J
device dysfunction. ASAIO J. 2011;57:32–7. Cardiovasc Imaging. 2010;26:88–92.
 Three-Chamber Function with
Cardiac CT 36
Jongmin Lee

Basic Concepts for deciding whether a patient is a suitable candidate for car-
diac transplant surgery.
Congestive heart failure can be defined as the inability of the Various methods to measure the cardiac output have been
heart to pump blood forward at a sufficient rate to meet the conceived and applied successfully. Direct measurement
metabolic demands of the body or the ability to do so only if techniques for the cardiac output are accurate but invasive.
the cardiac filling pressures are abnormally high. Congestive Based on the Fick principle, which calculates the flow vol-
heart failure is regarded as the final and most severe manifes- ume by the tissue consumption of a substance such as oxy-
tation of common forms of cardiac disease. To evaluate the gen, the cardiac output can be acquired as a function of
progress or risk of heart failure, a comprehensive assessment oxygen consumption, arterial and venous oxygen saturation,
of cardiac function is required. and blood hemoglobin concentration (Eq. 36.1).
The ability of the heart to pump blood is called the cardiac
X tc
function. Traditionally, the mechanical function of heart has Q= (36.1)
been described by the pressure, volume, and flow changes

{ }
[ X ]a − [ X ]v  × [ Hb ] × 1.36 × 10
during a cardiac cycle [1]. The volumes, mass, and function
of cardiac chambers provide important diagnostic and prog- Q, cardiac output; Xtc, tissue consumption of oxygen; [x]a,
nostic information for patients with ischemic or nonischemic arterial concentration of oxygen; [x]v, venous concentration
cardiomyopathies [2]. of oxygen; [Hb], concentration of hemoglobin [3]
Indicator dilution methods are other accurate options for
measuring the cardiac output, but they are also invasive
Cardiac Output methods. Right heart catheterization-based thermodilution
technique has been used as a gold standard method to deter-
A representative and comprehensive parameter of whole-­ mine cardiac output [4].
heart function is cardiac output. Cardiac output is the sum of Indirect methods for the cardiac output measurement have
stroke volumes during 1  min, which can be expressed as been developed for a less invasive and simpler application. If
“stroke volume x heart rate.” Cardiac output is a parameter the relationship between the pulse pressure and the stroke
representing a ventricular pumping function normalized by volume is assumed to be linear, the fractional change of the
the heart rate. Since theoretically the heart rate influences the cardiac output can be acquired by the “fractional change of
ventricular ejection fraction and the stroke volume as a pulse pressure x fractional change of heart rate.” For example,
reverse correlation with the diastolic filling time, heart rate if both the pulse pressure and the heart rate increase up to
should be normalized for the objective comparison of a car- 10%, the cardiac output increases 21% by the calculation as
diac output function. The cardiac output is a relevant diag- “1.1 * 1.1 = 1.21” [1]. A more simplified indirect solution is
nostic and prognostic factor in various cardiovascular suggested by van Lieshout et  al. [5]. Cardiac output is
diseases and an important surrogate marker for a global car- acquired based on an empirical assumption that each 1 mmHg
diac function. In addition, cardiac output is a crucial factor pulse pressure is responsible for approximately 2 ml of stroke
volume (CO = HR * PP * 2 ml). An additional empirical fact
is that body surface area has a greater influence on cardiac
J. Lee (*)
output than body weight or body mass index. Therefore,
Department of Radiology, Kyungpook National University,
School of Medicine, Daegu, South Korea patient-tailored cardiac output index based on body surface
e-mail: [email protected] area is suggested as a useful marker [1].

© Humana Press 2019 423

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_36
424 J. Lee

Cardiac Function with different heart rates, the resistance against the blood flow
changes continuously and periodically. To explain the flow
The influencing factors of the cardiac pumping function can resistance in fluctuating arterial flow, a complex-valued gener-
be classified as preload, afterload, and myocardial inotropic alization of resistance is used, which is the “impedance.”
functions. Abnormal ventricular function can be classified as The ventricular afterload is influenced by the intraluminal
systolic and diastolic dysfunctions. The systolic dysfunction pressure of great arteries, ventricular wall stress and tension,
means an impaired myocardial contractility due to either a and end-diastolic volume and strain. An excessive ventricular
myocardial weakening or an increased afterload. Diastolic dilatation state increases myocardial stress by Laplace law (i.e.,
dysfunction typically refers to an impaired chamber filling tension = radius × pressure) and requires higher inotropic reflex
with an increasing afterload of the chamber upstream. by Frank-Starling law, which result in the excessive afterload
of the ventricle. If the ventricular luminal volume increases, the
Preload perpendicular pressure to the myocardium increases due to the
The preload of cardiac ventricles can be evaluated using some increased mass of contained blood. This pressure is a burden
morphometric markers such as end-diastolic volume (ventricu- against the myocardial contraction and hence the afterload.
lar end-diastolic fiber length), stroke volume, ventricular dilata- The aortic pressure is related to the peripheral vascular
tion, intrathoracic blood volume, and atrial ejection parameters. resistance (PVR), the arterial compliance, and the volume of
The atrial and ventricular preload may be influenced by the dis- blood. Based on the Ohm’s law, the PVR can be simply
tribution of blood between the intrathoracic and extrathoracic acquired by the ratio of a pulse pressure (PP) and a stroke
compartments. The blood volume distribution is influenced by volume (i.e., PVR = PP/SV).
the body position, the intrathoracic pressure, the intrapericardial
pressure, the venous tone, and the pumping action of skeletal Inotropic State
muscle. Atrial contraction is a particularly important preload in The ventricular inotropic activity is influenced by the end-­
the case of ventricular diastolic dysfunction. diastolic volume (preload), sympathetic nerve activity, circu-
How does the preload influence to a downstream func- lating catecholamines, exogenously administered inotropic
tion? The atrial stroke volume may control the downstream agents, physiologic depressants, pharmacologic depressants,
ventricular ejection function. The Frank-Starling law defined loss of ventricular substance, and intrinsic myocardial
that the stroke volume of the cardiac chamber increases in depression. The sympathetic nerve activity is increased by
response to the increase in the volume of blood filling the norepinephrine released at sympathetic nerve endings and
chamber (the end-diastolic volume) when all other factors increases cardiac output. Therefore, the end-diastolic vol-
remain constant. This reflective physiology is explained as a ume may not be changed because of the increased ventricu-
neurohormonal activation mechanism dominating myocar- lar contractility and heart rate. The circulating catecholamines
dium. The Frank-Starling law explains how the preload can may be produced from adrenal gland and extracardiac sym-
influence to the inotropic function (Fig. 36.1). pathetic ganglia. The exogenous inotropic agents are cardiac
glycosides, isoproterenol, sympathomimetic agents, cal-
Afterload cium, caffeine, theophylline, etc. The physiologic depres-
The ventricular afterload can be estimated using markers such sants are myocardial hypoxia, hypercapnia, ischemia, and
as end-systolic volume, stroke volume, ventricular mass, aortic acidosis, and the pharmacologic depressants are guanidine,
distensibility (AD), aortic volume, and systemic vascular resis- procainamide, barbiturates, anesthetics, and so on [6].
tance or impedance. Since the arterial flow pattern is pulsatile The in vivo ventricular contractility may be difficult to mea-
sure directly. Instead, the common markers representing myo-
cardial inotropic function are stroke volume, ejection fraction,
Cardiac Output

and their secondary derivatives. The ventricular ejection frac-

tion is an extremely useful indirect clinical marker representing
Inotropy
the myocardial contractility and is widely accepted for a routine
clinical application. However, since the ejection fraction value
may be confounded by the ventricular afterload, this cannot be
an absolute myocardial contractility marker. In addition, a sim-
ple ratio between end-­ diastolic and end-systolic volumes
doesn’t reflect the size variability of ventricles, which influences
to the stroke volume. For example, the athlete’s heart can main-
End Diastolic Volume
tain the stroke volume and the left ventricle diastolic function in
Fig. 36.1  The Frank-Starling law of the heart explains a positive provo- spite of the decreased ejection fraction in a resting state.
cation of end-diastolic volume (x-axis) to cardiac output (y-axis). By the As an indirect but invasive method, the maximum pressure
ventricular inotropic function, the curve shifts to upward or downward gradient rate (dP/dtmax) can be measured by the intraventricular
36  Three-Chamber Function with Cardiac CT 425

catheterization. This value is usually acquired during isovolumet- reproducibility compared with MRI [12]. In addition, a
ric contraction. The normal range of the left ventricular dP/dtmax phantom study validated dual-source CT as the most accu-
has been suggested to be 1500–2000 mmHg/s [1]. Another poten- rate modality for the left ventricular function evaluation,
tial indirect and noninvasive marker for the ventricular inotro- which was better than 64-slice MDCT and MRI [13]. A
pic function is the strain. The myocardial strain imaging is large-scale prospective multinational observational cohort
available using echocardiography and cardiac MRI by tracking study using CONFIRM registry reported that an additional
speckles and tagging matrix. However, it is limited to register measurement of left ventricular function on CT coronary
speckles or matrix on CT images directly. Currently, based on angiography could improve the risk stratification and identi-
the voxel tracking of a motion coherence algorithm, CT strain fication of patients at risk for incident mortality. In this study,
imaging was used to evaluate three-dimensional myocardial the LVEF, LVEDV, and LVESV demonstrated significant
deformation during a cardiac cycle, and favorable results were diversions of mortality curves [14]. Therefore, the cardiac
reported to show myocardial inotropic function [7]. CT function study, at least for left ventricle, should be in a
consensus of clinical feasibility.
Due to the technical advances in MDCT, high-quality
Cardiac CT cine images can be acquired, and three-dimensional images
of all cardiac chambers can be generated. In addition, many
There has been uncertainty regarding the comprehensive post-processing software applications supply an option to
evaluation of the left ventricular function using MDCT [8, automatically evaluate three-chamber function. Since the
9]. However, from the beginning of MDCT, MDCT-derived automated segmentation of the right atrium has a limitation
cardiac function parameters have shown a significant corre- until now, the other three chambers are now candidates for
lation with reference values demonstrating its potency for automated segmentation (Fig.  36.2). The automated three-­
routine clinical applications [10, 11]. A geometric chamber chamber function evaluation using MDCT has been exten-
function evaluation by CT shows a significant accuracy and sively validated with cardiac MR data [2, 15].

LV RV LA
Parameter Value (Index) Value (Index) Value (Index)
Ejection Fraction% 69 60 33
End Diastolic Volume(ml) 142 87 150 92 97 59
End Systolic Volume(ml) 45 27 61 37 65 39
Stroke Volume (ml) 97 59 90 55 32 20
Cardiac Output(L/min) 6.0 3.7 5.6 3.4

LV LA
Parameter Value (Index) Parameter Value (Index)
Myocardial Mass (g) 107 Maximal Volume (ml) 123 75
Myocardial Volume (ml) 102 Cyclic Volume Change (ml) 58 36
LV/RV Regurgitation Fraction(%) 7.9 Reservoir Volume(ml) 27

Heart Rate: 62 bpm

BSA: 1.6 m2

Vol(ml) Trigger Delay (msec)

150
ES ED

125

100

75

50

25

0
0 10 20 30 40 50 60 70 80 90
Phase %

Fig. 36.2  The reports of the automated three-chamber function evalu- In this software, the left atrial ejection fraction calculates the ejected
ation. The left atrial cyclic volume change (58  ml) is lower than left volume fraction of active pumping
ventricular stroke volume (97 ml) due to the existence of conduit flow.
426 J. Lee

Although cardiac MR is the reference standard in deter- on. The indices are normalized values by the body surface
mining left ventricular function, based on the clinical feasi- area in a unit of square meters.
bility and high spatial resolution cardiac CT, CT-based
function evaluation is becoming more feasible. In this chap-
ter, the chamber function evaluation using MDCT will be Markers for Preload
discussed for each chamber. In addition, the comprehensive
multichamber function evaluation will be discussed in view- To evaluate the preload of the left ventricle, the LVEDV can
points of hemodynamic ergonomics and morphometric be used. In a retrospectively ECG-gated cardiac CT, all dia-
markers. stolic phases are included. When ten phase images are recon-
structed during one cardiac cycle, the image with phase
number 9 or 10 would be the end-diastolic image. Since
Left Ventricular Function images with phase number 7–10 show better image quality
than systolic phases due to less motion, the LVEDV mea-
The left ventricle is the most important chamber for the com- surement is more accurate and reproducible in automated
prehensive cardiac pumping function, as the left ventricle is post-processing software. The LVEDV may be influenced by
the final thrust for systemic arterial blood circulation against the left ventricular myocardial compliance or strain, the size
relatively high flow impedance. During one cardiac func- of the subject or total blood volume, the left atrial stroke vol-
tional cycle, the left ventricular volume changes in a multi- ume, and the intrathoracic blood volume.
modal pattern. During systole, left ventricular functional Beta-adrenergic inhibitor such as propranolol has a nega-
phases comprise of isovolumic contraction and ejection. tive inotropic action and is used for reducing heart rate dur-
During diastole, isovolumic relaxation, rapid inflow, diasta- ing cardiac CT.  However, there is a concern about its
sis, and atrial systole phases are included. If CT images are influence on other functional cardiac parameters. Since beta-­
reconstructed using ten phases for one cardiac cycle, the blockers decrease heart rate, the left atrial volume may be
end-­systolic phase and the diastasis phase match fourth and increased as the preload for left ventricle. Subsequently, the
seventh to eighth phases, respectively (Fig. 36.3). left ventricular end-diastolic volume and the stroke volume
Morphometric markers representing left ventricular pre- will be increased by the Frank-Starling law. Several studies
load are the left ventricular end-diastolic volume index reported that a beta-blocker has no effect on the systolic and
(LVEDVI) and the left atrial systolic volume index (LASVI). diastolic blood pressure; however it increases the end-­
The markers for the left ventricular contractility includes left systolic volume and decreases the ejection fraction. However,
ventricular ejection fraction (LVEF), left ventricular stroke minor controversy exists regarding the effect of beta-­blockers
volume index (LVSVI), left ventricular cardiac output index on the end-diastolic volume [16–19]. Recently, more papers
(LVCI), myocardial strain, and adjusted left ventricular emp- suggested that the beta-blocker effect on end-diastolic vol-
tying fraction (adjLVeFr). The markers for the afterload of left ume is insignificant [20]. This means that the beta-blocker
ventricle are LVESVI, left ventricular muscle mass index may not significantly influence the preload of the left
(LVMMI), PVR, aortic AD, and aortic blood volume and so ventricle.

Fig. 36.3  Left ventricular Isovolumic Isovolumic

time-volume curve during the contraction relaxation
cardiac cycle shows Ejection Rapid Diastasis Atrial
multimodal pattern. The usual inflow systole
cardiac phases for cardiac CT
are diastasis and end-systolic
phase, which match seventh Left ventricular
to eighth and fourth series volume
among reconstructed
ten-phase cardiac CT image
sets during a cardiac cycle
ECG

Systole Diastole

1 2 3 4 5 6 7 8 9 10 Cardiac CT phases
0 10 20 30 40 50 60 70 80 90 100%
36  Three-Chamber Function with Cardiac CT 427

Markers for Inotropy Q∗ K ∗ f corr ∗ Q

CO = ∞
= (36.2)
c ( t ) dt A
The ejection fraction, defined as the fraction of blood ejected ∫ 0
from a ventricle with each heartbeat, is the most common
parameter representing the left ventricular contractility and Q the amount of indicator injected; c(t) indicator con-
systolic function. The global ejection fraction of the left ven- centration as a function of time; K the conversion factor
tricle can be determined reliably by measuring the systolic between HU and Iodine concentration; A the area under the
and diastolic luminal volumes on the cardiac CT. In the cases time-HU curve; Q*(=fcorr*Q) the deconvolution correction
of localized cardiomyopathy such as ischemic heart disease, factor
due to compensatory hyperkinesia of normal myocardium, To avoid a complicated calculation for the cardiac out-
the global ejection fraction may be normal in spite of the put, a simple and indirect marker was recommended. A
regional wall motion abnormality [21]. simple measurement of right-to-left cardiac transit time of
Zeb et  al. [22] evaluated the significance of the CT contrast media during single-level bolus-timing scans pre-
regional ejection fraction to depict the regional wall motion sented a significant correlation with cardiac output index by
abnormality. By comparing CT results with SPECT myocar- MRI (R2 = 0.70). A cut-off value of 20.5 s predicted cardiac
dial perfusion and invasive coronary angiography, they sug- dysfunction with 100% specificity and positive predictive
gested that the CT regional ejection fraction was comparable value [25].
with SPECT perfusion imaging for depicting ischemic seg- Strain is a concept of the continuum mechanics meaning
ments. In addition, they suggested that the CT regional ejec- “a deformation of a body without breakage from a reference
tion fraction showed a better performance than SPECT to configuration to a current configuration” [26]. In the case of
determine significant coronary artery stenosis. Due to higher the heart, a myocardial movement is driven by bimodal
spatial resolution and image contrast, CT may detect a subtle ergonomics, active systolic contraction, and passive dia-
change of regional wall motion. In addition, the regional wall stolic relaxation. In addition, the myocardial movement is
motion abnormality detected on CT may be a sensitive pre- influenced by preload and its afterload. This complicated
sentation of myocardial ischemia. myocardial activity is expressed as the strain in cardiac
As discussed in the basic concept section, the cardiac out- imaging. In actuality, cardiac strain represents only the
put is an important and common comprehensive marker for range of myocardial movement but cannot directly represent
the cardiac pumping function. In cardiac CT, noninvasive myocardial contractility or expansibility. However, the
measurement of the cardiac output has been attempted by strain rate, which is the time derivative of the strain, may
adopting several complex mathematical and empirical for- reflect active and passive motions of the myocardium.
mulae. Yee et  al. [23] measured a left ventricular output Currently, two-dimensional or three-dimensional myocar-
using the bolus-timing scan for CT pulmonary angiography dial strain is applied to evaluate the myocardial inotropic
in a routine clinical setting. Using input variables including function.
ascending and descending aortic radii, the geometric dis- Buss et al. [27] compared the endocardial strain and the
tance and the contrast media concentration between the strain rate between two-dimensional (2D) cardiac CT and
ascending and descending aortas, a mathematically approxi- echocardiography. Using a commercial tracking software,
mated left ventricular cardiac output was acquired. Although multidirectional endocardial strain was evaluated throughout
a clear validation of the method was not conducted, the fea- one cardiac cycle. The results showed significant correlation
sibility on clinical practice is acceptable due to its simplicity with echocardiography and NT-pro-BNP level. Since the 2D
without any additional risks to the patient. cardiac CT strain imaging revealed significantly shorter pro-
From the preliminary bolus-timing scan data, a time-­ cessing time, high reproducibility, and lack of observation
attenuation curve can be produced. Mahnken et  al. [24] window limitation, non-inferiority to 2D echocardiography
applied the Stewart-Hamilton equation to calculate a car- was disclosed.
diac output using the area under the time-attenuation curve In a recent study by Tanabe et al. [7] using cardiac CT and
and the injected amount of contrast media as input vari- dedicated reconstruction software, 10-phase image sets are
ables. In addition, a conversion factor between the attenua- interpolated to 50 time frame image sets, and a three-­
tion and real iodine concentration was included as an input dimensional left ventricular myocardial model is generated
constant (Eq. 36.2). To simulate the first pass flow without by image segmentation. From the mesh of the ventricular
a recirculated contrast enhancement, a gamma variate func- model, motion vectors and subsequent strain values can be
tion was additionally applied. As results, a significant cor- acquired along longitudinal, radial, and circular directions.
relation of the cardiac output and stroke volume data This strain evaluation presented an additional benefit of CT
between a test-­bolus technique and a geometric measure- coronary angiography to detect an infarcted segment of
ment was achieved. myocardium.
428 J. Lee

Markers for Afterload LV End- LV End-

diastole systole Active
pumping

Volume
Just like typical fluid pumps, heart function may be influ- phase
Passive
enced by the downstream flow impedance, referred to as emptying &
afterload. Left ventricular afterload can be described as the conduit
Reservoir
aortic capacity to receive the left ventricular stroke volume. phase
phase
This aortic capacity may be related with intra-aortic blood Vmax
mass, aortic wall stiffness or elasticity, and aortic PEV
Vpre-A
morphology. aEV
Vmin
The aortic wall elasticity is a key factor for dampening the
pulsatile energy of the entry flow and converting it to stabi-
lized steady flow for the end organs. The aortic wall stiffness
0 10 20 30 40 50 60 70 80 90 100%
is an opposite concept to the dampening function and
Phase
increases flow impedance throughout the aorta. Left ventric-
ular output function is influenced by the aortic flow imped- Fig. 36.4  Time-volume curve of left atrium shows bimodal pattern
ance, which is the afterload of left ventricular performance. along functional phases. The reservoir phase is left atrial diastolic phase
Ganten et al. [28] suggested the aortic AD as a marker for during left ventricular systole. The passive emptying and conduit and
the active pumping phases generate weak bimodal curve during left
aortic wall stiffness (Eq.  36.3). Although the research was
atrial systole. The timing of the maximum volume can be beyond the
applied on abdominal aorta, the concept of the aortic AD left ventricular end-systole due to the isovolumic relaxation and super-
may be applied on AAo and aortic arch for the left ventricu- imposed conduit flow from pulmonary vein. Vpre-A is the volume imme-
lar afterload evaluation. diately before the late atrial contraction. pEV is the passive emptying
volume same as reservoir volume. aEV is the active emptying volume
∆A same as pumping volume
D= (36.3)
A0 ⋅ ∆p

left ventricle through functional phases of reservoir, passive
D, aortic distensibility; A, aortic cross-sectional area; A0, emptying and conduit, and active pumping (Fig. 36.4) [30].
minimum aortic cross sectional area during the cardiac cycle; The relative contribution to the left ventricular diastolic fill-
p, blood pressure ing by left atrial phases comprises of 40%, 35%, and 25% of
Lee et al. [29] suggested the AWSI, which shows aortic passive emptying (reservoir) volume, conduit volume, and
wall stiffness consistently without any influence by different active pumping volume, respectively [31]. At an early stage
intraluminal pressures and aortic sizes. In ex vivo and in vivo of left ventricular diastolic dysfunction, the relative contri-
experiments, AWSI showed a significantly higher accuracy bution of active emptying increases with a reciprocally
than other stiffness factors including a classic compliance, decreasing passive emptying. This change is noted as a
Young’s modulus of elasticity, beta-stiffness index, and aor- decreased E/A ratio during the trans-mitral Doppler
tic AD. ultrasonography.
The reservoir phase of left atrium is during the left ven-
  tricular systolic phase. This phase may be influenced by the
 −2b 
 ⋅ mBP  left atrial wall compliance, the descending motion of the left
AWSI = ln  a 2 1
⋅  (36.4) ventricular base during systole, and the LVESVI [32].
2 2 V
   mBP   d  The passive emptying and conduit flow phase is shown
 1 +    
   a    during an early period of the left ventricular diastolic phase.
In this phase, the left ventricular diastolic negative pressure
AWSI, aortic wall stiffness index; mBP, mean blood pres- allows left atrial blood to flow in. The passive decrease of the
sure; Vd, diastolic volume; a, b, model-specific constants left atrial volume by a passive emptying can be described as
the “reservoir volume.” A direct inflow from pulmonary vein
to left ventricle occurs through left atrium, which serves as a
Left Atrial Function conduit role. The flow and its volume without left atrial vol-
ume change are called as a “conduit flow” and a “conduit
Left Atrial Physiology volume,” respectively [33].
During the conduit flow, the flow volume should not be
The left atrium comprises of three morphologically differ- related to the left atrial volume decrease. The sum of the pas-
ent compartments: anterior, venous, and appendage com- sive emptying (reservoir) volume and the conduit volume is
partments. The left atrium modulates diastolic filling of the higher than the left atrial geometric volume decrease,
36  Three-Chamber Function with Cardiac CT 429

because of the direct conduit flow through widely opened Table 36.1  Nominations of fractional blood flow volumes from left
space from pulmonary vein to negatively loaded left ventri- atrium to left ventricle
cle [34]. The conduit flow is influenced by a paradoxical Volumes Synonyms Definition
relationship between the left ventricular and left atrial relax- Reservoir Passive emptying Outflow in left atrial passive
volume volume emptying phase
ation capacities. The high left ventricular compliance will
Conduit Direct flow from pulmonary
increase the conduit flow, whereas the high left atrial compli- volume vein to left ventricle
ance will increase the reservoir volume with the relatively Pumping Active emptying Outflow in left atrial active
decreased conduit volume. volume volume, ejection pumping phase
During the passive emptying of left atrium, both the con- volume
duit volume and the reservoir volume fill the relaxing left Stroke Cyclic volume Reservoir volume + pumping
volume change volume
ventricle. The left atrial output volume should be the sum of Output Reservoir volume + conduit
the conduit volume, the reservoir volume, and the active volume volume + pumping volume
ejection volume. However, the geometric volume measure-
ment cannot present the total amount of the left atrial output
volume due to the isovolumetric conduit phase. In this con- Morphometric Markers for Left Atrial Function
text, during the morphometric evaluation of a left atrial
stroke volume, a simple subtraction of the end-diastolic vol- The morphometric marker representing the preload of left
ume by the end-systolic volume underestimates the left atrial atrium may be pulmonary vascular volume, LAVImax, and left
output volume. The existence of the conduit flow is the rea- atrial expansion index. The markers for the left atrial systolic
son why CT-based geometric left atrial stroke volume is ability will be the left atrial ejection fraction (LAEF), LAaEF,
shown to be lower than the left ventricular stroke volume and adjLAeFr. The markers showing the afterload of left
(Fig.  36.5). The sane meaning of the morphometric ”left atrium are the minimum LAVI (LAVImin), the left atrial pas-
atrial stroke volume“ should be the sum of reservoir and sive ejection fraction (LApEF), LVESVI, LVMMI, LVSVI,
pumping volumes, which equals to the cyclic volume change and LVCI.
(Table 36.1).
The active pumping function of left atrium is during the Reservoir Function
late atrial contraction, which contributes up to 30% of the To evaluate the left atrial reservoir function, the measured
left ventricular stroke volume [35]. The left atrial active ejec- left atrial volume can be utilized. The left atrial volume
tion may be governed by the Frank-Starling mechanism increase has been known as a strong predictor for adverse
same as ventricles [36]. Therefore, the left atrial active emp- cardiovascular outcomes [37]. The left atrial volume change
tying is influenced by the systemic venous return (preload) can be evaluated only with retrospective ECG gating since
as well as the left ventricular end-diastolic pressure the pre-defined cardiac phase by prospective gaiting may
(afterload). miss the true maximum and minimum sizes of left atrium.
Recently, Agner et al. [38] successfully validated CT-based
left atrial volumetry with references by cardiac MR.  They
End End found that the cardiac CT overestimated the maximum and
diastolic systolic
minimum left atrial volumes compared to cardiac MR (8%
Volume

and 10%, respectively). This inter-modality discrepancy may

Right ventricle
be due to the difference between the spatial and temporal
Left ventricle resolutions and image sharpness.
Since the venous compartment of left atrium possesses
relatively less myocardial tissue and more venous wall
attached to pericardium, the motion and contraction of the
Left atrium
venous wall are weaker than the other compartments.
Whereas, the venous compartment conducts a role of reser-
voir for the pulmonary venous return flow. The appendage
0 10 20 30 40 50 60 70 80 90 100%
compartment comprises of a thicker wall and pectinate mus-
Phase
cles interconnecting the opposing walls. Due to the active
Fig. 36.5  The time-volume curves of three chambers. Along the car- contractile function and thick wall, the appendage is more
diac phases, ventricles and left atrium reveal reciprocal change of their resistant to overload than the other compartments; therefore,
volumes. The right ventricular volume is usually higher than that of left
ventricle. The left atrial curve height is shorter than that of left ventricle
it is relatively insensitive to overloading. Paradoxically, if the
reflecting lower morphometric left atrial stroke volume than left ven- appendage compartment shows dilatation with wall thinning,
tricular stroke volume due to the conduit volume this feature may imply a long-standing volume overload of
430 J. Lee

the left atrium, typically due to the left ventricular diastolic To evaluate the passive emptying function of left atrium,
dysfunction with or without atrial fibrillation. the passive ejection fraction (pEF) can be used. The pEF is
As another marker for left atrial volumetry, the LAVImax the passive emptying volume (pEV) per maximum left atrial
has been suggested. This index is a normalized maximum volume [(LAVmax − LAVpre-A)/LAVmax], where the LAVpre-A is
volume by the body surface area. According to the results of the left atrial volume immediately before atrial contraction.
echocardiography, the LAVImax is the best parameter for car- For the active pumping function of left atrium, the active
diovascular outcome prediction and risk stratification [37]. ejection fraction can be acquired as the active emptying vol-
Choi et al. [39] reported that LAVImax showed a significant ume (LAVpre-A–LAVmin) over LAVpre-A (Fig. 36.4) [30].
predictability for the silent myocardial ischemia in a high-­ Im et al. [42] suggested the adjLAeFr as a marker for emp-
risk group. The LAVImax from echocardiography presented a tying function evaluation. The maximum and minimum left
higher adjusted odds ratio than LVEF, E/E’ ratio, LVMI, and atrial volume indices were normalized by LAEF (LAVImax/
the diastolic left ventricular dysfunction. LAEF and LAVImin/LAEF). The adjLAEFr showed the best
The left atrial expansion index is another dynamic marker predictability for the recurrent atrial fibrillation as compar-
for the left atrial volumetry. The left atrial expansion capac- ing with conventional markers including LVEF, LAVmax,
ity may reflect the reservoir function during left atrial dias- LAVmin, LAEF, LAVImax, and LAVImin.
tole. To quantitate the expansion capacity, the left atrial
expansion index was suggested by Hoit et  al. [30]. The Afterload
expansion index was calculated as the left atrial volume The left atrial dilatation is a classical marker for left ven-
change per minimum left atrial volume [(Vmax – Vmin)/Vmin]. tricular diastolic dysfunction, which is an increased afterload
of left atrium. The left atrial volume increase can be either
Emptying Function reversible by the instant volume overload or irreversible due
The three compartments of left atrium present different func- to an eccentric wall remodeling. The aging, neurohumoral
tional activities. The appendage shows the highest ejection activation, and chronic atrial stretch may provoke the left
fraction, whereas the venous compartment shows the lowest atrial dilatation with histological changes including wall
ejection fraction [40]. The LAA ejection fraction is not hypertrophy and fibrosis as eccentric left atrial remodeling
strongly related to the global and anterior compartmental [43]. In this background, eccentric remodeling may trigger
function, whereas LAA function is reported to be more episodes of atrial fibrillation.
related to the left ventricular functional status [41]. The ante- The LAVImin is related to the afterload of left atrium. The
rior compartment of left atrium has a smooth and movable LAVImin may be influenced by the left atrial elasticity and
wall and is not adhered by pericardium or venous structures. LVEDP and was reported as the best predictor for the major
The anterior compartment presents a less compromised wall adverse cardiac events (MACE) in a study using 3D echocar-
motion by surrounding structures and may properly reflect diography [44].
the left atrial performance. In the short-axis plane, the ante- With Doppler ultrasonography of the trans-mitral flow,
rior compartment demonstrates a left atrial wall motion the flow spectrum shows bimodal peaks due to the left atrial
conveniently. passive emptying flow (E velocity) and the active pumping
The LAEF may be calculated by a simple formula: 100 * flow (A velocity), whereas with cardiac CT with multiphase
(LAVmax − LAVmin)/LAVmax, where LAV is the left atrial vol- reconstruction during a cardiac cycle, the left ventricular
ume. With some software, the LAEF shows the volume frac- time-volume curve and its time-differential curve can be
tion by active pumping during systole (Fig. 36.2). The left acquired. The left ventricular volume gap between two adja-
atrial cyclic volume change implies the gap between the cent phases shows the bimodal peaks in the time domain. If
maximum and the minimum volumes. However, the maxi- the volume gap is divided by the mitral valvular orifice area
mum and the minimum volumes may not always match the at the concordant phase, velocity values can be acquired. In
left ventricular end-diastolic (100% of R-R interval) and the the bimodal time-velocity curve, E and A velocities and
end-systolic (40% of R-R interval) phases. During the left E/ACT can be measured on cardiac CT [45]. The E/ACT repre-
ventricular early passive filling phase, left atrium can be sents the proportion of passive and active emptying volume
more instantly dilated than the left atrial end-diastolic phase during the left atrial systolic function. The E/ACT is related
due to a superimposed conduit flow directly from pulmonary with the left atrial afterload or the diastolic function of the
vein to the left ventricle (Fig. 36.4). Due to a concern about left ventricle.
mismatching between cardiac phases and left atrial volumes,
the gap between the maximum and the minimum volumes Global Function
may be suitable for calculating the ejection fraction, in spite The left atrial time-volume curve provides several potential
of a different concept from the ejection fraction by the left markers to evaluate the atrial systolic and diastolic functions.
atrial active contraction. However, since the volume is a three-dimensional scalar
36  Three-Chamber Function with Cardiac CT 431

value, two-dimensional wall strain evaluation may be more Right Ventricular Contractility
sensitive to the geometric change of the left atrial wall. The
total left atrial strain measured by echocardiography has The ejection fraction is a reliable marker for right ventricular
been reported to be the best predictor of exercise capacity systolic function. In patients with acute pulmonary thrombo-
[46]. In light of the fact that the strain imaging of the left embolism combined by pulmonary arterial hypertension,
ventricular wall has been attempted using cardiac CT data, RVEF and right ventricular end-systolic volume have been
study results of CT-derived left atrial strain are expected in discovered to have significant implications [49]. With car-
the literature soon. diac CT right ventricular volumetry requires more resources
but may be less reliable than left ventricular volumetry due
to prominent trabeculation, atypical morphology, indistinct
Right Ventricular Function basal boundary, and motion artifacts.
Sato et al. [50] suggested a simplified method to predict
Right Ventricular Physiology RVEF in patients with pulmonary hypertension. Using a ret-
rospectively gated cardiac CT, the tricuspid annular plane
The right ventricular function is different from the left ven- systolic excursion (TAPSE) was measured during a cardiac
tricular function only by the magnitude of the peak systolic cycle. Using a fitting equation between the TAPSE and ejec-
pressure due to the lower pulmonary vascular resistance tion fraction values acquired by cardiac MRI, a simplified
(PLVR). The right ventricle comprises of thin myocardial prediction equation for RVEF was suggested and sufficiently
wall that has a highly compliant nature. Due to this compli- validated. The RVEF in the ml-equivalent unit was acquired
ance, the right ventricle accommodates the changes of pre- as double value of TAPSE in millimeter. In pulmonary
load well. However the right ventricle is limited in overcoming hypertension patients, 19.7 mm cut-off value of TAPSE was
the increased afterload due to its limited contractility. In this suggested to predict a reduced RVEF.
context, a common cause of the right ventricular failure is left The right ventricular cardiac output or the cardiac index
ventricular failure. If right ventricular failure occurs due to may be direct markers for the systolic function as left ven-
the primary pulmonary process, it is called as the cor pulmo- tricle. In a study using a test-bolus scan data, the time-­
nale [1]. attenuation curves were drawn on main pulmonary arteries,
Right ventricular failure has effects on right atrium and and the right ventricular cardiac output was acquired using
systemic venous return. The clinical presentation of the the modified Stewart-Hamilton equation (Eq.  36.2). Based
right ventricular failure contains systemic venous conges- on the reference standard data acquired during right heart
tion noted as jugular vein dilatation, hepatomegaly, periph- catheterizations and thermodilution, the right ventricular
eral edema, IVC dilatation, and right atrial dilatation. The cardiac output acquired from dynamic CT showed a signifi-
right ventricular dysfunction superimposed on left ventric- cant correlation and clinical feasibility [51].
ular pathology is regarded as additional poor prognostic
factor. Di Bella et al. [47] showed the combined right ven-
tricular dysfunction after acute myocardial infarction sig- Afterload of the Right Ventricle
nificantly deteriorated patient prognosis. Right ventricular
function may be additional information on left ventricular The pulmonary arterial pressure is an essential afterload of
function during cardiac pathology evaluation using the car- the right ventricle, and a pulmonary hypertension will
diac CT. increase the afterload of right ventricle. The pulmonary arte-
In cases of acute pulmonary arterial hypertension such as rial hypertension escalates the right ventricular end-diastolic
acute massive pulmonary thromboembolism, a quick, sim- pressure, which relates to the dilatation of the right ventricle
ple, and noninvasive method for estimating right ventricular with a limited contractility.
afterload may be indispensable to urgently inform therapeu- The dilated right ventricle may shift interventricular sep-
tic decision-making [48]. The morphometric markers repre- tum to present an eccentric configuration of left ventricle in
senting the preload of right ventricle are RVEDVI and right the short-axis plane. Yamasaki et al. [52] presented a simple
atrial stroke volume index (RASVI). The markers for the predictor for pulmonary arterial hypertension among the
right ventricular contractility evaluation may be the right patients with congenital heart diseases. The “eccentricity
ventricular ejection fraction (RVEF), right ventricular stroke index” is acquired as a ratio between the orthogonal long and
volume index (RVSVI), and right ventricular output index short diameters of left ventricular lumen in the short-axis
(RVCI). The markers showing the afterload of right ventricle plane. The eccentricity index values show a significant cor-
may comprise of RVESVI, right ventricular MMI (RVMMI), relation with the right ventricular systolic pressure and the
PLVR, pulmonary circulation volume, and pulmonary AD mean pulmonary arterial pressure from right heart cather-
(AD). ization results. Also the eccentricity index correlates with
432 J. Lee

cardiac MR functional parameters representing the right [53]. Müller et  al. [53] measured the PLVR indirectly and
ventricular afterload such as RVESV and RVEF. noninvasively using the electron-beam CT data and Hagen-­
If the PLVR is measured, the evaluation of the right ven- Poiseuille equation (Eq. 36.5). This study revealed a signifi-
tricular afterload should be more convenient. The pulmonary cant correlation between the noninvasively measured PLVR
arterial flow impedance increases in patients with pulmonary and catherization-based reference values (r2 = 073, regres-
thromboembolism, chronic obstructive pulmonary disease, sion coefficient = 1.001 after removing intercept).
and emphysema resulting in pulmonary arterial hypertension

803 ⋅ LB[m]3
PLVR [ Wood] = ⋅ 5 + 10 ⋅ ( HCTL / L − 0.52 )  (36.5)
( )
2
CO[2L / min ] ⋅ PTTM [ s ] + 6.14

LB, patient’s heights used for vascular length; HCT, hema- outflow tract dilatation and wall thickening, and flattening or
tocrit used for blood viscosity; PTT, pulmonary transit time; leftward convexity of interventricular septum are classic
empirically acquired constants image markers representing the elevated right ventricular
Since this study is based on Hagen-Poiseuille law, inher- end-diastolic pressure and pulmonary arterial hypertension.
ent limitation by assumption such as a straight cylindrical The right ventricular free wall thickness greater than 6 mm,
tube, homogenous fluid (Newtonian fluid), laminar flow, RV/LV wall thickness ratio greater than 0.32, RV/LV luminal
steady non-pulsatile flow, and rigid wall without arterial diameter ratio greater than 1.28, and MPA/AAo diameter
compliance may not be overcome. However, the results dem- ratio greater than 0.84 have been implicated with pulmonary
onstrate a feasibility of this technique in clinical practice. arterial hypertension [56]. Lastly, the finding of bronchial
Stiffening of pulmonary arterial wall attenuates its damp- artery hypertrophy greater than 1.5  mm was reported as a
ing function against the pulsatile right ventricular outflow marker for pulmonary arterial hypertension [57].
and subsequently increases the flow impedance as afterload.
Using cardiac MDCT, Revel et al. [54] measured right ven-
tricular and pulmonary arterial functional parameters to Right Atrial Function
diagnose pulmonary hypertension. Image-based right pul-
monary AD (area change/maximum area) has shown the best Theoretically, the morphometric markers representing the
diagnostic value for pulmonary hypertension and has been preload of right atrium are RAVImax, vena cava size, intratho-
useful for risk stratification. racic pressure state, and intrapericardial pressure state. The
Morphological alteration of pulmonary vasculature and markers for the right atrial contractility may be RAEF and
the lungs on CT images may suggest increased right ven- RAaEF. The afterload of right atrium may be evaluated by
tricular afterload. Pulmonary thromboembolism and exten- RAVImin, RA passive EF, RVESVI, RVMMI, RVSVI, and
sive pulmonary parenchymal lesions including interstitial RVCI.  In spite of these theoretically conceived markers,
lung disease and chronic obstructive pulmonary disease may automated measurement of right atrial volume is limited to
provoke pulmonary arterial hypertension. few commercial softwares due to the insufficient and
Pulmonary arterial dilatation is a predictor for pulmonary ­nonhomogeneous enhancement, the irregular morphology,
hypertension. If the main pulmonary arterial diameter is and motion artifacts.
greater than 29 mm at a distance of 2 cm from a pulmonary Systemic venous congestion state is a type of increased
valve, pulmonary arterial hypertension could be predicted preload of the right atrium. Although the exact volumetry is
with 84% sensitivity and 75% specificity. If the main pulmo- not feasible, a morphological change may reflect the increased
nary artery shows the maximum transverse diameter larger preload. A vena caval enlargement used to be a sign for sys-
than the adjacent AAo, pulmonary arterial hypertension temic venous congestion. However, this finding is related to
could be suggested with 70% sensitivity and 92% specificity. many confounders such as the breath-holding capabilities, the
The underline condition is that the AAo should not show the strength of the Valsalva maneuver, and the original size of the
aneurysmal dilatation. In addition, the segmental pulmonary venous structure. The retrograde vena caval or hepatic venous
arterial enlargement greater than 1.25 times the caliber of the enhancement has been regarded as a sign representing the
adjacent bronchus is another marker for pulmonary arterial right heart dysfunction during a conventional slow-infusion
hypertension [55]. contrast-enhanced CT [58]. However, dynamic contrast-
A right ventricular deformation may suggest pulmonary enhanced CT with a high injection rate (>3 mL/s) could not
arterial hypertension as well. Right ventricular enlargement, maintain the feasibility of this systemic venous reflux sign
36  Three-Chamber Function with Cardiac CT 433

[59]. In cardiac CT, the sign of the retrograde opacification of Frank-Starling’s law (Fig. 36.1). In this context, when evalu-
inferior vena cava and hepatic vein may not be reliable for the ating global cardiac function, the function of the interrelated
prediction of the right cardiac dysfunction. cardiac chambers is best when considered collectively.
The TAPSE is known as a simple marker for the right ven- To evaluate the function of the cardiac chambers simulta-
tricular function [50]. Sivak et al. [60] used the TAPSE for neously and comprehensively, the “three-chamber function”
evaluating the right atrial function using cohorts of a normal option can be used with commercially available post-­
and a pulmonary arterial hypertension. In echocardiography, processing software. By plotting the time-volume curves of
the TAPSE is divided to two parts; TAPSERA during the right left ventricle, left atrium, and right ventricle along cardiac
atrial contraction and TAPSERV before the right atrial contrac- phases (time domain), qualitative comparisons of the three-­
tion after tricuspid valve opening representing the right atrial chamber function and the quantitative evaluation of the func-
passive emptying function. In a pulmonary arterial hyperten- tional parameters are possible (Figs. 36.2 and 36.5).
sion group, TAPSERV shows a significant decrease, while the For a quantitative evaluation of multichamber functional
change of TAPSERA is not significant. The late atrial contrac- parameters, reference values are indispensable. However, the
tion may be more preserved than the passive emptying func- normal values of multichamber functional parameters show
tion in the case of right atrial afterload increase. significant differences by the reports in the literature. These
The longitudinal strain of the right atrial wall evaluated normal values show different ranges depending on the mea-
by speckle-tracking echocardiography demonstrated its fea- suring techniques, ethnicity, age, sex, and cohort properties
sibility to predict increased afterload due to the right ven- (Table  36.2); since the references may be different among
tricular diastolic dysfunction [61]. Although no article about cardiac CT centers, large-scale multicenter studies account-
CT strain evaluation for right atrial function has been pub- ing for potential confounders are expected [62–65].
lished, studies are expected in the near future. During the evaluation of the three-chamber function
results, the major left ventricular functional parameters
such as LVEDVI, LVESVI, LVEF, LVSVI, LVCI, and
Global Cardiac Function LVMMI are checked first. Subsequently the major left
atrial functional parameters such as LAVImax, LAVImin,
The four cardiac chambers are working in a tight interrelated LASVI, and LAaEF are evaluated. Finally, the right ven-
mode. Left ventricular diastolic dysfunction may increase tricular functional parameters such as RVEF, Qp/Qs,
the afterload of left atrium. An erroneous left atrial emptying RVEDVI, and RVSV are assessed (Figs. 36.6 and 36.7). In
function may provoke the right ventricular afterload increase. the scope of global cardiac function, the morphometric
An increased left atrial emptying function will stimulate the markers supporting the preload may be the vena cava size;
left ventricular inotropic function, which is described as the however, its reliability is not accepted. The global inotropic

Table 36.2  An example of normal reference ranges for cardiac CT functional parameters in a single institute (Kyungpook National University
Hospital)
Left ventricle Left atrium Right ventricle
Male Female Male Female Male Female
EDV(ml) 100–180 108–122 127–223
ESV (ml) 28–66 28–52 53–111
LVMM(mg) 129–205 100–158
EF(%) 57–75 59–73 42–74
SV(ml) 50–70 40–84
CO(L/min) 4.0–8.0 3.2–9.6
EDVI(ml/m2) 57–88 47–74
ESVI(ml/m2) 16–33 11–24
SVI(ml/m2) 40–50 11–25 13–27 35–60
CI(L/m2) 2.5–4.5 2.0–5.4
MMI(mg/m2) 31–81
LAVImax 32–56 39–63
LAVImin 17–39 19–43
EFI(%/m2) 26–50 28–52
LARsrV(ml) 16–28
LACnV(ml) 18–25
LAPmpV(ml) 13–18
LARsrV left atrial reservoir volume, LACnV left atrial conduit volume, LAPmpV left atrial active pumping volume
434 J. Lee

a b

c d

Fig. 36.6  A 64-year-old man presented atypical chest pain and was m2) and LVESVI (48  mL/m2) are increased. Global LVEF (46%)
evaluated using cardiac CT. (a) Calcified atherosclerotic plaques decreased slightly. Further calculated LAVImax (68  mL/m2) and
(arrows) are noted in proximal and middle levels of left anterior LAVImin(41  mL/m2) are increased. LAEFI (23  mL/m2) is decreased.
descending artery (LAD) and left main coronary artery. (b) In en face RVEF (40%) is slightly decreased. All the other parameters are in nor-
view of LAD, significant luminal stenosis (arrow) is noted. (c, d) Short-­ mal ranges. The conclusion is a significant coronary artery stenosis with
axis cine images at mid-diastolic and end-systolic phases show left ven- LV systolic dysfunction accompanied by LA and RV systolic
tricular hypokinesia. (e) The three-chamber function report reveals dysfunction
upward shifting of ventricular time-volume curves. LVEDVI (89 mL/
36  Three-Chamber Function with Cardiac CT 435

e LV RV LA
Parameter Value (Index) Value (Index) Value (Index)
Ejection Fraction% 46 40 14
End Diastolic Volume(ml) 154 89 151 88 81 47
End Systolic Volume(ml) 82 48 91 53 70 41
Stroke Volume (ml) 71 41 60 35 11 6.4
Cardiac Output(L/min) 5.2 3.0 4.4 2.6

LV LA
Parameter Value (Index) Parameter Value (Index)
Myocardial Mass (g) 107 1.8 Maximal Volume (ml) 115 67
Myocardial Volume (ml) 102 Cyclic Volume Change (ml) 45 26
LV/RV Regurgitation Fraction(%) 15 Reservoir Volume(ml) 0.0

Heart Rate: 73 bpm

BSA: 1.7 m2

Vol(ml) Trigger Delay (msec)

200 ED ES
175
150
125
100
75
50
25
0
0 10 20 30 40 50 60 70 80 90
Phase %

Fig. 36.6 (continued)

function can be evaluated by the markers such as LVEF, et  al. [66] successfully validated Qp/Qs values, acquired
LVSVI, LVCI, and Qp/Qs. The markers for the afterload of using a 256-slice MDCT based on the right heart
whole heart may be LVESVI, PVR, AD, and aortic volume catheterization.
(Table 36.3). The myocardial strain of cardiac chambers may be a
As an integrated marker to show a comprehensive cardiac potential marker for a comprehensive multichamber function
function, the Qp/Qs may be used to compare stroke volumes evaluation. Using the speckle-tracking echocardiography,
between right and left ventricles. In managing congenital four-chamber strain can be evaluated. When the mean and
heart diseases, the Qp/Qs is a reliable marker to diagnose and peak longitudinal strains are plotted in one graph, the rela-
monitor a left-to-right shunt and to decide treatment strategy. tionship among four-chamber wall motions can be evalu-
The CT-based geometric measurement of the right and left ated, and a comprehensive functional assessment of the heart
ventricular stroke volumes can present the Qp/Qs values may become possible. If the strain ratios between ventricles
under the condition of an absent regurgitation in atrioven- and between atria, the abnormal wall motion and its compen-
tricular and ventriculoarterial valves. Recently, Yamasaki satory mechanism may be analyzed feasibly [67].
436 J. Lee

a b

c d

Fig. 36.7  A 47-year-old male complains resting dyspnea and chest dis- are decreased. LVSVI and LVCI are not decreased. LAVImax(105) and
comfort. The patient has suffered from Behçet’s disease, and valvulo- LAVImin(84) are increased. The left atrial active pumping flow is not
plasties were performed 6  years ago. (a, b) Volume-rendered images detected. The left atrial conduit flow volume seems to be increased with
along the left ventricular outflow tract reveal a well-functioning pros- maintaining the fraction of reservoir volume. LAEFI [10] is decreased.
thetic aortic valve (arrows). Focal sinus of Valsalva aneurysm is noted The other functional parameters including right ventricle are within nor-
(arrow heads) (c, d) Multiplanar longitudinal images show a well-­ mal ranges. The prosthetic aortic and mitral valves are mechanically
functioning prosthetic mitral valve (arrows). (e, f) In four-chamber functioning well. However the left ventricle shows eccentric hypertro-
plane, left atrial dilatation and limited emptying are noted. Left ventricle phy and dilatation with systolic dysfunction. Subsequently the left
(LV) shows dilatation. (g) The three-chamber function analysis report atrium dilates with systolic dysfunction. No left atrial active pumping
reveals upward shifting of all ventricular and left atrial time curves (right function is detected suggesting the risk of atrial fibrillation. In conclu-
lower corner) suggesting chamber dilatation. LVEDVI(185), sion, a functional aortic valvular stenosis should be considered if active
LVESVI(86), and LVMMI(95) are increased. LVEF [54] and PVR [5] Behçet’s myocarditis can be excluded clinically
36  Three-Chamber Function with Cardiac CT 437

e f

g LV RV LA
Parameter Value (Index) Value (Index) Value (Index)
Ejection Fraction % 54 46 0.0
End Diastolic Volume (ml) 370 186 257 130 167 84
End Systolic Volume (ml) 172 87 140 70 167 84
Stroke Volume (ml) 198 100 118 59 0.0 0.0
Cardiac Output (L/min) 14.5 7.3 8.6 4.3

LV LA
Parameter Value (Index) Parameter Value (Index)
Myocardial Mass (g) 189 2.4 Maximal Volume (ml) 210 106
Myocardial Volume (ml) 179 Cyclic Volume Change (ml) 43 22
LV/RV Regurgitation Fraction (%) 41 Reservoir Volume (ml) 0.0

Heart Rate: 73 bpm

BSA: 2.0 m2

Vol (ml) Trigger Delay (msec)

400
ED ES
350

300
250

200

150
100

50

0
0 10 20 30 40 50 60 70 80 90
Phase %

Fig. 36.7 (continued)
438 J. Lee

Table 36.3  Functional parameters of cardiac chambers

Left ventricle Left atrium Right ventricle Right atrium
Preload LVEDVI, LASVI LAVImax, LAEI, pulmonary RVEDVI, RASVI RAVImax, VC size,
vessel size, intrathoracic pressure
Inotropy LVEF, LVSVI, LVCI, LAEF, LAaEF, adjLAeFr RVEF, RVSVI, RVCI RAEF, RAaEF
strain, adjLVeFr
Afterload LVESVI, LVMMI, LAVImin, LApEF, LVESVI, RVESVI, RVMMI, PLVR, PTT, PAD, RAVImin, RApEF, RVESVI,
PVR, AD, AWSI, aortic LVMMI, LVSVI, LVCI, E/ LAEI, RV, and pulmonary vascular RVMMI, RVSVI, RVCI,
volume ACT deformation TAPSE
LAEpI left atrial expansibility index, LVCI left ventricular cardiac output index, PTT pulmonary transit time, PAD pulmonary arterial
distensibility

c­ omputed tomography for identification of myocardial infarction.

Summary Eur Radiol. 2017;27(4):1667–75.
8. Mahnken AH, Katoh M, Bruners P, Spuentrup E, Wildberger JE,
Based on the high spatial resolution and acceptable temporal Gunther RW, et al. Acute myocardial infarction: assessment of left
ventricular function with 16-detector row spiral CT versus MR
resolution, cardiac CT provides reliable information about imaging–study in pigs. Radiology. 2005;236(1):112–7.
the function of the cardiac chambers. Since the functions of 9. Hundt W, Siebert K, Wintersperger BJ, Becker CR, Knez A, Reiser
four cardiac chambers are interrelated, the functional abnor- MF, et  al. Assessment of global left ventricular function: com-
mality can be double checked by the adjacent chamber func- parison of cardiac multidetector-row computed tomography with
angiocardiography. J Comput Assist Tomogr. 2005;29(3):373–81.
tion. A functional compensatory mechanism among 10. Raman SV, Shah M, McCarthy B, Garcia A, Ferketich AK. Multi-­
chambers can be evaluated. In addition, by analyzing the detector row cardiac computed tomography accurately quantifies
multiple chamber function simultaneously, a comprehensive right and left ventricular size and function compared with cardiac
magnetic resonance. Am Heart J. 2006;151(3):736–44.
global cardiac function can be evaluated.
11. Yamamuro M, Tadamura E, Kubo S, Toyoda H, Nishina T,

The main expectation from current cardiac CT may com- Ohba M, et  al. Cardiac functional analysis with multi-detector
prise epicardial coronary arterial patency, the myocardial row CT and segmental reconstruction algorithm: comparison
perfusion status, and the cardiac function as a circulation with echocardiography, SPECT, and MR imaging. Radiology.
2005;234(2):381–90.
pump. The multichamber function evaluation will provide
12. Maffei E, Messalli G, Martini C, Nieman K, Catalano O, Rossi A,
more valuable information about the cardiac performance in et al. Left and right ventricle assessment with cardiac CT: valida-
addition to CT coronary angiography. Based on the morpho-­ tion study vs. cardiac MR. Eur Radiol. 2012;22(5):1041–9. Pubmed
functional results, more reliable explanation about the Central PMCID: 3321142.
13. Groen JM, van der Vleuten PA, Greuter MJ, Zijlstra F, Oudkerk
patient’s discomfort will be possible.
M.  Comparison of MRI, 64-slice MDCT and DSCT in assessing
functional cardiac parameters of a moving heart phantom. Eur
Radiol. 2009;19(3):577–83. Pubmed Central PMCID: 2816236.
14. Arsanjani R, Berman DS, Gransar H, Cheng VY, Dunning A,

References Lin FY, et al. Left ventricular function and volume with coronary
CT angiography improves risk stratification and identification of
1. Mohrman DE, Heller LJ. Cardiovascular physiology, McGraw- patients at risk for incident mortality: results from 7758 patients
Hill, New York; 2006. p. 47–70. in the prospective multinational CONFIRM observational cohort
2. Fuchs A, Kuhl JT, Lonborg J, Engstrom T, Vejlstrup N, Kober L, et al. study. Radiology. 2014;273(1):70–7.
Automated assessment of heart chamber volumes and function in 15. Mao SS, Li D, Vembar M, Gao Y, Luo Y, Lam F, et al. Model-based
patients with previous myocardial infarction using multidetector com- automatic segmentation algorithm accurately assesses the whole
puted tomography. J Cardiovasc Comput Tomogr. 2012;6(5):325–34. cardiac volumetric parameters in patients with cardiac CT angiog-
3. Stouffer GA. Cardiovascular hemodynamics for the clinician, raphy: a validation study for evaluating the accuracy of the worksta-
Wiley Blackwell, Oxford; 2008. p. 3-100. tion software and establishing the reference values. Acad Radiol.
4. Galie N, Hoeper MM, Humbert M, Torbicki A, Vachiery JL, 2014;21(5):639–47.
Barbera JA, et al. Guidelines for the diagnosis and treatment of pul- 16. Dell'Italia LJ, Walsh RA. Effect of intravenous metoprolol on left
monary hypertension: the task force for the diagnosis and treatment ventricular performance in Q-wave acute myocardial infarction.
of pulmonary hypertension of the European Society of Cardiology Am J Cardiol. 1989;63(3):166–71.
(ESC) and the European Respiratory Society (ERS), endorsed 17. Port S, Cobb FR, Jones RH. Effects of propranolol on left ventricu-
by the International Society of Heart and Lung Transplantation lar function in normal men. Circulation. 1980;61(2):358–66.
(ISHLT). Eur Heart J. 2009;30(20):2493–537. 18. Silke B, Verma SP, Frais MA, Reynolds G, Taylor SH. Comparative
5. van Lieshout JJ, Wesseling KH. Continuous cardiac output by pulse effects of metoprolol and celiprolol on cardiac hemodynamics and
contour analysis? Br J Anaesth. 2001;86(4):467–9. left ventricular volume at rest and during exercise-induced angina.
6. Longo DL, Fauci AS, Kasper DL, Hauser SL, Jameson JL, Loscalzo Clin Pharmacol Ther. 1986;39(1):5–14.
J. Harrison’s principles of internal medicine: McGraw-Hill; New 19. Tsusaki H, Yonamine H, Tamai A, Shimomoto M, Kuwano K, Iwao
York, 2012. H, et al. Left ventricular volume and function in cynomolgus mon-
7. Tanabe Y, Kido T, Kurata A, Sawada S, Suekuni H, Kido T, keys using real-time three-dimensional echocardiography. J Med
et  al. Three-dimensional maximum principal strain using cardiac Primatol. 2007;36(1):39–46.
36  Three-Chamber Function with Cardiac CT 439

20. Mo YH, Jaw FS, Wang YC, Jeng CM, Peng SF. Effects of proprano- resonance imaging, 320-slice multi-detector computed tomogra-
lol on the left ventricular volume of normal subjects during CT cor- phy, and transthoracic echocardiography. Eur Heart J Cardiovasc
onary angiography. Korean J Radiol. 2011;12(3):319–26. Pubmed Imaging. 2014;15(5):532–40.
Central PMCID: 3088849. 39. Choi MJ, Kim JK, Kim SG, Yoon JW, Koo JR, Kim HJ, et al. Left
21. Martin GV, Sheehan FH, Stadius M, Maynard C, Davis KB, Ritchie atrial volume index is a predictor of silent myocardial ischemia in
JL, et  al. Intravenous streptokinase for acute myocardial infarc- high-risk patients with end-stage renal disease. Int J Cardiovas Imag.
tion. Effects on global and regional systolic function. Circulation. 2013;29(7):1433–9. PubMed PMID: WOS:000325716000004.
1988;78(2):258–66. English.
22. Zeb I, Li D, Nasir K, Gupta M, Kadakia J, Gao Y, et al. Computerized 40. Feuchtner GM, Dichtl W, Bonatti JO, Jodocy D, Muller S,

left ventricular regional ejection fraction analysis for detection Hintringer F, et  al. Diagnostic accuracy of cardiac 64-slice com-
of ischemic coronary artery disease with multidetector CT angi- puted tomography in detecting atrial thrombi. Comparative study
ography. Int J Cardiovasc Imaging. 2013;29(3):685–92. Pubmed with transesophageal echocardiography and cardiac surgery.
Central PMCID: 4277887. Investig Radiol. 2008;43(11):794–801.
23. Yee S, Scalzetti EM. Measurement of cardiac output from dynamic 41. Christiaens L, Varroud-Vial N, Ardilouze P, Ragot S, Mergy J,
pulmonary circulation time CT. Med Phys. 2014;41(6):061904. Bonnet B, et al. Real three-dimensional assessment of left atrial and
24. Mahnken AH, Klotz E, Hennemuth A, Jung B, Koos R,
left atrial appendage volumes by 64-slice spiral computed tomog-
Wildberger JE, et al. Measurement of cardiac output from a test-­ raphy in individuals with or without cardiovascular disease. Int J
bolus injection in multislice computed tomography. Eur Radiol. Cardiol. 2010;140(2):189–96.
2003;13(11):2498–504. 42. Im SI, Na JO, Kim SW, Choi CU, Kim JW, Yong HS, et al. Adjusted
25. Vanhoenacker PK, Van Hoe LR. A simple method to estimate car- left atrial emptying fraction as a predictor of procedural outcome
diac function during routine multi-row detector CT exams. Eur after catheter ablation for atrial fibrillation. Tex Heart Inst J.
Radiol. 2007;17(11):2845–51. 2015;42(3):216–25. Pubmed Central PMCID: 4473613.
26. Truesdell C, Noll W, Antman SS. The non-linear field theories of 43. Nattel S, Burstein B, Dobrev D. Atrial remodeling and atrial fibrilla-
mechanics. 3rd ed. Berlin, New York: Springer; 2004. xxix, 602 p. tion: mechanisms and implications. Circ Arrhythm Electrophysiol.
p. 2008;1(1):62–73.
27. Buss SJ, Schulz F, Mereles D, Hosch W, Galuschky C, Schummers 44. Caselli S, Canali E, Foschi ML, Santini D, Di Angelantonio E,
G, et al. Quantitative analysis of left ventricular strain using cardiac Pandian NG, et  al. Long-term prognostic significance of three-­
computed tomography. Eur J Radiol. 2014;83(3):e123–30. dimensional echocardiographic parameters of the left ventricle and
28. Ganten MK, Krautter U, von Tengg-Kobligk H, Bockler D,
left atrium. Eur J Echocardiogr. 2010;11(3):250–6.
Schumacher H, Stiller W, et al. Quantification of aortic distensibil- 45. Boogers MJ, van Werkhoven JM, Schuijf JD, Delgado V,

ity in abdominal aortic aneurysm using ECG-gated multi-detector El-Naggar HM, Boersma E, et  al. Feasibility of diastolic func-
computed tomography. Eur Radiol. 2008;18(5):966–73. tion assessment with cardiac CT: feasibility study in com-
29. Seo J, Choi D, Rienmueller R, Lim JG, Chang Y, Lee J. Suggestion parison with tissue Doppler imaging. J Am Coll Cardiol Img.
for a new image-based aortic wall stiffness evaluation tech- 2011;4(3):246–56.
nique: arterial wall stiffness index. Int J Cardiovasc Imaging. 46. Kusunose K, Motoki H, Popovic ZB, Thomas JD, Klein AL,

2009;25(Suppl 1):83–94. Marwick TH.  Independent association of left atrial function with
30. Hoit BD. Left atrial size and function: role in prognosis. J Am Coll exercise capacity in patients with preserved ejection fraction. Heart.
Cardiol. 2014;63(6):493–505. 2012;98(17):1311–7.
31. Prioli A, Marino P, Lanzoni L, Zardini P.  Increasing degrees of 47. Di Bella G, Siciliano V, Aquaro GD, De Marchi D, Rovai D, Carerj
left ventricular filling impairment modulate left atrial function in S, et al. Right ventricular dysfunction: an independent and incre-
humans. Am J Cardiol. 1998;82(6):756–61. mental predictor of cardiac deaths late after acute myocardial
32. Barbier P, Solomon SB, Schiller NB, Glantz SA. Left atrial relax- infarction. Int J Cardiovasc Imaging. 2015;31(2):379–87.
ation and left ventricular systolic function determine left atrial res- 48. Goldhaber SZ, Elliott CG.  Acute pulmonary embolism: part

ervoir function. Circulation. 1999;100(4):427–36. II: risk stratification, treatment, and prevention. Circulation.
33. Boudoulas KD, Paraskevaidis IA, Boudoulas H, Triposkiadis
2003;108(23):2834–8.
FK.  The left atrium: from the research laboratory to the clinic. 49. Dogan H, Kroft LJ, Huisman MV, van der Geest RJ, de Roos
Cardiology. 2014;129(1):1–17. A.  Right ventricular function in patients with acute pulmonary
34. Schweitzer A, Agmon Y, Aronson D, Abadi S, Mutlak D, Carasso embolism: analysis with electrocardiography-synchronized multi-­
S, et  al. Assessment of left sided filling dynamics in diastolic detector row CT. Radiology. 2007;242(1):78–84.
dysfunction using cardiac computed tomography. Eur J Radiol. 50. Sato T, Tsujino I, Oyama-Manabe N, Ohira H, Ito YM, Sugimori H,
2015;84(10):1930–7. et al. Simple prediction of right ventricular ejection fraction using
35. Rahimtoola SH, Ehsani A, Sinno MZ, Loeb HS, Rosen KM,
tricuspid annular plane systolic excursion in pulmonary hyperten-
Gunnar RM.  Left atrial transport function in myocardial infarc- sion. Int J Cardiovasc Imaging. 2013;29(8):1799–805.
tion. Importance of its booster pump function. Am J Med. 51. Pienn M, Kovacs G, Tscherner M, Johnson TR, Kullnig P,

1975;59(5):686–94. Stollberger R, et al. Determination of cardiac output with dynamic
36. Hoit BD, Shao Y, Gabel M, Walsh RA.  In vivo assessment of contrast-enhanced computed tomography. Int J Cardiovasc
left atrial contractile performance in normal and pathological Imaging. 2013;29(8):1871–8.
conditions using a time-varying elastance model. Circulation. 52. Yamasaki Y, Nagao M, Kamitani T, Yamanouchi T, Kawanami S,
1994;89(4):1829–38. Yamamura K, et al. Clinical impact of left ventricular eccentricity
37. Tsang TS, Abhayaratna WP, Barnes ME, Miyasaka Y, Gersh BJ, index using cardiac MRI in assessment of right ventricular hemo-
Bailey KR, et al. Prediction of cardiovascular outcomes with left dynamics and myocardial fibrosis in congenital heart disease. Eur
atrial size: is volume superior to area or diameter? J Am Coll Radiol. 2016;26(10):3617–25.
Cardiol. 2006;47(5):1018–23. 53. Muller HM, Tripolt MB, Rehak PH, Groell R, Rienmuller R,

38. Agner BF, Kuhl JT, Linde JJ, Kofoed KF, Akeson P, Rasmussen Tscheliessnigg KH. Noninvasive measurement of pulmonary vas-
BV, et al. Assessment of left atrial volume and function in patients cular resistances by assessment of cardiac output and pulmonary
with permanent atrial fibrillation: comparison of cardiac magnetic transit time. Investig Radiol. 2000;35(12):727–31.
440 J. Lee

54. Revel MP, Faivre JB, Remy-Jardin M, Delannoy-Deken V, Duhamel tracking echocardiography in patients with right ventricular myo-
A, Remy J.  Pulmonary hypertension: ECG-gated 64-section CT cardial infarction. Int J Cardiovasc Imaging. 2017;33(1):47–56.
angiographic evaluation of new functional parameters as diagnostic 62. Lin FY, Devereux RB, Roman MJ, Meng J, Jow VM, Jacobs A,
criteria. Radiology. 2009;250(2):558–66. et al. Cardiac chamber volumes, function, and mass as determined
55. Tan RT, Kuzo R, Goodman LR, Siegel R, Haasler GB, Presberg by 64-multidetector row computed tomography: mean values
KW. Utility of CT scan evaluation for predicting pulmonary hyper- among healthy adults free of hypertension and obesity. J Am Coll
tension in patients with parenchymal lung disease. Medical College Cardiol Img. 2008;1(6):782–6.
of Wisconsin Lung Transplant Group. Chest. 1998;113(5):1250–6. 63. Stojanovska J, Cronin P, Patel S, Gross BH, Oral H, Chughtai K,
56. Chan AL, Juarez MM, Shelton DK, MacDonald T, Li CS, Lin TC, et  al. Reference normal absolute and indexed values from ECG-­
et al. Novel computed tomographic chest metrics to detect pulmo- gated MDCT: left atrial volume, function, and diameter. AJR Am J
nary hypertension. BMC Med Imaging. 2011;11:7. Pubmed Central Roentgenol. 2011;197(3):631–7.
PMCID: 3073886. 64. Stolzmann P, Scheffel H, Leschka S, Schertler T, Frauenfelder T,
57. Remy-Jardin M, Duhamel A, Deken V, Bouaziz N, Dumont
Kaufmann PA, et al. Reference values for quantitative left ventricu-
P, Remy J.  Systemic collateral supply in patients with chronic lar and left atrial measurements in cardiac computed tomography.
thromboembolic and primary pulmonary hypertension: assess- Eur Radiol. 2008;18(8):1625–34.
ment with multi-detector row helical CT angiography. Radiology. 65. Budoff MJ, Pagali SR, Hamirani YS, Chen A, Cheu G, Gao Y, et al.
2005;235(1):274–81. Sex-specific biatrial volumetric measurements obtained with use
58. Collins MA, Pidgeon JW, Fitzgerald R.  Computed tomog-
of multidetector computed tomography in subjects with and with-
raphy manifestations of tricuspid regurgitation. Br J Radiol. out coronary artery disease. Tex Heart Inst J. 2014;41(3):286–92.
1995;68(814):1058–60. Pubmed Central PMCID: 4060347.
59. Yeh BM, Kurzman P, Foster E, Qayyum A, Joe B, Coakley
66. Yamasaki Y, Nagao M, Kawanami S, Kamitani T, Sagiyama K,
F. Clinical relevance of retrograde inferior vena cava or hepatic vein Yamanouchi T, et al. One-stop shop assessment for atrial septal
opacification during contrast-enhanced CT. AJR Am J Roentgenol. defect closure using 256-slice coronary CT angiography. Eur Radiol.
2004;183(5):1227–32. 2017;27(2):697–704.
60. Sivak JA, Raina A, Forfia PR. Assessment of the physiologic contri- 67. Addetia K, Takeuchi M, Maffessanti F, Nagata Y, Hamilton J, Mor-­
bution of right atrial function to total right heart function in patients Avi V, et al. Simultaneous longitudinal strain in all 4 cardiac cham-
with and without pulmonary arterial hypertension. Pulm Circ. bers: a novel method for comprehensive functional assessment of
2016;6(3):322–8. the heart. Circ Cardiovasc Imaging. 2016;9(3):e003895.
61. Nourian S, Hosseinsabet A, Jalali A, Mohseni-Badalabadi R.

Evaluation of right atrial function by two-dimensional speckle-
 Part VIII
Where We Are: Cardiac Imaging Outside the
Coronaries
 Diseases of the Myocardium
and Pericardium 37
Ana Paula S. Lima and Karen G. Ordovas

Computed tomography images of the heart can be used

­clinically for assessment of pericardial and myocardial dis-
eases. Delineation of a normal pericardium and characteriza-
tion of simple and complex pericardial collections can be
performed with cardiac-gated CT. Several abnormalities of
the myocardium can also be assessed with cardiac CT, par-
ticularly regarding cardiac morphology and perfusion, but
also functional abnormalities. More commonly, myocardial
diseases are identified as incidental findings on a chest CT
for other clinical purposes and can aid to clinical manage-
ment and risk stratification.

Pericardial Disease

Pericardial Anatomy

The pericardium is composed of an inner visceral layer,

attached to the myocardium and epicardial fat, and an outer
parietal layer containing serous fluid, normally up to
50  mL.  On CT studies, the normal pericardium is usually Fig. 37.1  Abnormally thick pericardium on cardiac CT
less than 2 mm thick and if greater than 4 mm is considered
abnormal (Fig. 37.1).
At the pericardial insertion points, reflections of pericar- Pericardial sinuses and recesses can be confused for
dium are formed and may contain larger amount of serous lymphadenopathy and masses, and therefore knowledge of
fluid, the so-called sinuses and recesses. Two sinuses are its normal appearance on cross-sectional images is essential.
formed by the pericardial reflections: the transvers sinus and The posterior extension of the superior pericardial recess
the oblique sinus. The pericardial recesses are named the (Fig.  37.2) often mimics right paratracheal adenopathy,
superior and inferior aortic recesses, left and right pulmonic whereas the anterior extension can be confused with aortic
recesses, posterior pericardial recess, left and right pulmo- intramural hematoma, adenopathy, or mediastinal mass.
nary vein recesses, and postcaval recess. Another common pitfall is the pulmonary vein recesses, par-
ticularly on the right, which can be mistaken for lymphade-
nopathy or pulmonary nodules (Fig. 37.3).
A. P. S. Lima
Department of Radiology and Biomedical Imaging, University of
California San Francisco, San Francisco, CA, USA Pericardial Effusion
K. G. Ordovas (*)
Department of Radiology, University of California San Francisco, Accumulation of more than 50  ml of pericardial fluid is
San Francisco, CA, USA
­considered pericardial effusion. Pericardial effusions can rep-
e-mail: [email protected]

© Humana Press 2019 443

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_37
444 A. P. S. Lima and K. G. Ordovas

Fig. 37.4  Simple pericardial effusion

Fig. 37.2  Posterior aspect of the superior pericardial recess

Fig. 37.5  Loculated pericardial effusion

Fig. 37.3  Pulmonary vein recesses

resent transudate or exudate and have multiple mechanisms date from exudates is not always possible based on imaging
including inflammation, infection, trauma and iatrogenic insult. appearance alone. However, increased and heterogeneous
The main role of cardiac CT in the assessment of pericar- fluid density and areas of nondependent collections suggest
dial effusion is to differentiate simple from complex collec- exudates (Fig. 37.5).
tions, to delineate the distribution of loculated effusions, and The characterization of a hemorrhagic effusion, known as
to guide interventional procedures. hemopericardium, may be very important for guiding clini-
Simple pericardial effusion (transudate) has a density cal management. Hemopericardium can be seen in the set-
close to water on CT (Fig. 37.4). Differentiation of transu- ting of malignancy, trauma, or aortic rupture, and the CT
37  Diseases of the Myocardium and Pericardium 445

Fig. 37.6 Hemopericardium Fig. 37.7  Cardiac tamponade

Pericarditis
appearance depends on the age of the bleeding. Acute hemo-
pericardium has a high density on CT, whereas subacute Pericarditis is characterized by inflammation of the pericar-
hematomas present with a lower heterogeneous density dial layers, usually without a definite etiology, but most com-
(Fig. 37.6). monly due to viral infections. Other conditions associated
with pericarditis are thoracic radiation, collagen vascular
diseases, uremia, myocardial infarction, and cardiac surgery.
Pericardial Tamponade Tuberculosis is still a very common cause of pericarditis
worldwide.
Pericardial tamponade is characterized by impaired diastolic The classical clinical presentation of pericarditis is sharp
relaxation and ventricular filling due to accumulation of peri- and positional chest pain, often preceded by viral illness, dif-
cardial fluid that increases the intrapericardial pressure and fuse ST segment elevation on ECG, and pericardial friction
causes an acute constrictive effect in the heart. Usually the rub on physical examination.
hemodynamic effect is more pronounced on the lower pres- The most common cardiac CT finding of pericarditis is
sure right atrium and ventricle and can compromise the pericardial thickening and enhancement, which is usually
venous return and cardiac output. accompanied by effusion. Rarely, increased density of the
The volume and rapidity of fluid accumulation within the epicardial and pericardial fat can be visualized. Pericardial
pericardium will determine patients’ hemodynamic instabil- enhancement is better appreciated in a late venous phase,
ity. In the acute setting, a volume as low as 150–200 ml can usually more than 75 s after contrast injection (Fig. 37.8).
lead to tamponade, whereas a gradual accumulation in a The presence of pericardial calcification and adhesions
chronic situation allows the pericardium to accumulate more are suggestive of a chronic pericarditis and can be clearly
than 1 L of fluid without hemodynamic instability. delineate in a cardiac CT.
The method of choice to interrogate the presence of car-
diac tamponade is echocardiography. Cardiac CT showing a
large effusion with compression of the right cardiac cham- Constrictive Pericarditis
bers can be diagnostic in the correct clinical scenario and can
indeed be the first indication of a life-threatening diagnosis Patients with chronic pericarditis may develop pericardial
(Fig. 37.7). In addition, very commonly cardiac CT can iden- fibrosis that impairs ventricular relaxation and leads to heart
tify the cause for the bleeding and expedite patient’s clinical failure, a condition called constrictive pericarditis. Although
workup. many conditions can lead to constrictive pericarditis, the
446 A. P. S. Lima and K. G. Ordovas

Fig. 37.9  Pericardial calcification and atrial enlargement in constric-

Fig. 37.8  Pericarditis with pericardial enhancement tive pericarditis

probability of this complication is related to the etiology and patient with constrictive/restrictive physiology is highly sug-
is high in bacterial pericarditis (20–30%), intermediate in gestive of constrictive pericarditis, and the specificity
immune-mediated and neoplastic pericarditis (2–5%), and increases when thickness is greater than 6 mm. In approxi-
low in viral pericarditis (<1%) [1]. mately 20% of the cases, constrictive pericarditis is seen in
Whereas in developing countries tuberculosis remains the the absence of a thick pericardium [3].
main cause of constrictive pericarditis, the most common Additional morphologic abnormalities seen on cardiac CT
pericardial constriction insult in Europe and North America are enlargement of the atria and, in advanced disease, narrow-
is cardiac surgery. ing and elongation of ventricles. Dilatation of the inferior vena
Constrictive pericarditis may have a subacute presenta- cava and hepatic veins can also be seen. Contrast enhancement
tion or a more chronic and insidious clinical presentation on late venous phases and large pericardial effusion suggests
with typical symptoms of dyspnea on exertion, fatigue, lower the inflammatory and effusive subtypes, respectively, whereas
extremity swelling, and hepatic congestion. Occasionally, thick adhesions between the pericardial layers are characteris-
clinical presentation is atypical with symptoms resembling tics of adhesive subtype of constrictive pericarditis. Although
those of primary liver disease [2]. a retrospectively gated cardiac CT can demonstrate an early
The most common type of constrictive pericarditis is the diastolic septal bounce, echocardiography and cardiac MRI
chronic fibrous type, characterized by thick fibrous pericar- are better methods to identify this abnormality.
dium, with calcifications in approximately 30% of the cases,
which usually involve the atrioventricular grooves. Additional
types of constrictive pericarditis include acute inflammatory, Myocardium
effusive-constrictive, and adhesive [3].
Echocardiography is often the first imaging test per- CT as a Primary Tool for Myocardial Diseases
formed for evaluation of constrictive pericarditis, and classi-
cally demonstrates preserved left ventricular contractility, CT is rarely used as the primary technique for evaluation of
and signs of ventricular interdependence such as an early myocardial diseases. However, evaluation of the myocar-
diastolic ventricular septal bounce. However, very rarely dium is frequently performed in a comprehensive cardiac CT
echocardiography is able to document the presence of peri- for ischemic heart disease assessment. Although coronary
cardial thickening. imaging is the most important application of CT in the
Cardiac CT is an excellent method to evaluate the pericar- assessment of cardiac diseases, modern scanners allow for
dium in its entirety and to detect the presence of calcification visualization of perfusion defects and myocardial contrac-
(Fig.  37.9). Pericardial thickening of 4  mm or greater in a tion abnormalities that can be used to characterize the hemo-
37  Diseases of the Myocardium and Pericardium 447

Fig. 37.10  Apical type of hypertrophic cardiomyopathy

Fig. 37.11  Dilated cardiomyopathy on non-gated cardiac CT

dynamic significance of coronary lesions. In addition,
delayed cardiac CT images can demonstrate areas of delayed Calcifications
myocardial enhancement that have been shown to correlate Myocardial calcifications can be easily identified on chest
with the presence of myocardial scar and irreversible myo- CT examinations and usually indicate sequela of a previous
cardial injury. Detailed discussion of the use of cardiac CT myocardial insult [4]. The most common type of myocardial
for myocardial perfusion and viability are discussed on a calcification seen on CT are dystrophic calcifications within
dedicated chapter of this book. areas of necrotic scar due to an old myocardial infarction,
In patients with nonischemic myocardial disease, cardiac which are usually subendocardial following a coronary terri-
CT can provide additional information to echocardiography tory (Fig. 37.12). Frequently, the presence of myocardial cal-
particularly in patients with hypertrophic cardiomyopathy. cification is diagnostic of an old myocardial infarction which
Cross-sectional imaging is commonly used in HCM patients may be unknown to the patient and the medical team and
to characterize the distribution and quantify the severity of would indicate the need for secondary prevention measures.
myocardial hypertrophy, which could aid in risk stratifica- Myocardial calcifications can also result from previous myo-
tion and guide therapy decisions (Fig. 37.10). It can also help carditis, trauma, or iatrogenic injuries. Rarely, metastatic
differentiate between severe myocardial hypertrophy and myocardial calcifications may occur in the setting of impaired
myocardial mass. In patients with dilated cardiomyopathy, calcium-phosphorus metabolism, usually in patients with
cardiac CT can provide precise quantification of ventricular chronic renal failure or hyperparathyroidism [5].
volumes and function, which can be challenging for echocar-
diography when ventricular volumes are very large Wall Thinning and Aneurysm
(Fig. 37.11). A common cardiac CT application in patients Similar to myocardial calcifications, the presence of myocar-
with dilated cardiomyopathy is assessment of ventricular dial wall thinning reflects the presence of a previous insult to
thrombi suspected on echocardiography evaluation. the cardiac muscle, which can be ischemic or nonischemic in
etiology. Areas of wall thinning that match a coronary terri-
tory are classically due to sequel of a chronic myocardial
Incidental Cardiac CT Findings on Chest CT infarction (Fig. 37.13). Diffuse myocardial wall thinning can
be seen in the setting of dilated cardiomyopathy from several
Very commonly myocardial abnormalities are seen as inci- etiologies. Retrospectively cardiac-gated CT examinations
dental findings on chest CT or coronary CT studies and may can aid in functional characterization of areas of myocardial
in fact be important for establishing the diagnosis and thinning. This functional assessment can be particularly use-
informing patient management. ful in differentiating normal apical thinning, which should
448 A. P. S. Lima and K. G. Ordovas

a b

Fig. 37.12 (a, b) Myocardial calcification due to old myocardial infarction

ity and mortality. True ventricular aneurysms usually involve

the left apical and anterolateral walls and have a wide point
of attachment with the chamber. These aneurysms increase
the risk of thrombus formation and consequent embolic
events. False ventricular aneurysms are most commonly seen
in the diaphragmatic surface of the ventricle and impose a
high risk of cardiac rupture, characterizing a cardiac urgency.

Hypoperfusion
In patients with an old myocardial infarction or myocardial
ischemia at rest, areas of decreased subendocardial perfusion
can be identified in cardiac CT and chest CT scans. This
finding is particularly meaningful if matching the coronary
territory of a vessel with evidence of atherosclerotic disease
(Fig. 37.14). Although perfusion defects are easily identifi-
able in cardiac-gated CT for coronary artery disease evalua-
tion, incidental areas of subendocardial perfusion
abnormalities may be visualized on non-gate chest CT
angiograms for pulmonary artery or aorta investigation.
However, non-gated chest CT studies have a greater potential
to be affected by motion and hard beaming artifact, which
Fig. 37.13  Left ventricular wall thinning with associated calcification
could limit the assessment of subendocardial perfusion
defects [7].
show normal contractility, from an apical myocardial
­infarction [6].
When associated with outpouching of the ventricular Conclusion
walls in diastole, areas of wall thinning are characteristic of
ventricular aneurysms. Detection of ventricular aneurysms Cardiac CT has great utility for evaluation of pericardia dis-
may have a significant impact in clinical management as they eases, particularly as an aid to echocardiography in charac-
are associated with increased risk of cardiovascular morbid- terizing complex collections, loculation, and pericardial
37  Diseases of the Myocardium and Pericardium 449

a b

Fig. 37.14 (a, b) Subendocardial perfusion defect at rest in a patient with old myocardial infarction

2. Garcia MJ.  Constrictive pericarditis versus restrictive cardiomy-

constriction. Cardiac and chest CT can characterize ischemic opathy? J Am Coll Cardiol. 2016;67(17):2061–76.
myocardial injury as an area of wall thinning, aneurysm, 3. Cummings KW, Green D, Johnson WR, et al. Imaging of pericar-
abnormal contractility, impaired perfusion, and calcifica- dial diseases. Semin Ultrasound CT MR. 2016;37(3):238–54.
4. Austin CO, Kramer D, Canabal J, et al. A heart of stone: a case of
tions. More commonly, these myocardial abnormalities are acute development of cardiac calcification and hemodynamic col-
incidental findings on a non-gated chest CT and could pro- lapse. J Cardiovasc Comput Tomogr. 2013;7(1):66–8.
vide important insight for patients’ diagnosis and 5. Mana M, Sanguineti F, Unterseeh T, et al. Petrified myocardium:
management. the age of stone? Circulation. 2013;126(9):1139–42.
6. Ferencik M, Abbara S, Hoffmann U, et  al. Left ventricular thin-­
point detection using multidetector spiral computed tomography.
Am J Cardiol. 2004;93(7):949–51.
References 7. Shriki JE, Shinbane J, Lee C, et al. Incidental myocardial infarct
on conventional nongated CT: a review of spectrum of findings
1. Fardman A, Charron P, Imazio M, et  al. European guidelines on with gated CT and cardiac MRI correlation. AJR Am J Roentgenol.
pericardial diseases: a focused review of novel aspects. Curr 2012;198(3):496–504.
Cardiol Rep. 2016;18(5):46.
 CT of Cardiac and Paracardiac Masses
38
Harold Goerne and Prabhakar Rajiah

There is a broad spectrum of cardiac and paracardiac masses, Imaging of Cardiac Masses
which includes nonneoplastic as well as neoplastic lesions,
both benign and malignant lesions. Paracardiac masses such Imaging plays an important role in the evaluation of cardiac
as those involving the esophagus, tracheobronchial tree, masses, since most of these are not amenable to biopsy due
lungs, pericardium, diaphragm, and lymph nodes are more to their location and there is high risk of complications.
common than the cardiac masses. The most common cardiac Imaging is necessary to confirm the mass; determine the
mass is a thrombus. The most common neoplastic lesion to exact location, extension, and relationships; distinguish
involve the heart is metastasis, which is 20–40 times more benign and malignant lesions, although a precise diagnosis
common than primary neoplasms [1]. While the overall prev- is not always possible [4]; identify masses at risk of embo-
alence of cardiac metastasis is 1.23% [2], primary cardiac lism and obstruction; and determine management options
neoplasms have a prevalence of 0.001–0.056% [2, 3]. In including biopsy, surgery, and serial follow-up. There are
adults, 75% of the primary cardiac neoplasms are benign, several imaging options available including echocardiogra-
with myxoma being the most common (45–50%) type, fol- phy, magnetic resonance imaging (MRI), CT, and nuclear
lowed by fibroelastoma (15%), lipoma (5–10%), and fibroma medicine. Transthoracic echocardiogram (TTE) is often the
(2–4%). The most common primary cardiac malignancy is first imaging modality used in the evaluation of cardiac
angiosarcoma (10%), followed by sarcomas with myoblastic mass, with a mass often incidentally detected in this modal-
or fibroblastic differentiation (5%) and rhabdomyosarcoma ity. It is widely available, inexpensive, and portable, has
(2–5%) [4, 5]. Pediatric cardiac neoplasms are usually excellent temporal and good spatial resolutions, and can be
benign (90%) [6, 7], with rhabdomyoma (70%) being the performed in hemodynamically unstable patients, and there
most common neoplasm followed by fibroma (10%) [7, 8]. is no radiation exposure. However, it is operator-dependent,
Malignant tumors are extremely rare in children, most of limited in patients with poor acoustic windows, and limited
which are sarcomas, followed by lymphoma. The clinical in imaging certain parts of the heart such as the right ven-
manifestations of cardiac masses depend on the pathology tricle and apical regions, has limited field of view to image
and other features. Even benign cardiac neoplasms can be extracardiac structures, and does not have tissue character-
life-threatening depending on their location, size, and move- ization capabilities [10]. Transesophageal echocardiogram
ment due to hemodynamic compromise, arrhythmia, or (TEE) can be used to overcome some of these limitations,
thromboembolism [9]. In addition, there are several normal particularly improving the spatial resolution, but is an inva-
variants that may mimic masses. sive method [4]. MRI is an important modality in the evalu-
In this chapter, we review the role of computed tomogra- ation of cardiac masses. It has excellent spatial and good
phy (CT) in the evaluation of cardiac masses, CT protocols, temporal resolution, wide field of view, multiplanar imaging
and the imaging appearances of common and uncommon capabilities, and also tissue characterization capabilities,
cardiac and paracardiac masses. primarily due to inherent soft tissue contrast which can be
augmented by several sequences and contrast administra-
tion, which help in distinguishing normal myocardium,
H. Goerne
Department of Radiology, Cardiothoracic Imaging, UT fluid, edema, fat, fibrosis, and hemorrhage. Disadvantages
Southwestern Medical Center, Dallas, TX, USA of MRI include long acquisition times (30 min–1 h.), lim-
P. Rajiah (*) ited availability, need for higher technical expertise, inabil-
Department of Radiology, Cardiothoracic Imaging Division, UT ity to perform in hemodynamically unstable patients, and
Southwestern Medical Center, Dallas, TX, USA contraindications including claustrophobia, several cardiac

© Humana Press 2019 451

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_38
452 H. Goerne and P. Rajiah

devices, and severe renal dysfunction [4]. Nuclear medicine hence a longer scan range may be necessary in such patients.
techniques, primarily 18F-FDG-PET scan, are useful in For cardiac masses, it is advantageous to use a triphasic
assessing the metabolic activity of cardiac lesions, which injection protocol which enables visualization of all the
helps in distinguishing benign from malignant lesions, stag- ­cardiac chambers as well as reduction of streak artifacts in
ing of malignant lesions, and assessing response to therapy. the SVC and right atrium. This involves initial contrast
Nuclear medicine techniques are associated with radiation injection at flow rate of 5–7 ml/s to opacify the left heart,
and have lower spatial resolution. followed by either a contrast injection at a slower rate of
2–4  ml/s or injection of a mixture of contrast and saline
(50:50 or 60:40 ratio) at same flow rate of 5-7 ml/s to opacify
CT in Cardiac Masses the right heart and finally by saline bolus injection to mini-
mize streak artifact. Additional non-contrast phase may be
CT has emerged as an important imaging modality in the acquired if there is a concern of calcification, and a delayed
evaluation of cardiac and paracardiac neoplasms. CT has phase may be acquired if there is a need for distinguishing
several advantages including wide availability and rapid thrombus from slow flow or for delayed enhancement.
acquisition time. The excellent isotropic spatial resolution Images are reconstructed at submillimeter thickness (0.5–
and multiplanar reconstruction capabilities of CT enable 0.75 mm), with small field of view (200–250 mm) for the
excellent, three-dimensional rendering of the masses, which heart. In addition, full field-of-view 2 × 1 mm axial images,
is essential for surgical planning. The wide field of view of 2 × 2 mm coronal and sagittal multiplanar images, and 7 ×
CT enables visualization of paracardiac and other extracar- 2 mm MIP axial images are reconstructed for the evaluation
diac structures, which is essential for staging malignancies. of lungs and other thoracic structures. With retrospective
CT is primarily used in the evaluation of cardiac masses in ECG-gated acquisitions, cine images of the heart can then
patients who cannot have MRI scan due to contraindications. be reconstructed in any desired plane.
It is also useful in specific situations where MRI or echocar-
diography cannot provide optimal information, for instance,
the presence of calcification or evaluation of vascular supply
of tumors. CT can also provide functional dynamic informa- Approach to Cardiac Masses
tion. The principal disadvantages of CT are the use of ioniz-
ing radiation with its theoretical risks and iodinated contrast The first step in evaluation of a cardiac mass is to determine
media which is associated with nephrotoxicity in patients if it is a normal variant or a real mass. There are several nor-
with renal dysfunction. mal variants discussed in detail below, which occur in char-
acteristic locations with characteristic appearance. Awareness
of these normal variants will prevent misdiagnosis as a car-
CT Protocol diac mass. For masses, further determination includes if it is
a nonneoplastic or neoplastic lesion, more particularly dis-
The cardiac CT protocol should be customized based on the tinguishing benign from malignant lesion. Age, gender,
specific clinical question, type, and location of the mass. All symptoms, and other clinical history are important in helping
scans are performed with ECG gating to minimize cardiac to make this distinction. For instance, a mass adjacent to an
motion. Prospective ECG triggering is the default mode due akinetic segment in a patient with prior history of myocardial
to lower radiation exposure, but retrospective ECG gating is infarction is highly suggestive of a thrombus. A cardiac mass
recommended if ventricular functional estimation, mass in a patient with known primary malignancy should alert the
mobility, and valvular motion need to be evaluated. It is also possibility of metastasis. There are several imaging features
opted if coronary arteries are evaluated in patients with high that help in distinguishing benign from malignant lesions.
or irregular heart rates. The radiation dose is higher for a Some features of benign lesions are:
retrospective gated mode, but it can be minimized by using More common on the left; Smaller in size; Smooth, well-
ECG-based tube current modulation [4]. Usually, there is no defined margins; Involvement of one compartment/chamber;
need for beta-blocker and nitroglycerine predication for the No invasive features; May have a pedicle; No feeding artery;
evaluation of cardiac masses. The tube voltage and current Calcification is rare; Pericardial effusion and metastasis are
are selected based on patient size. The scan range depends not seen.
on the localization and extent of the mass. Occasionally the Features of malignant lesions are; More common in the
total extent of the mass may be seen only after the study, and right, especially the atrium; Larger in size; Ill-defined,
38  CT of Cardiac and Paracardiac Masses 453

lobulated, invasive, and infiltrative margins; Involvement

of multiple chambers or compartments (myocardium, peri-
cardium, extracardiac); Broad base of attachment; Feeding
artery may be present; Calcification may be seen in osteo-
sarcoma; Pericardial effusion and metastasis may be seen.

Normal Variants

Prominent Crista Terminalis

Crista terminalis is a fibromuscular ridge located in the

inner posterior aspect of the right atrium (RA), which
divides this chamber into a trabeculated anterolateral wall
and a smooth posterior wall [11]. Crista terminalis repre-
sents the junction of embryological primitive right atrium
and sinus venosus and is caused by regression of the septum
spurium, while the sinus venosus is incorporated into the
right atrial wall. This regression shows a wide range of vari-
ations and can be prominent [12]. A prominent crista termi-
nalis is one of the most common anatomical variants, which Fig. 38.1  Prominent crista terminalis. Axial CT scan shows a focal
can mimic atrial neoplasm, thrombus, or vegetation [13]. On protrusion from the posterolateral aspect of the right atrium (arrow),
CT, it is seen as a linear and well-defined filling defect in the characteristic of a prominent crista terminalis
posterior wall of the right atrium (Fig. 38.1), which crosses
the atrial wall from superior to inferior, with smooth borders
and similar attenuation as that of the atrial wall with no con- Thebesian valves and the crista terminalis [17]. It is
trast enhancement, which is well appreciated in four-cham- believed to be caused by incomplete resorption of the right
ber and axial planes. side of the sinus venosus valve [17, 18]. The estimated
prevalence is about 2% [17, 19] and is often detected inci-
dentally in imaging, particularly echocardiogram. A patent
Persistent Eustachian Valve foramen ovale is present in 83% of cases [17], with higher
prevalence of atrial septal aneurysm and paradoxical embo-
Eustachian valve is located at the junction between the infe- lism [18]. CT plays a role when echocardiogram is incon-
rior vena cava (IVC) and right atrium (RA) and directs oxy- clusive and differential diagnosis such as thrombus or cor
genated blood flow from the IVC toward the foramen ovale triatriatum dexter is considered. On CT, the network is hard
during fetal life [14]. After birth, with the closure of foramen to visualize when there is heterogeneous contrast in the
ovale, the valve tends to regress as it has no specific function right atrium.
[15]. However, this valve can occasionally persist until adult
life and when prominent, can be mistaken for mass on echo-
cardiography. On CT, it is easy to recognize this variant,
whose features depend on the size of the valve and can either Left Lateral Ridge
be rigid and elongated or an undulating membrane [16].
Thebesian valve is located at the junction between the coro- Left lateral ridge is located in the left atrium (LA) between
nary sinus and RA. the left superior pulmonary vein and left atrial appendage
(LAA) [20]. This ridge is also called the Coumadin/warfa-
rin ridge because it used to be often mistaken for thrombus
Chiari Network and anticoagulation therapy was initiated. This is also
called the ligament of Marshall and contains the oblique
The Chiari network is a reticulum of fibers resembling a vein of Marshall, which is an embryological remnant of a
web located in the right atrium, between the Eustachian and left superior vena cava that drains into the coronary sinus
454 H. Goerne and P. Rajiah

[21, 22]. On CT, it is easy to recognize this structure, since Nonneoplastic Masses
it usually has a rounded tip that gives a pedunculated
appearance (Fig. 38.2). Thrombus

Thrombus is the most common intracardiac mass [23].

Predisposing factors include wall motion abnormalities
(myocardial infarction/dilated cardiomyopathy), atrial fibril-
lation, intracardiac devices, and hypercoagulable states [3].
Thrombus is most common in the left heart chambers. Left
atrial thrombus is associated with atrial fibrillation and mitral
valvular abnormalities and is most commonly seen in the pos-
terior wall or the atrial appendage. On CT, thrombus is seen
as a round or oval filling defect with low attenuation
(Fig. 38.3). A left atrial appendage thrombus is often difficult
to distinguish from contrast admixture. An attenuation thresh-
old of 65 HU has sensitivity, specificity, and positive and
negative predictive values of 94%, 97%, 94%, and 97%,
respectively, in the diagnosis of thrombus [24]. Other studies
have shown that there is no significant difference in the mean
CT numbers of thrombus and slow flow (91.7 ± 11.6 HU vs.
85.2 ± 10.9 HU, respectively, P = 0.241), but the mean iodine
concentration (mg/ml) as measured in a dual-energy CT is
significantly different (3.53  ±  0.72 vs. 1.37  ±  0.31, respec-
tively, P < 0.001) [25]. Ratio of left atrial appendage attenua-
Fig. 38.2  Coumadin ridge. Axial CT scan shows a prominent ridge tion/ascending aorta of >0.75 has 100% negative predictive
between the left superior pulmonary vein and the body of the left value in excluding thrombus [26]. A delayed phase (30 s from
atrium, which is consistent with a Coumadin ridge (arrow) early phase) can distinguish slow flow from thrombus since a

a b

Fig. 38.3  Thrombus. (a) Axial CT shows a hypoattenuating filling filling defect in the LV apex (arrow), consistent with a fresh thrombus.
defect in the left atrial appendage (arrow) consistent with a thrombus. Note the subendocardial hypoattenuation in the apex related to hypo-
(b) Four-chamber CT reconstruction shows a small hypoattenuating perfusion due to prior infarction (arrowhead)
38  CT of Cardiac and Paracardiac Masses 455

a b

Fig. 38.4  Contrast admixture. (a) Axial CT scan shows a hypoattenu- age with disappearance of the previously seen hypoattenuation. The
ating filling defect in the left atrial appendage, which has an attenuation mean attenuation in the delayed phase is 115 HU, and the left atrial
of 44.8 HU in the arterial phase (arrow). This could be either due to appendage/ascending aorta ratio is 0.9. All these indicate that this is
thrombus or contrast admixture. (b) Delayed phase axial CT scan in the contrast admixture and not a thrombus
same patient shows homogeneous opacification of the left atrial append-

thrombus is present in both the phases with low attenuation,

but contrast admixture is present only in the early phase
(Fig. 38.4) [10]. Also, in the late phase, the left atrial append-
age/ascending aorta attenuation ratio is lower in thrombus
(0.29HU  ±  0.12) than circulatory stasis (0.85 HU  ±  0.12)
[27]. Thrombus in the left ventricle has a crescentic shape
with broad base and low attenuation but in some cases can be
pedunculated [3]. It is closely associated with wall motion
abnormalities. In the right chambers, thrombus is associated
with central catheters or deep vein thrombosis and should be
considered pulmonary emboli in transit [23]. Other causes
include ARVD, Behcet’s disease, metastases, and trauma.
Chronic thrombus can be calcified and occasionally show
some contrast enhancement due to vascularization (Fig. 38.5).
Thrombus should also be distinguished from neoplasms such
as myxoma. Features in favor of thrombus are - more com-
mon in ventricles than atria; common in left atrial appendage
when located in the left atrium; smaller in size; rarely mobile
or prolapsing; polypoidal; may be calcified; and no contrast
enhancement. Features in favor of myxoma are - more com-
mon in the left atrium; particularly attached to the fossa ova- Fig. 38.5  Chronic thrombus. Axial CT scan shows a densely calcified
lis, larger in size; villous or polypoid; usually pedunculated chronic thrombus (arrow) in the right atrium
456 H. Goerne and P. Rajiah

and mobile; may prolapse into AV valve; rarely calcified; and Caseous Mitral Annular Calcification
shows variable contrast enhancement [10, 27].
Caseous mitral annular calcification (CMAC) is a rare type
of mitral annular calcification (MAC), with an estimated
Lipomatous Hypertrophy of Interatrial Septum incidence of 2.7% of MAC [36], that is caused by degenera-
tion and liquefaction necrosis, resulting in peripheral dense
Lipomatous hypertrophy of interatrial septum (LHIS) is a calcifications and central caseous admixture of liquid cal-
rare (1–8%) condition [28, 29] characterized by abnormal cium, fatty acids, and cholesterol [37, 38]. It is more ­prevalent
fatty deposition in the interatrial groove and inferior pyrami- in elderly patients and in hypertension [39]. It can result in
dal space, which lies between the atrial walls surrounding the abnormal mitral flow, bradyarrhythmia, and AV blocks [39].
fossa ovalis (true septum) [30–32]. Although the exact etiol- It is more common in the posterior mitral annulus and seen
ogy is not known, it is more common in obese patients with as a focal ovoid mass with a “shell” of peripheral calcifica-
a high amount of epicardial fat and hence likely a metabolic tion (>500 HU) with a slightly less hyperattenuating central
condition. Histologically, there is increase in the number of component (350–500 HU) that represents the liquefied state
adipocytes (hyperplasia) [33], both mature adipocytes and [37, 39] (Fig. 38.7). There is no contrast enhancement, which
fetal fat cells (brown fat). Hypertrophy has also been shown distinguishes caseous MAC from neoplasm and abscess.
[34, 35]. There is no capsule and hence this is not a true Calcification also distinguishes it from abscess [40].
lipoma [29, 35]. On CT, there is increased thickness of inter-
atrial septum >2 cm [32, 33, 35], with fat attenuation and the
pathognomonic “dumbbell” shape that spares the fossa ova- Pericardial Cyst
lis, which is devoid of fat (Fig. 38.6). The fat is contiguous
with the epicardial fat, and the thickening is more extensive Pericardial cyst is a congenital abnormality caused by a
in cephalad than caudad region. There is no contrast enhance- pinched-off portion of the pericardium during development
ment, but this area shows uptake in FDG-PET scan, probably [41]. The wall of the cyst is composed of a thin layer of
due to the presence of brown fat [29] or inflammation [32]. fibrous tissue and a single layer of mesothelial cells [42, 43].
Occasionally, accumulation of fat can be extensive and may It is most commonly (80%) seen in the right anterior
obstruct the right atrial inflow [32]. ­cardiophrenic angle [44]. It is usually asymptomatic and

Fig. 38.6  Lipomatous hypertrophy of interatrial septum. Axial CT Fig. 38.7  Caseous MAC. Short-axis CT image shows coarse circum-
scan shows fatty infiltration of interatrial septum with sparing of the ferential calcification of mitral annulus and heterogeneous lower-­
fossa ovalis (arrow), characteristic of lipomatous hypertrophy of the density central core with associated mass effect, consistent with caseous
interatrial septum mitral annular calcification (arrow)
38  CT of Cardiac and Paracardiac Masses 457

bers [46–48]. On CT, it is seen as a well-defined thin-walled

unilocular spherical lesion. The content is usually serous
with attenuation similar to water, but occasionally it can be
complex with soft tissue attenuation making it difficult to be
distinguished from soft tissue mass. Air may be seen inside
the cyst either due to infection or communication with tra-
cheobronchial tree. There is no contrast enhancement of the
cyst content, but the wall may enhance. It is usually asymp-
tomatic but may produce symptoms when large due to com-
pression of adjacent structures.

Hiatal Hernia

Hiatal hernia is a protrusion of the stomach into the thoracic

cavity, with four reported types. Type I (sliding hernia) is the
most common (85%), where the gastric cardia slides up into
the posterior mediastinum [49]. Types II, III, and IV repre-
sent paraesophageal hernias, in which a part of the stomach
lies anterior or lateral to the esophagus. In Type II, the GE
Fig. 38.8  Pericardial cyst. Axial CT scan shows a well-defined, uni- junction remains in its normal position. Type III is a mixed
locular, thin-walled, homogeneous cystic mass in the right cardio- hernia with features of Types I and II.  Type IV is a giant
phrenic angle (arrow) in continuity with pericardium, consistent with a paraesophageal hernia with one-third to one-half of the
pericardial cyst
stomach protruding into the thoracic cavity [49]. It is com-
mon in patients >50  years. When large, hiatal hernia may
incidentally encountered in imaging but may occasionally produce posterior indentation of the left atrium and may be
present with chest pain or dyspnea. On CT, it is seen as a confused with a mass in echocardiogram. It is easy to iden-
well-defined unilocular lesion with smooth border and tify in radiograph and barium swallow, but sometimes CT
homogenous density similar to water (−10 to 10 HU) scan (Fig. 38.9) is required for establishing the diagnosis and
(Fig. 38.8). Higher attenuation may be seen with infection or distinguishing other sinister lesions in the posterior medias-
hemorrhage, and there is no contrast enhancement. tinum. Gastric folds are seen within the hiatal hernia.
Differential diagnosis includes other cystic mediastinal
lesions, including bronchogenic cyst, esophageal duplication
cyst, neurenteric cyst, thymic cyst, intrathoracic meningo- Other Nonneoplastic Masses
cele, mature cystic teratoma, and lymphangioma [43], which
can be distinguished based on location. For example, dupli- Coronary artery aneurysms and pseudoaneurysms can pro-
cation cyst is located adjacent to the esophagus, usually in a duce mass effect on the heart. Coronary artery aneurysm
cervical location, and bronchogenic cyst is located around refers to a focal dilatation that is at least 1.5 times the normal
the carina, whereas a pericardial cyst is characteristically diameter of the coronary artery and involves less than 50% of
seen in the cardiophrenic angle. If the cyst communicates its length [50]. The most common cause of coronary aneu-
with the pericardial sac, it is then called a diverticulum. rysm is atherosclerosis in adults and Kawasaki disease in
childhood, with other causes including inflammatory disor-
ders, connective tissue diseases, mycotic aneurysm, congeni-
Bronchogenic Cyst tal (ALCAPA) or trauma/iatrogenic [50]. A giant coronary
artery aneurysm can be seen with luminal thrombosis. CT is
Bronchogenic cyst is a congenital lesion resulting from used in the diagnosis as well as establishment of relationship
abnormal budding of the embryonic ventral foregut [45].The with adjacent structures (Fig.  38.10a). Pseudoaneurysms
wall of the cyst resembles the tracheobronchial epithelium, with or without luminal thrombus may also mimic a cardiac
i.e., pseudostratified columnar respiratory epithelium, with mass (Fig. 38.10b). Coronary artery bypass graft aneurysm
variable amounts of smooth muscle, mucous glands, and car- can also produce mass effect on the heart. These are related to
tilage [43]. It is most commonly seen in the mediastinum accelerated atherosclerosis, seen after 5 years of bypass sur-
(85%) [45], usually adjacent to the carina or along the tra- gery [51], whereas pseudoaneurysms may be present after
cheobronchial tree. It can also be seen in lungs and rarely in 6  months, especially in the anastomotic sites [51].
other locations such as interatrial septum or in cardiac cham- Complications of CABG aneurysm include thrombosis, fistu-
458 H. Goerne and P. Rajiah

a b

Fig. 38.9  Hiatal hernia. (a) Axial CT scan in a patient with suspected reconstruction in another patient shows a paraesophageal hernia Type
left atrial mass in echocardiogram shows indentation of posterior wall III (arrow) that displaces the left atrium and coronary sinus (arrowhead)
of the left atrium by a large hiatal hernia (arrow). (b) Short-axis CT cranially

las, and myocardial infarction. Interatrial septal aneurysm is also be seen in the free wall of atrium, AV valve, or ventri-
a focal saccular deformity in the region of fossa ovalis caused cle. Multiple myxomas are seen in 10% of cases, more
by mobile interatrial tissue. It is diagnosed when the maximal common in younger men and those with familial syn-
septal excursion into either atrium is >1.5 cm. It is associated dromes. In Carney syndrome, myxomas often occur outside
with a patent foramen ovale in 33% and ASD in 19%. the left atrium and have higher recurrence rates [56].
Although increased risk of stroke has been reported in patients Pathologically, myxoma is composed of myxoid stromal
with atrial septal aneurysm and PFO, recent studies have cells. There are two anatomical types, solid and papillary,
shown no increased risk [52]. Vegetations of cardiac valves with the former having firm and the latter having a friable
are seen in infective endocarditis. Predisposing factors gelatinous consistence [57]. The clinical manifestations
include endocardial damage, prosthetic material, bacteremia, depend on the type and size of the tumor, with the solid
and congenital heart disease. The most common etiological type producing obstruction of flow at AV valve and conges-
agents are Staphylococcus aureus (31%) and Streptococcus tive cardiac failure, whereas the papillary type produces
viridans (17%) [53]. On CT, vegetation is seen as an irregu- embolic/neurologic symptoms [57]. Fever, malaise, and
lar- or round-shaped hypoattenuating mass attached to a car- weight loss may also be seen. Systemic emboli may be seen
diac valve (Fig. 38.10c), usually in the low pressure side of as a complication of left atrial myxomas, with most of them
the valve or in the subvalvular apparatus or in the ascending located in the central nervous system. Malignant behavior
aorta. The motion of the vegetation and valvular dysfunction has been reported, characterized by recurrence, locally
can be evaluated in retrospective gated cine images [54]. invasive features, extension outside heart and distal metas-
tasis [58]. This has been shown to be faciliated by several
cytokines and growth factors [58]. On CT, myxoma typi-
Benign Neoplasms cally has round or ovoid shape and lobular or smooth bor-
ders with a narrow pedicle in 2/3 (Fig.  38.11a) and
Myxoma broad-­based villous appearance in 1/3 [3] (Fig.  38.11b).
Heterogeneous attenuation is seen. Calcifications are rare,
Myxoma is the most common primary cardiac neoplasm, more common in right atrial lesions. Heterogeneous con-
accounting for 50% of these neoplasms. It is more common trast enhancement is seen in a small number of cases [55].
in women and those between 30 and 60  years [3]. The CT shows the motion and change in position of the mass,
majority (60%) originate in the left atrium [55], typically particularly prolapse of the mass into the ventricle
­
from the fossa ovalis with a small pedicle or stalk. It can (Fig. 38.11c), which is essential for surgical planning.
38  CT of Cardiac and Paracardiac Masses 459

a b

Fig. 38.10  Nonneoplastic masses. (a) Coronary aneurysm. Coronal thrombosed pseudoaneurysm (arrow) as a complication of endocarditis
CT reconstruction shows a giant right coronary aneurysm measuring up in prosthetic aortic valve (arrowhead). (c) Vegetation. Coronal CT
to 10 cm in diameter (arrow) that is producing extrinsic impression on reconstruction shows a linear and irregular mass related to a vegetation
the right ventricle. Note the diffuse and irregular dilatation in the proxi- (arrow) in the ventricular aspect of the right leaflet of aortic valve in a
mal segment of RCA (arrowhead). (b) Pseudoaneurysm. Three-­ patient with endocarditis
chamber CT reconstruction in another patient shows a partially

Fibroelastoma attached to the endocardial surface. Histologically, there is

an avascular matrix of connective tissue composed of proteo-
Papillary fibroelastoma is the second most common primary glycans, collagen, and elastic fibers surrounded by a layer of
neoplasm of the heart [4, 59, 60]. It can occur at any age but hypertrophied endothelial cells [61]. Almost 80% of fibro-
is more common in middle age and older adults with no sex elastomas are located in the surface of a valve [62], with pre-
predilection. The tumor usually is small in size (usually dilection for aortic and mitral valves, usually in the
<2  cm) and has a cauliflower shape with a small stalk non-ventricular surface, sparing the leaflet tips. However, it
460 H. Goerne and P. Rajiah

a b

Fig. 38.11 Myxoma. (a) Four-chamber CT scan shows a characteristic attached to the interatrial septum. (c) Two-chamber view in the same
polypoid left atrial mass (arrow) attached by a stalk to the fossa ovalis, patient as Fig. 38.11b shows the myxoma (arrow) prolapsing through
consistent with a myxoma. (b) Four-chamber CT reconstruction in the mitral valve into the left ventricle. There is near occlusion of the
another patient shows a broad-based left atrial myxoma (arrow) that is inlet of LV and mildly dilated left atrium

can occur in other locations such the chordae tendineae, LV cences (fibrous strands at sites of valve closure), and hyper-
apex, or LVOT.  The friable composition of this tumor is a trophy of Arantius nodules (seen at free margin of aortic
risk factor for rupture and embolic disease, and eventually it valve cusps). Vegetations are irregular with valvular destruc-
can obstruct one of the coronary ostia, resulting in angina tion and systemic signs and symptoms. Usually fibroelas-
[62]. Usually the diagnosis is made by echocardiogram, due toma arises in the midportion of the leaflet, whereas the
to the small size and mobility. On CT, it appears as a small, Lambl’s excrescence arises from the line of closure [63].
homogeneous hypoattenuating mass with a thin stalk that is Arantius nodules are small nodules in the tip of the ventricu-
attached to the valve, sparing the free edges (Fig.  38.12). lar side of the leaflets of semilunar valves, at exact point of
Lesions have been reported up to 70 mm [10]. Differential coaptation. Fibroelastoma is surgically resected if it is symp-
diagnosis includes vegetation, thrombus, Lambl’s excres- tomatic or mobile in the left heart.
38  CT of Cardiac and Paracardiac Masses 461

Lipoma tion, and it is more common in middle and older ages. It can
originate anywhere in the heart, either from the epicardial or
Lipoma is a benign neoplasm consisting of mature adipose subendocardial layers and rarely intramyocardial, and may
tissue surrounded by capsule. There is no gender predilec- extend into pericardial sac. Endocardial lipoma is seen in the
LV or RA roof or along the septum. It can be asymptomatic
but when larger can cause symptoms related to obstruction,
compression, or arrhythmias [56]. It is usually solitary but
may be multiple in tuberous sclerosis [56]. On CT, a lipoma
is seen as a well-defined mass with homogeneous fat attenu-
ation (<−50 UH) and broad base, with no contrast enhance-
ment (Fig. 38.13). A lipoma of the interatrial septum can be
distinguished from lipomatous hypertrophy by the involve-
ment of the fossa ovalis and absence of dumbbell shape.

Hemangioma

Hemangioma is a benign vascular tumor and has three histo-

logic types, capillary, cavernous, and arteriovenous [64],
although some tumors have mix composition of the three
types. It is usually a focal lesion, but diffuse angiomatosis
may be occasionally seen [65]. Seventy-five percent of car-
diac hemangiomas occur in adults [7, 66]. Seventy percent
are seen in the ventricles, and 23% affect the right atrium,
with left atrium rarely affected [64, 67]. Common locations
Fig. 38.12 Fibroelastoma. Three-chamber CT reconstruction shows a
are lateral wall of the LV and anterior wall of the RV. It can
small round mass attached by a stalk to the edge of the right leaflet of affect the endocardium, myocardium, or epicardium and can
aortic valve facing the aortic side, consistent with a fibroelastoma (arrow) be intracavitary. On CT, it is seen as a well-defined round or

a b

Fig. 38.13 Lipoma. (a) Axial CT scan shows a focal ill-defined fat Sagittal reconstruction in the same patient shows the lipoma in the right
attenuation mass in the right atrioventricular groove that displaces the atrioventricular groove displacing the acute angle of the right coronary
right coronary artery laterally (arrow), consistent with a lipoma. (b) artery laterally (arrow)
462 H. Goerne and P. Rajiah

oval tumor when intracavitary but may mold the shape to intense (Fig. 38.14) and heterogenous in hemangioma, simi-
adjacent structures when it is intramyocardial or epicardial. lar to other hypervascular tumors such as paragangliomas
The attenuation is usually heterogeneous with some fat inter- and sarcomas. Fat and calcium may help in distinguishing
spersed between the vascular components. Calcification may hemangioma from these lesions. Slow-flow hemangiomas
be present due to phleboliths. Contrast enhancement is may show little or no contrast enhancement. Although spon-
taneous regression has been reported, basically all hypervas-
cular tumors require surgical excision.

Paraganglioma

Paraganglioma is a rare tumor originating from chromaffin

cells in sympathetic or parasympathetic chains, with only
approximately 300 cases reported in the literature [68]. It can
either be functioning (extra-adrenal pheochromocytoma) or
nonfunctioning (chemodectoma) [56]. Ten percent of pheo-
chromocytomas occur extra-adrenally. Paraganglioma is more
common in young adults. It is usually encapsulated but can be
infiltrative. It is more commonly seen in relation to great ves-
sels than the heart and is commonly seen to originate from the
paraganglia in posterior wall or roof of the left atrium, atrial
septum, or the surrounding coronary arteries or aortic root. On
CT, it is seen as a well-defined round or oval mass with hetero-
geneous attenuation and avid contrast enhancement
(Fig. 38.15). Sometimes the contrast enhancement is hetero-
geneous due to central necrosis. Aggressive features with poor
Fig. 38.14  Hemangioma. Axial CT scan shows an intensely enhanc-
ing mass in between the right and left atrium (arrow), which was proven
margins and extracardiac extension may also be seen. The use
to be a hemangioma of nuclear medicine is helpful in ­localization of extra-adrenal

a b

Fig. 38.15  Paraganglioma. (a) Axial CT scan shows an intensely enhancing intrapericardial mass (arrow) in the posterior aspect of the
enhancing mass in the posterior aspect of the ascending aorta (arrow). ascending aorta, which was biopsy proven to be a paraganglioma
(b) Coronal CT reconstruction in the same patient shows an intensely
38  CT of Cardiac and Paracardiac Masses 463

paragangliomas with the use of I-131 or I-123 metaiodoben- Fibroma

zylguanidine (MIGB), a norepinephrine analogue [56].
Surgical resection may be performed after embolization of Fibroma is the second most common tumor in infants and
feeding arteries to minimize bleeding. children. It has also been called cardiac fibromatosis or
fibrous or fibroelastic hamartoma to suggest that it is not a
true neoplasm [56, 65]. It is a solitary mass composed of col-
Rhabdomyoma lagen matrix, sometimes with elastic fibers [65]. In early
ages, there is predominant cellularity (fibroblasts), but in ado-
Rhabdomyoma is the most common cardiac neoplasm (75%) lescent and adults, it is predominantly acellular with only few
in children. 60–90% of these tumors are multiple [7, 69]. layers of peripheral cells [66]. Calcification is common but
There is a strong association with tuberous sclerosis complex hemorrhage, necrosis, and cysts are rare. It is often diagnosed
(TSC), with 80% incidence of TSC in cardiac rhabdomyo- in the first year of life with symptoms depending on the size
mas and 50% of patients with TSC having rhabdomyomas and location, including obstruction, valvular dysfunction, or
[7, 66]. Rhabdomyoma is considered a hamartoma and is arrhythmias that can lead to syncope or sudden cardiac death
well circumscribed but without a capsule and contains myo- [66]. It is more common in the left ventricle, commonly
cytes with disproportionate size and vacuoles that give the located in the LV free wall, ventricular septum, and RV wall
classic “spider cell” appearance [65]. Symptoms depend on [70], with the atrium rarely affected, particularly in patients
the location and size of the tumor. It involves the ventricles with syndromes such as familial adenomatous polyposis,
and atria, equally on right and left. A common feature of this Gardner, and Gorlin syndromes. On CT, it is seen as an intra-
tumor is spontaneous regression [7], as a result of which sur- myocardial, homogeneous iso or hypoattenuating lesion with
gery is not required, unless it is justified by clinical variable margins (Fig.  38.16). Central calcification may be
­manifestations [65] such as arrhythmia or cardiac failure; seen in 25% of the cases [66]. Contrast enhancement is vari-
otherwise serial imaging follow-up is recommended. For the able, ranging from no enhancement to heterogeneous
same reason, this lesion is not seen in adults. On CT, the enhancement, often delayed [10, 56]. Fibroma is ­distinguished
appearance is that of the normal myocardium, both in pre from rhabdomyoma since it is usually solitary and has calci-
and post contrast. Hence, small lesions without mass effect fication, low attenuation, and delayed enhancement. Surgical
can be difficult to visualize. Unlike fibroma, rhabdomyomas resection is performed in symptomatic cases to prevent sud-
do not calcify.

a b

Fig. 38.16  Fibroma. (a) Four-chamber CT reconstruction shows a Small calcifications are noted near the anterior surface of the tumor
large hypoattenuating mass with central calcification in a child (arrow), (arrow). The mid segment of LAD is surrounded by the tumor, showing
consistent with a fibroma. (b) Sagittal reconstructed CT image in the the appearance of a myocardial bridge (arrowhead)
same patient shows the large fibroma in the interventricular septum.
464 H. Goerne and P. Rajiah

den cardiac death. There is no spontaneous regression and no cers; Wilms tumor; pheochromocytoma, endometrial, IVC
reports of recurrence even after partial resection [65]. leiomyosarcoma (Fig.  38.17e), invasive uterine leiomyoma
(Fig.  38.17f)) or pulmonary vein (lung cancer). Renal cell
carcinoma is the most common lesion in transvenous spread,
Less Common Benign Neoplasms often involving the right kidney [75], with 10% of the RCC
with IVC extension reaching the right atrium [75]. The CT
Less common benign tumors described in the literature features of metastasis depend on the type of primary neo-
include teratoma, lymphangioma, inflammatory myofibro- plasm and route of extension. CT not only provides evalua-
blastic tumor, mesothelioma of the atrioventricular node, and tion of the heart and pericardium but also provides evaluation
Purkinje cell hamartoma (histiocytoid cardiomyopathy). of the entire chest including assessment of lungs, mediasti-
Some of these neoplasms are less common in the heart but num, bones, and soft tissue.
more common in the mediastinum, such as teratoma and
lymphangioma. Cardiac teratoma is located in the pericar-
dium (90%) [3] and has heterogeneous attenuation due to Sarcoma
different components.
Primary cardiac malignancies are rare and represent 25% of
primary cardiac neoplasms, with most of them being sarcomas.
Malignant Neoplasms Angiosarcoma is the most common subtype, with others being
rhabdomyosarcoma, undifferentiated sarcoma (Fig.  38.18a),
Metastasis myxofibrosarcoma, osteosarcoma, leiomyosarcoma, and lipo-
sarcoma [10]. Rhabdomyosarcoma is the most common pedi-
Metastasis is more common than primary neoplasms of the atric cardiac primary malignancy, usually the embryonic type.
heart. Cardiac metastasis is present in about 10–12% in It is more common in adults, where it is of pleomorphic type,
autopsies with known malignancy [71]. Routes of spread to whereas in children, embryonic type is more common.
the heart include direct local spread and lymphatic, hematog- Angiosarcoma typically arises in the right atrium, whereas
enous, and transvenous routes. Lung cancer represents the osteosarcoma typically arises in the left atrium [76]. Sarcoma
most common primary neoplasm in patients with cardiac/ has infiltrative nature with lobulated borders, broad base, and
pericardial metastasis, accounting for 37% of these patients heterogeneous enhancement with necrosis and hemorrhage.
[72], due to its proximity with the heart [73]. Other primary Calcifications/ossifications are often seen in osteosarcoma
neoplasms to involve the heart are nonsolid malignancies (Fig. 38.18b). Liposarcoma has fatty attenuation and soft tissue
(Kaposi’s sarcoma, leukemia) (20%), breast (7%), esopha- components. Extension of tumor and infiltration of adjacent
gus, (6%) and skin (4.5%) [72]. Direct extension occurs from structures are vital for surgical planning, which CT provides.
lung (Fig.  38.17a), esophageal, and other mediastinal neo- Multiple lesions may be seen, and pulmonary metastasis is
plasms (Fig. 38.17b). Pericardium is the most frequent loca- often present. Total resection is not possible in most of the
tion of metastatic cardiac involvement due to rich lymphatics cases, so surgery is just palliative.
in visceral pericardium and manifests as pericardial effusion,
thickening, irregularity, and nodules. Pericardial effusion is
often the earliest finding, but it is difficult to diagnose a Mesothelioma
malignant effusion based on CT, and hence pericardiocente-
sis is often required [73]. Hematogenous extension through Mesothelioma of the pericardium is a rare tumor of pericardial
coronary arteries or veins often manifests with myocardial mesothelial cells, and its association with prior asbestos expo-
nodules or infiltration (Fig. 38.17c). This is usually associ- sure is not clear, with only few cases with prior exposure [77].
ated with systemic metastases particularly lungs. Melanoma There are three histological types: epithelial, spindle cell, and
spreads to the heart in 64% of cases in an autopsy series [74]. mixed [78]. The appearance of this tumor is very similar to peri-
Transvenous extension to the heart can occur either through cardial metastasis with thickening/irregularity, diffuse plaque,
the SVC (lung, lymphoma, thyroid cancers), IVC (renal solitary nodule (Fig. 38.19), or multiple masses. Calcifications
(Fig.  38.17d), hepatic, adrenal cortex, or endometrial can- may be seen. Contrast enhancement is usually present.
38  CT of Cardiac and Paracardiac Masses 465

a b

c d

Fig. 38.17  Metastasis. (a) Axial CT image shows a bronchogenic neo- h­ eterogeneous mass (arrow) in the inferior vena cava, which was an
plasm (arrow), which is encasing a pulmonary artery (arrowhead) and extension of renal cell carcinoma that invades the right atrium and the
infiltrating the lateral wall of the left ventricle. (b) Axial CT scan in inlet of the right ventricle. (e) Coronal CT reconstruction in another
another patient with a metastatic clear cell sarcoma shows a large het- patient shows a large invasive intravascular tumor in the infrarenal seg-
erogeneous and lobulated mass (arrow) compressing the right superior ment of IVC (arrow) that grows cranially until the IVC-right atrium
pulmonary vein and right pulmonary artery, with infiltration of the peri- junction. The mass shows heterogeneous enhancement. This is an intra-
cardium and epicardial fat pad. Note the moderate pericardial effusion vascular leiomyosarcoma. (f) Coronal CT reconstruction shows a long
(arrowhead) and left lung nodules, which also represent metastases. (c) tumor protruding into the right atrium from the inferior vena cava. Note
Four-chamber CT scan shows a well-defined intracavitary mass in the the large and tortuous vessels (arrow) into the mass resembling an
right ventricle (arrow), consistent with metastasis in a patient with umbilical cord. This is a case of invasive leiomyoma from the uterus
known breast cancer. (d) Coronal CT reconstruction shows a large
466 H. Goerne and P. Rajiah

e f

Fig. 38.17 (continued)

a b

Fig. 38.18 Sarcoma. (a) Axial CT scan shows a large lobulated tumor shows a densely ossified mass which is infiltrating into the right supe-
in the left atrium with invasion of posterior wall (arrow), which was rior pulmonary vein (arrow), consistent with an osteosarcoma of the left
proven to be an intimal sarcoma. (b) Axial CT scan in another patient atrium
38  CT of Cardiac and Paracardiac Masses 467

Lymphoma

Lymphomatous involvement of the heart is usually second-

ary to systemic disease. Primary lymphoma of the heart is
extremely rare, accounting to 1.3% of primary cardiac neo-
plasms [10]. These primary lymphomas are usually non-­
Hodgkin B-cell type with aggressive behavior [10], usually
associated with HIV/AIDS. Right atrium is the cavity that is
most commonly involved. Pericardium, subepicardial fat,
and atrioventricular groove are also affected. The tumor is
ill-defined, diffuse, and infiltrative (Fig.  38.20a), and the
spread is usually through the epicardial surface [79] or a
large focal mass or multiple nodules with heterogeneous
contrast enhancement. Pericardial effusion is a common fea-
ture. Although lymphoma can involve the coronary arteries,
it typically encases the vessels without compression [80].
Leukemia is seen as diffuse infiltration (Fig. 38.20b).

Fig. 38.19  Mesothelioma. Axial CT scan shows a large pericardial

mass (arrow) adjacent to the right atrioventricular groove with mass
effect and slight displacement of the right chambers consistent with
pericardial mesothelioma

a b

Fig. 38.20 Lymphoma. (a) Axial CT scan shows a large irregular para- tenuating intramyocardial masses in the apical-septal (arrow) and
cardiac mass (arrow) that infiltrates the posterior wall of the right mid-lateral (arrowhead) walls, related to cardiac involvement in a
atrium, with minimal enhancement, which was proven to be a lym- patient with known leukemia
phoma. (b) Axial CT scan in another patient shows multiple hypoat-
468 H. Goerne and P. Rajiah

Conclusion 17. Schneider B, Hofmann T, Justen MH, Meinertz T. Chiari’s network:

normal anatomic variant or risk factor for arterial embolic events? J
Am Coll Cardiol. 1995;26(1):203–10.
CT is a valuable modality in the evaluation of cardiac masses 18. Islam AKMM, Sayami LA, Zaman S. Chiari network: a case report
and provides excellent information with regard to the location, and brief overview. J Saudi Heart Assoc. 2013;25(3):225–9.
extension, shape, size, attachment, and compromise of adjacent 19. Hightower JS, Taylor AG, Ursell PC, LaBerge JM. The Chiari net-
work: a rare cause of intracardiac guide wire entrapment. J Vasc
structures. CT is primarily used in the evaluation of cardiac Interv Radiol. 2015;26(4):604–6.
masses in patients who cannot have MRI scan due to contrain- 20. Lodhi AM, Nguyen T, Bianco C, Movahed A.  Coumadin ridge:
dications. It is also useful in specific situations where MRI or an incidental finding of a left atrial pseudotumor on transthoracic
echocardiography cannot provide optimal information, for echocardiography. World J Clin Cases. 2015;3(9):831–4.
21. Kim DT, Lai AC, Hwang C, et al. The ligament of Marshall: a struc-
instance, the presence of calcification or evaluation of feeding tural analysis in human hearts with implications for atrial arrhyth-
arteries. CT can also provide functional dynamic information. mias. J Am Coll Cardiol. 2000;36(4):1324–7.
22. Terpenning S, White CS.  Imaging pitfalls, normal anatomy, and
anatomical variants that can simulate disease on cardiac imaging as
References demonstrated on multidetector computed tomography. Acta Radiol
Short Rep. 2015;4(1):1–15.
1. Burke A, Virmani R.  Tumors of the heart and great vessels. 23. Kim DH, Choi SI, Choi J-A, et al. Various findings of cardiac thrombi
Washington, DC: Amer Registry of Pathology; 1996. on MDCT and MRI. J Comput Assist Tomogr. 2006;30(4):572–7.
2. Lam KY, Dickens P, Chan AC. Tumors of the heart. A 20-year expe- 24. Bittencourt MS, Achenbach S, Marwan M, et  al. Left ventricular
rience with a review of 12,485 consecutive autopsies. Arch Pathol thrombus attenuation characterization in cardiac computed tomog-
Lab Med. 1993;117(10):1027–31. raphy angiography. J Cardiovasc Comput Tomogr. 2012;6(2):121–6.
3. Kassop D, Donovan MS, Cheezum MK, et al. Cardiac masses on 25. Hong YJ, Hur J, Kim YJ, et al. Dual-energy cardiac computed tomogra-
cardiac CT: a review. Curr Cardiovasc Imaging Rep [Internet]. phy for differentiating cardiac myxoma from thrombus. Int J Cardiovasc
2014;7(8):9281. Available from: http://www.ncbi.nlm.nih.gov/ Imaging. Springer Netherlands. 2014;30(Suppl 2(2)):121–8.
pmc/articles/PMC4090749/pdf/12410_2014_Article_9281.pdf. 26. Patel A, Au E, Donegan K, et  al. Multidetector row computed
4. Hoey E, Ganeshan A, Nader K, Randhawa K, Watkin R. Cardiac tomography for identification of left atrial appendage filling defects
neoplasms and pseudotumors: imaging findings on multidetector in patients undergoing pulmonary vein isolation for treatment of
CT angiography. Diagn Interv Radiol. 2012;18(1):67–77. atrial fibrillation: comparison with transesophageal echocardiogra-
5. Hoey ET, Mankad K, Puppala S, Gopalan D, Sivananthan MU. MRI phy. Heart Rhythm. 2008;5(2):253–60.
and CT appearances of cardiac tumours in adults. Clin Radiol. 27. Hur J, Kim YJ, Lee H-J, et  al. Left atrial appendage thrombi in
2009;64(12):1214–30. stroke patients: detection with two-phase cardiac CT angiog-
6. Kumar N, Agarwal S, Ahuja A, Das P, Airon B, Ray R. Spectrum of raphy versus transesophageal echocardiography 1. Radiology.
cardiac tumors excluding myxoma: experience of a tertiary center 2009;251(3):683–90.
with review of the literature. Pathol Res Pract. 2011;207(12):769–74. 28. Czekajska-Chehab E, Tomaszewska M, Olchowik G, Tomaszewski
7. Tao TY, Yahyavi-Firouz-Abadi N, Singh GK, Bhalla S.  Pediatric M, Adamczyk P, Drop A.  Lipomatous hypertrophy of the inter-
cardiac tumors: clinical and imaging features. Radiographics. atrial septum in ECG-gated multislice computed tomography of the
Radiological Society of North America. 2014;34(4):1031–46. heart. Med Sci Monit. 2012;18(7):MT54–9.
8. Beghetti M, Gow RM, Haney I, Mawson J, Williams WG, Freedom 29. Fan C-M, Fischman AJ, Kwek BH, Abbara S, Aquino SL.

RM.  Pediatric primary benign cardiac tumors: a 15-year review. Lipomatous hypertrophy of the interatrial septum: increased uptake
Am Heart J. 1998;134(6):1107–14. on FDG PET. AJR Am J Roentgenol [Internet]. 2005;184(1):339–
9. Motwani M, Kidambi A, Herzog BA, Uddin A, Greenwood JP, 342. Available from: http://www.ajronline.org/doi/pdfplus/10.2214/
Plein S.  MR imaging of cardiac tumors and masses: a review of ajr.184.1.01840339.
methods and clinical applications. Radiology. 2013;268(1):26–43. 30. Anderson RH, Anderson RH, Webb S, Brown NA, Webb S, Brown
10. Rajiah P, Kanne JP, Kalahasti V, Schoenhagen P. Computed tomog- NA.  Clinical anatomy of the atrial septum with reference to its
raphy of cardiac and pericardiac masses. J Cardiovasc Comput developmental components. Clin Anat. 1999;12(5):362–74.
Tomogr. 2011;5(1):16–29. 31. Sánchez-Quintana D, Ho SY, Cabrera JA, Farré J, Anderson

11. Anavekar NS, Bonnichsen CR, Foley TA, et al. Computed tomog- RH. Topographic anatomy of the inferior pyramidal space: relevance
raphy of cardiac pseudotumors and neoplasms. Radiol Clin N Am. to radiofrequency catheter ablation. J Cardiovasc Electrophysiol.
2010;48(4):799–816. 2001;12(2):210–7.
12. Akcay M, Bilen ES, Bilge M, Durmaz T, Kurt M. Prominent crista 32. Laura DM, Donnino R, Kim EE, Benenstein R, Freedberg RS,
terminalis: as an anatomic structure leading to atrial ­arrhythmias Saric M.  Lipomatous atrial septal hypertrophy: a review of its
and mimicking right atrial mass. J Am Soc Echocardiogr. anatomy, pathophysiology, multimodality imaging, and rel-
2007;20(2):197.e9–e10. evance to percutaneous interventions. J Am Soc Echocardiogr.
13. Salustri A, Bakir S, Sana A, Lange P, Mahmeed Al WA. Prominent 2016;29(8):717–23.
crista terminalis mimicking a right atrial mass: case report. 33. Heyer CM, Kagel T, Lemburg SP, Bauer TT, Nicolas V. Lipomatous
Cardiovasc Ultrasound. 6 ed. BioMed Central. 2010;8(1):47. hypertrophy of the interatrial septum: a prospective study of
14. Gad A, Mannan J, Chhabra M, Zhang XXY, Narula P, Hoang incidence, imaging findings, and clinical symptoms. Chest.
D. Prominent Eustachian valve in newborns: a report of four cases. 2003;124(6):2068–73.
AJP Rep. 2016;6(1):e33–7. 34. O’Connor S, Recavarren R, Nichols LC, Parwani AV. Lipomatous
15. Watson T, Kakar P, Srivastava S, Dhanjal TS. Eustachian valve rem- hypertrophy of the interatrial septum: an overview. Arch Pathol Lab
nant. Cardiol J. 2007;14(5):508–9. Med. 2006;130(3):397–9.
16. Yavuz T, Nazli C, Kinay O, Kutsal A. Giant Eustachian valve with 35. Xanthos T, Giannakopoulos N, Papadimitriou L.  Lipomatous

echocardiographic appearance of divided right atrium. Tex Heart hypertrophy of the interatrial septum: a pathological and clinical
Inst J. 2002;29(4):336–8. approach. Int J Cardiol. 2007;121(1):4–8.
38  CT of Cardiac and Paracardiac Masses 469

36. Pomerance A. Pathological and clinical study of calcification of the phy and other imaging modalities. Heart. BMJ Publishing Group
mitral valve ring. J Clin Pathol. 1970;23(4):354–61. Ltd and British Cardiovascular Society. 2011;97(13):1101–9.
37. Shriki J, Rongey C, Ghosh B, et  al. Caseous mitral annular cal- 60. Burke A, Jeudy J, Virmani R. Cardiac tumours: an update. Heart.
cifications: multimodality imaging characteristics. World J Radiol. 2008;94(1):117–23.
2010;2(4):143–7. 61. Barbier G, Vazquez Figueroa JG, Rinehart S, et al. Tissue character-
38. Harpaz D, Auerbach I, Vered Z, Motro M, Tobar A, Rosenblatt ization of a papillary fibroelastoma on the aortic valve by contrast-­
S. Caseous calcification of the mitral annulus: a neglected, unrec- enhanced 320-detector row computed tomography. J Cardiovasc
ognized diagnosis. J Am Soc Echocardiogr. 2001;14(8):825–31. Comput Tomogr. 2010;4(5):345–7.
39. Elgendy IY, Conti CR. Caseous calcification of the mitral annulus: a 62. Mariscalco G, Bruno VD, Borsani P, Dominici C, Sala A. Papillary
review. Clin Cardiol. Wiley Periodicals, Inc. 2013;36(10):E27–31. fibroelastoma: insight to a primary cardiac valve tumor. J Card
40. Gilbert HM, Grodman R, Chung MH, Hartman G, Krieger KH, Surg. 2010;25(2):198–205.
Hartman BJ.  Sterile, Caseous mitral valve “abscess” mimicking 63. Daveron E, Jain N, Kelley GP, et  al. Papillary fibroelastoma and
infective endocarditis. Clin Infect Dis. 1997;24(5):1015–6. Lambl's excrescences: echocardiographic diagnosis and dif-
41. Wang ZJ, Reddy GP, Gotway MB, Yeh BM, Hetts SW, Higgins ferential diagnosis. Echocardiography. Blackwell Science Inc.
CB.  CT and MR imaging of pericardial disease. Radiographics. 2005;22(5):461–3.
2003;23(suppl_1):S167–80. 64.
Hrabak-Paar M, Hübner M, Stern-Padovan R, Lušić
42. Kim JH, Goo JM, Lee HJ, et al. Cystic tumors in the anterior medi- M. Hemangioma of the interatrial septum: CT and MRI features.
astinum. J Comput Assist Tomogr. 2003;27(5):714–23. Cardiovasc Intervent Radiol. 2011;34(Suppl 2(S2)):S90–3.
43. Jeung M-Y, Gasser B, Gangi A, et al. Imaging of cystic masses of 65. Becker AE.  Primary heart tumors in the pediatric age group: a
the mediastinum. Radiographics. 2002;22(suppl_1):S79–93. review of salient pathologic features relevant for clinicians. Pediatr
44. Rajiah P, Kanne JP. Computed tomography of the pericardium and Cardiol. Springer-Verlag. 2000;21(4):317–23.
pericardial disease. J Cardiovasc Comput Tomogr. 2010;4(1):3–18. 66. Burke A, Virmani R.  Pediatric heart tumors. Cardiovasc Pathol.
45. McAdams HP, Kirejczyk WM, Rosado-de-Christenson ML,
2008;17(4):193–8.
Matsumoto S.  Bronchogenic cyst: imaging features with clinical 67. Esmaeilzadeh M, Jalalian R, Maleki M, Givtaj N, Mozaffari K,
and histopathologic correlation. Radiology. 2000;217(2):441–6. Parsaee M.  Cardiac cavernous hemangioma. Eur J Echocardiogr.
46. Seo N, Kang J-W, Lim C-H, Kim B, Lee HJ, Lim T-H. CT findings The Oxford University Press. 2007;8(6):487–9.
of an intracardiac bronchogenic cyst. Int J Cardiovasc Imaging. 68. Wang J-G, Han J, Jiang T, Li Y-J. Cardiac paragangliomas. J Card
2011;27(5):701–4. Surg. 2015;30(1):55–60.
47. Forcillo J, Dion D, Sauvageot C, Jeanmart H.  Intraventricular
69. Sparrow PJ, Kurian JB, Jones TR, Sivananthan MU. MR imaging
bronchogenic cyst: a rare congenital anomaly. Ann Thorac Surg. of cardiac tumors. Radiographics. 2005;25(5):1255–76.
2015;100(3):1101–3. 70. Beroukhim RS, Prakash A, Buechel ERV, et  al. Characterization
48. Wang J, Zhu Q, Liang B, Shi H, Han P, Kong X. Left ventricular of cardiac tumors in children by cardiovascular magnetic reso-
bronchogenic cyst. Ann Thorac Surg. 2016;101(2):744–6. nance imaging: a multicenter experience. J Am Coll Cardiol.
49. Dean C, Etienne D, Carpentier B, Gielecki J, Tubbs RS, Loukas 2011;58(10):1044–54.
M. Hiatal hernias. Surg Radiol Anat. 2012;34(4):291–9. 71. Abraham KP, Reddy V, Gattuso P.  Neoplasms metastatic to the
50. Díaz-Zamudio M, Bacilio-Pérez U, Herrera-Zarza MC, et  al.
heart: review of 3314 consecutive autopsies. Am J Cardiovasc
Coronary artery aneurysms and ectasia: role of coronary CT angi- Pathol. 1990;3(3):195–8.
ography. Radiographics. 2009;29(7):1939–54. 72. Klatt EC, Heitz DR. Cardiac metastases. Cancer. 1990;65(6):1456–9.
51. Frazier AA, Qureshi F, Read KM, Gilkeson RC, Poston RS, White 73. Chiles C, Woodard PK, Gutierrez FR, Link KM.  Metastatic

CS.  Coronary artery bypass grafts: assessment with multidetec- involvement of the heart and pericardium: CT and MR imag-
tor CT in the early and late postoperative settings. Radiographics. ing. Radiographics. Radiological Society of North America.
2005;25(4):881–96. 2001;21(2):439–49.
52. Rajiah P, Kanne JP.  Computed tomography of septal defects. J 74. Glancy DL, Roberts WC.  The heart in malignant melanoma. A
Cardiovasc Comput Tomogr. 2010;4(4):231–45. study of 70 autopsy cases. Am J Cardiol. 1968;21(4):555–71.
53. Murdoch DR, Corey GR, Hoen B, et al. Clinical presentation, eti- 75. Oto A, Herts BR, Remer EM, Novick AC.  Inferior vena cava
ology, and outcome of infective endocarditis in the 21st century: tumor thrombus in renal cell carcinoma: staging by MR imag-
the International Collaboration on Endocarditis-Prospective Cohort ing and impact on surgical treatment. AJR Am J Roentgenol.
Study. Arch Intern Med. 2009;169(5):463–73. 1998;171(6):1619–24.
54. Murillo H, Restrepo CS, Marmol-Velez JA, et  al. Infectious dis- 76. Agaimy A, Rösch J, Weyand M, Strecker T. Primary and metastatic
eases of the heart: pathophysiology, clinical and imaging overview. cardiac sarcomas: a 12-year experience at a German heart center.
Radiographics [Internet]. 2016;36(4):963–983. Available from: Int J Clin Exp Pathol. 2012;5(9):928–38.
http://pubs.rsna.org/doi/pdf/10.1148/rg.2016150225. 77. Thomason R, Schlegel W, Lucca M, Cummings S, Lee S. Primary
55. Grebenc ML, Rosado-de-Christenson ML, Burke AP, Green
malignant mesothelioma of the pericardium. Case report
CE, Galvin JR.  From the archives of the AFIP.  Radiographics. and literature review. Tex Heart Inst J.  Texas Heart Institute.
2002;22(3):673–89. 1994;21(2):170–4.
56. Araoz PA, Araoz PA, Mulvagh SL, et al. CT and MR imaging of 78. Suman S, Schofield P, Large S.  Primary pericardial mesothe-

benign primary cardiac neoplasms with echocardiographic correla- lioma presenting as pericardial constriction: a case report. Heart.
tion. Radiographics. 2000;20(5):1303–19. 2004;90(1):e4.
57. Swartz MF, Lutz CJ, Chandan VS, Landas S, Fink GW. Atrial myx- 79. Jeudy J, Kirsch J, Tavora F, et  al. From the radiologic pathology
omas: pathologic types, tumor location, and presenting symptoms. archives: cardiac lymphoma: radiologic-pathologic correlation.
J Card Surg. Blackwell Publishing Inc. 2006;21(4):435–40. Radiographics. 2012;32(5):1369–80.
58. Amano J, Kono T, Wada Y, et al. Cardiac myxoma: its origin and tumor 80. Shah RN, Simmons TW, Carr JJ, Entrikin DW.  Primary car-

characteristics. Ann Thorac Cardiovasc Surg. 2003;9(4):215–21. diac lymphoma diagnosed by multiphase-gated cardiac CT and
59. Auger D, Pressacco J, Marcotte F, Tremblay A, Dore A, Ducharme CT-guided percutaneous trans-sternal biopsy. J Cardiovasc Comput
A.  Cardiac masses: an integrative approach using echocardiogra- Tomogr. 2012;6(2):137–9.
 Valvular Heart Disease and Prostheses
39
Gudrun M. Feuchtner

 alvular Imaging by Cardiac CTA:

V cardiac cycle from R wave to R wave is fundamental for
From Dream to Reality dynamic imaging. In order to obtain systolic and diastolic
views of the valves and 4D cine loops of their function,
Over decades, imaging of cardiac valves has been the domain image reconstruction of axial thin-slice source images (of
of echocardiography [1], while the “shades-of-gray” 0.75–1 mm) at either 5 or 10% increments of the RR interval
obscured computed tomography (CT). From 2005 onward, is recommended.
multislice CT technology improved in terms of higher tem-
poral and spatial resolution, along with the advent of ECG
gating, which has made a dream come true: functional 4D Retrospective ECG Gating
cine imaging of cardiac valves. While the first steps with
16-slice CT in 2005 were limited to patients with low heart Retrospective ECG gating is the technique of first choice.
rates, nowadays, scanners with highest temporal resolution Some vendors provide dedicated ECG-gating technology,
of >75 ms (in 2009, the second-generation 128-slice and in such as for the “dual-source” CT (DSCT). Functional evalu-
2014 the third-generation 296-slice dual-source CT) have ation of valves [2] with prospective ECG triggering at a
approached a temporal resolution to that of echocardiogra- reduced mean radiation dose of 3.8 mSv is feasible. A DSCT
phy and created new horizons in the dynamic assessment of scanner consists of two X-ray tubes operating at the same
cardiac valves and prosthetic heart valves (PHV), which are time. While one tube is turned “on” (full mAs, 100%), the
highlighted in this chapter. second tube covers the entire cardiac cycle at 20% reduced
Further, explanations of how to examine cardiac valves by tube mA, enabling the reconstruction of multiphase CT
computed tomography (CTA) and how to tailor a cardiac CT image datasets. The first tube (100% mAs) can be pulsed into
scan protocols to cardiac valves are provided. Based on cur- any arbitrary phase of the cardiac cycle, either during the
rent scientific evidence, the integration of CTA into the clini- end-diastolic (70% of RR interval) in low heart rates <65 bpm
cal work-up of patients in conjunction with other imaging or end-­systolic phase (40% of RR interval) in patients with
modalities such as echocardiography or 18flourdeoxyglucose high heart rates >65 bpm), in order to ensure motion artifact-­
(FDG) positron emission tomography (PET)/CT is discussed. free visualization of coronary arteries. More recently intro-
duced CT systems are capable of “single-beat” scans either
“by using high-pitch” (pitch, 3.2–3.4) DSCT technology, or
Technical CT Prerequisites by “volume CT” approach: Equipped with a broad detector
width of up to 16 cm, the entire volume of the heart is cov-
For the evaluation of cardiac valve function and anatomy, ered with one rotation during one heartbeat, enabling func-
64-slice CT or more advanced technology such as dual-­ tional assessment of cardiac valves during one cardiac cycle.
source CT or volume CT scanners equipped with up to 296–
320- slices is required. Furthermore, ECG gating including
the acquisition of a multiphase CT dataset covering the entire Iodine Contrast Media Injection

For visualization of all four cardiac valves (right and left

G. M. Feuchtner (*) sided), a “biphasic split” contrast agent injection protocol is
Department of Radiology, Innsbruck Medical University, recommended [3]. A monophasic contrast agent bolus
Innsbruck, Austria
­injection at a high flow rate of 4–6 mL/s, which is commonly
e-mail: [email protected]

© Humana Press 2019 471

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_39
472 G. M. Feuchtner

used for coronary CT angiography, typically results in lack (Fig. 39.1). Open surgical aortic valve replacement (AVR) is
of right chamber enhancement (due to significant “wash- the only effective treatment in cases of severe aortic stenosis.
out”) without appropriate delineation of the tricuspid and With up to 40% of patients being declined for conventional
pulmonary valve. In contrast, by using a “biphasic split pro- open-heart AVR due to an elevated risk profile, the new
tocol,” homogenous right ventricular and atrial enhancement emerging transcatheter aortic valve implantation (TAVI) pro-
is feasible, due to a prolonged bolus transit time. The “bipha- cedure provides an efficient minimally invasive treatment
sic split protocol” consists of a contrast agent bolus at high option with excellent long-term survival and outcome results.
flow rate (5–6 cc/s), followed by a second bolus with a lower Cardiac and aortoiliacal CTA is the modality of choice for
flow rate (3.0–4 cmL/s) [3] (volume 80% high flow versus planning of TAVI, in order to define the appropriate trans-
20% low flow). The second bolus may be mixed with saline catheter access route (transaortal, transfermoral, transaxil-
solution or followed by a saline chaser. lary, or transapical) and to determine the optimal prosthesis
The final administration of a saline solution optimizes bolus size and in order to avoid complications such as coronary
transit and geometry. A further advantage of a biphasic split pro- ostium overstenting or annulus rupture.
tocol is the reduction of streak artifacts within the right chambers Beyond planning of TAVI, cardiac CTA allows for sizing
due to lower injection flow rate during the second phase [3]. of the aortic valve orifice area (AVA) and grading of aortic
stenosis severity. For sizing of the AVA, the CTA images
must be reconstructed during mid-systole (5–25% of RR
Post-processing interval), and the “best phase” with the smallest AVA should
be selected (typically 25% of RR interval). Additionally, all
Dedicated advanced 3D post-processing workstations short-axis views of the aortic valve from the tip of the leaf-
equipped with multiplanar reformations (MPR) in three lets caudally toward the annulus should be reviewed, in
planes are required for the generation of valvular planes. order to select the smallest AVA at the level of the orifice
Thin-slice MPR (1  mm slice width) are advantageous over (Fig. 39.1).
thicker slices (>5  mm) and maximum intensity projections Cardiac CTA has shown a good correlation with other
(MIP) as well are suitable, in order to ensure best display and modalities such as TTE or invasive catheterization for mea-
highest resolution of thin structured such as valvular leaflets. suring the AVA [4–7], while a tendency towards AVA “over-
The following standardized planes for valves should be gen- estimation” by CTA as compared to transthoracic
erated: for the aortic valve, left sagittal oblique, left coronal echocardiography (TTE) has been observed. This can be
oblique, and orthogonal cross-sectional axial oblique views (see explained by inter-modality differences: CTA provides direct
figures in following sections) through the valve and, for the anatomic AVA sizing, while TTE estimates the effective AVA
mitral valve, four-chamber, three-chamber, and two-chamber based on transvalvular pressure gradient (VTI, velocity time
views, as well as perpendicular short-axis mitral valve views. integral) assuming left ventricular outflow tract (LVOT) cir-
cularity. However, LVOT is most commonly elliptic, thus
explaining the inter-modality differences.
3D Volume Rendering Technique, MIP, and 4D In practice, high-grade high-flow “classic” aortic stenosis
(AS) is robustly diagnosed by TTE based on increased trans-
Three-dimensional volume rendering technique (VRT) valvular flow and velocity. However, in patients with “low-­
allows for the visualization of valvular leaflets, calcifications flow-­low-gradient” (paradoxical and classic, with preserved
(Fig.  39.1), and heart valve prostheses. Maximum intensity and impaired left ventricular ejection fraction) aortic steno-
projections (MIP) are less useful to fully visualize valves, sis, diagnosis by TTE is more difficult and typically requires
since leaflets are thin. In such instances, the full spatial reso- stress echocardiography for affirmation. In those patients,
lution of thin-slice MPR are advantageous. Finally, the gen- anatomic sizing by AVA by cardiac CTA is a useful “true”
eration of 4D cine imaging video loops during the entire anatomic parameter for estimation of the true severity of
cardiac cycle allows for evaluation of leaflet function and, in AS. Notably, in patients in whom TAVI is considered, CTA
particular, prosthesis malfunction. provides the advantage of concomitant aortic root and annu-
lus sizing for TAVI planning and also provides an evaluation
of coronary artery disease (CAD).
The Aortic Valve Furthermore, CTA allows for quantification of the aortic
valve calcification (AVC) score (Agatston, volume, and mass
Aortic Stenosis score) on native ECG-gated unenhanced CT scan, as stan-
dardized performed for coronary artery calcium scores
Degenerative aortic stenosis (AS) is common in the elderly (CACS). The AVC score is a valuable prognostic marker,
population with a prevalence of >2% and carries a poor prog- which predicts outcome in patients with severe asymptom-
nosis with a 1- and 5-year survival of 60 and 32%, respec- atic AS [8] and assists decision-making in terms of whether
tively. Its hallmark is severe aortic valve calcifications to treat patients with surgical AVR or whether to postpone
39  Valvular Heart Disease and Prostheses 473

Fig. 39.1  Aortic stenosis

(AS). An 82-year-old female
a b
with severe high grade AS III,
severely calcified leaflets, and
narrowing of the inner orifice
area (a, b, white arrows)
during mid-systolic phase,
tricuspid aortic valve
(“Mercedes-star-like”) (a)
with partial degenerative
fusion of the left/right
coronary cusp. Aortic valve
by 3D volume rendering
technique (VRT) (a, b).
Multiplanar reformation
(MPR) (c–d) of the aortic
valve: C = left sagittal
oblique (three-chamber
view), D = axial oblique view
of aortic valve, and E = left
coronal oblique view. Image
D allows for sizing of the c d
inner aortic valve orifice area
(AVA) (black arrows denote
severely stenosed AVA during
mid-systole, 25% of the RR
interval). Black lines (c, e)
indicate the level of
cross-­sectional view of aortic
valve (d)

AVR and recommend optimal medical therapy. Both asymp- mal tricuspid valves. During systolic phase, bicuspid valves
tomatic severe AS and low-flow-low-gradient AS [9] are are characterized by a “fish-mouth”-like opening of leaflets,
considered as “Class IIA” indications for surgery [1]. with a round or oval shape (Fig. 39.2). Bicuspid valves are
Accordingly, the decision whether or not to perform conven- classified as either “type 0” (no raphe) or “type 1” (with
tional AVR or TAVI is usually influenced by a variety of raphe, 85%), with either fusion of left and right (L/R) (most
prognostic markers, including the AVC score. common), right and non-coronary (R/N), or left and non-­
coronary (L/N) cusps. During diastole, the “linear sign”
(Fig. 39.2) is pathognomonic for bicuspid valves, indicating
Aortic Valve Morphology a more-or-less symmetrical appearance, while minor devia-
tions from absolute linearity are natural. A special designa-
The normal aortic valve consists of three leaflets (tricuspid), tion is reserved for “type 2,” the original tricuspid valve,
the non-, left, and right coronary cusps, which symmetrically which develops a “functional bicuspid” configuration of the
merge with a “Mercedes-star”-like appearance on axial slices leaflet opening due to secondary fusion of two leaflets of an
by cardiac CTA during the diastolic phase. originally tricuspid valve, due to degenerative aortic stenosis
The most common congenital abnormality is the “bicus- (most common) (Figs. 39.1a and 39.2a) or rheumatic disease
pid” valve type (Fig. 39.2), where only two of the three leaf- (less common).
lets are present. One of those leaflets may or may not show a For an accurate diagnosis of a bicuspid valve, cardiac
“raphe,” indicating the fusion site of two original leaflets. CTA yields a high sensitivity of 94% and a specificity of
The “raphe” type is more common (80%). Due to the smaller 100% [10], but functional CCTA datasets are required,
orifice and a tendency toward degeneration, bicuspid valve including a systolic phase, to define whether leaflet fusion or
malfunctions such as stenosis are more common than in nor- raphe is present or not (type 1 and type 2). In contrast, the
474 G. M. Feuchtner

Fig. 39.2  Bicuspid valve

types. (a) VRT with severe
a b
calcifications and
degeneration “linear sign”
during end-diastole, (a) axial
and (b) three-chamber view,
congenital type 0 (no raphe).
(c) “Linear sign” during
end-diastole (70% of RR
interval) (type 0) and B “fish
mouth” opening during
mid-systole (25%RR interval)
(type 0). Bicuspid valve
congenital type 0 (without
raphe, panel C/B/D/E) and
type 1 (with raphe, f). (d)
“Fish mouth” during
mid-systole, mild
degenerations, and fusion of
left/right coronary cusp (R/L)
during end-diastole (type 1 c d
R/L). (e) “Linear sign” with
severe calcification during
end-diastole with fusion of
non- (N) and right (R)
coronary cusp (type 0 N/R)

e f

bicuspid type 0 (no raphe) can be diagnosed during diastolic sural” or “unicommissural type,” consisting of one leaflet
phase based on the “linear sign” [11]. and a “hole-like” appearance, which often require surgery
In rare cases, a “quadricuspid” aortic valve may be pres- early after birth or in early adulthood due to a stenotic
ent, consisting of four leaflets with an estimated incidence of component.
0.003–0.043% [12]. Another extremely rare congenital mal-
formation is the “unicuspid valve,” either the “acommis-
39  Valvular Heart Disease and Prostheses 475

Fig. 39.3 Aortic
regurgitation (AR). A
a b
72-year-old male with
moderate AR. During
end-diastolic phase, small
central incomplete leaflet
adaption is shown (“poke
hole”-like) (a) on 3D VRT
(AB) and MPR (CDE). MPR
in left sagittal oblique
(three-chamber view) (c),
axial oblique (d), and left
coronal oblique (e)
reconstructions. Note
incomplete co-adaptation of
leaflets (c, e, d). A central
regurgitation orifice area
(black arrow, d) can be
visualized during end-diastole
(70% of RR interval). Black
lines (c) indicate the level of
cross-sectional oblique view
of the aortic valve (d)
c d

Aortic Regurgitation ROA is recommended on oblique axial cross-sectional slices

reformatted for aortic valve views. Mild and moderate AR
Diagnosis of aortic regurgitation (AR) is primarily estab- [13] may create only tiny “poke hole”-like central leakages
lished by TTE in clinical practice based on Doppler regurgi- (Fig. 39.3) which may be difficult to visualize and limit the
tant flow jet length and width and the pressure half-time accuracy of CTA [14].
(PHT) method. AR is further stratified into mild, moderate, The most common request for cardiac CTA in clinical
and severe AR by TTE. practice however is clarifying the etiology of AR such as aor-
In patients with AR, cardiac CTA allows for direct visual- tic root aneurysm sizing, the characterization of cardiac
ization of incomplete leaflet margins adaption, and a “central masses such as vegetations causing AR, or the preoperative
regurgitant orifice area” (ROA) during the end-diastolic evaluation of CAD.
phase may be depicted (Fig. 39.3). Such findings should be Furthermore, in patients with singular aortic valve dis-
included into a structured cardiac CTA reports. The anatomic ease and no other dysfunctional heart valves, the AR volume
ROA provides an estimate of the severity of AR; with (mL) and fraction (%) can be calculated as the difference
<25  mm2 ROA as cutoff for mild and >75  mm2 ROA for between left ventricular (LV) and right ventricular (RV)
severe AR [13]. CT image reconstruction is recommended stroke volume. Notably, the absence of any other valvular
between 60% and 80% of the RR interval; however, the car- dysfunction (mitral, pulmonary, or tricuspid) is an important
diac phase with least cardiac motion and optimal image qual- prerequisite for the accurate quantification of the AR vol-
ity should be used when determining the ROA. Sizing of the ume and fraction [15].
476 G. M. Feuchtner

Mitral Valve close to the annulus, patients may have an increased risk for
fatal injury via perforation and bleeding while inserting
Mitral Stenosis mitral annuloplasty devices during open-heart surgery or
during new investigative percutaneous techniques.
Mitral stenosis (MS) most commonly develops in the course
of rheumatic disease on the valvular level due to nodular
and/or diffuse thickening of the leaflets and chordae. Mitral Valve Prolapse
Transthoracic echocardiography (TTE) enables an accurate
functional diagnosis of MS and its severity (stratified into During mid-end-systolic phase (25–40% of RR interval),
mild, MS I; moderate, MS II; and severe, MS III) depending direct visualization of mitral valve prolapse (MVP) by CTA
on the increase in transvalvular flow velocity and pressure [19] is feasible, and the exact measurement of leaflet excur-
gradient. However, nodular and diffuse thickening and leaflet sion below the annulus is possible. A deviation of approxi-
calcifications may be found incidentally on coronary CTA mately >3  mm is usually regarded as “significant” MVP
exams and should be reported (Fig. 39.4). with functional relevance, causing mitral regurgitation. The
In addition, cardiac CT allows for sizing of the mitral ori- recommended views for the sizing of prolapse excursion
fice area (MOA) during the end-diastolic phase, which pro- are two- and three-chamber views. Two different types of
vides a reliable estimate of MS severity [16]. MVP are distinguished, the “bowing” (“billowing”) type,
In patients with secondary MS due to orifice obstruction which most commonly caused by primary myxomatous
and atrial myxoma prolapse, CTA and MRI are the cross-­ degenerative disease (“Barlow’s disease”) due to redundant
sectional imaging modalities of choice to further character- leaflets forming round-shaped lesions and the “flail leaflet”
ize masses after an initial echocardiography screening. Atrial type. A flail leaflet is characterized by free leaflet margin
myxoma, which typically appears as round-shaped prolapse and usually occurs along with ruptured chordae,
hypodense pedunculated masses arising from interatrial sep- commonly due to rheumatic or inflammatory disease, or
tum (e.g., fossa ovalis) with minor contrast uptake, may during acute myocardial infarct involving a papillary mus-
cause obstruction of the MOA during the diastolic phase. cle. The exact anatomic assignment of the prolapse to the
Prior to surgery, CTA allows for preoperative rule out of corresponding leaflet section (anterior A1, upper; A2, mid;
coronary artery disease and is able to accomplish two tasks and A3, caudal segment of the anterior leaflet or poste-
with a single scan. rior P1, upper; P2, mid; and P3, caudal part of the posterior
leaflet) is critical for planning surgical reconstruction.
While 3D TEE usually provides this information, TTE typi-
Mitral Regurgitation cally does provide sufficient image quality. Beyond exact
anatomic assignments, CTA also adds information about
Similar to aortic regurgitation, diagnosis of mitral regurgita- leaflet morphology such as whether calcifications are pres-
tion (MR) is firmly established by Doppler flow jet assess- ent or not (Fig. 39.5).
ment during TTE, and cardiac CTA does not play a role in
the primary diagnosis. Due to the saddle shape of mitral
valve with irregular leaflet margins, the direct anatomic visu- Degenerative Mitral Annular Calcification
alization of the mitral regurgitant orifice area (M-ROA) dur-
ing end-systole is more difficult compared to the aortic valve, CTA plays a major role in establishing diagnosis of degen-
though technically feasible with a good correlation to TTE as erative mitral annular calcification (MAC) based on its clear
shown in one study of 19 patients [17]. detection of calcified masses. On TTE, large inhomogenous
In patients with functional MR (FMR) due to heart failure caseous/liquefaction necrosis type of MAC may mimic
and LV enlargement, the anatomy of the subvalvular appara- tumors due to their mass-like appearance, with potential
tus and valve geometry is of interest for planning of novel invasion into periannular structures such as the myocardium.
minimally invasive or percutaneous mitral valve intervention In such patients, cardiac CT (both unenhanced and contrast-­
techniques for annuloplasty. The subvalvular apparatus is enhanced CTA) is the modality of choice for ascertaining
highly variable due to anatomic variations in the posterior diagnosis, based on lack of contrast agent uptake and high
papillary muscle (PM)  – both single and multiple heads  – CT densities (>130 HU) seen on unenhanced CT [20]
and insertions are often found. In patients with heart failure (Fig. 39.6).
and FMR, increased tenting heights, mitral valve sphericity
index, and more outward displacement of the PMs are found, Pulmonary and Tricuspid Valve
which are indicators of FMR severity [18]. Beyond, CTA The visualization of right-sided valves is a challenge, due to
allows for 3D visualization of left circumflex artery (CX) streak artifacts from contrast agent inflow. Such artifacts are
course in relation to the mitral annulus. If the CX is located minimized by using the biphasic contrast agent injection
39  Valvular Heart Disease and Prostheses 477

a b

c d

Fig. 39.4  Mitral stenosis/regurgitation, rheumatic disease. A 49-year-­ axis (b, d) views by MPR (a, b) and 3D VRT (c, d). Arrows (a, c)
old female with rheumatic disease and combined MS/MR dysfunction. denote mitral valve orifice = O. Pm, papillary muscle thickening result-
Diffuse A,  anterior and nodular; P,  posterior mitral leaflet thickening ing in leaflet retraction and restricted leaflet motion
(involving P2/3 segment) (arrows) on two-chamber (a, c) and short-­

protocol described above [3] and designed to ensure homog- “definite,” “possible,” or “rejected.” TTE and TEE are the
enous attenuation of right cardiac chambers. primary imaging tools during initial patient triage. In patients
with suspected paravalvular involvement and insufficient
Infective Endocarditis TTE/TEE imaging, cardiac CTA has been recently recog-
Diagnosis of infective endocarditis (IE) is made by applying nized as a “Class IIa” indication [1] by the American Heart
major and minor Duke criteria, consisting of a combination Association (AHA) 2014 guidelines for the management of
of laboratory and imaging findings and either stratified as patients with valvular disease.
478 G. M. Feuchtner

Fig. 39.5 Mitral
a b
regurgitation/prolapse. (a) A
66-year-old female with
severe regurgitation (R) of the
mitral valve caused by leaflet
prolapse (B bowing and F flail
leaflet), leading to central
incomplete closure of leaflets
(arrow, right upper panel B)
prior to surgery. The
regurgitant (R) orifice (black
arrows) shows irregular
borders (c). Three-chamber
view (a), four-chamber view
(b), and short axis through
mitral valve plane (c). Black
line indicates mitral annulus
and orthogonal line the
tenting height to maximal c
leaflet deviation below the
annulus plane toward the LA
(Bowing = B; A)

Similar to TTE and TEE, cardiac CTA allows for the visu- Fistula
alization of specific lesions developing in the course of IE A communication between cardiac chambers and the aortic
[21] with high accuracy. Specific lesions include (1) vegeta- root may develop due to inflammatory arrosion, most
tions, defined as hypodense, round, irregular, or longitudinal ­commonly on atrioventricular valvular, annular, or the mem-
masses attached to the valvular leaflets or the endocardium branous ventricular septum level. Contrast agent continuity
(“mural vegetations”). Consisting of inflammatory cells and is revealed on cardiac CTA.
granulomatous tissue, such lesions may uptake contrast at
later stages, indicating the formation of organized connective  aravalvular Abscess and Pseudoaneurysm
P
tissue. Vegetations not treated by antibiotics or removed by Paravalvular abscesses and pseudoaneurysm may be found in
surgical resection may calcify. The size of vegetations is patients with severe IE. Such lesions, which more often occur
varying from few millimeters up to >1 cm, with larger veg- after prosthetic heart valve endocarditis (PVE) than in native val-
etations often having greater mobility. The size and mobility vular IE, are described in greater detail in the next paragraph.
serve as a pivotal marker for a patients’ individual risk for
embolization into the systemic circulation, thus influencing
the clinical decision management in terms of antibiotic con- Prosthetic Heart Valves (PHV)
ventional treatment versus surgical resection of large (>1 cm)
and mobile vegetations. The imaging of prosthetic valves is a major challenge in clin-
ical practice. The early and precise diagnosis of dysfunc-
Leaflet Perforations tional valves is pivotal in order to avoid fatal complications
During the course of inflammation, irregular or round-shaped and adverse outcome. In prosthetic valve endocarditis (PVE),
lesions may occur within the cusp leading to a “torn” – aspect the 1-year mortality rate is high with 24.8–50% with
of leaflets typically causing severe valvular insufficiency improved survival for those undergoing early surgery [23].
(Fig. 39.7). A specific term is designated for the aneurysmal Prosthetic heart valve (PHV) thrombosis carries an annual
leaflet perforation type (“AR jet lesion”) [22], occurring within risk of thromboembolic events of 1–2% for mechanic and
the anterior mitral cusp, caused by an eccentric AR jet which 1.46% [24] up to recently reported 13–40% for bioprosthesis
create hole-like lesions in patients with chronic subacute. detected by CCTA [25].
39  Valvular Heart Disease and Prostheses 479

a b c

d e

Fig. 39.6 Degenerative mitral annular calcification (MAC). An 3D VRT (e) reconstruction of MAC.  Lack of contrast agent uptake
80-year old female with 0-shape circumferential MAC (e) and an inner within the mass (a, c versus b, d) allowed for a firm diagnosis of
caseous/liquefaction necrosis (LN) aspect with lesser dense areas and a MAC.  Oblique axial (a, c) and short-axis (b, d) views of the mitral
denser outer calcified shell. Mass-like aspect with myocardial invasion valve
posterior (c). Contrast-enhanced (a, b) and non-contrast CT (c, d) and

TTE and TEE are the primary imaging tools. However, outperform TEE, such as indicated in a study on transcatheter
reverberation artifacts from metal create individual device-­ bioprosthetic valve thrombosis [25, 31].
related limitations in image quality of the paravalvular
territory.
The advantage of CCTA is in the 3D visualization and 4D Spectrum of CTA Findings in PHD
dynamics of leaflet motion, in addition to the complementary
evaluation of coronary artery disease [26] or bypass graft Masses are hypodense, round, longitudinal, or irregular-­shaped
patency prior to surgery. lesions, representing either a “vegetation” in cases of infection
Previous studies investigating CCTA in prosthetic valve and clinical signs of endocarditis, “thrombus/pannus,” or
dysfunction (PVD) reported good correlation [27–29] with “undermined areas requiring further work-up.” CT attenuation
up to 100% agreement in small sample-sized studies [27] is a valuable parameter to distinguish between thrombus and
involving surgery. In PHV endocarditis [30], the best diag- pannus, with increasing Hounsfield units (HU) indicating orga-
nostic performance was observed for a combined CCTA/TEE nization with neoangiogenesis and connective tissue prolifera-
approach. In contrast, for the evaluation of the paravalvular tion of a thrombus [31–33]. A HU threshold of 145 provided
territory, CCTA performed superior to [30, 31] to TEE. Recent high sensitivity (87.5%) and specificity (95.5%) in discriminat-
advances in CCTA technology in terms of ­temporal resolu- ing pannus from thrombus (Fig.  39.8). Furthermore, masses
tion further improved the visualization of small mobile with a lower attenuation of HU <90 HU were associated with a
masses attached to devices. Most recent data reveals improved higher success rate of complete thrombolysis, as compared to
performance of CCTA for mass detection [31] that may even masses with attenuation >145 HU [32]. Similarly, the CT atten-
480 G. M. Feuchtner

a b c

d e f

Fig. 39.7  Infective endocarditis. A 69-year-old man with septic shock (abscess) confirmed by emergency surgery. Aortic valve was replaced.
and organ failure due to IE.  V,  vegetations, longitudinal hypodense CTA was performed primarily for coronary artery evaluation prior to
masses, mobile, floating in the LVOT, and P, perforation of the left coro- surgery. Three-chamber view (a, c), left sagittal oblique (b, d), and left
nary cusp. Note small pseudodiverticula (2 mm) of ascending aorta at coronal oblique and cross-sectional axial plane through aortic valve (c,
sino-tubular junction, indication infection uprising and periaortic fluid e) with MPR (a–c) and 3D VRT (d–e)

Fig. 39.8 (a and b) PHV 1.

a b
Thrombus/pannus at posterior
circumference of mechanic
bileaflet aortic prosthesis
causing leaflet opening
restriction with transvalvular
pressure gradient increase
corresponding to severe
stenosis grade III. Hypodense
mass was detected
incidentally on CTA
performed for coronary artery
disease evaluation prior to
surgery
39  Valvular Heart Disease and Prostheses 481

uation of vegetations is variable depending on lesion age and Paravalvular abscess can be diagnosed by CTA based on
degree of vascularization. Vegetations may or may not show the detection of a dense paravalvular infiltration with liquid-­
minor uptake of contrast agent. equivalent Hounsfield units (HU) of 0–30 HU and a sur-
Special types of masses are lesions originating from the rounding layer of tissue with contrast uptake (“definite
prosthetic valve apparatus, such as “degenerative lesions,” abscess”). Such a vascularized tissue layer may be present
which occur if biomaterial from a porcine or bovine biopros- during arterial contrast injection or missing and appearing
thesis degenerates, which also appear “mass-like” with during the late phase (70 s after contrast agent injection). The
rather irregular or round contours. Degenerated bioprosthe- formation of an outer vascularized layer with contrast uptake
ses are prone to develop infections, and the mass-like lesions is often classified by cardiac surgeons as the “abscess cavity
are classified as “vegetations” when associated with symp- wall” and indicates connective fibrous tissue formation. In
toms of infection such as fever, C-reactive protein elevation, patients with an acute recent onset of infection, exclusively
and/or positive blood cultures (Figs. 39.9 and 39.10). pure paravalvular liquid is seen on cardiac CTA.  If the CT
Masses frequently cause functional PHV impairment densities are not clearly related to liquid but reach above
such as stenosis or regurgitation. Patients with thrombi that 30–40 HU, paravalvular abscesses may be difficult to dis-
do not respond to lysis and that cause PHV dysfunction are criminate from periprosthetic postsurgical fibrosis or material
candidates for surgical resection. used to fix the device, such as the “fibroglue.” Accordingly, in
A visual representation of a pseudoaneurysm on CT is patients with an unclear focus of infection, an abscess should
defined as paravalvular cavity filled with contrast media. They be suspected if a “dense” paravalvular infiltration with posi-
mostly originate from a PHV annulus and grow large, often tive HU (loss of periaortic fatty tissue layer) is seen and 18flu-
larger than 1  cm. On TTE, paravalvular flow jet is usually ordeoxyglucose (FDG) positron emission tomography (PET)
detected, while the exact size, extent, and anatomic position of (18FDG-PET/CT) [34] appended. 18FDG-PET/CT increased
a pseudoaneurysm often exceed the acoustic window on the number of true positive cases diagnosed with PVE and
TTE.  Therefore, CCTA is a valuable imaging tool to fully improved the sensitivity from 70% to 97% (p  =  0.008) if
determine the size and paravalvular territory involved implemented as “major Duke” criterion into the diagnostic
(Fig.  39.10) and is also beneficial for precise surgical triage of patients with possible infective endocarditis. In such
planning. patients, the diagnosis of periprosthetic abscesses and dehis-
Paravalvular leaks are often visualized on CT as contrast cence is of paramount importance, as surgical treatment is
agent outflow at the annulus level, or flow from above to almost always required to ensure a favorable outcome.
below, and seen in both infected and non-infected devices. Dehiscence, defined as the loosening of the prosthesis
Common causes are the undersizing of the PHV in relation anchor within the annulus, is defined as rocking valve motion
to the annulus, which was observed in the early years (2008– >10 degree above annulus and found in both infected and non-
2011) of transcatheter aortic valve implantation (TAVI) infected devices (e.g., due to suture loosening). Due to its abil-
resulting in moderate up to severe paravalvular leakage with ity to generate three-dimensional interactive oblique planes,
adverse prognosis. CCTA is superior to TTE and TEE in showing the full circum-
ferential extent of “dehiscence” and the degree of cranial or

Fig. 39.9  PHV2. A

a b
75-year-old male with
structural bioprosthetic valve
degeneration and infection.
Mass-like vegetation (V),
arrow, on left sagittal coronal
planes (a) and corresponding
cross-sectional axial oblique
plaque through the aortic
bioprosthesis (b). Upper
white arrow denotes the
stented prosthesis and
attached diffuse thickening
and pseudoaneurysm (lateral
white arrows)
482 G. M. Feuchtner

a b

c d

Fig. 39.10  PHV 3. Infected and degenerated bioprosthetic valve: large leaflets and calcified bioprosthetic valve due to inflammation and
pseudoaneurysm (a) (white and black arrows) surrounding the aortic degeneration. (a, c) 3D VRT; (b, d) MPR; A, B, axial views; (c, d)
root and causing dehiscence and leaflet malfunction. Note thickened three-chamber view; LA, left atrium; LV, left ventricle

caudal displacement of the device. Exact measurement of the and/or the PHV apparatus in fully biological devices such as
“displacement angle” is feasible, and “rocking valve motion” aortic root grafts (e.g., the Freestyle™, Medtronic, or stent-­
can be visualized on 4D cine loops. Due to the risk of PHV less valves); SVD is often seen as (1) thickening of the leaf-
device embolization, urgent surgery is necessary in “rocking lets and/or (2) calcification and/or (3) “restricted leaflet
valves” and with a high degree of annulus loosening, in order motion” visualized on 4D cine images. The effective orifice
to avoid a fatal outcome. area (EOA) can be sized [35] by CCTA, if image quality
Structural bioprosthetic valve degeneration (SVD) is due allows for measurement during early mid-systolic phase and
to degenerative processes of the porcine or bovine biomate- serves as diagnostic parameter on CCTA with a good corre-
rial on cellular level, affecting the leaflets (in stented PHV) lation to the EOA measured by TTE.
39  Valvular Heart Disease and Prostheses 483

In fully biological aortic root grafts or new models of  ifferentiation of Paravalvular Abscess
D
stent-less tissue valves, severe destruction may be found and Periprosthetic Fibrosis
including paravalvular leak, aneurysmata, and dehiscence
(Figs. 39.9 and 39.10). Dense tissue with CT attenuation values around 30–40 HU
“Stuck valve” (mechanical) is restricted leaflet motion or above may create difficulties in distinguishing peripros-
with <70 degrees of opening angle during the systolic phase, thetic fibrosis from abscess. 18FDG-PET/CT [34] has higher
measured in left coronal oblique and left sagittal oblique pro- accuracy and recently showed an increase in sensitivity for
jections [36]. With mechanic valves, the opening angle diagnosis of PVE if a positive 18FDG-PET/CT was recog-
depends on the individual device implanted, while less than nized as “major Duke criterion.”
70° can be taken as a generalized valuable cutoff for restricted
orifice opening. In one study, opening angles were found to
be 64 vs 79 degrees in dysfunctional vs normal aortic valve Renal Dysfunction
prostheses (P  <  0.0001) [36]. Common causes are patient-­
prosthesis mismatch (PPM) due to mal-sizing and the A non-enhanced native CT scan in retrospective ECG-
implantation of an oversized PHV or broken mechanic leaf- gated mode allows for the evaluation of the mechanic leaf-
lets (Fig. 39.8). let function and serves as valuable alternative imaging
One of the most common findings in PVD and PVE dur- option for the assessment of mechanic leaflet function in
ing TTE are increased transvalvular pressure gradients, or patients with impaired renal function. However, unen-
transvalvular or paravalvular regurgitation. Such functional hanced cardiac CTA does not permit the detection of masses
abnormalities may be related to morphological findings such such as thrombi or panni.
as thrombi, panni [31–33], or vegetation. In cases of unclear In conclusion, cardiac CTA is a valuable 3D imaging
dysfunction, CCTA should be always requested to clarify the technique in the diagnostic work-up of patients with PHV
etiology of PVD. New data indicate a superior performance dysfunction and should be systematically integrated after
of CCTA over TTE for the detection of PHV masses [31]. initial TTE/TEE screening. Specifically, the evaluation of the
paravalvular territory such as the degree of the circumferen-
tial involvement in 3D views and to clarify etiology of PVD,
Limitations of CCTA Artifacts such as the characterization of masses such as thrombi or
panni, are indications for performing cardiac CTA. CTA
Among the variety of PHV models, artifacts depend on the offers the advantage of providing coronary artery disease
specific material composition. Fortunately, most devices, in evaluation prior to surgery.
particular the most commonly implanted bileaflet type and
the stent struts of bioprosthesis create only minor artifacts.
However, hypodense beam hardening artifacts may mimic Summary
the presence of a pannus or thrombus apposition on the pros-
thesis rings. Only one mechanical monoleaflet tilting disc Cardiac CTA has created new horizons in the evaluation of
valve (Björk-Shiley, the “convexo-concave,” composed of cardiac valves over the past 10 years. CTA offers an innova-
51% cobalt, 20% chromium, 15% tungsten, and 15% nickel) tive 3D complementary imaging modality with added value
causes severe artifacts, preventing the assessment of the in the clinical work-up algorithm in various specific
paravalvular space and device. This type of PHV is rarely settings.
seen nowadays, as it was withdrawn from the market in 1986
due to outlet strut fractures (OSFs) with escape of the disc,
resulting in embolization with fatal outcomes (0.7%). References
Notably, the average degree of opening angle was 60° for the
first device (1979) and 70° for the second. 1. Nishimura RA, Otto CM, Bonow RO, et  al. 2014 AHA/ACC
Furthermore, recent advances in CT technology (iterative guideline for the management of patients with valvular heart dis-
reconstruction techniques [37] and prospective ECG trigger- ease: executive summary: a report of the ACC/AHA Task Force on
Practice Guidelines. J Am Coll Cardiol. 2014;63(22):2438–88.
ing [38] allow for beam hardening artifact reduction. 2. Feuchtner G, Göetti R, Plass A, Baumueller S, Stolzmann P,
Prospective ECG triggering is recommended only in patients Scheffel H, Wieser M, Marincek B, Alkadhi H, Leschka S. Dual-­
with regular heart rates. Arrhythmia, which is frequent in step prospective ECG-triggered 128-slice dual-source CT for evalu-
patients after PHV implantation, may result in misalignment ation of coronary arteries and cardiac function without heart rate
control: a technical note. Eur Radiol. 2010;20(9):2093–9.
artifacts hampering image quality due to RR interval 3. Vrachliotis TG, Bis KG, Haidary A, Kosuri R, Balasubramaniam M,
variability. Gallagher M, Raff G, Ross M, O'neil B, O'neill W. Atypical chest
pain: coronary, aortic, and pulmonary vasculature enhancement at
484 G. M. Feuchtner

biphasic single-injection 64-section CT angiography. Radiology. 20. Srivatsa SS, Taylor MD, Hor K, Collins DA, King-Strunk M,

2007;243(2):368–76. Pelberg RA, Mazur W. Liquefaction necrosis of mitral annular cal-
4. Feuchtner GM, Dichtl W, Friedrich GJ, et al. Multislice computed cification (LNMAC): review of pathology, prevalence, imaging and
tomography for detection of patients with aortic valve stenosis and management: proposed diagnostic imaging criteria with detailed
quantification of severity. J Am Coll Cardiol. 2006;47(7):1410–7. multi-modality and MRI image characterization. Int J Cardiovasc
5. Shah RG, Novaro GM, Blandon RJ, Whiteman MS, Asher CR, Imaging. 2012;28(5):1161–71. https://doi.org/10.1007/s10554-
Kirsch J.  Aortic valve area: meta-analysis of diagnostic perfor- 011-9935-x. Epub 24 Aug 2011.
mance of multi-detector computed tomography for aortic valve area 21. Feuchtner GM, Stolzmann P, Dichtl W, et al. Multislice computed
measurements as compared to transthoracic echocardiography. Int J tomography in infective endocarditis: comparison with transesoph-
Cardiovasc Imaging. 2009;25(6):601–9. ageal echocardiography and intraoperative findings. J Am Coll
6. Ropers D, Ropers U, Marwan M, Schepis T, Pflederer T, Wechsel Cardiol. 2009;53(5):436–44.
M, Klinghammer L, Flachskampf FA, Daniel WG, Achenbach 22. S M, Hangler H, Feuchtner G, Bartel T.  Multimodality imaging
S. Comparison of dual-source computed tomography for the quan- reveals regurgitant jet lesion in aneurysmal anterior mitral leaflet: a
tification of the aortic valve area in patients with aortic stenosis severe complication of endocarditis. Eur Heart J. 2014;35(10):623.
versus transthoracic echocardiography and invasive hemodynamic 23. Lalani T, Chu VH, Park LP, Cecchi E, Corey GR, Durante-Mangoni
assessment. Am J Cardiol. 2009;104(11):1561–7. E, Fowler VG Jr, Gordon D, Grossi P, Hannan M, Hoen B, Muñoz P,
7. Li X, Tang L, Zhou L, Duan Y, Yanhui S, Yang R, Wu Y, Kong Rizk H, Kanj SS, Selton-Suty C, Sexton DJ, Spelman D, Ravasio V,
X.  Aortic valves stenosis and regurgitation: assessment with Tripodi MF, Wang A, International Collaboration on Endocarditis–
dual source computed tomography. Int J Cardiovasc Imaging. Prospective Cohort Study Investigators. In-hospital and 1-year
2009;25(6):591–600. mortality in patients undergoing early surgery for prosthetic valve
8. Rosenhek R, Binder T, Porenta G, et  al. Predictors of out- endocarditis. JAMA Intern Med. 2013;173(16):1495–504.
come in severe, asymptomatic aortic stenosis. N Engl J Med. 24. Egbe AC, Pislaru SV, Pellikka PA, et al. Bioprosthetic valve throm-
2000;343:611–61. bosis versus structural failure: clinical and echocardiographic pre-
9. Pibarot P, Clavel MA. Management of paradoxical low-flow, low-­ dictors. J Am Coll Cardiol. 2015;66(21):2285–94.
gradient aortic stenosis: need for an integrated approach, including 25. Makkar RR, Fontana G, Jilaihawi H, et  al. Possible subclinical
assessment of symptoms, hypertension, and stenosis severity. J Am leaflet thrombosis in bioprosthetic aortic valves. N Engl J Med.
Coll Cardiol. 2015;65(1):67–71. 2015;373(21):2015–24.
10. Alkadhi H, Leschka S, Trindade PT, Feuchtner G, Stolzmann P, 26. Bettencourt N, Rocha J, Carvalho M, Leite D, Toschke AM, Melica
Plass A, Baumueller S. Cardiac CT for the differentiation of bicus- B, Santos L, Rodrigues A, Gonçalves M, Braga P, Teixeira M,
pid and tricuspid aortic valves: comparison with echocardiography Simões L, Rajagopalan S, Gama V. Multislice computed tomogra-
and surgery. AJR Am J Roentgenol. 2010;195(4):900. phy in the exclusion of coronary artery disease in patients with pre-
11. Murphy DJ, McEvoy SH, Iyengar S, Feuchtner G, Cury RC,
surgical valve disease. Circ Cardiovasc Imaging. 2009;2(4):306–13.
Roobottom C, Baumueller S, Alkadhi H, Dodd JD. Bicuspid aor- 27. Tsai IC, Lin YK, Chang Y, Fu YC, Wang CC, Hsieh SR, Wei HJ,
tic valves: diagnostic accuracy of standard axial 64-slice chest CT Tsai HW, Jan SL, Wang KY, Chen MC, Chen CC. Correctness of
compared to aortic valve image plane ECG-gated cardiac CT. Eur J multi-detector-row computed tomography for diagnosing mechani-
Radiol. 2014;83(8):1396–401. cal prosthetic heart valve disorders using operative findings as a
12. Gouveia S, Martins JD, Costa G, Paramés F, Freitas I, Rebelo M, gold standard. Eur Radiol. 2009;19(4):857–67.
Trigo C, F Pinto F. Quadricuspid aortic valve - 10-year case series 28. Symersky P, Budde RP, de Mol BA, Prokop M.  Comparison of
and literature review. Rev Port Cardiol. 2011;30(11):849–54. multidetector-row computed tomography to echocardiography and
13. Alkadhi H, Desbiolles L, Husmann L, Plass A, Leschka S,
fluoroscopy for evaluation of patients with mechanical prosthetic
Scheffel H, Vachenauer R, Schepis T, Gaemperli O, Flohr TG, valve obstruction. Am J Cardiol. 2009;104(8):1128–34.
Genoni M, Marincek B, Jenni R, Kaufmann PA, Frauenfelder 29. Fagman E, Perrotta S, Bech-Hanssen O, Flinck A, Lamm C,

T. Aortic regurgitation: assessment with 64-section CT. Radiology. Olaison L, Svensson G. ECG-gated computed tomography: a new
2007;245(1):111–21. role for patients with suspected aortic prosthetic valve endocarditis.
14. Feuchtner GM, Dichtl W, Müller S, et al. 64-MDCT for diagnosis Eur Radiol. 2012;22(11):2407–14.
of aortic regurgitation in patients referred to CT coronary angiogra- 30. Habets J, Tanis W, Reitsma JB, van den Brink RB, Mali WP,
phy. Am J Roentgenol. 2008;191:W1–7. Chamuleau SA, Budde RP. Are novel non-invasive imaging tech-
15. Feuchtner GM, Spoeck A, Lessick J, Dichtl W, Plass A, Leschka niques needed in patients with suspected prosthetic heart valve
S, Mueller S, Klauser A, Scheffel H, Wolf F, Jaschke W, Alkadhi endocarditis? A systematic review and meta-analysis. Eur Radiol.
H.  Quantification of aortic regurgitant fraction and volume with 2015;25(7):2125–33.
multi-detector computed tomography comparison with echocar- 31. Feuchtner G, Plank F, Mueller S, Schachner T, Bonaros N, Burghard
diography. Acad Radiol. 2011;18(3):334–42. P, Wolf F, Alhassan D, Blanke P, Leipsic J, Alkadhi H, Plass A,
16. Messika-Zeitoun D, Serfaty JM, Laissy JP, et al. Assessment of the Spandorfer A, Felmly LM, De Cecco CN, Schoepf UJ.  Cardiac
mitral valve area in patients with mitral stenosis by multislice com- computed tomography angiography for evaluation of prosthetic
puted tomography. J Am Coll Cardiol. 2006;48:411–3. valve dysfunction. JACC Cardoivasc Imaging. 2017;10(1):91–3.
17. Alkadhi H, Wildermuth S, Bettex DA. Mitral regurgitation: quan- 32. Gündüz S, Özkan M, Kalçik M, Gürsoy OM, Astarcioğlu MA,
tification with 16-detector row CT—initial experience. Radiology. Karakoyun S, Aykan AÇ, Biteker M, Gökdeniz T, Kaya H, Yesin M,
2006;238:454–63. Duran NE, Sevinç D, Güneysu T. Sixty-four-section cardiac com-
18. Delgado V, Tops L, Schujif JD, De Roos A, Brugada J, Schalij MJ, puted tomography in mechanical prosthetic heart valve dysfunction:
Thomas JD, Bax JJ. Assessment of mitral valve anatomy and geom- thrombus or pannus. Circ Cardiovasc Imaging. 2015;8(12):e003246.
etry with multislice computed tomography. J Am Coll Cardiol Img. 33. Tanis W, Habets J, van den Brink RB, Symersky P, Budde RP,
2009;2(5):556–65. Chamuleau SA.  Differentiation of thrombus from pannus as the
19. Feuchtner G, Alkadhi H, Karlo C, et al. Cardiac CT angiography cause of acquired mechanical prosthetic heart valve obstruction
for the diagnosis of mitral valve prolapse: comparison with echo- by non-invasive imaging: a review of the literature. Eur Heart J
cardiography. Radiology. 2010;254:374–83. Cardiovasc Imaging. 2014;15(2):119–29.
39  Valvular Heart Disease and Prostheses 485

34. Saby L, Laas O, Habib G, Cammilleri S, Mancini J, Tessonnier L, rate measurement of mechanical prosthetic heart valve leaflet clos-
Casalta JP, Gouriet F, Riberi A, Avierinos JF, Collart F, Mundler ing angles compared with fluoroscopy. J Comput Assist Tomogr.
O, Raoult D, Thuny F.  Positron emission tomography/com- 2014;38(3):451–6.
puted tomography for diagnosis of prosthetic valve endocarditis: 37. Symersky P, Habets J, Westers P, de Mol BA, Prokop M, Budde
increased valvular (18)F-fluorodeoxyglucose uptake as a novel RP.  Prospective ECG triggering reduces prosthetic heart valve-­
major criterion. J Am Coll Cardiol. 2013;61(23):2374–82. induced artefacts compared with retrospective ECG gating on 256-­
35. Chenot F, Montant P, Goffinet C, Pasquet A, Vancraeynest D,
slice CT. Eur Radiol. 2012;22(6):1271–7.
Coche E, Vanoverschelde JL, Gerber BL. Evaluation of anatomic 38. Suchá D, Willemink MJ, de Jong PA, Schilham AM, Leiner T,
valve opening and leaflet morphology in aortic valve bioprosthesis Symersky P, Budde RP.  The impact of a new model-based itera-
by using multidetector CT: comparison with transthoracic echocar- tive reconstruction algorithm on prosthetic heart valve related arti-
diography. Radiology. 2010;255(2):377–85. facts at reduced radiation dose MDCT. Int J Cardiovasc Imaging.
36. Suchá D, Symersky P, Vonken EJ, Provoost E, Chamuleau SA, 2014;30(4):785–93.
Budde RP. Multidetector-row computed tomography allows accu-
 Cardiac Devices
40
Ian R. Drexler, Alan C. Legasto, Daniel B. Green,
and Quynh A. Truong

Intracardiac devices and implants are commonly encoun-  aged Ball Mechanical Heart Valve
C
tered in cardiac diagnostic imaging. These may range from The caged ball mechanical heart valve (Fig.  40.1) was the
prosthetic heart valves to ventricular assist devices. original type of prosthetic heart valve used. The Starr-­
Familiarity by the interpreting physician with the normal and Edwards valve was the first such valve, initially used as a
abnormal appearance of each device is essential for proper prosthetic mitral valve. Although the caged ball valve is no
detection of device complications. While devices may be longer used, thousands of patients still have them, so it is
detected via computed tomography, beam hardening and important to be familiar with their appearance [1]. In the
streak artifacts may limit their full appearance, which is why closed position, the ball, or poppet, is at the base of the cage,
it is important to know both the cross-sectional appearance creating a seal to prevent retrograde blood flow. Conversely,
and the scout or radiographic appearance of a device. This the ball is near the apex when it is open, allowing forward
chapter reviews the normal appearance of common intracar- flow. The position of the poppet depends on the pressure dif-
diac devices as well as well as frequently seen artifacts. ferential between the two chambers the valve is traversing. A
disadvantage of this design is that blood must be pumped
around the ball, rather than centrally, which forces the left
Prosthetic Heart Valves and Valve Closure ventricle to work harder. It is important to note that the ball
in the caged ball valve may be either radiolucent or radi-
Since the introduction of valve replacement surgery in the opaque depending on the poppet material.
1960s, approximately 90,000 valve substitutes have been
implanted in the United States, and 280,000 have been  ingle Tilting Disc and Bileaflet Mechanical Heart
S
implanted worldwide each year [1]. Currently, valves are Valves
classified as either mechanical, bioprosthetic, or percutane- The single tilting disc valve offers advantages over the caged
ous. The patient’s age and surgical risk typically dictate what ball valve, including improved hemodynamics and structural
type of valve is placed. stability. It consists of a metallic ring with a single circular
disc connected via metal struts [4]. Like the caged ball valve,
the disc opens when the pressure of one chamber exceeds
Mechanical Heart Valves that of the more distal chamber, such as when the left ven-
tricular pressure exceeds aortic pressure during ventricular
Mechanical valves were the first type of prosthetic valve systole. However, risks of this type of valve include potential
used clinically. They are designed to last many years, though severe hemodynamic instability due to either pannus forma-
require lifelong anticoagulation due to risk of clot formation tion or valve thrombosis. For this reason, bileaflet valves are
[2]. Thus, in younger patients who will require a functioning now the most commonly inserted mechanical valve [3].
valve for several decades, the mechanical heart valve is pre-
ferred. Mechanical heart valves may be either caged ball, Bileaflet Mechanical Valve
single tilting disc, or bileaflet [3]. The bileaflet mechanical valve (Fig.  40.2) is composed of
two semicircular discs that pivot about mechanical struts.
This design provides improved hemodynamic flow com-
pared to the tilting disc and caged ball valves, though it is
I. R. Drexler · A. C. Legasto · D. B. Green · Q. A. Truong (*)
Department of Radiology, Weill Cornell Medicine, susceptible to regurgitation.
New York, NY, USA
e-mail: [email protected]

© Humana Press 2019 487

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_40
488 I. R. Drexler et al.

a b c

Fig. 40.1  Caged ball mechanical valve: Frontal radiograph (a) demon- artifact, making evaluation limited. The caged ball device has also been
strates the typical appearance of a caged ball mitral valve replacement. used as an aortic valve replacement, as seen in a frontal chest radio-
On CT transaxial imaging (b), the cage causes metallic beam hardening graph (c)

a c d

b e f

Fig. 40.2  Bileaflet mechanical valve: Scout image (a) and volume-­ images show a bileaflet mechanical aortic valve in the closed position
rendered image (b) demonstrate mechanical valves in the aortic (white as during ventricular diastole and in the open position (e, f) as during
arrow) and mitral (black arrow) positions. Axial (c) and (d) coronal CT ventricular systole

Bioprosthetic Valves used in elderly patients, who do not need their valves to last
as long as younger patients’ valves and who are at risk of
A major advantage of using bioprosthetic valves (Fig. 40.3) falls, which can lead to major bleeding events if using
over mechanical valves is the lower incidence of anticoagulation.
antithromboembolic-­related hemorrhages. Conversely, a dis- Bovine pericardial prostheses, a type of bioprosthetic
advantage is the decreased longevity of the bioprosthetic valve, have been used for aortic valve replacements (AVR)
valve due to a combination of tissue degeneration and calci- since the first-generation valves were introduced in the early
fication [5]. For these reasons, bioprosthetic valves can be 1970s. The second-generation valves, such as the Carpentier-­
40  Cardiac Devices 489

a b e

c d

Fig. 40.3  Bioprosthetic aortic valve: CT images demonstrate en face and coronal views of a bioprosthetic aortic valve in the closed position
(a, b) and in the open position (c, d). A volume-rendered image (e) demonstrates the bioprosthetic valve in place

Edwards pericardial bioprosthesis, allowed for a more those patients who are not surgical candidates for standard
advanced method of leaf mounting, resulting in more sym- aortic valve replacements. The device is delivered arterially
metrical opening of the valve compared to its first-generation and is opened within the diseased aortic valve, displacing
predecessor [6]. and replacing it [7].
Cardiac CT is typically obtained prior to TAVR place-
ment for planning purposes. Important measurements to take
Percutaneous Heart Valves include aortic annular size, the distance between the annulus
and coronary ostia, and aortic valve area. The number of
Most recent advancement in percutaneous interventions cusps and degree of cusp calcification should also be
for structural heart disease includes transcatheter aortic assessed. In addition to imaging of the heart, the aorta and
valve replacement (TAVR) for severe aortic stenosis and iliofemoral arterial vessels are assessed for their dimensions
percutaneous mitral clip for severe mitral regurgitation. and degree of atherosclerosis, which helps plan the interven-
Less commonly performed is transcatheter pulmonary tional approach [8]. More recently, TAVRs have been used
valve implantation for congenital heart disease patients on patients who have had prior aortic valve replacements that
with right ventricular outflow tract (RVOT) dysfunction, have failed, creating a valve in valve. Post-TAVR CT
either severe pulmonic stenosis or regurgitation, to mini- (Fig. 40.4) may be obtained to evaluate for complications or
mize the number of open-heart surgeries over their causes of device failure (Fig. 40.5), which include peripros-
lifetimes. thetic aortic regurgitation, prosthetic endocarditis, valve
migration, and valve injury [9, 10].
 ranscatheter Aortic Valve Replacement [TAVR]
T
Aortic stenosis is the most common valvular heart disease.
Patients with aortic stenosis develop increased left ventricu- Mitral Valve Clip
lar wall thickness and contractility, resulting in decreased
diastolic compliance and myocardial fibrosis. In some cases, Mitral regurgitation (MR) is the second most common val-
sudden cardiac death may result as a consequence of arrhyth- vular heart disease after aortic stenosis. If untreated, severe
mias, low cardiac output, and ischemia. Transcatheter aortic MR can lead to left ventricular failure, pulmonary hyperten-
valve replacement (TAVR) was developed as a treatment for sion, atrial fibrillation, and death. Nearly half of patients
490 I. R. Drexler et al.

a b

c d

Fig. 40.4 Transcatheter aortic valve replacement (TAVR): Scout CT. Note the native aortic valve calcifications (white arrows) that have
image (a) demonstrates the metallic TAVR overlying the spine (white been peripherally displaced. Volume-rendered image (d) shows the
arrow). En face (b) and sagittal (c) views demonstrate the TAVR on valve in position

a b c

Fig. 40.5  TAVR complications: Multiplanar reformat CT images from three different patients demonstrate paravalvular leak (a), pseudoaneurysm
formation (b), and restricted motion of the non-coronary cusp (c)
40  Cardiac Devices 491

Fig. 40.6  Mitral Clip: Scout

image (a) and four-chamber
a b
view CT slice (b) demonstrate
a MitraClip device in place
(white arrows)

Fig. 40.7  Mitral valve

a b
annuloplasty: Multiplanar
reformat CT images (a, b)
show a mitral valve
annuloplasty

H
Fig. 40.8 Tricuspid a b
annuloplasty: Chest
radiograph (a) and volume-­ Aortic
rendered image (b) Prosthesis
demonstrate tricuspid
annuloplasty as well as a Mitral
pacemaker lead, mitral Prosthesis
prosthesis, and aortic
prosthesis
L

Tricuspid PPM Lead

Annuloplasty
Ring

with MR in need of valve repair or replacement are consid- Valvular Ring Annuloplasty
ered high risk for surgical intervention. The MitraClip
(Abbot Vascular, Abbot Park, IL) (Fig. 40.6) is a percutane- The concept of an annuloplasty (Figs.  40.7 and 40.8) first
ous method of repairing the mitral valve through mechani- appeared in the 1950s. The idea that fixing the mitral annulus
cal coaptation of the mitral leaflets [11]. Potential risks stems from the concept that functional mitral regurgitation is
include clip embolization and detachment from a single in part due to abnormal enlargement of the mitral annulus,
mitral leaflet [12]. which can lead to displacement of the normal coaptation
492 I. R. Drexler et al.

between the anterior and posterior mitral leaflets [13]. confusion. The device’s casing, or can, contains its software,
Annuloplasty is commonly seen with repair of the tricuspid hardware, and battery. The can, combined with the hardware,
valve for severe tricuspid regurgitation. The annuloplasty including the battery, capacitor, circuitry, and connector con-
ring is incomplete, to avoid interfering with the heart’s sensi- figuration, forms a fingerprint that is unique to each manu-
tive conduction system. Complications of annuloplasty facturer, which facilitates identification on radiography [15].
placement that can be assessed with CT include ring dehis- There may be anywhere from one to three leads terminating
cence and systolic anterior motion of the mitral valve [14]. in the heart, which defines single chamber, dual chamber,
and biventricular devices. Risks of placement include infec-
tion, lead dislodgement, perforation, and clot formation.
 acemakers, Implantable Cardioverter-­
P
Defibrillators, and Other EP Devices Implantable Pacemakers
A single-chamber device (Fig. 40.9) is identified by the pres-
Cardiac Rhythm Management Devices ence of a single lead terminating either within the right atrial
appendage or right ventricle, whereas a dual-chamber device
Implantable pacemakers and implantable cardioverters-­ has two leads, typically terminating in the right atrium and
defibrillators are commonly seen on CT. It is often easier to right ventricle. Positioning of the right atrial and right ven-
assess the type of electrophysiologic device by looking at the tricular leads are the same whether the device is a single- or
scout view or, alternatively, a frontal radiograph, as metallic dual-chamber system [16]. A biventricular device for cardiac
beam hardening artifact on cross-sectional images may cause resynchronization therapy (Figs. 40.10 and 40.11) has leads

Fig. 40.9 Single-lead
a b
pacemaker: Scout frontal
image (a) demonstrates a left
chest wall single-lead
pacemaker. A CT multiplanar
reformat maximum intensity
projection image (b)
demonstrates the lead
coursing from the right atrium
and terminating in the right
ventricular apex, where there
is excessive metallic streak
artifact

a b

Fig. 40.10  Biventricular ICD/pacemaker: Scout image (a) demon- nates in one of its branches, such as the posterolateral branch. The right
strates a biventricular ICD/pacemaker. The right ventricular lead termi- atrial lead (solid white arrow) terminates in the right atrial appendage.
nates in the right ventricular apex (solid black arrow) and contains two Axial CT image (b) demonstrates the same leads in the right ventricular
shock coils (dashed black arrow). The left ventricular lead (dashed apex (solid black arrow), right atrial appendage (solid white arrow),
white arrow) enters the coronary sinus via the right atrium and termi- and the coronary sinus (dashed white arrow)
40  Cardiac Devices 493

Fig. 40.11 Cardiac
resynchronization therapy.
a b
Volume-rendered images
before (a) and after (b)
placement of a left ventricular
lead into a branch of the
coronary sinus. Axial cardiac
CT images (c, d) showing the
LV lead (arrows) course from
the right atrium and into the
coronary sinus

c d

at the right atrial appendage, right ventricle, and a branch of

the coronary sinus to pace the left ventricle.

Subcutaneous ICD
Patients who require an ICD over decades may require one
or more lead extractions and replacements, each with a risk
for morbidity. The presence of the lead itself in the vascula-
ture imposes its own risks, including occlusion and infection.
For these reasons, subcutaneous ICD (Fig. 40.12) insertion
has become more commonplace [17].

Leadless Pacemaker
The risks of complication from conventional pacemakers,
including pocket hematoma, pneumothoraces, and vascular
obstruction, have led to the rise of leadless pacemakers
(Fig.  40.13). Leadless pacemakers are endovascularly
inserted into the right ventricle, where they are affixed to the
myocardium [18]. While measuring less than 1 cm3 in vol-
Fig. 40.12  Subcutaneous ICD: Scout image demonstrates a left chest
ume, they possess the same functionality as conventional wall subcutaneous ICD, with shock coil and lead tip (arrow) overlying
pacemakers. the left heart shadow
494 I. R. Drexler et al.

Implantable Loop Recorder or, depending on the size, a VSD or PDA.  The ASO
(Figs. 40.15 and 40.16) is composed of a braided nitinol wire
The implantable loop recorder (ILR) (Fig. 40.14) is a single-­ mesh and is shaped into two flat discs connected at a waist
lead ECG monitor that is subcutaneously inserted into the [21, 22]. The size of the occluder is dependent upon the size
patient. The device stores ECG data either in response to of the defect it is meant to close. With good positioning
patient activation or to a significant arrhythmia [19]. The bat- across an ASD, the left atrial disc will be in the left atrium,
tery lasts several years, and insertion is used for patients with and the right atrial disc will be in the right atrium. An appro-
unexplained syncope or with arrhythmias that may not be priately sized device should be disc-shaped, rather than
captured using shorter term ECG monitoring, such as with a mushroom-shaped [23]. An ASD or VSD may present with
Holter monitor [20]. The procedure is minimally invasive, signs of left-to-right shunt, with symptoms typically worsen-
with minimal risk of site infection and hematoma. ing with larger defects and, in turn, larger pulmonary to sys-
temic shunt fractions (Qp/Qs).
Additional occluders include the CardioSEAL septal
Atrial Septal and Ventricular Septal Closure occluder and the Gore Helex septal occluder. The
Devices CardioSEAL septal occluder was developed as an updated
design of the double-umbrella device (Clamshell occluder)
Numerous devices have been used for closure of atrial and [24]. The Gore Helex septal occluder (HSO) is composed
ventricular septal defects. The Amplatzer septal occluder of polytetrafluorethylene membrane connected by a niti-
(ASO) (St Jude Medical, St Paul, MN) is a percutaneously nol wire. Like the ASO, the Gore HSO has a double-disc
placed, self-expanding device that is used to close an ASD shape [22].

Fig. 40.13 Leadless
pacemaker: Scout image (a)
a b
and axial CT (b) demonstrate
the typical appearance and
position of leadless
pacemaker (arrows), which is
inserted into the right
ventricular apex. Extensive
metallic beam hardening
artifact on CT limits
evaluation of the device itself,
but its position and location
are easily discernable

Fig. 40.14  Implantable loop a b

recorder: Scout (a) and axial
CT (b) images demonstrate
implantable loop recorder
(arrows), which is implanted
in the subcutaneous tissues of
the left anterior chest wall
40  Cardiac Devices 495

a b c

Fig. 40.15  Redo PFO closure: CT multiplanar reformat image (a) demonstrates an Amplatzer septal occluder that was placed to close off a patent
foramen ovale after a CardioSEAL occluder (b) failed. A volume-rendered image (c) shows both devices together

Fig. 40.16  Amplatzer septal occluder: Axial CT image demonstrates Fig. 40.17  WATCHMAN: A multiplanar reformat CT image demon-
an Amplatzer septal occluder traversing the interatrial septum at the site strates a WATCHMAN device in a patient’s left atrial appendage. Note
of an atrial septal defect that there is thrombus contained within the WATCHMAN, which is pre-
venting the thrombus from potentially embolizing into the systemic
circulation, which could lead to cerebral infarction

Left Atrial Appendage Devices

WATCHMAN Device
Atrial fibrillation (AF) affects as many as 6 million people The WATCHMAN device (Boston Scientific, Natick, MA)
in the United States. Due to the increased risk of embolic (Fig. 40.17) is a self-expanding device designed for LAA clo-
stroke, which is thought to most often arise from thrombi sure. It is composed of a nitinol structure covered by a porous
in the left atrial appendage (LAA), patients are often kept polyethylene terephthalate membrane. The device is intro-
on long-term anticoagulation therapy [25]. Prior to plac- duced via a vascular sheath inserted into the femoral vein, and
ing an atrial appending closure device, such as a has the means to cross through the interatrial septum [26, 27].
WATCHMAN or LARIAT, a cardiac CT is performed to Complications of the WATCHMAN device placement include
interrogate the left atrial appendage’s anatomy, including device failure secondary to incomplete closure, infection, peri-
its shape, size, and position relative to the pulmonary cardial effusion, air embolism, ­ thrombus formation during
artery [26]. implantation, device embolization, and leak [28].
496 I. R. Drexler et al.

a b c

Fig. 40.18  AtriCure Clip: Scout image (a) demonstrates an AtriCure Clip (arrow), which is also seen on cross-sectional CT (b) and on volume-­
rendered imaging (c)

Fig. 40.19 Intra-aortic
balloon pump: An intra-aortic
a b
balloon pump (IABP) tip is
seen on a scout image (a).
Note that it is the only
radiopaque portion seen
overlying the heart (arrow).
Axial CT (b) demonstrates
the balloon filled with air in
the descending thoracic aorta
(arrow)

AtriCure Clip Circulator Assist Devices

An AtriCure Clip (Fig. 40.18) is an atrial appendage excluder
that is placed epicardially at the base of the appendage. It is Short-Term
composed of two parallel titanium tubes with elastic nitinol
springs. Complications of placement include occlusion of the IABP
circumflex coronary artery and pulmonary arteries, bleeding, Since 1967, when the first intra-aortic balloon pump (IABP)
incomplete LAA closure, and device migration [29]. was used, the IABP (Fig. 40.19) has become a ubiquitous cir-
culatory support device in hospitals. During ventricular dias-
LARIAT Device tole, the balloon is inflated in the descending thoracic aorta,
The LARIAT snare device (SentreHEART, Redwood augmenting perfusion of the coronary arteries and cerebral cir-
City, CA) uses a transcatheter approach to exclude the culation. Collapse during systole creates negative intra-aortic
LAA in patients with AF.  The device utilizes magnetic- pressure, which decreases afterload and in turn reduces heart
based guidewires in the LAA and pericardial space to strain [32–34]. The distal tip of the IABP is radiopaque, allow-
capture and close off the LAA at its ostium, under fluoro- ing for its position to be easily detectible on radiography. If the
scopic and echocardiographic guidance [30]. balloon is too proximal within the aorta, it can occlude the great
Contraindications to LARIAT placement include pericar- vessels, resulting in cerebral infarction or upper extremity isch-
dial thickening, calcification, or adhesions which can emia. If the balloon is too distal, it can occlude the superior
interfere with pericardial access, a LAA width of mesenteric arteries, resulting in bowel ischemia, or the renal
>40  mm, a retro-pulmonary artery, pectus excavatum, arteries, resulting in renal failure. Thus, the ideal position of the
and a posteriorly rotated heart [31]. balloon tip is 2–4 cm distal to the left subclavian artery.
40  Cardiac Devices 497

Impella and TandemHeart weaned from cardiopulmonary bypass. Since that time,

For years, the only mechanical circulatory device in use was LVADs have evolved and today are on their third generation.
the intra-aortic balloon pump. More recently, several devices The three situations LVADs are used today are as a bridge to
have come to the market. The Impella (Abiomed, Danvers, heart transplantation, as destination therapy in patients who
MA) is a nonpulsatile axial flow pump used to move blood cannot undergo a heart transplant, and in patients who require
from the left ventricle into the ascending aorta. The Impella myocardial recovery, such as after an episode of viral myo-
(Fig. 40.20) is typically placed into the femoral artery, either carditis [37]. First-generation LVADs utilized a pulsatile
percutaneously or via surgical cutdown [35]. flow, while second-generation LVADs utilized continuous
The TandemHeart (Cardiac Assist, Inc., Pittsburgh, PA) is flow. The latter led to improved survival and decreased mor-
a percutaneously placed mechanical circulatory assist device bidity associated with first-generation LVADs, such as
that creates a left atrium to femoral artery bypass. An inflow thromboembolism and ventricular arrhythmias [38].
cannula is placed into the femoral artery and advanced into Complications of LVADs are varied and include cannula
the left atrium via a transseptal puncture. Oxygenated blood obstruction, pericardial tamponade, thrombus formation,
is then pumped into the contralateral femoral artery. This aortic stenosis and insufficiency, right-sided heart failure,
results in reduced cardiac workload and oxygen demand [36]. and infection [39].

HeartWare Pump
Long-Term Assist Devices The HeartWare (HeartWare Inc., Miami Lakes, FL)
(Fig.  40.21) is a third-generation LVAD.  Not only are the
Left ventricular assist devices (LVADs) have been used since HeartWare and other third-generation LVADs smaller than
the 1960s, initially to support patients who could not be their predecessors, but they also function better by utilizing

Fig. 40.20  Impella: Scout

radiograph (a) demonstrates
a b
catheter in the aorta, attached
to the Impella, which has
radiopaque markers
proximally (white arrow) and
distally (black arrow). A
multiplanar reformat CT
image (b) shows what the
Impella looks like in cross
section

Fig. 40.21  HeartWare pump:

Scout image (a) and axial CT
a b
image (b) demonstrate a
normal appearance of the
HeartWare pump in the LV
apex
498 I. R. Drexler et al.

magnetic levitation [40]. The device consists of the pump, to the ascending aorta. The outflow cannula is often not visi-
which is attached to a cuff at the left ventricular apex, and the ble on conventional radiography and scout CT images.
outflow graft, which is attached to the ascending aorta.

HeartMate II Other Devices

The HeartMate II (Thoratec Corp, Pleasanton, CA)
(Fig.  40.22) is a second-generation, continuous flow LVAD Swan-Ganz
[41]. The radiopaque inflow cannula, placed in the left
­ventricular apex, diverts blood from the left ventricle into the The Swan-Ganz catheter (Fig. 40.23) is a pulmonary artery
pump via the bend relief. The purpose of the bend relief is to catheter that is used for hemodynamic monitoring [42]. The
allow the pump to be accommodated within the upper abdo- ideal position of the catheter tip is in either the left or right
men without kinking of the cannula. Note that the junction of main pulmonary arteries. The tip should not go past the prox-
the inflow cannula and bend relief is radiolucent. The blood imal interlobar pulmonary arteries, which extend approxi-
exits the pump via the outflow cannula, which is anastomosed mately 2 cm from the hilum. Complications for placement of
the catheter tip too distal include arterial occlusion leading to

a b c

Fig. 40.22  HeartMate II: A scout image (a) shows the normal position (white arrow) at the level of its anastomosis with the ascending aorta.
of a HeartMate II LVAD. Notable structures include the inflow cannula An additional CT image (c) demonstrates the outflow cannula (white
(white arrow), bend relief (white dashed arrow), and outflow cannula arrow) anterior the right atrioventricular groove and the inflow cannula
(black arrow). Axial CT image (b) demonstrates the outflow cannula (black arrow) in the left ventricular apex

Fig. 40.23 Swan-Ganz
catheter: Scout image (a)
a b
demonstrates a catheter (white
arrows) overlying the left
brachiocephalic vein and
coursing into the superior
vena cava, right atrium, right
ventricle, and then right
ventricular outflow tract. A
CT maximum intensity
projection multiplanar
reformat image (b)
demonstrates the same
catheter in the SVC, right
atrium, right ventricle, and
right ventricular outflow tract
40  Cardiac Devices 499

3. Tokunaga S, Tominaga R. Current status of the mechanical valve

and bioprosthesis in Japan. J Artif Organs. 2008;11:53–9.
4. Pai GP, Ellison RG, Rubin JW, et al. Disc immobilization of Bjork-­
Shiley and Medtronic Hall valves during and immediately after
valve replacement. Ann Thorac Surg. 1987;44:73–6.
5. Riess F-C, Bader R, Cramer E, et al. Hemodynamic performance
of the Medtronic Mosaic porcine bioprosthesis up to ten years. Ann
Thorac Surg. 2007;83:1310–8.
6. Puvimanasinghe J, Takkenberg J, EIijkemans M, et al. Comparison
of Carpentier-Edwards pericardial and supraannular biopros-
theses in aortic valve replacement. Eur J Cardio-thorac Surg.
2006;29:374–9.
7. Bonow RO, Leon MB, Doshi D, et  al. Management strate-
gies and future challenges for aortic valve disease. Lancet.
2016;387:1312–23.
8. Holmes DR, MacK MJ, Kaul S, et  al. 2012 ACCF/AATS/SCAI/
STS expert consensus document on transcatheter aortic valve
replacement. J Am Coll Cardiol. 2012;59:1200–54.
9. Mylotte D, Piazza N. Transcatheter aortic valve replacement fail-
ure. Circ Cardiovasc Interv. 2015;8(4). pii: e002531.
10. Pislaru SV, Nkomo VT, Sandhu GS. Assessment of prosthetic valve
function after TAVR. JACC Cardiovasc Imaging. 2016;9:193–206.
11. Wan B, Rahnavardi M, Tian DH, et al. A meta-analysis of MitraClip
system versus surgery for treatment of severe mitral regurgitation.
Ann Cardiothorac Surg. 2013;2:683–92.
12. Stewart MH, Jenkins JS. The evolving role of percutaneous mitral
valve repair. Ochsner J. 2016;16:270–6.
Fig. 40.24  PDA Clip: Multiplanar reformat CT image demonstrates a 13. Schwartz CF, Gulkarov I, Bohmann K, et al. The role of annulo-
PDA clip (arrow) between the aortic arch and pulmonary artery plasty in mitral valve repair. J Cardiovasc Surg. 2004;45:419–25.
14. Maisano F, Skantharaja R, Denti P, et  al. Mitral annuloplasty.

pulmonary infarction. Risks also include pulmonary artery Multimed Man Cardiothorac Surg. 2009;2009(918) MMCTS
2009:mmcts.2008.003640.
rupture, pulmonary artery dissection, and pseudoaneurysm
15. Jacob S, Shahzad MA, Maheshwari R, et al. Cardiac rhythm device
formation [43]. identification algorithm using X-rays: CaRDIA-X. Heart Rhythm.
2011;8:915–22.
16. Costelloe CM, Murphy WA, Gladish GW, Rozner MA. Radiography
of pacemakers and implantable cardioverter defibrillators. Am J
PDA Clip
Roentgenol. 2012;199:1252–8.
17. Aziz S, Leon AR, El-Chami MF. The subcutaneous defibrillator: a
A patent ductus arteriosus (PDA) (Fig. 40.24) is the persis- review of the literature. J Am Coll Cardiol. 2014;63:1473–9.
tent postnatal patency of an embryologic structure, the duc- 18. Reynolds D, Duray GZ, Omar R, et al. A leadless intracardiac trans-
catheter pacing system. N Engl J Med. 2016;374:533–41.
tus arteriosus, which connects the pulmonary artery to the
19. Paruchuri V, Adhaduk M, Garikipati NV, et  al. Clinical utility of
aorta, allowing for movement of oxygenated blood while in a novel wireless implantable loop recorder in the evaluation of
utero. Failure of the PDA to close after birth can lead to left- patients with unexplained syncope. Heart Rhythm. 2011;8:858–63.
to-right shunting and, if untreated, can cause chronic strain 20. Shanmugam N, Liew R. The implantable loop recorder-an impor-
tant addition to the armentarium in the management of unexplained
on the heart, possibly resulting in Eisenmenger’s syndrome.
syncope. Ann Acad Med Singap. 2012;41:115–24.
A PDA clip is a device that is surgically placed for perma- 21. Bilkis AA, Alwi M, Hasri S, et al. The Amplatzer duct occluder:
nent closure of the PDA [44], which can readily be seen in experience in 209 patients. J Am Coll Cardiol. 2001;37:258–61.
cardiac imaging. 22. Tobis J, Shenoda M.  Percutaneous treatment of patent foramen
ovale and atrial septal defects. J Am Coll Cardiol. 2012;60:1722–32.
23. Lee T, Tsai I-C, Fu Y-C, et  al. MDCT evaluation after closure
Acknowledgments  The authors would like to extend our gratitude to of atrial septal defect with an Amplatzer septal occluder. Am J
Benjamin W. Cobb for his assistance with imaging acquisition. Roentgenol. 2007;188:W431–9.
24. Kaulitz R, Paul T, Hausdorf G. Extending the limits of transcath-
eter closure of atrial septal defects with the double umbrella device
(CardioSEAL). Heart. 1998;80:54–9.
References 25. Chatterjee S, Herrmann HC, Wilensky RL, et  al. Safety and pro-
cedural success of left atrial appendage exclusion with the lariat
1. Pibarot P, Dumesnil JG.  Prosthetic heart valves: selection of the device. JAMA Intern Med. 2015;175:1104.
optimal prosthesis and long-term management. Circulation. 26. Wang Y, Di Biase L, Horton RP, et al. Left atrial appendage studied
2009;119:1034–48. by computed tomography to help planning for appendage closure
2. Habets J, Mali WPTM, Budde RPJ. Multidetector CT angiography device placement. J Cardiovasc Electrophysiol. 2010;21:973–82.
in evaluation of prosthetic heart valve dysfunction. Radiographics. 27. Waksman R, Pendyala LK.  Overview of the Food and Drug

2012;32:1893–905. Administration circulatory system devices panel meetings on
500 I. R. Drexler et al.

WATCHMAN left atrial appendage closure therapy. Am J Cardiol. 36. Brown JL, Estep JD.  Temporary percutaneous mechanical cir-

2015;115:378–84. culatory support in advanced heart failure. Heart Fail Clin.
28. Iskandar S, Vacek J, Lavu M, Lakkireddy D. Left atrial appendage 2016;12(3):385–98.
closure for stroke prevention. Cardiol Clin. 2016;34:329–51. 37. Kirklin JK, Naftel DC. Mechanical circulatory support: registering
29. Ailawadi G, Gerdisch MW, Harvey RL, et al. Exclusion of the left a therapy in evolution. Circ Heart Fail. 2008;1:200–5.
atrial appendage with a novel device: early results of a multicenter 38. Lalonde SD, Alba AC, Rigobon A, et  al. Clinical differences

trial. J Thorac Cardiovasc Surg. 2011;142:1002–1009.e1. between continuous flow ventricular assist devices: a compari-
30. Bartus K, Han FT, Bednarek J, et  al. Percutaneous left atrial
son between HeartMate II and HeartWare HVAD.  J Card Surg.
appendage suture ligation using the LARIAT device in patients 2013;28:604–10.
with atrial fibrillation: initial clinical experience. J Am Coll Cardiol. 39. Carr CM, Jacob J, Park SJ, et  al. CT of left ventricular assist
2013;62:108–18. devices. Radiographics. 2010;30:429–44.
31. Ismail TF, Panikker S, Markides V, et al. CT imaging for left atrial 40. Sawa Y. Current status of third-generation implantable left ventric-
appendage closure: a review and pictorial essay. J Cardiovasc ular assist devices in Japan, Duraheart and HeartWare. Surg Today.
Comput Tomogr. 2015;9:89–102. 2015;45:672–81.
32. Parissis H, Graham V, Lampridis S, et al. IABP: history-­evolution-­ 41. Rodriguez LE, Suarez EE, Loebe M, Bruckner BA.  Ventricular
pathophysiology- indications: what we need to know. J Cardiothorac assist devices (VAD) therapy: new technology, new hope?
Surg. 2016;11:122. Methodist Debakey Cardiovasc J. 2013;9:32–7.
33. Hurwitz LM, Goodman PC. Intraaortic balloon pump location and 42. Hunter TB, Taljanovic MS, Tsau PH, et al. Medical devices of the
aortic dissection. AJR Am J Roentgenol. 2005;184:1245–6. chest. Radiographics. 2004;24:1725–46.
34. Ihdayhid AR, Chopra S, Rankin J. Intra-aortic balloon pump. Curr 43. Godoy MCB, Leitman BS, de Groot PM, et al. Chest radiography
Opin Cardiol. 2014;29:285–92. in the ICU: part 2, evaluation of cardiovascular lines and other
35. Rihal CS, Naidu SS, Givertz MM, et al. 2015 SCAI/ACC/HFSA/ devices. Am J Roentgenol. 2012;198:572–81.
STS clinical expert consensus statement on the use of percutane- 44. Mandhan PL, Samarakkody U, Brown S, et  al. Comparison of
ous mechanical circulatory support devices in cardiovascular care. suture ligation and clip application for the treatment of patent
J Am Coll Cardiol. 2015;65:e7–e26. ductus arteriosus in preterm neonates. J Thorac Cardiovasc Surg.
2006;132:672–4.
 Part IX
Where We Are: Transcatheter Therapy Planning
 CT in the Context of Transcatheter
Aortic Valve Replacement 41
Eli Konen, Orly Goitein, and Arik Wolak

Aortic valve stenosis (aortic stenosis, AS) is the most com- tic balloon dilation is reserved for patients who are not
mon cause of left ventricular outflow obstruction in children eligible for surgical aortic valve replacement (SAVR) or
and adults [1]. Severe AS is defined as maximum aortic valve serves as a bridge to valve replacement procedure [2].For
velocity ≥4 m/s with an aortic valve area ≤1 cm2, and/or the more than 5 decades, SAVR, using either biological or
mean transvalvular gradient exceeds 40  mmHg [2]. mechanical valves, has been the gold standard treatment for
Symptomatic AS includes a combination of the following: severe symptomatic AS, severe AS with ejection fraction
chest discomfort, heart failure, decreased exercise tolerance, <50%, and severe asymptomatic AS in patients undergoing
effort dyspnea, and syncope. The main AS etiologies are cal- CABG [2, 10]. Although it offers an effective treatment
cific disease of a trileaflet valve or a congenitally abnormal option, it is still an extensive surgical procedure with an
valve (bicuspid or unicuspid) and rheumatic valve disease. In overall mortality rate ranging between 1.9% and 17%
the Western world, calcific aortic disease of either a tricuspid depending on the patient’s risk group [11, 12]. As a rule, bio-
or bicuspid valve is the most common cause. AS prevalence prosthetic valves tend to be less durable, while mechanical
increases with age [3]. AS in younger patients is usually valves require lifelong anticoagulation treatment, both being
associated with congenitally abnormal valves, whereas prone to infections [13]. Due to high surgical risk or surgical
trileaflet valve is more common in the elderly patients [4]. contraindications, a significant portion of potential SAVR
Usually, patients with AS and normal left ventricular systolic candidates will not benefit from the procedure. Therefore,
function remain asymptomatic until the stenosis is severe during the last decade, percutaneous aortic valve replace-
[5]. Once symptomatic, patients with severe AS who do not ment is becoming an increasingly growing alternative to
undergo valve replacement have a poor prognosis [6, 7]. SAVR, mainly for these high surgical risk patients [14–16].
Echocardiography is currently the main tool in the initial AS Encouraging new data suggest that intermittent risk patients
diagnosis and evaluation. Cardiac catheterization is recom- may also benefit from a percutaneous procedure [17].
mended only if the noninvasive evaluation is nondiagnostic. Transcatheter aortic valve replacement (TAVR) is the
The role of multidetector computed tomography (MDCT) in replacement of the aortic valve without open heart surgery.
the initial evaluation of AS is minimal and limited to aortic The valve is delivered and positioned inside the diseased aor-
valve calcification quantification [8], potentially beneficial in tic valve through a blood vessel or through the apex of the
“paradoxical low-flow low-gradient” AS and in excluding heart using a delivery system, which is removed once the
pseudo-severe AS [9]. MDCT plays a key role in decision-­ prosthetic valve is implanted. The catheter-based delivery
making and procedure planning once procedure is indicated system can be inserted using several approaches including
as will be discussed later. The poor prognosis of untreated transfemoral, transapical, axillary-subclavian, direct aortic
symptomatic AS patients, together with enhanced survival (through the ascending aorta), and transcaval (through the
rates following valve replacement, is the rational for the rec- femoral vein instead of the femoral artery and across from
ommendation for early valve replacement. Percutaneous aor- the inferior vena cava into the adjacent abdominal aorta)
[18–20]. The transfemoral approach is performed under mild
E. Konen · O. Goitein (*) sedation while the delivery system is inserted into the femo-
Department of Diagnostic Imaging, Chaim Sheba Medical Center, ral artery through a small incision in the groin and progressed
Ramat Gan, Israel
e-mail: [email protected] along the artery to the correct position at the aortic valve.
The transapical approach is performed under general anes-
A. Wolak
Department of Cardiology, Shaare Zedek Medical Center, thesia. A minimal surgical incision is made between the ribs
Jerusalem, Israel followed by a small puncture of the cardiac apex where the

© Humana Press 2019 503

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_41
504 E. Konen et al.

Fig. 41.1  Aortic valve

replacement TAVR can be
performed using several
approaches: open heart
surgery, transaortic,
transapical, and transfemoral

delivery system is introduced into the left ventricle and lus) that is formed by joining the basal attachment of the
through the left ventricular outflow tract to the correct posi- valvar leaflets. The second is the anatomical ventriculoaortic
tion at the aortic valve (Fig. 41.1). Regardless of the approach, junction (the “surgical” annulus). The third ring is the sino-
the procedure requires constant echocardiography guidance tubular junction. An additional fourth ring is seen in three
and pacemaker insertion for rapid pacing during the pros- dimensions, as the valve leaflets take the form of a three-­
thetic valve final positioning and as backup if intra-­procedural pronged coronet, with the hinges from the supporting ven-
heart block occurs. Currently, there are several available tricular structures forming the “crown-like” ring (Fig. 41.3)
TAVR systems (Fig.  41.2). The fundamental difference [23]. The dimensions of the aortic root components vary in
between these systems is in the aortic prostheses deployment size and shape (circular vs. elliptic) between different indi-
method being either a self-expandable (SE) or balloon-­ viduals and change throughout the cardiac cycle by up to
expandable (BE) system [21]. 20% [24]. The aortic valve is usually trileaflet, yet bicuspid
Understanding the anatomy of the aortic valve compo- aortic valve is a common abnormality and rarely unicuspid
nents and surroundings  – the aortic root, valvular leaflets, valve can be found. As for the aortic root, marked variations
left ventricular outflow tract, origin of the coronary arteries, exist among individuals in the dimensions of all elements.
the mitral valve, and the conduction system – is essential for Such disparities in the size of the leaflets can cause inaccu-
optimal patient selection and for achieving best TAVR result rate measurements of the aortic annulus especially when
with minimal procedural complications. The aortic root is acquiring measurements, using a two-dimensional modality.
the anatomical location of TAVR procedure. It is located The left main and the right coronary arteries most often arise
between the subpulmonary infundibulum anteriorly, the ori- within the left and right sinuses of Valsalva just below the
fice of the mitral valve and the muscular ventricular septum sinotubular junction, yet many times they can arise from a
posteriorly. It extends from the origin of the valvular leaflets lower or higher position in the aortic root. Postmortem of
at the left ventricle all the way distally to the sinotubular normal hearts demonstrated an average distance of 12.6 and
junction [22]. Four circumferential rings outline the aortic 13.2 mm from the orifice of the left main and right coronary
root. The first, the most basal one, is a virtual ring (the annu- arteries to the aortic annulus, respectively [23, 25] . The left
41  CT in the Context of Transcatheter Aortic Valve Replacement 505

a b c d e f g h

Fig. 41.2 Currently, eight transcatheter aortic valve replacement (JVT Research & Development Corp, Irvine, CA); (e) St. Jude Medical
(TAVR) systems are commercially available in Europe (a–h), whereas Portico Valve (St. Jude Medical, St. Paul, MN); (f) Direct Flow Medical
two TAVR systems are approved by the US Food and Drug Valve (Direct Flow Medical, Inc., Santa Rosa, CA); (g) Medtronic
Administration in the United States (a, b). (a) Edwards Lifesciences Engager Valve (Medtronic, Minneapolis, MN); and (h) Boston
Sapien 3 Valve (Edwards Lifesciences, Irvine, CA); (b) Medtronic Scientific Lotus Valve (Boston Scientific, Marlborough, MA). (From
CoreValve Evolut R (Medtronic, Minneapolis, MN); (c) Symetis Vahl et al. [27], with permission)
Acurate neo Valve (Symetis, Ecublens VD, Switzerland); (d) JenaValve

impact their management strategy. Therefore, it is crucial

that a multidisciplinary team will choose the best treatment
approach on an individual basis, considering age, frailty, and
anatomic and clinical features. The current general approach
is to prefer SAVR for lower-risk young patients, TAVR for
high-risk or inoperable patients, and medical palliative treat-
ment to prohibitive-risk patients (Fig.  41.4) [27]. TAVR is
currently evolving in three selected severe AS patients’
groups: valve-in-valve for bioprosthetic valve failure, bicus-
pid aortic valve, and AS with concomitant valvular or coro-
nary artery disease [27]. Per-patient TAVR procedure
tailoring is the current standard of care. A dedicated “heart
team” of collaborating physicians, which includes cardiolo-
gists, cardiac surgeons, and cardiac imaging experts (echo-
cardiography, MDCT, and MRI), facilitates better
communication and understanding of the complex anatomy
at question [27–29].
Fig. 41.3  Aortic root and anatomic location of aortic annulus. The MDCT offers comprehensive pre-procedural imaging
three red rings represent the aortic valve cusps, with the green circle and understanding of the aortic root structures relevant to
placed at the nadir of the aortic cusps denoting the annular plane. the TAVR procedure, including the left ventricular outflow
LC left coronary cusp, NC noncoronary cusp, RC  right coronary cusp.
(From Leipsic et al. [28], with permission) tract, aortic annulus, aortic sinuses, sinotubular junction,
ascending aorta, coronary ostia height, procedure access
route, and non-cardiac pathology which may affect the
ventricular outflow tract (LVOT) is a portion of the left ven- procedure (Table  41.1, Figs.  41.5, 41.6, 41.7, 41.8, and
tricle through which blood passes to the aorta. LVOT charac- 41.9) [27, 28, 30–32]. Correct device sizing is key in pre-
teristics which may hinder device positioning include venting complications, including coronary occlusion,
changing angulation with age and the presence of subaortic annular rupture, and paravalvular leak, severely affecting
asymmetrical septal hypertrophy [23].The aortic valve is patient outcome [33, 34].
located in proximity to all cardiac chambers including the
conduction system. Therefore, it is only understandable why
aortic valve procedures may cause heart block or intraven- Acquisition Protocols
tricular conduction abnormalities. The length of the mem-
branous septum and the presence of basal septum In order to evaluate the entire spectrum of aortic pathology
calcifications were documented as associated with post-­ and procedure access route, the scanning protocol should
TAVR conduction abnormalities [26]. Patients with severe extend from the clavicles to the groin. A separate non-­
AS present with a wide variety of comorbidities that may enhanced ECG-gated scan (“calcium score” scan) of the
506 E. Konen et al.

FUTURE MANAGEMENT STRATEGIES FOR PATIENTS WITH SYMPTOMATIC SEVERE AORTIC STENOSIS

‘Prohibitive risk’ patients Extreme risk or ‘inoperable’ patients ‘High risk’ patients ‘Lower risk’ patients*

Surgical aortic valve

SAVR SAVR SAVR (preferred)
replacement (SAVR)
Transcatheter aortic TAVR TAVR (preferred) TAVR
valve replacement (TAVR)

• Both SAVR and TAVR • SAVR suboptimal • Both SAVR and TAVR expected • Both SAVR and TAVR expected to
considered ‘futile’ to improve survival and QoL improve survival and QoL
• TAVR expected to improve survival
• Focus on symptom relief and quality of life (QoL) • TAVR preferred unless age, • May consider TAVR in absence of
and palliation anatomical or other patient factors anatomic or unfavorable clinical
make SAVR the superior option characteristics with emphasis on
patient age and valve durability

Fig. 41.4  TAVR, future expectations and barriers. Patients with severe formance standard is set by surgical aortic valve replacement (SAVR).
aortic stenosis present with a spectrum of comorbidities that influence TAVR has evolved as the preferred treatment for high-risk and inoper-
the treatment options available to them. The heart team will choose the able patients. Because the durability of transcatheter heart valves and
optimal treatment strategy for individual patients based on their age, outcomes in lower-risk patients require further studies, SAVR remains
frailty, and anatomic and clinical characteristics. For patients undergo- the treatment of choice for such patients (From Vahl et al. [27], with
ing transcatheter aortic valve replacement (TAVR), the bar for the per- permission)

Table 41.1  Pre-procedural TAVR CT report ­ brillation is present. Due to the nature of severe aortic
fi
Aortic annulus stenosis, administration of beta-blocker is not recom-
Diameters (shortest, longest) mended, since it may result in further decrease of left ven-
Annulus calcifications tricular function [35]. Isosorbide dinitrate spray is
Area and area-based diameters relatively contraindicated as well. Non-­gated acquisition
Perimeter and perimeter-derived diameters can be used for scanning the entire aorta and femoral ves-
Aortic annulus plane for fluoroscopy sels, thus decreasing both scan time and volume of admin-
Left ventricular outflow tract calcifications istered contrast. Spatial resolution should be high with a
Aortic root
reconstructed slice width  ≤1.0  mm throughout the entire
Sinotubular junction diameters
scan. Acquisition protocols vary and are determined by the
Sinus of Valsalva width and height
Coronary ostia to aortic annulus distance specific scanner platform. Wide detector systems allow
Aorta capture of the entire dataset with ECG gating. Spiral high-
Ascending aorta, aortic arch, descending aorta width pitch ECG-gated scanning offers rapid cranio-­caudal cov-
Tortuosity, kinking, dissection, obstruction erage, with a potential for further decrease in the contrast
Aortic valve volume needed for adequate imaging [28, 30, 35–37].
Aortic cusps length, calcifications
Cuspidity (bicuspid, tricuspid)
Vascular access Contrast Media and Radiation Exposure
Iliofemoral arteries: minimal diameter
Tortuosity, angulation
Obtaining a satisfactory quality scan with a minimum con-
Calcifications
Subclavian arteries – minimal diameter trast amount is of utmost importance in the elderly popula-
Noncardiac findings tion of TAVR patients. It has been recently shown that
Findings that will affect procedure (change, postpone, cancel) contrast media volume can be substantially reduced, while
Modified from Achenbach et al. [30], with permission maintaining adequate image quality, utilizing low-tube volt-
age acquisitions, and enhancing contrast conspicuity [36,
38–40].The risk for contrast-induced nephropathy following
heart with a limited field of view offers the ability to quan- TAVR is non-negligible and is related to repeated contrast
tify the calcification burden in each of the aortic root administration [41, 42]. In the TAVR patient population, it is
­components. It is imperative to scan the aortic root using associated with poor outcome and increased mortality [41,
ECG gating, which can be either retrospective or prospec- 43]. Radiation exposure is of less significance when the
tive or high-pitch mode. The ECG gating allows procure- elderly TAVR population is considered, since the c­ arcinogenic
ment of accurate, motion-free images. Retrospective effects of radiation exposure require a lag time of at least
gating may be helpful when high heart rates or atrial 10 years [43]. When addressing TAVR procedures in younger
41  CT in the Context of Transcatheter Aortic Valve Replacement 507

a b c

d e f

Fig. 41.5  Aortic annulus. (a) The aortic annulus obtained using two (red). (f) Volume rendering of the aortic root demonstrating the aortic
perpendicular pales through the nadir of the aortic cusps (d, e). Annulus “rings”: the left ventricular outflow tract (LVOT), the annulus (orange),
diameters longest and shortest (in orange). (b) Aortic annulus perimeter the sinotubular junction, and the ascending aorta. The origin of the left
and derived diameters (in orange). (c) Aortic annulus area and derived main coronary artery (white asterisk), origin of the right coronary artery
diameters (in orange). (d, e) Multiplanar reformats (MPR) serving for (red asterisk)
correct location at the annular plane at the nadir of the aortic valves

patients, radiation exposure issues cannot be overlooked and MDCT are, by far, superior to 2D modalities such as 2D
should be addressed appropriately. transthoracic echocardiography (TTE) or angiography [23,
24, 28, 30, 46–50]. Thus, MDCT is considered the gold stan-
dard for device sizing [27, 51, 52] . The anatomical struc-
Aortic Annulus Evaluation tures at the aortic root change throughout the cardiac cycle.
The annulus endures pulsatile changes as well as contour
The aortic annulus is a complex three-dimensional virtual deformity caused by the adjacent structures such as the aor-
ring at the base of the aortic cusps [23]. The annulus is oval tomitral tissue and left atrium [23, 53]. These annular cyclic
in nature [23, 44, 45], and hence 3D imaging modalities changes need to be taken into consideration when choosing
including 3D transesophageal echocardiography (TEE) and the device that needs to withstand them. Recently, some
508 E. Konen et al.

a b c d

Fig. 41.6 (a–c) Cross section at the levels of the left ventricular outflow tract (LVOT) (a), sinuses of Valsalva (b), and the sinotubular junction (c).
(d) Volume rendering of the aortic root demonstrating the corresponding aortic “rings” (a–c)

a b c

Fig. 41.7  Distance from the annulus to the origin of the coronary arteries. (a) Distance to the left main (LM) (red asterisk LM origin). (b) Distance
to the right coronary artery (RCA) (black asterisk RCA origin). (c) Left coronary sinus height

authors have advocated the utilization of systolic phases in annular perimeter and area, and the calculated derived vir-
order to refrain from prosthesis under-sizing [24, 54]. tual diameters from both perimeter and area (Fig. 41.5) [28,
However, if systolic phases are suboptimal, it is important to 30, 32, 35, 46, 57]. MDCT perimeter- and area-derived vir-
ensure adequate image quality using the best available phase. tual diameters have been shown to change the least during
Current CT scanners offer isotropic data acquisition allow- the cardiac cycle and are considered to be the most reliable
ing precise, reproducible nonoperator-dependent measure- diameter measurement [28, 30, 32, 35, 46, 57]. Dedicated
ments [46, 55, 56]. Accurate annular sizing should be application platforms for facilitating these measurements
performed using two orthogonal planes along the aortic root. exist and are utilized in daily practice [54, 57–60]. Cases in
The perpendicular plane immediately below the nadir of the which the annular size is borderline (for a certain device
aortic cusps represents the annulus. At this location, mea- size) are particularly influenced by the geometric changes
surements should include diameters (shortest and longest), between systole and diastole. A comprehensive multimodal-
41  CT in the Context of Transcatheter Aortic Valve Replacement 509

Fig. 41.8 Fluoroscopic Optimal Angles Graph

projection angle predicted by
a b
LAO = 11 Caudal = 8
MDCT. (a) Volume rendering
of the aortic root aligning all 40
three cusps in the same plane.
(b) The “line of
perpendicularity”
demonstrating the preferred
fluoroscopic angel for the
20
procedure

Caudal/Cranical [deg]
0

-20

-40

-40 -20 0 20 40
PHILIPS RAO/LAO [deg]

ity approach incorporating MDCT, 3D TEE, and balloon siz- roscopic (C-arm) angle in which all three coronary cusps are
ing (during the procedure) is recommended when evaluating aligned in the same plane allows ease of utilization with the
challenging cases [28, 54, 61]. Prosthesis sizing is under- physical constraints of a catheterization suite (Fig.  41.9).
taken using dedicated tables for each device. When sizing a This can be achieved both manually or using dedicated
device, one should consult with the most updated manufac- automated software. It is of high value for accurate device
turer instructions. positioning and lowering the amount of aortograms and
hence the contrast media volume needed for procedure com-
pletion [30, 46, 59, 63, 64].
Aortic Root Evaluation

Other components of the aortic root should be addressed, Calcification Burden

including the distance of the coronary ostia location
(Fig. 41.6), aortic cusp length and calcification, aortic sinuses The presence, quantity, and distribution of aortic root calcifi-
width and degree of calcification, and sinotubular junction cation are related to procedure outcome [65–69]. During a
and ascending aorta size. These data are essential in fitting TAVR procedure, the native aortic valve leaflets are com-
the appropriate device to the specific anatomy [30]. Prosthesis pressed by the inserted device rather than resected as in
inappropriate sizing is associated with coronary occlusion, SAVR. TAVR prosthesis anchoring requires some degree of
annular rupture, paravalvular regurgitation, and device calcification; however, high calcification burdens in the aor-
embolization [62]. tic valve and LVOT are strong predictors for paravalvular
aortic regurgitation (PAR) [65, 67–69]. Greater calcium bur-
den in the left coronary cusp has been implicated as an inde-
Aortic Annulus Plane for Fluoroscopy pendent predictor for the need for permanent pacemaker
following the procedure [70]. LVOT calcification adjacent to
The specific aortic root orientation is different for every the noncoronary cusp was demonstrated as a predictor for
patient with a general anterior right-sided angulation in aortic root injury (Fig. 41.6) [71]. Other authors found a high
most cases. The accurate position for each patient is easily total aortic calcification burden to be an integrative predictor
depicted using MDCT data. Thus, providing a correct fluo- for cardiac and all-cause mortality [66].
510 E. Konen et al.

a b c

Fig. 41.9  Vascular access by MDCT (90-year-old TAVI patient). (a) dicular slices for measurement of vascular diameters at the levels of
Volume rendering (VR) of the entire aorta and femoral arteries demon- right common iliac artery (c), right external iliac artery (d), and right
strating extensive atherosclerotic disease. (b) Curved multiplanar refor- common femoral artery (e). Note, the circumferential calcifications at
mats (CMPR) of the aorta and the right iliac and femoral arteries this level, representing an unfavorable vascular access
demonstrating atherosclerotic disease. (c–e) Cross-sectional perpen-

Aortic Valve Cuspidity stenosis and dilatation of the proximal aorta segments
[72]. Recently, a TAVR-oriented classification was sug-
Bicuspid aortic valve (BAV) is the most common congeni- gested, dividing BAV into tri-commissural, bi-commis-
tal cardiac anomaly. Inherent abnormalities in the aortic sural raphe type and bi-commissural non-raphe type [73].
valve are related to progressive calcification which leads to Short term outcome (within 30 days) of TAVR in patients
41  CT in the Context of Transcatheter Aortic Valve Replacement 511

with BAV was similar to that reported for those with nor- incidence) new conduction disorders requiring placement
mal, tricuspid valves although a higher rate of permanent of a new permanent pacemaker (reported more in SE
pacemaker insertion was demonstrated for both BE and SE devices), major vascular and bleeding events, acute kidney
device types [73]. Some authors have reported a higher injury, major stroke, and myocardial infarction [77]
incidence of post-TAVR PAR in patients with BAV; how- (Figs. 41.10 and 41.11). As cumulative experience increases
ever recent published data does not support this associa- and technical aspects of TAVR progress, a significant
tion [73, 74]. reduction in post-procedural complications is apparent,
especially when comparing results between the earliest
reports and the current literature [78, 82].The new-genera-
Procedure Access tion BE and SE devices incorporate several advanced tech-
nical alterations. Slimmer-profile systems (14–16F
Transfemoral arterial approach for TAVR was introduced expandable sheaths) are related to less vascular injury and
in 2006 and is considered the preferred approach ever offer procedure availability for patients, who, previously,
since [75]. Single-plane angiography is considered a mini- were not eligible for transfemoral approach. Ease of device
mal requirement for pre-TAVR vascular evaluation; how- manipulation is related to the decrease in adverse events
ever atherosclerotic burden assessment is considerably such as device malpositioning, the need for a second valve-
limited using this approach [76]. Major vascular complica- in-valve intervention, urgent surgery, or the need of cardio-
tions are not uncommon and were described with a higher pulmonary support [78]. The current ability for device to
incidence especially in women [77–79]. Target vessel repositioning or retrieval permits real-time procedure mod-
dimensions, tortuosity, and extent of calcifications are ifications. The wider range of valve sizes for both SE and
important factors when addressing procedure access BE devices allows better patient device adjustment.
(Fig. 41.9). Contrast-enhanced MDCT has been shown to Paravalvular aortic regurgitation (PAR) is a much-discussed
have a greater predicative value for post-TAVR vascular issue. Moderate (and greater) post-TAVR PAR is reported
complications as compared with angiography [18]. as associated with an increase in the overall mortality risk
Potential contraindications for transfemoral approach [78, 84, 85]. Some authors have found a higher incidence of
include external sheath diameter exceeding the minimal PAR when comparing SE to BE valve; however this is not
arterial diameter, significant vessel tortuosity, and exten- reported by all authors [21, 58, 82]. Adherence to MDCT
sive peripheral vascular disease [75, 80]. Calcifications, in sizing lowers the incidence of PAR [59, 86]. Reduced bio-
particular circumferential or near-circumferential, limit prosthesis leaflet motion was demonstrated utilizing four-
arterial expandability, thus increasing the potential risk for dimensional MDCT but resolved following anticoagulation
perforation or dissection [77, 80]. Routine implementation administration and was not associated with increased clini-
of pre-TAVR CT, along with ongoing decrease in the deliv- cal events [87].
ery system size requirements, holds promise for further
lowering vascular complication rates [80].
Transfemoral approach is associated with lower mortality Catastrophic Procedure Complications
rate (30-day and 1-year) as compared with non-transfemoral
approach [19, 20]. However, in the presence of prohibitive Coronary occlusion is rare and caused mainly by displace-
transfemoral access, alternative procedure routes exist, ment of the calcified native valves over the coronary ostia.
including transapical, transaortic, transcaval, and subclavian It was documented as more prevalent in women, heavily
approaches. calcified cusps, balloon-expandable devices (twice as
high), and “valve-in-valve” procedures [34]. A coronary
height <12 mm, relatively small sinuses of Valsalva diam-
Post-TAVR Outcomes eter (<30 mm), long aortic cusps, and narrow sinotubular
junction were demonstrated in the majority of cases [30,
Successful device implantation is reported in more than 34, 88] (Fig.  41.7). Annular rupture is rare, described
90% of patients [16, 33, 77, 81–83]. New-generation mostly as case reports, some of which suggest relation to
balloon-­expandable devices have been introduced, signifi- prosthesis oversizing [62, 89, 90]. Other adverse outcomes
cantly lowering mortality, with reported 1-year survival include aortic dissection, device embolization, and tam-
rates of 85.6% and 74.7% for the transfemoral and the ponade. All of these are rare, acute complications requir-
transapical or transaortic approaches, respectively [78]. ing immediate intervention and are uniformly associated
Post-TAVR complications include (in descending order of with poor outcome [88, 91, 92].
512 E. Konen et al.

Fig. 41.10 Post-TAVI
pseudoaneurysm. (a)
a b
Multiplanar reformats (MPR)
of the left ventricular outflow
tract (LVOT) before TAVI
demonstrating heavy
calcifications at the aortic
annulus (dotted arrow), at the
LVOT (black arrow), and at
the mitral annulus (MAC)
(white arrow). (b) MPR of the
LVOT after TAVI
demonstrating a
pseudoaneurysm (white
asterisk) originating at the
aortic annulus. TAVI device is
in place (black star). (c, d)
MIP and volume rendering of
the LVOT demonstrating the
pseudoaneurysm (white
asterisk), TAVI device (black
star), MAC (white arrow), and
LVOT calcifications (black
arrow)

c d

Valve-in-Valve: New Options for TAVR been developed in order to guide accurate valve-in-valve
proper sizing [95].
In recent years, the utilization of aortic bioprosthetic
valves is favored over mechanical valves due to their
lower thrombogenicity [93]. However, bioprosthesis Evaluation of the Coronary Anatomy
deterioration usually occurs by 10–15 years. TAVR valve-
in-valve offers a noninvasive solution for failed biopros- Assessment of the coronary arteries using the data collected by
thetic surgical valves. An overall 1-year survival of 83% TAVR study is feasible yet challenging. In the majority of these
was reported in a large international registry using both patients, it is recommended to avoid beta-blockers and nitrates
SE and BE devices [75, 94]. Smartphone platforms have prior to the acquisition; therefore, only limited rate control and
41  CT in the Context of Transcatheter Aortic Valve Replacement 513

Fig. 41.11 Post-TAVI
pseudoaneurysm follow-up
a b
(same patient as in
Fig. 41.10). (a) Volume
rendering (VR) pre-TAVI –
normal study. (b) VR
post-TAVI (2 days)
demonstrating a
pseudoaneurysm (white
asterisk). (c) VR post-TAVI
(2 months) demonstrating a
decrease in size of the
pseudoaneurysm (white
asterisk). (d) VR post-TAVI
(6 months) demonstrating
resolution of the
pseudoaneurysm with
conservative treatment

c d

coronary dilation are achieved during the scan. Additionaly mentation prior to TAVR truly embodies the ideal of a com-
owing to the nature of the TAVR target population (elderly and prehensive “heart team” incorporating imaging experts,
frail), it is usual to encounter breathing artifacts, blooming arti- cardiologists, and cardiac surgeons for optimal patient care.
facts due to heavy calcification, and motion artifact due to
arrhythmia (mostly atrial fibrillation and relative tachycardia).
In one study of TAVR patients, sensitivity, specificity, and posi- References
tive and negative predictive value of 96%, 73%, 72%, and 96%,
respectively, were documented for the evaluation of only seven 1. American College of C, American Heart Association Task Force
on Practice G, Society of Cardiovascular A, Bonow RO, Carabello
proximal segments (left main, proximal, and middle left ante- BA, Chatterjee K, et al. ACC/AHA 2006 guidelines for the man-
rior descending artery, proximal and middle right coronary agement of patients with valvular heart disease: a report of the
artery, and proximal and middle circumflex artery) or bypass American College of Cardiology/American Heart Association task
grafts [96]. Currently, it is a reasonable approach to perform an force on practice guidelines (writing committee to revise the 1998
guidelines for the management of patients with valvular heart dis-
initial TAVR CTA study and to proceed to diagnostic invasive ease) developed in collaboration with the Society of Cardiovascular
coronary angiography only if the data acquired is insufficient. Anesthesiologists endorsed by the Society for Cardiovascular
In conclusion, MDCT prior to TAVR provides imaging Angiography and Interventions and the Society of Thoracic
information which, when used properly, serves to reduce Surgeons. J Am Coll Cardiol. 2006;48(3):e1–148.
2. Nishimura RA, Otto CM, Bonow RO, Carabello BA, Erwin JP 3rd,
procedure-specific complications. It allows tailored, per-­ Guyton RA, et al. 2014 AHA/ACC guideline for the management
patient procedure planning including device type, size, pro- of patients with valvular heart disease: a report of the American
cedure access route, procedure fluoroscopic angulation, and College of Cardiology/American Heart Association task force on
post-procedure follow-up. Multimodality imaging imple- practice guidelines. J Am Coll Cardiol. 2014;63(22):e57–185.
514 E. Konen et al.

3. Eveborn GW, Schirmer H, Heggelund G, Lunde P, Rasmussen valve implantation: a systematic review and meta-analysis. Can J
K.  The evolving epidemiology of valvular aortic stenosis. The Cardiol. 2015;31(12):1427–38.
Tromso study. Heart. 2013;99(6):396–400. 20. Thourani VH, Jensen HA, Babaliaros V, Suri R, Vemulapalli

4. Roberts WC, Ko JM. Frequency by decades of unicuspid, bicuspid, S, Dai D, et  al. Transapical and transaortic transcatheter aor-
and tricuspid aortic valves in adults having isolated aortic valve tic valve replacement in the United States. Ann Thorac Surg.
replacement for aortic stenosis, with or without associated aortic 2015;100(5):1718–26. discussion 26-7.
regurgitation. Circulation. 2005;111(7):920–5. 21. Abdel-Wahab M, Mehilli J, Frerker C, Neumann FJ, Kurz T, Tolg R,
5. Otto CM, Nishimura RA. New ACC/AHA valve guidelines: align- et al. Comparison of balloon-expandable vs self-expandable valves
ing definitions of aortic stenosis severity with treatment recommen- in patients undergoing transcatheter aortic valve replacement: the
dations. Heart. 2014;100(12):902–4. CHOICE randomized clinical trial. JAMA. 2014;311(15):1503–14.
6. Kitai T, Honda S, Okada Y, Tani T, Kim K, Kaji S, et al. Clinical 22. Anderson RH.  Clinical anatomy of the aortic root. Heart.

outcomes in non-surgically managed patients with very severe ver- 2000;84(6):670–3.
sus severe aortic stenosis. Heart. 2011;97(24):2029–32. 23. Piazza N, de Jaegere P, Schultz C, Becker AE, Serruys PW,

7. Leon MB, Smith CR, Mack M, Miller DC, Moses JW, Svensson Anderson RH. Anatomy of the aortic valvar complex and its impli-
LG, et  al. Transcatheter aortic-valve implantation for aortic ste- cations for transcatheter implantation of the aortic valve. Circ
nosis in patients who cannot undergo surgery. N Engl J Med. Cardiovasc Interv. 2008;1(1):74–81.
2010;363(17):1597–607. 24. Jurencak T, Turek J, Kietselaer BL, Mihl C, Kok M, van Ommen
8. Messika-Zeitoun D, Aubry MC, Detaint D, Bielak LF, Peyser PA, VG, et al. MDCT evaluation of aortic root and aortic valve prior to
Sheedy PF, et  al. Evaluation and clinical implications of aortic TAVI. What is the optimal imaging time point in the cardiac cycle?
valve calcification measured by electron-beam computed tomogra- Eur Radiol. 2015;25(7):1975–83.
phy. Circulation. 2004;110(3):356–62. 25. Cavalcanti JS, de Melo NC, de Vasconcelos RS.  Morphometric
9. Cueff C, Serfaty JM, Cimadevilla C, Laissy JP, Himbert D, Tubach and topographic study of coronary ostia. Arq Bras Cardiol.
F, et al. Measurement of aortic valve calcification using multislice 2003;81(4):359–62, 5–8.
computed tomography: correlation with haemodynamic severity of 26. Hamdan A, Guetta V, Klempfner R, Konen E, Raanani E, Glikson
aortic stenosis and clinical implication for patients with low ejec- M, et  al. Inverse relationship between membranous septal length
tion fraction. Heart. 2011;97(9):721–6. and the risk of atrioventricular block in patients undergoing
10. Vahanian A, Alfieri O, Andreotti F, Antunes MJ, Baron-Esquivias transcatheter aortic valve implantation. JACC Cardiovasc Interv.
G, Baumgartner H, et al. Guidelines on the management of valvular 2015;8(9):1218–28.
heart disease (version 2012): the joint task force on the management 27. Vahl TP, Kodali SK, Leon MB. Transcatheter aortic valve replace-
of Valvular heart disease of the European Society of Cardiology ment 2016: a modern-day "through the looking-glass" adventure. J
(ESC) and the European Association for Cardio-Thoracic Surgery Am Coll Cardiol. 2016;67(12):1472–87.
(EACTS). Eur J Cardiothorac Surg. 2012;42(4):S1–44. 28. Leipsic J, Gurvitch R, Labounty TM, Min JK, Wood D, Johnson M,
11. Mohr FW, Holzhey D, Mollmann H, Beckmann A, Veit C, Figulla et  al. Multidetector computed tomography in transcatheter aortic
HR, et  al. The German aortic valve registry: 1-year results from valve implantation. JACC Cardiovasc Imaging. 2011;4(4):416–29.
13,680 patients with aortic valve disease. Eur J Cardiothorac Surg. 29. Kappetein AP, Head SJ, Genereux P, Piazza N, van Mieghem NM,
2014;46(5):808–16. Blackstone EH, et al. Updated standardized endpoint definitions for
12. Thourani VH, Suri RM, Gunter RL, Sheng S, O'Brien SM, Ailawadi transcatheter aortic valve implantation: the valve academic research
G, et al. Contemporary real-world outcomes of surgical aortic valve Consortium-2 consensus document (VARC-2). Eur J Cardiothorac
replacement in 141,905 low-risk, intermediate-risk, and high-risk Surg. 2012;42(5):S45–60.
patients. Ann Thorac Surg. 2015;99(1):55–61. 30. Achenbach S, Delgado V, Hausleiter J, Schoenhagen P, Min JK,
13. Bach DS, Kon ND, Dumesnil JG, Sintek CF, Doty DB.  Ten- Leipsic JA.  SCCT expert consensus document on computed
year outcome after aortic valve replacement with the freestyle tomography imaging before transcatheter aortic valve implanta-
stentless bioprosthesis. Ann Thorac Surg. 2005;80(2):480–6. tion (TAVI)/transcatheter aortic valve replacement (TAVR). J
discussion 6-7. Cardiovasc Comput Tomogr. 2012;6(6):366–80.
14. Siemieniuk RA, Agoritsas T, Manja V, Devji T, Chang Y, Bala 31. Wood DA, Tops LF, Mayo JR, Pasupati S, Schalij MJ, Humphries
MM, et al. Transcatheter versus surgical aortic valve replacement K, et al. Role of multislice computed tomography in transcatheter
in patients with severe aortic stenosis at low and intermediate risk: aortic valve replacement. Am J Cardiol. 2009;103(9):1295–301.
systematic review and meta-analysis. BMJ. 2016;354:i5130. 32. Apfaltrer P, Henzler T, Blanke P, Krazinski AW, Silverman JR,
15. Indraratna P, Tian DH, Yan TD, Doyle MP, Cao C. Transcatheter Schoepf UJ. Computed tomography for planning transcatheter aor-
aortic valve implantation versus surgical aortic valve replacement: tic valve replacement. J Thorac Imaging. 2013;28(4):231–9.
a meta-analysis of randomized controlled trials. Int J Cardiol. 33. Ludman PF, Moat N, de Belder MA, Blackman DJ, Duncan

2016;224:382–7. A, Banya W, et  al. Transcatheter aortic valve implantation in
16. Siontis GC, Praz F, Pilgrim T, Mavridis D, Verma S, Salanti G, et al. the United Kingdom: temporal trends, predictors of outcome,
Transcatheter aortic valve implantation vs. surgical aortic valve and 6-year follow-up: a report from the UK Transcatheter aortic
replacement for treatment of severe aortic stenosis: a meta-analysis valve implantation (TAVI) registry, 2007 to 2012. Circulation.
of randomized trials. Eur Heart J. 2016;37(47):3503–12. 2015;131(13):1181–90.
17. Leon MB, Smith CR, Mack MJ, Makkar RR, Svensson LG, Kodali 34. Ribeiro HB, Webb JG, Makkar RR, Cohen MG, Kapadia SR,
SK, et  al. Transcatheter or surgical aortic-valve replacement in Kodali S, et  al. Predictive factors, management, and clinical out-
intermediate-risk patients. N Engl J Med. 2016;374(17):1609–20. comes of coronary obstruction following transcatheter aortic valve
18. Okuyama K, Jilaihawi H, Kashif M, Takahashi N, Chakravarty T, implantation: insights from a large multicenter registry. J Am Coll
Pokhrel H, et al. Transfemoral access assessment for transcatheter Cardiol. 2013;62(17):1552–62.
aortic valve replacement: evidence-based application of computed 35. Blanke P, Schoepf UJ, Leipsic JA. CT in transcatheter aortic valve
tomography over invasive angiography. Circ Cardiovasc Imaging. replacement. Radiology. 2013;269(3):650–69.
2014;8(1) 36. Matsumoto S, Yamada Y, Hashimoto M, Okamura T, Yamada M,
19. Chandrasekhar J, Hibbert B, Ruel M, Lam BK, Labinaz M, Glover Yashima F, et  al. CT imaging before transcatheter aortic valve
C. Transfemoral vs non-transfemoral access for transcatheter aortic implantation (TAVI) using variable helical pitch scanning and its
41  CT in the Context of Transcatheter Aortic Valve Replacement 515

diagnostic performance for coronary artery disease. Eur Radiol. ment: a multicenter retrospective analysis. J Am Coll Cardiol.
2017;27(5):1963–70. 2012;59(14):1287–94.
37. Bittner DO, Arnold M, Klinghammer L, Schuhbaeck A, Hell MM, 52. Yang TH, Webb JG, Blanke P, Dvir D, Hansson NC, Norgaard BL,
Muschiol G, et  al. Contrast volume reduction using third gen- et  al. Incidence and severity of paravalvular aortic regurgitation
eration dual source computed tomography for the evaluation of with multidetector computed tomography nominal area oversiz-
patients prior to transcatheter aortic valve implantation. Eur Radiol. ing or undersizing after transcatheter heart valve replacement with
2016;26(12):4497–504. the Sapien 3: a comparison with the Sapien XT. JACC Cardiovasc
38. Felmly LM, De Cecco CN, Schoepf UJ, Varga-Szemes A,
Interv. 2015;8(3):462–71.
Mangold S, McQuiston AD, et  al. Low contrast medium-volume 53. Hamdan A, Guetta V, Konen E, Goitein O, Segev A, Raanani E,
third-generation dual-source computed tomography angiography et al. Deformation dynamics and mechanical properties of the aor-
for transcatheter aortic valve replacement planning. Eur Radiol. tic annulus by 4-dimensional computed tomography: insights into
2017;27(5):1944–53. the functional anatomy of the aortic valve complex and implica-
39. Formosa A, Santos DM, Marcuzzi D, Common AA, Prabhudesai tions for transcatheter aortic valve therapy. J Am Coll Cardiol.
V.  Low contrast dose catheter-directed CT angiography (CCTA). 2012;59(2):119–27.
Cardiovasc Intervent Radiol. 2016;39(4):606–10. 54. Murphy DT, Blanke P, Alaamri S, Naoum C, Rubinshtein R, Pache
40. Kok M, Turek J, Mihl C, Reinartz SD, Gohmann RF, Nijssen EC, G, et al. Dynamism of the aortic annulus: effect of diastolic versus
et  al. Low contrast media volume in pre-TAVI CT examinations. systolic CT annular measurements on device selection in trans-
Eur Radiol. 2016;26(8):2426–35. catheter aortic valve replacement (TAVR). J Cardiovasc Comput
41. Abe M, Morimoto T, Akao M, Furukawa Y, Nakagawa Y, Shizuta Tomogr. 2016;10(1):37–43.
S, et  al. Relation of contrast-induced nephropathy to long-term 55. Blanke P, Euringer W, Baumann T, Reinohl J, Schlensak C, Langer
mortality after percutaneous coronary intervention. Am J Cardiol. M, et al. Combined assessment of aortic root anatomy and aortoil-
2014;114(3):362–8. iac vasculature with dual-source CT as a screening tool in patients
42. Jochheim D, Schneider VS, Schwarz F, Kupatt C, Lange P, Reiser evaluated for transcatheter aortic valve implantation. AJR Am J
M, et  al. Contrast-induced acute kidney injury after computed Roentgenol. 2010;195(4):872–81.
tomography prior to transcatheter aortic valve implantation. Clin 56. Schmidkonz C, Marwan M, Klinghammer L, Mitschke M,

Radiol. 2014;69(10):1034–8. Schuhbaeck A, Arnold M, et  al. Interobserver variability of CT
43. Hausleiter J. Reduced contrast media volumes for CT imaging of angiography for evaluation of aortic annulus dimensions prior
TAVI candidates. Eur Heart J Cardiovasc Imaging. 2016;17(1):11. to transcatheter aortic valve implantation (TAVI). Eur J Radiol.
44. Al Ali AM, Altwegg L, Horlick EM, Feindel C, Thompson CR, 2014;83(9):1672–8.
Cheung A, et  al. Prevention and management of transcatheter 57. Korosoglou G, Gitsioudis G, Waechter-Stehle I, Weese J,

balloon-expandable aortic valve malposition. Catheter Cardiovasc Krumsdorf U, Chorianopoulos E, et al. Objective quantification of
Interv. 2008;72(4):573–8. aortic valvular structures by cardiac computed tomography angiog-
45. Messika-Zeitoun D, Serfaty JM, Brochet E, Ducrocq G, Lepage L, raphy in patients considered for transcatheter aortic valve implanta-
Detaint D, et al. Multimodal assessment of the aortic annulus diam- tion. Catheter Cardiovasc Interv. 2013;81(1):148–59.
eter: implications for transcatheter aortic valve implantation. J Am 58. Athappan G, Patvardhan E, Tuzcu EM, Svensson LG, Lemos PA,
Coll Cardiol. 2010;55(3):186–94. Fraccaro C, et  al. Incidence, predictors, and outcomes of aortic
46. Gurvitch R, Webb JG, Yuan R, Johnson M, Hague C, Willson regurgitation after transcatheter aortic valve replacement: meta-­
AB, et al. Aortic annulus diameter determination by multidetector analysis and systematic review of literature. J Am Coll Cardiol.
computed tomography: reproducibility, applicability, and implica- 2013;61(15):1585–95.
tions for transcatheter aortic valve implantation. JACC Cardiovasc 59. Binder RK, Webb JG, Willson AB, Urena M, Hansson NC,

Interv. 2011;4(11):1235–45. Norgaard BL, et  al. The impact of integration of a multidetector
47. Hansson NC, Thuesen L, Hjortdal VE, Leipsic J, Andersen HR, computed tomography annulus area sizing algorithm on outcomes
Poulsen SH, et  al. Three-dimensional multidetector computed of transcatheter aortic valve replacement: a prospective, multi-
tomography versus conventional 2-dimensional transesophageal center, controlled trial. J Am Coll Cardiol. 2013;62(5):431–8.
echocardiography for annular sizing in transcatheter aortic valve 60. Van Linden A, Kempfert J, Blumenstein J, Mollmann H, Kim WK,
replacement: influence on postprocedural paravalvular aortic regur- Alkaya S, et al. Manual versus automatic detection of aortic annulus
gitation. Catheter Cardiovasc Interv. 2013;82(6):977–86. plane in a computed tomography scan for transcatheter aortic valve
48. Mylotte D, Dorfmeister M, Elhmidi Y, Mazzitelli D, Bleiziffer implantation screening. Eur J Cardiothorac Surg. 2014;46(2):207–
S, Wagner A, et  al. Erroneous measurement of the aortic annular 12. discussion 12.
diameter using 2-dimensional echocardiography resulting in inap- 61. Patsalis PC, Al-Rashid F, Neumann T, Plicht B, Hildebrandt HA,
propriate CoreValve size selection: a retrospective comparison Wendt D, et  al. Preparatory balloon aortic valvuloplasty during
with multislice computed tomography. JACC Cardiovasc Interv. transcatheter aortic valve implantation for improved valve sizing.
2014;7(6):652–61. JACC Cardiovasc Interv. 2013;6(9):965–71.
49. Shibayama K, Mihara H, Jilaihawi H, Berdejo J, Harada K, Itabashi 62. Blanke P, Reinohl J, Schlensak C, Siepe M, Pache G, Euringer W,
Y, et  al. 3D assessment of features associated with transvalvular et al. Prosthesis oversizing in balloon-expandable transcatheter aor-
aortic regurgitation after TAVR: a real-time 3D TEE study. JACC tic valve implantation is associated with contained rupture of the
Cardiovasc Imaging. 2016;9(2):114–23. aortic root. Circ Cardiovasc Interv. 2012;5(4):540–8.
50. Wichmann JL, Varga-Szemes A, Suranyi P, Bayer RR, Litwin
63. Hell MM, Biburger L, Marwan M, Schuhbaeck A, Achenbach
SE, De Cecco CN, et  al. Transcatheter aortic valve replacement: S, Lell M, et  al. Prediction of fluoroscopic angulations for trans-
imaging techniques for aortic root sizing. J Thorac Imaging. catheter aortic valve implantation by CT angiography: influ-
2015;30(6):349–58. ence on procedural parameters. Eur Heart J Cardiovasc Imaging.
51. Willson AB, Webb JG, Labounty TM, Achenbach S, Moss R, 2017;18(8):906–14.
Wheeler M, et  al. 3-dimensional aortic annular assessment by 64. Holzamer A, Sitka E, Hengstenberg C, Schmid C, Debl K, Maier
multidetector computed tomography predicts moderate or severe L, et al. Multislice computed tomography-based prediction of the
paravalvular regurgitation after transcatheter aortic valve replace- implantation plane in transcatheter aortic valve implantation: deter-
516 E. Konen et al.

mination of the line of perpendicularity and the implanter's views. and inoperable patients with severe aortic stenosis. Circulation.
Eur J Cardiothorac Surg. 2015;48(6):879–85. discussion 85-6. 2016;134(2):130–40.
65. Azzalini L, Ghoshhajra BB, Elmariah S, Passeri JJ, Inglessis I, 79. Kodali S, Williams MR, Doshi D, Hahn RT, Humphries KH, Nkomo
Palacios IF, et al. The aortic valve calcium nodule score (AVCNS) VT, et  al. Sex-specific differences at presentation and outcomes
independently predicts paravalvular regurgitation after transcathe- among patients undergoing transcatheter aortic valve replacement:
ter aortic valve replacement (TAVR). J Cardiovasc Comput Tomogr. a cohort study. Ann Intern Med. 2016;164(6):377–84.
2014;8(2):131–40. 80. Toggweiler S, Gurvitch R, Leipsic J, Wood DA, Willson AB,

66. Harbaoui B, Montoy M, Charles P, Boussel L, Liebgott H, Girerd Binder RK, et al. Percutaneous aortic valve replacement: vascular
N, et al. Aorta calcification burden: towards an integrative predic- outcomes with a fully percutaneous procedure. J Am Coll Cardiol.
tor of cardiac outcome after transcatheter aortic valve implantation. 2012;59(2):113–8.
Atherosclerosis. 2016;246:161–8. 81. Barbash IM, Finkelstein A, Barsheshet A, Segev A, Steinvil A,
67. Jilaihawi H, Makkar RR, Kashif M, Okuyama K, Chakravarty T, Assali A, et al. Outcomes of patients at estimated low, intermediate,
Shiota T, et  al. A revised methodology for aortic-valvar complex and high risk undergoing transcatheter aortic valve implantation for
calcium quantification for transcatheter aortic valve implantation. aortic stenosis. Am J Cardiol. 2015;116(12):1916–22.
Eur Heart J Cardiovasc Imaging. 2014;15(12):1324–32. 82. Chieffo A, Buchanan GL, Van Mieghem NM, Tchetche D,

68. Khalique OK, Hahn RT, Gada H, Nazif TM, Vahl TP, George Dumonteil N, Latib A, et al. Transcatheter aortic valve implanta-
I, et  al. Quantity and location of aortic valve complex calcifi- tion with the Edwards SAPIEN versus the Medtronic CoreValve
cation predicts severity and location of paravalvular regurgita- Revalving system devices: a multicenter collaborative study: the
tion and frequency of post-dilation after balloon-expandable PRAGMATIC plus initiative (pooled-RotterdAm-Milano-Toulouse
transcatheter aortic valve replacement. JACC Cardiovasc Interv. in collaboration). J Am Coll Cardiol. 2013;61(8):830–6.
2014;7(8):885–94. 83. Kapadia SR, Svensson LG, Roselli E, Schoenhagen P, Popovic
69. Koh EY, Lam KY, Bindraban NR, Cocchieri R, Planken RN, Koch Z, Alfirevic A, et al. Single center TAVR experience with a focus
KT, et al. Aortic valve calcification as a predictor of location and on the prevention and management of catastrophic complications.
severity of paravalvular regurgitation after transcatheter aortic valve Catheter Cardiovasc Interv. 2014;84(5):834–42.
implantation. Interact Cardiovasc Thorac Surg. 2015;20(3):345–50. 84. Beohar N, Kirtane AJ, Blackstone E, Waksman R, Holmes D Jr,
70. Fujita B, Kutting M, Seiffert M, Scholtz S, Egron S, Prashovikj E, Minha S, et al. Trends in complications and outcomes of patients
et al. Calcium distribution patterns of the aortic valve as a risk fac- undergoing transfemoral transcatheter aortic valve replacement:
tor for the need of permanent pacemaker implantation after trans- experience from the PARTNER continued access registry. JACC
catheter aortic valve implantation. Eur Heart J Cardiovasc Imaging. Cardiovasc Interv. 2016;9(4):355–63.
2016;17(12):1385–93. 85. Takagi H, Umemoto T, Group A.  Impact of paravalvular aortic
71. Hansson NC, Norgaard BL, Barbanti M, Nielsen NE, Yang TH, regurgitation after transcatheter aortic valve implantation on sur-
Tamburino C, et  al. The impact of calcium volume and distribu- vival. Int J Cardiol. 2016;221:46–51.
tion in aortic root injury related to balloon-expandable trans- 86. Popma JJ, Gleason TG, Yakubov SJ, Harrison JK, Forrest JK,
catheter aortic valve replacement. J Cardiovasc Comput Tomogr. Maini B, et al. Relationship of annular sizing using multidetector
2015;9(5):382–92. computed tomographic imaging and clinical outcomes after self-­
72. Sievers HH, Schmidtke C.  A classification system for the bicus- expanding corevalve transcatheter aortic valve replacement. Circ
pid aortic valve from 304 surgical specimens. J Thorac Cardiovasc Cardiovasc Interv. 2016;9(7)
Surg. 2007;133(5):1226–33. 87. Makkar RR, Fontana G, Jilaihawi H, Chakravarty T, Kofoed KF, de
73. Jilaihawi H, Chen M, Webb J, Himbert D, Ruiz CE, Rodes-Cabau Backer O, et al. Possible subclinical leaflet thrombosis in biopros-
J, et al. A bicuspid aortic valve imaging classification for the TAVR thetic aortic valves. N Engl J Med. 2015;373(21):2015–24.
era. JACC Cardiovasc Imaging. 2016;9(10):1145–58. 88. Holmes DR Jr, Mack MJ, Kaul S, Agnihotri A, Alexander KP,
74. Mylotte D, Lefevre T, Sondergaard L, Watanabe Y, Modine T, Dvir Bailey SR, et al. ACCF/AATS/SCAI/STS expert consensus docu-
D, et al. Transcatheter aortic valve replacement in bicuspid aortic ment on transcatheter aortic valve replacement: developed in col-
valve disease. J Am Coll Cardiol. 2014;64(22):2330–9. labration with the American Heart Association, American Society
75. Webb JG, Chandavimol M, Thompson CR, Ricci DR, Carere RG, of Echocardiography, European Association for Cardio-Thoracic
Munt BI, et al. Percutaneous aortic valve implantation retrograde Surgery, Heart Failure Society of America, mended hearts, Society
from the femoral artery. Circulation. 2006;113(6):842–50. of Cardiovascular Anesthesiologists, Society of Cardiovascular
76. Vahanian A, Alfieri O, Al-Attar N, Antunes M, Bax J, Cormier Computed Tomography, and Society for Cardiovascular Magnetic
B, et  al. Transcatheter valve implantation for patients with aortic Resonance. J Thorac Cardiovasc Surg. 2012;144(3):e29–84.
stenosis: a position statement from the European Association of 89. Himbert D, Brochet E, Serfaty JM, Vahanian A. Contained aortic
Cardio-Thoracic Surgery (EACTS) and the European Society of root rupture after transcatheter aortic valve implantation. Eur Heart
Cardiology (ESC), in collaboration with the European Association J. 2010;31(24):2995.
of Percutaneous Cardiovascular Interventions (EAPCI). Eur Heart 90. Oda T, Okamura T, Miyazaki Y, Nakamura T, Mikamo A, Wada
J. 2008;29(11):1463–70. Y, et  al. 1-year follow-up of contained aortic root rupture after
77. Genereux P, Head SJ, Van Mieghem NM, Kodali S, Kirtane AJ, transcatheter aortic valve replacement. JACC Cardiovasc Interv.
Xu K, et  al. Clinical outcomes after transcatheter aortic valve 2016;9(3):295–6.
replacement using valve academic research consortium definitions: 91. Eggebrecht H, Schmermund A, Kahlert P, Erbel R, Voigtlander
a weighted meta-analysis of 3,519 patients from 16 studies. J Am T, Mehta RH. Emergent cardiac surgery during transcatheter aor-
Coll Cardiol. 2012;59(25):2317–26. tic valve implantation (TAVI): a weighted meta-analysis of 9,251
78. Herrmann HC, Thourani VH, Kodali SK, Makkar RR, Szeto
patients from 46 studies. EuroIntervention. 2013;8(9):1072–80.
WY, Anwaruddin S, et  al. One-year clinical outcomes with 92. Rezq A, Basavarajaiah S, Latib A, Takagi K, Hasegawa T, Figini
SAPIEN 3 transcatheter aortic valve replacement in high-risk F, et al. Incidence, management, and outcomes of cardiac tampon-
41  CT in the Context of Transcatheter Aortic Valve Replacement 517

ade during transcatheter aortic valve implantation: a single-center 94. Dvir D, Webb JG, Bleiziffer S, Pasic M, Waksman R, Kodali S,
study. JACC Cardiovasc Interv. 2012;5(12):1264–72. et al. Transcatheter aortic valve implantation in failed bioprosthetic
93. Brennan JM, Edwards FH, Zhao Y, O'Brien S, Booth ME, Dokholyan surgical valves. JAMA. 2014;312(2):162–70.
RS, et  al. Long-term safety and effectiveness of mechanical ver- 95. Bapat V. Valve-in-valve apps: why and how they were developed
sus biologic aortic valve prostheses in older patients: results from and how to use them. EuroIntervention. 2014;10(Suppl U):U44–51.
the Society of Thoracic Surgeons adult cardiac surgery National 96. Hamdan A, Wellnhofer E, Konen E, Kelle S, Goitein O, Andrada B,
Database. Circulation. 2013;127(16):1647–55. et al. Coronary CT angiography for the detection of coronary artery
stenosis in patients referred for transcatheter aortic valve replace-
ment. J Cardiovasc Comput Tomogr. 2015;9(1):31–41.
 CT for Minimally Invasive Repair
of Mitral Valve and Other Structural 42
Heart Diseases

John F. Mooney, Philipp Blanke, Shaw Hua Kueh,

Stephanie Sellers, and Jonathon A. Leipsic

Minimally invasive procedures on the mitral valve and other replacement the gold standard for treating significant mitral
cardiac structures offer a feasible alternative when the risk of valve disorders [2]. However the risk of adverse peri- or
surgery is unacceptably high. In particular, the success of post-­operative morbidity and mortality may be unaccept-
transcatheter aortic valve replacement (TAVR) has led to the ably high in some patients. Risk categorization is commonly
development of transcatheter mitral valve replacement expressed with risk prediction tools like the Society of
(TMVR) implantation. With a high burden of mitral disease Thoracic Surgeons (STS) risk prediction model [3]. This
in an ageing population with significant comorbidities to includes pre-­surgical morbidities like advanced age, under-
contraindicate surgery, percutaneous repair or replacement is lying kidney failure, pre-existing heart failure as well as
predicted to progressively expand. acute heart failure requiring active treatment with inotropes
The mitral valve poses a challenge to imaging and device and balloon pump. In addition to the STS risk model, other
implantation due to its complex shape and proximity to the risk factors that may preclude surgical intervention include
left ventricular outflow tract (LVOT). CT offers excellent patient frailty, highly calcified aorta, severe pulmonary
spatial and temporal resolution to define the anatomy of hypertension and severe liver disease [4]. Patient character-
interest and relationship to surrounding structures. It can istics also increase risk of morbidity and mortality, with
help determine accurate sizing for devices, provide images older patients, those with low LVEF and those with an
that correspond with fluoroscopy and screen for adverse risk increased Charlson comorbidity index score more likely to
and contraindication to procedures. Similar to the experience be denied surgery [5].
with TAVR, pre-procedural CT has the potential to accu- Percutaneous repair or replacement of the mitral valve
rately identify those at risk of complications and aid plan- presents as a viable option for patients deemed too high a
ning to minimize adverse outcomes. In addition to mitral surgical risk. Current American College of Cardiology/
procedures, CT is also used prior to structural procedures on American Heart Association (ACC/AHA) guidelines [4] list
atrial and septal defect, planning for pulmonary vein isola- criteria for patients with mitral regurgitation that are poten-
tion and pulmonary valve replacement. tially suitable for minimally invasive approach and include:

• Severely symptomatic heart failure (NYHA class III/IV)

 inimally Invasive Repair or Replacement
M despite medical therapy.
of the Mitral Valve • Chronic mitral regurgitation caused by leaflet degenera-
tion (primary MR) as opposed to MR resulting from left
Procedural Overview ventricular remodelling (secondary MR).
• MR graded as moderate/severe to severe on
Mitral regurgitation is prevalent in over 10% of individuals echocardiography.
aged greater than 75  years [1], with surgical repair or • The anatomy permits repair or replacement.
• The patient otherwise has a reasonable life expectancy.
• Comorbidities place the patient as an unacceptable surgi-
J. F. Mooney · P. Blanke · S. H. Kueh · S. Sellers
Department of Radiology, University of British Columbia, cal risk.
Vancouver, BC, Canada
J. A. Leipsic (*)
Transcutaneous mitral valve procedures have predomi-
Department of Radiology, St. Paul’s Hospital, Providence nantly involved mitral repair rather than replacement. The
Healthcare, Vancouver, BC, Canada MitraClip™ device (Abbott Vascular) has a widespread use, is
e-mail: [email protected] the only FDA-approved device and has ongoing registry-­based

© Humana Press 2019 519

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_42
520 J. F. Mooney et al.

evidence to its utility [6]. The concept of the MitraClip is Technologies, Irvine CA, USA), Tendyne™ (Tendyne
derived from the Alfieri surgical repair of the mitral valve, Holdings, Roseville, MN, USA) and Twelve™ (Twelve Inc.
where a double stitch between the anterior and posterior leaf- Redwood City CA, USA). The devices incorporate distinct
lets fixes the cusps where regurgitation is maximal. A double features to aid positioning. These include atrial skirts to aid
orifice is thereby created with two smaller inflow orifices. apposition against the left atrial wall, tabs to engage the
Access is gained by the femoral vein and transeptal puncture. myocardium, anchors to engage the annulus and leaflets and
Experience with the MitraClip device has been presented in paddles to engage the leaflets or apical tether. Transeptal
several registries and trials. The EVEREST I trial assessed fea- implantation has not been successful, and the devices are
sibility and followed 24 patients with clips. Of these 13 had designed for transapical delivery [10].
reduced MR sustained at 6  months [7]. The subsequent
EVEREST II trial randomly assigned 279 patients to either sur-
gical repair or percutaneous repair in a 2:1 ratio [8]. Outcomes  T for Minimally Invasive Repair or
C
were assessed at 30  days and 12  months with primary end- Replacement of the Mitral Valve
points being survival, 30-day adverse outcomes, requirement
for additional surgery and continued grade 3/4 MR. The overall Pre-procedural CT is commonly used to determine anatomi-
findings were of greater efficacy in the surgical cohort (73% vs cal feasibility and to help with device sizing, deployment and
55%), though with less adverse outcomes within the percutane- vascular access for TAVR. It can also aid in defining adverse
ous repair group (15% vs 48%). The safety of percutaneous features such as subannular calcification, risk of coronary
repair was confirmed in an analysis of high-risk candidates obstruction and excess annular dilation. Integration of CT
from the EVEREST II cohort. A total of 78 patients with esti- imaging has led to improved outcomes after TAVR, with
mated surgical mortality of 12% or greater were compared to a reduction in paravalvular leaks [12, 13] and vascular compli-
cohort that was screened but not enrolled [9]. The 12-month cations [14, 15]. The successful utility of CT in optimizing
survival was 78% in the high-­risk group receiving the device sizing for TAVR has led to its adoption for planning in TMVR.
and 55% in the matched comparison group. In the treatment Planning for mitral valve procedures using CT offers sev-
arm, there were improvements at 12 months in MR grade ≤2 eral advantages. It can accurately define annular dimensions
(78%), left ventricular end-­diastolic volume (172 ml to 140 ml, in a reproducible fashion, calcification, and the anatomical
p = 0.001), NYHA class (89% III/IV at baseline, 74% I/II at landing zone and help with apical access and the detailed
12 months, p < 0.0001) and rate of hospitalization (from 0.59 anatomy of the LVOT. It can also provide coplanar projec-
to 0.32, p = 0.034) [10]. Current guidelines recommend percu- tion angles to assist implantation under fluoroscopy. Imaging
taneous mitral repair in patients with chronic primary MR, who considerations for a transcatheter mitral valve repair include
are severely symptomatic despite optimal medical therapy and a planimetric area  >4.0  cm2, minimal leaflet calcification
at prohibitive risk for surgery [4]. where the device will grasp, limited flail width (<15 mm and
The safety and efficacy of the deployment of a transcatheter flail gap <10 mm) and large central jet >6 cm2 or >30% of the
aortic valve for the relief of native mitral stenosis in the setting left atrial area (generally defined by echocardiogram) [8].
of severe mitral annular calcification (MAC) have been pre- Mitral valve replacement requires determination of annular
sented by a global registry to measure technique and outcomes area and risk of LVOT obstruction. Oversizing can result in
[11]. A total of 64 patients with severe mitral annular calcifica- LVOT obstruction as well as potential rupture, and
tion (NYHA III–IV) and an average Society of Thoracic ­undersizing can result in valve embolization [11].
Surgeons score of 14.4, from 32 centres, underwent compas-
sionate deployment of a balloon-expandable transcatheter aor-
tic device. The transapical (43.8%) followed by the transeptal CT Acquisition
(40.6%) and transatrial (15.6%) route was used. The device
was deployed successfully in 72% of cases, with 11 requiring a CT provides image quality of high-spatial and high-temporal
second valve. A total of six patients had severe left ventricular resolution which is necessary for accurate sizing of the mitral
outflow tract obstruction causing haemodynamic compromise, valve. The constant motion of the mitral apparatus during the
and of these two died. The 30-day all-cause mortality was cardiac cycle requires rapid image acquisition with multide-
29.7% (12.5% cardiovascular cause, 17.2% non-cardiovascu- tector computed tomography (MDCT), typically with at least
lar cause), and of the survivors, 84% had improvement in a 64-slice scanner. Image acquisition is synchronized to the
symptoms (NYHA I–II). The frail state of many of the patients ECG cycle and retrospective acquisition throughout the car-
was observed as a significant contributor to mortality. diac cycle performed. Contrast enhancement is required to
Devices specific for TMVR are currently undergoing enhance tissue boundaries. Retrospective acquisition allows
feasibility trials. These include Tiara™ (Neovasc Inc.
­ the whole cardiac cycle to be imaged and any time point to
Richmond BC, Canada), CardiAQ™ (CardiAQ Valve be reconstructed. This allows assessment throughout the car-
42  CT for Minimally Invasive Repair of Mitral Valve and Other Structural Heart Diseases 521

diac cycle albeit at higher radiation dose [16]. Image analysis Two papillary muscles (anterolateral and postero-medial)
is performed with multiplanar reformation (MPR) to provide arise from the free wall of the left ventricle. Fibrous chordae
thin two-dimensional planes through the three-dimensional run from each papillary muscle to the free edge of both leaf-
data set allowing reconstruction in any plane [16]. lets. During systole, papillary muscle contraction brings the
leaflets to coapt and prevent flow into the left atrium. In dias-
tole, papillary muscle relaxation allows the leaflets to open.
Mitral Valve Anatomy

The mitral valve poses a challenge to both imaging and pros- Determining the Mitral Annular Area
thetic implantation due to its complex anatomy, larger size
and proximity to the left ventricular outflow tract. It is a com- MDCT also plays a critical role, owing to its exceptional spatial
plex three-dimensional structure composed of a saddle-­ resolution, for the determination of the appropriateness of indi-
shaped annulus, anterior and posterior leaflets, papillary vidual anatomy for TMVR. In particular, in the early days of
muscles and chordae. TMVR, there are limited device sizes making accurate annular
The annulus is a three-dimensional, nonplanar structure sizing and segmentation essential. Annular area can vary
with two distinct peaks. It serves as the conduit between the widely between patients, with estimates of annular area rang-
left atrium and left ventricle and is incorporated in the left ing from 7 to 10 cm2 in healthy subjects [19] and between 11
ventricle myocardium [17]. The posterior peak is at the inser- and 20 cm2 in patients with LV dilatation and functional MR.
tion of the posterior leaflet and extends from the lateral to [19] This is reflected chiefly in an increased distance between
medial trigone. The anterior peak is defined by the intervalvu- the septal and lateral, also known as A2 to P2, distance com-
lar fibrosa (aka aorto-mitral curtain or aortic continuity) that pared to the intercommisural distance [20]. Accurately defining
lies between the mitral and aortic annulus. Between these mitral dimensions has improved with 3D imaging techniques.
peaks lies the nadir at the trigones. The intervalvular fibrosa Traditional 2D imaging techniques rely on accurate orientation
lies between the mitral and aortic annulus and is anchored to to account for the nonplanar structure of the mitral apparatus,
the mitral annulus at the lateral and medial trigones [18]. and estimated dimensions varied dependent on orientation
The mitral valve is unique as it is the only cardiac valve to [21]. A lack of consensus on defining the anterior boundary has
possess two leaflets – the anterior and posterior leaflets. The also limited consistency in 2D measurement [22].
anterior leaflet is semi-circular in shape, whereas the poste- The use of 3D segmentation has improved accurate depic-
rior leaflet is rectangular. At each end are clefts, or commis- tion of the mitral annulus [23] and is achieved by generating
sures, which divide the leaflets. The posterior leaflet is a cubic spline of manually placed points along the 3D con-
divided further by clefts, creating three scallops labelled P1, tour. The LV long axis is identified, and seeding points are
P2 and P3 from lateral to medial. Though less distinct, by placed manually along the insertion of the posterior mitral
convention, the anterior leaflet is also categorized into cor- leaflet and along the fibrous continuity with stepwise rota-
responding scallops A1, A2 and A3. tion at 22.5° to insert each point. Figure 42.1 illustrates the

a b

Fig. 42.1  Acquisition of the annular plane. (a) The long axis of the LV tion of the posterior mitral leaflet and contour of the fibrous continuity,
is identified and from this the short axis view obtained. (b) Rotation with rotation of the LV long axis by 22.5° in stepwise manner
around the long axis allows positioning of seed points along the inser-
522 J. F. Mooney et al.

acquisition of the long-axis and short-axis views, whereby step is to identify the trigones, where the anterior mitral
rotation around the long axis allows planting of seed points valve leaflet separates from the atrioventricular junction to
in the short axis. In Fig. 42.2 the seed points have been set, follow the fibrous continuity. Figure 42.3 demonstrates how
and a method of least squares allows projection of the 3D the trigones are identified using the 3D saddle-shaped model
image onto a 2D plane by defining the geometric plane. reconstruction in short- and long-axis views.
The saddle-shaped annulus has long been shown to best With the trigones defined, the anterior and posterior seg-
define the geometry of the mitral annulus, but if one were to ments can be demarcated and allow comparison between
size a transcatheter device on the basis of this saddle-shaped saddle-shaped and D-shaped models. The saddle-shaped
annulus, the device would almost certainly result in LVOT model includes the anterior segment for the measurement of
obstruction. As a result, based on CT analyses, it has been dimensions, and the D-shaped model excludes structure
proposed that the annulus be truncated anteriorly at the med- anterior to the trigone-to-trigone line, including the aorto-­
ical and lateral trigones to create a D-shaped annulus that mitral continuity and anterior aortic peaks. The intercommi-
excludes the annular apparatus [24]. This ‘D’-shaped model sural distance is parallel to the trigone-to-trigone distance
uses CT to define the trigones and then the reconstructed 2D and perpendicular to the septal-to-lateral distance, transect-
dimensions using a method of least squares plane. The initial ing it through the virtual plane at the centroid [24].

a b

c d

Fig. 42.2  Segmentation of the annulus. (a) Seeding points demon- posed on long-axis image. (d) Saddle-shaped annulus on short-axis
strated in long axis. (b) The 16 seeding points in short axis. (c) view showing intercommisural (IC) and septa-lateral (SL) lines
3-­dimensional reconstruction of the saddle-shaped annulus superim-
42  CT for Minimally Invasive Repair of Mitral Valve and Other Structural Heart Diseases 523

a b

c d

Fig. 42.3  Identification of trigones and definition of the D-shaped view with lateral (blue dot) and septal (green dot) trigones. (c, d)
annulus. (a) Rotation on the long axis identifies trigones, where the Construction of D-shaped annulus with omission of anterior segment in
mitral annulus separates from the intervalvular fibrosa. (b) Short-axis both long-axis (c) and short-axis (d) views

A D-shaped model is considered to have advantages over tion due to dilated cardiomyopathy. Those with primary
the saddle-shaped model. Firstly, it offers a standardized and MR (in this case mitral valve prolapse) were found to have
reproducible method, with low interobserver variability, of an absolute annular area 18% larger than those with sec-
estimating annular dimensions. Secondly, it resembles the ondary (functional) MR.  Whether this in part relates to
cross-sectional area of TMVR devices which is of great rel- annular disjunction or frank septal-to-lateral enlargement is
evance for correct sizing and implantation. This in turn uncertain and will require further investigation. This is
reduces the risk of over- and undersizing, with the inherent despite those patients with functional MR having overall
risks of LVOT obstruction or paravalvular leak and device larger LV volumes [20].
embolization, respectively. Thirdly, compared to the saddle-­
shaped model, there is less projection into the LVOT and the
risks this poses [24]. Landing Zone Characteristics
The D-shaped method has identified variation in annular
area between primary and secondary MR [20]. Primary The surrounding anatomy of the annulus can be accurately
causes relate to leaflet dysfunction and include leaflet pro- defined with CT, including insertion at the atrioventricular
lapse, degeneration and destruction due to endocarditis or junction, calcification and leaflets. The anatomy of the atrio-
fibrosis. Secondary causes relate most commonly due to LV ventricular junction can vary between primary and second-
dilatation with or without prior infarction with tethering of ary MR.  Those with primary MR related to mitral valve
the posterior leaflet with chordae rupture or annular dilata- prolapse may have mitral valvular disjunction, where the
524 J. F. Mooney et al.

mitral leaflet insertion is displaced back into the atrium [25]. and the LVOT long axis (aorto-mitral angle) poses risk where
In secondary MR, changes in the surrounding myocardium the angle is increased, and near-parallel angles pose minimal
can lead to leaflet tethering, reduced coaptation of leaflets risk, whereas perpendicular angulation poses maximal risk.
and creation of a posterior myocardial shelf due to basal Secondly, a smaller overall left ventricular size may not be
remodelling. Figure 42.4 demonstrates anatomical variation able to accommodate an implanted valve without threatening
between valvular disjunction present in mitral valve prolapse the LVOT.  A third contribution to LVOT obstruction is
(primary MR) and the presence of a prominent atrioventricu- hypertrophy at the basal septum that can accentuate the
lar shelf in secondary mitral regurgitation [20]. aorto-mitral angle and consequently diminish the LVOT
CT can also readily demonstrate calcification and leaflet area. Finally, the prosthetic device can protrude, or flare, into
anatomy. Calcification on the mitral apparatus is a relatively the LVOT and limit the cross-sectional area for systolic
common finding with advanced age. Extent of calcification blood flow. Figure 42.5 compares a neo-LVOT with low risk
can range from localized and mild to severe extending onto of obstruction against one with very high risk of
leaflets and rarely to caseous formation. The exact impact of obstruction.
annular calcification on self-expanding TMVR devices Determining the risk of LVOT obstruction is very impor-
remains to be established, though significant calcification tant prior to TMVR.  CT can provide insight by modelling
particularly when irregular is generally considered a contra- device deployment and segmenting the residual LVOT with
indication to percutaneous valve implantation. Similarly, the virtual TMVR stent in place [27]. This predicted area of
leaflet length, position and relationship to surrounding anat- the neo-LVOT incorporates the influence of the aorto-mitral
omy can be presented by CT. The length and anatomy of the angle, LV size and basal hypertrophy and can indicate over-
leaflets can be determined in mid-diastole. The distance of all risk of obstruction. This approach is however limited as
the papillary muscles to the leaflets is also relevant, as closely no discrete LVOT area has been established at which the risk
tethered leaflets or direct papillary insertion to the leaflets of obstruction increases. It also will vary with loading condi-
may interfere with device anchoring [26]. tions, residual LV function and the integrity of the choral
apparatus. The stage of the cardiac cycle is also influential on
obstruction, yet its role in determining LVOT remains
Prediction of LVOT Obstruction unclear – minimal LVOT area occurs at end-systole – though
this may have no influence on obstruction as maximal flow
LVOT obstruction was a key cause of morbidity and mortal- occurs early to mid-systole, proposing this stage may be
ity in the global registry for TMVR [11]. Proximity of the more influential on risk on obstruction.
mitral annulus combined with unsuitable anatomy can lead
to flow limitation or obstruction after implantation of pros-
thetic MV. Specific risk for LVOT obstruction relates to the Fluoroscopic Angulation
anatomical relationship of the aorto-mitral angle, the LV
size, the configuration of the interventricular septum and the CT can also assist valve implanters with coplanar views to
device itself. Firstly, the angle between the mitral outflow assist coaxial device deployment, already utilized in TAVR

a b

Fig. 42.4  The atrioventricular junction in (a) mitral valve prolapse with disjunction between MV insertion and atrioventricular junction (red
arrow) and (b) functional mitral regurgitation with formation of an atrioventricular shelf adjacent to the MV insertion point (yellow arrow)
42  CT for Minimally Invasive Repair of Mitral Valve and Other Structural Heart Diseases 525

1 2

3 4

Fig. 42.5  CT data sets demonstrating prediction of neo-LVOT obstruc- shows an optimal LVOT tract and image 2 shows obstruction. In short-­
tion. Comparison between low- (1,3) and high-(2,4)risk cases for LVOT axis views, the cross-sectional area of the neo-LVOT is shown. Image 3
obstruction. The green cylinder simulates a 29  mm diameter device shows an optimal neo-LVOT area (highlighted in red) and image 4
implant. The long-axis views demonstrate the neo-LVOT (shown by red shows complete obstruction
bar), and the green bar represents the cross-sectional area. Image 1

[28]. Plane segmentation on CT imaging can guide on the guide the proceduralist as to how to position and direct his/
optimal angle (cranial or caudal) that corresponds to fluoro- her hands to allow for a perpendicular deployment. In
scopic projection, presenting implanters with an anatomical Fig. 42.7, the ‘en face’ images and required C-arm projec-
‘roadmap’ prior to the procedure of what they can expect to tions are demonstrated, and Fig. 42.8 shows the projections
see for coaxial deployment at corresponding C-arm angles required to emulate the septal-to-lateral line. It is worth not-
[29]. The native annulus lies in an anterior-superior plane ing that fluoroscopic angulation between the TT and SL
with angulation to the right. Projection angles have been views (as indicated along the red sigmoidal line in the fig-
assessed that correspond to the virtual lines that transect the ure); small changes in RAO angulation require correspond-
annular plane  – the septal-to-lateral line and trigone-to-­ ing larger changes in cranial or caudal angulation to maintain
trigone line. Figure 42.6 shows corresponding CT and fluo- a coplanar angle.
roscopic views with superimposed CT-derived annular The addition of a coronary sinus wire can aid in deploy-
geometry. Angulation of the arm to show the trigone-to-­ ment, acting as a landmark by aligning with P1 along the
trigone line requires steep angulation (in this case RAO 77.3 floor of the coronary sinus and with P2. The guidewire path
CAU 51.8) that is not clinically practical. can also be represented in pre-procedural CT to help define
In addition to the S curve of the mitral annulus, an en face the patient-specific relationship between the coronary sinus
angle is also commonly provided. This angle is meant to help and P2 rather than relying on assumptions [29].
526 J. F. Mooney et al.

a b

c d

Fig. 42.6  Predicting fluoroscopic angulation along the trigone-to-­ sponding fluoroscopy. (d) curve demonstrating optimal fluoroscopic
trigone line. (a) shows transection of the annular plane at A2/P2. (b) projection angles as identified by yellow dot – note the T-T projection
demonstrates the en face view of the annular plane, with the D-shaped angles here are RAO 77.3, CAU 51.8 which can be clinically impracti-
annulus shown. (c) superimposes the reconstructed annulus on corre- cal to emulate

Defining Access Point Pulmonary Valve Implantation

With current devices for TMVR often deployed transapi- Pulmonary valve replacement is generally required with
cally, pre-procedural CT can provide imaging of the anat- worsening right function that can occur in patients with pre-
omy required for catheter entry. A trajectory that is coaxial vious congenital heart surgery. It offers an alternative
to the annular plan and bisects the centroid of the annulus approach to redo surgery, angioplasty or stenting. The indi-
can be provided that extends to the chest wall (Fig.  42.9) cations for replacement include severe pulmonary regurgita-
[10]. The relationship of this line to the corresponding line tion, severe RV dilatation, severe RV dysfunction or
from the centroid to the ‘true’ apex can also be defined – this worsening exercise function.
is often projected medial or inferior to the coaxial line. The Pre-procedural imaging with CT can be used to assess the
trajectory line can then be assessed for proximity to papil- anatomy and aid with sizing [30]. Morphology of the RV
lary muscles, myocardial scar, the left anterior descending outflow tract can be assessed for type and suitability for
(LAD) coronary artery and any coronary grafts. The most implant. Funnel-shaped outflow tracts (type 1) are high risk
appropriate intercostal space for chest wall access can also for valve embolization and not suitable, whereas tubular-­
be determined. shaped tracts offer better support. The conduit length from
42  CT for Minimally Invasive Repair of Mitral Valve and Other Structural Heart Diseases 527

Fig. 42.7  Fluoroscopic angulation aimed to guide the positioning of the proceduralist’s hands and the direction to proceed with the delivery
catheter to achieve a perpendicular deployment (LAO 52.3, CAU 26.7)

the pulmonary bifurcation can also be determined, as can the anatomical structure. A rim of <3 mm is considered a contra-
internal diameter of the conduit. indication for deployment, and for PFOs the minimum dis-
Finally, proximity to the coronary vessels can be determined tance to the aorta or SVC should not exceed 9 mm [31].
to avoid coronary compression with device deployment. CT can also be used for sizing prior to percutaneous ven-
tricular septal defect (VSD) closure. The size and shape can
be determined, measured in end-diastole. Proximity to the
 ercutaneous Closure of Atrial Septal
P valves and the AV node can also be determined, which may
Defects and Ventricular Septal Defects contradict closure [32].

CT can be used for planning of atrial septal defect (ASD) or

patent foramen ovale (PFO) closure. Percutaneous devices Pulmonary Vein Isolation
employ a central waist with two umbrella-like discs at either
end to fix the device to the atrial wall. With closure, transep- Radiofrequency ablation to isolate arrhythmogenic foci sur-
tal flow and clot embolization can be disrupted. rounding the pulmonary veins is an interventional procedure
Pre-procedural CT can determine defect size, rim dis- to treat atrial fibrillation. CT has a role in defining the pulmo-
tance, distorted or thinned septum and the presence of throm- nary vein anatomy, presence of thrombus [33] and proximity
bus. Accurate sizing is required to determine device size, and to the oesophagus [34]. Additionally, CT imaging can be
coronal oblique images in end-systole are used. Rim mea- merged with electrical mapping systems to define the atrial
surement is also essential from the defect to surrounding border and vein location for the proceduralist [35].
528 J. F. Mooney et al.

Fig. 42.8  Projection of the septal-to-lateral line. Note the predicted fluoroscopic angulation (RAO 24, CRA 25.2) is more clinically applicable

a b

Fig. 42.9  Defining transapical access. (a) Two-chamber view demon- lus. (b) Volume-rendered image depicting true axis (purple line) and
strating ‘true axis’ and coaxial trajectory. The green line represents the coaxial trajectory (blue line). Proximity to the left anterior descending
true long axis, and the blue line represents the trajectory that is coaxial artery also demonstrated
to the annular plan and bisects the centroid of the reconstructed annu-
42  CT for Minimally Invasive Repair of Mitral Valve and Other Structural Heart Diseases 529

The number, position and presence of accessory veins can 8. Feldman T, Foster E, Glower DD, et  al. Percutaneous
be determined in pre-procedural CT. Most commonly there repair or surgery for mitral regurgitation. N Engl J Med.
2011;364:1395–406.
are four, two each side. Occasionally there can be a single 9. Whitlow PL, Feldman T, Pedersen WR, et al. Acute and 12-month
ostium with the merger of the veins. Knowledge of vein anat- results with catheter-based mitral valve leaflet repair: the EVEREST
omy is required to prevent injury and ensure all veins drain- II (Endovascular valve edge-to-edge repair) high risk study. J Am
ing into the left atrium are isolated. Ostial size can also be Coll Cardiol. 2012;59:130–9.
10. Blanke P, Naoum C, Webb J, et al. Multimodality imaging in the
determined, with smaller ostia more susceptible to stenosis. context of Transcatheter mitral valve replacement: establishing
The presence of thrombus in the left atrium or left atrial consensus among modalities and disciplines. J Am Coll Cardiol
appendage can be shown on CT. Delayed phase imaging can Img. 2015;8:1191–208.
discriminate filling defects due to thrombus from defects due 11. Guerrero M, Dvir D, Himbert D, et al. Transcatheter mitral valve
replacement in native mitral valve disease with severe mitral annu-
to slow contrast filling as thrombus will persist in delayed lar calcification: results from the first multicenter global registry.
phases. The presence of thrombus contraindicates proceed- JACC Cardiovasc Interv. 2016;9:1361–71.
ing with ablation. 12. Willson AB, Webb JG. Labounty TM, et al. 3-dimensional aortic
Proximity to the oesophagus can be determined, with the annular assessment by multidetector computed tomography pre-
dicts moderate or severe paravalvular regurgitation after transcath-
left pulmonary veins commonly lying anterior. There is the eter aortic valve replacement: a multicenter retrospective analysis.
potential risk of atrial-oesophageal fistula formation with J Am Coll Cardiol. 2012;59:1287–94.
ablation of ostia in close proximity to the oesophagus. 13. Jilaihawi H, Kashif M, Fontana G, et al. Cross-sectional computed
tomographic assessment improves accuracy of aortic annular siz-
ing for transcatheter aortic valve replacement and reduces the
incidence of paravalvular aortic regurgitation. J Am Coll Cardiol.
Conclusions 2012;59:1275–86.
14. Toggweiler S, Gurvitch R, Leipsic J, et  al. Percutaneous aortic
valve replacement: vascular outcomes with a fully percutaneous
CT imaging prior to minimally invasive cardiac procedures procedure. J Am Coll Cardiol. 2012;59:113–8.
is emerging as a vital tool. Excellent spatial and temporal 15. Hayashida K, Lefevre T, Chevalier B, et  al. Transfemoral aortic
resolution with wide-field views allows an understanding of valve implantation new criteria to predict vascular complications.
JACC Cardiovasc Interv. 2011;4:851–8.
the target and surrounding anatomy. Accurate sizing can 16. Schoenhagen P, Numburi U, Halliburton SS, et  al. Three-­

determine appropriate device to minimize complications. dimensional imaging in the context of minimally invasive and
Finally, higher-risk features can be identified to allow appro- transcatheter cardiovascular interventions using multi-detector
priate patient selection and minimization of morbidity and computed tomography: from pre-operative planning to intra-­
operative guidance. Eur Heart J. 2010;31:2727–40.
mortality. 17. Levine RA, Triulzi MO, Harrigan P, Weyman AE. The relationship
of mitral annular shape to the diagnosis of mitral valve prolapse.
Circulation. 1987;75:756–67.
18. Berdajs D, Zund G, Camenisch C, Schurr U, Turina MI, Genoni
References M. Annulus fibrous of the mitral valve: reality or myth. J Card Surg.
2007;22:406–9.
1. Nkomo VT, Gardin JM, Skelton TN, Gottdiener JS, Scott 19. Lee AP, Hsiung MC, Salgo IS, et  al. Quantitative analysis of

CG, Enriquez-Sarano M.  Burden of valvular heart diseases: a mitral valve morphology in mitral valve prolapse with real-time
population-­based study. Lancet. 2006;368:1005–11. 3-­dimensional echocardiography: importance of annular saddle
2. Maisano F, Alfieri O, Banai S, et  al. The future of transcatheter shape in the pathogenesis of mitral regurgitation. Circulation.
mitral valve interventions: competitive or complementary role of 2013;127:832–41.
repair vs. replacement? Eur Heart J. 2015;36:1651–9. 20. Naoum C, Leipsic J, Cheung A, et  al. Mitral annular dimensions
3. Shroyer AL, Coombs LP, Peterson ED, et  al. The Society of and geometry in patients with functional mitral regurgitation and
Thoracic Surgeons: 30-day operative mortality and morbidity risk mitral valve prolapse: implications for transcatheter mitral valve
models. Ann Thorac Surg. 2003;75:1856–64. discussion 64-5. implantation. J Am Coll Cardiol Img. 2016;9:269–80.
4. Nishimura RA, Otto CM, Bonow RO, et  al. 2014 AHA/ACC 21. Hyodo E, Iwata S, Tugcu A, et  al. Accurate measurement of
guideline for the management of patients with valvular heart dis- mitral annular area by using single and biplane linear mea-
ease: executive summary: a report of the American College of surements: comparison of conventional methods with the
Cardiology/American Heart Association Task Force on Practice three-dimensional planimetric method. Eur Heart J Cardiovasc
Guidelines. J Am Coll Cardiol. 2014;63:2438–88. Imaging. 2012;13:605–11.
5. Mirabel M, Iung B, Baron G, et  al. What are the characteristics 22. Hahn RT, Abraham T, Adams MS, et al. Guidelines for performing
of patients with severe, symptomatic, mitral regurgitation who are a comprehensive transesophageal echocardiographic examination:
denied surgery? Eur Heart J. 2007;28:1358–65. recommendations from the American Society of Echocardiography
6. Nishimura RA, Vahanian A, Eleid MF, Mack MJ.  Mitral valve and the Society of Cardiovascular Anesthesiologists. J Am Soc
disease--current management and future challenges. Lancet. Echocardiogr. 2013;26:921–64.
2016;387:1324–34. 23. Grewal J, Suri R, Mankad S, et al. Mitral annular dynamics in myx-
7. Feldman T, Wasserman HS, Herrmann HC, et  al. Percutaneous omatous valve disease: new insights with real-time 3-dimensional
mitral valve repair using the edge-to-edge technique: six-month echocardiography. Circulation. 2010;121:1423–31.
results of the EVEREST Phase I Clinical Trial. J Am Coll Cardiol. 24. Blanke P, Dvir D, Cheung A, et al. A simplified D-shaped model of
2005;46:2134–40. the mitral annulus to facilitate CT-based sizing before t­ ranscatheter
530 J. F. Mooney et al.

mitral valve implantation. J Cardiovasc Comput Tomogr. 30. Ruiz CE, Kliger C, Perk G, et  al. Transcatheter therapies for the
2014;8:459–67. treatment of valvular and paravalvular regurgitation in acquired
25. Hutchins GM, Moore GW, Skoog DK.  The association of floppy and congenital valvular heart disease. J Am Coll Cardiol.
mitral valve with disjunction of the mitral annulus fibrosus. N Engl 2015;66:169–83.
J Med. 1986;314:535–40. 31. Webb G, Gatzoulis MA.  Atrial septal defects in the adult: recent
26. Cheung A, Webb J, Verheye S, et al. Short-term results of transapi- progress and overview. Circulation. 2006;114:1645–53.
cal transcatheter mitral valve implantation for mitral regurgitation. 32. Wahl A, Meier B.  Patent foramen ovale and ventricular septal
J Am Coll Cardiol. 2014;64:1814–9. defect closure. Heart. 2009;95:70–82.
27. Blanke P, Naoum C, Dvir D, et  al. Predicting LVOT obstruction 33. Romero J, Husain SA, Kelesidis I, Sanz J, Medina HM, Garcia
in transcatheter mitral valve implantation: concept of the Neo-­ MJ. Detection of left atrial appendage thrombus by cardiac com-
LVOT. J Am Coll Cardiol Img. 2017;10(4):482–5. puted tomography in patients with atrial fibrillation: a meta-­
28. Gurvitch R, Wood DA, Leipsic J, et  al. Multislice computed
analysis. Circ Cardiovasc Imaging. 2013;6:185–94.
tomography for prediction of optimal angiographic deployment 34. Xu B, Gooley R, Seneviratne SK, Nasis A. Clinical utility of multi-­
projections during transcatheter aortic valve implantation. JACC detector cardiac computed tomography in structural heart interven-
Cardiovasc Interv. 2010;3:1157–65. tions. J Med Imaging Radiat Oncol. 2016;60:299–305.
29. Blanke P, Dvir D, Naoum C, et al. Prediction of fluoroscopic angu- 35. Burke MC, Roberts MJ, Knight BP. Integration of cardiac imaging
lation and coronary sinus location by CT in the context of trans- and electrophysiology during catheter ablation procedures for atrial
catheter mitral valve implantation. J Cardiovasc Comput Tomogr. fibrillation. J Electrocardiol. 2006;39:S188–92.
2015;9:183–92.
 Cardiac CT: Electrophysiological
Applications 43
Joan M. Lacomis, Iclal Ocak, Friedrich Knollmann,
Andrew Voigt, and Raveen Bazaaz

Current electrophysiological applications of cardiac CT are Treatment of AF

primarily focused on evaluating patients with atrial fibrilla-
tion (AF) who are either a candidate for or who have been Management of AF has long represented a significant c­ linical
treated with pulmonary vein isolation (PVI) or left atrial challenge. Since the publication of the AFFIRM trial in 2003
appendage occlusion. In this chapter, we will review the [6], it has been recognized that there is no clear clinical
importance of AF, describe methods of pulmonary vein iso- advantage to a rhythm control (vs a rate control) strategy
lation and LAA occlusion, and explore the role of CT imag- among older patients who are minimally symptomatic.
ing for these procedures. However, among symptomatic patients, AF can severely
diminish quality of life. This has been used as an important
justification for pursuing a sinus rhythm aggressively. For
Importance of Atrial Fibrillation years, cardioversion and pharmacologic therapy were the
mainstay treatments for AF and continue to play important
Atrial fibrillation is the most common sustained cardiac roles in its management. AF is usually treated first with anti-
arrhythmia with a current prevalence of 6.1 million people arrhythmic medications [7]. Recently, the association
affected in the USA and, with the incidence still increasing, between AF, obesity, and obstructive sleep apnea (OSA) has
an expected prevalence of 16 million people by 2050 [1, 2]. been recognized. There is now emerging data suggesting that
AF typically progresses from a syndrome of paroxysmal to intensive lifestyle/weight loss intervention and therapy of
persistent and thereon to permanent AF (Table 43.1). OSA may reduce the symptomatic burden of AF [8, 9]. If a
AF is a significant cause of chronic left ventricular dys- patient is hemodynamically compromised, electrical cardio-
function and remains one of the leading causes of debilitat- version is the treatment of choice [10]. However, AF is often
ing stroke imposing a tremendous burden on affected patients refractory to or recurrent despite these treatments. The lim-
and the healthcare delivery system. Twenty-thirty percent of ited efficacy and known toxicity of antiarrhythmic drugs has
all strokes are due to AF [3–5]. led to the development of non-pharmacologic approaches.
The original Cox-Maze and its later modifications are sur-
gical therapies for the treatment of AF. The newer catheter-­
based strategies for treating AF and/or preventing stroke in
patients with AF represent nonsurgical options for patients
who have failed traditional pharmacologic methods. The
catheter-based strategies to treat AF include pulmonary vein
isolation (PVI) with radiofrequency catheter ablation
(RFCA) or cryoballoon ablation (CBA). More recently, LAA
J. M. Lacomis (*) · I. Ocak occluders have been deployed to reduce the risk of stroke in
Department of Radiology, Thoracic Imaging Division, University AF patients at risk for cardioembolic stroke who are also at
of Pittsburgh Medical Center, University of Pittsburgh School of high risk for bleeding as an alternative strategy to long-term
Medicine, Pittsburgh, PA, USA
e-mail: [email protected] oral anticoagulation [11].
Multidetector computed tomography (MDCT) is fre-
F. Knollmann
Department of Radiology, UCDavis Health, Sacramento, CA, USA quently used for preprocedural planning and identification of
post-procedural complications. Imagers need to possess a
A. Voigt · R. Bazaaz
Heart and Vascular Institute, University of Pittsburgh Medical thorough understanding of the procedures, including
Center, Pittsburgh, PA, USA potential complications, the CT anatomy of the target
­

© Humana Press 2019 531

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_43
532 J. M. Lacomis et al.

Table 43.1  Types of atrial fibrillation (AF) der of the PLA.  Occasionally, another contiguous line of
AF category Defining characteristic ablation lesions will be placed across the roof of the LA,
Paroxysmal Terminates spontaneously or with intervention extending between the vestibules [16–21], though the value
within 7 days of onset of additional lines of ablation beyond PVI has been ques-
Persistent Fails to self-terminate within 7 days and requires tioned by a more recent study [22].
pharmacologic or electrical cardioversion to restore
sinus rhythm
Permanent Lasts longer than a year, and sinus rhythm is not Cryoballoon Ablation
possible CBA is currently the most widely used alternative to RFCA
Lone Occurs in adults <60 years of age without underlyingand was developed in part because of the often tedious and
cardiopulmonary disease technically challenging nature of RFCA [23]. In addition,
Isolated Occurs without associated atrial tachycardia or cryoablation may preserve the tissue architecture and be asso-
aflutter
ciated with less thrombus formation as compared to RFCA
[24], though this has not been clearly established. Whereas
regions, the scanning and postprocessing techniques neces- RFCA is applied in a point-by-point mode and results in cel-
sary to provide the needed anatomical information, and the lular necrosis by tissue heating, CBA is applied to the PV
pertinent report content for scans for these procedures. antral tissue with a balloon in a single-step mode and leads to
cellular necrosis by freezing [25]. Like RFCA, the LA can
only be accessed using a transfemoral approach followed by
Pulmonary Vein Isolation a trans-atrial septal puncture. Then, a balloon catheter is
advanced over a guidewire into the ostium of a PV where it is
Radiofrequency Catheter Ablation inflated to completely occlude the ostium. Balloon occlusive
In 1998, Haissaguerre and his colleagues identified the mus- pulmonary venography and electroanatomic mapping are
cular sleeves of the distal PVs (adjacent to the left atrium often utilized to avoid positioning the cryoballoon within the
(LA)) as an important source of ectopic beats initiating AF PV during ablation. Balloon selections are based on PV ostial
[12]. Over 90% of the ectopic beats initiating AF arise in the sizes on the preprocedural MDCT, though as experience with
PVs with the culprit beats most often arising in the left supe- the technique has grown, in the vast majority of cases, the
rior pulmonary vein (LSPV) [12, 13]. The understanding of larger 28 mm cryoballoon is utilized. The CBA of each PV is
this relationship led to the development of the Cox-Maze-­ performed in at least two 3–4 min freezing cycles. The opera-
type surgeries followed by the interventional therapies of tor monitors the temperature at the distal aspect of the bal-
RFCA and CBA directed at causing anatomic scars to dis- loon, with higher nadir (warmer) temperatures suggesting
rupt the electrical pathways between the initiating PVs and poor overall contact with antral tissue/inefficacy and lower
the substrate for propagation of AF, the posterior left atrium nadir (colder) temperatures being worrisome for injury to col-
(PLA) [14]. RFCA was the first of these therapies and is still lateral structures. The intracellular freezing is followed by
the most common. The technique itself has been modified injury during thawing due to a combination of ice crystals
over the years moving away from point ablation within the expanding and contracting in the intra- and extracellular
distal PVs to the current target of the pulmonary vein antral spaces, thereby damaging cell walls and causing microvascu-
tissue in the PLA. lar occlusions and ischemia. The combined local hemorrhagic
RFCA is a catheter-based procedure which is technically and inflammatory responses lead to permanent cellular necro-
complex, time-consuming, and performed using general sis [25]. The balloon is deflated, and the single-­step freezing
anesthesia at many centers. For the catheter-based proce- process is repeated for each of the remaining PVs. Cryoballoon
dures, the LA can only be accessed via a transfemoral (CB) ostial ablation lesions form well-­demarcated circumfer-
approach, followed by trans-atrial septal puncture. RFCA is ential scars, electrically isolating the PVs from the LA [26].
carried out by using an electrical current with alternating fre- In 2013, the STOP-AF trial, comparing CBA vs specific
quencies and results in hyperthermic cellular injury through antiarrhythmic agents, demonstrated the safety and effec-
a combination of coagulation and tissue necrosis [15]. The tiveness of CBA for the treatment of PAF, with a 70% rate of
ablation catheter is navigated circumferentially around the freedom of AF recurrence at 1 year in the CBA group [23].
bilateral PV inflow vestibules (i.e., PV antral regions) with- The more recent landmark “FIRE AND ICE” trial, compar-
out entering any pulmonary venous ostia; using a point-by-­ ing RFCA vs CBA, found that CBA was non-inferior to
point technique, multiple adjacent, approximately 4 RFCA [25]. Single center trials and registry data showed
millimeters in diameter, point ablation lesions are created. freedom from AF of 45–72% for RFCA compared to 55–77%
When complete, the discreet point lesions collectively form for CBA at follow-up ranging from 6 months to 2 years [27].
bilateral contiguous, linear, circumferential lesions to elec- Two smaller single-center studies showed CBA in persistent
trically isolate the bilateral antral PV tissue from the remain- AF patients has a 60–67% success rate at 1 year [25, 28].
43  Cardiac CT: Electrophysiological Applications 533

The long-term outcomes for PV catheter ablation are not question of pseudothrombus from mixing artifact from slow
as good as initially hoped [29] with PV reconnection and filling vs thrombus. Contrast dose and kvp are tailored for
repeat ablation procedures being common. This has prompted patient-specific weight, BMI, and renal function; the use of
a search for additional targets for ablation and further inves- dual-barrel injectors with contrast timing, contrast/saline
tigation into the underlying LA substrate for AF. Although bolus profile, and injection rates can be the same as coronary
the majority of foci triggering AF have been mapped to the CTA and can be done with a test bolus or automated bolus
PVs, it has long been known that AF may also be triggered tracking [38–43].
by arrhythmogenic foci originating from the right atrium Scans that are obtained post-PVI to evaluate for compli-
(RA), LA, LAA, coronary sinus (CS), superior vena cava cations should image the entire chest as postop complica-
(SVC), or the vein of Marshall [12, 30–33]. The optimum tions (or other pathologies accounting for patient’s
treatment strategy including employing hybrid catheter-­ presentation) are not limited to the heart. ECG gating is
based and surgical techniques is an ongoing area of dependent on a variety of factors including not only whether
investigation. the scanner is sufficient for patient’s rate and rhythm, but
patient’s clinical status which can range from minor symp-
toms to obtundation (see “Interpretation post-PVI”) will also
CT Technique dictate the ability to ECG gate. The contrast volume is
increased to cover the entire region, similar to an aortic CTA
Image Acquisition evaluation.
The main goals of imaging are to delineate and display the
PLA and its adjacent anatomy, typically in a combination of Postprocessing
multiplanar reformats and 3D models, to identify any find- Postprocessing includes multiplanar reformats (MPRs) and
ings that may contraindicate or increase the complexity of three-dimensional (3D) volume rendered models of the LA,
the ablation, and to identify any significant incidental find- LAA, and distal PVs. LA segmentation protocols for this
ings that warrant further investigation or intervention either purpose are available on all of the major vendor worksta-
pre- or post-ablation. Of note, it is increasingly common in tions. The 3D models are displayed from both epicardial
recent years to perform pre-PVI anatomic mapping on (extra-atrial) and endocardial (intra-atrial) vantages; the
patients who have had prior ablations and now have recurrent esophagus can also be included on the epicardial views to
AF with a second ablation procedure planned. In such show its location along the posterior atrial wall (Fig. 43.1a–
patients, it is also important to identify lasting complications c). Electrophysiologists reference these models for the abla-
from the prior ablation which may be unsuspected such as tion thus rendering the need for retrograde pulmonary
PV stenosis or occlusion. venography obsolete. Typically, navigation of the catheter
Since AF patients who are candidates for RFCA or CBA during ablation is performed using a combination of an elec-
typically have PAF or less commonly persistent AF, their troanatomic mapping system, intracardiac echo, and fluo-
rhythm may be AF or sinus at the time of CT which is typi- roscopy. At some centers, the epicardial 3D volume
cally obtained within 24 h pre-ablation [34–36]. Pre-ablation rendering is merged with the intraprocedural electroana-
anatomic mapping of the PLA was initially described on four tomic map so the progress of the ablation lines can be
detector scanners without ECG gating [37]. As with other tracked in real time [44–47] (Fig. 43.2a, b).
cardiac CT, modern scanning is now done on a minimum of PV ostial diameters are measured utilizing the MPRs to
64 slice scanners. Since PLA anatomy can be accurately obtain true orthogonal obliques [42, 43]. Maximum PV ostial
characterized with or without ECG gating, patients who are diameters occur during late atrial diastole (approximately
in AF can be successfully scanned without ECG gating, and 85% R-R), and minimum diameters occur during atrial sys-
retrospective gating is generally discouraged due to the tole (approximately 15% R-R) [44–48]. PV ostial diameters
higher radiation dose [36]. However, AF, especially if rate are needed to choose the appropriate size of the cryoballoon.
controlled, is less of a challenge for high-quality low-­ Because of the dynamic nature of the PVs, measurements
radiation ECG-gated acquisitions on the newest generation from ECG-gated studies should be obtained from the same
>64 slice MDCT scanners. phase on any subsequent post-ablation scans [48, 49].
Regardless if an ECG-gated or a nongated acquisition is
done, the z-axis coverage is increased from that of a coro-
nary CTA to start the scan at the mid-aortic arch level to Interpretation Pre-PVI
include the distal third of the superior pulmonary veins and
the entire LAA. Although TEE is routinely done to exclude Interpretation of CT scans prior to PVI requires evaluation of
LAA thrombus, a limited repeat acquisition through the axial source images, MPRs, and 3D models. Necessary ana-
LAA can be done 60–90 s post-contrast injection if there is a tomic information includes evaluation of the number and
534 J. M. Lacomis et al.

Fig. 43.1 (a–c): Standard 3D

posterior left atrial models:
a b
Volume rendered PA
epicardial view (a),
endocardial view (b), and
epicardial view including the
esophagus (arrow) (c) of the
posterior left atrium. (LS left
superior pulmonary vein, LI
left inferior pulmonary vein,
RS right superior pulmonary
vein, RI right inferior
pulmonary vein, RM right
middle lobe ostial branch,
ssRLL superior segment RLL
ostial branch, LM left middle
pulmonary vein). In 2a, ostial
c
branches are labeled in white

a b

Fig. 43.2 (a, b): Electroanatomic mapping: In (a) a CARTO RFCA, the red areas represent low voltage, and the purple areas repre-
(BioSense, Diamond Bar, CA) electroanatomic map from a PA view of sent normal voltage (viable tissue). Notice how wide the antral area is
the PLA post-RFCA delineates the completed ablation lines. The red targeted for ablation surrounding the PVs. The overlying coiled objects
octagons represent the ablation points which collectively encircle the are superimposed snapshots of the positions of the tip of the circular
pulmonary vein antral tissue with the goal of electrical isolation. In (b) mapping catheter, used to validate electrical isolation of each of the
the corresponding CARTO endocardial voltage map, PA view post-­ veins

location of PVs, including the presence of conjoined or from the LA, starting with the V1 segment. The V1 segment
accessory veins, identification of ostial branches, the relative is between the PV ostium and the first branch point [35, 39]
location of the esophagus along the posterior left atrial wall, . If this length is within 5 mm of the ostium, the first branch
and identification of any findings that may contraindicate or is termed an ostial branch [35]. Ostial branches are impor-
increase the complexity of the ablation. tant because they are at increased risk for stenosis. There is
Each PV can be divided into longitudinal segments much variation between the different PVs in ostial diame-
between branch points in a retrograde manner extending ters, angulations, and V1 segment lengths [42, 43].The atrial
43  Cardiac CT: Electrophysiological Applications 535

wall between ipsilateral PVs is termed the intervenous sad- right middle lobe (RML); the right inferior (RIPV) drains the
dle or carina [35]. There are two funnel-shaped, PV inflow right lower lobe (RLL) and occasionally all or part of the
vestibules or antral regions, right and left; each consists of RML; the left superior (LSPV) drains the left upper lobe
all the PV ostia on the ipsilateral side and their intervenous (LUL) including the lingula; and the left inferior (LIPV)
saddles. The V1 segments of the PVs, bilateral vestibules, drains the left lower lobe (LLL) [32, 34, 36, 37, 40, 43, 53].
and the interposed atrial wall comprise the PLA [35]. The If variant PV anatomy is present, the right side tends to be
remainder of the LA is divided into the roof, the appendage, more complex, often with accessory PVs, while the left side
the isthmus, and the septum [50]. The extensive electrical tends to be more simplified, as in the case of a conjoined
conduction system within the roof of the LA is termed vein. A conjoined or common vein occurs when the superior
Bachman’s bundle; there is a complex network of intercon- and inferior veins merge proximal to the LA and enter the
necting electrical pathways in the PLA. Of the more recently LA as a single trunk of varying lengths [34, 40]. The result-
identified culprits in either the initiation or sustenance of ing single pulmonary venous ostium is typically broad
AF, the identification of regions of functional reentry within (Fig.  43.4a, e). Deciding if there is a short common trunk
the LA termed rotors is of most interest to imagers due to with large superior and inferior ostial branches versus two
their association with underlying atrial fibrosis and the abil- adjacent but separate left-sided veins with a small interve-
ity of cardiac magnetic resonance imaging to delineate nous saddle can be difficult and may account for some of the
fibrosis [51, 52]. variation in reported incidence of conjoined veins. While
Normally, all of the PVs drain into the LA; although vari- conjoined veins are usually on the left, they also rarely occur
ation in number and configuration of PVs entering the LA is on the right, bilaterally or inferiorly (combined RIPV and
common (Fig.  43.3), PVs which drain to structures other LIPV) [34, 39, 40, 54]. The anatomic variant of a left com-
than the LA (i.e., the superior or inferior vena cavae, bra- mon PV (LCPV) is often seen in clinical practice and has
chiocephalic vein, RA) are considered anomalous. relevance for preprocedural planning for electrophysiolo-
The classic description of four separate PVs with a supe- gists. LCPV anatomy can be particularly challenging for
rior and inferior ostium on the right and a superior and infe- operators contemplating a CBA approach, as the size of the
rior on the left is the most common anatomic configuration common segment is sometimes larger than the diameter of
and is referred to as conventional pulmonary venous anat- the cryoballoon itself, forcing the operator to either switch to
omy (Fig. 43.1a). In several studies, some of which included RFCA or to ablate using a more challenging segmental
only pre-RFCA AF patients, four PVs reportedly occur approach with multiple applications of cryothermal energy
60–82% of the time, but large population studies have not at the antral tissue surrounding the PV.
been done, and the true incidence is unknown [32, 34, 36, 37, An accessory vein has its own independent atrio-­
40, 43, 53]. Typically, the right superior pulmonary vein pulmonary venous junction separate from the superior and
(RSPV) drains the right upper lobe (RUL) and usually the inferior PVs on that side. Accessory veins are important as

Fig. 43.3  Normal pulmonary

vein variations: Top row (left
to right): conventional
anatomy with left superior
(LS) and left inferior (LI)
pulmonary veins widely
separated, conventional
anatomy with LS and LI
LS RS LS RS LCV RS LCV RS
immediately adjacent to each
other, short left common vein LI RI LI RI RI RI
(LCV), long LCV. Bottom
row (left to right): accessory
right middle vein (accR), two
accessory right middle veins,
top vein (TV) with or without
a right middle vein, right
common vein or left middle
vein. Rarely, there can be
LS RS LS RS LS accR LS RCV
bilateral common veins (not
drawn) LI accR LI accR LI (TV) accL
RI accR accR LI
RI RI
536 J. M. Lacomis et al.

their ostia may be smaller than those of the superior and infe- the RSPV (Figs.  43.4c and 43.5b) [34, 44]. The top vein
rior veins, exposing them to an increased risk of stenosis. It drains either the ssRLL or the posterior segment right upper
is important to know what segments or lobes of lung acces- lobe (psRUL) or a combination of the two. This should not
sory veins are draining, because if complications such as be confused with another frequently seen normal structure,
hemodynamically significant pulmonary venous stenosis or an accessory LAA.  An accessory LAA is a small-sized,
thrombosis occur, the associated pulmonary parenchymal blind-ending, diverticula-like outpouching which contains a
abnormalities will be reflected in the segment or lobe of lung variable number of digitations [58]. The accessory LAA,
drained by the affected vein. which arises from the anterosuperior aspect of the LA roof,
There may be one or more accessory veins, typically on can be near the RSPV, more midline, or closer to the LSPV
the right or toward the right posterior side of the roof of the (Fig. 43.5a, b) [50].
LA.  The frequency of accessory veins has been reported In addition to evaluating PLA anatomy, scans are rou-
between 12% and 33% [39, 40, 43, 54]. The two most com- tinely evaluated for other significant findings. A well-­
mon accessory veins are the right middle lobe vein (which opacified LAA can be assessed for thrombus, a
drains all or part of the right middle lobe (RML), accounting contraindication to the procedure [15–21]. However, if the
for 55–93% of accessory right-sided veins (Fig. 43.4b), and LAA is dilated and has slow flow, particularly if TEE shows
the superior segment right lower lobe (ssRLL) vein account- spontaneous echo contrast indicating slow filling, it may be
ing for 28% of accessory right-sided veins (Fig.  43.4c). difficult to differentiate poor filling from thrombus. In fact,
However, note that any third vein interposed between the the two may look identical. There have been several pro-
RSPV and RIPV, regardless if it’s draining the RML or posed methods to differentiate pseudothrombus due to
ssRLL, has often been generically termed a “middle vein” incomplete LAA opacification from LAA thrombus; the
[34, 37, 40, 53]. Rarely, accessory PVs can drain the basilar easiest is to obtain a second set of delayed images through
segments of the RLL. In addition, the occasional accessory the LAA at approximately 60–90  s post-contrast injection
vein on the left, which typically drains the lingula, is also (Fig. 43.6a, b) [59].
commonly referred to as a “middle vein” (Fig. 43.4d) [32, While TEE is the standard of reference and the only
43, 44, 55–57]. approved imaging modality to assess for LAA thrombus, in
The other interesting accessory vein is the “top vein” a meta-analysis of 19 studies with 2955 patients, cardiac CT
which enters the roof of the LA posteriorly, superomedial to with the addition of delayed images was shown to be as

Fig. 43.4 (a–d): Normal

pulmonary vein variations: (a)
a b
left common vein (LCV),
accessory right middle
pulmonary vein (RM)
draining the entire right
middle lobe; (b) two
accessory “right middle”
veins, separately draining the
medial (RMM) and lateral
(RML) segments of the right
middle lobe; (c) accessory
superior segment right lower
lobe pulmonary vein draining
that lobe and accessory “top
c d
pulmonary vein” draining the
posterior segment of the right
upper lobe; (d) endocardial
view showing the broad
ostium of a LCV (LAA, left
atrial appendage; RS, right
superior pulmonary vein;
RI, right inferior pulmonary
vein; ** in 4d denotes the
Ridge of Marshall)
43  Cardiac CT: Electrophysiological Applications 537

Fig. 43.5 (a, b): Accessory

LAA: Both an accessory LAA
a b
and a top pulmonary vein
(TV) typically arise in close
proximity to the RSPV. In (a)
the accessory LAA (black
arrow) arises from the anterior
aspect of the roof of the LA,
whereas in (b), an accessory
TV (white arrow) arises from
the posterior aspect of the LA

Fig. 43.6 (a, b): LAA

a b
pseudothrombus: Particularly
as the LAA dilates, filling
defects in the tip of the LAA
are common on CTA (a) and
could represent thrombus or
pseudothrombus from slow
LAA filling. Complete
opacification of the LAA on
delayed images (b) excludes
the presence of true thrombus

accurate as TEE.  In the meta-analysis, cardiac CT had an ume rendering, specifying its location in the report in rela-
overall accuracy of 99% with a mean sensitivity of 100%, tion to the PLA is recommended. Anatomically, the
specificity of 99%, PPV of 92%, and NPV of 100% [59]. esophagus is adjacent and posterior to the LA for approxi-
Any other abnormality of the LA, the RA, the atrial sep- mately 5 cm of the length of the esophagus, either running
tum, or the azygous continuation of the IVC that may inter- midline, right or left para-midline, or obliquely from the
fere with access to or ability to navigate a catheter within the LSPV to the RIPV.
PLA from a transfemoral approach should be reported [60]. Since catheter ablation for AF requires either general
Any abnormality of a structure in close proximity to the anesthesia or intravenous conscious sedation, incidental
PLA, particularly if the abnormality exerts mass effect on the findings such as pneumonia, including aspiration after
LA, such as a descending aortic aneurysm, esophageal dila- TEE, and coronary artery disease also need to be specifi-
tation, paraesophageal hernia, or mediastinal mass, should cally addressed as their presence may warrant postponing
also be reported [42]. The esophagus warrants specific atten- or canceling the ablation. Similarly, findings such as a
tion for RFCA cases as it is at risk for iatrogenic injury dur- ­suspected lung carcinoma or other malignancies may affect
ing the procedure due to its close proximity to the PLA. In the decision to proceed with an ablation until they are
addition to including the esophagus on the 3D epicardial vol- ­evaluated [25, 26, 38, 43].
538 J. M. Lacomis et al.

Interpretation Post-PVI some areas, the esophageal wall is in direct continuity with
the posterior left atrial wall, and in other areas, there is a thin
The major role of CT post-PVI is to assess for complications. layer of fat interposed between the posterior wall of the LA
The overall major complication rate for PVI is estimated at and the anterior wall of the esophagus (Fig. 43.7a) [56, 74,
3.9–4.5% [24]. Both major and minor complications, listed 75]. If the thermal injury is severe enough, the esophagus can
in Table 43.2, can occur, most with either RFCA or CB abla- perforate. If the esophageal perforation goes undetected, the
tion [21, 61–73]. Many are well-known complications that ensuing mediastinitis can extend to the LA wall which can
can occur with other interventional catheter-based cardiac also perforate, creating fistulous communication between the
procedures; there are also the risks associated with sedation two [66–69, 72, 73, 76].
and general anesthesia that are not included. Early signs and symptoms of AEF are nonspecific and
Of note, cardiac perforation and associated tamponade include chest pain, dysphagia, fever, and an elevated white
can occur during trans-atrial septal puncture, when manipu- blood cell count [61, 62, 69, 70, 77]. Patients may rapidly
lating the catheters in the LA and distal pulmonary veins or deteriorate as mediastinitis progresses over the span of
when performing the ablations, but are acutely apparent hours to a few days. The diagnosis is easily missed in its
from clinical parameters and echocardiographic findings early clinical stages. When the posterior left atrial wall is
during these procedures and not reliant on MDCT for diag- perforated, air from the esophagus may enter the LA, and
nosis. There are three complications noteworthy to PVI pro- signs of endocarditis, septic or air emboli, and overwhelm-
cedures: phrenic nerve injuries (PNI), atrio-esophageal ing sepsis may develop. Patients are at risk for mental status
fistula (AEF), and PV stenosis, the latter two of which are changes, stroke, seizures, massive hematemesis, and rapid
reliant on MDCT for diagnosis. Rapid and accurate interpre- cardiovascular collapse with death. The mortality of >50%
tation of subtle CT findings of severe injury is paramount. is in part attributed to significant delays in or missed diagno-
sis [61, 62, 71–75, 78].
Atrio-esophageal Fistula
AEF is a rare but devastating complication only reported MDCT
with RFCA. AEF occurs due to thermal injury to the esopha- MDCT is the primary imaging modality to assess for compli-
gus through the LA wall. First described in 2004, there have cations post-ablation [53, 66]. Early CT findings are those of
only been a few published cases. The true incidence of mediastinitis such as fat stranding or a small mediastinal
unpublished occurrences is unknown, and many suspect the fluid collection and/or extraluminal gas locules in the region
incidence is likely higher [57–61, 63, 64, 66]. In the pub- of the PLA or esophagus [14, 70, 72, 73] (Fig. 43.7b).
lished cases, patients presented from 2 days to 5 weeks post-­ As the process progresses to a fistula, the posterior wall of
ablation, and the incidence of mortality was over 50%. the LA may become thickened and irregular or develop a
The underlying problem is a thermal injury during the pseudoaneurysm. Gas locules can occur within the pericar-
PLA ablation to the fragile anterior esophageal wall due to dial space or within the wall or lumen of the PLA [14, 70, 72,
the anatomic close proximity of the esophagus and PLA. In 73]. Small pleural or pericardial effusions are relatively com-
mon minor complications post-RFCA; however, rapidly
accumulating collections or pleural or pericardial enhance-
Table 43.2  Reported complications of RFCA or CBA ment should raise suspicion for AEF [14, 71] (Fig. 43.8a, b).
Minor Confident MDCT diagnosis of an AEF is important to
Pleural effusion ensure adequate emergent treatment which requires surgical
Pericardial effusion exploration with both esophageal and atrial repairs; TEE and
Hemodynamically insignificant pulmonary vein stenosis
esophagoscopy with CO2 insufflation are contraindicated
Femoral vein catheterization complication: hematoma,
arteriovenous fistula, pseudoaneurysm when AEF is suspected [21, 62, 77].
Major
Pulmonary veins: dissection, perforation, severe stenosis, Pulmonary Vein Stenosis
thrombosis, venous infarct, veno-occlusive disease with resultant Pulmonary vein stenosis (PVS) is a significant complication
PA hypertension of catheter ablation for AF. There is significant variation in
Cardiac perforation: hemopericardium, tamponade
the reported incidence of severe, hemodynamically signifi-
LA intramural hematoma, dissection
cant pulmonary vein stenosis, since post-PVI follow-up
Coronary arteries: spasm, ischemia, infarct
Cardiac dysrhythmias imaging is not always obtained. Also, the pulmonary veins
Thromboembolic: stroke, transient ischemic attack (TIA), undergo structural remodeling after ablation with a mean
pulmonary or systemic emboli reduction in cross-sectional areas of roughly 5–10% [49]. A
Phrenic nerve palsy, paralysis recent systematic review suggested that the incidence of
Atrio-esophageal fistula, esophageal perforation, mediastinitis PVS remains approximately 2% [49, 79]. Pathophysiology
43  Cardiac CT: Electrophysiological Applications 539

a b

Fig. 43.7 (a, b): Atrio-esophageal fistula, early findings: The axial new, subtle fluid collection (arrow) in the mediastinum that is insepa-
image from a CT obtained pre-ablation (a) shows the normal esophagus rable from the esophagus and abuts the posterior left atrial wall near the
(arrow) toward the left side of the LA, immediately posterior and in LSPV ostium with infiltration of the adjacent fat. There are also large
direct continuity to the atrio-pulmonary venous junction of a long left right and small left pleural effusions; the former showed signs of locu-
common vein. Image (b) was obtained 5  days post-RFCA when the lation on more superior images (not shown). (From Lacomis et al. [14],
patient presented with chest pain. Note the signs of mediastinitis with a with permission)

a b

Fig. 43.8 (a, b): Atrio-esophageal fistula, late findings: The patient the pericardial space (a, arrows). A large pericardial effusion has devel-
continued to clinically deteriorate with ongoing chest pain, new dys- oped, and the pericardium is enhancing (a, b). The right pleural effu-
pnea, fever, and elevated white blood cell count and underwent CT of sion has markedly increased. In (b), there is a pseudoaneurysm (arrow)
the chest, abdomen, and pelvis 48 h later which showed progression of of the posterior left atrial wall, inferoposterior to the RIPV. A nasogas-
findings and additional evidence of an atrio-esophageal fistula. There tric tube is in the esophagus. (From Lacomis et  al. [14], with
are new extraluminal gas locules adjacent to the esophagus and within permission)

of PVS after catheter ablation remains unclear but most MDCT

likely due to irreversible progressive inflammation and col- Acutely, an awareness of a recent history of PVI is neces-
lagen deposition in the PV wall adjacent to an ablation sary to avoid misdiagnosing the pulmonary findings of acute
lesion [79]. PV stenosis is related to catheter position. The PV infarcts as pneumonia or pulmonary arterial infarcts
more distal positioning the catheter from the ostium, the (Fig. 43.9a–d) [19, 80–83]. CT can be more definitive, with
higher the risk. direct assessment of the PVs, particularly when the pre-­
PV stenosis has been reported up to 2 years post-RFCA ablation CT is available for comparison. Obtaining the PV
and, when severe, can result in PV thrombosis, venous CT with ECG gating is recommended. This provides the
infarct, pulmonary veno-occlusive disease, and pulmonary best images for multiplanar reformatting and measurement
artery hypertension. Symptoms are often nonspecific and of the ostial diameters. If ECG gating is not possible, non-
insidious in the onset such as vague chest pain, dyspnea, or ECG-­gated CT images should be adequate for diagnosis of
cough, but patients can also present more acutely with sharp severe PV stenosis although diagnosis of mild or moderate
pleuritic chest pain and hemoptysis [19, 80]. areas of stenosis should be done more cautiously on non
540 J. M. Lacomis et al.

a b

c d

Fig. 43.9 (a–d): Pulmonary vein stenosis and occlusion: Epicardial (a, (c) shows large areas of airspace consolidation in the left mid- and
b) views of the PLA on a pre-ablation CT showing unexpected occlu- lower lung zones with a moderate left pleural effusion. The initial chest
sion of the LSPV (white arrow) and stenoses of the RSPV and RIPV CT (d) shows thrombus in the LSPV (white arrow) and large areas of
(black arrows). By history, the patient had an ablation 11 years earlier, ground glass opacity with intralobular septal thickening consistent with
followed 2  weeks later by a 6-week hospitalization for an “unusual wedge-shaped venous infarcts in the lingula and left lower lobe with a
pneumonia” with low-grade fevers and white count; all cultures and left pleural effusion
pleural fluid were negative. The initial PA radiograph from 11 years ago

ECG-gated CT images due to the variability of the PV diam- vein stenosis following ablation with balloon angioplasty
eters with different phases of the cardiac cycle [44, 48]. has been used as a treatment option for symptomatic patients
Both the minimum diameter and length of a stenosis should although data suggest restenosis rates are relatively high
be assessed. Imaging criteria for the diagnosis of PVS ranging from 47% to 57% [19, 85].
include a reduction in PV long-axis diameter of >50% com-
pared to pre-ablation study, corresponding to a reduction in  hrenic Nerve Injury
P
PV cross-sectional area of >75%. Symptoms typically pres- PNI during PVI is more common in CBA but can also occur
ent when the severity of stenosis reaches 60% reduction in in RFCA. It can result in either a palsy or paralysis and may
diameter [49]. A stenosis is likely more clinically significant be transient or permanent. A right-sided PNI is the most fre-
if ancillary findings of fat infiltration in the adjacent medias- quently observed complication with CBA. The risk of PNI
tinum, enlarged reactive lymph nodes, peripheral lung opac- due to CBA is reported between 6% and 24% [86]. The
ities, and localized septal thickening due to veno-occlusive reported incidence of transient and persistent PNI resulting
disease and localized pulmonary hypertension are found from ablation with the first-generation CB was 6% and 4%,
[19, 38, 80, 81, 84]. Noninvasive treatment of pulmonary respectively. The risk of PNI with the second-generation CB
43  Cardiac CT: Electrophysiological Applications 541

is even higher, with the reported incidence of transient and HAS-BLED score [94–97]. Patients with both a high risk of
persistent PNI being 16–24% and 5–7%, respectively [87]. stroke (CHA2DS2-VASc score of at least 3) and elevated
The underlying problem is a hypothermic (CBA) or ther- risks of bleeding on oral AC represent the target population
mal (RFCA) injury to the right phrenic nerve because of the for LAA occlusion procedures. However, because the
proximity between the right phrenic nerve and the right PVs CHA2Ds2-VASc score and other currently used risk scores
[87–89]. The right pericardiophrenic bundle, consisting of have a poor predictive value with a C-statistic of 0.55–0.64,
the right phrenic nerve, pericardiophrenic artery, and pericar- new thoughts of risk stratification for stroke are focused on
diophrenic vein, courses along the lateral margin of the SVC the concept of assessing the LA for an underlying atrial
passing close to the RSPV ostium [89–91]. Patients with uni- myopathy including the presence of atrial fibrosis, dilatation,
lateral PNI are ­typically asymptomatic, but dyspnea, particu- and loss of contractile function which leads to decreased
larly orthopnea, can occur. LAA emptying velocities and blood stasis, thus increasing
thrombogenicity [94, 98]. An interrelationship between com-
MDCT plex LAA shapes, risk of thrombogenicity, and stroke has
The right pericardiophrenic bundle is not consistently visual- also been postulated [99–102]. Thus echocardiography (LA
ized on MDCT. New unilateral elevation of the right hemi- strain), cardiac MRI (LA and LAA flow, LAA morphology,
diaphragm post-CBA is a sign of possible PNI and can be fibrosis, LAA shape), and cardiac CT (LAA shape) have
further evaluated with a fluoroscopic sniff test. In addition, potential emerging roles in stroke risk stratification [103].
Stroker et  al. showed pre-CBA evaluation of the PVs with Although oral anticoagulation with a traditional vitamin
MDCT as a useful tool for predicting the risk of PNI. They K antagonist or a novel non-vitamin K antagonist drugs are
found that large PV dimensions, ostial branches, and a short effective therapies, they are not without risk. Only 65% of
distance between the SVC and any of the right-sided PVs patients on oral warfarin are therapeutic on the agent, and
were associated with increased risk of PNI [86]. If the dis- there exists risk of a major hemorrhage in all patients on oral
tance between the RSPV and right pericardiophrenic bundle anticoagulation with an elevated HAS-BLED score [96].
is less than 10 mm, then there is a higher risk of PNI [89]. The LAA is the primary source for thromboembolism in
AF. In a review of 23 studies in patients with non-valvular
I njury to Pulmonary Parenchymal Tissue AF, 91% of clots were localized to the LAA [104]. The role
CBA has recently been associated with acute injury involv- of occlusion of this structure to prevent thromboembolic
ing the left mainstem bronchus which is often in close prox- complications from AF has been the subject of clinical inves-
imity to the ostium of the LSPV. Knight et al. demonstrated tigation over the last decade. Increasing scientific evidence is
cryothermal injury to the left mainstem bronchus using a leaning to a viable role for appendage occlusion in stroke
bronchoscope during CBA at the LSPV in three of ten prevention in non-valvular AF. In 2016, the US Centers for
patients [92]. The clinical relevance of this remains unclear, Medicare and Medicaid Services approved a device for
but it is postulated to be a cause for persistent dry cough seen endocardial LAA occlusion (LAAO) in patients considered
in some patients post-CBA who do not have PNI.  Case suboptimal candidates for oral AC therapy. In Europe, per
reports have also described hematomas adjacent to the RIPV, the 2016 European Society of Cardiology Management
identified by CTA post-CBA in patients presenting with Guidelines for the management of AF, LAAO for stroke pre-
hemoptysis [93]. vention in patients with AF and contraindications for long-­
term AC therapy has been given a class IIb recommendation
[105]. The goal of LAA occlusion is to prevent thrombi
Left Atrial Appendage Occluder Procedures formed in the LAA cavity from access to the LA and thus the
systemic circulation.
Justification for LAA Occlusion

Many patients with AF face both symptoms resulting in dim- Current LAA Occlusion Methods
inution in quality of life (addressed by lifestyle interven-
tions, anti-arrhythmic drugs, and catheter ablation) and an There are three categories of LAA occlusion: surgical-based
increased risk of stroke. Oral anticoagulation (AC) is the epicardial, catheter-based endocardial, and catheter-based
mainstay of prevention of stroke in AF patients. Currently, endoepicardial. In addition, there are hybrid techniques com-
although there are several different risk scores that are used bining surgical- and catheter-based techniques. Any of the
to identify patients with a high risk for stroke, patients with techniques that include an epicardial component result in
non-valvular AF are most commonly stratified for yearly both electrical and mechanical exclusions of the LAA, so
stroke risk using the CHADS2 or the newer CHA2DS2-­ they have the added benefit of potential eventual elimination
VASc score and are stratified for bleeding risk using the of LAA-initiated AF.
542 J. M. Lacomis et al.

Surgical-Based Epicardial approved by the FDA in 2009 for LAA exclusion during other
Traditional open surgical techniques include stapling, open cardiac surgeries in patients with a CHADS2 score of ≥2
suture ligation, and amputation done during other open and is now undergoing evaluation as a stand-­alone minimally
heart surgical procedures. The main interest in the devel- invasive thoracoscopic procedure in CHADS2 score of ≥2
opment of device-based LAA occluders is due to the low patients who do not need other cardiac surgeries [107]. The
overall success rate of only 40% complete LAA exclu- AtriClip is a self-closing clip made of two parallel rigid tita-
sion by these traditional surgical methods (0% for sta- nium rods encasing elastic nitinol springs and wrapped in a
ples, 23% for sutures, 73% for amputation) resulting in braided polyester sheath that is placed epicardially around the
“incomplete exclusion” or recanalization and ongoing, base of the LAA and externally compresses to exclude the
even increased, risk of LAA source thromboemboli LAA cavity from the systemic circulation (Fig. 43.11).
[106].
Multidetector CT is not routinely used in the preoperative Catheter-Based Endocardial
assessment of the LAA for the traditional surgically based pro- Catheter (percutaneous) endocardial device-based LAA
cedures; however, an incompletely surgically excluded LAA is occluders were originally described in 2001. This is a highly
occasionally encountered postoperatively even as an incidental evolving field with discontinuation of some of the earlier
finding when performing CT for other reasons (Fig. 43.10). devices, ongoing investigation of current utilized devices,
The AtriClip, or Gillinov-Cosgrove clip (Atricure, West and rapid introduction of new ones. Today, worldwide, the
Chester, OH), is a surgical device-based epicardial technique most widely used are the WATCHMAN device (Boston

Fig. 43.10  Incomplete LAA

exclusion: Multiplanar
reformat in a two-chamber
view of the heart and
corresponding volume
rendering showing an
incompletely excluded LAA
postsurgical epicardial LAA
suturing (arrow) with focal
central outpouching of
contrast from the LA into
the LA

Fig. 43.11  AtriClip LAAO:

Volume rendering and thick
MIP reconstruction of the
heart with a frontal cut
showing the typical
appearance of an AtriClip
placed across the neck of the
LAA during coronary artery
bypass grafting; there is no
residual LAA distal to the clip
consistent with complete
epicardial occlusion,
thrombosis, and atrophy of
the LAA. (Images courtesy of
Pal Suranyi MD, PhD;
Medical University of South
Carolina, Charleston, SC,
USA)
43  Cardiac CT: Electrophysiological Applications 543

Scientific, Natick, MA), the original Amplatzer Cardiac Plug trial failed to show noninferiority of the WATCHMAN to
(ACP), and its newer version, the Amplatzer Amulet (Abbott warfarin at preventing a combination of primary endpoints,
Medical, Chicago, IL). The LAmbre (Lifetech Scientific met predefined procedural safety outcomes, and showed
Corp., Shenzhen, China) and the WaveCrest (J&J BWI noninferiority to warfarin for the prevention of long-term
Cardiovascular & Specialty Solutions Group, Diamond Bar, ischemic events. FDA approval for the device was based on
CA) are some examples of more recently developed devices the combined data from the PROTECT, PREVAIL, and con-
and will be briefly mentioned. The access to the LAA for any tinued access protocols [11, 108, 109].
of the endocardial occluders is done via transeptal LA access Clinical eligibility for a WATCHMAN requires that a
using fluoroscopic and TEE guidance. patient has a demonstrable nonoptimal candidacy for antico-
The WATCHMAN device is CE mark approved and, in agulation, a CHADS score of ≥2/CHA2DS2-VASc score of
the USA as of July 2016, the only FDA-approved percutane- ≥3, and the ability to tolerate short-term anticoagulation.
ous method for LAA occlusion. The efficacy and safety of The device consists of a radial, self-expanding nitinol frame
the WATCHMAN were evaluated in two prospective, ran- with a 160 micron polyethylene terephthalate (PET) cap that
domized clinical trials, most notably the PROTECT-AF and blocks emboli from crossing it. Multiple fixation anchors
the PREVAIL trials [11, 108]. The PROTECT-AF trial around the perimeter of the frame attach the occluder to the
showed noninferiority of the WATCHMAN to warfarin and LAA endothelium. The WATCHMAN is positioned by both
for the first time implicated the LAA in the pathogenesis of intraprocedural TEE and fluoroscopy to accommodate the
stroke in AF; however, there were periprocedural safety con- largest device that can optimally occlude the LAA orifice
cerns including a 5% pericardial effusion rate. The PREVAIL [109, 110] (Fig. 43.12).

a b

c d

PET cap

Fixation Anchors

Fig. 43.12 (a–d): WATCHMAN LAAO: The WATCHMAN is an within the LA is abnormal and may warrant short-term anticoagulation
endocardial type of LAA occluder and is made of a self-expanding niti- (not shown). In (c), note the close proximity of the left main coronary
nol frame with a PET cap and ten fixation anchors around its periphery artery (arrow) and the LSPV to the LAA. (A, B, courtesy of Orly
to attach to the LAA endothelium (d). The appropriate landing zone for Goitein, MD, Sheba Medical Center. C, Image c courtesy of Pal
the WATCHMAN is the neck of the LAA (a, b). Thrombus within or Suranyi, MD, PhD, University of South Carolina. D, image d courtesy
distal to the WATCHMAN in the lobar LAA is an expected finding and of Boston Scientific. © 2017 Boston Scientific Corporation or its affili-
can fill the device (b) or be more laminar (a, thin white arrows); identi- ates. All rights reserved)
fication of thrombus on the outer surface (LAA side) of the PET cap or
544 J. M. Lacomis et al.

The ongoing Amulet IDE trial is a prospective, random- shaped device with a nitinol frame with anchors around the
ized, multicenter active control worldwide trial, designed to frame; an occlusive, polytetrafluoroethylene membrane cov-
evaluate the safety and effectiveness of the Amulet ers the LA side of the device, and a polymer foam around the
LAAO. Subjects will be randomized in a 1:1 ratio between device faces the LAA side.
the Amulet (treatment) and a WATCHMAN (control). The
trial will be conducted at up to 150 sites worldwide including Catheter-Based Endoepicardial
US sites [111]. The Lariat procedure is CE mark approved, currently only
The first-generation ACP was CE mark approved and FDA approved for suture placement and knot tying in surgi-
widely used in Europe [112]. The device design is based on cal procedures (i.e., soft tissue closure), but has not under-
the Amplatzer plugs that have a long-term success in atrial gone any clinical trials for stroke prevention in AF patients.
septal defect closure and other vascular procedures. The The Lariat procedure is used off-label when AF patients with
ACP has several components and is made of self-expanding a high risk of stroke have an absolute contraindication to
nitinol mesh anchored by multiple fixation hooks on a even short-term oral anticoagulation. There is an ongoing
polyester-­covered lobe positioned in the neck of the LAA FDA-approved clinical trial assessing the utility of LAA
with a flexible connecting waist extending to a proximal electrical isolation with the LARIAT device in conjunction
polyester-covered disc which fits across to cover the LAA with catheter-based pulmonary vein isolation in patients with
ostium on the LA side [113] (Fig.  43.13). The second-­ persistent AF to see whether freedom from AF is improved
generation ACP, the Amulet, is similar in design with sizing over conventional techniques [111]. It is a percutaneous,
and anchoring modifications made to accommodate larger catheter-based LAA epicardial suture ligation system that
LAA ostia and improve stability. requires both endocardial (trans-atrial septal) and epicardial
The LAmbre, CE mark approved and used in Europe, is (subxiphoid pericardial) access with the assistance of both
another self-expanding LAA occluder made of a nitinol endo- and epicardial-based magnets to stabilize the LAA for
umbrella with hooks around its perimeter connected by a placement.
short central waist to a PET-filled disc-shaped cover. Similar
to the ACP, the disc-shaped cover is positioned on the LA
side of the LAA ostium; the umbrella-shaped portion is posi- CT Technique
tioned approximately 5 mm distal to the LAA orifice within
the LAA neck. The umbrella portion can be collapsed to During deployment, typically TEE and fluoroscopy are used
facilitate repositioning during placement. The WaveCrest is for guidance. However, preprocedural planning relies on a
CE mark approved and used in Europe. It is another umbrella-­ combination of TEE and increasingly MDCT to assess the

Fig. 43.13  Amplatzer Amulet LAAO: The Amulet is an endocardial Amulet, and St. Jude Medical are trademarks of St. Jude Medical, LLC,
type of LAA occluder made of a lobe (white arrow), with multiple fixa- or its related companies. Reproduced with permission of St. Jude
tion hooks (*) positioned in the neck of the LAA, a disc positioned on Medical © 2017. All rights reserved; CT images of the Amulet courtesy
the LA side of the LAA ostium (black arrow), and a flexible rod waist of Orly Goitein, MD, Sheba Medical Center)
(open black arrows) connecting the lobe and disc. (AMPLATZER,
43  Cardiac CT: Electrophysiological Applications 545

interatrial septal, LA and LAA anatomy, and the anatomic The LAA is a complex and highly variable structure
relationship of the LAA to adjacent anatomy and exclude h­eterogeneously lined by fine pectinate muscles which
LAA thrombus. should not be mistaken for intracardiac thrombus. The LAA
The CT scanner requirements (minimum of 64 slice wall is thin; between the pectinate muscles, it is often exceed-
MDCT scanners) and scan technique including the use of a ingly paper thin, so it is at risk for perforation during the
second delayed acquisition through the LAA are the same as manipulation of catheters and device seating. The LAA is
previously detailed for pre-PVI and post-PVI imaging with a composed of three major components: an ostium, a neck
few minor modifications. (clinically referred to as the landing zone), and a lobar por-
As with pre-PVI patients, the z-axis of the scan starts at the tion (Fig. 43.14).
mid-aortic arch, so the most cephalad extent of the LAA is
imaged and covers the heart or the xiphoid process whichever  he LAA Ostium
T
is more distal (for Lariat evaluation). Similar to pre-PVI The LAA ostium is the well-defined junctional orifice
imaging, the decision to ECG gate or not gate the acquisition between the LA and the LAA.  LAA ostia can be ovoid,
will be site dependent on the ability of the available scanner round, triangular, or waterdrop like or foot shaped [114]. The
to obtain high-quality images of the LA and LAA in the set- shape of the LAA ostium may influence device selection.
ting of AF.  Although, in normal sinus rhythm, maximum The less round ostia are potentially greater risks for peride-
LAA dimensions occur during atrial diastole (ventricular sys- vice leaks [117]. Maximum ostial diameters and circumfer-
tole), in persistent AF, the LAA dimensions change little over ence are key components for occluder selection, since the
the course of the cardiac cycle. While a dedicated FOV for the occluders are available in fixed sizes (Fig. 43.14).
heart is preferred, post-scan, reconstructing the data set with The LAA ostium is positioned above the left atrioven-
the FOV opened to include the anterior chest wall, is needed. tricular groove, typically at the level of the LSPV but can
arise superiorly or inferiorly to the LSPV (Figs. 43.14 and
43.15). Inferior LAA ostia are also more posteriorly located
Postprocessing and are notable because of their close proximity to the
mitral valve [114].
Postprocessing for all occluders includes epicardial vantage The lateral ridge (i.e., Ridge of Marshall, “Coumadin,” or
3D volume renderings of the LA and distal PVs with the “Warfarin” ridge) is the infolding of the LA wall that sepa-
LAA and of the LAA by itself. For potential Lariat patients, rates the LAA ostium from the adjacent PVs (Fig.  43.4b)
additional 3D volume renderings include the LA and distal [118]. The ridge is of variable widths, dependent on the
PV model with addition of the MPA and with addition of the amount of epicardial fat in its sulcus or groove on the extra-­
imaged sternum including the xiphoid and at least the atrial side, so the distance between the LAA ostium and the
­anterior ribs. left PVs varies between patients. The left PVs, particularly
LAA ostial sizing and additional LAA sizing parameters the LSPV, are at increased risk for compression by LAA
delineated below are most accurately measured on MDCT occluders [118]. There are no current recommendations for
utilizing the MPRs to obtain true orthogonal obliques, providing CT measurements of the lateral ridge or measuring
although measurements have been required for some devices the distance between the left PVs and the LAA ostium.
on the volume renderings [114]. It is important to be aware Of note, other vascular structures can course between the
that measurements can vary with patient volume status. In LAA and LSPV on the epicardial side of the lateral ridge.
addition, early experience with 3D printed models has shown These include a persistent left-sided SVC and the S-shaped
their potential utility in preoperative planning for cardiovas- sinoatrial nodal (SAN) artery [91, 119].
cular surgical procedures; others have suggested that, simi-
larly, 3D printing of the LA and LAA may be helpful in  he LAA Neck
T
pre-LAAO planning [115, 116]. The neck of the LAA is the tubular junction between the
ostium and the major lobar region and is the landing zone for
anchoring the endocardial LAA occluders (Fig. 43.14). It is
Interpretation Pre-LAA Occluder typically relatively smooth, but occasionally smaller second-
ary lobes can arise from the neck. Early branching of second-
Left Atrial Appendage ary lobes from the neck can interfere with device seating or,
when proximal to the landing zone of the WATCHMAN
The specific LAA measurements mentioned below are pri- (<10 mm from LAA ostium), may lead to peridevice leak.
marily obtained using TEE; however, MDCT measurements The angle between the ostium, neck, and main lobe is
are increasingly being used as adjuncts so familiarity with reported on CT and TEE and influences the choice of deliv-
these measurements is necessary. ery sheath; sharp angulation between the neck and main lobe
546 J. M. Lacomis et al.

a b c

d e f

Fig. 43.14 (a–f): Left atrial appendage: This windsock shape LAA is orthogonal oblique perpendicular to the ostium; diameters are derived
divided into the ostium (a, dashed black line), interposing neck which from the perimeter or can be measured directly. In (a–c), note all of the
is the landing zone for occluders (a, solid white line shows length of the vascular structures in close proximity to the LAA neck and ostium. The
neck; dashed white line denotes distal margin) and the lobar region (a, LAA ostium comes in varying shapes, including oval and tear-dropped
solid black line). Although standardly obtained on TEE, pre-LAAO (d). The endocardial view (e) shows the superior location of this LAA
measurements include the length of the largest lobe (a, solid black line) ostium compared to the LSPV ostium, separated by the Ridge of
and angle between the ostium, neck, and dominant lobe (b); in (c) Marshall (**)
perimeter of the LAA ostium (black dashes) is obtained from a true

can interfere with endocardial device manipulation to the difficult to categorize. Furthermore, as AF becomes persistent,
landing zone. the LAA undergoes structural remodeling, and some of the
complexity of the lobulations is lost as the entire structure
 he LAA Lobar Region
T dilates and loses muscle mass and contractility.
The lobar portion of the LAA is the longest part and can Lobar anatomy, such as maximum available working
consist of single, double, or multiple lobes with varying depth for device deployment and potential prominent pecti-
degrees of trabeculation. In 2010, Wang et  al. did an ana- nate musculature, may affect suitability for a WATCMAN as
tomical study of LAA morphology on MDCT and developed the feet are situated within the lobar region. Conversely,
a classification system, based on the presence of absence of devices such as the Amulet are positioned more ostially. As
an obvious bend in the LAA [114]. detailed earlier in this chapter, assessment of the LAA for
An LAA with an obvious bend in the proximal or middle thrombus vs pseudothrombus from underfilling and mixing
part of the dominant lobe is classified as a “chicken wing” artifacts requires scrutiny of both the early and late images.
shape. LAA without obvious bends in the main lobar region Pectinate muscles, fluid in adjacent pericardial recesses, and
are classified as windsock (one dominant lobe), cauliflower fat interposed between the LSPV and LAA should not be
(limited overall length with more complex internal structures), mistaken for thrombus.
and cactus (dominant central lobe with secondary lobes Orientation of the LAA behind the main pulmonary artery
extending outward both superiorly and inferiorly) shapes and maximum transverse approach dimension of 45 mm or
[114] (Fig. 43.15). There is an overlap in LAA morphological greater, which exceeds the size of the snare, are also relevant
features among the different categories making some LAA contraindications for the Lariat procedure (Fig. 43.16).
43  Cardiac CT: Electrophysiological Applications 547

Fig. 43.15  Wings and things

(a–d): LAA lobar
a b
morphology is classified into
shapes depending on the
presence or absence of an
obvious bend within the main
lobar region. LAA ostial
shapes and position in
relationship to the LSPV are
variable. (a) Chicken wing
(bend in the lobar region), (b)
windsock (bend is at the
junction of neck and lobar
region), (c) cauliflower, (d)
cactus. Note that in these
examples, the LAA ostium is
positioned inferior to the
LSPV ostium, whereas in the c d
earlier Fig. 43.1b, the LAA
ostium is at the level of the
PV’s and in Figs. 43.4d and
43.14f, the LAA ostium is
superior to the left common
vein ostium

Fig. 43.16  Suitability for

a b
Lariat procedure: The position
of the tip of the LAA (arrow)
in relationship to the main
pulmonary artery (MPA) is
one of the anatomic factors
determining suitability for a
Lariat procedure. In (a) same
patient as in Fig. 43.14a, the
tip of the LAA is posterior to
the MPA, so it is not
reachable from a subxiphoid
approach precluding
candidacy for a Lariat,
whereas in (b), same patient
as in 43.14b, the LAA is
directed anteriorly and to the
left of the MPA which is a
favorable LAA tip location
for a Lariat

Extra-atrial Anatomy thickening or calcification, sternal-pericardial adhesions,

Extra-atrial anatomy needs to be scrutinized for potential and unfavorable sternal or chest wall anatomy such as pec-
Lariat candidates, as there are additional exclusions for the tus excavatum or severe scoliosis. Patients with a history of
Lariat system involving extra-atrial structures that can prior cardiac surgery or thoracic radiation with the excep-
interfere with the pericardial puncture or the epicardial tion of localized breast radiation are excluded prior to
access route: a posteriorly oriented heart, any pericardial imaging.
548 J. M. Lacomis et al.

Interpretation Post-LAA Occluder catheters, or devices themselves and usually require emer-
gency treatment [117, 121].
Complications of the endocardial occluders include all of the During the Lariat procedure, since the pericardium is
same complications related to vascular and trans-atrial septal directly punctured for access, pericarditis and pericardial
access as the catheter-based ablation procedures and some effusion are common; a pericardial drain is typically placed
specific to the LAAO. Echocardiography is the primary for 24 h post-procedure, and short-term colchicine therapy is
imaging modality for the assessment of cardiac-related routinely administered [121, 122]. Postop CT should include
LAAO complications, but some are encountered on CT, so evaluation of the pericardium for the accumulation of the
familiarity with possible complications is prudent. pericardial effusion post-drain removal, hemopericardium,
ongoing CT signs of pericarditis, or post-pericardiectomy
 trial Septal Defect
A syndrome as well as a search for other etiologies for chest
Similar to the ablation procedures, trans-atrial septal access pain and dyspnea.
can result in a persistent iatrogenic atrial septal defect, which
can be seen on CT post-LAAO; the defect is usually small,  evice Embolization and Incomplete LAA
D
<5  mm, usually being followed on echocardiography, and Occlusion
typically doesn’t require treatment (Fig. 43.17) [120]. They Undersizing or inadequate seating of the endocardial devices
may persist longer than a year but frequently eventually can result in device migration or embolization requiring
resolve spontaneously [120]. catheter or surgical retrieval; rates are low, reported between
0% and 3% [108, 117]. Similarly, peridevice leaks can result
 ardiac Perforation and Pericardial
C in residual flow into the LAA between the margin of the
Complications device and the wall of the LAA; if the peridevice leak is less
The more serious complication of cardiac perforation pres- than 5  mm wide, it is considered small and followed with
ents during the LAAO procedure. Similar to the ablation pro- echocardiography [117]. Peridevice leaks are common. A
cedures, incorrect location of the trans-atrial septal puncture, substudy of the PROTECT trial showed that 32% of the
beyond the region of the fossa ovalis, is one of the causes of WATCHMAN patients had at least some degree of peride-
cardiac perforation and periprocedural significant pericardial vice flow at 1 year [123]; multicenter experience in Europe
effusion/hemopericardium and possible tamponade. Cardiac reported 16% of ACP patients had a small peridevice leak at
perforation, hemopericardium, and tamponade can also 6  months [124]. The Lariat suture may also incompletely
occur with perforation of the LA or LAA by the guidewires, occlude the LAA, but similar to other epicardial-based tech-
niques, the leak is central. Malpositioned or oversized
devices may also overlap the contiguous LSPV or the mitral
valve leaflet.

Thrombus Formation
Over time, the endocardial devices endothelialize. CT identi-
fication of thrombus on the LAA side distal to an endocardial
occluder or epicardial closure site is an expected finding.
Thrombus on the LA side may indicate a need for short-term
anticoagulation therapy. In the PROTECT trial, the incidence
of some thrombus on the LA side of the WATCHMAN was
approximately 4%. A higher incidence of thrombus on the
flexible connecting waist of the ACP of approximately 14%
was one of the reasons the device was redesigned.

Conclusion

The increasing disease burden caused by atrial fibrillation

has led to new interventions aimed at restoring normal sinus
rhythm and preventing embolic complications. Cardiac CT
Fig. 43.17  Persistent atrial septal defect: A small, <5 mm ASD (arrow) can help identify suitable candidates, plan the intervention,
and left to right shunt is persistent 6  months post-trans-atrial septal
puncture for CBA; these small ASDs can occur with any of the proce- and diagnose complications. In some institutions, this appli-
dures utilizing trans-atrial septal puncture including RFCA and LAAO cation has become the most common indication for the
43  Cardiac CT: Electrophysiological Applications 549

c­ linical use of cardiac CT. Consequently, expertise in mak- 16. Pappone C, et  al. Circumferential radiofrequency ablation of
ing the pertinent observations outlined in this chapter is of pulmonary vein ostia: a new anatomic approach for curing atrial
fibrillation. Circulation. 2000;102(21):2619–28.
critical importance for the success of any cardiac CT pro- 17. Lee R, et al. The closed heart MAZE: a nonbypass surgical tech-
gram and offers a tremendous opportunity for interdisciplin- nique. Ann Thorac Surg. 1999;67(6):1696–702.
ary collaboration. 18. Haissaguerre M, et al. Successful catheter ablation of atrial fibril-
lation. J Cardiovasc Electrophysiol. 1994;5(12):1045–52.
19. Packer DL, et al. Clinical presentation, investigation, and manage-
ment of pulmonary vein stenosis complicating ablation for atrial
References fibrillation. Circulation. 2005;111(5):546–54.
20. Schwartzman D, Bazaz R, Nosbisch J. Catheter ablation to sup-
1. Go AS, et al. Prevalence of diagnosed atrial fibrillation in adults: press atrial fibrillation: evolution of technique at a single center. J
national implications for rhythm management and stroke preven- Interv Card Electrophysiol. 2003;9(2):295–300.
tion: the AnTicoagulation and Risk Factors in Atrial Fibrillation 21. Pappone C, et al. Pulmonary vein denervation enhances long-term
(ATRIA) Study. JAMA. 2001;285(18):2370–5. benefit after circumferential ablation for paroxysmal atrial fibril-
2. Miyasaka Y, et  al. Secular trends in incidence of atrial fibrilla- lation. Circulation. 2004;109(3):327–34.
tion in Olmsted County, Minnesota, 1980 to 2000, and impli- 22. Verma A, et al. Approaches to catheter ablation for persistent atrial
cations on the projections for future prevalence. Circulation. fibrillation. N Engl J Med. 2015;372(19):1812–22.
2006;114(2):119–25. 23. Packer DL, et  al. Cryoballoon ablation of pulmonary veins for
3. Grond M, et  al. Improved detection of silent atrial fibril- paroxysmal atrial fibrillation: first results of the North American
lation using 72-hour Holter ECG in patients with isch- Arctic Front (STOP AF) pivotal trial. J Am Coll Cardiol.
emic stroke: a prospective multicenter cohort study. Stroke. 2013;61(16):1713–23.
2013;44(12):3357–64. 24. Lustgarten DL, Keane D, Ruskin J. Cryothermal ablation: mecha-
4. Henriksson KM, et al. Comparison of cardiovascular risk factors nism of tissue injury and current experience in the treatment of
and survival in patients with ischemic or hemorrhagic stroke. Int J tachyarrhythmias. Prog Cardiovasc Dis. 1999;41(6):481–98.
Stroke. 2012;7(4):276–81. 25. Kuck K-H, et al. Cryoballoon or radiofrequency ablation for par-
5. Kishore A, et  al. Detection of atrial fibrillation after ischemic oxysmal atrial fibrillation. N Engl J Med. 2016;374(23):2235–45.
stroke or transient ischemic attack: a systematic review and meta-­ 26. Khoueiry Z, et al. Outcomes after cryoablation vs. radiofrequency
analysis. Stroke. 2014;45(2):520–6. in patients with paroxysmal atrial fibrillation: impact of pulmo-
6. Investigators, T.A.F.F.-u.I.o.R.M.  A comparison of rate control nary veins anatomy. Europace. 2016;18(9):1343–51.
and rhythm control in patients with atrial fibrillation. N Engl J 27. Ang R, et  al. The hot and the cold: radiofrequency versus
Med. 2002;347(23):1825–33. cryoballoon ablation for atrial fibrillation. Curr Cardiol Rep.
7. Fuster V, et  al. ACC/AHA/ESC guidelines for the manage- 2015;17(9):631.
ment of patients with atrial fibrillation: executive summary a 28. Guhl EN, et  al. Efficacy of cryoballoon pulmonary vein isola-
report of the American College of Cardiology/American Heart tion in patients with persistent atrial fibrillation. J Cardiovasc
Association Task Force on Practice Guidelines and the European Electrophysiol. 2016;27(4):423–7.
Society of Cardiology Committee for Practice Guidelines and 29. Rillig A, et al. Six-year clinical outcomes after catheter ablation
Policy Conferences (Committee to Develop Guidelines for the of atrial fibrillation in patients with impaired left ventricular func-
Management of Patients With Atrial Fibrillation) developed in tion. J Cardiovasc Electrophysiol. 2015;26:1169.
collaboration with the North American society of pacing and elec- 30. Kholova I, Kautzner J.  Morphology of atrial myocardial exten-
trophysiology. Circulation. 2001;104(17):2118–50. sions into human caval veins: a postmortem study in patients with
8. Pathak RK, et  al. Impact of CARDIOrespiratory FITness and without atrial fibrillation. Circulation. 2004;110(5):483–8.
on arrhythmia recurrence in obese individuals with atrial 31. Kurotobi T, et  al. Marshall vein as arrhythmogenic source in
fibrillation: the CARDIO-FIT study. J Am Coll Cardiol. patients with atrial fibrillation: correlation between its anatomy
2015;66(9):985–96. and electrophysiological findings. J Cardiovasc Electrophysiol.
9. Pathak RK, et al. Long-term effect of goal-directed weight man- 2006;17(10):1062–7.
agement in an atrial fibrillation cohort: a long-term follow-up 32. Scharf C, et al. Anatomy of the pulmonary veins in patients with
study (LEGACY). J Am Coll Cardiol. 2015;65(20):2159–69. atrial fibrillation and effects of segmental ostial ablation ana-
10. Kim MH, et al. Clinical outcomes and costs associated with a first lyzed by computed tomography. J Cardiovasc Electrophysiol.
episode of uncomplicated atrial fibrillation presenting to the emer- 2003;14(2):150–5.
gency room. Am J Cardiol. 2001;88(1):A7. 74-6. 33. Dill T, et  al. Pulmonary vein diameter reduction after radiofre-
11. Holmes DR Jr, et  al. Prospective randomized evaluation of the quency catheter ablation for paroxysmal atrial fibrillation evalu-
Watchman Left Atrial Appendage Closure device in patients with ated by contrast-enhanced three-dimensional magnetic resonance
atrial fibrillation versus long-term warfarin therapy: the PREVAIL imaging. Circulation. 2003;107(6):845–50.
trial. J Am Coll Cardiol. 2014;64(1):1–12. 34. Lacomis JM, et al. Multi-detector row CT of the left atrium and
12. Haissaguerre M, et al. Spontaneous initiation of atrial fibrillation pulmonary veins before radio-frequency catheter ablation for
by ectopic beats originating in the pulmonary veins. N Engl J atrial fibrillation. Radiographics. 2003;23 Spec No:S35–48. dis-
Med. 1998;339(10):659–66. cussion S48–50.
13. Shah DC, et al. Electrophysiologically guided ablation of the pul- 35. Lacomis J, et  al. Direct comparison of CT & MRI for charac-
monary veins for the curative treatment of atrial fibrillation. Ann terization of PLA morphology. J Interv Card Electrophysiol.
Med. 2000;32(6):408–16. 2006;16(1):7–13.
14. Lacomis JM, et al. CT of the pulmonary veins. J Thorac Imaging. 36. Kato R, et  al. Pulmonary vein anatomy in patients undergoing
2007;22(1):63–76. catheter ablation of atrial fibrillation: lessons learned by use of
15. Oral H, et al. Catheter ablation for paroxysmal atrial fibrillation: magnetic resonance imaging. Circulation. 2003;107(15):2004–10.
segmental pulmonary vein ostial ablation versus left atrial abla- 37. Schwartzman D, Lacomis J, Wigginton WG.  Characterization
tion. Circulation. 2003;108(19):2355–60. of left atrium and distal pulmonary vein morphology using
550 J. M. Lacomis et al.

multidimensional computed tomography. J Am Coll Cardiol. 57. Wittkampf FH, et al. Pulmonary vein ostium geometry: analysis by
2003;41(8):1349–57. magnetic resonance angiography. Circulation. 2003;107(1):21–3.
38. Cronin P, et al. MDCT of the left atrium and pulmonary veins in 58. Duerinckx AJ, Vanovermeire O. Accessory appendages of the left
planning radiofrequency ablation for atrial fibrillation: a how-to atrium as seen during 64-slice coronary CT angiography. Int J
guide. AJR Am J Roentgenol. 2004;183(3):767–78. Cardiovasc Imaging. 2008;24(2):215–21.
39. Lacomis J, et al. 3D-Multidetector CT and variations in pulmo- 59. Romero J, et al. Left atrial appendage closure devices. Clin Med
nary venous and left atrial anatomy; implications in atrial fibril- Insights Cardiol. 2014;8(4251-CMC-Left-Atrial-Appendage-
lation patients undergoing ablative therapy. Radiology. 2002;S Closure-Devices.pdf):45–52.
225:631. 60. Lazoura O, et al. Prevalence of left atrial anatomical abnormalities
40. Marom EM, et al. Variations in pulmonary venous drainage to the in patients with recurrent atrial fibrillation compared with patients
left atrium: implications for radiofrequency ablation. Radiology. in sinus rhythm using multi-slice CT.  J Cardiovasc Comput
2004;230(3):824–9. Tomogr. 2012;6(4):268–73.
41. Cronin P. 2D or Not 2D that is the question, but 3D is the answer. 61. Sosa E, Scanavacca M. Left atrial-esophageal fistula complicating
Acad Radiol. 2007;14(7):769–71. radiofrequency catheter ablation of atrial fibrillation. J Cardiovasc
42. Cronin P, et al. Reliability of MDCT in characterizing pulmonary Electrophysiol. 2005;16(3):249–50.
venous drainage, diameter and distance to first bifurcation: an 62. Scanavacca MI, et  al. Left atrial-esophageal fistula following
interobserver study. Acad Radiol. 2007;14(4):437–44. radiofrequency catheter ablation of atrial fibrillation. J Cardiovasc
43. Cronin P, et  al. Normative analysis of pulmonary vein drainage Electrophysiol. 2004;15(8):960–2.
patterns on multidetector CT with measurements of pulmonary 63. Kelly S, Bicknell SG, Sharma S.  Left atrial wall hematoma
vein ostial diameter and distance to first bifurcation. Acad Radiol. after radiofrequency ablation for atrial fibrillation. AJR Am J
2007;14(2):178–88. Roentgenol. 2006;186(5):1317–9.
44. Lickfett L, et  al. Changes of pulmonary vein orifice size and 64. Ramakrishna G, et  al. Endocardial flap of left atrial dissection
location throughout the cardiac cycle: dynamic analysis using following radiofrequency ablation. Pacing Clin Electrophysiol.
magnetic resonance cine imaging. J Cardiovasc Electrophysiol. 2003;26(8):1771–3.
2005;16(6):582–8. 65. Park HW, et al. Disseminated intravascular coagulation as a com-
45. Hsieh MH, et al. Double multielectrode mapping catheters facili- plication of radiofrequency catheter ablation of atrial fibrillation. J
tate radiofrequency catheter ablation of focal atrial fibrillation Cardiovasc Electrophysiol. 2005;16(9):1011–3.
originating from pulmonary veins. J Cardiovasc Electrophysiol. 66. Sonmez B, Demirsoy E, Yilmaz O. Atrioesophageal fistula: is it
1999;10(2):136–44. an unavoidable complication of radiofrequency ablation? J Thorac
46. Zhong H, Lacomis JM, Schwartzman D.  On the accuracy of Cardiovasc Surg. 2003;126(5):1662–3. author reply 1663.
CartoMerge for guiding posterior left atrial ablation in man. Heart 67. Gillinov AM, Pettersson G, Rice TW.  Esophageal injury during
Rhythm. 2007;4(5):595–602. radiofrequency ablation for atrial fibrillation. J Thorac Cardiovasc
47. Bhakta D, Miller JM.  Principles of electroanatomic mapping. Surg. 2001;122(6):1239–40.
Indian Pacing Electrophysiol J. 2008;8(1):32–50. 68. Doll N, et  al. Esophageal perforation during left atrial radiofre-
48. Choi SI, et  al. Variation of the size of pulmonary venous ostia quency ablation: is the risk too high? J Thorac Cardiovasc Surg.
during the cardiac cycle: optimal reconstruction window at ECG- 2003;125(4):836–42.
gated multi-detector row CT. Eur Radiol. 2005;15(7):1441–5. 69. Bunch TJ, et al. Temporary esophageal stenting allows healing of
49. Merchant FM, et al. Pulmonary vein remodeling following atrial esophageal perforations following atrial fibrillation ablation pro-
fibrillation ablation: implications for the radiographic diagnosis of cedures. J Cardiovasc Electrophysiol. 2006;17(4):435–9.
pulmonary vein stenosis. J Atr Fibrillation. 2016;9(2):1453. 70. Schley P, Gulker H, Horlitz M.  Atrio-oesophageal fistula fol-
50. Wongcharoen W, et  al. Morphologic characteristics of the left lowing circumferential pulmonary vein ablation: verification
atrial appendage, roof, and septum: implications for the ablation of of diagnosis with multislice computed tomography. Europace.
atrial fibrillation. J Cardiovasc Electrophysiol. 2006;17(9):951–6. 2006;8(3):189–90.
51. Gal P, Marrouche NF.  Magnetic resonance imaging of atrial 71. Pappone C, et al. Atrio-esophageal fistula as a complication of per-
fibrosis: redefining atrial fibrillation to a syndrome. Eur Heart J. cutaneous transcatheter ablation of atrial fibrillation. Circulation.
2017;38(1):14–9. 2004;109(22):2724–6.
52. Calkins H, et al. HRS/EHRA/ECAS expert consensus statement 72. Malamis AP, Kirshenbaum KJ, Nadimpalli S.  CT radio-
on catheter and surgical ablation of atrial fibrillation: recom- graphic findings: atrio-esophageal fistula after transcatheter
mendations for personnel, policy, procedures and follow-up. A percutaneous ablation of atrial fibrillation. J Thorac Imaging.
report of the Heart Rhythm Society (HRS) Task Force on cath- 2007;22(2):188–91.
eter and surgical ablation of atrial fibrillation. Heart Rhythm. 73. Ripley KL, et  al. Time course of esophageal lesions after cath-
2007;4(6):816–61. eter ablation with cryothermal and radiofrequency ablation:
53. Ghaye B, et  al. Percutaneous ablation for atrial fibrillation: the implication for atrio-esophageal fistula formation after cath-
role of cross-sectional imaging. Radiographics. 2003;23 Spec eter ablation for atrial fibrillation. J Cardiovasc Electrophysiol.
No:S19–33. discussion S48–50. 2007;18(6):642–6.
54. Lacomis J, et  al. Ablation imaging of left atrial and pulmonary 74. Lemola K, et al. Computed tomographic analysis of the anatomy
venous anatomy in atrial fibrillation patients: comparison of of the left atrium and the esophagus: implications for left atrial
3D-multidetector CT angiography and magnetic resonance angi- catheter ablation. Circulation. 2004;110(24):3655–60.
ography. Radiology. 2002;S 225:626. 75. Cury RC, et al. Relationship of the esophagus and aorta to the left
55. Mangrum JM, et  al. Intracardiac echocardiography-guided, ana- atrium and pulmonary veins: implications for catheter ablation of
tomically based radiofrequency ablation of focal atrial fibril- atrial fibrillation. Heart Rhythm. 2005;2(12):1317–23.
lation originating from pulmonary veins. J Am Coll Cardiol. 76. Gillinov AM, et al. Esophageal perforation during left atrial radio-
2002;39(12):1964–72. frequency ablation: is the risk too high? J Thorac Cardiovasc Surg.
56. Tsao HM, et al. Role of right middle pulmonary vein in patients 2003;126(5):1661–2. author reply 1662.
with paroxysmal atrial fibrillation. J Cardiovasc Electrophysiol. 77. Pappone C, et al. Atrial electroanatomic remodeling after circum-
2001;12(12):1353–7. ferential radiofrequency pulmonary vein ablation: efficacy of an
43  Cardiac CT: Electrophysiological Applications 551

anatomic approach in a large cohort of patients with atrial fibrilla- 99. Burrell LD, et al. Usefulness of left atrial appendage volume as a
tion. Circulation. 2001;104(21):2539–44. predictor of embolic stroke in patients with atrial fibrillation. Am
78. Bunch TJ, et al. Outcomes after cardiac perforation during radio- J Cardiol. 2013;112(8):1148–52.
frequency ablation of the atrium. J Cardiovasc Electrophysiol. 100. Di Biase L, et al. Atrial fibrillation ablation strategies for parox-
2005;16(11):1172–9. ysmal patients: randomized comparison between different tech-
79. Rostamian A, et  al. The incidence, diagnosis, and management niques. Circ Arrhythm Electrophysiol. 2009;2(2):113–9.
of pulmonary vein stenosis as a complication of atrial fibrillation 101. Anselmino M, et al. Left atrial appendage morphology and silent
ablation. J Interv Card Electrophysiol. 2014;40(1):63–74. cerebral ischemia in patients with atrial fibrillation. Heart Rhythm.
80. Ravenel JG, McAdams HP.  Pulmonary venous infarction 2014;11(1):2–7.
after radiofrequency ablation for atrial fibrillation. AJR Am J 102. Yamamoto M, et  al. Complex left atrial appendage morphology
Roentgenol. 2002;178(3):664–6. and left atrial appendage thrombus formation in patients with
81. Saad EB, et  al. Pulmonary vein stenosis after catheter ablation atrial fibrillation. Circ Cardiovasc Imaging. 2014;7(2):337–43.
of atrial fibrillation: emergence of a new clinical syndrome. Ann 103. Fluckiger JU, et al. Left atrial flow velocity distribution and flow
Intern Med. 2003;138(8):634–8. coherence using four-dimensional FLOW MRI: a pilot study
82. Burgstahler C, et  al. Visualization of pulmonary vein stenosis investigating the impact of age and Pre- and Postintervention
after radio frequency ablation using multi-slice computed tomog- atrial fibrillation on atrial hemodynamics. J Magn Reson Imaging.
raphy: initial clinical experience in 33 patients. Int J Cardiol. 2013;38(3):580–7.
2005;102(2):287–91. 104. Blackshear JL, Odell JA. Appendage obliteration to reduce stroke
83. Yu WC, et  al. Acquired pulmonary vein stenosis after radio- in cardiac surgical patients with atrial fibrillation. Ann Thorac
frequency catheter ablation of paroxysmal atrial fibrillation. J Surg. 1996;61(2):755–9.
Cardiovasc Electrophysiol. 2001;12(8):887–92. 105. Kirchhof P, et  al. 2016 ESC Guidelines for the management of
84. Saad EB, et al. Pulmonary vein stenosis after radiofrequency abla- atrial fibrillation developed in collaboration with EACTS.  Eur
tion of atrial fibrillation: functional characterization, evolution, and Heart J. 2016;37(38):2893–962.
influence of the ablation strategy. Circulation. 2003;108(25):3102–7. 106. Kanderian AS, et al. Success of surgical left atrial appendage clo-
85. Mohr FW, et  al. Curative treatment of atrial fibrillation with sure: assessment by transesophageal echocardiography. J Am Coll
intraoperative radiofrequency ablation: short-term and midterm Cardiol. 2008;52(11):924–9.
results. J Thorac Cardiovasc Surg. 2002;123(5):919–27. 107. Ellis CR, Aznaurov SG. The left atrial appendage and thromboem-
86. Stroker E, et  al. Anatomic predictors of phrenic nerve injury in bolic stroke in atrial fibrillation: rationale for ligation by an epi-
the setting of pulmonary vein isolation using the 28-mm second-­ cardial approach: the AtriClip stroke trial. J Innov Card Rhythm
generation cryoballoon. Heart Rhythm. 2016;13(2):342–51. Manag. 2015;6:2052–6.
87. Hai JJ, Tse HF. Phrenic nerve palsy in cryoballoon ablation: can it 108. Holmes DR, et al. Percutaneous closure of the left atrial append-
be prevented? Heart Rhythm. 2016;13(2):352–3. age versus warfarin therapy for prevention of stroke in patients
88. Andrade JG, et al. The biophysics and biomechanics of cryobal- with atrial fibrillation: a randomised non-inferiority trial. Lancet.
loon ablation. Pacing Clin Electrophysiol. 2012;35(9):1162–8. 2009;374(9689):534–42.
89. Horton R, et al. Locating the right phrenic nerve by imaging the 109. Reddy VY, et  al. Percutaneous left atrial appendage closure for
right pericardiophrenic artery with computerized tomographic stroke prophylaxis in patients with atrial fibrillation: 2.3-Year
angiography: implications for balloon-based procedures. Heart Follow-up of the PROTECT AF (Watchman Left Atrial Appendage
Rhythm. 2010;7(7):937–41. System for Embolic Protection in Patients with Atrial Fibrillation)
90. Matsumoto Y, et al. Detection of phrenic nerves and their relation Trial. Circulation. 2013;127(6):720–9.
to cardiac anatomy using 64-slice multidetector computed tomog- 110. Belgaid DR, et  al. Prospective randomized evaluation of the
raphy. Am J Cardiol. 2007;100(1):133–7. watchman left atrial appendage closure device in patients with
91. Cummings KW, et  al. Cross-sectional imaging anatomy and atrial fibrillation versus long-term warfarin therapy: the PREVAIL
pathologic conditions affecting thoracic nerves. Radiographics. trial. Int J Cardiol. 2016;219:177–9.
2017;37(1):73–92. 111. Lee RJ, et  al. Percutaneous alternative to the Maze procedure
92. Verma N, et al. Bronchial effects of cryoballoon ablation for atrial for the treatment of persistent or long-standing persistent atrial
fibrillation. Heart Rhythm. 2017;14(1):17–8. fibrillation (aMAZE trial): rationale and design. Am Heart J.
93. Bessiere F, Chevalier P.  Pulmonary vein hematoma after atrial 2015;170(6):1184–94.
fibrillation cryoablation: a new complication. Heart Rhythm. 112. Tzikas A, Sievert H, Holmes DR Jr. Percutaneous left atrial append-
2013;10(9):1359. age occlusion in 2016. EuroIntervention. 2016;11(14):e1576–8.
94. Lip GYH, et al. Refining clinical risk stratification for predicting 113. Khattab AA, Meier B.  Transcatheter left atrial append-

stroke and thromboembolism in atrial fibrillation using a novel age exclusion, gold or fool's gold? Eur Heart J Suppl.
risk factor-based approach: The euro heart survey on atrial fibril- 2010;12(suppl_E):E35–40.
lation. Chest. 2010;137(2):263–72. 114. Wang Y, et al. Left atrial appendage studied by computed tomog-
95. Kosior DA. Risk stratification schemes for stroke in atrial fibril- raphy to help planning for appendage closure device placement. J
lation: the predictive factors still undefined. Pol Arch Med Wewn. Cardiovasc Electrophysiol. 2010;21(9):973–82.
2015;125(12):889–90. 115. Pellegrino PL, et al. Left atrial appendage closure guided by 3D
96. Pisters R, et al. A novel user-friendly score (HAS-BLED) to assess printed cardiac reconstruction: emerging directions and future
1-year risk of major bleeding in patients with atrial fibrillation: the trends. J Cardiovasc Electrophysiol. 2016;27(6):768–71.
Euro Heart Survey. Chest. 2010;138(5):1093–100. 116. Giannopoulos AA, et  al. Cardiothoracic applications of
97. Steinberg BA, et  al. Higher risk of death and stroke in patients 3-­dimensional printing. J Thorac Imaging. 2016;31(5):253–72.
with persistent vs. paroxysmal atrial fibrillation: results from the 117. Wunderlich NC, et  al. Percutaneous interventions for left atrial
ROCKET-AF Trial. Eur Heart J. 2015;36(5):288–96. appendage exclusion: options, assessment, and imaging using
98. Lee G, Sanders P, Kalman JM. Catheter ablation of atrial arrhyth- 2D and 3D echocardiography. JACC Cardiovasc Imaging.
mias: state of the art. Lancet. 2012;380(9852):1509–19. 2015;8(4):472–88.
552 J. M. Lacomis et al.

118. Cabrera JA, Saremi F, Sanchez-Quintana D.  Left atrial append- 122. Stone D, Byrne T, Pershad A.  Early results with the LARIAT
age: anatomy and imaging landmarks pertinent to percutaneous device for left atrial appendage exclusion in patients with atrial
transcatheter occlusion. Heart. 2014;100(20):1636–50. fibrillation at high risk for stroke and anticoagulation. Catheter
119. Saremi F, et  al. Arterial supply to sinuatrial and atrioven- Cardiovasc Interv. 2015;86(1):121–7.
tricular nodes: imaging with multidetector CT.  Radiology. 123. Viles-Gonzalez JF, et  al. The clinical impact of incomplete left
2008;246(1):99–107; discussion 108-9. atrial appendage closure with the Watchman Device in patients
120. Singh SM, Douglas PS, Reddy VY. The incidence and long-term with atrial fibrillation: a PROTECT AF (Percutaneous Closure of
clinical outcome of iatrogenic atrial septal defects secondary to the Left Atrial Appendage Versus Warfarin Therapy for Prevention
transseptal catheterization with a 12F transseptal sheath. Circ of Stroke in Patients With Atrial Fibrillation) substudy. J Am Coll
Arrhythm Electrophysiol. 2011;4(2):166–71. Cardiol. 2012;59(10):923–9.
121. Price MJ, et al. Early safety and efficacy of percutaneous left atrial 124. Tzikas A, et al. Left atrial appendage occlusion for stroke prevention
appendage suture ligation: results from the U.S. transcatheter in atrial fibrillation: multicentre experience with the AMPLATZER
LAA ligation consortium. J Am Coll Cardiol. 2014;64(6):565–72. Cardiac Plug. EuroIntervention. 2016;11(10):1170–9.
 Part X
Where We Are: Congenital Heart Disease
 Special Technique Considerations
for Congenital Heart Disease Imaging 44
Anthony M. Hlavacek

While the evaluation of patients with acquired coronary Patient Preparation

artery disease has been the major driving force behind many
of the innovations in cardiac computed tomography (CT), Imaging Environment
this modality has proven to be a very useful imaging modal-
ity for patients with congenital heart disease as well. The Most patients with CHD will be sent for cardiac CT with a
high spatial resolution and isotropic nature of the resulting known cardiac diagnosis, based on prior imaging and/or car-
datasets allow viewing in multiple two-dimensional or curvi- diac interventions. As a result, the CT is generally requested
linear planes, along with the ability for excellent three-­ in order to address a specific clinical question to help guide
dimensional reconstructions. This inherent three-dimensional medical or interventional management. In order to provide
property of cardiac CT makes it useful for imaging tortuous useful diagnostic information, each exam must be tailored
vessels and defining complex anatomical relationships. With for both a specific diagnosis and specific clinical question.
a scan time of only a few seconds or less, it is possible to This requires an in-depth knowledge of the patient’s history,
image most children without sedation. Rapid image acquisi- prior intervention(s), and common hemodynamic sequela of
tion with current CT scanners results in minimal motion arti- the patient’s form of CHD. The acquisition protocol, techni-
facts from breathing and poor cooperation. Unlike cal parameters, and scan range will vary depending on the
echocardiography, CT is window independent, has a large patient characteristics and clinical indication. For instance, a
field of view, and provides excellent visualization of airway 2-year-old patient with a presumed vascular ring might be
structures. In contrast to MRI, CT imaging is less likely to optimally scanned using a non-gated protocol with minimal
require sedation or anesthesia and is minimally affected by or no sedation and aggressive dose reduction techniques,
metallic devices, such as stents, coils, pacemakers, and defi- while a patient requiring optimal imaging of the coronary
brillators. For these reasons, cardiac CT is being increasingly arteries will require a breath hold, slow heart rate, and the
utilized in patients with CHD [1], and recent expert census highest temporal and spatial resolution possible on the scan-
and guideline documents recommend its use for a variety of ner platform. The imager should review the patient history
these patients [2, 3]. Given the varied imaging indications and relevant clinical issues well before the patient arrives for
and complex anatomy and physiology in many of these the scan. He or she should be actively involved in guiding
patients, however, special techniques are often necessary to institutional imaging practices, patient preparation, image
optimize the diagnostic information obtained [4]. This chap- acquisition, and interpretation for each CT scan performed
ter assumes basic competence in cardiac CT imaging, includ- on patients with CHD. Ideally, the imager should be physi-
ing dose reduction strategies, but will provide cally present during scan acquisition. Proficiency in cardiac
recommendations regarding patient preparation, scan tech- CT for patients with CHD requires an expertise in both con-
niques, and reporting that can be utilized to optimize the genital cardiology and cardiac imaging; and the optimal
diagnostic utility of cardiac CT in patients with CHD. imaging environment includes strong collaboration between
radiologists, cardiologists, and cardiac surgeons.

Sedation Requirements
A. M. Hlavacek (*) In patients with congenital heart disease, the risk of sedation
Department of Pediatrics, Division of Pediatric Cardiology,
Medical University of South Carolina, Charleston, SC, USA
must be weighed against the diagnostic information
e-mail: [email protected] ­necessary. Depending on the available scanner platform and

© Humana Press 2019 555

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_44
556 A. M. Hlavacek

clinical indication, many patients can be scanned without ence of a parent in the room or child life services may be
sedation. Centers with newer-generation scanners utilizing helpful. Increasing experience with scanners utilizing volu-
ultrahigh pitch or volumetric scan modes can often achieve metric acquisition or ultrahigh pitch scanning modes should
scan times of less than a second. With these techniques, it is result in a decrease in the need for sedation in children of all
usually possible to adequately visualize anatomy without ages, due to the short acquisition time.
sedation or a breath hold [5, 6]. Patients requiring detailed Scans that require coronary artery or functional analysis
evaluation of the coronary arteries or evaluation of function are more susceptible to motion artifact. Neonates that only
will generally need to be able to cooperate with breathing require identification of the coronary artery origins prior to
instructions, however. The imager must take into account surgical repair can often be adequately imaged without
local resources, clinical indications, and patient age when sedation [6, 7]. However, when detailed coronary artery
determining the need for sedation. imaging (including detailed ostial anatomy) is indicated in
Scans that do not require detailed coronary imaging or a patient under 5–6 years of age, general anesthesia is fre-
functional analysis are less susceptible to motion artifact. quently necessary to cooperate with breath holding
For these scans, infants <6 months of age can generally be (Table  44.1). Older children can often cooperate, particu-
swaddled and imaged without sedation (Fig. 44.1). The use larly if breath holding is practiced with the patient prior to
of a pacifier and an oral dextrose solution can be helpful to the scan. Practicing breath holds is also helpful to assess
keep the infant calm. Patients between 6 months and 3 years any respiratory variations in the heart rate, which is com-
of age often require sedation to lie still in the scanner but can mon in pediatric patients.
usually be imaged free breathing. Most children ≥3 years of
age can cooperate with holding still in the scanner without
sedation, unless they are developmentally delayed. The pres- Intravenous Access

Peripheral IV access is most commonly utilized for contrast

injection with congenital cardiac CT. Many central lines can
be used as well, provided that manufacturer’s recommended
injection guidelines are utilized. Umbilical catheters should
generally be avoided, however, given the frequently subopti-
mal and unpredictable contrast intensity caused by reflux of
contrast into the liver (Fig. 44.2). Special precautions should
be taken to avoid air bubbles in patients with CHD, since
intracardiac shunting in patients with complex congenital
heart disease can result in a systemic arterial embolus. In
general, the largest gauge IV cannula feasible for the patient
body size should be utilized.
The ideal location for contrast injection varies by cardiac
defect and should be determined in advance. While the right
antecubital vein is commonly utilized, an alternative location
might be beneficial in certain diagnoses. Injections via the
right upper extremity are generally preferred over the left
arm, in order to avoid streak artifact in the arch vessels
caused by residual high-density contrast in the left

Table 44.1  Sedation and anesthesia for cardiac scans in children

Detailed coronary/intracardiac
Sedation need anatomy or function indicated Other indications
No sedationa ≥ 5–6 years <6 months
or ≥ 3 yr
Sedation, free n/a 6 months – 3
breathing yearsb
General ≤ 5–6 years n/a
anesthesia
a
Assuming normal developmental status
Fig. 44.1  This is an example of a method using blankets, tape, and a b
Consider scanning without sedation if ultrahigh pitch or volumetric
pacifier to swaddle and scan a 4-month-old infant without sedation scan modes are available
44  Special Technique Considerations for Congenital Heart Disease Imaging 557

Fig. 44.2  Image (a) is a

chest radiograph, in which an
a b
umbilical venous catheter
(arrow) appears well placed
within the atrium. Image (b)
is a coronal reconstruction of
a CT performed 3 h later.
Contrast was injected via the
catheter, but significant reflux
of contrast into the liver
resulted in a nondiagnostic
study

b­ rachiocephalic vein. Conversely, the left upper extremity Acquisition Techniques

would be preferable in a patient with situs inversus and might
also be considered in someone with a persistent left superior Injection Protocol
vena cava (SVC). Injection via a lower extremity vein is
often advantageous in neonates and infants, to avoid residual Adequate opacification of cardiac structures can be attained
high-­density contrast in the SVC, particularly when an via manual or power injection, although the latter has been
abnormality in the vicinity of the SVC is suspected (right shown to provide superior image quality [14]. Although
pulmonary artery/veins, right coronary, ascending aorta) [8]. extravasation is a concern, power injection has been safely uti-
Consideration of IV location can also be beneficial in patients lized in IV sizes as small as 24 gauge [15]. When attempting
with systemic venous abnormalities or occlusions, in order power injection via a small IV, however, a pressure-­limited
to optimize opacification of the structure of interest. For protocol and saline test injection are recommended. When
instance, injection via a lower extremity could definitively injecting via a central line, manufacturer guidelines regarding
exclude a congenital interruption of the inferior vena cava, flow rate and maximum pressure should be reviewed prior to
rather than relying on venous recirculation. On the other use. When in doubt, hand injection should be considered.
hand, the presence of an occlusive thrombus in the inferior A contrast volume of 1.5–2  ml/kg is generally used in
vena cava might be better visualized with an upper extremity pediatric patients, up to standard adult contrast volumes.
injection and acquisition during venous recirculation, due to When utilizing aggressive dose reduction strategies, particu-
the likelihood of incomplete mixing and beam-hardening larly in small children and infants, contrast dilution with
artifact in the IVC resulting from a lower extremity saline (2:1 contrast/saline ratio) can minimize beam-­
injection. hardening artifacts. Contrast dilution also allows lengthening
of the injection time, which creates a higher likelihood of
optimal contrast enhancement of both venous and arterial
Beta Blockade structures in patients with complex anatomy or variable con-
trast transit times. Numerous injection protocols have been
It is well established that heart rate control improves image identified for patients with CHD.  The most often utilized
quality in the evaluation of coronary arteries and valve struc- protocols are as follows:
tures [9, 10]. The higher temporal resolution afforded by
newer-generation scanners allows diagnostic images at • Dual-phase injection—with this protocol, contrast is
higher rates, but the quality remains linked to heart rate [11]. injected at a consistent rate, followed by a saline flush. Due
Beta blockade can be safely used in most pediatric and adult to its simplicity, it is the most frequently utilized protocol.
patients [10, 12], although caution should be utilized in It is ideal for patients with straightforward anatomy and a
patients with potentially unstable hemodynamics. When specific clinical question. Image acquisition is timed to the
newer-generation scanners are utilized, the coronary origins vessel of interest. For example, this protocol would be ben-
can usually be identified in infants without beta blockade eficial in a patient with an isolated coarctation of the aorta,
[13]. Heart rate control should be considered in CHD patients in which only the aortic arch anatomy is required. In
when detailed coronary or valve imaging is needed but is younger patients, utilizing a slow injection rate with this
unnecessary when extracardiac structures are the primary protocol often allows adequate opacification of most car-
area of interest. diovascular anatomy regardless of clinical indication.
558 A. M. Hlavacek

• Triphasic injection—also known as a “biventricular pro- then be timed to initiate once the contrast bolus is complete.
tocol,” this method consists of a two-phase contrast injec- The potential inability to detect IV malfunction prior to scan
tion followed by a saline flush, ensuring simultaneous acquisition is a limitation to this method, however. Regardless
opacification of right and left heart structures. The two-­ of the triggering method, the imager should factor in the time
phase contrast injection can be achieved by giving the required for scan initiation and acquisition time. This is par-
second phase with diluted contrast or via a slower rate ticularly important in infants and small children, who have
(e.g., 100% contrast followed by 30:70% contrast/saline very fast circulation times. With these patients, it is fre-
ratio at constant rate vs. 100% contrast at 4  cc/sec fol- quently necessary to initiate the scan within a few seconds of
lowed by 2 cc/sec). The acquisition is timed for opacifica- completion of the contrast injection. If the acquisition time
tion of the aorta. This method is most useful in adult of the scanner mode is longer than average (>3–4  s), the
patients with repaired CHD, in whom evaluation of both injection time should be lengthened and consideration should
coronary and right-sided cardiac structures is indicated. be given to initiation of the scan before completion of the
• Venous two-phase injection protocol—with this method, contrast bolus.
a portion of the contrast is given (30–50%), followed by a
30–60 s pause and then injection of the remainder of the
contrast via the dual-phase protocol above. This typically Sequence Selection
results in simultaneous opacification of systemic arterial
and venous anatomy in a single acquisition, regardless of The discussion of scan selection below assumes a detailed
IV location. This method is useful in the initial evaluation understanding of available sequences, which is well
of infants with heterotaxy syndromes, in whom complex described in the literature [16]. The first issue to consider
arterial and venous connections are common. when selecting a scan protocol for patients with congenital
heart disease is whether ECG gating should be utilized.
Functional assessment always requires cardiac gating. ECG
Scan Triggering gating is generally necessary to optimally evaluate structures
that are prone to cardiac motion artifact (intracardiac anat-
The complex physiology that is present in patients with CHD omy, coronary arteries, and the aortic root), although newer
often makes timing of the scan acquisition difficult. In adult ultrahigh pitch sequences are often able to accurately image
CHD patients with significant ventricular dysfunction, valve these structures without cardiac gating [6, 17]. Most other
regurgitation, or history of venous occlusion, a pre-scan tim- patients can be scanned with a non-gated sequence, although
ing bolus can be utilized to optimize the acquisition time. In image quality is often inferior to those obtained with ECG-­
most patients with congenital heart disease, however, auto- gated sequences [18]. In addition to clinical indication,
mated bolus tracking is usually sufficient. Using this method, patient age and cooperation level should be considered.
a region of interest (ROI) is selected, which will automati- Aside from prospectively triggered high-pitch sequences,
cally initiate the scan upon reaching a pre-defined Hounsfield ECG-gated sequences are generally more prone to patient
unit level within the relevant cardiac structure. The monitor- movement artifact. Therefore, one might choose a non-gated
ing sequence should be set to initiate at least 2–3  s before sequence to image poorly cooperative patients, rather than
completion of the contrast bolus. In patients with more com- using sedation to perform an ECG-gated scan.
plex anatomy, however, it can be challenging to identify the After deciding to use an ECG-gated sequence, one must
relevant cardiac structure on the pre-monitoring slice. In choose whether to use a retrospectively gated helical
those patients, the ROI can be placed outside of the body. sequence, prospectively triggered axial sequence, or a pro-
Instead of automatic triggering, the scan is manually initi- spectively triggered ultrahigh pitch helical sequence.
ated upon visualization of opacification during the monitor- Retrospectively gated sequences are the most robust in
ing sequence. This method, however, requires practice and a patients with elevated heart rates or arrhythmias but also
strong understanding of cardiac physiology. Each of the result in the highest radiation exposure [19]. Therefore,
methods above requires a monitoring sequence to time scan when other sequences are available, its use should be lim-
acquisition, which increases the radiation exposure. For this ited to those with significant arrhythmias or elevated heart
reason, some advocate image acquisition at a fixed time from rates. Radiation exposure is significantly decreased when a
injection. This method can be useful in infants and small prospectively triggered axial sequence is utilized. This
children, in whom a combination of higher heart rate and scan mode was initially limited to patients with a heart rate
small IV size results in diffuse opacification of cardiovascu- below 70, but it has been shown to generate diagnostic
lar structures throughout the chest by the time the contrast coronary imaging in children with heart rates as high as
bolus is complete. With this method, the contrast bolus 140 using dual-source technology [20, 21]. More recently,
should be lengthened to at least 10–15  s. The scan should prospectively gated volumetric scanning using a 320-slice
44  Special Technique Considerations for Congenital Heart Disease Imaging 559

scanner was shown to generate diagnostic coronary images Systemic Venous Abnormalities
in a group of patients ≤3 years of age (mean HR 111 ± 19),
although diagnostic quality was compromised in patients Venous imaging can be performed with a venous ­two-phase
≤12.5  months or 5.6  kg [22]. Prospectively triggered contrast bolus (described above), a longer contrast injec-
­ultrahigh pitch sequences offer the potential for further tion with late image acquisition, or with a delayed scan
reduction in radiation dose but require a heart
­ (venous phase). The timing of venous recirculation is
rate  <  60–70  in order to consistently provide a detailed dependent on the size of the patient, cardiac output, and
coronary ­evaluation [11, 23]. intracardiac shunting. When needed, a delayed scan in the
venous phase can be performed 30–60  s after the initial
scan. If a patient has poor cardiac output, adequate venous
Lesion-Specific Recommendations opacification may take 120  s or longer. Superior central
venous anatomy will be visualized earlier than the inferior
The general recommendations above should assist cardiac vena cava. A higher contrast load (2.5–3  cc/kg) may be
imaging specialists in obtaining highly diagnostic images on beneficial for optimal venous imaging, particularly in
the majority of patients with CHD. Detailed, lesion-specific those requiring visualization of the inferior vena cava.
recommendations for all congenital heart defects are beyond These methods can also be helpful to evaluate the systemic
the scope of this chapter but are available in a recent expert venous baffles in patients who have undergone a Mustard
consensus document published by the Society of or Senning procedure for transposition.
Cardiovascular CT [4]. Specific technical recommendations
for selected lesions are listed below.
Cyanotic Heart Lesions

Thoracic Arterial Abnormalities Before imaging patients with cyanotic congenital heart
disease, one must consider the physiology present in the
Straightforward pulmonary or aortic abnormalities should be suspected lesion. Many patients with unrepaired cyanotic
scanned with the contrast bolus timed to the area of interest, heart disease have right to left shunting within the heart,
using an injection lasting approximately 12–15 s. If abnor- limiting pulmonary blood flow. In these patients, contrast
malities of both pulmonary and aortic anatomy are sus- injected into the heart will quickly enter the aortic circula-
pected, extending the contrast bolus to ~20 s, with the scan tion, with limited opacification of the pulmonary circula-
timed to the aorta, should allow simultaneous opacification tion. This physiology, for example, will be present in a
of both circulations. One drawback of a lengthened contrast patient with tetralogy of Fallot with significant pulmonary
injection is that dense contrast entering the right atrium and outflow obstruction (prior to repair). Patients with transpo-
ventricle at the time of image acquisition can obscure visual- sition physiology, on the other hand, usually have normal
ization of the right coronary artery, however. An alternative (or elevated) pulmonary blood flow. Rather, the presence
option in this scenario is to use the triphasic injection proto- of cyanosis in these patients is a result of poor mixing
col described above. between the systemic and pulmonary circulations, which
exist in parallel. In either scenario, the aorta will often
opacify before the pulmonary arteries; and opacification of
Pulmonary Venous Anomalies the pulmonary circulation will usually be proportional to
the patient’s oxygen saturation level. The contrast bolus
Optimal opacification of pulmonary venous anomalies is should be lengthened in order to allow time for sufficient
obtained by timing image acquisition to either the left contrast to enter the pulmonary arteries. Timing of pulmo-
atrium or ascending aorta. If the pulmonary venous return nary arterial opacification will also depend on the source
is obstructed, opacification may occur later than expected. of pulmonary blood supply. In neonates/infants with pul-
In this scenario, a longer contrast injection with later monary atresia or severe pulmonary stenosis, the pulmo-
image acquisition may be necessary. The scan range nary flow may originate from the ductus arteriosus or a
should be adjusted to include the suspected area of pul- systemic to pulmonary shunt (e.g., BT shunt), and contrast
monary venous return, which will often require including will reach the pulmonary arteries simultaneously with the
the majority of the thoracic cavity. If infradiaphragmatic aorta. In any patient with cyanotic heart disease, special
TAPVR is suspected, the scan range should include the precautions should be made to avoid air embolism during
upper abdomen. contrast injection.
560 A. M. Hlavacek

Single Ventricle Physiology

Patients with single ventricle physiology generally follow a

three-stage surgical palliation, which is described in Chap.
47. Many patients that follow this pathway require their first
surgical intervention within the first 1–2 weeks of life. The
first surgical procedure varies, dependent on the native anat-
omy, but often includes placement of a systemic to pulmo-
nary shunt (e.g., Blalock-Taussig shunt). When imaging
these patients before or after their first surgical procedure,
one should note that the pulmonary arteries and aorta will
opacify simultaneously.
The second procedure is called a Glenn (or Hemi-Fontan)
procedure, and is generally performed at 4–6 months of age.
With this, the superior vena caval flow is sent directly into
the pulmonary arteries (“bilateral Glenn” if there are bilat-
eral SVCs). In preparation for this procedure, CT is occa-
sionally warranted to evaluate the upper systemic venous
anatomy (right and/or left SVC) and pulmonary arteries.
Therefore, the goal of CT imaging at this stage is simultane- Fig. 44.3  Axial image from a cardiac CT performed on a patient with
ous venous and arterial opacification. This is achieved by a history of a bilateral Glenn procedure. This is an example of how
incomplete mixing of contrast within the pulmonary arteries in these
using a long injection (~15–18 s), with acquisition near the patients can be mistaken for thrombus (arrows)
end of the injection. If there is high likelihood of upper cen-
tral venous obstruction, consider injection in the foot to visu-
alize the superior venous anatomy without streak artifact one can bolus track on the aorta using the appropriate acquisi-
from dense contrast injection. Alternatively, one could per- tion protocol. Evaluation of the Fontan circuit is complex due
form a delayed image acquisition (venous phase) to ensure to incomplete mixing within the circuit, differential streaming
that the systemic veins are all opacified. of into the pulmonary arteries, and variable timing of opacifi-
After the Glenn procedure, an injection in the upper extrem- cation of the superior and inferior venous system. These fac-
ity will directly opacify the pulmonary arteries through the tors can result in nondiagnostic or false-­positive evaluations
anastomosis between the superior vena cava and pulmonary for thrombus formation [25], which is common in this popula-
artery. Incomplete mixing of contrast, aortopulmonary collat- tion. Numerous approaches have been recommended to
eral flow, and bilateral superior vena cavae will result in dif- address this issue. Regardless of strategy, one should consider
ferential branch pulmonary artery opacification. This can result using a larger amount of contrast (2.5–3 ml/kg, up to 150 mL).
in a false-positive finding of pulmonary embolus (Fig. 44.3). Traditionally, simultaneous injection in lower and upper
Some patients develop collateral vessels from the upper sys- extremity veins has been recommended in this patient group
temic venous system to the pulmonary veins or lower systemic [26]. In order to get an adequate flow rate, this method usually
venous system. These are called “veno-venous collaterals,” and requires placement of a femoral venous catheter. While this
their opacification is dependent on IV location. If the goal of usually results in opacification of the entire Fontan circuit,
the study is to evaluate the Glenn anastomosis or pulmonary most patients will still have incomplete mixing of contrast due
arteries, injection in the lower extremity with bolus tracking in to swirling and unopacified hepatic venous inflow. This is par-
the superior vena cava is recommended, allowing opacification ticularly a problem in patients with bilateral superior vena
of the upper systemic veins via venous recirculation [24]. If the cavae. Some authors advocate a single acquisition in the
injection is lengthened, this method will simultaneously venous phase, using a delay of 120–180 s after initiation of
opacify the systemic arterial system and upper venous system contrast injection [27]. This will result in homogenous opaci-
(Fig. 44.4). If the injection is performed via an upper extremity, fication, allowing for identification of clots. However, the
then a scan in the venous phase would be required to ensure image quality will be inadequate for coronary imaging. A
homogeneous opacification of the pulmonary arteries, particu- venous two-phase contrast protocol (as described in previous
larly in the presence of bilateral superior vena cavae. sections but using 50% of the contrast in the first phase) will
The third surgical stage is called a Fontan procedure, which allow opacification of the entire Fontan circuit while providing
involves connection of the inferior vena caval flow into the diagnostic imaging of the aorta and coronary arteries.
pulmonary arteries. If the primary clinical question involves However, since the opacification will not be homogeneous, a
the aortic arch, intracardiac anatomy, or ventricular function, second scan in the venous phase will be necessary to more
44  Special Technique Considerations for Congenital Heart Disease Imaging 561

a b c

Fig. 44.4  In this patient with a history of Norwood and Glenn proce- 3D volume-rendered images viewed from anterior (b) and rightward (c)
dures for hypoplastic left heart syndrome, contrast was injected slowly perspectives). The connection between the superior vena cava and pul-
via the foot, allowing opacification of the upper systemic veins and pul- monary arteries can be seen (white arrow), along with stenosis of the
monary arteries via venous recirculation, along with arterial opacifica- left pulmonary artery (yellow arrow). The reconstructed aorta can be
tion of the aorta in one acquisition (a) oblique coronal reconstruction; visualized in detail as well (Ao)

definitively evaluate for a thrombus in the Fontan circuit. If an • Cardiac position: [The heart is positioned leftward/mid-
evaluation of function is needed, a multiphase acquisition will line/rightward, with apex pointing to the left, right,
allow visualization of contrast streaming and can differentiate inferiorly.]
thrombus versus venous admixture better than a single cardiac • Segments: [Atrial situs, ventricular looping, great arterial
phase acquisition. Dual-energy scanning has also been shown orientation, atrioventricular, and ventriculo-arterial
to be useful in this population when screening for thrombus connections.]
formation [28]. • Systemic veins: [Right superior vena cava, presence/absence
of left superior vena cava. If left SVC, connections/bridging
vein. Inferior vena cava is intact/interrupted.]
Interpretation and Reporting • Pulmonary veins: [Normal/abnormal pulmonary venous
connection with # pulmonary veins seen entering the
Given the complexity within the spectrum of native and left atrium. Anomalous pulmonary venous connections, if
repaired CHD, a collaborative approach between radiology present.]
and pediatric/adult congenital cardiology is recommended • Atria: [Size of right/left atrium. Interatrial septum is intact
for interpretation of these studies. The use of multiplanar vs. atrial septal defect.]
and 3D reconstruction software is also recommended when • AV valves: [Appearance of tricuspid and mitral valves
interpreting CT studies on patients with CHD, particularly +/− motion.]
when attempting to describe extracardiac abnormalities • Ventricles: [Size of right/left ventricle. Septum is intact
[29–31]. Furthermore, 3D reconstructions are often benefi- vs. ventricular septal defect.]
cial in communicating CT findings to cardiologists and sur- • Ventricular outflows: [Description of subaortic and sub-
geons that are less accustomed to visualizing axial CT pulmonic outflows.]
images. • Semilunar valves: [Appearance of pulmonary and aortic
A segmental approach to interpretation and reporting has valves. +/− motion.]
long been recognized as a fundamental concept in cardiac • Pulmonary arteries: [Description main and branch pul-
imaging for CHD [32–34]. A detailed description of segmen- monary arteries/stenosis/embolus.]
tal analysis can be found in Chap. 26 (“CT Spectrum of • Aorta: [Aortic root and ascending aorta. Left/right arch
Congenital Heart Disease”). Briefly, the segmental approach with (normal/abnormal) branching pattern. Evidence of
involves a sequential evaluation of the three segments of the coarctation.]
heart (atria, ventricles, and great arteries) and their connec- • Coronary arteries: [Coronary origins and branching pat-
tions (atrioventricular and ventriculo-arterial connections). tern, stenosis, filling defects.]
This analysis should be utilized and applied in creating a • Ductus arteriosus: [Patent/not patent, appearance.]
structured report for each cardiac CT performed in patients • Collateral arteries: [Aortopulmonary collaterals, veno-­
with CHD. A recommended report template for cardiac find- venous collaterals.]
ings with CHD is as follows: • Pericardium: [Normal, thickened, calcified, effusion.]
562 A. M. Hlavacek

References raphy angiography in the pediatric population. J Cardiovasc

Comput Tomogr. 2012;6(4):252–9.
13. Cui Y, Huang M, Zheng J, Li J, Liu H, Liang C. Assessments of
1. Han BK, Lesser AM, Vezmar M, Rosenthal K, Rutten-Ramos S,
coronary artery visibility and radiation dose in infants with congen-
Lindberg J, Caye D, Lesser JR. Cardiovascular imaging trends in
ital heart disease on cardiac 128-slice CT and on cardiac 64-slice
congenital heart disease: a single center experience. J Cardiovasc
CT. Pediatr Cardiol. 2016;37(1):135–43.
Comput Tomogr. 2013;7(6):361–6.
14. Saake M, Lell MM, Rompel O, Gloeckler M, May M, Eller A,
2. Cohen MS, Eidem BW, Cetta F, Fogel MA, Frommelt PC, Ganame
Achenbach S, Uder M, Wuest W. Contrast medium application in
J, Han BK, Kimball TR, Johnson RK, Mertens L, Paridon SM,
pediatric high-pitch cardiovascular CT angiography: manual or
Powell AJ, Lopez L. Multimodality imaging guidelines of patients
power injection? J Cardiovasc Comput Tomogr. 2014;8(4):315–22.
with transposition of the great arteries: a report from the American
15. Amaral JG, Traubici J, BenDavid G, Reintamm G, Daneman

Society of Echocardiography developed in collaboration with the
A.  Safety of power injector use in children as measured by inci-
Society for Cardiovascular Magnetic Resonance and the Society
dence of extravasation. AJR Am J Roentgenol. 2006;187(2):580–3.
of Cardiovascular Computed Tomography. J Am Soc Echocardiogr.
16. Halliburton S, Arbab-Zadeh A, Dey D, Einstein AJ, Gentry R,
2016;29(7):571–621.
George RT, Gerber T, Mahesh M, Weigold WG. State-of-the-art in
3. Han BK, Rigsby CK, Hlavacek A, Leipsic J, Nicol ED, Siegel MJ,
CT hardware and scan modes for cardiovascular CT. J Cardiovasc
Bardo D, Abbara S, Ghoshhajra B, Lesser JR, Raman S, Crean
Comput Tomogr. 2012;6(3):154–63.
AM.  Computed tomography imaging in patients with congeni-
17. Beeres M, Schell B, Mastragelopoulos A, Herrmann E, Kerl JM,
tal heart disease part I: rationale and utility. An expert consensus
Gruber-Rouh T, Lee C, Siebenhandl P, Bodelle B, Zangos S, Vogl
document of the Society of Cardiovascular Computed Tomography
TJ, Jacobi V, Bauer RW. High-pitch dual-source CT angiography of
(SCCT): endorsed by the Society of Pediatric Radiology (SPR)
the whole aorta without ECG synchronisation: initial experience.
and the North American Society of Cardiac Imaging (NASCI). J
Eur Radiol. 2012;22(1):129–37.
Cardiovasc Comput Tomogr. 2015;9(6):475–92.
18. Shuman WP, Leipsic JA, Busey JM, Green DE, Pipavath SN, Hague
4. Han BK, Rigsby CK, Leipsic J, Bardo D, Abbara S, Ghoshhajra B,
CJ, Koprowicz KM. Prospectively ECG gated CT pulmonary angi-
Lesser JR, Raman SV, Crean AM, Nicol ED, Siegel MJ, Hlavacek
ography versus helical ungated CT pulmonary angiography: impact
A. Computed tomography imaging in patients with congenital heart
on cardiac related motion artifacts and patient radiation dose. Eur J
disease, part 2: technical recommendations. An expert consensus
Radiol. 2012;81(9):2444–9.
document of the Society of Cardiovascular Computed Tomography
19. Jin KN, Park EA, Shin CI, Lee W, Chung JW, Park JH. Retrospective
(SCCT): endorsed by the Society of Pediatric Radiology (SPR)
versus prospective ECG-gated dual-source CT in pediatric patients
and the North American Society of Cardiac Imaging (NASCI). J
with congenital heart diseases: comparison of image quality and
Cardiovasc Comput Tomogr. 2015;9(6):493–513.
radiation dose. Int J Cardiovasc Imaging. 2010;26(Suppl 1):63–73.
5. Han BK, Overman DM, Grant K, Rosenthal K, Rutten-Ramos
20. Duan Y, Wang X, Cheng Z, Wu D, Wu L. Application of prospective
S, Cook D, Lesser JR.  Non-sedated, free breathing cardiac CT
ECG-triggered dual-source CT coronary angiography for infants
for evaluation of complex congenital heart disease in neonates. J
and children with coronary artery aneurysms due to Kawasaki dis-
Cardiovasc Comput Tomogr. 2013;7(6):354–60.
ease. Br J Radiol. 2012;85(1020):e1190–7.
6. Bridoux A, Hutt A, Faivre JB, Flohr T, Duhamel A, Pagniez J, Remy
21. Paul JF, Rohnean A, Elfassy E, Sigal-Cinqualbre A. Radiation dose
J, Remy-Jardin M.  Coronary artery visibility in free-­ breathing
for thoracic and coronary step-and-shoot CT using a 128-slice dual-­
young children on non-gated chest CT: impact of temporal resolu-
source machine in infants and small children with congenital heart
tion. Pediatr Radiol. 2015;45(12):1761–70.
disease. Pediatr Radiol. 2011;41(2):244–9.
7. Vastel-Amzallag C, Le Bret E, Paul JF, Lambert V, Rohnean A, El
22. Tada A, Sato S, Kanie Y, Tanaka T, Inai R, Akagi N, Morimitsu
Fassy E, Sigal-Cinqualbre A.  Diagnostic accuracy of dual-source
Y, Kanazawa S.  Image quality of coronary computed tomogra-
multislice computed tomographic analysis for the preoperative
phy angiography with 320-row area detector computed tomog-
detection of coronary artery anomalies in 100 patients with tetral-
raphy in children with congenital heart disease. Pediatr Cardiol.
ogy of Fallot. J Thorac Cardiovasc Surg. 2011;142(1):120–6.
2016;37(3):497–503.
8. Yang M, Mo XM, Jin JY, Zhang J, Liu B, Wu M, Teng GJ. Image
23. Alkadhi H, Stolzmann P, Desbiolles L, Baumueller S, Goetti R,
quality and radiation exposure in pediatric cardiovascular CT
Plass A, Scheffel H, Feuchtner G, Falk V, Marincek B, Leschka
angiography from different injection sites. AJR Am J Roentgenol.
S.  Low-dose, 128-slice, dual-source CT coronary angiography:
2011;196(2):W117–22.
accuracy and radiation dose of the high-pitch and the step-and-­
9. Dewey M, Vavere AL, Arbab-Zadeh A, Miller JM, Sara L, Cox
shoot mode. Heart. 2010;96(12):933–8.
C, Gottlieb I, Yoshioka K, Paul N, Hoe J, de Roos A, Lardo AC,
24. Ghadimi Mahani M, Agarwal PP, Rigsby CK, Lu JC, Fazeli

Lima JA, Clouse ME. Patient characteristics as predictors of image
Dehkordy S, Wright RA, Dorfman AL, Krishnamurthy R. CT for
quality and diagnostic accuracy of MDCT compared with conven-
assessment of thrombosis and pulmonary embolism in multiple
tional coronary angiography for detecting coronary artery stenoses:
stages of single-ventricle palliation: challenges and suggested pro-
CORE-64 Multicenter International Trial. AJR Am J Roentgenol.
tocols. Radiographics. 2016;36(5):1273–84.
2010;194(1):93–102.
25. Prabhu SP, Mahmood S, Sena L, Lee EY. MDCT evaluation of pul-
10. Mahabadi AA, Achenbach S, Burgstahler C, Dill T, Fischbach R,
monary embolism in children and young adults following a lateral
Knez A, Moshage W, Richartz BM, Ropers D, Schroder S, Silber S,
tunnel Fontan procedure: optimizing contrast-enhancement tech-
Mohlenkamp S. Safety, efficacy, and indications of beta-adrenergic
niques. Pediatr Radiol. 2009;39(9):938–44.
receptor blockade to reduce heart rate prior to coronary CT angiog-
26. Sandler KL, Markham LW, Mah ML, Byrum EP, Williams

raphy. Radiology. 2010;257(3):614–23.
JR.  Optimizing CT angiography in patients with Fontan physiol-
11. Gordic S, Husarik DB, Desbiolles L, Leschka S, Frauenfelder T,
ogy: single-center experience of dual-site power injection. Clin
Alkadhi H. High-pitch coronary CT angiography with third genera-
Radiol. 2014;69(12):e562–7.
tion dual-source CT: limits of heart rate. Int J Cardiovasc Imaging.
27. Park EA, Lee W, Chung SY, Yin YH, Chung JW, Park JH. Optimal
2014;30(6):1173–9.
scan timing and intravenous route for contrast-enhanced computed
12. Han BK, Lindberg J, Overman D, Schwartz RS, Grant K, Lesser
tomography in patients after Fontan operation. J Comput Assist
JR. Safety and accuracy of dual-source coronary computed tomog-
Tomogr. 2010;34(1):75–81.
44  Special Technique Considerations for Congenital Heart Disease Imaging 563

28. Grewal J, Al Hussein M, Feldstein J, Kiess M, Ellis J, Human D, anomalies in children: comparison of axial, multiplanar, and 3D
Leipsic J. Evaluation of silent thrombus after the Fontan operation. images. AJR Am J Roentgenol. 2011;196(5):1040–6.
Congenit Heart Dis. 2013;8(1):40–7. 32. Brandt PW, Calder AL.  Cardiac connections: the segmental

29. Di Sessa TG, Di Sessa P, Gregory B, Vranicar M.  The use of approach to radiologic diagnosis in congenital heart disease. Curr
3D contrast-enhanced CT reconstructions to project images of Probl Diagn Radiol. 1977;7(3):1–35.
vascular rings and coarctation of the aorta. Echocardiography. 33. Lapierre C, Dery J, Guerin R, Viremouneix L, Dubois J, Garel
2009;26(1):76–81. L.  Segmental approach to imaging of congenital heart disease.
30. Lee EY, Siegel MJ, Hildebolt CF, Gutierrez FR, Bhalla S, Fallah Radiographics. 2010;30(2):397–411.
JH.  MDCT evaluation of thoracic aortic anomalies in pediatric 34. Shinebourne EA, Macartney FJ, Anderson RH. Sequential chamber
patients and young adults: comparison of axial, multiplanar, and localization--logical approach to diagnosis in congenital heart dis-
3D images. AJR Am J Roentgenol. 2004;182(3):777–84. ease. Br Heart J. 1976;38(4):327–40.
31. Lee EY, Tracy DA, Mahmood SA, Weldon CB, Zurakowski D,
Boiselle PM.  Preoperative MDCT evaluation of congenital lung
 CT of Coronary Artery Anomalies
45
Long Jiang Zhang, Shahryar M. Chowdhury,
and Guang Ming Lu

Coronary artery anomalies are defined as any coronary arte- potentially involved vascular structures. For example, in
rial pattern with a feature (e.g., origination, course, and ter- patients with the coronary to pulmonary artery fistula or
mination) rarely encountered in the general population. other extracardiac termination, longer scanning range should
Although coronary anomalies are rare, they are important to be included. (3) Including systolic and diastolic cardiac
detect because they have been implicated in chest pain, syn- phases into the acquisition, and the resultant reformatted CT
cope, myocardial infarctions, and even sudden death, espe- images, can be very important to detect dynamic compres-
cially in young adults [1–4]. Electrocardiogram-gated sion of abnormal coronary segments in coronaries that dis-
multi-detector row computed tomography (MDCT) of the play myocardial bridging or abnormal origin of coronary
heart allows the accurate and noninvasive depiction of con- artery from the opposite coronary sinus with an interarterial
genital coronary anomalies in patients of all ages [5, 6]. course. (4) Post-processing techniques such as multiplanar
MDCT has been shown to be superior to conventional coro- reformation (MPR) including curved planar reformation
nary angiography in showing the origin and proximal path of (CPR), maximal intensity projection (MIP), and volume ren-
anomalous coronary arteries [7]. This chapter briefly dering (VR) should be comprehensively used to show the
describes MDCT-reformatted techniques, normal coronary anatomical relations between large vessels and heart
artery anatomy, MDCT coronary angiography findings in chambers.
congenital coronary anomalies, and the clinical implications
of these anomalies.
Normal Anatomic Features

 pecial Considerations for Assessing

S Normally, myocardial blood flow is supplied by two main
Coronary Artery Anomalies coronary arteries, the right coronary artery (RCA) and the
left coronary artery (LCA) (Fig. 45.1) [8, 9].
Important considerations should be taken when preparing
and performing MDCT angiography for the assessment of
coronary artery anomalies, which include the following: (1) Normal Coronary Artery Branches and Courses
Non-contrasted coronary artery calcium scans can suggest
abnormal coronary origins in adults. However, delineating The RCA originates from the right coronary sinus of
important details about the course, such as the presence of an Valsalva, descends in the anterior right atrioventricular
intramural segment, is very difficult using non-contrasted groove, and usually gives off the conus branch (50–60%) as
scans alone. Thus, MDCT coronary angiography should be its first branch and then two or three large right ventricular
performed in these patients even if an abnormality is found wall branches. The acute marginal branch is the first large
on coronary artery calcium scoring scans. (2) CT scanning branch, which occasionally continues to the apex.
range should be properly adjusted to appropriately cover the The LCA originates from the left coronary sinus of
Valsalva, which bifurcates into the left circumflex (LCX) and
left anterior descending (LAD) arteries or trifurcates with an
L. J. Zhang (*) · G. M. Lu
Department of Medical Imaging, Jinling Hospital, Medical School additional intermediate ramus branch. The LAD usually
of Nanjing University, Nanjing, Jiangsu, China descends in the anterior interventricular groove, giving off
S. M. Chowdhury septal branches into the interventricular septum and diagonal
Division of Pediatric Cardiology, Department of Pediatrics, branches descending toward the lateral margin of the left
Medical University of South Carolina, Charleston, SC, USA

© Humana Press 2019 565

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_45
566 L. J. Zhang et al.

a b

c d

Fig. 45.1  Normal anatomy of the coronary arteries by MDCT coro- dominant system, the RCA gives rise to the PDA.  The PDA appears
nary angiography. Volume-rendered CT angiogram with 3D reconstruc- somewhat small, likely associated with the LAD continuing around the
tion highlighting normal coronary anatomy. (a) Normal course of the apex of the heart to supply some of the inferior septum – also known as
RCA in the right interventricular groove shows an early CA branch. “wraparound” LAD, a normal variant. RCA  =  right coronary artery;
The second branch is the SNA.  Afterwards AM branches supply the CA = conus artery; AM = acute marginal artery; SNA = sinoatrial nodal
right ventricle. (b) The LAD runs in the interventricular sulcus with artery; LCA  =  left coronary artery; LAD  =  left descending artery;
diagonal arteries supplying the lateral wall of the left ventricle. (c) The D  =  diagonal; OM  =  obtuse marginal artery; LCX  =  left circumflex
LCA splits into the LAD and LCX. The LCX runs in the left interven- artery; PDA = posterior descending artery; LPVB = left ventricular pos-
tricular groove and gives off a number of OM branches. (d) In this right terior branch

ventricular wall. The LCX enters the left atrioventricular this territory (codominant supply). In 10% of patients (left
groove to supply obtuse marginal branches to the lateral and coronary dominant patients), the LCX supplies both the PDA
posterolateral walls of the left ventricle. and the posterior left ventricular branch [10]. In these
patients, hypoplasia of the RCA is often found.

Coronary Dominance
Sinoatrial Nodal Supply
In 70% of patients, the RCA gives off the posterior descend-
ing artery (PDA) and posterior left ventricular branches Anatomic variations in the origin and course of the sinoatrial
(right dominant supply). In the subjects with right coronary nodal artery (SNA) are important to recognize as it is occa-
dominance, dysplasia or absence of the LCX can be often sionally misdiagnosed as a coronary artery fistula [11].
observed. In these cases, if the RCA supplies the typical Sinoatrial nodal artery can be grouped into three types
LCX course, the RCA is referred to as “superdominant” according to the origin, course, and termination [12]. (1)
(Fig. 45.1). In approximately 20% of patients, the PDA origi- Right SNA (49.1%) originates from the RCA or right coro-
nates from the RCA, but branches from the LCX also supply nary sinus and courses between the right atrium and the root
45  CT of Coronary Artery Anomalies 567

of aorta toward the ostium of the superior vena cava (SVC). Coronary Artery Anomalies of Origin
(2) Left SNA (42.5%) originates from the proximal segment
of the LCX and courses along the anterior wall of the left High Takeoff
atrium, terminating at the ostium of the SVC. (3) Posterior
SNA (8.4%) originates from the posterolateral segment of High takeoff of a coronary artery is defined as origin of a
the LCX, LCA, or left ventricular posterior branch of the coronary artery more than 1 cm above the sinotubular junc-
RCA.  Entrance into the SVC can be classified as precaval tion [10]. High takeoff of the RCA (Fig. 45.2a–c) is reported
(the SNA approaches the sinoatrial node anterior to the much more commonly than high takeoff of the LCA
SVC), retrocaval (the SNA approaches the sinoatrial node (Fig. 45.2d). This variant has no hemodynamic significance,
posterior to the SVC), and pericaval (through multiple but it may make selective invasive angiography more diffi-
branches surrounding the SVC). cult [10, 15]. In addition, cardiac surgeons should be aware
of this anomaly to avoid accidental cross-clamping or tran-
section during surgery. In rare instances, high takeoff of a
 lassification, Prevalence, and Clinical
C coronary artery may be associated with a malignant interar-
Relevance of Coronary Anomalies terial or intramural course – similar to those associated with
origin of the coronary artery from the opposite sinus as
Coronary artery anomalies are found in 0.2–5.8% of indi- described below (Fig. 45.2b–c).
viduals [2]. The majority of coronary artery anomalies are
benign (anomalies that are usually not clinically signifi-
cant). However, while having a low prevalence, malignant Multiple Ostia
coronary artery anomalies can cause life-threatening
sequelae such as cardiorespiratory arrest and sudden death. Multiple ostia have no significant clinical relevance. The
Various classification systems exist for coronary artery LAD and LCX may arise separately with no LCA (Fig. 45.3a).
anomalies. No classification system is universally accepted. This occurs in a small percentage (0.41%) of individuals,
In this chapter, coronary anomalies are divided into the fol- and invasive coronary angiography is technically difficult
lowing three broad categories: anomalies of origin, anoma- [14, 15, 17]. The RCA and the conus branch may arise sepa-
lies of course, and anomalies of termination according to rately from the right sinus (Fig.  45.3b). An aberrant conus
Kim et  al.’s simplified classification of coronary artery artery arising separately from the RCA is particularly at risk
anomalies (Table 45.1) [13–15]. Regardless of the classifi- for injury from ventriculostomy or other heart surgery.
cation system used, it is most important to delineate lesions
that are benign versus those that are malignant especially
when evaluating patients with concerning clinical symp-  rigin of Coronary Artery from the Posterior
O
toms such as angina, syncope, myocardial infarct, or sud- Sinus of Valsalva
den cardiac death [3, 14, 16].
An origin of the RCA or the LCA from the posterior sinus of
Valsalva is rare. This anomaly has no clinical relevance
Table 45.1  Classification of coronary artery anomalies
(Fig. 45.4) unless accompanied by an interarterial or intra-
Anomalies of origin mural course.
 High takeoff
 Multiple ostia
 Single coronary artery
Single Coronary Artery
 Anomalous origin of coronary artery from the pulmonary arterya
 Origin of coronary artery or branch from opposite or ­
non-­coronary sinus and an anomalous (retroaortic, interarteriala, Single coronary artery is a rare congenital anomaly in which
prepulmonic, septal [subpulmonic]) course only one coronary artery arises from the aorta and supplies
Anomalies of course blood to the entire heart [18]. It is seen in only 0.002–0.04%
 Myocardial bridginga of the population [2, 18, 19]. A single coronary artery may
 Duplication of arteries originate from either the right or left aortic sinus. In most
Anomalies of termination
cases, this single artery will give off two branches each with
 Coronary artery fistulaa
 Coronary arcade
distributions of the typical RCA and LCA. Rarely, different
 Extracardiac termination distributions from that of the normal coronary artery tree
From Greenberg et al. [13], with permission
may be observed, such as LCX from the RCA. In most cases
Hemodynamically significant anomalies, which may contribute to
a this is a benign variant. However, patients with single
myocardial perfusion abnormalities
568 L. J. Zhang et al.

a b

c d

Fig. 45.2  High takeoff of the coronary arteries. (a) 3D volume refor- course, these arteries may display an acute angle of takeoff and nar-
matted CT angiogram showing the RCA arising from the ascending rowed proximal course (such as in B) suggesting an intramural segment
aorta above the sinotubular junction superior to the right sinus. (b) 3D which is also a malignant finding. (d) A coronary artery tree image that
volume reformatted CT angiogram showing the RCA arising from the shows the LCA arising from the ascending aorta above the sinotubular
ascending aorta above the sinotubular junction superior to the left sinus junction with a 90° angle from the aorta and widely patent proximal
in close proximity to the left/right commissure. (c) Curved planar refor- course  – there is no evidence of an intermural course, and this is a
matted image showing the RCA arising from the ascending aorta cours- benign finding. AO = aorta; LAD = left descending artery; LCA = left
ing between the ascending aorta and pulmonary artery trunk, conferring coronary artery; LCX = left circumflex artery; RCA = right coronary
increased risk for sudden cardiac death. Even if there is no interarterial artery

Fig. 45.3 Multiple ostia of

the coronary arteries. (a) 3D
a b
volume reformatted CT
angiogram shows the LAD
and LCX originate separately
from the left coronary sinus
of Valsalva. (b) Coronary
artery tree image shows the
right coronary artery and
conus artery separately
originate from the right
coronary sinus of Valsalva.
CA = conus artery; LAD = left
descending artery;
RCA = right coronary artery;
LCX = left circumflex artery
45  CT of Coronary Artery Anomalies 569

a b

Fig. 45.4  Anomalous left coronary artery arising from posterior coro- otherwise unremarkable course. In the absence of an intramural or
nary sinus of Valsalva. (a) Axial contrast-enhanced CT angiography interarterial course, this is a benign lesion. AO  =  aorta; LAD  =  left
shows the left coronary artery arising from the non-coronary sinus of descending artery; LCA = left coronary artery; LCX = left circumflex
Valsalva. (b) 3D volume rendering reformatted CT image shows the left artery; RCA = right coronary artery
coronary artery arising from the non-coronary sinus of Valsalva with an

c­ oronary artery are at increased risk for sudden death if a coronary artery (Fig.  45.6). In addition, the LCX or LAD
major coronary branch crosses between the pulmonary artery may arise from the right coronary sinus of Valsalva or oppo-
and the aorta. site coronary artery [15, 16]. An anomalous LCX most com-
monly arises from a separate ostium within the right sinus of
Valsalva or as a proximal branch of the RCA (approximately
Coronary Origin Stenosis and Atresia 0.32–0.67% of the population) [2, 14, 21]. Anomalous LCX
arteries most often take a retroaortic course, but more rarely
Congenital coronary ostial atresia or hypoplasia is very it may take either an interarterial or a prepulmonic course. It
rarely reported. It may be isolated or be associated with other may be an isolated anomaly with the LAD originating nor-
congenital heart disease such as truncus arteriosus. Coronary mally from the left coronary sinus or may be associated with
ostial atresia or hypoplasia is usually associated with myo- other branch anomalies, such as origin of the LAD from the
cardial ischemia; the degree of severity is dependent upon anomalous LCX. This anomaly has not been associated with
the development of collaterals from the patent coronary sys- sudden death. In all of these anomalies, the coronary ostium
tem. The coronary ostium may also be occluded by a valve-­ may be at the normal level, or the involved artery may have a
like ridge [20]. This entity is often associated with congenital high or low takeoff [20, 21].
supra-aortic stenosis in patients with Williams syndrome and A coronary artery arising from the contralateral sinus of
predisposes these patients to sudden cardiac death. Valsalva or opposite coronary artery can take any of four
common courses, depending on the anatomic relationship of
the anomalous vessel to the aorta and the pulmonary trunk:
Origin of the Coronary Artery (a) interarterial (i.e., between the aorta and the pulmonary
from the Opposite Sinus of Valsalva artery) (Figs. 45.5 and 45.6), (b) retroaortic, (c) prepulmonic,
or Opposite Coronary Artery or (d) septal (subpulmonic) [2, 3, 10, 14, 15, 22]. It is of great
clinical importance which course is taken. Retroaortic, pre-
Origin of the coronary artery from the contralateral sinus of pulmonic, and septal (subpulmonic) courses are generally
Valsalva or contralateral coronary artery has multiple sub- benign, while an interarterial course is malignant carrying a
types. Most commonly, the RCA may arise from the left risk for sudden cardiac death. The mechanism behind sudden
coronary sinus of Valsalva or contralateral coronary artery in cardiac death in patients with an interarterial course is
0.03–0.17% of patients (Fig. 45.5) [3, 10, 15]. The LCA can ­controversial. Some have argued that compression of the
arise from the right coronary sinus of Valsalva or opposite coronary artery between the aorta and pulmonary artery
570 L. J. Zhang et al.

Fig. 45.5  Right coronary

artery arising from the left
a b
coronary sinus of Valsalva
(interarterial type). (a) Thin-
slab maximum intensity
projection CT image shows
the right coronary artery
arising from the left coronary
sinus of Valsalva and coursing
between the aortic root and
pulmonary artery. (b) 3D
volume-­rendered reformatted
CT reconstruction shows the
separate origins of the LCA
and RCA. (c) Multiplanar
reformatted CT image shows
the slit-like, eccentric shape
of the proximal RCA. (d) 3D
volume-rendered reformatted
CT reconstruction can be used c d
to visualize the eccentric
RCA origin (arrow). In all
images, note the acute angle
of takeoff from the sinus,
proximity to the left/right
commissure, and slit-like
orifice and proximal course,
suggesting an intramural
course. AO = aorta;
PA = pulmonary artery;
LAD = left descending artery;
LCX = left circumflex artery;
RCA = right coronary artery

a b

Fig. 45.6  Anomalous left coronary artery arising from right coronary missure, and slit-like orifice and proximal course, suggesting an intra-
sinus of Valsalva. (a) Axial contrast-enhanced CT angiography shows mural course. The intramural course ends when the LCA returns to
the left coronary artery arising from the right coronary sinus of Valsalva normal size (arrow). (b) 3D volume rendering reformatted CT image
and coursing between the aortic root and pulmonary trunk. Note the shows the left coronary artery arising from the right coronary sinus of
acute angle of takeoff from the sinus, proximity to the left/right com- Valsalva near the intercoronary commissure (*)
45  CT of Coronary Artery Anomalies 571

d­ uring exercise contributes to ischemia and sudden death, MDCT angiography can be important in guiding surgical
while others have rejected that hypothesis [23–25]. “unroofing” procedure [33]. In addition, intramural coronar-
Patients with an anomalous coronary origin are at risk for ies are important to identify in patients who require coronary
having an abnormal intramural proximal course – that is, the origin transfer procedures, such as the arterial switch proce-
proximal coronary artery runs within the wall of the aorta. dure in patients with transposition of the great arteries or in
An intramural course is associated with increased risk for patients undergoing the Ross procedure.
sudden death – higher in anomalous LCA versus RCA [2, 3,
21, 22, 26–32]. The vast majority of anomalous coronary
arteries with an intramural course also display an interarte-  rigin of the Coronary Artery
O
rial course, though not all anomalous coronaries with an from the Pulmonary Trunk
interarterial course have an intramural segment. A number of
MDCT angiographic features are noted in patients with an Anomalous origin of the coronary artery from the pulmonary
intramural course: (1) The coronary origin and proximal artery is one of the most clinically significant coronary artery
course are narrow or slit-like, (2) the coronary originates anomalies. Anomalous origin of the LCA from the pulmo-
eccentrically from the aortic sinus near the commissure, and nary artery (ALCAPA), also called Bland-White-Garland
(3) there is an acute angle of takeoff from the aortic sinus. syndrome, is the most common type constituting 0.25–0.5%
One can determine where the intramural segment ends at the of all congenital heart disease (Fig.  45.7a, b) [15, 34–36].
point when the distal coronary artery returns to a circular Other types of this coronary artery anomaly include origin of
shape. Identifying the length of the intramural course by the RCA from the pulmonary trunk (Fig. 45.7c, d), origin of

Fig. 45.7 Anomalous
coronary artery arising from
a b
the pulmonary artery. (a)
Axial contrast-enhanced CT
image shows the LCA arising
from the left-facing sinus of
the pulmonary artery root. (b)
3D volume reformatted CT
image shows the LCA arising
from the pulmonary artery
root. Note the coronary
arteries appear dilated
consistent with collateral
development from the RCA
system to the LCA. (c)
Multiplanar reformatted CT
image shows the RCA arising
from the right-facing sinus of
the pulmonary artery root. (d)
3D volume reformatted CT c d
image shows the RCA arising
from the pulmonary artery
root. AO = aorta; LAD = left
descending artery; LCX = left
circumflex artery;
PA = pulmonary artery;
RCA = right coronary artery
572 L. J. Zhang et al.

LAD from pulmonary artery, and origin of LCX from pul- Coronary Artery Anomalies of Course
monary artery. There are rare reports of coronaries also origi-
nating from the branch pulmonary arteries. Myocardial Bridging
ALCAPA can be classified into two types: infant and
adult types. Clinical presentation is influenced by the devel- Coronary arteries generally course epicardially.
opment of collaterals from the RCA to the LCA system. In Myocardial bridging is a congenital anomaly character-
infants with poor collateralization, symptoms manifest in the ized by muscular encasement of a coronary artery seg-
first few months of life with symptoms due to myocardial ment that has dipped into the myocardium [15, 36, 37]. It
ischemia – extreme fussiness with feeds, poor feeding, and is reported the prevalence of myocardial bridging ranges
poor growth. Without surgical repair, rapid death ensues in from 15% to 85% in autopsy studies; however, it is only
up to 90% of patients within weeks or months of birth [34]. seen in 0.5–2.5% of invasive angiographic studies [36].
In patients with greater collateralization, older infants and MDCT coronary angiography displays better sensitivity
toddlers may present with symptoms of left-to-right shunting than invasive angiography with a reported incidence of
from the collateral retrograde through the LCA into the pul- 3.5–30% [36–38]. Myocardial bridging is most commonly
monary artery. These patients display poor growth, tachy- localized in the middle segment of the LAD; but it can be
pnea, and left heart enlargement. These patients often have found in other coronary arteries. Multiple myocardial
poor ventricular function and atrioventricular valve regurgi- bridges can be observed in the same or different coronary
tation. In adult type with significant collateralization, arteries and their branches. Additionally, segments of the
ALCAPA syndrome may be asymptomatic. Alternatively, coronary artery proximal to the myocardial bridge are
adults with ALCAPA may develop myocardial infarction, more vulnerable to atherosclerotic plaques due to abnor-
left ventricular dysfunction, mitral valve regurgitation, or mal blood flow profiles, increasing the risk of myocardial
silent myocardial ischemia, leading to sudden cardiac death. ischemia [36, 37].
MDCT coronary angiography can directly visualize the Myocardial bridges can be divided into two types: super-
LCA arising from the main pulmonary artery, which is the ficial bridges (75%) (Fig.  45.8) and deep bridges (25%)
diagnostic hallmark of ALCAPA syndrome (Fig. 45.7a, b). (Fig. 45.9) although no uniform depth criteria have been set
Most often, the LCA originates from the left-facing pulmo- to classify myocardial bridges [38]. Deep myocardial bridge
nary sinus. Less often, the LCA originates from the often denotes the completely encased coronary artery, which
­non-­facing pulmonary sinus. Other indirect signs, such as often results in compression of involved coronary segment
the dilated and tortuous RCA and LCA, dilated intercoro- during the systolic phase. Superficial myocardial bridges are
nary collateral vessels, left ventricular hypertrophy and encased incompletely by myocardium; however, long super-
dilatation, and dilated bronchial arteries, can also be found ficial MB can result in severe stenosis during the cardiac
in this entity [35]. cycle (Fig. 45.7).
Surgical correction is the standard treatment for the Myocardial bridging is asymptomatic in the vast majority
patients with ALCAPA syndrome. The aim of surgery is to of patients [8]. However, there is evidence that myocardial
restore a two-coronary-artery circulation system. In neonates bridging can be responsible for angina pectoris, myocardial
with poor collateralization, early correction is required. In infarction, life-threatening arrhythmias, or even death in cer-
asymptomatic adults with adequate collateralization, surgery tain patients, especially those with deep myocardial bridge
should be considered when extensive delayed subendocardial or long superficial bridging [39, 40].
enhancement caused by myocardial infarction is seen on The MDCT coronary angiogram can clearly show the
magnetic resonance imaging [35]. The preferred surgery presence, course, and anatomical features of intramuscular
includes a coronary button transfer, the Takeuchi procedure, coronary arteries [36, 41]. When myocardial bridging is
as it is the most anatomic correction and has excellent long-­ suspected, systolic and diastolic phase reformation is rec-
term results. In adult patients with adequate collateralization, ommended to assess the luminal narrowing during the sys-
alternative approaches include placement of a coronary artery tolic phase. Other indirect findings include a “milking”
bypass graft (CABG) or ligation of the origin of the LCA to effect which can be found during the systolic and diastolic
stop competitive flow. A combined approach in adults may be phase reformatted CT images or 4D cine and a “step-
preferred. After surgery, MDCT angiography is often used to down-step-­up” phenomenon induced by systolic compres-
ensure there is no kinking of the LCA after button transfer sion of the tunneled segment in sagittal MPR or VR
and to assess the patency of bypass grafts if symptoms recur. reformatted images.
45  CT of Coronary Artery Anomalies 573

a b c

Fig. 45.8  Superficial myocardial bridging of the left anterior descend- matted CT image and (c) coronary artery tree image show moderate
ing artery. (a) Curved planar reformatted image shows the superficial stenosis of intramyocardial segment of the LAD artery (arrow).
intramural course (arrow) of the mid-segment of the left anterior LAD = left descending artery; LCX = left circumflex artery; RCA = right
descending artery. Note the LAD is not entirely encased in myocardium coronary artery
consistent with a superficial myocardial bridge. (b) 3D volume refor-

a b

RCA

LAD

Fig. 45.9  Deep myocardial bridging of the left anterior descending determining the superficial and deep myocardial bridging. (b) 3D vol-
artery. (a) Curved planar reformatted image shows deep intramural ume reformatted CT image shows the covering myocardium in the mid-
course with 4 mm covering myocardium (arrow) of the mid-segment of segment of the left anterior descending artery (arrow). RCA = right
the left anterior descending artery. Note 2 mm is the cutoff value for coronary artery; LAD = the left anterior descending artery
574 L. J. Zhang et al.

Fig. 45.10  Coronary artery

duplication. (a) 3D volume
a b
reformatted CT image
showing a duplicated right
coronary artery. (b) 3D
volume reformatted CT
angiogram showing a short
LAD (LAD1) coursing and
terminating in the anterior
interventricular sulcus
without reaching the apex and
a long LAD (LAD2) reaching
the apex. CA = conus artery;
LAD = left descending artery;
LCX = left circumflex artery;
RCA = right coronary artery

Coronary Artery Duplication phy [43–45]. Although most patients have no symptoms,
some patients may have exertional dyspnea and fatigue.
Coronary artery duplication can occur in both the RCA and Symptom severity depends on the degree of shunt and left-
LCA. Duplication of the RCA is rare; few cases are reported or right-heart overload, i.e., the size of the origin of the fis-
in the literature (Fig. 45.10a) [15]. LAD duplication is more tula and the drainage sites. Large coronary artery fistulae can
common occurring in 0.13–1% of the general population cause myocardial ischemia, congestive heart failure, endo-
[15, 42]. LAD duplication consists of a short LAD coursing carditis, thrombosis, pulmonary hypertension, and/or
and terminating in the proximal interventricular sulcus embolic events. These symptoms may not develop until late
­without reaching the apex and a secondary LAD originating adulthood.
from the LCA (Fig. 45.10b) or RCA and entering the distal Coronary fistulae more commonly involve the RCA (60%
anterior interventricular sulcus and coursing to the apex [15, of cases) than the LCA (40%) [15]. In less than 5% of cases,
42]. LAD duplication should not be confused with an LAD fistulae originate from both the LCA and the RCA. The most
and a diagonal branch running parallel to each other. Such a common site of drainage is the pulmonary artery (Fig. 45.11a,
parallel diagonal branch does not reenter the anterior inter- b), followed by the right ventricle (Fig.  45.11c, d) and the
ventricular sulcus and take over the course of the distal right atrium (Fig.  45.12a, b). Drainage into the left atrium
LAD. Coronary artery duplication is a benign anomaly, but it (Fig. 45.12c, d) or left ventricle occurs in less than 10% of
may complicate surgical intervention when aortocoronary cases [45–47].
bypass or other coronary artery surgeries are performed. It is On MDCT coronary angiography, the involved coronary
also important to identify in congenital heart diseases where artery is often dilated or tortuous, and the site of drainage can
transection of the right ventricular outflow tract would be be clearly visualized. In coronary artery fistula to pulmonary
considered (i.e., tetralogy of Fallot) as an LAD off the RCA artery, the contrast ejection sign can be found in pulmonary
would be disrupted in such a case. artery trunk.

Coronary Artery Anomalies of Termination Coronary Arcade

Coronary Artery Fistula Coronary arcade is a direct communication between the

RCA and the LCA, which is large enough to visualize by
Coronary artery fistula is a condition in which a communica- invasive coronary angiography or MDCT coronary angiogra-
tion exists between a coronary artery and a cardiac chamber phy but without significant coronary artery stenosis.
and a great vessel or other vascular structure [15]. Coronary Coronary arcade should be differentiated from collateral ves-
artery fistulae are rare, with the reported prevalence of sels on the basis of the prominent straight connection
0.002% in the general population and approximately 0.1– between the two unobstructed major arteries, often at or near
0.2% of all patients undergoing selective coronary angiogra- the level of the crux, which is different from the tortuous
45  CT of Coronary Artery Anomalies 575

Fig. 45.11  Right coronary artery a b

fistulae. (a) Thin-slab maximum-
intensity projection image shows
tortuous vessels (blue arrow) in the
vicinity of the pulmonary artery trunk.
The fistula orifice displays the typical
ejection sign into the pulmonary artery
(yellow arrow). (b) 3D volume
reformatted CT angiogram shows the
tortuous vessels (blue arrow) near the
excluded pulmonary artery. (c) Thin-slab
maximum-­intensity projection image
shows the dilated proximal RCA and
large fistula (blue arrow) terminating
into the right ventricle (yellow arrow).
(d) 3D volume reformatted CT
angiogram shows the dilated proximal
RCA and large fistula (blue arrow) c d
terminating into the right ventricle.
AO = aorta; LAD = left descending
artery; LCX = left circumflex artery;
RCA = right coronary artery

Fig. 45.12  Left coronary artery fistulae. a b

(a) 3D volume reformatted CT
angiogram shows a tortuous vessel
(arrows) traveling from the LAD to the
RA. (b) Thin-slab maximum-intensity
projection image shows the tortuous
vessel (arrows) displaying the typical
ejection sign into the right atrium. (c)
Thin-slab maximum-intensity projection
image shows a tortuous vessel (arrow)
traversing from the LCX to the LA. (d)
3D volume reformatted CT angiogram
shows a tortuous vessel (arrows)
traveling from the LCX to the
LA. LA = left atrium; LAD = left
descending artery; LCX = left circumflex
artery; RA = right atrium; RCA = right
coronary artery c d
576 L. J. Zhang et al.

c­ ollateral vessels [15]. In addition, coronary arcade should 4. Angelini P.  Novel imaging of coronary artery anomalies to
be differentiated from coronary to pulmonary artery fistula; assess their prevalence, the causes of clinical symptoms, and
the risk of sudden cardiac death. Circ Cardiovasc Imaging.
the former has no direct communication between the coro- 2014;7(4):747–54.
nary artery and pulmonary artery. 5. van Ooijen PM, Dorgelo J, Zijlstra F, Oudkerk M.  Detection,
visualization and evaluation of anomalous coronary anatomy on
16-slice multidetector-row CT. Eur Radiol. 2004;14(12):2163–71.
6. Datta J, White CS, Gilkeson RC, Meyer CA, Kansal S, Jani ML,
Extracardiac Termination Arildsen RC, Read K.  Anomalous coronary arteries in adults:
depiction at multi-detector row CT angiography. Radiology.
Systemic arterial termination of a coronary artery is uncom- 2005;235(3):812–8.
mon. Connections may exist between the coronary arteries 7. Shi H, Aschoff AJ, Brambs HJ, Hoffmann MH.  Multislice
CT imaging of anomalous coronary arteries. Eur Radiol.
and extracardiac vessels such as the bronchial, internal mam- 2004;14(12):2172–81.
mary, pericardial, anterior mediastinal, superior and inferior 8. Young PM, Gerber TC, Williamson EE, Julsrud PR, Herfkens
phrenic, and intercostal arteries and the esophageal branch of RJ.  Cardiac imaging: Part 2, normal, variant, and anomalous
the aorta. This entity can be distinguished from coronary ­configurations of the coronary vasculature. AJR Am J Roentgenol.
2011;197(4):816–26.
artery fistula in that the coronary artery is not usually 9. Dodd JD, Ferencik M, Liberthson RR, Cury RC, Hoffmann U,
enlarged and tortuous. Most coronary-systemic arterial com- Brady TJ, Abbara S.  Congenital anomalies of coronary artery
munications are not significant, except for the presence of origin in adults: 64-MDCT appearance. AJR Am J Roentgenol.
atherosclerosis in the native coronary artery originating from 2007;188(2):W138–46.
10. Shriki JE, Shinbane JS, Rashid MA, Hindoyan A, Withey JG,
systemic arterial communication [15]. DeFrance A, Cunningham M, Oliveira GR, Warren BH, Wilcox
It may be difficult to identify a coronary artery terminat- A.  Identifying, characterizing, and classifying congenital anoma-
ing into extracardiac artery by MDCT coronary angiography lies of the coronary arteries. Radiographics. 2012;32(2):453–68.
due to the small caliber of the vessels involved. Therefore, 11. Schoepf UJ.  CT of the heart: principles and applications. In:

Contemporary cardiology. Totowa: Humana Press; 2005.
ensuring adequate scanning range and careful slice by slice 12. Zhang LJ, Wang YZ, Huang W, Chen P, Zhou CS, Lu

observation in thin-slice axial contrast-enhanced CT images GM. Anatomical investigation of the sinus node artery using dual-­
is key for diagnosing these coronary-systemic arterial source computed tomography. Circ J. 2008;72(10):1615–20.
communications. 13. Greenberg MA, Fish BG, Spindola-Franco H. Congenital anoma-
lies of coronary artery: classification and significance. Radiol Clin
North Am. 1989;27:1127–46.
14. Kang JW, Seo JB, Chae EJ, Jang YM, Do KH, Lee JS, Song
Summary JW, Lim TH.  Coronary artery anomalies: classification and
electrocardiogram-­ gated multidetector computed tomographic
findings. Semin Ultrasound CT MR. 2008;29(3):182–94.
In conclusion, coronary artery anomalies include anomalies 15. Kim SY, Seo JB, Do KH, Heo JN, Lee JS, Song JW, Choe YH, Kim
of origin, course, and termination. These anomalies can TH, Yong HS, Choi SI, Song KS, Lim TH. Coronary artery anoma-
clearly and accurately be visualized by state-of-the-art lies: classification and ECG-gated multi-detector row CT findings
MDCT coronary angiography owing to high spatial resolu- with angiographic correlation. Radiographics. 2006;26(2):317–33.
discussion 333-314
tion and three-dimensional imaging capability. It is impor- 16. Raimondi F, Bonnet D.  Imaging of congenital anomalies of the
tant to identify and report those anomalies that are malignant, coronary arteries. Diagn Interv Imaging. 2016;97(5):561–9.
such as interarterial or intramural course in patients with an 17. Danias PG, Stuber M, McConnell MV, Manning WJ. The diagnosis
anomalous coronary origin. Due to its superior ability to of congenital coronary anomalies with magnetic resonance imag-
ing. Coron Artery Dis. 2001;12(8):621–6.
depict coronary artery anomalies than conventional invasive 18. Aldana-Sepulveda N, Restrepo CS, Kimura-Hayama E. Single cor-
coronary angiography, MDCT coronary angiography is the onary artery: spectrum of imaging findings with multidetector CT. J
first-line choice for imaging coronary artery anomalies. Cardiovasc Comput Tomogr. 2013;7(6):391–9.
19. Desmet W, Vanhaecke J, Vrolix M, Van de Werf F, Piessens J,
Willems J, de Geest H.  Isolated single coronary artery: a review
of 50,000 consecutive coronary angiographies. Eur Heart J.
References 1992;13(12):1637–40.
20. Frescura C, Basso C, Thiene G, Corrado D, Pennelli T, Angelini
1. Angelini P.  Coronary artery anomalies: an entity in search of an A, Daliento L. Anomalous origin of coronary arteries and risk of
identity. Circulation. 2007;115(10):1296–305. sudden death: a study based on an autopsy population of congenital
2. Perez-Pomares JM, de la Pompa JL, Franco D, Henderson D, Ho SY, heart disease. Hum Pathol. 1998;29(7):689–95.
Houyel L, Kelly RG, Sedmera D, Sheppard M, Sperling S, Thiene 21. Bunce NH, Lorenz CH, Keegan J, Lesser J, Reyes EM, Firmin
G, van den Hoff M, Basso C. Congenital coronary artery anoma- DN, Pennell DJ. Coronary artery anomalies: assessment with free-­
lies: a bridge from embryology to anatomy and pathophysiology- breathing three-dimensional coronary MR angiography. Radiology.
-a position statement of the development, anatomy, and pathology 2003;227(1):201–8.
ESC Working Group. Cardiovasc Res. 2016;109(2):204–16. 22. Lim JC, Beale A, Ramcharitar S, Medscape. Anomalous origina-
3. Sundaram B, Kreml R, Patel S. Imaging of coronary artery anoma- tion of a coronary artery from the opposite sinus. Nat Rev Cardiol.
lies. Radiol Clin N Am. 2010;48(4):711–27. 2011;8(12):706–19.
45  CT of Coronary Artery Anomalies 577

23. Barth CW 3rd, Roberts WC. Left main coronary artery originating tricular function after dual coronary repair. J Am Coll Cardiol.
from the right sinus of Valsalva and coursing between the aorta and 1997;30(2):547–53.
pulmonary trunk. J Am Coll Cardiol. 1986;7(2):366–73. 35. Pena E, Nguyen ET, Merchant N, Dennie C. ALCAPA syndrome:
24. Lorenz EC, Mookadam F, Mookadam M, Moustafa S, Zehr
not just a pediatric disease. Radiographics. 2009;29(2):553–65.
KJ.  A systematic overview of anomalous coronary anatomy and 36. Hazirolan T, Canyigit M, Karcaaltincaba M, Dagoglu MG, Akata
an examination of the association with sudden cardiac death. Rev D, Aytemir K, Besim A. Myocardial bridging on MDCT. AJR Am J
Cardiovasc Med. 2006;7(4):205–13. Roentgenol. 2007;188(4):1074–80.
25. Angelini P, Walmsley RP, Libreros A, Ott DA. Symptomatic anom- 37. Corban MT, Hung OY, Eshtehardi P, Rasoul-Arzrumly E,

alous origination of the left coronary artery from the opposite sinus McDaniel M, Mekonnen G, Timmins LH, Lutz J, Guyton RA,
of valsalva. Clinical presentations, diagnosis, and surgical repair. Samady H.  Myocardial bridging: contemporary understanding of
Tex Heart Inst J. 2006;33(2):171–9. pathophysiology with implications for diagnostic and therapeutic
26. Ashrafpoor G, Danchin N, Houyel L, Ramadan R, Belli E, Paul strategies. J Am Coll Cardiol. 2014;63(22):2346–55.
JF.  Anatomical criteria of malignancy by computed tomography 38. Konen E, Goitein O, Sternik L, Eshet Y, Shemesh J, Di Segni
angiography in patients with anomalous coronary arteries with an E. The prevalence and anatomical patterns of intramuscular coro-
interarterial course. Eur Radiol. 2015;25(3):760–6. nary arteries: a coronary computed tomography angiographic study.
27. Krupinski M, Urbanczyk-Zawadzka M, Laskowicz B, Irzyk M, J Am Coll Cardiol. 2007;49(5):587–93.
Banys R, Klimeczek P, Gruszczynska K, Baron J. Anomalous ori- 39. Morales AR, Romanelli R, Tate LG, Boucek RJ, de Marchena
gin of the coronary artery from the wrong coronary sinus evaluated E.  Intramural left anterior descending coronary artery: sig-
with computed tomography: “high-risk” anatomy and its clinical nificance of the depth of the muscular tunnel. Hum Pathol.
relevance. Eur Radiol. 2014;24(10):2353–9. 1993;24(7):693–701.
28. Zhang LJ, Wu SY, Huang W, Zhou CS, Lu GM. Anomalous origin 40. Ferreira AG Jr, Trotter SE, Konig B Jr, Decourt LV, Fox K, Olsen
of the right coronary artery originating from the left coronary sinus EG. Myocardial bridges: morphological and functional aspects. Br
of valsalva with an interarterial course: diagnosis and dynamic Heart J. 1991;66(5):364–7.
evaluation using dual-source computed tomography. J Comput 41. Lu GM, Zhang LJ, Guo H, Huang W, Merges RD. Comparison of
Assist Tomogr. 2009;33(3):348–53. myocardial bridging by dual-source CT with conventional coronary
29. Zhang LJ, Yang GF, Huang W, Zhou CS, Chen P, Lu GM. Incidence angiography. Circ J. 2008;72(7):1079–85.
of anomalous origin of coronary artery in 1879 Chinese adults on 42. Bozlar U, Ugurel MS, Sari S, Akgun V, Ors F, Tasar M. Prevalence
dual-source CT angiography. Neth Hear J. 2010;18(10):466–70. of dual left anterior descending artery variations in CT angiogra-
30. Cheezum MK, Ghoshhajra B, Bittencourt MS, Hulten EA, Bhatt A, phy. Diagn Interv Radiol. 2015;21(1):34–41.
Mousavi N, Shah NR, Valente AM, Rybicki FJ, Steigner M, Hainer 43. Saboo SS, Juan YH, Khandelwal A, George E, Steigner ML,

J, MacGillivray T, Hoffmann U, Abbara S, Di Carli MF, DeFaria Landzberg M, Rybicki FJ.  MDCT of congenital coronary artery
Yeh D, Landzberg M, Liberthson R, Blankstein R. Anomalous ori- fistulas. AJR Am J Roentgenol. 2014;203(3):W244–52.
gin of the coronary artery arising from the opposite sinus: preva- 44. Said SA, el Gamal MI, van der Werf T. Coronary arteriovenous fis-
lence and outcomes in patients undergoing coronary CTA.  Eur tulas: collective review and management of six new cases--chang-
Heart J Cardiovasc Imaging. 2017;18(2):224–35. ing etiology, presentation, and treatment strategy. Clin Cardiol.
31. Lee HJ, Hong YJ, Kim HY, Lee J, Hur J, Choi BW, Chang HJ, 1997;20(9):748–52.
Nam JE, Choe KO, Kim YJ. Anomalous origin of the right coro- 45. Zhang LJ, Zhou CS, Wang Y, Jin Z, Yu W, Zhang Z, Zhang B, Fang
nary artery from the left coronary sinus with an interarterial course: X, Cui X, Li K, Huang W, Zheng L, Ji XM, Hoffman C, Schoepf
subtypes and clinical importance. Radiology. 2012;262(1):101–8. UJ, Lu GM. Prevalence and types of coronary to pulmonary artery
32. Chaitman BR, Lesperance J, Saltiel J, Bourassa MG. Clinical, angi- fistula in a Chinese population at dual-source CT coronary angiog-
ographic, and hemodynamic findings in patients with anomalous raphy. Acta Radiol. 2014;55(9):1031–9.
origin of the coronary arteries. Circulation. 1976;53(1):122–31. 46. Zhou CS, Zhang LJ, Wang YN, Yu W, Huang W, Zhang B, Zhao YE,
33. Gulati R, Reddy VM, Culbertson C, Helton G, Suleman S, Reinhartz Chen X, Zheng L, Zhang ZQ, Jin ZY, Lu GM. Incidence and classi-
O, Silverman N, Hanley FL.  Surgical management of coronary fication diagnosis of coronary artery fistula at dual-source CT coro-
artery arising from the wrong coronary sinus, using standard and nary angiography. Zhonghua Yi Xue Za Zhi. 2012;92(47):3336–40.
novel approaches. J Thorac Cardiovasc Surg. 2007;134(5):1171–8. 47. Lim JJ, Jung JI, Lee BY, Lee HG. Prevalence and types of coronary
34. Schwartz ML, Jonas RA, Colan SD.  Anomalous origin of left artery fistulas detected with coronary CT angiography. AJR Am J
coronary artery from pulmonary artery: recovery of left ven- Roentgenol. 2014;203(3):W237–43.
 CT Spectrum of Congenital Heart
Disease 46
David Steflik and Anthony M. Hlavacek

Congenital heart disease (CHD) is a relatively common lack of the functional capabilities of MRI, the radiation
problem, affecting approximately 0.8% of newborn infants exposure, and the necessity for intravenous administration
[1]. An understanding of the CT findings of congenital heart of contrast material. Nevertheless, CT is enjoying broader
diseases and their postoperative complications is important use because of its ease of use and widespread availability,
for the radiologist, since patients with palliated or repaired and thus, an understanding of the CT features of CHD is
congenital heart disease are living longer and many of them essential to ensure a correct diagnosis.
will require lifelong care and imaging. After acquisition of the image set, the data should be
reviewed in a segmental fashion, as discussed below. This
chapter reviews the native CT appearances of common con-
Imaging Algorithms genital heart malformations. The CT spectrum of postopera-
tive congenital heart disease is covered in Chap. 46. The CT
The initial imaging algorithm for patients with suspected techniques for evaluating congenital heart disease in children
CHD usually includes an echocardiogram and possible chest are discussed in Chap. 43.
radiograph [2]. However, echocardiography is limited by
acoustic windows, relatively small field of view, and inabil-
ity to image airway structures. Magnetic resonance imaging Sequential Segmental Analysis
(MRI) also has been added to the imaging armamentarium.
MRI provides excellent anatomic and functional information Sequential segmental analysis forms the basis of morpho-
and is particularly valuable in the evaluation of valvular and logic assessment of the heart. There are three segments
myocardial function, but it is time-consuming, and it is con- within the heart: atria, ventricles, and great arteries. Each
traindicated in many patients with pacemakers and in some segment is defined by the morphology of its structures, rather
with implantable cardioverter-defibrillators [3–6]. than their relative location within the chest or connections to
Cardiac CT is being increasingly used in the evaluation other structures. The connections of these different anatomic
of patients with congenital heart disease and has several segments must also be described, which will then correlate
advantages over MRI [7–9]. It is more available and less with the physiology seen by the clinician at the bedside.
time-­consuming than MRI [10, 11]. Fast scan protocols The first structures that must be objectively evaluated are
limit the need for sedation, which is a significant risk in the atria. The most reliable ways to assess the atria are its
patients with congenital heart disease [12]. Current scanners appendages, which are easily seen on cardiac CT [13]. The
provide high spatial resolution, enabling visualization of right atrial appendage is broad and triangular, with pectinate
small structures; high temporal resolution, which can mini- muscles extending into the body of the atrium (Fig. 46.1a).
mize respiratory and cardiac motion artifacts; and isotropic The left atrial appendage is narrow and “fingerlike,” with
voxels, allowing for reconstructions with excellent resolu- pectinate muscles confined to the appendage (Fig  46.1b).
tion. CT allows a more complete evaluation of lung paren- The right atrium usually receives the inferior vena cava,
chyma than MRI.  CT, unlike MRI, is not hampered by unless the intrahepatic portion is interrupted. In the tradi-
postoperative metal artifacts. Limitations of CT include the tional classification system, the usual relationship of the atria
is termed “situs solitus.” The mirror-image relationship, with
the morphologic right atrium leftward of the morphologic
D. Steflik · A. M. Hlavacek (*) right atrium, is termed “situs inversus.” When there are two
Department of Pediatrics, Division of Pediatric Cardiology,
Medical University of South Carolina, Charleston, SC, USA
right or left atrial appendages, which is found in atrial
e-mail: [email protected] isomerism, or heterotaxy, the arrangement is classically
­

© Humana Press 2019 579

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_46
580 D. Steflik and A. M. Hlavacek

Fig. 46.1  Axial images

displaying the typical
a b
characteristics of right and
left atrial appendages. Note
that the right atrial appendage
(a) has a broad base with
pectinate muscles extending
into the body of the atrium,
while the left atrial appendage
(b) is generally narrow with
pectinate muscles confined to
the body of the appendage
(RAA right atrial appendage,
LAA left atrial appendage)

called “situs ambiguous,” although defining the atrial the pulmonary artery. The ventriculoarterial connection can
arrangement as “right” or “left atrial isomerism” is more also be described as “double outlet” (usually from the RV)
descriptive [14]. or “common,” as is found with truncus arteriosus. The ori-
The next structures that must be systematically assessed entation of the great arteries is generally defined by the
are the ventricles. The right ventricle is most readily identi- location of the aortic valve relative to the pulmonary valve.
fied by coarse apical trabeculations and a moderator band, In a normal heart, the aortic valve is oriented posterior and
while the left ventricle has relatively fine apical trabecula- rightward to the pulmonary valve. With usual transposition
tions. When the respective atrioventricular valves are patent, (“D-transposition”), the aortic valve is generally anterior
the morphologic right ventricle is noted to have a muscular and rightward of the pulmonic valve, but is usually anterior
infundibulum separating the tricuspid valve from the semilu- and leftward in congenitally corrected transposition
nar valve, while the left ventricle has fibrous continuity (“L-transposition”).
between the mitral valve and semilunar valve (if present).
The usual relationship between the ventricles is termed
“D-looped,” while the mirror-image arrangement is termed Atrial Septal Defects
“L-looped.” In the rare instance that the ventricles cannot be
described in a right-left relationship, the arrangement is Atrial septal defects represent 8–10% of congenital heart
often termed “X-looped.” defects [15]. Secundum, primum, and sinus venosus defects
Once the atrial and ventricular orientations are defined, constitute the three most common types of ASD. Secundum
the atrioventricular connections can be described. If the right defects are the most common type of atrial septal defect,
atrium (RA) is connected to the right ventricle (RV) and the constituting 75% of defects [16]. A secundum ASD is a
left atrium (LA) is connected with the left ventricle (LV), the defect in the central portion of the atrial septum, or the area
atrioventricular connections are termed “concordant.” containing the fossa ovalis (Fig. 46.2). These defects can be
“Discordant” atrioventricular connections exist when the RA large, and are usually solitary, but can also consist of multi-
connects to the LV and the LA connects with the RV. When ple smaller communications. Primum defects reside in the
atrial isomerism is present, the connections are termed apical (inferior) portion of the atrial septum and are contigu-
“mixed.” A “univentricular” connection exits when there is ous with the atrioventricular valves. These defects are gener-
double inlet to one ventricle (generally left) or there is the ally considered to fall within the spectrum of atrioventricular
absence of one atrioventricular connection (as in classical septal defects and are generally accompanied by a “cleft”
tricuspid atresia). mitral valve. Sinus venosus defects involve a deficiency of
Finally, the connections and orientation of the great the wall that separates the right pulmonary veins from the
arteries must be assessed. A concordant ventriculoarterial superior vena cava (or infrequently inferior vena cava) and
connection exists when the morphologic RV connects to thus the right atrium, creating an interatrial communication
the pulmonary artery and the morphologic LV connects (Fig.  46.3). All sinus venosus defects are associated with
with the aorta. A discordant ventriculoarterial connection anomalous drainage of at least one of the right pulmonary
(or “transposition”) exists when the morphologic RV con- veins, usually with connection of the right superior pulmo-
nects to the aorta and the morphologic LV connects with nary vein to the superior vena cava [17]. The least common
46  CT Spectrum of Congenital Heart Disease 581

Fig. 46.2 (a) “Four-chamber

view” displaying a secundum
a b
atrial septal defect (*). The
right atrium and ventricle are
dilated. (b) Axial image
revealing dilation of the main
pulmonary artery, which is
common in atrial septal
defects (RA right atrium, RV
right ventricle, LA left atrium,
LV left ventricle, MPA main
pulmonary artery)

blood seen in the defect. Cardiac CT can reliably define the

location and size of atrial septal defects in larger children and
adults [18], particularly when cardiac gating is applied, but
will occasionally be inaccurate in non-gated studies or with
smaller children and infants. When compared to transtho-
racic and transesophageal echocardiography in adult patients,
cardiac CT has been shown to be superior in determining
candidacy for device closure in the catheterization laboratory
[19] and can define device protrusion or malposition after
closure [20, 21]. It can also be helpful for preoperative
assessment of pulmonary venous drainage in patients with
sinus venosus defects [22, 23]. Patients with a hemodynami-
cally significant defect will display dilation of the right ven-
tricle and main pulmonary artery on cardiac CT. Ventricular
volume measurements can be utilized to quantify the degree
of left to right shunt [18, 24].
The differential diagnostic consideration for secundum
ASD is a patent foramen ovale, in which there is incomplete
fusion between the septum primum and septum secundum
which creates a flap-like opening at the foramen ovale. There
Fig. 46.3  Axial image revealing the typical location of a superior sinus is no structural deficiency of septal tissue. The patent fora-
venosus defect (*) men is small and stays closed as long as left atrial pressure is
greater than right atrial pressure, but it can reopen when right
atrial pressure increases, such as with a Valsalva maneuver,
form of ASD is the coronary sinus defect, in which the coughing, or sneezing. In this case, the blood can flow from
­coronary sinus is at least partially “unroofed,” resulting in a the right to left atrium.
communication from the LA to RA.  Comprising <1% of
ASD, it is usually seen in association with heterotaxy syn-
dromes with bilateral superior vena cavae, although it can
also occur in isolation. The major determinant of the magni- Ventricular Septal Defects
tude and direction of any atrial-level shunt is the relative
compliance of the ventricles. Significant left to right shunts Ventricular septal defects (VSD) are the most common con-
can lead to right atrial and ventricular enlargement and flow- genital heart lesions identified during childhood, making up
related elevations in pulmonary arterial pressures but rarely 20–30% of congenital heart defects [25]. There are multiple
result in irreversible pulmonary hypertension. classification schemes for ventricular septal defects, but the
The CT features of atrial septal defects are discontinuity most straightforward method is to define the defect as mus-
of a portion of the interatrial septum with contrast-enhanced cular, perimembranous, or subarterial. Muscular defects are
582 D. Steflik and A. M. Hlavacek

Fig. 46.4 (a) “Four-chamber

a b
view” in a patient with a
perimembranous ventricular
septal defect. Note that the
defect (*) is in fibrous
continuity with the tricuspid
and aortic valve (black
arrow). (b) Axial image in a
patient whose
perimembranous defect has
been closed spontaneously by
aneurysmal tricuspid valve
tissue (*). (c) “Four-chamber
view” in an infant with an
apical muscular ventricular
septal defect (*). (d) 3D
volume-rendered image from
the same patient in which the
free wall of the right ventricle
has been cropped, providing c d
an “en-face” view of the
ventricular septum. The defect
is inferior to the moderator
band (white arrow) (RV right
ventricle, LV left ventricle)

completely surrounded by muscular septum (Fig.  46.4). relative pulmonary and systemic vascular resistances. Small
Based on the location on the right ventricular aspect of the defects are generally well tolerated and often close spontane-
septum, muscular defects may be further described as inlet, ously. Significant pulmonary overcirculation from a large
mid-muscular, apical, or outlet. Perimembranous defects ventricular-level shunt results in left atrial and ventricular
(also known as membranous, conoventricular, or infracristal dilation, poor growth, and tachypnea. If left unrepaired, large
defects) border the membranous septum, usually resulting in defects will result in irreversible pulmonary hypertension.
fibrous continuity between the tricuspid and aortic valves
(Fig. 46.4). These defects are the most common type of VSD,
accounting for up to 80% of defects identified at surgery or Patent Ductus Arteriosus
autopsy [26]. Subarterial defects (also known as supracristal,
conal, infundibular, subpulmonary, or doubly committed) A patent ductus arteriosus is a normal finding in nearly every
result in fibrous continuity between the pulmonary and aortic fetus, with usual functional closure within 24  h after full-­
valves, due to an absence of a subpulmonary infundibulum. term birth [27]. Excluding premature infants, the ductus
Perimembranous and subarterial defects can be accompanied remains patent in about 1 in 2000 live births, accounting for
by aortic valve prolapse into the VSD, resulting in insuffi- about 7% of congenital heart disease [28]. With decreasing
ciency. Subaortic membranes and muscular subpulmonary gestation or other congenital abnormalities, the incidence
stenosis are also more frequent in patients with a perimem- increases dramatically [29].
branous defect. With CT, an isolated ductus is usually visualized as a
The degree of left to right shunt through a ventricular sep- tubular structure extending from the underside of the aortic
tal defect is dependent upon the size of the defect and the arch just below the origin of the left subclavian artery to the
46  CT Spectrum of Congenital Heart Disease 583

Fig. 46.5  Axial (a) and 3D

a b
volume-rendered (b) images
in an infant with a patent
ductus arteriosus (arrow)
connecting the main
pulmonary artery to the
descending aorta (MPA main
pulmonary artery, Ao aorta)

left pulmonary artery (Fig.  46.5). The communication The presence of significant AV valve regurgitation will
between the pulmonary arteries is best seen on sagittal or worsen any heart failure symptoms present. CT findings
oblique reformations. After birth, it usually begins constrict- include atrial enlargement, ventricular dilation (right, left, or
ing at the pulmonary artery end, resulting in a conical shape. biventricular), and dilation of the pulmonary arteries.
However, numerous ductal shapes have been described [30].
The ductus arteriosus may become aneurysmal and calcified,
which may lead to rupture [30, 31]. The degree of left to right Pulmonary Stenosis
shunt through a patent ductus is dependent upon its diameter
and length. The shunt is also dependent on the difference Pulmonary stenosis most commonly occurs at the level of the
between the pulmonary and systemic vascular resistances valve but can also be intracavitary (within the body of the
[16]. The physiology is similar to VSD, with large shunts RV), subvalvar (in the pulmonary infundibulum), supraval-
resulting in left atrial and ventricular dilation, poor growth, var (at the sinotubular junction), or more distal in the main or
and tachypnea. branch pulmonary arteries. It is imperative to determine the
level(s) of obstruction, as treatment differs accordingly.
Typical pulmonary valve stenosis results in doming of the
Atrioventricular Septal Defects pulmonary valve in systole, but the valve can also be dys-
plastic with markedly thickened leaflets. Subvalvar obstruc-
Atrioventricular septal defects (AVSD), also known as atrio- tion can occur in isolation but is usually found in the presence
ventricular canal or endocardial cushion defects, account for of other abnormalities. “Double-chambered right ventricle”
5% of congenial heart disease. Atrioventricular septal defects is a specific form of pulmonary obstruction resulting from
include a wide array of anatomical variants that all have a hypertrophied muscle bundles that septate the RV into a
defect of the atrioventricular septum and abnormalities of the high-pressure proximal chamber and a low-pressure distal
atrioventricular (AV) valves. Many of these patients will chamber. It is mostly associated with ventricular septal
have a common AV valve, with shunting at both the atrial and defects of the perimembranous variety. Peripheral pulmonic
ventricular levels (Fig. 46.6). Others will have separate AV stenosis occurs mainly with other cardiac defects but can be
valves, with shunting at the atrial level, ventricular level, an isolated lesion. Diffuse pulmonary artery hypoplasia/ste-
both, or neither. The defect in the atrioventricular septum nosis is found in both Alagille and Williams syndromes. In
results in an aortic valve that is displaced anteriorly, creating addition to visualization of the level of pulmonary outflow
a “gooseneck deformity” that predisposes these patients to obstruction (Fig. 46.7), cardiac CT will reveal right ventricu-
subaortic stenosis. The physiology and clinical manifesta- lar hypertrophy.
tions of atrioventricular septal defects are determined by the
location and amount of left to right shunting that occurs
through the corresponding septal defects, along with the Tetralogy of Fallot
degree of AV valve regurgitation. Patients with primarily
atrial-level shunting will have physiology similar to a secun- Tetralogy of Fallot (TOF) is the most common form of cya-
dum ASD, while patients with significant ventricular-level notic congenital heart disease. The multiple anatomic fea-
shunting will have physiology similar to a VSD (see above). tures of Tetralogy of Fallot result from a single anatomic
584 D. Steflik and A. M. Hlavacek

Fig. 46.6  Axial (a) and

“four-chamber view” (b) in
a b
two infants with an
atrioventricular septal defect.
In the “complete” form, there
is a common atrioventricular
valve (white arrow) along
with a large septal defect
(black arrow), allowing
shunting at the atrial and
ventricular levels (RA right
atrium, RV right ventricle, LA
left atrium, LV left ventricle)

Fig. 46.7 (a) Oblique

sagittal image in a child with
a b
valvar pulmonary stenosis.
Note the thickened valve
(arrows). (b) 3D volume-­
rendered image views from an
anterior perspective, revealing
diffuse hypoplasia of the right
pulmonary artery, along with
multiple focal stenoses, in a
child with Williams syndrome
(RPA right pulmonary artery)

abnormality: the anterior displacement of the outlet septum. becomes more apparent as a result of increased right to left
This ­anterior displacement results in subvalvar pulmonary shunting through the VSD.
stenosis, an anterior malalignment VSD, “over-riding” aorta,
and right ventricular hypertrophy (Fig. 46.8). The pulmonary
valve is often hypoplastic or stenotic. The most extreme form Aortic Stenosis
of this abnormality is TOF with pulmonary atresia (also
known as pulmonary atresia with VSD), in which the pulmo- The precise incidence of aortic stenosis is unknown, given
nary circulation is supplied by a patent ductus arteriosus or that minor malformations such as bicuspid aortic valve
aortopulmonary collateral arteries (Fig. 46.8d). CT has been are common and generally asymptomatic in childhood.
shown to be highly accurate in the identification of pulmo- The incidence of congenital bicuspid valve is 1.3% of the
nary arterial supply in these patients [32, 33]. In 3–5% of population, making it the most common congenital heart
patients with TOF, the pulmonary valve is absent, resulting defect. Approximately 60–75% of patients with congeni-
in massive dilation of the main and branch pulmonary arter- tal aortic stenosis are at the valve level [16]. Among those
ies during fetal development. The physiology and clinical with congenital aortic stenosis, about 20% will have some
manifestations of TOF are primarily influenced by the degree other form of congenital heart disease, most commonly
of pulmonary outflow obstruction. In those with no or little VSD, coarctation, or PDA [34]. Bicuspid aortic valve is
obstruction to pulmonary blood flow, the oxygen saturations common in patients with Turner’s syndrome, while
may be normal, and the physiology is similar to a large VSD, patients with supravalvar stenosis often have Williams
with pulmonary overcirculation and heart failure symptoms. syndrome. Similar to pulmonary stenosis, aortic stenosis
As the degree of pulmonary obstruction increases, cyanosis can occur in the subvalvar, valvar, or supravalvar region
46  CT Spectrum of Congenital Heart Disease 585

Fig. 46.8 (a) Axial image in

an infant with tetralogy of
a b
Fallot. Anterior deviation of
the outlet septum (*) results
in narrowing of the
pulmonary outflow. (b)
Oblique sagittal slice in the
same infant, displaying an
aorta that “overrides” the
ventricular septum. (c) 3D
volume-rendered image from
the same patient in which the
free wall of the right ventricle
has been cropped, providing
an “en-face” view of the
ventricular septum. The
anteriorly deviated outlet
septum (*) is seen, resulting
in pulmonary outflow
obstruction and a ventricular c d
septal defect (X). (d) 3D
volume-rendered
reconstruction viewed from
an anterior perspective in an
infant with tetralogy of Fallot
with pulmonary atresia and
multiple AP collaterals. Three
collateral vessels are seen
originating from the
descending aorta, with one
vessel supplying the right
lung and two collateral
vessels supplying the left lung
(Ao aorta, P pulmonary
outflow, RV right ventricle, LV
left ventricle, DA descending
aorta)

Fig. 46.9 (a) 3D volume-­

a b
rendered reconstruction in a
“three-chamber view”
demonstrating a discrete
subaortic membrane (arrows).
(b) Maximum intensity
projection image in an
oblique sagittal plane
revealing supravalvar aortic
stenosis consisting of tubular
hypoplasia of the ascending
aorta (*)
586 D. Steflik and A. M. Hlavacek

(Fig. 46.9). Abnormal septal tissue growth, as in hypertro- coarctation of the aorta can present in cardiogenic shock
phic cardiomyopathy, and subaortic membranes can cause once the ductus arteriosus closes. CT has been shown to be
obstruction below the level of the valve. Valvar aortic ste- highly accurate in describing the anatomy prior to surgical
nosis is most commonly due to abnormal fusion of valve repair [35, 36].
commissures, resulting in a b­ icuspid or, rarely, unicuspid
aortic valve. Older patients with aortic stenosis often have
thickened, dysplastic valves. The most common form of a Interrupted Aortic Arch
bicuspid aortic valve results from fusion of the right and
left leaflets. Supravalvar stenosis is the least common Interrupted aortic arch (IAA) is defined as the lack of luminal
form and commonly occurs at the level of the sinotubular continuity between the ascending and descending aortas.
junction, although diffuse hypoplasia of the ascending [37]. The type of interrupted arch is classified by the location
aorta can occur as well. In addition to visualization of the of the discontinuity. Type A interruptions are located distal to
level of obstruction, cardiac CT will reveal left ventricular the left subclavian artery and are associated with aortopulmo-
hypertrophy. nary window defects [38]. If the interruption occurs between
the left carotid and subclavian arteries, it is called a type B
interruption. Type B interruptions frequently have a ventricu-
Coarctation of the Aorta lar septal defect with posterior malalignment of the outlet
septum and are the most common form of IAA. An interrup-
Coarctation of the aorta makes up 6–8% of congenital heart tion between the brachiocephalic and left carotid arteries (or
disease [16]. Genetically, it is associated with Turner’s syn- between the carotid arteries) is called a type C interruption,
drome. There is also a strong association between coarcta- which is the rarest form of IAA. Like other critical left-sided
tion and bicuspid aortic valve. Coarctation of the aorta obstructive lesions (critical aortic stenosis, coarctation, hypo-
typically occurs in the region distal to the left subclavian plastic left heart syndrome, etc.), neonates with IAA are
artery (also known as the aortic isthmus), near the insertion dependent on ductal patency for systemic blood flow
of the ductus arteriosus. CT will often reveal a discrete nar- (Fig.  46.11). Those that are not identified and treated after
rowing with a prominent posterior “shelf” (Fig.  46.10). birth present with cardiogenic shock as the ductus arteriosus
This type is classically called the “juxtaductal” or “adult” closes. Management is thus aimed at maintaining patency of
form, although it is often seen in infants as well. Older chil- the ductus arteriosus until ultimate surgical intervention. CT
dren and adults will often display aortic dilatation proximal has been shown to be useful in this diagnosis and often plays
and distal to the coarctation and collateral vessel formation an important role in the preoperative evaluation [39].
on CT. Collateral circulation is usually via the internal
mammary arteries and the intercostal arteries. Increased
collateral flow through intercostal vessels can cause notch- Ebstein’s Anomaly
ing of the posterior third through eighth ribs. The “infan-
tile” or “preductal” form results in long-segment hypoplasia Ebstein’s anomaly is a rare disorder, making up <1% of con-
of the aortic isthmus (Fig.  46.10). Neonates with severe genital heart defects. It has been associated with maternal

Fig. 46.10  Both images are a

3D volume-rendered
b
reconstructions in an oblique
sagittal “candy cane view.”
Image (a) displays the
“infantile form” of
coarctation, with tubular
hypoplasia of the aortic
isthmus (*). Image (b) reveals
the “adult form” of
coarctation, with discrete
narrowing (arrow) and
numerous engorged spinal
and intercostal arteries due to
collateral flow
46  CT Spectrum of Congenital Heart Disease 587

a b

Fig. 46.11  These images are from a neonate with a type B interrupted subclavian arteries. A patent ductus arteriosus provides blood flow to
aortic arch and aberrant right subclavian artery. (a) Maximum intensity the descending aorta, which supplies an aberrant right subclavian artery
projection image in a coronal plane displaying a hypoplastic ascending (*) (AA ascending aorta, DA descending aorta, LCCA left common
aorta that supplies only the right and left common carotid arteries. (b) carotid artery, LSCA left subclavian artery, PDA patent ductus
3D volume-rendered reconstruction viewed from a rightward perspec- arteriosus)
tive, revealing the interruption between the left common carotid and left

Fig. 46.12  Axial image (a)

a b
and 3D volume-rendered
reconstruction in a “four-­
chamber view” (b) in two
patients with Ebstein’s
anomaly. Note the apical
displacement of the septal
leaflet of the tricuspid valve
(arrow) and dilation of the
right atrium (RA)

lithium use, but most cases are idiopathic. The typical CT erate adequate flow across the pulmonary valve, resulting in
finding of Ebstein’s anomaly consists of apical and rotational right to left shunting at the atrial level and cyanosis. Those
displacement of the tricuspid valve toward the right ventricu- with the most extreme forms are ductal dependent at birth,
lar outflow tract, resulting in an “atrialization” of the right requiring single-ventricle palliation strategies [41]. Those
ventricle proximal to the displaced tricuspid valve [40] with the mildest disease experience minimal symptoms.
(Fig. 46.12). There is also a tethering and fenestration of the Other patients with greater severity require surgical interven-
TV anterior leaflet with significant TV annular dilation. As a tion later in childhood or early adulthood due to progressive
result of tricuspid regurgitation during fetal life, neonates exercise intolerance.
with severe Ebstein’s anomaly often display massive dilation
of the right atrium and atrialized right ventricle.
The physiology and clinical presentation in Ebstein’s Transposition of the Great Arteries (D-TGA)
anomaly depend on the degree of tricuspid regurgitation and
atrialization of the right ventricle. Neonates with severe tri- This anomaly is characterized by concordant atrioventricu-
cuspid regurgitation and atrialization lack the ability to gen- lar connections with discordant ventriculoarterial connec-
588 D. Steflik and A. M. Hlavacek

Fig. 46.13  Images (a) (axial)

a b
and (b) (oblique sagittal) are
from an adult with
transposition of the great
arteries (D-TGA). The aorta
and main pulmonary artery
leave the heart in a parallel
fashion, with the aortic valve
residing anterior and
rightward of the pulmonary
valve (Ao aorta, PA main
pulmonary artery, RV right
ventricle, LV left ventricle)

tions. The aorta arises from the morphologic right ventricle vast majority of these patients will have associated lesions,
and is anterior and to the right of the pulmonary artery, including VSD, pulmonary stenosis, and/or tricuspid valve
which arises from the morphologic left ventricle (Fig. 46.13). abnormalities [42]. In the context of usual atrial arrangement
Many patients with TGA also have an ASD and/or VSD as (“situs solitus”), the morphologic right ventricle will be left
well. In transposition physiology, the systemic and pulmo- sided (“L-looped”) causing the aortic valve to be positioned
nary circulations are in parallel, whereas in the usual anterior and leftward to the pulmonary valve (Fig.  46.14).
­circulation, these circulations are in series. As a result, pul- Hence, this defect is occasionally called “L-TGA.” However,
monary venous blood (oxygenated) is pumped back into the approximately 5% of these patients have a mirror-image atrial
pulmonary circulation, while systemic venous return (deox- arrangement (“situs inversus”), which would cause the aortic
ygenated) is pumped back into the systemic circulation. In valve to be positioned anterior and rightward. For this reason,
order to be compatible with life, there must be mixing of CCTGA is the preferred terminology.
these two circulations to allow oxygenated blood to reach Viewed in simplistic terms, the physiology in CCTGA is
the systemic circulation and deoxygenated blood to reach normal, given that systemic venous return is sent to the lungs
the lungs. Mixing is most effective at the atrial level, but and pulmonary venous return is sent to the body. This is sup-
shunting at the ventricular level or patent ductus arteriosus ported by the fact that some patients are not identified until
can contribute as well. Neonates with inadequate mixing of their teenage years or adulthood. However, the right ventricle
the two ­circulations are born with severe hypoxemia, which is not morphologically designed to pump against systemic
can result in metabolic acidosis due to inadequate oxygen afterload, and many patients will eventually develop heart
delivery. These patients are usually taken for emergent bal- failure symptoms due to right ventricular systolic dysfunc-
loon septostomy in the catheterization laboratory to create a tion. Presentation and symptomology during childhood are
larger atrial communication, allowing increased mixing. largely dependent on associated abnormalities (VSD, pul-
Coronary abnormalities are common and must be defined monary stenosis, or complete heart block).
prior to surgical repair. Neonatal repair is the standard
approach, with rare exceptions.
Double Outlet Right Ventricle

Congenitally Corrected Transposition (L-TGA) In double outlet right ventricle, both great arteries arise from
the right ventricle (≥50% of the circumference of both arte-
Congenitally corrected transposition of the great arteries rial valves). Usually, both arterial valves are separated from
(CCTGA) results when the atrioventricular and ventriculoar- the mitral valve by a muscular infundibulum. The orientation
terial connections are both discordant. Hence, CCTGA is of the great arteries is variable. There is always a VSD, which
occasionally called “double discordance.” This results in sys- is conventionally described in relation to the nearest arterial
temic venous return emptying normally into the morphologic valve: subaortic, subpulmonary, doubly committed (equally
RA and then passing through a mitral valve into the morpho- close to both arterial valves), or noncommitted (distant from
logic LV and then to the lungs via the pulmonary artery. This both arterial valves). Often, the arterial valve that is furthest
pulmonary venous return then empties normally into a mor- from the VSD has subvalvar or valvar stenosis, but there is
phologic LA, then through a tricuspid valve and a morpho- great variability with this lesion. The most common variant
logic RV, and then to the body via the aorta (Fig. 46.14). The of DORV has a subaortic VSD and pulmonary stenosis
46  CT Spectrum of Congenital Heart Disease 589

Fig. 46.14  These images are

a b
from a patient with
congenitally corrected
transposition (L-TGA). (a) In
this axial image, the
discordant atrioventricular
connections can be seen. (b)
Oblique sagittal image
revealing “double
discordance,” with the left
atrium supplying the right
ventricle, which is connected
to the aorta. (c) Oblique
coronal image, also
displaying “double
discordance” (right
atrium→left
ventricle→pulmonary artery).
As is common with this
defect, one can identify a c d
repaired ventricular septal
defect, subpulmonary
obstruction, parallel great
arteries, and a pacemaker lead
in the left ventricle due to
complete heart block. (d) This
axial image reveals that the
aortic valve is anterior and
leftward to the pulmonary
valve (RA right atrium, RV
right ventricle, LA left atrium,
LV left ventricle, MPA main
pulmonary artery, Ao aorta)

a b

Fig. 46.15 (a) Axial image in an infant with double outlet right ven- the right ventricle has been cropped, providing an “en-face” view of
tricle and a subaortic ventricular septal defect. The pulmonary outflow the ventricular septum. Note the subaortic ventricular septal defect (*)
(arrow) reveals both subvalvar stenosis and valvar hypoplasia. (b) 3D and pulmonary outflow obstruction (arrow). (Ao aorta, P pulmonary
volume-rendered image from the same patient in which the free wall of artery)
590 D. Steflik and A. M. Hlavacek

Fig. 46.16  These images are

from an infant with double
a b
outlet right ventricle and
subpulmonary ventricular
septal defect. Images (a, b)
are coronal slices (a is
anterior to b), revealing a
hypoplastic aortic outflow and
widely patent pulmonary
outflow from the right
ventricle. The aorta is seen
diving posteriorly to the right
of the pulmonary artery.
Images (c, d) (oblique sagittal
plane and 3D volume-­
rendered reconstruction),
showing both great arteries
originating from the right
ventricle. The only egress
from the left ventricle is via a c d
ventricular septal defect (*).
Note the narrow aortic
outflow and hypoplastic aortic
arch (Ao aorta, P pulmonary,
RV right ventricle, LV left
ventricle)

(Fig.  46.15), providing physiology similar to TOF (see valve stenosis and/or regurgitation is common. Truncus
above), depending on the degree of pulmonary outflow arteriosus is classified into three major types. Type 1 is the
obstruction. The type with a subpulmonary VSD is c­ ommonly most common, with a short main pulmonary artery segment
called a “Taussig-Bing anomaly” and is associated with aor- giving rise to both branch pulmonary arteries. Type II is the
tic hypoplasia and/or coarctation (Fig. 46.16). A ­biventricular next most common, with both branch pulmonary arteries
repair is feasible in most cases, with baffling of the VSD arising separately from the posterior wall of the common
patch to one of the arterial valves. trunk. In type III, both pulmonary arteries separate from the
lateral walls of the common trunk. A type IV was described
in the original classification scheme, but is no longer felt to
Truncus Arteriosus represent a form of truncus arteriosus [44]. Aortic arch
abnormalities are a common associated finding, with
Truncus arteriosus is a rare cardiac anomaly, accounting for 21–36% of patients having a right aortic arch and approxi-
<1% of congenital heart disease. It is frequently associated mately 11% displaying an interrupted arch [45]. Truncus
with DiGeorge syndrome, especially when it is paired with arteriosus is one of the few congenital heart defects in which
an interrupted aortic arch. With truncus arteriosus (also a ductus arteriosus is not present during fetal life, unless
known as “common arterial trunk”), there is a single artery there is an interrupted aortic arch.
supplying blood flow to the pulmonary, systemic, and coro-
nary circulations (Fig. 46.17). A subarterial-type ventricular
septal defect is always present, although there may be addi- Vascular Rings
tional septal defects. The single semilunar valve, often
called the “truncal valve,” can be comprised of one to five Vascular rings represent approximately 1–2% of congenital
leaflets, although it is most often tricuspid [43]. Truncal heart defects, although this is likely an underestimate, since
46  CT Spectrum of Congenital Heart Disease 591

Fig. 46.17  Images (a, b) b

a
(axial slice, 3D volume-­
rendered reconstruction
viewed from posterior
perspective) are from an
infant with truncus arteriosus
type I. Aortic and pulmonary
blood flow originate from a
common trunk. A short main
pulmonary artery (arrow)
supplies a right and left
pulmonary artery, with the
aorta continuing superiorly to
provide systemic blood flow.
Images (c, d) are from an
infant with truncus arteriosus
and an interrupted aortic arch
(axial slice, 3D volume-­
rendered reconstruction
viewed from posterior c d
perspective). A common trunk
supplies the pulmonary
arteries and a hypoplastic
ascending aorta. The
descending aorta is supplied
by a patent ductus arteriosus
(Ao aorta, RPA right
pulmonary artery, LPA left
pulmonary artery, AAo
ascending aorta, PDA patent
ductus arteriosus, DAo
descending aorta)

many patients are asymptomatic [46]. Vascular rings are lum) can be seen originating from the leftward aspect of the
defined as an arch anomaly in which the trachea and esopha- descending aorta behind the trachea and esophagus, continu-
gus are completely encircled by vascular structures (or liga- ing as the left subclavian (Fig. 46.19). Also originating from
mentous structures with a vascular origin). There is this diverticulum, a ductus arteriosus or ligamentum connects
insufficient space to cover all forms of vascular rings, but the anteriorly to the main pulmonary artery to complete the ring.
two most common types are double aortic arch and right aor- With any type of confirmed or suspected vascular ring,
tic arch with an aberrant left subclavian and diverticulum of CT can accurately define the aortic arch anatomy and branch-
Kommerell. A more detailed coverage of vascular rings can ing pattern [48]. In addition, evidence of airway compression
be found in a review by Ramos-Duran [47]. can also be identified [49]. The caliber and course of vascular
A double aortic arch results from persistence of both the structures can help identify ligamentous structures, such as
right and left aortic arches, each providing ipsilateral carotid the ligamentum arteriosum or an atretic arch segment [47].
and subclavian arteries (Fig.  46.18). In 80% of cases, the
right arch is dominant, with the left arch being smaller and
more caudal [47]. Occasionally a segment of either arch is Anomalous Pulmonary Venous Connections
atretic (usually left), persisting only as a ligamentous
structure. By definition, a pulmonary venous connection is anoma-
A right aortic arch with an aberrant left subclavian and lous whenever it connects to a structure other than another
diverticulum of Kommerell is the second most commonly pulmonary vein or the left atrium. Totally anomalous pul-
identified vascular ring. In this lesion, after coursing to the monary venous connection (TAPVC) occurs when the
right of the trachea, an outpouching (Kommerell’s diverticu- entire lung drains anomalously. When at least one pulmo-
592 D. Steflik and A. M. Hlavacek

Fig. 46.18  Infant with a

double aortic arch. (a) Axial
a b
slice, showing a dominant
right aortic arch. (b) Axial
slice, displaying a smaller left
aortic arch. (c, d) 3D
volume-rendered
reconstructions viewed from
anterior (c) and superior (d)
perspectives, revealing both
aortic arches encircling the
trachea and esophagus. Note
that the right arch is superior
to the left arch, which is
typical. The double arch
results in significant
compression of the trachea in
this patient (RAA right aortic
arch, LAA left aortic arch, T
trachea, E esophagus) c d

nary vein connects normally to the left atrium, the condi- or the inferior vena cava. The mixed type contains a combi-
tion is considered partially anomalous pulmonary venous nation of any of the above, without any pulmonary veins
connection (PAPVC). CT performs particularly well in draining normally to the left atrium. In order to be compat-
the visualization of anomalous pulmonary venous connec- ible with life, there must be an atrial-level communication
tions and is being increasingly used in patients with these with right to left shunting. The clinical sequelae are primar-
lesions [50, 51]. ily determined by the presence/absence of obstruction to
The most widely accepted classification system of pulmonary venous return. Obstruction generally occurs
TAPVC splits the types into supracardiac, cardiac, infracar- within the pulmonary venous drainage itself (e.g., within
diac (or infradiaphragmatic), and mixed [46] (Fig. 46.20). the vertical vein) but can also occur at the atrial septum.
With exception of the mixed type, all of the pulmonary The presence of obstruction results in profound cyanosis
veins usually connect to a confluence posterior to the left and decreased cardiac output. Obstruction is most com-
atrium prior to reaching the ultimate location of drainage. monly seen in patients with infracardiac TAPVC and is
In the supracardiac type, this confluence drains superiorly rarely seen with the cardiac type, with the supracardiac
via a vertical vein into either the innominate vein, the azy- type falling somewhere between the two [52]. Although
gous vein, or the superior vena cava. In the cardiac type, the TAPVC is considered a “cyanotic heart defect,” with right
pulmonary venous system drains directly into the coronary to left shunting at the atrial level, the anomalous pulmonary
sinus or the right atrium. With the infracardiac type, the venous connection itself also results in a left to right shunt,
confluence drains inferiorly via a vertical vein, through the with oxygenated blood directed back into the pulmonary
diaphragm and into the portal venous system, hepatic veins, circulation. Therefore, patients with unobstructed TAPVC
46  CT Spectrum of Congenital Heart Disease 593

a b

c d

Fig. 46.19  Images (a–c) are from a child with a right aortic arch with caliber as the diverticulum transitions to the left subclavian artery, cre-
an aberrant left subclavian artery and diverticulum of Kommerell. (a) ated by the presence of a ligamentum tethering this area to the main
Axial image showing an aortic arch to the right of the trachea, supply- pulmonary artery. Image (d) is taken from a patient that also has a right
ing an aberrant left subclavian artery. Note the diverticulum (*), which aortic arch with an aberrant left subclavian. This patient did not have a
has a similar caliber to the aortic arch. (b) 3D volume-rendered recon- vascular ring. In contrast to image (c), note the absence of a diverticu-
struction viewed from a superior perspective, revealing the right aortic lum of Kommerell, resulting in a left subclavian artery with a uniform
arch and aberrant left subclavian partially encircling the trachea and caliber and course. This appearance of the left subclavian confirms the
esophagus. A ligamentum (not seen) will connect the diverticulum (*) absence of a ligamentum in this area and, therefore, the absence of a
and main pulmonary artery, completing the vascular ring. (c) 3D vascular ring (RAA right aortic arch, LSCA left subclavian artery, MPA
volume-­rendered reconstruction viewed from an anterior perspective. main pulmonary artery, T trachea, E esophagus)
The arrow highlights that there is an “elbow” and distinct change in

display a net physiology of a large left to right shunt, with drainage seen with the entity called “Scimitar syndrome.”
right ventricular enlargement, tachypnea, and poor weight The physiology of PAPVC is similar to an atrial septal
gain. defect, with dilation of the right ventricle and pulmonary
There are numerous forms of PAPVC, with descriptions arteries noted on CT. The degree of left to right shunting is
encompassing pulmonary venous connections to nearly any determined by the proportion of the lung draining anoma-
systemic vein in the thorax. The most common types of lously. Patients with anomalous drainage of a single seg-
PAPVC are drainage of the left upper pulmonary vein to the ment of the lung are often asymptomatic, while patients
innominate vein, right upper pulmonary vein to the supe- with numerous lung segments draining anomalously will
rior vena cava, and right pulmonary veins draining into the have symptoms similar to unobstructed TAPVC. Obstruction
inferior vena cava (Fig. 46.21). The last type is the typical is rare with PAPVC.
594 D. Steflik and A. M. Hlavacek

Fig. 46.20  These images are

all from infants with totally
a b
anomalous pulmonary venous
connections. (a) Axial image
displaying a confluence (*)
posterior to the left atrium,
which should alert the imager
to the presence of totally
anomalous pulmonary venous
connection. Images (b–d) are
3D volume-rendered
reconstructions viewed from a
posterior perspective. (b)
Supracardiac connection, with
a vertical vein connecting the
pulmonary venous confluence
(*) to the brachiocephalic
vein. (c) Cardiac connection,
with the confluence (*)
draining into the coronary c d
sinus. This patient has mixed
connections, as the left upper
pulmonary vein drains into
the brachiocephalic vein. (d)
Infracardiac connection, with
a vertical vein connecting the
pulmonary venous confluence
(*) to the portal vein (LA left
atrium, VV vertical vein, CS
coronary sinus)

a b

Fig. 46.21 (a) Axial image, showing an anomalous connection of the (arrow) connecting to the superior vena cava (*) superior to the right
right upper pulmonary vein (arrow) to the superior vena cava (*). (b) pulmonary artery. The other pulmonary veins can be seen draining nor-
3D volume-rendered reconstruction in the same patient viewed from a mally into the left atrium. A discrete coarctation of the aorta can be seen
posterior perspective, displaying the right upper pulmonary vein as well, which was the primary indication for the scan
46  CT Spectrum of Congenital Heart Disease 595

a b

Fig. 46.22 (a) “Two-chamber view” in a patient with cor triatriatum. the small orifice in the membrane (arrow), through which all pulmonary
A membrane (arrows) can be seen above the left atrial appendage, dis- venous return must flow (LAA left atrial appendage)
tinguishing it from a supramitral ring. (b) “Short axis” image, showing

Cor Triatriatum After birth, ductal patency is required for systemic perfu-
sion in patients with HLHS, with aortic arch and coronary
Cor triatriatum is a rare defect, making up less than 0.5% of flow occurring in a retrograde fashion. An unrestrictive atrial
all congenital heart disease [53]. In the classic form, there is communication is essential as well, particularly in patients
a discrete membrane within the left atrium separating the with mitral atresia. Pulmonary venous return is thus diverted
pulmonary venous return from the remainder of the left to the right ventricle, which must then supply both systemic
atrium. In these patients, the only route of egress for and pulmonary flows. There are no curative surgical options
­pulmonary venous blood toward the mitral valve is an open- for HLHS, but there is a staged single-ventricle palliative
ing within the membrane. This membrane is always located approach that is discussed in Chap. 47 (“The Use of
in the area between the pulmonary venous ostia and left atrial Cardiovascular CT in Repaired CHD”).
appendage (Fig. 46.22). This location differentiates it from a
supramitral ring, which is located between the left atrial
appendage and mitral valve. A variant from classical cor tria-  ulmonary Atresia with Intact Ventricular
P
triatum is called “partial cor triatriatum,” in which some, but Septum
not all, of the pulmonary veins are separated from the left
atrium by the membrane. The variable anatomy in pulmonary atresia with an intact
ventricular septum includes the size of the right ventricular
cavity, tricuspid valve, and coronary arteries. The right ven-
Hypoplastic Left Heart Syndrome tricular size correlates with the tricuspid valve annular size.
The tricuspid valve is frequently abnormal, varying from
Hypoplastic left heart syndrome (HLHS) includes a spec- small and stenotic to large and Ebstein’s-like. The tricuspid
trum of left ventricular hypoplasia that results in a small annulus size also correlates with the presence of coronary
left ventricle that is unable to support a full cardiac out- fistulae connecting to the right ventricle [56] (Fig.  46.24).
put. The most common type is a combination of mitral These fistulae can often be identified on CT and are a major
atresia and aortic atresia, followed by mitral and aortic prognostic factor in survival [57]. Some patients have a con-
stenosis [54]. In patients with aortic atresia, there remains dition termed “right ventricular-dependent coronary circula-
a small ascending aorta which must remain patent in order tion,” in which stenosis or atresia in the proximal coronary
for retrograde filling of the coronary arteries (Fig. 46.23). arteries results in coronary flow that is dependent on flow
While diffuse aortic hypoplasia is very common in from these fistulae. The pulmonary atresia is often complete
patients with HLHS, a discrete coarctation is present in muscular atresia of the right ventricular outflow tract, but a
80% of patients [55]. Special attention should be given in third of patients have a patent right ventricular outflow tract
evaluating the atrial septum, tricuspid valve, pulmonary with membranous atresia of the pulmonary valve leaflets
veins, and coronary arteries, as HLHS is associated with [56]. Regardless of the type of pulmonary atresia, there is
abnormalities in each. usually an intact main pulmonary artery segment.
596 D. Steflik and A. M. Hlavacek

a b

c d

Fig. 46.23 (a) “Four-chamber view” in an infant with mitral and aortic chamber view” in another infant with a less severe form of hypoplastic
atresia, resulting in hypoplasic left heart syndrome. The atretic mitral left heart, with mitral and aortic stenosis. The mitral valve is patent,
valve is seen (white arrow), with severe hypoplasia of the left ventricle. resulting in a mildly hypoplastic left ventricle. Contrast can be seen
An atrial septal defect is seen, which is the only route of egress from the streaming left to right through an atrial septal defect. (d) Maximum
left atrium. The right atrium and ventricle are dilated. (b) 3D volume-­ intensity projection in a “candy cane” view from the same patient, dis-
rendered reconstruction in the same patient viewed from a rightward playing a hypoplastic aortic arch with a coarctation. There is a patent
perspective. Aortic blood flow is supplied via a patent ductus arteriosus. ductus arteriosus, augmenting flow to the descending aorta (RA right
Note the severely hypoplastic ascending aorta (black arrow), which is atrium, RV right ventricle, LA left atrium, LV left ventricle, Ao aorta,
filled retrograde to supply blood flow to the coronary arteries. (c) “Four-­ PDA patent ductus arteriosus)

Fig. 46.24 (a) “Four-­

chamber view” in an infant
a b
with pulmonary atresia with
intact ventricular septum. The
tricuspid valve is patent, but
the lack of egress results in
the typical appearance of a
severely hypoplastic and
hypertensive right ventricle
(RV). (b) 3D volume-­
rendered reconstruction in
another patient with this
diagnosis, revealing a dilated
single coronary artery (arrow)
with a large fistulous
connection to the right
ventricle (*), which is
common in this entity
46  CT Spectrum of Congenital Heart Disease 597

ventricle is dependent upon the size of the VSD, which is

its only source of inflow. There is variability in the orienta-
tion of the ventricles and arterial arrangement. The most
common form of DILV has transposition of the great arter-
ies with L-looped ventricles (morphologic right ventricle
on the left) and a leftward and anterior aorta. The second
most common form has transposition of the great arteries
with D-looped ventricles (morphologic right ventricle on
the right) and a rightward and anterior aorta. Only 15% of
patients with DILV have normally related great arteries, an
anomaly also known as “Holmes heart.” Those with trans-
position are prone to also have aortic arch hypoplasia,
while those with normally related great arteries generally
display pulmonary stenosis [16]. All patients with DILV
are ultimately managed with staged single-ventricle pallia-
tion (see Chap. 45).

Systemic Venous Abnormalities

Fig. 46.25  “Four-chamber view” in a patient with tricuspid atresia.
The two most common systemic venous abnormalities are
Note the plate-like atresia (arrows), which is the typical appearance in
this lesion and represents the absence of a connection between the right persistent left superior vena cava (LSVC) and interrupted
atrium and ventricle. This patient has an intact ventricular septum, so no inferior vena cava. Neither abnormality results in symptoms,
contrast is seen in the right ventricle but their identification is important prior to interventional
catheterizations or surgical procedures requiring cardiopul-
monary bypass.
Tricuspid Atresia The incidence of LSVC is approximately 0.3% and is
much greater in those with congenital heart disease [59]. It
In patients with tricuspid atresia, there is no connection can be identified on CT, starting at the confluence of the left
between the right atrium and right ventricle, with no identifi- jugular and subclavian veins. It is variable in size and often
able tricuspid valve tissue (Fig. 46.25). This contrasts with empties into an enlarged coronary sinus (Fig.  46.27).
the mitral atresia that is seen with HLHS, in which there is a Occasionally, the coronary sinus is “unroofed,” and the
connection between the left atrium and left ventricle, but an LSVC is seen emptying directly into the left atrium. There is
atretic valve can generally be identified. The right ventricle is occasionally an innominate vein that serves as a communica-
generally hypoplastic, dependent on the presence and size of tion between the LSVC and RSVC.
a ventricular septal defect. If the ventricular septum is intact, An interrupted IVC carries an incidence of about 0.2%
there is usually pulmonary atresia [16]. If a VSD is present, and has a strong association with left atrial isomerism [60]. It
the right ventricular outflow is generally patent. Tricuspid results from an absence of a hepatic segment of the IVC. The
atresia is accompanied by concordant ventriculoarterial con- infra-hepatic segment of the IVC is routed via a dilated azy-
nections in 80% of patients, but transposed and double outlet gous system, which then empties into the atrium by way of
ventricular connections have been described [58]. Patients the superior vena cava. While identification of the hepatic
with tricuspid atresia who have transposed great arteries will portion of the IVC by CT is dependent upon the site of con-
often have aortic arch hypoplasia or coarctation. All patients trast injection, scan delay, and length of scan field, a dilated
with tricuspid atresia are ultimately managed with staged azygous can often be identified lateral to the descending
single-ventricle palliation (see Chap. 45). aorta regardless of the scan protocol (Fig. 46.28).

Double Inlet Left Ventricle

Heterotaxy Syndromes
Double inlet left ventricle (DILV) is comprised of both AV
valves emptying into the morphologic left ventricle Heterotaxy syndromes highlight the importance of the seg-
(Fig. 46.26). A diminutive right ventricle is always present mental approach to describing cardiac and extracardiac
but exists as simply an outlet chamber. The size of the right structures. Heterotaxy is generally described as an abnor-
598 D. Steflik and A. M. Hlavacek

a b

Fig. 46.26 (a) “Four-chamber view” in a patient with double inlet left ventricle that supplies the aorta (Ao). This segmental arrangement is
ventricle. Both atrioventricular valves (arrows) empty into a morpho- common with this entity, and the size of the ventricular septal defect
logic left ventricle. (b) “Short axis view” in the same patient, showing must be described in order to determine the adequacy for providing
a ventricular septal defect (*) leading to a left-sided morphologic right systemic blood flow

Fig. 46.27 (a) Maximum

a b
intensity projection in a
coronal plane. Contrast has
been injected in the left arm,
displaying a persistent left
superior vena cava (arrow)
with connection to the
coronary sinus. A bridging
innominate vein is absent. (b)
Axial image in another patient
with a persistent left superior
vena cava. Dilation of the
coronary sinus (*) provides a
clue to investigate for this
entity

Fig. 46.28 (a) Maximum

a b
intensity projection in a
sagittal plane. In this patient
with interruption of the
inferior vena cava, contrast
was injected in the foot,
opacifying a dilated azygous
vein (arrow). (b) Axial image.
Although contrast was
injected via an upper
extremity, this entity can be
identified by visualizing the
dilated azygous vein lateral to
the descending aorta (arrow)
46  CT Spectrum of Congenital Heart Disease 599

Fig. 46.29 (a) “Four-­

chamber view” in a patient
a b
with visceral heterotaxy and
an unbalanced atrioventricular
septal defect. Two broad-­
based atrial appendages are
seen (arrows), consistent with
right atrial isomerism. (b)
“Four-chamber view” in
another patient with visceral
heterotaxy and an unbalanced
atrioventricular septal defect.
In this patient, there are two
narrow atrial appendages (*),
consistent with left atrial
isomerism

mal arrangement of thoracic and abdominal viscera that is 7. Kulkarni A, Hsu HH, Ou P, Kutty S.  Computed tomography in
neither situs solitus (normal) nor situs inversus (mirror congenital heart disease: clinical applications and technical con-
siderations. Echocardiography. 2016;33(4):629–40. https://doi.
image). The complex of anomalies can be grouped into org/10.1111/echo.13147.
two categories, right atrial isomerism and left atrial isom- 8. Han BK, Rigsby CK, Hlavacek A, Leipsic J, Nicol ED, Siegel MJ,
erism, based on the appearance of the atrial appendages Bardo D, Abbara S, Ghoshhajra B, Lesser JR, Raman S, Crean
(Fig.  46.29). The former is also known as “asplenia syn- AM.  Computed tomography imaging in patients with congeni-
tal heart disease part I: rationale and utility. An expert consensus
drome” and the latter known as “polysplenia syndrome,” document of the Society of Cardiovascular Computed Tomography
although there is poor correlation between splenic anat- (SCCT): endorsed by the Society of Pediatric Radiology (SPR)
omy and cardiac findings. and the North American Society of Cardiac Imaging (NASCI).
Right atrial isomerism is associated with TAPVC, J Cardiovasc Comput Tomogr. 2015;9(6):475–92. https://doi.
org/10.1016/j.jcct.2015.07.004.
absent coronary sinus, common atrium, pulmonary atresia, 9. Han BK, Lesser AM, Vezmar M, Rosenthal K, Rutten-Ramos S,
DORV, and single-ventricle physiology [61]. Left atrial Lindberg J, Caye D, Lesser JR. Cardiovascular imaging trends in
isomerism is associated with an interrupted IVC, congeni- congenital heart disease: a single center experience. J Cardiovasc
tal heart block, and LVOT obstruction. Bilateral superior Comput Tomogr. 2013;7(6):361–6. https://doi.org/10.1016/j.
jcct.2013.11.002.
vena cavae and AVSD are common in both types of isom- 10. Goo HW, Park IS, Ko JK, Kim YH, Seo DM, Park JJ. Computed
erism. The physiology is dictated by the constellation of tomography for the diagnosis of congenital heart disease in pediat-
findings. ric and adult patients. Int J Cardiovasc Imaging. 2005;21(2–3):347–
65; discussion 367. https://doi.org/10.1007/s10554-004-4015-0.
11. Leschka S, Oechslin E, Husmann L, Desbiolles L, Marincek B,
Genoni M, Pretre R, Jenni R, Wildermuth S, Alkadhi H. Pre- and
References postoperative evaluation of congenital heart disease in children
and adults with 64-section CT. Radiographics. 2007;27(3):829–46.
1. Gatzoulis MA, Webb GD. Adults with congenital heart diseases: a https://doi.org/10.1148/rg.273065713.
growing population. Edinburgh: Churchill Livingstone; 2003. 12. Girshin M, Shapiro V, Rhee A, Ginsberg S, Inchiosa MA Jr. Increased
2. Kaemmerer H, Stern H, Fratz S, Prokop M, Schwaiger M, Hess risk of general anesthesia for high-risk patients undergoing magnetic
J.  Imaging in adults with congenital cardiac disease (ACCD). resonance imaging. J Comput Assist Tomogr. 2009;33(2):312–5.
Thorac Cardiovasc Surg. 2000;48(6):328–35. https://doi. https://doi.org/10.1097/RCT.0b013e31818474b8.
org/10.1055/s-2000-8344. 13. Wolla CD, Hlavacek AM, Schoepf UJ, Bucher AM, Chowdhury
3. Boxt LM.  Magnetic resonance and computed tomographic S.  Cardiovascular manifestations of heterotaxy and related situs
evaluation of congenital heart disease. J Magn Reson Imaging. abnormalities assessed with CT angiography. J Cardiovasc
2004;19(6):827–47. https://doi.org/10.1002/jmri.20077. Comput Tomogr. 2013;7(6):408–16. https://doi.org/10.1016/j.
4. de Roos A, Roest AA.  Evaluation of congenital heart disease by jcct.2013.11.008.
magnetic resonance imaging. Eur Radiol. 2000;10(1):2–6. https:// 14. Loomba RS, Hlavacek AM, Spicer DE, Anderson RH. Isomerism
doi.org/10.1007/s003300050002. or heterotaxy: which term leads to better understanding?
5. Gutierrez FR, Siegel MJ, Fallah JH, Poustchi-Amin M. Magnetic Cardiol Young. 2015;25(6):1037–43. https://doi.org/10.1017/
resonance imaging of cyanotic and noncyanotic congenital heart s1047951115001122.
disease. Magn Reson Imaging Clin N Am. 2002;10(2):209–35. 15. Hoffman JI, Kaplan S. The incidence of congenital heart disease. J
6. Haramati LB, Glickstein JS, Issenberg HJ, Haramati N, Crooke Am Coll Cardiol. 2002;39(12):1890–900.
GA. MR imaging and CT of vascular anomalies and connections 16. Allen HD, Driscoll DJ, Shaddy RE, Feltes TF. Moss and Adams’
in patients with congenital heart disease: significance in surgical heart disease in infants, children, and adolescents: including the
planning. Radiographics. 2002;22(2):337–47; discussion 348–339. fetus and young adult, vol. 1. 8th ed. Philadelphia: Lippincott
https://doi.org/10.1148/radiographics.22.2.g02mr09337. Williams & Wilkins; 2013.
600 D. Steflik and A. M. Hlavacek

17. Anderson RH, Spicer DE, Hlavacek AM, Cook AC, Backer
of experience. Eur J Pediatr. 2012;171(3):579–86. https://doi.
CL.  Wilcox’s surgical anatomy of the heart. 4th ed. Cambridge, org/10.1007/s00431-011-1621-4.
UK: Cambridge University Press; 2013. 33. Meinel FG, Huda W, Schoepf UJ, Rao AG, Cho YJ, Baker GH,
18. Osawa K, Miyoshi T, Morimitsu Y, Akagi T, Oe H, Nakagawa K, Hlavacek AM. Diagnostic accuracy of CT angiography in infants
Takaya Y, Kijima Y, Sato S, Kanazawa S, Ito H.  Comprehensive with tetralogy of Fallot with pulmonary atresia and major aor-
assessment of morphology and severity of atrial septal defects in topulmonary collateral arteries. J Cardiovasc Comput Tomogr.
adults by CT.  J Cardiovasc Comput Tomogr. 2015;9(4):354–61. 2013;7(6):367–75. https://doi.org/10.1016/j.jcct.2013.11.001.
https://doi.org/10.1016/j.jcct.2015.04.007. 34. Braunwald E, Goldblatt A, Aygen MM, Rockoff SD, Morrow

19. Ko SF, Liang CD, Yip HK, Huang CC, Ng SH, Huang CF, Chen AG. Congenital aortic stenosis. I. Clinical and hemodynamic find-
MC.  Amplatzer septal occluder closure of atrial septal defect: ings in 100 patients. II. Surgical treatment and the results of opera-
evaluation of transthoracic echocardiography, cardiac CT, and tion. Circulation. 1963;27:426–62.
transesophageal echocardiography. AJR Am J Roentgenol. 35. Xu J, Zhao H, Wang X, Bai Y, Liu L, Liu Y, Wei M, Li J, Zheng
2009;193(6):1522–9. https://doi.org/10.2214/AJR.09.2854. M.  Accuracy, image quality, and radiation dose of prospectively
20. Lee T, Tsai IC, Fu YC, Jan SL, Wang CC, Chang Y, Chen MC. MDCT ECG-triggered high-pitch dual-source CT angiography in infants
evaluation after closure of atrial septal defect with an Amplatzer and children with complex coarctation of the aorta. Acad Radiol.
septal occluder. AJR Am J Roentgenol. 2007;188(5):W431–9. 2014;21(10):1248–54. https://doi.org/10.1016/j.acra.2014.04.019.
https://doi.org/10.2214/ajr.06.0709. 36. Budoff MJ, Shittu A, Roy S.  Use of cardiovascular computed

21. Marini D, Ou P, Boudjemline Y, Kenny D, Bonnet D, Agnoletti tomography in the diagnosis and management of coarctation of the
G.  Midterm results of percutaneous closure of very large atrial aorta. J Thorac Cardiovasc Surg. 2013;146(1):229–32. https://doi.
septal defects in children: role of multislice computed tomography. org/10.1016/j.jtcvs.2013.01.024.
EuroIntervention. 2012;7(12):1428–34. https://doi.org/10.4244/ 37. Backer CL, Mavroudis C. Congenital Heart Surgery Nomenclature
eijv7i12a223. and Database Project: patent ductus arteriosus, coarctation of
22. Amat F, Le Bret E, Sigal-Cinqualbre A, Coblence M, Lambert V, the aorta, interrupted aortic arch. Ann Thorac Surg. 2000;69(4
Rohnean A, Paul JF.  Diagnostic accuracy of multidetector spiral Suppl):S298–307.
computed tomography for preoperative assessment of sinus venosus 38. Braunlin E, Peoples WM, Freedom RM, Fyler DC, Goldblatt A,
atrial septal defects in children. Interact Cardiovasc Thorac Surg. Edwards JE. Interruption of the aortic arch with aorticopulmonary
2011;12(2):179–82. https://doi.org/10.1510/icvts.2010.251298. septal defect. An anatomic review. Pediatr Cardiol. 1982;3(4):329–
23. Lembcke A, Razek V, Kivelitz D, Rogalla N, Rogalla P. Sinus veno- 35. https://doi.org/10.1007/bf02427036.
sus atrial septal defect with partial anomalous pulmonary venous 39. Yang DH, Goo HW, Seo DM, Yun TJ, Park JJ, Park IS, Ko JK,
return: diagnosis with 64-slice spiral computed tomography at low Kim YH.  Multislice CT angiography of interrupted aortic arch.
radiation dose. J Pediatr Surg. 2008;43(2):410–1. Pediatr Radiol. 2008;38(1):89–100. https://doi.org/10.1007/
24. Yamasaki Y, Nagao M, Kawanami S, Kamitani T, Sagiyama K, s00247-007-0662-3.
Yamanouchi T, Sakamoto I, Yamamura K, Yabuuchi H, Honda 40.
Schreiber C, Cook A, Ho SY, Augustin N, Anderson
H. One-stop shop assessment for atrial septal defect closure using RH.  Morphologic spectrum of Ebstein’s malformation: revisi-
256-slice coronary CT angiography. Eur Radiol. 2017;27:697. tation relative to surgical repair. J Thorac Cardiovasc Surg.
https://doi.org/10.1007/s00330-016-4407-2. 1999;117(1):148–55.
25. Reller MD, Strickland MJ, Riehle-Colarusso T, Mahle WT,
41. Attenhofer Jost CH, Connolly HM, Dearani JA, Edwards WD,
Correa A.  Prevalence of congenital heart defects in metropolitan Danielson GK. Ebstein’s anomaly. Circulation. 2007;115(2):277–
Atlanta, 1998–2005. J Pediatr. 2008;153(6):807–13. https://doi. 85. https://doi.org/10.1161/circulationaha.106.619338.
org/10.1016/j.jpeds.2008.05.059. 42. Allwork SP, Bentall HH, Becker AE, Cameron H, Gerlis LM,
26. Jacobs JP, Burke RP, Quintessenza JA, Mavroudis C.  Congenital Wilkinson JL, Anderson RH. Congenitally corrected transposition
Heart Surgery Nomenclature and Database Project: ventricular sep- of the great arteries: morphologic study of 32 cases. Am J Cardiol.
tal defect. Ann Thorac Surg. 2000;69(4 Suppl):S25–35. 1976;38(7):910–23.
27. Rudolph AM, Drorbaugh JE, Auld PA, Rudolph AJ, Nadas AS, 43. Van Praagh R, Van Praagh S.  The anatomy of common aortico-
Smith CA, Hubbell JP.  Studies on the circulation in the neona- pulmonary trunk (truncus arteriosus communis) and its embryo-
tal period. The circulation in the respiratory distress syndrome. logic implications. A study of 57 necropsy cases. Am J Cardiol.
Pediatrics. 1961;27:551–66. 1965;16(3):406–25.
28. Mitchell SC, Korones SB, Berendes HW.  Congenital heart dis- 44. Collett RW, Edwards JE.  Persistent truncus arteriosus; a clas-

ease in 56,109 births. Incidence and natural history. Circulation. sification according to anatomic types. Surg Clin North Am.
1971;43(3):323–32. 1949;29(4):1245–70.
29. Kitterman JA, Edmunds LH Jr, Gregory GA, Heymann MA,
45. Konstantinov IE, Karamlou T, Blackstone EH, Mosca RS, Lofland
Tooley WH, Rudolph AM.  Patent ducts arteriosus in premature GK, Caldarone CA, Williams WG, Mackie AS, McCrindle
infants. Incidence, relation to pulmonary disease and manage- BW.  Truncus arteriosus associated with interrupted aortic arch
ment. N Engl J Med. 1972;287(10):473–7. https://doi.org/10.1056/ in 50 neonates: a Congenital Heart Surgeons Society study. Ann
nejm197209072871001. Thorac Surg. 2006;81(1):214–22. https://doi.org/10.1016/j.
30.
Schneider DJ, Moore JW.  Patent ductus arteriosus. athoracsur.2005.06.072.
Circulation. 2006;114(17):1873–82. https://doi.org/10.1161/ 46. Keane JF, Lock JE, Fyler DC. Nadas’ pediatric cardiology. 2nd ed.
circulationaha.105.592063. Philadelphia: Saunders; 2006.
31. Sardesai SH, Marshall RJ, Farrow R, Mourant AJ. Dissecting aneu- 47. Ramos-Duran L, Nance JW Jr, Schoepf UJ, Henzler T, Apfaltrer
rysm of the pulmonary artery in a case of unoperated patent ductus P, Hlavacek AM.  Developmental aortic arch anomalies in infants
arteriosus. Eur Heart J. 1990;11(7):670–3. and children assessed with CT angiography. AJR Am J Roentgenol.
32. Lin MT, Wang JK, Chen YS, Lee WJ, Chiu HH, Chen CA, Chiu 2012;198(5):W466–74. https://doi.org/10.2214/AJR.11.6982.
SN, Wu ET, Lu CW, Huang SC, Chen SJ, Chiu IS, Chang CI, Wu 48. Leonardi B, Secinaro A, Cutrera R, Albanese S, Trozzi M,

MH.  Detection of pulmonary arterial morphology in tetralogy of Franceschini A, Silvestri V, Toma P, Carotti A, Pongiglione
Fallot with pulmonary atresia by computed tomography: 12 years G.  Imaging modalities in children with vascular ring and pulmo-
46  CT Spectrum of Congenital Heart Disease 601

nary artery sling. Pediatr Pulmonol. 2015;50(8):781–8. https://doi. drome: lessons learned from 115 consecutive patients. Circulation.
org/10.1002/ppul.23075. 2002;106(12 Suppl 1):I82–9.
49. Lee EY, Zurakowski D, Waltz DA, Mason KP, Riaz F, Ralph A, 55. Hawkins JA, Doty DB. Aortic atresia: morphologic characteristics
Boiselle PM.  MDCT evaluation of the prevalence of tracheoma- affecting survival and operative palliation. J Thorac Cardiovasc
lacia in children with mediastinal aortic vascular anomalies. J Surg. 1984;88(4):620–6.
Thorac Imaging. 2008;23(4):258–65. https://doi.org/10.1097/ 56. Hanley FL, Sade RM, Blackstone EH, Kirklin JW, Freedom RM,
RTI.0b013e31817fbdf7. Nanda NC.  Outcomes in neonatal pulmonary atresia with intact
50. Dyer KT, Hlavacek AM, Meinel FG, De Cecco CN, McQuiston ventricular septum. A multiinstitutional study. J Thorac Cardiovasc
AD, Schoepf UJ, Pietris NP.  Imaging in congenital pulmonary Surg. 1993;105(3):406–23, 424–407; discussion 423–404.
vein anomalies: the role of computed tomography. Pediatr Radiol. 57. Seguela PE, Houyel L, Loget P, Piot JD, Paul JF. Critical stenosis
2014;44(9):1158–68; quiz 1155–7. https://doi.org/10.1007/ of a right ventricle to coronary artery fistula seen at dual-source
s00247-014-3044-7. CT in a newborn with pulmonary atresia and intact ventricular sep-
51. Bonelli-Sica JM, de la Mora-Cervantes R, Diaz-Zamudio M,
tum. Pediatr Radiol. 2011;41(8):1069–72. https://doi.org/10.1007/
Castillo-Castellon F, Ramirez-Carmona R, Velazquez-Moreno s00247-011-2068-5.
J, Kimura-Hayama E.  Dual-source 256-MDCT for diagnosis of 58. Tandon R, Edwards JE. Tricuspid atresia. A re-evaluation and clas-
anomalous pulmonary venous drainage in pediatric population. AJR sification. J Thorac Cardiovasc Surg. 1974;67(4):530–42.
Am J Roentgenol. 2013;200(2):W163–9. https://doi.org/10.2214/ 59.
Sanders JM.  Bilateral superior vena cavae. Anat Rec.
ajr.11.8348. 1946;94:657–62.
52. Jonas RA, Smolinsky A, Mayer JE, Castaneda AR. Obstructed pul- 60. Mayo J, Gray R, St Louis E, Grosman H, McLoughlin M, Wise
monary venous drainage with total anomalous pulmonary venous D.  Anomalies of the inferior vena cava. AJR Am J Roentgenol.
connection to the coronary sinus. Am J Cardiol. 1987;59(5):431–5. 1983;140(2):339–45. https://doi.org/10.2214/ajr.140.2.339.
53. Krasemann Z, Scheld HH, Tjan TD, Krasemann T.  Cor tria-
61. Rubino M, Van Praagh S, Kadoba K, Pessotto R, Van Praagh
triatum: short review of the literature upon ten new cases. Herz. R.  Systemic and pulmonary venous connections in vis-
2007;32(6):506–10. https://doi.org/10.1007/s00059-007-2882-6. ceral heterotaxy with asplenia. Diagnostic and surgical con-
54. Tweddell JS, Hoffman GM, Mussatto KA, Fedderly RT, Berger S, siderations based on seventy-two autopsied cases. J Thorac
Jaquiss RD, Ghanayem NS, Frisbee SJ, Litwin SB. Improved sur- Cardiovasc Surg. 1995;110(3):641–50. https://doi.org/10.1016/
vival of patients undergoing palliation of hypoplastic left heart syn- s0022-5223(95)70095-1.
 The Use of Cardiovascular CT
in Repaired CHD 47
B. Kelly Han, Andrew Crean, and John R. Lesser

Advances in medical and surgical care of patients with con-  isk of Cardiovascular CT in Repaired
R
genital heart disease (CHD) have resulted in expected sur- and Palliated CHD
vival for even the most complex lesions [1]. There is an
increasing prevalence of CHD in all age groups, the majority The risks of cardiovascular CT include peripheral IV access,
of patients with CHD are now adults, and the number reach- exposure to iodinated contrast in almost all patients, limited
ing older adulthood is rapidly rising [2–4]. Patients with anesthesia in patients for whom a breath hold is required
repaired or palliated CHD require serial diagnostic evalua- when unable to cooperate due to hemodynamic or develop-
tions throughout their lives [5]. The improved spatial and mental status, and radiation exposure. Anesthesia risk
temporal resolution, rapid image acquisition, and radiation includes both the procedural risk of adverse event and the
dose reduction of newer generation scanners have dramati- risk for long-term adverse neurodevelopmental outcome for
cally increased the applicability of computed tomography repeated exposure at young ages [10–14]. Children with con-
(CT) to patients with congenital heart disease of all ages [6]. genital heart disease, particularly hospitalized patients or
Cardiovascular CT is increasingly used in patients with CHD those with unrepaired cyanotic heart disease, are at the high-
when echocardiography is not sufficient and cardiac mag- est risk for an adverse event with anesthesia [15]. When
netic resonance imaging (CMR) is contraindicated, unlikely anesthesia is needed for breath holding, it is usually limited
to provide adequate image quality due to artifact, or consid- to a single imaging sequence (high heart rate coronary imag-
ered high risk due to scan time or anesthesia when needed [7, ing or functional imaging acquired over several heartbeats)
8]. Cardiovascular CT in congenital heart disease will be and is of short duration.
optimally used in settings where all advanced diagnostic Cumulative radiation exposure is of particular concern for
modalities are available so that the test with the best informa- congenital heart disease patients requiring repeated diagnos-
tion and least risk can be chosen for each individual patient tic testing and intervention. These CHD patients may have a
and clinical indication. relatively high cumulative radiation exposure, where approx-
CT is an important diagnostic modality for select patients imately 75% is catheterization based in the current era [16].
with CHD after intervention, from the neonate to the adult Studies have shown that cardiovascular CT has lower radia-
patient. Evaluation of CHD is considered an appropriate tion exposure than catheterization when modern CT scanner
indication on the most recent consensus document on the use technology is available [17]. The reported radiation dose
of cardiovascular CT [9]. Table 47.1 lists consensus recom- from cardiac CT in patients with CHD varies from less than
mendations for the use CT in CHD [8]. 1 millisievert to 18 mSv [18–22]. Recent trends show many
CHD scans can be performed at a dose of approximately
1  mSv including anatomic evaluation and ECG-triggered
B. K. Han (*) exams for coronary evaluation [20, 21, 23]. The image qual-
Department of Advanced Congenital Cardiac Imaging, ity required for high-resolution coronary imaging is rarely
Minneapolis Heart Institute and the Children’s Hospitals and
needed for CHD CT studies. CT in CHD is optimally per-
Clinics of Minnesota, Minneapolis, MN, USA
e-mail: [email protected] formed in centers with the appropriate attention to CT dose
reduction and modern CT equipment. The scan planning is
A. Crean
Department of Cardiology, Ottawa Heart Institute, University of individualized and includes decisions about aggressive radi-
Ottawa, Ottawa, ON, Canada ation dose reduction with direct physician input concerning
J. R. Lesser patient preparation, data acquisition as well as interpretation
Minneapolis Heart Institute Foundation, Minneapolis, MN, USA of each scan [22, 24].

© Humana Press 2019 603

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_47
604 B. K. Han et al.

Table 47.1  Situations in which cardiovascular CT may be appropriate • Transposition of the great arteries, s/p atrial and arterial
in repaired CHD switch
Presence of CMR unsafe implant or foreign body (retained pacing • Tetralogy of Fallot, s/p palliation or complete repair
leads, non-MR compatible pacemaker/defibrillator, neurostimulator) • Univentricular heart disease through all stages of
Poor CMR image quality (known or expected) due to metallic
palliation
artifact
Unable to fit in MRI scanner due to obesity or severe claustrophobia
Neonate or young patient requiring evaluation of postoperative
complex anatomy, particularly if considered higher risk for adverse  iagnosis-Specific Preparation and CT Scan
D
event with sedation or anesthesia required for CMR, and the CT Protocols
scan can be performed with no or limited sedation
Critically ill patient with repaired CHD of any age that may not
tolerate breath holding or length of CMR scan
 oronary Artery Imaging After Intervention
C
Evaluation of ventricular assist device or ECMO cannula in CHD
positioning
Patient requiring CT for evaluation of extracardiac anatomy in Coronary imaging in patients with repaired CHD is recom-
addition to palliated or repaired CHD (e.g., lung parenchyma, mended for definition of coronary artery anatomy after surgi-
airway, skeletal abnormality)
cal manipulation, prior to repeat intervention on the RVOT,
Preoperative patients with prior sternotomy considered high risk for
vascular injury with sternal reentry due to an anterior coronary to assess the relationship to the RVOT and sternum, and for
artery, conduit, or sternal adhesions assessment of coronary lesions in symptomatic patients and
Evaluation of prosthetic valve function or structural integrity in certain adult patient undergoing repeat cardiac interven-
(calcification, stenosis, coaptation defect, leaflet immobility, tion to determine the need for concurrent coronary interven-
paravalvular leak, endocarditis, or clot)
tion [5] (Figs. 47.1 and 47.2). Coronary imaging is often one
Evaluation of calcification within vessels and surgical conduits prior
to catheter-based intervention (e.g., balloon angioplasty, aspect of a complete evaluation of congenital heart disease,
transcatheter valve replacement, stent placement) and the scan preparation will include consideration of all
Coronary artery imaging in CHD aspects of the patients CHD anatomy requiring evaluation
 (a) Patient needing detailed preoperative coronary artery for clinical management.
evaluation in addition to assessment of complex cardiac
anatomy
 (b) Patient with symptoms and signs suggestive of atherosclerotic  efinition of Coronary Artery Subsequent
D
coronary artery disease and a history of CHD, prior coronary to Coronary Intervention
intervention, or high-risk Kawasaki disease
 (c) Young symptomatic patients with known or suspected
Patients who have undergone surgical coronary intervention/
coronary anomaly, particularly if CMR is unlikely to provide manipulation in repair of a congenital heart defect include
complete assessment or more likely to require anesthesia patients who have undergone an arterial switch operation,
 (d) Delineation of coronary anatomy prior to surgical or the Nikaidoh or Ross procedure, and unroofing or reimplan-
percutaneous pulmonary valve implantation
 (e) Evaluation of coronary artery after any surgery requiring
tation of an anomalous origin of the coronary artery from the
coronary artery manipulation or reimplantation pulmonary artery or aortic root (ALCAPA or AAOCA). (See
Modified from Han et al. [8] section on TGA for coronary evaluation after intervention).

 he Use of Cardiovascular CT in Repaired

T  efinition of Coronary Artery Anatomy Prior
D
and Palliated CHD to Right Ventricular Outflow Tract (RVOT)
Intervention
Patients with repaired or palliated CHD referred for cardio- Coronary imaging is required prior to repeat RVOT inter-
vascular CT will have prior diagnostic studies and interven- vention for definition of the coronary artery relationship to
tions. The cardiac imager must know the underlying CHD both the sternum and the right ventricular outflow tract.
diagnosis, prior interventions and diagnostic studies, and the RVOT replacement is one of the most common repeat sur-
clinical question so that a targeted exam may be performed. gical procedures performed in patients of all ages with con-
Optimal CT imaging in CHD requires close collaboration genital heart disease, required in certain forms of tetralogy
between CT imagers, cardiologists, and surgeons. of Fallot, truncus arteriosus, the Rastelli procedure, Ross
The most common indications for cardiovascular CT after procedure, and in some patients after the arterial switch
intervention in CHD are: procedure [5, 25]. Coronary location can determine the
optimal sternal reentry and location of pulmonary conduit
• Coronary artery evaluation after surgical manipulation or in the setting of a coronary artery proximal to the RVOT or
reimplantation existing conduit. Transcatheter valves are currently used in
• Aortic arch evaluation after surgical or catheter-based existing conduits primarily for conduit stenosis, with pre-
intervention stenting in many cases to create a circumferential landing
47  The Use of Cardiovascular CT in Repaired CHD 605

Fig. 47.1 Coronary
relationship to the sternum
a b
and RVOT after arterial
switch or Rastelli procedure.
(a) Axial 2D image of the left
coronary artery as it arises
from the right-facing sinus in
this patient after Rastelli
procedure. The coronary
artery (arrow) runs between
the neo-aorta and the calcified
pulmonary homograft,
concerning for potential
coronary compression with
transcatheter valve placement.
(b and c) This coronal 3D
reconstruction (b) shows a
single coronary artery arising
c
from the right coronary sinus
of the neo-aortic root after
arterial switch procedure,
coursing anterior to the RVOT
after LeCompte maneuver.
The 2D image (c) from the
same patient shows the right
coronary artery coursing
anterior to the RVOT and
directly posterior to the
sternum (arrow). Note the
dilated neo-aortic root
additionally

zone for the valve. In a multi-institution study, 17% of

patients undergoing transcatheter pulmonary valve place-
ment were found to have abnormal coronary artery pattern,
and 5% of patients had coronary compromise on test bal-
loon prior to valve insertion [26, 27]. Transcatheter valves
within a stent cause significant artifact in MRI, and CT
angiography can visualize valvular detail or endocarditis/
thrombus if suspected [28–30].

 igh-Resolution Coronary Imaging in Patients

H
with CHD
As patients with CHD enter middle and late adulthood, they
may acquire coronary artery disease and require coronary
intervention [31]. High-resolution coronary imaging should
be pursued in patients with repaired or palliated CHD and
symptoms suggestive of ischemia. It is also recommended
that male patients over the age of 35 and postmenopausal
Fig. 47.2  Left main ostial coronary artery stenosis after the arterial women undergo coronary evaluation prior to planned cardiac
switch operation. 2D image showing the left main coronary artery is intervention to determine the need for concurrent coronary
severely narrowed at the ostium after arterial switch procedure (arrow).
The right coronary ostium is widely patent (*) intervention [5, 32].
606 B. K. Han et al.

Potential anatomic findings after coronary artery nique. If a shunt lesion is suspected, then a contrast den-
intervention: sity gradient is required across the heart and is achieved
through judicious manipulation of contrast: saline ratio in
• Ostial or proximal coronary artery narrowing after arterial each phase of a dual-phase injection [7].
switch operation or the Ross procedure
• Right coronary artery lesion after Nikaidoh (due to left-
ward translocation of the aorta) Aortic Arch
• Anomalous coronary artery may run adjacent or over the
RVOT in patient requiring re-intervention Aortic arch imaging is one of the most frequent indications
• Atherosclerotic coronary lesions in adult CHD patients for CT imaging in patients with congenital heart disease who
have previously undergone intervention (Figs.  47.3 and
 ediatric and Adult Congenital Coronary
P 47.4). The most common prior interventions are repair of
CT Scan Recommendations aortic coarctation and repair of interrupted aortic arch.
• The coronary artery detail needed for clinical manage- Advanced imaging is recommended every 5 years according
ment in CHD will vary by indication: to the current ACHD clinical guidelines [5]. Long-term sur-
–– For course and location of the coronary arteries prior vival for patients with repaired aortic coarctation is decreased
to RVOT intervention, minimal acquisition window compared to the general population, and many patients
and aggressive dose reduction should be used. On require re-intervention [33]. Younger age at repair results in
highest pitch or volumetric scanners, coronary course better long-term survival but a higher rate of re-intervention.
can often be determined with a younger patient quietly Preoperative hypertension and age over 20 years at the time
breathing. of repair are independent risk factors for mortality [34].
–– For symptomatic patients and those requiring detailed Aortic coarctation is repaired utilizing end-to-end anastomo-
coronary imaging to rule out a coronary lesion, opti- sis, left subclavian flap repair, interposition graft, or catheter-­
mal image quality is necessary and may require medi- based balloon angioplasty and stenting [35]. Patients with
cation to reduce heart rate. Small children may need aortic coarctation may also have a bicuspid aortic valve or
intubation for a breath hold. abnormal mitral valve and other left-sided obstructive
• Contrast injection should be modified so that coronary lesions. CT imaging is particularly relevant to patients who
artery anatomy and congenital anatomy may be evaluated have previously undergone stent placement in the aorta [36].
in the same scan acquisition whenever possible. This may Aortic coarctation is one of the lesions that may present in
require a prolonged or dual-contrast-phase injection tech- adulthood in a patient imaged for other reasons.

a b c

Fig. 47.3  Recurrent coarctation after stenting for thoracic aortic shows a 2D cross section of the distal stent in the same area. (c) 3D
coarctation (mid-aortic syndrome). (a and b) There is recurrent aortic reconstruction of an interposition graft (arrow) from the proximal
coarctation (*) in the mid thoracic aorta proximal to multiple stents descending thoracic aorta to the abdominal aorta to bypass the
placed to treat coarctation/mid-aortic syndrome. There is intimal pro- obstruction in the same patient
liferation causing obstruction at the distal end of the stents. Image B
47  The Use of Cardiovascular CT in Repaired CHD 607

 ommon Anatomic Findings After Aortic

C Scan Modifications
Arch Intervention • Standard IV placement and timing of contrast to the aorta
• Recurrent aortic arch obstruction (in stent stenosis, homo- unless other lesions require evaluation.
graft stenosis, or arch obstruction adjacent to the site or • Extend cranial scan range slightly; anomalous coronary
prior repair) arteries may arise from the proximal ascending aorta.
• Aortic aneurysm or pseudoaneurysm • Aggressive dose reduction may be used in most cases.
• Aortic dissection Heart rate-lowering medications are only needed if coro-
nary imaging is required.
• If a bicuspid valve needs evaluation in addition to the
aorta, systolic image acquisition is best (Sievers
classification).

Transposition Complexes

Transposition complexes include a wide variety of anatomic

and hemodynamic lesions requiring different surgical
approaches for repair or palliation. For isolated ventricular-­
arterial discordance, surgical pathway is determined by the
relationship of the great arteries to the ventricular chambers
and the ability of the valves to be used in the alternate circu-
lation, particularly the pulmonary valve suitability for the
aortic position. For patients with both atrioventricular and
ventricular-arterial discordance, the surgical approach is
determined by associated lesions, function of the right ven-
Fig. 47.4  Ascending aorta pseudoaneurysm after prior aortic valvot- tricle and tricuspid valve in the systemic circulation, and
omy. Unexpected severe dilation of the ascending aorta (arrow, 6.5 cm) concerns regarding long-term outcomes for a systemic right
at the site of aortic entry for prior aortic valvotomy to treat aortic steno- ventricle. Some centers consider primary transplant as an
sis. Note calcified aortic valve leaflets. Scan was performed to evaluate
the proximal descending aorta in the area of prior coarctation repair, option when indicated rather than a double switch (atrial and
which was unremarkable arterial switch) procedure (Figs. 47.5 and 47.6).

Transposition of the Great Arteries

Surgical Approach

Simple transposition (d-TGA)d Complex transposition Corrected transposition (L-TGA)

Atrial switch Arterial switch Rastelli Nikaidohh

Single Conventional
ventricle Double switch
repair (fixed
pathway (atrial and
associated
arterial switch)
anomalies)

Fig. 47.5  Surgical treatment approaches to transposition of the great arteries

608 B. K. Han et al.

d-TGA s/p Atrial Switch ciated cardiac defects, RV dysfunction, and tricuspid
­regurgitation [37]. These patients have a high rate of sinus
Until the 1980s most patients with isolated ventricular-­ node dysfunction requiring pacemaker placement. An
arterial discordance (most commonly d-TGA) underwent an implantable defibrillator is considered for patients with low
atrial switch operation with the systemic and pulmonary systemic ejection fraction [38].
venous return baffled to the opposite atrium, essentially cre-
ating physiologically corrected transposition with a systemic Common Lesions After the Atrial Switch
right ventricle and a subpulmonary left ventricle (Figs. 47.7, Operation
47.8, 47.9, and 47.10). Predictors of mortality include asso- • Systemic and pulmonary venous baffle obstruction (most
common in the superior vena cava baffle to the left atrium)
• Systemic RV dysfunction
• Tricuspid (systemic AV valve) regurgitation
• Subpulmonary obstruction due to septal displacement
into the LVOT

Scan Modifications
• Biventricular injection protocol (triphasic, two-phase
contrast and saline) (Fig. 47.11).
• Lengthen monitoring sequence if high suspicion for baf-
fle obstruction and the heart will fill via collaterals.
• ECG-gated functional scan if quantification of function is
needed (use aggressive dose reduction since function
analysis performed in 6–8 mm slices).
• May quantify regurgitation by stroke volume differences
if closely correlated to echocardiography.

d-TGA s/p Arterial Switch

Fig. 47.6  Systemic RV failure in L-TGA.  This image shows an The earliest patients who underwent arterial switch are now
enlarged and hypertrophied systemic right ventricle in a patient with in young adulthood. In most cases the coronary arteries are
L-TGA (also called physiologically corrected transposition). Heart transferred to the neo-aortic root (Figs. 47.12 and 47.13). An
block is common, and both epicardial and transvenous pacer leads are
exception is in a patient after an Aubert procedure or the
seen (arrows). Severe systemic tricuspid regurgitation required pros-
thetic valve replacement Nikaidoh procedure. In the Aubert procedure, the coronary is

Fig. 47.7  d-TGA after atrial

a b
switch operation (Senning).
(a) A patent pulmonary
venous baffle (arrow) to the
hypertrophied and enlarged
systemic right ventricle in a
patient after the atrial switch
operation. (b) Note the pacer
lead into the apex of the
thin-walled subpulmonary left
ventricle
47  The Use of Cardiovascular CT in Repaired CHD 609

a b

Fig. 47.8  d-TGA with Mustard palliation, systemic venous baffle simply early time point of imaging post contrast injection before recir-
stenting. (a) Coronal view to show the inferior limb of the systemic culation of opacified blood through the IVC. (b) Volume-rendered
venous baffle (asterisk) which has been stented open. Note that the rela- image of the same patient to demonstrate the clarity with which cardiac
tively low attenuation inside the stent does not imply thrombosis but CT can assess and depict stented structures

Fig. 47.9  Multilevel baffle

a b
obstruction in an adult with
Mustard palliation of
d-transposition of the great
arteries. (a) Coronal view
demonstrating focal
narrowing at the distal end of
the superior limb of the
systemic venous baffle
(arrow). (b) Coronal view
demonstrating similar
narrowing of the distal
inferior limb of the systemic
venous baffle (arrow). (c)
Axial view in the same
patient. The left-sided
pulmonary veins are occluded
(not shown); the right-sided
veins (asterisk) drain through
a stenosed baffle (arrow) into
the left atrium

c
610 B. K. Han et al.

Fig. 47.10  Baffle obstruction

a b
with azygous system
off-loading. (a) Sagittal and
(b) coronal maximum
intensity projection
reconstructions from the same
patient as the preceding image
(Fig. 47.9). The azygous
(white arrows) and
hemiazygous (black arrows)
systems are significantly
dilated. In a patient with
Mustard palliation of d-TGA,
this implies obstruction of the
superior venous limb. Blood
from the head and neck passes
down the SVC but – upon
meeting resistance to forward
flow through the baffle –
passes retrograde along the
azygous arch (black asterisks)
and down into the azygous
system. This collateral
pathway allows blood from
the upper body to enter the
inferior limb of the systemic
venous baffle via the IVC
instead

baffled through the neo-pulmonary valve, and in the Nikaidoh

procedure, the aorta is translocated leftward and closer to the
VSD. After the Nikaidoh the right coronary artery is at risk
for mid-vessel narrowing. Overall, the hemodynamic results
and quality of life after the arterial switch operation have
been excellent [39]. The risks for coronary lesions include a
stormy postoperative course, single or intramural coronary
artery, or two coronary ostia arising close to each other.
There is a bimodal risk of coronary events, in the early post-
operative era and then again around 10 years postoperatively.
Eight to 10% of patients will have coronary stenosis or
occlusion after the arterial switch operation [39, 40].
Cardiovascular CT has complete agreement with invasive
angiography for assessment of coronary artery anatomy after
the arterial switch procedure. Lesions are ostial or in the
proximal coronary artery course [41]. Some consider coro-
nary CT preferable for this indication since the catheter can
alter the coronary course and mask ostial narrowing during
selective coronary angiography.
Fig. 47.11  Ventricular function assessment by cardiovascular CT:
This ECG-gated and pulse-modulated CT dataset is reconstructed in a
Valvular and supravalvular PS are common in d-TGA
diastolic interval in the short axis plane for quantification of biventricu- post arterial switch, as the branch pulmonary arteries are
lar size and systolic function. Note the enlarged RV and the pacer lead stretched anteriorly over the neo-aortic root to the neo-­
artifact (arrow). A biventricular (triphasic, two-phase contrast and pulmonary valve. Repeat intervention on the RVOT is the
saline) injection protocol was used to opacify both the right and left
ventricle at the time of image acquisition
most common intervention after the arterial switch ­procedure.
47  The Use of Cardiovascular CT in Repaired CHD 611

a b

Fig. 47.12  d-TGA following arterial switch procedure. (a) Axial MIP LeCompte maneuver; in this instance there is a focal stenosis on the
and (b) volume render of CT dataset in a patient who has had the arte- proximal left pulmonary artery (arrows). Stenoses of this severity usu-
rial switch procedure for a diagnosis of d-TGA. Focal areas of narrow- ally impact flow and will generally be stented – subsequent stent sur-
ing in the pulmonary arteries are not uncommon following the veillance is then almost always carried out by CT rather than CMR

• Ostial coronary artery lesions after arterial switch

procedure
• Mid RCA lesions after the Nikaidoh procedure

Scan Modification
• Consider medication to control heart rate if high-­
resolution coronary artery imaging needed (at least once
in adulthood and patients with symptoms of ischemia).
• Biventricular (triphasic) contrast injection protocol to
opacify both the right and left side of the heart.
• Extend scan range to include the proximal ascending
aorta and branch pulmonary arteries.

Tetralogy of Fallot

Tetralogy of Fallot includes the most common form with

pulmonary stenosis and the less common forms of absent
Fig. 47.13  3D volume-rendered image of a single right coronary
artery with the left anterior descending coronary artery coursing ante- pulmonary valve and pulmonary artery atresia with aortopul-
rior to the RVOT (arrow) after LeCompte maneuver. Note the stents in monary collaterals. Repair of all forms of tetralogy involves
the proximal branch pulmonary arteries bilaterally (*) closure of the VSD with establishment of unobstructed pul-
monary blood flow. Relief of pulmonary obstruction is
The relationship of the coronary artery to the RVOT is criti- achieved with pulmonary valvotomy, transannular patch, and
cal for catheter-based intervention, and the relationship to RV-PA conduit. Patients with absent pulmonary valve require
both the RVOT and the sternum is critical for those undergo- pulmonary artery plication, and those with aortopulmonary
ing surgical RVOT placement. collaterals will have aortopulmonary collaterals ligated or
closed via catheter intervention. Tetralogy with absent pul-
 ost Common Residual Hemodynamic Lesions
M monary valve may also have issues with compression of sur-
• RVOT obstruction at the level of the neo-pulmonary valve rounding structures by the very enlarged pulmonary arteries.
or branch pulmonary arteries Cardiac CT is ideal for identification of complications
• Neo-aortic stenosis or insufficiency including both airway and coronary artery compression.
• Narrowing of the ascending aorta at the neo-aortic root All patients require interval postoperative evaluation to
anastomosis monitor right ventricular size and function, residual/recur-
• Neo-aortic root dilation rent pulmonary artery or conduit stenosis and/or ­insufficiency,
612 B. K. Han et al.

left ventricular dysfunction, and aortic root dilation. Risk of Scan Modifications
sudden death increases significantly approximately 25 years • Biventricular (triphasic) injection protocol to opacify
postoperatively. Risk factors include prior ventriculotomy, right and left side of the heart.
increased RVEDV and RVESV, decreased RV EF, and • Extend scan range to include proximal ascending aorta
ventricular ectopy. As patients enter adulthood, they are
­ and branch pulmonary arteries.
more likely to have EP devices and to be referred for CT • Use aggressive dose reduction for coronary evaluation if
imaging [42, 43]. lesions are not suspected.
Pulmonary valve placement is the most commonly per- • ECG-gated scan using aggressive dose reduction if ven-
formed procedure in adult congenital heart disease and is tricular volumes and function are needed for clinical
commonly performed in patients with prior repair of TOF management.
[44, 45]. An RVOT conduit is also required for certain forms
of transposition with pulmonary stenosis and insufficiency
and truncus arteriosus and for neo-pulmonary root stenosis Single Ventricle Heart Disease
not amenable to transcatheter therapy. Transcatheter valves
are placed in existing conduits. Research into placement in Single ventricle heart disease includes a variety of lesions
the native RVOT is underway but not widely performed. where either ventricle is inadequate to support a separate
Coronary compression with transcatheter valve placement pulmonary and systemic circulation. Surgical palliation cre-
has been reported to result in death, and pre-intervention ates separation of the venous and arterial circulation by
coronary evaluation is required. Coronary arteries running allowing passive flow of systemic venous flood to the lungs,
anterior to the RVOT or immediately adjacent to the RVOT with the single ventricle pumping pulmonary venous blood
are most problematic. Percutaneous valve endocarditis may to the body. Survival through all stages of palliation has
become a significant problem in follow-up. CT is uniquely improved in the last decades and is now 60–80% at 10 years,
suited to assess this complication given that all echo modali- with most mortality in the first year of life [48, 49]. Advanced
ties have difficulty imaging the anterior location of the pul- diagnostics are required in single ventricle (SV) heart dis-
monary valve, particularly if the RVOT was stented prior to ease between stages of palliation and for serial lifelong
valve placement. (Figures 47.14 and 47.15 show vegetation assessment. Echocardiography alone is inadequate for evalu-
on the pulmonary valve and hematoma in the RVOT). ation of thoracic vasculature [50]. At most centers, the cur-
rent standard is invasive catheterization before both stage 2
 ost Common Residual Lesions
M (Glenn) and stage 3 (Fontan) procedure. Catheterization and
• Main pulmonary artery or conduit stenosis and branch anesthesia have a relatively high rate of procedural adverse
pulmonary artery stenosis events in palliated single ventricle heart disease [15, 51].
• Pulmonary insufficiency Noninvasive evaluation with both cardiovascular magnetic
• Right ventricular dilation and dysfunction resonance (CMR) and cardiovascular CT has been shown to
• Tricuspid regurgitation be diagnostically equivalent to invasive evaluation prior to
• Aortic root dilation [46, 47] stage 2 SV palliation, with fewer procedural adverse events,
• Left ventricular dysfunction less sedation/anesthesia use, and peripheral rather than

Fig. 47.14  Pulmonary artery

hypoplasia in tetralogy of
a b
Fallot. (a) Axial and (b)
sagittal reconstructions which
show a normal-sized right
pulmonary artery but a
diffusely small left pulmonary
artery. Note that the LPA had
an origin stenosis at birth
which has subsequently been
stented (arrow)
47  The Use of Cardiovascular CT in Repaired CHD 613

a b

Fig. 47.15 (a and b) RVOT obstruction clarified by cardiac CT. (a) (arrows) – although the valve itself is tissue, the metallic valve struts
Sagittal oblique view of the right ventricular outflow tract by cine can cause significant artifact in a magnetic field. (c) Transcatheter pul-
CMR – there is a convex lesion (asterisks) bulging into, and narrowing, monary valve endocarditis visualized by CT: This 2D sagittal image is
the RVOTO. Beyond this is an area of signal dephasing (white oval), from a patient with a clinical diagnosis of endocarditis in the setting of
where the underlying anatomy is unclear. (b) CT reconstruction of the a rapidly increasing RVOT gradient after transcatheter pulmonary valve
same area demonstrates the convex lesion, which from attenuation val- placement. Vegetations are visualized as low attenuation material
ues can be inferred to be hematoma and demonstrates that the poorly within the stent (arrow), confirmed at the time of surgical valve
visualized area on CMR is in fact a pulmonary valve replacement replacement

c­ entral IV access [52, 53]. Diagnostic accuracy compared to • Central venous occlusion
surgical findings is excellent, and outcomes through the next • Aortopulmonary and veno-veno collateral vessels
stage of palliation and up to 8 years after the Fontan are simi-
lar [54]. A retrospective evaluation of pre-Fontan cardiac
 can Recommendations [7]
S
catheterizations determined that over 50% did not add infor-
Complete assessment of both systemic venous and arterial
mation [55], and future diagnostic algorithms will include
vasculature is needed for second-stage surgical planning
cardiovascular CT when patients face MRI contraindications
(Figs.  47.16 and 47.17). Central venous obstruction from
due to the high rate of pacemaker placement and ­interventions
previous surgical lines is common and may affect choice of
such as pulmonary artery stenting and coil embolization of
IV site.
aortopulmonary and veno-veno collateral vessels.

 ommon Anatomic Findings Between

C • Consider lower extremity or right arm IV to avoid scatter
Stages 1 and 2 artifact from dense contrast in the innominate vein.
• Aortopulmonary or Sano (RV-PA) shunt stenosis • Consider late image acquisition so arterial and venous
• Branch pulmonary artery stenosis vasculature can be obtained in a single dataset or a routine
• Aortic coarctation delayed venous scan.
614 B. K. Han et al.

Tetralogy of Fallot
Surgical Intervention

Pulmonary stenosis Pulmonary artery atresia

Absent pulmonary valve

Complete repair: pulmonary

valvotomy or transannular patch, Initial palliation with a shunt or
VSD closure Complete repair includes
“unifocalization” of aortopulmonary
RVOT conduit,
collateral arteries
pulmonary artery plication,
VSD closure

Complete repair includes

RV-PA conduit, VSD closure

Fig. 47.16  Surgical intervention for tetralogy of Fallot

a b

Fig. 47.17  Common findings after first-stage palliation of single ven- is also called a Sano shunt and is placed from the anterior right ventricle
tricle heart disease. (a) This 3D volume-rendered dataset is a posterior to the branch pulmonary arteries as part of the first-stage single ventri-
view of an aortopulmonary shunt to the main pulmonary artery (arrow), cle palliation, an alternative to an aortopulmonary shunt (see Figure a).
with proximal stenosis of the left pulmonary artery (*). Note the size (c) This 3D volume-rendered image shows aortic coarctation at the dis-
discrepancy between the reconstructed aorta and the proximal descend- tal Norwood anastomosis as the reconstructed aorta inserts into the
ing aorta at the distal Norwood anastomosis (short arrow). (b) This proximal descending aorta (arrow). The Sano shunt from the anterior
axial 2D image shows narrowing (arrow) at the distal RV-PA conduit surface of the right ventricle is also visualized (*)
(*) at the insertion to the branch pulmonary arteries. The RV-PA conduit
47  The Use of Cardiovascular CT in Repaired CHD 615

 ommon Anatomic Findings Between

C  ommon Anatomic Findings After
C
Stages 2 and 3 Third-Stage Palliation
• SVC to PA narrowing and branch pulmonary artery • Narrowing along the Fontan pathway (SVC, IVC, or
stenosis branch pulmonary arteries)
• Aortic coarctation, • Fenestration patency or occlusion
• Central venous occlusion • Aortic coarctation
• Aortopulmonary and veno-veno collateral vessels • Clot within the Fontan pathway, pulmonary arteries, or
rudimentary ventricle
Scan Recommendations • Aortopulmonary and veno-veno collaterals
• Consider lower extremity IV to avoid scatter artifact from • Assessment of ventricular function
dense contrast in the superior vena cava to pulmonary • Plastic bronchitis
artery connection (Fig. 47.18).
• Consider late image acquisition so SVC to PA connection is  can Acquisition Considerations
S
visualized in the venous phase. Since the head is larger rela- and Recommendations After Third-Stage
tive to body surface area in an infant, there is robust opacifi- Palliation (Fontan)
cation of the superior central venous system prior to inferior Streaming of contrast primarily into one pulmonary artery is
vena cava opacification. If detailed inferior vena cava anat- relatively common after the Fontan and can complicate
omy is needed (interrupted IVC), a lower extremity IV or a assessment of the contralateral pulmonary artery (Figs. 47.19,
routine delayed venous scan should be considered. 47.20, 47.21, and 47.22). Opacification of the inferior p­ ortion

Fig. 47.18  Pulmonary artery

aneurysm following BT shunt. a b
(a) CXR with normal
cardiopulmonary silhouette.
(b) CXR of same patient
10 years later, now showing
an abnormal configuration to
the right para-mediastinal
border. (c) Massive
aneurysmal dilatation of the
right main pulmonary artery
(black asterisk) at the
insertion of the BT shunt
(white asterisk). (d) Volume-­
rendered image to show the
size of the PA aneurysm
(asterisk) relative to
surrounding structures

c d
616 B. K. Han et al.

of the Fontan pathway is also difficult and variable, depend- sion with extensive veno-veno collateralization. Chest CT
ing on the type of Fontan connection and hemodynamics. performed acutely for chest pain or breathlessness may result
Older style atrio-pulmonary artery connections and dilated in radiologists unfamiliar with this kind of circulation report-
lateral tunnel Fontan pathways will fill much later than extra- ing large central “thrombi,” which in reality are streaming
cardiac or non-dilated lateral tunnel Fontan connections. artifacts due to incomplete mixing of opacified and unopaci-
Filling is also delayed in the presence of high pulmonary fied blood.
artery pressures, ventricular dysfunction, and venous occlu-
Scan Modifications
• Consider a two-phase contrast bolus with a 30–60 s pause
between phases. This will allow opacification of the
venous structures during image acquisition.
• A low-dose (70 or 80 kVp) ECG-triggered functional
scan with a narrow acquisition window can be used to
assess contrast flow in the superior vena cava to Fontan
pathway.
• Delayed scan is useful for evaluation of the Fontan path-
way, anywhere from 60 to 90 s after the initial contrast
injection depending on risk factors.
• The huge right atrium which can be seen in adult patients
with a classic RA-PA Fontan may rarely take over 5 min
for full mixing of iodine and blood to occur due to the
extremely slow swirling flow (referred to by echo as
“smoke”) that occurs.

 unctional Cardiovascular CT Imaging in CHD

F
Quantification of ventricular volumes and function is used in
clinical management for many indications in CHD. Examples
include the timing of RVOT conduit placement in patients
Fig. 47.19  Glenn superior vena cava to pulmonary artery anastomosis: with pulmonary stenosis and insufficiency and the placement
This 2D MIP of a venous phase angiogram shows the superior vena of defibrillators in patients with failing systemic and single
cava connection to the right pulmonary artery after second-stage single
ventricle palliation. The aortopulmonary shunt or Sano RV-PA conduit
right ventricles. On newer generation scanners with accept-
is taken down as part of this procedure able temporal resolution, cardiovascular CT is equivalent to

a b

Fig. 47.20  CT imaging after the Fontan procedure: (a) A venous phase Fontan pathway and fenestration status and quantify ventricular func-
angiogram that shows narrowing of the lateral tunnel Fontan as it inserts tion. The arrow points to unopacified blood passing from the lateral
into the undersurface of the RPA (arrow). The superior vena cava por- tunnel Fontan pathway, through a fenestration and into the atrium
tion of the pathway is widely patent. Note the pacer lead artifact and the which is opacified with contrast returning from the lungs. Note pacer
fenestration closure device between the Fontan and the atrium. (b) This lead on the epicardial surface of the heart
image is a single phase of a functional dataset obtained to evaluate the
47  The Use of Cardiovascular CT in Repaired CHD 617

Fig. 47.21  Thrombus and

pseudo-thrombus in a Fontan
a b
circuit. (a) Sagittal and (b)
axial images demonstrating
swirling unopacified blood
from the IVC entering the
right atrium (arrows). On
echo this may be mistaken for
thrombus on occasion. The
same error may be avoided on
CT by acquiring a delayed
phase several minutes after
injection of a sufficient
volume of contrast. (c, d)
Sagittal images in another
patient with an RA-PA Fontan
demonstrating thrombus in
the RA and proximal main PA
(asterisks) c d

a b

Fig. 47.22 (a) This figure shows retained pulmonary valve leaflets in disposes to clots (b), which are seen in another patient. Newer palliative
a patient who had ligation of the pulmonary valve as part of Fontan procedures oversew the pulmonary valve at the level of the valve leaf-
completion. The retained valve leaflets in the systemic circulation pre- lets to avoid this complication
618 B. K. Han et al.

CMR for right and left ventricular volumes and function as 3. Tutarel O, Kempny A, Alonso-Gonzalez R, Jabbour R, Li W,
long as contrast opacification is adequate for the chamber Uebing A, et  al. Congenital heart disease beyond the age of
60: emergence of a new population with high resource uti-
being evaluated [56–58]. Functional datasets are recon- lization, high morbidity, and high mortality. Eur Heart J.
structed in a short axis stack with 6–8  mm slices, and a 2014;35(11):725–32.
higher noise level is tolerated without loss of diagnostic 4. Baumgartner H. Geriatric congenital heart disease: a new challenge
accuracy. 70 and 80 kVp can be used for functional imaging in the care of adults with congenital heart disease? Eur Heart J.
2014;35(11):683–5.
in most patients with an estimated dose of 1–2.5 mSv [59]. 5. Warnes CA, Williams RG, Bashore TM, Child JS, Connolly HM,
Dearani JA, et al. ACC/AHA 2008 guidelines for the management
of adults with congenital heart disease: a report of the American
 ommon Indications for Functional
C College of Cardiology/American Heart Association task force on
practice guidelines (writing committee to develop guidelines on the
Analysis in CHD management of adults with congenital heart disease). Developed
in collaboration with the American Society of Echocardiography,
• Quantification of right ventricular volumes and ejection Heart Rhythm Society, International Society for Adult Congenital
fraction to determine the optimal timing of surgical or Heart Disease, Society for Cardiovascular Angiography and
Interventions, and Society of Thoracic Surgeons. J Am Coll
transcatheter pulmonary valve placement in patients with Cardiol. 2008;52(23):e143–263.
pulmonary stenosis or insufficiency and echocardio- 6. Han BK, Lesser AM, Vezmar M, Rosenthal K, Rutten-Ramos S,
graphic evidence of right ventricular enlargement or dys- Lindberg J, et al. Cardiovascular imaging trends in congenital heart
function (s/p TOF or any surgery requiring placement of a disease: a single center experience. J Cardiovasc Comput Tomogr.
2013;7(6):361–6.
pulmonary artery conduit). 7. Han BK, Rigsby CK, Leipsic J, Bardo D, Abbara S, Ghoshhajra B,
• Calculation of ejection fraction for patients with echocar- et al. Computed tomography imaging in patients with congenital heart
diographic evidence of systemic ventricular systolic dys- disease, part 2: technical recommendations. An expert consensus
function (single ventricle, systemic right or left ventricle) document of the Society of Cardiovascular Computed Tomography
(SCCT): endorsed by the Society of Pediatric Radiology (SPR)
to guide medical management of heart failure, placement and the North American Society of Cardiac Imaging (NASCI). J
of EP devices, and advanced heart therapies (VAD, Cardiovasc Comput Tomogr. 2015;9(6):493–513.
transplant). 8. Han BK, Rigsby CK, Hlavacek A, Leipsic J, Nicol ED, Siegel MJ,
• Evaluation of ejection fraction and ventricular dyssyn- et  al. Computed tomography imaging in patients with congeni-
tal heart disease part I: rationale and utility. An expert consensus
chrony in pacemaker-dependent patients to determine document of the Society of Cardiovascular Computed Tomography
need for biventricular pacing and to guide optimal lead (SCCT): endorsed by the Society of Pediatric Radiology (SPR)
placement (concomitant coronary sinus imaging is impor- and the North American Society of Cardiac Imaging (NASCI). J
tant for these cases) [60]. Cardiovasc Comput Tomogr. 2015;9(6):475–92.
9. Taylor AJ, Cerqueira M, Hodgson JM, Mark D, Min J, O’Gara P,
• Evaluation of prosthetic valve function in patients with et  al. ACCF/SCCT/ACR/AHA/ASE/ASNC/NASCI/SCAI/SCMR
unexplained systolic gradient, for evaluation of paraval- 2010 appropriate use criteria for cardiac computed tomography. A
vular leak or for possible mass/vegetation (clot/ report of the American College of Cardiology foundation appropri-
endocarditis) ate use criteria task force, the Society of Cardiovascular Computed
Tomography, the American College of Radiology, the American
• Evaluation of complex AV valve attachments in subarte- Heart Association, the American Society of Echocardiography,
rial obstruction, after AV canal defect, corrected transpo- the American Society of Nuclear Cardiology, the North American
sition, s/p Rastelli. Society for Cardiovascular Imaging, the Society for Cardiovascular
• Calculation of stroke volume difference to estimate regur- Angiography and Interventions, and the Society for Cardiovascular
Magnetic Resonance. J Cardiovasc Comput Tomogr. 2010;4(6):407
gitation (only possible in biventricular circulation). If e1–33.
more than one regurgitant lesion or shunt is present on the 10. Gottlieb EA, Andropoulos DB. Anesthesia for the patient with con-
echocardiogram, only the total stroke volume difference genital heart disease presenting for noncardiac surgery. Curr Opin
can be reported, and echocardiography must be used to Anaesthesiol. 2013;26(3):318–26.
11. Ing C, DiMaggio C, Whitehouse A, Hegarty MK, Brady J, von
determine severity of individual lesions. Ungern-Sternberg BS, et  al. Long-term differences in language
and cognitive function after childhood exposure to anesthesia.
Pediatrics. 2012;130(3):e476–85.
12. Flick RP, Katusic SK, Colligan RC, Wilder RT, Voigt RG, Olson
MD, et al. Cognitive and behavioral outcomes after early exposure
References to anesthesia and surgery. Pediatrics. 2011;128(5):e1053–61.
13. Ing CH, DiMaggio CJ, Whitehouse AJ, Hegarty MK, Sun M, von
1. Khairy P, Ionescu-Ittu R, Mackie AS, Abrahamowicz M, Pilote L, Ungern-Sternberg BS, et  al. Neurodevelopmental outcomes after
Marelli AJ. Changing mortality in congenital heart disease. J Am initial childhood anesthetic exposure between ages 3 and 10 years.
Coll Cardiol. 2010;56(14):1149–57. J Neurosurg Anesthesiol. 2014;26(4):377–86.
2. Marelli AJ, Ionescu-Ittu R, Mackie AS, Guo L, Dendukuri N, 14. DiMaggio C, Sun LS, Ing C, Li G. Pediatric anesthesia and neuro-
Kaouache M.  Lifetime prevalence of congenital heart ­ disease developmental impairments: a Bayesian meta-analysis. J Neurosurg
in the general population from 2000 to 2010. Circulation. Anesthesiol. 2012;24(4):376–81.
2014;130(9):749–56.
47  The Use of Cardiovascular CT in Repaired CHD 619

15. Ramamoorthy C, Haberkern CM, Bhananker SM, Domino

31. Stulak JM, Dearani JA, Burkhart HM, Ammash NM, Phillips SD,
KB, Posner KL, Campos JS, et  al. Anesthesia-related cardiac Schaff HV.  Coronary artery disease in adult congenital heart dis-
arrest in children with heart disease: data from the Pediatric ease: outcome after coronary artery bypass grafting. Ann Thorac
Perioperative Cardiac Arrest (POCA) registry. Anesth Analg. Surg. 2012;93(1):116–22; discussion 22–3.
2010;110(5):1376–82. 32. Bhatt AB, Foster E, Kuehl K, Alpert J, Brabeck S, Crumb S,
16. Downing TE, McDonnell A, Zhu X, Dori Y, Gillespie MJ, Rome JJ, et  al. Congenital heart disease in the older adult: a scientific
et al. Cumulative medical radiation exposure throughout staged pal- statement from the american heart association. Circulation.
liation of single ventricle congenital heart disease. Pediatr Cardiol. 2015;131(21):1884–931.
2015;36(1):190–5. 33. Pedersen TA. Late morbidity after repair of aortic coarctation. Dan
17. Watson TG, Mah E, Joseph Schoepf U, King L, Huda W, Hlavacek Med J. 2012;59(4):B4436.
AM. Effective radiation dose in computed tomographic angiogra- 34. Brown ML, Burkhart HM, Connolly HM, Dearani JA, Cetta F, Li
phy of the chest and diagnostic cardiac catheterization in pediatric Z, et al. Coarctation of the aorta: lifelong surveillance is mandatory
patients. Pediatr Cardiol. 2013;34(3):518–24. following surgical repair. J Am Coll Cardiol. 2013;62(11):1020–5.
18. Johnson JN, Hornik CP, Li JS, Benjamin DK Jr, Yoshizumi
35. Egan M, Holzer RJ.  Comparing balloon angioplasty, stenting

TT, Reiman RE, et  al. Cumulative radiation exposure and can- and surgery in the treatment of aortic coarctation. Expert Rev
cer risk estimation in children with heart disease. Circulation. Cardiovasc Ther. 2009;7(11):1401–12.
2014;130(2):161–7. 36. Holzer R, Qureshi S, Ghasemi A, Vincent J, Sievert H, Gruenstein
19. Podberesky DJ, Angel E, Yoshizumi TT, Toncheva G, Salisbury D, et  al. Stenting of aortic coarctation: acute, intermediate, and
SR, Alsip C, et al. Radiation dose estimation for prospective and long-term results of a prospective multi-institutional registry--Con-
retrospective ECG-gated cardiac CT angiography in infants and genital Cardiovascular Interventional Study Consortium (CCISC).
small children using a 320-MDCT volume scanner. AJR Am J Catheter Cardiovasc Interv. 2010;76(4):553–63.
Roentgenol. 2012;199(5):1129–35. 37. Khairy P, Landzberg MJ, Lambert J, O’Donnell CP. Long-term out-
20. Vastel-Amzallag C, Le Bret E, Paul JF, Lambert V, Rohnean A, comes after the atrial switch for surgical correction of transposition:
El Fassy E, et  al. Diagnostic accuracy of dual-source multislice a meta-analysis comparing the Mustard and Senning procedures.
computed tomographic analysis for the preoperative detection of Cardiol Young. 2004;14(3):284–92.
coronary artery anomalies in 100 patients with tetralogy of Fallot. J 38. Khairy P, Van Hare GF, Balaji S, Berul CI, Cecchin F, Cohen MI,
Thorac Cardiovasc Surg. 2011;142(1):120–6. et al. PACES/HRS expert consensus statement on the recognition
21. Han BK, Lindberg J, Overman D, Schwartz RS, Grant K, Lesser and management of arrhythmias in adult congenital heart disease:
JR. Safety and accuracy of dual-source coronary computed tomog- executive summary. Heart Rhythm. 2014;11(10):e102–65.
raphy angiography in the pediatric population. J Cardiovasc 39. Villafane J, Lantin-Hermoso MR, Bhatt AB, Tweddell JS, Geva
Comput Tomogr. 2012;6(4):252–9. T, Nathan M, et  al. D-transposition of the great arteries: the cur-
22. Young C, Taylor AM, Owens CM.  Paediatric cardiac computed rent era of the arterial switch operation. J Am Coll Cardiol.
tomography: a review of imaging techniques and radiation dose 2014;64(5):498–511.
consideration. Eur Radiol. 2011;21(3):518–29. 40. Ou P, Khraiche D, Celermajer DS, Agnoletti G, Le Quan Sang KH,
23. Meinel FG, Henzler T, Schoepf UJ, Park PW, Huda W, Spearman Thalabard JC, et  al. Mechanisms of coronary complications after
JV, et  al. ECG-synchronized CT angiography in 324 consecutive the arterial switch for transposition of the great arteries. J Thorac
pediatric patients: spectrum of indications and trends in radiation Cardiovasc Surg. 2013;145(5):1263–9.
dose. Pediatr Cardiol. 2015;36(3):569–78. 41. Ou P, Celermajer DS, Marini D, Agnoletti G, Vouhe P, Brunelle
24. Halliburton SS, Abbara S, Chen MY, Gentry R, Mahesh M, Raff F, et  al. Safety and accuracy of 64-slice computed tomography
GL, et al. SCCT guidelines on radiation dose and dose-­optimization coronary angiography in children after the arterial switch opera-
strategies in cardiovascular CT.  J Cardiovasc Comput Tomogr. tion for transposition of the great arteries. J Am Coll Cardiol Img.
2011;5(4):198–224. 2008;1(3):331–9.
25. Khairy P, Van Hare GF, Balaji S, Berul CI, Cecchin F, Cohen MI, 42. Khairy P, Aboulhosn J, Gurvitz MZ, Opotowsky AR, Mongeon
et al. PACES/HRS expert consensus statement on the recognition FP, Kay J, et  al. Arrhythmia burden in adults with surgically
and management of arrhythmias in adult congenital heart disease. repaired tetralogy of Fallot: a multi-institutional study. Circulation.
Heart Rhythm. 2014;11(10):e102–65. 2010;122(9):868–75.
26. Morray BH, McElhinney DB, Cheatham JP, Zahn EM, Berman 43. Khairy P, Dore A, Poirier N, Marcotte F, Ibrahim R, Mongeon FP,
DP, Sullivan PM, et  al. Risk of coronary artery compression et  al. Risk stratification in surgically repaired tetralogy of Fallot.
among patients referred for transcatheter pulmonary valve Expert Rev Cardiovasc Ther. 2009;7(7):755–62.
implantation: a multicenter experience. Circ Cardiovasc Interv. 44.
Geva T.  Indications for pulmonary valve replacement in
2013;6(5):535–42. repaired tetralogy of fallot: the quest continues. Circulation.
27. McElhinney DB, Hellenbrand WE, Zahn EM, Jones TK, Cheatham 2013;128(17):1855–7.
JP, Lock JE, et al. Short- and medium-term outcomes after trans- 45. Ferraz Cavalcanti PE, Sa MP, Santos CA, Esmeraldo IM, de

catheter pulmonary valve placement in the expanded multicenter Escobar RR, de Menezes AM, et al. Pulmonary valve replacement
US melody valve trial. Circulation. 2010;122(5):507–16. after operative repair of tetralogy of Fallot: meta-analysis and meta-­
28. Han BK, Moga FX, Overman D, Carter C, Lesser JR. Diagnostic regression of 3,118 patients from 48 studies. J Am Coll Cardiol.
value of contrast-enhanced multiphase computed tomography for 2013;62(23):2227–43.
assessment of percutaneous pulmonary valve obstruction. Ann 46. Le Gloan L, Mongeon FP, Mercier LA, Dore A, Marcotte F,

Thorac Surg. 2016;101(4):e115–6. Ibrahim R, et al. Tetralogy of Fallot and aortic root disease. Expert
29. Habets J, Tanis W, van Herwerden LA, van den Brink RB,
Rev Cardiovasc Ther. 2013;11(2):233–8.
Mali WP, de Mol BA, et  al. Cardiac computed tomography 47. Warnes CA, Child JS. Aortic root dilatation after repair of tetralogy of
angiography results in diagnostic and therapeutic change in Fallot: pathology from the past? Circulation. 2002;106(11):1310–1.
prosthetic heart valve endocarditis. Int J Cardiovasc Imaging. 48. d’Udekem Y, Xu MY, Galati JC, Lu S, Iyengar AJ, Konstantinov
2014;30(2):377–87. IE, et  al. Predictors of survival after single-ventricle palliation:
30. Wijesekera H, Seneviratne S, Woolley I. Cardiac computed tomog- the impact of right ventricular dominance. J Am Coll Cardiol.
raphy in endocarditis. Heart. 2010;96(19):1604. 2012;59(13):1178–85.
620 B. K. Han et al.

49. DiBardino DJ.  Long-term progression and survival following

55. Banka P, McElhinney DB, Bacha EA, Mayer JE Jr, Gauvreau
Norwood Single Ventricle Reconstruction. Curr Opin Cardiol. K, Geva T, et  al. What is the clinical utility of routine cardiac
2015;30(1):95–9. catheterization before a Fontan operation? Pediatr Cardiol.
50. Stern KW, McElhinney DB, Gauvreau K, Geva T, Brown
2010;31(7):977–85.
DW.  Echocardiographic evaluation before bidirectional Glenn 56. Aho MR, Gebregziabher M, Schoepf UJ, Suranyi P, Lee H, Gregg
operation in functional single-ventricle heart disease: com- D, et al. Impact of right ventricular contrast attenuation on the accu-
parison to catheter angiography. Circ Cardiovasc Imaging. racy of right ventricular function analysis at cardiac multi-detector-­
2011;4(5):498–505. row CT. Eur J Radiol. 2010;73(3):560–5.
51. Rabbitts JA, Groenewald CB, Mauermann WJ, Barbara DW,
57. Greupner J, Zimmermann E, Grohmann A, Dubel HP, Althoff TF,
Burkhart HM, Warnes CA, et al. Outcomes of general anesthesia Borges AC, et al. Head-to-head comparison of left ventricular func-
for noncardiac surgery in a series of patients with Fontan palliation. tion assessment with 64-row computed tomography, biplane left
Paediatr Anaesth. 2013;23(2):180–7. cineventriculography, and both 2- and 3-dimensional transthoracic
52. Brown DW, Gauvreau K, Powell AJ, Lang P, Colan SD, Del
echocardiography: comparison with magnetic resonance imaging as
Nido PJ, et al. Cardiac magnetic resonance versus routine cardiac the reference standard. J Am Coll Cardiol. 2012;59(21):1897–907.
catheterization before bidirectional glenn anastomosis in infants 58. Rizvi A, Deano RC, Bachman DP, Xiong G, Min JK, Truong
with functional single ventricle: a prospective randomized trial. QA. Analysis of ventricular function by CT. J Cardiovasc Comput
Circulation. 2007;116(23):2718–25. Tomogr. 2015;9(1):1–12.
53. Han BK, Vezmar M, Lesser JR, Michalak G, Grant K, Dassenko D, 59. Lesser AM, Newell MC, Samara MA, Gornick C, Grant K,

et al. Selective use of cardiac computed tomography angiography: Garberich R, et  al. Radiation dose and image quality of 70
an alternative diagnostic modality before second-stage single ven- kVp functional cardiovascular computed tomography imag-
tricle palliation. J Thorac Cardiovasc Surg. 2014;148(4):1548–54. ing in congenital heart disease. J Cardiovasc Comput Tomogr.
54. Brown DW, Gauvreau K, Powell AJ, Lang P, del Nido PJ, Odegard 2016;10(2):173–8.
KC, et  al. Cardiac magnetic resonance versus routine cardiac 60. Niazi I, Dhala A, Choudhuri I, Sra J, Akhtar M, Tajik AJ. Cardiac
catheterization before bidirectional Glenn anastomosis: long-term resynchronization therapy in patients with challenging anatomy
follow-up of a prospective randomized trial. J Thorac Cardiovasc due to venous anomalies or adult congenital heart disease. Pacing
Surg. 2013;146(5):1172–8. Clin Electrophysiol: PACE. 2014;37(9):1181–8.
 Part XI
Where We Are: The Heart and Beyond
 CT Imaging of the Heart-Lung Axis
48
Michelle C. Williams and Edwin J. R. van Beek

The heart and lungs are intimately related in terms of both optimised reconstruction algorithms or slice thickness/inter-
anatomy and physiology. Their close proximity in the thorax val have not been used. It may be possible to perform
means that CT imaging of one organ also visualises at least quantitative calcium scoring on these images, by application
part of the other. The shared physiology of respiration and of the Agatston or other calcium scoring methods. However
circulation has implications in both health and disease. differences in the acquisition and reconstruction of these
Diseases that affect one of these organs have secondary images will mean that the resulting calcium score may be
implications for the other. In addition shared aetiologies, overestimated [1, 2]. Nevertheless, these techniques have
such as smoking, or similar mechanistic pathways, such as been shown to correlate with traditional Agatston scores [1].
systemic inflammation, can lead to the coexistence of both Ordinal calcium scores can also be used to assess coronary
respiratory and cardiovascular diseases. Thus, CT imaging artery calcification on these non-optimised CT images [3, 4].
of the heart-lung axis is frequently encountered in clinical This method assesses the presence and quantity of calcifica-
practice and a holistic approach to cardiothoracic imaging is tion in the major epicardial vessels (none, mild, moderate,
therefore required. severe) and provides a summed score. Vascular calcification
This chapter will discuss some of the key imaging rela- is influenced by age and gender as well as cardiovascular
tionships between the heart and the lungs, including pulmo- risk; therefore, these results should be communicated in the
nary hypertension, thromboembolic disease, chronic lung context of the patient’s age and gender [5].
diseases and atherosclerosis. CT scans performed to assess respiratory disease can also
provide information on the heart structure and function.
These cardiovascular findings may be incidental or related to
CT Imaging Methods to Assess The Heart-­ the presenting complaint. This includes identification of
Lung Axis dilated cardiac chambers, dilated great vessels and valvular
calcification [6, 7]. Valvular calcification may indicate the
Cardiovascular disease may be identified on CT imaging presence of incidental aortic stenosis, particularly in elderly
performed to assess the lungs, such as CT pulmonary angi- populations [8]. A combination of cardiovascular CT find-
ography (CTPA) or high-resolution CT of the chest (HRCT, ings and traditional risk factors can be used to stratify patients
Fig. 48.1). As a rule, electrocardiogram (ECG) gating is not at risk, and this is associated with prognosis [9]. Evidence of
used for these types of studies, and therefore motion artefact congenital cardiovascular disease and malignancy can also
will impact the ability to assess the coronary arteries. be diagnosed and may lead to further dedicated imaging
Nevertheless, coronary artery calcification can be identified, examinations.
and a qualitative or semiquantitative assessment is usually Respiratory disease is frequently identified on CT imag-
feasible. ing performed to assess the heart [10]. It is now widely
Coronary artery calcium can be assessed on CT images accepted that for cardiac CT, a wide field of view to assess
which are not optimised for traditional calcium scoring [1]. the imaged lungs and other extra-cardiac findings within the
This includes images performed without ECG gating, scanned volume should be reconstructed and assessed [11,
contrast-­enhanced images and images where cardiac-­ 12]. For some patients the respiratory findings may be the
cause of their symptoms, such as patients with chest pain and
M. C. Williams (*) · E. J. R. van Beek a pulmonary embolism. For others the respiratory findings
Edinburgh Imaging Facility QMRI, University of Edinburgh, will require further assessment and follow-up, such as the
Edinburgh, UK identification of lung nodules (Fig. 48.2).
e-mail: [email protected]

© Humana Press 2019 623

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_48
624 M. C. Williams and E. J. R. van Beek

Fig. 48.1 Cardiovascular
a b
disease identified as coronary
artery calcification on (a)
CTPA in a patient with
pulmonary embolism (b)
high-resolution CT of the
chest in a patient with
idiopathic pulmonary fibrosis

Fig. 48.2 Extra-cardiac
a b
findings identified in the lungs
on CT coronary angiography
with (a) lung malignancy, (b)
pulmonary embolism, (c)
emphysema and (d)
pulmonary fibrosis

d
c

In addition to visual assessment of the lungs, a variety of contrast-enhanced CT of the lungs can be used to quantify
automated software tools can be applied. The extent of pulmonary perfusion [17].
emphysema can be quantified based on attenuation density A direct simultaneous assessment of the cardiovascular
[13, 14]. Interstitial lung disease can be identified based on and respiratory system can be planned, for instance, in
attenuation density and texture analysis [14, 15]. Automated patients referred for lung transplantation [18]. Patients pre-
software can also be used to identify lung nodules and track senting with acute chest pain may benefit from a combined
their dimensions and volume on follow-up imaging. Modern “triple-rule-out” scan in selected cases [19, 20]. This tech-
CT techniques can also be used to assess lung function. nique allows for simultaneous assessment of the coronary
Dual-energy imaging can be used to create pulmonary blood arteries for coronary artery disease, the pulmonary arteries
volume images, which can be used to assess pulmonary per- for the presence of pulmonary embolism and the aorta for
fusion in areas of infarction or in lung tumours [16]. Dynamic aortic dissection. However, this is a highly challenging
48  CT Imaging of the Heart-Lung Axis 625

t­echnique due to the requirement of contrast opacification in bidities such as heart disease, diabetes and hypertension
three separate time intervals, resulting in a high frequency of [33]. Secondary markers of inflammation such as elastin
non-diagnostic examinations, as well as exposing the patient degradation products are also increased in patients with
to an increased radiation dose and volume of contrast [19, COPD [34]. Inflammation is central to a number of other
20]. Therefore, the routine use of triple-rule-out CT for all respiratory diseases such as pulmonary fibrosis, bronchiecta-
patients attending to the emergency department is not sis and cystic fibrosis. In addition, inflammation has an
recommended. important role in the development of lung cancer, through
complex and incompletely understood mechanisms [35].
Thus, inflammation is a shared pathogenesis between a vari-
Shared Aetiology and Pathogenesis ety of respiratory and cardiovascular diseases and may repre-
sent a potential therapeutic target [36].
Respiratory and cardiovascular diseases share a number of However, a recent population study of 1154 patients iden-
factors in their aetiology and pathogenesis. The most impor- tified that, although airflow limitation was an independent
tant shared risk factor for cardiovascular and respiratory dis- predictor of atherosclerosis, systemic inflammation was not
eases is smoking. Globally over 6 million deaths per year can [37]. In addition, air flow limitation and endothelial dysfunc-
be linked to the effects of tobacco [21]. The proportion of tion appear to be unrelated, independent predictors of the
deaths attributable to tobacco is 71% for lung cancer, 42% presence of atherosclerosis [38]. Thus, the links between
for chronic obstructive pulmonary disease (COPD) and 10% respiratory and cardiovascular diseases and inflammation are
for cardiovascular disease [21]. In addition, smokers have an complex and remain incompletely understood.
excess of mortality not accounted for by these known asso- Overall, there is overwhelming evidence that there is a
ciations [22]. combined effect of cardiovascular disease and COPD, affect-
In addition to smoking, there is overlap between other tra- ing outcomes. The COPDGene study demonstrated comor-
ditional cardiovascular risk factors and the risk of respiratory bidities had a major impact on clinical outcomes [39], while
disease. A recent meta-analysis identified an increased prev- cardiovascular disease was directly associated with COPD
alence of risk factors for cardiovascular disease amongst severity, reduced functional status and quality of life [40]. In
patients with COPD [23]. In particular there was an increased a different cohort, the ECLIPSE study, it was also demon-
prevalence of hypertension (meta-odds ratio (OR) 1.33, 95% strated that patients with COPD had increased coronary
confidence interval (CI) 1.13–1.56, p  =  0.0007), diabetes artery risk with adverse outcomes in terms of morbidity and
(1.36, 95% CI 1.21–1.53, p < 0.0001) and a history of ever mortality [39–41].
having smoked (4.25, 95% CI 3.23–5.60, p  <0.0001) [23].
Interestingly, no association between the presence of COPD
and the prevalence of obesity or dyslipidaemia was demon- Pulmonary Hypertension and The Heart
strated [23]. Occupational exposure to asbestos is linked to
an increased risk of both respiratory and cardiovascular dis- Pulmonary hypertension is defined as a mean pulmonary
ease [24]. Exposure to airborne pollution or particulate mat- artery pressure of 25 mmHg or above on catheterisation of
ter is associated with increased respiratory and cardiovascular the right side of the heart [42]. There are multiple causes of
morbidity and mortality [25–27]. pulmonary hypertension, which the widely used clinical
Inflammation plays an important role in the development classification divides into five groups (pulmonary arterial
of both cardiovascular and respiratory diseases [28]. hypertension, Group 1; pulmonary hypertension due to left
Biomarkers of systemic inflammation are increased in heart disease, Group 2; pulmonary hypertension due to
patients with reduced lung function [29] and in patients with chronic lung disease and/or hypoxia, Group 3; chronic
symptomatic and asymptomatic cardiovascular disease [30, thromboembolic pulmonary hypertension, Group 4; and pul-
31]. Smoking is associated with an increase in inflammatory monary hypertension due to unclear multifactorial mecha-
biomarkers, even after smoking cessation [30]. Blood-borne nisms, Group 5) [43]. Thus, there is significant overlap in the
inflammatory cells make up an important part of the athero- requirement for imaging both the heart and the lungs in
sclerotic plaque, and inflammatory cytokines are involved in patients with suspected or known pulmonary hypertension.
the progression of atherosclerosis [32].
Lung inflammation is central to the development of
COPD, and low-grade systemic inflammation has been Acute Pulmonary Embolism
linked to an increase in complications, more rapid decline in
lung function and increased hospitalisation [29]. Systemic Acute pulmonary embolism can lead to changes in the heart
inflammation in patients with COPD is associated with and lungs that can be identified by echocardiography, CT and
poorer clinical outcomes and increased prevalence of comor- perfusion scintigraphy (Fig.  48.3). Right heart strain on
626 M. C. Williams and E. J. R. van Beek

a b I
VENT_AP VENT_PA VENT_RPO VENT_LPO

II
PERF_AP PERF_PA PERF_RPO PERF_LPO

Fig. 48.3  Examples of SPECT images of the lungs. Planar SPECT images (b) in a separate patient showing normal ventilation on the supe-
images (a) with normal ventilation images and multiple perfusion rior image and multiple perfusion defects on the inferior image
defects, indicating a high probability of PE.  Reconstructed SPECT

CTPA is indicated by the presence of dilation of the right and ventricular remodelling (Fig. 48.5). In addition changes
ventricle, dilation of the right atrium, an increase in the right in the lung parenchyma can be identified by CT including
ventricle to left ventricle ratio (RV/LV), bowing of the inter- infarction and mosaic perfusion [49]. Finally, initial studies
ventricular septum, dilation of the main and branch pulmo- have shown that CT perfusion may be useful in assessing
nary arteries and regurgitation of contrast into the inferior patients after large pulmonary embolism, as incomplete
vena cava or azygos vein (Fig. 48.4). An increased RV/LV peripheral resolution may lead to persistent perfusion abnor-
ratio is associated with an increased severity if pulmonary malities and increased risk of pulmonary hypertension [17].
embolism [44]. Increased mortality has been identified in
patients with an increased RV/LV ratio, with bowing of the
interventricular septum, regurgitation of contrast into the  ulmonary Hypertension Secondary to Lung
P
inferior vena cava and increased volume of the right ventricle Disease
[45]. Hence, these CT findings in the heart on CTPA may
identify patients who require more aggressive treatment with A variety of mechanisms lead to the development of pulmo-
fibrinolytic agents or follow-up. nary hypertension in patients with respiratory diseases such
as COPD or idiopathic pulmonary fibrosis (Fig. 48.6). COPD
causes obliteration of the pulmonary vascular bed through
Chronic Thromboembolic Pulmonary parenchymal destruction, and the cross-sectional area of
Hypertension small pulmonary vessels correlates with the severity of pul-
monary artery hypertension [50]. The vascular responses to
After treatment of an acute pulmonary embolism, these car- cigarette smoke include vasoconstriction, vascular remodel-
diac features can resolve following normalisation of the pul- ling, intima hyperplasia and muscular changes of the vessel
monary artery pressures. However, in a small proportion of wall [51]. Hypoxia leads to vasoconstriction and inhomoge-
patients, chronic thromboembolic pulmonary hypertension neous perfusion [52]. In addition hyperinflation in COPD
may develop [46]. After an acute embolic event, the cumula- can impair right and left ventricular filling [53].
tive incidence of developing chronic thromboembolic pul- These changes in the pulmonary vasculature lead to sub-
monary hypertension is between 0.1 and 9.1% at 2  years sequent alterations in the right ventricle. Patients with COPD
[47]. However, routine imaging follow-up after acute pulmo- are five times more likely to have disease of the pulmonary
nary embolism is not currently recommended [47, 48]. CT circulation compared to matched controls or the general pub-
features of chronic thromboembolic pulmonary hyperten- lic and two times more likely to develop heart failure [23].
sion include endothelialised thrombus or webs, pulmonary Enlargement of the pulmonary artery diameter compared to
artery stenosis or occlusion and features of right heart strain the diameter of the aorta in patients with COPD is associated
48  CT Imaging of the Heart-Lung Axis 627

Fig. 48.4  CTPA images from

two patients showing
a d
evidence of right heart strain.
In the first patient with a
saddle pulmonary embolism
CT shows (a) a dilated right
ventricle, an increased right
ventricle to left ventricle ratio
and bowing of the
interventricular septum, (b)
dilation of the main
pulmonary artery and (c) and
reflux of contrast into the e
azygos vein. The second b
patient has pulmonary
embolism in (d) the left and
right pulmonary arteries, (e)
dilation of the right ventricle
and right atrium and (f)
regurgitation of contrast into
the inferior vena cava

c f

with the development of pulmonary hypertension and includes thinning, dilation and hypertrophy of the right
increased mortality [54]. Furthermore, patients with features ­ventricle, septal bowing of the left ventricle, tricuspid regur-
of pulmonary hypertension are more likely to suffer repeated gitation and changes in cardiac output [58]. Eventually this
and more severe exacerbations of COPD [55]. However, will result in right heart failure and secondary changes within
there is significant capacity within the lungs and heart to the lung parenchyma.
compensate for these changes, and only 5–10% of patients
with severe COPD develop pulmonary hypertension [56].
Atherosclerosis and Respiratory Disease

Other Causes of Pulmonary Hypertension The coexistence of coronary artery disease and respiratory
disease can be identified on CT. The link between cardiovas-
Other causes of pulmonary hypertension include hereditary cular disease and COPD (chronic obstructive pulmonary dis-
causes and drug-induced and connective tissue diseases ease) has been extensively studied. Patients with COPD have
(Fig.  48.7). Pulmonary hypertension can occur in patients an increased frequency of cardiovascular disease, and car-
with congenital heart disease [57] (Fig.  48.8). In addition diovascular disease is associated with more severe disease,
idiopathic pulmonary hypertension may be diagnosed after hospitalisation, morbidity and mortality. In addition there is
exclusion of other causes. The combined endpoint of all a higher frequency of coronary artery disease in other respi-
causes of chronic pulmonary hypertension is morphological ratory diseases including bronchiectasis and interstitial lung
and functional changes in the right ventricle (Fig. 48.9). This disease. With the advancing use of low-dose CT to screen
628 M. C. Williams and E. J. R. van Beek

Fig. 48.5 Chronic
a b
thromboembolic pulmonary
hypertension with (a) mural
thrombus and web in the right
pulmonary artery branches
(arrows), (b) dilation of the
right ventricle, dilation of the
right atrium and an increased
right ventricle to left ventricle
ratio, (c) regurgitation of
contrast into the azygos vein
and (d) regurgitation of
contrast into the inferior vena
cava

c d

Fig. 48.6 Pulmonary
hypertension in a patient
a c
with idiopathic pulmonary
fibrosis (a), dilation of the
main pulmonary artery (b)
and dilation of the right
ventricle (c)

patients at risk of lung cancer, there will also be an opportu- of cardiovascular disease in patients with COPD [23].
nity to identify asymptomatic or undiagnosed coronary Subgroup analysis showed similar rates of coronary heart
artery disease in these patients. disease (meta-OR 1·86, 95% CI 1.51–2.30; p < 0.0001) and
myocardial infarction (2.71, 95% CI 1.69–4.35; p < 0.0001)
but increased rates of angina pectoris (8.16, 95% CI 3.08–
Cardiovascular Disease and COPD 21.59; p < 0.0001) [23]. Interestingly there was no increase
in cerebrovascular disease in patients with COPD (1.32;
Several cohort studies in patients with COPD have shown a 95% CI 0.99–1.76; p = 0.06) [23].
correlation between cardiovascular disease and COPD [59– However, studies relying on patient-reported frequency of
61]. A recent meta-analysis identified a 2.5-fold higher risk cardiovascular disease are potentially limited by referral or
48  CT Imaging of the Heart-Lung Axis 629

a b c

Fig. 48.7  Right pulmonary artery occlusion due to histoplasmosis (arrow) with (a) non-contrast CT, (b) contrast-enhanced CT and (c) lung
windows

Fig. 48.8  CT imaging in a

patient with an atrial septal
a b
defect with (a) hypertrophy of
the right ventricle and right
atrium, (b) dilation of the
pulmonary arteries and
pleural effusion, (c)
regurgitation of contrast into
the inferior vena cava (d) and
interstitial lung changes

c d

recollection bias. The assessment of coronary artery calcifi- were not independently associated with coronary artery cal-
cation on CT can provide a more robust assessment of car- cification [63]. However, these investigators did find an asso-
diovascular risk. Coronary artery calcification is a marker of ciation between spirometry and carotid intima-media
atherosclerosis, and patients with an increased volume of thickness amongst smokers and percent emphysema and
coronary artery calcification have a higher risk of myocardial ankle brachial index amongst smokers and non-smokers
infarction and mortality [62]. Several studies have shown [63]. Both COPD and cardiovascular diseases are a heteroge-
that patients with emphysema on CT or airflow limitation by neous mix of pathologies, potentially leading to these differ-
spirometry have more coronary artery calcification than ences in the assessment of coronary artery calcification.
those with normal lungs [37, 63]. However, in the Multi-­ Cardiovascular disease is linked with morbidity in patients
Ethnic Study of Atherosclerosis (MESA) where 3642 asymp- with COPD. In the COPDGene cohort, physician diagnosed
tomatic patients underwent cardiac CT, lung function self-reported cardiovascular disease was associated with
assessed with spirometry and emphysema assessed on CT reduced functional status and poorer quality of life [40].
630 M. C. Williams and E. J. R. van Beek

a d g

e
b
h

c
f

Fig. 48.9  Pulmonary hypertension and right heart failure in a patient hypertrophy, (e) dilation of the right ventricle, right atrium and pleural
with Klippel-Trenaunay-Weber syndrome with (a) mega vena cava con- effusions, (f) dilation of the main pulmonary artery, (g) perfusion
taining an inferior vena cava filter, (b) multiple enlarged cutaneous defects in the lungs identified on SPECT and (h) chest x-ray showing
veins, (c) enlarged iliac veins, (d) MRI of the brain showing hemi-­ cardiomegaly and enlarged pulmonary arteries

Patients with COPD have an increased risk of hospitalisation smokers), coronary artery calcification was associated with
for cardiovascular disease compared to age- and sex-matched reduced exercise capacity and increased mortality [41]. After
controls [60, 64]. Coronary artery calcification in patients with adjusting for age, gender and pack years, a calcium score per-
COPD is linked with an elevated risk of myocardial infarction centile above 90% was associated with a hazard ratio for mor-
[4]. Patients with COPD have a higher incidence of major tality of 1.77 (95% CI 1.04–3.01) [41].
adverse cardiovascular events compared to patients without Nevertheless, subclinical cardiovascular disease is under
COPD who have similar levels of coronary artery calcification diagnosed in patients with COPD [68, 69]. The overlapping
[65]. Patients with COPD also have an increased frequency of symptoms can make the diagnosis difficult. Therefore a high
atrial and ventricular arrhythmias [23, 66]. In addition the index of suspicion is required in the assessment of patients
presence of coronary artery calcification is associated with with COPD in order not to miss this important secondary
increased disease severity in patients with COPD [4]. diagnosis. The radiologist plays a vital role in bridging the
Cardiovascular disease is linked with increased mortality in gap between pulmonary medicine and cardiology in this
patients with COPD [64]. Compared to age- and sex-­matched respect.
controls, patients with COPD had an increased risk of cardio-
vascular mortality (risk ratio 2.07, 95% CI 1.82–2.36) and all-
cause mortality (risk ratio 2.82, 95% CI 2.61–3.05) [60]. In a  ardiovascular Disease and Other Respiratory
C
cohort of patients with COPD, Romme et  al. identified that Diseases
coronary artery calcification and thoracic aortic calcification
were associated with arterial stiffness and mortality [67]. In a In addition to COPD, an increased prevalence of cardiovas-
study of 942 patients from the ECLIPSE study cohort (672 cular disease has been identified in patients with a variety of
with COPD, 199 smokers with normal spirometry and 71 non- other respiratory diseases. In a study of 42 patients with
48  CT Imaging of the Heart-Lung Axis 631

fibrosing idiopathic interstitial pneumonias, the prevalence multidisciplinary team. Guidelines from the Fleischner
of asymptomatic coronary artery disease diagnosed with CT Society and the British Thoracic Society can be used to strat-
and myocardial perfusion scintigraphy was between 12% ify management based on nodule size and risk factors [87].
and 26% [70]. Invasive coronary angiography performed
prior to lung transplantation for patients with advanced
COPD or interstitial lung disease identified occult coronary Conclusion
artery disease in half of the patients [69]. The risk of cardio-
vascular disease and stroke is also higher for patients with Computed tomography imaging can assess both the heart and
bronchiectasis [71]. the lungs in a single, rapid diagnostic test. There is an important
overlap between cardiovascular and respiratory diseases, which
may be mediated through shared aetiologies, such as smoking,
 ardiovascular Disease and Lung Cancer
C or shared pathological mechanisms, such as inflammation.
Screening Cardiovascular disease may be identified on imaging per-
formed to assess respiratory disease and vice versa.
Low-dose non-contrast CT imaging is now widely used for Abnormalities of the right heart can be a common endpoint of
lung cancer screening [72]. It is possible to identify coronary a variety of causes of pulmonary hypertension including throm-
artery calcification on these CT scans, and the presence of boembolic, respiratory and cardiac diseases. Coronary artery
coronary artery calcification correlates with cardiovascular calcification can be used as a marker of the presence of coro-
and all-cause mortality [3, 73–75]. Traditional Agatston nary artery disease in patients being assessed with CT for respi-
scoring would be time consuming in these screening studies. ratory disease. Incidental extra-cardiac findings in the lungs are
However, an ordinal visual calcium score can be performed frequently identified on cardiac CT and may have important
which correlates with traditional Agatston score and with clinical and economic implications. Thus imaging of the heart-
mortality [3, 73, 74]. lung axis is an important feature of CT imaging of the heart,
where the radiologist plays a key role in providing insight and
helps triage the patient to appropriate care pathways.
 xtra-cardiac Findings in the Lungs
E
on Cardiac CT
References
Extra-cardiac incidental findings can be identified in the
lungs on cardiac CT [76–81]. This includes findings that 1. Budoff MJ, NASIR K, Kinney GL, Hokanson JE. Coronary artery
and thoracic calcium on noncontrast thoracic CT scans: comparison
may explain the patient’s symptoms of chest pain (such as of ungated and gated examinations in patients from the COPD gene
pulmonary embolism) or breathlessness (such as emphy- cohort. J Cardiovasc Comput Tomogr. 2011;5(2):113–8.
sema or pulmonary fibrosis). In addition, clinically occult 2. Steiner RM, Budoff MJ, Khurram N, Kinney GL, Hokanson JE,
diagnoses may be made such as the presence of previously Barr RG, Nath H, Lopez-Garcia C, Black-Shinn J, and Casaburi
R. Coronary artery calcium on non-contrast thoracic CT scans; com-
unknown lung cancer. Occasionally malignancy may be parison of ECG gated and non-gated examinations in patients from
the cause of chest pain, such as when a lung cancer the COPD gene cohort. In A2616–16. American Thoracic Society.
extends into the pericardium or with metastases in the 2011. doi:https://doi.org/10.1164/ajrccm-conference.2011.183.1_
heart. MeetingAbstracts.A2616.
3. Shemesh J, Henschke CI, Shaham D, Yip R, Farooqi AO, Cham
Incidental findings were identified in more than a third of MD, McCauley DI, et al. Ordinal scoring of coronary artery calci-
cardiac CT scans in a large Scottish population study [82]. fications on low-dose CT scans of the chest is predictive of death
However, significant findings that require further investiga- from cardiovascular disease. Radiology. 2010;257(2):541–8.
tion or treatment are much less common, representing only 4. OʼHare PE, Ayres JF, OʼRourke RL, Slaughter RE, Marshall
HM, Bowman RV, Fong KM, Yang IA.  Coronary artery calci-
16% [83]. Half of clinically significant extra-cardiac findings fication on computed tomography correlates with mortality in
occur in the lungs [83]. A meta-analysis identified that previ- chronic obstructive pulmonary disease. J Comput Assist Tomogr.
ously unknown malignancy was identified in 0.7% of cardiac 2014;38(5):753–9.
CT scans, with 70% of these being lung cancer [84]. Such 5. McClelland RL, Chung H, Detrano R, Post W, Kronmal
RA.  Distribution of coronary artery calcium by race, gender,
incidental findings can have important clinical and economic and age: results from the Multi-ethnic Study of Atherosclerosis
implications, particularly if further follow-up imaging is (MESA). Circulation. 2006;113(1):30–7.
required [85, 86]. 6. Foley PWX, Hamaad A, El-Gendi H, Leyva F. Incidental cardiac
Lung nodules represent a quarter of clinically significant findings on computed tomography imaging of the thorax. BMC Res
Notes. 2010;3(1):326; BioMed Central Ltd.
extra-cardiac findings identified on CT coronary angiogra- 7. McKie SJ, Hardwick DJ, Reid JH, Murchison JT. Features of car-
phy [83]. These may require follow-up with repeat CT scans, diac disease demonstrated on CT pulmonary angiography. Clin
positron emission tomography or referral to a respiratory Radiol. 2005;60(1):31–8.
632 M. C. Williams and E. J. R. van Beek

8. Liu F, Coursey CA, Grahame-Clarke C, Sciacca RR, Rozenshtein exposure to traffic-related air pollution with cardiovascular and
A, Homma S, Austin JHM. Aortic valve calcification as an inciden- respiratory hospital admissions in London, UK.  Occup Environ
tal finding at CT of the elderly: severity and location as predictors Med. 2016;73(5):300–7.
of aortic stenosis. AJR Am J Roentgenol. 2006;186(2):342–9. 26. Langrish JP, Bosson J, Unosson J, Muala A, David E, Newby NL,
9. Jairam PM, Gondrie MJA, Grobbee DE, Mali WPTM, Jacobs PCA, Mills AB, Sandström T.  Cardiovascular effects of particulate air
van der Graaf Y, PROVIDI Study Group. Incidental imaging find- pollution exposure: time course and underlying mechanisms. J
ings from routine chest CT used to identify subjects at high risk of Intern Med. 2012;272(3):224–39.
future cardiovascular events. Radiology. 2014;272(3):700–8. 27. Hoek G, Krishnan RM, Beelen R, Peters A, Ostro B, Brunekreef B,
10. Kim JW, Kang E-Y, Yong HS, Kim YK, Ok HW, Yu-Whan O, Ki Kaufman JD. Long-term air pollution exposure and cardio- respira-
YL, Han H.  Incidental extracardiac findings at cardiac CT angi- tory mortality: a review. Environ Health. 2013;12(1):43.
ography: comparison of prevalence and clinical significance 28. Van Eeden S, Leipsic J, Paul Man SF, Sin DD.  The relationship
between precontrast low-dose whole thoracic scan and postcontrast between lung inflammation and cardiovascular disease. Am J
retrospective ECG-gated cardiac scan. Int J Cardiovasc Imaging. Respir Crit Care Med. 2012;186(1):11–6.
2009;25(1):75–81; Springer Netherlands. 29. Gan WQ, Man SFP, Senthilselvan A, Sin DD.  Association

11. van Beek EJR, Thompson B. Should we include full field of view between chronic obstructive pulmonary disease and systemic
in assessment of cardiac CT? J Cardiovasc Comput Tomogr. inflammation: a systematic review and a meta-analysis. Thorax.
2008;2(1):64; author reply 64–5. 2004;59(7):574–80.
12. Mirsadraee S, van Beek EJR.  Cross-sectional cardiac imaging: 30. McEvoy JW, Nasir K, DeFilippis AP, Lima JA, Bluemke DA,
prevalence and significance of extracardiac findings. J Thorac Hundley WG, Barr RG, et  al. Relationship of cigarette smoking
Imaging. 2014;29(2):92–7. with inflammation and subclinical vascular disease: the multi-­
13. Galbán CJ, Han MLK, Boes JL, Chughtai KA, Meyer CR, Johnson ethnic study of atherosclerosis. Arterioscler Thromb Vasc Biol.
TD, Galbán S, et al. Computed tomography-based biomarker pro- 2015;35(4):1002–10.
vides unique signature for diagnosis of COPD phenotypes and dis- 31. Seifarth H, Schlett CL, Lehman SJ, Bamberg F, Donnelly P, Januzzi
ease progression. Nat Med. 2012;18(11):1711–5. JL, Koenig W, Truong QA, Hoffmann U. Correlation of concentra-
14. Balbinot F, da Costa Batista Guedes Á, Nascimento DZ, Zampieri tions of high-sensitivity troponin T and high-sensitivity C-reactive
JF, Alves GRT, Marchiori E, Rubin AS, Hochhegger B. Advances protein with plaque progression as measured by CT coronary angi-
in imaging and automated quantification of pulmonary diseases in ography. J Cardiovasc Comput Tomogr. 2014;8(6):452–8.
non-neoplastic diseases. Lung. 2016;194(6):871–9. 32. Falk E.  Pathogenesis of atherosclerosis. J Am Coll Cardiol.

15. Xu Y, van Beek EJR, Hwanjo Y, Guo J, McLennan G, Hoffman 2006;47(8 Suppl):C7–C12.
EA. Computer-aided classification of interstitial lung diseases via 33. Miller J, Edwards LD, Agustí A, Bakke P, Calverley PMA, Celli
MDCT: 3D adaptive multiple feature method (3D AMFM). Acad B, Coxson HO, et al. Comorbidity, systemic inflammation and out-
Radiol. 2006;13(8):969–78; Elsevier. comes in the ECLIPSE cohort. Respir Med. 2013;107(9):1376–84.
16. Otrakji A, Digumarthy SR, Lo Gullo R, Flores EJ, Shepard J-AO, 34. Maclay JD, McAllister DA, Rabinovich R, Haq I, Maxwell S,
Kalra MK.  Dual-energy CT: spectrum of thoracic abnormalities. Hartland S, Connell M, et al. Systemic elastin degradation in chronic
Radiographics. 2016;36(1):38–52. obstructive pulmonary disease. Thorax. 2012;67(7):606–12.
17. Mirsadraee S, Reid JH, Connell M, MacNee W, Hirani N,
35. Conway EM, Pikor LA, Kung SHY, Hamilton MJ, Lam S, Lam
Murchison JT, van Beek EJ.  Dynamic (4D) CT perfusion offers WL, Bennewith KL. Macrophages, inflammation, and lung cancer.
simultaneous functional and anatomical insights into pulmonary Am J Respir Crit Care Med. 2016;193(2):116–30.
embolism resolution. Eur J Radiol. 2016;85(10):1883–90. 36. King PT.  Inflammation in chronic obstructive pulmonary disease
18. Wild J, Arrigo M, Isenring BD, Buergi U, Kurowski T, Schuurmans and its role in cardiovascular disease and lung cancer. Clin Transl
MM, Huber LC, Benden C. Coronary artery disease in lung trans- Med. 2015;4(1):68.
plant candidates: role of routine invasive assessment. Respiration/ 37. Gudmundsson G, Margretardottir OB, Sigurdsson MI, Harris TB,
Int Rev Thorac Dis. 2015;89(2):107–11. Launer LJ, Sigurdsson S, Olafsson O, Aspelund T, Gudnason
19. Burris AC, Boura JA, Raff GL, Chinnaiyan KM.  Triple rule out V.  Airflow obstruction, atherosclerosis and cardiovascu-
versus coronary CT angiography in patients with acute chest pain: lar risk factors in the AGES Reykjavik study. Atherosclerosis.
results from the ACIC consortium. JACC Cardiovasc Imaging. 2016;252(September):122–7.
2015;8(7):817–25. 38. Chandra D, Gupta A, Strollo PJ, Fuhrman CR, Leader JK, Bon J,
20. Ayaram D, Bellolio MF, Murad MH, Laack TA, Sadosty AT, Erwin Slivka WA, et  al. Airflow limitation and endothelial dysfunction.
PJ, Hollander JE, Montori VM, Stiell IG, Hess EP. Triple rule-out Unrelated and independent predictors of atherosclerosis. Am J
computed tomographic angiography for chest pain: a diagnostic Respir Crit Care Med. 2016;194(1):38–47; American Thoracic
systematic review and meta-analysis. Edited by Alan E Jones. Acad Society.
Emerg Med. 2013;20(9):861–71. 39. Putcha N, Han MLK, Martinez CH, Foreman MG, Anzueto AR,
21. WHO. 2012. Mortality attributable to tobacco. http://www.who.int/ Casaburi R, Cho MH, et al. Comorbidities of COPD have a major
tobacco/publications/surveillance/rep_mortality_attributable/en/. impact on clinical outcomes, particularly in African Americans.
22. Carter BD, Abnet CC, Feskanich D, Freedman ND, Hartge P, Lewis Chronic Obstr Pulm Dis. 2014;1(1):105–14.
CE, Ockene JK, et al. Smoking and mortality--beyond established 40. Black-Shinn JL, Kinney GL, Wise AL, Regan EA, Make B, Krantz
causes. N Engl J Med. 2015;372(7):631–40. MJ, Graham Barr R, et al. Cardiovascular disease is associated with
23. Chen W, Thomas J, Sadatsafavi M, FitzGerald JM. Risk of cardio- COPD severity and reduced functional status and quality of life.
vascular comorbidity in patients with chronic obstructive pulmo- COPD. 2014;11(5):546–51; Taylor & Francis.
nary disease: a systematic review and Meta-analysis. Lancet Respir 41. Williams MC, Murchison JT, Edwards LD, Agustí A, Bakke P,
Med. 2015;3(8):631–9. Calverley PMA, Celli B, et  al. Coronary artery calcification is
24. Rong Y, Luo X, Zhang Z, Cui X, Liu Y, Chen W.  Occupational increased in patients with COPD and associated with increased
exposure to asbestos and cardiovascular related diseases: a meta-­ morbidity and mortality. Thorax. 2014;69(8):718–23.
analysis. Prev Med Rep. 2015;2:920–6. 42. Hoeper MM, Bogaard HJ, Condliffe R, Frantz R, Khanna D,

25. Samoli E, Atkinson RW, Analitis A, Fuller GW, Green DC,
Kurzyna M, Langleben D, et al. Definitions and diagnosis of pulmo-
Mudway I, Anderson HR, Kelly FJ.  Associations of short-term nary hypertension. J Am Coll Cardiol. 2013;62(25 Suppl):D42–50.
48  CT Imaging of the Heart-Lung Axis 633

43. Simonneau G, Gatzoulis MA, Adatia I, Celermajer D, Denton C, all-cause mortality. JACC Cardiovasc Imaging. 2010;3(12):1229–
Ghofrani A, Gomez Sanchez MA, et al. Updated clinical classifi- 36; Elsevier.
cation of pulmonary hypertension. J Am Coll Cardiol. 2013;62(25 63. Barr RG, Ahmed FS, Carr JJ, Hoffman EA, Jiang R, Kawut SM,
Suppl):D34–41. Watson K.  Subclinical atherosclerosis, airflow obstruction and
44. Reid JH, Murchison JT.  Acute right ventricular dilatation: a new emphysema: the MESA lung study. Eur Respir J. 2012;39(4):846–
helical CT sign of massive pulmonary embolism. Clin Radiol. 54; Alhaj EK, Alhaj NE, Bergmann SR, Hecht H, Matarazzo TJ,
1998;53(9):694–8. Smith S, Alhaj N, Alhaj M, Nelson S. Coronary artery calcification
45. Becattini C, Agnelli G, Germini F, Vedovati MC. Computed tomog- and emphysema. Can J Cardiol. 2008;24(5):369–72.
raphy to assess risk of death in acute pulmonary embolism: a meta-­ 64. Sidney S, Sorel M, Charles P, Quesenberry C dL, Lanes S, Eisner
analysis. Eur Respir J. 2014;43(6):1678–90; European Respiratory MD.  COPD and incident cardiovascular disease hospitalizations
Society. and mortality: Kaiser Permanente medical care program. Chest J.
46. Pengo V, Lensing AWA, Prins MH.  Incidence of chronic throm- 2005;128(4):2068–75.
boembolic pulmonary hypertension after pulmonary embolism. N 65. Gaisl T, Schlatzer C, Schwarz EI, Possner M, Stehli J, Sievi NA,
Engl J Med. 2004;350(22):2257–64. Clarenbach CF, et  al. Coronary artery calcification, epicardial fat
47. Konstantinides S, Torbicki A, Agnelli G. 2014 ESC guidelines on burden, and cardiovascular events in chronic obstructive pulmo-
the diagnosis and management of acute pulmonary embolism. Eur nary disease. Edited by James D Chalmers. PLoS One. 2015;10(5):
Heart J. 2014;35(45):3145–6. e0126613; Public Library of Science.
48. Exter d, Paul L, van Es J, Kroft LJM, Erkens PMG, Douma RA, 66. Konecny T, Park JY, Somers KR, Konecny D, Orban M, Soucek
Mos ICM, Jonkers G, et  al. Thromboembolic resolution assessed F, Parker KO, et  al. Relation of chronic obstructive pulmo-
by CT pulmonary angiography after treatment for acute pulmonary nary disease to atrial and ventricular arrhythmias. Am J Cardiol.
embolism. Thromb Haemost. 2015;114(1):26–34. 2014;114(2):272–7.
49. Ley S, Ley-Zaporozhan J, Pitton MB, Schneider J, Wirth GM, Mayer 67. Romme EAPM, McAllister DA, Murchison JT, van Beek EJR,
E, Düber C, Kreitner K-F. Diagnostic performance of state-of-the-­ Petrides GS, Price COS, Rutten EPA, Smeenk FWJM, Wouters
art imaging techniques for morphological assessment of vascular EFM, MacNee W. Associations between COPD related manifesta-
abnormalities in patients with chronic thromboembolic pulmonary tions: a cross-sectional study. Respir Res. 2013;14(1):129.
hypertension (CTEPH). Eur Radiol. 2012;22(3):607–16. 68. Ye C, Younus A, Malik R, Roberson L, Shaharyar S, Veledar E,
50. Matsuoka S, Washko GR, Dransfield MT, Yamashiro T, San Jose Ahmad R, et al. Subclinical cardiovascular disease in patients with
Estepar R, Diaz A, Silverman EK, Patz S, Hatabu H. Quantitative chronic obstructive pulmonary disease: a systematic review. QJM.
CT measurement of cross-sectional area of small pulmonary ves- 2016; https://doi.org/10.1093/qjmed/hcw135.
sel in COPD: correlations with emphysema and airflow limitation. 69. Reed RM, Eberlein M, Girgis RE, Hashmi S, Iacono A, Jones S,
Acad Radiol. 2010;17(1):93–9. Netzer G, Scharf S.  Coronary artery disease is under-­diagnosed
51. Barberà JA.  Mechanisms of development of chronic obstructive and under-treated in advanced lung disease. Am J Med.
pulmonary disease-associated pulmonary hypertension. Pulm Circ. 2012;125(12):1228.e13–22.
2013;3(1):160–4. 70. Cassagnes L, Gaillard V, Monge E, Faivre J-B, Delhaye C,

52. Kent BD, Mitchell PD, McNicholas WT.  Hypoxemia in patients Molinari F, Petyt G, et  al. Prevalence of asymptomatic coronary
with COPD: cause, effects, and disease progression. Int J Chron disease in fibrosing idiopathic interstitial pneumonias. Eur J Radiol.
Obstruct Pulmon Dis. 2011;6:199–208. 2015;84(1):163–71.
53. O'Donnell DE, K A W.  Lung hyperinflation in COPD: applying 71. Navaratnam V, Millett ERC, Hurst JR, Thomas SL, Smeeth L,
physiology to clinical practice. COPD Res Pract. 2015;1:4. Hubbard RB, Brown J, and Quint JK. 2016. Bronchiectasis and
54. Shin S, King CS, Whitney Brown A, Albano MC, Atkins M, Sheridan the risk of cardiovascular disease: a population-based study.
MJ, Ahmad S, et al. Pulmonary artery size as a predictor of pulmo- Thorax, BMJ Publishing Group Ltd and British Thoracic Society,
nary hypertension and outcomes in patients with chronic obstruc- 72:thoraxjnl-2015-208188.
tive pulmonary disease. Respir Med. 2014;108(11):1626–32. 72. National Lung Screening Trial Research Team, Church TR, Black
55. Wells JM, Morrison JB, Bhatt SP, Nath H, Dransfield MT. Pulmonary WC, Aberle DR, Berg CD, Clingan KL, Duan F, et al. Results of
artery enlargement is associated with cardiac injury during severe initial low-dose computed tomographic screening for lung Cancer.
exacerbations of COPD. Chest J. 2016;149(5):1197–204. N Engl J Med. 2013;368(21):1980–91.
56. Naeije R.  Pulmonary hypertension and right heart failure in
73. Chiles C, Duan F, Gladish GW, Ravenel JG, Baginski SG, Snyder
chronic obstructive pulmonary disease. Proc Am Thorac Soc. BS, DeMello S, Desjardins SS, Munden RF, NLST Study Team.
2005;2(1):20–2. Association of coronary artery calcification and mortality in the
57. Dimopoulos K, Wort SJ, Gatzoulis MA. Pulmonary hypertension national lung screening trial: a comparison of three scoring meth-
related to congenital heart disease: a call for action. Eur Heart J. ods. Radiology. 2015;276(1):82–90.
2014;35(11):691–700; The Oxford University Press. 74. Sverzellati N, Cademartiri F, Bravi F, Martini C, Gira FA, Maffei
58. Peña E, Dennie C, Veinot J, Muñiz SH. Pulmonary hypertension: E, Marchianò A, et al. Relationship and prognostic value of modi-
how the radiologist can help. Radiographics. 2012;32(1):9–32. fied coronary artery calcium score, FEV1, and emphysema in
59. Eriksson B, Lindberg A, Müllerova H, Rönmark E, Lundbäck
lung cancer screening population: the MILD trial. Radiology.
B.  Association of heart diseases with COPD and restrictive 2012;262(2):460–7. https://doi.org/10.1148/radiol.11110364.
lung function--results from a population survey. Respir Med. 75. Jacobs PC, Gondrie MJA, van der Graaf Y, de Koning HJ, Isgum
2013;107(1):98–106. I, van Ginneken B, Mali WPTM. Coronary artery calcium can pre-
60. Curkendall SM, deLuise C, Jones JK, Lanes S, Stang MR, Goehring dict all-cause mortality and cardiovascular events on low-dose CT
E, She D. Cardiovascular disease in patients with chronic obstruc- screening for lung cancer. Am J Roentgenol. 2012;198(3):505–11;
tive pulmonary disease, Saskatchewan Canada cardiovascular dis- American Roentgen Ray Society.
ease in COPD patients. Ann Epidemiol. 2006;16(1):63–70. 76. Koonce J, Schoepf JU, Nguyen SA, Northam MC, Ravenel

61. Huiart L, Ernst P, Suissa S. Cardiovascular morbidity and mortality JG.  Extra-cardiac findings at cardiac CT: experience with 1,764
in COPD. Chest J. 2005;128(4):2640–6. patients. Eur Radiol. 2009;19(3):570–6; Springer-Verlag.
62. Budoff MJ, J E Hokanson KNASIR, Shaw LJ, Kinney GL, Chow 77. Kirsch J, Araoz PA, Steinberg FB, Fletcher JG, McCollough

D, DeMoss D, et al. Progression of coronary artery calcium predicts CH, Williamson EE.  Prevalence and significance of incidental
634 M. C. Williams and E. J. R. van Beek

e­ xtracardiac findings at 64-multidetector coronary CTA. J Thorac HEART): an open-label, parallel-group, multicentre trial. Lancet.
Imaging. 2007;22(4):330–4. 2015;385(9985):2383–91.
78. Lehman SJ, Abbara S, Cury RC, Nagurney JT, Hsu J, Goela A, 83. Karius P, Schuetz GM, Schlattmann P, Dewey M.  Extracardiac
Schlett CL, et al. Significance of cardiac computed tomography inci- findings on coronary CT angiography: a systematic review. J
dental findings in acute chest pain. Am J Med. 2009;122(6):543–9; Cardiovasc Comput Tomogr. 2014;8(3):174–82. e1–6.
Elsevier. 84. Flor N, Di Leo G, Squarza SA, Tresoldi S, Rulli E, Cornalba G,
79. Schietinger BJ, Bozlar U, Hagspiel KD, Norton PT, Greenbaum Sardanelli F.  Malignant incidental extracardiac findings on car-
HR, Wang H, Isbell DC, et al. The prevalence of extracardiac find- diac CT: systematic review and meta-analysis. Am J Roentgenol.
ings by multidetector computed tomography before atrial fibrilla- 2013;201(3):555–64; American Roentgen Ray Society.
tion ablation. Am Heart J. 2008;155(2):254–9; Elsevier. 85. Showkathali R, Sen A, Brickham B, Dworakowski R, Wendler
80. Kawano Y, Tamura A, Goto Y, Shinozaki K, Zaizen H, Kadota O, MacCarthy P. ‘Incidental findings’ during TAVI work-
J. Incidental detection of cancers and other non-cardiac abnormali- up: more than just an inconvenience. EuroIntervention.
ties on coronary multislice computed tomography. Am J Cardiol. 2015;11(4):465–9.
2007;99(11):1608–9; Elsevier. 86. Goehler A, McMahon PM, Lumish HS, Carol CW, Munshi V,
81. Burt JR, Iribarren C, Fair JM, Norton LC, Mahbouba M, Rubin Gilmore M, Chung JH, et al. Cost-effectiveness of follow-up of pul-
GD, Hlatky MA, Go AS, Fortmann SP, Atherosclerotic Disease, monary nodules incidentally detected on cardiac computed tomo-
Vascular Function, Genetic Epidemiology (ADVANCE) Study. graphic angiography in patients with suspected coronary artery
Incidental findings on cardiac multidetector row computed tomog- disease. Circulation. 2014;130(8):668–75.
raphy among healthy older adults: prevalence and clinical corre- 87. MacMahon H, Austin JHM, Gamsu G, Herold CJ, Jett JR, Naidich
lates. Arch Intern Med. 2008;168(7):756–61; American Medical DP, Patz EF, Swensen SJ.  Guidelines for management of small
Association. pulmonary nodules detected on CT scans: a statement from the
82. Scot-Heart Investigators. CT coronary angiography in patients
Fleischner Society. Radiology. 2005;237(2):395–400.
with suspected angina due to coronary heart disease (SCOT-­
 Ischemic Stroke: The Role of Cardiac CT
49
Jin Hur and Byoung Wook Choi

Stroke is a leading cause of mortality and long-term disabil- sources in stroke patients is important for appropriate thera-
ity worldwide [1, 2]. While the etiology of ischemic stroke is peutic management. Cardioembolic sources can be divided
often found in the cervicocranial vasculature, approximately into two categories, major sources or minor sources
20% results from high-risk cardiac abnormalities [3, 4]. (Table  49.1) [17, 18]. Major sources of embolism have an
Because cardiac embolism has a treatable source, identifica- established causal relationship with stroke, and their identifi-
tion of cardioembolic sources in stroke patients is important cation is therefore usually clinically relevant. Minor sources
for appropriate therapeutic management. Currently, echocar- of embolism, on the other hand, have an ill-described or
diography is frequently used in the workup of patients with weak causal relationship with stroke, and their relevance in a
ischemic stroke to identify potential cardiac embolic sources given situation is much less certain. Therefore, in terms of
[5–7]. Transesophageal echocardiography (TEE) is highly therapeutic management, detection of major sources is more
accurate in the detection of left atrial thrombus, patent fora- important.
men ovale (PFO), atrial septal aneurysm (ASA), and aortic Echocardiography is indicated in patients in whom car-
atheroma [8–12]. dioembolism is suspected but unproven. Transthoracic
Cardiac computed tomography (CT) can be used to evalu- echocardiography (TTE) should be used as a screening test;
ate not only the coronary arteries but also other cardiac struc- although, in the absence of clinical or simple test evidence
tures such as the left atrial appendage (LAA), myocardium, of cardiac disease, the likelihood of discovering a major
valves, and septa [13–16]. Recent advances in multidetector embolic source is low [19]. TEE provides superior anatom-
computed tomography (MDCT) allow quick and accurate ical information in relation to the aortic arch, left atrium,
evaluation of cardiac anatomy and thoracic aorta by cardiac and the mitral and aortic valves. In patients in whom the
CT. Because the cause of embolic stroke is commonly either mechanism of stroke remains uncertain after initial workup,
from carotid atherosclerosis or from the heart, these potential recent reports suggest that TEE, performed soon after
sources of emboli need to be investigated to accurately pre- symptom onset, may identify a source of embolism in
scribe secondary stroke prevention. 30–45% of cases [9–12]. However, TEE is a semi-invasive
Given the clinical importance of ischemic stroke arising procedure that required special skills for proper perfor-
from the heart and aorta, we discussed the potential clinical mance and interpretation.
application cardiac CT in an ischemic stroke population, Cardiac CT is a well-established but not widely used
detailing on the diagnosis of cardioembolic sources based on technique for imaging cardioembolic sources. Cardiac CT
the latest available evidence and published guidelines. provides high-quality noninvasive images of the heart,
great vessels, and coronary vasculature [13–16].
Therefore, cardiac CT has the ability to provide high-
Cardioembolic Stroke quality noninvasive images of uniform sections of the
heart, thus allowing the possible detection of potential
Cardioembolic stroke is estimated to be the causative factor cardiac embolic sources. However, CT imaging has fun-
in 20 to 40% of all stroke cases [1]. Because cardiac embo- damental disadvantages including radiation dose and use
lism has a treatable source, identification of cardioembolic of iodine contrast medium [20–22].

J. Hur (*) · B. W. Choi

Department of Radiology, Yonsei University School of Medicine,
Severance Hospital, Seoul, South Korea
e-mail: [email protected]

© Humana Press 2019 635

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_49
636 J. Hur and B. W. Choi

Table 49.1  Proximal sources of embolism

Major sources Minor sources
Atrial dysrhythmias Interatrial septal abnormalities
 Atrial fibrillation  Patent foramen ovale
 Atrial flutter  Atrial septal defect
 Atrial septal aneurysm
Left atrial thrombus Spontaneous echo contrast (blood
stasis)
Left ventricular thrombus Pulmonary atriovenous malformation
Cardiac tumors Mitral valve prolapse
Vegetation Valvular calcification
 Infective  Mitral annular calcification
 Noninfective  Aortic valve stenosis
(marantic)
Prosthetic cardiac valve
Complex aortic atheroma

Cardioembolic Sources

Investigation of potential embolic sources is an important

diagnostic step in managing patients with acute ischemic
stroke or transient ischemic attack, especially when the Fig. 49.1  Image in a 48-year-old man with stroke. Axial cardiac CT
mechanism is considered to be embolic. image shows a large oval-shaped filling defect (arrow) in RA. The mass
was confirmed as a thrombus which was resolved on follow-up trans-
thoracic echocardiography after anticoagulation therapy. RA  right
atrium
Cardiac Thrombus
This is because pseudo-filling defect such as blood stasis can
The left atrium (LA) or LAA is an important location of also cause an apparent filling defect on CT images, thereby
thrombus formation and subsequent cardioembolic events, mimicking a thrombus.
especially in association with dysrhythmias, such as atrial Left ventricular (LV) thrombus formation is frequently
fibrillation (AF) [23, 24]. Currently, TEE has emerged as the associated with ischemic heart disease or dilated cardiomy-
most sensitive technique for the detection of intracardiac opathy, which is a potential source of embolic stroke [31].
thrombi and is believed to be the single best modality for Contrast-enhanced magnetic resonance imaging (MRI) has a
patients with suspected intracardiac thrombi [6–9]. Although higher sensitivity and specificity than TEE for LV thrombus
TEE is widely available, it is a semi-invasive test, usually detection, especially for apical thrombi, because echocar-
performed under conscious sedation. diography cannot clearly show the myocardium-thrombus
A noninvasive method with high reliability and accuracy interface at the apex [32]. Delayed-enhancement CMR
comparable to TEE for the identification of LA or LAA (DE-CMR) imaging has been validated as a sensitive method
thrombus would be of significant clinical value. Cardiac CT for LV thrombus detection [33]. Because DE-CMR identifies
is a sensitive tool for detecting intracardiac thrombi. In sev- a thrombus based on tissue characteristics rather than ana-
eral studies, investigators have reported that intracardiac tomic appearance, it enables a thrombus to be delineated
thrombi can be detected with CT (Figs. 49.1 and 49.2) [25– from the myocardium and chamber cavity irrespective of
30]. The diagnostic accuracy of cardiac CT has been exten- location or morphology (Fig.  49.3). A previous study
sively studied indicating a high negative predictive value reported that DE-CMR detected thrombi in 7% of subjects
(NPV). Nonetheless, there have been major discrepancies (55 patients), while cine-CMR identified thrombi in only
between reported positive predictive values (PPV) and accu- 4.7% (37 patients). Among the 55 patients with LV thrombi
racy estimates. A recent meta-analysis study demonstrated identified using DE-CMR, 44% (24 of 55) had cine-CMR
overall high accuracy of cardiac CT compared with TEE for analyses negative for thrombi [33]. Cardiac CT can also
the detection of LA or LAA thrombus in patients with clearly depict the endocardial border with high resolution
AF.  The investigators included 19 studies with a total of and can easily detect a LV thrombus. On CT, left ventricular
2955 patients, and the weighted mean sensitivity and speci- thrombus shows a significantly lower CT attenuation than
ficity were 96% (95% CI, 92–100%) and 92% (95% CI, normally perfused myocardial wall (Fig. 49.4) [34].
91–93%), whereas PPV and NPV were 41% (95% CI, Recently, left atrial occlusion devices have been recog-
37–44%) and 99% (95% CI, 99–100%), respectively [28]. nized as a treatment option for preventing embolic stroke
49  Ischemic Stroke: The Role of Cardiac CT 637

a b

Fig. 49.2  Images in a 41-year-old man with stroke and atrial fibrilla- image shows an oval-shaped echogenic thrombus (arrow) in the
tion. (a) Axial cardiac CT image shows an oval filling defect (arrow) in LAA.  LAA,  left atrial appendage; TEE,  transesophageal
the LAA. The left atrium was enlarged due to atrial fibrillation. (b) TEE echocardiography

a b

Fig. 49.3  Images in a 70-year-old man with stroke and myocardial thoracic echocardiography after anticoagulation therapy. (b) On four-­
infarction. (a) On short-axis DE-CMR, the mass shows black signal chamber DE-CMR, the LV myocardium shows diffuse delayed contrast
intensity without delayed enhancement (arrow) in LV apex. The mass enhancement in the apical wall (arrows) from myocardial infarction.
was confirmed as a thrombus which was resolved on follow-up trans- LV, left ventricle

in patients with AF who are intolerant to warfarin therapy can also be used to measure the size of the LAA, which is
[35]. With excellent NPV, cardiac CT has the ability to typically used before procedure for LAA occlusion, a pro-
potentially exclude thrombus in LAA and may obviate the cedure that can reduce the risk of stroke in patients with
need for pre-­procedural TEE in some cases. Cardiac CT AF [36].
638 J. Hur and B. W. Choi

Although delayed imaging with cardiac CT has excellent

sensitivity and specificity for thrombus, it requires rescan-
ning within 1  min. This would require preknowledge of
thrombus or high suspicion of LA or LAA thrombus as this
could double radiation exposure by requiring a second scan.
One concern regarding the use of cardiac CT for LAA evalu-
ation is radiation. However, CT has most recently undergone
technical and acquisition changes that substantially reduce
the dose of radiation [40].

Cardiac Tumor

Although cardiac tumors are rare, diagnosis and refined

characterization of these masses are important due to the
high morbidity and mortality associated with the functional
consequences of such tumors and the potential for arrhyth-
mia or emboli [41]. It is clinically important to differentiate
cardiac tumors from cardiac tumor-like lesions, such as
thrombi, because of the different therapeutic strategies used
Fig. 49.4  Image in a 62-year-old man with stroke and myocardial to treat these lesions.
infarction. Axial cardiac CT shows an oval-shaped filling defect (arrow) Myxomas are the most common cardiac tumor, and
in the LV apex, which was confirmed as a thrombus. LV, left ventricle they are associated with a high rate of embolic events
(30–40%) [42]. Imaging plays a key role in establishing
Spontaneous Echo Contrast (Blood Stasis) the diagnosis of patients with cardiac myxomas and
thrombi because the clinical presentation is often diverse
Spontaneous echo contrast (SEC) is a smoke-like echo phe- and nonspecific. Magnetic resonance imaging (MRI) is
nomenon with a swirling pattern of blood flow that can be presently the modality of choice to evaluate cardiac
observed on echocardiography, most often within the tumors (Fig. 49.7). Optimal tumor characterization may
LAA.  The reduced atrial contraction associated with atrial be achieved with cardiac MRI, which provides an unre-
fibrillation leads to blood stasis, which facilitates thrombus stricted field of view and superior soft-tissue depiction
formation in the LAA. without ionizing radiation [43, 44]. Cardiac CT can be
Spontaneous echo contrast (SEC) caused by slow blood used to detect cardiac tumors. Several useful imaging
flow during atrial fibrillation, as seen at ultrasonography, also characteristics can be found based on CT for differenti-
can cause an apparent filling defect on CT images and thereby ating cardiac myxomas from thrombi. Cardiac myxomas
mimic a thrombus (Fig.  49.5). These pseudo-filling defects tend to be larger than thrombi and originate in the fossa
seen on CT images represent incomplete mixing of CT con- ovalis or interatrial septum, whereas thrombi are smaller
trast material and blood. Therefore, it may be difficult to dif- than myxomas and are usually seen in the LAA
ferentiate between a thrombus and a pseudo-filling defect (Fig.  49.8) [45]. However, CT lacks the superior soft-
caused by blood stasis with a single-phase CT scanning. The tissue contrast abilities of MRI and is therefore not suit-
blood stasis phenomenon appears when LA dysfunction able for differentiating between cardiac myxomas and
causes an incomplete mixing of contrast agent and blood. thrombi. Therefore, it may be challenging to differenti-
Therefore, the delayed phase scanning will allow for the com- ate a cardiac myxoma from an atrial thrombus based on
plete mixing of contrast agent and blood (Fig. 49.6) [37–39]. CT imaging features because both can be localized at the
In sub-analysis of 753 patients from recent meta-­analysis left atrial septum and be irregular in shape. A recent
study, in which delayed images were obtained, sensitivity and study of dual-energy cardiac CT demonstrated that
specificity were 100% (95% CI, 96–100%) and 99% (95% iodine concentration obtained from the dual-energy CT
CI, 98–100%), respectively. Likewise, the PPV significantly data is a feasible quantitative parameter in differentiat-
increased to 92% while maintaining a high NPV of 100% and ing cardiac myxomas from thrombi [46]. This result sug-
an overall diagnostic accuracy of 99% (95% CI, 98–100%) gested that dual-energy cardiac CT can be a helpful
[28]. This result suggested that cardiac CT using delayed complementary tool to differentiate a myxoma from a
imaging is a reasonable alternative to TEE for evaluating thrombus in cases in which echocardiography or conven-
LAA or LA thrombus and differentiating from blood stasis. tional contrast CT is inconclusive.
49  Ischemic Stroke: The Role of Cardiac CT 639

a b

Fig. 49.5  Images in a 62-year-old woman with stroke and atrial fibril- image acquisition shows moderate spontaneous echo contrast (arrow)
lation. (a) Axial cardiac CT image shows triangular filling defect with no thrombus in the LAA. LAA, left atrial appendage; TEE, trans-
(arrow) in the LAA. (b) TEE image obtained on the same day after CT esophageal echocardiography

a b

Fig. 49.6  Images in a 68-year-old man with stroke. (a) Axial early-­ in LAA.  The defect caused by blood stasis disappeared at late-phase
phase cardiac CT image shows triangular filling defect (arrow) in LAA. imaging. LAA, left atrial appendage
(b) Axial late-phase cardiac CT image shows no filling defect (arrow)

Interarterial Septal Abnormalities A patent foramen ovale constitutes a potential conduit

for right-to-left shunting and has a high frequency of occur-
A paradoxical embolism is a type of stroke or arterial throm- rence in young adults with cryptogenic ischemic stroke
bosis caused by a thrombus of venous origin through a defect [47]. TEE with agitated saline injection and Valsalva
in the heart, such as a patent foramen ovale (PFO) or an atrial maneuver is the accepted reference standard modality for
septal defect (ASD), that creates the potential for right-to-­ the diagnosis of PFO [48, 49]. A previous study demon-
left shunting of blood. strated that CT has lower sensitivity compared to TEE in
640 J. Hur and B. W. Choi

a b

Fig. 49.7  Images in a 56-year-old man with stroke. (a) Axial cardiac DE-CMR, the mass shows strong contrast enhancement, suggestive of
CT shows a large filling detect (arrow) in LA. (b) A four-chamber tumor (arrow). The mass was pathologically confirmed after surgical
T2-weighted black blood double inversion spin echo CMR image resection as a myxoma
shows a high signal intensity mass (arrow) in LA. (c) On four-chamber

the detection of PFO [50]. This is because PFO requires a terized by redundant and excessively mobile tissue in the
provocative maneuver for diagnosis, which is impossible to region of the fossa ovalis. A previous study showed that a
do with CT. However, an interatrial septal channel with a PFO with concomitant ASA is associated with a high risk
contrast jet from the left atrium to the right atrium on CT is of paradoxical embolism [52].
very confirmative finding of a PFO (Fig. 49.9) [50]. A chan- An ASD is the most common congenital heart lesion
nel-like appearance of the interatrial septum and the char- found in adults. Patients with an ASD and a right-to-left
acteristic direction of the contrast jet toward the inferior shunt are at risk for a stroke due to paradoxical embolization.
vena cava help differentiate a PFO from an ASD [51]. An The typical feature of an ASD on cardiac CT studies is a
atrial septal aneurysm (ASA) is defined as exaggerated sharply defined defect in the septum with a contrast jet per-
excursion of the septum into the arterial chamber, charac- pendicular to the atrial septum (Fig. 49.10) [51].
49  Ischemic Stroke: The Role of Cardiac CT 641

a b

Fig. 49.8  Images in a 58-year-old man with stroke. (a) TTE image cally confirmed after surgical resection as a myxoma. LA, left atrium;
shows an oval-shaped echogenic mass in the LA. (b) Axial cardiac CT TTE, transthoracic echocardiography
shows a lobulated filling detect (arrow) in LA. The mass was pathologi-

a b

Fig. 49.9  Images in a 43-year-old man with stroke. (a) Axial cardiac of the interatrial septum (arrow) and contrast agent jets from LA to
CT image shows contrast agent jet (arrow) from LA to RA. (b) RA toward inferior vena cava (small arrows) consistent with patent
Oblique sagittal reformatted CT image shows the flap-like appearance foramen ovale
642 J. Hur and B. W. Choi

Fig. 49.11  Image in a 56-year-old man with stroke. Axial cardiac CT

image shows an oval-shaped filling defect in the mitral valve (arrow).
The mass was pathologically confirmed as vegetation
Fig. 49.10  Image in a 55-year-old woman with stroke. Oblique sagit-
tal reformatted CT image shows a contrast jet from LA and RA (arrow)
through the ASD. The direction of the jet is perpendicular to the inter-
atrial septum. ASD atrial septal defect

Valvular Vegetation

The incidence of ischemic stroke associated with infective

endocarditis is 15–20% [53]. Mitral valve involvement
­carries greater stroke risk than aortic valve involvement [54].
Embolic risk is also increased by mobility, consistency,
extent, and size [54, 55]. Currently, echocardiography is the
first-line imaging tool for the detection of valvular vegeta-
tion [56]. CT can be used to evaluate suspected vegetation or
periannular complications such as abscesses or mycotic
aneurysms (Fig. 49.11). A recent meta-analysis data proved
that the addition of cardiac CT to TEE can improve the
detection of periannular complications and vegetations, in
patients with prosthetic heart valve endocarditis [57].

Valvular Calcification Fig. 49.12  Image in a 74-year-old man with stroke. A short-axis refor-
matted cardiac CT image shows severe calcifications in the mitral annu-
Mitral annular calcification is a risk factor of stroke. A previ- lus (arrows)
ous study of population-based cohort of 1159 subjects in an
elderly demonstrated that mitral annular calcification was The common causes of aortic stenosis are congenital, cal-
associated with twice the risk of stroke, independent of tradi- cific degenerative, and rheumatic. Even in the absence of
tional risk factors for stroke [58]. Cardiac CT is useful hemodynamically significant left ventricular outflow
modality in the evaluation of the extent and location of mitral obstruction, there is a significant increase in the risk of death
annular calcification [59]. On CT, mitral annulus calcifica- from cardiovascular events, including stroke, which may be
tions are often found around the posterior annulus of the the result of embolic fragments detaching from the thickened
mitral valve on CT and may cause systemic embolism in valve leaflets. CT is sensitive modality to evaluate the pres-
association with infective endocarditis or liquefaction of cal- ence and extent of aortic valvular calcifications. Previous
cifications (Fig. 49.12). study demonstrated that cardiac CT enables quantification of
49  Ischemic Stroke: The Role of Cardiac CT 643

aortic valvular calcifications through a calcium-scoring echocardiographic method for visualizing atherosclerotic
method and grading of the stenosis degree by measuring the plaques in the thoracic aorta and remains the imaging stan-
valve area on mid-systolic images [60]. dard for this indication [64, 65]. Compared with TEE, CT
is much more useful for measuring plaque thickness, dis-
covering ulceration, and examining its components
Aortic Atheroma (Fig. 49.13). CT has some advantages in the evaluation of
aortic plaque because of the ability to reconstruct high-
Aortic atheroma of the proximal aortic arch is important quality multidimensional images and enable a retrospec-
source of embolism in patients with ischemic stroke [61, tive review with the same quality of reproducibility [66].
62]. Complex morphological features of the plaque, such Radiation dose exposure and contrast medium are the most
as ulcerations or superimposed thrombi, have been shown serious obstacles when using CT for patients with possible
to increase the risk of stroke [62, 63]. Echocardiography, embolic stroke. CT is also a static imaging modality that
CT, and MRI are the principal means of imaging athero- cannot be used to evaluate plaque mobility, an important
sclerotic plaques in the aorta. TEE has been the primary factor in risk stratification.

a b

Fig. 49.13  Images in a 71-year-old man with recurrent stroke. (a) Axial and (b) coronal reformatted cardiac CT images show an aortic plaque
(arrows) on the aortic arch. (c) TEE image shows an echogenic aortic plaque (arrow) on the aortic arch
644 J. Hur and B. W. Choi

Conclusion 14. Raff GL, Gallagher MJ, O’Neill WW, Goldstein JA.  Diagnostic
accuracy of noninvasive coronary angiography using 64-slice spiral
computed tomography. J Am Coll Cardiol. 2005;46:552–7.
Investigation of potential embolic sources is an important 15. Mollet NR, Cademartiri F, van Mieghem CA, et al. High-resolution
diagnostic step in managing patients with acute ischemic spiral computed tomography coronary angiography in patients
stroke or transient ischemic attack, especially when the referred for diagnostic conventional coronary angiography.
Circulation. 2005;112:2318–23.
mechanism is considered to be embolic. During the past 16. Boxt LM, Lipton MJ, Kwong RY, Rybicki F, Clouse ME. Computed
decade, cardiac CT has been tested and compared with TEE tomography for assessment of cardiac chambers, valves, myocar-
for the diagnosis of cardioembolic sources. Many studies dium and pericardium. Cardiol Clin. 2003;21:561–85.
showed that cardiac CT is a very useful and powerful modal- 17. Adams HP Jr, Bendixen BH, Kappelle LJ, et al. Classification of
subtype of acute ischemic stroke: definitions for use in a multi-
ity for the detection of cardioembolic sources in stroke center clinical trial. Stroke. 1993;24:35–41.
patients. However, based on current evidence, cardiac CT is 18. Doufekias E, Segal AZ, Kizer JR. Cardiogenic and aortogenic brain
not recommended for the use in the initial evaluation of embolism. J Am Coll Cardiol. 2008;51:1049–59.
intracardiac structures in stroke patients [67]. In addition, 19. Wolber T, Maeder M, Atefy R, et al. Should routine echocardiogra-
phy be performed in all patients with stroke? J Stroke Cerebrovasc
cardiac CT imaging has fundamental disadvantages includ- Dis. 2007;16:1–7.
ing radiation dose and use of iodine contrast media. 20. Hausleiter J, Meyer T, Hermann F, et al. Estimated radiation dose
Continued technological advances requiring less contrast associated with cardiac CT angiography. JAMA. 2009;301:500–7.
and ionizing radiation could increase the importance of car- 21. Heinrich MC, Häberle L, Müller V, Bautz W, Uder M. Nephrotoxicity
of iso-osmolar iodixanol compared with nonionic low-osmolar
diac CT in this field in the near future. contrast media: meta-analysis of randomized controlled trials.
Radiology. 2009;250:68–86.
22. Honoris L, Zhong Y, Chu E, et al. Comparison of contrast enhance-
References ment, image quality and tolerability in coronary CT angiogra-
phy using 4 contrast agents: a prospective randomized trial. Int J
Cardiol. 2015;186:126–8.
1. Bonita R. Epidemiology of stroke. Lancet. 1992;339:342–4.
23. Agmon Y, Khandheria BK, Gentile F, Seward JB. Echocardiographic
2. Sacco RL, Adams R, Albers G, et  al. Guidelines for prevention
assessment of the left atrial appendage. J Am Coll Cardiol.
stroke in patients with ischemic stroke or transient ischemic attack.
1999;34:1867–77.
Stroke. 2006;37:577–617.
24. Al-Saady NM, Obel OA, Camm AJ. Left atrial appendage: structure,
3. Albers GW, Amarenco P, Easton JD, Sacco RL, Teal
function and role in thromboembolism. Heart. 1999;82:547–55.
P.  Antithrombotic and thrombolytic therapy for ischemic stroke.
25. Achenbach S, Sacher D, Ropers D, et al. Electron beam computed
Chest. 2004;126:483S–512S.
tomography for the detection of left atrial thrombi in patients with
4. Kistler JP. Cerebral embolism. Compr Ther. 1996;22:515–30.
atrial fibrillation. Heart. 2004;90:1477–8.
5. McNamara RL, Lima JA, Whelton PK, Powe NR. Echocardiographic
26. Shapiro MD, Neilan TG, Jassal DS, et al. Multidetector computed
identification of cardiovascular sources of emboli to guide clinical
tomography for the detection of left atrial appendage thrombus:
management of stroke: a cost-effectiveness analysis. Ann Intern
a comparative study with transesophageal echocardiography. J
Med. 1997;127:775–87.
Comput Assist Tomogr. 2007;31:905–9.
6. Pearson AC, Labovitz AJ, Tatineni S, Gomez CR.  Superiority of
27. Singh NK, Nallamothu N, Zuck VP, Issa ZF.  Left atrial append-
transesophageal echocardiography in detecting cardiac source of
age filling defects on 64-slice multidetector computed tomography
embolism in patients with cerebral ischemia of uncertain etiology. J
in patients undergoing pulmonary vein isolation: predictors and
Am Coll Cardiol. 1991;17:66–72.
comparison to transesophageal echocardiography. J Comput Assist
7. Leung DY, Black IW, Cranney GB, et  al. Selection of patients
Tomogr. 2009;33:946–51.
for transesophageal echocardiography after stroke and systemic
28. Romero J, Husain SA, Kelesidis I, Sanz J, Medina HM, Garcia
embolic events: role of transthoracic echocardiography. Stroke.
MJ. Detection of left atrial appendage thrombus by cardiac com-
1995;26:1820–4.
puted tomography in patients with atrial fibrillation: a meta-­
8. Daniel WG, Mugge A. Transesophageal echocardiography. N Engl
analysis. Circ Cardiovasc Imaging. 2013;6:185–94.
J Med. 1995;332:1268–79.
29. Kim YY, Klein AL, Halliburton SS, et  al. Left atrial appendage
9. de Bruijn SF, Agema WR, Lammers GJ, et  al. Transesophageal
filling defects identified by multidetector computed tomography
echocardiography is superior to transthoracic echocardiography in
in patients undergoing radiofrequency pulmonary vein antral iso-
management of patients of any age with transient ischemic attack
lation: a comparison with transesophageal echocardiography. Am
or stroke. Stroke. 2006;37:2531–44.
Heart J. 2007;154:1199–205.
10. Harloff A, Handke M, Reinhard M, et al. Therapeutic strategies after
30. Hur J, Kim YJ, Nam JE, et al. Thrombus in the left atrial append-
examination by transesophageal echocardiography in 503 patients
age in stroke patients: detection with cardiac CT angiography: a
with ischemic stroke. Stroke. 2006;37:859–64.
preliminary report. Radiology. 2008;249:81–7.
11. Rauh R, Fischereder M, Spengel FA.  Transesophageal echocar-
31. Vaitkus PT, Barnathan ES.  Embolic potential, prevention and

diography in patients with focal cerebral ischemia of unknown
management of mural thrombus complicating anterior myocardial
cause. Stroke. 1996;27:691–4.
infarction: a meta-analysis. J Am Coll Cardiol. 1993;22:1004–9.
12. Strandberg M, Marttila RJ, Helenius H, Hartiala J. Transoesophageal
32. Srichai MB, Junor C, Rodriguez LL, et  al. Clinical, imaging,

echocardiography in selecting patients for anticoagulation after
and pathological characteristics of left ventricular thrombus: a
ischaemic stroke or transient ischaemic attack. J Neurol Neurosurg
comparison of contrast-enhanced magnetic resonance imaging,
Psychiatry. 2002;73:29–33.
transthoracic echocardiography, and transesophageal echocar-
13. Nikolaou K, Flohr T, Knez A, et al. Advances in cardiac CT imag-
diography with surgical or pathological validation. Am Heart J.
ing: 64-slice scanner. Int J Cardiovasc Imaging. 2004;20:535–40.
2006;152:75–84.
49  Ischemic Stroke: The Role of Cardiac CT 645

33. Weinsaft JW, Kim HW, Shah DJ, et al. Detection of left ventricular 52. Mas JL, Arquizan C, Lamy C, et  al. Recurrent cerebrovascular
thrombus by delayed-enhancement cardiovascular magnetic reso- events associated with patent foramen ovale, atrial septal aneurysm,
nance prevalence and markers in patients with systolic dysfunction. or both. N Engl J Med. 2001;345:1740–6.
J Am Coll Cardiol. 2008;52:148–57. 53. Mylonakis E, Calderwood SB. Infective endocarditis in adults. N
34. Bittencourt MS, Achenbach S, Marwan M, et  al. Left ven- Engl J Med. 2001;345:1318–30.
tricular thrombus attenuation characterization in cardiac com- 54. Cabell CH, Pond KK, Peterson GE, et  al. The risk of stroke and
puted tomography angiography. J Cardiovasc Comput Tomogr. death in patients with aortic and mitral valve endocarditis. Am
2012;6:121–6. Heart J. 2001;142:75–80.
35. Holmes DR, Reddy VY, Turi ZG, et al. PROTECT AF Investigators. 55. Di Salvo G, Habib G, Pergola V, et  al. Echocardiography pre-
Percutaneous closure of the left atrial appendage versus warfarin dicts embolic events in infective endocarditis. J Am Coll Cardiol.
therapy for prevention of stroke in patients with atrial fibrillation: a 2001;37:1069–76.
randomized non-inferiority trial. Lancet. 2009;374:534–42. 56. Daniel WG, Mugge A, Martin RP, et al. Improvement in the diag-
36. Garcia-Fernandez MA, Perez-David E, Quiles J, et al. Role of left nosis of abscesses associated with endocarditis by transesophageal
atrial appendage obliteration in stroke reduction in patients with echocardiography. N Engl J Med. 1991;324:795–800.
mitral valve prosthesis: a transesophageal echocardiographic study. 57. Habets J, Tanis W, Reitsma JB, et al. Are novel non-invasive imag-
J Am Coll Cardiol. 2003;42:1253–8. ing techniques needed in patients with suspected prosthetic heart
37. Hur J, Kim YJ, Lee HJ, et  al. Left atrial appendage thrombi valve endocarditis? A systematic review and meta-analysis. Eur
in stroke patients: detection with two-phase cardiac CT angi- Radiol. 2015;25:2125–33.
ography versus transesophageal echocardiography. Radiology. 58. Benjamin EJ, Plehn JF, D'Agostino RB, et al. Mitral annular calci-
2009;251:683–90. fication and the risk of stroke in an elderly cohort. N Engl J Med.
38. Kim SC, Chun EJ, Choi SI, et al. Differentiation between spontane- 1992;327:374–9.
ous echocardiographic contrast and left atrial appendage thrombus 59. Higgins J, Mayo J, Skarsgard P.  Cardiac computed tomography
in patients with suspected embolic stroke using two-phase multide- facilitates operative planning in patients with mitral calcification.
tector computed tomography. Am J Cardiol. 2010;106:1174–81. Ann Thorac Surg. 2013;95:e9–11.
39. Hur J, Kim YJ, Lee HJ, et al. Dual-enhanced cardiac CT for detec- 60. Budoff MJ, Takasu J, Katz R, et al. Reproducibility of CT measure-
tion of left atrial appendage thrombus in patients with stroke: a pro- ments of aortic valve calcification, mitral annulus calcification, and
spective comparison study with transesophageal echocardiography. aortic wall calcification in the multi-ethnic study of atherosclerosis.
Stroke. 2011;42:2471–7. Acad Radiol. 2006;13:166–72.
40. Budoff MJ.  Maximizing dose reductions with cardiac CT.  Int J 61. Amarenco P, Cohen A, Tzourio C, et  al. Atherosclerotic disease
Cardiovasc Imaging. 2009;25:279–87. of the aortic arch and the risk of ischemic stroke. N Engl J Med.
41. Castillo JG, Silvay G. Characterization and management of cardiac 1994;331:1474–9.
tumors. Semin Cardiothorac Vasc Anesth. 2010;14:6–20. 62. The French Study of Aortic Plaques in Stroke Group. Atherosclerotic
42. Bjessmo S, Ivert T. Cardiac myxoma: 40 years’ experience in 63 disease of the aortic arch as a risk factor for recurrent ischemic
patients. Ann Thorac Surg. 1997;63:697–700. stroke. N Engl J Med. 1996;334:1121–221.
43. Sparrow PJ, Kurian JB, Jones TR, Sivananthan MU. MR imaging 63. Tunick PA, Rosenzweig BP, Katz ES, Freedberg RS, Perez JL,
of cardiac tumors. Radiographics. 2005;25:1255–76. Kronzon I.  High risk for vascular events in patients with pro-
44. Kim EY, Choe YH, Sung K, Park SW, Kim JH, Ko YH. Multidetector truding aortic atheromas: a prospective study. J Am Coll Cardiol.
CT and MR imaging of cardiac tumors. Korean J Radiol. 1994;23:1085–90.
2009;10:164–75. 64. Rana BS, Monaghan MJ, Ring L, Shapiro LS, Nihoyannopoulos
45. Scheffel H, Baumueller S, Stolzmann P, et  al. Atrial myxomas P. The pivotal role of echocardiography in cardiac sources of embo-
and thrombi: comparison of imaging features on CT.  AJR Am J lism. Eur J Echocardiogr. 2011;12:25–31.
Roentgenol. 2009;192:639–45. 65. Pepi M, Evangelista A, Nihoyannopoulos P, et  al. European

46. Hong YJ, Hur J, Kim YJ, et  al. Dual-energy cardiac computed Association of Echocardiography. Recommendations for echocar-
tomography for differentiating cardiac myxoma from thrombus. Int diography use in the diagnosis and management of cardiac sources
J Card Imaging. 2014;30:121–8. of embolism: European Association of Echocardiography. Eur J
47. Kerut EK, Norfleet WT, Plotnick GD, et al. Patent foramen ovale: Echocardiogr. 2010;11:461–76.
a review of associated conditions and the impact of physiological 66. Ko Y, Park JH, Yang MH, et al. Significance of aortic atheroscle-
size. J Am Coll Cardiol. 2001;38:613–23. rotic disease in possibly embolic stroke: 64-multidetector row com-
48. Woods TD, Patel A.  A critical review of patent foramen ovale puted tomography study. J Neurol. 2010;257:699–705.
detection using saline contrast echocardiography: when bubbles lie. 67. Taylor AJ, Cerqueira M, Hodgson JM, Mark D, Min J, O'Gara P,
J Am Soc Echocardiogr. 2006;19:215–22. et al. ACCF/SCCTA/ACR/AHA/ASE/ASNC/NASCI/SCAI/SCMR
49. Pinto FJ. When and how to diagnose patent foramen ovale. Heart. 2010 appropriate use criteria for cardiac computed tomography. A
2005;91:438–40. report of the American college of cardiology foundation appropri-
50. Kim YJ, Hur J, Shim CY, et  al. Patent foramen ovale: diagnosis ate use criteria task force, the society of cardiovascular computed
with multidetector CT: comparison with transesophageal echocar- tomography, the American college of radiology, the American
diography. Radiology. 2009;250:61–7. heart association, the American society of echocardiography, the
51. Kim YJ, Hur J, Choe KO, et al. Interatrial shunt detected in coro- American society of nuclear cardiology, the north American society
nary computed tomography angiography: differential features of a for cardiovascular imaging, the society for cardiovascular angiogra-
patent foramen ovale and an atrial septal defect. J Comput Assist phy and interventions, and the society for cardiovascular magnetic
Tomogr. 2008;32:663–7. resonance. J Am Coll Cardiol. 2010;56:1864–94.
 Incidental Findings on CT Angiography
and How to Manage Them 50
Seung Min Yoo, Hwa Yeon Lee, and Charles S. White

Because of recent technologic advances, coronary CT angi- full field of view” provides full coverage (35–40 cm2) with
ography (CTA) has become a potent imaging tool in the thicker collimation of 2–5  mm encompassing the entire
evaluation of patients with atypical chest pain (i.e., low to transverse extent of the chest. The latter is an additional
intermediate risk for both stable angina and acute coronary reconstruction that may be obtained in order to provide and
syndrome) [1]. Coronary CTA is expected to be increasingly overread for extracardiac incidental findings [3, 4]. Because
used in clinical practice due to its high sensitivity and a nega- of the small size of the coronary arteries, reconstruction
tive predictive value that approaches 100% [2]. A major rea- with a small field of view is required to maximize spatial
son for performing coronary CTA is to evaluate coronary resolution by reducing pixel size, facilitating precise analy-
artery disease, and it is vital for interpreting physicians to sis of the degree of stenosis and characterize plaque in these
focus on identifying abnormalities in the coronary arteries. small structures. However, because all structures within a
However, neighboring organs such as central portions of the given CT slice have already been exposed to radiation, wide
lungs, mediastinum, aorta, esophagus, upper abdominal field of view images can be obtained without additional
organs, and thoracic skeleton are typically included in the radiation dose.
examination, regardless of whether a restricted field of view The prevalence of incidental extracardiac findings (Tables
(Fig. 50.1) or a full field of view is used for the evaluation of 50.1 and 50.2) demonstrated on coronary CTA is fairly high
coronary arteries. Thus, it is critical that all interpreting phy- ranging up to 67.0% on the wide field of view image with
sicians be careful not to overlook extracardiac incidental considerable variation in frequency due to study design and
findings (Fig.  50.2). This chapter will discuss the most definition [3–17]. Of these, the prevalence of clinically
important incidental findings and an approach to their fol- important findings needing further work-up or management
low-­up and management. is less common and ranges from 1.2% to 22.7% [18]. It is
certainly true that more malignant lesions (Fig. 50.3) will be
found in wide field of view because a substantial number of
 se of a Small Versus Wide Field of View
U lung cancers occur in the peripheral portion of the lungs, and
on Coronary CT Angiography most of breast tissue is excluded from a small field of view
on coronary CTA. Earlier detection of lung or breast cancer
A discussion of the concept of the small and wide fields of may lead to curable surgical resection and favorable out-
view is a prerequisite to understand the ongoing debate comes. In spite of this potential benefit, several reports [14,
regarding appropriate field of view in interpreting coronary 19–22] have advocated exclusive use of small of view,
CTA.  The “small or restricted field of view” is a primary although a larger group has favored the additional use of a
tool for analysis of coronary arteries using 16–25  cm2 of wide field of view in interpreting coronary CTA [3, 4, 9, 11,
coverage and thin collimation <1 mm. In contrast, “wide or 12, 15–18, 23–28]. There are differing opinions even within
specialties [4, 14, 18–22, 27, 29]. The main reason for the
S. M. Yoo controversy is the high prevalence of false-positive findings
Department of Radiology, CHA University Bundang Medical (i.e., low prevalence of malignancy less than 0.4%) on the
Center, Bundang, South Korea
larger field of view and lack of proven benefit of CT screen-
H. Y. Lee ing of lung cancer prior to the publication of the National
Smile Radiologic Clinic, Seoul, South Korea
Lung Screening Trial (NLST) [19, 20]. Work-up of numer-
C. S. White (*) ous indeterminate lung nodules that ultimately prove to be
Department of Diagnostic Radiology, University of Maryland
Medical Center, Baltimore, MD, USA
benign may lead to extra cost, morbidity related to down-
e-mail: [email protected] stream testing and intervention, a low but potential risk of

© Humana Press 2019 647

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_50
648 S. M. Yoo et al.

a c

Fig. 50.1  Representative small (a) and wide (b) field of views on dedi- includes the central lung parenchyma, aorta, mediastinum, bony struc-
cated coronary CTA compared with the wide field of view (c) of triple tures, and abdominal organs. The wide field of view on dedicated coro-
rule-out protocol. The small field of view on dedicated coronary CTA nary CTA includes about two-thirds of total lung volume

a b

Fig. 50.2  Various important extracardiac findings demonstrated on disease (arrowheads in c), anterior mediastinal mass (arrowheads in d),
small field of view coronary CTA image. Stanford type B intramural and central lung cancer (arrows in e) are demonstrated even using a
hematoma (arrowheads in a), central pulmonary emboli (arrowheads in small coronary CTA field of view. Thus, all interpreters should be
b), esophageal wall thickening in a patient with gastroesophageal reflux familiar with important extracardiac findings
50  Incidental Findings on CT Angiography and How to Manage Them 649

c d

Fig. 50.2 (continued)

radiation-induced cancer, and anxiety on the part of both to suggest that a wide field of view of coronary CTA be
physicians and patients. reconstructed and evaluated in the subset of patients who fit
Results from the NLST indicated a lung cancer and over- the entry criteria of the NLST.
all mortality benefit for low-dose chest CT screening com- However, even this suggestion should be viewed with cau-
pared with screening with chest radiography. The randomized tion because of differences in the patient selection. The indi-
trial of more than 53,000 asymptomatic subjects who had a viduals who enrolled in the NLST were entirely asymptomatic,
smoking history of more than 30 pack-years and were whereas nearly all patients who undergo coronary CTA are
55–74 years of age showed a 20.3% reduction of lung cancer referred for chest pain. Such a distinct referral pattern may
mortality in the CT arm compared with chest radiography affect the prevalence and characteristics of incidental extracar-
arm [30]. Based on the result of NLST, it seems reasonable diac findings. In addition, the specifics of imaging protocols
650 S. M. Yoo et al.

are different between low-dose chest CT and coronary Moreover, there is inconsistency in the use of wide field
CTA.  The former is a nonenhanced CT with low radiation of view in the radiologic practices depending on body part.
exposure, whereas the latter is obtained with intravenous con- Only small field of view images are typically reconstructed
trast media and a higher radiation exposure due to ECG gating, in other parts of body such as the orbits, inner ear, sinuses,
potentially permitting lung nodules and certain other incidental and thoracic spine CT, even though radiation is received
findings to be more easily identified. For example, aortic dis- through the entire transverse CT section [29]. Lastly, it
section would be directly visible on coronary CTA but would should be remembered that the typical z-axis coverage of
be unlikely to be detected on a lung cancer screening study. dedicated coronary CTA does not include the upper one-­
third of the entire chest volume (i.e., tracheal carina to lung
apices) [15]. This may lead to the false assumption that the
Table 50.1  Potentially clinically significant incidental extracardiac
findings entirety of both lungs is normal among patients undergoing a
normal dedicated coronary CTA and their treating clinicians.
Incidental finding Prevalence (%)
Pulmonary nodule ≥4 mm 0.4–16.5%
Therefore, until a randomized trial is available to provide an
Pulmonary consolidation 0.4–6.2% overview about cost-effectiveness and outcome benefits,
Marked mediastinal lymphadenopathy 0.1–2.3% reconstruction of only a small field of view may be accept-
Hiatal hernia 0.2–6.4% able in patients undergoing dedicated coronary CTA, partic-
Aortic dissection 0.0–0.3% ularly in those of different age and smoking habits as
Aortic aneurysm 0.3–1.6% compared with enrollees in NLST [29].
Pulmonary embolism 0.0–1.9%
Breast nodule 0.0–0.6%
Fracture 0.0–0.3% Table 50.2  Benign incidental extracardiac findings
Metastatic bone destruction Frequency not available Incidental finding Prevalence (%)
Pleural effusion 0.1–4.0% Pulmonary nodule <4 mm 1.7–9.3%
Adrenal nodule 0.0–0.8% Benign hepatic cyst 1.1–6.6%
Indeterminate hepatic nodule 0.0–2.3% Simple renal cyst 0.1–0.3%
Cholelithiasis 0.1–3.6% Benign adrenal adenoma 0.1–0.6%

a b

Fig. 50.3  Various important extracardiac findings identified only on nary artery bifurcation in a 42-year-old female patient. This nodule
a wide field of view coronary CTA image. (a, b) Solid pulmonary was not visible on a small field of view image. Breast ultrasonogra-
nodule 5  mm in diameter (arrow in a) is demonstrated in the left phy was recommended in the patient for further evaluation. D. Right
lower lobe on wide field of view image at the level of inferior cardiac seventh rib destruction (arrowheads on d) is noted on a wide field of
margin in a 42-year-old male smoker. Note that the nodule is not vis- view bone setting image at the level of the right main pulmonary
ible on a small field of view image (b). (c) Enhancing 8 mm nodule artery in an 81-year-old man. However, this lesion is not demon-
(arrows in c) is demonstrated in the lateral portion of the left breast strated on a small field of view image (e). The patient ultimately
on a wide field of view coronary CTA image at the level of pulmo- proved to have a rectal cancer
50  Incidental Findings on CT Angiography and How to Manage Them 651

c d

Fig. 50.3 (continued)

In the setting of acute chest pain, there is less debate than field of view [32], interpreters should be careful not to miss
in stable angina regarding the use of a wide field of view important extracardiac findings. Dedicated analysis of the
(either dedicated coronary CTA or triple rule-out protocol) coronary arteries with submillimeter collimation should be
because an alternative cause of acute chest pain such as performed with a small field of view, and interpretation with
peripheral pulmonary embolism (Fig.  50.4), pneumonia, three major image settings (mediastinal, lung, and bone win-
pneumothorax, rib fracture, cholecystitis, or pancreatitis can dow settings) should be routinely performed to analyze all
be identified when the coronary artery component is negative organs visualized within the field of view beyond the coro-
[13, 27]. According to one study, clinically important extra- nary arteries, even if only a small field of view image is used
cardiac findings that potentially changed patient manage- [23, 33]. In contrast, if wide field of view images are avail-
ment were identified up to 5% of patients presenting with able, the coronary arteries should be interpreted with small
acute chest pain. Thus, a stronger case can be made for a field of view image using submillimeter collimation, whereas
large field of view reconstruction in these patients [13, 31]. extracardiac findings should be separately analyzed on wide
field of view images with the three different window settings
using 2–5 mm collimation. The typical Hounsfield unit (HU)
 ractical Tips to Avoid Overlooking
P levels and widths for mediastinal, lung, and bone window
Important Incidental Findings settings are 50 and 350, −500 and 1800, and 500 and 2000,
respectively [33]. As a rule axial images are the best option
Because at least 8–22% (mean, 14%) of lung parenchyma, to identify extracardiac findings. Although it is not routinely
mediastinum, bone, chest wall, aorta, pulmonary arteries, used in clinical practice, additional reconstruction of coronal
and upper abdominal organs are included even in a small or sagittal views may provide additional information to
652 S. M. Yoo et al.

a b

Fig. 50.4  Peripheral pulmonary embolism in a segmental pulmonary field of view image (arrow on a), but not the small field of view image
artery in the right lower lobe is identified only on a wide field of view (b), thus showing the potential benefits of wide field of view imaging to
image at the level of the left atrium in a 71-year-old woman. The patient provide an alternative cause of chest pain. Although most central pul-
presented with acute chest pain and the primary concern was acute monary emboli are visible on a small field of view image, this may not
coronary syndrome. Dedicated coronary CTA showed normal coronary be the case for more peripheral emboli
arteries. Segmental pulmonary embolus is demonstrated on the wide

determine the precise location of a pulmonary nodule near a nodule. In contrast, eccentric calcification in a pulmonary
fissure or to identify vertebral pathology such as a compres- nodule does not necessarily indicate a benign etiology.
sion fracture [33]. If an extracardiac finding is encountered, If an indeterminate nodule is found on coronary CTA, the
a multidisciplinary approach can be made used to select the nodule can be followed by the Fleischner Society recommenda-
next appropriate step. Importantly, those who interpret car- tions. The recommendations classify lung nodules based on size
diac CTA should keep in mind that incidental extracardiac and the presence or absence of risk factors such as smoking
findings should always be compared on any previous studies history or known primary cancer. Follow-up CT is not recom-
to avoid unnecessary and potentially costly follow-up mended in small (<4 mm) pulmonary nodules in a nonsmoker.
examinations. In contrast, indeterminate pulmonary nodule more than 4 mm
irrespective of risk factors should be followed by low-dose chest
CT for up to 2 years to confirm resolution or stability. Stability
I ncidental Findings and How to  over 2 years typically indicates that a solid pulmonary nodule is
Manage Them benign. However, an indeterminate pulmonary nodule with
ground-glass attenuation should be followed more than 2 years
Lung because ground-­glass or part-solid nodules that prove to be
malignant typically have a longer doubling time [34].
An incidental pulmonary nodule is the most frequent finding
in the evaluation of wide field of view on coronary CTA. The
prevalence of all incidental pulmonary nodules and indeter- Mediastinum
minate pulmonary nodules needing further work-up is 0.9–
36.2% and 0.4–16.5%, respectively [18]. A pulmonary nodule The most frequent significantly abnormal finding in the medi-
is defined as a relatively well-defined area of increased atten- astinum is lymph node enlargement. Lymph node size less
uation less than 3 cm in diameter. Certain characteristics can than 10  mm in short diameter is usually benign. If there is
lead to a specific diagnosis. The presence of fat (HU of less lymph node enlargement ≥10 mm in short diameter (Fig. 50.6)
than −10) in the pulmonary nodule on nonenhanced CT is on coronary CTA, it should be further characterized by its
diagnostic of benign hamartoma (Fig. 50.5). Central, diffuse, shape and as well as any history of malignancy. The presence
lamellated, or popcorn shape calcification in a smoothly mar- of a central fatty hilum or calcification in an enlarged lymph
ginated nodule is a typical CT finding of benign pulmonary node or the concomitant presence of pneumonia may indicate
50  Incidental Findings on CT Angiography and How to Manage Them 653

a b

Fig. 50.5  Incidental pulmonary hamartoma in a 60-year-old female view, and magnified (arrows in c) images at the level of the main pul-
patient. Pulmonary nodule is demonstrated in the right middle lobe in monary artery. Note fat attenuation (arrows in c) indicative of pulmo-
enhanced (arrowheads in a), nonenhanced (arrows in b) wide field of nary hamartoma

a benign etiology. In contrast, lymph node enlargement that

lacks benign features should be followed by CT or further
evaluated [23, 33]. Conglomerate lymph node enlargement
with peripheral rim or diffuse enhancement can be a finding in
tuberculous or malignant lymph node enlargement. Another
frequent mediastinal finding is an esophageal abnormality
such as hiatal hernia or esophageal wall thickening. The gas-
troesophageal junction is located above the diaphragmatic hia-
tus on CT in patients with a hiatal hernia. This finding may
explain symptoms of chest pain due to gastroesophageal reflux
disease. If gastroesophageal reflux disease is suspected, endo-
scopic evaluation can be performed.

Aorta

The clinical presentation of aortic dissection and acute coro-

nary syndrome can be quite similar. Thus, the aorta should
Fig. 50.6  Indeterminate left hilar lymph node enlargement on a small be evaluated carefully on coronary CTA even when using a
field of view image in a 66-year-old male. Left hilar lymph node small field of view. Aortic dilatation or aneurysm is a f­ requent
enlargement (arrowheads) measuring about 12 mm in short diameter is incidental finding on coronary CTA.  An aortic aneurysm
demonstrated on a small field of view image of coronary CTA at the
level of the left atrium. Follow-up chest CT was recommended (Fig.  50.7) is defined as aortic dilatation more than 150%
654 S. M. Yoo et al.

Thoracic Skeleton and Pleura

It is important to identify the presence of acute rib fractures

because they may be an alternative cause of chest pain [33].
It is of value to remember that rib fractures can occur without
a history of blunt trauma, for example, due to chronic cough
or stress-related injury. It is also important to identify any
metastatic rib lesion (Fig. 50.3d) or vertebral destruction to
avoid unnecessary additional examinations to confirm the
diagnosis. Pleural effusion is another common incidental
finding. Pleural effusion in the dependent portion of the chest
with low density (<20 HU) and no enhancement usually
indicates a transudate. Although the presence of pleural lin-
ear or nodular enhancement, loculation, or high attenuation
of content (HU >20) may suggest exudative or complicated
effusion, the absence of such findings does not ­completely
Fig. 50.7  Ascending aortic aneurysm on a small field of view image in exclude an exudative or malignant pleural effusion [33].
a 60-year-old female patient. This is the same patient with Fig.  50.5
who has hamartoma (arrowheads) in the right middle lobe. Note ascend-
ing aortic aneurysm (arrows) measuring approximately 50  mm in
diameter
Upper Abdominal Organs

Frequent abdominal incidental findings are cysts (Fig. 50.8),

(often 5  cm in diameter) compared with size of a normal hemangiomas, adrenal nodules (Fig. 50.9), and gallbladder or
aorta. Six months to 1 year CT follow-up is recommended to renal stones. Most hepatic nodules less than 10 mm in diam-
evaluate changes in the aortic diameter. If there is greater eter are benign in asymptomatic patients. Standard recom-
than 1 cm aortic dilatation over the course of a year, surgical mendations such as the Fleischner guidelines for indeterminate
intervention should be considered. lung nodule do not exist for follow-up of incidental hepatic
nodules in low-risk patients [23]. Peripheral nodular enhance-
ment on the arterial phase of CT is typical for a hemangioma.
Occasionally, hepatocellular carcinoma or hypervascular
Pulmonary Artery metastasis such as melanoma may be a cause of hepatic nod-
ule with high attenuation on coronary CTA. Thus, the possi-
Most central pulmonary emboli may be visible even in a bility of malignant hepatic nodule should be considered in
small field of view image on coronary CTA. However, it patients with CT features of liver cirrhosis or history or a pri-
should be stressed that dedicated coronary CTA targets mary focus of cancer. According to the recently published
the left circulation but not the right. Thus, the visibility of Liver Imaging-Reporting and Data System (LI-RADS), a
central pulmonary emboli depends on the degree of nodule demonstrated on CT in patients at high risk for hepa-
enhancement in the pulmonary arteries. In contrast, tocellular carcinoma (e.g., liver cirrhosis) should be consid-
peripheral pulmonary embolism in segmental or subseg- ered at least as an indeterminate nodule (≥LR-3), irrespective
mental pulmonary arteries may only be noted on a wide of size. In this scenario, dynamic abdomen CT is recom-
field of view. In patients with pulmonary embolism dem- mended [35]. It should be kept in mind that coronary CTA is
onstrated on CTA, lower extremity Doppler ultrasonogra- not the optimal examination to identify hepatic nodules
phy can be recommended to evaluate for deep vein because it is not obtained in a parenchymal venous phase.
thrombosis. An incidental adrenal nodule is another common abdominal
finding within the field of view of coronary CTA. Hounsfield
units less than 10 on nonenhanced CT images often indicate a
Breast benign adenoma because most adenomas contain a fatty com-
ponent. However, adrenal adenomas with scanty fat often have
The prevalence of breast nodules found on coronary CTA is a higher CT attenuation on nonenhanced CT. In such cases, a
up to 0.6% [3, 4, 14, 26, 32]. If a breast nodule (Fig. 50.3c) dedicated abdominal CT study showing a characteristic wash-
is identified on a wide field of view coronary CTA image, out pattern is helpful to discriminate an adenoma from a malig-
ultrasonography and mammography should be performed to nant lesion. Greater than 40% washout on a 15 min delayed CT
exclude early breast cancer. image indicates an adenoma [23].
50  Incidental Findings on CT Angiography and How to Manage Them 655

a­ nalysis with a wide field of view may be valuable to identify

an alternative cause of chest pain. For patients with stable
angina, a small field of view image on coronary CTA may be
acceptable to evaluate extracardiac findings.

References
1. Taylor AJ, Cerqueira M, Hodgson JM, Mark D, Min J, O'Gara P,
Rubin GD, American College of Cardiology Foundation Appropriate
Use Criteria Task Force; Society of Cardiovascular Computed
Tomography; American College of Radiology; American Heart
Association; American Society of Echocardiography; American
Society of Nuclear Cardiology; North American Society for
Cardiovascular Imaging; Society for Cardiovascular Angiography
and Interventions; Society for Cardiovascular Magnetic Resonance,
Kramer CM, Berman D, Brown A, Chaudhry FA, Cury RC, Desai
MY, Einstein AJ, Gomes AS, Harrington R, Hoffmann U, Khare
Fig. 50.8  Hepatic cyst in a 50-year-old male patient. Hepatic cyst with R, Lesser J, McGann C, Rosenberg A, Schwartz R, Shelton M,
typical low attenuation (i.e., HU near zero) (arrowheads) is noted on a Smetana GW, Smith SC Jr. ACCF/SCCT/ACR/AHA/ASE/ASNC/
wide field of view image at the level of dome of liver NASCI/SCAI/SCMR 2010 appropriate use criteria for cardiac
computed tomography. A report of the American college of car-
diology foundation appropriate use criteria task force, the society
of cardiovascular computed tomography, the American college of
radiology, the American heart association, the American society
of echocardiography, the American society of nuclear cardiol-
ogy, the north American society for cardiovascular imaging, the
society for cardiovascular angiography and interventions, and the
society for cardiovascular magnetic resonance. J Am Coll Cardiol.
2010;56(22):1864–94.
2. Kim HR, Yoo SM, Rho JY, Lee HY, White CS. MDCT evaluation
of atherosclerotic coronary artery disease: what should radiologists
know? Int J Cardiovasc Imaging. 2014;30(Suppl 1):1–11.
3. Johnson KM, Dennis JM, Dowe DA. Extracardiac findings on coro-
nary CT angiograms: Limited versus complete image review. AJR
Am J Roentgenol. 2010;195(1):143–8.
4. Onuma Y, Tanabe K, Nakazawa G, Aoki J, Nakajima H, Ibukuro K,
Hara K. Noncardiac findings in cardiac imaging with multidetector
computed tomography. J Am Coll Cardiol. 2006;48:402–6.
5. Hunold P, Schmermund A, Seibel RM, Grönemeyer DH, Erbel
R.  Prevalence and clinical significance of accidental findings in
electron-beam tomographic scans for coronary artery calcification.
Eur Heart J. 2001;22(18):1748–58.
6. Horton KM, Post WS, Blumenthal RS, Fishman EK. Prevalence of
significant noncardiac findings on electron-beam computed tomog-
raphy coronary artery calcium screening examinations. Circulation.
2002;106(5):532–4.
Fig. 50.9  Nonspecific left adrenal nodule in a 50-year-old male 7. Schragin JG, Weissfeld JL, Edmundowicz D, Strollo DC, Fuhrman
patient. Nonspecific left adrenal nodule (arrows) measuring approxi- CR.  Non-cardiac findings on coronary electron beam computed
mately 10 mm is noted on a small field of view image at the level of tomography scanning. J Thorac Imaging. 2004;19(2):82–6.
adrenal gland. In this instance, the characteristic washout pattern is 8. Dewey M, Schnapauff D, Teige F, Hamm B. Non-cardiac findings
helpful to discriminate an adenoma from malignant lesion. Greater than on coronary computed tomography and magnetic resonance imag-
40% washout on a 15 min delayed CT image indicates an adenoma ing. Eur Radiol. 2007;17(8):2038–43.
9. Kirsch J, Araoz PA, Steinberg FB, Fletcher JG, McCollough CH,
Williamson EE. Prevalence and significance of incidental extracar-
diac findings at 64-multidetector coronary CTA. J Thorac Imaging.
Conclusion 2007;22(4):330–4.
10. Kawano Y, Tamura A, Goto Y, Shinozaki K, Zaizen H, Kadota
Many important extracardiac findings are demonstrated even J. Incidental detection of cancers and other non-cardiac abnormali-
ties on coronary multislice computed tomography. Am J Cardiol.
on the small field of view image of coronary CTA.  Thus, 2007;99(11):1608–9.
interpreting physicians irrespective of specialty should be 11. Northam M, Koonce J, Ravenel JG. Pulmonary nodules detected at
familiar with various important extracardiac findings. In the cardiac CT: comparison of images in limited and full fields of view.
acute chest pain setting, additional reconstructions and AJR Am J Roentgenol. 2008;191(3):878–81.
656 S. M. Yoo et al.

12. Kim JW, Kang EY, Yong HS, Kim YK, Woo OH, Oh YW, Lee KY, 24. Douglas PS, Cerqueria M, Rubin GD, Chin AS. Extracardiac find-
Han H. Incidental extracardiac findings at cardiac CT angiography: ings: what is a cardiologist to do? JACC Cardiovasc Imaging.
comparison of prevalence and clinical significance between precon- 2008;1(5):682–7.
trast low-dose whole thoracic scan and postcontrast retrospective 25. Greenberg-Wolff I, Uliel L, Goitein O, Shemesh J, Rozenman J, Di
ECG-gated cardiac scan. Int J Cardiovasc Imaging. 2009;25(Suppl Segni E, Konen E.  Extra-cardiac findings on coronary computed
1):75–81. tomography scanning. Isr Med Assoc J. 2008;10(11):806–8.
13. Lehman SJ, Abbara S, Cury RC, Nagurney JT, Hsu J, Goela
26. Yiginer O, Bas S, Pocan S, Yildiz A, Alibek S. Incidental findings
A, Schlett CL, Dodd JD, Brady TJ, Bamberg F, Hoffmann of cardiac MSCT: who might benefit from scanning the entire tho-
U.  Significance of cardiac computed tomography incidental find- rax on Ca score imaging? Int J Cardiol. 2010;140(2):239–41.
ings in acute chest pain. Am J Med. 2009;122(6):543–9. 27. Rumberger JA.  Noncardiac abnormalities in diagnostic cardiac

14. Machaalany J, Yam Y, Ruddy TD, Abraham A, Chen L, Beanlands computed tomography: within normal limits or we never looked! J
RS, Chow BJ.  Potential clinical and economic consequences of Am Coll Cardiol. 2006;48(2):407–8.
noncardiac incidental findings on cardiac computed tomography. J 28. Colletti PM.  Incidental findings on cardiac imaging. AJR Am J
Am Coll Cardiol. 2009;54(16):1533–41. Roentgenol. 2008;191(3):882–4.
15. Haller S, Kaiser C, Buser P, Bongartz G, Bremerich J.  Coronary 29. White CS.  The pros and cons of searching for extracardiac find-
artery imaging with contrast-enhanced MDCT: extracardiac find- ings at cardiac CT: use of a restricted field of view is acceptable.
ings. AJR Am J Roentgenol. 2006;187(1):105–10. Radiology. 2011;261(2):338–41.
16. Lee CI, Tsai EB, Sigal BM, Plevritis SK, Garber AM, Rubin
30. National Lung Screening Trial Research Team, Aberle DR, Berg
GD. Incidental extracardiac findings at coronary CT: clinical and CD, Black WC, Church TR, Fagerstrom RM, Galen B, Gareen IF,
economic impact. AJR Am J Roentgenol. 2010;194(6):1531–8. Gatsonis C, Goldin J, Gohagan JK, Hillman B, Jaffe C, Kramer BS,
17. Kim TJ, Han DH, Jin KN, Won Lee K.  Lung cancer detected at Lynch D, Marcus PM, Schnall M, Sullivan DC, Sullivan D, Zylak
cardiac CT: prevalence, clinicoradiologic features, and importance CJ. The national lung screening trial: overview and study design.
of full-field-of-view images. Radiology. 2010;255(2):369–76. Radiology. 2011;258(1):243–53.
18. Earls JP. The pros and cons of searching for extracardiac findings 31. Takakuwa KM, Halpern EJ. Evaluation of a “triple rule-out” coro-
at cardiac CT: studies should be reconstructed in the maximum nary CT angiography protocol: use of 64-Section CT in low-to-­
field of view and adequately reviewed to detect pathologic findings. moderate risk emergency department patients suspected of having
Radiology. 2011;261(2):342–6. acute coronary syndrome. Radiology. 2008;248(2):438–46.
19. Budoff MJ, Fischer H, Gopal A. Incidental findings with cardiac CT 32. Fantauzzi J, MacArthur A, Lu M, Jeudy J, White CS. Quantitative
evaluation: should we read beyond the heart? Catheter Cardiovasc assessment of percentage of lung parenchyma visualized on car-
Interv. 2006;68(6):965–73. diac computed tomographic angiography. J Comput Assist Tomogr.
20. Budoff MJ, Gopal A.  Incidental findings on cardiac computed 2010;34(3):385–7.
tomography. Should we look? J Cardiovasc Comput Tomogr. 33. Teague SD, Rissing S, Mahenthiran J, Achenbach S.  Learning to
2007;1(2):97–105. interpret the extracardiac findings on coronary CT angiography
21. Iribarren C, Hlatky MA, Chandra M, Fair JM, Rubin GD, Go examinations. J Cardiovasc Comput Tomogr. 2012;6(4):232–45.
AS, Burt JR, Fortmann SP.  Incidental pulmonary nodules on 34. Naidich DP, Bankier AA, MacMahon H, Schaefer-Prokop CM,
cardiac computed tomography: prognosis and use. Am J Med. Pistolesi M, Goo JM, Macchiarini P, Crapo JD, Herold CJ, Austin
2008;121(11):989–96. JH, Travis WD. Recommendations for the management of subsolid
22. Hlatky MA, Iribarren C.  The dilemma of incidental find-
pulmonary nodules detected at CT: a statement from the Fleischner
ings on cardiac computed tomography. J Am Coll Cardiol. society. Radiology. 2013;266(1):304–17.
2009;54(16):1542–3. 35. Mitchell DG, Bruix J, Sherman M, Sirlin CB.  LI-RADS (Liver
23. Sosnouski D, Bonsall RP, Mayer FB, Ravenel JG.  Extracardiac Imaging Reporting and Data System): summary, discussion, and
findings at cardiac CT: a practical approach. J Thorac Imaging. consensus of the LI-RADS management working group and future
2007;22(1):77–85. directions. Hepatology. 2015;61(3):1056–65.
 Part XII
Where We Are: The Bigger Picture
 Prognosis and Outcome: State
of the Evidence 51
Asim Rizvi, Hadi Mirhedayati Roudsari, James K. Min,
and Fay Y. Lin

Coronary artery disease (CAD) remains the leading cause of  rognostic Significance of CCTA for All-Cause
P
mortality in the developed world despite improvements in Mortality
diagnostic modalities and therapies targeted at reducing dis-
ease burden. The recent advent of coronary computed tomog- CCTA-evaluated CAD has demonstrated incremental prog-
raphy angiography (CCTA) has gained interest as a promising nostic value in chronic stable outpatients for ACM
noninvasive anatomic alternative to stress imaging tech- (Table 51.1). The earliest data from Min et al. determined the
niques for diagnosis of anatomically obstructive CAD. CCTA prognostic value of 16-detector row CCTA for ACM in 1127
has demonstrated high diagnostic accuracy for obstructive patients without known CAD [7]. After a 15-month follow-
CAD, with a sensitivity, specificity, positive predictive value ­up, the number of vessels affected by moderate or severe ste-
(PPV), and negative predictive value (NPV) of 95%, 83%, nosis as well as proximal left anterior descending artery
64%, and 99%, respectively, compared to invasive coronary stenosis demonstrated independent predictive value above
angiography (ICA) for >50% stenosis in the ACCURACY and beyond clinical risk factors (all p  <  0.001). Survival
(Assessment by Coronary Computed Tomographic decreased with higher Duke scores, with 96% survival for 2
Angiography of Individuals Undergoing Invasive Coronary moderate stenoses or 1 severe stenosis (p < 0.013) and 85%
Angiography) study and of 99%, 89%, 93%, and 100%, survival for moderate left main coronary artery stenosis
respectively, in the meta-analyses [1, 2]. More recently, the (p  <  0.001), respectively (Fig.  51.1). Notably, authors con-
PROMISE (PROspective Multicenter Imaging Study for cluded that a negative CCTA for CAD confers significantly
Evaluation of chest pain) study showed that CCTA is a safe decreased risk of ACM compared to the remainder of the
strategy for initial evaluation of CAD and demonstrated sim- population (0.3% vs. 4.8%; HR 0.12, 95% confidence interval
ilar downstream major adverse cardiovascular events [CI] 0.02–0.89) [7]. This study was followed up with a large
(MACE) with reduced risk of myocardial infarction (MI) or (n  =  5330), multicenter cohort utilizing 64-detector row
all-cause mortality (ACM) in comparison to stress imaging CCTA in patients without known CAD [8]. Over a period of
techniques (hazard ratio [HR], 0.66; 95% CI 0.44–1.00; 2.3 years, the HR for ACM in patients with obstructive CAD
p = 0.049) [3]. (≥70% stenosis) was 2.44 (95% CI 1.61–3.72, p  <  0.001).
Currently, scientific guidelines recommend the appropri- Risk increased with each additional diseased vessel, with an
ate use of CCTA for diagnosis of CAD in different patient HR of 2.23, 3.29, and 7.35 for 1-, 2-, and 3-­vessel diseases,
populations [4–6]. However, the prognostic value of CCTA respectively (all p < 0.001). Additionally, CCTA evaluation of
and the impact of CCTA on downstream outcomes are left ventricular ejection fraction (LVEF) improved prediction
important for patient risk stratification and management. The of ACM above and beyond CCTA measures of obstructive
purpose of this chapter is to discuss the current literature CAD, with worse outcomes for LVEF ≤50% (adjusted HR
using CCTA as a prognostic tool. [aHR] 1.56, 95% CI 1.04–2.36, p = 0.03) [8].
CCTA also demonstrates high diagnostic accuracy for
A. Rizvi (*) nonobstructive plaque, which has prognostic significance as
Department of Radiology, Dalio Institute of Cardiovascular well [9, 10]. Ostrom et al. examined ACM after 64-detector
Imaging, NewYork-Presbyterian Hospital and the Weill Cornell
Medical College, New York, NY, USA
row CCTA in 2538 symptomatic patients without known
CAD over 78  ±  12  months [11]. The survival rates for
Department of Radiology, Mayo Clinic, Rochester, MN, USA
e-mail: [email protected]; [email protected]
patients with 1-, 2-, and 3-vessel nonobstructive (<50% ste-
nosis) CAD versus obstructive (≥50% stenosis) CAD were
H. M. Roudsari · J. K. Min · F. Y. Lin
Department of Radiology, NewYork-Presbyterian Hospital and the
reported as 97.3%, 95%, and 93.1% versus 92.9%, 89.7%,
Weill Cornell Medical College, New York, NY, USA and 80%, respectively (Fig. 51.2). The risk-adjusted HR for

© Humana Press 2019 659

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_51
660 A. Rizvi et al.

Table 51.1  Studies evaluating the prognostic significance of coronary computed tomographic angiography for predicting all-cause mortality and
major adverse cardiovascular events
Patients
Author Year (n) Follow-up Study design Outcomes Population
Min et al. [7] 2007 1127 15 ± 4 months Retrospective ACM Patients with chest pain syndrome
Pundziute et al. [13] 2007 100 16 months Prospective Cardiac death, nonfatal Patients with known or suspected
MI, UA, CAD
revascularization
Ostrom et al. [11] 2008 5538 78 ± 12 months Retrospective ACM Patients with chest pain syndrome
Gaemperli et al. [14] 2008 220 14 ± 4 months Retrospective ACM, MI, UA, Patients without known CAD
revascularization
Carrigan et al. [16] 2009 227 2 ± 1 years Retrospective Cardiac death, nonfatal Patients with suspected CAD
MI, revascularization
Hadamitzky et al. [73] 2009 1150 18 months – Cardiac death, MI, UA, Patients with suspected CAD
revascularization
Gopal et al. [18] 2009 454 40 ± 9 months Prospective Cardiac death, MI Patients with suspected CAD
Abidov et al. [19] 2009 199 2 years Prospective Cardiac death, MI, UA, Patients with suspected CAD and
revascularization. inconclusive stress tests
Rubinshtein et al. [17] 2009 545 18 ± 6 months Prospective Cardiac death, MI, Patients with suspected CAD
revascularization
Motoyama et al. [67] 2009 1059 27 ± 10 months – Acute coronary Patients with suspected or known
syndrome CAD
Aldrovandi et al. [15] 2009 187 24 months Prospective Cardiac death, nonfatal Patients with suspected MI
MI, UA,
revascularization
Chow et al. [21] 2010 2076 16 ± 8 months Prospective Cardiac death, MI Patients with known or suspected
CAD
Min et al. [20] 2010 172 22 ± 5 months Prospective Cardiac death, MI, UA, Patients with known or suspected
revascularization CAD
Min et al. [8] 2010 5330 2 ± 1 years Retrospective ACM Patients without known CAD
Lin et al. [12] 2011 2583 3 ± 1 years Prospective ACM Patients without known CAD
Kristensen et al. [22] 2011 312 16 months Prospective Cardiac death, MI, UA, Patients with NSTEMI
revascularization
Beigel et al. [74] 2013 959 27 months Prospective ACM, UA, Patients with chest pain syndrome
revascularization
Dougoud et al. [28] 2014 218 6.9 years Prospective Cardiac death, nonfatal Patients with suspected CAD
MI, UA,
revascularization
Abbreviations: ACM all-cause mortality, MI myocardial infarction, UA unstable angina, CAD coronary artery disease

3-vessel nonobstructive CAD was elevated at 1.7 (95% CI <10% Framingham-estimated 10-year risk (3.4% over
1.49–2.05), as were those for 1-, 2-, and 3-vessel obstructive 3  years, log-­rank p  <  0.001) and among patients with no
CAD at 1.8 (95% CI 1.4–2.51), 2.3 (95% CI 1.91–2.93), and medically treatable CAD risk factors, including diabetes,
2.6 (95% CI 2–3.37), respectively (p < 0.001 for all). These dyslipidemia, and hypertension (6.7% over 3 years, log-rank
findings revealed that CCTA-diagnosed CAD was able to p < 0.001), with annualized mortality rates that would reclas-
predict ACM in symptomatic patients irrespective of age, sify these supposedly low-risk patients [12]. Thus, the iden-
sex, and conventional risk factors [11]. Lin et al. also evalu- tification of nonobstructive and obstructive CAD by CCTA
ated the independent predictive value of nonobstructive both portend elevated mortality above and beyond CAD risk
CAD diagnosed by 64-detector row CCTA in a multicenter factors and traditional risk scores.
cohort of 2583 patients without prior CAD over
3.1 ± 0.5 years [12]. An adjusted HR of 1.98 (95% CI 1.06–
3.69, p = 0.03) was observed for the presence of nonobstruc-  rognostic Significance of CCTA for Major
P
tive CAD overall, with the highest risk seen among patients Adverse Cardiovascular Events
with 3-vessel nonobstructive CAD (HR 4.75, 95% CI 2.10–
10.75, p < 0.001) or those with nonobstructive CAD affect- CCTA risk stratification for MACE is also well demon-
ing ≥5 segments (HR 5.12, 95% CI 2.16–12.10, p < 0.001). strated by study investigators (Table  51.1) [13–22]. The
Moreover, the presence of nonobstructive plaque was associ- earliest data on MACE was reported by Pundziute and col-
ated with higher risk of mortality even among patients with leagues [13]. Over a mean follow-up of 16  months, 100
 51  Prognosis and Outcome: State of the Evidence 661

1.00 None or Mild (<50%) Plaque (n=422)

≥2 Mild (30%-49%) Plaque with Proximal
Plaque in 1 Artery (n=64), p=0.192
1 Moderate (50%-69%) Plaque (n=212), p=0.065
2 Moderate (50%-69%) Plaque or 1 Severe (≥70%) Plaque
0.95 (n=101), p=0.013
3 Moderate (50%-69%) Plaque or 2 Severe (≥70%) Plaque or
Cumulative Survival

Severe (≥70%) Proximal LAD Plaque (n=145), p=0.002

3 Severe (≥70%) Stenoses or 2 Severe (≥70%) Stenoses with

0.90 Proximal LAD (n=86), p=0.001

0.85
Moderate or Severe (≥50%) Left Main Plaque (n=106), p<0.0001

0.80 X2=51, p<0.0001

0 0.5 1 1.5 2
Time to Follow-up (Years)

Fig. 51.1  Cumulative survival in patients exhibiting coronary plaque by the Duke prognostic coronary artery disease index. Risk-adjusted
p < 0.001 (adjusted for age, family history, and dyslipidemia). LAD left anterior descending artery. (From Min et al. [7], with permission)

Event Rate CTA diagnosed CAD

100
1.7% (19/1085) Normal
1 Vessel Non-obstructive Disease
2.7% (20/750) 2 Vessel Non-obstructive Disease
3 Vessel Non-obstructive Disease
4.6% (11/236)
1 Vessel Obstructive Disease
95 2 Vessel Obstructive Disease
6.9% (5/72) 3 Vessel Obstructive Disease

7.1% (22/310)
Event Free Survival

90
11.3% (7/68)

85

20% (3/15)
80

75

0 20 40 60 80 100 120
Follow Up (month)

Fig. 51.2  Risk-adjusted event-free survival by CCTA-diagnosed CAD artery calcium score. All-cause mortality increased significantly by
stratified by severity of disease and number of diseased coronary arteries. increasing severity of CCTA-diagnosed CAD and number of diseased
Risk adjustment for age, gender, hypercholesterolemia, diabetes mellitus, coronary arteries. CCTA coronary computed tomographic angiography,
smoking, hypertension, family history of premature CAD, and coronary CAD coronary artery disease. (From Ostrom et al. [11], with permission)
662 A. Rizvi et al.

symptomatic patients with known or suspected CAD were with an average of 1.96 years of follow-up for a composite
observed for MACE defined as cardiac death, unstable outcome of ACM, cardiac death, MI, unstable angina, or
angina, nonfatal MI, and revascularization. Patients with- revascularization [27]. Patients with obstructive CAD and
out CAD had an excellent prognosis compared to patients nonobstructive CAD had increased odds of cardiac death or
with any evidence of CAD, with reported first-year event MI with demonstrated odds ratios (OR) of 14.92 (95% CI
rates of 0% and 30%, respectively (log-rank p  <  0.005). 6.78–32.85) and 6.41 (95% CI 2.44–16.8), respectively,
The greatest risk of MACE was seen in patients with any when compared with those with no CAD [27].
coronary plaque (aHR 8.8; 95% CI 1.1–70), obstructive In a recent small study (n = 218) with a longer follow-up
CAD (aHR 28; 95% CI 3.3–239), and left main or left coro- duration of 6.9  years, Dougoud et  al. found annual event
nary artery disease (aHR 35; 95% CI 4.3–288) (p < 0.05 for rates for composite MACE (mortality, nonfatal MI, and late
all). In aggregate, a CCTA showing no evidence of CAD revascularization) in patients with obstructive CAD (≥50%),
could assure a high negative predictive value for MACE, nonobstructive CAD (<50%), and normal coronary arteries
with a dose effect evident for increasing extent and severity on CCTA to be 6.0%, 2.7%, and 0.3%, respectively (log-rank
of CCTA-visualized CAD [13]. p = 0.001) [28]. Thus, the prognostic value of the presence or
Most notably, the CONFIRM (The Coronary CT absence of CAD detected by CCTA is durable and incremen-
Angiography EvaluatioN For clinical Outcomes: An tal to clinical risk factors alone.
InterRnational Multicenter) registry has provided important
data regarding the incremental prognostic significance of
CCTA [23]. Using CONFIRM data, Leipsic and colleagues The Confirm Registry
evaluated the risk of MACE in 5262 patients with suspected
CAD but without medically modifiable CAD risk factors The CONFIRM registry is a large, prospective, dynamic
[24]. Over a mean follow-up of 2.3  ±  1.2  years, increased observational study and represents 27,125 consecutive
risk of MACE was observed in patients with obstructive patients who have undergone ≥64-detector row CCTA [23].
CAD (risk-adjusted HR 6.64; 95% CI 3.68–12.00; The study design and methods have been described else-
p ≤ 0.001). Furthermore, there was a dose-response relation- where [23]. The primary objective of this study is to assess
ship between the number of obstructed vessels and the risk of the prognostic significance of CCTA for the prediction of
MACE, with the greatest hazard seen in patients with future adverse CAD events including ACM, MI, unstable
3-­vessel obstructive CAD or left main CAD (HR 11.69; 95% angina (UA), target vessel revascularization, and CAD-­
CI 5.38, 25.4; p ≤ 0.001). Notably, an extremely low inci- related hospitalization [23]. The scale of the CONFIRM reg-
dence of MACE was reported in patients without any evi- istry has allowed detailed examination of the prognostic
dence of CAD on CCTA as compared with those with value of CCTA in multiple patient subgroups.
obstructive CAD (0.31% versus 2.06%) [24]. Further still,
the CONFIRM investigators demonstrated the prognostic
value of CCTA findings with long-term follow-up. Cheruvu  rognostic Significance of CCTA by Age
P
et al. followed 1884 patients without CAD for the outcomes and Gender
of ACM and MACE [25]. Over a mean follow-up of
5.6 ± 1.3 years, the risk of ACM increased with the severity The ability of CCTA to risk stratify younger and older
of CAD and the number of diseased vessels, with the highest men and women is fundamental to guide the appropriate
risk seen in those with 3-vessel or left main obstructive application of CCTA across diverse populations. To that
(>50% stenosis) CAD (HR: 2.87; 95% CI 1.57–5.23; end, Min and colleagues examined 24,775 patients with-
p = 0.001). During the same follow-up duration, an increased out known CAD for age- and sex-specific outcomes utiliz-
risk of MACE was observed in patients with obstructive ing the CONFIRM registry [29]. Over 2.3  years of
CAD (HR 6.63; 95% CI 3.91–11.26; p < 0.001) as compared follow-up, the risk of ACM was significantly higher in
to those with nonobstructive CAD (HR: 2.20; 95% CI 1.31– younger patients with CAD (<65 years) compared to older
3.67; p = 0.003). The incidence of MACE in patients with no patients with CAD (≥65  years), with an adjusted HR of
CAD was reported as 5.6% as compared to 13.24% and 4.00 versus 2.46 for 2-vessel CAD and 6.19 versus 3.10
36.28% in patients with nonobstructive and obstructive for 3-vessel CAD (p  <  0.001 for both) (Fig.  51.3).
CAD, respectively (p < 0.001) [25]. Moreover, gender-specific risk stratification revealed that
In addition, the prognostic value of CCTA for MACE has women had increased risk of ACM compared to men for
been confirmed in meta-analyses [26, 27]. One meta-­analysis 3-vessel CAD, with an adjusted HR of 4.21 versus 3.27
of 28 studies by Habib et al. comprised of 41,690 individuals (p < 0.001 for all) (Figure 51.4) [29].
51  Prognosis and Outcome: State of the Evidence 663

Fig. 51.3 Unadjusted a 1.00 Normal

all-cause 3-year Kaplan-­ Non-Obstructive
Meier survival by presence, p<01
1-Vessel, p<.01
extent, and severity of CAD
by CCTA as stratified by 2-Vessel, p<0.001
age <65 or ≥65 years.
Although rates of mortality in 0.95 3-Vessel/Left Main

Survival Probability
relationship to CAD extent p<0.001
are lower in patients
age <65 years (a), patients
age <65 years with 2- and
3-vessel CAD experience a
higher relative rate of 0.90
mortality referenced to
patients age <65 years with
no CAD in comparison with
patients age ≥65 years with
2- and 3-vessel CAD 0.85
referenced to patients
0.0 0.5 1.0 1.5 2.0 2.5 3.0
age ≥65 years with no CAD
Survival Time (Years)
(b). CAD coronary artery
At Risk Year 0 Year 1 Year 2 Year 3
disease, CCTA coronary
computed tomographic Normal 8646 8010 4693 2474

angiography. (From Min et al. Non-Obstructive 5103 4717 2606 1287

[29], with permission) 1-Vessel 1807 1693 1049 481

2-Vessel 692 640 419 180
3-Vessel/Left Main 501 466 306 151

b 1.00

Normal
Non-Obstructive, p<01
1-Vessel, p<.01
0.95
Survival Probability

2-Vessel, p<0.001

3-Vessel/Left Main
0.90 p<0.001

0.85
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Survival Time (Years)
At Risk Year 0 Year 1 Year 2 Year 3
Normal 1500 1347 837 433
Non-Obstructive 3011 2720 1475 643
1-Vessel 1311 1180 698 301
2-Vessel 654 588 323 144
3-Vessel/Left Main 629 568 358 173

 rognostic Significance of CCTA Based

P colleagues evaluated the prognostic significance of CCTA in
on Ethnic Differences a total of 16,451 patients among three subgroups of
Caucasian, East Asian, and African ethnicities [35]. When
Prior studies have reported that the epidemiology, pathology, comparing patients with obstructive CAD and those with
and prognosis of CAD may vary across different ethnicities nonobstructive CAD, the annual incidence of death or MI
[30–34]. Using data from CONFIRM registry, Hulten and was 2.2% versus 0.7% among Caucasians (aHR 2.77, 95%
664 A. Rizvi et al.

Fig. 51.4 Unadjusted a 1.00

all-cause 3-year Kaplan-­ Normal
Meier Survival by presence,
Non-Obstructive
extent, and severity of CAD p<0.001
by CCTA as stratified by Sex. 1-Vessel, p<0.001
Although rates of mortality in
relationship to CAD extent 0.95 2-Vessel, p<0.001

Survival Probability
are lower in women, women
with 3-vessel CAD experience
a higher relative rate of
mortality referenced to
women with no CAD (a) in 3-Vessel/Left Main
comparison with men with p<0.001
0.90
3-vessel CAD referenced to
men with no CAD (b). CAD
coronary artery disease,
CCTA coronary computed
tomographic angiography.
(From Min et al. [29], with 0.85
permission) 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Survival Time (Years)
At Risk Year 0 Year 1 Year 2 Year 3
Normal 5594 5179 3143 1545
Non-Obstructive 3232 2955 1564 650
1-Vessel 1214 1119 680 298
2-Vessel 459 433 253 113
3-Vessel/Left Main 347 317 218 107

b 1.00
Normal
Non-Obstructive
p<001
1-Vessel, p<.001

0.95 2-Vessel, p<0.001

Survival Probability

3-Vessel/Left Main
p<0.001

0.90

0.85
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Survival Time (Years)
At Risk Year 0 Year 1 Year 2 Year 3
Normal 4528 4160 2648 1357
Non-Obstructive 4849 4457 2513 1280
1-Vessel 1887 1743 1064 484
2-Vessel 881 790 489 211
3-Vessel/Left Main 777 712 444 217

CI 1.73–4.43, p < 0.001), 4.8% versus 1.1% among Africans  rognostic Significance of CCTA Above
P
(aHR 6.25, 95% CI 1.12–34.97, p = 0.037), and 0.8% versus and Beyond Coronary Artery Calcium Scoring
0.1% among East Asians (aHR 4.84, 95% CI 2.24–10.9,
p  <  0.001). Interestingly, East Asians had a lower risk of Coronary artery calcium (CAC) scoring derived from non-­
events (aHR 0.25, 95% CI 0.16–0.38, p < 0.001) when com- contrast CT has been shown to be a useful test for identifying
pared to Caucasians and Africans [35]. atherosclerotic plaque burden. In patients without chest pain
51  Prognosis and Outcome: State of the Evidence 665

symptoms, the absence of CAC has been associated with a In a recent investigation utilizing CONFIRM data, Blanke
lower risk of future cardiovascular adverse events. However, et  al. evaluated the long-term prognostic value of CCTA
the relationship between CAC scoring and CCTA for prog- among DM patients compared with non-DM individuals
nostication was uncertain prior to investigations within the [41]. A total of 1823 DM patients were identified and
CONFIRM cohort. First, Villines and the CONFIRM inves- propensity-­matched to 1823 non-DM patients with 5-year
tigators evaluated the prognostic value of CCTA in 10,037 follow-up. The presence of CAD on CCTA was graded as
symptomatic patients without coronary artery calcium with- none (0%), nonobstructive (1–49%), or obstructive (≥50%)
out known CAD [36]. Over a median follow-up of 2.1 years, stenosis. The study reported that DM patients did not display
no difference was noted in the rate of ACM in patients with increased risk of mortality compared with non-DM patients
CAC scores of 0, despite the presence of obstructive CAD in the absence of CAD (aHR of DM 1.32, 95% CI 0.78–2.24,
(p  =  0.7). MACE occurred in 3.9% of patients with CAC p = 0.30). However, DM patients did display heightened risk
scores of 0 and obstructive CAD (HR 5.7, 95% CI 2.5–13.1, of mortality compared with non-DM patients when they had
p < 0.001) compared to 0.8% of patients with CAC scores of nonobstructive CAD (HR 2.10, 95% CI 1.43–3.09,
0 and nonobstructive CAD. Further, CAC scoring in symp- p < 0.001). Notably, the mortality risk in this population was
tomatic patients did not provide incremental prognostic ben- even higher than non-DM patients with obstructive disease
efit for predicting MACE compared to clinical risk factors (p < 0.001). The authors concluded that among DM patients,
and CCTA alone (p = 0.84) [36]. Al-Mallah et al. examined CCTA-assessed nonobstructive and obstructive CAD was
the incremental prognostic value of CAC scoring and CCTA associated with significantly increased risk of ACM and
over clinical risk stratification in 8627 symptomatic patients MACE at 5-year follow-up, when compared with non-DM
without known CAD who were observed for MACE for a patients [41].
median follow-up duration of 25 months [37]. After adjust-
ing for clinical risks, a CAC score  >400 and detection of
obstructive CAD by CCTA were found to be significant pre-  rognostic Significance of CCTA Based on CAD
P
dictors of MACE (HR 4.8, 95% CI 2.9–8.1 versus HR 3.9; Severity and Left Ventricular Ejection Fraction
95% CI 2.7–5.5, respectively, p < 0.05 for both). Furthermore,
obstructive CAD detection by CCTA showed incremental CCTA acquisition protocols that span systole can evaluate
prognostic value over clinical risk stratification and CAC LVEF in addition to CAD severity. Utilizing data from the
scoring (C-statistics 0.82 vs. 0.79, p = 0.002) [37]. CONFIRM registry, Arsanjani and colleagues assessed the
Cho et al. examined CCTA and CAC scoring in asymp- gradations of LVEF and volumes measured with CCTA
tomatic patients [38]. During a median follow-up of among 7758 patients [42]. During a follow-up of 2.2 years,
2.5 years, patients with obstructive 2- and 3-vessel CAD or worse LVEF was independently associated with mortality
left main CAD had increased risk of ACM and MACE for moderately reduced (<45% but ≥35%) and severely
(p < 0.05 for both) compared to those without CAD. Both reduced (<35%) LVEF, with a demonstrated HR of 3.14 and
CAC scoring and CAD detection by CCTA improved perfor- 5.19, respectively (p  <  0.001). Moreover, LVEF demon-
mance over standard risk factors for predicting ACM and strated improved discrimination for mortality [area under the
MACE (likelihood ratio p  <  0.05 for all). However, incre- receiver operating characteristic curve (AUC) = 0.816] when
mental discriminatory ability was only shown for MACE. Net compared with CAD risk factors alone (AUC  =  0.781) or
reclassification improvement (NRI) was not observed when CAD risk factors combined with CAD extent and severity
CCTA was added to the risk factors plus CAC scoring model. (AUC  =  0.799) (p  <  0.001). The study concluded that left
Thus, among asymptomatic patients, CCTA is not useful for ventricular dysfunction improves risk prediction and dis-
risk stratification beyond combined assessment by risk fac- crimination for ACM [42].
tors with CACS [38].

 rognostic Significance of CCTA in Patients

P
 rognostic Significance of CCTA in Diabetic
P with Acute Chest Pain
Patients
Acute-onset chest pain is one of the most common presenta-
People with diabetes mellitus (DM) demonstrate increased tions for patient visits to the emergency department (ED).
risk of CAD and display adverse outcomes after surviving a Given the high sensitivity and NPV for the detection of coro-
CAD event [39]. The study investigators therefore assessed nary stenosis, CCTA may allow for timely diagnosis,
the prevalence, extent, severity, and prognosis of CAD in improved risk stratification, and appropriate triage of
DM and non-DM patients without known CAD who under- patients, especially those with low to intermediate risk.
went CCTA [40]. Accordingly, numerous studies have evaluated the long-term
666 A. Rizvi et al.

outcome of patients with negative CCTAs who were dis- recently published guidelines for the appropriate use of
charged from the ED [43]. CCTA in the diagnosis of acute chest pain in patients with
The CT-STAT (Coronary Computed Tomographic suspicion of ACS in the ED [48]. Utilizing CCTA in these
Angiography for Systematic Triage of Acute Chest Pain patients may lead to increased downstream invasive coro-
Patients to Treatment) was a multicenter trial that included 699 nary angiography (ICA) and coronary revascularization.
low-risk ED patients [44]. These prospectively enrolled patients The ACCF/SCCT/ACR/AHA/ASE/ASNC/NASCI/
were either randomly allocated to CCTA (n = 361) or to myo- SCAI/SCMR 2010 Appropriate Use Criteria for Cardiac
cardial perfusion scintigraphy (MPS) (n  =  338). The study Computed Tomography lists the use of CCTA as appropriate
investigators compared the efficiency, cost, and safety of CCTA for “detection of CAD in symptomatic patients without
in the evaluation of patients with acute chest pain and low risk known heart disease—acute symptoms with suspicion of
of ACS.  The primary outcome was time to diagnosis. The ACS (urgent presentation) (Appropriate, score 7)” in patients
investigators also showed a cost reduction in patients random- with the following [49]:
ized to CCTA. The study found that patients in the CCTA arm
had a 54% reduction in time to diagnosis and 38% reduction in • Normal ECG and cardiac biomarkers and low-­intermediate
costs. Notably, there was no difference in MACE between the pretest probability of CAD
two study groups [44]. Further, the ACRIN-PA (American • ECG uninterpretable and low-intermediate pretest proba-
College of Radiology Imaging Network-Pennsylvania) study bility of CAD
evaluated the safety (defined as the absence of MI or cardiac • Nondiagnostic ECG or equivocal cardiac biomarkers and
death during 30-day follow-­up) of this CCTA strategy in low- low-intermediate pretest probability of CAD
to intermediate-risk patients who presented to the ED [45]. Of
the 640 patients with negative CCTA, none of them died or had In theory, a triple rule-out (TRO)-CT protocol is attractive
a MI within 30 days of presentation. In addition, early CCTA for ED evaluation of chest pain because it may allow for
led to a shorter mean hospital stay (18 h versus 24.8 h) and an exclusion of ACS, aortic dissection (AD), and pulmonary
increased rate of ED discharge when compared to standard of embolus (PE) in a single scan [50]. However, TRO-CT has
care (50% versus 23%) [45]. potential disadvantages of increased radiation and contrast
ROMICAT (Rule Out Myocardial Infarction/Ischemia dose, higher imaging costs, and decreased image quality.
Using Computer Assisted Tomography) was a double-blind, Burris and colleagues compared the diagnostic yield of
prospective, observational study that demonstrated excellent TRO-CT with CCTA for assessment of obstructive (>50%
sensitivity and NPV of CCTA to rule out ACS in 368 patients stenosis) CAD, AD, and PE in 12,834 patients with acute
with inconclusive clinical evaluation of acute chest pain in the chest pain in the multicenter ACIC (Advanced Cardiovascular
ED [46]. The subsequent ROMICAT II study followed 368 ED Imaging Consortium) registry [51, 52]. The overall diagnos-
patients who presented with acute chest pain, negative initial tic yield was similar for TRO-CT and CCTA (18.3% versus
troponin, and ECG findings for 2 years [47]. The investigators 17.4%, p = 0.37). TRO-CT demonstrated no additional yield
found that the cumulative probability of 2-year MACE signifi- for CAD detection as compared to CCTA (15.5% versus
cantly increased with CCTA-stratified CAD categories, with 17.2%, p = 0.093). TRO-CT exhibited slightly higher diag-
0%, 5%, and 30% for no CAD, nonobstructive CAD, and nostic yield compared to CCTA in the detection of AD (1.7
obstructive CAD, respectively (log-rank p  <  0.001). Adding vs. 1.1%, p = 0.046) and PE (1.1 vs. 0.4%, p = 0.004), but the
evaluation of regional wall motion abnormalities (RWMA) clinical magnitude was very small. Therefore, they con-
improved risk stratification, with 2-year MACE events of 0.9%, cluded that there is no clear clinical benefit from routine use
15%, 10%, and 62% for no stenosis or RWMA, RWMA but no of TRO-CT in patients with acute chest pain [52].
stenosis, stenosis but no RWMA, and both stenosis and
RWMA, respectively (log-rank p < 0.001). The addition of CT
stenosis beyond the clinical Thrombolysis In Myocardial  rognostic Significance of CCTA in Patients
P
Infarction (TIMI) risk score improved discrimination for pre- with Known CAD
dicting MACE (0.84 vs. 0.61, p  <  0.001), and addition of
RWMA further improved the prediction (0.91, p < 0.001) com- Prior data have reported risk stratification of coronary artery
pared to combined TIMI risk score and CT stenosis alone [47]. bypass (CABG) patients utilizing ICA [53]. However, CCTA
In aggregate, these studies support the use of CCTA as an can risk stratify CABG patients noninvasively on the basis of
efficient, safe, and cost-effective alternative to the traditional native vessel anatomy and graft patency [54–57]. To date,
triage methods for low- and low-to-intermediate-risk ED limited data is available to assess the prognostic utility of
patients to exclude obstructive CAD as the etiology of chest CCTA in CABG patients.
pain while allowing for a faster ED discharge. Thus, the In an investigation by Small et al., the authors evaluated
Society of Cardiovascular Computed Tomography (SCCT) ACM in 657 CABG patients from a multicenter registry of
51  Prognosis and Outcome: State of the Evidence 667

10,628 patients who underwent CCTA, over 20  months of 2 Feature-Negative Plaques/No Plaques
follow-up [58]. CAD severity was categorized by two mod- 1.0
1 Feature-Positive Plaques
els: unprotected coronary territories (UCT) representing the
number of vessels with CAD and coronary artery protection .8

Cumulative event-free rate:

score (CAPS) defined as a summary of native vessel disease
and graft patency. In multivariate analysis, the UCT (HR 2 Feature-Positive Plaques
.6
1.35, 95% CI 0.98–1.84, p  =  0.004) and CAPS (HR 1.35,
95% CI 1.17–1.56, p < 0.001) were both independent predic-
tors of ACM. Furthermore, CAPS improved discrimination .4
and reclassification compared with clinical risk factors alone
(C-statistics, 0.75 vs. 0.64, p < 0.001; NRI, 27.2%, p = 0.003), .2
although UCT did not [58].

0 Logrank <0.001
 rognostic Significance of CCTA Plaque
P 0 10 20 30 40 50
Characteristics Month of follow-up

Previous studies have shown that approximately 60% of Fig. 51.5  Kaplan-Meier Curve for development of ACS on the basis of
plaque characteristics. Patient stratification according to the presence of
high-risk plaques are not associated with anatomically sig- 2- and 1-feature positive and 2-feature negative plaques/no plaques.
nificant stenosis [59, 60]. CCTA assessment of plaque char- The y-axis represents cumulative event-free rate (log-rank p < 0.001).
acteristics may permit early identification of vulnerable ACS acute coronary syndrome. (From Motoyama et  al. [67], with
plaque that would be missed by luminal evaluation alone permission)
[61–65]. Study investigators examined atherosclerotic
plaque characteristics by CCTA in patients with ACS and showed that presence of HRP on CCTA was a significant
stable angina, analyzing positive remodeling (PR), spotty predictor of ACS (OR 8.9, 95% CI 1.8–43.3, p = 0.006) after
calcification (SC), and consistency of non-calcified plaque adjusting for a  ≥50% coronary artery stenosis (OR 38.6,
(NCP <30 HU or 30 HU < NCP <150 HU) [66]. The study 95% CI 14.2–104.7, p < 0.001) [69].
found that among CCTA assessment of culprit lesions in These findings shed light on the prognostic power of
patients with ACS and stable angina, PR (87% vs. 12%), SC CCTA beyond the detection of stenosis severity and high-
(63% vs. 21%), and NCP <30 HU (79% vs. 9%) were signifi- light the importance of assessing non-stenotic coronary
cantly more prevalent in ACS patients (p  <  0.001 for all) artery morphology and atherosclerotic plaque characteris-
[66]. In a subsequent investigation, Motoyama et al. evalu- tics, which are difficult to evaluate through stress testing and
ated two atherosclerotic plaque characteristics, PR and low ICA luminology alone.
attenuation plaque (LAP) detected by CCTA, for predicting
acute coronary events in >1000 patients with established or
suspected CAD who were followed for 27 ± 10 months [67].  rognostic Significance of CCTA
P
Both features were found to be independent predictors of as Compared to Myocardial Perfusion
acute coronary events (HR 23, 95% CI 7–75, p < 0.001]. The Scintigraphy
authors reported that the likelihood of ACS in patients with
1- or 2-feature positive plaques was significantly higher As the prognostic value of myocardial perfusion scintigra-
compared to patients without these plaque features or phy (MPS) has been supported by a large body of literature,
patients with no plaques (22.2% vs. 3.7% vs. 0.5%, log-rank CCTA has often been compared to MPS for predictive value.
p  <  0.001) (Fig.  51.5). Importantly, the plaques associated In this regard, Shaw et al. propensity-matched 693 patients
with early ACS had larger LAP volume when compared with who underwent CCTA with 3067 patients who underwent
those associated with late ACS [67]. In addition, Imazeki and MPS for evaluation of new-onset chest pain. The study found
colleagues demonstrated that remodeling detected on CCTA similar risk stratification by Duke prognostic CAD index by
closely correlated with IVUS, and the remodeling index (RI) CCTA and by percentage of ischemic myocardium by MPS
was significantly larger in patients with ACS (1.19 ± 0.18) [70]. Van Werkhoven and colleagues evaluated the incremen-
than in those with stable angina (0.89  ±  0.10, P  <  0.001) tal prognostic significance of CCTA and MPS in 541 patients
[68]. Further, in a ROMICAT II subanalysis, the presence of over a median of 672 days [71]. CCTA and MPS were found
high-risk plaque (HRP) characteristics (PR, LAP, SC, and to be complementary, with an incremental predictive value
napkin-ring sign) was evaluated for prognosis of ACS in of obstructive CAD by CCTA above and beyond MPS and
patients with acute chest pain [69]. The study findings baseline risk factors. The annualized event rate was 1.0% in
668 A. Rizvi et al.

patients with a normal CCTA and MPS, 3.7–3.8% in those computed tomographic angiography for evaluation of coronary
with an abnormal CCTA or MPS, and 9.0% in those with artery stenosis in individuals without known coronary artery
disease: results from the prospective multicenter ACCURACY
both an abnormal CCTA and an abnormal MPS (log-rank (Assessment by Coronary Computed Tomographic Angiography
p < 0.005) [71]. of Individuals Undergoing Invasive Coronary Angiography) trial. J
Am Coll Cardiol. 2008;52:1724–32.
2. Mowatt G, Cook JA, Hillis GS, Walker S, Fraser C, Jia X, Waugh
N. 64-slice computed tomography angiography in the diagnosis and
Downstream Outcomes After CCTA assessment of coronary artery disease: systematic review and meta-­
analysis. Heart. 2008;94:1386–93.
PROMISE and SCOT-HEART (Scottish COmputed 3. Douglas PS, Hoffmann U, Patel MR, Mark DB, Al-Khalidi HR,
Tomography of the HEART) are two large trials that provide Cavanaugh B, Cole J, Dolor RJ, Fordyce CB, Huang M, Khan
MA, Kosinski AS, Krucoff MW, Malhotra V, Picard MH, Udelson
insight into the diagnosis, management, and outcomes in JE, Velazquez EJ, Yow E, Cooper LS, Lee KL, Investigators
patients with stable angina [3, 72]. The PROMISE study com- P. Outcomes of anatomical versus functional testing for coronary
pared the outcomes of CCTA and functional stress testing in artery disease. N Engl J Med. 2015;372:1291–300.
10,003 symptomatic patients without diagnosed CAD [3]. At 4. Taylor AJ, Cerqueira M, Hodgson JM, Mark D, Min J, O'Gara
P, Rubin GD, American College of Cardiology Foundation
a median follow-up of 25 months, there was no significant dif- Appropriate Use Criteria Task F, Society of Cardiovascular
ference in the incidence of MACE for CCTA (3.3%) and func- Computed T, American College of R, American Heart A, American
tional-testing (3.0%) groups (aHR 1.04, 95% CI 0.83–1.29, Society of E, American Society of Nuclear C, North American
p = 0.75). In addition, the CCTA group showed a reduced risk Society for Cardiovascular I, Society for Cardiovascular A,
Interventions, Society for Cardiovascular Magnetic R, Kramer
of death or nonfatal MI (HR 0.66, 95% CI 0.44–1.00, CM, Berman D, Brown A, Chaudhry FA, Cury RC, Desai MY,
p = 0.049) compared to the functional-­testing group. Despite Einstein AJ, Gomes AS, Harrington R, Hoffmann U, Khare R,
an increased rate of ICA in the CCTA group compared with Lesser J, McGann C, Rosenberg A, Schwartz R, Shelton M,
the functional-testing group (12.2% vs. 8.1%), there was a Smetana GW, Smith SC Jr. ACCF/SCCT/ACR/AHA/ASE/ASNC/
NASCI/SCAI/SCMR 2010 appropriate use criteria for cardiac
lower rate of ICA that showed no obstructive CAD within computed tomography. A report of the American college of car-
90 days in the CCTA group (3.4% vs. 4.3%, p = 0.02). The diology foundation appropriate use criteria task force, the society
study findings therefore revealed that CCTA was not associ- of cardiovascular computed tomography, the American college of
ated with better clinical outcomes than functional testing in radiology, the American heart association, the American society
of echocardiography, the American society of nuclear cardiol-
symptomatic patients with suspected CAD [3]. The SCOT- ogy, the north American society for cardiovascular imaging, the
HEART trial assessed 4146 patients with stable angina by society for cardiovascular angiography and interventions, and the
CCTA plus standard care or standard care alone [72]. In the society for cardiovascular magnetic resonance. J Am Coll Cardiol.
CCTA group, an increased certainty of diagnosis of angina 2010;56:1864–94.
5. Abbara S, Arbab-Zadeh A, Callister TQ, Desai MY, Mamuya W,
due to CAD was observed at 6 weeks (relative risk [RR] 1.79, Thomson L, Weigold WG.  SCCT guidelines for performance of
95% CI 1.62–1.96, p < 0.001). Furthermore, the CCTA group coronary computed tomographic angiography: a report of the
had a strong trend toward 38% reduction in cardiac death or Society of Cardiovascular Computed Tomography Guidelines
nonfatal MI related to CAD at a median follow-up of 1.7 years Committee. J Cardiovasc Comput Tomogr. 2009;3:190–204.
6. Raff GL, Abidov A, Achenbach S, Berman DS, Boxt LM, Budoff
(aHR 0.62, 95% CI 0.38–1.01, p = 0.053) [72]. MJ, Cheng V, DeFrance T, Hellinger JC, Karlsberg RP, Society of
Cardiovascular Computed T. SCCT guidelines for the interpretation
and reporting of coronary computed tomographic angiography. J
Conclusion Cardiovasc Comput Tomogr. 2009;3:122–36.
7. Min JK, Shaw LJ, Devereux RB, Okin PM, Weinsaft JW, Russo DJ,
Lippolis NJ, Berman DS, Callister TQ. Prognostic value of multi-
CCTA provides vital prognostic information beyond its high detector coronary computed tomographic angiography for predic-
diagnostic performance for CAD evaluation. A CCTA that tion of all-cause mortality. J Am Coll Cardiol. 2007;50:1161–70.
detects no coronary plaque accurately predicts a very low 8. Min JK, Lin FY, Dunning AM, Delago A, Egan J, Shaw LJ,
Berman DS, Callister TQ.  Incremental prognostic significance of
rate of major adverse cardiovascular events. Major studies left ventricular dysfunction to coronary artery disease detection
including the CONFIRM registry, the ROMICAT trial, the by 64-detector row coronary computed tomographic angiography
PROMISE study, and the SCOT-HEART study have pro- for the prediction of all-cause mortality: results from a two-centre
vided a robust evidence base for risk prediction with CCTA study of 5330 patients. Eur Heart J. 2010;31:1212–9.
9. Pundziute G, Schuijf JD, Jukema JW, Decramer I, Sarno G,
in diverse populations for broad application to patient care. Vanhoenacker PK, Boersma E, Reiber JH, Schalij MJ, Wijns W,
Bax JJ.  Evaluation of plaque characteristics in acute coronary
syndromes: non-invasive assessment with multi-slice computed
References tomography and invasive evaluation with intravascular ultrasound
radiofrequency data analysis. Eur Heart J. 2008;29:2373–81.
1. Budoff MJ, Dowe D, Jollis JG, Gitter M, Sutherland J, Halamert 10. Voros S, Rinehart S, Qian Z, Joshi P, Vazquez G, Fischer C, Belur
E, Scherer M, Bellinger R, Martin A, Benton R, Delago A, Min P, Hulten E, Villines TC.  Coronary atherosclerosis imaging by
JK.  Diagnostic performance of 64-multidetector row coronary coronary CT angiography: current status, correlation with intravas-
51  Prognosis and Outcome: State of the Evidence 669

cular interrogation and meta-analysis. JACC Cardiovasc Imaging. no modifiable risk factors: results from an international multicenter
2011;4:537–48. study of 5262 patients. Radiology. 2013;267:718–26.
11. Ostrom MP, Gopal A, Ahmadi N, Nasir K, Yang E, Kakadiaris I, 25. Cheruvu C, Precious B, Naoum C, et al. Long term prognostic util-
Flores F, Mao SS, Budoff MJ. Mortality incidence and the sever- ity of coronary CT angiography in patients with no modifiable cor-
ity of coronary atherosclerosis assessed by computed tomography onary artery disease risk factors: results from the 5 year follow-up
angiography. J Am Coll Cardiol. 2008;52:1335–43. of the CONFIRM international multicenter registry. J Cardiovasc
12. Lin FY, Shaw LJ, Dunning AM, Labounty TM, Choi JH, Weinsaft Comput Tomogr. 2016;10:22–7.
JW, Koduru S, Gomez MJ, Delago AJ, Callister TQ, Berman DS, 26. Abdulla J, Asferg C, Kofoed KF.  Prognostic value of absence or
Min JK.  Mortality risk in symptomatic patients with nonobstruc- presence of coronary artery disease determined by 64-slice com-
tive coronary artery disease: a prospective 2-center study of 2,583 puted tomography coronary angiography a systematic review and
patients undergoing 64-detector row coronary computed tomo- meta-analysis. Int J Cardiovasc Imaging. 2011;27:413–20.
graphic angiography. J Am Coll Cardiol. 2011;58:510–9. 27. Habib PJ, Green J, Butterfield RC, Kuntz GM, Murthy R, Kraemer
13. Pundziute G, Schuijf JD, Jukema JW, Boersma E, de Roos A, van DF, Percy RF, Miller AB, Strom JA. Association of cardiac events
der Wall EE, Bax JJ.  Prognostic value of multislice computed with coronary artery disease detected by 64-slice or greater coro-
tomography coronary angiography in patients with known or sus- nary CT angiography: a systematic review and meta-analysis. Int J
pected coronary artery disease. J Am Coll Cardiol. 2007;49:62–70. Cardiol. 2013;169:112–20.
14. Gaemperli O, Valenta I, Schepis T, Husmann L, Scheffel H,
28. Dougoud S, Fuchs TA, Stehli J, Clerc OF, Buechel RR, Herzog BA,
Desbiolles L, Leschka S, Alkadhi H, Kaufmann PA.  Coronary Leschka S, Alkadhi H, Kaufmann PA, Gaemperli O.  Prognostic
64-slice CT angiography predicts outcome in patients with known value of coronary CT angiography on long-term follow-up of 6.9
or suspected coronary artery disease. Eur Radiol. 2008;18:1162–73. years. Int J Cardiovasc Imaging. 2014;30:969–76.
15. Aldrovandi A, Maffei E, Palumbo A, Seitun S, Martini C, Brambilla 29. Min JK, Dunning A, Lin FY, et  al. Age- and sex-related differ-
V, Zuccarelli A, Tarantini G, Weustink AC, Mollet NR, Ruffini ences in all-cause mortality risk based on coronary computed
L, Crisi G, Ardissino D, de Feyter PJ, Krestin GP, Cademartiri tomography angiography findings results from the international
F.  Prognostic value of computed tomography coronary angiogra- multicenter CONFIRM (coronary CT angiography evaluation for
phy in patients with suspected coronary artery disease: a 24-month clinical outcomes: an international multicenter registry) of 23,854
follow-up study. Eur Radiol. 2009;19:1653–60. patients without known coronary artery disease. J Am Coll Cardiol.
16. Carrigan TP, Nair D, Schoenhagen P, Curtin RJ, Popovic
2011;58:849–60.
ZB, Halliburton S, Kuzmiak S, White RD, Flamm SD, Desai 30. Budoff MJ, Yang TP, Shavelle RM, Lamont DH, Brundage

MY. Prognostic utility of 64-slice computed tomography in patients BH.  Ethnic differences in coronary atherosclerosis. J Am Coll
with suspected but no documented coronary artery disease. Eur Cardiol. 2002;39:408–12.
Heart J. 2009;30:362–71. 31. Shaw LJ, Shaw RE, Merz CN, et  al., and American College of
17. Rubinshtein R, Halon DA, Gaspar T, Peled N, Lewis BS. Cardiac Cardiology-National Cardiovascular Data Registry I. Impact of eth-
computed tomographic angiography for risk stratification and pre- nicity and gender differences on angiographic coronary artery dis-
diction of late cardiovascular outcome events in patients with a ease prevalence and in-hospital mortality in the American college
chest pain syndrome. Int J Cardiol. 2009;137:108–15. of cardiology-national cardiovascular data registry. Circulation.
18. Gopal A, Nasir K, Ahmadi N, Gul K, Tiano J, Flores M, Young E, 2008;117:1787–801.
Witteman AM, Holland TC, Flores F, Mao SS, Budoff MJ. Cardiac 32. Detrano R, Guerci AD, Carr JJ, et al. Coronary calcium as a predic-
computed tomographic angiography in an outpatient setting: an tor of coronary events in four racial or ethnic groups. N Engl J Med.
analysis of clinical outcomes over a 40-month period. J Cardiovasc 2008;358:1336–45.
Comput Tomogr. 2009;3:90–5. 33. Ferdinand KC. Coronary artery disease in minority racial and eth-
19. Abidov A, Gallagher MJ, Chinnaiyan KM, Mehta LS, Wegner JH, nic groups in the United States. Am J Cardiol. 2006;97:12A–9A.
Raff GL. Clinical effectiveness of coronary computed tomographic 34. McClelland RL, Chung H, Detrano R, Post W, Kronmal

angiography in the triage of patients to cardiac catheterization RA. Distribution of coronary artery calcium by race, gender, and
and revascularization after inconclusive stress testing: results of a age: results from the multi-ethnic study of atherosclerosis (MESA).
2-year prospective trial. J Nucl Cardiol. 2009;16:701–13. Circulation. 2006;113:30–7.
20. Min JK, Feignoux J, Treutenaere J, Laperche T, Sablayrolles
35. Hulten E, Villines TC, Cheezum MK, et  al., and Investigators
J.  The prognostic value of multidetector coronary CT angiogra- C.  Usefulness of coronary computed tomography angiography
phy for the prediction of major adverse cardiovascular events: a to predict mortality and myocardial infarction among Caucasian,
multicenter observational cohort study. Int J Cardiovasc Imaging. African and east Asian ethnicities (from the CONFIRM [coronary
2010;26:721–8. CT angiography evaluation for clinical outcomes: an international
21. Chow BJ, Wells GA, Chen L, Yam Y, Galiwango P, Abraham A, multicenter] registry). Am J Cardiol. 2013;111:479–85.
Sheth T, Dennie C, Beanlands RS, Ruddy TD. Prognostic value of 36. Villines TC, Hulten EA, Shaw LJ, et  al., and Investigators

64-slice cardiac computed tomography severity of coronary artery CR. Prevalence and severity of coronary artery disease and adverse
disease, coronary atherosclerosis, and left ventricular ejection frac- events among symptomatic patients with coronary artery calcifi-
tion. J Am Coll Cardiol. 2010;55:1017–28. cation scores of zero undergoing coronary computed tomography
22. Kristensen TS, Kofoed KF, Kuhl JT, Nielsen WB, Nielsen MB, angiography: results from the CONFIRM (coronary CT angiogra-
Kelbaek H.  Prognostic implications of nonobstructive coronary phy evaluation for clinical outcomes: an international multicenter)
plaques in patients with non-ST-segment elevation myocardial registry. J Am Coll Cardiol. 2011;58:2533–40.
infarction: a multidetector computed tomography study. J Am Coll 37. Al-Mallah MH, Qureshi W, Lin FY, et al. Does coronary CT angi-
Cardiol. 2011;58:502–9. ography improve risk stratification over coronary calcium scoring
23. Min JK, Dunning A, Lin FY, et  al. Rationale and design of the in symptomatic patients with suspected coronary artery disease?
CONFIRM (Coronary CT angiography evaluation for clinical out- Results from the prospective multicenter international CONFIRM
comes: an international multicenter) registry. J Cardiovasc Comput registry. Eur Heart J Cardiovasc Imaging. 2014;15:267–74.
Tomogr. 2011;5:84–92. 38. Cho I, Chang HJ, Sung JM, et al., and Investigators C. Coronary
24. Leipsic J, Taylor CM, Grunau G, et al. Cardiovascular risk among computed tomographic angiography and risk of all-cause mortality
stable individuals suspected of having coronary artery disease with and nonfatal myocardial infarction in subjects without chest pain
670 A. Rizvi et al.

syndrome from the CONFIRM Registry (coronary CT angiography the society for cardiovascular magnetic resonance. J Cardiovasc
evaluation for clinical outcomes: an international multicenter regis- Comput Tomogr. 2010;4:407 e1–33.
try). Circulation. 2012;126:304–13. 50. Halpern EJ.  Triple-rule-out CT angiography for evaluation of

39. Buse JB, Ginsberg HN, Bakris GL, et al., American Heart A and acute chest pain and possible acute coronary syndrome. Radiology.
American Diabetes A.  Primary prevention of cardiovascular dis- 2009;252:332–45.
eases in people with diabetes mellitus: a scientific statement 51. Chinnaiyan KM, Depetris AM, Al-Mallah M, et  al. Rationale,

from the American Heart Association and the American Diabetes design, and goals of the advanced cardiovascular imaging consor-
Association. Circulation. 2007;115:114–26. tium (ACIC): a blue cross blue shield of Michigan collaborative
40. Rana JS, Dunning A, Achenbach S, et al. Differences in prevalence, quality improvement project. Am Heart J. 2012;163:346–53.
extent, severity, and prognosis of coronary artery disease among 52. Burris AC 2nd, Boura JA, Raff GL, Chinnaiyan KM. Triple rule out
patients with and without diabetes undergoing coronary computed versus coronary CT angiography in patients with acute chest pain:
tomography angiography: results from 10,110 individuals from results from the ACIC consortium. JACC Cardiovasc Imaging.
the CONFIRM (COronary CT angiography evaluation for clini- 2015;8:817–25.
cal outcomes): an international multicenter registry. Diabetes Care. 53. Liao L, Kong DF, Shaw LK, Sketch MH Jr, Milano CA, Lee KL,
2012;35:1787–94. Mark DB. A new anatomic score for prognosis after cardiac cath-
41. Blanke P, Naoum C, Ahmadi A, et al. Long-term prognostic utility eterization in patients with previous bypass surgery. J Am Coll
of coronary CT angiography in stable patients with diabetes mel- Cardiol. 2005;46:1684–92.
litus. JACC Cardiovasc Imaging. 2016;9(11):1280–8. 54. Weustink AC, Nieman K, Pugliese F, Mollet NR, Meijboom WB,
42. Arsanjani R, Berman DS, Gransar H, et  al., and Investigators
van Mieghem C, ten Kate GJ, Cademartiri F, Krestin GP, de Feyter
C.  Left ventricular function and volume with coronary CT angi- PJ. Diagnostic accuracy of computed tomography angiography in
ography improves risk stratification and identification of patients patients after bypass grafting: comparison with invasive coronary
at risk for incident mortality: results from 7758 patients in the angiography. JACC Cardiovasc Imaging. 2009;2:816–24.
prospective multinational CONFIRM observational cohort study. 55. Bassri H, Salari F, Noohi F, Motevali M, Abdi S, Givtaj N, Raissi K,
Radiology. 2014;273:70–7. Haghjoo M. Evaluation of early coronary graft patency after coro-
43. Samad Z, Hakeem A, Mahmood SS, Pieper K, Patel MR, Simel DL, nary artery bypass graft surgery using multislice computed tomog-
Douglas PS. A meta-analysis and systematic review of computed raphy angiography. BMC Cardiovasc Disord. 2009;9:53.
tomography angiography as a diagnostic triage tool for patients 56. Lee R, Lim J, Kaw G, Wan G, Ng K, Ho KT. Comprehensive non-
with chest pain presenting to the emergency department. J Nucl invasive evaluation of bypass grafts and native coronary arteries in
Cardiol. 2012;19:364–76. patients after coronary bypass surgery: accuracy of 64-slice mul-
44. Goldstein JA, Chinnaiyan KM, Abidov A, et al., and Investigators tidetector computed tomography compared to invasive coronary
C-S. The CT-STAT (coronary computed tomographic angiography angiography. J Cardiovasc Med (Hagerstown). 2010;11:81–90.
for systematic triage of acute chest pain patients to treatment) trial. 57. Hamon M, Lepage O, Malagutti P, Riddell JW, Morello R, Agostini
J Am Coll Cardiol. 2011;58:1414–22. D, Hamon M. Diagnostic performance of 16- and 64-section spi-
45. Litt HI, Gatsonis C, Snyder B, et al. CT angiography for safe dis- ral CT for coronary artery bypass graft assessment: meta-analysis.
charge of patients with possible acute coronary syndromes. N Engl Radiology. 2008;247:679–86.
J Med. 2012;366:1393–403. 58. Small GR, Yam Y, Chen L, et al. Prognostic assessment of coronary
46. Hoffmann U, Bamberg F, Chae CU, et  al. Coronary computed artery bypass patients with 64-slice computed tomography angiog-
tomography angiography for early triage of patients with acute chest raphy: anatomical information is incremental to clinical risk predic-
pain: the ROMICAT (rule out myocardial infarction using computer tion. J Am Coll Cardiol. 2011;58:2389–95.
assisted tomography) trial. J Am Coll Cardiol. 2009;53:1642–50. 59. Falk E, Shah PK, Fuster V. Coronary plaque disruption. Circulation.
47. Schlett CL, Banerji D, Siegel E, et al. Prognostic value of CT angi- 1995;92:657–71.
ography for major adverse cardiac events in patients with acute 60. Virmani R, Burke AP, Kolodgie FD, Farb A.  Vulnerable plaque:
chest pain from the emergency department: 2-year outcomes of the the pathology of unstable coronary lesions. J Interv Cardiol.
ROMICAT trial. JACC Cardiovasc Imaging. 2011;4:481–91. 2002;15:439–46.
48. Raff GL, Chinnaiyan KM, Cury RC, et  al. SCCT guidelines
61. Hoffmann MH, Shi H, Schmitz BL, et  al. Noninvasive coro-

on the use of coronary computed tomographic angiography nary angiography with multislice computed tomography. JAMA.
for patients presenting with acute chest pain to the emergency 2005;293:2471–8.
department: a report of the Society of Cardiovascular Computed 62. Motoyama S, Anno H, Sarai M, et al. Noninvasive coronary angi-
Tomography Guidelines Committee. J Cardiovasc Comput Tomogr. ography with a prototype 256-row area detector computed tomog-
2014;8:254–71. raphy system: comparison with conventional invasive coronary
49. Taylor AJ, Cerqueira M, Hodgson JM, Mark D, Min J, O'Gara P, angiography. J Am Coll Cardiol. 2008;51:773–5.
Rubin GD, American College of Cardiology Foundation Appropriate 63. Hoffmann U, Moselewski F, Nieman K, et al. Noninvasive assess-
Use Criteria Task F, Society of Cardiovascular Computed T, ment of plaque morphology and composition in culprit and stable
American College of R, American Heart A, American Society of lesions in acute coronary syndrome and stable lesions in stable
E, American Society of Nuclear C, North American Society for angina by multidetector computed tomography. J Am Coll Cardiol.
Cardiovascular I, Society for Cardiovascular A, Interventions and 2006;47:1655–62.
Society for Cardiovascular Magnetic R. ACCF/SCCT/ACR/AHA/ 64. Motoyama S, Kondo T, Anno H, Sugiura A, Ito Y, Mori K, Ishii
ASE/ASNC/NASCI/SCAI/SCMR 2010 Appropriate use criteria J, Sato T, Inoue K, Sarai M, Hishida H, Narula J. Atherosclerotic
for cardiac computed tomography. A report of the American col- plaque characterization by 0.5-mm-slice multislice computed
lege of cardiology foundation appropriate use criteria task force, tomographic imaging. Circ J. 2007;71:363–6.
the society of cardiovascular computed tomography, the American 65. Achenbach S, Moselewski F, Ropers D, et  al. Detection of cal-
college of radiology, the American heart association, the American cified and noncalcified coronary atherosclerotic plaque by
society of echocardiography, the American society of nuclear car- contrast-­enhanced, submillimeter multidetector spiral computed
diology, the north American society for cardiovascular imaging, tomography: a segment-based comparison with intravascular ultra-
the society for cardiovascular angiography and interventions, and sound. Circulation. 2004;109:14–7.
51  Prognosis and Outcome: State of the Evidence 671

66. Motoyama S, Kondo T, Sarai M, et al. Multislice computed tomo- emission computed tomography. J Cardiovasc Comput Tomogr.
graphic characteristics of coronary lesions in acute coronary syn- 2008;2:93–101.
dromes. J Am Coll Cardiol. 2007;50:319–26. 71. van Werkhoven JM, Schuijf JD, Gaemperli O, et  al. Prognostic
67. Motoyama S, Sarai M, Harigaya H, et al. Computed tomographic value of multislice computed tomography and gated single-photon
angiography characteristics of atherosclerotic plaques subse- emission computed tomography in patients with suspected coro-
quently resulting in acute coronary syndrome. J Am Coll Cardiol. nary artery disease. J Am Coll Cardiol. 2009;53:623–32.
2009;54:49–57. 72. Investigators S-H. CT coronary angiography in patients with sus-
68. Imazeki T, Sato Y, Inoue F, Anazawa T, Tani S, Matsumoto
pected angina due to coronary heart disease (SCOT-HEART):
N, Takayama T, Uchiyama T, Saito S.  Evaluation of coronary an open-label, parallel-group, multicentre trial. Lancet.
artery remodeling in patients with acute coronary syndrome 2015;385:2383–91.
and stable angina by multislice computed tomography. Circ J. 73. Hadamitzky M, Freissmuth B, Meyer T, et al. Prognostic value of
2004;68:1045–50. coronary computed tomographic angiography for prediction of
69. Puchner SB, Liu T, Mayrhofer T, et al. High-risk plaque detected cardiac events in patients with suspected coronary artery disease.
on coronary CT angiography predicts acute coronary syndromes JACC Cardiovasc Imaging. 2009;2:404–11.
independent of significant stenosis in acute chest pain: results from 74. Beigel R, Brosh S, Goitein O, et al. Prognostic implications of non-
the ROMICAT-II trial. J Am Coll Cardiol. 2014;64:684–92. obstructive coronary artery disease in patients undergoing coronary
70. Shaw LJ, Berman DS, Hendel RC, Borges Neto S, Min JK, Callister computed tomographic angiography for acute chest pain. Am J
TQ.  Prognosis by coronary computed tomographic angiography: Cardiol. 2013;111:941–5.
matched comparison with myocardial perfusion single-photon
 Cardiac CT: Comparative
Cost-­Effectiveness 52
Christopher L. Schlett

The Need for Comparative Cost-Effectiveness unclear compared to the existing, standard-of-care practice,
like it was for cardiac CT.  Comparisons can become com-
Across the world, but particularly in developed countries, we plex when more than one outcome is being considered; for
are facing the problem of rising health-care costs in the era of example, a new diagnostic test is more effective but has sub-
limited resources. Between 2005 and 2015, the health-care stantially higher severe side effects and associated costs.
expenditure as portion of the gross domestic product Today, CEA is one of the multiple considerations that form
increased by 12% within the OECD countries while the USA reimbursement decisions by different institutions across the
remains the country with the highest expenditure (2015: $3.2 world including, e.g., the National Institute for Health and
trillion, 16.9% of the gross domestic product) [1, 2]. Since Care Excellence (NICE) in the UK and the Institute for
the increase in spending is limited at some point, we have to Quality and Efficiency in Health Care in Germany. However,
make decisions on different levels. Unlike most daily deci- from an industry point of view, it can be argued that CEA is
sion, many health-care decisions have significant conse- a form of price control, which leads to reduced returns to
quences and involve essential uncertainties and trade-offs. industry and to lower research and development expenditure
However, high health-care expenditures are not necessarily and stifles future innovations. Nevertheless, CER and CEA
inappropriate, and any level of spending might be considered became a popular and important tool in medicine including
“appropriate,” depending on the value gained by that the evaluation of novel diagnostic test strategies like cardiac
investment. CT.
The aim of comparative effectiveness research (CER) in
general is to compare alternative approaches to patient care.
Using randomized controlled trials (RCTs) – one of the three A Primer for Cost-Effectiveness Analysis
important types of CER – broadly accepted evidence can be
generated to support one or the other strategy; however, A few basic elements should briefly be reviewed for a better
RCTs also have their limitations, e.g., measuring long-term understanding of the results derived from CEA regarding
effects. Cost-effectiveness analysis (CEA) is another method cardiac CT. While costs seem to be a more simple measure,
in the broader CER framework although comparative cost the quantification of health outcomes is complex  – a typi-
data are currently not routinely evaluated in most countries, cally used metric is quality-adjusted life-year (QALY). It
including the USA [3]. CEA is one of the five major cost takes into account both, the quality and the quantity of life
analysis types (cost minimization, cost-benefit, cost-­ lived, using the following equation: “QALY = time * utility.”
consequence, cost-effectiveness, cost-utility) and the most The utility itself covers multiple health metrics which are
common type published in medical literature beside cost-­ combined into a single overall measure and ranges from 0
utility analysis. In general, CEA evaluates the difference in (equals to “dead”) to 1 (equals to “perfect health”). For all
health outcomes for two or more interventions or diagnostic other health states in-between, there is a deduction [4]; for
strategies in relation to their cost difference. example, for major cardiac events, commonly used utility
The need for CER, and particularly CEA, usually arises weights range from 0.70 to 0.85 [5], but utility weights vary
when the value of a new intervention or diagnostic test is by many factors (e.g., gender) and must be carefully
chosen.
Another challenge is to derive accurate cost estimates,
C. L. Schlett (*) and both measuring and modeling of this parameter can be
Department of Diagnostic and Interventional Radiology, University
Medical Center Freiburg, Freiburg, Germany
complex, too. The costs of a cardiac CT exam include the
e-mail: [email protected] fees associated with performing the scan as well as

© Humana Press 2019 673

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_52
674 C. L. Schlett

d­ ownstream costs associated with the imaging test (includ- over time (e.g., nonfatal myocardial infarcts) and/or for
ing additional diagnostic tests, therapeutic procedures, etc.). ­modeling predictable events that occur over time (e.g., devel-
Even the actual cost just for the CT scan is difficult to deter- oping coronary stenosis). Central elements of all Markov
mine since it is often a fraction of the charged price and var- models are mutually exclusive health states and the transi-
ies among hospitals, states, and countries. This inherent lack tion between states, which reflect events (Fig.  52.2). The
of consistency in costs can be a stumbling block for accurate time spent in each health state determines overall expected
CEA. In the USA, many CEAs use cost estimates typically outcome. Analyzing the model often includes sensitivity
derived from the US Centers for Medicare and Medicaid analysis and/or Monte Carlo simulations to model the given
Services (CMS) average payments. uncertainty.
A frequently used tool in CEA is the incremental cost-­
effectiveness ratio (ICER). It was initially described in 1977
by Drs. Weinstein and Stason and serves as a summary mea-  omparative Cost-Effectiveness of Cardiac
C
sure in the comparison of two different interventions or CT for Acute Chest Pain Patients
diagnostic strategies (e.g., A vs. B) while it is defined as
“ICER = [cost-A - cost-B] / [outcome-A - outcome-B]” [6]. Two major RCTs have been conducted assessing comparative
The ICER can be used as a decision rule in resource alloca- effectiveness and cost of cardiac CT in patients with suspi-
tion if a willingness-­to-pay value can be determined; e.g., in cion of acute coronary syndrome  – ACRIN-PA (American
the UK, NICE has formally acknowledged a range from College of Radiology Imaging Network, Pennsylvania) and
£20,000 to £30,000 per QALY, but higher thresholds can be ROMICAT II (Rule Out Myocardial Infarction using
argued for special circumstances such as treatments for rare Computer Assisted Computed Tomography) [9, 10].
conditions; in other European countries, a threshold of Optimally, a strategy including early CCTA would reduce
€80,000 per QALY and in the USA a threshold of $50,000 costs and provide more cost-effective care in the triage of
per QALY are often considered as appropriate. However, patients with suspicion of acute coronary syndrome. Both
these thresholds may not necessarily reflect a society’s will- studies found a significant reduction in length of stay and a
ingness to pay, and alternative methods have been sug- higher rate of direct discharge from the emergency depart-
gested, including thresholds based on per capita incomes or ment with no alteration in safety [9, 10]. In ROMICAT II, no
benchmark interventions [7]. reduction in overall costs was observed [9]; costs in the emer-
Also, the chosen time horizon can impact the results sig- gency department decreased ($2101 vs. $2566 in CCTA vs.
nificantly. CEA assessing cardiac CT for stable chest pain standard of care), while in contrast, the costs during hospital-
patients are more likely to choose a lifetime horizon given ization slightly increased using CCTA ($4026 vs. $3874),
the chronic disease character, while CT for acute chest pain potentially due to the higher number of revascularizations [9].
would focus more on short-term costs/effects. Similarly, the Cost data for ACRIN-PA have not yet been released.
chosen perspective has great impact on the CEA results. Only a few CEA studies, limited to the USA, are available
Each stakeholder within the health-care system (e.g., gov- comparing different triage strategies in patients with suspicion
ernments, insurance companies, hospitals, physicians, of acute coronary syndrome. In a study by Khare et al. [11],
employers, patients) has a different perspective with regard the authors found that a triage strategy including CCTA would
to cost and effect; thus, what could be considered cost-­ dominate other strategies, such as stress echocardiography or
effective from a societal standpoint is not necessarily cost-­ stress ECG, regarding its cost-effectiveness (ICER, $29,738/
effective from the perspective of an individual hospital. QALY for CCTA, if compared with stress echocardiography,
Nevertheless, most CEA studies apply the societal perspec- and $7332/QALY, if compared with the stress ECG).
tive in the academic environment [8]. In another study by Ladapo et al. using first-order Monte
Before deriving the CEA results, the therapeutic and diag- Carlo microsimulations for 55-year-old men and women, a
nostic strategies must be formulated. One way are decision triage strategy including early CCTA resulted in an ICER of
trees, which are a schematic representation of all of the clini- $6400/QALY for men, and it was actually cost-saving for
cally and policy relevant features of the decision problem. women ($6630 for CCTA and $7010 for standard of care)
They include “Decision Nodes” (a choice between alterna- [12]. This finding was driven by their observation that the
tives strategies), “Chance Nodes” (possible events deter- CCTA strategy was cheaper in women than in men ($6630
mined by probability), and “Terminal Nodes” (outcomes vs. $10,190). These findings were sensitive to several param-
associated with a given pathway) as illustrated in Fig. 52.1. eters, including correctly classifying a patient as healthy by
Another approach are Markov models, which represent a CCTA or the time horizon over which patients could return
mathematical modeling technique derived from matrix alge- to the emergency department, but the authors found that
bra. Those are recursive decision trees that are used for mod- CCTA remained in the range of what is typically considered
eling conditions that have events which may occur repeatedly cost-effective in most sensitivity analyses [12].
52  Cardiac CT: Comparative Cost-­Effectiveness 675

Fig. 52.1  Decision Tree. A True-Positive

simplified version of a Outcome-1
decision tree comparing
different diagnostic tests vs. Test-A Positive
no testing. The decision tree
consists typically of three
False-Positive
node types: “Decision Test-A Outcome-2
Nodes,” representing a choice
between alternatives
strategies; “Chance Nodes,”
possible events determined by Test-A Negative
a probability; and “Terminal
Nodes,” outcomes associated
with a given pathway. By
“Rolling Back” the decision Test-B
tree, costs and health outcome
can be determined for each
test strategy
No Testing
Outcome-999

Decision Node Chance Node Terminal Node

Initial Healthy
Healthy
Post-
Cycle 1 Healthy
Event
Dead

Post- Post-
Event Cycle 2 Healthy
Event
Dead

Post-
Cycle 3 Healthy
Event
Dead

Dead
Post-
Cycle 4 Healthy
Event
Dead

Fig. 52.2  Example of a Markov model and Monte Carlo simulation. The to evaluate a Markov model. The simulation determines the prognoses of a
Markov model (left part) consists of different health states with one starting large number of individual patients. Each patient begins in the starting state,
state (e.g., “Healthy”) and one absorbing state (e.g., “Dead”), which is a and at the end of each cycle, the transition to another health state is based
state that, once entered, cannot be left. A patient is always in one of a finite on a probability. The patient is given credit for each cycle spent in a non-
number of discrete health states. All events are represented as transitions absorbing state, and each state may be adjusted for quality of life. The
from one state to another. Monte Carlo simulation (right part) can be used simulation is stopped when the patient enters the absorbing state

Thus, CCTA may have the potential to be a cost-effective [13, 14]. Regarding the effectiveness determined as occur-
alternative compared to a traditional triage of patients with rence of major cardiac events, PROMISE revealed low
suspicion of acute coronary syndrome; however, it has to event rates, which showed no difference between the two
hold up against other, newer strategies such as using high-­ strategies at 2.1 years (3.3% vs. 3.0% for cardiac CTA vs.
sensitivity troponin in future research. functional testing, p = 0.75). In SCOT-HEART, event rate
(overall death or MI) was about twofold higher than for the
PROMISE trial and showed a reduction of nearby 40% by
 omparative Cost-Effectiveness Analysis
C cardiac CT vs. standard of care (HR, 0.62; 95% CI, 0.38–
of Cardiac CT for Stable Chest Pain 1.01; p = 0.053). In general, the use of diagnostic testing
improves angina and quality of life; however, this improve-
Comparative effectiveness and cost of cardiac CT to other ment does not vary significantly between the randomized
testing strategies for patients with stable chest pain have testing strategy, which was also observed in other, smaller
been evaluated in several randomized controlled trials, studies assessing health status by, e.g., the Seattle Angina
including the two large trials PROMISE and SCOT-HEART Questionnaire (SAQ) [15].
676 C. L. Schlett

Assessing the comparative costs for the diagnostic testing Nevertheless, many CEA studies suggested that strate-
strategies, SCOT-HEART trial showed slightly higher costs gies which began with CCTA and were followed by cardiac
in the CCTA arm at 6 months (difference, $462, p < 0.0001), stress imaging were cost-effective for patients with stable
which mainly reflects the difference in upfront procedural chest pain [22–24]. However, in sensitivity analyses, the
cost. There were neither differences in cost associated with optimal diagnostic imaging strategy depends strongly on the
outpatient and inpatient services nor medication use (all pretest CAD probability (Fig. 52.3). For CCTA, cost-effec-
p ≥ 0.16). Similarly, in the PROMISE trial, the costs were tiveness was often reached in cohorts with low-intermediate
not different through 3 years of follow-up (difference: $627, pretest probability of CAD; in contrast, ICA becomes a
p = NS) [16]. cost-­effective strategy in patients with a high pretest CAD
Although 3 years is a relative long follow-up for a random- probability (≥70%). But sensitivity analysis revealed also
ized controlled trial, it cannot reflect a lifelong horizon, which other influencing factors, such as test characteristics and
is relevant for decision-making. More importantly, stable costs of a test.
chest pain must be often considered as a chronic disease, and In summary, results from different CERs suggest that
therefore a lifelong horizon becomes even more relevant. CCTA may serve as an initial gatekeeper test resulting in
Several CEAs have been conducted, including long-term cost-effectiveness under varying assumptions in patients pre-
outcome modeling such as Markov cohort simulation and senting with stable chest pain with a low to intermediate pre-
state-transition microsimulation [17]. These studies demon- test probability of CAD. Similarly, the 2016 updated NICE
strate a high heterogeneity; the most relevant studies have guidelines, which considered already the results of the
been listed in Table  52.1. One of the most comprehensive PROMISE and SCOT-HEART trial, recommend CCTA as a
CEAs was published in 2015 by Genders et al. which consid- first-line modality for patients presenting with new-onset
ered most but not all possible imaging tests as well as several chest pain due to suspected CAD in the UK [25].
national perspectives, including the USA, the UK, and the
Netherlands [18]. They found that a strategy that began with
CCTA, continued with cardiac stress imaging if CCTA found Cost-Effectiveness Analysis
at least coronary stenosis, and that ended with invasive coro- Regarding Reporting Incidental
nary angiography (ICA) if stress imaging induced any isch- Findings on Cardiac CT
emia, maximized QALYs and was cost-effective in the USA
(ICER, $22,000/QALY for men and $21,000/QALY for While most CEAs regarding cardiac CT mainly focus
women) and the Netherlands (ICER, €38,000/QALY for whether CCTA is a cost-effective diagnosis strategy for
men and €18,000/QALY for women). For the UK, the results chest pain patients, other questions remain open and can
were more divergent. For men, the preferred strategy with an be addressed with similar statistical techniques. One
ICER of £7000/QALY began with CCTA and continued with example is a CEA study by Goehler et  al. [26]. Their
optimal medical therapy without ICA, if CCTA found only objective was to assess from a society perspective and life-
moderate CAD, or if stress imaging induced only mild isch- long horizon – whether reporting pulmonary nodules, inci-
emia. In women, the preferred strategy was stress echocar- dentally detected in CCTA, would be cost-effective. Using
diography followed by ICA if echocardiography induced a validated lung cancer simulation model, they found for a
mild or moderate ischemia (ICER £7000/QALY). typical chest pain cohort that reporting and following up
A comparable study using gender-stratified analysis by pulmonary nodules according to the guidelines of the
Ladapo et  al. [19] using a typical case of a 55-year-old international societies led to a reduction in cumulative
female or male patient with stable chest pain found that lung cancer mortality (4.6% relative reduction) and con-
CCTA with stress testing was the performed strategy with an secutively to higher QALY; however, also downstream
ICER ranging from $26,200/QALY in men to $35,000/ testing increased. Thus, the ICER for the overall cohort
QALY in women; however, differences in health outcomes was $154,700/QALY, which is significantly above the nor-
were small, and CCTA raised overall costs, partly through mally assumed willingness-to-pay threshold. One reason
the follow-up of incidental findings. Although results from for this observation is the ­phenomenon of competing risks
Medicare data showed a decrease of overall CAD-related where the majority of patients undergoing cardiac CT die
costs after 9 months if patients underwent CCTA rather than of CAD and other causes, rather than of lung cancer [26,
imaging stress test [20], a higher rate of invasive downstream 27]. Even if restricting the cohort to smokers, an ICER of
testing after CCTA has been reported, for example, in the $129,800/QALY remained above the assumed willingness-
SPARC registry and the PROMISE trial [13, 21]. The long-­ to-pay threshold.
term value regarding costs and, more importantly, regarding Although these CEA data do not support reporting pul-
health outcome of this is quite controversial and only par- monary nodules incidentally detected on cardiac CT, the
tially included in the CEA studies [17]. international guidelines still recommend to report incidental
Table 52.1  Cost-effectiveness analysis of cardiac CT for stable chest pain
Year of Suggested initial
publication, Threshold imaging test for
first author, Country of Imaging algorithms willingness to pay low-intermediate Optimal strategy
Reference Time frame Perspective analysis compared ($ or € /QALY) Reference case analysis pretest probability depends on
2009, Ladapo Lifetime Health-care USA CCTA + xECG/ICA, US$50,000/ 55-year-old men and CCTA Pretest probability,
[19] system xECG + CCTA/ICA, QALY women with pretest (risks of) radiation
TCA ± ICA, xECG ± ICA, probability of 70% and exposure, and
xECHO ± ICA, 30%, respectively sensitivity of CCTA
SPECT ± ICA, ICA, vs. no
testing
2009, Lifetime Patient, Netherlands, CCTA ± ICA vs. ICA €80,000/QALY 60-year-old men and CCTA Optimization criterion
Genders [23] physician, USA women (men pretest (i.e., outcomes),
hospital, probability of 79%; pretest probability,
health-care women pretest probability sensitivity/specificity
52  Cardiac CT: Comparative Cost-­Effectiveness

system, society of 65%) of CCTA

2009, Kreisz Short term Health-care Australia CCTA ± ICA vs. ICA AUS$50,000 / Patients referred for ICA CCTA Pretest probability
[28] (10 years) system QALY with pretest probability of
10–90%
2010, Min Short term Health-care USA CCTA ± ICA, US$50,000 / 55-year-old men with CCTA CCTA sensitivity,
[29] and payer CCTA ± SPECT/ICA, QALY pretest probability of 30% SPECT sensitivity,
lifetime SPECT ± ICA, pretest probability,
SPECT ± CCTA/ICA, cost tests
vs. ICA
2012, Meyer Short term Health-care USA CCTA (dual energy) vs. n/a 51- to 71-year-old men and CCTA Sensitivity, age, cost
[30] payer SPECT women with pretest tests, pretest
probability of 80% probability
2013, Short term Health-care Netherlands CCTA vs. xECG €80,000/QALY 46- to 66-year-old men and CCTA Pretest probability
Genders [31] and system women with pretest
lifetime probability of 70%
2015, Lifetime Health-care USA, UK, CCTA, Stress Test (CMR, For USA: 60-year-old men and CCTA (for USA, Pretest probability,
Genders [18] system, society Netherlands SPECT, xECHO), CCTA ± $50,000/QALY women with low to Netherlands, UK False-positive rate
Stress Test (CMR , SPECT, For UK: £25,000/ intermediate probability men)xECHO (for
xECHO), vs. ICA QALY for CAD UK women)
For Netherlands:
€80,000/QALY
2016, Agus Short term Health-care UK CCTA vs. Stress Test £20,000/QALY 57.8 ± 10.0 years, 56% CCTA Pretest probability
[32] (1 year) system male; 42/22/36% of low-/
intermediate-/high- pretest
probability of CAD
2016, Lifelong Health-care Brazilian CCTA, Stress Test (xECG, I$11,909-­35,727/ 60 years, intermediate CCTA Cost tests, pretest
Bertoldi [33] system xECHO, CMR) vs. ICA QALY (50%) pretest probability probability
of CAD
Abbreviations used: CMR cardiac magnetic resonance imaging, ICA invasive angiography, CCTA coronary CT angiography, QALY quality-adjusted life-year, ICER incremental cost-effectiveness
ratio, xECG stress electrocardiography, xECHO stress echocardiography
677
678 C. L. Schlett

Fig. 52.3  Sensitivity analysis 100%

regarding pretest probability
for CAD. The bar graphs
demonstrate under which 90%
assumed pretest probabilities
for CAD CCTA remain the
preferred strategy. USA 80%
denotes the United States, UK
the United Kingdom, NL the m w

Pre-Test Probability for CAD

Netherlands, m male patients, 70%
w female patients

60%

m
50%

40%

w m
30%

20%

w
10%

US UK NL
0%
2015, Genders [18]
2009, Genders [23]

2009, Kreisz [28]

2010, Min [29]

2016, Agus [32]

2016, Bertoldi [33]

findings, and most radiologists follow these r­ ecommendations 7. Marseille E, et  al. Thresholds for the cost-effectiveness of inter-
across the word, illustrating that decision-making is not only ventions: alternative approaches. Bull World Health Organ.
2015;93:118–24.
a process of economic evaluation. 8. Russell LB, et al. The role of cost-effectiveness analysis in health
and medicine. Panel on cost-effectiveness in health and medicine.
JAMA. 1996;276(14):1172–7.
References 9. Hoffmann U, et al. Coronary CT angiography versus standard eval-
uation in acute chest pain. N Engl J Med. 2012;367(4):299–308.
10. Litt HI, et  al. CT angiography for safe discharge of patients

1. OECD.Stat, Health expenditure and financing. Joint OECD, with possible acute coronary syndromes. N Engl J Med.
EUROSTAT and WHO Health Accounts SHA Questionnaires 2012;366(15):1393–403.
(JHAQ). https://stats.oecd.org/index.aspx?DataSetCode=SHA 11. Khare RK, et al. Sixty-four-slice computed tomography of the cor-
2. US Centers for Medicare & Medicaid Services National Health onary arteries: cost-effectiveness analysis of patients presenting to
Expenditures 2015 Highlights. https://ccf.georgetown.edu/wp-con- the emergency department with low-risk chest pain. Acad Emerg
tent/uploads/pdfs/highlights.pdf Med. 2008;15(7):623–32.
3. Goehler A, Gazelle GS.  Examining the use of comparative and 12. Ladapo JA, et  al. Cost-effectiveness of coronary MDCT in the
cost-effectiveness analyses in radiology. AJR Am J Roentgenol. triage of patients with acute chest pain. AJR Am J Roentgenol.
2014;203(5):939–44. 2008;191(2):455–63.
4. Torrance GW. Measurement of health state utilities for economic 13. Douglas PS, et al. Outcomes of anatomical versus functional test-
appraisal. J Health Econ. 1986;5(1):1–30. ing for coronary artery disease. N Engl J Med. 2015;372(14):
5. Hanmer J, et  al. Report of nationally representative values for 1291–300.
the noninstitutionalized US adult population for 7 health-related 14.
Investigators, S-H.  CT coronary angiography in patients
quality-­of-life scores. Med Decis Mak. 2006;26(4):391–400. with suspected angina due to coronary heart disease (SCOT-­
6. Weinstein MC, Stason WB.  Foundations of cost-­ effectiveness HEART): an open-label, parallel-group, multicentre trial. Lancet.
analysis for health and medical practices. N Engl J Med. 2015;385(9985):2383–91.
1977;296(13):716–21.
52  Cardiac CT: Comparative Cost-­Effectiveness 679

15. Min JK, et al. Coronary CT angiography versus myocardial perfu- in patients with an intermediate pretest likelihood for coronary
sion imaging for near-term quality of life, cost and radiation expo- artery disease. Heart. 2012;98(6):460–7.
sure: a prospective multicenter randomized pilot trial. J Cardiovasc 25. Moss AJ, et  al. The updated NICE guidelines: cardiac CT as the
Comput Tomogr. 2012;6(4):274–83. first-line test for coronary artery disease. Curr Cardiovasc Imaging
16. Fordyce CB, Newby DE, Douglas PS. Diagnostic strategies for the Rep. 2017;10(5):15.
evaluation of chest pain: clinical implications from SCOT-HEART 26. Goehler A, et  al. Cost-effectiveness of follow-up of pulmonary
and PROMISE. J Am Coll Cardiol. 2016;67(7):843–52. nodules incidentally detected on cardiac computed tomographic
17. van Waardhuizen CN, et al. Diagnostic performance and compara- angiography in patients with suspected coronary artery disease.
tive cost-effectiveness of non-invasive imaging tests in patients Circulation. 2014;130(8):668–75.
presenting with chronic stable chest pain with suspected coro- 27. Lu MT, et al. Lung cancer screening eligibility in the community:
nary artery disease: a systematic overview. Curr Cardiol Rep. cardiovascular risk factors, coronary artery calcification, and car-
2014;16(10):537. diovascular events. Circulation. 2016;134(12):897–9.
18. Genders TS, et al. The optimal imaging strategy for patients with 28. Kreisz FP, et  al. The pre-test risk stratified cost-effectiveness of
stable chest pain: a cost-effectiveness analysis. Ann Intern Med. 64-slice computed tomography coronary angiography in the detec-
2015;162(7):474–84. tion of significant obstructive coronary artery disease in patients
19. Ladapo JA, et  al. Clinical outcomes and cost-effectiveness of
otherwise referred to invasive coronary angiography. Heart Lung
coronary computed tomography angiography in the evaluation of Circ. 2009;18(3):200–7.
patients with chest pain. J Am Coll Cardiol. 2009;54(25):2409–22. 29. Min JK, et al. Cost-effectiveness of coronary CT angiography ver-
20. Min JK, et al. Costs and clinical outcomes in individuals without sus myocardial perfusion SPECT for evaluation of patients with
known coronary artery disease undergoing coronary computed chest pain and no known coronary artery disease. Radiology.
tomographic angiography from an analysis of Medicare category 2010;254(3):801–8.
III transaction codes. Am J Cardiol. 2008;102(6):672–8. 30. Meyer M, et  al. Cost-effectiveness of substituting dual-

21. Hachamovitch R, et al. Patient management after noninvasive car- energy CT for SPECT in the assessment of myocardial perfu-
diac imaging results from SPARC (study of myocardial perfusion sion for the workup of coronary artery disease. Eur J Radiol.
and coronary anatomy imaging roles in coronary artery disease). J 2012;81(12):3719–25.
Am Coll Cardiol. 2012;59(5):462–74. 31. Genders TS, et al. Coronary computed tomography versus exercise
22. Genders TS, et al. Coronary computed tomography versus exercise testing in patients with stable chest pain: comparative effectiveness
testing in patients with stable chest pain: comparative effectiveness and costs. Int J Cardiol. 2013;167(4):1268–75.
and costs. Int J Cardiol. 2013;167(4):1268–75. 32. Agus AM, et al. The cost-effectiveness of cardiac computed tomog-
23. Genders TS, et al. CT coronary angiography in patients suspected of raphy for patients with stable chest pain. Heart. 2016;102(5):356–62.
having coronary artery disease: decision making from various per- 33. Bertoldi EG, et al. Long-term cost-effectiveness of diagnostic tests
spectives in the face of uncertainty. Radiology. 2009;253(3):734–44. for assessing stable chest pain: modeled analysis of anatomical and
24. Dorenkamp M, et  al. Direct costs and cost-effectiveness of dual-­ functional strategies. Clin Cardiol. 2016;39(5):249–56.
source computed tomography and invasive coronary angiography
 Barriers to Greater Clinical
Implementation 53
David C. Levin

Coronary artery disease (CAD) is a ubiquitous and danger- the primary codes for MPI and SE and are not affected by
ous disease. It was estimated that in the United States in code bundling that occurred in echocardiography in 2009 or
2010, almost 1.5 million individuals either died of CAD or MPI in 2010.
experienced a new nonfatal myocardial infarction (MI), a Figure 53.1 shows the trends in MPI. The total rate per
recurrent MI, or a new silent MI [1]. Other than in patients 1000 rose sharply from 63.4 in 2001 to peak at 88.0 in 2006,
who present with an acute MI and can be diagnosed by an an increase of 39% in only 5 years. The next 3 years saw a
ECG and/or enzyme measurements, the way to definitively gradual progressive decline. Beginning in 2010, the decline
establish the presence of CAD is through the use of noninva- accelerated, and by 2014 the rate had dropped to 58.7, i.e.,
sive imaging  – principally radionuclide myocardial perfu- below what it had been 13 years earlier. Most MPI exams are
sion imaging (MPI), stress echocardiography (SE), or performed by cardiologists. Their utilization rate trend gen-
coronary CT angiography (CCTA). None of the federal or erally parallels the trend for total MPI exams and shows a
commercial payers have yet agreed to cover the use of these sharp increase through 2006, followed by a relatively flat
techniques for screening asymptomatic individuals, but period for the next 3  years and then followed in turn by a
given the nature of CAD, clinicians should have a low thresh- prolonged decline starting in 2010. The role of radiologists is
old for ordering these tests in patients with suggestive symp- much less and never showed the same sort of increase in the
toms and risk factors. The question is, which test to use first? early 2000s. Beginning in 2006, the radiologists’ rate trend
In the author’s opinion, the answer in most patients is CCTA. demonstrated a gradual progressive decline.
Our research group (the Center for Research on Utilization One can speculate about why participation by cardiolo-
of Imaging Services  – CRUISE  – in the Department of gists rose so sharply between 2001 and 2006 and then
Radiology at Thomas Jefferson University Hospital) has dropped so sharply in more recent years. Most of the use by
studied utilization patterns of these three tests in the US cardiologists occurred in private offices. In the early 2000s,
Medicare fee-for-service population from 2001 through it was the best noninvasive test available for suspected CAD
2014. We use the nationwide Medicare Part B databases, (CCTA was in its infancy). The literature to that point had
covering over 37 million fee-for-service enrollees, but not generally supported its efficacy, and cardiologists were
including over 15 million people in managed Medicare enthusiastic about it. It was a well-reimbursed procedure that
Advantage plans. The results are surprising and in some did not require much image post-processing and was not dif-
ways disturbing. The data are presented as utilization rates ficult to read. In addition to the primary CPT-4 code for the
per 1000 Medicare fee-for-service enrollees. Total rates are MPI exam itself, there were “add-on” codes almost always
shown, as well as those by cardiologists and radiologists – used for determination of left ventricular wall motion and
the two specialties most heavily involved in cardiac imaging. ejection fraction (EF). These made the procedure even more
Although the data come from the Medicare population lucrative, since a claim for services then included three sepa-
(mostly over age 65), there is no reason to believe that the rate codes. However, beginning in 2010, the Centers for
trends are different in the younger, commercially insured Medicare and Medicaid Services (CMS) made an important
population. It should be noted that the trend lines reflect just change by bundling the add-on codes for wall motion and EF
together with the primary MPI code. From then on, when an
MPI exam was performed along with wall motion and EF
D. C. Levin (*) studies, only the new single-bundled code could be used
Department of Radiology, Thomas Jefferson University Hospital,
Philadelphia, PA, USA
when the claim was filed. The global reimbursement for the
e-mail: [email protected] single-bundled code was 36% lower than the sum of the

© Humana Press 2019 681

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_53
682 D. C. Levin

Fig. 53.1  Utilization rates of 100

radionuclide myocardial
perfusion imaging (MPI) in
the Medicare fee-for-service 90
population, 2001 through
2014. Vertical axis shows 80
examinations per 1000
beneficiaries. In addition to
the total rates, data are shown 70
for cardiologists and
radiologists, who together
provide the vast majority of 60
MPI studies. The small
proportions of examinations 50
performed by other specialists
are not shown. (Adapted from
Levin et al. [24], with 40
permission)
30

20

10

0
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Total Cardiologists Radiologists

reimbursements for the three codes prior to 2010. Since trend again closely parallels the total trend. Radiologists
SPECT cameras are expensive to purchase and maintain, and have essentially no role in SE, and therefore no trend line is
technologists must be hired to operate them, the new lower shown.
payment levels posed financial difficulties for office prac- The CCTA trends are shown in Fig. 53.3. Note that the
tices. It appears that many cardiologists closed their offices scale is far lower than that for MPI (Fig. 53.1). The year 2006
or sold them to hospitals and became hospital employees. is the first year shown because that was the first year in which
Another factor was the rise of the radiology benefits manage- specific codes for CCTA were available. In that first year, the
ment companies (RBMs). These companies and the com- total CCTA utilization rate was 0.99 per 1000. In 2007, the
mercial insurers who were their clients imposed rate more than doubled to 2.1. Indeed, it seemed as if this
preauthorization requirements for expensive high-tech pro- new and exciting procedure was poised for a rapid rise in
cedures like MPI.  No longer could a cardiologist simply use. But to the surprise and disappointment of many in the
order an MPI exam and get it done. Now, he/she had to go field, nothing of the sort happened. Instead, a sharp decline
through the process of obtaining preauthorization, a process was seen, reaching a nadir of 1.07 in 2012 and 2013. A small
that was often time-consuming and inconvenient and could increase to 1.11 occurred in 2014, but it is too early to tell if
result in a denial. With the “double whammy” of a drop in this represents the start of a more favorable trend. Utilization
reimbursement and the hassle of the preauthorization pro- rates among both cardiologists and radiologists peaked in
cess, it isn’t surprising that enthusiasm for MPI diminished. 2007 (at 1.26 and 0.67, respectively) but then declined in
The trends for SE are shown in Fig. 53.2. SE is generally subsequent years. The decline was steeper among cardiolo-
not included in preauthorization programs  – i.e., it can be gists, and their rate had dropped to only 0.47 by 2014.
ordered and performed without preauthorization. The total Among radiologists, the utilization rate reached its nadir in
utilization rate stayed relatively unchanged from 2001 2010 but then began to increase slowly, reaching 0.59  in
through 2010 but then began a 4-year period of decline. The 2014. Comparing the overall data for MPI and CCTA reveals
rate had been 12.6  in 2010 but declined to 10.2 by 2014. that in 2014, there were 53 times as many MPI studies per-
Cardiologists perform the vast majority of SEs, and their formed as CCTAs in this population.
53  Barriers to Greater Clinical Implementation 683

Fig. 53.2  Utilization rates of

stress echocardiograms (SEs) 14
in the Medicare fee-for-­
service population, 2001
through 2014. Vertical axis
shows examinations per 1000 12
beneficiaries. Total rates and
rates for cardiologists are
shown. Radiologists have
essentially no role in SE. The 10
differences between the total
rates and those for
cardiologists are attributable
8
to SEs performed by a variety
of other specialists,
particularly primary care
physicians and internists. 6
(Adapted from Levin et al.
[24], with permission)

0
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Total Cardiologists

Fig. 53.3  Utilization rates of 2.5

coronary CT angiography
(CCTA) in the Medicare
fee-for-service population,
2006 through 2014. 2006 was
the first year in which specific 2
codes for CCTA were
available. Vertical axis shows
examinations per 1000
beneficiaries. In addition to
total rates, data are shown for 1.5
radiologists and cardiologists,
who together provide the vast
majority of CCTAs. The small
proportions of examinations
provided by other specialists 1
are not shown. (Adapted from
Levin et al. [24], with
permission)

0.5

0
2006 2007 2008 2009 2010 2011 2012 2013 2014

Total Cardiologists Radiologists

684 D. C. Levin

Why Has CCTA Failed to Grow More Rapidly? i­ nvasive coronary angiography (ICA) as the gold standard in
230, 291, and 360 patients, respectively. All these patients
The answer to this question is not clear-cut and is probably had been worked up by experienced cardiologists and
multifactorial. Here are some of the reasons: (1) Multi-­ referred for ICA because of high clinical concern for
detector row CT scanners are expensive, particularly those of CAD. Many of the workups had included MPI. One of the
the latest generations. (2) CCTA can take up considerable most striking findings of the three studies was the high per-
scanner time and may slow throughput of other patients centage of patients who proved to have normal coronary
undergoing noncardiac scans. Busy hospital radiology arteries or nonobstructive disease on ICA. In the ACCURACY
departments may have difficulty allocating the necessary trial, 75% of the patients had either normal vessels or nonob-
time for cardiac exams. (3) The procedure is often labor-­ structive disease (≤50% stenosis) [2]. In the CORE 64 trial,
intensive for the interpreting physician. Formatting and post-­ 44% had either normal vessels or nonobstructive disease [3].
processing take time, and often multiple phases must be In the Dutch trial [4], there were 151 patients who were diag-
examined to get a complete exam. (4) Some patients are dif- nosed clinically with typical angina pectoris, yet 30% were
ficult to image, particularly those with heavily calcified found to have normal coronary arteries or nonobstructive
plaques; those who are obese, have rapid heart rates, or disease on ICA. There were 77 sicker patients who were
arrhythmias; or those who are unable to hold their breath or thought clinically to have unstable angina, yet 25% of them
remain still. (5) Preauthorization requests are sometimes had normal vessels or nonobstructive disease. The sickest
denied by the RBMs, and the hassle the ordering physician group of all was 50 patients thought to have non-ST-eleva-
must endure can be discouraging. RBMs and payers are tion myocardial infarction, but even in this group, 16% had
more likely to deny recently developed tests like CCTA than normal coronary arteries or nonobstructive disease. What
older ones. (6) The procedure is not well reimbursed, given these striking data from the three trials show is that the clini-
the time it takes and its labor-intensive nature. For example, cal workup of individuals with suspected CAD, which gener-
Medicare’s national professional component reimbursement ally includes the use of more traditional imaging techniques
for CCTA is $121, compared with $80 for a SPECT MPI. The like MPI or SE, is not reliable. Too many patients who do not
physician time required for the former may be 3–4× that for need to go to the cardiac catheterization lab are being sent
the latter. (7) Cardiologists may be reluctant to interpret there for ICA.
CCTAs because it would require them to report any abnor- Even more striking data comes from the study of Patel
malities in the lungs or mediastinum, areas in which they et al. [5] which reported on almost 400,000 patients in the
may not feel comfortable. (8) Old habits sometimes die hard; National Cardiovascular Data Registry of the American
cardiologists who have become accustomed to doing MPI or College of Cardiology. All received elective ICA, and 84%
SE on their patients may be unwilling to give up those had undergone prior noninvasive testing (primarily with
procedures. MPI) before ICA. Among patients with positive noninvasive
tests, 59% had either no disease or nonobstructive
CAD.  Conversely, among those with negative noninvasive
test results, 28% actually did have obstructive CAD on
Is CCTA Being Used Appropriately? ICA. In other words, the clinical workups of these patients,
most of which included noninvasive testing, yielded many
In this author’s opinion, the answer to the question is no,
false negatives and even more false positives. Here again is
CCTA is not being used appropriately. It should be used far
further evidence that too many patients are being sent for
more often than it is. In many instances, it should replace
ICA who do not need it.
MPI as the first imaging test to be used in patients with sus-
pected CAD.  There are number of reasons for this, as we
discuss below.
The Inability of MPI to Rule Out CAD

A number of studies have shown that negative results on an

 he Inadequacy of the Clinical Workup
T MPI exam do not exclude the presence of significant CAD
of Suspected CAD when It Relies on More [6–12]. Patients with multiple nonobstructive coronary
Traditional Noninvasive Imaging plaques may have normal MPI studies because such plaques
do not restrict coronary blood flow, even during exercise. But
In 2008, three major multicenter trials of CCTA were plaques of this sort can usually be detected on CCTA. This
reported – the ACCURACY trial, the CORE 64 trial, and the will allow the treating physician to begin aggressive medical
Dutch university hospital trial of Meijboom et  al. [2–4]. therapy, in an effort to avert future disease progression.
These studies reported on the accuracy of CCTA, using False-negative MPI results can also occur when obstructing
53  Barriers to Greater Clinical Implementation 685

stenoses are present in all three major coronary arteries, as can provide valuable prognostic information. Chow et  al.
this can lead to balanced ischemia. In addition, in patients [16] followed 2076 patients after CCTA for a mean of
with dominant left coronary arteries, left main coronary ste- 16 months for the occurrence of major adverse cardiac events
nosis can lead to uniform reduction of perfusion of the entire (MACE). The annual MACE rate was 0.13% in patients
left ventricle and a false-negative MPI exam. whose CCTA exam had demonstrated normal coronary arter-
By contrast, the negative predictive value of CCTA is ies, 0.52% in those with only nonobstructing plaques, 1.44%
close to 100% [2, 4, 10, 12]. This means it will be a better in those with single-vessel obstructive CAD, 4.3% in those
gatekeeper to the catheterization lab than MPI. If a patient with two-vessel obstructive CAD, and 9.79% in those with
with suspected CAD has an adequate CCTA with a negative three-vessel obstructive CAD.
result, the treating physician can be assured the patient does The results of the large CONFIRM registry, reported by
not need to undergo ICA. Min et  al. [17], provided more evidence of the prognostic
value of CCTA. They studied over 24,000 patients without
known CAD who underwent CCTA and were followed for
Direct Comparison of CCTA and MPI all-cause mortality for a mean of 2.3 years. Compared with
patients having normal coronary arteries by CCTA, those
A recent study by Arbab-Zadeh et  al. directly compared with nonobstructive disease had a higher risk of mortality
CCTA with MPI [13]. Patients in the study all underwent the (HR 1.62). Additional mortality risk was associated with
two noninvasive imaging tests and then proceeded to have single-vessel obstructive CAD (HR 2.00), two-vessel
ICA as the gold standard. Among 245 patients without a pre- obstructive CAD (HR 2.92), and three-vessel or left main
vious history of CAD, the per-patient results for CCTA and obstructive CAD (HR 3.70). Similar findings have been
MPI, respectively, were as follows: sensitivity 0.91 vs 0.55, demonstrated by others as well [18–20]. Taken together,
specificity 0.80 vs 0.70, PPV 0.81 vs 0.63, and NPV 0.91 vs these studies point to another great strength of CCTA – its
0.63. The authors concluded that CCTA was more accurate ability to stratify risk and enhance the decision-making pro-
than MPI for the diagnosis of CAD. Another important study cess of clinical cardiologists. The studies have shown that
was that of Moscariello et al. [14]. They evaluated 185 con- while patients with nonobstructive CAD by CCTA may not
secutive patients with chest pain and a positive MPI exam. need PCI or CABG, they are at higher risk than patients with
All patients also underwent CCTA and ICA. Of the 185, ICA completely normal coronary arteries. They need to be given
revealed that, despite the positive MPI, 110 of them (59%) aggressive medical therapy and, if necessary, counseled
had either normal coronary arteries or mild nonobstructive about lifestyle modification to prevent disease progression.
plaques. MPI obviously had a high rate of false positives. CCTA also noninvasively allows assessment of risk of one-,
But in comparison, on a per-patient basis, CCTA had a sen- two-, and three-vessel obstructive disease, so decisions can
sitivity of 100%, specificity of 93.6%, PPV of 91.5%, and be made about whether to send the patient for ICA and pos-
NPV of 100%. The better performance of CCTA, and par- sible percutaneous or surgical intervention.
ticularly its high NPV, suggests that it ought to be the first
imaging test performed in patients with suspected CAD, in
preference to MPI or SE.  he Advantages of CCTA as the First Imaging
T
Test in Patients with Acute Chest Pain
Presenting to the Emergency Department
The Value of CCTA as a Prognostic Examination
Approximately 8 million patients present to EDs each year
MPI is not able to accurately detect nonobstructing coronary complaining of acute chest pain [21, 22]. It is the second
artery plaques but CCTA can. Several CCTA studies have most common chief complaint in ED patients. As much as $8
compared Kaplan-Meier survival in three groups of patients – billion is spent on those patients for subsequent hospitaliza-
those with completely normal coronary arteries, those with tions and evaluations that turn out to be negative. Conversely,
nonobstructive CAD, and those with obstructive lesions 2–8% of these cases have acute coronary syndromes that are
(≥50% stenosis). Lin et  al. [15] found that among 2583 missed on the initial evaluations. This is obviously a signifi-
patients who underwent CCTA and were followed for over cant public health problem, and it is important to improve
3 years, the presence of any nonobstructive plaque was asso- efficiency and accuracy of the workup of these ED cases.
ciated with higher mortality (hazard ratio [HR] 1.98) com- Recent studies have shown that such improvement is possi-
pared with individuals with completely normal coronary ble through greater use of CCTA.
arteries. Mortality was progressively greater in patients with Goldstein et al. [23, 24] randomized 699 acute chest pain
nonobstructive CAD in 1, 2, and 3 vessels. Thus, even in patients to CCTA or MPI as part of their workup in the
patients with mild, low-risk atherosclerotic disease, CCTA ED. They found that in the CCTA group, time to diagnosis
686 D. C. Levin

was reduced by 54%, and cost of ED care was reduced by whereas CT is highly sensitive in doing so. Since these small
38%. In the patients whose index ED imaging test was nor- calcified plaques rarely narrow the lumen, their presence will
mal, there was no difference between the two groups in the often be missed on ICA when in fact they may be the early
6-month incidence of MACE. In another ED study, Litt et al. harbingers of nonobstructive CAD in what otherwise would
[24, 25] randomized 908 acute chest pain patients to CCTA appear to be “normal” coronary arteries. Thus ICA may
and another 462 to traditional care, in which the patient’s underestimate a patient’s atherosclerotic disease burden [28].
physicians determined which tests to order. Patients in the
CCTA cohort were more frequently discharged directly from
the ED (50% vs 23%), had shorter length of stay in the ED Conclusions
(18 h vs 25 h), and had a higher rate of detection of CAD (9.0
vs 3.5%). Of the patients undergoing CCTA, 640 had normal The cardiac imaging community is facing a peculiar para-
results, and none of them died or experienced an MI within dox. There is abundant evidence showing that CCTA is the
30 days. Hoffmann et al. [26] studied 1000 ED patients with most appropriate initial imaging test in patients with sus-
acute chest pain but no ECG or troponin abnormalities. The pected CAD. The evidence has been presented at length here
patients were randomized to early CCTA or standard evalua- and elsewhere in this textbook. Yet CCTA is far less com-
tion. Mean length of stay was shorter by 7.6 h in the CCTA monly utilized than older, less effective tests like MPI and
group. More patients were discharged directly from the ED SE (recall that in the Medicare population, MPI is used 53
(47% vs 12%). There was no significant difference between times as often as CCTA). Possible reasons for this have been
the two groups in incidence of MACE at 28 days. Poon et al. discussed earlier. Hopefully some of the obstacles prevent-
[27] randomized 894 ED patients each to CCTA or standard ing greater use of CCTA will be resolved as time goes on.
evaluation (serial ECGs and troponins). The CCTA cohort The imaging equipment will undoubtedly become faster,
had a 14% rate of admission to the hospital, compared with more efficient, and more automated, thus allowing radiologic
40% among the standard evaluation cohort. Mean length of technologists to do most of the work without requiring the
stay in the ED was 7.7 h among the CCTA group, compared presence of a physician. Advances in software and machine
with 11.5 h among the standard evaluation group. Symptoms learning are already occurring and will allow for computer-­
requiring return to the hospital within 30  days occurred in aided identification of exams that are normal. We must do a
1.3% in the CCTA group, compared with 3.6% in the stan- better job at educating our primary care colleagues and the
dard evaluation group. The latter group was more than seven nurse practitioners and physician assistants who work for
times as likely to subsequently undergo ICA without revas- them about the advantages accruing to early use of CCTA in
cularization. There were no cardiac deaths in either group their chest pain patients. Cardiologists and other cardiac
during 30 days of follow-up. Three acute MIs occurred in the imagers who have become overly accustomed to using MPI
CCTA group, compared with six in the standard evaluation (perhaps because they own the equipment) should be encour-
group. It seems quite apparent from these four studies of aged to change their approach. As Arbab-Zadeh has pointed
patients presenting to EDs with acute chest pain that early out [28], CCTA has come a long way and is ready for prime
CCTA is the preferred pathway to expeditious diagnosis and time. Cardiologists and radiologists must work together to
disposition. make it happen.

Advantages of CCTA References

Although ICA is still considered the gold standard in diag- 1. Go AS, Mozaffarian D, Roger VL, et  al. Executive summary:
heart disease and stroke statistics – 2014 update. A report from the
nosing CAD because of its higher spatial resolution, it is American Heart Association. Circulation. 2014;129:399–410.
important to recognize that CCTA offers some substantial 2. Budoff MJ, Dowe D, Jollis JG, et  al. Diagnostic performance
advantages. It is noninvasive, has less risk and cost, and can of 64-multidetector row coronary computed tomographic angi-
be done on an outpatient basis. Reconstructed CCTA images ography for evaluation of coronary artery stenosis in individu-
als without known coronary artery disease. J Am Coll Cardiol.
of the coronary arteries can be rotated and viewed from an 2008;52:1724–32.
infinite number of angles, whereas ICA images can only be 3. Miller JM, Rochitte CE, Dewey M, et  al. Diagnostic perfor-
viewed from the six or seven angles at which the image mance of coronary angiography by 64-row CT.  N Engl J Med.
intensifier was positioned during the contrast injections. On 2008;359:2324–36.
4. MeijboomWB MMFL, Schuijf JD, et  al. Diagnostic accuracy of
CCTA, the coronary arteries can be viewed in cross section, 64-slice computed tomography coronary angiography. J Am Coll
whereas that is not possible with ICA. Lastly, small calcified Cardiol. 2008;52:2135–44.
plaques are usually not visualized on ICA because the X-ray 5. Patel MR, Peterson ED, Dai D, et al. Low diagnostic yield of elec-
beam is not sensitive in detecting small calcific deposits, tive coronary angiography. N Engl J Med. 2010;362:886–95.
53  Barriers to Greater Clinical Implementation 687

6. Schuijf JD, Wijns W, Jukema JW, et al. Relationship between non- 17. Min JK, Dunning A, Lin FY, et al. Age- and sex-related differences
invasive coronary angiography with multi-slice computed tomog- in all-cause mortality risk based on coronary computed tomography
raphy and myocardial perfusion imaging. J Am Coll Cardiol. angiography findings. J Am Coll Cardiol. 2011;58:849–60.
2006;48:2508–14. 18. Russo V, Zavalloni A, Reggiani MLB, et al. Incremental prognostic
7. Weustink AC, Moller NR, Neefjes LA, et al. Diagnostic accuracy and value of coronary CT angiography in patients with suspected coro-
clinical utility of noninvasive testing for coronary artery disease. nary artery disease. Circulation Cardiovasc Imaging. 2010;3:351–9.
Ann Intern Med. 2010;152:630–9. 19. Van Werkhoven JM, Cademartiri F, Seitun S, et al. Diabetes: prog-
8. Mark DB, Berman DS, Budoff MJ, et al. ACCF/ACR/AHA/NASCI/ nostic value of CT coronary angiography – comparison with a non-
SAIP/SCAI/SCCT expert consensus document on coronary com- diabetic population. Radiology. 2010;256:83–92.
puted tomographic angiography. Circulation. 2010;121:2509–43. 20. Naisis A, Meridith IT, Sud PS, Cameron JD, Troupis JM, Seneviratne
9. Dorbala S, Hachamovitch R, Carli MF, et al. Myocardial perfusion SK. Long-term outcome after CT angiography in patients with pos-
imaging and multidetector computed tomographic coronary angi- sible acute coronary syndrome. Radiology. 2014;272:674–82.
ography. J Am Coll Cardiol. 2006;48:2515–7. 21. Galperin-Aizenberg M, Cooke TS, Hollander JE, Litt HI. Cardiac
10. Marwick TH, Cho I, Hartaigh B, Min JK. Finding the gatekeeper to CT angiography in the emergency department. AJR Am J
the cardiac catheterization laboratory: coronary CT angiography or Roentgenol. 2015;204:463–74.
stress testing? J Am Coll Cardiol. 2015;65:2747–56. 22. Levin DC, Parker L, Halpern EJ, Julsrud PR, Rao VM. The lack of
11. Dilsizian V, Taillefer R.  Journey in evolution of nuclear cardiol- growth in use of coronary CT angiography: is it being appropriately
ogy. Will there be another quantum leap with the F-18 labeled myo- used? AJR Am J Roentgenol. 2011;196:862–7.
cardial perfusion tracers? J Am Coll Cardiol Cardiovasc Imaging. 23. Goldstein JA, Chinnaiyan KM, Abidov A, et al. The CT-STAT (cor-
2012;5:1269–84. onary computed tomographic angiography for systematic triage
12. Chinnaiyan KM, Raff GL, Goraya T, et  al. Coronary computed of acute chest pain patients to treatment) trial. J Am Coll Cardiol.
tomography angiography after stress testing. Results from a multi- 2011;58:1414–22.
center statewide registry, ACIC (Advanced Cardiovascular Imaging 24. Levin DC, Parker L, Halpern EJ, Rao VM. Recent trends in imaging
Consortium). J Am Coll Cardiol. 2012;59:688–95. for suspected coronary artery disease: what is the best approach? J
13. Arbab-Zadeh A, Di Carli MF, Cerci R, et  al. Accuracy of com- Am Coll Radiol. 2016;13:381–6.
puted tomographic angiography and single-photon emission com- 25. Litt HI, Gatsonis C, Snyder B, et al. CT angiography for safe dis-
puted tomography-acquired myocardial perfusion imaging for the charge of patients with possible acute coronary syndromes. N Engl
diagnosis of coronary artery disease. Circ Cardiovasc Imaging. J Med. 2012;366:1393–403.
2015;8:e003533. 26. Hoffmann U, Truong QA, Schoenfeld DA, et  al. (ROMICAT-II)
14. Moscariello A, Vliegenthart R, Schoepf UJ, et  al. Coronary CT coronary CT angiography versus standard evaluation in acute chest
angiography versus conventional cardiac angiography for therapeu- pain. N Engl J Med. 2012;367:299–308.
tic decision making in patients with high likelihood of coronary 27. Poon M, Cortegiano M, Abramowicz AJ, et  al. Associations

artery disease. Radiology. 2012;265:385–92. between routine coronary computed tomographic angiography and
15. Lin FY, Shaw LJ, Dunning AM, et al. Mortality risk in symptom- reduced unnecessary hospital admissions, length of stay, recidivism
atic patients with nonobstructive coronary artery disease. J Am Coll rates, and invasive coronary angiography in the emergency depart-
Cardiol. 2011;58:510–9. ment triage of chest pain. J Am Coll Cardiol. 2013;62:543–52.
16. Chow BJW, Wells GA, Chen L, et  al. Prognostic value of
28. Arbab-Zadeh A. What will it take to retire invasive coronary angi-
64-slice cardiac computed tomography. J Am Coll Cardiol. ography? J Am Coll Cardiol Img. 2016;9:565–7.
2010;55:1017–102817.
 A Test on the Move: Cardiac CT in China
as a Case Study 54
Bin Lu, Weihua Yin, Xinshuang Ren, and Siyu Chen

Cardiovascular Diseases in China decreased from 3.89‰ in 1990 to 3.07‰ in 2013, but the
cardiovascular mortality number increased by 46% as a
With the rapid economic growth and increased life expec- result of the increase in the elder population.
tancy, China’s population is experiencing rapid transitions;
the leading health problems and the challenges imposed on
the health system by epidemiological and demographic fac- Cardiovascular Risk Factors in China
tors have also changed in recent years. An epidemic of car-
diovascular diseases (CVD) in China is emerging as a result Hypertension
of lifestyle changes, urbanization, and the accelerated pro- According to the 2015 Chinese residents’ nutrition and
cess of aging. The incidence of CVD is continuously increas- chronic disease status report, incidence of hypertension had
ing and will remain as an upward trend in the next decades. increased to 25.2% in residents older than 18 years old, total-
In 2014, cardiovascular disease was the leading cause of ing about 270 million people who suffered from hyperten-
death in China among both urban and rural areas. sion [4]. In 2010, 2043 patients died of hypertension in
Cardiovascular mortality was higher in rural areas than urban China, accounting for 24.6% of all-cause death. In the past
areas since 2009. Among rural areas in 2014, cardiovascular 50  years, four nationwide hypertension sampling surveys
mortality was 2.96‰, while cardiac mortality was 1.44‰, have been done in China; the incidence of hypertension
and cerebrovascular mortality was 1.52‰ (cerebral hemor- increased among the population of individuals over 15 years
rhage 0.75‰, cerebral infarction 0.45‰). Cardiovascular old, with rates of 5.1%, 7.7%, 13.6%, and 17.6% in 1959,
mortality in urban areas was 2.62‰, while cardiac mortality 1979, 1991, and 2002, respectively [5]. Although many stud-
was 1.36‰, and cerebrovascular mortality was 1.26‰ (cere- ies have suggested the control of hypertension is an effective
bral hemorrhage 0.52‰, cerebral infarction 0.42‰) [1, 2]. strategy to prevent cardiovascular events, systemic analysis
Figure  54.1 shows major disease mortality changes from of serial cross-sectional health survey data showed that only
1990 to 2014 in rural and urban areas throughout China [3]. a small percentage of hypertensive patients achieved the goal
The causes of mortality in different provinces of China were of blood pressure control. In 1991, the awareness rate, treat-
systemically analyzed from 1990 to 2013 [3]. Lower respira- ment rate, and control rate of hypertension in China were
tory infarction was the primary cause of death in 16 prov- only 22.4%, 12.0%, and 3.0%, respectively [6]. With the
inces, and another 15 provinces had cerebrovascular disease increase of government expenditure on health and improve-
as the main cause in 1990. Furthermore, in 2013, cerebrovas- ment of residential health awareness, these same rates
cular disease had been the primary cause of death in 27 prov- increased to 26.1%, 22.8%, and 6.1% in 2009 (Fig. 54.2) [6].
inces and ischemic heart disease in another 5 provinces. Prior studies showed anxiety, alcohol, salt intake, and
Chronic noninfectious diseases including ischemic heart dis- obesity were important risk factors of hypertension. A
ease, stroke, chronic obstructive pulmonary disease, and meta-­analysis, including eight prospective studies,
neoplasm have a major impact on the human lifespan. The showed the odds ratio of hypertension in people with anx-
age-adjusted mortality rate of cardiovascular disease iety compared to those without was 1.55 (95%CI 1.24–
1.94) [7]. Salt intake is higher in Chinese residents in
comparison to people from western countries. A study in
B. Lu (*) · W. Yin · X. Ren · S. Chen Shanghai including 19,519 residents, aged 35–91  years
Department of Radiologic Imaging, Fuwai Hospital, Chinese
Academy of Medical Sciences and Peking Union Medical College, old, showed the odds ratio to be 1.117 (95%CI 1.102–
National Center of Cardiovascular Diseases, Beijing, China 1.223) comparing hypertension in people who intake

© Humana Press 2019 689

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_54
690 B. Lu et al.

Fig. 54.1  Mortality rates of 350

CVD in urban and rural urban rural
Chinese populations (China:
300
1990–2013). (Courtesy of
National Center of
Cardiovascular Disease of 250
China.)
200

150

100

50

0
1990 1995 2000 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Fig. 54.2  Awareness rate, 30

treatment rate, and control awareness rate treatment rate control rate
rate of hypertension from
1991 to 2009. (Courtesy of 25
National Center of
Cardiovascular Disease of
China.) 20

15

10

0
1991 1993 1997 2000 2004 2006 2009

more than 12  g salt per day and a control group that 52.9%, while the rate was 2.4% in females [11, 12]. Smoking
intakes 6–12 g salt per day [8]. rates in male doctors and teachers were 40% and 36.5%,
It is important to emphasize secondary hypertension as respectively, which were higher than the rates in most other
well, considering more than 10% of hypertension results countries. A 2012 survey collecting information on a par-
from renal disease, endocrine dysfunction, and obstructive ticular population (18–59  years old) in China showed the
sleep apnea syndrome [9]. A study in the Xinjiang province smoking rate was 55.3% in males and 1.4% in females.
showed that 14.76% of hypertensive patients were diagnosed Males in the construction industry had the highest smoking
with secondary hypertension; the proportion increased to rate (58.6%), with the average amount smoked was 15.6
21.9% in young patients, and it decreased to 9.85% in elder cigarettes per day (15.7% in male, 10.3  in female) [13].
patients [10]. Renal hypertension was the most common type Smoking in adolescents is a serious problem in China. A
of secondary hypertension. meta-analysis summarized the smoking prevalence state in
adolescents aged 12–17 years old from 1981 to 2010. The
Smoking smoking rate among male students fluctuated between
China is known to have one of the highest smoking rates for 39.04% and 46.03%, while the smoking rate among female
adult men since 1984 among countries worldwide. In 2010, students, increasing rapidly in recent years, was between
a global survey covering 28 provinces in China showed that 2.47% and 19.44% [14]. Current smoking rate in 2010 was
the smoking rate of Chinese males over 15  years old was 17.4% in male students and 3.26% in female students.
54  A Test on the Move: Cardiac CT in China as a Case Study 691

Tendency of smoking at young age warrants attention. In 1.21  mmol/l in women, respectively [19]. Incidence of
2010, GATS survey showed the proportion of students who hyperlipidemia (TC >6.22 mmol/l) was highest in men aged
smoked a whole cigarette before 13 years old in relation to 45–59 years old and in women older than 60 years. The dis-
those who were currently smoking was 55.9% in males and tribution of hyperlipidemia was in accordance to the area.
57.0% in females; furthermore, 52.7% of current smokers Incidence of hyperlipidemia was higher in urban and eastern
aged 20–34 years old became daily smokers before 20 years areas compared to rural, middle, and western areas [19].
old. Given the high smoking rate in China, secondhand Dyslipidemia in teenagers is a major health problem in
smoke exposure (SHS) reached up to 74.1% in males and China. China’s 2002 Nutrition and Health Survey showed
71.6% in females in 2010. About 738 million people suf- that the incidence of hypercholesterolemia (TC >220 mg/dl
fered from SHS [15]. A prospective study in China showed or 5.72 mmol/l) in children and teenagers (3–18 years old)
smoking to be important risk factor of mortality in adults. was 0.8% (1.4% in urban areas, 0.6% in rural areas). The
The relative risk (RR) and population attribute risk (PAR) of incidence rate of hypertriglyceridemia among this popula-
all smokers were 1.23 (95%CI 1.18–1.27) and 7.9%. RR tion (TG >150  mg/dl or 1.70  mmol/l) was 2.8% (2.5% in
and PAR were 1.18 (95%CI 1.13–1.23) and 10.0% in male urban areas, 2.9% in rural areas) [20]. A meta-analysis
smokers and 1.27 (95%CI 1.19–1.34) and 3.5% in female including data from 2001 to 2011 showed incidence of
smokers [16]. Respiratory disease, cardiovascular disease, hyperlipidemia in teenagers increased with time. Obese teen-
and cancer were common causes of death in smokers; 24.1% agers were more likely to have hyperlipidemia, especially
of male smokers and 4.0% of female smokers died from car- those with family history of hyperlipidemia [21].
diovascular disease.
The smoking cessation rate increased from 9.42% in 1996 Diabetes
to 16.9% in 2010, and about 16.1% of smokers intended to China’s chronic disease monitor report, published in 2013,
quit in the next 12 months [15]. The smoking cessation rate reflects the prevalence of diabetes in recent years. Based on
for rural areas within Midwest China was very low (1.9%). 98,658 samples from 162 monitor centers in 31 provinces,
Though smoking cessation rate increased, there was no the incidence of diabetes in China was 9.7% [22]. Incidence
change in relapse rate, which was 32.5% in 2002 and 33.1% increased with age, and it was higher in urban areas than in
in 2010 [11, 15]. In order to reduce potential damage caused rural areas regardless of gender. Regarding people younger
by SHS, regulations on cigarette control took effect in than 60 years old, incidence was higher in males; however,
Beijing on June 1, 2015. Per these regulations, smoking was incidence was higher in females among people older than
banned in enclosed public locations, and the advertisement, 60 years.
sales promotion, and sponsorship of cigarettes were prohib- In order to investigate the effect of diabetes on mortality
ited; violators will be fined up to 10,000 yuan. in the Chinese population, 690 individuals with newly diag-
nosed diabetes and 519 nondiabetic individuals were enrolled
Hyperlipidemia and monitored for 23 years [23]. In total, 338 (56.5%) dia-
Multiple prospective cohort studies certified that serum total betic patients and 100 (20.3%) nondiabetic patients died dur-
cholesterol (TC) and low-density lipoprotein (LDL) are ing the follow-up. Cardiovascular disease was the major
important risk factors associated with cardiovascular events. cause of death for patients with diabetes; collectively, 47.5%
Total triglyceride (TG) was also shown to be predictive of of males and 49.7% of females died of cardiovascular dis-
cardiovascular disease. ease, with nearly half of them passing due to stroke (52.3%
According to several international studies, total choles- of males, 42.3% of females). Incidence of all-cause mortality
terol and low-density lipoprotein (LDL) are lower in the per year in patients with diabetes was 3 times that of the
Chinese population compared to those in western countries. incidence in nondiabetic individuals (male, 36.9‰ vs
In 2001, a cross-sectional survey of a nationally representa- 13.3‰; female, 27.1‰ vs 9.2‰). Diabetic females had
tive sample of 15,540 Chinese adults, ages 35–74, showed higher risk of cardiovascular mortality than diabetic males
that the prevalence of hypercholesterolemia was relatively (HR 6.9 vs 3.5).
high (24.8%), especially among citizens in developed areas, The Daqing Diabetes Prevention Study was the first and
as well as in middle-aged and elderly individuals [17, 18]. longest-running lifestyle intervention research study for dia-
The percentage of adults with controlled blood cholesterol betes worldwide. Individuals in the experimental group were
was low in China; only 3.5% of men and 1.5% of women under positive lifestyle intervention for 6  years. After
were under control [17]. Chronic disease monitoring in 20  years of follow-up, the results showed that lifestyle
China reported the average levels of serum total cholesterol ­intervention reduces the risk of diabetes by 51%. The cumu-
(TC) and total glyceride (TG) in people older than the age of lative incidence rate of diabetes in people with lifestyle inter-
18 from 31 provinces. The average level of TC and TG was vention was 80% and a rate of 93% in the control group.
4.06 mmol/l and 1.45 mmol/l in men and 4.03 mmol/l and Adjusted to age and grouping random factors, incidence of
692 B. Lu et al.

diabetes in the experimental group was still lower than that showed individuals with central obesity (WHtR >0.50) had
in control group, and the odds ratio for lifestyle intervention increased risk of cardiovascular disease.
was 0.57 (95%CI 0.41–0.81). The onset of diabetes in the
experimental group was 3.6 years later than that in control  ack of Physical Activity
L
group [24]. In addition, lifestyle intervention decreases the A cross-sectional study including 460,000 adults, aged
risk of severe retinopathy (laser treatment or blindness) by 35–74  years old, showed that the average physical activity
47% [25]. For patients with impaired glucose tolerance in the level was 21.7 MET-h/d, mainly attributed to occupational
Daqing Diabetes Prevention Study, lifestyle intervention activity (62%) and housework (26%) [33]. China’s Nutrition
decreased the all-cause mortality by 29% and cardiovascular and Health Survey reported citizen physical activity
mortality by 41% [26]. decreased from 1991 to 2011, especially occupational activ-
The 3B study approach investigated risk factors for car- ity. Male occupational activity decreased from 382 MET-h/w
diovascular disease among type 2 diabetics [27]. In total, to 264 MET-h/w in 2011, while female occupational activity
25,817 diabetic patients from 104 hospitals were included. decreased from 420 MET-h/w to 243 MET-h/w [34]. The
The average age was 62.6 years old, and 47% of the enrolled percentage of exercise remains low with less than 7 MET-­-
patients were male. In all, 72% of these patients had hyper- h/w in males and less than 7 MET-h/w in females. A 2011
tension or hyperlipidemia. Incidence of cardiovascular survey of students aged 11–18 years old showed that the rate
events in diabetic patients with combined hypertension or of reaching standard physical activity was only 19.9%,
hyperlipidemia was six times that of patients with isolated 25.1% in boys and 14.6% in girls [35].
diabetes. Independent predictors of cardiovascular events Results from an additional study showed that physical
included the following: low BMI (<24 kg/m2), no smoking or activity negatively relates to incidence of diabetes. This
drinking, high-level education, and course of, diabetes for Chinese multicenter study showed morbidity of diabetes
less than 5 years. decreased with increased physical activity. Mortality risk of
ischemic heart disease (HR 1.20–1.71), ischemic stroke (HR
Overweight and Obesity 1.05–1.53), and diabetes (HR1.11–1.45) was higher in peo-
Incidence of individuals who are overweight (BMI 24.0– ple with lack of physical activity than those with sufficient
27.9 kg/m2) or obese (BMI >28.0 kg/m2) increased in recent physical activity [36].
years with changes in lifestyle and development of economy.
China’s 2002 Nutrition and Health Survey showed the over- Unreasonable Diet
weight rate was 17.6%, and the obesity rate was 5.6% [28]. Dietary structure in China has changed in recent years [37].
Nutrition and health status surveillance of nine provinces Characteristics of the new dietary structure that are against
showed overweight and obesity rates increased rapidly over the prevention of cardiovascular disease include the follow-
the past 20 years, reaching 44% in 2011 [29]. The Chronic ing: decrease in cereal intake, obvious increase in fat intake,
Disease Monitor Program in China reports that in 2010, the low intake of vegetables and fruits, and high intake of salt
overweight rate, obesity rate, and central obesity rate were (15.9 g/d). The Chinese Nutrition and Health Survey showed
30.6%, 12.0%, and 40.7%, respectively. These rates were that the percentage of people getting more than 30% energy
much higher than the rates reported in 2002 [30]. Sample from fat increased from 35.8% in 1989 to 55.0% in 2009.
surveys in the Guangdong province showed central obesity Sodium intake decreased in recent years but remains at a
rate increased from 12.9% in 2002 to 23.7% in 2010, which high level (4.7  g/d). Potassium intake increased in recent
implied a change of obesity type [31]. The prevalence of years, yet it is still below recommended dose (2  g/d). A
obesity in adolescents was not optimistic. Data from the nationwide disease survey of 970,000 people showed that
National Survey on Chinese Students’ Physical Fitness and more than 60% of citizens are aware of the negative effects
Health between 1985 and 2010 showed an obvious increase of salt on health and 40% of citizens had taken measures to
of overweight and obesity rate in students aged 7–18 years decrease salt intake.
old. The overweight and obesity rates in 2010 were 8.7 times A prospective research study conducted in Shanghai
and 38.1 times those in 1985 (overweight, 9.6% vs 1.1%; showed that compliance score to diet recommendation nega-
obesity, 5.0% vs 0.1%) [32]. tively relates to all-cause mortality and cardiovascular mor-
Obesity, especially central obesity, was an important car- tality [38]. The amount of vegetable and fruits consumed in
diovascular risk factor. A prospective study including the diet negatively relates to coronary artery disease.
270,000 participants, aged 35–74 years old, showed positive Cardiovascular risk decreased by 38% in the females with a
correlation between waistline and new-onset diabetes. high intake of vegetables and fruits (medium 814  g/d)
Waistline/height ratio (WHtR) is a simple indicator evaluat- ­compared with those with a low intake (medium 274  g/d)
ing the extent of central obesity. Several Chinese studies [39]. There was no significant difference found among male
citizens. In addition, the amount of salt intake positively
54  A Test on the Move: Cardiac CT in China as a Case Study 693

relates with onset hypertension; however, potassium intake concentration of PM2.5 in people older than 65 was stronger
negatively relates with this condition. A meta-analysis than those younger than 65.
showed that low-sodium intake can remarkably decrease Studies showed long-term exposure to outdoor air pollu-
systolic blood pressure by 4.9  mmHg and diastolic blood tion had a greater impact on cardiovascular disease. In a
pressure by 1.5 mmHg [40]. cohort study, researchers retrieved air condition data from an
air condition monitoring station and analyzed the relation-
Air Pollution ship between particulate matter exposure and cardiovascular
A great number of studies show that particulate matter (PM) mortality according to the results of a 10-year (1991–2000)
in the air, especially PM2.5, presents as an independent risk follow-up [50]. Baseline PM exposure dose was correlated
factor of cardiovascular disease [41]. Recent studies evalu- with cardiovascular mortality: every 10 ug/m3 increase in
ated the relationship between exposure duration and onset of concentration of total PM, SO2, and NOx leads to an increase
cardiovascular disease, as well as cardiovascular mortality. of cardiovascular mortality risk by 0.9% (95%CI 0.3–1.5%),
Particular matters such as PM2.5, SO2, and NOx prove to 3.2% (95%CI 2.3–4.0%), and 2.3% (95%CI 0.6–4.1%),
relate with cardiovascular disease. The following chart sum- respectively. A cohort study in Shenyang including 24,845
marizes studies that evaluated the effect of PM2.5 people showed that every 20 ug/m3 increase in SO2 leads to a
(Table 54.1) [42–48]. From 2010 to 2012, a study was con- 19% (OR 1.19, 95%CI 1.05–1.34) increase in hypertension
ducted in Beijing to monitor air condition and showed that among males. No correlation was found in female partici-
ischemic heart disease onset and mortality increased by pants. Moreover, a cohort study conducted in Hong Kong
0.27% (95%CI 0.21–0.33%) and 0.25% (95%CI 0.10– evaluated the relationship between PM exposure and cardio-
0.40%), respectively, with every 10 ug/m3 increase in PM2.5. vascular disease mortality in older participants (>65  years
PM2.5 had a hysteresis effect on the onset of ischemic heart old) during 10–13-year follow-up. The results showed that
disease as the onset rate increased at the first, second, and cardiovascular mortality increased by 22% (HR 1.22, 95%CI
third day after exposure of PM2.5. Also, the results showed 1.08–1.39), ischemic heart disease mortality by 42% (HR
that people older than 65 years were more sensitive to PM2.5; 1.42, 95%CI 1.16–1.73), and cerebrovascular mortality by
the correlation between onset ischemic heart disease and the 24% (HR 1.24, 95%CI 1.00–1.53) with every 10 ug/m3

Table 54.1  Relationship between PM2.5 concentrations and CVD mortality

PM2.5
Monitoring concentrations
time Region (mean) Outcome indicator/No. CVD risk (95% CI)
2004–2005 Shanghai [42] 56.4 μg/m3 Total mortality/79,530 Every 10 μg/m3 increase, mortality of CVD increases
0.41% (0.01–0.82%)
2004–2008 Xian [43] 176.7 μg/m3 —— Every 10 μg/m3 increase, mortality of CVD increases
0.27% (0.08–0.46%), mortality of coronary heart
diseases increases 0.39% (0.14%, 0.65%)
2004–2008 Xian [44] Median Mortality of CVD/22,051 Every 100 μg/m3 increases, mortality of CVD, coronary
2004 heart diseases, and stroke increases 6.18%, 8.23%, and
184.9 μg/m3 5.13%, respectively
2005
194.9 μg/m3
2006
206.8 μg/m3
2007
193.7 μg/m3
2008
179.1 μg/m3
2006–2008 Shenyang [45] 75 μg/m3 Total mortality/60,938 Every 10 μg/m3 increases, mortality of CVD increases
0.53% (0.09–0.97%);
mortality of CVD in population aged ≥75 years old
increases 0.64% (0.02–1.24%)
2007–2008 Guangzhou [46] 70.1 μg/m3 Total mortality/58,400 Every 10 μg/m3 increases, mortality of CVD increases
1.22% (0.63–1.68%)
2006–2011 Shanghai [47] 55 μg/m3 Mortality of out-of-hospital Every 10 μg/m3 increases, mortality of out-of-hospital
coronary heart diseases/18,202 coronary heart diseases increases 0.68% (0.14–1.21%)
2010–2012 Beijing [48] 96.2 μg/m3 Mortality of ischemic heart Every 10 μg/m3 increases, mortality of ischemic heart
diseases/53,247 diseases increases 0.25% (0.10–0.40%)
From National Health and Family Planning Commission of the People’s Republic of China [49], with permission
694 B. Lu et al.

increase in PM2.5. A large amount of epidemiologic evi-  oronary Artery Disease

C
dence proves the correlation between air pollution and car- According to the nationwide health survey, incidence of
diovascular disease; but, as most of the studies were ischemic heart disease in 2008 was 15.9‰ in urban areas and
observational, more experimental studies were needed for 4.8‰ in rural areas, which was much higher than incidence
further evaluation. rates in 2003 (12.4‰ in urban areas, 4.6‰ in rural areas)
[52].
Coronary artery mortality was 1.08‰ in urban areas and
Cardiovascular Disease 1.05‰ in rural areas for the year 2014; furthermore, mortal-
ity rate in men was higher than that in women. Coronary
Cerebrovascular Disease artery disease mortality rate increased in recent years.
The Department of Health in China has organized a nation- Mortality rate in rural areas increased rapidly since 2012 and
wide health survey every 5 years since 1993. According to closely approached the rate in urban areas in 2014 (Fig. 54.4).
the reports, incidence of cerebrovascular disease in China Rate of acute myocardial infarction had an upward trend
increased in recent years, and the incidence in urban areas from 2002 to 2014. The mortality rate rose fast since 2005,
was higher than that in rural areas. China Health Statistics and mortality rate in rural areas went over the rate in urban
Yearbook 2015 reported cerebrovascular mortality in urban areas since 2007 (Fig. 54.5) [52].
areas was 1.26‰, cerebral hemorrhage mortality was 0.52‰,
and cerebral infarction mortality was 0.42‰ for the year Heart Failure
2014. For that same year, cerebrovascular mortality in rural In 2000, the prevalence of chronic heart failure among the
areas was 1.52‰, while cerebral hemorrhage mortality was Chinese population aged 35–74  years was 0.9%: 0.7% in
0.75‰, and cerebral infarction mortality was 0.45‰. The men and 1.0% in women. Prevalence was higher in the north
related figures suggest a higher mortality rate in rural areas than the south and higher in urban areas. The prevalence of
than in urban areas. Cerebrovascular mortality rate increased heart failure increased significantly with age. During the past
with age, and the rate was higher in men than in women after two to three decades, the main cause for heart failure has
adjusted to age, regardless of the region (Fig. 54.3). shifted from rheumatic valvular heart disease to coronary
China’s 2010 Chronic Disease and Risk Factor heart disease [53].
Surveillance investigated 98,658 adults to figure out the inci- The preliminary results from the China Heart Failure
dence of transient ischemic attack adjusted to age was 2.27%. Registry Study (China-HF) showed that currently the aver-
Transient ischemic attack (TIA) was common in older female age age of patients with heart failure is 66 ± 15 years old
with low education, smoking history, hypertension, myocar- and has been rising. From this population, 54.5% of HF
dial infarction, or diabetes. Only 3.08% of adults were aware patients are male, and 84.7% of the patients were at III–IV
of the disease, and only 4.07% of TIA patients received rec- levels according to the NYHA functional classification cri-
ommended therapy [51]. teria. The main comorbidities with heart failure have
changed significantly: the proportion with valvular diseases

Fig. 54.3  Trends in stroke 160

mortality (China: 2003–
2014). (Courtesy of National 140
Center of Cardiovascular
Disease of China.)
120

100

80

60

40

20
urban rural
0
2003 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
54  A Test on the Move: Cardiac CT in China as a Case Study 695

mortality,
(1/100, 000)
urban rural
120

100.86 107.5
105.37
100 93.17
94.96 95.97 98.68
91.41
86.34
80
75.72
71.27 68.62
69.24
64.67
60
57.1
51.89
46.1 46.27 45.29
42.1
40 39.56
33.74
27.57 28.68
24.89
22.2
20

0 year
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Fig. 54.4  Coronary heart disease mortality trends in urban and rural areas between 2002 and 2014. (Courtesy of National Center of Cardiovascular
Disease of China.)

mortality,
urban rural
(1/100, 000)
80

70 68.6
66.62

60
55.32
50 48.53 51.46
47.36
44.95
43.14 43.19 42.5
40 39.72 39.32
38.09
34.12
30
28.1
25.53
23.2 23.16 23.42
20 21.5
17.36 18.4
16.46 15.77
12 11.3
10

0 year
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Fig. 54.5  Acute myocardial infarction mortality trends in urban and rural areas between 2002 and 2014. (From National Health and Family
Planning Commission of the People’s Republic of China [49], with permission.)
696 B. Lu et al.

has gradually decreased; coronary artery disease (49.4%), lipid-lowering drugs, calcium channel blockers, and angio-
hypertension (54.6%), and chronic kidney disease (29.7%) tensin receptor blockers.
have become the most common comorbidities. Infection
(45.9%) continues to be the primary reason for the onset of
heart failure, followed by fatigue or stress (26.0%) and then  urrent Practice of Cardiac Imaging (CT)
C
myocardial ischemia (23.1%). The usage of diuretics in in China
patients during hospitalization has not changed signifi-
cantly. The usage of digoxin showed a downward trend due High-Technology Medical Equipment
to the effects of international clinical trials. The use of
angiotensin II receptor antagonist (24.6%), aldosterone Many researchers have reported on inequities in health sta-
receptor antagonist (55.4%), and beta-receptor blocker tus, healthcare, health insurance, and healthcare workforce
(50.6%) has increased ­significantly [54]. in China; yet, only a small number of studies have focused
on equity in distribution of high-technology medical equip-
 edical Care Expenditure [55–57]
M ment, such as computed tomography (CT). He et al. exam-
Since 1980, the amount of inpatients increased and amplified ined the issue using panel data from multiple provinces in
rapidly after 2000. The medical cost of cardiovascular dis- 2006 and 2009, providing information not only about the
ease increased correspondingly. The growth rate of medical numbers of CTs and magnetic resonance imaging (MRI) but
costs was even higher than the rate of GDP, which resulted also detailed characteristics about these types of equipment
from the increasing need for inpatient services and high pro- such as machine model, price, and government subsidies for
portion of unreasonable medicine use. purchasing the machine. To inform China policymaking on
The number of discharged patients diagnosed with cardio- allocation of CTs and MRIs, He et al. compare the growth of
vascular disease was 17,938,600  in 2014, accounting for the high-technology equipment between the study provinces
12.75% of total inpatients. From this inpatient total, 6.63% and a group of selected industrialized countries [58].
were patients with heart disease, and 6.12% were patients with Examining equity in distribution of CTs and MRIs in China
cerebrovascular disease. Ischemic heart disease (6,553,700 is particularly important for several reasons. First, unneces-
per year), acute myocardial infarction (541,400 per year), and sary allocation and overuse of these two types of high-technol-
cerebral infarction (5,319,900 per year) were the main diagno- ogy medical equipment have been a known part of the medical
ses among the patients; the others were hypertension arms race in industrialized countries. Given the rapid eco-
(2,526,100 per year), intracranial hemorrhage (1,299,900 per nomic growth in China, with the world’s second largest econ-
year), and rheumatic heart disease (250,100 per year). omy, some experts were concerned that a similar phenomenon
The growth rate of inpatients diagnosed with cardiovas- of medical race might occur in China, although they have not
cular disease was 10.10% per year, faster than average conducted empirical analyses to quantify their concerns.
inpatient growth rate (6.33%). The growth rate rank in Second, availability of healthcare resources, including CT and
patients with cardiovascular disease included the follow- other types of high-technology medical equipment, is impor-
ing: cerebral infarction (12.3%), ischemic heart disease tant in defining the healthcare system’s performance. One
(11.74%), intracranial hemorrhage (9.76%), acute myocar- analysis of all 30 Organization for Economic Cooperation and
dial infarction (8.12%), hypertension (8.06%), hyperten- Development (OECD) countries concluded that distributions
sive cardiomyopathy and nephropathy (5.82%), and of CTs positively correlate with total health expenditure per
rheumatic heart disease (1.43%). capita and economic incentives to hospitals [59]. Researchers
Hospitalization costs for acute myocardial infarction in from developing countries have also paid attention to CT dis-
2014 totaled $1.34 billion, while intracranial hemorrhage seminations. Third, there is a huge gap in socioeconomic
was $2.08 billion and cerebral infarction $4.70 billion. The development among eastern, middle, and western regions of
growth rates of these conditions were 32.02%, 18.90%, and China. Because the eastern region has more healthcare
24.96% per year, respectively, since 2004. Single hospital- resources, an arms race among hospitals may exacerbate the
ization cost of acute myocardial infarction, intracranial hem- equity problem in China’s allocation of healthcare resources
orrhage, and cerebral infarction was 24,706 yuan, 15,929.7 across regions. Fourth, one study reported that the number of
yuan, and 8,841.4 yuan, respectively. The growth rates were CTs in China increased 55.4% and the number of MRIs
8.72%, 6.63%, and 2.81% per year. increased 90.2% from 2002 to 2005 [60]. While the newer ver-
Drug expenses for cardiovascular disease in hospitals sions of the equipment quickly provide higher-quality images
with more than 100 beds were 65.6 billion yuan. The top five at greater prices, they also raise issues of equity regarding
expenses were drugs improving cardiovascular circulation, which hospitals ­purchased them and whether the government
myocardial nutrition and drugs improving coronary flow, subsidized the purchase [61].
54  A Test on the Move: Cardiac CT in China as a Case Study 697

Finally, in response to the heavy investment in high-­ Table 54.3  Number of CTs in the study sites and the selected OECD
technology medical equipment by hospitals, China’s central countries
government implemented a policy in 2005 called the Chinese Study sites in China 2009 Change from 2006 to 2009 (%)
Certificate of Need (CON), which aims to improve “appro- Shanghai 7.6 62.7
priate allocation and efficient use of large medical equip- Zhejiang 8.3 31.0
ment” through regional health planning and quota control. Shanxi 8.5 63.8
Hunan 6.0 42.2
Nevertheless, it remains an unsettled question whether
OECD countries
implementation of the policy results in reasonable growth
Australia 38.8 −30.7
and more equitable distributions of CT and MRI. Austria 29.9 1.0
As shown in Table 54.2, there are substantial geographic, Canada 13.9 15.8
demographic, and economic differences among the study Denmark 23.7 50.0
sites analyzed by He et  al. [58]. Shanghai has the highest Finland 20.5 38.5
population density and lowest rural population percentage, Greece 33.9 28.4
and it was ranked first in GDP per capita among the four cit- Ireland 15.8 23.4
ies in 2009. Israel 7.7 42.6
Table 54.3 shows CT distributions in the selected OECD Iceland 34.5 42.6
Japan 97.3 5.1
countries. Among the selected countries, Japan had the high-
New Zealand 14.6 18.7
est number of CTs in 2009 (97.3%), while the UK had the
R.O. Korea 37.1 10.1
lowest level (7.4%). In comparison, the study sample in UK 7.4 −1.3
China is lower than that of the UK. USA 34.3 6.5
Most of the selected OECD countries had a growing num- Entire sample 41.8 1.04
ber of CTs between 2006 and 2009 except Australia and the
UK, both of which had negative growth of CTs. In compari-
son, all the study sites in China had a positive growth rate of Despite the improvement in equity status of CTs in some
CTs from 2006 to 2009, and the growth rate was higher than of the study sites, He et al. found that the distributions of CTs
most of the selected OECD countries. positively correlated with economic development level
He et al. found that China had lower numbers of CTs per across all cities in the four study sites in either 2006 or 2009.
million people in 2009 than most of the selected OECD He et  al. also found that the vast majority of CTs were
countries, despite the increases in its CT numbers from 2006 imported and only a small percentage were made by domes-
to 2009 being higher than most of the OECD countries. tic Chinese firms, indicating that China is still at a low level
These findings suggest that China is the largest developing of high-technology medical equipment development [58].
country but still lags behind the developed countries in high-­
technology medical equipment, such as CTs. On the other
hand, with the rapid economic development and large popu- Cardiac CT Imaging
lation, the number of CTs in China increased substantially in
these years [58]. Cardiac CT and its main application, coronary CT angiogra-
phy (CTA), remain a challenge because of rapid cardiac
motion in combination with small dimensions and complex
anatomy. It requires a dedicated CT system, ECG-­
Table 54.2  Description of the study sites in 2009 synchronized data acquisition or image reconstruction, as
Study site Shanghai Zhejiang Shanxi Hunan well as advanced post-processing and interpretation.
Location Eastern Eastern Northwest Central Nationwide surveys describing cardiac CT utilization have
China China China China
been performed in the USA and Germany, but there is a lack
Number of cities / 14 13 17
within site of data from China. While the use of coronary CTA has
Area (10,000 km2) 0.63 10.18 20.58 21.18 increased sharply, there are inadequate data describing the
Population 19 47 38 69 number, indications, and acquisition techniques of coronary
Population density 3049 463 183 326 CTA examinations. Also, there may be significant regional
(person/km2) imbalance regarding utilization of cardiac CT, data acquisi-
Percentage of rural 11.4 42.1 56.5 56.8 tion, contrast injection protocols, radiation exposure, and
population (%)
GDP per capita 78,326 48,196 21,659 19,479
diagnostic analytic skills across China.
(CNY) Fuwai Hospital launched a national survey to investigate
Rank of GDP per 1 5 16 20 the utilization of cardiac CT in China and to develop a
capita national coronary CTA information database, allowing
698 B. Lu et al.

c­omparisons between our findings and those from other ease in asymptomatic individuals in 95% (145/152); exclu-
countries. A 25-item questionnaire was designed. Initial sion of coronary anomalies in 64% (98/152); evaluation for
questions were used for the following: to identify a contact pericardial disease in 39% (59/152); follow-up of patients
person at each hospital; to confirm the availability of at least with coronary stents in 28% (43/152); diagnosis of coronary
one advanced, dedicated CT scanner; and to confirm that the artery disease in patients with a high pretest likelihood in
volume of coronary CTA procedures exceeded 100 in the last 21% (32/152); and triple rule-out in 18% (28/152). Maurer
12  months. Only hospitals fulfilling these criteria were listed coronary CTA indications in Germany and other coun-
included in the subsequent analysis. This survey was con- tries, which were different from our findings (Table 54.4).
ducted from December 2014 to February 2015. The remain- Beta-blockers were routinely used by 30.3% (46/152) of
ing questions pertained to the type of facility, the indications hospitals regardless of the heart rates before coronary CTA
for cardiac CT, patient preparation, and technical aspects of and administered based on heart rate limits in the remaining
the acquisition protocols. Additional questions referred to sites. The respective thresholds were 65–74 beats/minute at
storage and interpretation of cardiac CT data sets and future 57 of 152 sites (38%), 75–84 beats/minute at 34 (22%) sites,
needs for cardiac CT. and 85–100 beats/minute at 15 (10%) hospitals. Routine
Hospitals in China are organized according to a three-tier administration of sublingual nitroglycerin prior to cardiac
system. They are designated as primary, secondary, or ter- CT was reported by 131/152 respondents (86%). In all hos-
tiary institutions that are recognized by the hospital’s ability pitals (100%), patient breath-hold capacity was tested before
to provide different levels of medical care and medical edu- the scan. The majority (96/152, 63%) of respondents stated
cation and conduct medical research. Based on the level of that they did not perform cardiac CT in patients with atrial
service, size, medical technology, and available medical fibrillation, 35 (23%) did so occasionally and depending on
equipment, the three tiers are further subdivided into three circumstances, and 21 (14%) stated they did routinely.
subsidiary levels A, B, and C, resulting in a total of nine lev- Tube current was adjusted to the patient’s body weight or
els. Our survey focused on tertiary hospitals, because pri- body mass index by 90/152 sites (59%), and 123 sites (81%)
mary and secondary hospitals in China typically do not used lead aprons to shield the thyroid gland and/or genitals.
perform cardiac CT. The final minimal sample size was 240, Prospectively ECG-triggered image acquisition was used in
to include a buffer of 5%. Finally, after the removal of hospi- more than 90% of cases by 20/152 hospitals (13%), between
tals that refused participation, there were 152 hospitals that 50% and 89% of all cases by 47 hospitals (31%), between
participated in this study with at least 1 dedicated CT (64
rows or above) performing more than 100 CTA in the last
year (Fig.  54.6). The median annual number of coronary
Table 54.4 Coronary CTA indications in China and in other
CTA procedures performed at these institutions was 1037 countries
(range, 150–8072). Frequency% Frequency% in
Common clinical indications for coronary CTA included Frequency% in in Germany worldwide
the following: exclusion of coronary artery disease in patients Indications China [62] (Maurer [63]) countries)
with low to intermediate likelihood of disease in 100% of Exclusion of 100 (152/152) 91 (41/45) 97 (164/169)
centers (152/152); screening to rule out coronary artery dis- CAD in patients
with low-­
intermediate
1,442 pretest likelihood
hospitals • Minimal Exclusion of 95 (145/152) 20 (9/45) 34 (58/169)
• All hospitals
from the • Level A&B sample size CAD in healthy
NHFPC list tertiary calculation people
hospitals with buffer Follow-up of 66 (101/152) 78 (35/45) 76 (128/169)
24, 709 (LABEL) 240 coronary bypass
hospitals candidates Exclusion of 64 (98/152) 80 (36/45) 92 (156/169)
coronary anomaly
Evaluation of 39 (59/152) 4 (2/45) Not applicable
168 pericardial disease
• performing hospitals • Removal of Follow-up of 28 (43/152) 36 (16/45) 33 (56/169)
≥100 cCTA • with at least hospitals patients’ coronary
in the last 1 dedicated that refused stents
year CT (≥64 participation Diagnosis of 21 (32/152) 22 (10/45) 33 (55/169)
152 rows) 237 CAD in patients
hospitals respondents with a high pretest
likelihood
Fig. 54.6  A total of 152 tertiary hospitals in China were randomized Triple rule-out 18 (28/152) 53 (24/45) 44 (75/169)
enrolled for the nationwide survey of cardiac CT imaging
54  A Test on the Move: Cardiac CT in China as a Case Study 699

11% and 49% by 70 hospitals (46.0%), and in 10% or less by Only 46/152 centers (30%) reported that they also evaluated
15 sites (10%). In addition, 57 hospitals (38%) had experi- pulmonary and other non-cardiac structures. In most hospi-
ence with the use of novel coronary CTA acquisition tech- tals (123/152, 81%), coronary artery calcium scores
niques, including single-shot snap, high-pitch, or prospective (Agatston scores) were not calculated. Instead, a qualitative
acquisition. Iterative reconstruction algorithms were used by degree of calcification (i.e., mild, moderate, and severe) was
125 hospitals (82%). In our survey, 24 of 152 hospitals (16%) reported.
identified >15  mSv as the modal value of radiation dose, To our knowledge, this is the first nationwide survey
64/152 (42%) stated 10–15 mSv, 48 (32%) cited a range of describing cardiac CT, and especially coronary CTA utiliza-
6–9 mSv, 16 hospitals (11%) quoted 3–5 mSv, and no hospi- tion, in mainland China. We used random sampling of hospi-
tals stated <3 mSv (Fig. 54.7). The median dose among all tals in all Chinese provinces in order to reflect geographic
hospitals was 10–15 mSv, but after adjustment by coronary differences in population and economic status. Routine com-
CTA volume, the median dose among all patients was ponents of cardiac CT including scanner technology, clinical
between 6 and 9 mSv. indications, image acquisition techniques, contrast media
Regarding contrast media suppliers, we investigated the injections protocols, image reconstruction and interpreta-
market shares of domestic and foreign vendors and found tion, and clinical challenges are described. This study pro-
that 43/152 hospitals (28%) used foreign products for almost vides a comprehensive picture of current cardiac CT
all patients (≥90%), 44 (29%) used foreign products in most utilization in China.
of their patients (50–89%), 32 hospitals (21%) used foreign It was initiated and conducted by the Chinese International
contrast agents in less than half of their studies (11–49%), Regional Committee of the Society of Cardiovascular
and 33 hospitals (22%) used foreign products in few or no Computed Tomography (SCCT IRC China). In order to limit
cases (≤10%). Only 24/152 sites (16%) adjusted the contrast expenses, questionnaires were collected by either the market-
injection rate based on patient body weight or body mass ing team of GE Healthcare China or the authors themselves.
index. Over 97% used biphasic or triphasic injection proto- However, random verification of questionnaires provided by
cols for contrast injection (biphasic, 74%; triphasic, 23%; GE addressed concerns regarding commercial bias.
and monophasic, 3%). Together with other professional organizations, SCCT
The vast majority of respondents (145/152; 95%) stored released appropriate use criteria for cardiac CT in 2010 and
image data in PACS, and the remaining 7 sites stored them two guidelines for the interpretation and reporting of coro-
on compact disks. The average duration of a coronary CTA nary CTA in 2009 and 2014 [62]. The authors of this report,
examination was estimated to be 14.2 min (median 15 min; as members of the cardiothoracic group of Chinese Society
range 5–30  min, Fig.  54.8), the average estimated post-­ of Radiology, published a Chinese coronary CTA guideline
processing time was 13.6  min (median 15  min; range in January 2011  in the form of an expert consensus docu-
5–25 min), and the estimated time required for interpretation ment. These documents played a significant role in promot-
and reporting was 18.0 min (median 20 min; range 8–30 min). ing the development of cardiac CT in China. Nevertheless,

Estimated Average Duration

Estimated Radiation Dose of CCTA Acquistion
among Hospitals
min 30
>15mSv
25

10-15mSv 20

15
6-9mSv
10
3-5mSv
5

<3mSv 0
0 19 38 57 76 95 114 133 152 0 19 38 57 76 95 114 133 152
Order of CCTA amount Order of CCTA amount

Fig. 54.7  Estimated radiation dose among hospitals were 6–9 mSv in Fig. 54.8  The average duration of a coronary CTA examination was
median estimated to be 15 min (range 5–30 min)
700 B. Lu et al.

the survey results suggest that the utilization of cardiac CT in 85 min. Many factors may cause these inequalities, including
mainland China is still limited by several factors. First, high-­ whether a site was a teaching versus non-teaching hospital,
end cardiac CT scanners are not uniformly available. the local economic level, and the interest of hospital leaders
Although the hospitals targeted by our survey were of the and radiology directors. We believe that the best approach to
highest level in China, one third of candidate hospitals did alleviate this diversity is continuing education [69]. The
not have a dedicated, high-end cardiac CT scanner available. Chinese Society of Radiology and SCCT should promote
Participating hospitals were all public; thus obtaining such education and training programs similar to those of other
an expensive scanner (64+ CT) via government procurement academic societies [72].
is difficult. Furthermore, a percentage of the participating We believe that our data provides an overview of the cur-
hospitals are centers specialized on cancer care, pediatric rent practice of cardiac CT in mainland China and provides
hospitals, maternity hospitals, mental hospitals, and tradi- interesting insights into the different practical aspects of its
tional Chinese medicine hospitals, and as such, cardiovascu- routine use. The results of our survey show that cardiac CT
lar diseases are not their major concern. Second, several of with modern CT systems has become an established radio-
the clinical scenarios which were considered appropriate and logical modality in China after nearly 10 years of evolution,
frequent indications for coronary CTA by the hospitals in but at the same time, the method would benefit from more
this survey are actually inappropriate by current clinical sophisticated and uniform application and performance.
guidelines. For example, 95% of respondents stated that they Education is most likely the key to achieve further improve-
performed coronary CTA in clinically healthy individuals, ment. Being the first study of its kind, the study will serve as
which is rated as inappropriate in current guidelines [65]. a baseline for future studies to judge changes in coronary
While a few experts recommend the use of coronary CTA as CTA as well as other related procedures in China.
a screening test, these so-called “health check-up” indica-
tions may reflect an undesirable trend of aggressive market-
ing of cardiac CT [66]. Chinese physicians may not always  urrent Scientific Research of Cardiovascular
C
refuse a patients’ “unreasonable” request, especially in areas CT in China
of rapidly evolving healthcare and in certain socioeconomic
environments [67]. Third, the estimated radiation dose was As cardiovascular disease is widely distributed in China, the
relatively high, reflecting choices of acquisition parameters. scientific research of cardiovascular CT is also increasing.
In a seminal paper, Hausleiter et al. reported dose data from Here we investigated the different types of cardiovascular
the PROTECTION I study and demonstrated that the esti- CT research from July 2015 to June 2016 (Table 54.5).
mated radiation dose of coronary CTA on a global scale in
2007 was 12 mSv, as compared to 15 mSv in East Asia [68].
In the past 8 years, several effective strategies to reduce radi- Table 54.5  Published papers on cardiovascular CT in China from July
ation dose have become available, with submillisievert coro- 2015 to June 2016
nary CTA feasible in selected patients [69]. Nevertheless, Radiology Cardiology Other In
estimated doses differed significantly among sites. Some of Research topics journal journal journals total
the most effective strategies for dose reduction, including Dose and image 21 2 5 28
prospectively ECG-triggered acquisition, tube current quality
adjustment, particularly in combination with iterative recon- CAD diagnosis 4 2 6
struction algorithms, were not frequently used [70]. These CT-guided PCI 3 2 1 6
CT plaque imaging 2 1 3
approaches for dose reduction are largely dependent on CT
Functional CT 2 2 4
hardware and software. Even if available on the scanner plat- imaging
form, technicians may require a more user-friendly operating Congenital heart 3 1 4
interface, supporting “point-of-care” decisions to use low-­ diseases
dose techniques, rather than uniform use of “one-size-fits-­ Pulmonary vessel 2 3 5
all” protocols ordered by supervisors. diseases
Valvular diseases 1 2 3
Fourth, we found significant variability in all aspects of
Adipose tissue 1 1 2 4
cardiac CT performance, including the number of annual volume of
coronary CTA procedures, which varied from 150 to over pericardium
8000 per year; radiation dose, which varied from 3–5 mSv to Biomarkers and 6 6 12
>15 mSv; as well as cumulative time required for scanning, CCTA
processing, and reporting, which varied from 18  min to Other rare diseases 2 1 3
In total 39 17 22 78
54  A Test on the Move: Cardiac CT in China as a Case Study 701

Information Sources Congenital Heart Disease

SCI-EXPANDED was searched electronically from July 1, The congenital heart disease is a constant topic in China due
2015 to June 31, 2016, utilizing the keywords and word to the huge population. Four articles were focused on con-
variant of CT, cardiac CT, and coronary heart disease. The genital heart disease last year, including the tetralogy of
search criteria was as follows: (TS=(((computed tomogra- Fallot, the vascular ring, the aortic arch branching variation,
phy) OR ((multi-slice spiral CT) OR (multi-slice CT) OR and the variation of the left atrial appendage.
(multi-row spiral CT))) AND ((heart OR cardiac OR (heart
disease) OR cardiovascular) OR ((coronary artery) OR
(coronary artery disease) OR (coronary heart disease)))) Pulmonary Disease
AND CU=(CHINA)) AND type: (Article). The article
number was 167. Exclusion criteria include case reports, Five articles of pulmonary disease on cardiovascular CT
reviews, meta-analysis, comments, and letters. We also were published last year, mostly focused on risk factors of
excluded research for cardiovascular US, MRI, SPECT, acute and chronic pulmonary embolism.
and PET-CT of other diseases and research not done in
China. Finally, 78 articles were included.
Valvular Disease

 ublished Papers on Cardiovascular CT by

P On evaluation of valvular disease, computed tomography
Chinese Researchers in 2016 had become one of the standard imaging modalities for pre-­
procedural annular sizing prior to transcatheter aortic valve
The research field can be concluded to radiation dose and replacement (TAVR) and mitral annulus repair. Three arti-
image quality, coronary artery disease, congenital heart dis- cles were published on the cardiac CT evaluation of valvular
ease, pulmonary disease, valvular disease, epicardial and disease last year. They might be used for choosing valve
pericoronary adipose tissue, biomarkers, and other rare dis- device and complication treatment.
eases. For the departments of the author, 39 authors came
from radiology, 17 came from cardiology, and 22 came from
other departments. The 22 authors from other departments Epicardial and Pericoronary Adipose Tissue
are distributed in surgery, ICU, fundamental school, CT
companies, etc. Epicardial and pericoronary adipose tissue was found to cor-
relate with coronary heart disease. Thus, an increasing inter-
est of relationship between adipose tissue and coronary heart
Radiation Dose and Image Quality disease was studied. In the last year, four articles were pub-
lished on this region.
In the last year, articles about radiation dose and image qual-
ity of cardiovascular CT were 28 in total, including the con-
trast medium dose, image processing and three-dimensional Biomarkers
reconstruction, and image quality control. The prospective
ECG-triggered high-pitch coronary CT angiography at 70 Biomarkers of coronary artery calcification were also stud-
kVp in a clinical setting was investigated. ied. Instead of radiologists, most studies were done by cardi-
ologists and fundamental researchers. There were 12 articles
on this topic. The biomarkers for coronary artery calcifica-
Coronary Artery Disease tion were associated with cystatin C, big endothelin-1, and
osteocalcin-positive endothelial progenitor cells. The coro-
There were 19 articles focused on the diagnosis of coronary nary artery calcification was associated with low serum tes-
artery disease. Among them, six studies were about CAD tosterone in elderly men with stable CAD and serum
diagnosis, six about CT-guided PCI, three about plaque gamma-glutamyltransferase in type 2 diabetes mellitus.
imaging, and four about cardiac functional CT (myocardial Hyperphosphatemia, angiotensin-converting enzyme 2, and
CT perfusion, FFR-CT). FGF-23 were correlated with coronary artery calcification in
702 B. Lu et al.

Fig. 54.9  Comparisons of 50 48

different impact factors and
paper volumes of published 45
papers between 2013 and
2016 on cardiovascular CT in 40
China
35

30

25
21
19
20 18
17

15 13

10

0
IF < 1.0 IF = 1-3 IF > 3
2013 2016

peritoneal dialysis patients. Genetic circulating microRNAs References

miR-134, miR-3135b, and miR-2861 were correlated with
coronary artery calcification in symptomatic patients. SNPs 1. Ministry of Health of the People’s Republic of China. Statistics
that are associated with CAD may also increase the risk of Year Book of Health of the People’s Republic of China. Beijing:
Pecking Union Medical College Press; 2009–2012.
calcification with aortic arch. This might suggest utilization 2. National Health and Family Planning Commission of the People’s
of radiomics in the future, which combines the radiology Republic of China. Year Book of Health and Family Planning in
feature with their clinical features, gene information, and the People’s Republic of China. Beijing: Pecking Union Medical
pathology information. College Press; 2015.
3. Zhou MG, Wang HD, Zhu J, et  al. Cause-specific mortality for
240 causes in China during 1990–2013: a systematic subnational
analysis for the Global Burden of Disease Study 2013. Lancet.
2016;387:251–72.
Other Rare Diseases 4. Report of Nutrition and Chronic Diseases in China. 2015. http://
www.labagd.com/Item/14257.aspx.
Three other studies include detection of aortic rupture using 5. Li LM, Rao KQ, Kong LZ, et  al. A description on the Chinese
national nutrition and health survey in 2002. Chin J Epidemiol.
postmortem computed tomography, detection of coronary
2005;07:478–84.
artery dissection, and detection of cardiac tumor. 6. Xi B, Liang Y, Reilly KH, et al. Trends in prevalence, awareness,
treatment, and control of hypertension among Chinese adults 1991-
2009. Int J Cardiol. 2012;158(2):326–9.
7. Pan Y, Cai W, Cheng Q, et  al. Association between anxiety and
I mpact Factors of Published Papers hypertension: a systematic review and meta-analysis of epidemio-
logical studies. Neuropsychiatr Dis Treat. 2015;11:1121–30.
on Cardiovascular CT 8. Survey on prevalence and risk factors for primary hypertension
among communities in Changing District, Shanghai. Geriatr Health
For published papers on cardiovascular CT in China last Care. 2012;18(1):19–22.
year, there were 13 original papers for SCI-IF <1.0, 48 papers 9. Wang JG, Li LH.  Hypertension series: the differential diagnosis
and treatment for secondary hypertension- continued. Chin Circ J.
for SCI-IF 1.0–3.0, and 19 papers for SCI-IF >3.0. Compared
2012;27(2):85–6.
with 2013, although the quality of papers was not obviously 10. Li NF, Wang L, Zhou KM, et al. Analysis of etiology of the patients
increased, the number of SCI-IF 1.0–3.0 papers increased with hypertension from the People’s Hospital of Xinjiang Uygur
from 21 to 48. This suggested that much more research had Autonomous Region. Chin J Cardiol. 2007;35(9):865–8.
11. Yang GH, Fan LX, Tan J, et al. Smoking in China: Findings of the
been done on cardiovascular CT in China. We also hope that
1996 National Prevalence Survey. JAMA. 1999;282(13):1247–53.
the high-quality research will come soon, which will be 12. Yang GH, Li Q, Hsia J. Prevalence of smoking in China in 2010. N
mentioned in the articles below (Fig. 54.9). Engl J Med. 2011;364(25):2469–70.
54  A Test on the Move: Cardiac CT in China as a Case Study 703

13. Huang ZJ, Wang LM, Zhang M, et al. Smoking behavior among the 35. Chen Y, Zheng Z, Yi J, et al. Associations between physical inac-
Chinese employed floating population aged 18-59 in 2012. Chin J tivity and sedentary behaviors among adolescents in 10 cities in
Epidemiol. 2014;35(11):1192–7. China. BMC Public Health. 2014;14(1):744.
14. Han J, Che XG. A meta-analysis of cigarette smoking prevalence 36. Fan S, Chen J, Huang J, et al. Physical activity level and incident
among adolescents in China: 1981-2010. Int J Environ Res Public type 2 diabetes among Chinese adults. Med Sci Sports Exerc.
Health. 2015;12:4617–30. 2015;47(4):751–6.
15. Yang GH.  Global Adult Tobacco Survey (GATS)-China Report. 37. Zhai FY, Yan XG.  Survey Report on the Status of Nutrition and
Beijing: Three-Gorgas Press; 2010, 2011.11. Health of the Chinese People in 2002: Series Two, Status of Diet
16. Gu D, Kelly TN, Wu X, et al. Mortality attributable to smoking in and Nutrients consumption, vol. 7. Beijing: People’s Medical
China. N Engl J Med. 2009;360(2):150–9. Publishing House; 2006.
17. He J, Gu D, Reynolds K, et al. Serum total and lipoprotein choles- 38. Yu D, Zhang X, Xiang YB, et  al. Adherence to dietary guide-
terol levels and awareness, treatment, and control of hypercholes- lines and mortality: a report from prospective cohort studies
terolemia in China. Circulation. 2004;110(4):405–11. of 134,000 Chinese adults in urban Shanghai. Am J Clin Nutr.
18. Yang WY, Xiao JZ, Yang ZJ, et al. Serum lipids and lipoproteins in 2014;100(2):693–700.
Chinese men and women. Circulation. 2012;125:2212–21. 39. Yu D, Zhang X, Gao YT, et al. Fruit and vegetable intake and risk
19. Li JH, Mi SQ, Li YC, et al. The levels and distribution of serum lipids of CHD: results from prospective cohort studies of Chinese adults
in Chinese adults, 2010. Chin J Prev Med. 2012;46(7):607–12. in Shanghai. Br J Nutr. 2014;111(2):353–62.
20. Zhao WH, Zhang J, Li Y. Report of Chinese nutrition and health sta- 40. Du S, Neiman A, Batis C, et  al. Understanding the patterns and
tus: blood lipid 2002 [J]. People’s Medical Publishing House. 2002. trends of sodium intake, potassium intake, and sodium to potassium
21. Ding WQ, Dong HB, Mi J, et  al. Prevalence of dyslipidemia
ratio and their effect on hypertension in China. Am J Clin Nutr.
in Chinese children and adolescents: a Meta-analysis. Chin J 2014;99(2):334–43.
Epidemiol. 2015;36(1):71–7. 41. Brook RD, Rajagopalan S, Pope CA 3rd, et al. Particulate matter air
22. Xu Y, Wang L, He J, et  al. Prevalence and control of diabetes in matter air pollution and cardiovascular disease: an update to the sci-
Chinese adults. JAMA. 2013;310(9):948–58. entific statement from the American Heart Association. Circulation.
23. An YL, Zhang P, Wang JP, et al. Cardiovascular and all-cause mor- 2010;121(21):2331–78.
tality over a 23-year period among Chinese with newly diagnosed 42. Kan H, London SJ, Chen G, et al. Differentiating the effects of fine
diabetes in the Da Qing IGT and Diabetes Study. Diabetes Care. and coarse particles on daily mortality in Shanghai, China. Environ
2015;38(7):1365–71. Int. 2007;33:376–84.
24. Li GW, Zhang P, Wang J, et al. Long-term effect of lifestyle interven- 43. Huang W, Cao J, Tao Y, et al. Seasonal variation of chemical species
tions to prevent diabetes in the China Da Qing Diabetes Prevention associated with short-term mortality effects of PM(2.5) in Xi’an, a
Study: a 20-year follow-up study. Lancet. 2008;371:1783–9. Central City in China. Am J Epidemiol. 2012;175(6):556–66.
25. Gong Q, Gregg EW, Wang J, et  al. Long-term effects of a ran- 44. Zhao K, Cao JY, Wen XM. Correlation between PM2.5 pollution in
domised trial of a 6-year lifestyle intervention in impaired glucose air and mortality of residents in urban area, Xi’an. [Chinese]. J Prev
tolerance on diabetes-related microvascular complications: the Med Inf. 2011;27(4):257–62.
China Da Qing Diabetes Prevention Outcome Study. Diabetologia. 45. Ma Y, Chen R, Pan G, et al. Fine particulate air pollution and daily
2011;54(2):300–7. mortality in Shenyang, China. Sci Total Environ. 2011;409:2473–7.
26. Li GW, Zhang P, Wang JP, et al. Cardiovascular mortality, all-cause 46. Yang C, Peng X, Huang W, et al. A time-stratified case-crossover study
mortality, and diabetes incidence after lifestyle intervention for of fine particulate matter air pollution and mortality in Guangzhou,
people with impaired glucose tolerance in the Da Qing Diabetes China. Int Arch Occup Environ Health. 2012;85(5):579–85.
Prevention Study: a 23-year follow-up study. Lancet Diabetes 47. Dai J, Chen R, Meng X, et al. Ambient air pollution, temperature
Endocrinol. 2014;2(6):474–80. and out-of-hospital coronary deaths in Shanghai, China. Environ
27. Ji L, Hu D, Pan C, et  al. Primacy of the 3B approach to control Pollut. 2015;203:116–21.
risk factors for cardiovascular disease among type 2 diabetes. Am J 48. Xie W, Li G, Zhao D, et al. Relationship between fine particulate
Med. 2013;126(10):925e11–22. air pollution and ischaemic heart disease morbidity and mortality.
28. Ma GS, Li YP, Wu YF, et al. The prevalence of body weight and Heart. 2015;101:257–63.
obesity and its change among Chinese people during 1992 to 2002. 49. National Health and Family Planning Commission of the People’s
Chin J Prev Med. 2005;39(5):311–5. Republic of China. China Health and Family Planning Statistics
29. Gordon Larsen P, Wang H, Popkin BM. Overweight dynamics in Yearbook. Beijing: Peking Union Medical College Press; 2015.
Chinese children and adults. Obes Rev. 2014;15(S1):37–48. 50. Cao J, Yang C, Li J, et al. Association between long-term exposure
30. Li XY, Jiang Y, Hu N, et al. Prevalence and characteristic of over- to outdoor air pollution and mortality in China: a cohort study. J
weight and obesity among adults in China, 2010. Chin J Prev Med. Hazard Mater. 2011;186:1594–600.
2012;46(8):683–6. 51. Wang Y, Zhao X, Jiang Y, et  al. Prevalence, knowledge, and

31. Lao XQ, Ma WJ, Sobko T, et  al. Overall obesity is leveling-off treatment of transient ischemic attacks in China. Neurology.
while abdominal obesity continues to rise in a Chinese popu- 2015;84(23):2354–61.
lation experiencing rapid economic development: analysis of 52. Ministry of Health Statistics Center. 2008 China health services
serial cross-sectional health survey data 2002–2010. Int J Obes. survey – 4th Family health survey. China Statistics Press. 2009.
2015;39(2):288–94. 53. Gu DF, Huang GY, He J, et al. Investigation of prevalence and dis-
32. Ma J, Cai XH, Wang HJ, et al. The trend analysis of overweight and tributing feature of chronic heart failure in Chinese adult popula-
obesity in Chinese students during 1985-2010. Chin J Prev Med. tion. Chin J Cardiol. 2003;31(1):3–6.
2012;46(9):776–80. 54. Zhang J, Zhang YH, et al. China heart failure registry study–a mul-
33. Du HD Li LM, Whitlock G, et al. Patterns and socio-demographic ticenter, prospective investigation for preliminary analysis on etiol-
correlates of domain-specific physical activities and their associa- ogy, clinical features and treatment in heart failure patients. Chin
tions with adiposity in the China Kadoorie Biobank Study. BMC Circ J. 2015;30(5):413–6.
Public Health. 2014;14:826. 55. Ministry of Health of People’s Republic of China. National health
34. Ng SW, Howard AG, Wang HJ, et al. The physical activity transition statistics annual report 1980–2001 [J], vol. 01: China Statistics
among adults in China: 1991–2011. Obes Rev. 2014;15(S1):27–36. Press; 1985. p. 56.
704 B. Lu et al.

56. Ministry of Health of People’s Republic of China. China health sta- Guidelines Committee. J Cardiovasc Comput Tomogr.
tistics yearbook [J]. China Statistics Press. 2002–2012. 2014;8:342–58.
57. National Health and Family Planning Commission of the People’s 65. Kim YJ, Yong HS, Kim SM, Kim JA, Yang DH, Hong YJ. Korean
Republic of China. China statistical yearbook of health and fam- guidelines for the appropriate use of cardiac CT. Korean J Radiol.
ily planning [J]. Beijing: People’s Medical Publishing House; 2015;16:251–85.
2013–2015. 66. Muhlestein JB, Lappe DL, Lima JA, et al. Effect of screening for
58. He D, Yu H, Chen Y. Equity in the distribution of CT and MRI in coronary artery disease using CT angiography on mortality and
China: a panel analysis. Int J Equity Health. 2013;12:39. cardiac events in high-risk patients with diabetes: the FACTOR-64
59. Oh E-H, Imanaka Y, Evans E. Determinants of the diffusion of com- randomized clinical trial. JAMA. 2014;312:2234–43.
puted tomography and magnetic resonance imaging. Int J Technol 67. Blumenthal D, Hsiao W. Lessons from the East—China’s rapidly
Assess Health Care. 2005;21:73–80. evolving health care system. N Engl J Med. 2015;372:1281–5.
60. Ma J, Lu M, Quan H.  From a national, centrally planned health 68. Hausleiter J, Meyer T, Hermann F, et al. Estimated radiation dose
system to a system based on the market: lessons from China. Health associated with cardiac CT angiography. JAMA. 2009;301:500–7.
Aff (Millwood). 2008;27:937–48. 69. Stehli J, Fuchs TA, Bull S, et al. Accuracy of coronary CT angiog-
61. X-ray computed tomography. Wikipedia 2012. http://en.Wikipedia. raphy using a submillisievert fraction of radiation exposure: com-
org/wik/x-ray_computed_tomography. parison with invasive coronary angiography. J Am Coll Cardiol.
62. Liu K, Hsieh C, Zhuang N, Gao Y, Li ZN, Ren XS, Yang L, Zhang 2014;64:772–80.
JY, Budoff MJ, Lu B.  Current utilization of cardiac computed 70. Zhao L, Plank F, Kummann M, et al. Improved non-calcified plaque
tomography in mainland China: a national survey. J Cardiovasc delineation on coronary CT angiography by sonogram-affirmed
Comput Tomogr. 2016;10(1):76–81. https://doi.org/10.1016/j. iterative reconstruction with different filter strength and relation-
jcct.2015.11.004. PMID26699712. ship with BMI. Cardiovasc Diagn Ther. 2015;5:104–12.
63. Maurer MH, Hamm B, Dewey M.  Survey regarding the clinical 71. Schindera ST, Treier R, von Allmen G, et  al. An education and
practice of cardiac CT in Germany: indications, scanning technique training programme for radiological institutes: impact on the reduc-
and reporting. RöFo. 2009;181:1135e1143. tion of the CT radiation dose. Eur Radiol. 2011;21:2039–45.
64. Leipsic J, Abbara S, Achenbach S, et  al. SCCT guidelines for 72. Garcia MJ, Blankstein R, Budoff MJ, et al. COCATS 4 Task Force
the interpretation and reporting of coronary CT angiography: a 7: Training in Cardiovascular Computed Tomographic Imaging. J
report of the Society of Cardiovascular Computed Tomography Am Coll Cardiol. 2015;65:1810–21.
 Part XIII
Where We Are Going: The Genes and the Heart
 Differences and Disparities
in Cardiovascular Medicine Related 55
to Gender, Race, and Ethnicity:
The Role of Cardiac CT

John W. Nance

Knowledge of demographic differences in physiology, History

pathology, and therapeutic response is necessary for optimal
patient diagnosis and treatment. However, there are demo- There have been significant changes in the global burden of
graphic, socioeconomic, and geographic differences in cardiovascular disease over the past several decades, which
healthcare access, disease occurrence, and resultant morbid- are beyond the scope of this chapter. Interestingly, however,
ity/mortality that transcend biology; these differences consti- there is a dramatic and ongoing reevaluation of heart disease
tute healthcare and health status disparities. Some differences burden in varied populations. Some of this shift is due to
in healthcare quality are due to clinical appropriateness/need changing risk profiles, treatment strategies, and actual dis-
and patient preferences, which may be appropriate, but there ease prevalence; however, there is likely also a component of
is also well-established discrimination as well as differences prior and ongoing information bias. For example, hyperten-
in the operation of healthcare systems leading to unneces- sion and hypertensive heart disease were reported as twice as
sary, unjust, and avoidable inequities. An understanding of frequent in American blacks compared to whites in the
both unavoidable biological differences and avoidable dis- 1920s, but coronary heart disease (CHD) was considered
parities in cardiovascular health and care is necessary to uncommon in blacks, possibly due to missed cases [1]. This
appropriately select patients for cardiac CT, interpret the generally accepted medical “fact” was relatively slow to
findings, and guide further management. Differences and change, until the increased focus on disparities in minority
disparities, as defined above, will be discussed concurrently. health over the past three decades. The US Department of
A brief note is warranted on the categorizations discussed Health and Human Services Secretary’s Task Force on Black
throughout this chapter. Race is a sociopolitical construct, and Minority Health was published in 1985 and provided
rather than a biological or genotypic construct, and conclu- one of the earlier comprehensive surveys of heart health in
sions regarding racial differences in pathophysiology should various populations, showing higher average annual age-­
be interpreted accordingly. Likewise, ethnicity (in this chap- adjusted death rates for heart disease in blacks versus whites,
ter, Hispanic or not Hispanic) is a cultural distinction with particularly in the younger and female subsets. While a sig-
tremendous intra-ethnic heterogeneity. Shifts in the defini- nificant portion of this difference can be attributed to differ-
tions of race and ethnicity can change faster than shifts in ences in hypertension, the study also found that CHD
disease burden or pathophysiology, necessitating constant mortality rates were in fact similar in black and white men
surveillance to identify relevant differences and eliminate and higher in black than white women [2]. These differences
disparities. Sex will be used to denote biological (i.e., chro- were attributed to disparities in care in the pivotal release of
mosomal) differences, whereas gender differences are socio- the Institute of Medicine’s report, Unequal Treatment:
cultural. Finally, the volume and quality of data on Confronting Racial and Ethnic Disparities in Health Care,
cardiovascular health for different populations are variable, in 2002 [3]. Disparities specific to cardiovascular care were
leading to unfortunate but unavoidable gaps in the following likewise highlighted in the Henry J.  Kaiser Family
discussion. Rectifying the disparities in both healthcare qual- Foundation’s report, Racial/Ethnic Differences in Cardiac
ity and healthcare knowledge should remain a high priority. Care: The Weight of the Evidence, also released in 2002.
The history of cardiovascular disease in women suffers
from similar flaws: While past data suggested that cardiovas-
J. W. Nance (*) cular disease, and CHD in particular, were predominantly
Department of Radiology, Medical University of South Carolina, diseases of men, there was likely a component of informa-
Charleston, SC, USA
tion bias given historic gender disparities in clinical trial
e-mail: [email protected]

© Humana Press 2019 707

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_55
708 J. W. Nance

inclusions and the more common atypical presentation of p­revalence and expression that are independent of risk
acute coronary syndromes in women. Indeed, CHD has ­factors, implying more complex underlying pathophysio-
killed more American women than men on an annual basis logic mechanisms. On a provider level, different races and
since 1984 [4]. Furthermore, there are gender differences in ethnicities are subject to varying types and levels of risk fac-
anatomic burden of disease that do not correlate linearly with tor modulation, diagnostic strategies, and therapeutic inter-
clinical outcome (discussed below), further obfuscating ventions. Access to care, infrastructure, and insurance
interpretations of historic studies on disease prevalence and coverage constitutes the major system-level disparities
extent. While management recommendations are similar in ­currently encountered.
men and women, some of these long-standing biases appar-
ently persist, and there are continued disparities in time to
diagnosis and implementation of recommended medical and Patients of African Descent
surgical treatment [5].
Black patients overall have worse risk factor profiles than
whites, particularly hypertension, diabetes, and obesity (the
Race and Ethnicity latter most pronounced in women). The predictive value of
some risk factors (e.g., hypertension and diabetes) is greater
The burden of cardiovascular disease spares no race or eth- than whites, while the value of others (such as lipoprotein
nicity, but significant heterogeneity exists between and even [a]) is lower [1]. The consequent burden of cardiovascular
within different subpopulations (Table 55.1). Outcome dis- disease, particularly hypertensive heart disease, is dramatic,
parities can be attributed to patient-, provider-, and system-­ with three to five times higher cardiovascular mortality asso-
level problems. On a patient level, there are differences in the ciated with hypertension in blacks compared to whites.
rates and prognostic implications of cardiovascular disease The spectrum and burden of CHD in blacks are more
risk factors, but there are also differences in disease nuanced. For example, the prevalence of myocardial infarc-
tion is higher in white males, but incidence is higher in black
males (prevalence and incidence are higher in black females
Table 55.1  Racial and ethnic differences in cardiovascular diseasea
compared to white counterparts). Blacks are more likely to
African origin Asian origin Hispanic origin present with atypical symptoms, seek care later than whites,
Risk Much higher rates More Nearly twice the and are subject to greater delays prior to receiving medical or
factors and lower control nontraditional rate of diabetes
of hypertension risk factors and and metabolic surgical treatments. Perhaps most unsettling, the outcomes
Twice as likely to smoking in syndrome, of CHD are significantly worse in blacks, with higher mor-
have diabetes certain particularly tality at younger ages [1].
subpopulations among Surprisingly, black patients have less angiographically
Mexican-­
Americans significant coronary artery disease (CAD) than white coun-
Disease Approximately Approximately Approximately terparts, despite worse risk factor profiles and outcomes [6].
burden 30% higher 50% lower 30% lower The reasons for this apparent paradox are unclear; one of the
mortality rate mortality rate mortality rate leading theories is that the overall worse risk factor profile
from from from
and relatively increased rates of hypertensive heart disease
cardiovascular cardiovascular cardiovascular
disease disease disease and left ventricular hypertrophy provide a more malignant
Worse outcomes Higher rates of substrate for adverse events. Cardiac CT, both coronary cal-
in the presence of heart failure cium scoring and coronary CT angiography (cCTA), has pro-
cardiovascular Worse outcomes
vided another possible mechanism: blacks tend to have
disease in the presence
of cardiovascular overall decreased coronary artery calcium and calcified
disease plaques, while having relatively more noncalcified disease
Imaging Less coronary Less coronary Less coronary [7, 8], possibly reflecting pathophysiologic differences in the
findings artery calcium artery calcium artery calcium stability of underlying atherosclerosis (Fig.  55.1). There is
and Lower left Less
angiographically ventricular mass nonobstructive conflicting data on the prognostic implications of these find-
obstructive Less and obstructive ings, with some studies reporting differing outcomes associ-
disease nonobstructive CAD ated with coronary artery calcium burden (as measured by
Greater left and obstructive unenhanced CT calcium scoring) in blacks and others show-
ventricular mass CAD
Relatively more ing no significant difference [9, 10]. The overall prognostic
noncalcified patterns in cCTA seem to hold across populations, with very
plaque low risk of future events in patients without any atheroscle-
a
Figures are relative to non-Hispanic whites rotic disease and progressively worse outcomes in patients
55  Differences and Disparities in Cardiovascular Medicine Related to Gender, Race, and Ethnicity: The Role of Cardiac CT 709

with nonobstructive and obstructive CAD. Interestingly, data How do we apply this information to daily cardiac CT
from the large CONFIRM registry also showed that patients practice? While there are clearly differences in risk factor
with African ethnicity had the highest relative increase in profiles, there is no evidence to suggest that these differences
adverse events (specifically, all-cause mortality and myocar- should be exploited in cardiac CT ordering patterns.
dial infarction) when obstructive versus nonobstructive dis- However, the relatively lower amounts of calcified plaque
ease was present, again arguing for an increased clinical and greater amounts of noncalcified plaque in blacks, com-
susceptibility to disease in blacks [11]. bined with worse outcomes in the presence of any disease,

a b

c d

Fig. 55.1  A 59-year-old black woman presented to the emergency (arrows in b and c). The corresponding LAD territory was thinned and
department with atypical chest pain. Calcium scoring revealed a single severely hypokinetic on functional images (arrows in d). The patient
punctate focus of calcium (arrow in a), corresponding to an Agatston was subsequently taken to catheterization, where angioplasty and stent-
score of 1. Despite the low Agatston score, cCTA revealed occlusion of ing of the LAD occlusion (arrow in e) were performed
the proximal LAD at the level of the takeoff of the first diagonal branch
710 J. W. Nance

delays before treatment and suffer worse outcomes com-

e
pared to non-Hispanic whites [13]. Similar to other immi-
grant populations, Hispanics in the United States suffer
increasing levels of cardiovascular disease as acculturation
increases. Interestingly, Hispanics with low acculturation
have been shown to demonstrate a paradoxical relationship
between socioeconomic status and endpoints of cardiovascu-
lar disease relative to other Western populations, i.e., worse
outcomes with higher socioeconomic status. This relation-
ship reverts back to the expected, inverse relationship
between socioeconomic status and outcomes as accultura-
tion increases [14].
Imaging studies have shown that cCTA has good sensitiv-
ity and specificity for the detection of coronary artery steno-
sis [15] and that coronary artery calcium scoring and cCTA
findings independently predict death or myocardial infarc-
tion in Hispanics [11]. There is no available evidence to sug-
gest that cardiac CT should be ordered, acquired, or
interpreted differently in Hispanic compared to non-Hispanic
white patients.
Fig. 55.1 (continued)

Patients of Asian Descent

might warrant more heavy weighting for cCTA over simple
coronary artery calcium scoring in clinical situations in Comparative racial and ethnic research involving Asian pop-
which both would be appropriate. Noncalcified and nonob- ulations is less studied. Asians, particularly East Asians, tend
structive disease also warrant specific mention in cCTA to have better risk profiles than whites and other racial/ethnic
reports. Finally, the optimal clinical management of black minorities, with correspondingly less subclinical and clinical
patients who have undergone cardiac CT is likely a high-­ cardiovascular disease. East Asians are unique among racial
yield area for future research, particularly in the blooming and ethnic minorities in having similar outcomes compared
era of personalized medicine. For example, do black patients to whites following adverse cardiac events. Calcium scoring
with noncalcified nonobstructive disease but increased myo- and cCTA demonstrate good sensitivity and specificity and
cardial mass require more aggressive follow-up or medical hold independent prognostic value in Asians; however, East
management? Future studies should help validate personal- Asians have been shown to have less than expected mortality
ized treatment algorithms. and nonfatal myocardial infarction in the presence of athero-
sclerosis on cCTA [11]. Nevertheless, there is no evidence to
suggest that race-specific ordering, acquisition, or interpreta-
Hispanics tion should be applied to Asian populations.

Among the well-studied racial and ethnic minorities,

Hispanics show the greatest underlying heterogeneity, with Gender
varying regions of origin, acculturation levels, and contribu-
tions of European, African, and Native American ancestry. Cardiovascular disease is the leading killer of women in the
As a group Hispanics have an overall worse risk profile than United States, and while overall mortality from heart dis-
non-Hispanic whites, with particularly high levels of diabe- ease has decreased since 2001, mortality in women under
tes and metabolic syndrome. Despite this increased risk, 45 years old has increased. CHD prevalence is lower in all
Hispanics suffer lower levels of subclinical cardiovascular women compared to all men, but this differential decreases
disease on imaging (coronary artery calcium and left ven- with advancing age, and in absolute numbers, more women
tricular mass) [7, 12] and have lower rates of total cardiovas- than men die of cardiovascular disease annually [4]. Global
cular disease and CHD; this phenomenon is sometimes epidemiological figures are similar  – heart disease is the
termed the “Hispanic paradox.” As in patients of African leading cause of death in women in all major high-income
descent, Hispanic patients presenting with acute coronary economies and the majority of middle- and low-income
syndrome and myocardial infarction experience longer economies [16].
55  Differences and Disparities in Cardiovascular Medicine Related to Gender, Race, and Ethnicity: The Role of Cardiac CT 711

a b

Fig. 55.2  A 54-year-old white woman presented to the emergency mid-LAD (arrow in a). Contrast with the obstructive mixed calcified
department with chest pain. There was no detectable coronary calcium. and noncalcified plaque in this 57-year-old white man presenting with
cCTA revealed nonobstructive noncalcified plaque in the proximal to chest pain (arrow in b)

Unfortunately, awareness of heart disease in women lags propensity for plaque rupture in women >50  years old.
behind the actual burden, both among patients and providers, Women under 50 years old have significantly higher rates of
which can lead to suboptimal management or delays in pre- plaque erosion (versus rupture) compared to older counter-
sentation and diagnosis. Despite similar guidelines for the parts. In women of all ages, there is less obstructive CAD
prevention and management of cardiovascular disease in compared to men, both in an asymptomatic population and
men and women, there is disparity in their implementation. in those presenting with acute coronary syndrome. Up to
Symptomatic women are less likely to be referred for nonin- 20% of women presenting with acute coronary syndrome
vasive imaging, angiography, or revascularization (Fig. 55.2). have angiographically normal coronary arteries, with isch-
Women with angiographic evidence of CHD are less likely emia caused by microvascular disease. For this reason, the
to receive optimal primary and secondary therapy. Women term “ischemic heart disease” (IHD), which includes
have poorer outcomes following acute coronary syndrome obstructive CAD, nonobstructive CAD, and microvascular
and higher risks of intracranial bleeding following thrombo- disease, is preferred over “CAD.”
lytics. They are less likely to receive and wait longer for Traditional, emerging, and sex-specific risk factors play a
diagnostic and invasive procedures while admitted and less role in comprehensive risk stratification in women.
likely to receive recommended pharmacotherapy upon dis- Application of global risk scores (such as the Framingham
charge. Finally, adverse outcomes are more common in Risk Score) is a class I indication in asymptomatic adults
women following revascularization [5]. without a clinical history. The American Heart Association’s
Differences in pathophysiology, risk factor profile and (AHA) women-specific guidelines further specifies that
relative risks, and presentation of cardiovascular disease asymptomatic women should be classified into three differ-
between women and men can help explain some, but not all, ent risk categories – high risk, at risk, and ideal cardiovascu-
of the variations in management and outcomes. Estrogen is lar health – with the reasoning that traditional scoring (e.g.,
thought to play a protective role in premenopausal women, Framingham Risk Score) is relevant in identifying high-risk
helping to explain the gap in CHD prevalence in younger women but suboptimal to ensure low risk given the high life-
women compared to younger men; recall that the gap nar- time burden of cardiovascular disease in women (nearly one
rows in older women and men. Indeed, postmenopausal in two) [17]. Symptomatic women should also be classified
changes in atherosclerotic plaque composition have been into low, intermediate, and high risk of IHD. Because of the
identified, with higher rates of lipid-laden plaques with higher prevalence of atypical symptoms in women, risk
necrotic cores, more inflammatory changes, and a greater ­stratification is more heavily weighted toward age and risk
712 J. W. Nance

f­actors rather than clinical presentation. In general, women To date, routine clinical practice has underexploited the
are considered low risk if premenopausal (e.g., 50s or unique ability of cCTA to noninvasively identify nonobstruc-
younger), intermediate risk in their 60s, and high risk in their tive CAD in women. There are no societal guidelines on rec-
70s. Women with multiple cardiac risk factors, functional ommended therapy for patients with signs and symptoms of
disability, and extensive comorbidities should be elevated IHD and nonobstructive CAD. Ischemia on stress testing is
one risk category, and women with peripheral arterial disease often dismissed as “false positive” due to lack of obstructive
or long-­standing, poorly controlled diabetes are considered disease on anatomic imaging, and secondary preventative
high-risk equivalents [18]. strategies are rarely implemented [23]. cCTA may have a
Current consensus statements are mixed on the appropri- more prominent role with evolving diagnostic and treatment
ate use of cCTA in varied populations. For example, the paradigms for IHD other than obstructive CAD, which tends
AHA consensus statement on the role of noninvasive testing to disproportionately affect women.
in women with suspected IHD states that imaging evaluation
of women with low risk for IHD (in the acute or stable set-
tings) is inappropriate and that cCTA is only appropriate in References
women with intermediate to high risk with an abnormal
12-lead rest ECG [18]. In contrast, the ACCF/AHA/ACP/ 1. Clark LT.  Issues in minority health: atherosclerosis and coro-
nary heart disease in African Americans. Med Clin North Am.
AATS/PCNA/SCAI/STS guidelines for the diagnosis and 2005;89(5):977–1001, 994.
management of patients with stable IHD states that cCTA 2. Savage DD. Highlights from the Department of Health and Human
may be appropriate in patients with intermediate pretest Services Task Force on Black and Minority Health. J Natl Black
probability of IHD who are able to exercise and in patients Nurses Assoc. 1986;1(2):16–23.
3. Institute of Medicine (US) Committee on Understanding and
with low to intermediate probability who are not able to Eliminating Racial and Ethnic Disparities in Health Care. Ethnic
exercise [19]. ACCF/SCCT/ACR/AHA/ASE/ASNX/ disparities in health. In: Smedley BD, Stith AY, Nelson AR, edi-
NASCI/SCAI/SCMR appropriate use criteria for cCTA state tors. Unequal treatment: confronting racial and ethnic disparities
that cCTA is also acceptable in low to intermediate probabil- in health care. Washington, DC: National Academies Press (US);
2003.
ity patients presenting with acute chest pain, low probability 4. Mozaffarian D, et  al. Executive summary: heart disease and
asymptomatic patients with family history of premature stroke statistics--2016 update: a report from the American Heart
CAD, and intermediate probability asymptomatic patients Association. Circulation. 2016;133(4):447–54.
without known CAD [20]. Clearly, further work needs to be 5. Shah T, Palaskas N, Ahmed A. An update on gender disparities in
coronary heart disease care. Curr Atheroscler Rep. 2016;18(5):28.
done to improve the quality of evidence, including sex-­ 6. Liao Y, et  al. The relative effects of left ventricular hypertrophy,
specific evidence, and refine optimal diagnostic strategies. coronary artery disease, and ventricular dysfunction on survival
The diagnostic accuracy of cCTA is comparable in women among black adults. JAMA. 1995;273(20):1592–7.
and men, with >90% sensitivity and specificity in the 7. Bild DE, et  al. Ethnic differences in coronary calcification: the
Multi-Ethnic Study of Atherosclerosis (MESA). Circulation.
70–90% range for obstructive CAD [18]. As a general rule, 2005;111(10):1313–20.
calcified and obstructive CAD prevalence and extent in 8. Nance JW Jr, et al. Coronary atherosclerosis in African American
women lag behind men by approximately 10 years; accord- and white patients with acute chest pain: characterization with cor-
ingly, women undergoing cCTA tend to be older, with higher onary CT angiography. Radiology. 2011;260(2):373–80.
9. Detrano R, et  al. Coronary calcium as a predictor of coro-
risk according to traditional risk stratification methods. nary events in four racial or ethnic groups. N Engl J Med.
Notably, over twice as many women (60–70%) as men 2008;358(13):1336–45.
(~30%) with suspected IHD undergoing invasive coronary 10. Doherty TM, Tang W, Detrano RC. Racial differences in the signifi-
angiography have nonobstructive CAD [21]. cance of coronary calcium in asymptomatic black and white subjects
with coronary risk factors. J Am Coll Cardiol. 1999;34(3):787–94.
Several studies have demonstrated sex-specific differ- 11. Hulten E, et  al. Usefulness of coronary computed tomogra-

ences in the prognostic value of various cCTA parameters. phy angiography to predict mortality and myocardial infarction
While women and men have similar outcomes with regard to among Caucasian, African and East Asian ethnicities (from the
extent and severity of obstructive CAD, only women show CONFIRM [Coronary CT Angiography Evaluation for Clinical
Outcomes: an International Multicenter] Registry). Am J Cardiol.
elevated risk-adjusted mortality for the presence and extent 2013;111(4):479–85.
of nonobstructive CAD [18] and may therefore benefit from 12. Gardin JM, et  al. Do differences in subclinical cardiovascular

more aggressive risk factor modification and medical ther- disease in Mexican Americans versus European Americans help
apy. Accordingly, nonobstructive CAD (particularly when explain the Hispanic paradox? Am J Cardiol. 2010;105(2):205–9.
13. Graham G.  Racial and ethnic differences in acute coronary syn-
>2 major epicardial vessels are involved) should be high- drome and myocardial infarction within the United States: from
lighted in cCTA reports. Importantly, the absence of CAD demographics to outcomes. Clin Cardiol. 2016;39(5):299–306.
(obstructive or nonobstructive) portends a very favorable 14. Gallo LC, et  al. Do socioeconomic gradients in subclinical ath-
prognosis in both men and women [22]. erosclerosis vary according to acculturation level? Analyses of
55  Differences and Disparities in Cardiovascular Medicine Related to Gender, Race, and Ethnicity: The Role of Cardiac CT 713

Mexican-Americans in the multi-ethnic study of atherosclerosis. 20. Taylor AJ, et  al. ACCF/SCCT/ACR/AHA/ASE/ASNC/NASCI/

Psychosom Med. 2009;71(7):756–62. SCAI/SCMR 2010 appropriate use criteria for cardiac computed
15. Jug B, et al. Influence of race and ethnicity on diagnostic perfor- tomography. A report of the American College of Cardiology
mance of 64-slice multidetector coronary computed tomographic Foundation Appropriate Use Criteria Task Force, the Society of
angiography. Int J Cardiol. 2013;168(2):1521–3. Cardiovascular Computed Tomography, the American College of
16. Gholizadeh L, Davidson P. More similarities than differences: an Radiology, the American Heart Association, the American Society
international comparison of CVD mortality and risk factors in of Echocardiography, the American Society of Nuclear Cardiology,
women. Health Care Women Int. 2008;29(1):3–22. the North American Society for Cardiovascular Imaging, the
17. Mosca L, et  al. Effectiveness-based guidelines for the preven-
Society for Cardiovascular Angiography and Interventions, and
tion of cardiovascular disease in women--2011 update: a guide- the Society for Cardiovascular Magnetic Resonance. J Am Coll
line from the American Heart Association. J Am Coll Cardiol. Cardiol. 2010;56(22):1864–94.
2011;57(12):1404–23. 21. Jespersen L, et  al. Stable angina pectoris with no obstructive

18. Mieres JH, et al. Role of noninvasive testing in the clinical evalu- coronary artery disease is associated with increased risks of
ation of women with suspected ischemic heart disease: a consen- major adverse cardiovascular events. Eur Heart J. 2012;33(6):
sus statement from the American Heart Association. Circulation. 734–44.
2014;130(4):350–79. 22. Min JK, et  al. Age- and sex-related differences in all-cause mor-
19. Fihn SD, et  al. 2012 ACCF/AHA/ACP/AATS/PCNA/SCAI/
tality risk based on coronary computed tomography angiography
STS Guideline for the diagnosis and management of patients findings results from the International Multicenter CONFIRM
with stable ischemic heart disease: a report of the American (Coronary CT Angiography Evaluation for Clinical Outcomes:
College of Cardiology Foundation/American Heart Association an International Multicenter Registry) of 23,854 patients without
Task Force on Practice Guidelines, and the American College of known coronary artery disease. J Am Coll Cardiol. 2011;58(8):
Physicians, American Association for Thoracic Surgery, Preventive 849–60.
Cardiovascular Nurses Association, Society for Cardiovascular 23. Pepine CJ, et al. Emergence of nonobstructive coronary artery dis-
Angiography and Interventions, and Society of Thoracic Surgeons. ease: a woman's problem and need for change in definition on angi-
J Am Coll Cardiol. 2012;60(24):e44–e164. ography. J Am Coll Cardiol. 2015;66(17):1918–33.
 Cardiac CT Radiomics
56
Márton Kolossváry and Pál Maurovich-Horvat

Medical imaging modalities have undergone great RADIOLOGICAL RADIOMIC BIG DATA INFERENCE
­improvements in the past decades. State-of-the-art scanners IMAGES FEATURES
can depict anatomical structures at previously unimaginable
detail, while also providing functional information regarding
pathologies. No matter how sophisticated our imaging tech-
niques might be, our interpretation of results is still based
Feature Datamining
predominantly on visual inspection. Qualitative assessment extraction
of radiological images discards vast amount of information,
while relying on greatly subjective and hard to reproduce Fig. 56.1  Pipeline of radiomics-based patient analysis. After image
expert opinions [1, 2]. In the era of personalized medicine acquisition, new novel radiomics-based image characteristics are
where subtle differences in the genome are seen as opportu- extracted to quantify different lesion properties. The hundreds of vari-
nities for new personalized therapies [3], superficial visual ables are joined together to create “big data” databases. Machine learn-
ing is used to find new meaningful connections between the parameters
assessment of pathologies seems outdated. and the clinical outcome data. Based on the results, new imaging bio-
Radiomics is the process of extracting numerous quantita- markers can be identified which have the potential to increase the diag-
tive image-based features from radiological examinations, to nostic accuracy of radiological examinations
create large datasets where each abnormality is characterized
by hundreds of different parameters [4, 5]. These parameters
are used to create “big data” datasets, which can be analyzed atherosclerotic lesion on CT might correlate with known
using machine learning techniques to find new meaningful ­histological morphologies and thus could give additional
patterns and relationships in the data (Fig. 56.1). ­prognostic information above stenosis severity.
Several qualitative imaging markers representing coro-
nary plaque heterogeneity have been identified on coronary
Concept of Radiomics CT angiography (CTA). These have been shown to correlate
with histological abnormalities and major adverse cardiac
In radiological images, each voxel represents a separate mea- events (MACE) to some degree [6, 7]. Being qualitative
surement of some physical characteristic in that given area or markers, they are prone to interobserver variations and
volume. Basically, radiological examinations are a vast num- require much experience. Furthermore, there are aspects of
ber of physical measurements done in a spatial coordinate heterogeneity which are difficult to comprehend by visual
system. In CT, the amount of radiation absorption in the inspection. Subtle differences in HU values are undetectable
given voxel is utilized to create the image (for details see by eye sight, while they might hold valuable information of
Chap. 2). Different tissue components all absorb radiation to underlying tissue architecture. Much of radiomics is about
a different extent; thus the Hounsfield units (HU) mirror the mathematically quantifying these fine textural properties,
underlying tissue structure of a given voxel. Therefore, it is shape of the lesion, and the spatial distribution of HU values
reasonable to assume that spatial heterogeneity of a coronary on radiological images.

M. Kolossváry (*) · P. Maurovich-Horvat

Cardiovascular Imaging Research Group, Heart and Vascular
Center, Semmelweis University, Budapest, Hungary
e-mail: [email protected]

© Humana Press 2019 715

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_56
716 M. Kolossváry and P. Maurovich-Horvat

Classification of Radiomic Techniques  arameters Representing the Shape

P
of the Distribution
Radiomic techniques can be divided into four major groups: The concept of distribution shape is grasped using so-called
(1) intensity-based metrics, which measure different proper- moments. Moments are defined as the average of the values
ties of the distribution of HU values; (2) texture-based analy- (xi) minus a given value (c) raised to a given power (q). If
sis, which quantifies the spatial heterogeneity of the lesion; c = 0 and q = 1, the moment equals the mean. If c = the mean
(3) shape-based measures, which quantify the three-­ (μ), we call our moments central moments. If c  =  μ and
dimensional geometry of an abnormality; and (4) transform-­ q = 2, then we obtain the variance, which informs us about
based metrics, which transform the image from the spatial how spread-out our data is from the mean. The square root of
domain to the frequency or scale domain. Summary of the variance is the standard deviation (SD), which also quan-
radiomic techniques can be found in Fig. 56.2. tifies the spread of the data, and in cases of normally distrib-
uted data, it gives information where approximately 68% of
the data is situated around the mean. If we divide our central
Radiomic Features moments with the SD on the qth power, then we obtain stan-
dardized or normalized central moments. If c = μ and q = 3
Intensity-Based Metrics and we standardize by SD3, then we define skewness, which
quantifies the asymmetry of the distribution. Negative skew
Intensity-based metrics are often also called first-order means the left tail is longer and a considerable proportion of
­statistics, meaning that statistics are derived from individual the distribution is shifted to the right, while positive skew
voxels themselves, discarding all spatial and relational indicates just the opposite. If c = μ and q = 4 and we normal-
information. These metrics can easily be calculated by ize by SD4, we come to the kurtosis, which basically quanti-
­
­segmenting out the given voxels and then calculating the sta- fies the number of outliers in the data. The smaller the value,
tistics on the array of voxel values. This is also referred to as the more data points are in the proximity of the mean, while
histogram analysis, since the calculated statistics resemble higher values indicate that a larger proportion of the data
some property of the distribution of the voxel values. points are located beyond one SD distance of the mean, as
compared to the normal distribution. Higher degree moments
Parameters Representing the Average can also be defined but are seldom used due to exhaustive
and Spread of the Data calculations and difficult interpretation.
The following statistics are the most common metrics used in
practice: mean which is the closest point to all other points, Parameters Representing the Diversity of Values
median which is the middle element of an ordered array of While previous statistics quantified the spread and the shape of
values, minimum and maximum which reflect the most extreme the distribution, they do not give any information regarding the
points in the distribution, percentiles which are cutoff points diversity of the data, meaning how dissimilar the data points
for a given proportion of the distribution, and interquartile are. Voxel values can also be viewed as signals with different
range, which are two distinguished data points, that define the intensities. Concepts from signal processing—a subfield of
25th and the 75th percentile, which indicate the spread of mid- information theory—can be implemented to quantify heteroge-
dle 50% of the data points. While previous statistics represent neity of voxel values. Energy is the sum of all squared elements
information regarding the values and the range of the distribu- and quantifies the overall magnitude of voxel intensities in the
tion, they do not express any information regarding the shape image. Uniformity is the sum of squared probabilities of each
of the distribution. There are an infinite number of different given voxel value. If there is only one value present, then uni-
distributions which have the same mean or interquartile range formity is 1; the more d­ ifferent values there are, the smaller the
but have very different shapes (Fig. 56.3). probabilities of each value; thus the squared sum will be closer

Fig. 56.2  Classification of Radiomics

different radiomic techniques.
(From Kolossvary et al. [5],
with permission) INTENSITY-BASED TEXTURE-BASED SHAPE-BASED TRANSFORM-BASED
METRICS METRICS METRICS METRICS

Average and spread Second-order statistics 1, 2, 3-Dimensional Fourier transform

mean, SD, etc. GLCM length, area, etc.
Shape Higher-order statistics Minkowski functionals Gabor transform
skew, kurtosis, etc. GLRLM, GLSZM, etc. area, genus, etc.
Diversity Laws’ texture energy Fractal dimension Wavelet transform
entropy, energy, etc. box-counting, etc.
56  Cardiac CT Radiomics 717

Fig. 56.3  Pipeline for

calculating first-order CORONARY HISTOGRAM STATISTICS
statistics on two SEGMENTATION ANALYSIS
representative examples of
coronary lesions. First the
coronary arteries need to be

Relative frequency
segmented. Then histograms Mean 273.8
need to be created showing SD 189.3
the relative frequency of given
Skewness -0.37
HU values. From these
different statistics can be Kurtosis 3.14
calculated. The image also
justifies the use of several
different parameters to reflect Hounsfield unit
a lesion, since the average
attenuation values and the
standard deviations are the

Relative frequency
same, while only higher Mean 273.8
moments can differentiate
SD 189.3
between these two plaques.
(From Kolossvary et al. [5], Skewness 0.73
with permission) Kurtosis 1.91

Hounsfield unit
Outer vessel wall
Inner vessel wall

to 0. Shannon entropy is a measure to quantify the amount of Agatston score [9]. The calcium score of a lesion is com-
information contained in the data [8]. In cases where there are puted by calculating the area of voxels with attenuation val-
two outcomes and both are equally probable (e.g., a coin toss), ues greater than 130 HU and multiplying it by a weighting
then we are at maximum uncertainty about the outcome of a factor based on the maximum attenuation value of the given
toss, since both occur with the same probability; thus the unpre- lesion (for details see Chap. 2). This simple idea of taking the
dictability is at maximum. However, if one outcome is more area of calcium in a lesion and multiplying it by a number is
probable than the other (biased coin), then we have some one of the best cardiovascular risk stratification methods we
assumption about what the outcome might be, since one of the have to date [10, 11]. It is also feasible to quantify volumetric
events is more probable; thus our uncertainty is smaller. This quantities of plaque components using coronary CTA [12].
shows that higher probability (p) events carry less information, Unfortunately, there is only limited information regarding
because we could have guessed what might happen, while rare the prognostic value of different quantitative plaque compo-
events carry more information, since their occurrence is scarce nents on MACE. Versteylen et al. have shown total plaque
and highlights only a few instances. Entropy quantifies this volume, total noncalcified volume, per-plaque maximal vol-
uncertainty or randomness, by weighting the information con- ume, noncalcified percentage, and plaque burden to be all
tent of an event (which is defined as log2p) with its relative significantly larger in acute coronary syndrome patients, as
frequency in the dataset and then adding up all these values for compared to non-acute coronary syndrome controls [13].
all the events and multiplying it by minus one. The result is Furthermore, adding quantitative plaque characteristics to
called the commonly used term: bits. It quantifies the minimum the Framingham risk score (FRS) and qualitative CTA read-
number of required binary elements to describe all possible ing results significantly improves the diagnostic performance
entries in the data. Higher values indicate that more bits are of the model, as compared to FRS and qualitative informa-
required to describe the data suggesting higher uncertainty tion (area under the curve: 0.64 vs. 0.79, p  <  0.05,
meaning greater heterogeneity, while smaller values indicate respectively).
just the opposite. Overall, high SD and entropy and low unifor- As attractive it may seem to measure quantitative met-
mity all represent increased heterogeneity. rics regarding coronary plaques, there are major potential
pitfalls that we must not forget. These metrics are all
 tilization and Potential Pitfalls of Intensity-­
U based on absolute HU values. It has been shown in ex vivo
Based Metrics studies that using a different CT scanner for the measure-
Several of the abovementioned metrics have been used to ments changed the observed median Agatston score sig-
quantify coronary lesions. One of the first attempts to utilize nificantly (Philips, 353; Toshiba, 410; GE, 469; Siemens,
quantitative information from cardiac CT scans was the 332; p  <  0.05). Furthermore, the limited inter-platform
718 M. Kolossváry and P. Maurovich-Horvat

reproducibility resulted in up to 6.5% of the intermediate- Second-Order Statistics

risk patients reclassified into either low- or high-risk Haralick et al. proposed the idea of gray-level co-occurrence
groups [14]. Moreover, using different image reconstruc- matrix (GLCM) [22]. GLCMs or gray-tone spatial depen-
tion algorithms on the same examination also had a sig- dencies matrix (GTSDM) is a second-order statistics mean-
nificant effect on observed values and risk prediction [15]. ing that the statistics calculated from these values are based
Same tendencies can be observed for coronary CTA scans on the relationship of two voxels. The main idea is not to
as well. It seems there is also a considerable amount of look at intensity values themselves, but to examine how
inter-vendor variation [16], and higher levels of iteration many times similar intensity values occur next to each other
during image reconstruction may also have an effect on at a given distance and direction. To achieve this, first our
observed values [17, 18]. Furthermore, there is no consen- image values must be discretized, meaning that they need to
sus on the cutoff values for identifying different tissue be partitioned into n non-overlapping groups. Next raw
components, which is mainly due to the aforementioned GLCMs are created. For simplicity, we will continue assum-
reasons, and because of different plaque components, HU ing we have a 2D image, for example, a cross section of a
ranges overlap significantly, due to the similar molecular lesion, not a 3D dataset. To derive a GLCM, two parameters
composition and limited spatial- and low-contrast resolu- need to be given: distance and direction. Distance refers to
tion of current CT scanners [19, 20]. Therefore, several how far away we look from our reference voxel, for example,
different cutoff values have been proposed [12, 21]. In the voxels next to a reference voxel are at a distance of 1.
addition, these are all population-based results; thus indi- Direction is usually an angle, where the voxels to the east of
vidual variations can be significantly larger. Furthermore, a reference voxel are at an angle of 0°, ones to the northeast
most of these statistics are related to the mean, which is a are at 45°, ones to the north are at 90°, and ones to the north-
very robust statistic since it is calculated using each voxel west are at 135°. As we will see, there is no need to define the
value. Therefore, if such alterations can be observed aver- remaining four directions, since they can be derived from
aging hundreds of thousands of voxel values, then indi- one of the earlier defined directions.
vidual voxel values show considerably larger variations, In a GLCM there are as many rows as there are column,
which would have a great effect on personalized patient which is equal to the number of intensity groups (n) to which
management based on these metrics. we discretized our image. The rows are labeled as i, and the
Overall, intensity-based metrics can quantify the distribu- columns as j. A value in the ith row and jth column (location
tion and heterogeneity of our data. However, there are major defined as [i, j]) in the raw GLCM enumerates how many
concerns, due to the lack of quantification standards, the con- times in our image a voxel of value j is located at a given
cerns of reproducibility, the effects of noise, etc. Finally, distance and in a certain direction with respect to any one of
these metrics all discard spatial information; therefore mea- the voxels in the image that have a value of i. For example, in
sures of diversity only reflect properties of the distribution, a raw GLCM where distance is 1 and the angle is 0°, the
which has nothing to do with spatial heterogeneity. value in the 3rd column and the 2nd row is equal to the num-
ber of times a value of 3 occurs to the right of a value of 2 in
the whole image. As stated earlier, one needs to only calcu-
Texture-Based Metrics late these GLCMs in four directions since the other four can
be derived from the opposite GLCM by transposing the
Earlier metrics discarded the spatial relation of the voxels matrix, which means we simply mirror the values on the
to each other. This seems conceptually flawed since we main diagonal, which goes through the matrix from the top-­
know plaque composition has a significant effect on plaque left to the bottom-right. For example, the value in the ith row
vulnerability, and plaque composition is expressed by the and jth column in a GLCM of distance 1 and angle 0° is
spatial interrelationship of the voxels on coronary equal to the value in the jth row and ith column in the GLCM
CTA. Even so, it is very hard to conceptualize this spatial of distance 1 and an of angle 180°. This is because it is equiv-
interaction between the image points. This problem first alent to ask how many times a value j occurs to the right of
emerged from the analysis of satellite imagery in the 1970s value i and to ask how many times does value i occur to the
and is referred to as texture analysis. Texture is a broad left of j. Therefore, we can simply add up these correspond-
concept, which tries to describe patterns in an image. ing GLCMs to receive four symmetrical (value in ith row and
Patterns are basically repetitions of different characteris- jth column equals the value in jth row and ith column)
tics, such as shape, color, intensity, etc. Texture analysis ­matrices. Since the absolute number of occurrences is not
tries to mathematically define these ideas by calculating very informative, because we do not know whether, for
different statistics which are based on the spatial relation- example, 42 is a large value or not, we change the values in
ship of two or more image points and not simply on the the raw GLCM to relative frequencies, which we get by sim-
voxel values themselves. ply dividing each entry by the sum of the entries.
56  Cardiac CT Radiomics 719

These matrices already tell us a lot of information. The calculated by multiplying the elements of the GLCM by a
values on the main diagonal represent the probability of two weight and then summing all the values. The weight for a
voxels having the same value at a given direction and angle. value in the ith row and jth column of the GLCM equals the
The further away we go from the diagonal, the greater the difference between the intensity i and j squared. This empha-
difference between the two voxels. Two extremes would be a sizes values where there is a large difference in the intensities
matrix with only elements on the diagonals, which means of neighboring voxels, because elements on the diagonal
that there are only identical intensity values in the given (which means the two pixels have the same intensity) receive
direction in our image. On the other extreme, if each entry in a weight of 0, while the further away we move from the main
the GLCM is the same, then our intensity values occur at diagonal (where there are bigger differences between the two
random in our picture in the given direction. voxels), the higher the weights, which increases exponen-
Haralick originally proposed 14 different metrics that can tially. Therefore, higher values of contrast indicate bigger
be derived from the GLCMs, but many more are used [23]. differences between the voxels at a given direction and dis-
Basically, all parameters are derived by weighting the entries tance, while smaller values indicate uniformity between the
of the matrix based on some objective and then summing voxel values. Homogeneity/inverse difference moment is just
them up. These values are the following, but not limited to: the opposite of contrast. We want to express the amount of
angular second moment/uniformity/energy is calculated just uniformity in the image; thus we want to give bigger weights
as is the first-order case, by squaring the elements of the to elements close to the main diagonal of the GLCM and
GLCM and then summing them up. Values close to 1 indi- smaller weights to values farther away from it. To achieve
cate few highly probable values in the GLCM, which means this, we take the same weight as before and multiply the val-
that the same two values occur usually next to each other, ues by the reciprocal of this weight. Since the denominator
while a uniformity close to 0 would mean that different val- would be 0 for elements on the diagonal, we add 1 to it for all
ued voxels occur next to each other at random. Contrast is weights. Pipeline for deriving GLCMs is shown in Fig. 56.4.

CORONARY EXTRACTION OF DISCRETIZATION

SEGMENTATION VOXELS

5 2
3 2 2 2
1 2 3 3 3 4
3 5 4 4
1 3 2 4
1 3 3
1 2 4

Outer vessel wall

Inner vessel wall

0.00 0.05 0.05 0.00 0.00 0 2 2 0 0 0 2 2 0 0

0.05 0.11 0.08 0.05 0.03 2 4 3 2 1 0 2 1 2 0
0.05 0.08 0.16 0.03 0.03 2 3 6 1 1 0 2 3 1 1
0.00 0.05 0.03 0.05 0.03 0 2 1 2 1 0 0 0 1 0
0.00 0.03 0.03 0.03 0.00 0 1 1 1 0 0 1 0 1 0

NORMALIZED GLCM GLCM RAW GLCM

Fig. 56.4  Pipeline for calculating gray-level co-occurrence matrices and jth column of the raw GLCM. To achieve symmetry, the transpose
(GLCM). First the coronary arteries need to be segmented. Then the is added to the raw GLCM. Next, the matrix is normalized by substitut-
voxels need to be extracted from the images. Next the images need to be ing each value by its frequency; this results in the normalized
discretized into n different value groups. Then a given direction and GLCM.  Afterward, different statistics can be calculated from the
distance is determined to calculate the GLCM (distance 1, angle 0°). GLCMs. To get rotationally invariant results, statistics are calculated in
Raw GLCMs are created by calculating the number of times a value j all four directions and then averaged. (From Kolossvary et al. [5], with
occurs to the right of value i. This value is then inserted into the ith row permission)
720 M. Kolossváry and P. Maurovich-Horvat

These and several other measures can be calculated from emphasizing long runs. Gray-level nonuniformity highlights
the GLCMs. The values are a function of the angle at which the dissimilarity of the voxels in the picture. It is calculated
we examined neighboring voxels. It is rational to suppose by adding up the number of different run lengths for a given
that texture is rotationally invariant. If we rotate our image intensity value and then squaring and summing them. If the
by 90°, we should still get the same value for contrast or runs are equally distributed throughout the HU values, then
homogeneity. Texture is an intrinsic property of the image; it the value is minimal, while unevenly distributed gray-level
cannot have high and low contrast at the same time. To solve intensities create large values. Run length nonuniformity
this discrepancy, the statistics calculated in all four direc- adds up the number of times the given run length occurred
tions are averaged. for all gray values and then squares and adds them up. Instead
of measuring intensity nonuniformity, this measures run
Higher-Order Statistics length nonuniformity. These metrics are usually divided by
While second-order statistics looked at the relationship of the total number of runs to normalize it. The last statistics,
two voxels, higher-order statistics assess the relationship of run percentage is the ratio of the total number of runs to the
three or more voxels. Galloway proposed the idea of gray-­ number of voxels. It has the lowest value in cases of linear
level run length matrix (GLRLM), which not only looks at structures, where we have few long runs. Pipeline for deriv-
pairs of voxels, but rather assesses how many identical-value ing GLRLMs can be found in Fig. 56.5.
voxels occur in a given direction next to each other [24]. Several other concepts have also been introduced, such as
Similarly, we can create such matrices in four different direc- the gray level gap length matrix (GLGLM), which is similar
tions in 2D.  The rows of the GLRLM represent again the to GLRLMs, but instead of quantifying how long same value
given attenuation values, while the columns represent the voxels occur, it calculates how many voxels we need to travel
given run lengths. Thus, an entry x in the ith row and jth col- in a direction to find a voxel with the same value [25]. The
umn would mean that in the image it occurs x times that i gray-level size zone matrix (GLSZM) is also similar to
intensity voxels are next to each other j times in the given GLRLM but quantifies how many voxels of the same value
direction. are connected to each other, irrespective of direction [26].
Galloway proposed five easy statistics that can be calcu- The neighborhood gray-tone difference matrix (NGTDM)
lated from these matrices, but many more exist [23]. Short tries to mimic human perception mechanisms by including a
runs emphasis takes each value in the matrix and divides it whole neighborhood of voxels for calculating statistics [27].
by its run length squared and then adds up all the values; thus Methods to eliminate the problem of image discretization
short runs are emphasized, while long run lengths are dimin- have also been proposed, such as the multiple gray-level size
ished. Long runs emphasis does just the opposite; it multi- zone matrix (MGLSZM), which calculates the metrics for
plies each value of the matrix by its run length squared several different gray-level quantizations, which are summed
instead of dividing it and then sums all the values, thus using a weighted average [28].

CORONARY EXTRACTION OF DISCRETIZATION GLCM

SEGMENTATION VOXELS

5 2
3 2 2 2 4 0 0
1 2 3 3 3 4
3 5 4 4 4 0 1
1 3 2 4 3 1 1
1 3 3 3 1 0
1 2 4 2 0 0

Outer vessel wall

Inner vessel wall

Fig. 56.5 Pipeline for calculating gray-level run length matrices times a i value voxels occur next to each other in the given direction. The
(GLRLMs). First the coronary arteries need to be segmented. Next the ith row and jth column of the GLRLM represent how many times it
voxels need to be extracted from the images. Then the images need to be occurs in the image and that i value voxels are next to each other j times.
discretized into n different value groups. Next a given direction (angle To get rotationally invariant results, the statistics calculated in different
0°) is determined. GLRLMs are created by calculating the number of directions are averaged. (From Kolossvary [5], with permission)
56  Cardiac CT Radiomics 721

 aws’ Texture Energy Measures

L Shape-Based Metrics
Laws proposed a different method for quantifying different
texture properties [29]. Instead of calculating metrics based Coronary lesions are complex 3D objects. As the ratio and
on the relationship of neighboring voxels, he suggested filter- the spatial distribution of different plaque components
ing the image for specific structures. This is done by convolu- change, also does the shape of the lesion. Shape is an abstract
tion, which simply means that we replace each pixel in the concept trying to quantify different geometrical properties of
image with a weighted value of its neighbors. The values that objects. Shape-based metrics can be grouped based on the
we use as weights for filtering are stored in the kernel matrix. dimension in which we do our measurements.
The convolution technique is used in daily practice to recon-
struct images with sharper or smoother kernels. The sharp  ne-Dimensional (1D) Metrics
O
kernel improves spatial resolution, enhances edges, and 1D metrics are basically based on the diameters of the lesion.
reduces metallic or blooming artifacts (e.g., stents, heavy cal- The diameter can be measured on the longest axes or the
cification). A sharper kernel generates images with higher shortest axes or in any arbitrary direction. They can also be
spatial resolution but increases the image noise, whereas a compared to other entities, as in the case of diameter steno-
smoother kernel generates images with lower noise but with sis, where the diameter of the lesion is divided by the vessel
reduced spatial resolution. diameter on cross-sectional images to assess stenosis sever-
If we choose our kernel values carefully, we can empha- ity on coronary CTA. Furthermore, from these values many
size different textures, such as edges or ripples, while cancel- different metrics can be calculated, such as the ratio of lon-
ing out other effects. Laws proposed five one-dimensional gest diameter to the shortest diameter, which is close to 1 in
filters which can be combined to achieve two- or three-­ case of cylindrical objects and gets bigger as the abnormality
dimensional complex filters. Through convolution of our becomes elliptic, or the difference between the longest and
original image and the kernel, we receive a new image, the shortest diameter which also quantifies roundness.
which we can then analyze, usually calculating the energy as
proposed earlier in the chapter.  wo-Dimensional (2D) Metrics
T
2D metrics are based on cross sections of the lesion. From
 tilization and Potential Pitfalls of Texture-Based
U these images several different parameters can be calculated.
Metrics These are mainly based on different areas or on different bor-
Texture-based methods have been implemented to classify der lengths. In most cases these 2D metrics are used to
tumor heterogeneity [30], but cardiac implementations are approximate more complex 3D metrics, since they are sim-
scares. This is mainly due to several difficulties specific to pler to calculate. For example, lesion area is a measure used
cardiac imaging that are less prominent in tumors. to describe the size of a lesion. Most common in case of
Atherosclerotic lesions are very small; therefore only a coronary CTA is the area stenosis, which is the ratio of the
limited number of voxels depict the abnormality. These vessel wall area to the vessel area. This metric is commonly
statistics are robust in cases when there are a sufficient used to assess positive remodeling.
number of data points. While the submillimeter resolution
of high-tech scanners is adequate for clinical decision-  hree-Dimensional (3D) Metrics
T
making and even volumetric analysis, it might not be 3D metrics aim to quantify different geometrical properties
enough for texture analysis. Furthermore, while anatomi- of the lesion. Many of these concepts originate from rigid
cal planes are suitable for heterogeneity analysis of tumors, body mechanics. All 3D objects have three so-called princi-
since they are directionally invariant, coronary arteries are pal axes or eigenvectors, which are mutually perpendicular
complex 3D structures which have a specific direction; to each other and intersect at the center of mass. Torque
thus simple anatomical slices are inadequate for analysis. applied around these principal axes acts independently,
Finally, while a 2D plane can resemble the heterogeneity meaning that if we push or rotate our object along one prin-
of a tumor, it may not be sufficient for atherosclerotic cipal axis, it will not rotate in any other direction. These
lesions, since they are complex 3D structures, where het- eigenvectors also have corresponding eigenvalues, which
erogeneity needs to be assessed in 3D, on the course of the can be interpreted as weights proportional to the amount of
coronary artery. Despite these potential limitations, there mass or HU intensities located in that direction.
are promissing results of appling radiomics to identify vul- Several different parameters can be derived from the
nerable plaques from coronary CTA images. Kolossvary et eigenvectors, such as: asymmetry is defined as 1 minus the
al showed that texture-based radiomic statistics signifi- square root of the ratio of the smallest and the largest eigen-
cantly outperformed conventional volumetric quantitative values. In cases where HU values are distributed equally in
parameters to identify napkin-ring sign plaques from coro- all directions of the principal axes, then asymmetry is 0,
nary CTA images [23]. while in cases where the distribution is imbalanced in the
722 M. Kolossváry and P. Maurovich-Horvat

three directions, then it is close to 1. Compactness is a scaled sion is the easiest, we simply calculate how many voxels
product of the three eigenvalues divided by the volume of the contain the image, and then we repeat this many times with
lesion. The less spread-out our lesion is, the higher the a finer grid. We then plot the number of occupied voxels
­compactness. Roundness is the difference of the biggest versus the reciprocal of the scale on a log-log plot. The
­eigenvalue of the smallest enclosing and the largest enclosed ­box-counting dimension will be equal to the slope of the
ellipse. It describes how closely similar the lesion is to an ­regression line that connects the points. The information
ellipse. Small values indicate similarity to an ellipse, while and the correlation dimension can also be calculated from
larger values indicate more rounded shape. Rényi dimensions [34, 35].

Minkowski Functionals  tilization and Potential Pitfalls of Shape-Based

U
Minkowski functionals originate from the area of integral Metrics
geometry [31] and quantify different properties of com- To date, mainly the effect of 1D parameters on later MACE
plex geometrical structures. They can be implemented to rates has been investigated. Conventionally, stenosis severity
calculate characteristics on 2D and 3D images from the is expressed by diameter stenosis in the daily routine. Several
number of connected and non-connected voxels, edges, studies have shown the prognostic value of diameter stenosis
and vertices. The results of some functionals are com- on later MACE rates, but long-term effects still remain a
monly known parameters such as area and circumference question [36]. Furthermore, lesion length has also been
in 2D or volume and surface area in 3D [32]. Furthermore, shown to effect later outcomes in patients undergoing percu-
the Euler characteristic or genus can be calculated which taneous coronary intervention [37]. Most commonly used
can be seen as a connectivity parameter representing the 2D shape-based metric is the cross-sectional area, which is
number of separate voxel groups containing a signal used to calculate positive remodeling, which has been shown
minus the number of completely enclosed regions where to be a vulnerability feature of rupture-prone plaques [38].
there is no signal. These values can be calculated on sev- The main concern with shape-based metrics trying to
eral images derived from the original image using differ- quantify some geometrical aspect of a coronary lesion is that
ent thresholds to receive an array of different values which plaques grow along the coronary arteries, which have com-
can be statistically analyzed. plex 3D shapes. The difficulty is that the length of a lesion is
not equal to the distance measured between the origin and
Fractal Dimension the end of lesion on the image, but instead equal to the dis-
Fractal geometry examines the self-similarity of objects. tance along the coronary artery between the two points.
Fractals exhibit expanding symmetry by repeating patterns Therefore, if we measure shape-based metrics not account-
that are present at each scale. Conventional objects showing ing for coronary geometry, then we are also including geo-
no fractal geometry scale their characteristics by a scalar on metrical information regarding the arteries themselves. This
the power of the dimension meaning: if we scale a 1D object, might not be a problem, since coronary geometry and shape-­
for example, a line by two, then the length of the line will based metrics can be very much interrelated, but little is
increase by 21. If we double the edge length of a 2D polygon, known how they might affect each other. Finally, complex
for example, a square, then the area will increase by 22. If we shape-based radiomic features require complex computa-
scale the radius of a 3D object such as a sphere, then its vol- tional algorithms which are currently not implemented in
ume will increase by 23. For such geometrical objects, the clinical software.
fractal dimension is equal to the topological dimension and it
is an integer. However, fractals behave differently. For exam-
ple, a line whose topological dimension is 1 can have a frac- Transform-Based Metrics
tal dimension anywhere between 1 and 2. If we were to scale
this line by 2, then its length would not be twice as long. Images are two- or three-dimensional datasets where the pix-
Actually, its length is undefined, since if we were to enlarge els are positioned in the spatial domain. Transform-based
our image, then we would see more and more detail which metrics convert these images from the spatial domain to a
affects the length of the line, but this continues infinitely as different coordinate system, for example, the frequency
we magnify our image. domain, which uses information taken from the rate at which
Fractal dimension is a measure for quantifying the spa- image intensity values change to describe the image, instead
tial complexity of an object and is proportional to how the of using spatial coordinates for this purpose.
detail of the image changes as we alter the scale [33]. Rényi
dimensions generalize the concept of fractal dimensions Fourier and Gabor Transforms
from which several different definitions can be gained. The Fourier and Gabor transforms originate from signal process-
box-­counting dimension or Minkowski-Bouligand dimen- ing. Fourier transform takes a 1D signal and decomposes it
56  Cardiac CT Radiomics 723

to a series of sinusoidal waves. Applying 1D Fourier trans- fields such as computer vision and data science. Radiomics
forms in both the x and the y imaging planes, we can trans- together with genetics, proteomics, and metabolomics has
form our image to a 2D frequency domain. Gabor transform the potential to change our understanding of human diseases.
is similar to the Fourier transform but first filters the image We are at the dawn of a new era, where illnesses are no
by applying a Gaussian filter. This can be useful for image ­longer represented by a single value, but by a myriad of
analysis, since the image can be filtered from specific fre- ­parameters. Thus, medical professionals need to get com-
quencies such as noise, or different characteristics can be fortable with relying more and more on quantitative data,
emphasized on the picture, for instance, edges. Using the rather than our medical assumptions. If we are unwilling to
inverse Fourier transform, the original image can be recon- join this transition, we might be left behind.
structed from these sine and cosine frequencies. After trans-
forming back our images, we can extract different statistics
from our filtered images, such as texture-based metrics men- References
tioned earlier.
1. Jones BK, Buckwalter AJ, McCarthy FE, DeYoung RB, El-Khoury
YG, Dolan L, Gannon HF, Inwards YC, Klein JM, Kyriakos M,
Wavelet Transforms McCarthy FE, Rosenberg EA, Siegal PG, Unni KK, Fayad L,
While Fourier transform can be used to analyze the fre- Kransdorf JM, Murphey DM, Panicek MD, Rubin AD, Sundaram
quency component of a signal or an image, we lose all our M, Vanel D. Reliability of histopathologic and radiologic grading
spatial information. Wavelet transform on the other hand not of cartilaginous neoplasms in long bones. J Bone Joint Surg Am.
2007;89(10):2113–23. https://doi.org/10.2106/JBJS.F.01530.
only contains frequency information but also includes spatial 2. Rutkow IM. Surgical decision making. The reproducibility of clini-
information as well [39]. Basically we can sacrifice fre- cal judgement. Arch Surg. 1982;117(3):337–40.
quency precision to gain spatial information. By setting the 3. Barabasi AL, Gulbahce N, Loscalzo J.  Network medicine: a
values of two parameters, we can weight frequency over network-based approach to human disease. Nat Rev Genet.
2011;12(1):56–68. https://doi.org/10.1038/nrg2918.
temporal information or vice versa. Therefore, a family of 4. Gillies RJ, Kinahan PE, Hricak H.  Radiomics: Images Are More
transformed pictures can be obtained from which statistical than Pictures, They Are Data. Radiology. 2016;278(2):563–77.
parameters can be calculated. https://doi.org/10.1148/radiol.2015151169.
5. Kolossvary M, Kellermayer M, Merkely B, Maurovich-Horvat
P.  Cardiac computed tomography radiomics: a comprehensive
 tilization and Potential Pitfalls of Transform-­
U review on radiomic techniques. J Thorac Imaging. 2017;33(1):26–
Based Metrics 34. https://doi.org/10.1097/RTI.0000000000000268.
Fourier and Gabor transforms are extensively used for image 6. Feuchtner G, Kerber J, Burghard P, Dichtl W, Friedrich G, Bonaros
filtering to increase quality. Their main application is not to N, Plank F.  The high-risk criteria low-attenuation plaque <60
HU and the napkin-ring sign are the most powerful predictors
produce quantitative parameters for image classification; of MACE: a long-term follow-up study. Eur Heart J Cardiovasc
thus their utilization as a classifier is scarce. Wavelet trans- Imaging. 2016. https://doi.org/10.1093/ehjci/jew167.
forms are becoming increasingly popular since they incorpo- 7. Otsuka K, Fukuda S, Tanaka A, Nakanishi K, Taguchi H, Yoshikawa
rate spatial and frequency information as well. With the J, Shimada K, Yoshiyama M.  Napkin-ring sign on coronary CT
angiography for the prediction of acute coronary syndrome. JACC
tuning of the parameters, infinite number of statistics can be Cardiovasc Imaging. 2013;6(4):448–57. https://doi.org/10.1016/j.
derived from them. This poses a potential problem, since the jcmg.2012.09.016.
selection of high accuracy classifiers from numerous poten- 8. Shannon CE.  A mathematical theory of communica-
tial parameters is prone to overfitting. tion. Bell Syst Tech J. 1948;27(3):379–423. https://doi.
org/10.1002/j.1538-7305.1948.tb01338.x.
9. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M
Jr, Detrano R. Quantification of coronary artery calcium using ultra-
Conclusions fast computed tomography. J Am Coll Cardiol. 1990;15(4):827–32.
10. McClelland RL, Jorgensen NW, Budoff M, Blaha MJ, Post WS,
Kronmal RA, Bild DE, Shea S, Liu K, Watson KE, Folsom AR,
Radiology has undergone unprecedented developments over Khera A, Ayers C, Mahabadi AA, Lehmann N, Jockel KH, Moebus
the past years. While medical images were only seen as S, Carr JJ, Erbel R, Burke GL. 10-Year coronary heart disease risk
rough representations of underlying pathologies in the past, prediction using coronary artery calcium and traditional risk factors:
the improvement of spatial and temporal resolution all leads derivation in the MESA (Multi-Ethnic Study of Atherosclerosis)
with validation in the HNR (Heinz Nixdorf Recall) study and the
to increased image quality allowing us to view radiologists DHS (Dallas Heart Study). J Am Coll Cardiol. 2015;66(15):1643–
as in vivo pathologists. This also means that radiologists are 53. https://doi.org/10.1016/j.jacc.2015.08.035.
subjected to vast amounts of information, to the point that 11. Polonsky TS, McClelland RL, Jorgensen NW, Bild DE, Burke
might not be comprehensible by humans. Radiomics has the GL, Guerci AD, Greenland P. Coronary artery calcium score and
risk classification for coronary heart disease prediction. JAMA.
potential to vastly increase our knowledge of the underlying 2010;303(16):1610–6. https://doi.org/10.1001/jama.2010.461.
pathological processes, but this requires new specialists who 12. Brodoefel H, Reimann A, Heuschmid M, Tsiflikas I, Kopp

are familiar with new concepts which incorporate scientific AF, Schroeder S, Claussen CD, Clouse ME, Burgstahler
724 M. Kolossváry and P. Maurovich-Horvat

C.  Characterization of coronary atherosclerosis by dual-source 24. Galloway MM.  Texture analysis using gray level run lengths.

computed tomography and HU-based color mapping: a pilot Comput Graphics Image Process. 1975;4(2):172–9. https://doi.
study. Eur Radiol. 2008;18(11):2466–74. https://doi.org/10.1007/ org/10.1016/S0146-664X(75)80008-6.
s00330-008-1019-5. 25. Xinli W, Albregtsen F, Foyn B.  Texture analysis using gray level
13. Versteylen MO, Kietselaer BL, Dagnelie PC, Joosen IA, Dedic A, gap length matrix. Paper presented at the Selected papers from
Raaijmakers RH, Wildberger JE, Nieman K, Crijns HJ, Niessen WJ, the  9th Scandinavian conference on Image analysis: theory and
Daemen MJ, Hofstra L. Additive value of semiautomated quanti- applications of image analysis II: theory and applications of image
fication of coronary artery disease using cardiac computed tomo- analysis II, Uppsala, Sweden; 1995.
graphic angiography to predict future acute coronary syndrome. J 26. Thibault G, Fertil B, Navarro C, Pereira S, Cau P, Levy N, Sequeira
Am Coll Cardiol. 2013;61(22):2296–305. https://doi.org/10.1016/j. J, Mari J-L. Shape and texture indexes application to cell nuclei clas-
jacc.2013.02.065. sification. Int J Pattern Recognit Artif Intell. 2013;27(01):1357002.
14. Willemink MJ, Vliegenthart R, Takx RA, Leiner T, Budde RP, Bleys https://doi.org/10.1142/s0218001413570024.
RL, Das M, Wildberger JE, Prokop M, Buls N, de Mey J, Schilham 27. Amadasun M, King R. Textural features corresponding to textural
AM, de Jong PA. Coronary artery calcification scoring with state-­ properties. IEEE Trans Syst Man Cybern. 1989;19(5):1264–74.
of-­the-art CT scanners from different vendors has substantial effect https://doi.org/10.1109/21.44046.
on risk classification. Radiology. 2014;273(3):695–702. https://doi. 28. Thibault G, Angulo J, Meyer F. Advanced statistical matrices for
org/10.1148/radiol.14140066. texture characterization: application to cell classification. IEEE
15. Szilveszter B, Elzomor H, Karolyi M, Kolossvary M, Raaijmakers Trans Biomed Eng. 2014;61(3):630–7. https://doi.org/10.1109/
R, Benke K, Celeng C, Bartykowszki A, Bagyura Z, Lux A, TBME.2013.2284600.
Merkely B, Maurovich-Horvat P.  The effect of iterative model 29. Laws KI. Textured image segmentation. A Dissertation Presented
reconstruction on coronary artery calcium quantification. Int J to the Faculty of the Graduate School. University of Southern
Cardiovasc Imaging. 2016;32(1):153–60. https://doi.org/10.1007/ California; 1980.
s10554-015-0740-9. 30. O'Connor JP, Rose CJ, Waterton JC, Carano RA, Parker GJ, Jackson
16. Symons R, Morris JZ, Wu CO, Pourmorteza A, Ahlman MA, Lima A.  Imaging intratumor heterogeneity: role in therapy response,
JA, Chen MY, Mallek M, Sandfort V, Bluemke DA. Coronary CT resistance, and clinical outcome. Clin Cancer Res. 2015;21(2):249–
angiography: variability of CT scanners and readers in measure- 57. https://doi.org/10.1158/1078-0432.CCR-14-0990.
ment of plaque volume. Radiology. 2016:161670. https://doi. 31. Minkowski H. Volumen und Oberfläche. Math Ann. 1903;57:447.
org/10.1148/radiol.2016161670. 32. Larkin TJ, Canuto HC, Kettunen MI, Booth TC, Hu DE, Krishnan
1 7. Precht H, Kitslaar PH, Broersen A, Dijkstra J, Gerke O,
AS, Bohndiek SE, Neves AA, McLachlan C, Hobson MP, Brindle
Thygesen J, Egstrup K, Lambrechtsen J.  Influence of adap- KM. Analysis of image heterogeneity using 2D Minkowski func-
tive statistical iterative reconstruction on coronary plaque tionals detects tumor responses to treatment. Magn Reson Med.
analysis in coronary computed tomography angiography. J 2014;71(1):402–10. https://doi.org/10.1002/mrm.24644.
Cardiovasc Comput Tomogr. 2016. https://doi.org/10.1016/j. 33. Mandelbrot B.  How long is the coast of Britain? statistical self-­
jcct.2016.09.006. similarity and fractional dimension. Science. 1967;156:636–8.
18. Fuchs TA, Fiechter M, Gebhard C, Stehli J, Ghadri JR,
34. Rose CJ, Mills SJ, O'Connor JP, Buonaccorsi GA, Roberts C,
Kazakauskaite E, Herzog BA, Husmann L, Gaemperli O, Kaufmann Watson Y, Cheung S, Zhao S, Whitcher B, Jackson A, Parker
PA. CT coronary angiography: impact of adapted statistical itera- GJ. Quantifying spatial heterogeneity in dynamic contrast-enhanced
tive reconstruction (ASIR) on coronary stenosis and plaque com- MRI parameter maps. Magn Reson Med. 2009;62(2):488–99.
position analysis. Int J Cardiovasc Imaging. 2013;29(3):719–24. https://doi.org/10.1002/mrm.22003.
https://doi.org/10.1007/s10554-012-0134-1. 35. Lopes R, Betrouni N.  Fractal and multifractal analysis: a review.
19. Pohle K, Achenbach S, Macneill B, Ropers D, Ferencik
Med Image Anal. 2009;13(4):634–49. https://doi.org/10.1016/j.
M, Moselewski F, Hoffmann U, Brady TJ, Jang IK, Daniel media.2009.05.003.
WG.  Characterization of non-calcified coronary athero- 36. Kolossváry M, Szilveszter B, Merkely B, Maurovich-Horvat

sclerotic plaque by multi-detector row CT: comparison to P. Plaque imaging with CT – a comprehensive review on coronary
IVUS.  Atherosclerosis. 2007;190(1):174–80. https://doi. CT angiography based risk assessment. Cardiovasc Diagn Ther.
org/10.1016/j.atherosclerosis.2006.01.013. 2017. https://doi.org/10.21037/cdt.2016.11.06.
20. Sun J, Zhang Z, Lu B, Yu W, Yang Y, Zhou Y, Wang Y, Fan 37. Claessen BE, Smits PC, Kereiakes DJ, Parise H, Fahy M, Kedhi
Z.  Identification and quantification of coronary atherosclerotic E, Serruys PW, Lansky AJ, Cristea E, Sudhir K, Sood P, Simonton
plaques: a comparison of 64-MDCT and intravascular ultra- CA, Stone GW.  Impact of lesion length and vessel size on clini-
sound. AJR Am J Roentgenol. 2008;190(3):748–54. https://doi. cal outcomes after percutaneous coronary intervention with evero-
org/10.2214/AJR.07.2763. limus- versus paclitaxel-eluting stents pooled analysis from the
21. Takahashi S, Kawasaki M, Miyata S, Suzuki K, Yamaura M, Ido SPIRIT (Clinical Evaluation of the XIENCE V Everolimus Eluting
T, Aoyama T, Fujiwara H, Minatoguchi S.  Feasibility of tissue Coronary Stent System) and COMPARE (Second-generation
characterization of coronary plaques using 320-detector row com- everolimus-­eluting and paclitaxel-eluting stents in real-life practice)
puted tomography: comparison with integrated backscatter intra- Randomized Trials. JACC Cardiovasc Interv. 2011;4(11):1209–15.
vascular ultrasound. Heart Vessel. 2016;31(1):29–37. https://doi. https://doi.org/10.1016/j.jcin.2011.07.016.
org/10.1007/s00380-014-0577-5. 38. Motoyama S, Kondo T, Sarai M, Sugiura A, Harigaya H, Sato T,
22. Haralick RM, Shanmugam K, Dinstein I.  Textural features for Inoue K, Okumura M, Ishii J, Anno H, Virmani R, Ozaki Y, Hishida
image classification. IEEE Trans Syst Man Cybern. 1973;SMC-­ H, Narula J.  Multislice computed tomographic characteristics of
3(6):610–21. https://doi.org/10.1109/TSMC.1973.4309314. coronary lesions in acute coronary syndromes. J Am Coll Cardiol.
23. Kolossváry M, Karády J, Szilveszter B, Kitslaar P, Hoffmann U, 2007;50(4):319–26. https://doi.org/10.1016/j.jacc.2007.03.044.
Merkely B, Maurovich-Horvat P. Radiomic features are superior 39. Dettori L, Semler L.  A comparison of wavelet, ridgelet, and

to conventional quantitative computed tomographic metrics to curvelet-­based texture classification algorithms in computed
identify coronary plaques with napkin-ring sign. Circ Cardiovasc tomography. Comput Biol Med. 2007;37(4):486–98. https://doi.
Imaging. 2017;10(12):e006843. org/10.1016/j.compbiomed.2006.08.002.
 Advanced Methods for Coronary Artery
Plaque Analysis 57
Pál Maurovich-Horvat and Udo Hoffmann

The quest to identify vulnerable patients and perhaps even Plaque Characterization
individual atherosclerotic coronary plaques that will cause
future acute coronary events remains to be one of the It has been demonstrated that vulnerable plaques harbor
greatest challenges of modern cardiovascular medicine the same morphologic characteristics as ruptured plaques,
[1–3]. Current guidelines on diagnostic strategies of with the only difference lying in the intact thin fibrous cap
patients with stable coronary artery disease (CAD) focus [8]. Vulnerable plaques are generally large and contain a
on the detection of significant coronary artery stenosis or large necrotic core covered by markedly attenuated and
myocardial ischemia [4]. Needless to say that such a strat- inflamed fibrous caps; therefore these plaques are identi-
egy will not identify the patients who die suddenly due to fied as thin-­cap fibroatheromas (TCFA) in histology [9,
acute coronary events without prior clinical manifestations 17]. Thin fibrotic cap is defined as a cap thickness of
of CAD [5]. It has been demonstrated by postmortem <65 μm, which is beyond the spatial resolution limit of the
investigations that majority of acute coronary syndromes state-of-the-art CT scanners (≈0.4 mm); therefore the non-
are caused by plaque rupture and subsequent sudden lumi- invasive morphometric assessment of the fibrotic cap is
nal thrombosis [6–8]. Rupture-prone atherosclerotic currently not feasible [8]. However, histopathologic inves-
lesions have been termed as vulnerable plaques, which tigations have revealed that vulnerable plaques prone to
have distinct characteristics from stable lesions. The dif- rupture need to grow in a sizeable extent in all three spatial
ferences in morphology and functional and metabolic dimensions [17, 18]. Furthermore, it has been shown that
characteristics provide a unique opportunity for non-inva- the necrotic core length of vulnerable plaques ranges
sive cardiac imaging to identify these lesions before they between 2 and 17  mm (mean 8  mm) and the area of the
cause acute events (Fig.  57.1) [9–11]. Moreover, it has necrotic core in 80% of vulnerable plaques is larger than
been demonstrated that on a patient level, the total quantity 1.0  mm2 [17]. These spatial dimensions are above the
of plaques (i.e., plaque burden) has a strong predictive plaque detection threshold of coronary CTA (>1  mm
value for myocardial infarction [12, 13]. Therefore, it plaque thickness) [19]. In addition, 95% of vulnerable
appears that the assessment of coronary plaque morphol- plaques are confined to the proximal coronary segments,
ogy and plaque burden is potentially very important for where coronary CTA has the best accuracy due to the large
prediction of future cardiovascular events [7, 14, 15]. vessel diameter and less frequent motion artifacts [17, 20].
Coronary computed tomography angiography (CTA) is a Therefore, the non-invasive assessment of coronary
unique imaging technique that allows for the non-invasive plaques and the identification of high-risk lesions with
depiction of the global coronary plaque burden, not just coronary CTA might be feasible. The conventional coro-
the individual plaques and luminal stenosis [16]. nary CTA plaque classification scheme is based on the
presence and quantity of calcific plaque components; thus
the following plaque categories can be differentiated: cal-
cified, partially calcified (mixed), and non-calcified plaque
P. Maurovich-Horvat (*) (NCP). Partially calcified lesions can be further divided
Cardiovascular Imaging Research Group, Heart and Vascular
into predominant calcified and predominant non-calcified
Center, Semmelweis University, Budapest, Hungary
e-mail: [email protected] plaques (Fig. 57.2). In clinical reports, the description of
plaques as “non-calcified” is preferable to “soft” or “lipid
U. Hoffmann
Department of Cardiovascular Imaging, Massachusetts General rich” as low CT density (HU) levels do not necessarily cor-
Hospital, Boston, MA, USA relate closely with pathology or ­biochemistry [21].

© Humana Press 2019 725

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_57
726 P. Maurovich-Horvat and U. Hoffmann

Fig. 57.1  The morphologic a Normal FFR Physiologic ESS Normal FFR
and functional characteristics
of stable and vulnerable
plaques. (a) Fibrocalcific
lesion with calcification and
small lipid pools considered to Lumen
be a stable plaque. The plaque
causes mild luminal narrowing
without ischemia (FFR >0.8;
green color). The ESS in the
vicinity of the plaque is in the Fibrous
normal range. (b) A thin-cap Calcium tissue Lipid Vessel wall
fibroatheroma with a large
lipid-rich necrotic core, thin
fibrous cap, neovascularization, b Normal FFR Thin fibrous cap High ESS Low ESS
spotty calcium, and presence
of inflammatory cells is a Lumen
vulnerable plaque. Despite the
positively remodeled vessel
wall at the site of the plaque, Necrotic
the lesion causes luminal core
narrowing and ischemia (FFR
Necrotic
<0.8; red color). The
core
downstream plaque region with
low ESS promotes plaque Macrophages
growth, whereas the high ESS
at the most stenotic part can ESS FFR
Vasa vasorum High 1.0
trigger plaque rupture.
proliferation
Abbreviations: ESS,
endothelial shear stress; FFR, Spotty calcium 0.9
fractional flow reserve. (From
Maurovich-Horvat et al. [11], Vessel wall 0.8
with permission)
Abnormal FFR Low 0.7

Fig. 57.2  The conventional

coronary CTA plaque
Calcified

classification scheme is based

on the presence and quantity
of calcium. The following
plaque categories can be
differentiated: calcified,
partially calcified (mixed),
Predominantly

and non-calcified plaque.

Calcified

Partially calcified lesions can

Partially Calcified

be further divided into

predominantly calcified and
predominantly non-calcified
plaques. The figure shows
Predominantly
Non-Calcified

cross-sectional images and

curved multiplanar
reconstructions of various
plaque types
Non Calcified
57  Advanced Methods for Coronary Artery Plaque Analysis 727

Low-Attenuation Plaque if they were associated with plaque progression and sug-
gested that among plaques with features of vulnerability,
Despite the challenges associated with CT density-based those with a higher degree of progression have much greater
plaque assessment, several investigators have attempted to likelihood to result in ACS [34, 35].
correlate HU measurements with plaque lipid content using
intravascular ultrasound (IVUS) or histology as reference
standard. NCPs with high CT attenuation correlated with Napkin-Ring Sign
fibrous tissue, and those with low densities correlated with
necrotic core and fibro-fatty tissue [22]. In general, plaques A histopathologic study demonstrated that among plaque fea-
with a mean attenuation of <30 HU are considered low-­ tures that are potentially accessible by non-invasive imaging
attenuation plaques and have been associated to increased techniques, the size of necrotic core (>3.5 mm2) is one of the
risk of acute coronary events [23]. However, almost all inves- best discriminators between vulnerable plaques and stable
tigators reported a substantial overlap of CT densities lesions [10, 14]. Therefore, the detection of large necrotic
between fibrous and lipid-rich plaques, which precludes the core might be a promising target for coronary CTA to risk-
reliable identification of high-risk lesions based solely on stratify individual coronary plaques. Since lipid-rich tissue is
HU measurements [22, 24]. Furthermore, it is important to associated with low attenuation in coronary CTA, plaque
note that the measurement of plaque attenuation is affected cross section with centrally located area of low attenuation
by several technical factors, such as the adjacent intraluminal might be indicative of the presence of large lipid-rich necrotic
iodinated contrast attenuation, tube voltage, slice thickness, core. In an ex vivo coronary CTA investigation, the authors
and reconstruction filter [25–28]. Despite of these limitations demonstrated that plaques with a large necrotic core have a
associated to HU measurements, low CT attenuation is a distinct attenuation pattern, which was termed as the napkin-
well-established high-risk coronary plaque feature. ring sign (NRS) [36]. The NRS is a qualitative plaque feature,
and it can be defined in a non-calcified plaque cross section
by the presence of two features: (1) a central area of low CT
Positive Remodeling attenuation that is apparently in contact with the lumen, (2)
which is surrounded by a ring-like higher attenuation plaque
The seminal work of Glagov demonstrated that the growth of tissue (Fig.  57.3) [11, 36]. A histopathologic investigation
the lipid-rich atherosclerotic plaques initially manifests with showed that the area of necrotic core was more than twice as
outward plaque expansion (i.e., positive vessel remodeling) large in NRS plaques than in non-NRS plaques (median
to preserve the coronary lumen and flow [29]. Therefore, it 1.10 mm2 vs. 0.46 mm2, p = 0.05) [37]. Furthermore the spec-
should be noted that positively remodeled plaques might be ificity of NRS to identify TCFA was excellent (94.1%) [38].
missed if assessed solely by luminal narrowing measure- Two large prospective studies demonstrated that the presence
ments, such as invasive coronary angiography. Positive of NRS is an independent predictor of future acute coronary
remodeling is associated with plaque vulnerability, abun- events (independent of the presence of positive remodeling
dance of macrophages, and increased necrotic core [30]. and low attenuation) [39, 40]. The NRS appears to be a signa-
Coronary CTA can visualize the vessel wall; therefore it is ture of plaques with large necrotic core; furthermore longitu-
well suited to measure cross-sectional diameter and area. The dinal investigations show strong prognostic value of this
remodeling index is calculated as the vessel cross-­sectional plaque feature to predict future coronary events.
area at the site of stenosis divided by the average of proximal
and distal reference areas [31, 32]. A ≥1.1 remodeling index
was suggested for the definition of positive remodeling in Spotty Calcium
coronary CTA, which corresponds to a 10% increase in cross-
sectional area at the site of the stenosis relative to the refer- Calcification is a condicio sine qua non of advanced athero-
ence sites [32, 33]. The pioneering work of Motoyama and sclerosis [41]. However, the effect of calcification on plaque
colleagues demonstrated a strong association between ACS instability is not fully understood [42–44]. Histopathology
and the presence of positively remodeled plaques as assessed studies have demonstrated that most plaque ruptures contain
by coronary CTA [23]. Furthermore, positive remodeling had some calcification; however, approximately two-thirds of
the best diagnostic performance among other high-risk plaque these are microcalcifications that are not detectable by coro-
features (low attenuation, spotty calcification) to identify the nary CTA [45]. In coronary CTA, spotty calcification is
culprit lesions in ACS patients (sensitivity 87%, specificity defined as a small (less than 3 mm), hyperdense plaque com-
88%, PPV 89%, NPV 85%) [23]. The same group demon- ponent surrounded by non-calcified plaque tissue [23]. Spotty
strated in a follow-up study using serial coronary CTA that calcifications can be classified as small (<1 mm), intermedi-
the high-risk plaques only resulted in an acute coronary event ate (1–3 mm), and large (>3 mm) calcifications, and it seems
728 P. Maurovich-Horvat and U. Hoffmann

a b c

Fig. 57.3  Cross sections of a non-calcified coronary plaque with large necrotic core. The stars indicate the necrotic core, which corre-
napkin-­ring sign. The circumferential outer rim of the plaque (red lates with the low-attenuation plaque core on the CT images, whereas
dashed line) has a higher CT attenuation in both the non-contrast (panel the outer portion of the plaque contains fibrous plaque tissue correlating
A) and contrast-enhanced (panel B) images as compared to the attenu- to the high-attenuation CT rim. Notably, the low-attenuation central
ation within the central part of the plaque. The corresponding histo- part of the plaque is apparently in contact with the lumen (panel B). L:
pathological section (panel C) demonstrates a high-risk plaque with lumen (From Maurovich-Horvat et al. [36], with permission)

that small spotty calcium shows the s­trongest association observations, a recent study showed that patients with five
with vulnerable plaques [46]. Several large studies demon- or more than five coronary segments with non-obstructive
strated that spotty calcium is a sensitive but not very specific atherosclerotic plaque have a similar risk of developing
marker of high-risk coronary plaques [23, 47–50]. adverse coronary event as patients with obstructive CAD
with less than five coronary segments involved [12].
Segment stenosis score (SSS) and segment involvement
Plaque Quantification score (SIS) are semiquantitative metrics to quantify plaque
burden [55]. The SSS is calculated by grading all coronary
Rapture-prone plaques and culprit lesions in patients who segments as 0, no stenosis/no plaque; 1, <50% stenosis; 2,
suffered ACS tend to have large size; therefore it is reasonable 50–69% stenosis; and 3, ≥70% stenosis, whereas SIS is the
to hypothesize that coronary CTA plaque quantification might number of segments that contain any plaque. Both SSS and
have an incremental value in risk stratification over traditional SIS are independent predictors of coronary events [56–59].
CTA plaque assessment [9, 11]. In line with this notion, coro- Similarly, results from the CONFIRM (Coronary CT
nary CTA studies of ACS and stable angina pectoris (SAP) Angiography Evaluation for Clinical Outcomes: an
patients demonstrated that the culprit plaques in patients with International Multicenter) registry also showed SIS to be an
ACS have larger plaque volume than stable lesions in patients independent predictor of major adverse events (hazard ratio,
with SAP [50]. In addition, patients with unstable angina 1.22; CI, 1.03–1.44) [60]. The CONFIRM risk score is a
quantitative coronary CTA analysis revealed that plaques combination of clinical cardiovascular risk, the number of
with morphological features of plaque disruption (e.g., intra- proximal coronary segments with stenosis greater than 50%,
plaque contrast dye penetration) had larger volume and con- and the number of proximal segments with partially calci-
tained more low CT attenuation plaque components as fied or calcified plaques. The combined risk score outper-
compared to plaques with no signs of disruption [51]. forms all clinical risk prediction scores and significantly
Invasive and non-invasive imaging studies have demon- improves prediction of all-cause mortality [60]. Automated
strated that coronary atherosclerosis is a dynamic process. software tools are now available for plaque quantification
Plaques change size, morphology, and composition over and characterization and have the potential to improve the
few months, gaining or losing high-risk features [52–54]. reproducibility, accuracy, and efficiency of coronary CTA-
The probability that a plaque remains unstable and eventu- based plaque burden assessment (Fig. 57.4). A recent study
ally causes a clinical event increases with the number of demonstrated that patients who developed ACS had a higher
plaques. Several longitudinal studies have demonstrated total plaque volume and total NCP volume at the baseline;
that a larger plaque burden is associated with adverse out- therefore it seems that semiautomatic plaque quantification
come [12, 13]. A sub-study of the COURAGE trial (Clinical provides additional prognostic value for ACS over both clin-
Outcomes Utilizing Revascularization and Aggressive Drug ical risk factors and traditional CT reading (including cal-
Evaluation) showed that plaque burden is a stronger predic- cium score, segment stenosis score, lesion severity, and
tor of future major adverse cardiovascular events than isch- number of segments with NCP) [61]. The accuracy of auto-
emic myocardium burden [13]. In line with these mated quantification of coronary plaque by coronary CTA
57  Advanced Methods for Coronary Artery Plaque Analysis 729

A B C D E F

5 cm

A B C D E F

Dense Calcium
Necrotic Core
Fibrous Fatty
Fibrous
Media

A B C D E F
15
Area (mm2)

10

35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130

Distance from the orifice (mm)

Fig. 57.4  Semiautomated plaque characterization and quantification indicates the lumen. The middle panel shows plaque cross sections
using a dedicated software tool. The upper panel represents a stretched along the vessel without and with color overlay. The lower panel illus-
view of a segmented left anterior descending coronary artery. trates the areas of different plaque components. The letters from A to F
The orange line indicates the outer vessel wall, whereas the yellow line represent the location of the cross-sectional images

was successfully validated against gray-scale intravascular when compared with the reference standard invasive FFR
ultrasound (IVUS) and virtual histology IVUS [22, 62, 63]. measurement [66, 67]. Importantly, FFRCT can be derived
It is important to note that automated plaque quantification from coronary CTA dataset acquired in a standard fashion,
software tools provide a good intra-platform reproducibility without the need for additional imaging, extra radiation, or
but poor inter-platform reproducibility; therefore it is rec- any medication (Fig.  57.5). It has been demonstrated that
ommended to use the same software for serial or compara- FFR identifies hemodynamically significant lesions likely to
tive ­assessments [64]. produce ischemia-related symptoms; however, it is less
clear why FFR might predict the subsequent risk for ACS
resulting from plaque rupture and coronary thrombosis [68].
Functional Plaque Assessment FFRCT The presence of a large plaque volume, large low-attenua-
and ESSCT tion plaque volume, and higher positive remodeling index
was found to be strongly predictive of reduced FFR regard-
Fractional flow reserve (FFR) derived from coronary CTA less of the degree of stenosis on multivariable analysis [69].
(FFRCT) is a promising non-invasive marker of coronary It is proposed that the presence of plaques with large necrotic
physiology [65]. The diagnostic performance of FFRCT is cores may be associated with the inability of the vessel to
superior to anatomical coronary stenosis assessment alone dilate and may predispose to ischemia and abnormal FFR
730 P. Maurovich-Horvat and U. Hoffmann

CCTA FFR-CT ICA

Case 1

FFRCT 0.84

Case 2

FFRCT 0.76

Fig. 57.5  Both cases (upper and lower panels) represent severe steno- arrows indicate severe luminal narrowing both in coronary CTA and
sis (approximately 70%) as depicted by coronary CTA. FFR-CT dem- ICA.  Abbreviations: CTA, computed tomography angiography; FFR,
onstrates no lesion-specific ischemia (0.86) in the first case, whereas the fractional flow reserve; ICA, invasive coronary angiography (Courtesy
FFR-CT shows ischemia (0.70) in the second case. Both coronary CTA of HeartFlow Inc., Redwood City, CA, USA)
stenosis severities and the FFR-CT values were confirmed by ICA. The

[68]. Therefore, FFR seems to be an important determinant Abnormal FFR, the flow perturbations, the altered ESS,
of lesion vulnerability. and lesion strain might be responsible for the development of
Plaques develop at specific areas of coronary arteries rupture-prone lesion [74, 75]. Furthermore, patients with an
where flow is disturbed, such as the outer walls of ­bifurcations, obstructive coronary plaque may in fact develop an ACS due
in side branches, and in the inner curve of the arteries. These to the high ESS-induced thrombus formation [76]. Therefore
predilection sites are characterized by low and oscillatory the determination of blood flow-related functional parame-
endothelial shear stress [70, 71]. On the other hand, it seems ters might improve the atherosclerotic lesion characteriza-
that high endothelial shear stress (ESS) is a key component tion [77]. The three-dimensional dataset acquired by
in atherosclerotic plaque destabilization, arterial wall remod- coronary CTA allows for computational fluid dynamics and
eling, and atherosclerosis progression [72]. ESS is the tan- the simulation of patient-specific hemodynamic parameters
gential force generated by the friction of flowing blood on including FFRCT and ESSCT, which highly likely will improve
the endothelial surface of the arterial wall [73]. the accuracy of coronary CTA to detect vulnerable plaques.
57  Advanced Methods for Coronary Artery Plaque Analysis 731

Spectral Plaque Assessment spectral CT imaging of gold-labeled high-density lipoprotein

(Au-HDL) nanoparticles designed to target activated macro-
Conventional single-energy coronary CTA faces a significant phages showed promising results in atherosclerotic mice
challenge in differentiating components of non-calcified model [83]. The results in animal studies are encouraging and
plaques (e.g., lipid-rich vs. fibrous) [78]. Several studies have might open new avenues in multi-contrast spectral imaging.
shown considerable overlap in Hounsfield units between The capabilities of spectral CT imaging are promising; how-
lipid-rich and fibrous non-calcified plaques inherent to the ever, further studies are warranted to determine the proper
limited spatial and soft-tissue resolution of CT [79]. Spectral methods for plaque assessment by this emerging technology.
or multi-energy CT imaging has a great potential to improve
morphological plaque assessment. Spectral CT allows for
both monochromatic energy imaging and material basis Coronary CTA Radiomics
decomposition. Materials and tissues can be distinguished by
evaluating material density images [80]. The tissue character- For long, radiological images have been regarded as pic-
ization capabilities of spectral CT may improve the identifi- tures of the inner body. However, radiological images are
cation of lipid-rich components of non-calcified plaque tissue in fact multidimensional datasets consisting of voxels, and
[81]. Furthermore, density and atomic number histograms of each voxel value represents a specific measurement based
materials can be derived within regions of interests. It has on physical tissue characteristics. Radiomics is the process
been demonstrated that this histogram analysis is able to dif- of obtaining quantitative parameters from these spatial
ferentiate between components of renal stones and may also datasets, in order to create “big data” databases, where
be applied to evaluate components of coronary calcifications each lesion is characterized by hundreds of different
such as calcium oxalate and hydroxyapatite [82]. In addition, parameters  (Fig.  57.6) [84, 85]. These features aim to

a c

0.0 0.25 0.50 0.75 1.0

Fig. 57.6  Panel A shows volume rendered and cross-sectional images calcified plaque; NRS, napkin-ring sign. Panel C: heatmap of a covari-
of a plaque with napkin-ring sign (NRS). Panel B represents a plaque ance matrix of more than 7000 radiomic features. A dendrogram of the
with no NRS. Green dashed lines indicate the location of cross-­sectional corresponding hierarchical clustering can be seen on the right side of
planes. Colors indicate different CT attenuation values. NCP, non-­ the image [89]
732 P. Maurovich-Horvat and U. Hoffmann

References
1. Braunwald E.  Epilogue: what do clinicians expect from
imagers? J Am Coll Cardiol. 2006;47(8 Suppl):C101–3.
doi:S0735-1097(06)00163-X [pii]. https://doi.org/10.1016/j.
jacc.2005.2.072.
2. Heidenreich PA, Trogdon JG, Khavjou OA, Butler J, Dracup K,
Ezekowitz MD, Finkelstein EA, Hong Y, Johnston SC, Khera A,
Lloyd-Jones DM, Nelson SA, Nichol G, Orenstein D, Wilson PW,
Woo YJ, American Heart Association Advocacy Coordinating
Committee; Stroke Council; Council on Cardiovascular Radiology
and Intervention; Council on Clinical Cardiology; Council on
Epidemiology and Prevention; Council on Arteriosclerosis;
Thrombosis and Vascular Biology; Council on Cardiopulmonary;
Critical Care; Perioperative and Resuscitation; Council on
Cardiovascular Nursing; Council on the Kidney in Cardiovascular
Disease; Council on Cardiovascular Surgery and Anesthesia,
and Interdisciplinary Council on Quality of Care and Outcomes
Research. Forecasting the future of cardiovascular disease in
the United States: a policy statement from the American Heart
Association. Circulation. 2011;123(8):933–44. https://doi.
org/10.1161/CIR.0b013e31820a55f5.
3. Waxman S, Ishibashi F, Muller JE.  Detection and treatment of
vulnerable plaques and vulnerable patients: novel approaches
to prevention of coronary events. Circulation. 2006;114(22):
2390–411.
4. Libby P.  Mechanisms of acute coronary syndromes and their
implications for therapy. N Engl J Med. 2013;368(21):2004–13.
5. Writing Group Members, Mozaffarian D, Benjamin EJ, Go AS,
Arnett DK, Blaha MJ, Cushman M, Das SR, de Ferranti S, Despres
JP, Fullerton HJ, Howard VJ, Huffman MD, Isasi CR, Jimenez MC,
Fig. 57.7  Combined image of functional and anatomical plaque fea-
Judd SE, Kissela BM, Lichtman JH, Lisabeth LD, Liu S, Mackey
tures represented in a three-dimensional artistic illustration
RH, Magid DJ, DK MG, Mohler ER 3rd, Moy CS, Muntner P,
Mussolino ME, Nasir K, Neumar RW, Nichol G, Palaniappan L,
Pandey DK, Reeves MJ, Rodriguez CJ, Rosamond W, Sorlie PD,
quantify ­morphological characteristics difficult or impos- Stein J, Towfighi A, Turan TN, Virani SS, Woo D, Yeh RW, Turner
sible to comprehend by visual assessment [86]. As the MB, American Heart Association Statistics Committee; Stroke
number of examinations grows, the limitations of human Statistics Subcommittee. Heart disease and stroke statistics-2016
update: a report from the American Heart Association. Circulation.
visual perception influence the overall diagnostic accuracy
2016;133(4):e38–360.
of radiological examinations [87]. This results in increased 6. Burke AP, Farb A, Malcom GT, Liang YH, Smialek J, Virmani
inter- and intra-reader variability [88]. The quantitative R.  Coronary risk factors and plaque morphology in men
nature of radiomics has the potential to greatly decrease with coronary disease who died suddenly. N Engl J Med.
1997;336(18):1276–82.
this variability and therefore increase the diagnostic accu-
7. Falk E, Nakano M, Bentzon JF, Finn AV, Virmani R.  Update on
racy of radiological examinations. Therefore, one of the acute coronary syndromes: the pathologists’ view. Eur Heart J.
main advantages of radiomics in coronary CTA imaging 2013;34(10):719–28.
might be to decrease the reader dependency of determining 8. Virmani R, Kolodgie FD, Burke AP, Farb A, Schwartz SM. Lessons
from sudden coronary death: a comprehensive morphological clas-
high-risk plaque features [89]. Furthermore, radiomics has
sification scheme for atherosclerotic lesions. Arterioscler Thromb
the potential to identify new imaging biomarkers and Vasc Biol. 2000;20(5):1262–75.
improve risk stratification and early disease detection 9. Narula J, Garg P, Achenbach S, Motoyama S, Virmani R, Strauss
based on coronary CTA datasets. Please see for more HW.  Arithmetic of vulnerable plaques for noninvasive imaging.
Nat Clin Pract Cardiovasc Med. 2008;5(Suppl 2):S2–10. https://
details section 56 on Cardiac CT radiomics.
doi.org/10.1038/ncpcardio1247.
A holistic approach of combining functional and mor- 10. Finn AV, Nakano M, Narula J, Kolodgie FD, Virmani R. Concept
phological parameters with clinical and genetic informa- of vulnerable/unstable plaque. Arterioscler Thromb Vasc Biol.
tion is a promising technique to change our understanding 2010;30(7):1282–92.
11. Maurovich-Horvat P, Ferencik M, Voros S, Merkely B, Hoffmann
and treatment of cardiovascular diseases and can facilitate
U.  Comprehensive plaque assessment by coronary CT angiogra-
personalized risk assessment to identify vulnerable patients phy. Nat Rev Cardiol. 2014;11(7):390–402.
and vulnerable plaques (Fig. 57.7).
57  Advanced Methods for Coronary Artery Plaque Analysis 733

12. Bittencourt MS, Hulten E, Ghoshhajra B, O'Leary D, Christman H, Narula J.  Multislice computed tomographic characteristics of
MP, Montana P, Truong QA, Steigner M, Murthy VL, Rybicki FJ, coronary lesions in acute coronary syndromes. J Am Coll Cardiol.
Nasir K, Gowdak LH, Hainer J, Brady TJ, Di Carli MF, Hoffmann 2007;50(4):319–26. https://doi.org/10.1016/j.jacc.2007.03.044.
U, Abbara S, Blankstein R. Prognostic value of nonobstructive and 24. Pohle K, Achenbach S, Macneill B, Ropers D, Ferencik

obstructive coronary artery disease detected by coronary computed M, Moselewski F, Hoffmann U, Brady TJ, Jang IK, Daniel
tomography angiography to identify cardiovascular events. Circ WG.  Characterization of non-calcified coronary athero-
Cardiovasc Imaging. 2014;7(2):282–91. sclerotic plaque by multi-detector row CT: comparison to
13. Mancini GB, Hartigan PM, Shaw LJ, Berman DS, Hayes SW, Bates IVUS. Atherosclerosis. 2007;190(1):174–80.
ER, Maron DJ, Teo K, Sedlis SP, Chaitman BR, Weintraub WS, 25. Achenbach S, Boehmer K, Pflederer T, Ropers D, Seltmann M,
Spertus JA, Kostuk WJ, Dada M, Booth DC, Boden WE. Predicting Lell M, Anders K, Kuettner A, Uder M, Daniel WG, Marwan
outcome in the COURAGE trial (clinical outcomes utilizing revas- M.  Influence of slice thickness and reconstruction kernel on the
cularization and aggressive drug evaluation): coronary anatomy computed tomographic attenuation of coronary atherosclerotic
versus ischemia. JACC Cardiovasc Interv. 2014;7(2):195–201. plaque. J Cardiovasc Comput Tomogr. 2010;4(2):110–5.
14. Narula J, Nakano M, Virmani R, Kolodgie FD, Petersen R,
26. Cademartiri F, Mollet NR, Runza G, Bruining N, Hamers R,

Newcomb R, Malik S, Fuster V, Finn AV. Histopathologic charac- Somers P, Knaapen M, Verheye S, Midiri M, Krestin GP, de Feyter
teristics of atherosclerotic coronary disease and implications of the PJ. Influence of intracoronary attenuation on coronary plaque mea-
findings for the invasive and noninvasive detection of vulnerable surements using multislice computed tomography: observations in
plaques. J Am Coll Cardiol. 2013;61(10):1041–51. an ex vivo model of coronary computed tomography angiography.
15. Ferencik M, Schlett CL, Ghoshhajra BB, Kriegel MF, Joshi
Eur Radiol. 2005;15(7):1426–31.
SB, Maurovich-Horvat P, Rogers IS, Banerji D, Bamberg F, 27. Ferencik M, Chan RC, Achenbach S, Lisauskas JB, Houser SL,
Truong QA, Brady TJ, Nagurney JT, Hoffmann U.  A computed Hoffmann U, Abbara S, Cury RC, Bouma BE, Tearney GJ, Brady
tomography-­based coronary lesion score to predict acute coronary TJ. Arterial wall imaging: evaluation with 16-section multidetec-
syndrome among patients with acute chest pain and significant tor CT in blood vessel phantoms and ex  vivo coronary arteries.
coronary stenosis on coronary computed tomographic angiogram. Radiology. 2006;240(3):708–16.
Am J Cardiol. 2012;110(2):183–9. 28. Suzuki S, Furui S, Kuwahara S, Kaminaga T, Yamauchi T, Konno
16. Achenbach S, Friedrich MG, Nagel E, Kramer CM, Kaufmann PA, K, Yokoyama N, Isshiki T. Accuracy of attenuation measurement
Farkhooy A, Dilsizian V, Flachskampf FA. CV imaging: what was of vascular wall in  vitro on computed tomography angiography:
new in 2012? J Am Coll Cardiol Img. 2013;6(6):714–34. effect of wall thickness, density of contrast medium, and mea-
17. Virmani R, Burke AP, Farb A, Kolodgie FD. Pathology of the vul- surement point. Investig Radiol. 2006;41(6):510–5. https://doi.
nerable plaque. J Am Coll Cardiol. 2006;47(8 Suppl):C13–8. org/10.1097/01.rli.0000209662.24569.c7.
18. Kolodgie FD, Burke AP, Farb A, Gold HK, Yuan J, Narula J, Finn 29. Glagov S, Weisenberg E, Zarins CK, Stankunavicius R, Kolettis
AV, Virmani R. The thin-cap fibroatheroma: a type of vulnerable GJ.  Compensatory enlargement of human atherosclerotic coro-
plaque: the major precursor lesion to acute coronary syndromes. nary arteries. N Engl J Med. 1987;316(22):1371–5. https://doi.
Curr Opin Cardiol. 2001;16(5):285–92. org/10.1056/NEJM198705283162204.
19. van der Giessen AG, Toepker MH, Donelly PM, Bamberg F,
30. Varnava AM, Mills PG, Davies MJ.  Relationship between coro-
Schlett CL, Raffle C, Irlbeck T, Lee H, van Walsum T, Maurovich-­ nary artery remodeling and plaque vulnerability. Circulation.
Horvat P, Gijsen FJ, Wentzel JJ, Hoffmann U.  Reproducibility, 2002;105(8):939–43.
accuracy, and predictors of accuracy for the detection of coronary 31. Achenbach S, Ropers D, Hoffmann U, MacNeill B, Baum U,
atherosclerotic plaque composition by computed tomography: an Pohle K, Brady TJ, Pomerantsev E, Ludwig J, Flachskampf FA,
ex  vivo comparison to intravascular ultrasound. Investig Radiol. Wicky S, Jang IK, Daniel WG.  Assessment of coronary remod-
2010;45(11):693–701. eling in stenotic and nonstenotic coronary atherosclerotic lesions
20. Hoffmann U, Ferencik M, Cury RC, Pena AJ. Coronary CT angi- by multidetector spiral computed tomography. J Am Coll Cardiol.
ography. J Nucl Med. 2006;47(5):797–806. 2004;43(5):842–7.
21. Leipsic J, Abbara S, Achenbach S, Cury R, Earls JP, Mancini GJ, 32. Mintz GS, Nissen SE, Anderson WD, Bailey SR, Erbel R,

Nieman K, Pontone G, Raff GL.  SCCT guidelines for the inter- Fitzgerald PJ, Pinto FJ, Rosenfield K, Siegel RJ, Tuzcu EM, Yock
pretation and reporting of coronary CT angiography: a report of PG.  American College of Cardiology Clinical Expert Consensus
the Society of Cardiovascular Computed Tomography Guidelines Document on Standards for Acquisition, Measurement and
Committee. J Cardiovasc Comput Tomogr. 2014;8(5):342–58. Reporting of Intravascular Ultrasound Studies (IVUS). A report of
22. Voros S, Rinehart S, Qian Z, Vazquez G, Anderson H, Murrieta the American College of Cardiology Task Force on Clinical Expert
L, Wilmer C, Carlson H, Taylor K, Ballard W, Karmpaliotis D, Consensus Documents. J Am Coll Cardiol. 2001;37(5):1478–92.
Kalynych A, Brown C 3rd. Prospective validation of standard- 33. Gauss S, Achenbach S, Pflederer T, Schuhback A, Daniel WG,
ized, 3-dimensional, quantitative coronary computed tomographic Marwan M. Assessment of coronary artery remodelling by dual-­
plaque measurements using radiofrequency backscatter intra- source CT: a head-to-head comparison with intravascular ultra-
vascular ultrasound as reference standard in intermediate coro- sound. Heart. 2011;97(12):991–7.
nary arterial lesions: results from the ATLANTA (assessment of 34. Motoyama S, Ito H, Sarai M, Kondo T, Kawai H, Nagahara Y,
tissue characteristics, lesion morphology, and hemodynamics Harigaya H, Kan S, Anno H, Takahashi H, Naruse H, Ishii J,
by angiography with fractional flow reserve, intravascular ultra- Hecht H, Shaw LJ, Ozaki Y, Narula J. Plaque characterization by
sound and virtual histology, and noninvasive computed tomogra- coronary computed tomography angiography and the likelihood of
phy in atherosclerotic plaques) I study. JACC Cardiovasc Interv. acute coronary events in mid-term follow-up. J Am Coll Cardiol.
2011;4(2):198–208. 2015;66(4):337–46.
23. Motoyama S, Kondo T, Sarai M, Sugiura A, Harigaya H, Sato T, 35. Ahmadi A, Leipsic J, Blankstein R, Taylor C, Hecht H, Stone
Inoue K, Okumura M, Ishii J, Anno H, Virmani R, Ozaki Y, Hishida GW, Narula J.  Do plaques rapidly progress prior to myocardial
734 P. Maurovich-Horvat and U. Hoffmann

i­nfarction? The interplay between plaque vulnerability and pro- lesions in acute coronary syndromes using coronary dual-source
gression. Circ Res. 2015;117(1):99–104. CT angiography. Atherosclerosis. 2010;211(2):437–44.
36. Maurovich-Horvat P, Hoffmann U, Vorpahl M, Nakano M, Virmani 51.
Madder RD, Chinnaiyan KM, Marandici AM, Goldstein
R, Alkadhi H. The napkin-ring sign: CT signature of high risk cor- JA.  Features of disrupted plaques by coronary computed tomo-
onary plaques? JACC Cardiovasc Imaging. 2010;3(4):440–05. graphic angiography: correlates with invasively proven complex
37. Seifarth H, Schlett CL, Nakano M, Otsuka F, Karolyi M, Liew lesions. Circ Cardiovasc Imaging. 2011;4(2):105–13.
G, Maurovich-Horvat P, Alkadhi H, Virmani R, Hoffmann 52. Arbab-Zadeh A, Fuster V.  The myth of the “vulnerable plaque”:
U. Histopathological correlates of the napkin-ring sign plaque in transitioning from a focus on individual lesions to atherosclerotic
coronary CT angiography. Atherosclerosis. 2012;224(1):90–6. disease burden for coronary artery disease risk assessment. J Am
38. Maurovich-Horvat P, Schlett CL, Alkadhi H, Nakano M, Otsuka Coll Cardiol. 2015;65(8):846–55.
F, Stolzmann P, Scheffel H, Ferencik M, Kriegel MF, Seifarth H, 53. Kubo T, Maehara A, Mintz GS, Doi H, Tsujita K, Choi SY, Katoh
Virmani R, Hoffmann U. The napkin-ring sign indicates advanced O, Nasu K, Koenig A, Pieper M, Rogers JH, Wijns W, Bose D,
atherosclerotic lesions in coronary CT angiography. J Am Coll Margolis MP, Moses JW, Stone GW, Leon MB.  The dynamic
Cardiol Img. 2012;5(12):1243–52. nature of coronary artery lesion morphology assessed by serial
39. Otsuka K, Fukuda S, Tanaka A, Nakanishi K, Taguchi H, Yoshikawa virtual histology intravascular ultrasound tissue characterization.
J, Shimada K, Yoshiyama M.  Napkin-ring sign on coronary CT J Am Coll Cardiol. 2010;55(15):1590–7.
angiography for the prediction of acute coronary syndrome. J Am 54.
Iwasaki K, Matsumoto T.  Dynamic change of high-risk
Coll Cardiol Img. 2013;6(4):448–57. plaque detected by coronary computed tomographic angiog-
40. Feuchtner G, Kerber J, Burghard P, Dichtl W, Friedrich G, Bonaros raphy in patients with subclinical coronary artery disease. Int
N, Plank F.  The high-risk criteria low-attenuation plaque <60 J Cardiovasc Imaging. 2016;32:1667. https://doi.org/10.1007/
HU and the napkin-ring sign are the most powerful predictors s10554-016-0957-2.
of MACE: a long-term follow-up study. Eur Heart J Cardiovasc 55. Min JK, Shaw LJ, Devereux RB, Okin PM, Weinsaft JW, Russo
Imaging. 2016;18:772. https://doi.org/10.1093/ehjci/jew167. DJ, Lippolis NJ, Berman DS, Callister TQ.  Prognostic value
41. Otsuka F, Finn AV, Virmani R.  Do vulnerable and ruptured
of multidetector coronary computed tomographic angiogra-
plaques hide in heavily calcified arteries? Atherosclerosis. phy for prediction of all-cause mortality. J Am Coll Cardiol.
2013;229(1):34–7. 2007;50(12):1161–70.
42. Huang H, Virmani R, Younis H, Burke AP, Kamm RD, Lee
56. Andreini D, Pontone G, Mushtaq S, Bartorelli AL, Bertella E,
RT. The impact of calcification on the biomechanical stability of Antonioli L, Formenti A, Cortinovis S, Veglia F, Annoni A,
atherosclerotic plaques. Circulation. 2001;103(8):1051–6. Agostoni P, Montorsi P, Ballerini G, Fiorentini C, Pepi M.  A
43. Maldonado N, Kelly-Arnold A, Vengrenyuk Y, Laudier D, Fallon long-term prognostic value of coronary CT angiography in sus-
JT, Virmani R, Cardoso L, Weinbaum S. A mechanistic analysis of pected coronary artery disease. JACC Cardiovasc Imaging.
the role of microcalcifications in atherosclerotic plaque stability: 2012;5(7):690–701.
potential implications for plaque rupture. Am J Phys Heart Circ 57. Dougoud S, Fuchs TA, Stehli J, Clerc OF, Buechel RR, Herzog BA,
Phys. 2012;303(5):H619–28. Leschka S, Alkadhi H, Kaufmann PA, Gaemperli O.  Prognostic
44. Mauriello A, Servadei F, Zoccai GB, Giacobbi E, Anemona
value of coronary CT angiography on long-term follow-up of 6.9
L, Bonanno E, Casella S.  Coronary calcification identi- years. Int J Cardiovasc Imaging. 2014;30(5):969–76.
fies the vulnerable patient rather than the vulnerable plaque. 58. Hadamitzky M, Taubert S, Deseive S, Byrne RA, Martinoff

Atherosclerosis. 2013;229(1):124–9. https://doi.org/10.1016/j. S, Schomig A, Hausleiter J.  Prognostic value of coronary com-
atherosclerosis.2013.03.010. puted tomography angiography during 5 years of follow-up
45. Burke AP, Weber DK, Kolodgie FD, Farb A, Taylor AJ, Virmani in patients with suspected coronary artery disease. Eur Heart J.
R.  Pathophysiology of calcium deposition in coronary arteries. 2013;34(42):3277–85.
Herz. 2001;26(4):239–44. 59. Al-Mallah MH.  Does coronary CT angiography improve risk

46. van Velzen JE, de Graaf FR, de Graaf MA, Schuijf JD, Kroft LJ, stratification over coronary artery calcium scoring in symptom-
de Roos A, Reiber JH, Bax JJ, Jukema JW, Boersma E, Schalij atic patients with a low pre-test probability of coronary artery dis-
MJ, van der Wall EE.  Comprehensive assessment of spotty cal- ease and a CAC of zero? Reply. Eur Heart J Cardiovasc Imaging.
cifications on computed tomography angiography: comparison to 2014;15(2):232–3.
plaque characteristics on intravascular ultrasound with radiofre- 60. Hadamitzky M, Achenbach S, Al-Mallah M, Berman D, Budoff
quency backscatter analysis. J Nucl Cardiol. 2011;18(5):893–903. M, Cademartiri F, Callister T, Chang HJ, Cheng V, Chinnaiyan
47. Kim SY, Kim KS, Seung MJ, Chung JW, Kim JH, Mun SH, Lee K, Chow BJ, Cury R, Delago A, Dunning A, Feuchtner G,
YS, Lee JB, Ryu JK, Choi JY, Chang SG. The culprit lesion score Gomez M, Kaufmann P, Kim YJ, Leipsic J, Lin FY, Maffei E,
on multi-detector computed tomography can detect vulnerable Min JK, Raff G, Shaw LJ, Villines TC, Hausleiter J, Investigators
coronary artery plaque. Int J Cardiovasc Imaging. 2010;26(Suppl C.  Optimized prognostic score for coronary computed tomo-
2):245–52. graphic angiography: results from the CONFIRM registry
48. Kitagawa T, Yamamoto H, Horiguchi J, Ohhashi N, Tadehara F, (COronary CT Angiography EvaluatioN For Clinical Outcomes:
Shokawa T, Dohi Y, Kunita E, Utsunomiya H, Kohno N, Kihara An InteRnational Multicenter Registry). J Am Coll Cardiol.
Y. Characterization of noncalcified coronary plaques and identifi- 2013;62(5):468–76.
cation of culprit lesions in patients with acute coronary syndrome 61. Versteylen MO, Kietselaer BL, Dagnelie PC, Joosen IA, Dedic A,
by 64-slice computed tomography. JACC Cardiovasc Imaging. Raaijmakers RH, Wildberger JE, Nieman K, Crijns HJ, Niessen
2009;2(2):153–60. WJ, Daemen MJ, Hofstra L.  Additive value of semiautomated
49. Ozaki Y, Okumura M, Ismail TF, Motoyama S, Naruse H, Hattori quantification of coronary artery disease using cardiac computed
K, Kawai H, Sarai M, Takagi Y, Ishii J, Anno H, Virmani R, Serruys tomographic angiography to predict future acute coronary syn-
PW, Narula J. Coronary CT angiographic characteristics of culprit drome. J Am Coll Cardiol. 2013;61(22):2296–305.
lesions in acute coronary syndromes not related to plaque rupture as 62. Voros S, Rinehart S, Qian Z, Joshi P, Vazquez G, Fischer C, Belur
defined by optical coherence tomography and angioscopy. Eur Heart P, Hulten E, Villines TC.  Coronary atherosclerosis imaging by
J. 2011;32(22):2814–23. https://doi.org/10.1093/eurheartj/ehr189. ­coronary CT angiography: current status, correlation with intravas-
50. Pflederer T, Marwan M, Schepis T, Ropers D, Seltmann M,
cular interrogation and meta-analysis. JACC Cardiovasc Imaging.
Muschiol G, Daniel WG, Achenbach S. Characterization of culprit 2011;4(5):537–48.
57  Advanced Methods for Coronary Artery Plaque Analysis 735

63. Boogers MJ, Broersen A, van Velzen JE, de Graaf FR, El-Naggar AF, Serruys PW. Strain distribution over plaques in human coro-
HM, Kitslaar PH, Dijkstra J, Delgado V, Boersma E, de Roos A, nary arteries relates to shear stress. Am J Phys Heart Circ Phys.
Schuijf JD, Schalij MJ, Reiber JH, Bax JJ, Jukema JW. Automated 2008;295(4):H1608–14.
quantification of coronary plaque with computed tomography: 76. Yong AS, Pennings GJ, Chang M, Hamzah A, Chung T, Qi M,
comparison with intravascular ultrasound using a dedicated reg- Brieger D, Behnia M, Krilis SA, Ng MK, Lowe HC, Kritharides
istration algorithm for fusion-based quantification. Eur Heart J. L. Intracoronary shear-related up-regulation of platelet P-selectin
2012;33(8):1007–16. and platelet-monocyte aggregation despite the use of aspirin and
64. Oberoi S, Meinel FG, Schoepf UJ, Nance JW, De Cecco CN, clopidogrel. Blood. 2011;117(1):11–20.
Gebregziabher M, Costello P, Weininger M.  Reproducibility of 77. Han D, Starikov A, Ó Hartaigh B, Gransar H, Kolli KK, Lee
noncalcified coronary artery plaque burden quantification from JH, Rizvi A, Baskaran L, Schulman-Marcus J, Lin FY, Min
coronary CT angiography across different image analysis plat- JK. Relationship between endothelial wall shear stress and high-­
forms. AJR Am J Roentgenol. 2014;202(1):W43–9. risk atherosclerotic plaque characteristics for identification of
65. Taylor CA, Fonte TA, Min JK.  Computational fluid dynamics coronary lesions that cause ischemia: a direct comparison with
applied to cardiac computed tomography for noninvasive quan- fractional flow reserve. J Am Heart Assoc. 2016;5(12) https://doi.
tification of fractional flow reserve: scientific basis. J Am Coll org/10.1161/JAHA.116.004186.
Cardiol. 2013;61(22):2233–41. 78. Danad I, Fayad ZA, Willemink MJ, Min JK.  New applications
66. Min JK, Leipsic J, Pencina MJ, Berman DS, Koo BK, van
of cardiac computed tomography: dual-energy, spectral, and
Mieghem C, Erglis A, Lin FY, Dunning AM, Apruzzese P, Budoff molecular CT imaging. JACC Cardiovasc Imaging. 2015;8(6):
MJ, Cole JH, Jaffer FA, Leon MB, Malpeso J, Mancini GB, Park 710–23.
SJ, Schwartz RS, Shaw LJ, Mauri L.  Diagnostic accuracy of 79. Leber AW, Knez A, Becker A, Becker C, von Ziegler F, Nikolaou K,
fractional flow reserve from anatomic CT angiography. JAMA. Rist C, Reiser M, White C, Steinbeck G, Boekstegers P. Accuracy
2012;308(12):1237–45. of multidetector spiral computed tomography in identifying
67. Norgaard BL, Leipsic J, Gaur S, Seneviratne S, Ko BS, Ito H, and differentiating the composition of coronary atherosclerotic
Jensen JM, Mauri L, De Bruyne B, Bezerra H, Osawa K, Marwan plaques: a comparative study with intracoronary ultrasound. J Am
M, Naber C, Erglis A, Park SJ, Christiansen EH, Kaltoft A, Lassen Coll Cardiol. 2004;43(7):1241–7.
JF, Botker HE, Achenbach S, Group NXTTS.  Diagnostic perfor- 80. Willemink MJ, Leiner T, Maurovich-Horvat P. Cardiac CT imag-
mance of noninvasive fractional flow reserve derived from coronary ing of plaque vulnerability: hype or Hope? Curr Cardiol Rep.
computed tomography angiography in suspected coronary artery 2016;18(4):37. https://doi.org/10.1007/s11886-016-0714-0.
disease: the NXT trial (analysis of coronary blood flow using CT 81. Machida H, Tanaka I, Fukui R, Shen Y, Ishikawa T, Tate E,

angiography: next steps). J Am Coll Cardiol. 2014;63(12):1145–55. Ueno E.  Current and novel imaging techniques in coronary
68. Ahmadi A, Stone GW, Leipsic J, Serruys PW, Shaw L, Hecht CT. Radiographics. 2015;35(4):991–1010.
H, Wong G, Norgaard BL, O'Gara PT, Chandrashekhar Y, 82. Kulkarni NM, Eisner BH, Pinho DF, Joshi MC, Kambadakone AR,
Narula J.  Association of coronary stenosis and plaque morphol- Sahani DV. Determination of renal stone composition in phantom
ogy with fractional flow reserve and outcomes. JAMA Cardiol. and patients using single-source dual-energy computed tomogra-
2016;1(3):350–7. phy. J Comput Assist Tomogr. 2013;37(1):37–45.
69. Park HB, Heo R, Ó Hartaigh B, Cho I, Gransar H, Nakazato R, 83. Cormode DP, Roessl E, Thran A, Skajaa T, Gordon RE,

Leipsic J, Mancini GB, Koo BK, Otake H, Budoff MJ, Berman Schlomka JP, Fuster V, Fisher EA, Mulder WJ, Proksa R,
DS, Erglis A, Chang HJ, Min JK. Atherosclerotic plaque charac- Fayad ZA.  Atherosclerotic plaque composition: analysis with
teristics by CT angiography identify coronary lesions that cause multicolor CT and targeted gold nanoparticles. Radiology.
ischemia: a direct comparison to fractional flow reserve. JACC 2010;256(3):774–82.
Cardiovasc Imaging. 2015;8(1):1–10. 84. Gillies RJ, Kinahan PE, Hricak H.  Radiomics: images are more
70. Stary HC, Blankenhorn DH, Chandler AB, Glagov S, Insull W Jr, than pictures, they are data. Radiology. 2016;278(2):563–77.
Richardson M, Rosenfeld ME, Schaffer SA, Schwartz CJ, Wagner 85. Kolossvary M, Kellermayer M, Merkely B, Maurovich-Horvat
WD, et al. A definition of the intima of human arteries and of its P.  Cardiac computed tomography radiomics: a comprehensive
atherosclerosis-prone regions. A report from the Committee on review on radiomic techniques. J Thorac Imaging. 2017;33:26.
Vascular Lesions of the Council on Arteriosclerosis, American https://doi.org/10.1097/RTI.0000000000000268.
Heart Association. Circulation. 1992;85(1):391–405. 86. Davnall F, Yip CS, Ljungqvist G, Selmi M, Ng F, Sanghera B,
71. Chatzizisis YS, Coskun AU, Jonas M, Edelman ER, Feldman
Ganeshan B, Miles KA, Cook GJ, Goh V.  Assessment of tumor
CL, Stone PH.  Role of endothelial shear stress in the natural heterogeneity: an emerging imaging tool for clinical practice?
history of coronary atherosclerosis and vascular remodeling: Insights Imaging. 2012;3(6):573–89.
molecular, cellular, and vascular behavior. J Am Coll Cardiol. 87. Krupinski EA, Berbaum KS, Caldwell RT, Schartz KM, Kim

2007;49(25):2379–93. J. Long radiology workdays reduce detection and accommodation
72. Samady H, Eshtehardi P, McDaniel MC, Suo J, Dhawan SS,
accuracy. J Am Coll Radiol. 2010;7(9):698–704.
Maynard C, Timmins LH, Quyyumi AA, Giddens DP.  Coronary 88. Puchner SB, Liu T, Mayrhofer T, Truong QA, Lee H, Fleg JL,
artery wall shear stress is associated with progression and transfor- Nagurney JT, Udelson JE, Hoffmann U, Ferencik M.  High-risk
mation of atherosclerotic plaque and arterial remodeling in patients plaque detected on coronary CT angiography predicts acute
with coronary artery disease. Circulation. 2011;124(7):779–88. coronary syndromes independent of significant stenosis in acute
73. Slager CJ, Wentzel JJ, Gijsen FJ, Schuurbiers JC, van der Wal chest pain: results from the ROMICAT-II trial. J Am Coll Cardiol.
AC, van der Steen AF, Serruys PW.  The role of shear stress in 2014;64(7):684–92.
the generation of rupture-prone vulnerable plaques. Nat Clin Pract 89. Kolossváry M, Karády J, Szilveszter B, Kitslaar P, Hoffmann U,
Cardiovasc Med. 2005;2(8):401–7. Merkely B, Maurovich-Horvat P. Radiomic features are superior
74. Fearon WF. Is a myocardial infarction more likely to result from to conventional quantitative computed tomographic metrics to
a mild coronary lesion or an ischemia-producing one? Circ identify coronary plaques with napkin-ring sign. Circ Cardiovasc
Cardiovasc Interv. 2011;4(6):539–41. Imaging. 2017;10(12).
75. Gijsen FJ, Wentzel JJ, Thury A, Mastik F, Schaar JA, Schuurbiers
JC, Slager CJ, van der Giessen WJ, de Feyter PJ, van der Steen
 Part XIV
Where We Are Going: Risk Prediction and
Management: The Next Wave
 Coronary CT Angiography
for Screening, Risk Stratification, 58
and Management of Asymptomatic
Patients: State of the Evidence

Felix G. Meinel and Matthias Renker

Unlike coronary artery calcium scoring, coronary CTA is Table 58.1  Wilson and Jungner classic screening criteria
generally not established as a tool for screening and risk The condition should be an important health problem
stratification of asymptomatic individuals. The most recent There should be a treatment for the condition
multi-society appropriate use guidelines rate coronary CTA Facilities for diagnosis and treatment should be available
as “rarely appropriate” in asymptomatic patients with low to There should be a latent stage of the disease
intermediate risk for CAD [1]. They suggest that coronary There should be a test or examination for the condition
The test should be acceptable to the population
CTA “may be appropriate” in asymptomatic patients with a
The natural history of the disease should be adequately understood
high risk for CAD (>20% ten-year CAD risk or CAD equiva- There should be an agreed policy on whom to treat
lent such as diabetes or peripheral artery disease) [1]. The total cost of finding a case should be economically balanced in
Compared to coronary artery calcium scoring, coronary CTA relation to medical expenditure as a whole
is associated with a higher radiation dose and the risks of Case finding should be a continuous process, not just a “once and
contrast material administration. These risks are small but for all” project
real [2, 3] and need to be carefully weighed when consider- From Wilson and Jungner[4], with permission
ing coronary CTA as a screening test in large populations of
individuals with no signs or symptoms of CAD. Table 58.2  Synthesis of emerging screening criteria proposed over the
Commissioned by the World Health Organization, Wilson past 40 years
and Jungner in 1968 published ten criteria for diseases that The screening program should respond to a recognized need
may be amenable to successful screening programs [4]. They The objectives of screening should be defined at the outset
are now considered the “classic” WHO criteria, also known There should be a defined target population
as Wilson’s criteria (Table  58.1). Criteria 1–7 on this list There should be scientific evidence of screening program
effectiveness
clearly apply to CAD and indeed make it seem like an obvi-
The program should integrate education, testing, clinical services,
ous candidate for screening. Item 8 is a challenge: There is and program management
far-reaching consensus whom to treat in symptomatic There should be quality assurance, with mechanisms to minimize
patients with CAD, but this is not so clear for asymptomatic potential risks of screening
patients. The paragraph “Role for Management” in this The program should ensure informed choice, confidentiality, and
respect for autonomy
chapter discusses the practical consequences of finding evi-
The program should promote equity and access to screening for the
dence of CAD in an asymptomatic individual. Even more entire target population
challenging are the requirements for a successful screening Program evaluation should be planned from the outset
program that were formulated in the “modified” WHO crite- The overall benefits of screening should outweigh the harm
ria published in 2008 [5] (Table 58.2). The key items on that From Andermann et al.[5], with permission
list are 4 and 10. For screening to be justified, there should be
evidence (in the form of randomized controlled trials) that the program is effective and that the overall benefits of
screening outweigh the harm. As will be discussed in the
F. G. Meinel (*) paragraph “Impact” below, such evidence does not currently
Department of Diagnostic and Interventional Radiology, Rostock exist with regard to coronary CTA screening for CAD.
University Medical Center, Rostock, Germany
e-mail: [email protected]
M. Renker
Department of Cardiology, Kerckhoff Heart and Thorax Center,
Bad Nauheim, Germany

© Humana Press 2019 739

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_58
740 F. G. Meinel and M. Renker

Utilization  ssociation with Coronary Artery Calcium

A
Score
Although the benefit/risk profile of coronary CTA in asymp-
tomatic individuals remains controversial, it is not infrequently The amount of coronary artery calcium as determined by CT
performed. The CONFIRM registry enrolled 27,125 patients calcium scoring is predictive of coronary CTA findings in
who underwent coronary CTA in six countries between 2003 asymptomatic individuals. In a 2015 analysis of the
and 2009; at least 7590 (28%) of these individuals were CONFIRM registry, the prevalence of obstructive CAD was
asymptomatic for coronary artery disease [6]. In a 2009 survey 42% in patients with a CACS of >100 compared to 9% in
of 169 institutions providing cardiac CT in 38 countries, 34% patients with a CACS of ≤100 [16].
reported performing cardiac CT “for the exclusion of CAD in
clinically healthy patients” [7]. A more recent survey (con-
ducted 2014–2015) analyzed the patterns of coronary CTA Association with Clinical Risk Factors
utilization in 152 tertiary care hospitals with advanced CT sys-
tems across China [8]. As many as 95% of these hospitals Not surprisingly, the diagnostic yield of coronary CTA
listed “screening to rule out coronary artery disease in asymp- increases in cohorts with cardiovascular risk factors. This
tomatic individuals” as one of their clinical indications for per- has been shown for smoking, diabetes, hypertension, and
forming coronary CTA studies. The use of coronary CTA in chronic kidney disease (Table 58.4). The prevalence of sub-
asymptomatic patients as part of self-referred health “checkup” clinical CAD in asymptomatic patients is also associated
examinations has indeed become a growing trend, particularly with less well-established cardiovascular risk factors such as
in Asia. But also in Western countries, it is not uncommon to the inflammatory marker C-reactive protein and subclinical
see radiology or cardiology practices advertising coronary CT hypothyroidism (Table  58.4). Thus, larger cohorts of coro-
angiography to the public for “checkup” examinations. nary CTA performed in asymptomatic individuals provide an
opportunity to investigate the effect of established cardiovas-
cular risk factors on subclinical coronary atherosclerosis and
Diagnostic Yield identify novel risk factors for CAD.

 symptomatic Individuals with Low

A
Cardiovascular Risk Table 58.4  Coronary CTA findings in asymptomatic individuals: cor-
relation with cardiovascular risk factors
Table 58.3 provides an overview of published cohorts of low-­ Correlation with coronary CTA
risk, asymptomatic patients who underwent coronary Risk factor findings References
CTA. In individuals with a calcium score of zero, a p­ revalence Chronic kidney Yes [17]
disease
of nonobstructive plaque on the order of 6% and a prevalence
Arterial stiffness Yes [18]
of obstructive CAD from non-calcified plaque of <1% have
Smoking Yes [19, 20]
been reported (Table  58.3). Similar prevalence rates were Hypertension Yes [21]
found in a cohort of young adults under the age of 45 (mean Diabetes Yes [22]
age 40, Table  58.3). Others have reported a prevalence of Metabolic syndrome Yes [23]
33–58% for nonobstructive CAD and 4–5% for obstructive C-reactive protein Yes [24]
CAD even in such relatively low-risk groups as normal- Subclinical Yes [25]
weight individuals and middle-aged sportsmen (Table 58.3). hypothyroidism

Table 58.3  Major publications on the diagnostic yield of coronary CTA in asymptomatic low-risk patients
Age Obstructive CAD
Author Year Target population n (yearsa) No CAD(%) Nonobstructive CAD(%) (%) Reference
Braber TL 2016 Middle-aged sportsmen 318 54.7 36.8 57.9 5.3 [9]
Kim S 2015 Normal weight 2078 53.4 63.1 33.2 3.7 [10]
Lee MS 2013 Zero calcium score 6531 49.8 93.2 6.1 0.7 [11]
Cho I 2013 Zero calcium score 4491 48 93.0 6.2 0.8 [12]
Kim KJ 2013 Low risk 2133 48.7 87.3 11.4 1.3 [13]
Jin KN 2012 Young adults (<45 years) 914 40.4 90.6 8.4 1.0 [14]
Yoo DH 2011 Zero calcium score 6040 50.1 93.1 6.1 0.8 [15]
Publications with n ≥ 200 are listed
Mean or median
a
58  Coronary CT Angiography for Screening, Risk Stratification, and Management of Asymptomatic Patients: State… 741

Specific High-Risk Groups CONFIRM registry, a population of 7590 subjects without

chest pain and without a history of coronary artery disease
Studies on coronary CTA screening for asymptomatic CAD was followed for a median period of 24 months [6]. This was
have been performed in a number of specific high-risk groups a mixed population of patients with Framingham risk scores
(Table 58.5) with the largest number of publications focusing ranging from low to high. All-cause mortality as well as a
on patients with diabetes mellitus. In patients with diabetes composite of all-cause mortality and nonfatal myocardial
without symptoms of CAD, only 20–31% have normal coro- infarction served as endpoints in this analysis. While both
nary arteries on coronary CTA, while 36–48% show nonob- CACS and coronary CTA improved risk stratification beyond
structive CAD and 23–44% obstructive CAD (Table 58.6). standard clinical risk factors, the incremental benefit of
Patients with familial hypercholesterolemia represent ­adding coronary CTA to clinical risk factors and CACS was
another specific high-risk group with an exceptionally high minimal and not considered clinically meaningful [6].
expected diagnostic yield of coronary CTA.  In a cohort of
101 statin-treated asymptomatic patients with familial hyper-
cholesterolemia, coronary CTA demonstrated nonobstruc-  ssociation with Coronary Artery Calcium
A
tive CAD in 59% of patients and obstructive CAD in 26% Score
[34]. HIV infection has been recognized as a major risk fac-
tor for accelerated atherosclerosis. Indeed, in one cohort of The CONFIRM registry investigators also analyzed the
55 HIV-infected individuals with low cardiovascular risk (as prognostic value of coronary CTA findings in relationship to
determined by traditional risk factors), coronary CTA the calcium score [16]. In their analysis of 3217 asymptom-
revealed a 29% prevalence of obstructive CAD [44]. atic patients, coronary CTA had incremental prognostic
value over clinical risk factors (Framingham risk score) for
patients with a CACS of >100 but not for those with a CACS
Prognostic Value of ≤100 [16]. In the group of asymptomatic individuals with
a CACS of >100, the incremental benefit of coronary CTA
General Asymptomatic Population for risk prediction was greatest in patients with an intermedi-
ate CACS (101–400) and declined in patients with higher
Does coronary CTA have incremental value over CACS in CACS (>400). The primary conclusion of this study was that
asymptomatic subjects? In one publication from the coronary CTA for risk stratification in asymptomatic patients
may be appropriate in patients with an intermediate CACS
(101–400) but is probably not useful in patients with lower
Table 58.5  Publications on coronary CTA in asymptomatic individu- or higher CACS.
als from specific high-risk groups
Risk group References
Diabetes mellitus [26–33]
Association with Clinical Risk Factors
Familial hypercholesterolemia [34–36]
Peripheral or cerebral artery disease [37–40]
Status post mediastinal radiotherapy for Hodgkin’s [41, 42] Other authors looked at more selected groups of patients with
lymphoma cardiovascular risk factors. One published cohort consisted of
Morbid obesity [43] 711 asymptomatic patients with a “high a priori risk of CAD”,
HIV positive [44–46] which was defined as either a high estimated cardiovascular
Smokers [19, 20] lifetime risk or a borderline or mildly abnormal ECG-
Family history of early-onset coronary artery disease [47] treadmill stress test result [48]. The primary e­ ndpoint was a

Table 58.6  Major publications on the diagnostic yield of coronary CTA in asymptomatic patients with diabetes
Age Obstructive CAD
Author Year n (yearsa) Diabetes duration(yearsa) No CAD(%) Nonobstructive CAD(%) (%) Reference
Kang SH 2016 591 62.2 12.5 28.4 39.9 31.6 [27]
Halon DA 2016 630 63.5 10.1 20.6 48.3 31.1 [28]
Kim JJ 2015 933 63.4 11.7 20.7 39.1 40.1 [29]
Kim JJ 2015 284 64.9 13.1 20.1 36.3 43.7 [30]
Dedic A 2015 378 56 Not stated 22.8 44.4 32.8 [31]
Idilman IS 2015 273 58.6 4.7 24.2 40.6 35.2 [32]
Min JK 2014 400 60.4 Not stated 30.0 42.2 27.8 [33]
Muhlestein JB 2014 336 61.5 12.3 31.3 46.1 22.6 [26]
Publications with n ≥ 200 are listed
Mean or median
a
742 F. G. Meinel and M. Renker

composite of cardiac death and nonfatal myocardial infarc- s­creening as well, so that aggressive risk factor reduction
tion. Findings at coronary CTA were significant predictors of appears to be the most important management approach in
events during a mean follow-up of 2.7 years. In particular, a high-risk individuals. Beyond lifestyle changes (e.g., dietary
segment involvement score of ≥5 was significantly associated measures, physical activity, weight reduction, and avoidance
with adverse events. However, this analysis did not adjust for of any tobacco exposure), optimal medical treatment is rec-
CACS and therefore does not demonstrate an incremental ommended to meet specific treatment targets that are adjusted
prognostic value of coronary CTA over CACS. to proposed risk categories. For low-density lipoprotein cho-
lesterol (LDL-C, the primary lipid analysis with respect to
management), the treatment target approach is presently a
Specific High-Risk Groups consensus recommendation comprised in the 2016 European
guidelines on the management of dyslipidemias for applica-
Six relatively large cohorts have investigated the prognostic bility and compliance reasons, while its benefit is not conclu-
value of coronary CTA in asymptomatic patients with diabe- sively proven [50]. Explicit LDL-C target levels were
tes (Table 58.7). All found a substantially increased hazard discontinued in the 2013  US national guidelines [51]. The
for adverse events in asymptomatic patients with obstructive rationale and details for non-lipid treatment strategies and
CAD at CTA (Table  58.7), and all but one also reported a goals for cardiovascular disease prevention are covered in
(smaller) risk increase in patients with nonobstructive CAD recent national guidelines such as those on the assessment of
compared to those with no CAD.  Three of these cohorts cardiovascular risk [52], lifestyle, diet and exercise [53], and
adjusted the prognostic value of coronary CTA for results of management of obesity [54].
CACS and found an incremental prognostic value of coro-
nary CTA findings over CACS in asymptomatic individuals
with diabetes. One cohort study investigated the prognostic  reatment Implications by Coronary Calcium
T
value of coronary CTA in 70 asymptomatic patients on Scoring
hemodialysis [49]. In this relatively small cohort, cumulative
event rates of 36% vs. 0% in patients with vs. without As indicated above, CAC testing has been validated by mul-
obstructive CAD at CTA were reported. tiple studies and has been shown to be a very robust predictor
of adverse cardiovascular events in the general population, as
well as in specific risk groups such as the elderly and
Role for Management ­diabetics. Importantly, CAC has been shown to help guide
medical therapy. A CAC score ≥75th percentile (for age and
 rimary Prevention in Individuals at Risk
P gender) or ≥300 Agatston units is considered high risk by
for Coronary Artery Disease means of the 2013 ACC/AHA guideline on the treatment of
blood cholesterol to reduce atherosclerotic cardiovascular
Implementing current state of knowledge, national and risk in adults and warrant high-dose statins [52]. In the ran-
European guidelines advocate primary prevention in terms domized St. Francis Heart study, atorvastatin reduced cardio-
of cardiovascular risk factor reduction as the most important vascular events by 42% in those with a CAC score >400
strategy in any individual being at risk for CAD. These Agatston units (number needed to treat for the reduction of 1
guidelines therefore pertain to potential candidates of myocardial infarction or death, 17) [55].

Table 58.7  Major publications on the prognostic value of coronary CTA in asymptomatic patients with diabetes
HR for
obstructive
Age Diabetes CAD(most
Author Year n (yearsa) duration(yearsa) Obstructive CAD (%) adjusted) 95%-CI Reference
Kang SH 2016 591 62.2 12.5 31.6 14.4 1.6–134.7 [27]
Halon DA 2016 630 63.5 10.1 31.1 6.6 3.3–13.1 [28]
Kim JJ 2015 933 63.4 11.7 40.1 2.0 1.2–3.3 [29]
Dedic A 2015 378 56 Not stated 32.8 9.5 0.9–104.2 [31]
Min JK 2014 400 60.4 Not stated 27.8 1.8b 1.2–2.82 [33]
Muhlestein JB 2014 336 61.5 12.3 22.6 5.4 1.0–27.7 [26]
Publications with n ≥ 200 are listed
a
Mean or median
b
The HR in the paper by Min et al. refers to each grade increment in maximum stenosis severity (0%, 1–49%, 50–69%, >70%)
58  Coronary CT Angiography for Screening, Risk Stratification, and Management of Asymptomatic Patients: State… 743

Management recommendations based on current guide- Impact

lines are as follows [52]:
Only one randomized controlled trial has evaluated the effec-
–– CAC score ≥75th percentile or ≥300 Agatston units tiveness of coronary CTA screening to reduce cardiovascular
should be treated with high-dose statins. morbidity and mortality in asymptomatic individuals. In the
–– CAC score <75th percentile and<300 Agatston units FACTOR-64 trial, 900 asymptomatic individuals with diabe-
should be treated with low−/moderate-dose statins. tes were randomized to CAD screening with coronary CTA
–– CAC score  =  0 should be considered for lifestyle or optimal diabetes care according to established standards
modification. [26]. In the coronary CTA arm, the CT results were used for
clinical decision-making. Individuals with normal coronary
In the multiethnic study of atherosclerosis (MESA), arteries were recommended to continue standard diabetes
individuals with a CAC score ≥100 Agatston units showed mellitus care. Participants with evidence of CAD by coro-
an estimated benefit for acetylsalicylic acid regardless of nary CTA or a coronary artery calcium score >10 were rec-
risk factors (number needed to treat = 92, estimated 5-year ommended to begin a more aggressive risk factor modification
number needed to harm  =  442 for major bleeding) [56]. including tighter control of LDL-C, triglycerides, HbA1C,
Reversely, a benefit for acetylsalicylic acid in individuals and blood pressure. Additionally, patients with moderate
with CAC of 0 Agatston units was very unlikely (5-year (50–69%) stenosis were recommended to receive stress car-
number needed to treat = 2036 for individuals with low car- diac imaging, and patients with severe (≥70%) stenosis were
diovascular risk and 808 for elevated traditional risk status; recommended to undergo diagnostic coronary angiography.
5-year number needed to harm = 442 for major bleeding). Patients were followed for a mean of 4 years. The trial’s pri-
Comparable data exist for angiotensin-converting enzyme mary outcome was a composite endpoint of all-cause mortal-
inhibitors. ity, nonfatal myocardial infarction, and unstable angina
requiring hospitalization.
The trial’s main result was that coronary CTA screening
 reatment Implications by Coronary CT
T and CTA-directed management did not significantly reduce
Angiography event rates compared to the control group (6.2% vs. 7.6%;
hazard ratio, 0.80 [95% confidence interval, 0.49–1.32];
Through optimized downstream prescription of preventive P = 0.38). Although discouraging, this result should not be
therapies and altered use of myocardial revascularization regarded as refuting the idea of screening coronary CTA in
procedures, the SCOT-HEART trial has demonstrated out- high-risk asymptomatic individuals once and for all. The
come modification by coronary CTA in patients presenting event rate in the FACTOR-64 trial was low, much lower than
with suspected angina due to CAD [57]. Altogether, coro- the estimation that was used for sample size calculations.
nary CTA was associated with a reduction of fatal/nonfatal Hence, the trial was not powered to detect a moderate risk
myocardial infarction of approximately 50% in this symp- reduction. The rate of coronary CTA-driven revasculariza-
tomatic population. tions was also low (5.8%). These observations suggest that
For asymptomatic patients, there is currently no conclu- the trial participants had a relatively low-risk profile and/or
sive evidence that coronary CTA-directed medical manage- that the standard care was unusually stringent and effective.
ment or revascularization reduces morbidity or mortality Therefore, further randomized trials with larger sample sizes
(see paragraph “Impact” below). Likewise, there are no and higher-risk inclusion criteria (e.g., longer-standing dia-
guidelines available on how coronary CTA findings should betes) are warranted. There was also some “contamination”
guide the management of asymptomatic patients. If coro- in the trial as only 395 (87.4%) of the 452 participants in the
nary CTA is performed in asymptomatic patients at risk for CTA arm actually received a CT scan. In an as-treated analy-
CAD, it appears reasonable to manage these patients similar sis, the event rates for the primary endpoint were 5.6% vs.
to patients with stable chest pain. Patients with moderate 7.9% (HR, 0.69 [95% CI, 0.41–1.16]; P = 0.16) indicating a
(50–69%) stenosis should be recommended to receive stress trend toward a benefit of CTA screening.
cardiac imaging (stress MRI or nuclear stress test) to detect
myocardial ischemia, and patients with severe (≥70%) ste-
nosis should be recommended to undergo diagnostic coro- Conclusions
nary angiography. Additionally, aggressive risk factor
modification and low-dose acetylsalicylic acid should be Screening for CAD in asymptomatic individuals is generally
considered in all patients with more than minimal CAD at not considered an appropriate application of coronary CTA
coronary CTA. by existing guidelines. Nevertheless, the practice is not
744 F. G. Meinel and M. Renker

uncommon. The diagnostic yield of coronary CTA in low-­ 8. Liu K, et al. Current utilization of cardiac computed tomography in
mainland China: a national survey. J Cardiovasc Comput Tomogr.
risk asymptomatic patients is low with obstructive coronary 2016;10(1):76–81.
artery disease found in approximately 1% of low-risk indi- 9. Braber TL, et  al. Occult coronary artery disease in middle-aged
viduals in most studies. The prevalence of asymptomatic sportsmen with a low cardiovascular risk score: the Measuring
CAD increases with the presence of established cardiovascu- Athlete's Risk of Cardiovascular Events (MARC) study. Eur J Prev
Cardiol. 2016;23(15):1677–84.
lar risk factors such as smoking, hypertension, diabetes, and 10. Kim S, et al. Normal-weight obesity is associated with increased risk
chronic kidney disease. Coronary CTA has a particularly of subclinical atherosclerosis. Cardiovasc Diabetol. 2015;14:58.
high diagnostic yield in specific high-risk subgroups of 11. Lee MS, et al. Asymptomatic subjects with zero coronary calcium
asymptomatic individuals such as patients with diabetes or score: coronary CT angiographic features of plaques in event-prone
patients. Int J Cardiovasc Imaging. 2013;29(Suppl 1):29–36.
familial hypercholesterolemia. In asymptomatic patients 12. Cho I, et al. Prevalence and prognostic implication of non-calcified
with diabetes, only 20–31% have normal coronary arteries plaque in asymptomatic population with coronary artery calcium
on coronary CTA with 36–48% showing nonobstructive score of zero. Korean Circ J. 2013;43(3):154–60.
CAD and 23–44% showing obstructive CAD.  Findings at 13. Kim KJ, et al. The prevalence and characteristics of coronary ath-
erosclerosis in asymptomatic subjects classified as low risk based
coronary CTA have incremental prognostic value over coro- on traditional risk stratification algorithm: assessment with coro-
nary artery calcium scoring in high-risk groups (in particular, nary CT angiography. Heart. 2013;99(15):1113–7.
patients with diabetes), while the incremental prognostic 14. Jin KN, et al. Subclinical coronary atherosclerosis in young adults:
value is probably not meaningful in a more mixed popula- prevalence, characteristics, predictors with coronary computed
tomography angiography. Int J Cardiovasc Imaging. 2012;28(Suppl
tion. The only randomized controlled trial on the impact of 2):93–100.
coronary CTA screening in asymptomatic diabetic patients 15. Yoo DH, et  al. Significance of noncalcified coronary plaque in
did not show a significant reduction in cardiovascular events. asymptomatic subjects with low coronary artery calcium score:
However, screening for CAD in asymptomatic high-risk assessment with coronary computed tomography angiography. Int J
Cardiovasc Imaging. 2011;27(Suppl 1):27–35.
individuals – in particular, patients with diabetes – remains a 16. Cho I, et  al. Incremental prognostic utility of coronary CT angi-
possible future application of coronary CTA. More random- ography for asymptomatic patients based upon extent and sever-
ized trials are needed to assess whether coronary CTA ity of coronary artery calcium: results from the COronary CT
screening can decrease morbidity and mortality in these Angiography EvaluatioN For Clinical Outcomes InteRnational
Multicenter (CONFIRM) study. Eur Heart J. 2015;36(8):501–8.
patients. 17. Cho I, et  al. Coronary atherosclerosis detected by coronary CT
angiography in asymptomatic subjects with early chronic kidney
disease. Atherosclerosis. 2010;208(2):406–11.
References 18. Nam HJ, et  al. Association between brachial-ankle pulse wave
velocity and occult coronary artery disease detected by multi-­
detector computed tomography. Int J Cardiol. 2012;157(2):227–32.
1. Wolk MJ, et  al. ACCF/AHA/ASE/ASNC/HFSA/HRS/SCAI/
19. Yi M, et al. Coronary CT angiography findings based on smoking
SCCT/SCMR/STS 2013 multimodality appropriate use criteria
status: do ex-smokers and never-smokers share a low probability
for the detection and risk assessment of stable ischemic heart dis-
of developing coronary atherosclerosis? Int J Cardiovasc Imaging.
ease: a report of the American College of Cardiology Foundation
2015;31(Suppl 2):169–76.
Appropriate Use Criteria Task Force, American Heart Association,
20. Kim JA, et al. Relationship between amount of cigarette smoking
American Society of Echocardiography, American Society of
and coronary atherosclerosis on coronary CTA in asymptomatic
Nuclear Cardiology, Heart Failure Society of America, Heart
individuals. Int J Cardiovasc Imaging. 2013;29(Suppl 1):21–8.
Rhythm Society, Society for Cardiovascular Angiography and
21. Im TS, et al. Grade-response relationship between blood pressure
Interventions, Society of Cardiovascular Computed Tomography,
and severity of coronary atherosclerosis in asymptomatic adults:
Society for Cardiovascular Magnetic Resonance, and Society of
assessment with coronary CT angiography. Int J Cardiovasc
Thoracic Surgeons. J Am Coll Cardiol. 2014;63(4):380–406.
Imaging. 2014;30(Suppl 2):105–12.
2. Meinel FG, et al. Radiation risks from cardiovascular imaging tests.
22. Gurudevan S, et  al. Impaired fasting glucose is associated with
Circulation. 2014;130(5):442–5.
increased severity of subclinical coronary artery disease compared
3. Meinel FG, et al. Contrast-induced acute kidney injury: definition,
to patients with diabetes and normal fasting glucose: evaluation
epidemiology, and outcome. Biomed Res Int. 2014;2014:859328.
by coronary computed tomographic angiography. BMJ Open.
4. Wilson JMG, Jungner G. Principles and practice of screening for
2016;6(8):e005148.
disease. Public Health Paper Number 34. 1968 [cited 18 Oct 2016];
23. Ryu J, et  al. Relation of coronary atherosclerosis and metabolic
Available from: http://www.who.int/iris/handle/10665/37650.
syndrome in asymptomatic subjects: evaluation with coronary
5. Andermann A, et al. Revisiting Wilson and Jungner in the genomic
CT angiography. Int J Cardiovasc Imaging. 2013;29(Suppl 2):
age: a review of screening criteria over the past 40 years. Bull
101–7.
World Health Organ. 2008;86(4):317–9.
24. Yang DH, et al. Association between C-reactive protein and type of
6. Cho I, et al. Coronary computed tomographic angiography and risk
coronary arterial plaque in asymptomatic patients: assessment with
of all-cause mortality and nonfatal myocardial infarction in subjects
coronary CT angiography. Radiology. 2014;272(3):665–73.
without chest pain syndrome from the CONFIRM registry (coro-
25. Park YJ, et al. Impact of subclinical hypothyroidism on the coro-
nary CT angiography evaluation for clinical outcomes: an interna-
nary artery disease in apparently healthy subjects. Eur J Endocrinol.
tional multicenter registry). Circulation. 2012;126(3):304–13.
2011;165(1):115–21.
7. Maurer MH, et  al. Indications, imaging technique, and reading
26. Muhlestein JB, et al. Effect of screening for coronary artery disease
of cardiac computed tomography: survey of clinical practice. Eur
using CT angiography on mortality and cardiac events in high-risk
Radiol. 2012;22(1):59–72.
58  Coronary CT Angiography for Screening, Risk Stratification, and Management of Asymptomatic Patients: State… 745

patients with diabetes: the FACTOR-64 randomized clinical trial. c­ omputed tomography angiography: results and risk factor analy-
JAMA. 2014;312(21):2234–43. sis. Int J Radiat Oncol Biol Phys. 2014;89(1):59–66.
27. Kang SH, et al. Long-term prognostic value of coronary CT angiog- 43. Lubanski MS, et al. Subclinical coronary atherosclerosis identified
raphy in asymptomatic type 2 diabetes mellitus. JACC Cardiovasc by coronary computed tomographic angiography in asymptomatic
Imaging. 2016;9(11):1292–300. morbidly obese patients. Heart Int. 2010;5(2):e15.
28. Halon DA, et  al. Coronary computed tomography (CT) angiog- 44. d’Ettorre G, et al. Significant coronary stenosis detected by coro-
raphy as a predictor of cardiac and noncardiac vascular events in nary computed angiography in asymptomatic HIV infected sub-
asymptomatic type 2 diabetics: a 7-year population-based cohort jects. J Infect. 2012;64(1):82–8.
study. J Am Heart Assoc. 2016;5(6):e003226. 45. Fitch KV, et  al. Subclinical myocyte injury, fibrosis and strain in
29. Kim JJ, et  al. Impact of diabetes duration on the extent and
relationship to coronary plaque in asymptomatic HIV-infected indi-
severity of coronary atheroma burden and long-term clinical viduals. AIDS. 2016;30(14):2205–14.
outcome in asymptomatic type 2 diabetic patients: evaluation 46. Ketlogetswe KS, et al. Lower adiponectin is associated with sub-
by coronary CT angiography. Eur Heart J Cardiovasc Imaging. clinical cardiovascular disease among HIV-infected men. AIDS.
2015;16(10):1065–73. 2014;28(6):901–9.
30. Kim JJ, et  al. A prospective two-center study on the associations 47. Kral BG, et  al. Noncalcified coronary plaque volumes in healthy
between microalbuminuria, coronary atherosclerosis and long-term people with a family history of early onset coronary artery disease.
clinical outcome in asymptomatic patients with type 2 diabetes Circ Cardiovasc Imaging. 2014;7(3):446–53.
mellitus: evaluation by coronary CT angiography. Int J Cardiovasc 48. Plank F, et al. The diagnostic and prognostic value of coronary CT
Imaging. 2015;31(1):193–203. angiography in asymptomatic high-risk patients: a cohort study.
31. Dedic A, et  al. Prognostic value of coronary computed tomogra- Open Heart. 2014;1(1):e000096.
phy imaging in patients at high risk without symptoms of coronary 49. de Bie MK, et  al. CT coronary angiography is feasible for the
artery disease. Am J Cardiol. 2016;117(5):768–74. assessment of coronary artery disease in chronic dialysis patients,
32. Idilman IS, et  al. Nonalcoholic fatty liver disease is associ-
despite high average calcium scores. PLoS One. 2013;8(7):e67936.
ated with significant coronary artery disease in type 2 diabetic 50. Catapano AL, et al. 2016 ESC/EAS Guidelines for the Management
patients: a computed tomography angiography study 2. J Diabetes. of Dyslipidaemias: The Task Force for the Management of
2015;7(2):279–86. Dyslipidaemias of the European Society of Cardiology (ESC)
33. Min JK, et al. Incremental prognostic value of coronary computed and European Atherosclerosis Society (EAS)Developed with
tomographic angiography over coronary artery calcium score for the special contribution of the European Association for
risk prediction of major adverse cardiac events in asymptomatic Cardiovascular Prevention & Rehabilitation (EACPR). Eur Heart J.
diabetic individuals. Atherosclerosis. 2014;232(2):298–304. 2016;37(39):2999–3058.
34. Neefjes LA, et  al. CT coronary plaque burden in asymp-
51. Stone NJ, et  al. 2013 ACC/AHA guideline on the treatment of
tomatic patients with familial hypercholesterolaemia. Heart. blood cholesterol to reduce atherosclerotic cardiovascular risk in
2011;97(14):1151–7. adults: a report of the American College of Cardiology/American
35. Ten Kate GJ, et al. The effect of LDLR-negative genotype on CT Heart Association Task Force on Practice Guidelines. J Am Coll
coronary atherosclerosis in asymptomatic statin treated patients Cardiol. 2014;63(25 Pt B):2889–934.
with heterozygous familial hypercholesterolemia. Atherosclerosis. 52. Goff DC Jr, et al. 2013 ACC/AHA guideline on the assessment of
2013;227(2):334–41. cardiovascular risk: a report of the American College of Cardiology/
36. Neefjes LA, et  al. Accelerated subclinical coronary atherosclero- American Heart Association Task Force on Practice Guidelines.
sis in patients with familial hypercholesterolemia. Atherosclerosis. Circulation. 2014;129(25 Suppl 2):S49–73.
2011;219(2):721–7. 53. Eckel RH, et  al. 2013 AHA/ACC guideline on lifestyle manage-
37. Hromadka M, et al. Incidence of severe coronary stenosis in asymp- ment to reduce cardiovascular risk: a report of the American
tomatic patients with peripheral arterial disease scheduled for major College of Cardiology/American Heart Association Task Force on
vascular surgery. Int Angiol. 2016;35(4):411–7. Practice Guidelines. J Am Coll Cardiol. 2014;63(25 Pt B):2960–84.
38. Calvet D, et  al. Predicting asymptomatic coronary artery disease 54. Jensen MD, et al. 2013 AHA/ACC/TOS guideline for the manage-
in patients with ischemic stroke and transient ischemic attack: the ment of overweight and obesity in adults: a report of the American
PRECORIS score. Stroke. 2014;45(1):82–6. College of Cardiology/American Heart Association Task Force on
39. den Dekker MA, et  al. Prevalence of severe subclinical coronary Practice Guidelines and The Obesity Society. J Am Coll Cardiol.
artery disease on cardiac CT and MRI in patients with extra- 2014;63(25 Pt B):2985–3023.
cardiac arterial disease. Eur J Vasc Endovasc Surg. 2013;46(6): 55. Arad Y, et  al. Coronary calcification, coronary disease risk fac-
680–9. tors, C-reactive protein, and atherosclerotic cardiovascular dis-
40. Yoon YE, et  al. Incidence of subclinical coronary atherosclerosis ease events: the St. Francis heart study. J Am Coll Cardiol.
in patients with suspected embolic stroke using cardiac computed 2005;46(1):158–65.
tomography. Int J Cardiovasc Imaging. 2011;27(7):1035–44. 56. Miedema MD, et al. Use of coronary artery calcium testing to guide
41. Daniels LA, et  al. Screening for coronary artery disease after
aspirin utilization for primary prevention: estimates from the multi-­
mediastinal irradiation in Hodgkin lymphoma survivors: ethnic study of atherosclerosis. Circ Cardiovasc Qual Outcomes.
phase II study of indication and acceptancedagger. Ann Oncol. 2014;7(3):453–60.
2014;25(6):1198–203. 57. Williams MC, et al. Use of coronary computed tomographic angi-
42. Girinsky T, et  al. Prospective coronary heart disease screening ography to guide management of Patients with coronary disease.
in asymptomatic Hodgkin lymphoma patients using coronary J Am Coll Cardiol. 2016;67(15):1759–68.
 Part XV
Where We Are Going: Lesion-specific Ischemia,
Infarction, and Viability
 Transluminal Attenuation Gradient
and Other CT Techniques for Gauging 59
Lesion Significance

Yeon Hyeon Choe, Jin-Ho Choi, and Sung Mok Kim

Coronary CT angiography (CCTA) has emerged as the Theoretical Background of TAG

noninvasive modality of choice for imaging of the coro-
nary arteries and serves as a gatekeeper to interventional Enhancement in the vessel lumen depends on various ­factors
treatment of coronary artery disease (CAD) [1, 2]. including contrast volume and concentration, injection
However, severe motion and blooming artifacts are associ- speed, CT scan protocols including tube voltage (kVp), as
ated with elevated heart rates, calcifications, and stents, well as patient body habitus and sex [8]. Contrast attenuation-­
respectively, and can often times hinder the evaluation of time curves in a certain anatomical region can depict
image sets. Despite the high diagnostic performance of upslope, peak, and downslope (Fig. 59.1). Blood in the cor-
CCTA for the detection of stenosis, not all lesions are onary artery propagates from the proximal segment to distal
hemodynamically significant. Current CT technology does segments as time lapses. The normal mean flow velocity in
not allow for clear vessel wall-lumen delineation due to the coronary arteries is 40 ± 19 cm/s (peak diastolic veloc-
limited spatial resolution; however, the use of commercial- ity, 64 ± 26 cm/s; peak systolic velocity, 34 ± 22 cm/s) using
ized software tools can help radiologists detect stenotic Doppler guidewire technique [9]. The peak diastolic veloc-
lesions semiautomatically [3, 4]. Notably, even with the ity in the distal left anterior descending coronary artery
use of software, manual interaction is necessary to define (LAD) is 21.2 ± 7.9 cm/s using transthoracic Doppler echo-
the luminal contour. Thus, efforts have been made to cardiography [10].
enhance the diagnostic capability of CCTA through the use On CCTA image sets, small decreases in contrast opaci-
of CT fractional flow reserve (CT-FFR), vasodilator stress fication were observed going from proximal to distal seg-
myocardial CT perfusion (CTP), and transluminal attenua- ments of the coronary arteries [11]. The attenuation
tion gradients (TAG) [5–7]. gradients were steeper in patients with obstructive coronary
artery diseases [11–13]. According to Steigner et al. [11],
three distinct coronary contrast opacification gradients
(gradients according to distance from the coronary ostium
[G(d)], lumen cross-­sectional area [G(a)], and lumen short-
axis diameter [G(s)]) could be measured within a single
heartbeat using a 320-­slice CT scanner. Notably, gradient
variations between cardiac phases, heart rates, body habi-
Y. H. Choe (*)
Department of Radiology, HVSI Imaging Center, Heart Vascular tus, and readers were low. Gradients in patients with lesions
Stroke Institute, Samsung Medical Center, Sungkyunkwan were significantly different (p < 0.021) than in patients con-
University School of Medicine, Seoul, South Korea sidered normal at CCTA.  For all patients, the gradients
e-mail: [email protected]
defined with respect to the coronary lumen cross-sectional
J.-H. Choi area and short-axis diameters were highly linear, not sig-
Division of Cardiology, Department of Medicine (J-H C), Samsung
nificantly influenced by the coronary artery (left anterior
Medical Center, Sungkyunkwan University School of Medicine,
Seoul, South Korea descending artery [LAD] versus left circumflex [LCX] ver-
sus right coronary artery [RCA]), and only have minute
S. M. Kim
Department of Radiology, HVSI Imaging Center, Heart Vascular variations with respect to individual patient parameters.
Stroke Institute, Seoul, South Korea While 64-slice CT results also showed abnormalities in

© Humana Press 2019 749

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_59
 750 Y. H. Choe et al.

600 Current CT software analysis techniques allow for easy

Cornary Artery Attenuation (HU)

analysis of TAG assuming vessel centerlines are drawn cor-

500
rectly. Automated gradient calculation software shows no
Proximal LAD difference between manual and automated methods despite
400
Proximal RCA varying body mass index categories, presence and severity of
300 coronary artery disease, plaque composition, or reconstruc-
tion algorithms [18].
200 Notably, the LAD coronary artery is different from the
RCA in TAG, as the LAD has more bifurcations such as sep-
100
tal and diagonal branches [14, 17]. Therefore, TAG may vary
0 according to coronary anatomy. For example, the LAD and
6 8 10 12 14 16 18 20 22 24 26 28 LCX coronary arteries taper more rapidly along the vessel
Time After Contrast Injection (sec) length than the RCA. TAG may be associated with branching
of coronary arteries and is relatively constant in the segments
Fig. 59.1  Time-attenuation curves of the coronary arteries. The graph
shows time-attenuation curves of the coronary arteries (left anterior between the branches, while TAGs of larger and smaller ves-
descending branch, LAD; right coronary artery, RCA) obtained from sels may be different.
two different patients, respectively. Dynamic CT was performed under
adenosine stress after a 50 mL bolus of contrast material (400 Ig/mL)
was administered at 5 mL/s
Coronary Flow Measurement with CT

TAG values in patients with obstructive CAD [12], Park Lackner et al. [19] suggested that the CT data in an experi-
et al. claimed that a decrease in TAG is associated with a mental model simultaneously acquired for the coronary and
decrease in coronary luminal diameter (transluminal diam- aorta time-density curves enables qualitative and semiquan-
eter gradient [TDG]) [14]. titative assessment of stenotic changes in flow. In a study by
Bovenschulte et  al., the ratio of upslopes of the coronary
artery and aorta, derived from the time-density curve data
Techniques for TAG using a 64-slice CT, was used as a measure of the difference
in the slope of density increase in the coronary artery and
Theoretically, whole-heart coverage at a single gantry rota- aorta [20]. All examinations of hemodynamically relevant
tion in a single heartbeat using 256-slice or 320-slice CT is stenosis (≥70%) and insignificant stenoses produced ratios
advantageous for TAG measurement. Attenuation along a of ≤0.55 and ≥0.77, respectively.
vessel does not reflect contrast density at a single time point Lardo et al. [21] proposed a new method for noninvasive
using a 64-slice CT [12] (Fig. 59.2). measurement of absolute coronary blood flow (CBF)
Contrast enhancement should be strong enough to allow termed transluminal attenuation flow encoding (TAFE).
automatic detection of vessel luminal contour using dedi- They applied TAFE to calculate absolute CBF using four
cated software. TAG can be affected by scan timing [15], vessel input parameters including TAG, cross-sectional
rendering scans optimal before the peak enhancement of area, length, and the contrast bolus duration derived from
the coronary artery [15, 16]. Additionally, it is desirable to the arterial input function in animal studies (Fig.  59.4).
obtain CT images during the upslope period (ascending TAFE-derived CBF divided by myocardial mass strongly
leg) of the time-attenuation curve of CCTA [15], as dynamic correlated with microsphere myocardial blood flow
scanning shows increasing TAG at time points after peak (R2  =  0.90, p  <  0.001). In human studies, TAFE-derived
attenuation in the coronary arteries (Fig.  59.3) [14]. TAG CBF in the LAD, LCX, and RCA was 26.4 ± 10.7 mL/min,
may fluctuate along the course of the coronary tree due to 20.1 ± 13.0 mL/min, and 43.2 ± 40.9 mL/min, respectively.
CT technical causes, for example, motion artifacts and vari- CBF per unit mass was 0.93 ± 0.48 mL/g/min. Interobserver
ations in photon flux. Chow et  al. [17] used corrected variability was minimal with excellent correlation
­coronary opacification (CCO) within coronary lumen nor- (R  =  0.96, p  <  0.0001) and agreement (mean difference,
malized to the aorta. 4.2 mL/min).
59  Transluminal Attenuation Gradient and Other CT Techniques for Gauging Lesion Significance 751

Fig. 59.2  TAG and TDG measurement. (a) Transluminal attenuation TDG was 0.162 mm/cm. (b) LAD segmental lesion measured 60% ste-
gradient (TAG) and transluminal diameter gradient (TDG) were calcu- nosis (arrow) by quantitative coronary angiography, and fractional flow
lated in the left anterior descending branch (LAD) using a single-­ reserve with adenosine-induced hyperemia was 0.64
heartbeat acquisition with a 320-slice CT. TAG was −16.7 HU/cm and
752 Y. H. Choe et al.

b Contrast Density Difference

Contrast density difference (CDD) is defined as the decline

in luminal contrast attenuation over a coronary lesion
(Fig. 59.5) [22]. CDD was significantly greater in hemody-
namically relevant lesions, as defined by invasive fractional
flow reserve (FFR), compared to nonsignificant lesions
(26.0 ± 20.2% vs. 16.6 ± 10.9%; p = 0.013). At a threshold
of ≥24%, CDD predicted hemodynamically significant
lesions with a specificity of 75%, sensitivity of 33%, PPV
of 35%, and NPV of 73% [22]. TAG showed no significant
difference between hemodynamically significant and non-
significant lesions (−14.2  ±  14.2  HU/cm vs.
−11.0 ± 12.8 HU/cm; p = 0.45). In the receiver operating
characteristic (ROC) curve analysis, AUCs for CDD, TAG
(threshold, −6.5 HU/cm), and visual assessment of CCTA
were 0.67 (95% CI, 0.54–0.8), 0.56 (95% CI, 0.41–0.71),
and 0.61 (95% CI, 0.48–0.74), respectively. The AUC for
CDD was not significantly higher compared to TAG and
visual stenosis assessment.

Fig. 59.2 (continued)

40
(HU/10 mm distance)

LAD_normal
30
LCx_normal
LCx_stenosis
20
LCx_occlusion

10

Scan number
0
-5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8

-10
TAG

-20

-30

-40

-50

Fig. 59.3  Changes in TAG over time in each coronary artery with artery (LAD) at each time point (all, p > 0.05). Significant difference in
dynamic scanning in an animal model. Scan number was coded as zero TAG values was only found between normal LAD and occluded LCX
when attenuation of the ascending aorta reached its peak. There were no in the fifth to eighth repeated scans and between stenotic LCX and
significant differences in transluminal attenuation gradient (TAG) val- occluded LCX in the third to eighth repeated scans (all, p  <  0.05).
ues between stenotic and normal left circumflex coronary artery (LCX) (From Park et al. [14], with permission)
or between stenotic LCX and normal left anterior descending coronary
59  Transluminal Attenuation Gradient and Other CT Techniques for Gauging Lesion Significance 753

Fig. 59.4  Implementation of a c

transluminal attenuation flow
encoding (TAFE). After 3D
isotemporal acquisition of the typical location
whole heart (a), multiplanar for bolus
reformations of the three tracking
primary coronary vessels are
generated (b). TAG is
computed (b, lower panel) Flow
region
between two user-defined
of
proximal and distal boundary interest
locations. The average
distance between the
user-defined point(s) and the LV
average cross-sectional area Coronary
vessel
(a) over the user-defined
length is automatically
calculated. The contrast bolus
duration (Td) is the temporal d 400
element of contrast dispersion b 1 HUmax
and is derived from AIF
measured in the descending 0.9

Attenuation Intensity (HU)

300
aorta (c, d). Once these four Measured
0.8
HU(s)/HUostium

CT-based parameters have Model

been isolated, an equation 0.7
(bottom) can be used to 200
TAG*
0.6
calculate coronary flow in ts
mL/min. AIF arterial input 0.5 Td
function, TAG transluminal y = -0.0044x + 0.926
100
attenuation gradient. (From 0.4
HUmin
Lardo et al. [21], with 0.3
permission) 0 20 40 60 80 100 120 0
0 5 10 15 20 25 30
Distance, s (mm) time (sec)

 = 0.0505cm2, ŝ = 6cm Td = 15sec = 0.25 min

TAG* = 0.044 cm-1 p ŝ
Qª
Td –TAG*
ª 7.40 ml/min

Corrected Coronary Attenuation significantly greater in arteries with obstructive CAD (diam-
eter stenoses ≥50%) (0.191 ± 0.214; median, 0.106; inter-
Coronary attenuation variability after contrast material quartile range [IQR], 0.042–0.296) (p  =  0.004), and the
injection may occur due to the lack of temporal uniformity proportion of abnormal CCO differences increased with
on 64-slice CT systems. The corrected coronary attenuation worsening diameter stenosis (p  <  0.001). Similarly, CCO
(CCO) is calculated as the quotient of the attenuation in the differences were greater in arteries with TIMI flow grade < 3
coronary artery and aorta (CCO,  coronary artery attenua- (0.406  ±  0.226) compared to coronaries with normal flow
tion/aorta attenuation) (Fig. 59.6) [17]. According to Chow (0.078 ± 0.078, p < 0.001). With CCO differences (thresh-
et al., in normal arteries, the mean CCO was 0.979 ± 0.070 old, >0.184 [mean CCO difference  +  2 standard devia-
(median, 0.965; interquartile range [IQR], 0.940–1.005) tions]), abnormal coronary flow (TIMI flow grade < 3) was
[17]. Normal CCO variability (difference between highest identified with a sensitivity, specificity, positive predictive
CCO and lowest CCO) was measured as 0.100  ±  0.042 value, and negative predictive value of 83.3% (95% confi-
(median, 0.099; IQR, 0.073–0.128). Changes in CCO across dence interval [CI], 57.7–95.6%), 91.2% (95% CI, 75.2–
coronary stenoses seem to predict abnormal resting coro- 97.7%), 83.3% (95% CI, 57.7–95.6%), and 91.2% (95% CI,
nary blood flow (Thrombolysis in Myocardial Infarction 75.2–97.7%), respectively. Notably, the accuracy of this
[TIMI] flow grade < 3) [17]. Compared with the CCO vari- method was 88.5% with very good agreement (kappa = 0.75;
ability in nonobstructive arteries, the CCO difference was 95% CI, 0.55–0.94).
754 Y. H. Choe et al.

Fig. 59.5  Coronary density difference. (a) Coronary density difference measures 82% for the focal severe stenosis with soft plaques in left cir-
cumflex artery. (b) Invasive coronary angiography reveals severe stenosis (arrow) in distal vessel with FFR = 0.4
59  Transluminal Attenuation Gradient and Other CT Techniques for Gauging Lesion Significance 755

b Transluminal Diameter Gradient

Coronary artery intraluminal attenuation was shown to

decrease with diminution of vessel diameters (Fig.  59.7)
[14]. In a study by Park et  al., a moderate correlation was
found between TAG and transluminal diameter gradient
(TDG) (r = 0.580; p < 0.0001) in 152 coronary arteries of 62
patients [14]. Separate analysis of significant and nonsignifi-
cant stenosis groups also showed a significant correlation
between TAG and TDG (r = 0.610 [p < 0.0001] for nonsig-
nificant stenosis; r = 0.565 [p = 0.0001] for significant steno-
sis) (Fig. 59.8). In addition, TAG-positive arteries exhibited
significantly greater TDG values in groups with (p < 0.0001)
and without significant stenosis (p < 0.0001). There were no
significant differences in TAG values between groups with
significant stenosis (diameter stenosis ≥50%) and without
significant stenosis (p = 0.884), whereas a significant differ-
ence was found in TDG values (−0.379 [IQR, 0.377] versus
−0.273 [IQR, 0.191], p  =  0.021). However, it should be
noted that coronary artery attenuation also depends on the
coronary flow physiology and CT technical factors.
Fig. 59.5 (continued)

Fig. 59.6  Corrected coronary

attenuation (CCO). CCO was
calculated as 1.02, 0.79, 0.78,
and 0.84, respectively, in four
sequential segments of right
coronary artery on images
obtained from a 64-slice CT
system. Note the stepwise
changes (arrows) in the
attenuation curve according to
each imaging slab
756 Y. H. Choe et al.

Fig. 59.7  Comparison of TAG and TDG. (a) TAG was −23.5 HU/cm (−3.7%/cm), and TDG was −0.12/cm (−4.3%/cm). (b) Invasive coronary
angiography shows severe stenosis (arrow) in the mid-RCA
 59  Transluminal Attenuation Gradient and Other CT Techniques for Gauging Lesion Significance 757

mixture and scanned using 100 kVp. Even though all holes

20
were filled with the same contrast mixture, intraluminal
attenuation was observed to gradually decline along with
0
decreasing vessel diameter. The reduction of intraluminal
attenuation was much greater as the diameters became
TAG (HU/10 mm Distance)

-20 smaller. The same findings were observed regardless of the

acquisition’s tube voltage or the concentration of contrast
-40 mixture.
In an animal study conducted by the same investigator, a
-60 dynamic CT scan was performed in 3 canines using a 320-­
row CT system (Fig. 59.3) [14]. Small distal coronary arter-
-80 ies exhibited only 55–78% of the maximum enhancement of
proximal segments, regardless of which coronary artery was
being evaluated. The animal study revealed that the length of
-100
R = 0.565 the coronary artery was an important factor in calculating
p <0.0001 TAG. For example, normal vessels that were short with rapid
-120
changes in diameter exhibited higher TAG values.
-1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0
TDG (mm/10 mm Distance)
Clinical Evidences of TAG
Fig. 59.8  Correlation between TAG and TDG. Significant correlation
was found in patients with significant stenosis. (From Park et al. [14],
with permission)
The diagnostic estimates and AUCs of ROC analyses for
TAG, CCTA, and the combination of CCTA and TAG in the
literature are presented in Table 59.1.
 xperimental Studies: Effects of Cardiac
E In a study by Choi et al. that used invasive coronary angi-
Output and Vessel Diameters on TAG ography as the reference standard, TAG decreased as steno-
sis severity increased from −2.37  ±  4.67  HU/cm for a
Funama et  al. investigated TAG values at different scan diameter stenosis between 0% and 49% to −13.46 ± 9.59 HU/
acquisition points along the time-density curve (TDC) after cm for a diameter stenosis of 100% (p < 0.0001) (Fig. 59.9)
contrast material injection using a flow phantom and a [12]. The accuracy in determining stenosis severity classifi-
­320-­row CT scanner [15]. At a cardiac output of 2.0 and cation by CCTA, as compared with ICA, improved signifi-
4.0  L/min with 0% stenosis, TAG exhibited smaller varia- cantly when TAG was considered (c-statistic 0.932 ± 0.012
tions at each time point along the TDC (−3.02 to +0.55 HU/ vs. 0.951 ± 0.010, p = 0.001; global chi-square 254.381 vs.
cm at 2.0 L/min, −2.63 to +0.43 HU/cm at 4.0 L/min) com- 265.899, p = 0.001). The improved diagnostic accuracy was
pared to 70% stenosis [15]. Compared with a cardiac output mainly driven by increased specificity (89.7% [95% CI,
of 2.0 L/min with 70% stenosis, the TAG curve for a cardiac 83.0–94.4] vs. 93.7% [95% CI, 87.9–97.2]). Additionally,
output of 4.0 L/min gradually changed with time (−6.64 to the accuracy of stenosis severity classification of calcified
+1.18  HU/cm at 2.0  L/min vs. −3.46 to +2.75  HU/cm at stenoses by CCTA significantly increased when TAG was
4.0 L/min). The TAG value was affected by acquisition tim- taken into account (c-statistic, p < 0.0001; global chi-square,
ing after contrast material injection and cardiac output; how- p = 0.001). The addition of TAG resulted in a reclassification
ever, the size of the connecting tube (with and without of stenosis severity in a significant number of vessels with
stenosis) was identical. calcified lesions (net reclassification improvement [NRI]
In a study by Park et al. [14], a vessel phantom was con- 0.095, net proportion of patients reclassified 3.15%;
structed with 25 tubular holes in a 5-cm-thick polyethylene p = 0.046). TAG also improved the accuracy for determining
disk. The diameter of each hole was 5.0, 4.5, 4.0, 3.5, 3.0, stenosis severity caused by noncalcified plaques (c-statistic,
and 2.9–1.0 mm with 0.1 mm intervals, simulating vessels of p = 0.048; global chi-square, p = 0.012). However, the addi-
different diameters. The phantom was immersed in two dif- tion of TAG did not result in significant stenosis severity
ferent mixtures of an iodine contrast agent and saline, reclassification in vessels with noncalcified lesions (NRI
approximating attenuations of 800 and 600 HU at 100 kVp, −0.006; p = 0.56) or the entire cohort (NRI 0.036; p = 0.06).
with all tubular holes completely filled with contrast mix- According to Wong et  al., median TAG320 in FFR-­
ture. CT scans were performed with the phantom immersed significant vessels was significantly lower when compared
in the 600 HU contrast mixture using 80, 100, and 120 kVp. with nonsignificant vessels (−19 [−26 to −13] vs. −10 [−16
Then, the phantom was immersed in the 800 HU contrast to −5] HU/cm, p < 0.001) [23].
758 Y. H. Choe et al.

Table 59.1  Diagnostic estimates of TAG, CCTA, and combination of TAG and CCTA according to studies
Reference
Number Number of TAG standard Positive Negative Diagnostic
Authors of vessels cutoff (threshold Sensitivity Specificity predictive predictive accuracy
(reference) patients analyzed CT scanner (HU/cm) value) (%) (%) value (%) value (%) (%) AUC
Choi et al. 126 370 64-slice −1.8 CAG 92.6 69.1 85.3 82.9 0.860a
[12] 84.0 89.7 94.0 74.3 0.932a
59 94 83 82 0.951a
Zheng 107 309 128-slice −11.33 CAG 73.4 92.1 89.3 79.6 0.827
et al. [24] DS 94.7 77.7 83.9 92.3 0.924
94.0 93.8 90.2 96.3 0.983
Choi et al. 63 97 64-slice −6.54 FFR 47.5 91.2 79.2 71.2 73.2 0.696a
[26] DS or (0.80) 92.5 52.6 57.8 90.9 69.0 0.726a
64-slice 90 63.2 63.2 90.0 74.2 0.809a
SS
Yoon et al. 62 82 64-slice −6.54 FFR 37.5 88.0 66.7 68.8 0.63
[6] DS or (0.80) 71.9b 68.0b 59.0b 79.1b 0.73b
64-slice
SS
Wang 32 32 64-slice −15.1 FFR 37 58 23 73 0.67
et al. [28] DS or (0.80) 100 54 42 100
128-slice
DS
Nakanishi 103 146 64-slice −6.54 FFR 63.5 28.7 33.0 58.7 41.1 0.54
et al. [27] or higher (0.80) 84.6 39.4 43.6 82.2 55.5 0.62
0.64
Hell et al. 59 72 128-slice −6.5 FFR 57 61 28 31 0.56
[22] DS (0.8) 0.61
Stuijfzand 85 253 256-slice −7.51 FFR 69 44 83 27 NA
et al. [25] (0.80) 95 75 98 54 0.85
95 76 98 54 0.87
Wong 75 97 320-slice −15.1 FFR 71 77 63 83 NA
et al. [23] (0.80) 89 65 57 92 0.77a
73 97 92 87 0.84a
Ko et al. 51 82 320-slice −15.37 FFR 58.3 86.2 63.6 83.3 0.72
[13] (0.80) 79.2 58.6 44.2 87.2 0.69
45.8 98.3 91.7 81.4 0.72
Ko et al. 27 51 320-slice −15.1 FFR 76.5 85.3 72.2 87.9 82.4 0.78
[29] (rest) (0.80) 76.5 55.9 46.4 82.6 62.7 0.66
58.8 94.1 83.3 82.1 82.4 0.76
Ko et al. 27 35 320-slice −10 FFR 75 61 50 82 65.7 0.75
[29] (stress) (0.80) 50 78 55 75 68.6 0.64
Numbers in the second lines represent diagnostic values with CCTA for diameter stenosis >50%. Numbers in italic represent diagnostic values with
combination of CCTA and TAG. All results were analyzed on a per-vessel basis
Abbreviations: AUC area under curve, CAG invasive coronary angiography, CCTA coronary CT angiography, DS dual source, HU Hounsfield unit,
FFR fractional flow reserve, NA not assessed, SS single source, TAG transluminal attenuation gradient
a
C-statistic
b
CCTA stenosis >70%

In a study by Zheng et al., the impact of plaque composi- lesions, TAG decreased progressively as Agatston scores
tion, Agatston scores, and lesion length ratio on TAG was increased, from −13.93 ± 14.84 HU/cm for scores <100 to
analyzed in 107 patients [24]. TAG decreased progressively −28.64 ± 18.26 HU/cm for scores >300 (p = 0.000). Notably,
as ICA-determined stenosis severity increased, from there was no significant difference in TAG between scores
−6.10  ±  6.97 HU/cm for stenosis <50% to <100 and scores between 100 and 300 (−13.93 ± 14.84 vs.
−20.45 ± 13.69 HU/cm for stenosis of 70–99% (p = 0.002). −17.25 ± 13.91, p = 0.198). TAG improved the diagnostic
This tendency was similar in calcified vessels. In calcified accuracy of CCTA (c-statistic =0.982 vs. 0.942, p = 0.0001)
59  Transluminal Attenuation Gradient and Other CT Techniques for Gauging Lesion Significance 759

in calcified lesions. In calcified vessels, CCTA showed mod- sels with calcified lesions (NRI 0.093, net proportion of
erate sensitivity [70%; 95% CI =57–82%] and low specific- patients reclassified 5.31%; p = 0.022); however, the addi-
ity (66%; 95% CI =54–76%), while TAG showed moderate tion of TAG to CCTA did not result in significant stenosis
sensitivity (72%; 95% CI =64–79%) and high specificity severity reclassification among all vessels (NRI 0.009;
(91%; 95% CI =86–95%). The addition of TAG to CCTA p = 0.09). TAG also decreased as lesion length ratio (LLR)
markedly improved the diagnostic performance of CCTA increased, from −10.90  ±  10.19  HU/cm for LLR <1/3 to
for calcified lesions and resulted in increased sensitivity −32.61  ±  16.64  HU/cm for LLR between 1/2 and 2/3
(84%; 95% CI =  72–92%) and specificity (89%; 95% CI (p  =  0.000). Interestingly, the TAG for LLR >2/3 was
=77–96%) (Fig. 59.10). Adding TAG to CCTA resulted in a −14.51  ±  10.58 HU/cm, which was not lower than the
reclassification of stenosis severity in a large number of ves- cohort with LLR between 1/2 and 2/3.

a
(1) (2)
900
800

Luminal Attenuation (HU)

700
600
500
400
300
200 Y = 0.142* X + 467

100 TAG = 1.42 (HU/10 mm)

0
0 20 40 60 70 100 120 140
QCA: MLD = 2.46 mm, DS = 28.7% Length from Ostium (mm)

(3)

480 HU 827 HU 447 HU 476 HU

10.4 mm2 0.9 mm2 7.6 mm2 4.5 mm2
3.6 mm 1.1 mm 3.1 mm 2.4 mm

Fig. 59.9 Transluminal attenuation gradient measurement. (a) LAD artery does not decrease, demonstrating no significant obstruc-
Calcified lesion in the mid-left anterior descending (LAD) artery that tion. (b) CCTA demonstrates calcified lesions in the proximal LAD
was indeterminate by CCTA, but diameter stenosis (DS) was 28.7% by artery and moderate stenosis in the mid-LAD artery, confirmed by inva-
quantitative coronary angiography (1). Red arrows indicate the most sive coronary angiography (1). The linear correlation coefficient of
severe stenotic sites. Gray dots represent intervals that were excluded black dots, TAG, is negatively sloped (−11.95 HU/cm) (2). Decrease of
because of significant calcification or significant (DS 50%) stenosis (2). intraluminal attenuation in the distal vessel is shown (3). (From Choi
Cross-sectional views with gray border and sloped legend in italics rep- et al. [12], with permission)
resent excluded intervals. The intraluminal attenuation in the distal
760 Y. H. Choe et al.

b
(1) (2)

500

Luminal Attenuation (HU)

400

300

200
Y = -1.195* X + 506
100 TAG = -11.95 (HU/10 mm)

0
0 40 80 120 160
QCA: MLD = 1.30 mm, DS = 52.7% Length from Ostium (mm)

(3)

522 HU 503 HU 624 HU 301 HU 344 HU 316 HU

6.6 mm2 7.9 mm2 4.1 mm2 2.9 mm2 3.2 mm2 2.8 mm2
2.9 mm 3.2 mm 2.3 mm 1.9 mm 2.1 mm 1.7 mm

Fig. 59.9 (continued)

Comparison of TAG and CCO formed for vessels with a diameter stenosis ≥50% on
CCTA (n = 104, p = 0.07).
In a study by Stuijfzand et al., TAG did not provide incre- Choi et al. compared TAG and CCO of CCTA with inva-
mental diagnostic value over 256-slice coronary CTA sively measured FFR in 63 patients [26]. The overall diag-
alone for assessing the hemodynamic significance of a nostic performance of TAG and CCO was similar and
coronary stenosis. Step-and-shoot 256-slice CCTA was moderate on a per-vessel basis (c-statistic = 0.696 vs. 0.637,
used to evaluate TAG, TAG-CCO (relating coronary den- p  =  0.29). The sensitivity, specificity, and positive and
sity to corresponding descending aortic opacification), negative predictive values of TAG using a cutoff
­
and TAG-ExC (excluding calcified coronary segments) ≤−6.54 HU/cm for FFR < 0.80 were 47.5, 91.2, 79.2, and
analyses in 85 patients [25]. TAG and TAG-ExC did not 71.2%. Using CCO and a cutoff >0.063 provided sensitiv-
discriminate between vessels with or without hemody- ity, specificity, and positive and negative predictive values
namically significant lesions with FFR values >0.8 or of 65.0, 61.4, 54.2, and 71.4%, respectively. TAG demon-
≤0.8 (−13.5  ±  17.1  HU/cm vs. –11.6  ±  13.3  HU/cm, strated incremental value to the diagnostic performance of
p = 0.36; and –13.1 ± 15.9 HU/cm vs. –11.4 ± 11.7 HU/ CCTA alone; however, the added value was not observed
cm, p  =  0.77, respectively). TAG-­CCO was significantly for CCO (c-statistic = 0.726 vs. 0.809, p = 0.025; c-statis-
lower in vessels with a hemodynamically significant tic = 0.726 vs. 0.784, p = 0.09). In NRI analysis, the addi-
lesion (−0.050  ±  0.051/cm vs. –0.036  ±  0.034/cm, tion of TAG to CCTA did not result in significant
p = 0.03), and TAG-ExC resulted in a slight improvement reclassification (NRI = 1.0%, p = 0.41), while the addition
of NRI (0.021, p < 0.05). However, TAG-CCO was no lon- of CCO to CCTA resulted in negative reclassification
ger significantly lower when subgroup analysis was per- (NRI = −9.3%, p = 0.036).
59  Transluminal Attenuation Gradient and Other CT Techniques for Gauging Lesion Significance 761

a 100 b 100

80 80

60 60

Sensitivity
Sensitivity

40 40

20 20

0 0
0 20 40 60 80 100 0 20 40 60 80 100
100.Specificity 100.Specificity
CCTA 0.924 ± 0.02 CCTA 0.942 ± 0.01
CCTA+TAG 0.983 ± 0.01 * 0.0006 * 0.011 CCTA+TAG 0.982 ± 0.01 * 0.0001 * 0.0001
TAG
0.827 ± 0.03 * < 0.0001 TAG
0.839 ± 0.02 * < 0.0001

Fig. 59.10  Predicted probability of TAG in addition to CCTA stenosis stenosis severity on a reference of results from CAG. Area under the
severity. (a) Predicted probability of TAG in addition to CCTA stenosis ROC curve is shown as mean ± SD. TAG transluminal attenuation gra-
severity in total relevant vessels on a reference of results from CAG. (b) dient, CAG coronary angiography, CCTA coronary computed tomogra-
Predicted probability of TAG in calcified vessels in addition to CCTA phy angiography. (From Zheng et al. [24], with permission)

1.0
Comparison of CT-FFR and TAG

Noninvasive FFR computed from CCTA provides a better

diagnostic performance for the diagnosis of lesion-specific 0.8
ischemia compared to CCTA stenosis and TAG [6, 13, 27].
Ko et  al. compared the diagnostic performance of
320-detector row CCTA-derived computed FFR (FFRCT), 0.6
Sensitivity

TAG (TAG320), and CCTA alone to diagnose hemodynami-

cally significant stenosis as determined by invasive FFR
[13]. The median TAG320 was significantly lower in vessels
0.4
with hemodynamically significant stenoses compared to ves- Source of the
Curve
sels without significant stenosis, −17.2 HU/cm (IQR, −20.0
to −6.40) versus −8.86  HU/cm (IQR, −13.2 to −3.83) TAG320
(p = 0.002). By using TAG320, a modest yet statistically sig- 0.2 FFRct
nificant correlation with invasive FFR was demonstrated CTA
(Spearman ρ  =  0.47; p  <  0.001). FFRCT demonstrated a Reference Line
strong and statistically significant correlation with invasive 0.0
FFR (Spearman ρ = 0.78; p < 0.001). The ROC curve analy- 0.0 0.2 0.4 0.6 0.8 1.0
sis for TAG320 demonstrated an AUC of 0.72 (p  =  0.002), 1 - Specificity

which was not found to be superior to the AUC of CCTA Fig. 59.11  Areas under the ROC curve (AUC) of CT angiography
(0.72 vs. 0.68; p = 0.67) (Fig. 59.11). NRI for TAG320 when (CTA), TAG320, and FFRCT, with a reference line for comparison.
compared with CCTA was 0.89  ±  0.24 (standard error; (From Ko et al. [13], with permission)
p < 0.001). There was a slight difference in AUC when CCTA
with TAG was compared with CCTA alone (0.72 vs 0.69; Overall per-vessel accuracy, sensitivity, specificity, and posi-
p = 0.59). By using a FFRCT threshold ≤0.80, FFRCT demon- tive and negative predictive values for TAG320 (<−15.37)
strated an AUC of 0.93 (p < 0.001), which was significantly were 78%, 58%, 86%, 64%, and 83%, respectively; the same
superior to both CCTA (p = 0.008) and TAG320 (p = 0.003). values for FFRCT were 83%, 92%, 79%, 65%, and 96%,
762 Y. H. Choe et al.

respectively. NRI for FFRCT compared with CCTA was for FFRCT, respectively [27]. On a per-patient basis, the AUC
1.42 ± 0.24 (p < 0.001). by ROC curve analysis for FFRCT (0.77) demonstrated
According to Nakanishi et  al., FFRCT allowed for the greater performance for predicting hemodynamically signifi-
identification of lesion-specific ischemia with greater diag- cant ischemia compared to the AUC of obstructive stenosis
nostic accuracy than TAG, CCTA, or the combination of the on CCTA (0.55, p  <  0.0001 vs. FFRCT) and TAG +
two with invasive FFR used as the reference standard. In 103 CCTA.  AUCs for TAG analyses with varying thresholds
patients with suspected or known CAD, the sensitivity, spec- were as follows: ≤−11/cm (0.52, p  =  0.0001 vs. FFRCT),
ificity, positive predictive value, and negative predictive ≤−6.54/cm (0.56, p  =  0.0006 vs. FFRCT), and ≤−15.1/cm
value were 53.8, 45.7, 35.4, and 64.2% for TAG; 84.6, 39.4, (0.54, p  <  0.0001 vs. FFRCT) (Fig.  59.12). Compared to
43.6, and 82.2% for CCTA; and 82.7, 74.5, 64.2, and 88.6% CCTA alone (AUC, 0.62), the combination of TAG and

a b

Fig. 59.12  Integration of CCTA, TAG, and stress CT perfusion imag- (c) Adenosine-stress MRI confirms the presence of myocardial isch-
ing. (a) TAG measures −20.3 HU/cm. (b) Adenosine-stress CT perfu- emia as a perfusion defect (arrow). (d) Invasive coronary angiography
sion imaging shows a perfusion defect (arrow) in the anteroseptal wall. shows severe stenosis in the proximal LAD branch with FFR = 0.48
59  Transluminal Attenuation Gradient and Other CT Techniques for Gauging Lesion Significance 763

CCTA did not significantly improve the diagnostic value and a mild decrease in specificity compared with CCTA + CTP
(AUC, 0.63 for ≤−11/cm, p = 0.76; AUC, 0.64 for ≤−6.54/ and CCTA + TAG320. Functionally significant coronary steno-
cm, p = 0.43; AUC, 0.64 for ≤−15.1/cm, p = 0.36). sis was defined as FFR ≤ 0.8. The cohort included 75 patients
According to Wang et  al., TAG with an AUC of 0.67 (age 64.1 ± 10.8 years, 52 men) and 44 (35%) FFR-significant
(p = 0.152) was unable to discriminate between vessels with or vessels. In 127 vessels, CCTA predicted FFR-significant steno-
without hemodynamically significant lesions [28]. The authors sis with 89% sensitivity and 65% specificity compared with
calculated lesion length/minimal luminal diameter4 (LL/MLD4), MDCT-IP, which showed 88% sensitivity and 83% specificity.
TAG, CCO, and CT-FFR in 32 patients. A Poiseuille-based Using the generalized estimating equation, the predictive val-
index (LL/MLD4) was calculated as the LL divided by the 4th ues of CTP + CCTA (p = 0.003) and TAG320 + CCTA (p = 0.002)
power of the minimal luminal diameter (MLD4). A ratio of LL/ were comparable. The c-statistics for the combined assessment
MLD4 > 3.86 was used to define a hemodynamically significant of CTP + CCTA (0.845) and TAG320 + CCTA (0.844) were also
coronary stenosis. Compared to invasive FFR, the per-vessel comparable (p = 0.98). The c-statistic for MDCT-IP was 0.905.
sensitivity and specificity of CCTA, CT-FFR, LL/MLD4, CCO, The diagnostic accuracy of MDCT-IP (AUC = 0.91) was supe-
and TAG (≤−15.1 HU/cm) for detecting hemodynamically sig- rior to both TAG320  +  CCTA and CTP  +  CCTA (p  =  0.01).
nificant lesions were 100% and 54%, 100% and 91%, 85% and There was an incremental value of adding CTP and TAG320 to
92%, 66% and 88%, and 37% and 58%, respectively. ROC CCTA assessment (MDCT-IP) for the detection of significant
analysis resulted in an AUC of 0.91 for CT-FFR (p = 0.0005), FFR for NRI (1.46, p < 0.0001) and the integrated discrimina-
0.88 for LL/MLD4 (p < 0.0001), and 0.85 for CCO (p < 0.0001). tion improvement index (IDI) (0.26, p < 0.0001). In addition,
CT-FFR, LL/MLD4, and CCO provided enhanced diagnostic MDCT-IP also had incremental value over TAG320  +  CCTA
performance over CCTA analysis alone for the discrimination assessment for NRI (0.93, p  <  0.0001) and IDI (0.12,
of hemodynamically significant coronary stenosis. p  <  0.0001). MDCT-IP also had incremental value over
According to Yoon et  al. [6], the TAG was significantly CTP + CCTA assessment for NRI (1.45, p < 0.0001) and IDI
lower in vessels with ischemia (−5.8 ± 8.0 HU/cm) compared (0.10, p < 0.0001).
to vessels without ischemia (−2.3 ± 4.8 HU/cm, p = 0.029).
ROC curve analysis for TAG showed an AUC of 0.63
(p = 0.037) with an optimal cutoff value of ≤−6.54 HU/cm. Rest and Stress TAG
The sensitivity, specificity, PPV, NPV, positive likelihood
ratio, and negative likelihood ratio of TAG ≤−6.54 HU/cm Ko et al. compared TAG values (TAG320) during rest and vaso-
were 37.5% (21.7–56.3%), 88.0% (75.0–95.3%), 66.7% dilator stress conditions on a 320-slice scanner [29]. Stress
(41.2–85.6%), 68.8% (55.8–79.4%), 3.13 (1.30–7.49), and TAG320 was not interpretable in 16 vessels (31.4%). ROC anal-
0.71 (0.54–0.93), respectively (Fig. 59.12). The ROC curve ysis showed a comparable AUC for rest/stress TAG320 (0.78 and
analysis for FFRCT demonstrated an AUC of 0.94 (p < 0.001) 0.75), which was higher than CCTA alone (0.68) and rest/stress
with an optimal cutoff value of ≤0.77. The sensitivity, contrast opacification (CO) difference (0.76 and 0.67). There
­specificity, PPV, NPV, positive likelihood ratio, and negative was incremental predictive value when rest TAG320 was added
likelihood ratio of FFRCT ≤ 0.77 were 81.3% (63.0–92.1%), to CCTA assessment based on the IDI (0.33, p < 0.0001) and
94.0% (82.5–98.4%), 89.7% (71.5–97.3%), 88.7% (76.3– NRI (1.24, p < 0.0001). There was incremental predictive value
95.3%), 13.54 (4.46–41.08), and 0.20 (0.10–0.41), respec- when stress TAG320 was added to CCTA assessment based on
tively. In all vessels, the AUC for FFRCT (0.94) was NRI (0.72, p  =  0.04). The mean rest CO difference using a
significantly greater than CCTA (0.73, p < 0.001) and TAG threshold of 55 HU predicted FFR ≤ 0.80 with a sensitivity,
(0.63, p < 0.001). In a subgroup with noncalcified or partially specificity, PPV, and NPV of 82%, 44%, 42%, and 83%,
calcified plaque (n  =  58), the AUC for FFRCT (0.94) was respectively. The corresponding values for mean stress CO dif-
greater compared to TAG (0.63, p < 0.001) and CCTA (0.70, ference using a threshold of 62 HU were 75%, 65%, 53%, and
p < 0.001). In a subgroup with calcified plaque (n = 24), the 83%. The ROC AUC for mean rest/stress CO difference was
AUC for FFRCT (0.92) was not significantly different from the 0.76 (95% CI 0.60–0.93) and 0.67 (95% CI 0.48–0.86), respec-
AUC for TAG (0.75, p = 0.168) and CCTA (0.80, p = 0.195). tively, which were not significantly different (p = 0.37).

Comparison of TAG with CT Perfusion  everse Attenuation Sign and TAG

R
in Patients with Chronic Total Occlusion,
Wong et al. compared the diagnostic accuracy of combined CT Stent, and Myocardial Bridging
perfusion (CTP) + CCTA, TAG by 320-detector row CT
(TAG320, cutoff value of −15.1 HU/cm)  +  CCTA, and The reverse attenuation gradient (RAG) sign is defined as a
CTP + TAG320 + CCTA (CT-integrated protocol [MDCT-IP]) reverse intraluminal opacification gradient of vessels distal
assessment in predicting significant FFR (Fig. 59.12) [23]. The to occlusive lesions (Fig.  59.13). Therefore, the RAG sign
MDCT-IP approach resulted in a marked increase in sensitivity represents the retrograde collateral flow distal to an occlusive
764 Y. H. Choe et al.

According to Choi et  al., coronary arteries with CTO

showed longer occlusion length (cutoff ≥15 mm), higher dis-
tal TAG (cutoff ≥ −0.9 HU/cm), more frequent side branches,
blunted stump, cross-sectional calcification ≥50%, and col-
lateral vessels compared with coronaries with STO
(p < 0.001, all) [32]. The combination of these findings could
distinguish CTO from STO (c-statistics = 0.88 [95% confi-
dence interval 0.94–0.90], sensitivity 83%, specificity 77%,
positive predictive value 55%, negative predictive value
93%; p < 0.001). TAG from the ostium to the distal vessel
(TAGall) [CTO −11.0 (−19.3 to −4.5) HU/cm vs. STO −12.9
(−21.6 to −6.9) HU/cm; p  =  0.006] and TAG distal to the
occlusion (TAGdistal) [CTO −0.1 (−3.2 to −2.2) HU/cm vs.
STO −3.0 (−4.7 to −1.1) HU/cm; p < 0.001] were higher in
CTO versus STO. However, according to Zheng et  al., the
TAG (−17.96 ± 20.36 HU/cm) for CTO was not significantly
lower than TAG for stenosis between 70% and 99%
(−20.45 ± 13.69 HU/cm) [24].
The degree of invasive coronary angiography-assessed
systolic compression of myocardial bridging (MB) signifi-
cantly correlates with TAG, but not MB depth or length [33].
In a study by Li et al., TAG was lowest (−17.4 ± 6.7 HU/cm)
in patients with significant dynamic compression (systolic
compression ≥50%) and highest in patients without MB
compression (−9.5 ± 4.3 HU/cm, p < 0.001). Linear correla-
tion revealed a moderate relationship between the percentage
of systolic compression and TAG (Pearson correlation,
r = −0.52; p < 0.001), but no significant relationship between
Fig. 59.13  Reverse attenuation sign and TAG in a patient with chronic the percentage of systolic compression and MB depth or
total occlusion of RCA.  The degree of attenuation in the more distal length. ROC curve analysis determined the optimal cutoff
segment is higher than that of the more proximal segment and TAG in value of TAG as −14.8HU/cm (area under curve  =  0.813;
this segment distal to occlusion measures 13.4 HU/cm
95% confidence interval = 0.764–0.855; p < 0.001), which
yielded high diagnostic accuracy (82.1%, 248/302).
lesion. This sign seems to be useful for the detection of
chronic total occlusion (CTO) and helps differentiate CTO
from subtotal occlusion (STO). In a study by Li et  al., the Summary and Conclusions
CTO group had the RAG sign significantly more frequently
than the STO group (65% [32/49] vs. 7% [3/45]; p < 0.001) TAG depends on vessel diameter, vessel-specific branching
[30]. Similarly, a significant difference in measurements of patterns, luminal stenosis, scan timing, and cardiac output.
the attenuation gradient (5.1  HU/cm  ±  13.4 vs. –13.4  HU/ The range of thresholds of TAG to predict lesion-specific
cm ± 8.7; p < 0.001) and lesion length (23.6 mm ± 22.7 vs. ischemia has varied widely, ranging from −1.8 HU/cm (64-­
6 mm ± 3; p < 0.001) was noted between the groups. All seg- slice CT) to −15.1  HU/cm (320-slice CT). TAG in FFR-­
ments with RAG at CCTA were shown by means of ICA to significant vessels may be significantly lower when compared
be supplied by retrograde collateral vessels. When a combi- with nonsignificant vessels, and TAG may improve diagnos-
nation of all those parameters was used for the diagnosis of tic accuracy of CCTA in calcified lesions. TAG and RAG
CTO, sensitivity and specificity were 90% (44/49) and 93% sign may be useful in the diagnosis of CTO and stent evalu-
(42/45), respectively. ation; however, TAG assessment is susceptible to the influ-
According to Li et  al. [31], RAG sign was present in ence of calcification and artifact throughout the entire
59.3% (35/59) of all stent occlusions. Superior diagnostic coronary artery course. Notably, the diagnostic utility of
performance was confirmed by ROC analysis with an AUC TAG remains uncertain in branched vessel disease. In vessels
of 0.898 when a combination of the traditional diagnosing without significant calcification or artifact, the combination
criterion and RAG sign was employed. of TAG and CCTA may provide comparable diagnostic accu-
59  Transluminal Attenuation Gradient and Other CT Techniques for Gauging Lesion Significance 765

racy for functional assessment of coronary artery stenosis. 13. Ko BS, Wong DT, Norgaard BL, Leong DP, Cameron JD, Gaur S,
et al. Diagnostic performance of transluminal attenuation gradient
Regarding the current literature, the clinical value of TAG and noninvasive fractional flow reserve derived from 320-­detector
and TAG-CCO is controversial and still requires validation; row CT angiography to diagnose hemodynamically significant cor-
however, TAG-CCO may overcome the temporal difference onary stenosis: an NXT substudy. Radiology. 2016;279:75–83.
in coronary attenuations among imaging slabs on multi-beat 14. Park EA, Lee W, Park SJ, Kim YK, Hwang HY.  Influence of
coronary artery diameter on intracoronary transluminal attenua-
CT acquisitions. There is a significant correlation between tion gradient during CT angiography. JACC Cardiovasc Imaging.
TAG and TDG. CT-FFR provides superior diagnostic perfor- 2016;9:1074–83.
mance for the diagnosis of lesion-specific ischemia com- 15. Funama Y, Utsunomiya D, Oda S, Shimonobo T, Nakaura T,

pared to CCTA stenosis and TAG in noncalcified lesions. Mukunoki T, et al. Transluminal attenuation-gradient coronary CT
angiography on a 320-MDCT volume scanner: effect of scan tim-
ing, coronary artery stenosis, and cardiac output using a contrast
medium flow phantom. Phys Med. 2016;32(11):1415–21.
References 16. Krug KB, Bovenschulte H, Geissler HJ, Flesch M, Schneider T,
Brackrogge D, et al. In-vivo measurements of coronary blood flow
1. Dweck MR, Williams MC, Moss AJ, Newby DE, Fayad using 16-slice multidetector spiral computed tomography (MDCT)
ZA.  Computed tomography and cardiac magnetic resonance in in a porcine model. Rofo. 2009;181:220–9.
ischemic heart disease. J Am Coll Cardiol. 2016;68:2201–16. 17. Chow BJ, Kass M, Gagne O, Chen L, Yam Y, Dick A, et al. Can
2. Kim YJ, Yong HS, Kim SM, Kim JA, Yang DH, Hong YJ. Korean differences in corrected coronary opacification measured with
guidelines for the appropriate use of cardiac CT. Korean J Radiol. ­computed tomography predict resting coronary artery flow? J Am
2015;16:251–85. Coll Cardiol. 2011;57:1280–8.
3. Min JH, Kim SM, Lee S, Choi JH, Chang SA, Choe YH. Diagnostic 18. Chatzizisis YS, George E, Cai T, Fulwadhva UP, Kumamaru KK,
performance of algorithm for computer-assisted detection of signif- Schultz K, et al. Accuracy and reproducibility of automated, stan-
icant coronary artery disease in patients with acute chest pain: com- dardized coronary transluminal attenuation gradient measurements.
parison with invasive coronary angiography. AJR Am J Roentgenol. Int J Cardiovasc Imaging. 2014;30:1181–9.
2014;202:730–7. 19. Lackner K, Bovenschulte H, Stutzer H, Just T, Al-Hassani H, Krug
4. Rief M, Kranz A, Hartmann L, Roehle R, Laule M, Dewey B. In vitro measurements of flow using multislice computed tomog-
M.  Computer-aided CT coronary artery stenosis detection: com- raphy (MSCT). Int J Cardiovasc Imaging. 2011;27:795–804.
parison with human reading and quantitative coronary angiography. 20. Bovenschulte H, Krug B, Schneider T, Schwabe H, Kabbasch C,
Int J Cardiovasc Imaging. 2014;30:1621–7. Bangard C, et  al. CT coronary angiography: coronary CT-flow
5. Koo HJ, Yang DH, Kim YH, Kang JW, Kang SJ, Kweon J, et al. quantification supplements morphological stenosis analysis. Eur J
CT-based myocardial ischemia evaluation: quantitative angiog- Radiol. 2013;82:608–16.
raphy, transluminal attenuation gradient, myocardial perfusion, 21. Lardo AC, Rahsepar AA, Seo JH, Eslami P, Korley F, Kishi S, et al.
and CT-derived fractional flow reserve. Int J Cardiovasc Imaging. Estimating coronary blood flow using CT transluminal attenuation
2016;32(Suppl 1):1–19. flow encoding: formulation, preclinical validation, and clinical fea-
6. Yoon YE, Choi JH, Kim JH, Park KW, Doh JH, Kim YJ, et  al. sibility. J Cardiovasc Comput Tomogr. 2015;9:559–566. e551.
Noninvasive diagnosis of ischemia-causing coronary stenosis using 22. Hell MM, Dey D, Marwan M, Achenbach S, Schmid J, Schuhbaeck
CT angiography: diagnostic value of transluminal attenuation gra- A.  Non-invasive prediction of hemodynamically significant coro-
dient and fractional flow reserve computed from coronary CT angi- nary artery stenoses by contrast density difference in coronary CT
ography compared to invasively measured fractional flow reserve. angiography. Eur J Radiol. 2015;84:1502–8.
JACC Cardiovasc Imaging. 2012;5:1088–96. 23. Wong DT, Ko BS, Cameron JD, Leong DP, Leung MC, Malaiapan
7. Yoon YE, Lim TH. Current roles and future applications of cardiac Y, et  al. Comparison of diagnostic accuracy of combined assess-
CT: risk stratification of coronary artery disease. Korean J Radiol. ment using adenosine stress computed tomography perfusion +
2014;15:4–11. computed tomography angiography with transluminal attenuation
8. Zhu X, Zhu Y, Xu H, Tang L, Xu Y. The influence of body mass gradient + computed tomography angiography against invasive
index and gender on coronary arterial attenuation with fixed iodine fractional flow reserve. J Am Coll Cardiol. 2014;63:1904–12.
load per body weight at dual-source CT coronary angiography. 24. Zheng M, Wei M, Wen D, Zhao H, Liu Y, Li J, et al. Transluminal
Acta Radiol. 2012;53:637–42. attenuation gradient in coronary computed tomography angiogra-
9. Ofili EO, Kern MJ, Labovitz AJ, St Vrain JA, Segal J, Aguirre FV, phy for determining stenosis severity of calcified coronary artery: a
et al. Analysis of coronary blood flow velocity dynamics in angio- primary study with dual-source CT. Eur Radiol. 2015;25:1219–28.
graphically normal and stenosed arteries before and after endolumen 25. Stuijfzand WJ, Danad I, Raijmakers PG, Marcu CB, Heymans
enlargement by angioplasty. J Am Coll Cardiol. 1993;21:308–16. MW, van Kuijk CC, et al. Additional value of transluminal atten-
10. Youn HJ, Foster E.  Demonstration of coronary artery flow using uation gradient in CT angiography to predict hemodynamic sig-
transthoracic Doppler echocardiography. J Am Soc Echocardiogr. nificance of coronary artery stenosis. JACC Cardiovasc Imaging.
2004;17:178–85. 2014;7:374–86.
11. Steigner ML, Mitsouras D, Whitmore AG, Otero HJ, Wang C, 26. Choi JH, Koo BK, Yoon YE, Min JK, Song YB, Hahn JY, et  al.
Buckley O, et al. Iodinated contrast opacification gradients in nor- Diagnostic performance of intracoronary gradient-based methods
mal coronary arteries imaged with prospectively ECG-gated single by coronary computed tomography angiography for the evaluation
heart beat 320-detector row computed tomography. Circ Cardiovasc of physiologically significant coronary artery stenoses: a validation
Imaging. 2010;3:179–86. study with fractional flow reserve. Eur Heart J Cardiovasc Imaging.
12. Choi JH, Min JK, Labounty TM, Lin FY, Mendoza DD, Shin DH, 2012;13:1001–7.
et  al. Intracoronary transluminal attenuation gradient in coronary 27. Nakanishi R, Matsumoto S, Alani A, Li D, Kitslaar PH, Broersen
CT angiography for determining coronary artery stenosis. JACC A, et al. Diagnostic performance of transluminal attenuation gradi-
Cardiovasc Imaging. 2011;4:1149–57. ent and fractional flow reserve by coronary computed tomographic
766 Y. H. Choe et al.

angiography (FFR(CT)) compared to invasive FFR: a sub-group flow--a useful sign for differentiating chronic total occlusion from
analysis from the DISCOVER-FLOW and DeFACTO studies. Int subtotal occlusion. Radiology. 2013;266:766–72.
J Cardiovasc Imaging. 2015;31:1251–9. 31. Li M, Zhang J, Zhang Q, Pan J, Lu Z, Wei M.  Coronary stent
28. Wang R, Renker M, Schoepf UJ, Wichmann JL, Fuller SR, Rier JD, occlusion: reverse attenuation gradient sign observed at computed
et al. Diagnostic value of quantitative stenosis predictors with coro- tomography angiography improves diagnostic performance. Eur
nary CT angiography compared to invasive fractional flow reserve. Radiol. 2015;25:568–74.
Eur J Radiol. 2015;84:1509–15. 32. Choi JH, Kim EK, Kim SM, Kim H, Song YB, Hahn JY, et  al.
29. Ko BS, Seneviratne S, Cameron JD, Gutman S, Crossett M,
Noninvasive discrimination of coronary chronic total occlusion and
Munnur K, et al. Rest and stress transluminal attenuation gradient subtotal occlusion by coronary computed tomography angiography.
and contrast opacification difference for detection of hemodynami- JACC Cardiovasc Interv. 2015;8:1143–53.
cally significant stenoses in patients with suspected coronary artery 33. Li Y, Yu M, Zhang J, Li M, Lu Z, Wei M. Non-invasive imaging
disease. Int J Cardiovasc Imaging. 2016;32:1131–41. of myocardial bridge by coronary computed tomography angiog-
30. Li M, Zhang J, Pan J, Lu Z. Obstructive coronary artery disease: raphy: the value of transluminal attenuation gradient to predict sig-
reverse attenuation gradient sign at CT indicates distal retrograde nificant dynamic compression. Eur Radiol. 2017;27(5):1971–9.
 CT Angiography-Derived Fractional
Flow Reserve 60
Adriaan Coenen, Frank Gijsen, and Koen Nieman

Coronary CT angiography allows for noninvasive evaluation measurement of flow, FFR is in fact the pressure ratio over a
of the coronary anatomy and angiographic assessment of stenotic lesion during hyperemia. The FFR procedure con-
coronary artery disease. Limitations are the tendency to over- sists of placement of a wire that allows for the measurement
estimate stenosis severity in comparison to invasive angiogra- of pressure across the lesion of interest. A vasodilator, often
phy and its inability to assess the hemodynamic severity of adenosine, is administrated directly into the coronary artery
stenosis. Invasive fractional flow reserve (FFR) is widely or intravenously to temporarily create a hyperemic state.
regarded as the clinical reference for the hemodynamic evalu- Adenosine causes maximum vasodilation of the microvascu-
ation and therapeutic management of coronary artery disease. lature and increases flow through the coronary arteries by a
Computational fluid dynamics (CFD) is a field of mathemat- factor of 4–5. Blood pressure is measured simultaneously in
ics and fluid mechanics concerned with the simulation of the coronary artery and in the aorta through the guiding cath-
fluid motion. CFD is applied in various industries, to model eter. The FFR is a lesion-specific index calculated as the
the interaction between flow and structures. In addition to fraction between the pressure in the coronary artery and the
various industries, CFD has been used to investigate blood aorta. In case of multiple lesions, a pullback pressure curve
flow in the human body and in the context of various patho- can be performed to assess the area of most severe hemody-
physiologies. With the spatial resolution provided by contem- namic obstruction. Several landmark trials have established
porary computed tomography, it has become possible to use the clinical value of FFR in the management of coronary
computational fluid dynamics to simulate the spatial distribu- artery disease. The DEFER trial demonstrated that in patients
tion of flow and pressure in the coronary arteries from ana- with an intermediate-grade stenosis and normal FFR value
tomical information. CT angiography-derived fractional flow (>0.75), intervention can be deferred with comparable short-
reserve (CT-FFR) is a new technique that allows for assess- and long-term outcome. The landmark FAME trial random-
ment of the hemodynamic significance of epicardial coronary ized patients between revascularization guided by
artery disease from standard CT angiographic images. angiography and FFR and showed that the latter was associ-
ated with a lower adverse event rate (death, myocardial
infarction, and revascularization). The use of FFR to assess
Invasive FFR hemodynamic severity of stable coronary artery disease is
supported by European and American guidelines, particu-
Invasive measurement of the FFR has become the clinical larly when the presence of myocardial ischemia has not been
reference to determine the hemodynamic significance of epi- established by noninvasive tests [1, 2].
cardial coronary artery disease. While the name suggests

A. Coenen
Principles
Departments of Radiology and Cardiology, Erasmus University
Medical Center, Rotterdam, The Netherlands CFD is focused on the simulation of the flow of fluids.
F. Gijsen Computational fluid dynamics is applied in various areas of
Department of Biomedical Engineering, Erasmus University engineering and industrial design, for instance, the behavior
Medical Center, Rotterdam, The Netherlands of a flying aircraft or the blood flow interaction of mechani-
K. Nieman (*) cal heart valves or ventricular assist devices. For the purpose
Stanford University, School of Medicine, Cardiovascular Institute, of CT-FFR, CFD is applied to the anatomy from a CTA scan
Stanford, CA, USA
to simulate coronary blood flow. This simulation allows for
e-mail: [email protected]

© Humana Press 2019 767

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_60
768 A. Coenen et al.

Fig. 60.1  Overview of the

a b
CT-FFR procedure. The CTA
images are transferred to the
central HeartFlow facilities or
loaded onto a local CT-FFR
application (panel a and b).
From the CTA images, the
coronary artery lumen
contours are segmented (panel
c and d). The left ventricular
myocardium is segmented
from the CTA (panel e). The
resting total coronary blood
flow is estimated based on the
left ventricular myocardium
volume (panel e). The
microvascular resistance is
reduced and a hyperemic state
is simulated. The Navier-­
Stokes equations are solved
either for a full 3D model in c d
the HeartFlow solution or in a
simplified model for the
locally operated applications.
The CT-FFR outcome is
presented as a color-coded
map (panel f)

e f

computation of the flow, velocity, and pressure. Information a­ lgorithms have been described in the public domain, all of
needed to simulate flow are (1) 3D geometrical contours of which apply CFD in a slightly different manner. However, to
the lumen of the artery; (2) the inflow and boundary compute the fractional flow reserve from a CT angiogram,
­conditions, for instance, pressure; and (3) characteristics of most algorithms share a number of fundamental steps
the fluid, for instance, blood viscosity. Several CT-FFR (Fig. 60.1) [3, 4].
60  CT Angiography-Derived Fractional Flow Reserve 769

Obtaining the Anatomical Structure tions need to be performed. To reduce the computational

time, parallel supercomputers are being used [3]. While the
From the CT angiogram, the lumen of the entire coronary HeartFlow algorithm is a full 3D computational flow simula-
artery tree, to the extent the spatial resolution of the scan tion throughout the coronary artery tree, other applications
allows, needs to be segmented. While efforts are made to incorporate simplifications to allow for on-site CT-FFR sim-
automate the coronary segmentation process, a substantial ulations on more or less regular workstations. These locally
degree of manual correction is often still required. In the case performed CT-FFR applications with simplified models
of the algorithm by HeartFlow Inc., the segmentation, as reduce the computational intensity based on the principal
well as other steps, is performed at the central facilities after that for computation of a pressure drop, it is not necessary to
transferring the CT data. The HeartFlow algorithm applies a compute the entire flow field at each point in the coronary
subvoxel resolution technique to improve segmentation of artery [8–10]. This approach does have limitations as, for
the coronary arteries [5, 6]. Other algorithms require local example, shear stress computations are no longer possible.
operators and expertise to perform and/or correct the coro-
nary segmentation, which can be a time-consuming process.
 imulation of Hyperemia and CT-FFR
S
Computation
 stimation of the Global and Local
E
Resting Flow The CTA was acquired in a resting state, while the FFR is
normally measured during hyperemia. Wilson et al. investi-
First, the total coronary flow is estimated from the total myo- gated the effect of adenosine on the coronary circulation and
cardial mass. Allometric scaling laws describe the relation- demonstrated that adenosine reduces the microvascular
ship between the size of an organ and the required blood resistance by a factor of 0.21–0.24 [11]. To simulate a state
supply. The main work to determine the relationship between of hyperemia, the microvascular resistance of the coronary
the resting coronary blood flow and the size of the heart was arteries is virtually reduced, mimicking the vasodilatory
done by Choy et al. [7]. From the CTA the left ventricular effect of adenosine. By performing the CFD simulation and
myocardium can be segmented automatically, and based on computing the coronary blood flow and pressure during vir-
this mass, the total resting coronary blood flow can be esti- tual hyperemia, an FFR value can be calculated by dividing
mated [5, 8]. The total coronary blood flow is then distrib- the intracoronary pressure by the aortic pressure. Pressure
uted over the 3D coronary model. The pattern of distribution and FFR values are calculated throughout the coronary artery
is determined by the diameter size of the coronary branches tree and often displayed onto a three-dimensional coronary
based on the principle that form (diameter) will follow func- model using color coding.
tion (flow demand).

CT-FFR Applications
 omputation of the Coronary Blood Flow
C
and Pressure Performance of CFD onto coronary CT images including the
aortic root was developed by Charles Taylor and colleagues
Simulated flow is governed by the Navier-Stokes equations, at Stanford University and later commercialized by
and computation of the coronary blood flow requires solving HeartFlow Inc. (Redwood City, CA) [12]. HeartFlow offers
these equations throughout the 3D model. The Navier-Stokes CT-FFR as a remotely performed, full-service solution. The
equations describe the application of Newton’s second law client transfers the CT images and clinical information to
(force is determined by mass and acceleration) to fluid HeartFlow, where the complete workflow is performed. After
motion, combined with the conservation of mass. First the the CFD computation is completed, currently within 24  h,
3D model needs to be discretized, i.e., the model is divided the CT-FFR results are available by log-in into a secure
into small volumetric elements. The size of the elements may HeartFlow environment (Figs. 60.2 and 60.3). Several alter-
be varied, using larger elements around constant flow pat- native approaches to CT-FFR are under development.
terns for improved efficiency, and smaller elements where A  CT-FFR solution was developed by Siemens (cFFR,
flow alterations are expected. For each element in this volu- Forchheim, Germany), which can be performed locally on a
metric mesh, or finite-element model, the Navier-Stokes regular workstation (Fig.  60.4) [8]. Also Toshiba Medical
equations need to be solved, as the outcome from one ele- Systems (Odawara, Japan) developed a CT-FFR algorithm
ment is transferred to the adjacent elements. To do this for that can be processed on a regular computer, which was
the 3D coronary mesh, an enormous number of computa- recently evaluated in a clinical study [13]. Finally, Philips
770 A. Coenen et al.

Fig. 60.2  Patient with

a b c
extensive atherosclerosis and
moderate stenosis (arrows) of
the right (RCA, panel a) and
left anterior descending
coronary artery (LAD, panel
c) by coronary CT
angiography. CT-FFR
(HeartFlow) shows a decrease
in FFR toward the 0.80
threshold in both the LAD
and RCA (panel b). In the
absence of severe angina
complaints, the patient was
treated medically for the time
being

a b c

Fig. 60.3  Coronary CT angiography performed in a patient who sclerotic disease did result in an FFR of 0.71 by CT-FFR (panel b) and
underwent percutaneous coronary intervention of the left anterior 0.75 by invasive FFR (panel c). A moderate stenosis (arrow) in the right
descending coronary artery (LAD) with implantation of a bioresorbable coronary artery (RCA) did not cause a significant drop in CT-FFR
scaffold (arrow, panel b and c) 5 years before (panel b and c). Although (panel a and b)
there is no single lesion causing more than 50% stenosis, diffuse athero-

Healthcare recently presented a workstation-based FFR Validation of CT-Based Fractional Flow Reserve
model as well [14]. None of these on-site performed CT-FFR
solutions have been commercially released yet.  he HeartFlow Application
T
Recently, a new approach for FFR computation, where The CT-FFR solution by HeartFlow has been extensively
machine-learning is used to predict the hemodynamic impact evaluated (Table  60.1) [12, 16, 17]. The most recent NXT
based on angiographic anatomy instead of performing indi- trial demonstrated in a cohort of 251 patients that CT-FFR
vidual CFD computation, was introduced. The machine-­ correlates well with invasive FFR, with a per-patient and per-
learning approach greatly reduces the computational vessel sensitivity of 86% and 84%, specificity of 79% and
intensity and allows for CT-FFR computations in mere sec- 86%, positive predictive value of 65% and 61%, negative
onds [15]. Clinical validation of a machine-learning-based predictive value of 93% and 95%, and accuracy of 81% and
application is currently being undertaken. 86%, respectively. Restricted to lesions with an intermediate
stenosis severity (30–70%), where CT-FFR is most likely to
60  CT Angiography-Derived Fractional Flow Reserve 771

Fig. 60.4  Case example

a b
using the Siemens cFFR
application. A 40-year-old
man presented with exercise-­
related chest pain. In panel (a)
multiplanar reformation of the
CT angiogram shows
moderate stenosis of the distal
LAD just distal to the second
diagonal branch. Panel (b)
shows the 3D coronary model
segmented from the CTA. The
CT-FFR value was 0.93
indicating that the stenosis in c d
the LAD is not flow limiting
(panel c). Panel (d) displays
the invasive coronary
angiogram. Invasive FFR
measurements confirmed the
absence of a flow-limiting
stenosis (FFR 0.89)

Table 60.1  Per-vessel performance of CT angiography-derived FFR versus invasive FFR

Patient FFR Sensitivity Specificity Accuracy AUC
Design CT (vessel) ≤0.80 R (CTA) (CTA) (CTA) (CTA)
HeartFlow
Discover-flow [12] Multicenter prospective 64CT+ 103 36% 0.72 88% 82% 84% 0.90
(159) (91%) (40%) (59%) (0.75)
DEFACTO [16] Multicenter prospective 64CT+ 252 37% 0.63 80% 61% 69% 0.81
(407) (0.68)
NXT [17] Multicenter 64CT+ 251 21% 0.82 86% 84% 86% 0.93
prospective (484) (83%) (60%) (65%) (0.79)
Kawaji [45] Single-center 320CT 43 40% NA 93% 52% 69% 0.87
prospective (70)
Siemens
Renker [20] Single-center DSCT 53 30% 0.66 85% 85% – 0.92
retrospective (67) (90%) (34%) (0.72)
Coenen [21] Single-center DSCT 106 42% 0.59 88% 65% 75% 0.83
retrospective (189) (81%) (31%) (56%) (0.64)
Kruk [22] Single-center DSCT 90 43% 0.67 76% 72% 74% 0.84
retrospective (96) (100%) (2%) (44%) (0.66)
Yang [25] Single-center DSCT 72 39% 0.67 87% 77% 81% 0.89
retrospective (138) (94%) (66%) (78%) (0.85)
De Geer [24] Single-center DSCT 21 29% 0.77 83% 76% 78%
retrospective (23)
Kurata [23] Single-center DSCT 21 48% 0.74 100% 87% 93%
retrospective (29)
Toshiba
Ko [13] Single-center 320CT 30 48% 0.57 78% 87% 84% 0.88
prospective (58) (79%) (74%) (78%) (0.77)
Diagnostic performance reported on a per-vessel basis using invasive fractional flow reserve as the reference. Number of analyzed patients and
vessels with direct CT-FFR versus invasive FFR comparison. Sensitivity, specificity, and accuracy are on a per-vessel basis
FFR fractional flow reserve, R coefficient of correlation between CT-FFR and invasive FFR. Sensitivity, specificity, accuracy for CT-based FFR
(and for CT angiography). AUC area under the receiver operator curve (and for CTA), DSCT first- and second-generation dual-source CT
772 A. Coenen et al.

Per-patient Analysis Per-vessel Analysis

1 1

0.9 0.9
FFRCT
FFRCT
0.8 Site CT 0.8
Site CT
0.7 0.7

0.6 0.6
Sensitivity

Sensitivity
0.5 0.5

0.4 0.4

0.3 FFRCTAUC: 0.90 95% CI: 0.87, 0.94 0.3 FFRCTAUC: 0.93 95% CI: 0.91, 0.95
CT AUC: 0.81 95% CI: 0.76, 0.87 CT AUC: 0.79 95% CI: 0.74, 0.84
0.2 ∆AUC: 0.09 95% CI: 0.04, 0.14 0.2 ∆AUC: 0.14 95% CI: 0.09, 0.19
P=0.0008 P<0.0001
0.1 0.1

0 0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
1-Specificity 1-Specificity

Fig. 60.5  Receiver operator curves from the NXT trial based on 251 cantly larger area under the curve (AUC) compared with CTA, both on
patients, 484 vessels. In panel (a) the per-patient performance and in a per-patient and on a per-vessel basis. (From Norgaard et al. [17], with
panel (b) the per-vessel performance. CT-FFR resulted in a signifi- permission)

Planned Non-invasive Test Planned Invasive Angiography

NonObs CAD Obs CAD No ICA NonObs CAD Obs CAD No ICA
100% 100%
90% 90%
P = 0.95 P < 0.0001
80% 80%
70% 70%
60% 60%
50% 50%
40% 40%
30% 30%
20% 20%
10% 10%
0% 0%
Usual care FFRCT Usual care FFRCT
N (%): 6 (6.0) 13 (12.5) N (%): 137 (73.3) 24 (12.4)
P = 0.95 P < 0.0001

Fig. 60.6  Results from the platform trial, which evaluated the perfor- ography and CT-FFR in a proportion. Particularly in patients referred for
mance and diagnostic yield of invasive angiography, in patients either invasive angiography, the use of CTA and CT-FFR reduced the number
referred for noninvasive testing or invasive coronary angiography, in two of performed catheterizations and increased the proportion of positive
consecutive cohorts of which the second underwent coronary CT angi- invasive angiograms. (From Douglas et al. [18], with permission)

be implemented, sensitivity (80%) and specificity (85%) cost savings (Fig. 60.6) [18, 19]. The HeartFlow application
remained good. The diagnostic performance of CT-FFR was has been approved and is available for clinical use in many
superior to CTA alone (Table 60.1, Fig. 60.5) and allowed for parts of the world.
correct reclassification of 68% of patients with false-positive
CTA diagnoses [17]. The platform trial, which assessed the  orkstation-Based CT-FFR Applications
W
impact of CT-FFR on clinical management in consecutive The first generation of CT-FFR developed by Siemens was
cohorts, demonstrated that in patients referred for invasive evaluated in a number of single-center studies [20–25].
angiography, the use of CTA with CT-FFR reduced the pro- These cohorts were mostly retrospective cohorts of patients
portion of normal invasive ­angiograms and led to significant with a clinical indication for invasive FFR, which may have
60  CT Angiography-Derived Fractional Flow Reserve 773

increased the proportion of diffuse and/or intermediate were reported for CT-FFR, compared to relatively poor per-
severity angiographic disease (Fig. 60.7). Sensitivities in the formances for CTA, in particularly specificity, in these chal-
range of 76–88% and specificities in the range of 65–85% lenging populations. Outside the borderline CT-FFR range
of 0.75–0.85, the accuracy of CT-FFR can reach 95%, which
allows for confident exclusion or confirmation of hemody-
1.0 namic coronary disease without further need for testing [22].
The most recent CT-FFR solution by Siemens, which applies
a morphological analysis to predict hemodynamical signifi-
0.8 cance based on machine-learning, is under investigation in
the multicenter machine study. The first, overall promising
results using the CT-FFR application by Toshiba in a rela-
0.6
tively small, but prospective cohort were recently reported
[13]. Other techniques that use simplified models based on
CT-FFR

geometry data that can be extracted from CT are currently

under development [26].
0.4

 T-FFR Compared to Other Techniques

C
In a substudy of the NXT trial, CT-FFR demonstrated supe-
0.2
rior diagnostic performance in comparison to the translumi-
nal attenuation gradient along the coronary artery on CT
angiography [10]. CT-FFR by Siemens was shown to per-
0.0
form similarly well as static CT myocardial perfusion imag-
0.0 0.2 0.4 0.6 0.8 1.0
ing, while integration of CT-FFR with dynamic CT perfusion
invasive FFR
imaging provided incremental value over each technique
Fig. 60.7  Technical validation of an on-site performed CT-FFR appli- separately (Fig.  60.8) [25, 27]. Although in meta-analyses
cation in a patient’s cohort with clinically indicated coronary CT angi- CT-FFR performs well in comparison to other stress imaging
ography and invasive FFR measurements showed promising diagnostic techniques, no direct comparisons have been performed. No
performance. The correlation coefficient between CT-FFR and invasive
FFR was 0.59. CT-FFR improved the specificity of coronary CT angi- comparisons between the respective CT-FFR applications
ography from 31% to 65%. (From Coenen et al. [21], with permission) have been published either.

Fig. 60.8  Case example of a

a b
61-year-old man presenting
with stable chest pain. CTA
demonstrated a partially
calcified, severely stenosed
lesion in the LCX (panel a).
The CT-FFR simulation
demonstrated a significant
pressure drop (CT-FFR 0.33)
over the lesion (panel b).
CT-MPI, image fused with the
CT angiogram, shows a
perfusion defect in the basal
lateral wall (panel c). The
corresponding invasive
angiography confirmed the c d
short, severe stenosis, with an
invasively measured FFR of
0.48 (panel d)
774 A. Coenen et al.

Challenges and Limitations of percutaneous intervention can be predicted [30]. Simulated

stenting and treatment optimization would be especially of
While the CT-FFR approach represents an attractive means interest in patients with serial stenosis or complex bifurca-
to derive functional information without the need for addi- tion lesions where selection of the stenosis most benefiting
tional testing after CT angiography, there are also challenges. from revascularization is not straightforward. Surgical revas-
Generally, the algorithms rely on artifact-free imaging for cularization can be simulated in a similar fashion, although
reconstruction of the coronary artery models. Adherence to the accuracy would be limited by uncertainty about the status
CT imaging recommendations for optimal image quality is and dimensions of available grafting material.
explicitly recommended. In particular vessel discontinuity
between slabs as a result of arrhythmia or other displace-
ments affects the performance of CT-FFR negatively. Predicting Events
Although calcifications obscure the coronary lumen and
compromise delineation of the coronary lumen, a substantial Coronary endothelial cells respond to blood flow-induced
negative effect of calcifications on diagnostic performance shear stress; low shear stress regions are prone to plaque
could not be demonstrated [8, 28]. The diagnostic perfor- development. The mechanism behind atherosclerotic plaque
mance of CT-FFR in STEMI patients with multivessel dis- development and shear stress levels is multifactorial and cur-
ease was only modest [29], suggesting that the role of rently still being investigated. The initial response of the
CT-FFR in STEMI patients is limited. CT-FFR has not been arterial wall to plaque development is outward or positive
validated in patients with bypass graft, stents, prior myocar- remodeling, also called the Glagov effect [31]. However, if
dial infarction, congenital heart disease, or other out-of-the-­ plaque progression continues, also the coronary lumen will
ordinary cardiac morphologies. Momentarily, CT-FFR become affected. This reduction of the coronary lumen will
cannot be used to assess dynamic coronary compression result in an increase in local velocity and shear stress. While
(myocardial bridging, intra-arterial course of coronary the exact pathophysiology is still under debate [32], increased
anomaly), microvascular disease, or functional coronary shear stress is associated with increased plaque vulnerability
artery disease (spasm). [33, 34]. These findings are corroborated by clinical studies
The HeartFlow approach is most extensively validated that associate the location of plaque rupture with high wall
and currently the only commercially available CT-FFR shear stress [34–37]. Most clinical computation of wall shear
application. Datasets are transferred to and processed by the stress is based on intracoronary imaging, which can be sup-
HeartFlow company. Off-site processing has the advantage plemented by cardiac CT for vessel orientation in 3D space.
that no local expertise or time investment, beyond the acqui- Although coronary CT angiography does not provide the
sition of high-quality CT data, is required. Drawbacks same spatial resolution as intravascular ultrasound or optical
include the need to transfer patient data and the substantial coherence tomography, preliminary data suggest that shear
costs. Return times are currently around 24 h, but expected to stress can be computed based on CT alone [38, 39]. In the
shorten in the future. On-site CT-FFR approaches, as devel- EMERALD study, hemodynamic forces calculated with
oped by Siemens and Toshiba, leave the local investigator in CFD from CT scans performed prior up to 2  years before
control of the post-processing but demand substantial effort. myocardial infarction differentiated the ultimate culprit
User operation requires training and manual processing lesions from stable plaque. This suggests that the
times are still substantial. Segmentation of the coronary ­hemodynamic force onto coronary plaques can possibly pre-
arteries is particularly time-consuming, but may become dict future events [40].
more acceptable with the development of more automatic
segmentation tools. Although none of the locally performed
CT-FFR applications have been commercially released, costs Intracardiac Blood Flow Simulation
are expected to be less than the HeartFlow application.
CFD can be used to investigate contractile cardiac function
[41]. The high spatial resolution of the CT allows for precise
Further Development anatomical differentiation of the lumen in all cardiac phases.
CFD computations could be used to visualize the simulated
Revascularization Planning cardiac function and blood flow through the main cardiac
cavities and vessels. Better understanding of intracardiac
In addition to the present coronary status, simulations can flow patterns could provide valuable insights, for example,
also be performed on altered coronary models. By restoring in ventricular dyssynchrony [42]. Lantz et  al. presented a
dimensions of a stenotic vessel and simulation of the implan- framework for CFD simulation of the global cardiac func-
tation of a coronary stent, the potential hemodynamic benefit tion from time-resolved CT angiography and computed
60  CT Angiography-Derived Fractional Flow Reserve 775

blood flow in the pulmonary veins, atriums, left ventricle, 12. Koo BK, Erglis A, Doh JH, Daniels DV, Jegere S, Kim HS, et al.
Diagnosis of ischemia-causing coronary stenoses by noninvasive
aortic valve, aortic sinus, coronary arteries, and part of the fractional flow reserve computed from coronary computed tomo-
ascending aorta. Because of the high spatial resolution of graphic angiograms. Results from the prospective multicenter
CT, even flow patterns around the papillary muscles can be DISCOVER-FLOW (Diagnosis of Ischemia-Causing Stenoses
visualized [43]. CFD simulations of diseased cardiac valves Obtained Via Noninvasive Fractional Flow Reserve) study. J Am
Coll Cardiol. 2011;58(19):1989–97.
could improve our understanding of hemodynamic impact 13. Ko BS, Cameron JD, Munnur RK, Wong DT, Fujisawa Y,

[44]. CFD simulation can also be used to assess the hemo- Sakaguchi T, et  al. Noninvasive CT-derived FFR based on struc-
dynamic performance of prosthetic valves. However, large tural and fluid analysis: a comparison with invasive FFR for detec-
studies evaluating the clinical application of CFD computa- tion of functionally significant stenosis. JACC Cardiovasc Imaging.
2017;10(6):663–73.
tions of the cardiac and valvular function by CT are cur- 14. Nickisch H, Lamash Y, Prevrhal S, Freiman M, Vembar M, Goshen
rently lacking. L, et  al. Learning patient-specific lumped models for interactive
coronary blood flow simulations. In: Navab N, Hornegger J, Wells
WM, Frangi AF, editors. Medical image computing and computer-­
assisted intervention  – MICCAI 2015: 18th international confer-
References ence, Munich, Germany, October 5–9, 2015, proceedings, part
II. Cham: Springer International Publishing; 2015. p. 433–41.
1. Authors/Task Force members, Windecker S, Kolh P, Alfonso F, 15. Itu L, Rapaka S, Passerini T, Georgescu B, Schwemmer C,

Collet JP, Cremer J, et al. 2014 ESC/EACTS guidelines on myo- Schoebinger M, et al. A machine-learning approach for computa-
cardial revascularization: the task force on myocardial revascu- tion of fractional flow reserve from coronary computed tomogra-
larization of the European Society of Cardiology (ESC) and the phy. J Appl Physiol. 2016;121(1):42–52.
European Association for Cardio-Thoracic Surgery (EACTS) devel- 16. Min JK, Leipsic J, Pencina MJ, Berman DS, Koo BK, van Mieghem
oped with the special contribution of the European Association of C, et al. Diagnostic accuracy of fractional flow reserve from ana-
Percutaneous Cardiovascular Interventions (EAPCI). Eur Heart J. tomic CT angiography. JAMA. 2012;308(12):1237–45.
2014;35(37):2541–619. 17. Norgaard BL, Leipsic J, Gaur S, Seneviratne S, Ko BS, Ito H, et al.
2. Levine GN, Bates ER, Blankenship JC, Bailey SR, Bittl JA, Cercek Diagnostic performance of noninvasive fractional flow reserve
B, et al. 2011 ACCF/AHA/SCAI guideline for percutaneous coro- derived from coronary computed tomography angiography in sus-
nary intervention: a report of the American College of Cardiology pected coronary artery disease the NXT trial (analysis of coronary
Foundation/American Heart Association task force on practice blood flow using CT angiography: next steps). J Am Coll Cardiol.
guidelines and the Society for Cardiovascular Angiography and 2014;63(12):1145–55.
Interventions. Circulation. 2011;124(23):e574–651. 18. Douglas PS, Pontone G, Hlatky MA, Patel MR, Norgaard BL,
3. Taylor CA, Fonte TA, Min JK.  Computational fluid dynamics Byrne RA, et  al. Clinical outcomes of fractional flow reserve by
applied to cardiac computed tomography for noninvasive quantifi- computed tomographic angiography-guided diagnostic strategies
cation of fractional flow reserve: scientific basis. J Am Coll Cardiol. vs. usual care in patients with suspected coronary artery disease: the
2013;61(22):2233–41. prospective longitudinal trial of FFR(CT): outcome and resource
4. Morris PD, Narracott A, von Tengg-Kobligk H, Silva Soto DA, impacts study. Eur Heart J. 2015;36(47):3359–67.
Hsiao S, Lungu A, et al. Computational fluid dynamics modelling 19. Douglas PS, De Bruyne B, Pontone G, Patel MR, Norgaard BL,
in cardiovascular medicine. Heart. 2016;102(1):18–28. Byrne RA, et al. 1-year outcomes of FFRCT-guided care in patients
5. Min JK, Taylor CA, Achenbach S, Koo BK, Leipsic J, Norgaard with suspected coronary disease: the PLATFORM study. J Am Coll
BL, et al. Noninvasive fractional flow reserve derived from coro- Cardiol. 2016;68(5):435–45.
nary CT angiography: clinical data and scientific principles. JACC 20. Renker M, Schoepf UJ, Wang R, Meinel FG, Rier JD, Bayer RR,
Cardiovasc Imaging. 2015;8(10):1209–22. et al. Comparison of diagnostic value of a novel noninvasive coro-
6. Schaap M, van Walsum T, Neefjes L, Metz C, Capuano E, de nary computed tomography angiography method versus standard
Bruijne M, et  al. Robust shape regression for supervised ves- coronary angiography for assessing fractional flow reserve. Am J
sel segmentation and its application to coronary segmentation in Cardiol. 2014;114(9):1303–8.
CTA. IEEE Trans Med Imaging. 2011;30(11):1974–86. 21. Coenen A, Lubbers MM, Kurata A, Kono A, Dedic A, Chelu RG,
7. Choy JS, Kassab GS. Scaling of myocardial mass to flow and mor- et  al. Fractional flow reserve computed from noninvasive CT
phometry of coronary arteries. J Appl Physiol. 2008;104(5):1281–6. angiography data: diagnostic performance of an on-site clinician-­
8. Coenen A, Lubbers MM, Kurata A, Kono A, Dedic A, Chelu RG, operated computational fluid dynamics algorithm. Radiology.
et  al. Coronary CT angiography derived fractional flow reserve: 2015;274(3):674–83.
methodology and evaluation of a point of care algorithm. J 22. Kruk M, Wardziak L, Demkow M, Pleban W, Pregowski J,

Cardiovasc Comput Tomogr. 2016;10(2):105–13. Dzielinska Z, et  al. Workstation-based calculation of CTA-­
9. Itu L, Sharma P, Mihalef V, Kamen A, Suciu C, Lomaniciu D, based FFR for intermediate stenosis. JACC Cardiovasc Imaging.
editors. A patient-specific reduced-order model for coronary cir- 2016;9(6):690–9.
culation. 2012 9th IEEE international symposium on biomedical 23. Kurata A, Coenen A, Lubbers MM, Nieman K, Kido T, Kido T,
imaging (ISBI); 2012. 2–5 May 2012. et al. The effect of blood pressure on non-invasive fractional flow
10. Ko BS, Wong DT, Norgaard BL, Leong DP, Cameron JD, Gaur S, reserve derived from coronary computed tomography angiography.
et al. Diagnostic performance of transluminal attenuation gradient Eur Radiol. 2017;27(4):1416–23.
and noninvasive fractional flow reserve derived from 320-detector 24. De Geer J, Sandstedt M, Bjorkholm A, Alfredsson J, Janzon M,
row CT angiography to diagnose hemodynamically significant cor- Engvall J, et al. Software-based on-site estimation of fractional flow
onary stenosis: an NXT substudy. Radiology. 2016;279(1):75–83. reserve using standard coronary CT angiography data. Acta Radiol.
11. Wilson RF, Wyche K, Christensen BV, Zimmer S, Laxson
2016;57(10):1186–92.
DD. Effects of adenosine on human coronary arterial circulation. 25. Yang DH, Kim YH, Roh JH, Kang JW, Ahn JM, Kweon J, et al.
Circulation. 1990;82(5):1595–606. Diagnostic performance of on-site CT-derived fractional flow
776 A. Coenen et al.

reserve versus CT perfusion. Eur Heart J Cardiovasc Imaging. 36. Cicha I, Worner A, Urschel K, Beronov K, Goppelt-Struebe M,
2017;18(4):432–40. Verhoeven E, et  al. Carotid plaque vulnerability: a positive feed-
26. Schrauwen JT, Koeze DJ, Wentzel JJ, van de Vosse FN, van der back between hemodynamic and biochemical mechanisms. Stroke.
Steen AF, Gijsen FJ. Fast and accurate pressure-drop prediction in 2011;42(12):3502–10.
straightened atherosclerotic coronary arteries. Ann Biomed Eng. 37. Fujii K, Kobayashi Y, Mintz GS, Takebayashi H, Dangas G,

2015;43(1):59–67. Moussa I, et  al. Intravascular ultrasound assessment of ulcerated
27. Coenen A, Rossi A, Lubbers MM, Kurata A, Kono AK, Chelu RG, ruptured plaques: a comparison of culprit and nonculprit lesions
Segreto S, Dijkshoorn ML, Wragg A, van Geuns RM, Pugliese F, of patients with acute coronary syndromes and lesions in patients
Nieman K.  Integrating CT myocardial perfusion and CT-FFR in without acute coronary syndromes. Circulation. 2003;108(20):
the work-up of coronary artery disease. JACC Cardiovasc Imaging. 2473–8.
2017;10(7):760–70. 38. Zhang JM, Luo T, Tan SY, Lomarda AM, Wong AS, Keng FY, et al.
28. Norgaard BL, Gaur S, Leipsic J, Ito H, Miyoshi T, Park SJ, et al. Hemodynamic analysis of patient-specific coronary artery tree. Int
Influence of coronary calcification on the diagnostic performance of J Numer Method Biomed Eng. 2015;31(4):e02708.
CT angiography derived FFR in coronary artery disease a substudy 39. van der Giessen AG, Groen HC, Doriot PA, de Feyter PJ,

of the NXT trial. JACC Cardiovasc Imaging. 2015;8(9):1045–55. van der Steen AF, van de Vosse FN, et  al. The influence
29. Gaur S, Taylor CA, Jensen JM, Botker HE, Christiansen EH,
of boundary conditions on wall shear stress distribution in
Kaltoft AK, et al. FFR derived from coronary CT angiography in patients specific coronary trees. J Biomech. 2011;44(6):
nonculprit lesions of patients with recent STEMI. JACC Cardiovasc 1089–95.
Imaging. 2017;10(4):424–33. 40. Koo BK.  Haemodynamic force analysis improves non-invasive

30. Kim KH, Doh JH, Koo BK, Min JK, Erglis A, Yang HM, et al. A prediction of risk of ACS: the results of first-in-man EMERALD
novel noninvasive technology for treatment planning using virtual study (exploring the mechanism of the plaque rupture in acute coro-
coronary stenting and computed tomography-derived computed nary syndrome using coronary CT angiography and computational
fractional flow reserve. JACC Cardiovasc Interv. 2014;7(1):72–8. fluid dynamics). Imaging and functional assessment presented at:
31. Glagov S, Weisenberg E, Zarins CK, Stankunavicius R, Kolettis EuroPCR; May 16–19, 2016; Paris. 2016.
GJ. Compensatory enlargement of human atherosclerotic coronary 41. Chan BT, Lim E, Chee KH, Abu Osman NA. Review on CFD simu-
arteries. N Engl J Med. 1987;316(22):1371–5. lation in heart with dilated cardiomyopathy and myocardial infarc-
32. Wentzel JJ, Chatzizisis YS, Gijsen FJ, Giannoglou GD, Feldman CL, tion. Comput Biol Med. 2013;43(4):377–85.
Stone PH. Endothelial shear stress in the evolution of coronary ath- 42. Aguado-Sierra J, Krishnamurthy A, Villongco C, Chuang J,

erosclerotic plaque and vascular remodelling: current understand- Howard E, Gonzales MJ, et  al. Patient-specific modeling of dys-
ing and remaining questions. Cardiovasc Res. 2012;96(2):234–43. synchronous heart failure: a case study. Prog Biophys Mol Biol.
33. Gijsen FJ, Wentzel JJ, Thury A, Mastik F, Schaar JA, Schuurbiers 2011;107(1):147–55.
JC, et  al. Strain distribution over plaques in human coronary 43. Lantz J, Henriksson L, Persson A, Karlsson M, Ebbers T. Patient-­
arteries relates to shear stress. Am J Physiol Heart Circ Physiol. specific simulation of cardiac blood flow from high-resolution com-
2008;295(4):H1608–14. puted tomography. J Biomech Eng. 2016;138(12):121004.
34. Samady H, Eshtehardi P, McDaniel MC, Suo J, Dhawan SS,
44. Astorino M, Hamers J, Shadden SC, Gerbeau JF.  A robust and
Maynard C, et  al. Coronary artery wall shear stress is associated efficient valve model based on resistive immersed surfaces. Int J
with progression and transformation of atherosclerotic plaque Numer Method Biomed Eng. 2012;28(9):937–59.
and arterial remodeling in patients with coronary artery disease. 45. Kawaji T, Shiomi H, Morishita H, Morimoto T, Taylor CA, Kanao
Circulation. 2011;124(7):779–88. S, et al. Feasibility and diagnostic performance of fractional flow
35. Gijsen F, van der Giessen A, van der Steen A, Wentzel J.  Shear reserve measurement derived from coronary computed tomogra-
stress and advanced atherosclerosis in human coronary arteries. J phy angiography in real clinical practice. Int J Cardiovasc Imaging.
Biomech. 2013;46(2):240–7. 2017;33(2):271–81.
 CT Myocardial Perfusion Imaging:
Arterial First-Pass Imaging 61
Florian Schwarz, Amadeus Altenburger,
Michael Gebhard, and Christian Thilo

Coronary computed tomography angiography (CCTA) is an increase in coronary blood flow. For this purpose, coro-
well established for its high reliability in the exclusion of nary arteries have a substantial capacity for autoregulation
coronary artery disease (CAD) based on abundant high-class of blood flow. This autoregulation has long been recog-
evidence demonstrating excellent negative likelihood ratios nized as the reason why coronary stenoses under rest condi-
both in emergency and elective settings [1–4]. tions become flow-limiting (and thus create myocardial
On the other hand, when coronary stenoses are detected, ischemia) only when reaching at least 85% diameter steno-
an estimation of their hemodynamic relevance remains sis [8]. At lower degrees of stenosis, autoregulation will
­difficult, and typically, CCTA-based estimations of hemo- counteract any potentially hemodynamic effect under rest
dynamic relevance correlate only moderately with conditions.
­perfusion-based modalities [5]. On the other hand, when coronary blood flow increases
The recognition of the hemodynamic relevance of stenoses, due to increased cardiac oxygen demand or due to the
however, is of high importance as it has considerable prognos- administration of vasodilatory drugs (i.e., pharmacological
tic implications and directly influences treatment decisions [6, stress by application of dipyridamole, adenosine, or regad-
7]: nonobstructive CAD is associated with better outcomes enoson), the autoregulatory capacity will be “overwhelmed”
when treated with best medical therapy, while patients with at lower degrees of coronary stenosis with segmental myo-
flow-limiting stenoses usually benefit from angioplasty and cardial ischemia detectable at intermediate stenoses in the
stent implantation or coronary artery bypass surgery. 50% range [8, 9]. Therefore, the administration of pharma-
In particular, the detection of intermediate-degree coro- cological stress agents will significantly improve sensitivity
nary stenoses in CCTA frequently results in the recommen- for stenosis detection. Most traditional imaging modalities
dation for additional tests for the visualization of potential based on the direct visualization of myocardial ischemia
myocardial ischemia with the treatment consequences deter- include acquisitions of myocardial perfusion under stress as
mined only after this perfusion information is obtained. well as under rest to detect reversible and nonreversible
­perfusion defects.

Physiological Background
General Technical Considerations
The contractile function of the heart depends on aerobic
metabolism and thus oxygen supply via the coronary arter- It has long been hypothesized that the visualization and
ies. Due to the high oxygen extraction fraction of the heart quantification of segmental myocardial contrast uptake in
even under rest conditions, any excess oxygen demand of principle would be feasible with CT – initial experiments in
the heart during increased activity can only be satisfied by myocardial infarcts date back to the late 1970s [10–12] with
more specific reports on the quantification of myocardial
perfusion for the electron-beam CT platform [8, 13–17].
F. Schwarz (*) · A. Altenburger · M. Gebhard Approaches for the analysis of perfusion in CT are based
Department of Diagnostic and Interventional Radiology and
on the visualization of the first pass of iodinated contrast
Neuroradiology, Klinikum Augsburg, Augsburg, Germany
e-mail: [email protected] material after i.v. administration, similarly as in dynamic
contrast-enhanced MRI.
C. Thilo
Department of Cardiology, Klinikum Augsburg, Herzzentrum Clinically available iodinated contrast materials are
Augsburg-Schwaben, Augsburg, Germany nonionic small molecules which distribute freely in the

© Humana Press 2019 777

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_61
778 F. Schwarz et al.

Fig. 61.1  Overview of the different scan protocols which have been absolute quantification of iodine content within the myocardium which
applied to the visualization of myocardial perfusion in CT along with might serve as a surrogate marker for myocardial blood flow. Static CT
exemplary images and the major derivable parameter as the basis for perfusion approaches finally must rely on attenuation differences which
perfusion evaluation: Dynamic CT perfusion allows for a true quantifi- due to the linear relationship with iodine concentration allow for an
cation of myocardial blood flow. Static CT perfusion approaches imple- assessment of relative differences in myocardial iodine uptake
menting a dual-energy acquisition technique still allow for a precise

intravascular and with little delay in the extracellular space In general, technical solutions for the visualization of
but do not enter cells to any relevant extent. It has been myocardial perfusion in CT can be grouped into one of two
reported that the administration of iodinated contrast major approaches: dynamic or static CT perfusion imaging
agents has an influence on coronary blood flow, inducing (Fig. 61.1).
an initial reduction of blood flow followed by a hyperemic
response [18]. This, however, seems to play a considerably
smaller role for low-­osmolarity nonionic contrast agents Dynamic CT Perfusion
[19]. Importantly, there is significant diffusion of the con-
trast agents into the interstitial space. The first-pass extrac- “Dynamic CT perfusion” imaging refers to approaches in
tion of iodinated contrast is around 1/3 with maximal which repeated ECG-synchronized acquisitions of the left
vasodilation and even higher at lower flow rates [20]. ventricular myocardium are performed during the first pass
Since for a given tube voltage attenuation behaves virtu- of iodinated contrast material, similar to the scan sequences
ally linear to iodine concentration [21], hypoperfused areas underlying first-pass myocardial perfusion in MRI [23].
of the myocardium can be appreciated by hypoattenuation of Thereby the left ventricular myocardium is repeatedly sam-
ischemic myocardial segments in comparison with normal pled during the first pass of contrast material. After motion
myocardium. correction, the multiphasic dataset can be used to generate
Despite the mentioned early attempts, a widespread clini- time-attenuation curves for each voxel of the LV myocar-
cal application of myocardial CT perfusion imaging has dium on which the typical perfusion models can be applied
remained elusive until very recently. Only with the introduc- yielding quantitative perfusion parameters such as myocar-
tion of the recent scanner generations, the visualization of dial blood flow [24].
myocardial perfusion has become feasible, and the prospect The technical prerequisite for these approaches either is a
of coronary CT as a true “one-stop” shop modality has detector width covering the entire extension of the heart
spurred considerable scientific interest and research activity along the z-axis or  – in the case of narrow detectors  – a
in this direction [22]. shuttle-­
mode acquisition. For a full discussion of the
61  CT Myocardial Perfusion Imaging: Arterial First-Pass Imaging 779

u­nderlying technical prerequisites as well as the specific curves from which all mentioned curve features can be
advantages of these approaches, the reader is referred to derived, the static acquisition technique must rely on the dif-
Chap. 63 for more details. ferences in maximal attenuations (in HU), while other curve
parameters will remain unaddressed.
Consequentially, with the acquisition of only a singular
Static CT Perfusion timepoint, the correct acquisition timing in relation to the
contrast material bolus will be critical – much more so than
“Static CT perfusion” imaging on the other hand encom- for dynamic approaches where a wide temporal window is
passes approaches which are based on a single scan of the recorded beginning prior to the arrival of the contrast bolus
myocardium during contrast inflow. This group has also been in the myocardium. The optimal time of image acquisition
referred to as “arterial first-pass imaging” or “single-shot will be when the difference in attenuations between normal
acquisition.” and ischemic myocardium is most pronounced.
The rationale of this approach is to acquire the image at a In comparison with dynamic approaches, static CT perfu-
timepoint at which the hemodynamic effect of flow-limiting sion approaches are associated with considerably lower
stenosis is most apparent, i.e., in which due to the flow-­ effective doses, as low as 0.9 ± 0.1 mSv for static stress CT
limiting nature of the stenosis, contrast material will not have perfusion acquisitions [25]. Furthermore, only in static CT
washed into myocardial segments distal to stenoses, while perfusion, the acquisition of the coronary CTA dataset can be
myocardial segments not affected by coronary stenoses will combined with myocardial perfusion under rest conditions,
already show normal myocardial enhancement. while in dynamic perfusion this acquisition must be per-
In principle, all CT acquisition techniques are feasible for formed as a distinct scan.
static CT perfusion imaging (i.e., retrospective ECG gating, A technical variety of static CT perfusion is the use of a
prospective ECG triggering, and prospective ECG acquisi- dual-energy acquisition technique (Fig.  61.1). Dual-energy
tion with high-pitch factor). A closer look at the time-­ acquisition protocols promise to combine the higher contrast
attenuation curves of ischemic vs. normal myocardial resolution of a low-kV dataset with the lower noise of a high-
segments (Fig. 61.2) fosters an understanding of the concept ­kV dataset and thus facilitate the visualization of perfusion
of static vs. dynamic perfusion evaluation. In comparison heterogeneities within the myocardium, i.e., ischemic seg-
with normal myocardial segments, the enhancement in isch- ments [26–28]. It has been confirmed that the detection of
emic myocardium starts later, has a shallower upslope, and is ischemia is improved by use of a dual-energy acquisition
overall significantly less pronounced. While dynamic acqui- technique [29]. Furthermore, the attenuation characteristics
sitions permit an extensive sampling of time-attenuation at distinct photon energy spectra can be used to calculate the

Fig. 61.2 Time-attenuation
curves (TAC) of ischemic vs.
reference myocardium during
first pass of intravenously
administered contrast
material. TAC in ischemic
segments begins to rise later,
runs shallower, and has a later
peak. This creates a window
of approx. 8 s for the
visualization of ischemic
segments using static CT
perfusion protocols
780 F. Schwarz et al.

absolute concentration of iodine in each voxel [30]. It is con- tissue. When performing two perfusion acquisitions in
ceivable that the absolute iodine concentrations in the myo- sequence, there will be residual extracellular contrast mate-
cardium during first pass of contrast material might show rial both within the myocardium and within scar tissue. This
some correlation with MBF values thus permitting quantita- residual contrast material can be problematic as it will fur-
tive statements about myocardial perfusion. Again, a full dis- ther mitigate the already small attenuation difference
cussion of the physical background of the various technical between normal and ischemic myocardium in the second
implementations by different vendors and the available clini- scan potentially resulting in lower sensitivity for ischemia
cal evidence is beyond the scope of this chapter, and the detection. This phenomenon is to be expected for at least
reader is referred to Chap. 62. 15 min after contrast administration. Since the stress perfu-
sion scan should self-evidently have highest diagnostic accu-
racy for the detection (and thus also exclusion) of myocardial
Scan Protocol Considerations in Static CT perfusion defects, a stress-first approach would seem
Perfusion Imaging advantageous.
Similarly, a stress-first approach guarantees that the
To facilitate a distinction between ischemic myocardium and administration of both a beta-blocker and of sublingual
scar tissue, it has traditionally been considered a prerequisite nitrates  – encouraged prior to the CCTA/rest perfusion
to acquire perfusion scans under rest as well as under stress acquisition – will not interfere with the perfusion situation
conditions and thereby determine whether perfusion defects during stress perfusion. Particularly for beta-blockers, there
are “reversible” (which proves ischemia) or fixed (which is some evidence from the nuclear medicine myocardial per-
favors scar tissue, i.e., an old myocardial infarct). fusion literature that their application can hide or reduce the
In static CT perfusion imaging, careful planning of the size and/or severity of myocardial perfusion deficits with
scan delay allows the same dataset to be used for the CT potentially negative effects on the sensitivity of myocardial
coronary angiogram as well as the visualization of myocar- perfusion imaging [32]. However, in the CORE320 study in
dial perfusion under rest conditions (Fig. 61.3). As we will which an aggressive beta-blocker regime was applied at the
see in the following paragraphs, the myocardial contrast beginning of a rest-stress CT perfusion protocol, no negative
wash-in kinetics are such that this holds true even when scan effect of metoprolol on the diagnostic accuracy of CT perfu-
protocols are used that acquire data during several heart- sion was observed, and CT perfusion still had higher diag-
beats, i.e., prospective triggering or retrospective gating. nostic accuracy than SPECT myocardial perfusion imaging
The dataset used for the analysis of myocardial stress per- for the detection of significant CAD.
fusion, however, does currently not lend itself to the visual- A rest-stress approach (i.e., rest perfusion/CCTA first) has
ization of the coronary arteries themselves since the the main advantage that it more closely follows the logical
administration of vasodilatory drugs to trigger an increase in diagnostic pathway for CAD and reduces unnecessary stress
coronary blood flow usually results in a compensatory perfusion acquisitions: in all patients in whom the CCTA
increase in heart rate of approximately 20 bpms even when component already excludes obstructive CAD, a stress per-
beta-blockers had been administered previously [31]. fusion scan can be omitted, thus resulting in considerable
reduction of total effective dose. If the CCTA component
shows obstructive CAD, the stress perfusion can then be per-
Sequence of Stress and Rest Acquisitions formed with a specific diagnostic objective.
Furthermore, a rest-stress approach results in lower heart
According to the order of the acquisition of rest and stress rates during the CTA acquisition, which is desirable for high-
CT perfusion (with acquisitions at rest being simultaneously est possible image quality of the CTA scan: George et  al.
used for the visualization of the coronary arteries), a stress-­ showed that in the wake of a recent adenosine-stress perfu-
rest approach (stress perfusion first) and a rest-stress (rest sion acquisition, heart rates during CTA will be 7  bpms
perfusion first) approach can be distinguished which both higher than at baseline, suggesting that adenosine (or con-
have specific advantages and disadvantages. trast administration) triggers effects that persist despite the
The main advantage of the stress-rest approach is that the very short biological half-life of adenosine [33]. These
visualization of hypoattenuated ischemic myocardial seg- effects can be expected to be even more pronounced when
ments during stress perfusion is not compromised by any regadenoson is used due to its longer half-life.
prior administration of contrast material, a phenomenon also This conundrum of the optimal order of scans will most
referred to as contrast material contamination of the myocar- likely be resolved by integrating the pretest probability for
dium. Because of the small molecule nature of iodinated obstructive CAD, as first pointed out by Techasith et al. [34]
contrast material, there is extravasation into the extracellular and readily taken up by other authors [35]: in patients with
space, a phenomenon utilized for late enhancement in scar an intermediate to high likelihood of obstructive CAD, the
61  CT Myocardial Perfusion Imaging: Arterial First-Pass Imaging 781

a b

c d

Fig. 61.3  A 59-year-old patient with recurrence of stable angina Four-chamber view of rest perfusion shows very mild subendocardial
8  years after coronary artery bypass surgery. (a) Volume-rendering hypoattenuation in the apex. (d) Four-chamber view of stress perfusion
technique (VRT) reconstruction demonstrates ACVB to LAD. (b) acquired 3 min after starting a continuous infusion of adenosine with an
Curved planar reformation (CPR) of the venous bypass demonstrates increase in heart rate from 75 to 90 bpm demonstrates severe subendo-
normal vessel patency without any obvious high-grade stenoses – how- cardial hypoattenuation, consistent with a reversible perfusion defect in
ever, the exact status of the distal anastomosis remained uncertain. (c) the left ventricular anterior wall and apical region
782 F. Schwarz et al.

stress-rest approach (i.e., stress perfusion first) seems more CORE320 [31]. In this cohort undergoing a rest-stress static
advantageous as the proportion of unnecessary stress perfu- CT perfusion protocol prospectively triggered to 75%–95%
sion scans will be small  – defined as cases in which the of the RR interval, 80% of patients had excellent or good
stress perfusion acquisition will retrospectively prove image quality and only 3% of participants poor image qual-
unnecessary since CCTA excludes obstructive CAD. On the ity. Poor image quality was observed most frequently in
other hand, in patients with low pretest probability for heart rates ≥80  bpm (overall, 3%; in the subgroup with
obstructive CAD, the rest-stress approach clearly seems to HR  ≥80  bpm, 6%). The optimal diastolic phase for the
be favorable since in most of these patients, the stress perfu- assessment of myocardial perfusion was earlier in diastole
sion will be omissible. for lower heart rates and later in diastole for higher heart
rates (optimal phase: from 77% in patients with heart rates of
<65 bpm to 90% in patients with ≥80 bpm) [31].
 ptimal Scan Delay for Static CT Myocardial
O
Perfusion
Reconstruction and Postprocessing
Bischoff et  al. [36], Tanabe et  al. [37], and Pelgrim et  al.
[38] have investigated the optimal time window for image There is some evidence suggesting that the optimal postpro-
acquisition in relation to the contrast bolus by a detailed cessing tool differs between diastolic and systolic acquisi-
analysis of the time-attenuation curves derived from tions: comparing the effects of diastolic and systolic
dynamic stress CT perfusion datasets acquired in the con- reconstructions and various postprocessing modes (MPRs,
text of recent clinical studies. All three groups report a time MinIPs, MIPs) on ischemia conspicuity, Ghoshhajra et  al.
window of approximately 8–9 s, in which the difference in reported highest sensitivities in diastolic datasets using 8 mm
contrast enhancement between ischemic and normal myo- MPRs, while for systolic reconstructions, 3  mm MinIPs
cardial segments is optimal. Furthermore, the presented data showed highest specificity and accuracy [39].
support the feasibility of a bolus-triggering technique within One of the technical challenges associated with static
the ascending aorta and a delay of approx. 7 s after reaching CT perfusion imaging has traditionally been the higher
a threshold of 150 HU. susceptibility to beam-hardening artifacts created by the
Importantly, this relatively long time window of 8–9  s high attenuation within the left ventricle and particularly
implies the presence of an attenuation difference between within the descending aorta which lies directly adjacent to
ischemic and normal segments over several heartbeats. This a segment of the free wall of the left ventricle. It has been
demonstrates that even when selecting scan protocols that shown that by using dedicated beam-hardening correction
acquire data over two or three heartbeats, the evaluation of reconstruction algorithms, these artifacts can be reduced
myocardial perfusion should be feasible. significantly [40, 41].
It should be noted that in all three studies, the maximum Depending on the acquisition protocols of rest and stress
difference in attenuation was approximately 20–25 HU on a CT perfusion, the phases with least cardiac motion of the
background of approximately 100 HU, highlighting the rela- myocardium should be selected, with sharpest epi- and endo-
tively low contrast resolution or signal intensity that static cardial borders and least contour doubling [42].
CT perfusion techniques are associated with.

Image Analysis
 ptimal RR Phase for the Evaluation of Static
O
CT Myocardial Perfusion The images reconstructed from the perfusion datasets should
be evaluated in the three traditional cardiac axes, short axis,
Ghoshhajra et  al. analyzed the conspicuity of perfusion horizontal long axis, and vertical long axis, allowing for an
defects on systolic vs. diastolic reconstructions of retro- assessment of each myocardial segment in at least two
spectively acquired static CT stress perfusion datasets and planes. In general, a narrow window width and level
reported similar diagnostic accuracies of 61% for both car- (300/150) should be applied [42]. These window settings
diac phases with considerable higher sensitivity in end-­ maximize the differences between normal and ischemic
diastolic acquisitions and higher specificity in systolic myocardium and help to identify fat and vessels and calcifi-
acquisitions [39]. cations outside of this window width. For each patient, iden-
Using the data from the CORE320 study, Steveson et al. tical window settings should be applied for the analysis of
analyzed the influence of heart rate on image quality and rest and stress CT perfusion images.
optimal reconstruction phase for the evaluation of stress Unlike in research and clinical study environments, the
­perfusion within the confines of the diastolic acquisitions of routine clinical application of myocardial CT perfusion will
61  CT Myocardial Perfusion Imaging: Arterial First-Pass Imaging 783

certainly employ a combined interpretation of functional and with a large field of view on which they are t­ ypically revealed by
anatomic datasets. signs of contour doubling or haziness of anatomic edges. These
Coronary CTA is extensively validated as the most sensi- artifacts can mimic or mask perfusion deficits when they appear
tive noninvasive test for CAD, and due to its wide ­distribution, as focal hypo- and/or hyperdensities.
most imaging centers are well experienced in its perfor- Therefore, if the scan protocol selected for stress and rest
mance and interpretation. In general, findings should be perfusion permits distinct reconstructions across the RR
reported according to established guidelines [43]. interval, perfusion defects should be confirmed by their pres-
Once a potentially significant stenosis is identified on ence at several timepoints of the RR interval.
CCTA, the role of myocardial CT perfusion is (1) to unam- The only effective strategy against breathing artifacts
biguously prove or exclude its functional significance, (2) to remains their prevention by investing time in detailed instruc-
differentiate between reversible and fixed perfusion defects, tions to and training with the patient prior to performing the
and (3) to provide a quantitative parameter for the overall CT scan.
myocardial burden of ischemia [42].
Qualitative analysis of static CT perfusion is performed
by visual comparison of the images acquired during stress Beam Hardening
and rest conditions. Typically, standard cardiac long-axis and
short-axis planes are reconstructed as thick multiplanar This artifact occurs in parts of the myocardium lying adja-
reconstructions with 5–8  mm slice thickness for optimal cent to large hyperdense structures, in which low-energy
signal-to-noise ratio. These are then interpreted for the pres- photons are preferentially absorbed, resulting in a relative
ence and extent of perfusion defects [42]. increase in high-energy photons in the remaining photon
Several software tools are available to support the reader beam (i.e., beam hardening). Therefore, the x-ray photon
in the analysis of myocardial perfusion. Some software tools beam will have slightly different spectral properties during
derive secondary parameters from the myocardial tissue the sampling of these adjacent myocardial segments, which
attenuation values, such as the transmural perfusion ratio will cause differences in attenuation values attributed to the
(TPR) – this is a semiquantitative index parameter for each respective myocardial segments. This artifact is most
segment which relates subendocardial attenuation of a par- apparent in the basal inferior wall due to the contrast-filled
ticular segment with the mean epicardial attenuation of the descending aorta and the anterior wall of the left ventricle
entire heart. TPR has been initially validated for stress MRI which can be affected by the LV cavity or overlying ribs.
[44] but has been implemented to aid CT perfusion analysis Severe beam-hardening artifacts affecting the interventric-
[33, 45] and can best be visualized by bull’s eye plots. ular septum but also the midventricular and apical lateral
TPR is not reliable in the presence of prior infarcts or sig- wall of the left ventricle can also be observed in the pres-
nificant artifacts in several myocardial segments since both ence of prominent pacemaker electrodes in the right ven-
phenomena distort mean epicardial attenuation as the tricle (Fig. 61.4).
denominator for the calculation of TPR.  Therefore, the Special reconstruction algorithms have been developed to
radiologic interpretation should not rely on this parameter take beam hardening into account [40, 41]. In general, beam-­
alone but always be regarded as complementary to an initial hardening artifacts appear in the plane of the x-ray beam and
visual assessment of the perfusion images. can thus be identified by analyzing the distribution using
multiplanar reformations on the dataset during image inter-
pretation. Usually, beam hardening creates transmural
Typical Artifacts hypoattenuation and occurs directly adjacent to highly atten-
uating structures.
All CT perfusion scans should be evaluated for the presence
and extent of artifacts which should be included in the writ-
ten report. Common artifacts which can result in false-­ R  econstruction Artifacts
positive or false-negative evaluation of myocardial perfusion
are due to motion, beam-hardening, reconstruction, or partial The most relevant reconstruction artifact is the cone-beam
dataset misalignment artifacts. artifact which results from the fan configuration of the x-ray
beam along the z-axis. This becomes most relevant for mod-
ern wide-range detectors, in particular for the 320-detector
Motion CT with cone angles of up to 15.2° [46]. Cone-beam artifacts
typically present as low- and high-attenuation bands but usu-
Both the heartbeat and breathing are potential sources of motion ally extend beyond the range of the myocardium into adja-
artifacts. Both can be detected by analysis of ­reconstructions cent mediastinal and lung structures.
784 F. Schwarz et al.

a b c d

Fig. 61.4  A 62-year-old patient with a history of atypical chest pain entation, revealing subendocardial hypoattenuation in the anterolateral
underwent coronary CTA (a, b) followed by adenosine-stress CT perfu- myocardial segments, consistent with a reversible perfusion defect. The
sion (c, d). Coronary CTA demonstrated severe calcifications of the hypoattenuation of the interventricular septum in the stress image is due
LAD (a) whose hemodynamic significance remained uncertain. (b) to an artifact by the very prominent pacemaker electrode in the right
demonstrates four-chamber view of myocardial perfusion at rest. (c) ventricular apex – notice the hypodense bands originating at the tip of
shows an adenosine-stress myocardial perfusion dataset in the same ori- the electrode (d)

Partial Dataset Misalignment Artifacts  tress Image Interpretation

S
and the Assessment of Ischemia
As in coronary CTA, misalignment artifacts can occur when
the final image is composed of image stacks acquired during Stress CT perfusion images carefully need to be assessed for
different heartbeats, such as in step-and-shoot image acquisi- the presence of stress-induced subendocardial or transmural
tion protocols. More than when evaluating the coronary areas of hypoattenuation not present at rest, changes which
arteries, however, the differences in the exact acquisition are highly suggestive for (reversible) myocardial ischemia.
timepoint between adjacent image stacks can be problem- Furthermore, in large infarcts apparent on the rest images, a
atic, since they are acquired during different timepoints of relative enhancement may be appreciated on the stress
the contrast bolus and thus make comparisons of myocardial images if a rest-stress protocol is applied.
enhancement problematic.

Defect Severity and Ischemic Burden

Image Interpretation
Similarly to the method established in the nuclear medicine
 est Image Interpretation and the Assessment
R SPECT literature, the extent of segmental myocardial isch-
of Prior Infarcts emia can be expressed in a semiquantitative manner for each
myocardial segment [47], assigning a semiquantitative value
It is important to identify prior infarcts on rest perfusion expressing the severity of ischemia (0 = normal myocardial
images to be able to discern them from areas of ischemia perfusion; 1 = mild perfusion deficit, i.e., <1/3 transmurality;
during stress perfusion (Fig. 61.5). Typically, prior myocar- 2  =  moderate, i.e., <50% transmurality; 3  =  severe, i.e.,
dial infarcts are characterized by the presence of subendocar- >50% transmurality).
dial or transmural hypoattenuation during rest perfusion due The ischemic burden of the left ventricle can be qualita-
to the high proportion of scar tissue, by relative myocardial tively assessed by describing the myocardial segments
thinning and the myocardial inclusion of fat in chronic involved. Semiquantitative measures for global ischemic
infarcts. Older large infarcts often exhibit calcifications, burden can be generated by calculating the sum of all seg-
aneurysmal dilatation, or mural thrombus formation. mental ischemia values: the summed rest score (SRS),
61  CT Myocardial Perfusion Imaging: Arterial First-Pass Imaging 785

a b

c d

Fig. 61.5  A 55-year-old patient with a long-standing history of diabe- segment can be appreciated in diastolic and systolic reconstructions
tes and stable angina who underwent retrospectively ECG-gated coro- even under rest. Under adenosine stress, the hypoattenuated area is con-
nary CTA (a–d, a–b, diastolic reconstruction; c–d, systolic siderably larger (e, f), consistent with a partially reversible perfusion
reconstruction) followed by static adenosine-stress perfusion (e, f). defect and highly suggestive of on older non-transmural infarct with
Subendocardial hypoattenuation in the basal inferolateral myocardial significant adjacent ischemia
786 F. Schwarz et al.

e f

Fig. 61.5 (continued)

summed stress score (SSS), and the summed difference score d­ iagnostic accuracy and/or potentially incremental value of
(SDS). The SRS and the SSS are calculated by adding the static CT perfusion in diverse clinical scenarios.
ischemia scores for all segments at rest or stress, respec- Bettencourt et  al. have investigated 90 symptomatic
tively, and the SDS is the difference between both. The patients with suspected CAD who underwent static
­diagnostic and prognostic validity of this method still needs adenosine-­stress CT perfusion followed by CCTA and ICA
to be confirmed for CT perfusion. and report a significant increase in diagnostic accuracy for
the detection of stenoses ≥50% and ≥70% by the addition of
perfusion information [48]. Particularly for the detection of
 orrelation of Coronary Anatomy
C higher-grade CAD, the AUC increased from 0.8 for isolated
with Myocardial Perfusion Information CCTA to 0.93 for the integrated approach.
Wong et al. analyzed 75 symptomatic patients who under-
Ultimately, myocardial perfusion information will have to went CCTA/rest CT perfusion, static adenosine-stress CT
be matched to coronary anatomy both in respect to the indi- perfusion, and ICA/FFR assessment within 2  months at a
vidual vessel anatomy and stenoses identified on CCTA. single institution [49]. The authors compared the diagnostic
By  aligning anatomic with functional information, subtle performance for the prediction of FFR-relevant coronary
changes in myocardial perfusion will likely be detected with artery stenosis between CCTA only, CCTA + stress CT per-
higher sensitivity and perfusion defects without attributable fusion information, CCTA + transluminal attenuation gradi-
coronary stenosis easier identified as false-positive findings. ent information, and an integrated protocol combining all
three sources of information (i.e., CCTA + stress-rest CT
perfusion + transluminal attenuation gradient).
Evidence from Clinical Studies For CCTA only the authors reported sensitivities and
specificities of 89% and 65% for the prediction of FFR-­
By now substantial clinical evidence has been accumulated relevant coronary artery stenosis, while for CCTA + CT per-
demonstrating good diagnostic accuracy of static CT perfu- fusion, sensitivity and specificity were 88% and 83%.
sion for the detection of segmental myocardial ischemia, Diagnostic accuracy was highest for the integrated protocol
both from single-center and multicenter studies. (AUC = 0.91).
Yang et al. have recently reported results from a single-­
center study including 75 patients with suspected CAD
Single-Center Studies (ClinicalTrials.gov, NCT 01696006), in whom stress-rest
static CT perfusion using adenosine was compared with
Various single-center studies with often small sample sizes ICA  +  FFR [50]. The authors performed retrospectively
and heterogeneous reference standards have analyzed ECG-gated acquisitions for both the stress and rest CT
61  CT Myocardial Perfusion Imaging: Arterial First-Pass Imaging 787

p­ erfusion acquisition. The per-patient sensitivity and speci- followed by adenosine-stress CT perfusion imaging with
ficity of a visual assessment of myocardial perfusion CT data 20-min delay prior to stress perfusion acquisition. All
for all patients were 89% and 86%, respectively. In severely patients underwent rest-stress SPECT myocardial perfusion
calcified vessels, visual assessment of myocardial perfusion imaging as well as invasive coronary angiography within
CT data in combination with CCTA provided incremental 60 days.
value over CCTA alone for the detection of myocardial The authors report that the addition of myocardial perfu-
ischemia. sion to CCTA significantly increased the accuracy for the
detection of flow-limiting CAD as assessed by ICA-­
SPECT. This was demonstrated by a significant increase in
Meta-analyses the area under the curve (AUC) for the detection of ≥50%
stenosis from 0.82 to 0.87 by CCTA alone vs. CCTA-CTP
Some recent meta-analyses have reported pooled data from (p ≤ 0.001). The highest test performance was observed in
various single-center trials on myocardial perfusion the subgroup of patients without known CAD (AUC = 0.93).
CT.  Pelgrim et  al. pooled data from six studies (including The authors conclude that the combination of CTA and
one multicenter study) and reported for combined coronary perfusion correctly identifies patients with flow-limiting
CT angiography and static stress CT perfusion a vessel-­ CAD defined as ≥50% stenosis by ICA causing perfusion
based sensitivity and specificity for the detection of ICA-­ defect by SPECT/MPI.
based >50% stenoses of 84% and 93% [22]. Gonzalez et al. Furthermore, in head-to-head comparisons, the overall
pooled 18 studies in which CCTA was complemented with performance of stress CTP imaging in the diagnosis of ana-
stress CT perfusion or CT-FFR and report that CT perfusion tomic coronary stenosis ≥50% (with ICA as gold standard)
increased the specificity of CCTA from 0.43 to 0.77 for the was higher than that of SPECT, driven in part by the higher
detection of >50% stenoses as assessed by ICA [51]. It is sensitivity for left main and multivessel disease [53]. The
important to note, however, that Gonzalez et al. pooled data AUC for the detection of any coronary stenosis ≥50% was
regardless of whether a static or dynamic CT perfusion pro- 0.78 for rest-stress CT perfusion imaging and 0.69 for rest-­
tocol was used. stress SPECT imaging (p  =  0.001). In comparison with
SPECT, CT perfusion exhibited slightly lower specificity
(55% vs. 67%, p = 0.02) but considerably higher sensitivity
Multicenter Studies (88% vs. 62%, p < 0.001) [53, 54].
Radiation exposure of rest and stress CCTA was 3.16
CORE320 Study (www.clinicaltrials.gov, (2.82–3.63) and 5.31 (3.81–6.04) mSv, respectively, with
NCT00934037) higher doses required for the stress acquisition. This was
The CORE320 study (www.clinicaltrials.gov, NCT00934037) mostly due to the higher heart rates during adenosine stress
is the most comprehensive trial so far addressing the feasibil- (69 vs. 54  bpm). On the other hand, radiation exposure of
ity and accuracy of rest-stress static CT perfusion in symp- rest-stress SPECT and of ICA were 9.75 (9.1–13) mSv and
tomatic patients. 12.0 (7.6–18.0) mSv, respectively [55].
As a prospective, multicenter, multinational diagnostic
study, it included 381 patients and was designed for the main Regadenoson Crossover Study (www.
primary objective to evaluate the diagnostic accuracy of 320-­ clinicaltrials.gov, NCT01334918)
MDCT for detecting coronary artery luminal stenosis and Cury et  al. recently published the results of a multicenter,
corresponding myocardial perfusion deficits using a rest-­ multivendor, randomized, crossover clinical trial comparing
stress static CT perfusion approach in patients with suspected regadenoson stress-rest static CT perfusion with regadeno-
CAD. The reference standard was invasive coronary angiog- son stress-rest SPECT in 110 patients with known or sus-
raphy (ICA) and SPECT myocardial perfusion imaging [52]. pected CAD (moderate or high risk) and a clinical indication
The study population consisted of patients 45–85  years for stress-rest SPECT or CCTA [56].
old who were referred for clinically indicated conventional The vast majority of patients underwent a retrospectively
angiography for suspected or known CAD [52]. 30% of ECG-gated acquisition for stress CT perfusion after regade-
patients had had prior percutaneous intervention, and 28% noson administration followed by prospectively ECG-­
were status post coronary stent implantation. Prevalence of triggered rest CT perfusion/CCTA. Patients were randomized
anatomically significant CAD as assessed by quantitative into either having the stress-rest SPECT on the day before or
ICA was 60%. after stress-rest CT perfusion [57].
The CT protocol comprised an unenhanced coronary Due to the potentially higher adverse event rate of regad-
artery calcium score acquisition, followed by rest CCTA enoson upon initial administration (on the first day of the
(used for coronary CT angiography as well as rest perfusion) study), patients were randomized into either possible
after administration of beta-blocker and sublingual nitrates, sequence of scans (i.e., SPECT-CT vs. CT-SPECT).
788 F. Schwarz et al.

In this study, stress CT perfusion was noninferior to Conclusion

SPECT for detecting or excluding reversible ischemia with
an agreement rate of 0.87 and sensitivity and specificity of The high pace of technological innovations in cardiac CT has
0.9 and 0.84, respectively [56]. Total mean radiation dose extended its reach to the evaluation of ischemic cardiomy-
was significantly higher for stress-rest CT perfusion opathy, enabling pharmacological stress myocardial CT per-
(17.7  ±  6.8  mSv) compared with stress-rest SPECT fusion imaging. This extension promises to cover one of the
(11.2 ± 1.8 mSv, p = 0.001). traditionally weaker flanks of cardiac CT: the specificity and
positive predictive value in the detection of hemodynami-
Bischoff et al. 2016 cally significant stenoses. Furthermore, diagnostic accuracy
Results from a smaller multicenter study evaluating stress-­ in the assessment of stent patency and severely calcified
rest static CT perfusion in symptomatic patients with prior coronary trees may be increased.
revascularization in comparison with stress-rest SPECT and Future clinical studies will have to address the differences
ICA were recently reported by Bischoff et al. [25]. Thirty-six between static and dynamic acquisition protocols, as the for-
patients who presented with new onset symptoms requiring mer cannot provide fully quantitative data but come at con-
ICA were included and underwent a stress-rest static CT per- siderably lower radiation doses.
fusion protocol (with injection of the radionuclide for stress Regarding further technical progress, the best is yet to
SPECT MPI during stress CT perfusion imaging) immedi- come: further improvements in hardware will result in wider
ately followed by stress and rest SPECT acquisition and ICA detectors and higher temporal and spatial resolution.
within 1 week prior to or after the CT scan. Someday, photon-counting detectors with energy discrimina-
The stress CT protocol consisted of a prospectively ECG-­ tion capabilities might make their way into clinical practice.
triggered high-pitch spiral acquisition planned to scan the On the software side, much is to be expected in the short term
entire heart in the same systolic phase (commencing at 5% of from further refinements in reconstruction algorithms such as
the RR interval) during adenosine or after regadenoson injec- more advanced versions of iterative reconstruction and raw
tion. After a delay of 15–20 min, beta-blockers and ­sublingual data-based motion correction algorithms and in the medium
nitroglycerine were administered prior to a prospectively term from the full implementation of neural network-/deep
ECG-triggered CCTA with relatively broad acquisition of learning-based raw data postprocessing and data interpreta-
systole and diastole. tion. Furthermore, the full digitalization of the interface
The authors report that in this cohort of patients with between the image information and all other patient data in
known CAD and extensive vessel calcifications, the overall conjunction with the expected rapid implementation of artifi-
diagnostic accuracy for the detection of hemodynamically cial intelligence to a wide spectrum of clinical problems will
relevant coronary artery stenosis was only 31% for CCTA result in considerably improved and individually tailored
alone and could be increased to 78% when integrating the diagnostic and prognostic results for each and every patient.
information obtained from stress-rest CT perfusion. With the already proven excellent diagnostic accuracy of
Importantly, the high-pitch acquisition of the heart in systole stress CT perfusion for the detection of myocardial ischemia
had an effective dose of only 0.9 ± 0.1 mSv. and thus the hemodynamic relevance of coronary artery ste-
noses in conjunction with the unsurpassed capability of char-
acterizing coronary artery plaques among noninvasive
Clinical Studies Involving Protocol imaging modalities, the one-stop shop in cardiac imaging
Modifications that has been heralded so fervently and repeatedly now more
than ever seems to be just around the corner.
Nakamori et al. have recently suggested that in patients with
high calcium scores (>400 Agatston units) and/or prior
stents, a static CT stress perfusion-only acquisition might References
suffice [58]. In these patients, the authors argue, the rate of
nondiagnostic coronary artery segments will be not negligi- 1. Menke J, Kowalski J. Diagnostic accuracy and utility of coronary
ble for both the stress and the rest acquisition. Therefore, the CT angiography with consideration of unevaluable results: a sys-
tematic review and multivariate Bayesian random-effects meta-­
addition of an acquisition at rest yields comparably little analysis with intention to diagnose. Eur Radiol. 2016;26(2):451–8.
additional information. In 35 patients with high calcium 2. von Ballmoos MW, et al. Meta-analysis: diagnostic performance of
scores and/or stents who underwent stress-rest static CT per- low-radiation-dose coronary computed tomography angiography.
fusion as well as ICA/FFR, stress CT perfusion had an AUC Ann Intern Med. 2011;154(6):413–20.
3. Samad Z, et al. A meta-analysis and systematic review of computed
of 0.97 for the detection of hemodynamically significant ste- tomography angiography as a diagnostic triage tool for patients
noses which was not improved by the addition of the rest with chest pain presenting to the emergency department. J Nucl
perfusion/CCTA [58]. Cardiol. 2012;19(2):364–76.
61  CT Myocardial Perfusion Imaging: Arterial First-Pass Imaging 789

4. Hulten E, et  al. Outcomes after coronary computed tomography 25. Bischoff B, et  al. Myocardial ischemia detection with single-­

angiography in the emergency department: a systematic review and phase CT perfusion in symptomatic patients using high-pitch
meta-analysis of randomized, controlled trials. J Am Coll Cardiol. helical image acquisition technique. Int J Cardiovasc Imaging.
2013;61(8):880–92. 2017;33(4):569–76.
5. Di Carli MF, et al. Relationship between CT coronary angiography 26. Jin KN, et  al. Myocardial perfusion imaging with dual energy
and stress perfusion imaging in patients with suspected ischemic CT. Eur J Radiol. 2016;85(10):1914–21.
heart disease assessed by integrated PET-CT imaging. J Nuclear 27. De Cecco CN, et al. White paper of the Society of Computed Body
Cardiol. 2007;14(6):799–809. Tomography and Magnetic Resonance on dual-energy CT, part 3:
6. Pijls NHJ, et  al. Fractional flow reserve versus angiography for vascular, cardiac, pulmonary, and musculoskeletal applications. J
guiding percutaneous coronary intervention in patients with mul- Comput Assist Tomogr. 2017;41(1):1–7.
tivessel coronary artery disease: 2-year follow-up of the FAME 28. Meinel FG, et  al. First–arterial-pass dual-energy CT for assess-
(Fractional Flow Reserve Versus Angiography for Multivessel ment of myocardial blood supply: do we need rest, stress, and
Evaluation) study. J Am Coll Cardiol. 2010;56(3):177–84. delayed acquisition? Comparison with SPECT.  Radiology.
7. Shaw LJ, et al. Optimal medical therapy with or without percutane- 2014;270(3):708–16.
ous coronary intervention to reduce ischemic burden: results from 29. Arnoldi E, et  al. CT detection of myocardial blood volume defi-
the Clinical Outcomes Utilizing Revascularization and Aggressive cits: dual-energy CT compared with single-energy CT spectra. J
Drug Evaluation (COURAGE) trial nuclear substudy. Circulation. Cardiovasc Comput Tomogr. 2011;5(6):421–9.
2008;117(10):1283–91. 30. Koonce JD, et  al. Accuracy of dual-energy computed tomog-

8. Gould KL, Lipscomb K, Hamilton GW.  Physiologic basis for raphy for the measurement of iodine concentration using car-
assessing critical coronary stenosis. Instantaneous flow response diac CT protocols: validation in a phantom model. Eur Radiol.
and regional distribution during coronary hyperemia as mea- 2014;24(2):512–8.
sures of coronary flow reserve. Am J Cardiol. 1974;33(1): 31. Steveson C, et al. The effect of heart rate on exposure window and
87–94. best phase for stress perfusion computed tomography. J Comput
9. Uren NG, et  al. Relation between myocardial blood flow Assist Tomogr. 2017;41(2):242–8.
and the severity of coronary-artery stenosis. N Engl J Med. 32. Taillefer R, et al. Acute beta-blockade reduces the extent and sever-
1994;330(25):1782–8. ity of myocardial perfusion defects with dipyridamole Tc-99m ses-
10. Adams D, et al. Computed tomography of the normal and infarcted tamibi SPECT imaging. J Am Coll Cardiol. 2003;42(8):1475–83.
myocardium. Am J Roentgenol. 1976;126(4):786–91. 33. George RT, et al. Adenosine stress 64- and 256-row detector com-
11. Powell WJ, et al. Detection of edema associated with myocardial puted tomography angiography and perfusion imaging: a pilot
ischemia by computerized tomography in isolated, arrested canine study evaluating the transmural extent of perfusion abnormali-
hearts. Circulation. 1977;55(1):99–108. ties to predict atherosclerosis causing myocardial ischemia. Circ
12. Siemers PT, et al. Detection, quantitation and contrast enhancement Cardiovasc Imaging. 2009;2(3):174–82.
of myocardial infarction utilizing computerized axial tomography: 34. Techasith T, Cury RC.  Stress myocardial CT perfusion: an

comparison with histochemical staining and 99mTc-pyrophosphate update and future perspective. JACC Cardiovasc Imaging.
imaging. Investig Radiol. 1978;13(2):103–9. 2011;4(8):905–16.
13. Wolfkiel CJ, et al. Measurement of myocardial blood flow by ultra- 35. Seitun S, et  al. Stress computed tomography myocardial perfu-
fast computed tomography. Circulation. 1987;76(6):1262–73. sion imaging: a new topic in cardiology. Rev Esp Cardiol (English
14. Rumberger JA, Feiring AJ, et al. Use of ultrafast computed tomog- Edition). 2016;69(2):188–200.
raphy to quantitate regional myocardial perfusion: a preliminary 36. Bischoff B, et al. Optimal timing for first-pass stress CT myocardial
report. J Am Coll Cardiol. 1987;9(1):59–69. perfusion imaging. Int J Cardiovasc Imaging. 2013;29(2):435–42.
15. Rumberger JA, Stanford W, Marcus ML. Quantitation of regional 37. Tanabe Y, et  al. Optimal scan time for single-phase myocardial
myocardial perfusion by ultrafast CT: promises and pitfalls. Am J computed tomography perfusion to detect myocardial ischemia–
Card Imaging. 1987;1(4):336–43. derivation cohort from dynamic myocardial computed tomography
16. Gould RG. Perfusion quantitation by ultrafast computed tomogra- perfusion. Circ J. 2016;80(12):2506–12.
phy. Investig Radiol. 1992;27(Suppl 2):S18–21. 38. Pelgrim GJ, et  al. Optimal timing of image acquisition for arte-
17. Bell MR, Lerman LO, Rumberger JA.  Validation of minimally rial first pass CT myocardial perfusion imaging. Eur J Radiol.
invasive measurement of myocardial perfusion using electron beam 2017;86:227–33.
computed tomography and application in human volunteers. Heart. 39. Ghoshhajra BB, et al. A comparison of reconstruction and viewing
1999;81(6):628–35. parameters on image quality and accuracy of stress myocardial CT
18. Salerno M, Beller GA. Noninvasive assessment of myocardial per- perfusion. J Cardiovasc Comput Tomogr. 2011;5(6):459–66.
fusion. Circ Cardiovasc Imaging. 2009;2(5):412–24. 40. Kitagawa K, et  al. Characterization and correction of beam-­

19. Dawson P. Cardiovascular effects of contrast agents. Am J Cardiol. hardening artifacts during dynamic volume CT assessment of myo-
1989;64(9):2E–9E. cardial perfusion. Radiology. 2010;256(1):111–8.
20. Canty JM, et al. First-pass entry of nonionic contrast agent into the 41. Bucher AM, et al. Quantitative evaluation of beam-hardening arte-
myocardial extravascular space. Effects on radiographic estimates fact correction in dual-energy CT myocardial perfusion imaging.
of transit time and blood volume. Circulation. 1991;84(5):2071–8. Eur Radiol. 2016;26(9):3215–22.
21. Bae KT.  Intravenous contrast medium administration and scan
42. Mehra VC, Valdiviezo C, et al. A stepwise approach to the visual
timing at CT: considerations and approaches 1. Radiology. interpretation of CT-based myocardial perfusion. J Cardiovasc
2010;256(1):32–61. Comput Tomogr. 2011;5(6):357–69.
22. Pelgrim GJ, et al. The dream of a one-stop-shop: meta-analysis on 43. Raff GL, et al. SCCT guidelines for the interpretation and report-
myocardial perfusion CT. Eur J Radiol. 2015;84(12):2411–20. ing of coronary computed tomographic angiography. J Cardiovasc
23. Caruso D, et al. Dynamic CT myocardial perfusion imaging. Eur J Comput Tomogr. 2009;3(2):122–36.
Radiol. 2016;85(10):1893–9. 44. Keijer JT, et al. Magnetic resonance imaging of regional myocar-
24. Cannaò PM, et  al. Technical prerequisites and imaging protocols dial perfusion in patients with single-vessel coronary artery disease:
for dynamic and dual energy myocardial perfusion imaging. Eur J quantitative comparison with (201)Thallium-SPECT and coronary
Radiol. 2015;84(12):2401–10. angiography. J Magn Reson Imaging. 2000;11(6):607–15.
790 F. Schwarz et al.

45. Mehra VC, Ambrose M, et  al. CT-based myocardial perfusion

52. George RT, et  al. Diagnostic performance of combined noninva-
imaging-­ practical considerations: acquisition, image analy- sive coronary angiography and myocardial perfusion imaging using
sis, interpretation, and challenges. J Cardiovasc Transl Res. 320-MDCT: the CT angiography and perfusion methods of the
2011;4(4):437–48. CORE320 multicenter multinational diagnostic study. AJR Am J
46. Kitagawa K, et al. Prospective ECG-gated 320 row detector com- Roentgenol. 2011;197(4):829–37.
puted tomography: implications for CT angiography and perfusion 53. George RT, et al. Myocardial CT perfusion imaging and SPECT for
imaging. Int J Cardiovasc Imaging. 2009;25(2):201. the diagnosis of coronary artery disease: a head-to-head compari-
47. Palmas W, et  al. Incremental prognostic value of exercise thal- son from the CORE320 multicenter diagnostic performance study.
lium-­201 myocardial single-photon emission computed tomogra- Radiology. 2014;272(2):407–16.
phy late after coronary artery bypass surgery. J Am Coll Cardiol. 54. Magalhães TA, et  al. Combined coronary angiography and myo-
1995;25(2):403–9. cardial perfusion by computed tomography in the identification
48. Bettencourt N, et  al. Incremental value of an integrated adenos- of flow-limiting stenosis  – the CORE320 study: an integrated
ine stress-rest MDCT perfusion protocol for detection of obstruc- analysis of CT coronary angiography and myocardial perfusion.
tive coronary artery disease. J Cardiovasc Comput Tomogr. J Cardiovasc Comput Tomogr. 2015;9(5):438–45.
2011;5(6):392–405. 55. Rochitte CE, et al. Computed tomography angiography and perfu-
49. Wong DTL, et  al. Comparison of diagnostic accuracy of com- sion to assess coronary artery stenosis causing perfusion defects
bined assessment using adenosine stress computed tomography by single photon emission computed tomography: the CORE320
perfusion + computed tomography angiography with transluminal study. Eur Heart J. 2014;35(17):1120–30.
attenuation gradient + computed tomography angiography against 56. Cury RC, et  al. A randomized, multicenter, multivendor study of
invasive fractional flow reserve. J Am Coll Cardiol. 2014;63(18): myocardial perfusion imaging with regadenoson CT perfusion
1904–12. vs single photon emission CT.  J Cardiovasc Comput Tomogr.
50. Yang DH, et  al. Stress myocardial perfusion CT in patients sus- 2015;9(2):103–12–2.
pected of having coronary artery disease: visual and quantitative 57. Cury RC, et al. Regadenoson-stress myocardial CT perfusion and
analysis-validation by using fractional flow reserve. Radiology. single-photon emission CT: rationale, design, and acquisition
2015;276(3):715–23. methods of a prospective, multicenter, multivendor comparison.
51. Gonzalez JA, et  al. Meta-analysis of diagnostic performance of J Cardiovasc Comput Tomogr. 2014;8(1):2–12.
coronary computed tomography angiography, computed tomogra- 58. Nakamori S, et al. Feasibility of stress-alone cardiac CT for detect-
phy perfusion, and computed tomography-fractional flow reserve ing hemodynamically significant coronary stenosis in the presence
in functional myocardial ischemia assessment versus invasive frac- of high coronary calcium score and coronary stents. Cardiovasc
tional flow reserve. Am J Cardiol. 2015;116(9):1469–78. Imaging Asia. 2017;1(1):38–48.
 Myocardial Perfusion Imaging:
Dual-­Energy Approaches 62
Domenico De Santis, Marwen Eid, Taylor M. Duguay,
and Carlo N. De Cecco

The evaluation of patients presenting with symptoms sug- of the noninvasive diagnostic workup for the anatomic eval-
gestive of myocardial ischemia is one of the most common uation of the coronary arteries in patients with suspected
and challenging scenarios clinicians face. Despite consider- CAD ­[5–8]. A growing body of evidence has validated
able advances in treatment, more than 50% of acute myocar- CCTA as the noninvasive imaging technique with the high-
dial infarctions (AMI) resulting in death occur in patients est sensitivity and specificity in detecting CAD, with a
before undergoing cardiac catheterization. Thus, risk stratifi- pooled sensitivity and specificity of 98% and 89%, respec-
cation plays a central role in averting major adverse cardiac tively [6, 9]. These results compare favorably with alterna-
events [1]. tive noninvasive imaging tests, where SPECT reaches
The current WHO rating attributes more than 25% of deaths sensitivities and specificities of 88% and 61%, PET of 84%
worldwide to cardiovascular disease (CVD) [2]. Despite a and 81%, and cardiac ­magnetic resonance imaging (CMR)
decreasing trend in the last decade, CVD is the leading cause of of 89% and 76%, respectively [10].
death in the United States and worldwide. On average there is Although CCTA remains a morphological technique that
approximately one CVD-related death every 40 s, resulting in can accurately depict coronary anatomy and atherosclerotic
the death of over 2000 Americans each day. The estimated plaque burden, it is hampered by several limitations in the
direct and indirect cost of CVD in 2015 was $320.1 billion and assessment of the hemodynamic significant coronary stenosis.
is projected to be $918 billion by 2030 [3, 4]. The FAME and COURAGE trials [11, 12], two major studies
According to the current appropriate use criteria, coro- validating the impact of functional tests in coronary revascu-
nary CT angiography (CCTA) is a robust imaging technique larization, have shown that the hemodynamic relevance of
that provides a noninvasive, morphological assessment of coronary stenosis is not adequately predicted by purely ana-
the coronary arteries which can accurately depict coronary tomical tests. Additionally, without functional data, ICA and
anatomy and atherosclerotic plaque burden. Thanks to its CCTA can only provide limited correlation with myocardial
power to exclude significant coronary artery stenosis in perfusion defects [13, 14]. As revascularization should be
patients with low and intermediate coronary artery disease guided by information on the state of myocardial perfusion,
(CAD) risk profiles [5], CCTA has become an integral part increasing efforts aim at determining the functional relevance
of lesions by CCTA. Thus, noninvasive evaluation of patients
with suspected CAD has started to shift focus from morpho-
D. De Santis
logical CAD assessment to a complex, comprehensive mor-
Division of Cardiovascular Imaging, Department of Radiology and
Radiological Science, Medical University of South Carolina, phological and functional evaluation. Furthermore, patient
Charleston, SC, USA evaluation, management, and prognostication are more reli-
Department of Radiological Sciences, Oncological and able and effective when morphological and functional assess-
Pathological Sciences, University of Rome “Sapienza”, ments are used in concert [11, 12, 14, 15].
Latina, Italy Multiple CT techniques have the potential to provide a
T. M. Duguay (*) functional analysis. Some of these techniques are based on
Division of Cardiovascular Imaging, Department of Radiology and post-processing analysis of CCTA dataset and are focused on
Radiological Science, Medical University of South Carolina,
the direct assessment of coronary stenosis significance, such
Charleston, SC, USA
e-mail: [email protected] as CCTA-derived fractional flow reserve (CT-FFR) and
transluminal attenuation gradient (TAG). CT-FFR relies on
M. Eid ∙ C. N. De Cecco
Department of Radiology and Radiological Science, principles of computational fluid dynamics to calculate the
Medical University of South Carolina, Charleston, SC, USA ratio between the maximum coronary flow in the presence of

© Humana Press 2019 791

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_62
792 D. De Santis et al.

a coronary stenosis and the hypothetical maximum coronary uses two simultaneous acquisitions of the same body region
flow in absence of stenosis. Despite excellent results in terms at two different kV values to generate a similar snapshot of
of diagnostic accuracy, the only CT-FFR software that has the myocardial iodine distribution [22]. Since iodine concen-
been granted FDA approval to date requires complex offsite tration reflects myocardial perfusion, the per-voxel amount
analysis [16]. TAG represents the contrast attenuation gradi- of iodine can be used to quantify the myocardial blood pool
ent along the course of a coronary artery. The reliability of [23]. Iodine concentration is reduced in the myocardial areas
this technique is often hampered by extensive coronary cal- with hypoperfusion and/or reduced intravascular blood vol-
cifications or temporal inhomogeneity due to the acquisition ume [24]. This chapter will provide an overview of dual-­
window covering multiple heartbeats [17]. The correlation energy CTPMI.
between coronary density and the corresponding aortic
attenuation at the same axial slice, formally known as CCO
(corrected coronary opacification), has been proposed as a  ual-Energy CT Myocardial Perfusion
D
method to achieve more robust results. However, TAG and Imaging
CCO have inferior diagnostic performance when compared
to other functional tests [18]. Technical Considerations
Other techniques based on CT data are focused on direct
assessment of myocardial ischemia. Due to recent advance- Computed tomography is based on the emission of a beam
ments in CT technology, in fact, in addition to its role in of x-rays from the x-ray tube. Two main determinants are
assessing coronary morphology and left ventricular function, responsible for the x-ray beam characteristics: intensity
CCTA has been utilized in the evaluation of a third aspect in and energy. The intensity of the x-rays represents the
the diagnostic algorithm of ischemic heart disease – myocar- amount of photons produced and is related to the tube cur-
dial perfusion. Computed tomography myocardial perfusion rent (milliamperage, mA). The energy of the x-ray beam is
imaging (CTMPI) offers the possibility to directly detect the proportional to the difference of potential applied between
presence of perfusion defects in the myocardium following the anode and the cathode of the x-ray tube and is mea-
the administration of pharmacological stressing agent. sured in kilovolts (kV). When a certain kV value is
Providing diagnostic information for each of these three cor- selected, the chosen number represents the maximum kV
nerstones of ischemic heart disease workup, this emerging that will be reached by the x-ray tube. Therefore, the maxi-
technology has the potential to become the stand-alone mum energy of the photons is determined by the peak kV;
method for the evaluation of patients with suspected CAD however the generated x-ray beam can be represented as a
using a single imaging modality and within a single imaging broad spectrum characterized by different photon energies
session [19]. (Fig. 62.1).
Dual-energy CT, also called spectral CT, by means of dif-
ferent technical solutions obtains two x-ray spectra. Different
CTMPI: General Overview materials, such as iodine, calcium, and soft tissues, are char-
acterized by different attenuation profiles at different energy
CT myocardial perfusion imaging (CTMPI) uses the distri- levels (Fig.  62.2). Based on this tenant, dual-energy CT is
bution of iodinated contrast media (CM) in the myocardium able to differentiate materials with different atomic numbers,
as a surrogate for myocardial blood flow. Perfusion defects such as calcium and iodine, that look similar when scanned
are thus identified as hypo-attenuating areas containing in single energy. Different vendor-specific CT technologies
reduced amounts of CM [20]. For this purpose, two different have been developed to perform DE acquisitions for the eval-
CT techniques have been developed. Static CTMPI uses the uation of myocardial perfusion:
static myocardial distribution of CM during the early arte-
rial phase of the first-pass contrast enhancement to detect • Single source  – Sequential DECT (Canon Medical
myocardial attenuation abnormalities [21]. In contrast, Systems, Otawara, Japan)
Dynamic CTMPI uses several consecutive acquisitions • Single source  – Fast kV switching (GE Healthcare,
throughout the cardiac cycle to generate time attenuation Milwaukee, WI)
curves (TAC) of myocardial perfusion. Static CTMPI can be • Dual-source CT (Siemens Healthineers, Forchheim,
further categorized into single-energy and dual-energy Germany)
approaches. • Single source  – Dual-layer detector (Philips Medical
The single-energy technique provides a qualitative visual Systems, Cleveland, OH)
assessment of a snapshot of myocardial iodine contrast • Single source  – Twin beam (Siemens Healthineers,
attenuation at a single time point. Dual-energy CT (DECT) Forchheim, Germany)
62  Myocardial Perfusion Imaging: Dual-Energy Approaches 793

Fig. 62.1  Standard 120 kVp x 104

x-ray spectrum. The peak kV 10
determines the maximum 120 kVp
photon energy. However, the 9
x-ray beam can be represented
as a broad spectrum 8
characterized by different
photon energies. The 7
spectrum is formed by the
bremsstrahlung radiation 6

Photon output
(German word meaning
“braking radiation”) and two 5
spikes representing the
characteristic radiations of 4
tungsten, the main material
forming the anode target. The 3
lower-energy x-rays don’t
contribute to the image, but 2
they are responsible for part
of the radiation dose since 1
they are absorbed by the
tissues 0
0 20 40 60 80 100 120
Photon energy (keV)

x 104
18
70 kVp
16 80 kVp
90 kVp
100 kVp
14
110 kVp
120 kVp
12 130 kVp
140 kVp
Photon output

150 kVp
10

0
0 25 50 75 100 125 150
Photon energy (keV)

Fig. 62.2  Range of x-ray spectra that can be generated in dual-energy radiation moving to the right of the graph (better x-ray quality) and
CT. The data of a low-energy spectrum and those derived from a high-­ higher peak for characteristic radiations (higher quantity of x-rays). It
energy spectrum are simultaneously analyzed to derive dual-energy is worth noting that changes in kVp have no effect on the energies of the
datasets and material-specific information. The higher the kVp, the characteristic radiation and that the characteristic x-rays do not occur at
higher the number of output photons, reflecting in a bremsstrahlung kVp lower than 70 keV

The sequential DECT acquires DE data through ultrafast is moving, resulting in a pitch characterized by partial over-
sequential rotations acquired in volume or helical mode. In lap among the two kV levels [25] (Fig. 62.3a). The fast kV
volume mode, two single-rotation acquisitions at different switching technique relies on a single x-ray tube able to
kV levels are performed. In helical mode, the tube switches switch between high- and low-kV settings within the same
between the two kV levels at each rotation while the table gantry rotation coupled with detectors able to register data
794 D. De Santis et al.

a b c d e

Fig. 62.3  Different DECT technologies: single source, sequential DECT (a); single source, fast kV switching (b); dual-source CT (c); single
source, dual-layer detector (d); and single source, twin beam (e)

from both energy spectra (Fig.  62.3b) [26]. Dual-source low photon energies, approaching the iodine K-edge,
dual-energy CT scanners have two different x-ray tubes reduce the beam-­hardening artifacts and enhance iodine
mounted with an offset of approximately 90 degrees and attenuation, facilitating the identification of subtle attenua-
are simultaneously operated at low- and high-tube voltages. tion differences within the myocardium when compared to
Each tube is coupled with a dedicated detector layer 120 kVp dataset [35]. Figure 62.4 provides an overview on
(Fig.  62.3c). A third technical approach is represented by dual-energy workflow.
the dual-layer detector CT scanner (Fig. 62.3d) character-
ized by a single x-ray tube coupled with a detector com-
posed of two superimposed layers. The inner layer registers Acquisition Technique and Radiation Dose
data from low-energy photons, while the outer layer records
information from the high-energy photons [27]. Twin beam As in CCTA, temporal resolution is one of the most important
dual energy uses a single x-ray tube emitting a polychro- technical prerequisites for perfusion CT.  A high temporal
matic x-ray beam pre-filtered on the z-axis by means of two resolution is critical to reduce motion artifacts in consider-
different materials: tin (Sn) and gold (Au). The resultant ation of the increased heart rate observed during the stress
post-­filtered beam is split in high (Sn)- and low-energy phase. However, previous challenges in temporal resolution
(Au) spectra before reaching the patient (Fig. 62.3e). The have been overcome with second-generation (75 ms) [28] or
dual-­source approach is the most widely adopted technique third-generation (66  ms) [36] dual-source CT scanners.
[24, ­28–34]. To date, the state of the art of dual-source CT Similarly to the SE static technique, DE-CTMPI generates a
is represented by the third generation of CT scanners, snapshot of the myocardial iodine distribution at a single time
which are characterized by a rotation time of 0.25 s, a tem- point using two different kV acquisitions of the same body
poral resolution of 66 ms, and maximal volume coverage of region; this snapshot encompasses the entire left ventricle and
737 mm/s in flash mode and are able to offer multiple kV is acquired during the early arterial phase of first-pass con-
setting ranging from 70  kV up to 150  kV, in 10  kV trast enhancement. A reduction in beam-­hardening artifacts
increments. and direct visualization of myocardial iodine content gives
Dual-energy CT exams are composed of different datas- DECT multiple advantages over a single-­energy acquisition.
ets. Low-voltage and high-voltage spectra are combined For these reasons, perfusion defects and late enhancement are
with different blending proportion to obtain a dataset simi- often more easily recognized on DECT [37]. Both prospec-
lar to the conventional 120  kVp dataset that is routinely tive and retrospectively ECG-gated protocols are available for
used in clinical practice. Other unique characteristics of the DECT acquisition. In addition, hybrid image reconstruc-
dual-energy CT examination include the possibility to tion enables imaging even at high heart rates [28]. Scan time
obtain color-coded iodine maps, virtual non-contrast datas- and radiation dose for prospective ECG-triggered or retro-
ets, and virtual monoenergetic images. Iodine maps are spectively ECG-triggered acquisition are equivalent to those
overlaid to grayscale images, allowing both qualitative and of single-energy cardiac CT acquisition [38].
quantitative myocardial perfusion assessment. The virtual Since image acquisition must be performed at the moment
non-contrast dataset has the ability to overcome the need of peak contrast concentration in the coronary arteries to
for a true non-­enhanced scan and, thus, to lower the radia- detect early differences in contrast uptake, acquisition timing
tion exposure. Eventually, virtual monoenergetic images at is crucial to preserve diagnostic integrity. The optimal time
62  Myocardial Perfusion Imaging: Dual-Energy Approaches 795

Fig. 62.4  Dual-energy CT

workflow. High-kV and
low-kV data are blended in
order to simulate a
conventional polychromatic
dataset. Dual-energy CT
allows to generate colored
iodine maps that are
superimposed on the
grayscale images and lead to
a qualitative and qualitative
assessment of myocardial
perfusion. Virtual non-­
contrast (VNC) images can
overcome the need on the true
unenhanced scan, while
virtual monoenergetic images
(VMI) at low keV level,
approaching the k-edge of
iodine, emphasize the contrast
enhancement

frame for stress perfusion acquisition is between 8 s and 16 s (Fig. 62.5). Starting the examination with the rest phase has
after contrast enhancement in the aorta reaches 100 the advantage that a pure CCTA study is performed before
Hounsfield Units [21]. Acquisition timing may be further the evaluation of functional myocardial assessment [38].
enhanced using bolus tracking with automatic contrast detec- On the other hand, starting the protocol with the stress
tion in the proximal ascending aorta. The right acquisition phase avoids contamination of CM from the rest phase and
time is essential to detect the difference in early contrast enables maximal contrast difference between ischemic and
uptake in myocardial tissue to ensure diagnostic accuracy. normal myocardium [21]. However, it should be taken into
Furthermore, coronary artery assessment can be performed account that the elevated heart rate following stress admin-
during the rest acquisition [37], with no additional acquisi- istration could impair the quality of the subsequent rest
tion time or radiation necessary. acquisition [32].
If an assessment of myocardial perfusion at stress is Additionally, image acquisition can be repeated after
desired, an additional acquisition during pharmacological 8–10  min of CM administration in order to evaluate the
stress can be obtained. Multiple approaches to the time delayed myocardial enhancement, where a hyperattenuating
order of rest and stress acquisitions have been proposed myocardial pattern is indicative of nonviable myocardium
796 D. De Santis et al.

a Contrast media Contrast media

injection injection

Calcium ~5 Rest ~5 Stress 8 - 10 Delayed

Score minutes Phase minutes Phase minutes Phase

IV access B-blockers Stress agent

ECG-gating Nitrates administration

Contrast media Contrast media

b
injection injection

Calcium ~5 Stress ~5 Rest 8 - 10 Delayed

Score minutes Phase minutes Phase minutes Phase

IV access Stress agent B-blockers

ECG-gating administration Nitrates

Fig. 62.5  Timeline of rest/stress (a) and stress/rest (b) DE-CTMPI protocols

(Fig.  62.6). Acquisitions utilizing this delayed enhance- The radiation dose for comprehensive static CT perfusion
ment, performed after invasive revascularization by percuta- studies (including rest and stress perfusion imaging) varies
neous coronary intervention (PCI) procedures without greatly depending on the specific acquisition protocol.
additional administration of CM, have been proven to be a Effective dose values between 4.2 and 16.5 mSv have been
reliable method for evaluating transmurality and infarct reported (Table 62.1) with an average radiation exposure of
size, representing an interesting alternative to CMR [39– 9 mSv [43–55]. The radiation dose of CT myocardial perfu-
41]. However, compared with late enhancement in CMR sion imaging generally compares favorably with traditional
however, CT provides a lower signal-to-noise ratio, making nuclear imaging approaches [56–58] and is expected to
the detection of infarction more challenging [23]. decrease as new technologies are developed. Currently, the
Additionally, Meinel et al. [42] recently demonstrated that radiation dose of a comprehensive stress/rest static CTMPI
the addition of a delayed phase acquisition for late iodine protocol can be as low as 4.2 mSv [59, 60]. In terms of radia-
enhancement does not p­ rovide improvements in diagnostic tion dose, moreover, prospectively ECG-triggered or retro-
accuracy and can thus be safely omitted in order to reduce spectively ECG-triggered acquisitions are equivalent to
the radiation dose. those of CCTA acquisition, in which the rest and the stress
Data sets obtained at high and low kV values in DECT acquisitions deliver a combined average dose of 12  mSv.
can be used to create iodine concentration maps combining This value may be decreased using recent technological
the lower noise of high-energy acquisitions with the higher advancements, including iterative reconstruction and ­low-­kV
contrast resolution of lower-energy acquisitions [33]. acquisition [49, 61, 62].
62  Myocardial Perfusion Imaging: Dual-Energy Approaches 797

Fig. 62.6  A 63-year-old

female. (a) CTMPI acquired a b
10 min administration of
contrast media demonstrates a
transmural area of
hyperattenuating myocardial
pattern in the lateral wall and
is indicative of nonviable
myocardium (arrowhead).
The same patient underwent a
CMRI (b) showing an
extensive area of late
enhancement in the
corresponding myocardial
wall (arrow). Curved
multi-planar reformats of the
coronary arteries show no
significant stenosis at the c d e f
RCA (c), LAD (d), and LCx
(e). On the contrary, the first
obtuse marginal branch (f)
previously stented is affected
by proximal intimal
hyperplasia resulting in severe
in-stent stenosis (arrowhead).
(g) Cinematic Rendering
clearly depicts the stent
(asterisk) and the filiform
caudal segment (arrow),
poorly opacified by the
contrast media

g
 798

Table 62.1  Static dual-energy CT myocardial perfusion studies

Patient AverageCT dose Reference Sensitivity Specificity PPV NPV
Author population CT technology CT perfusionprotocol (mSv) technique Level of analysis (%) (%) (%) (%)
Ruzsics (2009) [46] 36 First DSCT Rest 14 SPECT Segment 92 93 83 97
Patient 97 67 93 80
Bauer (2010) [92] 36 First DSCT Rest 9.7 CMR Segment 77 97 85 96
Nagao (2010) [93] 10 First DSCT Stress NR SPECT, ICA Vessel 86 75 NR NR
Nance (2011) [41] 12 First DSCT Rest NR ICA Segment 91 95 82 98
Ko (2011) [44] 41 First DSCT Stress 8.6 CMR Segment 89 78 74 91
Wang (2011) [63] 31 First DSCT Rest 10.5 SPECT Segment 68 93 82 86
Vessel 81 92 89 85
Meyer (2012) [45] 50 Second DSCT Rest/stress/delayed 13.4 CMR Patient 90 71 NR NR
Ko (2012) [69] 45 First DSCT Rest/stress 16.5 ICA Vessel 89 74 80 85
Weininger (2012) [70] 20 Second DSCT Stress 12.8 SPECT Segment 94 98 92 94
CMR Segment 93 99 92 96
Delgado (2013) [71] 56 Second DSCT Stress 8.2 CMR Segment 76 99 89 98
Ko (2014) [59] 40 First DSCT Rest 4.2 CMR, ICA Vessela 42 83 59 70
Stress 4.6 Vesselb 87 79 71 91
Ko (2014) [60] 100 First DSCT Stress 4.2 CMR Segment 76 80 63 88
Vessel 89 74 73 90
Patient 97 36 82 82
Kim (2014) [68] 50 Second DSCT Stress/rest 11.4 CMR Segment 77 94 53 98
Patient 94 71 60 96
Zhao (2014) [94] 60 Second DSCT Rest NR ICA Segment 94 91 88 96
Kido (2014) [95] 21 First and second Stress 7.7 ICA Vessel 67 92 84 82
DSCT
D. De Santis et al.
De Cecco (2014) [75] 29 Second DSCT Rest 5.8 ICA Patient – 95c 50c NR NR
Stress 6.6 morphological 95d 50d
analysis 100e 33e
90f 67f
SPECT Patient – 91c 38c
hemodynamic 95d 75d
analysis 100e 38e
86f 75f
Delgado (2016) [72] 36 Second DSCT Stress 5.4 CMR Segment 72.6g 95.7g 73.6g 95.5g
72.6h 94.7h 68.8h 95.5h
Vessel 87.5g 86.2g 72.4g 94.5g
91.6h 85.9h 73.5h 96h
Patient 100g 90g 92g 100g
100h 91h 32h 100h
Abbreviations: PPV positive predictive value, NPV negative predictive value, CMR cardiovascular magnetic resonance, DSCT dual-source computed tomography, ICA invasive coronary angiogra-
phy, SPECT single-photon emission computed tomography, NR not reported
a
Combined CT angiography and rest CT perfusion
b
Combined CT angiography and stress CT perfusion
c
Rest
d
Stress
e
62  Myocardial Perfusion Imaging: Dual-Energy Approaches

Combined – either positive

f
Combined – both positive
g
Reader 1
h
Reader 2
799
800 D. De Santis et al.

Pharmacological Stress Agents Dipyridamole and dobutamine can also be used as stress
agents. Dipyridamole acts as an indirect vasodilator, increas-
Diseased coronary arteries have a limited compensatory ing the endogenous level of adenosine by blocking cellular
­dilation, since they are already operating within a dilatory uptake. It is also administered through continuous infusion
reserve capacity to compensate for reduced perfusion. On the with a dose 140 μg/kg/min for 4–6 min. Since its half-life is
contrary, healthy coronary vessels maintain a dilating capac- longer than adenosine, aminophylline, an adenosine receptor
ity when oxygen demand is increased [25–27]. Hypo-­ antagonist, is required to reverse its effects.
attenuated myocardium during pharmacological stress is Dobutamine is a synthetic catecholamine characterized
classified by comparison with images in rest conditions. by strong β1-receptor and mild α1- and β2-receptor agonist
Areas with constant hypo-attenuation during both phases cor- activity. When used in low doses, marked inotropic effects
respond to irreversible ischemia. Reversible ischemia can be are encountered, mediated by both α1- and β1-receptor stim-
suspected if hypo-attenuation is present only in stress phase ulation. These effects are used for treatment of heart failure
or if the hypo-attenuated area increases significantly [1]. and the identification of dysfunctional, but viable, cardiac
Since the reversible myocardial perfusion defects (MPD) muscle. When used in high doses, the heart rate is progres-
occur prior to fixed MPD, sensitivity for the detection of sively increased (mediated by β1-receptor stimulation).
reversible perfusion defects significantly improves under Despite a clear increase in cardiac output, systemic blood
pharmacological stress [26]. The various agents available to pressure usually only increases minimally due to a decrease
induce pharmacological stress include dobutamine, dipyri- in systemic vascular resistance because the peripheral vaso-
damole, adenosine, and regadenoson [28]. Due to safety con- constrictive effects (mediated by α1-receptor stimulation)
siderations, the two most commonly utilized agents in CT are overwhelmed by the vasodilative effects (mediated by
perfusion are adenosine and regadenoson. β2-receptor stimulation) [37]. As a result of the hemody-
The first adenosine stress myocardial perfusion CT study namic changes, there is an increase in oxygen demand result-
was conducted with an electron-beam CT scanner in an ing in a secondary dilation of the coronary arteries and, thus,
experimental dog model in 1987 [29]. Adenosine is a nonse- an increase in blood flow. Additionally, dobutamine may also
lective adenosine receptor agonist. Its action is a direct vaso- have a (minor) direct vasodilative effect on coronary vessels.
dilation of the coronary arteries and has an extremely short Table 62.2 provides a comprehensive overview of the phar-
half-life of a few seconds. It is administered in continuous macological stress agent aforementioned.
infusion with a dose of 140 μg/kg/min for at least 2 min to
induce a heart rate increase of 10–20 beats above the resting
heart rate. Regadenoson is a selective A2A receptor agonist. It Data Analysis
is advantageous as it can be administered in a single bolus,
making stress CT studies more time efficient [28]. For the evaluation of SE static CTMPI, myocardial perfusion
Regadenoson stress, moreover, causes fewer systemic side defects are assessed by visual evaluation of contrast enhance-
effects, which is beneficial for patients with asthma or ment in the left ventricle. According to the literature [46, 63],
chronic obstructive pulmonary disease [30, 31]. On the other a visual evaluation can be affected by bias, as interpretation
hand, to take into account the longer-lasting effects of regad- of defects in the myocardial blood pool is often highly user-­
enoson, sufficient time needs to be allowed before rest acqui- dependent when compared to the absolute quantitative
sition (15–20  min). Alternatively, theophylline can be assessment obtained with dynamic perfusion imaging.
administered as a countermeasure [32]. The use of regadeno- In the case of DECT, independent from the technology
son showed greater efficiency over dipyridamole and ade- used, an iodine distribution map can be generated analyzing
nosine in a stress/rest protocol [28]. This was attributed the differences in attenuation of iodine between both energy
mostly to the administration of the substance, because regad- spectra and by merging the low and high kV datasets [64,
enoson is available as a pre-drawn single-dose syringe. 65]. Dedicated software typically generates the color-coded
Typical side effects of both agents are headache, flushing, iodine distribution map overlay on top of a virtual non-­
and dyspnea. Both drugs may cause moderate to severe com- contrasted image. The iodine concentrations are normalized
plications [33, 34]. Ventricular tachycardia and transient to the myocardial areas with normal perfusion [24]. As
asymptomatic atrial-ventricular block occur in 0.14% of myocardial iodine concentration has a direct relationship
patients. Recent case reports indicate that adenosine and with myocardial perfusion, the myocardial blood pool can
regadenoson may cause acute myocardial infarction and be quantified based on the per-voxel amount of iodine
death [34–36]. (Fig. 62.7) [23].
Table 62.2  Overview of pharmacological agents used in stress myocardial perfusion
Common side
Mechanism of action Effect Dosage Contraindications Advantages effects
Adenosine Nonselective adenosine Coronary artery vasodilator 140 μg/kg/min for High-grade AV block Good sensitivity and Flushing
receptor agonist 3–6 min Asthma or COPD specificity Headache
Sinus bradycardia Dyspnea
Systemic hypotension May provoke
(BP < 90 mmHg) acute ischemia
Severe carotid stenosis
Regadenoson A2A-selective adenosine Coronary artery vasodilator Bolus of 400 μg in 10–20 s High-grade AV block Dose independent from Dyspnea
receptor agonist Sinus bradycardia weight Headache
Systemic hypotension Fewer side effects in Flushing
(BP < 90 mmHg) patients with asthma May provoke
62  Myocardial Perfusion Imaging: Dual-Energy Approaches

Severe carotid stenosis and COPD acute ischemia

Dobutamine Stimulator of β1-adrenergic Increases heart rate, blood High-dose protocol: IV Concomitant therapy with Physiological Tachycardia
receptors pressure, and myocardial dobutamine infusion in 3 min β-blockers mechanism May provoke
contractility stages (10, 20, 30, 40 μg/kg/min) Severe hypertension Increased oxygen acute ischemia
Low-dose protocol: (>220/120 mmHg) consumption in the
5–10 μg/kg/min Congestive heart failure myocardium
Unstable angina
Aortic valve stenosis (peak
gradient >50 mmHg)
Hypertrophic
cardiomyopathy
Complex arrhythmias
Myocarditis
Pericarditis
Dipyridamole Increases availability of Indirect coronary artery 140 μg/kg/min for 4 min High-grade AV block Good sensitivity and Headache
adenosine by inhibiting vasodilator Asthma or COPD specificity Hypotension
adenosine deaminase Sinus bradycardia Inexpensive Flushing
Systemic hypotension Dyspnea
(BP < 90 mmHg)
Severe carotid stenosis
801
802 D. De Santis et al.

Fig. 62.7  A 57-year-old

male. (a) Calcified eccentric a b
plaques of the proximal
LAD cause a 70% occlusion.
(b) Stress DE-CTMPI-derived
iodine map shows an area of
hypoperfusion in the basal
anterior myocardial segment,
suggestive for myocardial
infarct. Iodine map provides
both qualitative and
quantitative assessment of the
myocardial wall. Despite the
fact that the two ROIs return
slightly similar attenuation
values for mixed images (91.4
versus 122.3 HU), there is a
substantial difference in
iodine density between the
hypoperfused area (0.3 mg/
mL) and the remote
myocardium without visual
perfusion defect (5.7 mg/mL).
Cinematic rendering (c) of the
long axis plane provides
extremely detailed anatomic c
details, albeit unable to
visually demonstrate
attenuation differences within
the myocardial wall

A comprehensive CT myocardial perfusion examination that of the stress acquisition, and the distinction between
including both rest and stress acquisitions allows to differen- reversible and fixed defects is not possible (Fig.  62.10).
tiate reversible from fixed MPD [23]. Reversible ischemia is CTMPI datasets are evaluated in the left ventricle short-axis
suspected when hypo-attenuating myocardial areas are pres- view. The myocardium is schematically divided into 17 seg-
ent only in the stress phase (Fig. 62.8), whereas a myocardial ments based on the American Heart Association model [66].
infarction is characterized by a myocardial perfusion defect Although the blood supply of the left ventricle is character-
persisting at rest (Fig.  62.9). Flow-limiting stenosis can ized by a great variability, segments 1, 2, 7, 8, 13, 14, and 17
cause hypo-attenuation during the rest phase as well, but the are traditionally assigned to the left anterior descending
diagnostic sensitivity of the rest phase is lower compared to artery (LAD). Segments 3, 4, 10, and 15 are assigned to the
62  Myocardial Perfusion Imaging: Dual-Energy Approaches 803

a c d e

Fig. 62.8 (a, b) Short axis view of dual-energy CT myocardial perfu- reformats of the coronary arteries (c and d) show a calcified and nar-
sion of a 56-year-old male with multivessel disease and previous rowed RCA (c, asterisk), multiple calcified plaques of the LCx (d, white
CABG.  Rest phase (a) shows no perfusion defects in the ventricular arrowhead), and mixed plaques of the LAD (d, black arrowhead).
wall, while the stress phase (b) depicts an area of hypo-attenuation in Cinematic rendering (e) depicts the hypo-attenuation of the septal
the inferior and infero-septal mid-ventricular wall (white arrows), ­ventricular wall (black arrows)
suggestive for myocardial inducible ischemia. Curved multi-planar
­

a c d e

Fig. 62.9 (a, b) Short axis view of dual-energy CT myocardial perfu- calcified plaque of the proximal LAD (c, white arrow) causing approxi-
sion of a 65-year-old male. Both rest (a) and stress (b) phases show a mately a 50% stenosis and a diffuse atherosclerosis of the LCx (d, black
fixed perfusion defect in the lateral wall (arrowheads) and in the ante- arrow) responsible of a 70% vessel stenosis. Cinematic rendering (e)
rior papillary muscle (black asterisk). The stress phase also unveils depicts the diseased LAD (white arrow) and LCx (black arrow) with
inducible perfusion defect in the anteroseptal wall (white asterisk). high-anatomical details
Curved multi-planar reformats of the coronary arteries (c and d) show a
804 D. De Santis et al.

a c e

b d f

Fig. 62.10  Long-axis (a) and axial (b) maximum intensity projections a basal anterior- and anteroseptal perfusion defect (arrow) suggestive
of dual-energy CT myocardial perfusion of a 46-year-old male with for inducible ischemia. Stress myocardial SPECT (e) confirms the
unstable angina show an ostial occlusion of the LAD (arrowheads). inducible ischemia (asterisk). Cinematic rendering (f) clearly depicts
DE-CTMPI at rest (c) does not identify any myocardial perfusion the ostial LAD occlusion (arrowhead)
defect, while the stress phase with superimposed iodine map (d) unveils

Fig. 62.11  Model of 17 left

ventricular segmentation Left Anterior Descending 1
proposed by the American 1) Basal anterior
Heart Association with 2) Basal anteroseptal
corresponding feeding 7) Mid anterior 7
vessels. Segments assigned to 8) Mid anteroseptal
LAD are in red, segments 13) Apical anterior
assigned to RCA are in green, 14) Apical septal 2
and segments assigned to LCx 17) Apex 13 6
8
are in blue 12
Right Coronary Artery
3) Basal inferoseptal
4) Basal inferior 14 17 16
9) Mid inferoseptal
10) Mid inferior
15) Apical inferior
9 11
15 5
Left Circumflex 3
5) Basal inferolateral
6) Basal anterolateral 10
11) Mid inferolateral
12) Mid anterolateral
16) Apical lateral
4

right coronary artery (RCA) when this vessel is dominant. LAD, the inferior wall and septum are assigned to the RCA,
Segments 5, 6, 11, 12, and 16 are assigned to the left circum- and the lateral wall is assigned to the LCx. The correspon-
flex coronary artery (LCx) [67]. In clinical practice, the dence between the coronary arterial anatomy and the sup-
­anterior wall and the anterior septum are assigned to the plied myocardium is depicted in Fig. 62.11.
62  Myocardial Perfusion Imaging: Dual-Energy Approaches 805

Clinical Results stenosis greater than 50% against a standard of invasive angi-
ography in 35 patients with a high risk of CAD.
Dual-energy CTMPI studies performed during rest or stress The first multicenter trial [56] evaluating 381 patients in
protocols showed good accuracy compared to various refer- comparison with combined invasive coronary angiography
ence modalities (Table  62.1) [44, 46, 59, 60, 63, 68–73]. (ICA) and SPECT confirmed the higher diagnostic perfor-
Ruzsics et al. investigated the feasibility of rest myocardial mance of combined CCTA and CTMPI.  The sensitivity,
perfusion imaging using a first-generation dual-source CT specificity, positive predictive value, and negative predictive
system compared to SPECT in 36 patients and reported a value of the combined methods in the patient-based analysis
sensitivity and specificity of 92% and 93%, respectively were 80%, 74%, 65%, and 86%, respectively. Also the iden-
[46]. The incremental benefit of DE-CTMPI stress series tification of patients with hemodynamically significant CAD
was demonstrated in an early feasibility study by Blankstein was significantly improved by the addition of CTMPI. The
et al. [43] where sensitivity was increased from 76% to 92% patient-based diagnostic accuracy of combined CCTA and
and specificity was similarly increased from 67% to 85% static CTMPI in patients without prior myocardial infarction
against the gold standard of PCI. Diagnostic accuracy was or CAD was 90% and 93%, respectively, and 87% in the
found comparable to that of SPECT. In a meta-analysis per- whole patient population.
formed by Pelgrim et  al., stress DE-CTMPI reported a The DE technique demonstrated superior diagnostic per-
segment-­based sensitivity of 75% and a specificity of 95% in formance compared to SE due to the use of iodine maps for
comparison to stress MRI, based on a population of 196 the detection of perfusion defects with SPECT as the refer-
patients [74]. Ko et  al. demonstrated the incremental diag- ence standard [42]. In the existing literature, areas with
nostic value of combined CCTA and CT perfusion compared reduced myocardial iodine content at DECT perfusion stud-
to CCTA alone with a marked increase in sensitivity (91.8– ies are mainly evaluated with a qualitative approach, which
93.5%), specificity (67.7–85.5%), positive predictive value can introduce bias given that myocardial blood pool defect
(73.6–88.3%), and negative predictive value (87.5–91.4%) interpretation is often highly user-dependent compared to the
[69]. Compared to the reference standard of SPECT, Wang absolute quantitative assessment obtained with dynamic per-
et al. showed increased sensitivity, specificity, and accuracy fusion. Multiple quantitative evaluations performed in phan-
of detecting coronary stenosis >50% by addition of tom and clinical studies demonstrated high accuracy for
DE-CTMPI from 82%, 91%, and 86% with CCTA alone to myocardial iodine quantification and the role of quantitative
90%, 86%, and 88%, respectively [63]. A recent study measurement in the differentiation between healthy and
reported that the combined analysis of CCTA and DE-CTMPI ­ischemic or necrotic myocardium [23, 72, 77]. A recently
increases the specificity from 50% to 67% in a population at published experiment confirmed the accuracy of the state-of-
high risk for coronary artery disease (yielding a 17% higher the-art dual-source and dual-layer DE scanners in the quanti-
specificity than that of either individual test) and outper- fication of iodine concentration, with slightly more accurate
forms the anatomical test of CCTA alone for detection of results provided by the dual-source system operating with the
hemodynamically significant coronary artery stenosis [75]. 150Sn/70 kVp or 150Sn/80 kVp combinations compared to
A trial of 50 patients comparing DECT and SPECT for the the dual-layer scanner operating at 140 kVp [78].
detection of obstructive CAD against a gold standard of Additionally, in a direct comparison of the different
CMR reported higher sensitivity (90% compared to 85%) image datasets generated by DECT, iodine maps showed the
and specificity (71% vs. 58%) with DE-CTMPI than with highest sensitivity and specificity among all DECT datasets
SPECT (Fig.  62.12) [45]. The most recent studies by Ko for the detection of myocardial perfusion defects in the
et al. and Kim et al. demonstrated similar accuracy compared DE-CTMPI [79].
to the reference modality, cardiac CMR [59, 68]. Ko et al. In a porcine model, beam-hardening reduction through
additionally found fair agreement between rest and stress DE prospective rapid kV switching has been demonstrated
DECT iodine maps and described an incremental value of [80]. In a direct comparison with SE-CTMPI, the diagnostic
the stress protocol (accuracy of 83%) compared to rest pro- performance of DE-CTMPI remained unaffected by the
tocol (accuracy of 62%) for the detection of hemodynami- presence of beam-hardening artifacts [81]. Moreover, a
cally significant CAD [59]. A number of studies comparing novel algorithm for beam-hardening artifact correction was
DE-CTMPI with SPECT made the interesting observation proven to be effective in avoiding false-positive findings and
that the modalities may lead to divergent classification of increase the specificity of DE-CTMPI [82]. The DECIDE-
perfusion defects [55, 63, 76]. Nearly one-half of perfusion Gold trial, a multicenter randomized study with the objec-
defects that are reversible at SPECT are reclassified as fixed tive to determine a comprehensive assessment of the
with DECT [55]. Rocha-Filho et al. [73] demonstrated con- diagnostic accuracy of the DECT and DE-CTMPI in the
siderable improvements in sensitivity (from 83% to 91%), assessment of hemodynamic significance of CAD compared
specificity (from 71% to 91%), positive predictive value to ICA, is currently ongoing [83].
(from 66% to 86%), and negative predictive value (from 87% In a comparison between DE and SE-CTMPI in the detec-
to 93%) by adding a stress phase to CCTA for detection of tion of mixed perfusion defects with SPECT as the reference
806 D. De Santis et al.

a c e

b d f

g h i j

Fig. 62.12  A 68-year-old male. Myocardial SPECT at rest (a) and CMR (f) that clearly demonstrates the transmural delayed enhancement
peak stress (b) show a fixed perfusion defect in the anteroseptal wall with high contrast resolution in comparison with the healthy myocar-
(white asterisks). Dual-energy CT myocardial perfusion at rest (c) and dium. All these findings are suggestive of myocardial infarct. Curved
stress (d) phase with superimposed iodine maps confirms the fixed per- multi-­planar reformats of the LAD, LCx, and RCA (g, h, and i, respec-
fusion defect (black asterisks). The acquisition at the delayed phase tively) depict multiple calcified and mixed coronary plaques (arrows).
(e) shows an area of late enhancement in the corresponding left ven- Cinematic rendering (j) provides a detailed overview of the coronary
tricular segment (black arrowhead). The CTMPI is consistent with the arteries
62  Myocardial Perfusion Imaging: Dual-Energy Approaches 807

standard, DE-CTMPI showed better results in terms of sensi- effective first-line imaging modality in the workup of patients
tivity (91% versus 55%), negative predictive value (97% versus with CAD [45].
86%) and accuracy (93% versus 85%). However, SE-CTMPI
showed a higher specificity (98% versus 94%) [42].
Summary

Limitations A growing body of clinical evidence shows that combined

CCTA and CTMPI enable the morphological and functional
The main limitation of static SE- and DE-CTMPI techniques assessment of CAD with high accuracy in a single technique.
resides in their reliance upon the acquisition timing, which CTMPI is most likely to be used as an add-on to CCTA to
can significantly influence the diagnostic accuracy. As increase the specificity for hemodynamically relevant
single-­shot techniques only acquire a single data set, the lesions, in particular in the setting of a stenosis of intermedi-
peak contrast attenuation could be missed [21]. Furthermore, ate severity [91]. DE-CTMPI has been extensively investi-
the data collection from consecutive parts of the heart may gated and it has shown its usefulness in various clinical
be performed during different cardiac cycles depending on settings. As such, it may play a role in the short- and long-­
the CT system and acquisition technique, resulting in hetero- term management of patients with CAD. This is owed to the
geneous apicobasal attenuation. In addition, several types of fact that DE-CTMPI provides a comprehensive and accurate
artifacts may be associated with single-shot image acquisi- evaluation of coronary arteries and left ventricular function
tions, such as beam-hardening, motion, and partial scan arti- and ultimately enables physicians to provide the best and
facts [84–86]. most advanced therapy to their patients. However, future
studies are required to further evaluate the cost-effectiveness
of DE-CTMPI, its place in clinical management, and its role
Cost-Effectiveness in CAD diagnosis in comparison with dynamic CTMPI,
especially in light of new emerging noninvasive functional
As healthcare costs have continued to rise, the cost-­ coronary diagnostic techniques such as CT-FFR.
effectiveness of novel techniques has been an area of increas-
ing interest. This holds true across all fields of medicine,
including CT. There have been extensive investigations com- References
paring the cost-effectiveness of CCTA to that of SPECT. It is
well established that individuals without known CAD who 1. Bucher AM, De Cecco CN, Schoepf UJ, Wang R, Meinel FG,
Binukrishnan SR, et  al. Cardiac CT for myocardial ischaemia
undergo CCTA as an initial diagnostic test, compared with detection and characterization–comparative analysis. Br J Radiol.
those who underwent myocardial SPECT, incurred lower 2014;87:20140159.
healthcare costs [87, 88]. Moreover, CCTA is the most cost-­ 2. World Health Organization. The top 10 causes of death. http://www.
who.int/en/news-room/fact-sheets/detail/the-top-10-causes-of-death
effective diagnostic strategy for patients with suspected
3. Writing Group M, Mozaffarian D, Benjamin EJ, Go AS, Arnett
acute coronary syndrome and negative troponin [89]. The DK, Blaha MJ, et  al. Heart disease and stroke statistics-2016
state-of-the-art scanners available today in clinical practice update: a report from the American Heart Association. Circulation.
are also cost-effective for challenging patients with CAD, 2016;133:e38–360.
4. American Heart Association. Heart disease and stroke statistics–at-­
such as obese patients, patients with high or irregular heart-
a-­glance. 2015. https://www.heart.org/idc/groups/ahamah-public/@
beat, and patients who have high levels of coronary calcium wcm/@sop/@smd/documents/downloadable/ucm_470704.pdf
or a previous stent or bypass graft [90]. 5. Taylor AJ, Cerqueira M, Hodgson JM, Mark D, Min J, O’Gara P,
To date, only a single study comparing the cost-­ et  al.  ACCF/SCCT/ACR/AHA/ASE/ASNC/NASCI/SCAI/SCMR
2010 appropriate use criteria for cardiac computed tomography. A
effectiveness of DE-CTMPI versus SPECT was performed
report of the American College of Cardiology Foundation Appropriate
in the United States and published in 2012. Meyer et  al. Use Criteria Task Force, the Society of Cardiovascular Computed
investigated the performance of a stress-rest DE-CTMPI Tomography, the American College of Radiology, the American
protocol versus a stress-rest SPECT protocol in 50 patients Heart Association, the American Society of Echocardiography,
the American Society of Nuclear Cardiology, the North American
with CAD, using CMR as reference standard. DE-CTMPI
Society for Cardiovascular Imaging, the Society for Cardiovascular
resulted in lower costs in comparison to routine myocardial Angiography and Interventions, and the Society for Cardiovascular
perfusion SPECT ($2631 versus $2938, respectively; Magnetic Resonance. Circulation. 2010;122:e525–55.
P  <  0.001), better results in terms of quality-adjusted life-­ 6. De Cecco CN, Meinel FG, Chiaramida SA, Costello P, Bamberg
F, Schoepf UJ.  Coronary artery computed tomography scanning.
years (QALY) (14.13 versus 13.49 QALY), and, eventually,
Circulation. 2014;129:1341–5.
a more convenient cost profile of $3191 incremental cost-­ 7. Ohnesorge BM, Hofmann LK, Flohr TG, Schoepf UJ. CT for imag-
effectiveness ratio (ICER) in comparison to SPECT ($3557, ing coronary artery disease: defining the paradigm for its applica-
P < 0.001). Hence, DE-CTMPI has been proven to be a cost-­ tion. Int J Cardiovasc Imaging. 2005;21:85–104.
808 D. De Santis et al.

8. Flohr TG, Schoepf UJ, Ohnesorge BM.  Chasing the heart: new for the measurement of iodine concentration using cardiac CT pro-
developments for cardiac CT. J Thorac Imaging. 2007;22:4–16. tocols: validation in a phantom model. Eur Radiol. 2014;24:512–8.
9. von Ballmoos MW, Haring B, Juillerat P, Alkadhi H. Meta-analysis: 24. Kang DK, Schoepf UJ, Bastarrika G, Nance JW Jr, Abro JA,

diagnostic performance of low-radiation-dose coronary computed Ruzsics B.  Dual-energy computed tomography for integrative
tomography angiography. Ann Intern Med. 2011;154:413–20. imaging of coronary artery disease: principles and clinical applica-
10. Jaarsma C, Leiner T, Bekkers SC, Crijns HJ, Wildberger JE, Nagel tions. Semin Ultrasound CT MR. 2010;31:276–91.
E, et al. Diagnostic performance of noninvasive myocardial perfu- 25. Siegel MJ, Kaza RK, Bolus DN, Boll DT, Rofsky NM, De Cecco
sion imaging using single-photon emission computed tomography, CN, et al. White paper of the society of computed body tomography
cardiac magnetic resonance, and positron emission tomography and magnetic resonance on dual-energy CT, part 1: technology and
imaging for the detection of obstructive coronary artery disease: a terminology. J Comput Assist Tomogr. 2016;40:841–5.
meta-analysis. J Am Coll Cardiol. 2012;59:1719–28. 26. Matsumoto K, Jinzaki M, Tanami Y, Ueno A, Yamada M,

11. Pijls NH, Fearon WF, Tonino PA, Siebert U, Ikeno F, Bornschein B, Kuribayashi S.  Virtual monochromatic spectral imaging with
et al. Fractional flow reserve versus angiography for guiding per- fast kilovoltage switching: improved image quality as compared
cutaneous coronary intervention in patients with multivessel coro- with that obtained with conventional 120-kVp CT.  Radiology.
nary artery disease: 2-year follow-up of the FAME (Fractional Flow 2011;259:257–62.
Reserve Versus Angiography for Multivessel Evaluation) study. J 27. Gabbai M, Leichter I, Mahgerefteh S, Sosna J.  Spectral mate-
Am Coll Cardiol. 2010;56:177–84. rial characterization with dual-energy CT: comparison of com-
12. Shaw LJ, Berman DS, Maron DJ, Mancini GB, Hayes SW, Hartigan mercial and investigative technologies in phantoms. Acta Radiol.
PM, et al. Optimal medical therapy with or without percutaneous 2015;56:960–9.
coronary intervention to reduce ischemic burden: results from the 28. Nance JW Jr, Bastarrika G, Kang DK, Ruzsics B, Vogt S, Schmidt B,
Clinical Outcomes Utilizing Revascularization and Aggressive et al. High-temporal resolution dual-energy computed tomography
Drug Evaluation (COURAGE) trial nuclear substudy. Circulation. of the heart using a novel hybrid image reconstruction algorithm:
2008;117:1283–91. initial experience. J Comput Assist Tomogr. 2011;35:119–25.
13. Kern MJ, Samady H.  Current concepts of integrated coronary
29. Arnoldi E, Lee YS, Ruzsics B, Weininger M, Spears JR, Rowley
physiology in the catheterization laboratory. J Am Coll Cardiol. CP, et al. CT detection of myocardial blood volume deficits: dual-­
2010;55:173–85. energy CT compared with single-energy CT spectra. J Cardiovasc
14. Meijboom WB, Van Mieghem CA, van Pelt N, Weustink A,
Comput Tomogr. 2011;5:421–9.
Pugliese F, Mollet NR, et al. Comprehensive assessment of coro- 30. Blankstein R, Shturman LD, Rogers IS, Rocha-Filho JA, Okada
nary artery stenoses: computed tomography coronary angiography DR, Sarwar A, et al. Adenosine-induced stress myocardial perfu-
versus conventional coronary angiography and correlation with sion imaging using dual-source cardiac computed tomography. J
fractional flow reserve in patients with stable angina. J Am Coll Am Coll Cardiol. 2009;54:1072–84.
Cardiol. 2008;52:636–43. 31. Ko SM, Choi JW, Song MG, Shin JK, Chee HK, Chung HW, et al.
15. Sarno G, Decraemer I, Vanhoenacker PK, De Bruyne B, Hamilos Myocardial perfusion imaging using adenosine-induced stress
M, Cuisset T, et al. On the inappropriateness of noninvasive mul- dual-energy computed tomography of the heart: comparison with
tidetector computed tomography coronary angiography to trigger cardiac magnetic resonance imaging and conventional coronary
coronary revascularization: a comparison with invasive angiogra- angiography. Eur Radiol. 2011;21:26–35.
phy. JACC Cardiovasc Interv. 2009;2:550–7. 32. Meyer M, Nance JW Jr, Schoepf UJ, Moscariello A, Weininger
16. Stuijfzand WJ, Danad I, Raijmakers PG, Marcu CB, Heymans MW, M, Rowe GW, et  al. Cost-effectiveness of substituting dual-
van Kuijk CC, et al. Additional value of transluminal attenuation energy CT for SPECT in the assessment of myocardial perfusion
gradient in CT angiography to predict hemodynamic significance for the workup of coronary artery disease. Eur J Radiol. 2012;81:
of coronary artery stenosis. J Am Coll Cardiol Img. 2014;7:374–86. 3719–25.
17. De Cecco CN, Varga-Szemes A, Meinel FG, Renker M, Schoepf 33. Ruzsics B, Schwarz F, Schoepf UJ, Lee YS, Bastarrika G,

UJ. Beyond stenosis detection: computed tomography approaches Chiaramida SA, et  al. Comparison of dual-energy computed
for determining the functional relevance of coronary artery disease. tomography of the heart with single photon emission computed
Radiol Clin N Am. 2015;53:317–34. tomography for assessment of coronary artery stenosis and of the
18. Wang R, Renker M, Schoepf UJ, Wichmann JL, Fuller SR, Rier JD, myocardial blood supply. Am J Cardiol. 2009;104:318–26.
et al. Diagnostic value of quantitative stenosis predictors with coro- 34. Ruzsics B, Lee H, Powers ER, Flohr TG, Costello P, Schoepf
nary CT angiography compared to invasive fractional flow reserve. UJ.  Images in cardiovascular medicine. Myocardial ischemia
Eur J Radiol. 2015;84:1509–15. diagnosed by dual-energy computed tomography: correlation
19. Varga-Szemes A, Meinel FG, De Cecco CN, Fuller SR, Bayer RR with single-photon emission computed tomography. Circulation.
2nd, Schoepf UJ.  CT myocardial perfusion imaging. AJR Am J 2008;117:1244–5.
Roentgenol. 2015;204:487–97. 35. Fahmi R, Eck BL, Levi J, Fares A, Dhanantwari A, Vembar M, et al.
20. George RT, Silva C, Cordeiro MA, DiPaula A, Thompson DR, Quantitative myocardial perfusion imaging in a porcine ischemia
McCarthy WF, et al. Multidetector computed tomography myocar- model using a prototype spectral detector CT system. Phys Med
dial perfusion imaging during adenosine stress. J Am Coll Cardiol. Biol. 2016;61:2407–31.
2006;48:153–60. 36. Meinel FG, Canstein C, Schoepf UJ, Sedlmaier M, Schmidt B,
21. Bischoff B, Bamberg F, Marcus R, Schwarz F, Becker HC, Becker Harris BS, et al. Image quality and radiation dose of low tube volt-
A, et al. Optimal timing for first-pass stress CT myocardial perfu- age 3rd generation dual-source coronary CT angiography in obese
sion imaging. Int J Cardiovasc Imaging. 2013;29:435–42. patients: a phantom study. Eur Radiol. 2014;24:1643–50.
22. Cannao PM, Schoepf UJ, Muscogiuri G, Wichmann JL, Fuller SR, 37. Vliegenthart R, Pelgrim GJ, Ebersberger U, Rowe GW, Oudkerk M,
Secchi F, et al. Technical prerequisites and imaging protocols for Schoepf UJ. Dual-energy CT of the heart. AJR Am J Roentgenol.
dynamic and dual energy myocardial perfusion imaging. Eur J 2012;199:S54–63.
Radiol. 2015;84:2401–10. 38. George RT, Arbab-Zadeh A, Miller JM, Vavere AL, Bengel FM,
23. Koonce JD, Vliegenthart R, Schoepf UJ, Schmidt B, Wahlquist AE, Lardo AC, et  al. Computed tomography myocardial perfusion
Nietert PJ, et  al. Accuracy of dual-energy computed tomography imaging with 320-row detector computed tomography accurately
62  Myocardial Perfusion Imaging: Dual-Energy Approaches 809

detects myocardial ischemia in patients with obstructive coronary 55. Jin KN, De Cecco CN, Caruso D, Tesche C, Spandorfer A, Varga-­
artery disease. Circ Cardiovasc Imaging. 2012;5:333–40. Szemes A, et  al. Myocardial perfusion imaging with dual energy
39. Habis M, Capderou A, Sigal-Cinqualbre A, Ghostine S, Rahal S, CT. Eur J Radiol. 2016;85:1914–21.
Riou JY, et  al. Comparison of delayed enhancement patterns on 56. Rochitte CE, George RT, Chen MY, Arbab-Zadeh A, Dewey M,
multislice computed tomography immediately after coronary angi- Miller JM, et  al. Computed tomography angiography and perfu-
ography and cardiac magnetic resonance imaging in acute myocar- sion to assess coronary artery stenosis causing perfusion defects
dial infarction. Heart. 2009;95:624–9. by single photon emission computed tomography: the CORE320
40. Deseive S, Bauer RW, Lehmann R, Kettner M, Kaiser C, Korkusuz study. Eur Heart J. 2014;35:1120–30.
H, et  al. Dual-energy computed tomography for the detection of 57. Dey D, Slomka PJ, Berman DS. Achieving very-low-dose radiation
late enhancement in reperfused chronic infarction: a comparison exposure in cardiac computed tomography, single-photon emission
to magnetic resonance imaging and histopathology in a porcine computed tomography, and positron emission tomography. Circ
model. Investig Radiol. 2011;46:450–6. Cardiovasc Imaging. 2014;7:723–34.
41. Meinel FG, De Cecco CN, Schoepf UJ, Nance JW Jr, Silverman 58. Goldstein JA, Chinnaiyan KM, Abidov A, Achenbach S, Berman DS,
JR, Flowers BA, et  al. First-arterial-pass dual-energy CT for Hayes SW, et al. The CT-STAT (Coronary Computed Tomographic
assessment of myocardial blood supply: do we need rest, stress, Angiography for Systematic Triage of Acute Chest Pain Patients to
and delayed acquisition? Comparison with SPECT.  Radiology. Treatment) trial. J Am Coll Cardiol. 2011;58:1414–22.
2014;270:708–16. 59. Ko SM, Park JH, Hwang HK, Song MG.  Direct comparison of
42. Dole WP, Yamada N, Bishop VS, Olsson RA. Role of adenosine in stress- and rest-dual-energy computed tomography for detec-
coronary blood flow regulation after reductions in perfusion pres- tion of myocardial perfusion defect. Int J Cardiovasc Imaging.
sure. Circ Res. 1985;56:517–24. 2014;30(Suppl 1):41–53.
43. Gould KL, Lipscomb K. Effects of coronary stenoses on coronary 60. Ko SM, Song MG, Chee HK, Hwang HK, Feuchtner GM, Min
flow reserve and resistance. Am J Cardiol. 1974;34:48–55. JK.  Diagnostic performance of dual-energy CT stress myocar-
44. Sambuceti G, Marzilli M, Marraccini P, Schneider-Eicke J,
dial perfusion imaging: direct comparison with cardiovascular
Gliozheni E, Parodi O, et  al. Coronary vasoconstriction dur- MRI. AJR Am J Roentgenol. 2014;203:W605–13.
ing myocardial ischemia induced by rises in metabolic demand 61. Yin WH, Lu B, Hou ZH, Li N, Han L, Wu YJ, et  al. Detection
in patients with coronary artery disease. Circulation. 1997;95: of coronary artery stenosis with sub-milliSievert radiation dose by
2652–9. prospectively ECG-triggered high-pitch spiral CT angiography and
45. Kurata A, Mochizuki T, Koyama Y, Haraikawa T, Suzuki J,
iterative reconstruction. Eur Radiol. 2013;23:2927–33.
Shigematsu Y, et al. Myocardial perfusion imaging using adenos- 62. Zhang LJ, Qi L, Wang J, Tang CX, Zhou CS, Ji XM, et al. Feasibility
ine triphosphate stress multi-slice spiral computed tomography: of prospectively ECG-triggered high-pitch coronary CT angiogra-
alternative to stress myocardial perfusion scintigraphy. Circ J. phy with 30 mL iodinated contrast agent at 70 kVp: initial experi-
2005;69:550–7. ence. Eur Radiol. 2014;24:1537–46.
46. Rumberger JA, Feiring AJ, Lipton MJ, Higgins CB, Ell SR, Marcus 63. Wang R, Yu W, Wang Y, He Y, Yang L, Bi T, et al. Incremental value
ML. Use of ultrafast computed tomography to quantitate regional of dual-energy CT to coronary CT angiography for the detection of
myocardial perfusion: a preliminary report. J Am Coll Cardiol. significant coronary stenosis: comparison with quantitative coro-
1987;9:59–69. nary angiography and single photon emission computed tomogra-
47. Mahmarian JJ, Cerqueira MD, Iskandrian AE, Bateman TM,
phy. Int J Cardiovasc Imaging. 2011;27:647–56.
Thomas GS, Hendel RC, et al. Regadenoson induces comparable 64. Coursey CA, Nelson RC, Boll DT, Paulson EK, Ho LM, Neville
left ventricular perfusion defects as adenosine: a quantitative analy- AM, et al. Dual-energy multidetector CT: how does it work, what
sis from the ADVANCE MPI 2 trial. JACC Cardiovasc Imaging. can it tell us, and when can we use it in abdominopelvic imaging?
2009;2:959–68. Radiographics. 2010;30:1037–55.
48. Salgado Garcia C, Jimenez Heffernan A, Sanchez de Mora E, 65. Williams MC, Newby DE. CT myocardial perfusion imaging: cur-
Ramos Font C, Lopez Martin J, Rivera de los Santos F, et  al. rent status and future directions. Clin Radiol. 2016;71:739–49.
Comparative study of the safety of regadenoson between patients 66. Cerqueira MD, Weissman NJ, Dilsizian V, Jacobs AK, Kaul S,
with mild/moderate chronic obstructive pulmonary disease and Laskey WK, et  al. Standardized myocardial segmentation and
asthma. Eur J Nucl Med Mol Imaging. 2014;41:119–25. nomenclature for tomographic imaging of the heart. A statement
49. Becker A, Becker C. CT imaging of myocardial perfusion: possi- for healthcare professionals from the Cardiac Imaging Committee
bilities and perspectives. J Nucl Cardiol. 2013;20:289–96. of the Council on Clinical Cardiology of the American Heart
50. Iskandrian AE, Bateman TM, Belardinelli L, Blackburn B,
Association. Circulation. 2002;105:539–42.
Cerqueira MD, Hendel RC, et  al. Adenosine versus regadenoson 67. Donato P, Coelho P, Santos C, Bernardes A, Caseiro-Alves F.

comparative evaluation in myocardial perfusion imaging: results Correspondence between left ventricular 17 myocardial segments
of the ADVANCE phase 3 multicenter international trial. J Nucl and coronary anatomy obtained by multi-detector computed tomog-
Cardiol. 2007;14:645–58. raphy: an ex vivo contribution. Surg Radiol Anat. 2012;34:805–10.
51. Luu JM, Filipchuk NG, Friedrich MG.  Indications, safety and
68. Kim SM, Chang SA, Shin W, Choe YH. Dual-energy CT perfusion
image quality of cardiovascular magnetic resonance: experience in during pharmacologic stress for the assessment of myocardial per-
>5,000 North American patients. Int J Cardiol. 2013;168:3807–11. fusion defects using a second-generation dual-source CT: a com-
52. Shah S, Parra D, Rosenstein RS. Acute myocardial infarction dur- parison with cardiac magnetic resonance imaging. J Comput Assist
ing regadenoson myocardial perfusion imaging. Pharmacotherapy. Tomogr. 2014;38:44–52.
2013;33:e90–5. 69. Ko SM, Choi JW, Hwang HK, Song MG, Shin JK, Chee

53. Hsi DH, Marreddy R, Moshiyakhov M, Luft U.  Regadenoson
HK.  Diagnostic performance of combined noninvasive anatomic
induced acute ST-segment elevation myocardial infarction and mul- and functional assessment with dual-source CT and adenosine-­
tivessel coronary thrombosis. J Nucl Cardiol. 2013;20:481–4. induced stress dual-energy CT for detection of significant coronary
54. Geleijnse ML, Elhendy A, Fioretti PM, Roelandt JR. Dobutamine stenosis. AJR Am J Roentgenol. 2012;198:512–20.
stress myocardial perfusion imaging. J Am Coll Cardiol. 70. Weininger M, Schoepf UJ, Ramachandra A, Fink C, Rowe GW,
2000;36:2017–27. Costello P, et  al. Adenosine-stress dynamic real-time myocardial
810 D. De Santis et al.

perfusion CT and adenosine-stress first-pass dual-energy myocar- puted tomography for ischemia determination compared to “gold
dial perfusion CT for the assessment of acute chest pain: initial standard” non-invasive and invasive techniques (DECIDE-Gold):
results. Eur J Radiol. 2012;81:3703–10. a multicenter international efficacy diagnostic study of rest-stress
71. Delgado C, Vazquez M, Oca R, Vilar M, Trinidad C, Sanmartin dual-energy computed tomography angiography with perfusion.
M.  Myocardial ischemia evaluation with dual-source computed J Nucl Cardiol. 2015;22:1031–40.
tomography: comparison with magnetic resonance imaging. Rev 84. Nagao M, Matsuoka H, Kawakami H, Higashino H, Mochizuki T,
Esp Cardiol (Engl Ed). 2013;66:864–70. Ohshita A, et al. Detection of myocardial ischemia using 64-slice
72. Delgado Sanchez-Gracian C, Oca Pernas R, Trinidad Lopez C, MDCT. Circ J. 2009;73:905–11.
Santos Armentia E, Vaamonde Liste A, Vazquez Caamano M, 85. Busch JL, Alessio AM, Caldwell JH, Gupta M, Mao S, Kadakia J,
et  al. Quantitative myocardial perfusion with stress dual-energy et al. Myocardial hypo-enhancement on resting computed tomogra-
CT: iodine concentration differences between normal and isch- phy angiography images accurately identifies myocardial hypoper-
emic or necrotic myocardium. Initial experience. Eur Radiol. fusion. J Cardiovasc Comput Tomogr. 2011;5:412–20.
2016;26:3199–207. 86. Stenner P, Schmidt B, Bruder H, Allmendinger T, Haberland U,
73. Rocha-Filho JA, Blankstein R, Shturman LD, Bezerra HG, Okada Flohr T, et al. Partial scan artifact reduction (PSAR) for the assess-
DR, Rogers IS, et al. Incremental value of adenosine-induced stress ment of cardiac perfusion in dynamic phase-correlated CT.  Med
myocardial perfusion imaging with dual-source CT at cardiac CT Phys. 2009;36:5683–94.
angiography. Radiology. 2010;254:410–9. 87. Min JK, Kang N, Shaw LJ, Devereux RB, Robinson M, Lin F, et al.
74. Pelgrim GJ, Dorrius M, Xie X, den Dekker MA, Schoepf UJ, Costs and clinical outcomes after coronary multidetector CT angi-
Henzler T, et al. The dream of a one-stop-shop: meta-analysis on ography in patients without known coronary artery disease: compar-
myocardial perfusion CT. Eur J Radiol. 2015;84:2411–20. ison to myocardial perfusion SPECT. Radiology. 2008;249:62–70.
75. De Cecco CN, Harris BS, Schoepf UJ, Silverman JR, McWhite CB, 88. Min JK, Shaw LJ, Berman DS, Gilmore A, Kang N. Costs and clin-
Krazinski AW, et  al. Incremental value of pharmacological stress ical outcomes in individuals without known coronary artery disease
cardiac dual-energy CT over coronary CT angiography alone for undergoing coronary computed tomographic angiography from an
the assessment of coronary artery disease in a high-risk population. analysis of Medicare category III transaction codes. Am J Cardiol.
AJR Am J Roentgenol. 2014;203:W70–7. 2008;102:672–8.
76. Vliegenthart R, Henzler T, Moscariello A, Ruzsics B, Bastarrika G, 89. Goodacre S, Thokala P, Carroll C, Stevens JW, Leaviss J, Al Khalaf
Oudkerk M, et al. CT of coronary heart disease: part 1, CT of myo- M, et al. Systematic review, meta-analysis and economic modelling
cardial infarction, ischemia, and viability. AJR Am J Roentgenol. of diagnostic strategies for suspected acute coronary syndrome.
2012;198:531–47. Health Technol Assess. 2013;17:v-vi, 1–188.
77. Thilo C, Hanley M, Bastarrika G, Ruzsics B, Schoepf UJ. Integrative 90. Westwood M, Al M, Burgers L, Redekop K, Lhachimi S,

computed tomographic imaging of cardiac structure, function, per- Armstrong N, et al. A systematic review and economic evaluation
fusion, and viability. Cardiol Rev. 2010;18:219–29. of new-generation computed tomography scanners for imaging in
78. Pelgrim GJ, van Hamersvelt RW, Willemink MJ, Schmidt BT, Flohr coronary artery disease and congenital heart disease: Somatom
T, Schilham A, et al. Accuracy of iodine quantification using dual Definition Flash, Aquilion ONE, Brilliance iCT and Discovery
energy CT in latest generation dual source and dual layer CT. Eur CT750 HD. Health Technol Assess. 2013;17:1–243.
Radiol. 2017;27(9):3904–12. 91. Boiselle PM, Choe YH, Leipsic J, Pugliese F, Schoepf UJ,

79. Li W, Zhu X, Li J, Peng C, Chen N, Qi Z, et  al. Comparison of Vliegenthart R. Expert opinion: how and when to perform CT myo-
the sensitivity and specificity of 5 image sets of dual-energy com- cardial perfusion imaging. J Thorac Imaging. 2015;30:167–8.
puted tomography for detecting first-pass myocardial perfusion 92. Bauer RW, Kerl JM, Fischer N, Burkhard T, Larson MC, Ackermann
defects compared with positron emission tomography. Medicine. H, et al. Dual-energy CT for the assessment of chronic myocardial
2014;93:e329. infarction in patients with chronic coronary artery disease: com-
80. So A, Hsieh J, Imai Y, Narayanan S, Kramer J, Procknow K, et al. parison with 3-T MRI. AJR Am J Roentgenol. 2010;195:639–46.
Prospectively ECG-triggered rapid kV-switching dual-energy CT 93. Nagao M, Kido T, Watanabe K, Saeki H, Okayama H, Kurata A,
for quantitative imaging of myocardial perfusion. JACC Cardiovasc et al. Functional assessment of coronary artery flow using adenos-
Imaging. 2012;5:829–36. ine stress dual-energy CT: a preliminary study. Int J Cardiovasc
81. Carrascosa PM, Cury RC, Deviggiano A, Capunay C, Campisi R, Imaging. 2011;27:471–81.
Lopez de Munain M, et  al. Comparison of myocardial perfusion 94. Zhao RP, Hao ZR, Song ZJ. Diagnostic value of Flash dual-source
evaluation with single versus dual-energy CT and effect of beam-­ CT coronary artery imaging combined with dual-energy myocar-
hardening artifacts. Acad Radiol. 2015;22:591–9. dial perfusion imaging for coronary heart disease. Exp Ther Med.
82. Bucher AM, Wichmann JL, Schoepf UJ, Wolla CD, Canstein C, 2014;7:865–8.
McQuiston AD, et  al. Quantitative evaluation of beam-hardening 95. Kido T, Watanabe K, Saeki H, Shigemi S, Matsuda T, Yamamoto M,
artefact correction in dual-energy CT myocardial perfusion imag- et al. Adenosine triphosphate stress dual-source computed tomog-
ing. Eur Radiol. 2016;26:3215–22. raphy to identify myocardial ischemia: comparison with invasive
83. Truong QA, Knaapen P, Pontone G, Andreini D, Leipsic J,
coronary angiography. Springerplus. 2014;3:75.
Carrascosa P, et al. Rationale and design of the dual-energy com-
 Dynamic Myocardial CT Perfusion
Imaging 63
Marly van Assen, Gert Jan Pelgrim,
and Rozemarijn Vliegenthart

Several noninvasive imaging techniques are available for the disease (CAD). Computed tomography (CT), however, is
evaluation of myocardial perfusion. Nuclear techniques are able to evaluate both with state-of-the-art imaging tech-
positron emission tomography (PET) and single-photon emis- niques. Coronary CT angiography (CCTA) is often used for
sion computed tomography (SPECT). PET is the current gold noninvasive evaluation of coronary anatomy, yielding a very
standard for quantitative evaluation of myocardial perfusion, high sensitivity and negative predictive value close to 100%
while SPECT is the most commonly used imaging method for [4]. Combining CT myocardial perfusion imaging (CTMPI)
myocardial ischemia [1]. Cardiac magnetic resonance (MRI) with CCTA allows for anatomical and functional evaluation
perfusion imaging has, however, shown better accuracy for the of CAD using a single modality and examination.
detection of myocardial perfusion defects [2]. Dynamic CTMPI uses multiple acquisitions to capture
In a meta-analysis involving 11,826 patients [3], myocardial the first pass of contrast medium during its wash in and
perfusion evaluation with PET had a higher sensitivity for washout through the myocardium; this is in contrast to static
myocardial ischemia than evaluation with SPECT, 92.6% com- CT perfusion imaging which only shows contrast distribu-
pared to 88.3%, respectively. There was no significant differ- tion across the myocardium at a single point in time [5–8].
ence between the specificity of PET and SPECT, 81.3% The main advantage of dynamic CTMPI is the potential of
compared to 76.0%. Jaarsma et al. [1] showed in a recent meta- quantitative analysis of myocardial blood flow (MBF), myocar-
analysis (17,901 patients) a pooled sensitivity of 84%, 88%, dial blood volume (MBV), and other perfusion-­related param-
and 89% for PET, SPECT, and MRI, respectively, with a pooled eters directly from the CT images, for both the whole heart and
specificity of 81%, 61%, and 76%. They concluded that both for individual myocardial segments. Static CT does not provide
PET and MRI showed higher diagnostic accuracy than SPECT. a quantitative measure of perfusion but a relative map of single-
None of the abovementioned techniques allow simultane- time-point myocardial contrast distribution. Quantification of
ous anatomical and functional evaluation of coronary artery MBF could especially improve the detection of ischemia in
patients with three-vessel disease because this approach relies
on absolute values of MBF rather than on the differences
M. van Assen between normal and ischemic myocardium. Also, subtle sub-
Department of Radiology, University of Groningen, University
Medical Center Groningen, Groningen, The Netherlands clinical decreases in MBF can be detected by quantifying perfu-
sion, while not yet visible as gross perfusion defects [9].
Department of Radiology and Radiological Science, Division of
Cardiovascular Imaging, Medical University of South Carolina CTMPI was first studied in a number of animal studies.
Charleston, SC, USA Studies in pigs and dogs showed that dynamic CTMPI was
e-mail: [email protected] able to determine accurate MBF values, using 16- or 64-multi-
G. J. Pelgrim detector CT (MDCT) during rest and stress. CTMPI-determined
Department of Radiology, University of Groningen, University semiquantitative and quantitative MBF values as well as rest/
Medical Center Groningen, Groningen, The Netherlands stress MBF ratios were shown to have a high correlation with
R. Vliegenthart (*) microsphere-derived MBF [10–12]. In studies by Mahnken
Department of Radiology, University of Groningen, University et al. [13] and Bamberg et al. [14], a dual-source CT (DSCT)
Medical Center Groningen, Groningen, The Netherlands
system with alternating table positions (“shuttle mode”)
Department of Radiology and Radiological Science, Division of allowed for whole heart imaging in pigs, with a total coverage
Cardiovascular Imaging, Medical University of South Carolina
Charleston, SC, USA of 73 mm compared to 38 mm using only one table position.
e-mail: [email protected] DSCT scanners have an improved temporal resolution due to

© Humana Press 2019 811

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_63
812 M. van Assen et al.

the use of two X-ray sources, resulting in the possibility to scan table position, allowing a higher temporal sampling rate. These
a slab in less than 150 ms. Using DSCT in shuttle mode, they CT systems offer the same or even more coverage compared to
were able to show the hemodynamic effect of a coronary steno- DSCT scanners without the disadvantage of table movement in
sis with dynamic CTMPI in stress phase and concluded that shuttle mode, so far however at the disadvantage of higher radia-
CT-determined MBF measurements could differentiate isch- tion dose.
emic from nonischemic myocardium. However, MBF was
underestimated compared to microsphere-determined MBF.
Rossi et al. [15] showed that dynamic dual-source CTMPI is Cardiac Phase for Acquisition
able to distinguish regions with reduced MBF during stress
phase imaging with a good correlation with FFR m ­ easurements. Image acquisition is possible in systolic or diastolic phase.
They used a higher dose of adenosine as a stressor agent, Motwani et al. [17, 18] studied the effect of systolic and dia-
500 μg/kg/min, compared to Bamberg et al. and Mahnken et al. stolic acquisition on quantitative MRI perfusion imaging and
who used 140  μg/kg/min and 240  μg/kg/min, respectively. reported higher diastolic stress MBF values compared to sys-
Contrast agent was injected directly into the pulmonary vein, tolic stress MBF. However, diagnostic accuracy for myocar-
whereas in the studies of Bamberg et al. and Mahnken et al., the dial ischemia was similar for systolic and diastolic acquisition.
contrast agent was injected into a peripheral vein. Rossi et al. There are several important advantages for CT acquisition
[15] found higher MBF values than the studies of Bamberg during the systolic phase. Firstly, the heart is contracted during
et al. and Mahnken et al. 2.68 (2.31–2.81) ml/g/min compared systole, with maximal contraction at end systole, resulting in a
to 1.10(±0.25)  ml/g/min and 117.4(±18.6)  ml/100  ml/min in smaller total heart volume, in particular a shorter basal-apical
nonischemic myocardium during stress phase, possibly caused length. Thus, the range that needs to be covered for visualizing
by the differences in experimental setup. Apart from animal the entire heart within one scan cycle is reduced. Because of
validation studies, in recent years a number of patient studies the maximal contraction, the myocardium is thicker in systole
have been performed, evaluating the feasibility of CTMPI in a and allows for easier delineation during analysis. Secondly,
clinical setting. These studies will be discussed later in this although the systolic phase is shorter than the diastolic phase,
chapter, after discussing the technical issues in CTMPI. the systolic phase has a constant duration (200  ms approxi-
mately) independent of heart rate and is less sensitive to
arrhythmia. Thirdly, image acquisition during the systolic
Technical Information phase results in a lower-contrast dose in the left ventricle and,
thus, reduces beam-hardening artifacts [8, 18, 19].
Heart Coverage

The accuracy of dynamic CTMPI logically depends on the heart Radiation Dose
coverage of the available CT systems. The first systems used for
dynamic CTMPI in clinical patients were DSCT systems. Cardiac imaging is a major contributor to the average popula-
Second-generation DSCT scanners allow coverage of most of tion medical radiation exposure [7, 20]. Dynamic CTMPI is
the left ventricle (7.3 cm) using a shuttle mode in systolic phase. associated with a relatively high radiation dose since it requires
The shuttle mode consists of back-and-forth table movements multiple scan acquisitions during the first pass of contrast.
alternated with sequential scanning to cover the left ventricle in Effective radiation doses between 5 and 13 mSv, with an
two separate scans, which are later combined to reconstruct one average dose of 9.2  mSV, have been reported for dynamic
image. Third-generation DSCT systems have wider detectors CTMPI procedures (Tables 63.1 and 63.2). Cardiac patients
compared to second-­generation scanners, enabling whole heart often undergo multiple imaging procedures in their life,
imaging in systolic phase with a total range of 10.2 cm. Although increasing their cumulative radiation exposure [5, 6, 20].
the shuttle mode enables the coverage of a larger portion of the One particular advantage of CT, however, is that both coro-
heart, the movement between the two table positions lowers the nary stenosis and resulting myocardial ischemia can be eval-
temporal sampling rate, especially at high heart rates. A low uated with one imaging modality. In view of the significant
temporal sampling rate results in a decrease in information on reduction in radiation dose with newest CT systems for coro-
in- and outflow of contrast medium in the myocardium, the so- nary imaging, the total radiation dose of CTA plus CTMPI
called first-pass perfusion, and could thereby cause inaccurate does not need to exceed 10 mSv [39–41].
estimation of MBF values [16]. Another problem introduced by The effective radiation dose of CT perfusion procedures
the moving table positions is the possibility of motion artifacts, is within the range of nuclear perfusion imaging procedures.
decreasing the accuracy of the measurements. Single-­ tube SPECT perfusion studies show radiation doses of 6.6 mSv
multi-detector CT (MDCT) scanners with 256 or 320 detector for stress-only imaging and 11.3 mSv for both rest and stress
rows cover 8 and 16 cm, respectively. Wide detector coverage phase imaging. PET perfusion studies need a radiation rang-
enables (nearly) full left ventricle imaging with a stationary ing from 2.4 mSv to 13.5 mSv [42].
63  Dynamic Myocardial CT Perfusion Imaging 813

Table 63.1  Dynamic CT myocardial perfusion patient studies using visual analysis
Study Year Number of patients CT scanner Radiation dose (mSv) Reference Analysis Sens Spec
Yang [21] 2016 72 Second-generation DSCT 7.8 ICA + FFR Territory 79 91
Baxa [22] 2015 27 Second-generation DSCT 8.9 ICA Territory 97 95
Segmental 98 96
Weininger 2012 10 Second-generation DSCT 12.8 MRI Segmental 86 98
[23] SPECT Segmental 84 92
Wang [24] 2012 30 Second-generation DSCT 9.5 CCTA + SPECT Segment 100 76
SPECT
Bastarrika [25] 2010 10 Second-generation DSCT 18.8 MRI Segmental 86 98
Note: DSCT, dual-source CT; MDCT, multi-detector CT; ICA, invasive coronary angiography; FFR, fractional flow reserve; MRI, cardiac magnetic
resonance; SPECT, single-photon emission CT

Radiation dose reduction techniques in CTMPI show prom- Temporal Sampling Rate
ising results. Performing a scan only in stress instead of both in
rest and stress phase considerably reduces the radiation dose. A The temporal resolution of dynamic CTMPI needs to exceed
rest/stress phase combination, for example, in MR and PET the timescale of the fastest process observed; otherwise, the
perfusion imaging, is normally performed to acquire informa- perfusion parameters may be incorrect and most likely be
tion about the reversibility of a perfusion defect, which could underestimated. The fastest process in contrast medium kinet-
help to differentiate ischemia from myocardial infarction; see ics is typically the vascular transit time. For example, in
Fig. 63.1. However, in symptomatic patients without history of dynamic brain perfusion imaging, the minimal temporal sam-
myocardial infarction, the probability of a persistent perfusion pling rate is 2 s. Several articles imply that a low temporal sam-
defect is low. In these patients, the main reason for CT imaging pling rate in cardiac imaging, for example, in shuttle mode,
is to detect hemodynamically significant CAD, where presence leads to underestimation of the MBF in dynamic CTMPI [16,
of coronary stenosis is combined with evaluation of ischemia 46]. Temporal sampling rates of one acquisition every second
in the same vascular territory. Danad et al. [43] showed that the heartbeat for low heart rates, up to one acquisition per four
stress MBF value has a higher diagnostic accuracy than an heartbeats at high heart rates (which are common in stress situ-
index parameter comparing rest and stress acquisitions acquired ations), are reported using a DSCT system in shuttle mode [13,
using PET. This implies that a single MBF measurement dur- 16, 46]. These temporal sampling rates may be insufficient to
ing stress phase could be sufficient to evaluate the significance accurately capture the first-pass contrast enhancement curve.
of a coronary stenosis. Delayed enhancement scans are used to
detect myocardial scarring and also to determine the difference
between ischemic and infarcted myocardium in patients with Patient Preparation and Scanning Protocol
history of heart disease. Delayed enhancement CT falls outside
the scope of the current chapter. There are a number of points to consider in patient preparation.
Another option to reduce radiation dose is by lowering Patients should not use caffeine (coffee, tea, bananas, chocolate,
the tube voltage in patients with a normal body mass index etc.) for preferably 24 h or more before the examination, because
(BMI) from 100 kVp to 80 or 70 kVp [44]. This reduces of the interfering effects of caffeine on the effectiveness of the
the radiation dose up to 40% compared to 100 kVp, where stressor agent. During the CCTA, beta-­blockers may be needed
100  kVp can be reserved for patients with a BMI higher to lower the heart rate in order to optimize image quality. Some
than 25  kg/m [2, 44]. Kim et  al. [39] showed that auto- studies have found that the use of beta-blockers may have a
matic dose-­ modulation techniques combined with a negative effect on the detection of myocardial ischemia by
decreased scan duration during the first pass limit the radi- increasing the diastolic perfusion time [47, 48]. Sublingual
ation dose significantly without compromising the image nitrates, also commonly used in CCTA, have been shown to
quality. Of course, when decreasing scan duration, it is decrease the ischemic area on perfusion images [47].
important that the entire upslope of the contrast at first At this time, CTMPI is not yet clinically used apart from
pass is acquired. Iterative reconstruction techniques can be certain centers in Asia. The exact position of CTMPI in the
used to compensate the loss of image quality when using a work-up of CAD is still under investigation. When consider-
lower-tube current [45]. ing implementing CCTA with CTMPI, it is still a point of
Despite the developments in dose reduction techniques, debate whether CCTA or CTMPI should be the first in the
CTMPI procedures still yield a relatively high radiation examination order. In patients with a high probability or
exposure compared to other CT examinations. Patient-­ known CAD (after stenting or bypass grafting) in whom
tailored CTMPI protocols are essential to reduce unneces- myocardial ischemia is likely, CCTA can be performed after
sary radiation exposure. the stress CTMPI procedure; see Fig. 63.2. This eliminates
Table 63.2  Dynamic CT myocardial perfusion patient studies using (semi)quantitative analysis
814

Number of Radiation dose Cutoff

Study Year patients CT scanner (mSv) Reference Analysis Parameter Model used Sens Spec (ml/100 ml/min)
Tanabe [26] 2016 53 256 row MDCT 10.5 SPECT Segment MBF Singular value decomposition 80 86 0.92 mL/g/min
MRI 82 87 0.98 mL/g/min
Wichmann 2016 71 Second-generation 8.2 Visual CT Patient MBF Two-compartment model + 100 100 88
[27] DSCT (3 territories) upslope (Siemens)
MBF 81 91 15
(2 territories)
MBF 76 87 109
(1 territory)
MBV 100 88 16
(3 territories)
MBV 75 91 136
(2 territories)
MBV 86 45 20
(1 territory)
Wichmann 2015 137 Second-generation 643.8 (mGy. CCTA Territory MBF Patlak (Siemens) 82 81 103
[28] DSCT Cm) MBF ratio 91 93 0.71
Bamberg [29] 2014 31 Second-generation 11.1 MRI Territory MBF Patlak (Siemens) 100 75 88
DSCT Segment 78 75
Patient 100 75
Ebersberger 2014 37 Second-generation 9.6 SPECT Segment MBF Patlak (Siemens) 86 96 Individual
[30] DSCT thresholds
MBV 81 96
Kono [31] 2014 42 Second-generation 9.4 ICA + FFR Territory MBF Patlak (Siemens) 89 48 103
DSCT MBF ratio 98 70 0.85
Rossi [32] 2013 80 Second-generation 9.4 ICA Territory MBF Patlak (Siemens) 88 90 78
DSCT Patient 90 88
Huber [33] 2013 32 256 row MDCT 9.5 ICA + FFR Territory MBF Linear fit+upslope method 76 100 1.64 ml/g/min
Upslope – 83 88
Greif [34] 2013 65 Second-generation 9.7 ICA + FFR Territory MBF Patlak (Siemens) 95 74 75
DSCT Patient 97 66
So [35] 2012 26 64-MDCT 9.7 SPECT MPR Distributed parameter model 95 35 2.5 ml/g/min
MVR 97 24 1.4
Bamberg [36] 2011 33 Second-generation 10 ICA + FFR Territory MBF Patlak (Siemens) 93 87 75
DSCT Segment 91 98
Patient 95 64
Ho [37] 2010 35 Second-generation 9.2 ICA Segment MBF Patlak (Siemens) 83 78 –
DSCT
Kido [38] 2008 14 16- MDCT – QCA Territory MBF Patlak plot analysis 73 81 1.5 ml/g/min
SPECT 88 79
Note: DSCT, dual-source CT; MDCT, multi-detector CT; CCTA, coronary CT angiography; ICA, invasive coronary angiography; FFR, fractional flow reserve; QCA, quantitative coronary angiog-
raphy; MRI, cardiac magnetic resonance; SPECT, single-photon emission CT; MBF, myocardial blood flow; MBV, myocardial blood volume; MPR, myocardial perfusion reserve; MVR, myocar-
M. van Assen et al.

dial volume reserve

63  Dynamic Myocardial CT Perfusion Imaging 815

the influence of beta-blockers and nitroglycerin on the stress lower maximum heart rate during stress, which can lead to
CTMPI acquisition. In view of the relatively high radiation higher temporal ­sampling rate.
dose of CTMPI, in less-than-high probability patients it Compared to normal CCTA equipment, an additional
would be preferable to start with CCTA, in order to avoid infusion pump may be needed for the administration of the
unnecessary radiation dose in the case of no stenosis. This, stressor agent and an additional intravenous catheter unless
however, would require direct reading of CCTA, which may contrast medium and stressor are administered into the same
not be logistically feasible. An advantage could be that beta-­ arm such as in the case of regadenoson. Patients should be
blockers given for CCTA may, if CTMPI follows, result in observed continuously during the CTMPI procedure with a
12-lead ECG and a blood-pressure monitor.
An optional, delayed acquisition (10–15  min after last
Rest Stress
contrast bolus) can be considered in order to differentiate
between ischemic and infarcted myocardium in patients with
known CAD.  With a delayed enhancement (DE) scan,
infarcted myocardium can be identified owing to the delayed

SPECT
in- and outflow of contrast of infarcted myocardium. This
results in increased HU values in DE imaging. For this DE
scan, administration of contrast medium beyond that admin-
istered during the stress or rest phase is not needed. However,
DE scan acquisition might not be needed for the differentia-
tion between ischemia and infarction. Bamberg et  al. [29]
showed that myocardial blood volume (MBV) is decreased
in infarcted myocardium compared to ischemic myocar-
MRI

dium, while MBF is reduced in both conditions. Also, actual

MBF cutoff values may allow distinguishing between isch-
emia and infarction, as MBF in infarcted segments was
found to be lower than in merely ischemic segments accord-
ing to a study comparing CTMPI to MRI and SPECT [26].

Image Analysis

Analysis of dynamic CTMPI images is based on the distribu-

CT

tion of contrast medium throughout the myocardium. The

distribution of iodine contrast medium can be described in
two time-intensity curves. The tissue attenuation curve
(TAC) describes the concentration of contrast medium in the
myocardium over time, and the arterial input function (AIF)
Fig. 63.1  Rest and stress perfusion images made by SPECT and
describes the concentration of contrast medium in the sup-
MRI. For dynamic CTMPI stress imaging only could be sufficient plying artery (assessed at aorta or left ventricle).
Assessment of myocardial perfusion can be performed
visually, semiquantitatively, or quantitatively. It should be

Fig. 63.2 Schematic
representation of complete
IV acces,

minutes

minutes

Coronary CT Delayed
ECG

cardiac evaluation protocol,

5-15
3-5

Scout Image Stress Phase

including anatomical (CCTA) Angiography Phase
and functional (dynamic
CTMPI) evaluation. A third,
optional scan is the delayed
enhancement scan for the Stressor agent
detection of infarcted
Test Contrast Contrast
myocardium Bolus agent agent
Image Image Image Image
acquisition acquisition acquisition acquisition
816 M. van Assen et al.

cavity, and apical level. The basal and mid-cavity slices are
a
each divided into six equal segments, whereas the apical
slice is divided into four equal segments. The apex (segment
17) is often not taken into account for perfusion analysis
because the limited scan width of most scanners does not
allow the apex to be imaged [49]. Segmental CT perfusion
analysis can be compared to other perfusion modalities such
as MRI, PET, or SPECT. Perfusion analysis of vessel territo-
ries is based on a three-vessel division, namely, the left
descending artery (LAD), the right coronary artery (RCA),
and the circumflex artery (LCx) territory. Perfusion parame-
ters are calculated per segment, after which multiple seg-
ments are averaged to represent a territory. Segments 1, 2, 7,
8, 13, and 14 are assigned to the LAD territory; segments 3,
4, 9, 10, and 15 are assigned to RCA territory; and segments
b 5, 6, 11, 12, and 16 are generally assigned to the LCx terri-
Ischemic segment tory, depending on coronary dominance [49]. Anatomic vari-
ations in supplying arteries, however, may pose a problem.
Territory analysis can be additionally compared with meth-
ods which analyze CAD severity on vessel level such as ICA
and FFR methods. With nuclear modalities and MRI, it is
also possible to ­analyze perfusion on a territory level.
For clinical purposes it is important to know whether a
patient has myocardial ischemia and is in need of an inter-
Non-Ischemic segment vention. Therefore, per-patient analysis is important for clin-
140 ical diagnostics, while segment- and territory-based analyses
Non-ischemic segment
120 Ischemic segment are mostly used to validate the technique.
Enhancement [HU]

100
80
Qualitative, Visual Analysis
60
40 Qualitative analysis of dynamic CTMPI data can be done by
20 visually inspecting the enhancement of myocardial tissue during
0 the first pass of iodine contrast on dynamic series. Myocardium
1 3 5 8 10 12 14 16 18 21 23 25 27 29 31 34 36 38
Time [Seconds]
with reduced perfusion is hypo-attenuated and enhances later
compared to normally perfused myocardium on CTMPI scans.
Hypo-attenuation can in principle indicate both ischemic and
infarcted myocardium. Visual analysis of dynamic CTMPI
images is often done by analysis of color-­ coded maps; see
Fig. 63.3 (a) Color-coded perfusion map of myocardial blood flow.
(b) The color-coded map is based on quantitative data. Regions of inter- Fig. 63.4. It is important to realize that these color-coded maps
est (white circles) represent a tissue attenuation curve (TAC) specific are actually based on quantitative information of myocardial per-
for that area fusion. In contrast to SPECT, which primarily allows evaluation
of relative perfusion of myocardial areas within a patient, the
noted that, in contrast to static CTMPI, visual analysis of color map based on dynamic CTMPI represents actual MBF val-
dynamic CTMPI is often based on quantitative data, by ues, which allows ­assessment of globally reduced perfusion in
­creating color maps based on MBF data; see Fig. 63.3. In this three-vessel disease [50]. Below we describe the quantification
chapter qualitative analysis refers to visual analysis of of the underlying perfusion parameters.
CTMPI data without the use of any quantitative measure and
may include the visual assessment of color-coded maps.
Myocardial perfusion can be assessed on a segment, terri- S  emiquantitative Analysis
tory, or patient basis. For segmental analysis, the 16-segment
American Heart Association model is recommended [49]. Semiquantitative parameters can be derived from the TAC
This segmentation uses three imaging slices at basal, mid-­ curves. The upslope method, in which the upslope of the TAC
63  Dynamic Myocardial CT Perfusion Imaging 817

a b c

Fig. 63.4 (a) Mid ventricular slice of a dynamic CTMPI of a porcine cardium software (MMWP VA41A, Siemens). The AHA segmentation
heart with corresponding color-coded map representing (b) myocardial is projected onto the CT images. The red arrows point two the perfusion
blood flow (MBF) in ml/100/ml/min and (c) myocardial blood volume defects present in both hearts corresponding to both a decrease in MBF
(MBV) in ml/100 constructed with Volume Perfusion CT (VPCT) myo- as in MBV

curve is taken as an indirect measure of perfusion, is a popular Quantitative Analysis

method to analyze CTMPI data. The upslope is calculated by
making a linear fit of the upslope of the TAC curve. The main There are several methods to perform true quantitative
advantage of using the upslope is the possibility of shortening analysis of perfusion data, based on the AIF and TAC; see
the scan time and thereby reducing the radiation dose since Fig. 63.6. Of these methods, the model-dependent decon-
only knowledge of the wash in of contrast medium is required. volution method is most frequently used in recent litera-
For the upslope method, timing of the scan window is highly ture on cardiac and brain MRI perfusion analysis [51, 55].
important. When the entire upslope of the curve is not included, Considering that the contrast media currently used in MRI
the upslope method becomes inaccurate. (gadolinium) and CT (iodine) behave according to the
With the upslope method, MBF can be estimated with the same kinetic principles, the same approach can be used
following equation: [51, 55, 56].
Quantitative CT perfusion analysis using model-­
Maxupslope ( TAC )
MBF = (63.1) dependent convolution can be divided in two phases. First,
Maximum ( AIF ) the signal-time curves (AIF and TAC) should be transformed

into iodine concentration-time curves [10, 51]. In compari-
where the maximum upslope of the TAC (Maxupslope(TAC)) son with MRI, this is relatively easy for CT data since the
is divided by the maximum value of the AIF curve change in HU values is linearly related to the iodine concen-
(Maximum(AIF)); see Fig. 63.5. tration [10].
The upslope method accurately estimates MBF if the Thus, the iodine concentration is proportional to the sig-
maximum slope of the TAC curve is within the mean tissue nal enhancement:
transit time, the time that a certain blood volume spends in
the myocardium. In stress phase the mean tissue transit time c ( t ) = k * ( HU ( t ) - HU 0 ) (63.2)

is decreased, more so in normal myocardium than in isch-
emic myocardium. This decrease can cause an underestima- In this formula, c(t) is the iodine concentration over time,
tion of MBF [51]. Another issue arising with the use of the and HU0 is the baseline HU value (i.e., before iodine
upslope method is the inaccuracy of the AIF and TAC in the ­injection). K is an unknown scale constant that is automati-
case of low temporal sampling rates. When temporal sam- cally corrected for during the second phase, assuming that k
pling rate is too low, the AIF and TAC consist of only a few is tissue independent [51].
measurement points during upslope, and important charac- Second, a model-dependent deconvolution approach can
teristics of both AIF and TAC can be missed. be used to describe the perfusion process in the myocar-
Other semiquantitative parameters are peak enhancement, dium. This convolution theory-based approach is similar to
time to peak, and area under the curve, all derived from the deconvolution methods and models used in cardiac MRI
TAC. From these semiquantitative parameters, the upslope is studies [51, 57]. The iodine concentration in the myocar-
the most commonly used in MRI studies on semiquantitative dium over time is related to the iodine concentration in the
analysis of myocardial perfusion [52–54]. supplying artery, convoluted (⊗) by an impulse response
818 M. van Assen et al.

Fig. 63.5  The left figure shows AIF and TAC

the arterial input function (AIF) 700
and the tissue attenuation curves AIF
(TAC) of a nonischemic segment
and an ischemic segment. The 600
ischemic segment shows a max AIF TAC non-ischemic
decreased inflow of contrast
compared to the nonischemic 500 TAC ischemic
segments. On the right is an

HU-values
extended graph showing the TAC
in more detail. The maximum 400
AIF value and maximum upslope
of the TAC curve, used to
calculate the myocardial blood 300
flow, are indicated by black lines

200

100

0
8 10 12 14 16 18 21 23 25 27 29 31 34
Time (sec)

AIF and TAC

50
max upslope TAC
45 TAC non-ischemic

40
TAC ischemic
35

30
HU-values

25

20

15

10

0
8 10 12 14 16 18 21 23 25 27 29 31 34

Time (sec)

function (IRF). This relation is described by the following If IRF(t) is known, the TAC can be obtained as a summa-
equation [55, 56, 58]: tion of adjusted IRFs; see Fig.  63.7. In dynamic CTMPI
AIF and TAC are known parameters, measured from the CT
TAC ( t ) = MBF * AIF ( t ) Ä IRF ( t ) (63.3)
images, whereas IRF(t) is the unknown parameter. The
reversed process to reconstruct the IRF(t) is called decon-
where TAC(t) is the tissue attenuation curve over time (sec- volution. The main issue with a deconvolution approach is
onds) in HU values, MBF is the myocardial blood flow, that in contrast with convolution, deconvolution cannot
AIF(t) is the arterial input function over time in HU values, give a unique solution because there are multiple IRF(t)
and IRF(t) is the impulse response function over time. estimations possible that would result in the same TAC(t)
 63  Dynamic Myocardial CT Perfusion Imaging 819

a c

b 140
Non-ischemic segment
d 140
Non-ischemic segment
120 Ischemic segment 120 Ischemic segment
Enhancement [HU]

100 100
Enhancement [HU]
80 80

60 60

40 40

20 20

0 0
1 3 5 8 10 12 14 16 18 21 23 25 27 29 31 34 36 38 1 3 5 8 10 12 14 16 18 21 23 25 27 29 31 34 36 38
Time [Seconds] Time [Seconds]

Fig. 63.6 (a–c) both show a midventricular slice of a dynamic CTMPI (green) and of the ischemic segment with a perfusion defect (red) dur-
scan with a segmental overlay. A perfusion defect is visible in both ing multiple time points, where the TAC of the ischemic segments in
scans (red arrows). (b–d) give the corresponding tissue attenuation both scans shows a decreased enhancement
curve (TAC) of a nonischemic segment without the perfusion defect

or even a better approximation of the TAC(t). This problem Tracer-Kinetic Models

is solved in model-dependent deconvolution by determin- A wide variety of tracer-kinetic models can be used to repre-
ing a generic model to represent the IRF(t) and setting pre- sent IRF(t). Each of these models has its merits and limita-
defined boundaries for the estimated parameters describing tions; the optimal model for CTMPI analysis has yet to be
this model [55]. determined.
In the model-dependent deconvolution method, a The microcirculation of the myocardium is depicted in
tracer-­kinetic model is assumed to represent IRF(t), after Fig.  63.8a. Iodine is considered an extravasating contrast
which the model parameters can be optimized to best fit medium. The contrast medium distributes across the intra-
Eq. (63.3) to the measured TAC data [51, 55, 56]. From vascular space and the extracellular extravascular space,
the fitted model of the estimated IRF(t), the MBF value both defined by volume and transit time parameters. High-­
can be derived using Eq. 63.3, with known TAC(t), AIF(t), order perfusion models try to describe the complexity of
and IRF(t). these dynamics. The two-compartment model and the dis-
Convolution and deconvolution techniques are difficult tributed parameter model, both using four free parameters,
because noise in either the AIF or TAC data can influence the are examples of high-order perfusion models.
model optimization and result in an unstable solution for When only limited data are available or the quality of the data
IRF(t) and unreliable MBF values [58]. Noise in CTMPI is low, it becomes difficult to accurately assess all parameters in
data can be caused, for example, by inaccurate HU values, high-order models. Simplified models with fewer free parame-
measured during different cardiac phases. Reducing noise is ters are useful in these situations, for example, the extended Toft
therefore an important issue in dynamic CTMPI. model (three free parameters). These tracer-kinetic models fix
820 M. van Assen et al.

a Arterial Input Function Tissue Attenuation Curve

HU

HU
Time (t) Time (t)
b
HU

HU

Time (t) Time (t)

c
HU

HU

Time (t) Time (t)

Fig. 63.7 (a) When two identical bolus injections of the same concen- transit time. (b) When the bolus injections have different properties, for
tration (left) are given, then the IRF (right) will be the same for each example, different concentrations (left), the corresponding IRFs (right)
injection. For each bolus, the IRF shows an abrupt increase in HU values will be different. (c) Contrast medium inflow represented as a series of
(if the injection is given directly to arterial input); it then stabilizes for a bolus injections of different concentrations (left) and corresponding IRFs
period of time while the bolus passes through the tissue and finally shows (right). The tissue attenuation curve (TAC) is the sum of all IRFs corre-
a gradual return back to baseline level. The plateau represents the mean sponding to the bolus injections attenuated by the effects of blood flow
63  Dynamic Myocardial CT Perfusion Imaging 821

PS

Vp Ve

F
F
b c

Vp Ve
PS

Vp Ve PS
Vp Ve

F F

d e

PS

Vp Ve Vp

Fig. 63.8 Schematic representation of tracer-kinetic models. (a) model. The Vp and Ve spaces are represented by multiple small com-
Schematic representation of the myocardial microcirculation. The partments. The contrast medium can only exchange between neighbor-
blood plasma flows to the vascular space (Vp) driving the myocardial ing compartments. (d) Schematics of the extended Toft model, where
blood flow (F). The Vp compartment exchanges molecules via a flow infinite flow (F) is assumed. (e) Schematics of the Fermi model. This
(PS) with the extracellular extravascular space (Ve). (b) Schematics of mathematical model has only nonphysiological parameters and only the
the two-compartment model. The Vp space can only exchange contrast flow (F) can be derived
medium with the Ve space. (c) Schematics of the distributed parameter
822 M. van Assen et al.

one (or more) parameters at a constant value resulting in an accu- t­wo-compartment model combined with the upslope method.
rate parameter estimation with less free parameters. In the first phase, the TAC curve is reconstructed from the indi-
vidual measurements of iodine concentration in the myocar-
Two-Compartment Model dium over time, using a convolution approach. The Patlak
The two-compartment model describes the intravascular and method uses a least square fit method to fit a two-­compartment
the extracellular extravascular space as two compartments; model to the TAC curve. Subsequently the MBF is calculated
see Fig.  63.8b. This model does not take into account the using the upslope method; see Eq. 63.1. The maximal upslope
transit times; therefore, the IRF of a two-compartment model can be derived from the IRF(t) function, describing the TAC
does not have a stable plateau as the IRFs in Fig. 63.7. The curve. Because of the convolution approach to estimate an
transit time parameters represent the time a specific amount equation (IRF) describing the TAC, this method is ideal for CT
of blood volume is present in the tissue or capillaries. The data with low temporal sampling rates where the use of only an
peak value of the IRF corresponds to the volume transfer upslope method results in inaccurate MBF values because of
coefficient Ktrans. This parameter is defined as a product of the limited information on TAC and AIF curves. An example of
myocardial blood flow (MBF) and extraction fraction (E) a low temporal sampling scan mode is the ECG-triggered shut-
and represents the inflow into the extravascular extracellular tle mode with a temporal sampling rate of 2–3 s [51].
space and thereby the delivery of nutrients to tissue. This method substantially simplifies the mathematical
procedures of a model-dependent deconvolution approach.
Patlak Method The Volume Perfusion CT (VPCT) myocardium software
This method is used in the majority of dynamic CTMPI patient (MMWP VA41A, Siemens) adopted this method to calculate
studies (see Table  63.2) and is a hybrid method based on a the MBF in dynamic CTMPI data (Fig. 63.9). Although this

Fig. 63.9  An overview of the Volume Perfusion CT (VPCT) myocardium software (MMWP VA41A, Siemens). Multiple windows allow to visu-
alize different CT axis, show the signal intensity curves for the drawn regions of interest and give corresponding quantitative results
63  Dynamic Myocardial CT Perfusion Imaging 823

method gives a good estimation and is able to distinguish ­ athematical model providing only a functional representa-
m
ischemic myocardium from normal myocardium, MBF is tion of an IRF.  This model does not allow a physiological
substantially underestimated [16]. interpretation of the parameters used in the model; the
More accurate MBF values could be obtained by using parameters are simply used as shaping parameters; see
Eq.  63.3 to calculate the MBF directly instead of using a Fig. 63.8e. However, the flow (MBF) can still be estimated
hybrid approach with the upslope method (Eq. 63.1). However, by the above-described model-dependent deconvolution
this comes at the cost of higher computational complexity. technique [51, 55]. This model is successfully used in studies
on MRI analysis of myocardium perfusion [10, 59–62].
Distributed Parameter Model
The distributed parameter model is one of the more complex
models, taking into account all aspects of contrast medium Diagnostic Accuracy
kinetics. Contrast medium is assumed to exchange between
spaces. This model estimates volume, flow, and transit time Only a limited number of patient studies (n = 18) has been
parameters, providing a full description of the perfusion pro- published on dynamic CTMPI including a total of 805
cess [51]. In comparison with other models, the distributed patients. They are mostly small single-center studies with
parameter represents both the extracellular extravascular and large inter-study heterogeneity in protocols, scanner type,
the vascular space as a series of compartments, Fig. 63.8c. stressor agent, reference value, analysis method, and cutoff
Each extracellular extravascular space compartment inter- value, making it difficult to compare results.
acts only with the nearest vascular space compartment and
vice versa. The use of the distributed parameter model is lim-
ited by the temporal sampling rate of the CT system. If the Visual Analysis
temporal sampling rate is too low, accurate estimation of the
mean transit times becomes impossible. With higher perfu- Five studies used visual analysis to evaluate dynamic CTMPI
sion flow, the mean transit time decreases, requiring an even data, including a total of 149 patients with a median of 27
higher temporal sampling rate and a compact contrast patients per study. These dynamic CTMPI publications using
medium bolus. Because of the complexity of the distributed visual analysis are listed in Table 63.1. Sensitivity was found
parameter model, complex fitting methods are required, to range from 84% to 98% and specificity from 76% to 98%.
making the system more susceptible to errors [55]. So et al. All studies with visual analysis of dynamic CT data so far
[35] used the distributed parameter model to calculate myo- used a second-generation DSCT system.
cardial blood flow using 64-MDCT system. The myocardial Two studies analyzed individual segments and two stud-
perfusion ratio, a ratio between normal and remote myocar- ies analyzed vessel territories. Baxa et al. [22] analyzed both
dium, had 95% sensitivity and 35% specificity to identify segments and territories. They showed 97% sensitivity and
ischemic myocardium, with SPECT as reference method. 95% specificity for territory-based analysis, and 98% sensi-
tivity and 96% specificity for segment-based analysis, com-
Extended Toft Model paring visual analysis of CTMPI data with ICA. None of the
The extended Toft model has one fixed parameter and three studies performed with visual analysis of dynamic CTMPI
free parameters. It is a simplified variation of the two-­ data assessed diagnostic accuracy on a per-patient level.
compartment model and the distributed parameter model by Two visual analysis studies document interobserver agree-
assuming an infinite flow; see Fig. 63.8d. This means that the ment between dynamic CTMPI and other perfusion modalities.
MBF cannot be measured with this model; instead the vol- Weininger et  al. [23] compared visual analysis of dynamic
ume transfer coefficient Ktrans is obtained. Although describ- CTMPI to SPECT and MRI perfusion and obtained similar
ing a key part of the perfusion process, this model cannot be results. Interobserver agreement between CTMPI and MRI and
used to measure flow and should therefore not be compared CTMPI and SPECT was high with kappa values of 0.85 and
to flow-measuring models. However, in situations where the 0.82, respectively. Wang et al. [24] compared visual analysis of
flow cannot be measured accurately due to low temporal dynamic CTMPI analysis to SPECT and found high interob-
sampling rate, the extended Toft model can provide an alter- server agreement for detecting perfusion defects with CTMPI
native direct proxy measure for perfusion instead of MBF and SPECT with kappa values of 0.81 and 0.83, respectively.
[51, 55]. So far this model has not been applied in any pub-
lished dynamic CTMPI study in patients.
Semiquantitative Analysis
Fermi Model
The Fermi model assumes that the contrast medium does Huber et al. [33] is the only study using a semiquantitative
not  leave the intravascular space. The Fermi model is a parameter (upslope), and they compared it with a ­quantitative
824 M. van Assen et al.

parameter (linear fit+upslope method) using a 256 row p­ erfusion study that gender, age, and weight influence MBF
MDCT system with ICA and FFR measurements as a refer- values and that reference MBF values vary significantly
ence standard. The quantitative measured MBF yielded a within the general population. These results indicate that
sensitivity and specificity of 76% and 100%, while the semi- using a general cutoff value for perfusion parameters may be
quantitative upslope measure resulted in a reported sensitiv- sub-optimal.
ity and specificity of 83% and 88%, respectively.
 bsolute vs. Relative MBF Values
A
As mentioned previously, a wide variety of absolute MBF
Quantitative Analysis cutoff values have been reported for discriminating ischemic
from nonischemic myocardium, as well as a high heteroge-
A total of 13 dynamic CTMPI reports with quantitative anal- neity of MBF values in the normal population. This issue
ysis are shown in Table  63.2. Included were 656 patients could be avoided by using a relative instead of an absolute
with a median of 37 patients per study. Respective sensitivity measure of MBF.  A relative measure comparing normally
and specificity of MBF cutoffs for myocardial ischemia perfused myocardium with ischemic myocardium may be
ranged from 73% to 100% and 48% to 100% for MBF. For more accurate for the identification of myocardial ischemia
MBV, sensitivity varied between 75% and 100% and speci- in the individual patient. However, absolute values offer
ficity between 24% and 91%. Reference techniques include advantage in the diagnosis of global ischemia when normally
SPECT, MRI, and ICA  +  FFR.  It should be noted that of perfused myocardium is absent.
these reference techniques only MRI is able to provide a Wichmann et  al. [28] compared absolute MBF values
quantitative measure of MBF and MBV values. In studies of and relative MBF values to CCTA results in 137 patients
Tanabe et al. [26], Huber et al. [33], So et al. [35], and Kido using a second-generation DSCT.  Relative MBF values
et al. [38], a MDCT system was used; in all other studies, a yielded a higher diagnostic accuracy than absolute MBF
second-generation DSCT system in shuttle mode was used. values on a territory level, with a sensitivity and specificity
of 91% and 93% compared to 82% and 81%. Kono et al.
Segment, Territory, and Patient [31] reported similar results in 42 patients using a second-
Bamberg et al. [29, 36] determined the diagnostic accuracy generation DSCT, comparing relative MBF values and
of MBF on segment-, territory-, and patient-based analysis in absolute MBF values to combined ICA and FFR results.
two separate studies, compared to, respectively, MRI and They reported 98% sensitivity and 70% specificity for rela-
FFR. Segment-based analysis showed sensitivity of 78% and tive MBF values compared to 89% sensitivity and 48%
91%, respectively, compared to 100% and 93% on territory specificity for absolute MBF values. Both studies used a
basis and 100% and 95% on patient basis. Rossi et al. [32] two-compartment model combined with upslope analysis
analyzed their MBF data on both per-territory and per-patient to calculate MBF.
level resulting in 88–90% sensitivity and specificity, com-
pared to ICA alone. Myocardial Blood Volume
Myocardial blood volume (MBV) could be another parame-
Cutoff Values ter for the detection of myocardial perfusion defects. In
A wide range of cutoff values (75–103  ml/100  ml/min for ­ischemic myocardium vasodilation of the arterioles compen-
MBF) has been proposed to distinguish ischemic from non- sates for the decreased flow in the stenotic artery, thereby
ischemic myocardial segments. Studies using DSCT com- changing the volume of blood in the myocardium. Bamberg
bined with VPCT software have reported cutoff values in et  al. [29] showed that MBV values are lower in infarcted
ml/100  ml/min, while studies using MDCT have reported myocardium compared to ischemic myocardium and could
cutoff values in ml/g/min. Ebersberger et al. [30] used indi- therefore help to differentiate between the two states. Three
vidual cutoff values instead of a generic MBF cutoff value studies evaluated the diagnostic accuracy of both MBF and
and reported a sensitivity of 86% and a specificity of 96% MBV to detect ischemic myocardium. So et  al. [35] used
with SPECT as reference modality. The cutoff values were myocardial perfusion reserve (MPR) and myocardial volume
calculated by subtracting the standard deviation of all seg- reserve (MVR) to detect ischemic areas with SPECT as ref-
ments from the average value of the respective measurement. erence. MPR and MVR were defined as the ratio of MBV
Kim et  al. [63] assessed the range of CTMPI-determined and MBF values in stress and rest imaging. They reported
MBF in 19 healthy volunteers in rest and in stress using 128-­ sensitivity and specificity of 95% and 35% using MPR, and
slice DSCT.  Results showed considerable heterogeneity in 97% and 24% using MVR, compared to SPECT results.
absolute MBF values. Women had higher MBF values at rest Ebersberger et al. [30] determined absolute MBF and MBV
compared to men but had lower MBF values during stress values and reported a slightly lower sensitivity for MBV than
imaging. Danad et  al. [64] showed in a myocardial PET for MBF (81% compared to 86%) with comparable specific-
63  Dynamic Myocardial CT Perfusion Imaging 825

ity, comparing dynamic CTMPI results to SPECT. Wichmann Comparison of Static and Dynamic CTMPI
et al. [27] also determined absolute MBF and MBV values,
analyzing one-, two-, or three-vessel territories and com- In an animal study, Schwarz et  al. [65] concluded that
pared quantitative analysis of the CTMPI data to visual anal- dynamic CTMPI may be more sensitive for detection of
ysis of perfusion defects of the same CTMPI data. Global smaller perfusion defects compared to static
MBV values showed lower specificity compared to MBF CTMPI. However, Huber et al. [33] showed that diagnostic
values. Bamberg et al. [36] compared MBF and MBV values accuracy of the dynamic CTMPI parameter was similar to
to combined ICA and FFR measurements. They reported that that derived from static CTMPI data using combined ICA
MBF had a significantly higher discriminatory power than and FFR measurements as a reference. More studies are
MBV to detect myocardial ischemia. The combination of needed to investigate in which patients dynamic CTMPI has
MBF and MBV values was found to be useful in discriminat- additional value beyond visual analysis of myocardial isch-
ing ischemic and infarcted myocardium from normal emia based on myocardial blood supply imaging.
myocardium.

Comparison with Other Modalities

 omparison of Visual, Semiquantitative,
C
and Quantitative Analysis Currently there are no patient studies comparing dynamic
CTMPI and other ischemia imaging modalities against the
Huber et al. [33] showed a similar diagnostic accuracy for same reference standard.
semiquantitative parameters compared to a quantitative
measure of MBF in dynamic CTMPI, with combined ICA
and FFR results as a reference. The reported sensitivity Future Perspectives
and specificity were 83% and 88% for the semiquantitative
upslope parameter CTMPI, respectively, and 76% and The limited number of dynamic CTMPI studies in patients
100% for the quantitative MBF parameter. The range of shows promising results with regard to the diagnostic value
reported specificities among studies is larger for quantita- of dynamic CT for the detection of myocardial ischemia.
tive analysis, indicating that diagnostic accuracy for a However, several issues need to be addressed before clinical
quantitative approach is less robust. However, quantifica- implementation of this technique can be considered.
tion of myocardial perfusion may be useful in diagnosing The currently published patient studies are difficult to
three-vessel disease, where there is no nonischemic myo- compare due to heterogeneity in imaging protocols, refer-
cardium present as reference and in the case of global ence standards, and analysis techniques. An optimal and
hypoperfusion of the left ventricular myocardium. Three- robust protocol for dynamic CTMPI is yet to be determined.
vessel disease and global hypoperfusion are known to The main advantage of dynamic CTMPI compared to
cause false-negative results in SPECT.  Meinel et  al. [50] other perfusion imaging techniques is the possibility to truly
investigated whether quantification of global MBF is fea- quantify myocardial blood flow, making it possible to
sible and showed that global MBF is gradually lower in ­accurately identify perfusion defects even in the absence of
patients with increasing territorial perfusion defects and is normally perfused myocardium. Absolute perfusion values
correlated to the number of obstructed vessels. Global offer the possibility of improving CAD risk stratification,
MBF showed a moderate correlation with visual CTMPI diagnosing multivessel CAD, discriminating between
assessment and CCTA findings [50]. A study of Vliegenthart infarcted and ischemic myocardium, and identifying early
et  al. [9] investigated whether absolute global perfusion subclinically reduced myocardial perfusion such as in micro-
parameters as MBF and MBV could detect subclinical vascular disease.
changes in perfusion parameters in patients with hyperten- Although the quantification of perfusion parameters is
sion and diabetes. This offers the opportunity to use abso- assumed to have multiple advantages compared to visual
lute MBF values for the risk stratification of CAD prior to analysis, it is yet to be proven that more accurate MBF (or
the presence of evident myocardial ischemia. The diagno- MBV) values aid in the diagnosis of myocardial ischemia
sis of microvascular disease is another process that could and are able to improve patient outcome. Dynamic CTMPI
benefit from absolute quantification of MBF with dynamic can be analyzed visually by looking at the dynamic series
CTMPI.  Microvascular disease is characterized by a and quantitatively by either looking at color-coded maps
decrease in (global) perfusion without correlation to an based on absolute values or by analyzing the absolute perfu-
anatomical abnormality of the coronary arteries. Dynamic sion parameters directly. Until now there is no study compar-
CTMPI combined with CTA could possibly be used to ing the use of color-coded maps with absolute values to
diagnose microvascular disease and exclude CAD. identify ischemic myocardium. The question remains which
826 M. van Assen et al.

of the analysis methods yields highest diagnostic accuracy the multicenter SPECIFIC study is to determine the diagnos-
and best clinical feasibility. tic accuracy of CTMPI for the detection of hemodynamically
A disadvantage of dynamic CTMPI compared to other relevant coronary stenosis in patients with suspected or
modalities is the radiation dose, especially when CCTA and known CAD. A subgroup of patients will also undergo per-
dynamic CTMPI are combined. Patient-tailored protocols to fusion MRI. The diagnostic accuracy of MRI versus dynamic
reduce radiation dose should be developed. This could be CTMPI for the detection of perfusion defects will be com-
done by patient-specific tube current and kV modulation and pared to the reference standard, ICA plus FFR.
iterative image reconstruction. Static CTMPI is unable to In conclusion, the few patient studies focusing on dynamic
quantify MBF but has a lower radiation dose than dynamic CTMPI for myocardial ischemia detection show promising
CTMPI.  Whether the increased radiation dose of dynamic results. Absolute quantification of perfusion parameters offers
CTMPI weighs up to the benefits of absolute quantification, great potential, not only in the diagnosis of myocardial isch-
and in which patients, is yet to be determined. emia but potentially also in the detection of early signs of
Although the main benefit of dynamic CTMPI is the pos- reduced myocardial blood flow as well as the diagnosis of
sibility to quantify perfusion, several issues arise with the microvascular disease and three-vessel disease. With the advent
use of this quantitative imaging method. One of the chal- of new dose reduction techniques and new developments in CT
lenges with dynamic CTMPI is underestimation of MBF systems, resulting in faster scanning times and wider detectors,
compared to other quantitative modalities. Studies with PET-­ clinical implementation of dynamic CTMPI becomes closer.
determined MBF [66, 67] have reported stress values
between 3 and 5 ml/min/g, whereas dynamic CTMPI studies
[24, 32, 36]. report stress MBF values between 1.0 and References
1.4 ml/min/g. Studies of Bindschadler et al. [46] and Ishida
et al. [16] suggest that a limited temporal sampling rate is the 1. Jaarsma C, Leiner T, Bekkers SC, et  al. Diagnostic performance
of noninvasive myocardial perfusion imaging using single-photon
main cause of the underestimation of MBF determined by emission computed tomography, cardiac magnetic resonance, and
dynamic CTMPI.  Further research should investigate the positron emission tomography imaging for the detection of obstruc-
effect of temporal sampling rate on absolute MBF values and tive coronary artery disease. J Am Coll Cardiol. 2012;59(19):1719–
the influence on diagnostic accuracy. A second cause for 28. https://doi.org/10.1016/j.jacc.2011.12.040.
2. Schwitter J, Wacker CM, van Rossum AC, et  al. MR-IMPACT:
underestimation of MBF values with dynamic CTMPI could comparison of perfusion-cardiac magnetic resonance with single-­
be the use of combining a tracer-kinetic model with the photon emission computed tomography for the detection of coro-
Patlak method to calculate MBF. Possibly this method esti- nary artery disease in a multicentre, multivendor, randomized trial.
mates Ktrans instead of MBF [16]. By calculating the MBF Eur Heart J. 2008;29(4):480–9. https://doi.org/10.1093/eurheartj/
ehm617.
using a method purely based on deconvolution, MBF values 3. Parker MW, Iskandar A, Limone B, et al. Diagnostic accuracy of
could be more accurate compared to the most widely used cardiac positron emission tomography versus single photon emis-
hybrid method combining deconvolution and upslope calcu- sion computed tomography for coronary artery disease: a bivariate
lation. Each model has its own advantages and disadvan- meta-analysis. Circ Cardiovasc Imaging. 2012;5(6):700–7. https://
doi.org/10.1161/CIRCIMAGING.112.978270.
tages. Further research is needed to determine which 4. Hulten EA, Carbonaro S, Petrillo SP, Mitchell JD, Villines
tracer-kinetic model optimally approximates the true value TC.  Prognostic value of cardiac computed tomography angiogra-
of MBF in CTMPI. Currently, the effect of quantitative per- phy: a systematic review and meta-analysis. J Am Coll Cardiol.
fusion parameters on the diagnostic accuracy in the detection 2011;57(10):1237–47. https://doi.org/10.1016/j.jacc.2010.10.011.
5. Pelgrim GJ, Dorrius M, Xie X, et  al. The dream of a one-
of CAD and myocardial ischemia is still unknown. stop-­shop: meta-analysis on myocardial perfusion CT.  Eur
There are a few studies comparing the diagnostic accu- J Radiol. 2015;84(12):2411–20. https://doi.org/10.1016/j.
racy of different perfusion modalities [1, 3]; however, these ejrad.2014.12.032.
do not include dynamic CTMPI.  The performance of 6. Danad I, Szymonifka J, Schulman-Marcus J, Min JK.  Static and
dynamic assessment of myocardial perfusion by computed tomog-
(dynamic) CTMPI with regard to SPECT, PET, and MRI, raphy. Eur Hear J Cardiovasc Imaging. 2016;17:836–44. https://
potentially in combination with anatomical evaluation proce- doi.org/10.1093/ehjci/jew044.
dures, needs to be determined in large patient studies includ- 7. Williams MC, Newby DE. CT myocardial perfusion imaging: cur-
ing patient outcomes and analysis of cost-effectiveness. rent status and future directions. Clin Radiol. 2016;71(8):1–11.
https://doi.org/10.1016/j.crad.2016.03.006.
Large multicenter studies could also aid in the search for the 8. Caruso D, Eid M, Schoepf UJ, et  al. Dynamic CT myocardial
optimal perfusion cutoff values for several subpopulations in perfusion imaging. Eur J Radiol. 2016;85:1893. https://doi.
order to increase the diagnostic accuracy. It should be estab- org/10.1016/j.ejrad.2016.07.017.
lished if dynamic CTMPI in patients with previous coronary 9. Vliegenthart R, De Cecco CN, Wichmann JL, et al. Dynamic CT
myocardial perfusion imaging identifies early perfusion abnormali-
artery bypass graft surgery or stent placements has an added ties in diabetes and hypertension: insights from a multicenter reg-
value in the diagnostic process. The SPECIFIC study may istry. J Cardiovasc Comput Tomogr. 2016;10(4):301–8. https://doi.
answer some of these remaining questions. The objective of org/10.1016/j.jcct.2016.05.005.
63  Dynamic Myocardial CT Perfusion Imaging 827

10. George RT, Jerosch-Herold M, Silva C, et al. Quantification of myo- comparison with conventional catheter coronary angiography and
cardial perfusion using dynamic 64-detector computed tomogra- SPECT nuclear myocardial perfusion imaging. Am J Roentgenol.
phy. Investig Radiol. 2007;42(12):815–22. https://doi.org/10.1097/ 2012;198(3):521–9. https://doi.org/10.2214/AJR.11.7830.
RLI.0b013e318124a884. 25. Bastarrika G, Ramos-Duran L, Rosenblum MA, Kang DK,

11. Christian TF, Frankish ML, Sisemoore JH, et al. Myocardial per- Rowe GW, Schoepf UJ.  Adenosine-stress dynamic myo-
fusion imaging with first-pass computed tomographic imaging: cardial CT perfusion imaging: initial clinical experience.
measurement of coronary flow reserve in an animal model of Investig Radiol. 2010;45(6):306–13. https://doi.org/10.1097/
regional hyperemia. J Nucl Cardiol. 2010;17(4):625–30. https:// RLI.0b013e3181dfa2f2.
doi.org/10.1007/s12350-010-9206-6. 26. Tanabe Y, Kido T, Uetani T, et  al. Differentiation of myocardial
12. So A, Hsieh J, Li J-Y, Hadway J, Kong H-F, Lee T-Y. Quantitative ischemia and infarction assessed by dynamic computed tomog-
myocardial perfusion measurement using CT perfusion: a vali- raphy perfusion imaging and comparison with cardiac magnetic
dation study in a porcine model of reperfused acute myocardial resonance and single-photon emission computed tomography. Eur
infarction. Int J Cardiovasc Imaging. 2012;28(5):1237–48. https:// Radiol. 2016:1–12. http://www.embase.com/search/results?subacti
doi.org/10.1007/s10554-011-9927-x. on=viewrecord&from=export&id=L608199060.
13. Mahnken AH, Klotz E, Pietsch H, et al. Quantitative whole heart 27. Wichmann JL, Meinel FG, Schoepf UJ, et al. Semiautomated global
stress perfusion CT imaging as noninvasive assessment of hemody- quantification of left ventricular myocardial perfusion at stress
namics in coronary artery stenosis: preliminary animal experience. dynamic ct: diagnostic accuracy for detection of territorial myocar-
Investig Radiol. 2010;45(6):298–305. https://doi.org/10.1097/ dial perfusion deficits compared to visual assessment. Acad Radiol.
RLI.0b013e3181dfa3cf. 2016;23(4):429–37. https://doi.org/10.1016/j.acra.2015.12.005.
14. Bamberg F, Hinkel R, Schwarz F, Sandner TA, Baloch E, Marcus 28. Wichmann JL, Meinel FG, Schoepf UJ, et  al. Absolute versus
R, Becker A, Kupatt C, Wintersperger BJ, Johnson TR, Theisen relative myocardial blood flow by dynamic CT myocardial perfu-
DKM.  Accuracy of dynamic computed tomography adenosine sion imaging in patients with anatomic coronary artery disease.
stress myocardial perfusion imaging in estimating myocardial Am J Roentgenol. 2015;205(1):W67–72. https://doi.org/10.2214/
blood flow at various degrees of coronary artery stenosis using a AJR.14.14087.
porcine animal model. Investig Radiol. 2012;47(1):71–7. 29. Bamberg F, Marcus RP, Becker A, et al. Dynamic myocardial CT
15. Rossi A, Uitterdijk A, Dijkshoorn M, et al. Quantification of myo- perfusion imaging for evaluation of myocardial ischemia as deter-
cardial blood flow by adenosine-stress CT perfusion imaging in mined by MR imaging. JACC Cardiovasc Imaging. 2014;7(3):267–
pigs during various degrees of stenosis correlates well with cor- 77. https://doi.org/10.1016/j.jcmg.2013.06.008.
onary artery blood flow and fractional flow reserve. Eur Heart J 30. Ebersberger U, Marcus RP, Schoepf UJ, et al. Dynamic CT myo-
Cardiovasc Imaging. 2013;14(4):331–8. https://doi.org/10.1093/ cardial perfusion imaging: performance of 3D semi-automated
ehjci/jes150. evaluation software. Eur Radiol. 2014;24(1):191–9. https://doi.
16. Ishida M, Kitagawa K, Ichihara T, et al. Underestimation of myo- org/10.1007/s00330-013-2997-5.
cardial blood flow by dynamic perfusion CT: explanations by two-­ 31. Kono AK, Coenen A, Lubbers M, et al. Relative myocardial blood
compartment model analysis and limited temporal sampling of flow by dynamic computed tomographic perfusion imaging pre-
dynamic CT. J Cardiovasc Comput Tomogr. 2016;10:1–8. https:// dicts hemodynamic significance of coronary stenosis better than
doi.org/10.1016/j.jcct.2016.01.008. absolute blood flow. Investig Radiol. 2014;49(12):801–7. https://
17. Motwani M, Fairbairn TA, Larghat A, et  al. Systolic versus dia- doi.org/10.1097/RLI.0000000000000087.
stolic acquisition in myocardial perfusion MR imaging. Radiology. 32. Rossi A, Dharampal A, Wragg A, et  al. Diagnostic performance
2012;262(3):816–23. https://doi.org/10.1148/radiol.11111549. of hyperaemic myocardial blood flow index obtained by dynamic
18. Motwani M, Kidambi A, Sourbron S, et  al. Quantitative three-­ computed tomography: does it predict functionally significant
dimensional cardiovascular magnetic resonance myocardial perfu- ­coronary lesions? Eur Heart J Cardiovasc Imaging. 2014;15(1):85–
sion imaging in systole and diastole at 3.0 T. Heart. 2013;99:A56–7. 94. https://doi.org/10.1093/ehjci/jet133.
https://doi.org/10.1186/1532-429X-16-19. 33. Huber AM, Leber V, Gramer BM, et al. Myocardium: dynamic ver-
19. Rossi A, Merkus D, Klotz E, Mollet N, de Feyter PJ, Krestin sus single-shot CT perfusion imaging. Radiology. 2013;269(2):1–
GP.  Stress myocardial perfusion: imaging with multidetec- 8. https://doi.org/10.1148/radiol.13121441.
tor CT.  Radiology. 2014;270(1):25–46. https://doi.org/10.1148/ 34. Greif M, von Ziegler F, Bamberg F, et al. CT stress perfusion imag-
radiol.13112739. ing for detection of haemodynamically relevant coronary steno-
20. Einstein AJ.  Effects of radiation exposure from cardiac imaging: sis as defined by FFR.  Heart. 2013;99(14):1004–11. https://doi.
how good. J Am College Cardiol. 2012;59(6):553–65. https://doi. org/10.1136/heartjnl-2013-303794.
org/10.1016/j.jacc.2011.08.079.Effects. 35. So A, Wisenberg G, Islam A, et  al. Non-invasive assessment of
21. Yang DH, Kim Y-H, Roh JH, et al. Diagnostic performance of on-­ functionally relevant coronary artery stenoses with quantita-
site CT-derived fractional flow reserve versus CT perfusion. Eur tive CT perfusion: preliminary clinical experiences. Eur Radiol.
Heart J Cardiovasc Imaging. 2016;18:432. https://doi.org/10.1093/ 2012;22(1):39–50. https://doi.org/10.1007/s00330-011-2260-x.
ehjci/jew094. 36. Bamberg F, Becker A, Schwarz F, et al. Detection of hemodynami-
22. Baxa J, Hromádka M, Šedivý J, et al. Regadenoson-stress dynamic cally significant coronary artery stenosis : incremental diagnos-
myocardial perfusion improves diagnostic performance of CT tic value of dynamic CT-based. Radiology. 2011;260(3):689–98.
angiography in assessment of intermediate coronary artery steno- https://doi.org/10.1148/radiol.11110638/-/DC1.
sis in asymptomatic patients. Biomed Res Int. 2015;2015:105629. 37. Ho KT, Chua KC, Klotz E, Panknin C.  Stress and rest dynamic
https://doi.org/10.1155/2015/105629. myocardial perfusion imaging by evaluation of complete time-­
23. Weininger M, Schoepf UJ, Ramachandra A, et al. Adenosine-stress attenuation curves with dual-source CT. JACC Cardiovasc Imaging.
dynamic real-time myocardial perfusion CT and adenosine-stress 2010;3(8):811–20. https://doi.org/10.1016/j.jcmg.2010.05.009.
first-pass dual-energy myocardial perfusion CT for the assessment 38. Kido T, Kurata A, Higashino H, Inoue Y. Quantification of regional
of acute chest pain: initial results. Eur J Radiol. 2012;81(12):3703– myocardial blood flow using. Circ J. 2008;72:1086–91.
10. https://doi.org/10.1016/j.ejrad.2010.11.022. 39. Kim SM, Kim YN, Choe YH.  Adenosine-stress dynamic myo-
24. Wang Y, Qin L, Shi X, et  al. Adenosine-stress dynamic myocar- cardial perfusion imaging using 128-slice dual-source CT: opti-
dial perfusion imaging with second-generation dual-source CT: mization of the CT protocol to reduce the radiation dose. Int J
828 M. van Assen et al.

Cardiovasc Imaging. 2013;29(4):875–84. https://doi.org/10.1007/ 54. Schwitter J, Nanz D, Kneifel S, et  al. Assessment of myocardial
s10554-012-0138-x. perfusion in coronary artery disease by magnetic resonance: a com-
40. Deseive S, Pugliese Md F, Meave A, et al. Image quality and radia- parison with positron emission tomography and coronary angiogra-
tion dose of a prospectively electrocardiography-triggered high-­ phy. Circulation. 2001;103(18):2230–5.
pitch data acquisition strategy for coronary CT angiography: the 55. Pelgrim G, Handayani A, Dijkstra H, et al. Quantitative myocardial
multicenter, randomized PROTECTION IV study. J Cardiovasc perfusion with dynamic contrast-enhanced imaging in MRI and CT
Comput Tomogr. 2015;9:278–85. https://doi.org/10.1016/j. : theoretical models and current implementation. Biomed Res Int.
jcct.2015.03.001. 2016;2016:1.
41. Gordic S, Desbiolles L, Sedlmair M, et  al. Optimizing radiation 56. Lee TY. Functional CT: physiological models. Trends Biotechnol.
dose by using advanced modelled iterative reconstruction in high-­ 2002;20(8):3–10. https://doi.org/10.1016/S0167-7799(02)02035-8.
pitch coronary CT angiography. Eur Radiol. 2016;26(2):459–68. 57. Larghat AM, Maredia N, Biglands J, et al. Reproducibility of first-­
https://doi.org/10.1007/s00330-015-3862-5. pass cardiovascular magnetic resonance myocardial perfusion. J
42. Einstein AJ, Moser KW, Thompson RC, Cerqueira MD, Henzlova Magn Reson Imaging. 2013;37(4):865–74. https://doi.org/10.1002/
MJ.  Radiation dose to patients from cardiac diagnostic imaging. jmri.23889.
Circulation. 2007;116(11):1290–305. https://doi.org/10.1161/ 58. Gamel BJ, Katholi CR, Mesel E.  Pitfalls in digital computation
CIRCULATIONAHA.107.688101. of the impulse response of vascular beds from indicator-dilution
43. Danad I, Raijmakers PG, Appelman YE, et  al. Hybrid imaging curves. Circ Res. 1973;32(April):516–23.
using quantitative H215O PET and CT-based coronary angiog- 59. Papanastasiou G, Williams MC, Kershaw LE, et al. Measurement
raphy for the detection of coronary artery disease. J Nucl Med. of myocardial blood flow by cardiovascular magnetic resonance
2013;54(1):55–63. https://doi.org/10.2967/jnumed.112.104687. perfusion: comparison of distributed parameter and Fermi models
44. Fujita M, Kitagawa K, Ito T, et al. Dose reduction in dynamic CT with single and dual bolus. J Cardiovasc Magn Reson. 2015;17:17.
stress myocardial perfusion imaging: comparison of 80-kV/370-­ https://doi.org/10.1186/s12968-015-0125-1.
mAs and 100-kV/300-mAs protocols. Eur Radiol. 2014;24(3):748– 60. Futamatsu H, Wilke N, Klassen C, et  al. Evaluation of cardiac
55. https://doi.org/10.1007/s00330-013-3063-z. magnetic resonance imaging parameters to detect anatomi-
45. Gramer BM, Muenzel D, Leber V, et al. Impact of iterative recon- cally and hemodynamically significant coronary artery disease.
struction on CNR and SNR in dynamic myocardial perfusion imag- Am Heart J. 2007;154(2):298–305. https://doi.org/10.1016/j.
ing in an animal model. Eur Radiol. 2012;22(12):2654–61. https:// ahj.2007.04.024.
doi.org/10.1007/s00330-012-2525-z. 61. Maredia N, Plein S, Younger JF, et al. Detection of triple vessel cor-
46. Bindschadler M, Modgil D, Branch KR, La Riviere PJ, Alessio onary artery disease by visual and quantitative first pass CMR myo-
AM.  Comparison of blood flow models and acquisitions for cardial perfusion imaging in the CE-MARC study. J Cardiovasc
quantitative myocardial perfusion estimation from dynamic Magn Reson. 2011;13. http://www.embase.com/search/results?sub
CT. Phys Med Biol. 2008;141(4):520–9. https://doi.org/10.1016/j. action=viewrecord&from=export&id=L70465436:O29.
surg.2006.10.010.Use. 62. Sammut E, Zarinabad N, Wesolowski R, et  al. Feasibility of

47. Zoghbi GJ, Dorfman TA, Iskandrian AE.  The effects of medica- high-resolution quantitative perfusion analysis in patients with
tions on myocardial perfusion. J Am Coll Cardiol. 2008;52(6):401– heart failure. J Cardiovasc Magn Reson. 2015;17:13. https://doi.
16. https://doi.org/10.1016/j.jacc.2008.04.035. org/10.1186/s12968-015-0124-2.
48. Machecourt J, Longère P, Fagret D, et  al. Prognostic value of 63. Kim EY, Chung WJ, Sung YM, et al. Normal range and regional
thallium-201 single-photon emission computed tomographic heterogeneity of myocardial perfusion in healthy human myocar-
myocardial perfusion imaging according to extent of myo- dium: assessment on dynamic perfusion CT using 128-slice dual-­
cardial defect. Study in 1,926 patients with foilow-up at 33 source CT. Int J Cardiovasc Imaging. 2014;30(SUPPL. 1):33–40.
months. J Am Coll Cardiol. 1994;23(5):1096–106. https://doi. https://doi.org/10.1007/s10554-014-0432-x.
org/10.1016/0735-1097(94)90597-5. 64. Danad I, Raijmakers PG, Appelman YE, et al. Coronary risk fac-
49. Cerqueira MD, Weissman NJ, Dilsizian V, et al. Standardized myo- tors and myocardial blood flow in patients evaluated for coronary
cardial segmentation and nomenclature for tomographic imaging of artery disease: a quantitative [15O]H2O PET/CT study. Eur J Nucl
the heart. Circulation. 2002;105:539–42. Med Mol Imaging. 2012;39(1):102–12. https://doi.org/10.1007/
50. Meinel FG, Ebersberger U, Schoepf UJ, et  al. Global quantifica- s00259-011-1956-0.
tion of left ventricular myocardial perfusion at dynamic CT: fea- 65. Schwarz F, Hinkel R, Baloch E, et  al. Myocardial CT per-

sibility in a multicenter patient population. Am J Roentgenol. fusion imaging in a large animal model: comparison of
2014;203(2):174–80. https://doi.org/10.2214/AJR.13.12328. dynamic versus single-phase acquisitions. JACC Cardiovasc
51. Ingrisch M, Sourbron S.  Tracer-kinetic modeling of dynamic
Imaging. 2013;6(12):1229–38. https://doi.org/10.1016/j.
contrast-enhanced MRI and CT: a primer. J Pharmacokinet jcmg.2013.05.018.
Pharmacodyn. 2013;40(3, SI):281–300. https://doi.org/10.1007/ 66. Kajander SA, Joutsiniemi E, Saraste M, et  al. Clinical value of
s10928-013-9315-3. absolute quantification of myocardial perfusion with (15)O-water in
52. Al-Saadi N, Nagel E, Gross M, et al. Noninvasive detection of myo- coronary artery disease. Circ Cardiovasc Imaging. 2011;4(6):678–
cardial ischemia from perfusion reserve based on cardiovascular 84. https://doi.org/10.1161/CIRCIMAGING.110.960732.
magnetic resonance. Circulation. 2000;101(12):1379–83. 67. Bol A, Melin JA, Vanoverschelde JL, et  al. Direct comparison of
53. Al-Saadi N, Nagel E, Gross M, et al. Improvement of myocardial [13N]ammonia and [15O]water estimates of perfusion with quan-
perfusion reserve early after coronary intervention: assessment tification of regional myocardial blood flow by microspheres.
with cardiac magnetic resonance imaging. J Am Coll Cardiol. Circulation. 1993;87(2):512–25. https://doi.org/10.1161/01.
2000;36(5):1557–64. CIR.87.2.512.
 CT’s Role for Myocardial Viability
Assessment 64
Ahmed Hamdy and Kakuya Kitagawa

Why Assessing Myocardial Viability? cantly down to 3% compared to 16% in those treated medi-
cally. In the case of nonviable myocardium, however, the
 o Revascularize or Not to Revascularize, That
T opposite scenario was encountered where the annual death
Is the Question rate was 7.7% vs 6.2% in the revascularization and medical
treatment groups, respectively. The results of the analysis
If Shakespeare were to consider a career shift to become a signify the fact that revascularization for the nonviable myo-
cardiologist, that would be his most celebrated quote ever. cardium does not – at best – improve LV function nor sur-
The degree of left ventricular (LV) dysfunction is an undis- vival rates but carries an additional risk for the already
putable determinant of long-term survival in ischemic heart burdened patients. In short, myocardial viability assess-
disease (IHD) [1–5]. Broadly speaking, the LV dysfunction ment – whatever the tool used for the purpose – is critical for
following an ischemic event can be the result of one of four the clinical decision-making through proper selection of
possible scenarios depending on the stage and reversibility patients who will benefit most from revascularization.
state of the ischemic myocardium: (1) the acute irreversible
myocardial injury, i.e., myocardial necrosis in acute myocar-
dial infarction (MI), (2) the chronic irreversible injury, i.e., Why CT?
myocardial scarring in chronic/old MI, (3) the acute revers-
ible injury, i.e., myocardial stunning that refers to viable but  yocardial Assessment, Achilles Heel
M
acutely dysfunctional myocardium despite already restored of Cardiac CT?
myocardial perfusion following an acute ischemic event. The
condition can take up to several weeks to resolve, but eventu- So far, the main clinical application for cardiac CT is reserved
ally, the stunned myocardium restores its contractility, and for the evaluation of the coronary arteries where it dominates
(4) chronic reversible injury, i.e., myocardial hibernation that over other noninvasive methods and is regarded as a gate-
refers to viable but chronically dysfunctional myocardium keeper for invasive catheterization [8]. On the contrary, its
due to severe coronary artery disease (CAD) with perfusion role for myocardial viability assessment has always been
impairment at rest. Revascularization in the latter group of regarded premature and vulnerable. The very first attempts to
patients (those with viable hibernating myocardium) is asso- find a place for CT in myocardial evaluation date back to the
ciated with substantial survival benefit, symptomatic late 1970s and early 1980s [9–12]. These were animal stud-
improvement, and improved LV function [6]. For this reason, ies performed with early generations of CT scanners.
identifying viable and nonviable myocardium and hence dif- Frustrated with the slow machines, frequent artifacts and
ferentiating between reversible and irreversible LV dysfunc- low-resolution images, the modality was not further pursued
tion is of utmost importance. In a meta-analysis by Allman as a clinical tool. Alternatively, cardiac magnetic resonance
et al. involving more than 3000 patients, the authors demon- (CMR) emerged as a tissue characterizing modality that soon
strated a strong association between myocardial viability on became a mature clinical tool for viability assessment;
noninvasive testing and improved survival after revascular- mainly by identifying areas of late gadolinium enhancement
ization in patients with chronic CAD and LV dysfunction (LGE-MRI). Nearly 20  years later, the introduction of fast
[7]. In the case of viable myocardium, the annual death rate rotating, multi-slice CT (MSCT) technology made the door
in patients treated with revascularization was brought signifi- wide open for cardiac CT researchers to rediscover its poten-
tial in infarct imaging. This technical evolution accompanied
with the fact that both iodinated and gadolinium-based extra-
A. Hamdy · K. Kitagawa (*)
Department of Radiology, Mie University Hospital, Tsu, Japan cellular contrast materials used for CT and MRI respectively
e-mail: [email protected] share almost the same pharmacodynamics, led to renewed

© Humana Press 2019 829

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_64
830 A. Hamdy and K. Kitagawa

interest particularly in late enhancement CT (LE-CT) for The reasons why CMR is dominating the field of viabil-
viability assessment. In their prominent study, Gerber et al. ity assessment are its superior tissue characterization, lack
demonstrated good agreement (82%) between both LE-CT of ionizing radiation, and the relatively low nephrotoxicity
and LGE-MRI for the detection of the late-enhanced areas of its contrast agents. CMR, however, is not flawless. It
with an excellent agreement on localization [13]. On the his- requires a complex infrastructure, long scanning times, is
tological level, Lardo et al. found that areas of LE correlated expensive, and contraindicated in implanted electronic
well with histopathological findings in terms of detection, devices that are commonly encountered in patients with
localization, and transmurality assessment for both the acute ischemic cardiomyopathy.
and chronic infarctions [14]. These – among other studies – In the United States, centers providing cardiac CT are
are only examples of how cardiac CT has evolved beyond the nearly triple the number of available cardiac MRI centers.
scope of coronary vessels assessment into a potential clinical The same rule applies to the number of cardiac CT and MRI
tool for infarction imaging. scans performed annually. Statistically speaking, a quick
analysis of the Medicare Provider Utilization and Payment
Data for the year 2013 reveals that the number of cardiac CT
CT and Rivals, the Pros and Cons service providers in the United States was 1147 compared to
418 providers of cardiac MRI service. The number of cardiac
Several rivals compete to prove superiority in myocardial CT scans was 40,511 compared to 12,076 cardiac MRI scans
viability assessment, the list of which contains single-photon [16]. This obvious lead in service availability should act as
emission computed tomography (SPECT), fluorine-18-­ propelling force toward more clinical role for cardiac CT in
labeled deoxyglucose (FDG)-positron emission tomography viability assessment.
(PET), dobutamine echocardiography, CMR, and recently An important advantage of CT over MRI is the greatly
CT. Despite no consensus has been yet available, LGE-MRI reduced slice thickness. The widely available MSCT scan-
is regarded as the gold standard for infarction assessment ners provide slice thicknesses as thin as 0.5  mm, 10–20
[15]. The choice of the appropriate technique depends on times thinner than that in a typical MRI viability study.
many factors such as the availability, experience, cost, This has great clinical implications through reducing the
machine, and patient’s characteristics. The pros and cons as partial volume effects that occur when the slice thickness
well as the underlying mechanism by which these techniques exceeds the size of the imaged structure of interest.
assess viability also differ as summarized in Table  64.1. Another independent advantage of the reduced slice thick-
Comparing LE-CT to LGE-MRI is of particular interest ness is the greatly improved spatial resolution in the z axis
because both methods share the same pathophysiological which in turn enables the p­ roduction of true 3-dimen-
principle for infarction detection and because LGE-MRI is sional images and retrograde reconstruction of virtually
the current gold standard for viability assessment as previ- any arbitrary slice orientation from the original stack of
ously pointed out. axial images. LGE-MRI also allows production of any
arbitrary slice orientation, but this is only possible at the

Table 64.1  Summary of pathophysiological principles, pros and cons of the different viability assessment imaging modalities
Modality Principle Pros Cons
Dobutamine echocardiography Contractile reserve Most available Operator dependency
Validated Occasional acoustic window inadequacy
No radiation
SPECT Intact cell membrane Available Use of radioisotopes
 Thallium-201 Extensive validation and Low spatial resolution
 Technetium-99 m-sestamibi experience
FDG-PET Glucose metabolism High accuracy Use of radioisotopes.
Good spatial resolution Relatively limited availability
MRI Highest accuracy; considered Incompatible with implanted intra-cardiac
 Dobutamine cine MRI Contractile reserve gold standard devices or in claustrophobic patients.
 LGE-MRI AMI: Ruptured cell Best spatial resolution and
membranes contrast
Chronic MI: Expanded No radiation
LE-CT extracellular matrix Available Radiation
Rapid scan Limited validation and experience
Best spatial resolution
Combines CT coronary
angiography
64  CT’s Role for Myocardial Viability Assessment 831

time of data acquisition; a process that is heavily depen- sin-converting-enzyme inhibitors, statins) may promote the
dent on operator interaction and obviously time consum- development of adiposity within scar tissue which can
ing. This latter, i.e., scan time, is another disappointment explain the lack of reports before 1997. In addition, very
in LGE-MRI. A single slice would take about 20 s to be recently, Kami et  al. showed that cardiac pluripotent stem
acquired compared to 10  s for the entire myocardium cells differentiated into adipocytes through certain transcrip-
using the currently commercially available CT scanners. tion factors within an ischemic reperfused model [21].
The prognostic value of the deposited adipose tissue
and its impact on cardiac electrophysiology have been
 trategies for Infarct Detection
S totally unclear until very recently when researchers
and Myocardial Viability Assessment with CT proved that lipomatous tissue was associated with
increased propensity of ventricular tachycardia (VT) and
In the United States, 195,000 silent MIs are estimated to VT circuit sites [22, 23].
occur annually [17]. Updates from Framingham study From this standpoint, CT seems to be taking the lead in
showed that the 10-year mortality rate in unrecognized MI visualizing areas of lipomatous metaplasia. Echocardiography
patients was 45% compared to 39% in patients with recog- does not allow myocardial fat characterization, and CMR
nized MI [18]. These numbers demonstrate the importance requires acquisition of T1 pulse sequences, whereas fat can
of previous unrecognized MI detection and how this can help be clearly visualized throughout any phase of cardiac CT
provide the appropriate care and consequently decrease mor- including the non-contrast scan (Fig. 64.1). Ichikawa et al.
tality rates. CT can detect previous MI through identifying found a prevalence of 62% of cardiac adiposity in MI patients
areas of lipomatous metaplasia, myocardial calcification, compared to 3% of controls [24]. They also found a strong
wall thinning, or abnormal enhancement. correlation with infarct age of more than 3 years. Ahn et al.
found areas of lipomatous metaplasia in 22.4% of MI patients
and was more frequently associated with a longer post-­
Non-contrast or Non-cardiac CT infarct period, milder coronary artery stenosis, fewer number
of diseased vessels, and more severe regional wall motion
Lipomatous Metaplasia abnormalities [25]. Gupta et al. compared non-contrast car-
The term describes the process of adipose tissue/fat deposi- diac CT with SPECT for detection of chronic/old MI
tion in the myocardium, occurring most commonly second- depending on lipomatous metaplasia visualization on
­
ary to infarction, hence regarded as an indirect sign of CT.  Non-contrast CT yielded high sensitivity of 92% and
nonviable myocardium. The first report for identifying areas specificity of 72%. Based on a ROC curve, a reference
of fat metaplasia in explanted hearts dates back to 1997 [19]. Hounsfield unit (HU) of 21.7 achieved sensitivity of 97.4%
Ever since, the exact etiology and ­pathophysiology of the and specificity of 98.7% [26].
phenomenon remained largely unknown. In their study on
explanted hearts, Su et al. found that fat metaplasia prevailed  yocardial Scar Calcification
M
in 84% of the healed MI lesions [20]. They hypothesized that Previous myocardial infarction is the most common cause of
modern therapies in the management of IHD (e.g., angioten- extra-coronary cardiac calcification [27]. Cell necrosis and

Fig. 64.1 Lipomatous
metaplasia (arrows) in a a b
patient with long-standing
anteroseptal MI presenting as
subendocardial fat density in
both pre-contrast (a) and
LE-CT (b)
832 A. Hamdy and K. Kitagawa

membrane rupture during MI initiates the process of calcium tricular aneurysm which is readily visualized on CT
deposition which is then accentuated by the ischemia-­ (Fig. 64.3). No studies are available on the value of wall thin-
induced microenvironment. In 1975, calcification was ning evaluated with CT for prediction of viability and subse-
reported to be present in 8% of myocardial infarctions older quent functional recovery after revascularization, though
than 6 years using plain radiographs [28]. On echocardiogra- CMR studies show that left ventricular wall thickness more
phy, myocardial calcifications appear as echogenic foci caus- than 5.5 mm can predict viability with a sensitivity of 92%
ing back shadowing, while at CMR they appear as low-signal and a specificity of 56% [31].
foci in all pulse sequences. CT is obviously the modality of
choice for calcium detection. Old MI causes associate dys-
trophic calcification that typically appears as thin curvilinear
sheets of high-density calcium along the periphery of the
infarct [29] (Fig.  64.2). It is worth noting that myocardial
scar calcifications can appear on non-cardiac or non-gated
cardiac CT scans; though gated scans obviously offer the
best image quality. The prognostic value of myocardial cal-
cification occurring on top of previous MI is not yet known
and warrants research to resolve its relation specifically to
ventricular function and electrophysiology.

Contrast-Enhanced Cardiac CT

Wall Thinning
Following MI, the inflammatory process starts removing the
necrotic myocardium which ultimately leads to focal wall
thinning. It is easily expected that the infarcted part of the
ventricular wall becomes thinner with time. Nieman et  al.
confirmed that expectation, showing that patients with long-­
standing MI had significantly thinner myocardium that
reached 5 mm or less in about 50% of cases. In acute MI, no
significant differences were found when compared with con-
Fig. 64.3  LE-CT showing an apical LV aneurysm developing after
trols [30]. One complication of severe myocardial thinning anteroseptal MI involving the apex (block arrows). A hypodense throm-
following transmural infarction is the development of ven- bus is seen within the aneurysm (open arrow)

a b c

Fig. 64.2  Subendocardial curvilinear calcium density representing tous metaplasia can also be noted (open arrow). Note that calcium is not
dystrophic calcification (block arrows) in this long-standing anterosep- detected with LGE-MRI (c)
tal MI patient; (a) unenhanced CT and (b) LE-CT. Associated lipoma-
64  CT’s Role for Myocardial Viability Assessment 833

 irst-Pass Arterial Phase Scan

F strated a sensitivity of 83% and specificity of 91% compared
The first-pass arterial phase CT (basically a standard coro- to a clinical reference with mean attenuation value for the
nary CT angiography study or a myocardial perfusion CT) infarcted areas and normal myocardium being 38.9 ± 14 HU
can be used to detect infarcted myocardium. This is based on vs 104.0 ± 16 HU, respectively (p < 0.01) [34]. Nieman et al.
the fact that contrast enhancement of the myocardium during used arterial phase cardiac CT to differentiate between acute
arterial phase CT is directly related to myocardial perfusion. and chronic infarctions. They found significantly lower CT
Infarcted myocardium typically appears as hypoattenuated attenuation values in patients with long-standing MI
perfusion defects (Figs.  64.4 and 64.5). In one of the first (−13 ± 37 HU) than in those with acute MI (26 ± 26HU) and
studies on animal models, Hoffmann et al. found significant normal controls (73 ± 14 HU, p < 0.001) which can be attrib-
reductions in CT attenuation of the acute infarcts when com- uted to lipomatous metaplasia [30].
pared to reference myocardium (p  <  0.001). These differ- First-pass arterial cardiac CT was also compared to other
ences in attenuation also correlated well with differences in methods of viability assessment. When compared to LGE-­
microsphere-determined blood flow (p  <  0.01) [32]. In a MRI, it had a sensitivity of about 91%, specificity of 81%,
clinical study by Nikolaou et al., arterial phase cardiac CT and diagnostic accuracy of 86% but with tendency to under-
yielded a sensitivity of 85%, specificity of 91%, and accu- estimate infarcts sizes [35]. This underestimation was also
racy of 90% for detecting MIs [33]. The mean attenuation found by Mahnken et al. who found the mean infarct size to
value for the infarcted areas was 54 ± 19 HU vs 117 ± 28 HU be 24.5 ± 18.3% per slice for early-phase arterial MSCT and
for non-infarcted myocardium (p < 0.01). These results are 31.2 ± 22.5% on MRI[36]. The frequent underestimation is
in concordance with the study of Francone et al. who demon- probably because only a part of the infarcted tissue shows a

Fig. 64.4 Sagittal
a c
reformatted images of the
arterial phase (a) showing a
hypodense perfusion defect
(open arrows) and LE-CT (b)
showing an area of late
enhancement at the
anteroseptal wall of LV (block
arrows). Note that the
hypodense area appears
smaller than the
hyperenhanced area in LE-CT b
(b) and LGE-MRI (c). On the
other hand, excellent
correlation can be noted
between LE-CT and
LGE-MRI

a b c

Fig. 64.5  Short-axis reformatted images of stress perfusion CT(a) and LE-CT (b), showing good correspondence to short-axis LGE-MRI (c) for
the anterior (block arrows) and inferior (open arrows) walls MI
834 A. Hamdy and K. Kitagawa

perfusion defect on the arterial phase, while the whole phases (at 0, 7, 28, and 90 days after induced MI and reperfu-
infarcted tissue becomes hyperenhanced in late phase of the sion). On day 0, the mean MI size was 23.7 ± 11.8% on CT
study. When compared to SPECT, Hennman et al. showed an and 24.5 ± 10.6% on CMR. On day 90, MI sizes decreased
overall agreement reaching up to 96% between arterial phase significantly to 16.9 ± 8.4% and 18.9 ± 8.0%, respectively
CT and SPECT for detection of rest perfusion defects of old (p = 0.0019). On TTC staining the size of MI was 16.8 ± 8.2%.
MIs [37, 38]. Good agreement between LE-CT and LGE-­MRI was found
From a clinical prospective, the hypoattenuated areas are [43].
not exclusively seen in the infarcted myocardium. In fact, Human studies went side by side with animal studies.
they are encountered in other conditions including lipoma- Head to head comparison was made in most cases with LGE-­
tous metaplasia with or without chronic MI, microvascular MRI. Mahnken et  al. demonstrated excellent agreement of
obstruction (MO) in reperfused acute MI, and the severely acute reperfused MI size for LE-CT and LGE-MRI
ischemic but viable myocardium in critical coronary artery (33.3%  ±  23.8% compared with 31.2%  ±  22.5%, respec-
stenosis. Beam hardening artifacts from dense contrast mate- tively) [36]. The mean CT values of the normal and infarcted
rial and adjacent metalic objects as well as motion artifact myocardium at the late phase were 74.5 ± 11 and 108 ± 15.9
can also be the source of hypoattenuation. One should be HU, respectively. In their study on acute and chronic MI
aware of these myocardial hypoattenuation-causing condi- patients, Gerber et al. found good agreement (82%, k = 0.61,
tions. Combination of the hypoattenuated areas with the p  <  0.001) for the identification of the hyperenhanced
thinned myocardial walls, however, is in most cases an indic- infarcted regions between both techniques on segmental
ative of old MI of the involved segment. basis and also high correlation for infarct size estimation
[13]. The mean CT values of the normal and infarcted myo-
 ate Contrast Enhancement
L cardium at the late phase were 97 ± 11 and 131 ± 16 HU,
LE-CT shares the same pathophysiological principle with respectively. Choe et  al. also compared LE-CT and LGE-­
LGE-MRI which is current clinical gold standard for infarct MRI in 63 acute and chronic MI patients. LE-CT was able to
detection and viability assessment, owing to the proven fact detect the infarcted myocardium in all patients and corre-
that both iodinated and gadolinium-based contrast materials lated well with MRI for infarct’s volume fraction estimation
share almost the same pharmacodynamics [13]. In the acute (correlation coefficient of 0.81, p < 0.0001). The mean CT
phase of MI, the ruptured cell membranes allow the origi- values of the normal and infarcted myocardium at the late
nally extracellular, interstitial contrast material to enter the phase were 82.2  ±  23.2 and 115  ±  24.3  HU, respectively
necrotic myocytes with resultant increased volume of distri- [44]. These studies consistently reported good agreement
bution. In old MI, the necrotic tissue is replaced with fibrotic, between LE-CT and LGE-MRI and, at the same time, low
collagen-rich scar tissue that leads to the same result; an HU difference (≈30  HU) between normal and infarcted
increased distribution volume of contrast material (Figs. 64.4 myocardium which makes contrast material reduction diffi-
and 64.5). cult. For example, Nieman et  al. performed LE-CT with
Several animal studies were conducted to investigate fea- administration of smaller amount of contrast material than
sibility and diagnostic accuracy of LE-CT. One of the first previous studies (75–90 mL, 320 mg of iodine per mL). In
studies by Beucker et  al. using a 16-slice MSCT demon- spite of good correlation between LE-CT and LGE-MRI for
strated the feasibility of the technique for acute MI detection infarcted myocardium fraction estimation (13%  ±  9 vs
in porcine model, its ability to differentiate between reper- 15% ± 7; mean difference of −2% ± 6, p = 0.37) [45], late
fused and non-reperfused infarctions and good correlation enhancement was present in only 11 of the 15 patients exam-
with MRI and triphenyltetrazolium chloride (TTC) staining ined because of excessive image noise, streak artifacts, or
for infarct’s size estimation [39]. Same results were obtained non-detectable late enhancement. The mean CT values of the
with 64-slice MSCT by Baks et al. using an acute reperfused remote and infarcted myocardium at the late phase were
MI model, Brodoefel et al. using both acute and subacute MI 72 ± 10 and 93 ± 15 HU, respectively.
models, and Qu et  al. using acute and chronic MI models LE-CT was also compared to methods other than
[40–42]. Lardo et  al. investigated LE-CT using a 32-slice MRI. Habis et al. compared LE-CT to low-dose dobutamine
MSCT in both acute canine and chronic porcine MI models. echocardiography. In 36 patients with acute MI, 64-slice CT
They found excellent correlation with TTC staining regard- was used to acquire late CT images after invasive coronary
ing morphology, transmurality as well as infarct volume esti- angiography (mean delay of 24 ± 11 min) with no contrast
mation. Volumes by LE-CT compared to TTC were injection, whereas echocardiography was performed
21.1% ± 7.2% vs 20.4% ± 7.4% in acute infarcts (mean dif- 2–4 weeks later. LE-CT had a 98% sensitivity, 94% speci-
ference, 0.7%) and 4.15%  ±  1.93% vs 4.92  ±  2.06% in ficity, 97% accuracy, and 99% positive and 79% negative
chronic infarcts (mean difference, 20.76%) [14]. Mahnken predictive values for detecting viable myocardial segments
et  al. used LE-CT for assessment of infarction at different compared to echocardiography [46]. Chiou et al. compared
64  CT’s Role for Myocardial Viability Assessment 835

LE-CT to thallium SPECT and dobutamine e­ chocardiography c­ omplications, including dislodged microemboli and reper-
in patients with previous MI.  By per-patient analysis, CT fusion microvascular injury.
detected MI in 96% of the patients while SPECT detected The true power of ECV, however, lies in the detection and
only 87%. Among segments with late enhancement >75% quantification of the diffuse interstitial type of myocardial
segmental thickness, SPECT and echocardiography detected fibrosis, extending well beyond the confines of focal replace-
nonviable segments in 87.8% and 92.2%, respectively [47]. ment fibrosis usually found in infarct scars. For that purpose,
LE-CT was also compared with FDG-PET and SPECT by CT-derived ECV values were validated against MRI as the
Lee et al. where CT had 84.2% and 89.2% agreement with imaging reference standard. Nacif et al. found good agree-
SPECT and PET, respectively, for MI localization and sen- ment between both methods in their study on 24 healthy and
sitivity, specificity, and diagnostic accuracy of 73.5%, heart failure subjects. As expected, ECV values were higher
79.4%, and 78.4% compared with combined PET and in heart failure patients than in healthy individuals
SPECT [48]. However, Dwivedi et al. found less impressive (33.5%  ±  5.9 vs 29.3%  ±  2.7, respectively; p  =  0.03) and
agreement with PET, reaching a modest value of 70% with were inversely related to ejection fraction [52]. Bandula
qualitative analysis that increased to 77% with quantitative et al. also compared ECV values obtained by equilibrium CT
analysis [49]. and CMR and validated the results against histology in 23
When comparing SPECT or FDG-PET with LE-CT, one patients with aortic stenosis undergoing valve replacement
has to keep in mind that these techniques depend on different surgery [53]. There was a significant correlation between
principles for viability assessment. This has significant both CT- and MR-derived ECV and percentage of histologic
implications when comparing their sensitivities, specifici- fibrosis (r  =  0.71 [p  <  0.001] and r  =  0.84 [p  <  0.0001],
ties, and diagnostic accuracies. SPECT and PET, for exam- respectively). Also, CT-derived ECV was significantly cor-
ple, visualize the viable myocytes, while the nonviable related to MR-derived ECV (r = 0.73). In cardiac amyloido-
myocytes appear as defective, invisible areas on imaging, a sis, Triebel et  al. have shown good correlation between
result of decreased cellular metabolism. This is the opposite dynamic equilibrium CT- and MR-derived ECV with good
in case of LE-CT and LGE-MRI, which visualize the nonvi- tracking of clinical markers of amyloidosis severity [54].
able myocytes rather than the viable ones, a result of the Recently, we demonstrated significant correlation between
expanded extracellular matrix in the infarcted myocardium. CT- and MR-derived ECV (R2 = 0.84, p < 0.001) with a mean
difference of −3.3% [55]. We also demonstrated that CT
Extracellular Volume Fraction measurement of myocardial ECV in subjects without clinical
Based on the abovementioned fact, it would make perfect CAD is feasible with high inter- and intra-­observer repro-
sense to actually quantify the extracellular volume (ECV) ducibility [56]. In this study we reported that mean ECV in
fraction as an indirect indicator for the presence of MI. Pre- males was lower compared with females (25.5  ±  2.0% vs
and post-contrast T1 mapping MRI is a validated method for 27.1 ± 1.8%, p = 0.02) and that it was positively related to
ECV estimation. Using MRI, Sado et al. found higher ECV age (r = 0.46, p = 0.003). There was no statistically signifi-
values in previous MI patients than in healthy subjects, Fabry cant difference in ECV between different left ventricular
disease, dilated and hypertrophic cardiomyopathies, aortic segments.
stenosis, and cardiac amyloidosis patients [50]. Publications
on the value of CT for the same purpose are scarce. In a  ost-percutaneous Coronary Intervention (PCI) LE-CT
P
recent study by Jablonowski et al., CT was used to quantify Another strategy for LE-CT imaging is acquiring late-phase
ECV in animal models of acute MI of different forms (patchy scan within 30 min after conventional coronary angiography
microinfarcts induced with injection of 16 and 32  mm3 or PCI. The strategy makes perfect use of the intra-arterial
microemboli, contiguous massive infarct induced by left contrast material injected directly into coronary arteries
anterior descending artery ligation and, combined micro- (Fig. 64.6). While the earliest study to apply the idea used a
and massive infarcts) [51]. Interestingly, mean ECV values non-gated conventional chest CT [57], subsequent studies
correlated with the severity of infarcts ranging from 36% ± 3 used a dedicated ECG-gated cardiac CT scan that yielded
and 41% ± 3 for the 16 and 32 mm3 microemboli-induced 98% sensitivity, 94% specificity, and 97% accuracy for via-
microinfarcts, respectively, to 55% ± 5 and 56% ± 4 for the bility detection at a very early stage of acute MI [46].
contiguous massive and combined infarcts, respectively. Very recently, Jang et  al. used a hybrid interventional
Regression analysis also revealed excellent correlation therapeutic CT system in subacute MI swine model. Late
between CT-derived ECV values and histological postmor- scan images were obtained at 2, 5, 7, 10, 15, and 20 min
tem microscopy (r2 = 0.92). The results not only demonstrate after the end of angiography. Interestingly, the 2 min time
the potential use of ECV as a biomarker of myocardial via- point yielded the highest CT values for the nonviable and
bility but also as a discriminator between different degrees of viable myocardium as well as maximum mean difference
myocardial injury and as a detector of peri-procedural in CT attenuation between them. Such findings at such an
836 A. Hamdy and K. Kitagawa

Fig. 64.6  Short-axis (a) and

a b
sagittal (b) reformatted
images of cardiac CT
performed immediately after
primary PCI to right coronary
artery in this 81-year-old man
with inferior acute MI. Note
the hyperenhanced infarction
region (arrows)

early time point can be attributed to the intra-coronary nation of misregistration can potentially be expected com-
injection of the contrast material. When correlated to TTC pared to single-energy approach.
staining on histology, mean percentage difference of MI
size at 2 min (8.5% ± 1.8%) was also the lowest [58].
The Wikihow of Late Enhancement Cardiac CT

 ual-Energy CT for Late Enhancement Imaging

D How to Do It
Dual-energy imaging allows material decomposition and can
theoretically provide better detection of infarcted myocar- The perfect formula for a good LE-CT scan should be one
dium at late phase by exploiting the increased iodine contrast that achieves the best diagnostic image quality with the least
enhancement of low-kV monochromatic images or through exposure to radiation and minimal volume of contrast mate-
direct mapping of iodine distribution. rial. This formula is not secret but involves many technical
In a porcine model of reperfused chronic MI, Deseive aspects that – so far – are not standardized among different
et al. demonstrated a high accuracy and a good correlation institutes. In fact, variations of the technique among cardiac
of dual-energy LE-CT to LGE-MRI and histopathology CT centers are so evident owing to differences in the used
[59]. In a patient study, Wichmann et al. reported a sensi- CT scanners, contrast volume, injection protocols, timing
tivity of 89%, specificity of 98%, and accuracy of 96% for between injection and image acquisition, tube settings, and
monochromatic LE-CT with LGE-MRI as the reference image reconstruction methods. In many cases, it is just a
standard, while iodine distribution analysis provided infe- matter of time before cardiac CT staff build their own per-
rior performance (52% sensitivity, 88% specificity, 81% sonal experiences and preferences. Below, we will try to
accuracy) [60]. More recently, Truong et al. compared late explore the different techniques for late enhancement scan
enhancement by standard single-energy (at both 80 kV and and present our personal experiences, tips, and tricks.
100 kV) to dual-energy LE-CT. The authors used a chronic
MI porcine model and correlated results to histopathology. Contrast Material Volume
Interestingly, standard single-­energy LE-CT (particularly Despite advances in CT technology have made it possible to
at 100 kV) showed the best correlation to histopathology acquire coronary CT angiography studies using as low as
for scar volume and also the highest contrast-to-noise ratio 50 ml of contrast material, this is still not true for viability
while dual-energy LE-CT overestimated infarct’s size the imaging with late-phase scan. For myocardial infarct delin-
most [61]. In light of the results by Truong et al. and most eation, adequate concentration gradient of the contrast mate-
of the previous studies, the theoretical assumption of supe- rial has to be achieved across vascular and expanded
rior infarction detection with dual-energy CT over single- extracellular spaces of the infarct. Relatively high volumes
energy CT is yet to be confirmed. of contrast material are needed to achieve this gradient. In
In ECV estimation, a different story is told. Lee et al. used animal studies, 700–1500 mg iodine/kg was administered. In
iodine maps to estimate ECV in healthy subjects and in those a 75  kg human subject, this would equal an average of
with non-ischemic cardiomyopathy [62]. Interobserver 250 mL of contrast material. On the other hand, human stud-
agreement for ECV at CT was excellent (ICC = 0.987). In ies used quantities between 300 and 700 mg iodine/kg. Based
addition, Bland-Altman analysis between MRI and CT on experimental and clinical evidence, volumes from as low
showed good agreement, with 95% limits of agreement of as 120 mL to as high as 175 mL are commonly used.
−1.19 and 1.79. Since pre-contrast images are not required From our personal experience, ≥600  mg  iodine/kg pro-
to estimate ECV with the strategy, dose reduction and elimi- duces sufficient late enhancement [63]. For contrast materials
64  CT’s Role for Myocardial Viability Assessment 837

with iodine concentration of 370 mg/kg, this means a total of cant difference between the myocardial infarct area mea-
approximately 120 mL contrast agent. sured using LE-CT acquired at 5 and 10 min after injection
and MRI in ST elevated MI patients [66]. In Brodoefel et al.,
Injection Protocol the bolus injection protocol achieved maximal contrast
Two main protocols for contrast material delivery are used; between the viable and nonviable myocardium at 3 and
the bolus and the bolus-continuous infusion protocols. In the 5 min time points, while the infusion protocol achieved that
more commonly used bolus protocol, no further contrast at 10 and 15 min time points [64]. This is an example of how
material is injected after the contrast bolus of routine coro- injection protocol can affect the choice of the perfect time
nary CT angiography. The second protocol involves continu- point for MI visualization among other factors such as the
ous slow infusion of 30–90 mL of contrast material after the volume of the contrast agent and patient’s hemodynamics. In
bolus used for coronary angiography until the late-phase scan a busy clinical facility where the luxury of a dedicated car-
at a rate of 0.1–0.3  mL/sec. The theory behind continuous diac CT scanner is not available, it becomes more important
infusion technique is to maintain a persistently high intravas- to determine the earliest possible time point that allows for
cular concentration gradient that maintains a slow wash-in of accurate diagnosis.
the contrast material to extracellular spaces of the myocardial Based on our experience and yet unpublished data, 5 and
infarct. Eventually, this should result in accentuated contrast 7  min after contrast agent administration equally allow for
between the infarcted and healthy myocardium. adequate wash-in of the contrast material (Fig. 64.7). Earlier
Till this moment, there is no consensus on which injection time points are mostly too early for contrast accumulation in
protocol to use. However, Brodoefel et al. compared both pro- infarct scars, while later time points are accompanied with
tocols in reperfused acute and subacute porcine models and excessive washout. That said, some questions pop into one’s
reported that bolus-continuous infusion protocol yielded bet- mind. Is an early image acquisition time point totally useless?
ter image quality than did single-bolus protocol due to Can it act in the same way as early gadolinium enhancement
increased attenuation difference between viable and nonvia- in CMR to better detect MO or thrombi? Will imaging at mul-
ble myocardium (p  ≤  0.003) and between left ventricular tiple time points do better than a single one? And if so, will
blood pool and nonviable myocardium (p  <  0.001) [64]. the benefit overweigh the increased radiation dose? These and
When correlated to MRI for MI size estimation, good correla- other questions need to be raised and answered by dedicated
tion was found for both protocols although continuous infu- studies.
sion tended to have better correlation coefficient, reaching a
maximal of 0.96 compared to 0.92 in single-bolus technique.  ube Settings, Radiation Dose, and Enhancing
T
One matter should be taken into consideration. The con- Image Quality
tinuously infused contrast agent leads to simultaneously Acquiring a good image quality with a low-dose scan is a
increased LV blood pool attenuation which can mask small never-ending quest. A common strategy to lower radiation
subendocardial infarcts. To avoid this, additional waiting dose is to use a low tube voltage. Brodoefel et al. compared
time of about 5 min after infusion ends is recommended to tube currents of 80 and 120 kV for late-phase imaging in a
allow for adequate washout of the contrast material from the porcine model of reperfused acute MI [67]. The 80 kV proto-
left ventricular cavity. This increases the total time interval col not only resulted in 65% reduction of radiation dose but
between coronary and late scans to an average of 15 min and also provided better contrast-to-noise ratio than did the
makes it difficult to go below 10 min. 120 kV protocol. This, however, was not true for tube current.
The single-bolus technique, on the other hand, is easier,
more suitable for the clinical setting, and can save 5–10 min.
In fact, the vast majority of publications on LE-CT and its Table 64.2  Examples of reported time points for LE-CTacquisition in
different human studies
diagnostic accuracy use a single-bolus protocol.
In a comprehensive protocol to be discussed later, con- Study Timing
trast agent will be injected over the different phases of the Mahnken et al. (2005) [36] 15 min
Lardo et al. (2006) [14] 5 min
study including coronary angiography, perfusion scan and
Gerber et al. (2006) [13] 10 min
even the test bolus. This is similar in principle to the split- Lessick et al. (2007) [81] 6 min
dose protocol used by Goetti et al. [65]. Nieman et al. (2008) [45] 7 min
Boussel et al. (2008) [89] 10 min
 iming of Late Image Acquisition
T Choe et al. (2008) [44] 10 min
This is yet another point of huge debate in cardiac CT com- Kang et al. (2010) [90] 6 min
munity. Human studies reported variable time points for George et al. (2012) [91] 5 min
image acquisition ranging from 5 up to 15 min after contrast Kurobe et al. (2014) [63] 7 min
administration (Table 64.2). Jacquier et al. found no signifi- Kurita et al. (2016) [56] 7 min
838 A. Hamdy and K. Kitagawa

a b c

Fig. 64.7  Comparison between 3(a), 5(b), and 7(c) min time points for LE-CT acquisition in a patient with small anterior MI. Maximal contrast
accumulation and best scar visualization are seen at 7 min in this patient

When the authors compared 400  mAs to the standard took a further step toward reducing image noise by acquiring
800  mAs, they reported further lowering of radiation dose, 4 stacks of TSFF images in one breath hold and averaging
the cost of which was an unacceptable degradation of image them by means of non-rigid registration into one image data
quality that interfered with accurate diagnosis. Mahnken set [63]. Compared to conventional half-scan, TSFF with
et  al. also reported highest contrast-to-noise ratio and best averaging resulted in significantly lower prevalence of inho-
correlation with LGE-MRI and histology for the 80 kV proto- mogeneous myocardial signal (65% vs 90%; p  =  0.002),
col compared to 100 and 120 kV [68]. Reimann et al. demon- poor contrast between LV lumen and myocardium (15% vs
strated feasibility of low tube voltage and current protocol 63%; p  <  0.001), and streaking artifact (5% vs 63%;
with doses as low as 1.19 mSv [69]. p < 0.001). When compared to the gold standard, LGE-MRI,
Another strategy for dose lowering is using prospective TSFF with averaging also demonstrated better agreement,
electrocardiographic (ECG) gating. Prospective triggering higher interobserver agreement, and reproducibility than did
can significantly reduce radiation dose to as low as 1 mSv half-scan (Fig. 64.8). For best results, we use a combination
showing good agreement with both retrospective ECG gat- of all the above; low tube settings of 80 kV and 370 mAs,
ing and LGE-MRI [70, 71]. prospective ECG gating, cardiac shuttle mode for the dual-
In addition, the high-pitch mode available on second and source CT with TSFF reconstruction algorithm, and image
third dual-source CT scanners can lead to further reduction averaging. The result is much improved image quality with
in radiation dose. Using a high-pitch mode, Geotti et  al. relatively low-radiation dose of 1.8 mSv.
achieved ultra-low radiation dose of 0.89 ± 0.07 mSv with
sensitivity, specificity, and diagnostic accuracy of 90.0%, I maging for ECV Estimation
92.9%, and 91.7%, respectively, for MI diagnosis on per-­ ECV is calculated with the following equation:
patient analysis as compared to MRI [65]. ECV  =  (ΔHUm/ΔHUb)  ×  (1  −  Hct), where ΔHUm is the
These methods might have reduced radiation exposure to change in attenuation of the myocardium in Hounsfield unit
unpreceded levels, but this was not the case with the persis- (HU), ΔHUb is the change in attenuation of the blood pool,
tent poor signal-to-noise ratio, inhomogeneous myocardial and Hct is the hematocrit level. ΔHU  =  HUlate-HUpre,
signal, and streaking artifacts inherit for LE-CT. As a solu- where HUlate and HUpre are attenuation values at late p­ hase
tion, we adopted a novel image reconstruction algorithm and pre-contrast CT, respectively. This means that for ECV
known as targeted spatial frequency filtration (TSFF). TSFF calculation, we need to obtain a pre-contrast, post-­contrast
was originally developed for cardiac dynamic perfusion CT (late phase) studies and a hematocrit value.
using the cardiac shuttle mode which acquires images at two Same scanning parameters should be applied in pre- and
alternating table positions, to produce a hybrid algorithm of post-contrast scans to allow for the next step which is n­ on-­rigid
the conventional half-scan and full-­scan reconstruction in registration/subtraction of both scans. This is followed by cal-
dual-source CT [72]. The benefit of this algorithm is that it culating change in blood pool attenuation values by the used
can achieve both high temporal resolution and diminished software application and finally providing a hematocrit value
artifactual variations in CT number which means improved to obtain an ECV map. It is worth noting that the pre-­contrast
stability of myocardial signal and better image quality. We study can also be used for calcium scoring [52].
64  CT’s Role for Myocardial Viability Assessment 839

a b c d

Fig. 64.8  Comparison between c conventional half-scan (a), TSFF (b) and TSFF with averaging (c). Best contrast-to-noise ratio and scar visual-
ization is achieved by TSFF with averaging technique in comparison with LGE-MRI (d) (arrows)

of transmurality is advised in clinical setting (≤25%,

How to Interpret 26–50%, 51–75%, and ≥76%). The location of MI should
be accurately reported according to the 17-segment model
Image Post-processing of the American Heart Association and tracked back to its
Before starting reading of LE-CT scan, some image setting culprit artery.
manipulation can lead to improved MI visualization. The MO or no-reflow phenomenon is an established compli-
best combination of image settings includes: multiplanar ref- cation of coronary reperfusion therapy for acute MI that
ormation (MPR), slice thickness between 4 and 8 mm, aver- needs to be reported. MO can be seen at the core of acute MI
age intensity projection (AIP), and narrow window settings; as a subendocardial rim of hypodensity surrounded with the
window width (WW) of 150–200 and window level hyperenhanced infarcted tissue (Fig. 64.10). Chronic infarct-
(WL) ≃ 100 (Fig. 64.9). MPR allows assessment in different related complications such as ventricular aneurysms and
left ventricular axes including most importantly the short and thrombosis are also readily visible at LE-CT (Fig. 64.3).
long axes. Thick slices decrease image noise and allow for
better delineation of MI. AIP is preferred over maximum and  ide-by-Side Assessment with Arterial
S
minimum intensity projections (MIP and MinIP respec- Phase Scan
tively) as MIP causes strong contrast accentuation which can In general, three patterns which have different prognostic
overestimate infarct’s size while MinIP causes accentuation implications can be encountered when assessing LE-CT side
of the inherit low-density artifacts causing grainy image and by side with early arterial phase or perfusion CT scans. In the
infarct’s size underestimation. Finally, narrow window set- first, the arterial phase shows no perfusion defect, while the
tings achieve higher contrast-to-noise ratio than wide win- late phase shows a hyperenhanced lesion. In the second, the
dow settings. arterial phase shows a perfusion defect that disappears at late
phase to be replaced with hyperenhancement. In the third
The Basics pattern, an early perfusion defect persists within the core of
As already known, infarcted myocardium appears as areas late hyperenhancement [74]. The residual perfusion defect
of delayed hyperenhancement, hence the common saying described in the third scenario represents mainly post-­
“bright is dead.” As basic as it may sound, this mantra is not reperfusion injury in acute MI, i.e., microvascular obstruc-
entirely accurate. In many conditions, the late-enhanced tion. It may also represent a trickier condition where the
(bright) myocardium is not necessarily infarcted (dead) but residual perfusion defect is nothing but lipomatous metapla-
rather reflects an expanded extracellular matrix; a condition sia in long-standing MI.  Clinical history, myocardial wall
that is encountered in many myocardial diseases including thinning, and fat density on pre-contrast scans allow for easy
non-ischemic cardiomyopathies, amyloidosis, sarcoidosis, differentiation between MO and fat metaplasia. This side-by-
and many others. For this reason, differentiating ischemic side assessment is of utmost importance for prognosis pre-
from non-ischemic late enhancement is too fundamental to diction as will be discussed.
ignore. Late enhancement patterns on CT are identical to
those on LGE-MRI.  A typical infarction at late enhance-  ow to Predict Functional Recovery
H
ment study is limited to a vascular territory and always starts on Cardiac CT
at a subendocardial location because of the trans-myocar- The role of cardiac CT extends beyond its ability to detect
dial perfusion gradients known as the wavefront phenome- infarction to providing vital information on possible clinical
non of myocardial necrosis [73]. Depending on the duration outcomes and prognosis after acute MI.  Several studies
of coronary occlusion, variable degrees of transmural infarc- assessed one or more of the most important CT-derived prog-
tion can result. Semi-quantitative assessment of the degree nosticators including transmural extension, enhancement
840 A. Hamdy and K. Kitagawa

1 mm
thickness

WW 50
Min. IP WL 100

Average IP
6 mm thickness
WW200, WL100

Max. IP WW 500
WL 100

WW 200
WL 40

Fig. 64.9  Different post-processing image parameters. Best MI visualization is achieved with thick slices, average IP, and narrow window
settings (center)

a b c

Fig. 64.10  Reperfused anteroseptal acute MI in long-axis view showing hypodense perfusion defect at arterial phase CT (a), that persists through
LE-CT scan (b), representing post-reperfusion injury; MO (arrows). LGE-MRI performed 20 days later shows remnants of MO (c)
64  CT’s Role for Myocardial Viability Assessment 841

pattern revealed from side-by-side assessment of early arte- Infarct Size

rial and late-phase scans and finally infarct’s size. When assessing the impact of acute MI size as quantified
from cardiac CT on functional recovery, Lessick et al. found
Transmural Extent that the presence and size of early perfusion defects and late
Wada et al. compared transmural with subendocardial early enhancement were closely related to myocardial dysfunction
arterial perfusion defects in reperfused anterior acute MI on follow-up [81]. Sato et  al. also showed higher risk of
patients [75]. Although no functional differences were MACE in patients with large late enhancement size detected
observed at the acute phase, the 6-months’ scans showed by post-PCI LE-CT [82]. On arterial phase CT, Kühl et al.
poor recovery of LV function in transmural MI group and confirmed that the degree of transmural involvement and
good recovery in subendocardial MI group (p  =  0.03). extent of early rest perfusion defects are strongly linked to
These results were confirmed later on by Shapiro et al. [76]. adverse outcomes in non ST-segment elevation MI patients
On post-PCI LE-CT, Sato et  al. demonstrated that LV [83].
remodeling (p  =  0.001) and number of re-hospitalization
for heart failure (p  =  0.0017) were more significantly
observed in transmural MI compared to subendocardial MI  yocardial Viability Assessment as 
M
groups [77]. a Part of the Comprehensive Cardiac CT:
The One-­Stop Shop Around the Corner
Enhancement Pattern
Koyama et al. were the first to compare functional recovery In classic cardiology practice, most IHD patients still have
among different patterns of enhancement observed by com- to undergo multiple investigations for comprehensive
bined assessment of early- and late-phase scans. Infarctions assessment of function, morphology, perfusion, and viabil-
showing persistent hypodensity in late scan showed the ity prior to establishing a diagnosis and setting up a manage-
greatest wall thinning and lowest ejection fraction at inter- ment plan. Finding a single comprehensive assessment
mediate and long follow-up scans followed by those showing modality would significantly cut expenses, save time,
nonpersistent hypodense perfusion defects and finally those improve diagnosis, and make it easier for an already
with no perfusion defect at early scan [74]. Recently, this exhausted cardiac patient. Among the currently available
prognosticator gained more momentum and was evaluated cardiac imaging techniques, CMR and CT have the potential
by using post-PCI LE-CT. In one study by Ogasawara et al., to be such comprehensive tests. Despite CMR is superior in
17.5% of reperfused acute MIs showed hypodense areas at terms of function and viability assessment, its use for coro-
LE-CT; all of which were transmural infarctions with signifi- nary imaging is not as robust. On the contrary, cardiac CT’s
cantly higher infarct volumes. Both hypoenhanced lesions ability to visualize coronary arteries is surpassed only by
on late scans and larger MI sizes could predict major adverse invasive angiography and has proven – at the same time – to
cardiovascular events (MACE). However, the interesting part be of reliable diagnostic accuracy for function, perfusion,
is that the incidence of MACE was significantly higher in and viability assessment.
patients with hypoenhancement at delayed scans compared A typical comprehensive cardiac CT protocol comprises
to those without hypoenhancement regardless of infarct’s routine calcium scoring, coronary angiography, function or
volume (p  =  0.003) denoting a superior prognostic perfor- cine images, stress perfusion scan, and finally LE-CT. The
mance [78]. In another study, Watabe et al. also demonstrated protocol can be tailored to match personalized patient’s care,
that heterogeneous enhancement (defined by the authors as e.g., addition of pre-contrast scan for ECV estimation. Some
concomitant presence of hyper- and hypoenhancement centers also prefer to start with perfusion study before coro-
within the infarcted myocardium at LE-CT) along with nary angiography or to combine rest and stress-induced LV
infarct’s relative CT density of more than 2.2 could predict function assessment. In a comprehensive protocol, an aver-
MO and had a significant association with left ventricular age of 120 ml contrast material is used throughout different
remodeling [79]. phases of the study which is sufficient to perform a late
Kim et al. assessed both transmurality and enhancement ­phase viability scan with no need for additional administra-
patterns as prognostic factors and concluded that transmu- tion. Recent advances in CT technology have made it possi-
ral early perfusion defects had better correlation with fol- ble to perform a comprehensive cardiac CT study with
low-up myocardial dysfunction than did transmural radiation doses of approximately 10 mSv [63, 84, 85]. Even
delayed enhancement and that both had better correlation more, ultra-low dose of 2.24 mSv was reported [86]. A com-
than subendocardial abnormalities. They also confirmed prehensive protocol not only highlights functionally signifi-
that persistent hypodense defects in late-phase CT were cant coronary plaques but also improves both perfusion and
the most powerful, single-independent predictor of myo- viability assessment. As mentioned before, a perfusion defect
cardial dysfunction [80]. on perfusion CT can be attributed to ischemic but viable
842 A. Hamdy and K. Kitagawa

Fig. 64.11  Post-PCI, 66-years-old female with no history of previous detect an irreversible ischemic defect at the inferior wall due to an old
MI. Upper row of images represents dynamic perfusion CT and shows silent MI.  Bock arrows represent infarctions; open arrows represent
anterior and inferior walls perfusion defects. LE-CT (lower row) could reversible ischemic defects

myocardium, MO, or previous MI. Side-by-side interpreta- defects, and late enhancement (cine+ rest+ LE) was com-
tion of perfusion and late-phase scans helps identify false- pared to each individual component separately using the
positive perfusion defects particularly in post-revascularization universal definition of MI as the gold standard. The com-
patients and hence increase specificity of perfusion CT bined assessment yielded the highest diagnostic accuracy on
(Fig. 64.11). This is important as well in patients with previ- per-patient analysis (90%) with sensitivity of 88% and spec-
ous silent MI coming for initial assessment of their CAD. We ificity of 92%.
demonstrated that in patients with obstructive CAD but with-
out history of previous MI referred for comprehensive car-
diac CT, up to 20% of perfusion defects on dynamic perfusion Future Directions
scan were attributable to silent MI as revealed from LE-CT
[87]. However, Bettencourt et al. investigated whether diag- Cardiac CT is expected to play a greater role in viability
nostic accuracy of a comprehensive protocol to detect sig- assessment and to witness more involvement in everyday
nificant CAD is improved by adding a late-phase viability clinical practice, only when certain issues are to be addressed.
scan and reported that LE-CT did not have such additional The first of these is standardization of scan protocols includ-
value [88]. These results, however, are simply explained by ing contrast material volume, method of its delivery, and tim-
(1) the presence of concomitant reversible ischemic perfu- ing of image acquisition. Much effort has been and is still
sion defects in the majority of patients with previous MI, (2) being dedicated for the sake of radiation dose lowering. With
the presence of MI in few patients with normal coronary next generations of state-of-the-art CT scanners and with the
arteries, and (3) the lack of per-segment analysis in the study. help of more optimized protocols, radiation doses are
It is common sense to assume that diagnosis of MI expected to go even lower. Much attention has also to be paid
becomes more reliable when assessing different phases of for continuous image quality improvement through develop-
the comprehensive study. In an interesting research by ing effective image reconstruction algorithms. More valida-
Ghoshhajra et al., patients with intermediate-to-high proba- tion studies and large multicenter researches have to be
bility for CAD were recruited for comprehensive cardiac carried out to assess diagnostic performance and prognostic
CT protocol [84]. Assessment of infarctions using a combi- power of viability cardiac CT. ECV is also an attractive area
nation of regional wall motion abnormalities, rest perfusion of research in cardiac CT and holds potentials beyond viabil-
64  CT’s Role for Myocardial Viability Assessment 843

ity assessment. Comprehensive cardiac CT is indeed an 15. Swoboda PP, Plein S. Established and emerging cardiovascular mag-
netic resonance techniques for prognostication and guiding therapy
invaluable tool that can assess anatomy, function, perfusion, in heart failure. Expert Rev Cardiovasc Ther. 2014;12(1):45–55.
and viability in one scan and needs to be validated and tai- 16. Medicare, C.f. and M. Services. Medicare provider utilization and
lored in a fashion that is suitable for clinical setting. Finally, payment data. Retrieved from https://www.cms.gov/Research-
dual-energy CT is expected to gain more popularity, but Statistics-Data-and-Systems/Statistics-Trends-and-Reports/Medicare-
Provider-Charge-Data/Physician-and-Other-Supplier2013.html
extensive work is still needed to reduce artifacts and improve 17. Shriki JE, et al. Incidental myocardial infarct on conventional non-
image quality before it can be favored over conventional gated CT: a review of the spectrum of findings with gated CT and
single-energy CT for infarct detection. cardiac MRI correlation. Am J Roentgenol. 2012;198(3):496–504.
18. Kannel WB, Abbott RD. Incidence and prognosis of unrecognized
myocardial infarction: an update on the Framingham study. N Engl
J Med. 1984;311(18):1144–7.
References 19. Baroldi G, et al. Lipomatous metaplasia in left ventricular scar. Can
J Cardiol. 1997;13(1):65–71.
1. Hammermeister K, DeRouen T, Dodge H. Variables predictive of 20. Su L, Siegel JE, Fishbein MC. Adipose tissue in myocardial infarc-
survival in patients with coronary disease. Selection by univariate tion. Cardiovasc Pathol. 2004;13(2):98–102.
and multivariate analyses from the clinical, electrocardiographic, 21. Kami D, et al. Cardiac mesenchymal progenitors differentiate into
exercise, arteriographic, and quantitative angiographic evaluations. adipocytes via Klf4 and c-Myc. Cell Death Dis. 2016;7(4):e2190.
Circulation. 1979;59(3):421–30. 22. Pouliopoulos J, et al. Intramyocardial adiposity post-myocardial
2. Harris PJ, et al. Survival in medically treated coronary artery dis- infarction: new implications of a substrate for ventricular tachycar-
ease. Circulation. 1979;60(6):1259–69. dia. Circulation. 2013;128(21):2296–308. https://doi.org/10.1161/
3. Muhlbaier LH, et  al. Observational comparison of event-free CIRCULATIONAHA.113.002238
survival with medical and surgical therapy in patients with coro- 23. Sasaki T, et al. New insight into scar-related ventricular tachycar-
nary artery disease. 20 years of follow-up. Circulation. 1992;86(5 dia circuits in ischemic cardiomyopathy: fat deposition after myo-
Suppl):II198–204. cardial infarction on computed tomography--a pilot study. Heart
4. Emond M, et al. Long-term survival of medically treated patients in Rhythm. 2015;12(7):1508–18.
the Coronary Artery Surgery Study (CASS) Registry. Circulation. 24. Ichikawa Y, et  al. Adipose tissue detected by multislice com-

1994;90(6):2645–57. puted tomography in patients after myocardial infarction. JACC
5. Burns RJ, et al. The relationships of left ventricular ejection frac- Cardiovasc Imaging. 2009;2(5):548–55.
tion, end-systolic volume index and infarct size to six-month 25. Ahn SS, et  al. CT detection of subendocardial fat in myocardial
mortality after hospital discharge following myocardial infarction infarction. Am J Roentgenol. 2009;192(2):532–7.
treated by thrombolysis. J Am Coll Cardiol. 2002;39(1):30–6. 26. Gupta M, et  al. Non-contrast cardiac computed tomography can
6. Baker DW, et al. Management of heart failure: III. The role of revas- accurately detect chronic myocardial infarction: validation study. J
cularization in the treatment of patients with moderate or severe left Nucl Cardiol. 2011;18(1):96–103.
ventricular systolic dysfunction. JAMA. 1994;272(19):1528–34. 27. Nance JW, et  al. Myocardial calcifications: pathophysiology, eti-
7. Allman KC, et al. Myocardial viability testing and impact of revas- ologies, differential diagnoses, and imaging findings. J Cardiovasc
cularization on prognosis in patients with coronary artery dis- Comput Tomogr. 2015;9(1):58–67.
ease and left ventricular dysfunction: a meta-analysis. J Am Coll 28. Freundlich I, Lind T. Calcification of the heart and great vessels.
Cardiol. 2002;39(7):1151–8. CRC Crit Rev Clin Radiol Nucl Med. 1975;6(2):171–216.
8. Shaw LJ, et al. Coronary computed tomographic angiography as a 29. Gowda RM, Boxt LM. Calcifications of the heart. Radiol Clin N
gatekeeper to invasive diagnostic and surgical procedures: results Am. 2004;42(3):603–17.
from the multicenter CONFIRM (Coronary CT angiography evalu- 30. Nieman K, et  al. Differentiation of recent and chronic myocar-
ation for clinical outcomes: an international multicenter) registry. J dial infarction by cardiac computed tomography. Am J Cardiol.
Am Coll Cardiol. 2012;60(20):2103–14. 2006;98(3):303–8.
9. Gray WR, et  al. Computed tomography for localization and siz- 31. Baer FM, et  al. Comparison of low-dose dobutamine–gradient-­
ing of experimental acute myocardial infarcts. Circulation. echo magnetic resonance imaging and positron emission tomog-
1978;58(3):497–504. raphy with [18F] fluorodeoxyglucose in patients with chronic
10. Higgins CB, et  al. Evaluation of myocardial ischemic dam-
coronary artery disease a functional and morphological approach
age of various ages by computerized transmission tomogra- to the detection of residual myocardial viability. Circulation.
phy. Time-dependent effects of contrast material. Circulation. 1995;91(4):1006–15.
1979;60(2):284–91. 32. Hoffmann U, et  al. Acute myocardial infarction: contrast-­

11. Doherty PW, et al. Detection and quantitation of myocardial infarc- enhanced multi–detector row CT in a porcine model 1. Radiology.
tion in vivo using transmission computed tomography. Circulation. 2004;231(3):697–701.
1981;63(3):597–606. 33. Nikolaou K, et  al. Assessment of myocardial infarctions using
12. Huber DJ, Lapray J, Hessel S.  In vivo evaluation of experimen- multidetector-row computed tomography. J Comput Assist Tomogr.
tal myocardial infarcts by ungated computed tomography. Am J 2004;28(2):286–92.
Roentgenol. 1981;136(3):469–73. 34. Francone M, et al. ECG-gated multi-detector row spiral CT in the
13. Gerber BL, et  al. Characterization of acute and chronic myo-
assessment of myocardial infarction: correlation with non-invasive
cardial infarcts by multidetector computed tomography com- angiographic findings. Eur Radiol. 2006;16(1):15–24.
parison with contrast-enhanced magnetic resonance. Circulation. 35. Sanz J, et  al. Detection of healed myocardial infarction with

2006;113(6):823–33. multidetector-­row computed tomography and comparison with car-
14. Lardo AC, et al. Contrast-enhanced multidetector computed tomog- diac magnetic resonance delayed hyperenhancement. Am J Cardiol.
raphy viability imaging after myocardial infarction characteriza- 2006;98(2):149–55.
tion of myocyte death, microvascular obstruction, and chronic scar. 36. Mahnken AH, et  al. Assessment of myocardial viability in reper-
Circulation. 2006;113(3):394–404. fused acute myocardial infarction using 16-slice computed tomog-
844 A. Hamdy and K. Kitagawa

raphy in comparison to magnetic resonance imaging. J Am Coll 55. Kurita Y, et  al. Data on correlation between CT-derived and

Cardiol. 2005;45(12):2042–7. MRI-derived myocardial extracellular volume. Data Brief.
37. Henneman M, et al. Comprehensive cardiac assessment with mul- 2016;7:1045–7.
tislice computed tomography: evaluation of left ventricular func- 56. Kurita Y, et  al. Estimation of myocardial extracellular volume

tion and perfusion in addition to coronary anatomy in patients with fraction with cardiac CT in subjects without clinical coronary
previous myocardial infarction. Heart. 2006;92(12):1779–83. artery disease: a feasibility study. J Cardiovasc Comput Tomogr.
38. Henneman MM, et al. Comparison of multislice computed tomog- 2016;10(3):237–41.
raphy to gated single-photon emission computed t­omography 57. Kato M, et al. Plain computed tomography for assessment of early
for imaging of healed myocardial infarcts. Am J Cardiol. coronary microcirculatory damage after revascularization therapy
2008;101(2):144–8. in acute myocardial infarction. Circ J. 2006;70(11):1475–80.
39. Buecker A, et  al. A feasibility study of contrast enhancement of 58. Jang Y, et  al. Viability assessment after conventional coronary

acute myocardial infarction in multislice computed tomography: angiography using a novel cardiovascular interventional therapeu-
comparison with magnetic resonance imaging and gross morphol- tic CT system: comparison with gross morphology in a subacute
ogy in pigs. Investig Radiol. 2005;40(11):700–4. infarct swine model. J Cardiovasc Comput Tomogr. 2015;9(4):
40. Baks T, et al. Multislice computed tomography and magnetic reso- 321–8.
nance imaging for the assessment of reperfused acute myocardial 59. Deseive S, et al. Dual-energy computed tomography for the detec-
infarction. J Am Coll Cardiol. 2006;48(1):144–52. tion of late enhancement in reperfused chronic infarction: a com-
41. Brodoefel H, et  al. Sixty-four-MSCT in the characterization of parison to magnetic resonance imaging and histopathology in a
porcine acute and subacute myocardial infarction: determination porcine model. Investig Radiol. 2011;46(7):450–6.
of transmurality in comparison to magnetic resonance imaging and 60. Wichmann JL, et al. Diagnostic accuracy of late iodine–enhance-
histopathology. Eur J Radiol. 2007;62(2):235–46. ment dual-energy computed tomography for the detection of
42. Qu X, et  al. Acute and chronic myocardial infarction in a pig chronic myocardial infarction compared with late gadolinium–
model: utility of multi-slice cardiac computed tomography in enhancement 3-T magnetic resonance imaging. Investig Radiol.
assessing myocardial viability and infarct parameters. Eur J Radiol. 2013;48(12):851–6.
2012;81(4):e431–7. 61. Truong QA, et  al. Myocardial scar imaging by standard single-
43. Mahnken AH, et  al. Late-phase MSCT in the different stages
energy and dual-energy late enhancement CT: comparison with
of myocardial infarction: animal experiments. Eur Radiol. pathology and electroanatomic map in an experimental chronic
2007;17(9):2310–7. infarct porcine model. J Cardiovasc Comput Tomogr. 2015;9(4):
44. Choe YH, et al. Comparison of MDCT and MRI in the detection 313–20.
and sizing of acute and chronic myocardial infarcts. Eur J Radiol. 62. Lee H-J, et al. Myocardial extracellular volume fraction with

2008;66(2):292–9. dual-energy equilibrium contrast-enhanced cardiac CT in nonisch-
45. Nieman K, et  al. Reperfused myocardial infarction: contrast-­
emic cardiomyopathy: a prospective comparison with cardiac MR
enhanced 64-section CT in comparison to MR imaging 1. imaging. Radiology. 2016;280(1):49–57. https://doi.org/10.1148/
Radiology. 2008;247(1):49–56. radiol.2016151289
46. Habis M, et al. Acute myocardial infarction early viability assess- 63. Kurobe Y, et al. Myocardial delayed enhancement with dual-source
ment by 64-slice computed tomography immediately after coronary CT: advantages of targeted spatial frequency filtration and image
angiography: comparison with low-dose dobutamine echocardiog- averaging over half-scan reconstruction. J Cardiovasc Comput
raphy. J Am Coll Cardiol. 2007;49(11):1178–85. Tomogr. 2014;8(4):289–98.
47. Chiou K-R, et  al. Identification and viability assessment of
64. Brodoefel H, et al. Assessment of myocardial viability in a reper-
infarcted myocardium with late enhancement multidetector com- fused porcine model: evaluation of different MSCT contrast proto-
puted tomography: comparison with thallium single photon emis- cols in acute and subacute infarct stages in comparison with MRI. J
sion computed tomography and echocardiography. Am Heart J. Comput Assist Tomogr. 2007;31(2):290–8.
2008;155(4):738–45. 65. Goetti R, et  al. Delayed enhancement imaging of myocar-

48. Lee IH, et al. Comparison of multidetector CT with F-18-FDG-PET dial viability: low-dose high-pitch CT versus MRI.  Eur Radiol.
and SPECT in the assessment of myocardial viability in patients 2011;21(10):2091–9.
with myocardial infarction: a preliminary study. Eur J Radiol. 66. Jacquier A, et  al. Multidetector computed tomography in reper-
2009;72(3):401–5. fused acute myocardial infarction: assessment of infarct size and
49. Dwivedi G, et al. Scar imaging using multislice computed tomog- no-reflow in comparison with cardiac magnetic resonance imaging.
raphy versus metabolic imaging by F-18 FDG positron emission Investig Radiol. 2008;43(11):773–81.
tomography: a pilot study. Int J Cardiol. 2013;168(2):739–45. 67. Brodoefel H, et  al. Late myocardial enhancement assessed by

50. Sado DM, et al. Cardiovascular magnetic resonance measurement 64-MSCT in reperfused porcine myocardial infarction: diagnostic
of myocardial extracellular volume in health and disease. Heart. accuracy of low-dose CT protocols in comparison with magnetic
2012;98(19):1436–41. https://doi.org/10.1136/heartjnl-2012-302346 resonance imaging. Eur Radiol. 2007;17(2):475–83.
51. Jablonowski R, et  al. Multidetector CT measurement of myocar- 68. Mahnken AH, et al. Low tube voltage improves computed tomog-
dial extracellular volume in acute patchy and contiguous infarc- raphy imaging of delayed myocardial contrast enhancement in an
tion: validation with microscopic measurement. Radiology. experimental acute myocardial infarction model. Investig Radiol.
2014;274(2):370–8. 2007;42(2):123–9.
52. Nacif MS, et  al. Interstitial myocardial fibrosis assessed as
69. Reimann AJ, et al. Late enhancement using multidetector row com-
extracellular volume fraction with low-radiation-dose cardiac puter tomography: a feasibility study with low dose 80kV protocol.
CT. Radiology. 2012;264(3):876–83. Eur J Radiol. 2008;66(1):127–33.
53. Bandula S, et al. Measurement of myocardial extracellular volume 70. Chang H-J, et  al. Prospective electrocardiogram-gated delayed

fraction by using equilibrium contrast-enhanced CT: validation enhanced multidetector computed tomography accurately quanti-
against histologic findings. Radiology. 2013;269(2):396–403. fies infarct size and reduces radiation exposure. JACC Cardiovasc
54. Treibel TA, et al. Extracellular volume quantification by dynamic Imaging. 2009;2(4):412–20.
equilibrium cardiac computed tomography in cardiac amyloidosis. 71. Wang R, et al. Low dose prospective ECG-gated delayed enhanced
J Cardiovasc Comput Tomogr. 2015;9(6):585–92. dual-source computed tomography in reperfused acute myocar-
64  CT’s Role for Myocardial Viability Assessment 845

dial infarction comparison with cardiac magnetic resonance. Eur J 82. Sato A, et  al. Prognostic value of myocardial contrast delayed
Radiol. 2011;80(2):326–30. enhancement with 64-slice multidetector computed tomog-
72. Ramirez-Giraldo JC, et al. A strategy to decrease partial scan recon- raphy after acute myocardial infarction. J Am Coll Cardiol.
struction artifacts in myocardial perfusion CT: phantom and in vivo 2012;59(8):730–8.
evaluation. Med Phys. 2012;39(1):214–23. 83. Kühl JT, et  al. The transmural extent and severity of myocardial
73. Reimer KA, et  al. The wavefront phenomenon of ischemic cell hypoperfusion predicts long-term outcome in NSTEMI: an MDCT
death. 1. Myocardial infarct size vs duration of coronary occlusion study. JACC Cardiovasc Imaging. 2015;8(6):684–94.
in dogs. Circulation. 1977;56(5):786–94. 84. Ghoshhajra BB, et  al. Infarct detection with a comprehensive

74. Koyama Y, et  al. Assessment of reperfused acute myocardial
cardiac CT protocol. J Cardiovasc Comput Tomogr. 2012;6(1):
infarction with two-phase contrast-enhanced helical ct: ­prediction 14–23.
of left ventricular function and wall thickness 1. Radiology. 85. Williams M, et al. Cardiac and coronary CT comprehensive imag-
2005;235(3):804–11. ing approach in the assessment of coronary heart disease. Heart.
75. Wada H, et  al. Multi-detector computed tomography for imag- 2011;97(15):1198–205.
ing of subendocardial infarction-prediction of wall motion 86. Williams MC, et al. A low-dose comprehensive cardiac CT protocol
recovery after Reperfused anterior myocardial infarction. Circ J. assessing anatomy, function, perfusion, and viability. J Cardiovasc
2004;68(5):512–4. Comput Tomogr. 2013;7(1):69–72.
76. Shapiro MD, et al. Cardiac computed tomography for prediction of 87. Nakamura S, et al. Frequency of perfusion abnormality on quanti-
myocardial viability after reperfused acute myocardial infarction. J tative myocardial CT perfusion imaging in patients with no, non-­
Cardiovasc Comput Tomogr. 2010;4(4):267–73. obstructive, and obstructive coronary artery disease on coronary CT
77. Sato A, et al. Early validation study of 64-slice multidetector com- angiography. 11th annual scientific meeting abstracts, 2016. poster
puted tomography for the assessment of myocardial viability and no. 161.
the prediction of left ventricular remodelling after acute myocardial 88. Bettencourt N, et al. CAD detection in patients with intermediate-­
infarction. Eur Heart J. 2008 Feb;29(4):490–8. high pre-test probability: low-dose CT delayed enhancement
78. Ogasawara S, et al. Presence of myocardial hypoenhancement on detects ischemic myocardial scar with moderate accuracy but does
multidetector computed tomography after primary percutaneous not improve performance of a stress-rest CT perfusion protocol.
coronary intervention in acute myocardial infarction predicts poor JACC Cardiovasc Imaging. 2013;6(10):1062–71.
prognosis. Int J Cardiol. 2015;184:101–7. 89. Boussel L, et al. Assessment of acute myocardial infarction using
79. Watabe H, et  al. Enhancement patterns detected by multidetector MDCT after percutaneous coronary intervention: comparison with
computed tomography are associated with microvascular obstruc- MRI. Am J Roentgenol. 2008;191(2):441–7.
tion and left ventricular remodelling in patients with acute myocar- 90. Kang DK, et  al. Dual-energy computed tomography for integra-
dial infarction. Eur Heart J. 2016;37(8):684–92. tive imaging of coronary artery disease: principles and clinical
80. Kim T, et al. Assessment of myocardial viability using multidetec- applications. In: Seminars in ultrasound, CT and MRI. New York:
tor computed tomography in patients with reperfused acute myo- Elsevier; 2010.
cardial infarction. Clin Radiol. 2012;67(8):754–65. 91. George RT, et  al. Computed tomography myocardial perfusion
81. Lessick J, et al. Is functional improvement after myocardial infarc- imaging with 320-row detector computed tomography accurately
tion predicted with myocardial enhancement patterns at multidetec- detects myocardial ischemia in patients with obstructive coronary
tor CT? 1. Radiology. 2007;244(3):736–44. artery disease. Circ Cardiovasc Imaging. 2012;5(3):333–40.
 Part XVI
Where We Are Going: The Road Ahead
 Coronary CT Angiography
as the Gatekeeper to the Cath Lab: 65
Where Are We?

Christoph Artzner, Lynne M. Hurwitz,
and Fabian Bamberg

Coronary disease is highly prevalent, affecting approxi- (67.5% vs. 32.5%) [8]. Thus, it is not surprising that 18.3%
mately 18% of men and 11% of women over the age of of patients undergoing ICA had only evidence of moderate
65 years and 19% and 16% over the age of 75 years, respec- atherosclerosis (stenoses between 20% and 49%) and
tively [1]. For patients with stable chest pain symptoms, 40.1% of patients had minimal or no stenosis at all (all
accurately identifying patients with coronary artery disease ­stenoses <20%).
(CAD) is a central part of cardiovascular medicine. If there is Referral to ICA should be made carefully, since the
clinical concern for coronary artery disease, societal guide- interventional procedure is associated with risks and dis-
lines recommend determining patients’ pretest probability comfort for patients. Major complications while rare (less
based on history and physical examination and those catego- than 1% of procedures) include death, myocardial infarc-
rized as intermediate risk referred for noninvasive testing tion, cerebral infarction, or major embolization [9, 10].
(NIT) and those categorized as high risk referred for invasive Vascular access complications, reported to be as high as
coronary angiography [2–4]. 6% in some series, remain the leading cause of morbidity
Despite these guidelines and studies suggesting cost-­ after a cardiac catheterization procedure [11, 12].
effectiveness of selective referral to ICA [5–7], NIT prior Complications range from hematoma at femoral access,
to ICA is often not performed. Patel et al. evaluated 661,063 retroperitoneal hematoma, pseudoaneurysm, dissection,
patients undergoing elective catheterization without history arteriovenous fistulae, and infections [12]. Although stud-
of CAD in the National Cardiovascular Data Registry’s ies of closure devices have been developed to improve
CathPCI Registry between July 2009 and December 2011. patients comfort, safety and efficacy remain unclear [12].
Only 64% received at least one pre-procedural NIT [8]. The ICA with intra-arterial application of contrast media is
most common NIT was stress SPECT MPI (78.1%), fol- also associated with higher risk for contrast-­ induced
lowed by stress echo (11.6%), stress ECG (9.8%), CCTA nephropathy [13].
(2.1%), and stress CMR (0.8%) [8]. In this patient popula- While there are various NIT to assess for CAD, cardiac
tion, ICA demonstrated nonobstructive CAD (all stenoses CT angiography (CCTA) has evolved as a robust modality
<50%) in 58.4% of patients. Rates of obstructive CAD to directly detect and characterize CAD and in contrast to
exceeded rates of nonobstructive disease only among other noninvasive studies assesses the vessels directly and
those  patients with abnormal NIT and high-risk findings not the end muscle or electrical signals. Major technical
advances include substantially improved spatial (0.5  mm)
C. Artzner (*) and temporal resolution (66 msec) allowing optimal image
Department of Radiology, Duke University Medical Center, quality for a broad range of patients, as well as sensitive
Durham, NC, USA detector technology to improve IQ for calcified lesions and
Department of Diagnostic and Interventional Radiology, University additional low kV acquisitions, and iterative reconstruction
of Tuebingen, Tubingen, Germany approaches that have lowered radiation exposure to as low
e-mail: [email protected]
as one millisievert.
L. M. Hurwitz Given that great potential of a CCTA to detect and charac-
Department of Radiology, Duke University Medical Center,
Durham, NC, USA terize CAD, in this chapter, we critically review the current
scientific evidence showing whether CCTA can be effec-
F. Bamberg
Department of Diagnostic and Interventional Radiology, University tively utilized as a gatekeeper to ICA to reduce the number of
of Tuebingen, Tubingen, Germany unnecessary, purely diagnostic procedures.

© Humana Press 2019 849

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_65
850 C. Artzner et al.

 bility of CCTA to Depict

A infarction (0.19% vs. 0.43%). The data indicated that indi-
and Characterize CAD viduals who underwent CCTA in a non-acute setting were
more likely to undergo subsequent invasive cardiac proce-
Three prospective multicenter trials and hundreds of single-­ dures and have higher CAD-related spending than patients
center evaluations assessed the diagnostic performance of who underwent stress testing [21].
CCTA for identification of coronary stenoses. CCTA has a It was hoped that the recently published PROMISE trial
sensitivity from 94% to 99% with a specificity of 64–83% (Prospective Multicenter Imaging Study for Evaluation of
across a wide range of disease prevalence and stage of dis- Chest Pain) would elucidate CCTA regarding its gatekeeping
ease (acute and stable CP) for the detection of CAD. It has capabilities. It was designed to compare functional testing by
been shown that CCTA features a very favorable negative multiple modalities and anatomic evaluation of stenosis
predictive value (NPV) of 97–99% to exclude obstructive severity by CCTA [22]. In this study, more than 10,000
anatomic stenosis (<50%) which means that a CT-based symptomatic patients underwent CCTA or to functional test-
approach can effectively rule out anatomic CAD with good ing (exercise electrocardiography, nuclear stress testing, or
predictive capabilities for patients without coronary steno- stress echocardiography). The mean pretest likelihood of
ses [14–18]. Interestingly, sensitivity and specificity were obstructive CAD was 53 ± 21% and an anticipated event rate
reported not to be affected by body size or by heart rates, of 9%. The study was powered for a 20% reduction to show
whereas calcium scores >400 reduced the specificity signifi- superiority and a 10% margin for noninferiority. Over a
cantly [15, 19]. However, significant false-positive rates are median follow-up period of 25  months, the composite pri-
a result of the average to low specificity with overestimation mary end point (death, myocardial infarction, hospitalization
of CAD severity by CCTA [14, 15, 20]. This has mainly for unstable angina, or major procedural complication)
been attributed to cases with poor image quality and pres- occurred only in 3.3% of patients in the CTA group and 3.0%
ence of calcium. in the functional-testing group (adjusted hazard ratio, 1.04;
95% confidence interval, 0.83–1.29; p  =  0.75). However,
CTA was associated with fewer catheterizations showing no
Role of Morphologic CAD Assessment by obstructive CAD than was functional testing (3.4% vs. 4.3%,
CCTA as a Gatekeeper in Stable Angina p = 0.02), although more patients in the CTA group under-
went catheterization within 90  days after randomization
Based on the high diagnostic performance of CCTA to detect (12.2% vs. 8.1%). The median cumulative radiation expo-
CAD and the high negative predictive value to exclude CAD sure per patient was lower in the CTA group than in the
as compared with ICA [14–18], it has been assumed that the functional-­testing group (10.0  mSv vs. 11.3  mSv) if func-
presence of CAD can easily be excluded in a relevant portion tional NIT involved iodizing radiation. This study concluded
of patients. Therefore, it was expected that ICA rates would that initial CTA provides similar clinical outcomes over a
be decreased by the utilization of CCTA. However, despite median follow-up of 2 years [22].
single-center observations or case reports, existing studies Since PROMISE indicated similar clinical, economic,
based on anatomic assessment of CAD are controversial. A and safety-based outcomes for CCTA and functional testing,
clear trend toward gatekeeping capabilities of CCTA for ICA these results have been the subject of ongoing discussion,
was not found, relating to the low specificity regarding isch- and optimal NIT strategy has not yet been determined [22,
emia resulting from CAD and because optimal thresholds of 23]. Criticism of the PROMISE trial is that for routinely per-
stenoses might not have been found yet from a functional formed CCTA, stenosis evaluation alone discards important
perspective [14, 15]. functional data, and for perfusion analysis, myocardial per-
In 2011 a retrospective, observational cohort study by fusion scintigraphy (MPS) evaluation alone discards impor-
Shreibati et al. assessed Medicare records of 282,830 with- tant anatomic information, but a combined assessment may
out prior history of CAD who received nonemergent, nonin- identify a superior methodology [24].
vasive testing for CAD.  Compared with stress myocardial Additionally, a recent open-label, parallel-group, multi-
perfusion scintigraphy (MPS), CCTA was associated with an center trial of CCTA in patients with suspected angina due to
increased likelihood of subsequent cardiac catheterization coronary heart disease (SCOT-HEART) assessed the benefit
(22.9% vs. 12.1%), percutaneous coronary intervention of CCTA. The study included 4146 patients that were ran-
(7.8% vs. 3.4%), and coronary artery bypass graft surgery domly assigned either to standard care or standard care plus
(3.7% vs. 1.3%). CCTA was also associated with higher total CCTA. Of these participants, 47% had a baseline clinic diag-
healthcare spending, which was mainly attributable to nosis of coronary heart disease, and 36% had angina due to
CAD. At 180 days, CCTA was associated with a similar like- coronary heart disease. Compared to standard care, CTCA
lihood of all-cause mortality (1.05% vs. 1.28) and a slightly reclassified the diagnosis of coronary heart disease in 27%
lower likelihood of hospitalization for acute myocardial patients and the diagnosis of angina due to coronary heart
65  Coronary CT Angiography as the Gatekeeper to the Cath Lab: Where Are We? 851

disease in 23% patients vs. 1% and 1% (p < 0.0001). CCTA f­unctional testing, with over a quarter of these patients not
demonstrated increased certainty for both coronary heart dis- having obstructive CAD on invasive angiography [27]. Given
ease and the diagnosis of angina due to coronary heart dis- the great potential of a combined morphological and func-
ease. Interestingly the frequency of coronary heart disease tional assessment by CT for both anatomic stenosis and
increased, and the diagnosis of angina due to coronary functional significant of areas of stenosis, tremendous
showed a decreasing trend. In this study, CTA lead to an research efforts have been undertaken to develop tools and
increase of planned investigations (94 additional ICAs were techniques to also evaluate coronary and myocardial func-
indicated and 29 were cancelled) and treatments (375 initi- tion. Major developments include the derivation of the frac-
ated and 189 were cancelled) compared to standard care (8 tional flow reserve based on CCTA image data (FFRCT) as
and 1 and 95 and 14, respectively) but did not affect 6-week well as the assessment of myocardial perfusion by either
symptom severity or subsequent hospital admissions for static or dynamic scan acquisitions (Figs. 65.1 and 65.2).
chest pain [25]. In follow-up of these patients at 1.7 years, FFRCT is based on computational fluid dynamics and sim-
CTCA was associated with a 38% reduction in fatal and non- ulated maximal coronary hyperemia. Recent advances in
fatal myocardial infarction (26 vs. 42, HR 0.62, 95% CI computational fluid dynamics enable calculation of coronary
0.38–1.01; p = 0.053) [25]. The study concluded that CCTA flow and pressure fields from anatomic image data and vali-
clarifies the diagnosis, enables targeting of interventions, and dated it to invasive FFR performed during invasive catheter-
might reduce the future risk of myocardial infarction [25]. ization. This technology allows for calculation of FFRCT
Again, the topic of CCTA as a gatekeeper remains highly feasible without additional imaging or medications or radia-
controversial. A newly published prospective, single-center tion [28, 29]. Multiple studies showed superiority of CTA
study of 340 patients compared a CCTA-based approach and with FFRCT vs. CCTA.
a traditional primarily ICA-based approach. The study The DISCOVER-FLOW (Diagnosis of Ischemia-Causing
showed a decreased need for ICA in the group of patients Stenoses Obtained Via Noninvasive Fractional Flow Reserve)
that received CCTA imaging initially (14% of patients compared invasive FFR with FFRCT and CCTA in 103
received additional ICA) compared to the ICA group (100% patients. On a per-vessel basis, the accuracy (84.7 vs. 58.5%)
of patients received ICA). Over a median follow-up of and specificity (82.2 vs. 39.6%) and the positive predictive
3.3 years, rates of major adverse cardiovascular events were value (73.9 vs. 46.5%) improved substantially with similar
similar with 4.2% in the CCTA group and 3.7% in the coro- sensitivities and negative predictive values for FFR. The area
nary angiography group (adjusted hazard ratio 0.90, 95% under the receiver operating characteristic curve was 0.90 for
confidence interval 0.30–2.69, p = 0.86). Patient comfort and FFRCT and 0.75 for CCTA (p = 0.001) [30]. The DeFACTO
time to patient discharge were found to be superior for the (Determination of Fractional Flow Reserve by Anatomic
CCTA subgroup [26]. Computed Tomographic Angiography) trial evaluated
FFRCT against CTA for diagnostic accuracy of ischemia.
The study included 252 patients with 137 patients (54.4%)
 ole of Functional CAD Assessment by CCTA
R showing an abnormal FFR determined by ICA.  This study
as a Gatekeeper showed on a per-patient basis slightly lower values for diag-
nostic accuracy (73%), sensitivity (90%), specificity (54%),
The anatomical assessment of coronary artery stenosis does positive predictive value (67%), and negative predictive
not correlate well with functional significance of coronary value (84%) of FFRCT plus CT. When FFR findings were
stenosis. Thus, the hemodynamic significance of an angio- compared with obstructive CAD diagnosed by CT alone,
graphically intermediate stenosis remains relevant before area under the receiver operating characteristic curve was in
referral for revascularization treatment [20]. The goal of favor of FFR CT with AUC 0.81 (0.75–0.86) vs. AUC 0.68
CCTA like other vascular CTA exams has been for the (0.62–0.74; p < 0.001) [31]. Thus, these studies indicate that
assessment of anatomic stenosis with vessel narrowing of the inclusion of FFRCT might improve gatekeeping capabili-
50% or less considered non-flow limiting. This model for ties of CCTA.
management of vascular disease is limited in that the demon- To address the clinical utility of this new method and how
stration of end-organ ischemia is not determined from a its use may affect patient care and clinical outcomes, the
static anatomic dataset. A major criticism of the PROMISE PLATFORM (Prospective Longitudinal Trial of FFRCT
trial was that for routinely performed CCTA, stenosis evalu- Outcome and Resource Impacts) trial included 584 patients
ation alone discards important functional data [24]. Although at 11 sites with new-onset chest pain who were prospectively
CCTA has demonstrated excellent sensitivity and specificity assigned to receive either usual testing (n  =  287) or CTA/
for characterization of luminal stenosis, the moderate speci- FFRCT (n = 297). The primary end point was the percentage
ficity of CTA in the PROMISE trial increased the rate of of ICA with no significant obstructive CAD.  The patient
invasive catheterization by almost 50% compared with ­collective showed a pretest probability of obstructive CAD
 852

Summary: FFRCT is > 0.80 in all coronary vessels.

FFRCT
CORONARY ARTERIES & SYSTEMS FFRCT 0.00 0.20 0.40 0.60 0.80 1.00

Left Main Artery 0.99

Left Anterior Descending System 0.90

Left Circumflex System 0.91

Right Coronary Artery System 0.89

MAY BE FUNCTIONALLY SIGNIFICANT1,2,3

0.99
0.99

0.95
0.95
0.94

0.93
0.90
0.91
0.89
0.90
A B
FFRCT
FFRCT unavailable occlusion 0.00 0.60 0.70 0.80 0.90 1.00

MAY BE FUNCTIONALLY SIGNIFICANT MAY NOT BE SIGNIFICANT

Fig. 65.1  Example of a negative CCTA demonstrating the absence of CAD. FFRCT were calculated by HeartFlow, Redwood City, CA, USA
C. Artzner et al.
 Summary: The Left Anterior Descending System has an FFRCT £ 0.80.
FFRCT
CORONARY ARTERIES & SYSTEMS FFRCT 0.00 0.20 0.40 0.60 0.80 1.00

Left Main Artery —

Left Anterior Descending System 0.66

Left Circumflex System 0.88

Right Coronary Artery System 0.88

MAY BE FUNCTIONALLY SIGNIFICANT1,2,3

65  Coronary CT Angiography as the Gatekeeper to the Cath Lab: Where Are We?

0.70
0.88 0.88
0.71

0.68
0.66 0.66
a b
FFRCT
FFRCT unavailable occlusion 0.00 0.60 0.70 0.80 0.90 1.00

MAY BE FUNCTIONALLY SIGNIFICANT MAY NOT BE SIGNIFICANT

Fig. 65.2  Example of positive CCTA demonstrating stenosis in LAD. FFRCT were calculated by HeartFlow, Redwood City, CA, USA
853
854 C. Artzner et al.

of 49 ± 17%. Indication for ICA was FFRCT guided in 193 Therefore, it is questionable whether MPS will be able to
patients vs. 187 in usual care, and no obstructive CAD was identify vessels that would benefit from revascularization
found at ICA in 24 patients (12%) for CTA/FFRCT and 137 accurately [24, 37]. On the other hand, the addition of FFRCT
patients (73%) in the usual care arm (risk difference 61%, and atherosclerotic plaque characteristics (APCs) by coro-
95% confidence interval 53–69, p < 0.0001). Clinical event nary CCTA to coronary CTA may allow for a comprehensive
rates within 90 days were low for usual care and CTA with anatomic and functional assessment of CAD in a manner
FFRCT. Thus, the authors concluded CTA with FFRCT was a potentially promoting beneficial clinical and cost outcomes.
feasible and safe alternative to ICA and associated with a At the current stage, recruitment into the CREDENCE trial
significantly lower rate of invasive angiography showing no has been completed, and results of the study are highly
obstructive CAD [27]. ­anticipated [24].
These findings were also supported by the NXT (Analysis Future concepts to improve CCTA beyond anatomic
of Coronary Blood Flow Using CT Angiography: Next lesion assessment include APC evaluation, which has dem-
Steps). In this prospective multicenter trial, 254 patients with onstrated high agreement with invasive methods of plaque
suspected CAD were included and underwent clinically indi- assessment and has shown to improve discrimination of isch-
cated ICA. Coronary CTA was performed before ICA. The emia causing culprit coronary lesions [24, 38]. APC evalu-
area under the receiver operating characteristic curve for ates lesions regarding aggregate plaque volume (APV),
FFRCT was 0.90 versus 0.81 for coronary CTA (p = 0.0008). positive remodeling (PR), low attenuation plaque (LAP), and
Per-patient sensitivity and specificity to identify myocardial spotty calcification (SC) [39]. A single-center study by Park
ischemia were 86% and 79% for FFRCT versus 94% and et al. including 252 patients showed an association of APV
34% for coronary CTA and 64% and 83% for ICA, respec- and LAP with an increased risk of ischemia in stenosis
tively [32]. These results are very promising in that they indi- >50%, whereas PR was found to be a predictor of ischemia
cate the combination of CCTA and FFR in patients with independently of severity of stenosis [39].
CAD can decrease the rates of negative ICA procedures and As a second major development, CT perfusion imaging
thus act as a decision-maker for ICA. But to date, no study has been introduced recently. This approach becomes feasi-
addresses the gatekeeping capabilities of a comprehensive ble with further improvement of CT acquisition technique
CCTA CAD assessment including FFR when compared with and therefore higher temporal resolution to visualize myo-
functional diagnostics, like MPS, which is still widely cardial enhancement as the contrast medium passes through
accepted as standard of care. the heart [40, 41]. Based on these first feasibility studies, it
Subsequently, the Computed TomogRaphic Evaluation of remains unclear whether this technique will yield sufficient
Atherosclerotic DEtermiNants of Myocardial IsChEmia diagnostic performance and robustness to be potentially
(CREDENCE) trial was initiated as a prospective multi- implemented into clinical workflow. First prospective studies
center cross-sectional study to further investigate the perfor- are planned, but patient recruitment has not started.
mance of CCTA as a gatekeeper for ICA and to incorporate
latest CCTA data-based techniques beyond luminal stenosis
severity alone for the diagnosis of lesion-specific ischemia I mpact of CCTA in Patients with Acute Chest
including atherosclerotic plaque characteristics (APCs) and Pain on Diagnostic Cath Rates
calculation of fractional flow reserve (FFRCT) [24]. This
study is meant to mitigate the limitations and critics of the A study of Hoffmann et al. assessed 1000 patients with pre-
PROMISE trial. For this purpose, 618 participants under- sentation of acute coronary syndromes in the emergency
went clinically indicated CCTA, MPS, ICA, and FFR with department. ROMICAT-II was designed as a randomized,
either MPS or CCTA/FFRCT performed as part of the study. controlled, multicenter trial to compare an evaluation and
This design was chosen because despite availability of mul- management strategy facilitating CCTA as a first diagnostic
tiple functional imaging techniques (echocardiography, car- test with a standard emergency department evaluation for
diac magnetic resonance, MPS), MPS remained the clinical patients with acute chest pain and suggested acute coronary
standard for functional imaging to identify ischemia by syndrome. A total of 75 patients (8%) had a final diagnosis of
detecting stress-induced regional myocardial perfusion an acute coronary syndrome. ICA was performed in 54
defects [2, 3]. Although MPS can determine the severity and (11%) and 36 (7%) patients in the CCTA group and standard
extent of myocardial ischemia with high performance by group, respectively (p  =  0.06). Percutaneous interventions
single-photon emission computed tomography (SPECT) or were performed in 24 and 14 patients (p = 0.14) and bypass
positron emission tomography (PET) at a per-patient level surgery in 5 and 4 patients (p = 0.99) in the CCTA group and
[24, 33, 34]. However, its accuracy to correctly discriminate standard group, respectively. In this study, more diagnostic
ischemia on a per-vessel basis is less robust [24, 35] and testing was performed in the CCTA group than in the
compromised in patients with multi-vessel CAD [24, 36]. standard evaluation group (p  <  0.001), and CTA was
­
65  Coronary CT Angiography as the Gatekeeper to the Cath Lab: Where Are We? 855

d­ emonstrated to reduce the length of the stay by 7.6 h, and For stable angina, large, randomized trials found no supe-
less patients were admitted to an observation unit but dis- rior methodology comparing the standard of care and a
charged directly from the emergency department (47% vs. CCTA-based strategy but found slight increase of ICA rates
12%). From a cost-effectiveness standpoint, both strategies in the CCTA groups. For acute chest pain, both strategies
were evaluated as equivalent. Number of diagnostic tests and showed an identical safety profile with similar proportion of
radiation exposure were found to be higher in the CCTA purely diagnostic ICAs, but with the tendency to have less
group. No acute coronary syndromes were missed, and no negative findings in ICA and a higher likelihood to diagnose
significant differences in major adverse cardiovascular CAD compared with traditional standard of care, which in
events were found during 28 days of follow-up [42]. consequence does not result in functional apparent symp-
A further study assessing a similar scenario was per- toms in all subjects. It is hypothesized that these subjects
formed by Litt et al. In this study 1370 patients were included without functional impairment, who receive treated accord-
with a randomized assignment of 2:1 to a CCTA-based or a ingly to current guidelines, will have improved survival
conventional strategy. Of the 908 patients in the CCTA sub- rates. Nevertheless, this assumption is not supported by
group, 767 individuals received a CCTA. ICA was performed available scientific evidence.
in 37 of 908 patients in the CCTA subgroup vs. 18 of 462 Despite the sole use of CCTA for anatomic assessment
patients in the traditional care group at index visit. The rate of CAD, FFRCT along with anatomic assessment of CAD
of maximal stenosis <50% was lower in the CCTA subgroup are showing promising results and potential for more
than in the traditional care group, 9 of 37 and 10 of 18 robust decision-making potential for patient management
patients, respectively; and hence the rate of significant steno- and long-­term outcomes. Prospective, multicenter studies
sis was higher in the CCTA group than in the traditional care are currently under way, and results are expected to be
group, 28 of 37 and 8 of 18 patients, respectively. Within released soon.
30 days after index visit, the use of invasive angiography was
similar (5.1% and 4.2%; difference, 0.9%; 95% CI, −4.8 to
6.6) or in the rate of revascularization (2.7% and 1.3%; dif- References
ference, 1.4%; 95% CI, −4.3 to 7.0). Nevertheless, negative
findings on invasive angiography were less likely in the 1. Go AS, Mozaffarian D, Roger VL, Benjamin EJ, Berry JD, Blaha
MJ, Dai S, Ford ES, Fox CS, Franco S, Fullerton HJ, Gillespie C,
CCTA subgroup (29% vs. 53%; difference, −23.7% points; Hailpern SM, Heit JA, Howard VJ, Huffman MD, Judd SE, Kissela
95% CI, −48.8 to 3.3). There was no significant between-­ BM, Kittner SJ, Lackland DT, Lichtman JH, Lisabeth LD, Mackey
group difference in the likelihood of a repeat emergency RH, Magid DJ, Marcus GM, Marelli A, Matchar DB, McGuire DK,
department visit, hospitalization, or cardiologist office visit Mohler ER 3rd, Moy CS, Mussolino ME, Neumar RW, Nichol G,
Pandey DK, Paynter NP, Reeves MJ, Sorlie PD, Stein J, Towfighi
or rate of major cardiac events. The study showed similar A, Turan TN, Virani SS, Wong ND, Woo D, Turner MB, American
results compared to Hoffmann et  al. with shorter hospital Heart Association Statistics C, Stroke Statistics S.  Heart disease
stay and more direct discharges from emergency department. and stroke statistics–2014 update: a report from the American
Overall, coronary disease was more likely to be diagnosed in Heart Association. Circulation. 2014;129(3):e28–e292. https://doi.
org/10.1161/01.cir.0000441139.02102.80.
patients in the CCTA group than in patients in the traditional 2. Fraker TD Jr, Fihn SD, Chronic Stable Angina Writing C, American
care group (9.0% vs. 3.5%; difference, 5.6% points; 95% CI, College of C, American Heart A, Gibbons RJ, Abrams J, Chatterjee
0–11.2), but prognostic impact was not assessed [43]. K, Daley J, Deedwania PC, Douglas JS, Ferguson TB Jr, Gardin
JM, O’Rourke RA, Williams SV, Smith SC Jr, Jacobs AK, Adams
CD, Anderson JL, Buller CE, Creager MA, Ettinger SM, Halperin
JL, Hunt SA, Krumholz HM, Kushner FG, Lytle BW, Nishimura
Summary and Future Outlook R, Page RL, Riegel B, Tarkington LG, Yancy CW. 2007 chronic
angina focused update of the ACC/AHA 2002 guidelines for the
Accounting for the excellent negative predictive value of management of patients with chronic stable angina: a report of the
American College of Cardiology/American Heart Association Task
CCTA for CTA, it has been a long-standing goal to utilize it Force on Practice Guidelines Writing Group to develop the focused
as a gatekeeper for invasive coronary angiography to utilize update of the 2002 guidelines for the management of patients with
resources more efficiently and further reduce patient’s treat- chronic stable angina. J Am Coll Cardiol. 2007;50(23):2264–74.
ment burden. However, this idea is not supported by avail- https://doi.org/10.1016/j.jacc.2007.08.002.
3. Klocke FJ, Baird MG, Lorell BH, Bateman TM, Messer JV,
able scientific evidence in prospective, multicenter studies Berman DS, O’Gara PT, Carabello BA, Russell RO Jr, Cerqueira
contradicting numerous case reports and single-center MD, St John Sutton MG, DeMaria AN, Udelson JE, Kennedy
observations. This scientific evidence will be challenged by JW, Verani MS, Williams KA, Antman EM, Smith SC Jr, Alpert
the CONSERVE (Coronary Computed Tomographic JS, Gregoratos G, Anderson JL, Hiratzka LF, Faxon DP, Hunt SA,
Fuster V, Jacobs AK, Gibbons RJ, Russell RO, American College
Angiography for Selective Cardiac Catheterization) trial of C, American Heart A, American Society for Nuclear C.  ACC/
(NCT01810198), which is directly testing the concept of AHA/ASNC guidelines for the clinical use of cardiac radionuclide
CCTA as a gatekeeper to ICA [44]. imaging--executive summary: a report of the American College of
856 C. Artzner et al.

Cardiology/American Heart Association Task Force on Practice JJ, Prokop M, Doevendans PA, Hunink MG, Krestin GP, de Feyter
Guidelines (ACC/AHA/ASNC Committee to Revise the 1995 PJ.  Diagnostic accuracy of 64-slice computed tomography coro-
Guidelines for the Clinical Use of Cardiac Radionuclide Imaging). nary angiography: a prospective, multicenter, multivendor study. J
J Am Coll Cardiol. 2003;42(7):1318–33. Am Coll Cardiol. 2008;52(25):2135–44. https://doi.org/10.1016/j.
4. Cheitlin MD, Armstrong WF, Aurigemma GP, Beller GA, Bierman jacc.2008.08.058.
FZ, Davis JL, Douglas PS, Faxon DP, Gillam LD, Kimball TR, 15. Budoff MJ, Dowe D, Jollis JG, Gitter M, Sutherland J, Halamert
Kussmaul WG, Pearlman AS, Philbrick JT, Rakowski H, Thys DM, E, Scherer M, Bellinger R, Martin A, Benton R, Delago A, Min
Antman EM, Smith SC Jr, Alpert JS, Gregoratos G, Anderson JL, JK.  Diagnostic performance of 64-multidetector row coronary
Hiratzka LF, Faxon DP, Hunt SA, Fuster V, Jacobs AK, Gibbons computed tomographic angiography for evaluation of coronary
RJ, Russell RO, Acc, Aha, Ase. ACC/AHA/ASE 2003 guideline artery stenosis in individuals without known coronary artery
update for the clinical application of echocardiography: sum- disease: results from the prospective multicenter ACCURACY
mary article. A report of the American College of Cardiology/ (Assessment by Coronary Computed Tomographic Angiography
American Heart Association Task Force on Practice Guidelines of Individuals Undergoing Invasive Coronary Angiography) trial. J
(ACC/AHA/ASE Committee to Update the 1997 Guidelines Am Coll Cardiol. 2008;52(21):1724–32. https://doi.org/10.1016/j.
for the Clinical Application of Echocardiography). J Am Soc jacc.2008.07.031.
Echocardiogr. 2003;16(10):1091–110. https://doi.org/10.1016/ 16. Marwan M, Hausleiter J, Abbara S, Hoffmann U, Becker C,

S0894-7317(03)00685-0. Ovrehus K, Ropers D, Bathina R, Berman D, Anders K, Uder M,
5. Shaw LJ, Hachamovitch R, Berman DS, Marwick TH, Lauer Meave A, Alexanderson E, Achenbach S.  Multicenter Evaluation
MS, Heller GV, Iskandrian AE, Kesler KL, Travin MI, Lewin Of Coronary Dual-Source CT angiography in patients with inter-
HC, Hendel RC, Borges-Neto S, Miller DD.  The economic con- mediate Risk of Coronary Artery Stenoses (MEDIC): study design
sequences of available diagnostic and prognostic strategies for and rationale. J Cardiovasc Comput Tomogr. 2014;8(3):183–8.
the evaluation of stable angina patients: an observational assess- https://doi.org/10.1016/j.jcct.2014.04.006.
ment of the value of precatheterization ischemia. Economics of 17. Hadamitzky M, Taubert S, Deseive S, Byrne RA, Martinoff S,
Noninvasive Diagnosis (END) Multicenter Study Group. J Am Coll Schomig A, Hausleiter J.  Prognostic value of coronary com-
Cardiol. 1999;33(3):661–9. puted tomography angiography during 5 years of follow-up in
6. Nance JW Jr, Bamberg F, Schoepf UJ.  Coronary computed patients with suspected coronary artery disease. Eur Heart J.
tomography angiography in patients with chronic chest pain: 2013;34(42):3277–85. https://doi.org/10.1093/eurheartj/eht293.
systematic review of evidence base and cost-effectiveness. J 18. Hoffmann U, Bamberg F, Chae CU, Nichols JH, Rogers IS,

Thorac Imaging. 2012;27(5):277–88. https://doi.org/10.1097/ Seneviratne SK, Truong QA, Cury RC, Abbara S, Shapiro MD,
RTI.0b013e3182631c5c. Moloo J, Butler J, Ferencik M, Lee H, Jang IK, Parry BA, Brown
7. Toutouzas K, Synetos A, Karanasos A, Drakopoulou M, Tsiamis DF, Udelson JE, Achenbach S, Brady TJ, Nagurney JT. Coronary
E, Lerakis S, Stefanadis C. Percutaneous coronary intervention in computed tomography angiography for early triage of patients
chronic stable angina. Am J Med Sci. 2010;339(6):568–72. https:// with acute chest pain: the ROMICAT (Rule Out Myocardial
doi.org/10.1097/MAJ.0b013e3181d673d7. Infarction using Computer Assisted Tomography) trial. J Am
8. Patel MR, Dai D, Hernandez AF, Douglas PS, Messenger J, Coll Cardiol. 2009;53(18):1642–50. https://doi.org/10.1016/j.
Garratt KN, Maddox TM, Peterson ED, Roe MT. Prevalence and jacc.2009.01.052.
predictors of nonobstructive coronary artery disease identified 19. Mangold S, Wichmann JL, Schoepf UJ, Caruso D, Tesche C,

with coronary angiography in contemporary clinical practice. Steinberg DH, Varga-Szemes A, Stubenrauch AC, Bayer RR 2nd,
Am Heart J. 2014;167(6):846–52 e842. https://doi.org/10.1016/j. Biancalana M, Nikolaou K, De Cecco CN.  Diagnostic accuracy
ahj.2014.03.001. of coronary CT angiography using 3rd-generation dual-source CT
9. Segal AZ, Abernethy WB, Palacios IF, BeLue R, Rordorf G. Stroke and automated tube voltage selection: clinical application in a non-­
as a complication of cardiac catheterization: risk factors and clini- obese and obese patient population. Eur Radiol. 2016. https://doi.
cal features. Neurology. 2001;56(7):975–7. org/10.1007/s00330-016-4601-2.
10. Ammann P, Brunner-La Rocca HP, Angehrn W, Roelli H, Sagmeister 20. Meijboom WB, Van Mieghem CA, van Pelt N, Weustink A, Pugliese
M, Rickli H. Procedural complications following diagnostic coro- F, Mollet NR, Boersma E, Regar E, van Geuns RJ, de Jaegere PJ,
nary angiography are related to the operator’s experience and the Serruys PW, Krestin GP, de Feyter PJ. Comprehensive assessment
catheter size. Catheter Cardiovasc Interv. 2003;59(1):13–8. https:// of coronary artery stenoses: computed tomography coronary angi-
doi.org/10.1002/ccd.10489. ography versus conventional coronary angiography and correla-
11. Meyerson SL, Feldman T, Desai TR, Leef J, Schwartz LB,
tion with fractional flow reserve in patients with stable angina. J
McKinsey JF. Angiographic access site complications in the era of Am Coll Cardiol. 2008;52(8):636–43. https://doi.org/10.1016/j.
arterial closure devices. Vasc Endovasc Surg. 2002;36(2):137–44. jacc.2008.05.024.
12. Patel MR, Jneid H, Derdeyn CP, Klein LW, Levine GN, Lookstein 21. Shreibati JB, Baker LC, Hlatky MA. Association of coronary CT
RA, White CJ, Yeghiazarians Y, Rosenfield K, American Heart angiography or stress testing with subsequent utilization and spend-
Association D, Interventional Cardiac Catheterization Committee ing among Medicare beneficiaries. JAMA. 2011;306(19):2128–36.
of the Council on Clinical Cardiology CoCR, Intervention https://doi.org/10.1001/jama.2011.1652.
CoPVDCoCS, Anesthesia, Stroke C. Arteriotomy closure devices 22. Douglas PS, Hoffmann U, Patel MR, Mark DB, Al-Khalidi HR,
for cardiovascular procedures: a scientific statement from the Cavanaugh B, Cole J, Dolor RJ, Fordyce CB, Huang M, Khan
American Heart Association. Circulation. 2010;122(18):1882–93. MA, Kosinski AS, Krucoff MW, Malhotra V, Picard MH, Udelson
https://doi.org/10.1161/CIR.0b013e3181f9b345. JE, Velazquez EJ, Yow E, Cooper LS, Lee KL, Investigators
13. Wichmann JL, Katzberg RW, Litwin SE, Zwerner PL, De Cecco P. Outcomes of anatomical versus functional testing for coronary
CN, Vogl TJ, Costello P, Schoepf UJ. Contrast-induced nephropa- artery disease. N Engl J Med. 2015;372(14):1291–300. https://doi.
thy. Circulation. 2015;132(20):1931–6. https://doi.org/10.1161/ org/10.1056/NEJMoa1415516.
CIRCULATIONAHA.115.014672. 23. Mark DB, Federspiel JJ, Cowper PA, Anstrom KJ, Hoffmann

14. Meijboom WB, Meijs MF, Schuijf JD, Cramer MJ, Mollet NR, U, Patel MR, Davidson-Ray L, Daniels MR, Cooper LS, Knight
van Mieghem CA, Nieman K, van Werkhoven JM, Pundziute G, JD, Lee KL, Douglas PS, Investigators P.  Economic outcomes
Weustink AC, de Vos AM, Pugliese F, Rensing B, Jukema JW, Bax with anatomical versus functional diagnostic testing for coronary
65  Coronary CT Angiography as the Gatekeeper to the Cath Lab: Where Are We? 857

artery disease. Ann Intern Med. 2016;165(2):94–102. https://doi. JS, Gregoratos G, Anderson JL, Hiratzka LF, Faxon DP, Hunt SA,
org/10.7326/M15-2639. Fuster V, Jacobs AK, Gibbons RJ, Russell RO, American College of
24. Rizvi A, Hartaigh BO, Knaapen P, Leipsic J, Shaw LJ, Andreini C, American Heart Association Task Force on Practice G, American
D, Pontone G, Raman S, Khan MA, Ridner M, Nabi F, Gimelli A, Society for Nuclear C. ACC/AHA/ASNC guidelines for the clini-
Jang J, Cole J, Nakazato R, Zarins C, Han D, Lee JH, Szymonifika cal use of cardiac radionuclide imaging–executive summary: a
J, Gomez MJ, Truong QA, Chang HJ, Lin FY, Min JK. Rationale report of the American College of Cardiology/American Heart
and design of the CREDENCE trial: computed tomog raphic evalu- Association Task Force on Practice Guidelines (ACC/AHA/ASNC
ation of atherosclerotic determinants of myocardial IsChEmia. Committee to Revise the 1995 Guidelines for the Clinical Use of
BMC Cardiovasc Disord. 2016;16(1):190. https://doi.org/10.1186/ Cardiac Radionuclide Imaging). Circulation. 2003;108(11):1404–
s12872-016-0360-x. 18. https://doi.org/10.1161/01.CIR.0000080946.42225.4D.
25.
Investigators S-H.  CT coronary angiography in patients 34. Kim C, Kwok YS, Heagerty P, Redberg R.  Pharmacologic stress
with suspected  angina due to coronary heart disease (SCOT-­ testing for coronary disease diagnosis: a meta-analysis. Am Heart
HEART): an open-label, parallel-group, multicentre trial. J. 2001;142(6):934–44. https://doi.org/10.1067/mhj.2001.119761.
Lancet. 2015;385(9985):2383–91. https://doi.org/10.1016/ 35. Elhendy A, Schinkel AF, Bax JJ, van Domburg RT, Valkema R,
S0140-6736(15)60291-4. Biagini E, Feringa HH, Poldermans D. Accuracy of stress Tc-99m
26. Dewey M, Rief M, Martus P, Kendziora B, Feger S, Dreger H, Priem tetrofosmin myocardial perfusion tomography for the diagnosis and
S, Knebel F, Bohm M, Schlattmann P, Hamm B, Schonenberger E, localization of coronary artery disease in women. J Nucl Cardiol.
Laule M, Zimmermann E. Evaluation of computed tomography in 2006;13(5):629–34. https://doi.org/10.1016/j.nuclcard.2006.06.128.
patients with atypical angina or chest pain clinically referred for 36. Melikian N, De Bondt P, Tonino P, De Winter O, Wyffels E,

invasive coronary angiography: randomised controlled trial. BMJ. Bartunek J, Heyndrickx GR, Fearon WF, Pijls NH, Wijns W, De
2016;355:i5441. https://doi.org/10.1136/bmj.i5441. Bruyne B. Fractional flow reserve and myocardial perfusion imag-
27. Douglas PS, Pontone G, Hlatky MA, Patel MR, Norgaard BL, ing in patients with angiographic multivessel coronary artery
Byrne RA, Curzen N, Purcell I, Gutberlet M, Rioufol G, Hink U, disease. JACC Cardiovasc Interv. 2010;3(3):307–14. https://doi.
Schuchlenz HW, Feuchtner G, Gilard M, Andreini D, Jensen JM, org/10.1016/j.jcin.2009.12.010.
Hadamitzky M, Chiswell K, Cyr D, Wilk A, Wang F, Rogers C, 37. Beller GA, Ragosta M.  Decision making in multivessel coro-

De Bruyne B, Investigators P. Clinical outcomes of fractional flow nary disease: the need for physiological lesion assessment. JACC
reserve by computed tomographic angiography-guided diagnostic Cardiovasc Interv. 2010;3(3):315–7. https://doi.org/10.1016/j.
strategies vs. usual care in patients with suspected coronary artery jcin.2010.01.007.
disease: the prospective longitudinal trial of FFR(CT): outcome and 38. Voros S, Rinehart S, Qian Z, Joshi P, Vazquez G, Fischer C, Belur
resource impacts study. Eur Heart J. 2015;36(47):3359–67. https:// P, Hulten E, Villines TC.  Coronary atherosclerosis imaging by
doi.org/10.1093/eurheartj/ehv444. coronary CT angiography: current status, correlation with intravas-
28. Taylor CA, Fonte TA, Min JK.  Computational fluid dynamics
cular interrogation and meta-analysis. JACC Cardiovasc Imaging.
applied to cardiac computed tomography for noninvasive quantifi- 2011;4(5):537–48. https://doi.org/10.1016/j.jcmg.2011.03.006.
cation of fractional flow reserve: scientific basis. J Am Coll Cardiol. 39. Park HB, Heo R, o Hartaigh B, Cho I, Gransar H, Nakazato R, Leipsic
2013;61(22):2233–41. https://doi.org/10.1016/j.jacc.2012.11.083. J, Mancini GB, Koo BK, Otake H, Budoff MJ, Berman DS, Erglis
29. Kim HJ, Vignon-Clementel IE, Coogan JS, Figueroa CA, Jansen A, Chang HJ, Min JK. Atherosclerotic plaque characteristics by CT
KE, Taylor CA. Patient-specific modeling of blood flow and pressure angiography identify coronary lesions that cause ischemia: a direct
in human coronary arteries. Ann Biomed Eng. 2010;38(10):3195– comparison to fractional flow reserve. JACC Cardiovasc Imaging.
209. https://doi.org/10.1007/s10439-010-0083-6. 2015;8(1):1–10. https://doi.org/10.1016/j.jcmg.2014.11.002.
30. Koo BK, Erglis A, Doh JH, Daniels DV, Jegere S, Kim HS,
40. Bamberg F, Becker A, Schwarz F, Marcus RP, Greif M, von

Dunning A, DeFrance T, Lansky A, Leipsic J, Min JK. Diagnosis Ziegler F, Blankstein R, Hoffmann U, Sommer WH, Hoffmann VS,
of ischemia-causing coronary stenoses by noninvasive fractional Johnson TR, Becker HC, Wintersperger BJ, Reiser MF, Nikolaou
flow reserve computed from coronary computed tomographic K. Detection of hemodynamically significant coronary artery ste-
angiograms. Results from the prospective multicenter DISCOVER-­ nosis: incremental diagnostic value of dynamic CT-based myocar-
FLOW (Diagnosis of Ischemia-Causing Stenoses Obtained Via dial perfusion imaging. Radiology. 2011;260(3):689–98. https://
Noninvasive Fractional Flow Reserve) study. J Am Coll Cardiol. doi.org/10.1148/radiol.11110638.
2011;58(19):1989–97. https://doi.org/10.1016/j.jacc.2011.06.066. 41. Rossi A, Merkus D, Klotz E, Mollet N, de Feyter PJ, Krestin
31. Min JK, Leipsic J, Pencina MJ, Berman DS, Koo BK, van
GP.  Stress myocardial perfusion: imaging with multidetec-
Mieghem C, Erglis A, Lin FY, Dunning AM, Apruzzese P, Budoff tor CT.  Radiology. 2014;270(1):25–46. https://doi.org/10.1148/
MJ, Cole JH, Jaffer FA, Leon MB, Malpeso J, Mancini GB, Park radiol.13112739.
SJ, Schwartz RS, Shaw LJ, Mauri L.  Diagnostic accuracy of 42. Hoffmann U, Truong QA, Schoenfeld DA, Chou ET, Woodard PK,
fractional flow reserve from anatomic CT angiography. JAMA. Nagurney JT, Pope JH, Hauser TH, White CS, Weiner SG, Kalanjian
2012;308(12):1237–45. https://doi.org/10.1001/2012.jama.11274. S, Mullins ME, Mikati I, Peacock WF, Zakroysky P, Hayden D,
32. Norgaard BL, Leipsic J, Gaur S, Seneviratne S, Ko BS, Ito H, Goehler A, Lee H, Gazelle GS, Wiviott SD, Fleg JL, Udelson JE,
Jensen JM, Mauri L, De Bruyne B, Bezerra H, Osawa K, Marwan Investigators R-I. Coronary CT angiography versus standard evalu-
M, Naber C, Erglis A, Park SJ, Christiansen EH, Kaltoft A, Lassen ation in acute chest pain. N Engl J Med. 2012;367(4):299–308.
JF, Botker HE, Achenbach S, Group NXTTS.  Diagnostic perfor- https://doi.org/10.1056/NEJMoa1201161.
mance of noninvasive fractional flow reserve derived from coronary 43. Litt HI, Gatsonis C, Snyder B, Singh H, Miller CD, Entrikin DW,
computed tomography angiography in suspected coronary artery Leaming JM, Gavin LJ, Pacella CB, Hollander JE.  CT angiogra-
disease: the NXT trial (Analysis of Coronary Blood Flow Using CT phy for safe discharge of patients with possible acute coronary
Angiography: Next Steps). J Am Coll Cardiol. 2014;63(12):1145– syndromes. N Engl J Med. 2012;366(15):1393–403. https://doi.
55. https://doi.org/10.1016/j.jacc.2013.11.043. org/10.1056/NEJMoa1201163.
33. Klocke FJ, Baird MG, Lorell BH, Bateman TM, Messer JV,
44. Marwick TH, Cho I, OH B, Min JK. Finding the gatekeeper to the
Berman DS, O’Gara PT, Carabello BA, Russell RO Jr, Cerqueira cardiac catheterization laboratory: coronary CT angiography or
MD, St John Sutton MG, DeMaria AN, Udelson JE, Kennedy stress testing? J Am Coll Cardiol. 2015;65(25):2747–56. https://
JW, Verani MS, Williams KA, Antman EM, Smith SC Jr, Alpert doi.org/10.1016/j.jacc.2015.04.060.
 3D Printing from Cardiac CT Images
66
Karin E. Dill, Leonid Chepelev, Todd Pietila,
and Frank J. Rybicki

Medical 3D printing is a culmination of several steps, which models that can be fabricated in the real world and imple-
broadly involve acquisition of volumetric imaging datasets mented in a clinical workflow. Accurate 3D printable seg-
(in this case, cardiac CT studies), segmentation to extract the mentation and computer-aided design (CAD) will be
relevant anatomy, post-processing, and fabrication of ana- highlighted.
tomical structures using dedicated printing hardware. The This chapter is divided into three parts. The first part
resultant 3D printed models enable detailed visual inspection reviews fundamentals of 3D printing, with an introductory
and direct manipulation of cardiac anatomy and pathology overview designed to provide a cardiovascular imager with
depicted by CT in a manner that bypasses the limitations of the foundation to understand 3D printing from cardiac CT
two-dimensionality from a computer monitor. This enables DICOM images. The second part of this chapter uses leading
intuitive model exploration, allowing for presurgical plan- commercial software (Mimics Medical and Mimics inPrint,
ning and practice on the 3D printed models for a personal- Materialise) to illustrate the process of conversion of a
ized, patient-specific intervention strategy. While 3D printing DICOM dataset to a 3D printable model. The final part
has a rich history within medicine, recent hardware and soft- briefly discusses clinical applications.
ware advances have catalyzed an exponential growth in the
interest and practical implementation of this new modality in
the clinical arena. Already established in numerous technol- Fundamentals of 3D Printing
ogy sectors, 3D printing has reached medicine and, with car-
diac CT imaging advances, resulted in successes paralleling The essential steps in creating a 3D printed model are image
those seen in other specialties. Presently, cardiac CT acquisition, anatomy segmentation, model post-processing,
imaging-­derived applications of 3D printing primarily focus and model fabrication [1, 2]. Determining the clinical setting
on clinical use of these models for patient and physician edu- and appropriateness of such a model is a prerequisite and
cation, procedure simulation, and device design. guides subsequent steps. The functionality and accuracy of a
Elements to ensure a technically optimal cardiac CT 3D printed model is predicated by collaborative interactions.
acquisition and image segmentation are further detailed. The Imagers, referring specialists, technicians, engineers, and
crucial difference between traditional segmentation and seg- other parties involved in model creation need to work closely
mentation for 3D printing is highlighted. The latter relies on to ensure appropriate execution of each step [3]. Figure 66.1
a new paradigm focused on translating mathematical models illustrates a clinical cardiovascular 3D-printing workflow.
visualized as images on the computer screen to physical

Data Acquisition
K. E. Dill
Department of Radiology, UMass Medical Center,
Worcester, MA, USA 3D printing from medical images requires volumetric data
with a signal-to-noise ratio that allows segmentation of the
L. Chepelev · F. J. Rybicki (*)
Department of Radiology, University of Ottawa Faculty of anatomy of interest. In cardiovascular imaging, sufficient
Medicine, Ottawa, ON, Canada contrast opacification of the organ in question and reducing
Ottawa Hospital Research Institute, Ottawa, ON, Canada imaging artifact is required. The majority of CT studies pro-
e-mail: [email protected] duced for diagnostic purposes already satisfy these criteria
T. Pietila by utilizing improved scanning and reconstruction tech-
Materialise USA Biomedical Engineering, Plymouth, MI, USA niques, though insufficient contrast opacification may limit

© Humana Press 2019 859

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_66
860 K. E. Dill et al.

DICOM 3D Printed Model

Segmentation software compatible with 3D Printing CAD and Print Setup Software
PRIMARY SECONDARY STL CAD Software Printer-Specific
PROCESSING PROCESSING (As Needed) Software
(Technologist) (Radiologist) TERTIARY 3D PRINTING SETUP
PROCESSING (Engineer)
Pre-Processing STL
Example: Bone subtraction to
(Engineer+Medical Team)
Ensure Anatomy Generation
simplify vessel segmentation
Accuracy • Additional Editing • Ensure model accuracy
Segmentation
Example: Thresholding, • Smoothing Models • Orient models on build tray
• Edit Segmentation • • Combine models for
Region Growing Mirroring
• Sculpt/trim anatomy of • efficiency
Post-Processing Implantable Device Design
interest • • Select material
Example: sculpting, trimming, Extruding tissues (e.g.
hollowing lumens) • Select colors
smoothing

Fig. 66.1  Typical cardiac CT 3D printing workflow

segmentation accuracy. Specifically, minimized contrast digital subtraction angiography. In such cases, the CT imag-
doses, post-contrast saline flushes, and contrast admixture ing data typically serves as an overall scaffold, allowing
should be carefully considered in studies intended for 3D reconstruction of major cardiac chambers and major vessels,
printing since enhancement of the entirety of the imaged while echocardiography provides the data for valve segmen-
structure with high concentration contrast media is necessary tation and angiography (e.g., MRI or rotational angiography)
to optimize model quality. Contrast administration should be provides finer detail of vascular anatomy.
tailored to the imaged anatomy – in cases where the entirety
of the heart is included in the segmentation, all cardiac cham-
bers must be opacified, while cases focused on coronary Segmentation of Relevant Anatomy
arteries may require only limited opacification of the heart.
In some situations, referring physicians may request modifi- Segmentation is the second step in the 3D printing workflow
cations to the original model to include a broader portion of and involves the separation of the anatomy of interest from
the anatomy, which necessitates complete opacification of background anatomy. While many advances have been
the cardiovascular anatomy included in the field of view. achieved in developing semi- and fully automated software
In addition to appropriate contrast administration, other solutions for this task, no solution has yet been able to cir-
factors influence model quality [4]. Spatial resolution of CT cumvent the need for careful assessment by a human expert
image reconstruction is a key factor in deciding the resolu- with deep knowledge of anatomy, disease pathophysiology,
tion of the final segmented model. Ideally, the voxels within and imaging to ensure accuracy and appropriateness for a
a volumetric dataset should be isotropic with each voxel side specific clinical scenario. Optimal image quality and contrast
measuring no more than 1 mm, smaller if fine vascular detail opacification improve the automated segmentation accuracy,
requires depiction. The disadvantage of minimizing voxel thus allowing segmentation to focus on small adjustments to
sizes is the computational and manpower cost associated optimize quality. Where image quality is inferior, segmenta-
with manipulating datasets containing significantly more tion may involve bypassing all automation in favor of man-
images, some of which may not be amenable to automated ual selection of the relevant anatomy on every relevant image
segmentation. Minimizing the impact of cardiac motion, within the CT dataset. More frequently, with higher study
stair-step artifact, and breathing artifact will streamline quality, the task of segmentation involves thresholding (i.e.,
model segmentation. Finally, the post-processing itself selecting the upper and lower Hounsfield unit values for vox-
should not introduce artifact. els included in the model), followed by definition of a bound-
While the majority of relevant features of a cardiovascular ing box and manual refinement of the resultant collection of
3D printed model can be obtained from coronary CT images voxels using various sculpting tools, as discussed using a
that are now considered routine, results of CT-enabled car- practical example below.
diovascular anatomy segmentation may need to be fused Segmentation as outlined above is similar for clinical car-
with models obtained from multiple additional modalities, diac CT and by itself is not sufficient for 3D printing, since
including 3D transthoracic or transoesophageal echocar- the output remains an abstract collection of cubic or cuboid
diography, magnetic resonance imaging, or 3D rotational voxels. The transformation to a printable model occurs where
66  3D Printing from Cardiac CT Images 861

an interpolation between the vertices of the voxels is rather than the myocardium and vessel walls. To avoid tedious
­performed to define the various surfaces of a three-dimen- manual segmentation of vessel walls and the myocardium,
sional geometry. For example, a collection of cubic voxels CAD tools may be used to automatically construct walls sur-
may represent a sphere with a two-voxel-thick wall. rounding intraluminal contrast. These models can then be used
Converting this collection of voxels to a printable sphere within the same CAD software to design patient-specific
involves defining the surfaces of the outer and inner walls of instrumentation or to adjust the size of the existent general
this sphere and encoding this information into a file format tools and implants. Finally, manipulations of resultant models
that a 3D printer can convert directly or indirectly into a may be performed to optimize the printable model quality
series of instructions, such as printer head movement, fila- (wrapping or smoothing), adding or subtracting parts for bet-
ment extrusion, or advancing to the next layer. ter visualization of relevant pathology (trimming, coloring,
The most commonly used file format for 3D model repre- extruding) and planning for model presentation by adding
sentation is the Standard Tessellation Language (STL) file for- connectors between discrete isolated structures of interest.
mat. Although other file formats may supersede STL by Operations such as Boolean additions and subtractions allow
enabling more powerful model definition and annotation, model fusion and removal of a three-dimensional, precisely
essentially all medical 3D printing at the time of writing uses defined region of space, respectively. Boolean operations are
STL files. Just as typical workstations use Digital Imaging and indispensable in the creation and manipulation of additional
Communications in Medicine (DICOM) files to render patient model features superimposed upon the existing models.
data in the form of an image on a computer screen, 3D printers CAD tools capable of editing STL files to create any con-
use STL files to fabricate and describe geometric shapes. An ceivable geometry are available for a widely ranging audi-
STL file allows the definition of the surface of a 3D shape by ence and at a wide range of price points. These include
approximating it as a collection of triangles. The term tessella- industry-standard tools such as those provided by Materialise,
tion refers to the process of representing an arbitrary surface as as well as a large array of software packages that have been
a collection of connected triangles that form a mesh overlying approved by the FDA. Recent FDA commentary has clarified
this surface. Typical commercial software solutions provide that approved software is required for the major steps of
tools to minimize or eliminate possible printing failures. These medical modeling intended for clinical use [5, 6]. While free,
tools are typically seamlessly integrated with the modules open-source software is available and can be applied to med-
responsible for the conversion of voxel collections or regions ical images, it should not be used for clinical care, including
of interest to a printable model represented in an STL file. the generation of 3D models designed for patient education
One possible source of print failure is the incorporation of and patient consent for intervention. Commonly used tech-
geometry elements (non-manifold geometry) that can only niques and approaches to patient-specific 3D printed model
exist as abstract geometrical constructs. Some examples of design and creation will be practically demonstrated in the
this include shapes that meet at a single point or a single line, second part of this chapter using as an example the design of
self-intersecting surfaces, and disconnected surfaces or sur- a ventricular septal defect (VSD) patch.
faces with holes. In each of these cases, the incorporation of
these geometry elements precludes the definition of a repro-
ducible real-world shape that occupies a specific volume. Model Fabrication
Most standard software solutions currently provide rapid and
automatic built-in routines for model validation and error Model fabrication is the final step of the process. The manu-
correction. Careful review by human experts following auto- facturing process starts with the selection of the appropriate
mated adjustments is necessary to optimize and further refine materials and printing technologies to address model require-
model quality [5]. ments. For example, a model used for education alone can be
produced with lower-cost materials. Models used for clinical
practice on the other hand would have to reasonably match
Computer-Aided Design the flexibility and compliance of a human heart tissue, adhere
to a quality assurance review, and be sufficiently accurate to
Once the relevant cardiovascular anatomy has been seg- avoid providing misleading information to the clinician or
mented, converted to a three-dimensional object, and reviewed surgeon ultimately responsible for the intervention. The
by domain experts for accuracy, further model manipulation choice of a printing technology and material would therefore
typically takes place using computer-aided design (CAD) be guided by the application of the resultant models, in the
software. Since visualization of cardiac chambers and relevant context of additional factors, such as material and printer
vascular structures in cardiovascular CT imaging is achieved cost and the capabilities of the materials, ranging from
using intravenous contrast agents, segmentation of cardiovas- available colors, sterilization capability, biocompatibility,
­
cular anatomy typically reflects the intraluminal contrast and physical properties of the available materials.
862 K. E. Dill et al.

 tep-by-Step 3D Model Creation

S Additional image data reformatting may be performed at
and Printing Guide the loading step, including image realignment and voxel size
minimization using the existing voxel data to improve the
While theoretical discussion of 3D printing using cardiovas- resolution of the models resulting from subsequent segmen-
cular CT images is important, equally important is exposure tation steps.
to the specific steps involved in the preparation of a clinically
usable 3D printed model. Since the majority of time and
efforts in a 3D printing laboratory will be spent on segmenta- Segmentation of the Heart and Great Vessels
tion and additional image post-processing, we illustrate the
key concepts, typical tasks, and the overall software work- Segmentation is the essential next step to separate the rele-
flow with a case previously reported and used for education vant cardiac anatomy. Segmentation in the context of cardio-
at the 2015 annual meeting of the Radiological Society of vascular CT imaging is quite similar to segmentation in the
North America [7]. In this case, the patient in question has a established visualization applications and involves the appli-
double outlet right ventricle (DORV) with a ventricular sep- cation of manual or semi-automated approaches to separate
tal defect. Illustrated is a patient-specific customized patch the anatomical structure of interest from the background
using segmentation and computer-aided design software. anatomy.
The basic steps involved in this work will use common ter- In the most basic terms, the two key discriminators of rel-
minology [8] and will be illustrated using Materialise soft- evant voxels are typically voxel attenuation values and the
ware, including inPrint and 3-matic (Materialise Inc., presence of the selected voxels within a specific region of
Leuven, Belgium). space (a specific adjustable bounding box), typically in asso-
Segmentation software converts 2D cross-sectional imag- ciation with similar voxels (Fig. 66.4).
ing data to high-quality 3D printable models, operating on The segmentation of cardiac anatomy in the case of our
voxel data to generate 3D models represented as meshes, patient begins with the identification of the intravenous con-
typically in STL format, for printing or further manipulation. trast bolus and involves setting minimum (310 HU) and
The resultant models can also be optimized for further pro- maximum (3071 HU) Hounsfield unit thresholds to select
cessing in computational analysis, such as in flow dynamics only the voxels that correspond to intravenous contrast
for the assessment of computed fractional flow reserve. In (Fig.  66.5). This task is termed thresholding and results in
the final step, setting up model printing will be briefly dis- the selection of voxels that include those corresponding to
cussed in general terms, as 3D printer-specific software that the intravenous contrast. Unfortunately, skeletal structures
enables printing STL files from any 3D CAD application is within the bounding box are also selected.
ubiquitous and relatively simple, allowing printing tray Isolating the desired ROI from the background requires
placement, job cost estimation, and job control. the selection of a single voxel in a given ROI by simply
clicking on the desired anatomy to keep, in this case intralu-
minal contrast, in any of the CT reconstructions (Fig. 66.6).
Model Preparation Since the skeletal structures are irrelevant in this setting,
we shall remove them using the region growing semi-­
Model preparation should ideally be initiated at the point of automated algorithm. This algorithm uses a seed point placed
study protocolling and guided by the specific setting, for by the segmenter within the structure of interest to identify
example, an educational application or preoperative cardio- all the voxels that are contiguous with the voxel at the seed
vascular intervention. Once a DICOM file containing appro- point. A typical variation of algorithm includes the ability to
priately acquired CT image data is obtained, it is exported in consider only the 6 possible voxel neighbors connected face-­
a decompressed form from an institutional PACS system for to-­face or to include all 26 immediate neighbors of a given
further manipulation. DICOM decompression is an essential voxel.
step as specific image compression approaches may not be The resultant intermediate model is termed a region of
universally compatible with segmentation software. The interest (ROI) or a mask, to reflect the fact that at this point,
import of uncompressed DICOM data is a simple task in most the model is not a 3D printable object but rather a three-­
segmentation software solutions and is accessible through a dimensional collection of voxels that satisfies the three crite-
simple wizard in the inPrint software (Figs. 66.2 and 66.3). ria set above. Further manipulations are required to ensure
66  3D Printing from Cardiac CT Images 863

Fig. 66.2  Import wizard in

the inPrint software. The
specific series within a study
can be selected. The
associated images and
DICOM annotations can be
previewed to identify the
optimal series

Fig. 66.3  Loaded DICOM data within the inPrint software. The seg- and bone). The ROIs and parts are listed on the left. The three DICOM
mentation and 3D object manipulation selections are in the top left cor- image data reconstructions are shown on the right, with a window to
ner. The available operations are shown in the left middle panel (here, render 3D models on the bottom right, currently empty
guided segmentation operations including segmentation of the heart
864 K. E. Dill et al.

Fig. 66.4  Incremental refinement of cardiac segmentation. dle). Inclusion of adjacency criteria with a seed point in cardiac contrast
Thresholding in a wide bounding box (left) results in a wide region of segments isolated cardiovascular anatomy (right)
interest, which can be narrowed by decreasing bounding box size (mid-

Fig. 66.5 Thresholding
settings and resultant ROI.
The bounding box may be
adjusted by clicking and
dragging the outline of the
bounding box on the three
reconstructions (orange box
in the top right). Note the
ventricular septal defect on
the axial CT image (arrow)

3D printability. The quality of this ROI directly reflects the In the setting of a study that has not been tailored for 3D
quality of the image acquisition and the adequacy of contrast printing, further segmentation is required, akin to direct pla-
administration. Optimization of image acquisition to reduce nimetry in 3D visualization. Alternatively, unnecessary ana-
beam hardening, blooming, and motion artifact, optimize tomic parts may be trimmed manually after the 3D object has
contrast dose, and minimize isotropic voxel size is therefore been generated if a clear separation plane can be identified.
crucial in ensuring a relatively accurate and rapid For this example, high-quality acquisition has resulted in
segmentation. facile segmentation of the relevant cardiovascular anatomy,
66  3D Printing from Cardiac CT Images 865

Fig. 66.6  Application of the

region growing algorithm by
using the Isolate tool from the
Edit ROI menu (second on the
left)

Fig. 66.7  Create Part menu

(third from left in the top
row), offering options of
creating a solid or a hollow
part. In this case, a solid part
needs to be created using the
contrast-containing ROI

as defined by intraluminal contrast, and the creation of an taken to not smooth the printed part to an extent that erases
accurate ROI depicting intraluminal contrast. To enable fur- the relevant anatomy or closes relevant cavities within the
ther manipulations, an actionable 3D model needs to be gen- model, by selecting appropriate detail and filling parameters.
erated using the ROI outline (Fig. 66.7). While a smoothed model may be easy to manufacture, han-
The 3D models defined as a surface composed of trian- dle, and appreciate, physicians have to be cognizant of the
gles, describing a 3D object (as opposed to voxel collections dangers of excessive smoothing in model interpretation.
in ROI), are termed “parts.” When part creation is complete, The resultant object depicts a smoothed version of intra-
the ROI list collapses, and the part list is presented in the bot- luminal contrast and can be printed directly with or without
tom left. Furthermore, the annotations on the three CT an additional wrapping operation (typically used to further
images change to denote the outline of the created 3D object smooth model outlines, ensure printability, and fill any thin
(Fig. 66.8). holes). In this scenario, the intraluminal contrast is used
Since this object contains multiple surface imperfections merely to identify the extent and configuration of the ven-
and abrupt edges, it may be smoothed using the Smooth tool tricular septal defect in order to design a patient-specific
in the Edit Part menu (Fig. 66.9). Special care needs to be patch (Fig. 66.10).
866 K. E. Dill et al.

Fig. 66.8  Result of the solid

part creation operation. Note
the automatic switch to the
edit part menu with the
appearance of appropriate
options on the top left panel.
The parts list contains the
Contrast part. The depiction
of this part has changed on
the axial image to describe its
outer surface, which has been
smoothed during the part
creation process. The degree
of smoothing can typically be
controlled

Fig. 66.9  Smoothing operation, with minimal detail of 1 mm and option to fill cavities under 1 mm. Using the Contrast object in the selection
menu (top left), a new part called “Smooth Contrast” will be created
66  3D Printing from Cardiac CT Images 867

Fig. 66.10 Smoothed
contrast model, from the
cardiac apex (left) and from
the inferior view (right)
demonstrating the contrast
bridge between the left and
right ventricles,
corresponding to the
ventricular septal defect
(arrow)

Fig. 66.11  Using the Cut

tool from the Edit Part menu
on the intraluminal contrast
model will result in isolation
of the contrast that occupies
the septal defect

A reasonable approximation of a septal defect patch could minimally invasive procedures. For these applications,
be obtained by cutting out the contrast portion that corre- the simulation of cardiac and vessel walls is indispens-
sponds to the septal defect communication using the Cut tool able. This is enabled by the application of the Hollow
from the Edit Part menu (Fig.  66.11). The resultant model tool from the Edit Part menu. This tool creates a hollow
would provide an excellent representation of the septal defect model that completely envelops an existing model (intra-
extent and will allow implant sizing. luminal contrast), providing a reasonable simulation of
More extensive manipulations can be enabled by the walls surrounding the intraluminal contrast
emulating the myocardial and vascular walls, however. (Fig. 66.12).
The same model used for defect patch design may be Since the defect patch needs to be created from the inside,
used for further implant and tool sizing or simulating an approach needs to be made to visualize the defect. This
868 K. E. Dill et al.

Fig. 66.12  Hollow operation menu allowing the creation of a new part called “Contrast Walls” (left) by growing existing model outward, by
1.5 mm in this case. Note that the new part is larger and smoother than the previously shown smoothed contrast model

Fig. 66.13  Intermediate step in the application of the Cut tool from the menu), and points were selected at the apex and just outside both atria.
Edit Part menu. Note that in the selected model, “Contrast Walls” has By removing the outer area (dashed), the septal defect is exposed
been visualized using the bottom view (accessible from the View
66  3D Printing from Cardiac CT Images 869

Fig. 66.14  The septal defect is visualized following the cut operation (left). Multiple floating loose model fragments are removed by using the
Isolate tool from the Edit Part menu and selecting the model to keep (green, middle), resulting in a clean simulation of the ventricular septal defect

can be achieved by resecting the outer cardiac walls using

the Cut tool (Fig.  66.13). The resultant model lacks ven-
tricular sidewalls, thus exposing the septal defect
(Fig. 66.14).
Segmentation concludes with the creation of a model
depicting the relevant anatomy. Further manipulations to cre-
ate a medical device de novo are the domain of computer-­
aided design (CAD), which is carried out in the 3-matic
software. To bring the model to the 3-matic software, the
model can be exported using the Export tool from the Prepare
Print menu as an STL file (Fig. 66.15).
This model can be imported into the 3-matic software
using the built-in import wizard (Fig. 66.16). As the 3-matic
software is loaded, you will note that the 3-matic window
contains the menus and tabs on the top, the Scene Tree with
all 3D objects and their components in a hierarchical distri-
bution on the top right, the operations tab to set all relevant
operation parameters at the bottom right, and process logger
on the bottom.
The next step is to design the patch. This will be done
in three steps. First, a curve will be drawn to create the
outer outline of the septal defect patch ensuring ade-
quate defect coverage is provided (Fig.  66.17). This
curve may be partially drawn and the Close Curve algo-
rithm may be used to approximate the ends of the curve
(Fig. 66.18).
Fig. 66.15  Export model dialogue. The Contrast Walls model is being
exported to the C:\ directory, in a 1:1 scale and as an STL file, presently The next step in patch creation is the conversion of the
the most widespread file format for 3D printing newly created curve to a surface. For this, the curve is first
870 K. E. Dill et al.

Fig. 66.16  Importing the

heart simulation using the
Import Part wizard

Fig. 66.17  Creating the curve describing the defect patch from the drawn, avoiding free-hanging edges and creating good approximation
Curve > Create Curve operation. A smooth curve is created, with attract with the simulated wall. Note that the curve is incomplete here and can
curve option selected to attract the curve to the surface on which it is be closed automatically
66  3D Printing from Cardiac CT Images 871

Fig. 66.18  Closing the curve using the Curve > Close Curve operation. Please note the selection of the curve named “Curve” from the Scene Tree
object hierarchy at the top right. Once the Apply button is pressed, the curve closing operation is complete (left)

Fig. 66.19  Setup and result

of the Fill Hole Freeform
operation. Note that the curve
from the new part which was
renamed to “Patch” in the
Scene Tree (not shown) is
selected in the Entities list

separated into a new part (right click the curve in the Scene The final step in patch creation is the addition of thickness
Tree, Separate > Move to Part > Create New), followed by to the resultant abstract plane. This is achieved using the
the application of the Fill Hole Freeform operation to create Uniform Offset operation, which creates a predefined offset
a surface that describes a single plane of the patch based on a provided surface to generate a three-dimensional
(Fig. 66.19). object that can be printed (Fig. 66.20).
872 K. E. Dill et al.

Fig. 66.20  Generating a three-dimensional patch using the Design > from the plane used to create the patch. Please note the new Patch
Offset > Uniform Offset operation, with a distance of 0.75 mm to create object is selected in the Entities list
a 1.5 mm – thick patch as the offset applies in two directions outward

The septal defect patch is now ready to be printed. It may References

be exported from 3-matic using the File > Export operation,
in a process identical to that demonstrated for the cardiac 1. Mitsouras D, Liacouras P, Imanzadeh A, Giannopoulos A, Cai T,
Kumamaru KK, George E, Wake N, Caterson EJ, Pomahac B, Ho
wall simulation model. The resultant STL may then be V, Grant GT, Rybicki FJ.  Medical 3D printing for the radiologist.
printed after loading it in any of a wide range of printer-­ Radiographics. 2015;35(7):1965–88.
specific software solutions, arranging the object on the 2. Matsumoto JS, Morris JM, Foley TA, et  al. Three-dimensional

printer tray, selecting print materials, and finally submitting physical modeling: applications and experience at Mayo Clinic.
Radiographics. 2015;35(7):1989–2006.
the job to the printer. 3. Giannopoulos A, Mitsouras D, Yoo S-J, Liu P, Chatzizisis Y, Rybicki
FJ. Applications of 3D printing in cardiovascular diseases. Nat Rev
Cardiol. 2017;13(12):701–18.
Summary 4. George E, Liacouras P, Rybicki FJ, Mitsouras D.  Measuring and
establishing the accuracy & reproducibility of 3D-printed medical
models. Radiographics. 2017;37(5):1424–50.
The collective term 3D visualization refers to all image 5. Christensen A, Rybicki FJ.  Maintaining safety and efficacy for

post-­processing commonly used in current medical imag- 3D printing in medicine. 3D Print Med. 2017;3:1. https://doi.
ing, including cardiac CT.  These include multi-planar org/10.1186/s41205-016-0009-5.
6. Di Prima M, Coburn J, Hwang D, Kelly J, Khairuzzaman A, Ricles
reformatted images, maximum intensity projections, and L. Additively manufactured medical products–the FDA perspective.
volume rendering. All of these currently used manipula- 3D Print Med. 2016;2(1):1–6.
tions use DICOM images to render volumetric data on a 7. Giannopoulos AA*, Chepelev L*, Sheikh A, Wang A, Dang W,
2D computer screen. 3D printing represents a paradigm Akyuz E, Hong C, Wake N, Pietila P, Dydynski PB, Mitsouras D,
Rybicki FJ. (*Both authors contributed equally.) 3D printed ventric-
shift in representing imaging data, using novel formats to ular septal defect patch: a primer for the 2015 Radiological Society
define the surface of a 3D shape by approximating it as a of North America (RSNA) hands-on course in 3D printing. 3D Print
collection of triangles. These files can be modified and Med. 2015;1:3. https://doi.org/10.1186/s41205-015-0002-4.
printed using a variety of ­technologies for medical diagno- 8. Chepelev L, Giannopoulos A, Tang A, Mitsouras D, Rybicki

FJ.  Medical 3D printing: methods to standardize terminology and
sis, preprocedure planning, device creation, and report trends. 3D Print Med. 2017;3:4. https://doi.org/10.1186/
education. s41205-017-0012-5.
 Future Technological Advances
in Cardiac CT 67
Thomas G. Flohr, Thomas Allmendinger, Herbert Bruder,
Chris Schwemmer, Steffen Kappler,
and Bernhard Schmidt

Challenges in Cardiac CT typical SFOV of 50 cm) plus a transition angle for smooth
data weighting, in total 240–260° of fan-beam data. This cor-
Imaging of the heart with computed tomography (CT) is responds to a temporal resolution of about 2/3 of the CT
technically demanding. It is fair to say that the technological scanner’s gantry rotation time. In the center of the CT gantry,
advancement of CT during the past 15 years has been mainly where the heart is typically located, 180° of scan data (a half-­
driven by the ongoing refinement of cardiac CT. scan segment) plus the transition angle are sufficient for
Adequate visualization of the moving anatomy of the image reconstruction. In refined image reconstruction
heart and the coronary arteries requires a short “exposure approaches for cardiac CT, only the very minimum of scan
time” per image – technically speaking the CT scanner has to data needed to reconstruct that pixel is used for every image
provide high temporal resolution. The coronary arteries have pixel. Therefore, close to the isocenter, the temporal resolu-
diameters of only few millimeters, and plaque and stenosis tion is about half the gantry rotation time of the respective
need to be reliably and at best quantitatively evaluated. CT system [2].
Therefore, excellent spatial resolution is needed in addition. During the past 15  years, technical developments have
Synchronization of data acquisition with the patient’s aimed at improving the temporal resolution of CT scanners
electrocardiogram (ECG) is a prerequisite to image the heart to make cardiac CT more robust in clinical practice and
in a phase-consistent way – e. g., in the diastolic rest phase or applicable to a wider range of patients.
the end-systolic phase at higher heart rates. Finally, the radi- The most straightforward approach to reduce the “expo-
ation exposure to the patient has to be as low as reasonably sure time” per image is faster gantry rotation. Gantry rotation
achievable. A summary of challenges in cardiac CT and times dropped from 0.5 s with the first generation of four-­
recent developments may also be found in [1]. slice CT systems (resulting in a temporal resolution of
250 ms close to the isocenter) to 0.25 s with the latest genera-
tion of multi-detector row CT (MDCT) systems (correspond-
Temporal Resolution ing to 125  ms temporal resolution). With state-of-the-art
MDCT, robust visualization of the coronary arteries can be
In ECG-controlled CT examinations of the heart, dedicated achieved in clinical routine at moderate and reasonably regu-
image reconstruction approaches using only the very mini- lar heart rates but is still challenging in patients with high
mum of scan data are used to minimize the “exposure time” and irregular heart rates. Larger variability of the heart rate is
per image and to optimize temporal resolution. The mini- known to negatively affect the image quality of coronary CT
mum amount of scan data for image reconstruction in the angiography (CTA) [3]. Therefore, the administration of
entire scan field of view (SFOV) is a partial scan, comprising beta-blockers is frequently recommended to lower and stabi-
an angular range of 180° of fan-beam data (half a gantry lize the patients’ heart rates.
rotation) plus the total fan angle of the detector (~50° for a Even faster gantry rotation to further improve temporal
resolution is technically demanding because of significant
gravitational forces which are challenging in particular for
the x-ray tube design. Rotation times of less than 0.2  s,
T. G. Flohr · T. Allmendinger · H. Bruder · C. Schwemmer · S. Kappler needed to achieve a temporal resolution better than 100 ms,
B. Schmidt (*) appear to be beyond today’s mechanical limits. As an addi-
Department of Computed Tomography, Siemens Healthcare
GmbH, Forchheim, Germany tional challenge, the more the gantry rotation time is
e-mail: [email protected] decreased, the more the x-ray tube power has to be increased

© Humana Press 2019 873

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_67
874 T. G. Flohr et al.

to maintain adequate contrast-to-noise ratio (CNR) in images 88% on a per-patient basis, with no significant differences in
with shorter exposure time. As a consequence, extrapolating sensitivity or specificity in the subgroup of studies with and
from the power reserves of today’s MDCT systems about without heart rate control. The median radiation dose, how-
150 kW tube power would be needed for adequate CNR in ever, was smaller in the studies with heart rate control
the desired sub-mm images at rotation times of less than (1.6  mSv) than in the studies without heart rate control
0.2  s. Again, the required power densities in a small focal (8 mSv).
spot on the anode plate appear to be beyond today’s technical CT scanner designs with more than two x-ray tubes  –
limits. detector pairs [11], combined with refined approaches to
An alternative approach to improve temporal resolution improve temporal resolution even further by decoupling of
of mechanical CT systems is the dual-source principle. Dual-­ the source rotation from the detector rotation – appear to be
source CT (DSCT) systems are equipped with two x-ray mechanically too complex for a practical realization. An
tubes and two corresponding detectors offset by about 90° experimental CT design, referred to as “multisource interior
[4] (Fig. 67.1). Since 2005, three generations of DSCT sys- tomography” [12], has not left the concept stage so far.
tems have been commercially introduced. Enabled by simul- Multisource interior tomography intends to enable ultrafast
taneous data acquisition of both measurement systems, imaging by only irradiating a small region of interest, such as
DSCT scanners provide a temporal resolution close to a the heart, with narrow x-ray beams defined by many source –
quarter of the gantry rotation time [5] – 83 ms for the first-­ detector pairs surrounding the patient (Fig.  67.2). Images
generation DSCT, 75  ms for the second-generation DSCT, covering this region of interest can then be reconstructed
and 66 ms for the third-generation DSCT. Meanwhile, clini- using the interior tomography approach. Again, the practical
cal studies have demonstrated the potential of DSCT for realization of the proposed setup seems to be extremely chal-
clinically robust coronary CTA examinations with little or no lenging, and the clinical field of application of such systems
dependence of the diagnostic performance on the patient’s would be very limited.
heart rate [6–9]. In a meta-analysis of 33 studies published Yet another attempt to improve temporal resolution is the
between 2006 and 2011 which compared the diagnostic design of nonmechanical CT systems without mechanically
accuracy of first- and second-generation DSCT coronary rotating gantry ring. Already in 1984, an electron beam CT
angiography for the detection of >50% stenosis with invasive (EBCT) was introduced as a noninvasive imaging modality
catheter angiography as a reference standard [10], the authors for the diagnosis of coronary artery disease [13, 14]. It
found a pooled sensitivity of 98% and a pooled specificity of ­consisted of a stationary x-ray tube and a stationary detector

Fig. 67.1  DSCT with two

independent measurement
systems at an angle of 90°
(first generation) or 95°
(second and third generation)
67  Future Technological Advances in Cardiac CT 875

Fig. 67.2  Proposed setup for

a multisource interior
tomography system

ring. An electron beam was emitted from a powerful electron recently [16–21]. In essence two different strategies are pur-
gun and magnetically deflected to hit a semicircular anode sued. One attempt is based on using less scan data than theo-
surrounding the patient, thus generating an x-ray source that retically needed for image reconstruction (i.e., less than a
virtually rotated around the patient. Given the absence of half-scan interval at the scanner’s isocenter), thereby short-
mechanically moving parts, a sweep is needed as little as ening the exposure time per image. Pertinent examples are
50 ms. However, because of inherent disadvantages such as TRI-PICCS [16] or TRIM [17]. The image artifacts caused
excessive scattered radiation as a consequence of missing by the insufficient scan data range need to be corrected for,
anti-scatter grids or lack of sufficient tube power and there- e.g., by using a prior image based on a partial scan in an
fore suboptimal CNR in the images, EBCT systems never iterative image reconstruction approach and image regular-
made their way into clinical routine, and the EBCT principle ization terms that enforce image smoothness (as in TRI-­
has been abandoned by now. PICCS) or positivity (as in TRIM). The quantification of the
Unfortunately, no other convincing concept for a nonme- actual temporal resolution of these algorithms is difficult.
chanical CT design has been proposed so far, and a clinical Latest results suggest small improvements [22, 23].
introduction of such systems in the near future appears Another attempt is based on estimating the motion of the
unlikely. coronary arteries and compensating for it during image
From the very beginning of cardiac CT, approaches to reconstruction. Pertinent algorithms incorporate a motion
reduce the exposure time per image by mechanical advance- model into iterative reconstruction algorithms or approxi-
ments such as faster gantry rotation times, and design mate analytic reconstruction algorithms (see, e.g., [19, 24]).
changes such as dual-source CT, went hand in hand with The unknown motion of the coronary arteries during the CT
attempts to improve temporal resolution by algorithmic scan has to be estimated by suitable algorithms – this turns
approaches. Multi-segment reconstruction, the combination out to be a challenging problem. Most recent attempts aim at
of small scan data segments from multiple consecutive heart- obtaining a motion model of the coronary arteries by mini-
beats to build up the half-scan data segment needed for image mizing a cost function such as the image entropy [20].
reconstruction while reducing the data acquisition window Figure  67.3 shows an example of the performance of a
per heart cycle, was introduced in the first years of cardiac recently developed motion correction approach [25].
CT but is meanwhile only rarely used in clinical practice. It While algorithmic approaches to improve temporal reso-
can only improve image quality at higher heart rates when lution in cardiac CT beyond the mechanical system limits
the heart rate is regular, and it results in increased radiation appear conceptually appealing, the proposed algorithms are
dose (see, e.g., [15], who found an average radiation dose of still missing validation in large clinical studies. One approach
4.8 mSv for a 1-beat scan and 7.8 mSv for a 2-beat scan in (“Snapshot Freeze,” GE Healthcare, Waukesha, USA) has
the CORE320 trial). meanwhile been evaluated in several smaller studies. In a
Algorithmic approaches to achieve improved “virtual” study with 120 patients [26], improved image quality and
temporal resolution in the CT images beyond the mechanical coronary assessability were found in patients with higher
limits of the respective CT system have been proposed heart rate and heart rate variability. In another study [27]
876 T. G. Flohr et al.

a b

Fig. 67.3  Example for a coronary CTA image of a patient with a stent without (a) and with motion correction based on a minimization of the
in the right coronary artery (RCA), heart rate 72 bpm, systolic recon- image entropy in partial angle reconstructions (b). (See Ref. [25])
struction at 30% of the RR-interval, and temporal resolution 143 ms,

motion correction reduced the presence of motion artifacts about 14–18  lp/cm can be achieved. In practice, to avoid
and improved image quality, but did not influence the diag- excessive image noise and streak artifacts at high-contrast
nostic utility – it did not reduce the percentage of nondiag- structures, in-plane spatial resolution is reduced by smooth-
nostic images. In a randomized trial with 64 patients ing convolution kernels. Typically, about 0.4–0.5  mm in-­
receiving beta-blockers and 51 patients without heart rate plane resolution, corresponding to 10–12  lp/cm, is not
control, motion correction improved image quality and exceeded in coronary CTA.
reduced motion artifacts, but could not compensate for the Spatial resolution is not only important in the image plane
absence of beta-blockers [28]. Therefore, larger studies are but also perpendicular to it – then it is called through-plane
still needed to assess the clinical performance of current spatial resolution. Through-plane spatial resolution is mainly
motion correction algorithms, and it is currently not possible determined by the slice width of the reconstructed images
to decide on the clinical robustness of motion correction and and by the image increment. State-of-the-art ECG-controlled
on its potential to reliably improve temporal resolution in scanning of the heart with 64 or more slices with 0.5 mm,
cardiac CT. 0.6 mm, or 0.625 mm collimated slice width and overlapping
image reconstruction may result in 0.35–0.5  mm through-­
plane resolution and hence in nearly isotropic sub-millimeter
Spatial Resolution resolution to visualize the coronary arteries. Figure  67.4
illustrates the progress in spatial and temporal resolution
The spatial resolution of a CT system is limited by the size of from 4-slice CT to 64-slice CT and newer CT systems.
the individual detector pixels and by the size of the focal Yet, scanning of patients with severe coronary calcifica-
spot. With typically 700–900 detector pixels per detector row tions remains a challenge because of Ca blooming: partial
covering a SFOV of usually 500  mm in diameter, the in-­ volume artifacts as a consequence of still insufficient spatial
plane sampling distance in a modern CT system is about resolution make calcifications appear bigger as they are. This
0.28–0.36  mm at isocenter. By using techniques such as prevents reliable assessment of the coronary lumen and leads
quarter detector offset or flying focal spot to double the in-­ to overestimation of coronary artery stenosis. In a ­multicenter
plane sampling density, a maximum spatial resolution of coronary CTA trial [29], the presence of coronary calcifica-
67  Future Technological Advances in Cardiac CT 877

Fig. 67.4  Typical coronary

CTA examinations with (a) a a
4-slice CT scanner (temporal
resolution 250 ms, spatial
resolution about
0.6 × 0.6 × 1 mm3), (b) a
64-slice CT scanner (temporal
resolution 165 ms, spatial
resolution about
0.5 × 0.5 × 0.4 mm3), and (c)
a second-generation
dual-source CT (temporal
resolution 75 ms, spatial
resolution about
0.4 × 0.4 × 0.4 mm3). The
improvement of detail
visualization thanks to better
temporal and spatial b
resolution is obvious

tions with an Agatston score >1000 was the most relevant and on the CT equipment. Maintz et  al. [30, 31] found an
independent predictor of uninterpretable coronary segments, average lumen visibility of 50–59% for the majority of com-
followed by a heart rate >80  bpm and a body mass index monly used 2.5–4 mm coronary stents in a phantom experi-
>40  kg/m2. Furthermore, the limited ability of today’s CT ment carried out on 64-slice CT and first-generation DSCT,
scanners to reliably and unambiguously characterize plaque with extreme values of 3.3% for a tantalum stent and 90% for
composition (e.g., to distinguish lipid-rich and fibrous a magnesium stent. In a meta-analysis of 9 studies performed
plaques) is mainly a result of insufficient spatial resolution. with 64-slice CT [32], the authors concluded that still a rela-
Assessments of stent patency and in-stent restenosis are tively large proportion of coronary stents remains uninter-
related problems. Previous studies have demonstrated that pretable. Accordingly, only in selected patients, 64-slice CT
image quality strongly depends on the material of the stents
878 T. G. Flohr et al.

may serve as a potential alternative noninvasive method to tion. X-ray quanta that hit these separation zones between
rule out in-stent restenosis. the detector pixels do not contribute to signal generation in
Recent technical progress has resulted in improved spatial the detector. The relative area of these “dead zones” would
resolution in cardiac imaging to potentially overcome some increase significantly for further reduced dimensions of the
of the challenges of scanning patients with coronary calcifi- detector pixels, thereby reducing the radiation dose effi-
cations, plaque, and stents. The use of dedicated high-­ ciency of the detector to unacceptably low values
definition (HD) scan modes, aimed at improving both the (Fig. 67.6a).
angular and the in-plane sampling density, for the evaluation There is, however, light on the horizon. Photon-counting
of 25 coronary stents resulted in significantly lower in-stent detectors are a promising technology for future CT sys-
luminal attenuation. The mean measured in-stent luminal tems [35, 36]. They do not require optically in-transparent
diameter was significantly larger [33]. Improved in-stent layers between the individual detector pixels, which – as a
lumen assessment could be demonstrated with third-­ consequence – can be made much smaller than the detector
generation DSCT [34], mainly as a result of smaller detector pixels of today’s scintillation detectors (Fig. 67.6b). Small
pixels, smaller focal spot sizes, and refined iterative recon- detector sizes are also needed to mitigate physical effects
struction approaches. Two previously evaluated stents [30, such as pulse pileup; see section “New CT Concepts”.
31] presented a significantly improved in-stent lumen visibil- Therefore, CT systems with photon-counting detectors
ity of up to 76% and 83%, respectively, compared to visible hold promise of potentially providing improved spatial
diameters of 52% and 56.7% when using older CT equip- resolution  – at least theoretically up to a level of 0.2–
ment (Fig. 67.5). 0.25 mm comparable with catheter angiography. Increased
However, significant further improvements of spatial spatial resolution, however, goes hand in hand with
resolution with today’s CT scanner technology are unlikely. increased image noise. Consequently, to maintain adequate
The detector pixels of today’s medical MDCT systems can- CNR in the images, significantly increased radiation dose
not be made much smaller to further increase spatial resolu- is needed. This may be justified for detail examination of
tion. State-of-the-art CT detectors are based on scintillating plaques or evaluation of in-stent patency but problematic
ceramic materials that convert the x-rays into visible light. for routine cardiac scanning. From a technical perspective,
Optically reflecting material has to be inserted between the the required high power reserves to maintain adequate
individual detector pixels to prevent optical cross talk from CNR in high-resolution images are a challenge for the
one pixel to the next to avoid a degradation of spatial resolu- x-ray tube design.

Radiation Dose

As a consequence of the ongoing discussion of radiation

exposure by CT both in the public and in the scientific litera-
ture, the advancement of techniques to reduce radiation dose
in cardiac CT examinations has been a key topic for CT man-
ufacturers during the last years.
Radiation exposure of coronary CTA is significantly
influenced by the choice of the scan technique. Prospectively
ECG-triggered axial scanning has been proven to result in
significantly less radiation dose to the patient than retrospec-
tively ECG-gated spiral scanning. The pooled effective dose
in a meta-analysis of 20 coronary CTA studies was 3.5 mSv
with prospective triggering and 12.3 mSv with retrospective
gating [37]. A further reduction in radiation exposure to val-
ues below 1  mSv was demonstrated with ECG-triggered
DSCT high-pitch spiral scanning [38–41]. However, this
technique is limited to patients with low to moderate regular
heart rates. Another very efficient technique to lower radia-
Fig. 67.5  3 mm stent in a contrast-filled tube, scanned on a 64-slice tion exposure is the use of low-kV protocols, i.e., scan proto-
CT system (left) and on a third-generation DSCT system (right) in an
ECG-triggered routine scan mode, applying the highest available spa-
cols relying on 70–100 kV x-ray tube voltage instead of the
tial resolution in that scan mode. (Courtesy of T.  Gassenmaier, “standard” 120 kV setting. Because of the increased iodine
University of Würzburg, Germany) contrast at lower kV, the CNR in contrast-enhanced images
67  Future Technological Advances in Cardiac CT 879

Fig. 67.6  Schematic drawing of a X-rays TiO2 based reflector

a conventional solid-state
scintillation detector (a) used in
all commercially available
medical CT systems. The x-rays
are absorbed and converted into
visible light which is then
detected by photodiodes attached
GOS
to the backside of the detector
pixels. The detector pixels cannot
be made much smaller than today Glue
because optically in-transparent Photo-diodes
separation layers (based on TiO2)
between them are needed to
prevent optical cross talk. In b X-rays
novel photon-counting detectors,
(b) the individual detector pixels
are defined by a strong electric
field between cathode and
pixelated anodes. The absorbed Cathode
x-ray quanta directly produce
electron-hole pairs which induce
High voltage
fast current pulses. No optical
separation layers between the
detector pixels are necessary. Pixellated anodes
Therefore, the detector pixels can
be made smaller to improve
spatial resolution

increases at low kV if the radiation dose is kept constant As an add-on to the dose reduction approaches described
[42]. Vice versa, lower radiation dose is sufficient to ­maintain so far, iterative image reconstruction has found its way into
adequate CNR. Recent progress in x-ray tube design has led routine cardiac CT scanning. In an iterative reconstruction,
to the introduction of CT systems capable of providing high a correction loop is introduced into the image reconstruction
power reserves at low x-ray tube voltages of 70–100 kV. These process [46]. While the technical realization is highly
systems have the potential to enable coronary CTA at low kV vendor-­specific, all approaches aim at including the statisti-
even in obese patients without compromising CNR and thus cal properties of the acquired measurement data into the
reduce radiation dose by 49–68% [43]. image reconstruction process in a more optimal way than
To further reduce radiation dose in non-contrast-enhanced traditional filtered back projection. As a result, they main-
CT scans of the heart, e.g., for the detection and quantifica- tain or even improve high-contrast resolution and reduce
tion of coronary calcifications (“Ca-scoring”), spectral shap- image noise in low-contrast areas, which is the prerequisite
ing has been proposed [44, 45]. Tin pre-filtration of the x-ray for a potential radiation dose reduction. In particular, when
beam at 100  kV (Sn100  kV) removes unnecessary low-­ combined with other dose-efficient scan techniques, such as
energy quanta and results in a mean x-ray energy similar to ECG-triggered sequential scanning, ECG-triggered high-­
the 120  kV spectrum that is typically used for Ca-scoring. pitch scanning, and low-kV data acquisition, iterative recon-
Although the Agatston scores were systematically lower struction may enable coronary CTA at very low radiation
with Sn 100 kV than with 120 kV in a study with 70 patients dose levels [39, 47–49]. Figure 67.7 shows a representative
[45], the comparison of Agatston score categories and clinical example for a coronary CTA acquired at very low
percentile-­based cardiac risk categories showed excellent radiation dose by combining several dose reduction
agreement. Effective radiation dose was significantly lower technologies.
at Sn 100  kV than at 120  kV (0.19  mSv vs 0.82  mSv). Currently, alternative image reconstruction approaches
Currently, new approaches are being investigated to enable relying on machine learning and dictionary-based image rep-
Ca-scoring at other kV settings than 120 kV to reduce radia- resentation are being evaluated [50, 51]. In a nutshell, these
tion dose while maintaining the Agatston score. They rely on approaches decompose a low-dose image into small overlap-
dedicated beam-hardening corrections to correct the CT ping image segments and attempt to represent each of these
number of calcifications to their values at 120 kV. segments by a weighted combination of image primitives
880 T. G. Flohr et al.

r­elevant body materials, such as water, are three orders of

magnitude larger than the attenuation coefficients. This sug-
gests that phase-contrast CT should be much more sensitive
than absorption CT and allow for significantly reduced radi-
ation dose to the patient.
Unfortunately, this advantage can only be utilized at very
high spatial resolution [53], and spectacular results demon-
strating the potential of phase-contrast CT have so far been
limited to small specimen scanned at very high spatial reso-
lution; see the excised coronary artery in Fig. 67.8 [54].
At the resolution level of today’s medical CT, the relative
performance of phase-contrast CT is inferior to absorption
Fig. 67.7  Example of a coronary CTA examination in a patient with a
heart rate of 64 bpm, acquired at very low radiation dose of 0.21 mSv
CT. Increasing the resolution to a level at which the relative
by combining several dose reduction technologies: low-kV imaging at performance of phase-contrast CT is better than that of
70 kV, ECG-triggered high-pitch spiral scanning and iterative recon- absorption CT would require substantially increased radia-
struction. (Courtesy of Mannheim University, Germany) tion dose to maintain adequate CNR in the images. Therefore,
as of today, phase-contrast imaging is not an option for medi-
from a high-dose dictionary, thereby de-noising the original cal CT because of the predicted excessive radiation dose
low-dose image. The performance of such algorithms and requirements. The root cause of the disappointing dose per-
their potential to allow for reduced radiation dose in cardiac formance is the bad spatial coherence of today’s low-­
CT examinations even below the levels of “traditional” itera- brilliance x-ray sources [53]. Should new, compact x-ray
tive image reconstruction approaches will still have to be tube designs with much better spatial coherence, providing
evaluated. high x-ray photon flux from an extremely small focal spot,
To sum up, we will probably see moderate further become available the concept of phase-contrast CT is worth
improvements in the already very efficient approaches to revisiting.
reduce radiation dose in cardiac CT, but another break-
through cannot be expected. By adequate combination of the
available dose reduction techniques, CT scans of the heart CT Systems with Photon-Counting Detectors
can be performed at very low radiation dose levels. This may
open the potential to allow for scan protocols with somewhat Solid-state scintillation detectors used in today’s medical
higher radiation exposure in selected diagnostic situations, CT systems are a very mature technology. The two-step
when, e.g., high spatial resolution or spectral information is detection process, however, based on first converting the
needed to better characterize plaques in the coronary x-rays into visible light and then the light into an electrical
arteries. current (Fig. 67.6a), has certain disadvantages. Solid-state
scintillation detectors do not provide energy-resolved sig-
nals. As a consequence of the detection process, low-energy
New CT Concepts x-ray quanta that carry most of the low-contrast informa-
tion of the object are down-weighted in the signal. This
Phase-Contrast CT results in a nonoptimal CNR of the CT images in particular
in contrast-­enhanced CT scans. Furthermore, the spatial
Medical CT is based on measuring and displaying the local resolution of solid-state scintillation detectors is limited by
x-ray attenuation coefficients in a slice of the patient. Phase-­ their pixel size, which cannot be made much smaller than
contrast CT is an alternative CT technique that does not today.
measure the x-ray absorption but the phase shift of the Another type of detector that is currently being investi-
x-rays by the measurement object. The phase-contrast CT gated directly converts the absorbed x-rays into electrical
images show the local x-ray phase shift coefficients on a signals (Fig. 67.6b). The photon-counting detector is based
gray scale. Phase-contrast imaging has recently gained con- on a semiconductor such as cadmium telluride (CdTe) or
siderable attention in the scientific literature as a new cadmium-zinc-telluride (CZT). A review of photon-counting
method with potential applications in medical imaging, in detectors in medical x-ray imaging is available in [35, 36].
particular after its successful implementation using compact The absorbed x-rays directly induce short current pulses that
low-brilliance x-ray sources that are standard components are individually counted as soon as they exceed a threshold
of medical x-ray systems or CT systems [52]. For typical (Fig.  67.9). The pulse height is proportional to the x-ray
x-ray energies used in CT, the phase shift coefficients of energy  – a photon-counting detector can therefore provide
67  Future Technological Advances in Cardiac CT 881

B1

B2

B1 E1

B2 E2

B3

A B3 E3

Fig. 67.8  Phase-contrast CT scan of a left anterior descending coro- shows a large eccentric plaque with fibrosis, lipid, and calcifications
nary artery (A and B) with corresponding histopathology (E). The ves- causing high-grade stenosis. In the distal LAD, a calcified lesion can be
sel (arrow) is shown with surrounding epicardial fat (arrowhead) and found (B3 and E3). (Modified from Hetterich et  al. [54], with
myocardial muscle (chevron). The left main coronary artery (B1 and permission)
E1) shows mild intimal thickening. The proximal LAD (B2 and E2)

energy-resolved signals. The individual detector pixels are between the pixels are necessary. Therefore, the detector pix-
defined by the high electric field between common cathode els can be made much smaller to improve spatial resolution.
and pixelated anodes (Fig.  67.6b)  – in contrast to conven- The missing down-weighting of low-energy x-ray quanta
tional scintillation detectors, no additional separation layers has the potential to improve the CNR of the images, in par-
882 T. G. Flohr et al.

ticular in CT scans using iodinated contrast agent. In addi- resolved measurements, as illustrated in Figs.  67.9 and
tion, different energy thresholds for energy discrimination 67.10.
may be introduced. The detector can then simultaneously By implementing two energy bins for data readout,
provide CT raw data in different “energy bins” for spectrally photon-­counting detectors can provide dual-energy infor-
mation comparable to today’s dual-energy (DE) CT
approaches. Using more than two energy bins opens the
potential of refined spectral analysis, e.g., K-edge imaging
200
for material separation of three materials, with one of them
having a K-edge in the accessible energy range of about
150 40–90  keV.  In this way, different contrast agents such as
iodine and gadolinium could be separated in the same scan
[56, 57].
100 90keV Another potential application pertinent to cardiac CT is
75keV the detection of gold- and iron-based high-density lipoprotein
nanoparticle contrast agents that are claimed to attach to mac-
50 50keV
rophages accumulating in atherosclerotic plaques in the coro-
25keV nary arteries. Spectral CT could then provide information
0 about the macrophage content of plaques, with a high macro-
phage content associated with an elevated risk of plaque rup-
0 50 100 150 200 250 ture [58]. So far, however, feasibility has only been
demonstrated in small animal models, and the translation to
Fig. 67.9  The fast signal pulses at the anode are counted as soon as
they exceed a threshold. The pulse height is proportional to the x-ray human imaging is not straightforward, e.g., with regard to the
energy. By introducing more than one threshold (here, four energy very high amount of contrast agent that would have to be
thresholds are indicated), the detector provides spectral CT data. In this administered.
example, five x-ray quanta with an energy >25 keV are detected (blue
The spectral separation of a real CdTe- or CZT-based
arrows). Four of them are also detected in the energy band >50  keV
(green arrows), but only three exceed a threshold of 75 keV (red arrows) device is reduced by undesired but unavoidable physical
and only two the upper threshold of 90 keV (black arrows) effects, such as signal splitting at pixel borders (“charge shar-

Energy Bin = 25-140 keV Energy Bin = 75-140 keV

Energy Bin = 50-140 keV Energy Bin = 90-140 keV

Fig. 67.10  Clinical example of a contrast-enhanced CT scan of the sponding images are shown (see also Fig.  67.9). As expected, iodine
kidney in a 71-year-old woman, acquired on a preclinical photon-­ contrast decreases with increasing lower energy of the bin, and image
counting CT scanner at the NIH, Bethesda, USA (see also Ref. [55]). noise increases because less x-ray quanta are used for image
The detector provides four energy bins in a chess-pattern mode ([25– reconstruction
140  keV], [50–140  keV], [75–140  keV], [90–140  keV]); the corre-
67  Future Technological Advances in Cardiac CT 883

ing”) or energy loss of the x-ray quanta due to K-escape, tations of photon-counting CT in clinical practice and to solve
whereby the K-edges of the detector material lead to preferen- the remaining challenges for routine clinical use. A preclini-
tial absorption of some photon energies and the corresponding cal hybrid CT scanner based on a dual-source CT geometry
release of characteristic x-rays at lower energies within the with a conventional energy-integrating scintillation detector
detector itself. These mechanisms lead to a double counting of and a photon-counting CdTe detector was presented in 2012
x-ray quanta at the wrong energies and therefore to a reduction [61]. Meanwhile, the performance of the hybrid CT scanner
of spectral separation. For a realistic detector model, the has been evaluated [62, 63]. In an IRB-­approved patient study
energy discrimination potential is probably equivalent to that with 15 asymptomatic volunteers [55], the non-inferiority of
of a dual-kV scan with optimized pre-filtration [59]. this photon-counting device compared with standard CT has
A main limitation of photon-counting detectors today is been demonstrated for abdominal imaging. Figure  67.11
the finite pulse width of the detected x-ray pulses with a full illustrates the level of spatial resolution that can be achieved
width at half maximum (FWHM) of 10 ns and more. This with the preclinical photon-­counting CT scanner. Its mean-
leads to pulse pileup at high x-ray flux rates: overlapping ingful clinical use, however, depends on the availability of
low-energy pulses may be incorrectly registered as high-­ sufficient radiation dose reserves and the acceptance of
energy hits, and several overlapping pulses may be counted increased radiation dose to the patient.
as one hit only [60]. To reduce pulse pileup, the pixels of the In a nutshell, CT systems with photon-counting detector
detector need to be small: they are divided into sub-pixels of, have the potential to provide spectral information not as an
e.g., 0.25  ×  0.25  mm2 [59]. Another problem is count-rate add-on but as an integral part of each scan. In combination
drift at higher x-ray fluxes caused by nonhomogeneously with increased spatial resolution, new possibilities for the
distributed crystal defects in the sensor material. This may evaluation of coronary plaques or in-stent restenosis may be
lead to severe ring artifacts in the images at higher flux rates. opened. The CNR in contrast-enhanced scans will be
Photon-counting detectors are a very promising new improved, leading to a potential reduction of radiation dose
development in CT, yet the problems of pulse pileup and and/or contrast dose.
count-rate drift will have to be solved before these devices
can be introduced into clinical CT systems. Currently, pre-
clinical prototypes are used to evaluate the potential and limi-

Fig. 67.11  Coronary artery phantom with stent scanned with a conventional high-end CT scanner (left) and with a preclinical photon-counting
CT scanner in a high-resolution scan mode (right). Note the significantly increased spatial resolution
884 T. G. Flohr et al.

 ew Approaches Enhancing the Application

N First-pass enhancement studies provide a “snapshot
Spectrum of Cardiac CT image” of the myocardial blood volume at a single arterial
contrast phase by reusing the coronary CTA images to also
Coronary CTA has meanwhile been established as a nonin- evaluate the myocardium. Hypo-attenuation of the myocar-
vasive alternative to invasive catheter angiography. It has dium is interpreted as perfusion defect. To identify reversible
demonstrated a very high negative predictive value and is ischemia, first-pass enhancement scanning has not only been
primarily used to rule out coronary artery disease. However, performed at rest but also with the application of adenosine
as a consequence of still insufficient spatial resolution, it stress. In the recent CORE320 multicenter trial, CT first-pass
tends to overestimate the degree of coronary stenosis. rest and stress myocardial blood volume imaging showed a
Coronary CTA is a poor predictor of the hemodynamic rele- per-patient sensitivity and specificity for the diagnosis of
vance of stenosis and of myocardial ischemia [64]. CAD of 88% and 55% compared with invasive catheter angi-
During the last several years, efforts have been ongoing to ography as the gold standard. In those patients with a >50%
improve the positive predictive value of cardiac CT for rele- stenosis on coronary CTA, evaluation of CT first-pass
vant coronary stenosis, with the aim of positioning CT as a enhancement resulted in a sensitivity, specificity, positive
singular imaging modality to assess both coronary artery predictive, and negative predictive values of 80%, 74%,
morphology and the status of myocardial perfusion, at best 65%, and 86%, respectively [67, 68].
in a single examination [65]. First-pass enhancement scanning has also been performed
with the use of dual-energy acquisition techniques. Dual-­
energy data have been used to create “iodine maps” of the
Reduction of Ca Blooming myocardium as a surrogate parameter for the myocardial
blood supply. DE iodine maps are potentially more sensitive
Improved spatial resolution is needed for a better estimation for the detection of hypoperfused myocardium compared
of the actual degree of coronary stenosis, and for a reduction with the hypo-attenuation on single-energy CT images [69];
of Ca blooming, see section “Challenges in Cardiac CT”. In they allow for a potentially quantitative evaluation of the
addition, dual-energy-based approaches to remove calcified iodine content, and other sources of hypo-attenuation (e.g.,
plaques from the coronary arteries to better reveal the true fat) can be ruled out. Figure 67.13 shows a clinical example.
residual lumen are currently being investigated. Instead of Using first-generation dual-source CT for first-pass rest and
the common material decomposition into an iodine image stress DE CT scans in a group of 45 patients with known
and a soft tissue image, a modified material decomposition coronary artery disease, the authors found 93.2% sensitivity,
into an iodine image and a calcium image is performed. With 85.5% specificity, 88.3% PPV, and 91.4% NPV for the detec-
the choice of these two base materials, calcifications appear tion of significant coronary artery stenosis compared with
in the calcium image only and are suppressed in the iodine invasive catheter angiography (ICA) as the standard of refer-
image, thereby revealing the actual degree of coronary steno- ence [70].
sis. Because the two base materials iodine and calcium show Because of technical limitations, problematic image
much less difference in their dual-energy behavior than the quality at higher heart rates, and insufficient integration of
typical base materials iodine and water, the resulting mate- DE cardiac CT into routine clinical workflows, this tech-
rial images suffer from significantly increased noise which nique has so far been limited to first feasibility studies and
has to be reduced by refined image filtration. Figure 67.12 has not yet entered clinical routine. Future CT systems
shows an example of the performance achieved with this which provide spectral information on a routine basis as an
technique. Clinical studies are needed to evaluate whether integral part of each standard scan, such as the photon-
dual-energy-based calcium-removal has the potential to counting CT systems discussed in section “CT Systems
reduce the typical overestimation of the degree of coronary with Photon-­Counting Detectors”, combined with excellent
stenosis in coronary CTA. temporal resolution, will enable a more widespread utiliza-
tion of spectral cardiac CT and an intensified analysis of its
benefits and limitations on the way to routine clinical
 irst-Pass Enhancement Scanning
F application.
and Dynamic Perfusion CT The most refined CT technique to acquire information
about the blood supply of the myocardium is dynamic perfu-
Pertinent CT techniques to directly assess and visualize the sion scanning, characterized by time-resolved assessment of
local blood supply of the myocardium are first-pass enhance- the in- and outflow of iodinated contrast agent in the left ven-
ment scanning and dynamic perfusion CT. A review of cur- tricle. The entire myocardium can be examined by either per-
rently available CT techniques to assess myocardial perfusion forming repeated ECG-triggered axial scans using CT
can be found in [66]. systems with area detector technology (that provide 16 cm
67  Future Technological Advances in Cardiac CT 885

a b

Fig. 67.12  Dual-energy coronary CTA in a patient with calcifications reveal the true degree of stenosis from a dual-energy-based material
in the right coronary artery (RCA). Mixed images corresponding to a decomposition using iodine and calcium as the two base materials (b).
standard CT scan (a) and images with suppressed calcifications to (Courtesy of Medical University of South Carolina, USA)

z-axis coverage at isocenter) or by performing ECG-triggered dard. Most important, quantitative cutoff value of the MBF
axial scans at two alternating table positions, with the table index to detect functionally significant coronary lesions can
moving forward and backward between the two positions be defined: 78  mL/100  mL/min [72] or  – quite similar  –
[71]. As a benefit over qualitative first-pass enhancement 75 ml/100 mL/min [73].
techniques, dynamic perfusion CT provides quantitative Because of its quantitative nature, dynamic cardiac perfu-
results for relevant perfusion parameters such as the myocar- sion CT is particularly useful in cases where relative assess-
dial blood flow (MBF). Dynamic perfusion CT has the ment of the functional significance of a lesion is difficult, such
potential to improve the specificity of coronary CTA for the as in patients with multivessel disease (Fig. 67.14). In contrast
detection of functionally significant coronary stenosis. In a to first-pass enhancement approaches, the quantitative MBF
recent study [72], the specificity of visual coronary CTA index does not rely on the assumption that the best-­enhanced
(69%) and quantitative coronary CTA (77%) could be territory is normal and can be used as a reference [72].
improved by the subsequent use of the MBF index under The wide clinical applicability of dynamic cardiac perfu-
stress (89%), compared with invasive FFR as the gold stan- sion CT will secure its future role in the comprehensive
886 T. G. Flohr et al.

a c

b d

Fig. 67.13  ECG-gated dual-energy cardiac CT in a 60-year-old hyper- RCA. The DE energy iodine maps at rest (c) and during stress (d) indi-
tensive man with a history of smoking, using second-generation cate a reversible perfusion defect (arrows). (Courtesy of Vancouver
DSCT.  Both conventional catheter angiography (a) and CTA derived General Hospital, Vancouver, Canada)
from the rest DE scan (b) show chronic total occlusion of the proximal

workup of patients, despite simpler but less versatile processing, and the radiation dose can be further reduced.
approaches such as first-pass enhancement scanning or Ongoing technical progress, enabling the use of 80 kV and
CT-FFR (see next section), provided the complex workflow 70 kV tube voltages for dynamic cardiac perfusion CT, holds
of cardiac perfusion CT can be further simplified and stream- promise to reduce radiation dose to values of 5  mSv and
lined, both with regard to scan data acquisition and results below [74].
67  Future Technological Advances in Cardiac CT 887

a b c
MBF index (ml/100ml/min)
g
200 63

78

76

0
66

RCA LAD LCX

d e f

FFR = 0.71 FFR = 0.70

Fig. 67.14  Coronary CTA images (a–c) and conventional catheter angiography (d–f) in a patient with three-vessel disease. MBF derived from a
dynamic CT perfusion scan (g–i) is reduced in the whole myocardium (<78 mL/100 mL/min). (From Rossi et al. [72], with permission)

Meanwhile, first attempts have been made to also derive a fractional flow reserve (FFR), a measure of lesion-specific
prognostic value from quantitative perfusion parameters. ischemia [77].
Using data of 144 patients from a multicenter trial with fol- Standard coronary CTA images are used to create a patient-
low-­up for 6, 12, and 18 months, Meinel et al. [75] found that specific three-dimensional model of the coronary arteries.
global quantification of the left ventricular MBF at stress may Then, by applying the principles of CFD to this model, hyper-
have incremental predictive value for future major adverse emic coronary flow and pressure can be simulated from the
cardiac events (MACE) over clinical risk factors and assess- coronary CTA images at rest. These pressure and flow values
ment of stenosis at coronary CTA. Patients with global MBF are used to noninvasively compute the local fractional flow
of <121 mL/100 mL/min were at increased risk for MACE. The reserve (FFR), which is the ratio of maximal coronary blood
number of territories with perfusion defects was strongly pre- flow through a stenotic artery to the blood flow in the hypo-
dictive of MACE with adjusted hazard ratios of 1.41, 3.44, and thetical case that the artery was normal [78]. The end result is
4.76 for one, two, and three affected territories [76]. a color-coded map of computed local FFR values as an over-
Combining several parameters derived from cardiac CT lay to the three-dimensional model of the coronary arteries,
scans to improve and refine the prediction of MACE, poten- indicating regions of low FFR <0.8 corresponding to poten-
tially in combination with traditional risk factors and labora- tially hemodynamically relevant lesions in red (Fig. 67.15).
tory data of the patient, is a promising future research area A meta-analysis of 6 original studies comparing the per-
and may significantly influence the clinical value of cardiac formance of CT-FFR with the gold-standard invasive FFR
CT.  For the analysis of larger databases, machine learning (1354 vessels; 812 patients) in diagnosing hemodynamically
approaches will be needed. relevant coronary stenosis showed a pooled per-vessel sensi-
tivity and specificity of 0.84 [95% confidence interval (CI):
0.80–0.87] and 0.76 [95% CI: 0.73–0.79], respectively [79].
CT Fractional Flow Reserve The original studies were performed with two different
methods, an FDA-approved off-site approach requiring the
Another technique to assess the hemodynamic relevance of coronary CTA images to be sent to a centralized core lab to
stenosis is CT-FFR. It applies computational fluid dynamics perform coronary artery segmentation and CT-FFR compu-
(CFD) techniques to coronary CTA images to compute the tation (HeartFlow, Redwood City, California, USA) and a
888 T. G. Flohr et al.

Fig. 67.15  User-interface of a prototype preclinical on-site software to compute CT-FFR

preclinical on-site approach using software installed on a (Fig.  67.16). Correlation between machine learning and
local computer in the hospital (Siemens Healthcare GmbH, physics-based predictions is excellent [82], with a significant
Forchheim, Germany). A subgroup analysis of the studies reduction of the average execution time by machine learning,
performed with the preclinical on-site approach revealed a leading to near real-time assessment of FFR.
per-vessel sensitivity and specificity of 0.87 (95% CI: 079– In a study [83], on-site CT-FFR showed a comparable
0.93) and 0.71 (95% CI: 0.63–0.78), respectively. diagnostic accuracy as dynamic CT perfusion for identifying
According to Panchal et al. [79], the addition of CT-FFR to hemodynamically significant stenosis defined by invasive
coronary CTA improves the diagnostic ability of noninvasive FFR.  According to the authors, CT perfusion was more
CT to evaluate hemodynamically significant coronary lesions robust and could be successfully analyzed in all patients who
and might improve noninvasive clinical decision-­making for had to be excluded from CT-FFR evaluation due to technical
patients with CAD. In the PLATFORM study [80], equivalent problems. CT perfusion is a complementary modality in
clinical outcomes at lower cost were demonstrated for care patients with severe coronary calcifications and poor coro-
guided by coronary CTA and selective CT-FFR in patients nary CTA image quality. According to Coenen [84] diagnos-
with stable chest pain, compared with usual care over 1-year tic performance can be improved by combining CT-FFR and
follow-up. As a further step, the ADVANCE Registry is CT perfusion. A stepwise approach to improve diagnostic
planned as a multicenter, prospective registry enrolling performance is recommended, reserving CT perfusion for
approximately 5000 patients at 50 sites to evaluate utility, patients with intermediate CT-FFR results.
clinical outcomes, and resource utilization following CT-FFR- According to [85], the workflow of how CT-FFR will be
guided treatment in clinically stable, symptomatic patients used in clinical practice – whether on- or off-site – remains
diagnosed with CAD by coronary CTA [81]. to be determined.
Recently, machine learning has been used to compute In the future, machine learning is expected to go beyond
CT-FFR instead of computationally expensive CFD tech- providing CT-FFR numbers to predict the hemodynamic rel-
niques [82]. In the machine learning approach, the model is evance of stenosis. By combining CT-FFR with other param-
trained on a large database of synthetically generated coro- eters, such as plaque location and composition, machine
nary anatomies, where the target values are computed using learning-based approaches may have the potential for a more
the physics-based CFD model. The trained model predicts exact classification of patients and prediction of future
FFR at each point along the centerline of the coronary tree events.
67  Future Technological Advances in Cardiac CT 889

Training database with

synthetic geometries

Machine-learning
based FFR
algorithm

New dataset

Fig. 67.16  Principle of a machine learning approach using a trained model to compute CT-FFR

Prognostic Risk Assessment Multicenter Registry. Machine learning evaluating 25 clini-

cal and 44 coronary CTA parameters was found to predict
Traditional prognostic risk assessment to determine an indi- the 5-year all-cause mortality significantly better than exist-
vidual’s chances of developing cardiovascular disease is ing clinical or coronary CTA metrics alone.
based on evaluating few traditional risk factors, such as age, Future clinical studies will be needed to further evaluate
family history of CAD, hypertension, increased cholesterol the potential of machine learning for prognostic risk assess-
levels, or diabetes. Additional CT imaging findings, such as ment and to refine the prognosis – it is very likely, however,
coronary calcifications, coronary artery stenosis, and distri- that machine learning-based approaches will play a rapidly
bution and composition of coronary plaques, add incremen- increasing role in cardiac CT on its way to a comprehensive
tal improvement to the prediction of MACE (e.g., [86]). clinical imaging modality.
Machine learning is a rapidly evolving field that gives
computers the ability to make data-driven predictions or
decisions, through building a model from test inputs. References
Machine learning can identify patterns in large datasets with
a multitude of variables – hence it is consequent to apply it to 1. Flohr TG, De Cecco CN, Schmidt B, Wang R, Schoepf UJ, Meinel
FG.  Computed tomographic assessment of coronary artery dis-
large databases collecting CT imaging data, traditional risk ease: state-of-the-art imaging techniques. Radiol Clin N Am.
factors, clinical parameters, and follow-up data of patients. 2015;53(2):271–85.
Instead of evaluating just few selected clinical or imaging 2. Flohr T, Ohnesorge B. Heart rate adaptive optimization of spatial
parameters as in traditional approaches, machine learning and temporal resolution for ECG-gated multi-slice spiral CT of the
heart. JCAT. 2001;25(6):907–23.
can combine and rank a large number of features to poten- 3. Leschka S, Wildermuth S, Boehm T, Desbiolles L, Husmann L,
tially improve and refine the prediction of MACE. Plass A, Koepfli P, Schepis T, Marincek B, Kaufmann PA, Alkadhi
In a first feasibility study [87], the authors included H.  Noninvasive coronary angiography with 64-section CT: effect
10,030 patients with suspected coronary artery disease and of average heart rate and heart rate variability on image quality.
Radiology. 2006;241(2):378–85.
5-year follow-up from the CONFIRM International
890 T. G. Flohr et al.

4. Flohr TG, McCollough CH, Bruder H, Petersilka M, Gruber K, Süß 20. Rohkohl C, Bruder H, Stierstorfer K, Flohr T. Improving best-phase
C, Grasruck M, Stierstorfer K, Krauss B, Raupach R, Primak AN, image quality in cardiac CT by motion correction with MAM opti-
Küttner A, Achenbach S, Becker C, Kopp A, Ohnesorge BM. First mization. Med Phys. 2013;40(3):031901.
performance evaluation of a dual-source CT (DSCT) system. Eur 21. Fuchs TA, Stehli J, Dougoud S, Fiechter M, Sah BR, Buechel RR,
Radiol. 2006;16(2):256–68. Bull S, Gaemperli O, Kaufmann PA.  Impact of a new motion-­
5. McCollough CH, Schmidt B, Yu L, Primak A, Ulzheimer S, Bruder correction algorithm on image quality of low-dose coronary CT
H, Flohr TG. Measurement of temporal resolution in dual source angiography in patients with insufficient heart rate control. Acad
CT. Med Phys. 2008;35(2):764–8. Radiol. 2014;21(3):312–7.
6. Matt D, Scheffel H, Leschka S, Flohr TG, Marincek B, Kaufmann 22. Maass C, Hofmann C, Kachelriess M. TRI-PICCS in single source
PA, Alkadhi H. Dual-source CT coronary angiography: image qual- and dual source CT, in IEEE NSS-MIC (nuclear science sympo-
ity, mean heart rate, and heart rate variability. AJR Am J Roentgenol. sium- medical imaging conference). Knoxville: NSS/MIC; 2010.
2007;189(3):567–73. p. 2805–10.
7. Ropers U, Ropers D, Pflederer T, Anders K, Kuettner A, Stilianakis 23. Apfaltrer P, Schoendube H, Schoepf UJ, Allmendinger T,

NI, Komatsu S, Kalender W, Bautz W, Daniel WG, Achenbach Tricarico F, Schindler A, Vogt S, Sunnegårdh J, Stierstorfer K,
S. Influence of heart rate on the diagnostic accuracy of dual-source Henzler T, Fink C, Bruder H, Flohr TG, Ebersberger U. Enhanced
computed tomography coronary angiography. J Am Coll Cardiol. temporal resolution at cardiac CT with a novel CT image recon-
2007;50(25):2393–8. struction algorithm: initial patient experience. Eur J Radiol.
8. Scheffel H, Alkadhi H, Plass A, Vachenauer R, Desbiolles L, 2013;82(2):270–4.
Gaemperli O, Schepis T, Frauenfelder T, Schertler T, Husmann L, 24. Isola AA, Ziegler A, Koehler T, Niessen WJ, Grass M.  Motion-­
Grunenfelder J, Genoni M, Kaufmann PA, Marincek B, Leschka compensated iterative cone-beam CT image reconstruc-
S. Accuracy of dual-source CT coronary angiography: first expe- tion with adapted blobs as basis functions. Phys Med Biol.
rience in a high pre-test probability population without heart rate 2008;53(23):6777–97.
control. Eur Radiol. 2006;16(12):2739–47. 25. Hahn J, Bruder H, Allmendinger T, Stierstorfer K, Flohr T, et al.
9. Weustink AC, Neefjes LA, Kyrzopoulos S, van Straten M, Neoh Reduction of motion artifacts in cardiac CT based on partial angle
Eu R, Meijboom WB, van Mieghem CA, Capuano E, Dijkshoorn reconstructions from short scan data. Proceedings SPIE 9783, med-
ML, Cademartiri F, Boersma E, de Feyter PJ, Krestin GP, ical imaging 2016: physics of medical imaging, 97831A (March 25,
Mollet NR.  Impact of heart rate frequency and variability on 2016).
radiation exposure, image quality, and diagnostic performance 26. Andreini D, Pontone G, Mushtaq S, Bertella E, Conte E, Segurini
in dual-source spiral CT coronary angiography. Radiology. C, Baggiano A, Bartorelli AL, Annoni A, Formenti A, Petullà M,
2009;253(3):672–80. Beltrama V, Fiorentini C, Pepi M. Low-dose CT coronary angiogra-
10. Li M, Zhang GM, Zhao JS, Jiang ZW, Peng ZH, Jin ZT, Sun phy with a novel IntraCycle motion-correction algorithm in patients
G.  Diagnostic performance of dual-source CT coronary angiog- with high heart rate or heart rate variability. Eur Heart J Cardiovasc
raphy with and without heart rate control: systematic review and Imaging. 2015;16(10):1093–100.
meta-analysis. Clin Radiol. 2014;69(2):163–71. 27. Sheta HM, Egstrup K, Husic M, Heinsen LJ, Lambrechtsen

11. Besson GM. New CT system architectures for high temporal reso- J. Impact of a motion correction algorithm on quality and diagnos-
lution with applications to improved geometric dose efficiency and tic utility in unselected patients undergoing coronary CT angiogra-
cardiac imaging. Med Phys. 2015;42(5):2668–78. phy. Clin Imaging. 2016;40(2):217–21.
12. Wang G, Yu H, Ye Y. A scheme for multisource interior tomogra- 28. Sheta HM, Egstrup K, Husic M, Heinsen LJ, Nieman K,

phy. Med Phys. 2009;36(8):3575–81. Lambrechtsen J. Impact of a motion correction algorithm on image
13. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M, quality in patients undergoing CT angiography: a randomized con-
Detrano R. Quantification of coronary artery calcium using ultra- trolled trial. Clin Imaging. 2017;42:1–6.
fast computed tomography. JACC. 1990;15:827–32. 29. Vanhecke TE, Madder RD, Weber JE, Bielak LF, Peyser PA,

14. Budoff M, Georgiou D, Brody A, et  al. Ultrafast computed Chinnaiyan KM.  Development and validation of a predictive
tomography as a diagnostic modality in the detection of screening tool for uninterpretable coronary CT angiography results.
coronary artery disease: a multicenter study. Circulation. Circ Cardiovasc Imaging. 2011;4(5):490–7.
1996;93:898–904. 30. Maintz D, Seifarth H, Raupach R, Flohr T, Rink M, Sommer T,
15. Yoneyama K, Vavere AL, Cerci R, Ahmed R, Arai AE, Niinuma H, Ozgün M, Heindel W, Fischbach R. 64-slice multidetector coro-
Rybicki FJ, Rochitte CE, Clouse ME, George RT, Lima JA, Arbab-­ nary CT angiography: in vitro evaluation of 68 different stents. Eur
Zadeh A. Influence of image acquisition settings on radiation dose Radiol. 2006;16(4):818–26.
and image quality in coronary angiography by 320-detector volume 31. Maintz D, Burg MC, Seifarth H, Bunck AC, Ozgün M,

computed tomography: the CORE320 pilot experience. Heart Int. Fischbach R, Jürgens KU, Heindel W. Update on multidetector
2012;7(2):e11. coronary CT angiography of coronary stents: in  vitro evalua-
16. Tang J, Hsieh J, Chen GH.  Temporal resolution improvement in tion of 29 different stent types with dual-source CT. Eur Radiol.
cardiac CT using PICCS (TRI-PICCS): performance studies. Med 2009;19(1):42–9.
Phys. 2010;37(8):4377–88. 32. Carrabba N, Schuijf JD, de Graaf FR, et  al. Diagnostic accuracy
17. Schöndube H, Allmendinger T, Stierstorfer K, Bruder H, Flohr of 64-slice computed tomography coronary angiography for the
T. Evaluation of a novel CT image reconstruction algorithm with detection of in-stent restenosis: a meta-analysis. J Nucl Cardiol.
enhanced temporal resolution, in medical imaging 2011: physics of 2010;17(3):470–8.
medical imaging, vol. 7961. Lake Buena Vista/Florida: SPIE; 2011. 33. Fuchs TA, Stehli J, Fiechter M, Dougoud S, Sah BR, Gebhard C,
p. 79611N. Bull S, Gaemperli O, Kaufmann PA.  First in  vivo head-to-head
18. Bhagalia R, Pack JD, Miller JV, Iatrou M.  Nonrigid registration-­ comparison of high-definition versus standard-definition stent
based coronary artery motion correction for cardiac computed imaging with 64-slice computed tomography. Int J Cardiovasc
tomography. Med Phys. 2012;39(7):4245–54. Imaging. 2013;29(6):1409–16.
19. Tang Q, Cammin J, Srivastava S, Taguchi K.  A fully four-­
34. Gassenmaier T, Petri N, Allmendinger T, et  al. Next generation
dimensional, iterative motion estimation and compensation method coronary CT angiography: in vitro evaluation of 27 coronary stents.
for cardiac CT. Med Phys. 2012;39(7):4291–305. Eur Radiol. 2014;24(11):2953–61.
67  Future Technological Advances in Cardiac CT 891

35. Taguchi K, Iwanczyk JS. Vision 20/20: single photon counting x-ray 49. Schuhbaeck A, Achenbach S, Layritz C, Eisentopf J, Hecker F,
detectors in medical imaging. Med Phys. 2013;40(10):100901. Pflederer T, Gauss S, Rixe J, Kalender W, Daniel WG, Lell M,
https://doi.org/10.1118/1.4820371. Review. Ropers D. Image quality of ultra-low radiation exposure coronary
36. Taguchi K. Energy-sensitive photon counting detector-based X-ray CT angiography with an effective dose <0.1 mSv using high-pitch
computed tomography. Radiol Phys Technol. 2017;10(1):8–22. spiral acquisition and raw data-based iterative reconstruction. Eur
37. Menke J, Unterberg-Buchwald C, Staab W, Sohns JM, Seif Amir Radiol. 2013;23(3):597–606.
Hosseini A, Schwarz A. Head-to-head comparison of prospectively 50. Elad M, Aharon M. Image denoising via sparse and redundant rep-
triggered vs retrospectively gated coronary computed tomography resentations over learned dictionaries. IEEE Trans Image Process.
angiography: meta-analysis of diagnostic accuracy, image quality, 2006;15(12):3736–45.
and radiation dose. Am Heart J. 2013;165(2):154–63. 51. Ghadrdan S, Alirezaie J, Dillenseger JL, Babyn P.  Low-dose

38. Neefjes LA, Dharampal AS, Rossi A, Nieman K, Weustink
computed tomography image denoising based on joint wavelet
AC, Dijkshoorn ML, Ten Kate GJ, Dedic A, Papadopoulou SL, and sparse representation. Conf Proc IEEE Eng Med Biol Soc.
van Straten M, Cademartiri F, Krestin GP, de Feyter PJ, Mollet 2014;2014:3325–8.
NR. Image quality and radiation exposure using different low-dose 52. Pfeiffer F, Weitkamp T, Bunk O, David C. Phase retrieval and dif-
scan protocols in dual-source CT coronary angiography: random- ferential phase-contrast imaging with low-brilliance X-ray sources.
ized study. Radiology. 2011;261(3):779–86. Nat Phys. 2006;2:258–61.
39. Yin WH, Lu B, Hou ZH, Li N, Han L, Wu YJ, Niu HX, Silverman 53. Raupach R, Flohr T.  Performance evaluation of x-ray differential
JR, Nicola De Cecco C, Schoepf UJ. Detection of coronary artery phase contrast computed tomography (PCT) with respect to medi-
stenosis with sub-milliSievert radiation dose by prospectively cal imaging. Med Phys. 2012;39(8):4761–74.
ECG-triggered high-pitch spiral CT angiography and iterative 54. Hetterich H, Willner M, Habbel C, Herzen J, Hoffmann VS, Fill
reconstruction. Eur Radiol. 2013;23(11):2927–33. S, Hipp A, Marschner M, Schüller U, Auweter S, Massberg S,
40. Morsbach F, Gordic S, Desbiolles L, Husarik D, Frauenfelder T, Reiser MF, Pfeiffer F, Saam T, Bamberg F.  X-ray phase-contrast
Schmidt B, Allmendinger T, Wildermuth S, Alkadhi H, Leschka computed tomography of human coronary arteries. Investig Radiol.
S.  Performance of turbo high-pitch dual-source CT for coronary 2015;50(10):686–94.
CT angiography: first ex vivo and patient experience. Eur Radiol. 55. Pourmorteza A, Symons R, Sandfort V, Mallek M, Fuld

2014;24(8):1889–95. MK, Henderson G, Jones EC, Malayeri AA, Folio LR,
41. Chinnaiyan KM, Bilolikar AN, Walsh E, Wood D, DePetris A, Bluemke DA.  Abdominal imaging with contrast-enhanced
Gentry R, Boura J, Abbara S, Al-Mallah M, Bis K, Boswell G, photon-counting CT: first human experience. Radiology.
Gallagher M, Arunakul IO, Halliburton S, Jacobs J, Lesser J, 2016;279(1):239–45.
Schoepf UJ, Valeti US, Raff GL.  CT dose reduction using pro- 56. Roessl E, et  al. K-edge imaging in x-ray computed tomogra-

spectively triggered or fast-pitch spiral technique employed in phy using multi-bin photon counting detectors. Phys Med Biol.
cardiothoracic imaging (the CT dose study). J Cardiovasc Comput 2007;52:4679–96.
Tomogr. 2014;8(3):205–14. 57. Schlomka JP, Roessl E, Dorscheid R, Dill S, Martens G, Istel T,
42. McCollough CH, Primak AN, Braun N, Kofler J, Yu L, Christner Bäumer C, Herrmann C, Steadman R, Zeitler G, Livne A, Proksa
J. Strategies for reducing radiation dose in CT. Radiol Clin N Am. R.  Multienergy photon-counting K-edge imaging: potential for
2009;47(1):27–40. improved luminal depiction in vascular imaging. Radiology.
43. Meinel FG, Canstein C, Schoepf UJ, Sedlmaier M, Schmidt B, 2008;249(3):1010–6.
Harris BS, Flohr TG, De Cecco CN. Image quality and radiation 58. Cormode DP, Roessl E, Thran A, Skajaa T, Gordon RE,

dose of low tube voltage 3rd generation dual-source coronary Schlomka JP, Fuster V, Fisher EA, Mulder WJ, Proksa R,
CT angiography in obese patients: a phantom study. Eur Radiol. Fayad ZA.  Atherosclerotic plaque composition: analysis with
2014;24(7):1643–50. multicolor CT and targeted gold nanoparticles. Radiology.
44. Vonder M, Pelgrim GJ, Huijsse SE, Meyer M, Greuter MJ, Henzler 2010;256(3):774–82.
T, Flohr TG, Oudkerk M, Vliegenthart R.  Feasibility of spectral 59. Kappler S, Niederlöhner D, Stierstorfer K, Flohr T.  Contrast-­

shaping for detection and quantification of coronary calcifications enhancement, image noise and dual-energy simulations for
in ultra-low dose CT. Eur Radiol. 2017;27(5):2047–54. quantum-counting clinical CT.  Proc SPIE Med Imaging Conf.
45. Tesche C, De Cecco CN, Vliegenthart R, Albrecht MH, Varga-­ 2010;7622:76223H.
Szemes A, Duguay TM, Ebersberger U, Bayer RR 2nd, Canstein 60. Barber WC, et  al. Characterization of a novel photon counting
C, Schmidt B, Allmendinger T, Litwin SE, Morris PB, Flohr TG, detector for clinical CT: count rate, energy resolution, and noise
Hoffmann E, Schoepf UJ. Accuracy and radiation dose reduction performance. Proc SPIE Med Imaging Conf. 2009;7258:725824.
using low-voltage computed tomography coronary artery calcium 61. Kappler S, et  al. First results from a hybrid prototype CT scan-
scoring with tin filtration. Am J Cardiol. 2017;119(4):675–80. ner for exploring benefits of quantum-counting in clinical CT. Proc
46. Thibault JB, Sauer KD, Bouman CA, Hsieh J. A three-dimensional SPIE Med Imaging Conf. 2012;8313:83130X.
statistical approach to improved image quality for multislice helical 62. Yu Z, Leng S, Jorgensen SM, Li Z, Gutjahr R, Chen B, Halaweish
CT. Med Phys. 2007;34(11):4526–44. AF, Kappler S, Yu L, Ritman EL, McCollough CH. Evaluation of
47. Stehli J, Fuchs TA, Bull S, Clerc OF, Possner M, Buechel RR, conventional imaging performance in a research whole-body CT
Gaemperli O, Kaufmann PA.  Accuracy of coronary CT angiog- system with a photon-counting detector array. Phys Med Biol.
raphy using a submillisievert fraction of radiation exposure: com- 2016;61(4):1572–95.
parison with invasive coronary angiography. J Am Coll Cardiol. 63. Gutjahr R, Halaweish AF, Yu Z, Leng S, Yu L, Li Z, Jorgensen
2014;64(8):772–80. SM, Ritman EL, Kappler S, McCollough CH.  Human imaging
48. Layritz C, Schmid J, Achenbach S, Ulzheimer S, Wuest W, May with photon counting-based computed tomography at clinical dose
M, Ropers D, Klinghammer L, Daniel WG, Pflederer T, Lell levels: contrast-to-noise ratio and cadaver studies. Investig Radiol.
M.  Accuracy of prospectively ECG-triggered very low-dose 2016;51(7):421–9.
coronary dual-source CT angiography using iterative reconstruc- 64. Stolzmann P, Donati OF, Scheffel H, Azemaj N, Baumueller S, Plass
tion for the detection of coronary artery stenosis: comparison A, Kozerke S, Leschka S, Grünenfelder J, Boesiger P, Marincek B,
with invasive catheterization. Eur Heart J Cardiovasc Imaging. Alkadhi H. Low-dose CT coronary angiography for the prediction
2014;15(11):1238–45. of myocardial ischaemia. Eur Radiol. 2010;20(1):56–64.
892 T. G. Flohr et al.

65. Bastarrika G, Lee YS, Huda W, Ruzsics B, Costello P, Schoepf 77. Min JK, Berman DS, Budoff MJ, Jaffer FA, Leipsic J, Leon MB,
UJ. CT of coronary artery disease. Radiology. 2009;253(2):317–38. Mancini GB, Mauri L, Schwartz RS, Shaw LJ. Rationale and design
66. Bucher AM, De Cecco CN, Schoepf UJ, Wang R, Meinel FG, of the DeFACTO (determination of fractional flow reserve by ana-
Binukrishnan SR, Spearman JV, Vogl TJ, Ruzsics B. Cardiac CT for tomic computed tomographic AngiOgraphy) study. J Cardiovasc
myocardial ischaemia detection and characterization-comparative Comput Tomogr. 2011;5(5):301–9.
analysis. Br J Radiol. 2014;24:20140159. 78. Taylor CA, Fonte TA, Min JK.  Computational fluid dynamics

67. George RT, Arbab-Zadeh A, Miller JM, Vavere AL, Bengel FM, applied to cardiac computed tomography for noninvasive quantifi-
Lardo AC, Lima JA. Computed tomography myocardial perfusion cation of fractional flow reserve: scientific basis. J Am Coll Cardiol.
imaging with 320-row detector computed tomography accurately 2013;61(22):2233–41.
detects myocardial ischemia in patients with obstructive coronary 79. Panchal HB, Veeranki SP, Bhatheja S, Barry N, Mahmud E, Budoff
artery disease. Circ Cardiovasc Imaging. 2012;5(3):333–40. M, Lavine SJ, Mamudu HM, Paul TK. Fractional flow reserve using
68. George RT, Mehra VC, Chen MY, Kitagawa K, Arbab-Zadeh A, computed tomography for assessing coronary artery disease: a meta-­
Miller JM, Matheson MB, Vavere AL, Kofoed KF, Rochitte CE, analysis. J Cardiovasc Med (Hagerstown). 2016;17(9):694–700.
Dewey M, Yaw TS, Niinuma H, Brenner W, Cox C, Clouse ME, 80. Douglas PS, De Bruyne B, Pontone G, Patel MR, Norgaard BL,
Lima JA, Di Carli M.  Myocardial CT perfusion imaging and Byrne RA, Curzen N, Purcell I, Gutberlet M, Rioufol G, Hink
SPECT for the diagnosis of coronary artery disease: a head-to-head U, Schuchlenz HW, Feuchtner G, Gilard M, Andreini D, Jensen
comparison from the CORE320 multicenter diagnostic perfor- JM, Hadamitzky M, Chiswell K, Cyr D, Wilk A, Wang F, Rogers
mance study. Radiology. 2014;272(2):407–16. C, Hlatky MA, PLATFORM Investigators. 1-year outcomes of
69. Ruzsics B, Schwarz F, Schoepf UJ, et  al. Comparison of dual-­ FFRCT-guided care in patients with suspected coronary disease:
energy computed tomography of the heart with single photon The PLATFORM Study. J Am Coll Cardiol. 2016;68(5):435–45.
emission computed tomography for assessment of coronary 81. Chinnaiyan KM, Akasaka T, Amano T, Bax JJ, Blanke P, De

artery stenosis and of the myocardial blood supply. Am J Cardiol. Bruyne B, Kawasaki T, Leipsic J, Matsuo H, Morino Y, Nieman K,
2009;104(3):318–26. Norgaard BL, Patel MR, Pontone G, Rabbat M, Rogers C, Sand NP,
70. Ko SM, Choi JW, Song MG, et  al. Myocardial perfusion imag- Raff G. Rationale, design and goals of the heart flow assessing diag-
ing using adenosine-induced stress dual-energy computed tomog- nostic value of non-invasive FFRCT in coronary care (ADVANCE)
raphy of the heart: comparison with cardiac magnetic resonance registry. J Cardiovasc Comput Tomogr. 2017;11(1):62–7.
imaging and conventional coronary angiography. Eur Radiol. 82. Itu L, Rapaka S, Passerini T, Georgescu B, Schwemmer C,

2011;21:26–35. Schoebinger M, Flohr T, Sharma P, Comaniciu D.  A machine-­
71. Mahnken AH, Klotz E, Pietsch H, Schmidt B, Allmendinger T, learning approach for computation of fractional flow reserve
Haberland U, Kalender WA, Flohr T.  Quantitative whole heart from coronary computed tomography. J Appl Physiol (1985).
stress perfusion CT imaging as noninvasive assessment of hemody- 2016;121(1):42–52.
namics in coronary artery stenosis: preliminary animal experience. 83. Yang DH, Kim YH, Roh JH, Kang JW, Ahn JM, Kweon J, Lee JB,
Investig Radiol. 2010;45(6):298–305. Choi SH, Shin ES, Park DW, Kang SJ, Lee SW, Lee CW, Park SW,
72. Rossi A, Dharampal A, Wragg A, Davies LC, van Geuns RJ,
Park SJ, Lim TH.  Diagnostic performance of on-site CT-derived
Anagnostopoulos C, Klotz E, Kitslaar P, Broersen A, Mathur fractional flow reserve versus CT perfusion. Eur Heart J Cardiovasc
A, Nieman K, Hunink MG, de Feyter PJ, Petersen SE, Pugliese Imaging. 2017;18:432–40.
F.  Diagnostic performance of hyperaemic myocardial blood flow 84. Coenen A, Rossi A, Lubbers MM, Kurata A, Kono AK, Chelu RG,
index obtained by dynamic computed tomography: does it predict Segreto S, Dijkshoorn ML, Wragg A, van Geuns RM, Pugliese F,
functionally significant coronary lesions? Eur Heart J Cardiovasc Nieman K.  Integrating CT myocardial perfusion and CT-FFR in
Imaging. 2014;15(1):85–94. the work-up of coronary artery disease. JACC Cardiovasc Imaging.
73. Bamberg F, Becker A, Schwarz F, Marcus RP, Greif M, von
2017. pii: S1936-878X(16)30975–5. https://doi.org/10.1016/j.
Ziegler F, Blankstein R, Hoffmann U, Sommer WH, Hoffmann VS, jcmg.2016.09.028.
Johnson TR, Becker HC, Wintersperger BJ, Reiser MF, Nikolaou 85. Nakazato R, Park HB, Berman DS, Gransar H, Koo BK, Erglis
K. Detection of hemodynamically significant coronary artery ste- A, Lin FY, Dunning AM, Budoff MJ, Malpeso J, Leipsic J, Min
nosis: incremental diagnostic value of dynamic CT-based myocar- JK.  Noninvasive fractional flow reserve derived from computed
dial perfusion imaging. Radiology. 2011;260(3):689–98. tomography angiography for coronary lesions of intermediate ste-
74. Kim SM, Cho YK, Choe YH. Adenosine-stress dynamic myocar- nosis severity: results from the DeFACTO study. Circ Cardiovasc
dial perfusion imaging using 128-slice dual-source CT in patients Imaging. 2013;6(6):881–9.
with normal body mass indices: effect of tube voltage, tube current, 86. Chow BJ, Small G, Yam Y, Chen L, Achenbach S, Al-Mallah
and iodine concentration on image quality and radiation dose. Int J M, Berman DS, Budoff MJ, Cademartiri F, Callister TQ,
Cardiovasc Imaging. 2014;30(Suppl 2):95–103. Chang H-J, Cheng V, Chinnaiyan K, Delago A, Dunning A,
75. Meinel FG, Wichmann JL, Schoepf UJ, Pugliese F, Ebersberger Hadamitzky M, Hausleiter J, Kaufmann P, Lin F, Maffei E,
U, Lo GG, Choe YH, Wang Y, Tesche C, Segreto S, Kunz WG, Raff GL, Shaw LJ, Villines TC, Min JK. Incremental prognos-
Thierfelder KM, Bamberg F, De Cecco CN. Global quantification tic value of cardiac CT in CAD using CONFIRM (COroNary
of left ventricular myocardial perfusion at dynamic CT imaging: computed tomography a­ ngiography evaluation for clinical out-
prognostic value. J Cardiovasc Comput Tomogr. 2017;11(1):16–24. comes: an international multicenter registry). Circ Cardiovasc
76. Meinel FG, Pugliese F, Schoepf UJ, Ebersberger U, Wichmann Imaging. 2011;4:463–72.
JL, Lo GG, Choe YH, Wang Y, Segreto S, Bamberg F, De Cecco 87. Motwani M, Dey D, Berman DS, et al. Machine learning for pre-
CN. Prognostic value of stress dynamic myocardial perfusion CT in diction of all-cause mortality in patients with suspected coronary
a multicenter population with known or suspected coronary artery artery disease: a 5-year multicentre prospective registry analysis.
disease. AJR Am J Roentgenol. 2017;8:W1–9. Eur Heart J. 2017;38(7):500–7.
 Machine Learning and Artificial
Intelligence in Cardiovascular Imaging 68
Marwen Eid, James V. Spearman, Marly van Assen,
Domenico De Santis, Pooyan Sahbaee, Scott P. Landreth,
Brian Jacobs, and Carlo N. De Cecco

Introduction a substantial amount of patient information being recorded. This

available information will only increase in the future through
Artificial intelligence (AI) has captured the minds of science the use of bidirectional patient portals. Moreover, in the era of
fiction writers and the general public for quite some time. As evidence-based medicine, thousands of new evidence and data
advancements have been made in computer science and are being published daily. Going through such large volumes of
engineering research, much improved computational power data to determine what is clinically relevant and actionable can
and the creation of newer, more efficient algorithms such as be overwhelming, resulting in important information being
machine learning (ML) and deep learning (DL) have enabled missed by physicians. However, AI machines can now consis-
the feasibility of big data analysis. AI has moved from the tently perform repetitive tasks at maximum capacity, sometimes
realm of science fiction to applications used in everyday life, producing results faster and more efficiently than humans.
such as Tesla’s self-driving cars, Facebook’s facial recogni- Medicine is thereby a perfect testing ground for the application
tion, Amazon’s product recommendations, mobile check of ML, as these systems can augment the ability of physicians to
deposits, language translation software, and more. identify key information required for patient management while
As AI continues to improve, ML algorithms can now master presenting it in an understandable manner. In particular, because
tasks that were previously thought to be too complex for radiology directly involves extracting data consisting of specific
machines and are now even capable of detecting patterns that features seen on images and interpreting them through the
are beyond human perception [1]. This has led to a renewed and knowledge base acquired by the radiologist, the medical imag-
increased interest in ML as a useful tool in medical practice, ing field serves as an attractive arena for the incorporation of
particularly in the field of medical imaging [1]. Indeed, now ML systems.
more than ever, medicine has become a big data science, with As advanced AI and ML systems transition from fiction to
the introduction of electronic medical records (EMR) leading to reality and steadily approach their implementation into med-
ical and radiology practices, understanding the general meth-
ods, capabilities, and limitations of machine learning is of
M. Eid (*) · J. V. Spearman · M. van Assen · S. P. Landreth
B. Jacobs fundamental importance to physicians and radiologists for
Department of Radiology and Radiological Science, Medical the effective use of these systems. This chapter will intro-
University of South Carolina, Charleston, SC, USA duce some of the basic concepts of machine learning tech-
D. De Santis niques, provide a basic framework for their use, and highlight
Department of Radiological Sciences, Oncological and current and future applications in medicine and radiology
Pathological Sciences, University of Rome “Sapienza”,
with a special focus on cardiovascular imaging.
Latina, Italy
Division of Cardiovascular Imaging, Department of Radiology and
Radiological Science, Medical University of South Carolina,
Charleston, SC, USA Machine Learning and Artificial Intelligence
P. Sahbaee
Siemens Medical Solutions USA, Inc., Malvern, PA, USA Basics and Learning Process
C. N. De Cecco
Department of Radiology and Radiological Science, ML is the science of how computers learn from data [2, 3]; it
Medical University of South Carolina, Charleston, SC, USA encompasses a wide range of statistical analysis algorithms
Department of Radiology and Imaging Sciences, that iteratively improve with exposure to data, meaning that
Emory University, Atlanta, GA, USA the performance of the system improves with time and expe-

© Humana Press 2019 893

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7_68
894 M. Eid et al.

Machine Learning Workflow

Unsupervised

Finds Intrinsic
Raw Unlabeled Structure
Data within the Data

Feature Model
Extraction Development
Supervised

Predictive
Raw Labeled Model: Predicts Estimated Outcome
Data Labels of New
Data

Adjusted Cost Function:

Weights and Compares Real
Parameters Outcome to
Estimated Outcome

Fig. 68.1  Typical workflow of machine learning algorithms. In super- the other hand, in unsupervised learning, the algorithm is given unla-
vised learning, the algorithm is given labeled data, i.e., the true answer, beled data, i.e., it does not know the ground truth. The objective is for
based on which it will learn to classify this data. The estimated outcome the algorithm to find intrinsic meaning in the data and classify them into
is then compared with the true outcome, and the weights and parame- groups of similar characteristics
ters are adjusted accordingly until optimal performance is reached. On

rience [4]. Thus, ML systems aim to develop mathematical approximate the desired outcome and create the best possible
models that interpret the provided data in a meaningful way model for future predictions [1, 5].
in order to later predict the label of new data; in other words, Practically, there are two ways for ML systems to learn,
to create a model for autonomous predictions (Fig.  68.1). known as supervised and unsupervised learning [7]
Because ML can extract meaningful patterns from large (Fig. 68.2). In supervised learning, the input data is labeled
amounts of data, which is a component of human intelli- with the correct answer. These labels can be as specific or as
gence, it is considered to be a branch of AI. general as needed. The system is then given a series of vali-
All ML algorithms require a set of input data and desired dation sets, also called training sets, which will be used to
outputs [5]. Input data such as images and specific report text train the algorithm and improve its performance until it
are fed to the neural network, which is the name given to com- reaches a point of no further improvement. At this point, the
puter systems that simulate the way in which human brains machine is given a test dataset in order to test the model’s
think and solve problems. This data is transmitted through a performance. For example, in coronary artery calcium detec-
collection of nodes. The node or artificial neuron multiplies tion, a ML algorithm undergoing supervised learning would
each of these inputs by a weight. Then it adds the multiplica- be given a dataset of coronary arteries images with calcified
tions and passes the sum to an activation function. Notice that plaques labeled as calcium. Once the algorithm has reached
input nodes do not have activation functions. The weights its optimal performance, it would be given an unlabeled
indeed reflect importance of that node or feature relative to the dataset of coronary arteries with the aim of autonomously
task at hand, such that more important features receive higher detecting calcium deposits. Most machine learning algo-
weights than less important or irrelevant features [6]. Nodes rithms applicable to radiology use supervised learning
can be grouped into a hidden layer. When input data is given (Figs.  68.3 and 68.4). On the other hand, in unsupervised
to the algorithm, the hidden layers act as feature detectors that learning, the machine is exposed to unlabeled data. The
play important roles in decoding the inputs and transforming objective is for the algorithm to generate its own labels,
them into what the output layer expects. This is referred to as detect patterns, and organize the data in a consequential fash-
the forward propagation phase. After the outcome is com- ion. These new labels can then be used in a supervised learn-
puted, the difference error between the calculated outcome ing system to generate a useful prediction model; this is
and the given true outcome is calculated and propagated in a called semi-supervised learning.
backward direction to estimate the contribution of each node In a classic three-layer neural network (input, hidden, and
to the total error. This is known as the back propagation phase. output), inputs are analyzed by a single hidden layer in order
Based on that information, the algorithm adjusts the weights. to produce an output. With advancements in processing and
Through a series of forward and back propagation, minimiz- computational power, the combination of multiple hidden
ing the error, the algorithm iteratively learns how to best layers has now become feasible, creating deep learning (DL)
68  Machine Learning and Artificial Intelligence in Cardiovascular Imaging 895

Machine Learning Overview

Deep
Learning Machine
Learning

Supervised Learning Unsupervised Learning

Regression Classification Cluster Analysis Dimension Reduction

a. Decision Trees
a. Linear Regression b. K-nearest Neighbors
b. Ordinary Least c. Support Vector Machine a. Principal Component
Squares Regression a. K-Means Clustering
d. Logistic Regression b. Hierarchical Analysis
c. Local Regression b. Linear Discriminant
e. Naïves Bayes Clustering
d. Neural Networks Analysis
f. Random Forests
e. Convolutional Neural g. Convolutional Neural
Networks Networks

Fig. 68.2  Overview of machine learning algorithms. Machine learning with image feature detection and are commonly used in medical imag-
algorithms can be divided into two most commonly used learning meth- ing. Decision tress, k-nearest neighbors, support vector machines, naive
ods: supervised and unsupervised learning. The choice of the optimal Bayes, and random forest are also common algorithms employed in
algorithm greatly depends on the problem to be solved. Neural net- medicine and radiology
works and convolutional neural networks deliver good performance

Training: The algorithm iteratively learns until it finds the best model to classify calcified lesions
Assigned Labels:
Calcified plaque
Non-calcified plaque

Training Dataset
Machine Learning
Algorithm
Feature Identification:
Intensity (HU)
Lesion length Extracted Features
Lesion location
Shape...

Predicting: The model is applied to a set of new images

New Dataset

Feature Identification:
Intensity (HU)
Lesions length Optimal Predictive
Extracted Features
Lesion location Model
Shape...

Fig. 68.3  Example of a machine learning algorithm development and feature are adjusted as the training progresses until the optimal predictive
training. In the training phase, the algorithm is taught to detect specific model is reached. In the predicting phase, the algorithm will use the fea-
image features, based on the labeled data of the training set, which will be tures it learned and their corresponding weights to assign labels, calcified
used to subsequently correctly classify future data. The weights for each and noncalcified plaque in this example, to never before seen data
896 M. Eid et al.

a c
Normal / Normal / Normal / Normal /
Abnormal Abnormal Abnormal Abnormal

LSTM LSTM LSTM LSTM

CNN CNN CNN CNN ...

Image0 Image1 Image2 Image3

Vessel Cross-sections Vessel Centerline

Fig. 68.4  Illustration of a machine learning algorithm for the detection algorithm where the green portion shows automatically detected cal-
of coronary artery calcium. (a) Scheme of a recurrent neural network cium and the red portion the artery segments with no detected calcium.
with long short term memory used at our institution for the automatic As can be seen, the algorithm performs as accurately as the reader. (c)
detection of coronary artery calcium. (b) Curved multiplanar recon- Case example of a false positive where the algorithm detects calcium in
struction of the left anterior descending artery where the arrows and the mid and proximal parts of the right coronary artery as shown by the
annotation represent coronary artery calcium as manually noted by a green centerline when in fact no calcium is present. Image courtesy of
reader and the centerline represents the automatic segmentation by the Siemens Healthineers

neural networks (Fig.  68.5). DL networks allow one to imaging applications, has the potential to outpace computa-
approach and solve more complex and advanced problems. tional power. For example, a coronary CT angiography study
In deep learning, a convolutional neural network (CNN) is a from a high-definition scanner can consume approximately
class of deep neural networks, most commonly applied to half a gigabyte of storage space. A patient of interest to a
analyzing visual imagery such as medical images. study might have up to 3 or 4 of these scans and another
gigabyte of medical records data, yielding 3 gigabytes of
data per patient as a conservative estimate. A recent study by
Limitations of Machine Learning Algorithms Motwani et  al. analyzed 10,000 patients [10] which, using
the previous example, would account for approximately 30
Computer science literature is full of debate regarding the terabytes of data. The practical limitations of computational
selection of algorithms and how much testing is necessary resources require the pairing of data, as was done in this
before optimal efficiency is achieved [8, 9], largely because study, to generate in the near term a less comprehensive data-
the accuracy of the algorithm is a function of dataset size and set based on values derived from CT images or portions of
time (i.e., computational power). Dataset size, particularly in medical records. In that same regard, the availability of
68  Machine Learning and Artificial Intelligence in Cardiovascular Imaging 897

a b
Input Hidden Output Input Hidden Output

Classic Machine Learning Deep Learning

c
Input Hidden Output

Convolutional Neural Network

Fig. 68.5  Concepts of machine learning algorithms commonly used in lem solving. (c) Convolutional neural networks are a particular form of
medical imaging analysis. (a) Classic three-layer machine learning deep learning which perform particularly well with image analysis.
scheme where the first layer represents the input, the second layer is They apply mathematical transformations on individual pixels, evaluate
composed of the hidden nodes containing the weights and parameters, the outcome, and eventually identify specific portions of the image.
and the third layer is the calculated output. (b) With deep learning, all When a number of these convolutions are combined, they can recognize
three layers of the classic scheme become one hidden layer. A number more complex structures within the image
of hidden layers can be combined allowing much more complex prob-
898 M. Eid et al.

labeled data for algorithm training can also be an obstacle to  pplications of Machine Learning
A
ML implementation. Truly, the legal procedures to acquire in Medicine
medical data are complex, and even completely anonymized
data must gain institutional review board approval prior to its The clinical practice of medicine is one of the fields that can
distribution. Because ML algorithms continually improve as largely benefit from machine learning systems. Through
they learn from new data, these systems typically require tre- these algorithms, hundreds of thousands of variables such as
mendous amounts of data to be fully optimized for a specific patient information, demographics, lab values, and radiology
task [11]. A potential solution to this problem is the creation and pathology images can be analyzed alongside medical lit-
of publically available “banks” of data, containing massive erature and applied to patient care for risk stratification as
amounts of images that have been validated and annotated by well as to determine appropriate patient management and
experts readily available for ML training purposes [12–17]. prognosis.
Lastly, even with available data, the quality of the training
data must be optimal to reach the algorithm’s best perfor-
mance. This is a well-known principle coined “garbage in, Clinical Decision Support Tool
garbage out” by computer scientists, in which poor-quality
input will always yield poor-quality output. As most clinical work is conducted in accordance with an
ML systems are able to make highly accurate predications evidence-based medicine approach, it is important for clini-
or diagnoses analyzing complex relationships. However, an cians to keep abreast of novel published medical evidence.
important pitfall of these systems when considering their However, the sheer number of new peer-reviewed journal
application in clinical work or medical research is overfitting publications is increasing at an exponential rate, making it
(Fig. 68.6). Overfitting is when a statistical model tends to overwhelming for physicians to stay current with new evi-
describe random errors or noise rather than the true underly- dence while treating patients. ML systems can be of great
ing relationship between the data of a training set. This typi- assistance in this regard with the ability to constantly analyze
cally occurs when the model is more complex than it needs literature for new evidence which, combined with patient
to be to reach optimal efficiency. For example, having too information, can recommend an appropriate course of treat-
many parameters relative to the number of observations or ment. IBM’s AI software, Watson, is currently being investi-
using too few training examples during learning can cause gated for such use, specifically for applications in oncology.
the algorithm to adjust to very specific but random features Watson’s particular advantage is its capacity for natural lan-
of the data that have no causative relationship to the ground guage processing. This refers to the ability of a computer to
truth. The result is a model that is too specific to the exam- understand human language as it is spoken or written as
ples it was shown, rendering overly optimistic results for its opposed to requiring a highly precise programming language
accuracy while underperforming when exposed to unfamil- or very clear but brief voice commands. In a study by Seidman
iar real-world data. et al., the ability of Watson to recommend multidisciplinary

Under Fitting Appropriate Fitting Over Fitting

Feature 1

Feature 1

Feature 1

Feature 2 Feature 2 Feature 2

Fig. 68.6  Illustration of under- and overfitting. Underfitting occurs underlying trend in the data too well and is generally too good to be
when the algorithm is too simple to capture the underlying trend in the true, and the model shows low bias but high variance. This tends to hap-
data. Specifically, the algorithm shows low variance but high bias. An pen when the algorithm is excessively complex and is capturing noise
appropriate fit is seen when the algorithm performs well and in a bal- as part of the data. Validation and cross-validation with test sets is cru-
anced way on both training and test sets without being to flexible or cial to prevent overfitting
inflexible in fitting the data. Overfitting on the other hand captures the
68  Machine Learning and Artificial Intelligence in Cardiovascular Imaging 899

patient management in early stage breast cancer was investi- Machine learning systems learn in an analogous way to a
gated [18]. The software was able to correctly recommend resident in training, from observation. The more they see, the
radiotherapy in 98% of cases, genetic counseling in 94%, and better they get. However, one superior aspect of ML systems
fertility preservation in 91%. is that information is never forgotten. This is particularly
In a time when healthcare systems are facing physician important for rare diseases; which ML systems may poten-
shortages and increasing quantities and complexities of tially identify more efficiently than trained physicians to
patient data are being recorded with the introduction of elec- allow for timelier interventions. In this regard, Liu et  al.
tronic health records, the increased workload for healthcare investigated the accuracy of a ML algorithm in detecting
providers can result in oversight of important information. congenital cataracts from ocular images, compared with
Given that correct diagnosis and appropriate patient manage- detection by expert ophthalmologists [22]. The algorithm
ment are the physician’s most important tasks, ML systems achieved an accuracy of 99% and suggested the correct treat-
can be implemented in situations where they can serve as ment plan in 97% of cases. When applied to three sets of 100
highly efficient assistants by analyzing thousands of data in ocular images, each containing one case of congenital cata-
the form of clinical notes, lab results, radiology images, elec- ract, the algorithm successfully identified and provided accu-
trocardiogram tracings, and even genomic sequences. With rate treatment recommendations while making fewer
all of this information, ML systems can then infer clinical mistakes than the three expert ophthalmologists. Early detec-
causations that may eventually be presented to the physician tion of disease often translates into improved patient progno-
in real time to assist with clinical decision-making. This sis and is an area of medicine where ML can be of great of
application could be particularly helpful in acute care set- use. Investigators have used DL methods to detect skin
tings where physicians must make lifesaving decisions in a ­cancer, which can be lifesaving if diagnosed early. The abil-
timely manner. One such setting in which ML systems are ity of the DL algorithm to detect skin cancer was compared
being evaluated is the intensive care unit (ICU). Patients with 21 expert dermatologists [23]. The algorithm’s perfor-
admitted to the ICU typically present with acute illness and mance was comparable to the dermatologists, demonstrating
are at a higher risk of mortality. As large amounts of patient the ability of ML to accurately diagnose skin lesions. These
data are routinely recorded in the ICU, it seems an attractive two studies show the potential added value of ML applica-
setting to input data into ML models, infer causalities, and tions in particularly rural areas, where access to expert clini-
predict outcomes, with final predictions being presented to cians is not readily available and early diagnosis of rare or
physicians in a clinically actionable manner [19]. In a retro- fatal diseases is more likely to be missed.
spective study by Gultepe et al., the authors used ML algo-
rithms in septic patients in the ICU to predict rising lactate
levels, a sign of organ failure, as well as mortality with rou- Predicting Prognosis and Outcome
tinely measured data [20]. Results demonstrated the ability
to predict rising lactate levels by using vital signs and white Predicting prognosis and outcome is a central component of
blood cell counts as inputs, achieving a discriminatory power clinical practice as it is directly related to choosing an appro-
of 0.98 by area under the curve (AUC). Moreover, their ML priate management plan. Most scoring systems used are
system was able to predict mortality with a discriminatory based on a limited number of clinical, laboratory, or imaging
power and accuracy of 0.73. These results highlight the findings. ML systems have the ability to consider larger
potential of using ML algorithms for early detection of sep- quantities and complexities of variables, potentially generat-
sis and multiple organ failure, prompting a timely interven- ing more accurate prediction models based on novel or
tion and possible reduction of mortality. unknown predictors. As a result, the evaluation of ML algo-
In a more recent study by Gulshan et al., a DL algorithm rithms for the prediction of prognosis and outcomes in dif-
was used to screen for referable diabetic retinopathy and ferent clinical settings compared with routine scoring
macular edema from acquired retinal images. Using the systems is of great importance.
algorithm on two different datasets, they were able to reach One particular field of interest for this application is
an AUC of 0.991 and 0.990, sensitivities of 97.5% and oncology, where diagnosis and prognosis are often derived
96.1%, and specificities of 93.4% and 93.9% [21]. This study from histopathological analysis. In a study by Beck et al., the
demonstrated the potential of integrating ML systems to authors investigated the ability of a ML system named
enhance routine clinical workflow where simple validated “C-Path” to perform a semiquantitative analysis of breast
tests can be analyzed to infer a diagnosis. This can be par- histopathology slides and to infer patient survival rates [24].
ticularly helpful when a disease is very prevalent and a vali- The algorithm was able to detect 6642 morphological fea-
dated screening method requiring time and resources can be tures used to construct a prognostic model. The features rec-
delegated to ML systems, thus allowing clinicians to focus ognized by “C-Path” were independent from prognostic
on more important and urgent tasks. factors used in clinical routine and were significantly
900 M. Eid et al.

a­ ssociated with patient survival (P-value < 0.001). Of the 11 to predict treatment response in patients with ovarian cancer
most important features associated with patient survival, 3 based on individual genomic profiles was evaluated [29].
were related to the stroma and proved to be the strongest Gene expression data was found to provide accurate results
predictors of survival. Interestingly, most morphological fea- by correctly predicting treatment responsiveness with
tures used for prognosis in clinical practice are derived from 80–100% accuracy.
the epithelium. Since unsupervised learning allows the algorithm to
In a similar, yet more recent study by Yu et al., a ML algo- detect meaningful patterns from unlabeled data, it can be an
rithm was used to predict non-small cell lung cancer progno- additional method suitable for precision medicine. Indeed,
sis using a fully automated analysis of histopathologic through unsupervised learning, thousands of patient data
features [25]. Based on 2186 histopathology slides, the algo- regarding specific diseases with particularly heterogeneous
rithm was able to detect more than 9000 features to predict clinical presentations can be analyzed by ML algorithms to
patient prognosis. The top 80 features were first used to deduce more accurate risk stratification based on individual
accurately differentiate cancerous tissue from normal tissue, patient phenotypes. In a recent study by Shah et al., the clini-
achieving an AUC of 0.85. These features were also able to cal and radiological data of 397 patients diagnosed with
differentiate between adenocarcinoma and squamous cell heart failure with preserved ejection fraction were analyzed
carcinoma with an AUC of 0.85. For both adenocarcinoma by a ML algorithm [32]. The algorithm was able to divide the
and squamous cell carcinoma, the features detected by the patients into three clinically distinct subgroups based on
algorithm significantly correlated with patient survival. In clinical characteristics, cardiac structure and function, inva-
contrast, the tumor grade, a marker often used clinically, was sive hemodynamic parameters, and outcome. This study
not significantly correlated with patient survival. These two showed the potential of ML systems in creating accurate,
studies exemplify the ability of ML systems to predict sur- clinically similar subcategories of patients with rather het-
vival models based on a much greater number of features erogeneous diseases, allowing for improved risk ­stratification
with more complexities than previously unknown by clini- and ultimately defining therapeutically homogeneous
cians. This ability of ML has the potential to lead to a sub- subclasses.
stantial change in patient management.

 pplications of Machine Learning

A
 achine Learning: The Road to Precision
M in Radiology
Medicine?
ML software has had substantial success in image feature
An exciting aspect of using ML in medicine is the possibility detection and pattern recognition in daily life applications.
of precision medicine, defined by the national institute of With radiology being, at least in part, a science of detecting
health as “disease treatment and prevention that takes into clinically relevant features and patterns from acquired
account individual variability in genes, environment, and images, it naturally serves as a prime candidate for ML
lifestyle for each person” [26]. Indeed, while medicine to applications. At their earlier stages, these systems were ham-
date has been based on a population-wide diagnostic and pered by low computational power, limiting the amount of
treatment model, it is possible that in many cases the “stan- data that could be processed. However currently, with
dard of care” treatment may not be the optimal management increased computational power, graphics processing unit-­
plan. Integrating specific patient data from demographics, based computing, and the development of DL, the applica-
clinical reports, lab values, imaging results, and genomic tion of ML to radiology has become feasible and is gaining
sequences to be analyzed by the machine can be combined interest from physicians as well as computer scientists.
with decision support tools such as IBM’s Watson. Decision As radiology has become a central part of patient manage-
support tools are constantly updated with the most recent lit- ment, with new imaging technologies constantly creating
erature with the potential to provide improved, patient-­ new possibilities for solving clinical problems, radiologists
tailored recommendations for risk assessment, outcome often find themselves dealing with an abundant amount of
prediction, and subsequent treatment and management. studies to review. Much of the time spent during the assess-
An important area where such use is being extensively ment of these studies is dedicated to more trivial tasks such
researched is oncology, particularly in genomic and molecu- as image segmentation, searching for lung nodules, micro-
lar medicine. In these areas of study, the input of clinical data calcifications in breast tissue, or looking for small bone frac-
and patient genomic and molecular profiles into ML systems tures in cases that are often negative. A clinical practice
such as Watson is helping to provide a diagnosis as well as where these tasks can be delegated to a ML algorithm would
patient-tailored treatment options [27–31]. In a study by result in a much more enhanced workflow, allowing the radi-
McDonald et al., a support vector machine-based algorithm ologist to focus on more important tasks such as looking at
68  Machine Learning and Artificial Intelligence in Cardiovascular Imaging 901

truly emergent or complex cases and making a prompt diag- MRI [46], interstitial lung disease [47], pulmonary nodules
nosis. The fairly recent and rapid improvements to ML have detection [48, 49] and classification [50], pulmonary embo-
allowed these systems to handle much more complex tasks, lism [51], classification of chest radiographs [52], prostate
constantly learning and improving their performance more cancer [53], and breast cancer [54–57].
efficiently. As a result, completion of basic tasks using ML
systems has become feasible, and more applications for their
use, such as computer-aided diagnosis and image segmenta- Machine Learning in Cardiovascular Imaging
tion, are continuously being evaluated [33].
Image segmentation is an important step in reaching an Radiology should be at the forefront of medical specialties
accurate diagnosis for both radiologists and ML systems if embracing ML applications, and cardiac imaging has proven
they are to reach a point of automated image analysis in the to be one of the first specialties quickly adapting to that
future. However, it is a time-consuming process prone to change. The field has seen a wide range of ML uses such as
inter- and intraindividual variability. As imaging technolo- image segmentation, feature detection, CADx, risk, and
gies improve, more detailed anatomy is being captured, and prognosis prediction in a variety of modalities such as CT,
the need to perform time-consuming tasks such as segmenta- MRI, nuclear imaging, and echocardiography [58–62].
tion will only increase. While automating this process can be
beneficial by means of enhancing a radiologist’s workflow, it
remains a challenge for ML systems due to the high variation Computer-Aided Diagnosis
in human anatomy, thus requiring a tremendous amount of
data to reach optimal accuracy. However, with recent Machine-based evaluation of cardiac datasets involving
advancements in ML technology and particularly the devel- human-specified parameters is well established in clinical
opment of DL, novel algorithms are outperforming previous practice. However, new forms of analysis are emerging that
versions and are now capable of faster, more accurate, and allow the machine to evaluate what parameters should be
even fully automated image segmentation. Recent studies are measured based on a known endpoint. Early attempts at such
demonstrating this capability with applications for brain seg- systems struggled with the volume of data and the number of
mentation on MRI [34–36], rib cage, airway, pulmonary potential variables creating processing times that prohibited
lobes and nodules from chest CT [37–39], tumor and organs-­ large, high-quality evaluation. Despite these early shortcom-
at-­risk segmentation from CT images in radiation oncology ings, advances in both processing power and model sophisti-
[40, 41], third lumbar vertebra detection from CT [42], and cation are reaching a point where systems can be trained
fully automated abdominal analysis from CT [43]. using information from tens of thousands of patients with
A much anticipated application of ML in radiology is hundreds of thousands of data points being evaluated. This
computer-aided diagnosis (CADx). While computers have type of analysis requires both sophisticated ML algorithms
been helping radiologists for decades, it is expected that and specialized statistics to elucidate new parameters for
increased computational power and more sophisticated algo- clinical evaluation. While many of these methods are still in
rithms such as DL will render them faster, more efficient, and the nascent phase of research, some are now emerging as a
more accurate in completing this task [44]. Indeed, computer part of clinical workflows.
systems are now much better at detecting more complex Computational support of cardiac imaging utilizing basic
radiological findings and diagnosing a wider array of condi- algorithms is not a new concept. Quantifying image density
tions. An exciting example is IBM’s Watson. In acquiring for gross measurement of cardiac calcium scores is common
Merge Healthcare in 2015 to the tune of $1 billion, IBM was place and provides meaningful clinical guidance in many
able to provide Watson with over 30 billion images to learn diagnostic scenarios. Likewise, pattern recognition for the
from. Presented at the Radiological Society of North America identification and lengthwise evaluation of coronary arteries
conference in 2016, in an important step for precision medi- reduces the complexity of coronary artery stenosis. Similar
cine and CADx in radiology, Watson showcased its potential pattern recognition systems are being used for CT-FFR,
to analyze patient history and clinical data in addition to automated cardiac volume measurements, and wall motion
radiological images to reach a diagnosis in a number of con- characterization [63]. These algorithms are early forms of
ditions including aortic dissection, pulmonary embolism, and ML-based clinical support applications, as the data they gen-
breast lesions. Moreover, several studies have already shown erate are fed into human algorithms used to predict clinical
the ability of ML systems in detecting abnormalities with an risk. These applications are categorized as diagnostic sup-
accuracy comparable if not superior to that of expert radiolo- port tools.
gists for different conditions such as colonic polyps on CT Diagnostic support tools may rely on ML for their devel-
colonography [45], sclerotic bone metastases and enlarged opment, but differ fundamentally from true ML systems in
lymph nodes on body CT [44], cerebral microbleeds from that they are not adaptive. In other words, once the pattern
902 M. Eid et al.

recognition is established, it is fixed and cannot be retrained. improving. Previous studies have evaluated computer algo-
ML systems can potentially be able to adapt to unpredicted rithms for the detection of obstructive CAD from CCTA
patterns that generate errors in pattern recognition [69–73]. While these studies all reported high sensitivity val-
applications. ues, the corresponding specificity values were relatively low.
As mentioned earlier, accurate image segmentation is a In a recent study by Kang et al., a ML algorithm was evalu-
crucial part of image diagnosis and a critical step in reaching ated for the detection of both obstructive and nonobstructive
a fully automated analysis of radiological studies. As opposed CAD compared with three expert readers [74]. The proposed
to the current CADx software which follow a rigid set of algorithm achieved higher sensitivity, specificity, accuracy,
rules, ML systems could potentially learn as they are exposed and AUC values (93%, 95%, 94%, and 0.937, respectively)
to more data and are thus prone to adapt to the variation in compared with previous methods which reached a sensitivity
individual anatomies. This inherently results in a more accu- and specificity of 93% and 81%, respectively [69, 70, 73].
rate segmentation process across the board. The ability to autonomously diagnose both obstructive and
In a study by Avendi et al., the authors evaluated the per- nonobstructive CAD with high accuracy, as demonstrated in
formance of a ML algorithm using convolutional neural net- this study, brings the concept of automated diagnosis in car-
works in segmenting the left ventricle from magnetic diac imaging one step closer to a reality.
resonance imaging datasets for the evaluation of cardiac func- The determination of calcium burden in the coronary
tion [64]. Their proposed algorithm performed better than arteries through coronary artery calcium scoring (CACS) is
previous methods by reaching good contours in 96.69% of another important aspect of the diagnosis of CAD, risk strati-
cases and correlation coefficients of 0.98, 0.99, and 0.98 for fication for adverse events, and the subsequent direction of
end-diastolic volume, end-systolic volume, and ejection frac- appropriate patient management [75, 76]. CACS is usually
tion, respectively, when compared with manual evaluation. performed following a noncontrast cardiac scan with a semi-
Myocardial perfusion imaging (MPI) is another field of automatic calculation of the Agatston score, which often
cardiac imaging where ML is quickly gaining ground in requires manual intervention by a radiologist. In a study by
terms of its application. MPI plays an important role in the Wolterink et al., a convolutional neural network-based ML
diagnosis of coronary artery disease (CAD) as it provides algorithm was used to automatically derive calcium scores
important information regarding myocardial perfusion, from contrast-enhanced CCTA [77]. The study included 250
which is directly related to the significance of coronary patients who had undergone both CCTA and CACS scans.
artery stenoses as well as ventricular function information. A two-step system was used in which the primary algorithm
Although perfusion quantification with MPI is becoming would first identify voxels that are likely to be calcium, fol-
increasingly automated [65, 66], diagnosis remains visually lowed by a secondary algorithm that definitively classifies
based by physicians who often need to supervise the seg- the remaining voxels as either calcium or calcium like. Using
mentation process, as errors in segmentation may lead to this technique for calcium detection, a sensitivity of 71%
false positives [67]. In a study performed by Betancur et al., was achieved with an average of 0.48 false positives per
a support vector machine-based ML algorithm was used for scan, providing a correlation of 0.944 with expert readers for
a more precise automated definition of the mitral valve plane calcium scores. This study shows the possibility of using ML
during left ventricle segmentation in single-photon emission to accurately compute CACS from CCTA, potentially pre-
computed tomography (SPECT) [68]. The algorithm was cluding the need for a dedicated CACS scan. Furthermore,
tested on 392 patients who had undergone same-day SPECT application of such a ML algorithm may reduce, if not elimi-
and coronary computed tomography angiography (CCTA) nate, the need for manual intervention by a radiologist.
after a tenfold cross-validation process and was compared ML for diagnostic support has also been studied on car-
with the results of two expert readers. Results determined diac echocardiography. In their study, Sengupta et al. used an
that their ML algorithm performed just as well as the two associative memory classifier-based ML algorithm to differ-
expert readers regarding both mitral valve placement and entiate constrictive pericarditis from restrictive cardiomyop-
diagnostic accuracy for perfusion defects, with no significant athy on echocardiographic images [78]. An AUC of 0.892
difference observed in correlation or AUC between the algo- was achieved using 15 variables from speckle-tracking echo-
rithm and the expert readers. cardiography chosen by the algorithm. The addition of four
Reaching a point where machines can make a full diagno- echocardiographic features (early diastolic mitral valve
sis and radiology workflows can be fully automated remains velocity, ratio of early diastolic mitral flow velocity to early
a challenge, particularly in cardiac and chest imaging. diastolic mitral valve annular velocity, posterior and septal
Potential roadblocks include variation in patient anatomy as wall thickness) improved the AUC to 0.962. In comparison,
well as the high presence of motion artifacts, which make it the AUC of the clinically used markers of early diastolic
more challenging for automated systems to accurately detect mitral valve annular velocity and left ventricular longitudinal
pathology. However, the accuracy of ML systems is rapidly strain were 0.821 and 0.637, respectively.
68  Machine Learning and Artificial Intelligence in Cardiovascular Imaging 903

Prognosis and Outcome Prediction between high- and low-risk patients. This technique outper-
formed the current reference standard based on scar size and
As new technologies and imaging techniques are being intro- left ventricular function, which reached an accuracy and an
duced, radiology is increasingly relevant in appropriate AUC of 90.7% and 0.941, respectively.
patient management, not only for diagnostic purposes but In a study by Arsanjani et al., ML analysis was used to
also for the prediction of outcomes via incorporation of the integrate clinical and quantitative MPI data in order to pre-
many radiological features of diseases into prediction scores. dict revascularization in patients with suspected CAD, com-
However, those scores tend to rely on a limited number of pared with two expert readers and standalone measures of
factors, potentially omitting many unknown predictors from perfusion [86]. The algorithm analyzed data from 713
consideration. As big data analysis has become feasible patients, selected 33 features, and was tested through tenfold
through the recent advancements in ML, using these systems cross-validation. Given both stress and rest perfusion data,
to predict outcomes from radiological data has gained sig- the sensitivity of the algorithm for predicting early revascu-
nificant interest, particularly in the field of cardiac imaging larization was comparable to readers and standalone mea-
given the high morbidity and mortality rates associated with sures of perfusion (73.6%, 73.9%, and 75.5%, respectively,
cardiovascular disease. P > 0.05). However, its specificity was found to be superior
Indeed, coronary artery disease (CAD) is one of the most to both readers and standalone measures of perfusion (74.7
important causes of mortality and morbidity worldwide [79]. vs 67.2, 66, and 68.3, respectively), resulting in the highest
CCTA has been established as the modality of choice for the AUC for ML (0.81 vs 0.81, 0.72, and 0.77, respectively).
detection and characterization of atherosclerotic plaques, Thus, ML was found to be comparable, if not superior, to
and several studies have shown the ability to predict future expert readers in predicting early revascularization from
adverse events based on CCTA findings [80–83]. Thus, it is MPI data. This could result in a significant change in patient
of great interest to evaluate the application of ML in predict- management, particularly in reducing unnecessary revascu-
ing outcomes. In a recent prospective study including 10,030 larization procedures via improved patient risk stratification.
patients by Motwani et al., the authors investigated the abil- These studies not only suggest the ability of ML systems to
ity of a ML algorithm to predict 5-year all-cause mortality predict risk and outcome in a manner comparable, if not
based on 44 CCTA features and 25 clinical features, com- superior to routinely used scores, but also the possibility of
pared with established clinical and CCTA metrics [10]. The discovering new valid and potentially more accurate predic-
algorithm was found to be significantly superior to routinely tors that were previously unknown to clinicians and radiolo-
used scores, achieving an AUC of 0.79 vs 0.61 for the gists alike.
Framingham risk score, 0.64 for the segment stenosis score,
0.64 for the segment involvement score, and 0.62 for the
Duke index with all P-values reported <0.001. Adapting to Machine Learning
Dawes et al. also used a supervised ML algorithm on car-
diac MRI-derived three-dimensional wall function of the Although medicine has been slower than other fields in pur-
right ventricle to predict outcomes in patients with pulmo- suing ML applications, early studies show very promising
nary hypertension [84]. The algorithm identified certain wall results. As computational power and ML algorithms improve
contraction patterns associated with lower survival at 1 year. at an exponential speed, it is only a matter of time before AI
Particularly, loss of effective contraction in the septum and becomes a central part of medicine. Thus, it would seem
free wall as well as reduced basal longitudinal motion accu- wise for medical practices to start planning for such a future,
rately predicted poor outcomes. When these features were with pathology and radiology being at the forefront of this
added to routine hemodynamic, clinical, and functional change [87]. While some experts fear or even predict pathol-
markers, survival prediction significantly improved with an ogists and radiologists to be completely replaced by ML
AUC of 0.73 vs 0.60 (P < 0.001) and was able to more accu- applications, this is a premature statement due to the many
rately stratify patients between high- and low-risk groups remaining challenges in adapting to ML.
based on median survival time (13.8 vs 10.7 years; P < 0.001). In order for ML systems to be optimally efficient in ana-
In another study by Kotu et al., ML algorithms were used lyzing big data and integrated into clinical routine, they must
to extract image-based features from cardiac MRI for the be seamlessly interfaced with electronic medical records and
classification of post-myocardial infarction patients into high other established systems. For example, in the field of radiol-
and low arrhythmic risk groups [85]. The k-nearest neighbor ogy, ML systems must be integrated into image interpreta-
and support vector machine-based algorithm evaluated the tion systems to avoid impeding clinical workflow. This is
combination of 17 different factors for risk stratification. crucial for implementation, as it would not be feasible to
A combination of scar location and heterogeneity achieved an manually enter hundreds of data points for each patient into
accuracy of 94.44% and an AUC of 0.965 in differentiating the ML system to generate a recommendation or risk score.
904 M. Eid et al.

Furthermore, regarding the enhancement of clinical work- rithm was used to analyze patient ECG tracings in the ICU
flows, further multicenter large-scale clinical trials are [89]. The algorithm was able to detect atrial fibrillation, which
needed to validate the accuracy and efficiency of ML sys- was classified as clinical when already documented and sub-
tems reported in previous studies. Particularly, results shown clinical when undocumented by the clinicians. However,
in single-center trials must be validated in studies spanning upon further analysis, subclinical arrhythmia was not found
over a larger range of patient demographics, image qualities, to be related to clinical outcomes or worsened prognosis.
and disease patterns and prevalence, as these often vary in Although ML was better suited to detect subclinical atrial
different populations. The argument for further validation fibrillation, this had no clinical significance and could lead to
through multicenter trials centers on the idea that ML algo- overtreatment, possibly harming patients. As a result, ML
rithms will have to adapt to such variations to reach maximal data should be interpreted in light of clinical judgment.
accuracy in real-world scenarios. This then raises the impor- Finally, a potential obstacle for the implementation of ML
tant question: how is accuracy determined? Indeed, diagnos- systems in medical practice is physician resistance. As ML
tic accuracy is modeled on the reference standard used in systems exponentially improve and eventually overcome cur-
clinical practice. However, reference standards are not avail- rent shortcomings, physicians, particularly radiologists, fear
able in many situations, and even when it is, opinions vary that such systems could eventually replace them. Indeed, as
between experts regarding diagnosis and treatment manage- these systems perfect themselves overtime and are immune to
ment. Consequently, it will be important to determine the fatigue, a future where medical institutions favor machine
appropriate reference standard to evaluate the accuracy of analysis over human analysis is imaginable. In a recent arti-
ML systems. Furthermore, ML systems will need to learn cle, Chockley et al. coined the term “ultimate threat to radiol-
every “normal” in different clinical scenarios. For example, ogy” while describing ML, expressing the very real possibility
the brain of a 70-year-old patient might look completely dif- of the end of radiology as a specialty within the next 10 years.
ferent from that of a younger person without necessarily hav- Chockley particularly emphasized the public’s skepticism
ing clinical implications. about driverless cars not more than a decade ago, and how far
Another important challenge arises when considering that technology has reached today [90]. However, if the
both ethical and legal concerns. Electronic health records are extinction of radiology as a specialty were to occur, it is still
slowly becoming the standard in clinical practices, which quite far from becoming reality. In fact, a capable radiologist
raises concerns for patient security and privacy. Despite this is trained to detect 2613 findings and account for more than
software being advertised as secure and impenetrable, infor- 20,000 medical conditions across all imaging domains and
mation stored on servers is always at risk of breaches. In fact, modalities. As it takes substantial effort in terms of time and
recently targeted medical institutions all around the world data needed to create an algorithm accurate in the detection of
were compromised, with valuable patient information being a very specific finding on a single modality, it begs the ques-
held for ransom [88]. Adding more technological compo- tion: If ever possible, how long it will take to create algo-
nents such as ML systems to the medical field will only rithms that can accurately detect all findings across all
increase the complexity of this problem. Thus, finding an modalities so as to completely replace radiologists? In the
effective way of securing ML systems will be central to their coming years, it is likely that ML will at least displace some
integration into clinical workflows. Moreover, in a future of the more routine work done by radiologists that can easily
where machines could potentially diagnose and even recom- and safely be delegated to machines [91]. As Jha et al. stated,
mend treatment options, the questions of general patient a higher education is not essential for recognizing patterns
reception and legal responsibility for system errors remain and specific signs on an image [87]. However, they argue that
ambiguous. Would it be the treating physician who is respon- radiologists will still be needed as “information specialists”
sible for mistakes, the medical institution, the system’s com- to interpret and highlight the important information provided
pany, or the software’s developer? What happens if a by the machine while considering the patient’s clinical set-
physician chooses to follow his experience-based clinical ting. Furthermore, radiologists will remain desirable for the
judgment rather than a ML system recommendation, result- foreseeable future to help guide the treating team toward a
ing in an adverse event? These are just on the surface of diagnosis and further patient management.
potential obstacles that must be addressed before ML can
find its way into routine clinical practice.
Since ML will be able to integrate much more data than Summary
humanly possible, it has the potential to lead to new correla-
tions that were previously unknown to clinicians. However, The recent advancements in computational power and the
these new findings require interpretation with a critical eye in development of novel, more sophisticated ML algorithms
order to avoid overdiagnosing, overtreating, and potentially have made big data analysis in medicine feasible. Thus, ML
harming patients. In a recent study by Moss et al., a ML algo- is seeing an increasingly important role in clinical practice
68  Machine Learning and Artificial Intelligence in Cardiovascular Imaging 905

with a variety of applications such as clinical decision sup- evaluating coronary artery stenosis detection, stenosis quantification
and lumen segmentation algorithms in computed tomography angi-
port, precision medicine, risk stratification, and prognosis ography. Med Image Anal. 2013;17(8):859–76.
prediction. As these systems become particularly apt at rec- 16. Grand Challenges in Biomedical Imaging. 2016. Available from:
ognizing patterns from images, cardiovascular imaging is https://grand-challenge.org/All_Challenges/.
seeing an increased role of ML in clinical practice for image 17. Data Science Bowl. 2016. Available from: https://www.kaggle.

com/c/second-annual-data-science-bowl.
segmentation, CADx, and mortality prediction. The more 18. Seidman AD, Pilewskie ML, Robson ME, Kelvin JF, Zauderer MG,
likely applications in the coming years will be to enhance Epstein AS, et al. Integration of multi-modality treatment planning
clinical workflow through highly accurate tools to better for early stage breast cancer (BC) into Watson for Oncology, a
detect disease, assess risk, and stratify patients. Thus, ML Decision Support System: Seeing the forest and the trees. J Clin
Oncol. 2015;33(15_suppl):e12042–e.
should be viewed as an opportunity in a positive light, as 19. Johnson AE, Ghassemi MM, Nemati S, Niehaus KE, Clifton DA,
optimized systems will utilize all patient data to improve Clifford GD. Machine learning and decision support in critical care.
quality, efficiency, and outcomes, leading to optimized Proc IEEE. 2016;104(2):444–66.
patient and patient-specific care. 20. Gultepe E, Green JP, Nguyen H, Adams J, Albertson T, Tagkopoulos
I. From vital signs to clinical outcomes for patients with sepsis: a
machine learning basis for a clinical decision support system. J Am
Med Inform Assoc. 2014;21(2):315–25.
References 21. Gulshan V, Peng L, Coram M, Stumpe MC, Wu D, Narayanaswamy
A, et al. Development and validation of a deep learning algorithm
1. Erickson BJ, Korfiatis P, Akkus Z, Kline TL. Machine learning for for detection of diabetic retinopathy in retinal fundus photographs.
medical imaging. Radiographics. 2017;37(2):505–15. https://doi. JAMA. 2016;316(22):2402–10.
org/10.1148/rg.2017160130. 2
2. Long E, Lin H, Liu Z, Wu X, Wang L, Jiang J, et al. An artificial
2. Friedman J, Hastie T, Tibshirani R.  The elements of statistical intelligence platform for the multihospital collaborative manage-
learning: Springer series in statistics. Berlin: Springer; 2001. ment of congenital cataracts. Nat Biomed Eng. 2017;1:0024.
3. Abu-Mostafa YS, Magdon-Ismail M, Lin H-T. Learning from data. 23. Esteva A, Kuprel B, Novoa RA, Ko J, Swetter SM, Blau HM, et al.
New York: AMLBook; 2012. Dermatologist-level classification of skin cancer with deep neural
4. Jordan M, Mitchell T. Machine learning: trends, perspectives, and networks. Nature. 2017;542(7639):115–8.
prospects. Science. 2015;349(6245):255–60. 2
4. Beck AH, Sangoi AR, Leung S, Marinelli RJ, Nielsen TO, van de
5. Kohli M, Prevedello LM, Filice RW, Geis JR.  Implementing Vijver MJ, et al. Systematic analysis of breast cancer morphology
machine learning in radiology practice and research. AJR Am J uncovers stromal features associated with survival. Sci Transl Med.
Roentgenol. 2017;208(4):754–60. 2011;3(108):108ra13.
6. Martinez TR, Zeng X. Feature weighting using neural networks. In: 2
5. Yu KH, Zhang C, Berry GJ, Altman RB, Re C, Rubin DL, et  al.
Proceedings of the IEEE International Joint Conference on Neural Predicting non-small cell lung cancer prognosis by fully auto-
Networks IJCNN’04; 2004. p. 1327–30. mated microscopic pathology image features. Nat Commun.
7. Alpaydin E. Introduction to machine learning. Cambridge: The 2016;7:12474.
MIT Press; 2014. 2
6. Precision Medicine Initiative. 2016. Available from: https://www.
8. Langley P.  Machine learning as an experimental science. Mach nih.gov/news-events/news-releases/nih-awards-55-million-build-
Learn. 1988;3(1):5–8. million-person-precision-medicine-study https://www.nimhd.nih.
9. Özöğür-Akyüz S, Ünay D, Smola A.  Guest editorial: model gov/programs/collab/pmi/ https://syndication.nih.gov/multimedia/
selection and optimization in machine learning. Mach Learn. pmi/infographics/pmi-infographic.pdf.
2011;85(1):1. 27. Artificial Intelligence Used To Detect Rare Leukemia Type In

10. Motwani M, Dey D, Berman DS, Germano G, Achenbach S,
Japan. 2016. Available from: http://www.ndtv.com/health/artificial-
Al-Mallah MH, et al. Machine learning for prediction of all-cause intelligence-used-to-detect-rare-leukemia-type-in-japan-1440789.
mortality in patients with suspected coronary artery disease: a 2
8. Sharpless N.  As seen on 60 minutes: Watson accelerates preci-
5-year multicentre prospective registry analysis. Eur Heart J. sion oncology. 2016. Available from: https://www.ibm.com/blogs/
2017;38(7):500–7. think/2016/10/sharpless/.
11. Greenspan H, van Ginneken B, Summers RM.  Guest editorial 29. McDonald JF, Mezencev R, Long TQ, Benigno B, Bonta I, Priore

deep learning in medical imaging: Overview and future prom- GD.  Accurate prediction of optimal cancer drug therapies from
ise of an exciting new technique. IEEE Trans Med Imaging. molecular profiles by a machine-learning algorithm. J Clin Oncol.
2016;35(5):1153–9. 2015;33(15_suppl):e22182–e.
12. Petersen SE, Matthews PM, Bamberg F, Bluemke DA, Francis JM, 30. Ramarajan N, Badwe RA, Perry P, Srivastava G, Nair NS, Gupta
Friedrich MG, et  al. Imaging in population science: cardiovascu- S. A machine learning approach to enable evidence based oncol-
lar magnetic resonance in 100,000 participants of UK Biobank-­ ogy practice: ranking grade and applicability of RCTs to individual
rationale, challenges and approaches. J Cardiovasc Magn Reson. patients. J Clin Oncol. 2016;34(15_suppl):e18165–e.
2013;15(1):46. 31. Sidaway P. Immunotherapy: genomic and immunological features
13. Fonseca CG, Backhaus M, Bluemke DA, Britten RD, Do Chung predict a response. Nat Rev Clin Oncol. 2017;14(5):263.
J, Cowan BR, et al. The Cardiac Atlas Project—an imaging data- 32. Shah SJ, Katz DH, Selvaraj S, Burke MA, Yancy CW, Gheorghiade
base for computational modeling and statistical atlases of the heart. M, et  al. Phenomapping for novel classification of heart failure
Bioinformatics. 2011;27(16):2288–95. with preserved ejection fraction. Circulation. 2015;131(3):269–79.
14. del Toro OAJ, Goksel O, Menze B, Müller H, Langs G, Weber https://doi.org/10.1161/CIRCULATIONAHA.114.010637.
M-A, et  al. VISCERAL–VISual Concept Extraction challenge in 33. Wang S, Summers RM.  Machine learning and radiology. Med
RAdioLogy: ISBI 2014 challenge organization. Proceedings of the Image Anal. 2012;16(5):933–51.
VISCERAL Challenge at ISBI. 2014;1194:6–15. 34. Akkus Z, Galimzianova A, Hoogi A, Rubin DL, Erickson BJ. Deep
15. Kirişli H, Schaap M, Metz C, Dharampal A, Meijboom WB, learning for brain MRI segmentation: state of the art and future
Papadopoulou S-L, et  al. Standardized evaluation framework for directions. J Digit Imaging. 2017;30:449.
906 M. Eid et al.

35. Kamnitsas K, Ledig C, Newcombe VFJ, Simpson JP, Kane AD, 55. Kooi T, Litjens G, van Ginneken B, Gubern-Mérida A, Sánchez CI,
Menon DK, et  al. Efficient multi-scale 3D CNN with fully con- Mann R, et al. Large scale deep learning for computer aided detec-
nected CRF for accurate brain lesion segmentation. Med Image tion of mammographic lesions. Med Image Anal. 2017;35:303–12.
Anal. 2017;36:61–78. 56. Lassen B, van Rikxoort EM, Schmidt M, Kerkstra S, van Ginneken
36. Havaei M, Davy A, Warde-Farley D, Biard A, Courville A, Bengio B, Kuhnigk J-M. Automatic segmentation of the pulmonary lobes
Y, et al. Brain tumor segmentation with deep neural networks. Med from chest CT scans based on fissures, vessels, and bronchi. IEEE
Image Anal. 2017;35:18–31. Trans Med Imaging. 2013;32(2):210–22.
37. van Ginneken B. Fifty years of computer analysis in chest imaging: 57. Wang J, Yang X, Cai H, Tan W, Jin C, Li L. Discrimination of breast
rule-based, machine learning, deep learning. Radiol Phys Technol. cancer with microcalcifications on mammography by deep learn-
2017;10(1):23–32. ing. Sci Rep. 2016;6:27327.
38. Doel T, Gavaghan DJ, Grau V.  Review of automatic pulmonary 58. Dilsizian SE, Siegel EL.  Artificial intelligence in medicine and
lobe segmentation methods from CT. Comput Med Imaging Graph. cardiac imaging: harnessing big data and advanced computing to
2015;40:13–29. provide personalized medical diagnosis and treatment. Curr Cardiol
39. Mansoor A, Bagci U, Foster B, Xu Z, Papadakis GZ, Folio LR, Rep. 2013;16(1):441.
et al. Segmentation and image analysis of abnormal lungs at CT: 59. Krittanawong C, Zhang H, Wang Z, Aydar M, Kitai T.  Artificial
current approaches, challenges, and future trends. Radiographics. intelligence in precision cardiovascular medicine. J Am Coll
2015;35(4):1056–76. Cardiol. 2017;69(21):2657–64.
40. Ikushima K, Arimura H, Jin Z, Yabu-Uchi H, Kuwazuru J, Shioyama 60. Slomka PJ, Dey D, Sitek A, Motwani M, Berman DS, Germano
Y, et al. Computer-assisted framework for machine-learning-based G.  Cardiac imaging: working towards fully-automated
delineation of GTV regions on datasets of planning CT and PET/ machine analysis & interpretation. Expert Rev Med Devices.
CT images. J Radiat Res. 2017;58(1):123–34. 2017;14(3):197–212.
41. Ibragimov B, Xing L. Segmentation of organs-at-risks in head and 61. Douglas PS, Cerqueira MD, Berman DS, Chinnaiyan K, Cohen
neck CT images using convolutional neural networks. Med Phys. MS, Lundbye JB, et al. The future of cardiac imaging: report of a
2017;44(2):547–57. think tank convened by the American College of Cardiology. JACC
42. Belharbi S, Chatelain C, Herault R, Adam S, Thureau S, Chastan Cardiovasc Imaging. 2016;9(10):1211–23.
M, et  al. Spotting L3 slice in CT scans using deep convolutional 62. Nagueh SF. Unleashing the potential of machine-based learning for
network and transfer learning. Comput Biol Med. 2017;87:95–103. the diagnosis of cardiac diseases. Am Heart Assoc; 2016; 9(6).
43. Summers RM. Progress in fully automated abdominal CT interpre- 63. Tesche C, De Cecco CN, Albrecht MH, Duguay TM, Bayer RR
tation. AJR Am J Roentgenol. 2016;207(1):67–79. 2nd, Litwin SE, et al. Coronary CT angiography-derived fractional
44. Roth HR, Lu L, Liu J, Yao J, Seff A, Cherry K, et al. Improving flow reserve. Radiology. 2017;285(1):17–33.
computer-aided detection using convolutional neural net- 64. Avendi MR, Kheradvar A, Jafarkhani H. A combined deep-learning
works and random view aggregation. IEEE Trans Med Imaging. and deformable-model approach to fully automatic segmentation of
2016;35(5):1170–81. the left ventricle in cardiac MRI. Med Image Anal. 2016;30:108–19.
45. Näppi JJ, Hironaka T, Regge D, Yoshida H, editors. Deep transfer 65. Germano G, Kavanagh PB, Slomka PJ, Van Kriekinge SD, Pollard
learning of virtual endoluminal views for the detection of polyps in G, Berman DS.  Quantitation in gated perfusion SPECT imaging:
CT colonography. San Diego: Proc SPIE; 2016. the Cedars-Sinai approach. J Nucl Cardiol. 2007;14(4):433–54.
46. Qi D, Hao C, Lequan Y, Lei Z, Jing Q, Defeng W, et  al.
66. Garcia EV, Faber TL, Cooke CD, Folks RD, Chen J, Santana C. The
Automatic detection of cerebral microbleeds from MR images increasing role of quantification in clinical nuclear cardiology: the
via 3D convolutional neural networks. IEEE Trans Med Imaging. Emory approach. J Nucl Cardiol. 2007;14(4):420–32.
2016;35(5):1182–95. 67. Arsanjani R, Xu Y, Hayes SW, Fish M, Lemley M, Gerlach J, et al.
47. Anthimopoulos M, Christodoulidis S, Ebner L, Christe A,
Comparison of fully automated computer analysis and visual scor-
Mougiakakou S. Lung pattern classification for interstitial lung dis- ing for detection of coronary artery disease from myocardial perfu-
eases using a deep convolutional neural network. IEEE Trans Med sion SPECT in a large population. J Nucl Med. 2013;54(2):221–8.
Imaging. 2016;35(5):1207–16. 68. Betancur J, Rubeaux M, Fuchs T, Otaki Y, Arnson Y, Slipczuk L,
48. Bergtholdt M, Wiemker R, Klinder T, editors. Pulmonary nodule et  al. Automatic valve plane localization in myocardial perfusion
detection using a cascaded SVM classifier. SPIE Medical Imaging; SPECT/CT by machine learning: anatomical and clinical valida-
2016. International Society for Optics and Photonics. tion. J Nucl Med. 2017;58(6):961–7.
49. Setio AA, Ciompi F, Litjens G, Gerke P, Jacobs C, van Riel SJ, et al. 69. Arnoldi E, Gebregziabher M, Schoepf UJ, Goldenberg R, Ramos-­
Pulmonary nodule detection in CT images: false positive reduc- Duran L, Zwerner PL, et  al. Automated computer-aided stenosis
tion using multi-view convolutional networks. IEEE Trans Med detection at coronary CT angiography: initial experience. Eur
Imaging. 2016;35(5):1160–9. Radiol. 2010;20(5):1160–7.
50. Ciompi F, Chung K, Van Riel SJ, Setio AAA, Gerke PK, Jacobs C, 70. Halpern EJ, Halpern DJ. Diagnosis of coronary stenosis with CT
et  al. Towards automatic pulmonary nodule management in lung angiography: comparison of automated computer diagnosis with
cancer screening with deep learning. Sci Rep. 2017;7:46479. expert readings. Acad Radiol. 2011;18(3):324–33.
51. Rucco M, Sousa-Rodrigues D, Merelli E, Johnson JH, Falsetti L, 71. Goldenberg R, Eilot D, Begelman G, Walach E, Ben-Ishai E, Peled
Nitti C, et al. Neural hypernetwork approach for pulmonary embo- N.  Computer-aided simple triage (CAST) for coronary CT angi-
lism diagnosis. BMC Res Notes. 2015;8:617. ography (CCTA). Int J Comput Assist Radiol Surg. 2012;7(6):
52. Rajkomar A, Lingam S, Taylor AG, Blum M, Mongan J.  High-­ 819–27.
throughput classification of radiographs using deep convolutional 72. Kang K-W, Chang H-J, Shim H, Kim Y-J, Choi B-W, Yang W-I,
neural networks. J Digit Imaging. 2017;30(1):95–101. et  al. Feasibility of an automatic computer-assisted algorithm for
53. Rampun A, Tiddeman B, Zwiggelaar R, Malcolm P.  Computer the detection of significant coronary artery disease in patients pre-
aided diagnosis of prostate cancer: a texton based approach. Med senting with acute chest pain. Eur J Radiol. 2012;81(4):e640–e6.
Phys. 2016;43(10):5412. 73. Kang D, Slomka P, Nakazato R, Cheng VY, Min JK, Li D, et al.,
54. Dhungel N, Carneiro G, Bradley AP. A deep learning approach for editors. Automatic detection of significant and subtle arterial
the analysis of masses in mammograms with minimal user inter- lesions from coronary CT angiography. SPIE Medical Imaging;
vention. Med Image Anal. 2017;37:114–28. 2012. International Society for Optics and Photonics.
68  Machine Learning and Artificial Intelligence in Cardiovascular Imaging 907

74. Kang D, Dey D, Slomka PJ, Arsanjani R, Nakazato R, Ko H, 82. Hadamitzky M, Freißmuth B, Meyer T, Hein F, Kastrati A,

et  al. Structured learning algorithm for detection of nonobstruc- Martinoff S, et  al. Prognostic value of coronary computed tomo-
tive and obstructive coronary plaque lesions from computed graphic angiography for prediction of cardiac events in patients
tomography angiography. J Med Imaging (Bellingham, Wash). with suspected coronary artery disease. JACC Cardiovasc Imaging.
2015;2(1):014003. 2009;2(4):404–11.
75. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M 83. Hadamitzky M, Taubert S, Deseive S, Byrne RA, Martinoff S,
Jr, Detrano R. Quantification of coronary artery calcium using ultra- Schomig A, et al. Prognostic value of coronary computed tomog-
fast computed tomography. J Am Coll Cardiol. 1990;15(4):827–32. raphy angiography during 5 years of follow-up in patients with
76. Kondos GT, Hoff JA, Sevrukov A, Daviglus ML, Garside DB, suspected coronary artery disease. Eur Heart J. 2013;34(42):
Devries SS, et  al. Electron-beam tomography coronary artery 3277–85.
calcium and cardiac events: a 37-month follow-up of 5635 ini- 84. Dawes TJ, de Marvao A, Shi W, Fletcher T, Watson GM, Wharton
tially asymptomatic low- to intermediate-risk adults. Circulation. J, et  al. Machine learning of three-dimensional right ventricular
2003;107(20):2571–6. motion enables outcome prediction in pulmonary hypertension: a
77. Wolterink JM, Leiner T, de Vos BD, van Hamersvelt RW, Viergever cardiac MR imaging study. Radiology. 2017;283(2):381–90.
MA, Išgum I. Automatic coronary artery calcium scoring in cardiac 85. Kotu LP, Engan K, Borhani R, Katsaggelos AK, Orn S, Woie L,
CT angiography using paired convolutional neural networks. Med et  al. Cardiac magnetic resonance image-based classification of
Image Anal. 2016;34:123–36. the risk of arrhythmias in post-myocardial infarction patients. Artif
78. Sengupta PP, Huang YM, Bansal M, Ashrafi A, Fisher M, Shameer Intell Med. 2015;64(3):205–15.
K, et al. Cognitive machine-learning algorithm for cardiac imaging: 86. Arsanjani R, Dey D, Khachatryan T, Shalev A, Hayes SW, Fish
a pilot study for differentiating constrictive pericarditis from restric- M, et al. Prediction of revascularization after myocardial perfusion
tive cardiomyopathy. Circ Cardiovasc Imaging. 2016;9(6):e004330. SPECT by machine learning in a large population. J Nucl Cardiol.
79. Mahmood SS, Levy D, Vasan RS, Wang TJ. The Framingham Heart 2015;22(5):877–84.
Study and the epidemiology of cardiovascular disease: a historical 87. Jha S, Topol EJ.  Adapting to artificial intelligence: radi-

perspective. Lancet. 2014;383(9921):999–1008. ologists and pathologists as information specialists. JAMA.
80. Budoff MJ, Dowe D, Jollis JG, Gitter M, Sutherland J, Halamert 2016;316(22):2353–4.
E, et  al. Diagnostic performance of 64-multidetector row coro- 88. Chinthapalli K.  The hackers holding hospitals to ransom. BMJ.
nary computed tomographic angiography for evaluation of coro- 2017;357:j 2214.
nary artery stenosis in individuals without known coronary artery 89. Moss TJ, Calland JF, Enfield KB, Gomez-Manjarres DC, Ruminski
disease: results from the prospective multicenter ACCURACY C, DiMarco JP, et al. New-onset atrial fibrillation in the critically ill.
(Assessment by Coronary Computed Tomographic Angiography Crit Care Med. 2017;45(5):790.
of Individuals Undergoing Invasive Coronary Angiography) trial. 90.
Chockley K, Emanuel E.  The end of radiology? three
J Am Coll Cardiol. 2008;52(21):1724–32. threats to the future practice of radiology. J Am Coll Radiol.
81. Chow BJ, Small G, Yam Y, Chen L, Achenbach S, Al-Mallah M, 2016;13(12):1415–20.
et al. Incremental prognostic value of cardiac computed tomogra- 91. Obermeyer Z, Emanuel EJ.  Predicting the future  – big data,

phy in coronary artery disease using CONFIRMClinical perspec- machine learning, and clinical medicine. N Engl J Med.
tive. Circ Cardiovasc Imaging. 2011;4(5):463–72. 2016;375(13):1216–9.
 Index

A TIMI risk score, 331

Absolute myocardial blood flow values, 824 SPECT MPI, 341
Accessory LAA, 537 Acute myocardial infarction (AMI), 297, 791, 829
ACGME core competencies, 45 Acute pulmonary embolism, 625, 626
interpersonal and communication skills, 47 Adaptive intimal thickening (AIT), 212, 213
medical knowledge, 46 Adaptive iterative dose reduction (AIDR 3D), 82
patient care and procedural skills, 46 Adenosine, 767, 800
practice-based learning and improvement, 46 Adiponectin, 287
professionalism, 47 ADVANCE Registry, 888
systems-based practice, 46 Advanced visualization, 131
ACR Certificate of Advanced Proficiency (CoAP) examination, 47 Agatston score, 7, 136, 175, 631, 717, 742, 743, 758, 879, 903
ACR CT accreditation program Air pollution, 693–694
scanner requirement, 41, 42 ALCAPA syndrome, 572
technologist requirements, 42 Algebraic reconstruction theory (ART), 4, 5
ACR CT QC Manual, 42 Alkaptonuria, 351
Active bronchospastic disease, 104 Allometric scaling laws, 769
Acute aortic syndromes (AAS), 198 American College of Cardiology (ACC), 45, 355
Acute chest pain, 204, 854 American College of Cardiology/ American Heart
CEA, 674, 675 Association (ACC/AHA) guidelines, 519
prognostic significance, 665, 666 American College of Cardiology Foundation (ACCF), 245
Acute coronary syndrome (ACS), 415 American College of Radiology (ACR), 245
CAC, 334 American Heart Association (AHA), 211, 355, 711
CCTA (see Cardiovascular CT angiography) Amplatzer Amulet LAAO, 544
CMR, 341 Amplatzer septal occluder, 495
conventional catheter angiography, 341 Aneurysmal tricuspid valve tissue, 582
coronary artery spasm, 297 Angina
coronary atherosclerotic plaques ATP pump dysfunction, 320
calcified nodules, 336 clinical manifestation, 319
characterization, 335 occurrence of, 319
dual energy CT, 336 symptoms, 320
high risk plaque features, 336 Angiotensin-converting enzyme inhibitors, 743
multi-energy imaging techniques, 336 Annuloplasty, 492
NAFLD, 337 Anomalous pulmonary venous connections, 591–594
obstructive non-culprit plaques, 337 Anti-scatter grids (ASG), 59
plaque burden index, 337 Aorta, CTA, 653, 654
plaque rupture, 335, 336 Aortic annulus, 507
ROMICAT I trial, 337 Aortic arch imaging, CTA, 653, 654
ROMICAT II trial, 336, 337 Aortic atheroma, 643
thin cap fibroatheroma, 335 Aortic regurgitation (AR), 475
type III, 335 Aortic rupture, 702
type IV- V lesions, 335 Aortic stenosis (AS), 472, 473, 584–586
coronary stenosis, 334, 335 Aortic valve (AV), 232
epicardial fat measurement, 289 Aortic valve calcification (AVC), 472
hs-troponin assays, 339, 340 Aortic valve replacement (AVR), 472
MI, 341 Aortic valvotomy, 607
risk assessment Appropriateness criteria (ACR), 235, 236
cardiac biomarkers, 331 Area under the curve (AUC), 262, 752, 757, 761–764, 787
clinical presentations, 331 Area-length method, 411
diagnostic ECG findings, 331 Arrhythmia, 297
HEART score, 331 Arrhythmogenic right ventricular dysplasia (ARVD), 407, 418
low-risk populations, 332 Arterial first pass imaging, 779
negative serial biomarkers, 331 Arterial input function (AIF), 815, 817, 818

© Humana Press 2019 909

U. J. Schoepf (ed.), CT of the Heart, Contemporary Medical Imaging, https://doi.org/10.1007/978-1-60327-237-7
910 Index

Arterial switch procedure, 571, 604, 605, 607, 610, 611 Beyond Endorsed Lipid Lowering with EBT
Artificial intelligence (AI), 893, 894 Scanning (BELLES) trial, 277
See also Machine learning (ML) Biatrial approach, 391
Artificial neurons, 894 Bicaval approach, 391
Ascending aorta pseudoaneurysm, 607 Bicuspid aortic valve (BAV), 510, 584
Ascending aortic aneurysm, 654 Big data, 893
ASIR-V, 76 Bileaflet mechanical valve, 487, 488
As Low As Reasonably Achievable (ALARA) principles, 69 BioImage Study, 29
Asplenia syndrome, 599 Biomarkers, 701, 702
Asymptomatic patients, 739 Bioprosthetic aortic valve, 489
Atherosclerosis, heart-lung axis, 627 Bioprosthetic valves, 488
cardiovascular disease and COPD, 628–630 Biplane area- length method, 411
cardiovascular disease and lung cancer screening, 631 Biventricular ICD/pacemaker, 492
cardiovascular disease and respiratory diseases, 630 Biventricular injection protocol, 608
Atherosclerotic cardiovascular disease (ASCVD) Bland-Altman analysis, 836
at-risk individuals, 259 Bland-White-Garland syndrome, 571
high-risk approach, 259 Blooming artifact, 309
population approach, 259 Body mass index (BMI), 112, 352, 813
risk assessment algorithm Bolus geometry, 113
ACC/AHA CV Risk Calculator, 260 Boolean operations, 861
age and prediction algorithms, 260 Bouguer, Pierre, 3
AHA/ACC CV Risk Calculator, 260 Braking radiation, 793
CACS (see Coronary Artery Calcium score) Breath-hold training, 156–159
FRS, 259 Bremsstrahlung radiation, 793
lifetime risk models, 260, 261 Bronchogenic cyst, 457
low-/intermediate-risk, 260 Bruce Protocol treadmill exercise, 19
population-based prediction, 260 BT shunt, pulmonary artery aneurysm, 615
preventive imaging techniques, 261
Reynolds Risk Score, 260
risk markers, 261 C
SCORE, 259 Ca-blooming, reduction of, 884
Atherosclerotic plaque characteristics (APCs), 854 CAC, 326
Atretic mitral valve, 596 Cadmium-telluride (CdTe), 880, 882, 883
Atrial appendages, 580 Cadmium-zinc-telluride (CZT), 880, 882
Atrial fibrillation, 637, 639 CAD-RADS, 337, 338
Atrial septal aneurysm (ASA), 640 Caged ball mechanical heart valve, 487, 488
Atrial septal defect (ASD), 527, 580, 581, 629 Calcified eccentric plaques, 802
Atrial switch operation, 608 Calcified nodule, 217, 218
AtriClip, 542 Calcium, 623
AtriCure Clip, 496 Calcium channel blockers (CCBs), 104
Atrio-esophageal fistula (AEF), 538, 539 Calibration phantoms, 117
Atrioventricular septal defects (AVSD), 583, 584 Cardiac allograft vasculopathy (CAV)
Attenuating “unknown object”, 3 acute rejection/infection, 400
Atypical chest pain, 784 atherosclerotic plaques, 394, 395
Aubert procedure, 608 collateral vasculature, 394
Australasian Association of Nuclear Medicine conventional coronary angiography, 395, 398, 399
Specialists (AANMS), 44 coronary artery calcification, 398
Auto gating, 75 diffuse intimal thickening, 394, 395
Automated segmentation, 133, 134 early detection, 394
Automated tube voltage selection (ATVS) technology, 127 vs. intravascular ultrasound, 399, 400
Automatic exposure control (AEC), 155 IVUS, 395
Average intensity projection (AIP), 135 non-invasive stress tests, 395, 396
Azygous system off-loading, 610 pathophysiology, 394
prevalence of, 394
recommendations, 394, 395
B risk factors, 394
Back propagation phase, 894–896 Cardiac anatomy
Back-projection, 4 cardiac chambers, 231, 232
Baffle obstruction, 610 cardiac planes, 234
Barlow´s disease, 476 cardiac valves, 231–233
Beam-hardening artifacts, 90, 96, 782, 794 coronary arteries
Beer, August, 3 CCTA, 229, 230
Behçet's disease, 436, 455 LAD, 227
Beta-blockers, 557, 780, 813, 873, 876 LCx, 227
Better Evaluation of Acute Chest Pain with Coronary Computed left main, 227, 228
Tomography Angiography (BEACON) trial, 340 RCA, 227, 229
Index 911

SCCT, 230 left ventricular diastolic dysfunction, 428

coronary veins, 231 left ventricular diastolic filling, 428
pulmonary veins, 232–234 passive emptying and conduit flow phase, 428
visceral and parietal layers, 227 passive emptying function, 430
Cardiac and paracardiac masses reservoir function, 429
approach, 452, 453 reservoir phase of, 428
benign neoplasms reservoir volume, 428
fibroelastoma, 459, 461 trans-mitral Doppler ultrasonography, 428
fibroma, 463 left ventricle function
hemangioma, 461, 462 afterload, markers for, 428
less common benign tumors, 464 aortic AD, 428
lipoma, 461 AWSI, 428
myxoma, 458, 460 beta-stiffness index, 428
paraganglioma, 462, 463 bolus-timing scan, 427
rhabdomyoma, 463 ejection fraction, 427
Chiari network, 453 endocardial strain, 427
coumadin ridge, 454 geometric measurement, 427
CT, 452 invasive coronary angiography, 427
imaging of, 451, 452 isovolumic contraction, 426
left lateral ridge, 453 isovolumic relaxation, 426
malignant neoplasms myocardial movement, 427
lymphoma, 467 preload of, 426
mesothelioma, 467 SPECT myocardial perfusion, 427
mesotholioma, 464 time-attenuation curve, 427
metastasis, 464, 465 Young’s modulus of elasticity, 428
sarcoma, 464, 466 left ventricular time-volume curve, 426
neoplastic lesions, 451 maximum pressure gradient rate, 424
non-neoplastic masses MDCT, 425, 426
bronchogenic cyst, 457 preload of cardiac ventricles, 424
caseous MAC, 456 right atrial function, 432, 433
chronic thrombus, 455 right ventricular function
CMAC, 456 acute massive pulmonary thromboembolism, 431
contrast admixture, 455 Cor pulmonale, 431
coronary aneurysm, 459 eccentricity index, 431
coronary artery aneurysm, 457 ejection fraction, 431
hiatal hernia, 457, 458 Hagen-Poiseuille equation, 432
interatrial septal aneurysm, 458 morphological alteration, 432
LHIS, 456 pulmonary arterial dilatation, 432
lipomatous hypertrophy of interatrial septum, 456 right ventricular deformation, 432
MAC, 456 systemic venous congestion, 431
pericardial cyst, 456, 457 TAPSE, 431
pseudoaneurysm, 457, 459 time-attenuation curves, 431
thrombus, 454, 455 time-volume curve, 428, 429
vegetation, 458, 459 ventricular afterload, 424
persistent eustachian valve, 453 ventricular inotropic activity, 424
prominent crista terminalis, 453 Cardiac computed tomography (CT), 38, 740
Cardiac catheterization, 19 area detector CT, 57–59, 61
Cardiac chamber, 231, 232 Ca-blooming, reduction of, 884
atypical chest pain, 434 CT-FFR, 887–888
automated segmentaion, 425 DSCT system, 51
cardiac CT functional parameters, 433 advantages and disadvantages, 62
cardiac output, 423 clinical benefits, 62
fractional blood flow volumes, 429 cross-scattered radiation, 63
Frank-Starling law, 424 DE algorithms, 64
functional parameters of, 438 high-pitch scan mode, 62
global cardiac function, 433, 435 independent measurement systems, 61
in-vivo ventricular contractility, 424 model-based/measurement based correction approaches, 63
left atrial function motion artifacts, 64
active pumping function, 429 temporal resolution, 62
afterload of, 430 z-flying focal spot, 61
conduit flow, 428 EBCT, 52, 54
Doppler ultrasonography, 430 ECG-gated scanning techniques, 51
global function, 430 extra-cardiac findings in lungs, 631
LAA ejection fraction, 430 first pass enhancement scanning, 884–887
LAEF, 429, 430 functional recovery prediction, 839
LAVImax, 430 high-pitch scan modes, 52
912 Index

Cardiac computed tomography (CT) (cont.) Cardiac thrombus, 636, 637

irregular/arrhythmic heart rates, 51 Cardiac tumors, 638, 640
MDCT system Cardiac valves and prostheses, 231–233
ADC electronics and data transmission, 54 aortic valve
ECG-signal, 56, 57 aortic regurgitation, 475
fan-beam CT design, 54 aortic stenosis, 472
key requirements, 54 bicuspid valve, 473
quantum and electronic choice, 54 quadricuspid aortic valve, 474
16-slice CT detector, 55, 56 raphe type, 473
slip-ring designs, 54 bicuspid valve types, 474
solid-state scintillation detectors, 54 biphasic split protocol, 472
state-of-the art X-ray tubes, 54 degenerative MAC, 479
whole body scan field of view, 54 infective endocarditis, 480
z-flying focal spot, 56, 57 mitral regurgitation/prolapse, 478
medium-sharp reconstructions, 51 mitral stenosis/regurgitation, 477
multi-segment reconstructions, 51 mitral valve
myocardial viability assessment, 829, 841 aortic regurgitation, 476
phase-contrast CT, 880 infective endocarditis, 477, 478
photon-counting detectors, 880–883 leaflet perforations, 478
prognostic risk assessment, 889 MAC, 476
radiation dose, 878–880 MVP, 476
RVOT obstruction clarified, 613 paravalvular abscesses and pseudoaneurysm, 478
64-slice CT systems, 51 pulmonary and tricuspidal valve, 476
spatial resolution, 876–878 MPR, 472
stack artifacts, 51 PHV
temporal resolution, 873–876 CCTA artifacts, limitations of, 483
CNR, 874 CCTA technology, 479
DSCT, 874 degenerative lesions, 481
EBCT, 875 dehiscence, 481
exposure time per image, 873 paravalvular abscess, 481
gantry rotation times, 873 paravalvular leakage, 481
motion correction approach, 875, 876 vs. periprosthetic fibrosis, 483
multi-segment reconstruction, 875 pseudoaneurysm, 481
multisource interior tomography, 874 renal dysfunction, 483
SFOV, 873 ruputured chordae, 481
TRI-PICC/TRIM, 875 stuck valve, 483
virtual, 875 SVD, 482
Cardiac devices retrospective ECG-gating, 471
annuloplasty, 492 valvular imaging, 471
AtriCure Clip, 496 CardiAQTM, 520
bioprosthetic valves, 488 Cardioembolic stroke, 635
HeartMate II, 498 Cardiologists and radiologists collaboration, 38, 39
HeartWare, 497 CardioSEAL septal occluder, 494
ILR, 494 Cardio-selective β-adrenergic antagonist, 103
Impella, 497 Cardiovascular CT, repaired CHD
implantable pacemakers, 492 aortic arch imaging, 606, 607
LARIAT snare device, 496 d-TGA s/p atrial switch, 608, 611
leadless pacemakers, 493 indications, 618
mechanical valves intervention, coronary artery imaging, 604–606
bileaflet mechanical valve, 487 risk of, 603
caged ball mechanical heart valve, 487 single ventricle heart disease, 612–616, 618
single tilting disc valve, 487 situations in, 604
PDA clip, 499 Tetralogy of Fallot, 611, 612
subcutaneous ICD, 493 transposition complexes, 607
Swan-Ganz catheter, 498 use of, 604
TandemHeart, 497 Cardiovascular CT angiography (CCTA), 173
TAVR, 489 calcified plaque, 333, 334
WATCHMAN device, 495 coronary artery anomalies, 339, 340
Cardiac implanted electronic device (CIED), 184 CT myocardial perfusion, 337
Cardiac magnetic resonance (CMR) imaging, 341, 396, 791 data sampling, 333
Cardiac output (CO), 112, 414, 423 emergency department (ED)
Cardiac phase for acquisition, 812 diagnosis and hospital costs, 332
Cardiac planes, 234 exercise treadmill ECG-based care, 333
Cardiac resynchronization therapy, 493 intermediate lesions, 333
Cardiac Society of Australia and New Zealand (CSANZ), 44 intermediate-risk population, 332
Cardiac syndrome X, 322, 323 invasive coronary angiography, 332
Index 913

low-risk patients, 332 Cardiovascular risk stratification methods, 717

myocardial stress perfusion imaging, 332 Cardiovascular system, 4
prospective cohort trial, 332 CardoSEAL septal occluder, 494
randomized controlled study, 332 Carotid intima-media thickness (cIMT), 262
ROMICAT II trial, 332 Ca-scoring, 879
routine provocative testing, 332 Caseous mitral annular calcification (CMAC), 456
safety and efficiency, 333 Center for Research on Utilization of Imaging Services (CRUISE), 681
TIMI score, 332 Centers for Medicare and Medicaid Services (CMS), 681
trial data and observation, 332 Central pulmonary emboli, 648
fast-acting β-blocker, 333 Cerebrovascular accident (CVA), 355
FFR, 337 Cerebrovascular disease, 694
IV metoprolol, 333 Certification Board of Cardiovascular Computed Tomography
heart and coronary arteries, 69 (CBCCT), 45
heart rate control, 105 Challenging patient
myocardial bridging, 339 angina onset, 128
nitroglycerin, 333 anxious patient, 127, 128
Philips medical systems beta-blockers and/or sublingual nitroglycerine, 126
bariatric mode, 80 CCTA examinations, 125
Brilliance iCT (256), 79, 80 elevated heart rates and arrhythmias, 126, 128
64-slice version (Brilliance 64), 78, 79 heart ischemia, 128
prospective ECG-triggering, 333 obese patients, 127, 128
radiation doses, 333 pharmacological stress agents, 127–129
reporting, 337, 338 reduced renal function, 126
retrospective ECG-triggering, 333 Characteristic radiations, 793
Siemens Chest Pain (CATCH) study, 27
acquisition protocols selection, 74 Chiari network, 453
calcium Score, 75 China
coronary CT angiography, 75, 76 air pollution, 693–694
DSCT systems, 73 cardiac CT imaging
dual energy GSI, 77 aortic rupture, 702
GE Healthcare, 75 appropriate-and-frequent indications, 700
revolution CT, 75 average duration, 699
revolution EVO, 78 beta-blockers, 698
revolution HD (with/without GSI), 76 biomarkers, 701, 702
scan acquisition, 74 clinical indications, 698
single energy, 77 coronary artery disease, 701
SSCT systems, 73, 74 CT performance, 700
X-ray source/detector systems, 73 data acquisition, 697
TRO CT, 334 domestic and foreign shares, 699
Cardiovascular disease, 624, 628–630, 791 epicardial and pericoronany adipose tissue, 701
and lung cancer screening, 631 high-end cardiac CT scanners, 700
and respiratory diseases, 630, 631 image quality, 701
Cardiovascular Disease Prevention, see Atherosclerotic published papers, 700–702
cardiovascular disease pulmonary disease, 701
Cardiovascular health and care radiation dose, 699–701
African descent patients, 708–710 SCI-EXPANDED, 701
annual age-adjusted death rates, 707 3-tier system, 698
Asian descent patients, 710 25-item questionnaire, 698
gender differences, 707, 708 valvular disease, 701
healthcare quality, 707 cerebrovascular disease, 694
Hispanics, 710 coronary artery disease, 694, 695
mortality rates, 707 diabetes, 691, 692
patient selection, 707 heart failure, 694, 696
race and ethnicity, 707, 708 hyperlipidemia, 691
in women, 711 hypertension, 689, 690
ACCF/AHA/ACP/AATS/PCNA/SCAI/STS guidelines, 712 medical care expenditure, 696
adverse outcomes, 711 mortality, 689, 690
American Heart Association’s women-specific guidelines, 711 number of CTs studies, 697
cCTA, 712 obesity, 692
mortality, 710 overweight, 692
noninvasive testing, 712 physical activity, 692
plaque erosion, 711 PM2.5 concentrations and CVD mortality, 693
premenopausal women, 711 smoking, 690–691
risk categories, 711 study sites, 697
sex-specific risk factors, 711 unreasonable diet, 692, 693
Cardiovascular magnetic resonance (CMR), 407 China Health Statistics Yearbook 2015, 694
914 Index

China Heart Failure Registry Study (China-HF), 694 Computational fluid dynamics (CFD) technique, 767,
China’s 2010 Chronic Disease and Risk Factor Surveillance, 694 769, 770, 774, 775, 887, 888
Chinese Certificate of Needs (CON), 697 Computed Tomographic Angiography for the Systematic
Chinese Coronary CTA guideline, 699 Triage of Acute Chest Pain Patients to Treatment
Chinese International Regional Committee of the Society of (CT-STAT) trial, 332
Cardiovascular Computed Tomography, 699 Computed TomogRaphic Evaluation of Atherosclerotic
Chinese Nutrition and Health Survey, 692 DEtermiNants of Myocardial IsChEmia
Chinese Society of Radiology, 699 (CREDENCE) trial, 854
Chronic Active Epstein-Barr virus (CAEBV), 304, 305 Computed tomography angiography (CTA), 647
Chronic chest pain incidental findings and
adenosine stress-rest, 324 aorta, 653, 654
anomalous origin, 323, 324 breast, 654
atherosclerotic plaques, histologic characteristics, 320, 321 lung, 652
autopsy studies report, 319 mediastinum, 652, 653
CAC, 326 pulmonary artery, 654
CKD symptoms, 319 thoracic skeleton and pleura, 654
dual source CT ventricular function, 324 upper abdominal organs, 654
examination, 324, 326 practical tips, 651, 652
exercise ECG patients, 324 small vs. wide field, view on, 647, 649–651
FFR value, 326 Computed tomography angiography derived fractional
focal/diffuse spasm, 321, 322 flow reserve (CT-FFR), 140, 383, 386,
ischemic cardiomyopathy, 323 761, 791, 851, 854, 855, 887–889
lesion characterization, 326 applications
microvascular angina and vasospastic angina, 319 HeartFlow, 770, 772
microvascular dysfunction, 322, 323 machine-learning approach, 770
non-diagnostic scans, 326 on-site performed, 773
pathophysiology, 319, 320 workstation-based, 772, 773
pre-test probability, 324, 325 challenges and limitations, 774
prognostic value, coronary CTA, 326 intra-cardiac blood flow simulation, 774, 775
sex differences, 327 vs. invasive FFR, 771
SPECT and echocardiography, 324, 325 predicting events, 774
therapy for Chronic chest pain, 327 principles, 768
Chronic Disease Monitor Program in China, 692 anatomical structure, 769
Chronic kidney disease (CKD), 126, 319 CFD, 767
Chronic obstructive pulmonary disease (COPD), 625, 628–630 coronary blood flow and pressure, computation of, 769
Chronic thromboembolic pulmonary hypertension, 626, 628 global and local resting flow, estimation of, 769
Chronic total occlusion (CTO), 764 hyperemia, simulation of, 769
CCTA, 366, 367 simulate flow, 768
CTA appearance, 313, 314 revascularization planning, 774
CT-RECTOR score, 368, 369 Computed tomography myocardial perfusion
definition, 313 imaging (CTMPI), 811
guidewire-failure group, 368, 370 analysis, 782–783
HRP features assessment of prior infarcts, 784
adverse outcomes, 314 calcified coronary trees, 788
atherothrombosis, 315 CORE 320 study, 787
CT appearance, 314, 315 coronary anatomy, correlation with, 786
in-stent restenosis, 315, 316 datasets, 802
low-attenuation plaque, 314 defect severity and ischemic burden, 784
myocardium, 315, 316 dynamic CT perfusion, 778
napkin-ring sign, 314 iodinated contrast materials, 777
perivascular fat stranding, 315, 316 iodinated contrast media (CM), 792
positive remodeling, 314 ischemia, assessment, 784
spotty calcium, 314 meta-analyses, 787
intervention prediction, 366–368 physiological background, 777
revascularization, 313 protocol modifications, clinical studies, 788
Classifiers, 894 reconstruction and postprocessing, 782
Clinical Outcomes Utilizing Revascularization and Aggressive Drug regadenoson cross-over study, 787–788
Evaluation (COURAGE) trial, 728 rest image interpretation, 784
Close Curve algorithm, 869 scan protocols, 778
Cloud servers, 197 scanner generations, 778
Coarctation of the aorta, 586 single-center studies, 786–787
Collaborative approach, 38 static CT perfusion (see Static computed tomography perfusion)
Collateral circulation, 586 stent patency assessment, 788
Common arterial trunk, 590 stress Image Interpretation, 784
Comparative effectiveness research, 26–28, 673 tube voltage attenuation, 778
Complete cardiac evaluation protocol, 815 typical artifacts
Index 915

motion, 783 Conglomerate lymph node enlargement, 653

partial dataset misalignment artifacts, 784 Conjoint committee, 44, 45
reconstruction artifact, 783 Constrictive pericarditis, 445, 446
Computed tomography perfusion (CTP), 763, 773 Contamination, 743
Computed tomography vs. exercise testing in suspected coronary Contractile cardiac function, 774
artery disease (CRESCENT) trial, 273 Contrast density difference (CDD), 752
Computer-aided design (CAD), 197, 859, 861, 862, 869 Contrast enhanced cardiac computed tomography
Computerized tomography fractional flow reserve (FFR-CT), 363 dual-energy CT, delayed imaging, 836
Concordant ventriculoarterial connection, 580 extracellular volume fraction, 835
Cone-beam artifacts, 783 first pass arterial phase scan, 833, 834
CONFIRM registry, 662 late contrast enhancement, 834, 835
age and gender, prognostic significance, 662–664 post-percutaneous coronary intervention (PCI)
CAD severity and left ventricular ejection fraction, delayed imaging, 835, 836
prognostic significance, 665 wall thinning, 832
coronary artery calcium scoring, prognostic Contrast induced nephropathy (CIN), 127
significance, 664, 665 Contrast injection, 606
diabetic patients, prognostic significance, 665 Contrast material application, 109
ethnic differences, prognostic significance, 663, 664 Contrast material contamination of myocardium, 780
Congenital heart disease (CHD), 187, 205, 407, 579, 701 Contrast media (CM), 813, 849
anomalous pulmonary venous connections, 591–594 Contrast media injection protocols, 113, 114
aortic stenosis, 584–586 concentrations, 109
atrial septal defects, 580, 581 contrast material application, 109
AVSD, 583 flow rates, 109
coarctation of the aorta, 586 iodine delivery rate (IDR), 110
cor triatriatum, 595 iodine volume, 109
DILV, 597 patient care
double outlet right ventricle, 588–590 body mass index, 112
D-TGA, 587, 588 body weight, 112
Ebstein’s anomaly, 586, 587 cardiac output, 112
heterotaxy syndromes, 597, 599 extravasation, 111
HLHS, 595 site, 111
IAA, 586, 587 without the presence of negative side effects, 111
imaging algorithms, 579 pre-heating, 110
L-TGA, 588 saline chaser, 110
medical and surgical care, 603 scanner
patent ductus arteriosus, 582, 583 bolus geometry, 113
pulmonary atresia with intact ventricular septum, 595 lower tube voltage, 113
pulmonary stenosis, 583 reconstruction technologies, 113
sequential segmental analysis, 579, 580 total iodine load, 111
systemic venous abnormalities, 597 Contrast Walls, 868, 869
TOF, 583, 584 Contrast-enhanced computed tomography, 882
tricuspid atresia, 597 Contrast-induced acute kidney injury (CI-AKI), 126
truncus arteriosus, 590, 591 Contrast-to-noise ratio (CNR), 874, 875, 878–881
vascular rings, 590, 591 Conventional solid-state scintillation detector, 879
VSD, 581, 582 Convolution kernel, 119, 120
Congenital heart disease imaging Convolution technique, 819
acquisition techniques Cor triatriatum, 595
injection protocol, 557, 558 CORE 320 study, 787
scan triggering, 558 CORE320 multi-center trial, 884
sequence selection, 558 Coronary angiography (CAG), 301
interpretation and reporting, 561 Coronary artery aneurysm, 457
lesion-specific recommendations Coronary artery anomalies
cyanotic heart lesions, 559 anomalies of course, 567
pulmonary venous anomalies, 559 anomalies of termination, 567
single ventricle physiology, 560 assessment of, 565
systemic venous abnormalities, 559 axial contrast-enhanced CT, 570
thoracic arterial abnormalities, 559 classification of, 567
patient preparation congenital coronary ostial atresia, 569
beta blockade, 557 contralateral coronary artery, 569, 571
imaging environment, 555 contralateral sinus of Valsalva, 569
intravenous access, 556, 557 coronary anomalies, 567
sedation requirements, 556 coronary arcade, 574
Congenitally corrected transposition (L-TGA), 588, 589 coronary arteries, 568
Congenitally corrected transposition of the great coronary artery duplication, 574
arteries (CCTGA), 588 coronary artery fistula, 574
Congestive heart failure, 423 coronary dominance, 566
916 Index

Coronary artery anomalies (cont.) TECAB, 384

deep myocardial bridge, 573 temporal resolution, 382
extracardiac treatment, 576 thoracic vasculature, 383
high takeoff, 567 3D reconstruction, 383, 385
left coronary artery fistulae, 575 radiographic assessment, 381
life-threatening sequelae, 567 reoperative surgery, 381, 386–388
MDCT coronary angiography, 565, 566, 572 temporal resolution, 381, 382
multiple ostia, 567, 568 3rd generation DSCT, 382
myocardial bridging, 572 2/3-vessel disease, 327
non-coronary sinus of Valsalva, 569 Coronary artery calcification (CAC), 7, 8, 334
normal coronary artery branches and courses, 565 Coronary artery calcium score (CACS), 23–24, 79,
posterior sinus of Valsalva, 567 136, 203, 349, 472, 739–742, 898, 902–904
pulmonary artery, 571 ACC/AHA Expert Consensus document, 280
pulmonary trunk, 571, 572 ACC/AHA Pooled Cohort Equations, 262
right coronary artery fistulae, 575 ACCF/AHA Guideline for Assessment of Cardiovascular
single coronary artery, 567 Risk in Asymptomatic Adults, 280
sinus of Valsalva, 570 adherence to preventive therapies
SNA, 566, 567 aspirin and statin, 274
superficial myocardial bridging, 573 downstream diagnostic testing, 274
Coronary artery bypass grafting (CABG), 19, 457, 572, 777, 781, 826 EISNER study, 273, 274
acquisition time, 382 lifestyle modification and risk perception, 274
aortic cross-clamping, 381 medication utilization, 274
coronary calcium scanning, 381 motivational effects, 274
dual source acquisition, 382 RCTs, 274, 275
ECG-dyssynchrony, 382 ROBINSCA trial, 275
LIMA, 381 Agatston score, 280
minimally invasive techniques, 381 AHA Scientific Statement, 279
motion artifact, 382 AHA Writing Group, 279
myocardial revascularization, 381 American College of Radiology Appropriateness Criteria, 280
off-pump surgery, 381 asymptomatic patients
patency and long-term survival, 381 annualized all-cause mortality rate, 267
patency vs. occlusion, 382 BioImage Study, 268
postoperative planning CHD events, 267
arterial grafts, 385, 386 cumulative relative risk rate, 267
assessment, 385 implications, 267
CT-FFR, 386 JUPITER trial, 267, 268
ECG-gating, 385 lower-risk group, 267
meta-analysis, 385 negative risk markers, 268, 269
occlusion detection, 385 risk classification, 268, 269
procedural intervention, 386 CHD and stroke, 262
proximal and mid-vessel aspects, 384 cIMT, 262
saphenous vein grafts, 385 diabetes
spatial resolution, 386 adjusted hazard ratios, 265
venous grafts, 385 Diabetes Heart Study, 270
preoperative planning mortality, 265, 266
accuracy, 382 NCEP ATP III guidelines, 270, 272
aortocoronary anastomoses, 384 prevalence of, 265
calcium, 383 prognosis, 265
cardiac and mediastinal anatomy, 383 risk stratification, 272
cardiac function, 383 stable and acute chest pain, 272, 273
CT-FFR, 383 systematic review and meta-analysis, 265, 266
EBCT, 384 discrimination, 261, 262
initial diagnostic test, 382 ethnicity, gender, and age, 261
intramyocardial vessel course, 384 European Society of Cardiology Guidelines, 280
LIMA, 383 findings, 279
mammary arteries, 384 intermediate-risk individuals, 262
MIDCAB, 384 intra- and inter-scan reproducibility, 279
minimally invasive, 384 low- and high-risk patients, 268, 270–272
myocardial viability, 383 low-intermediate risk, 263
off-pump approach, 384 MESA risk score, 263, 264
on-pump surgery, 384 NCEP ATP III released guidelines, 279
percutaneous intervention, 384 noninvasive testing guidelines, 279
risk of stroke, 384 optimal risk factor status, 263
routine catheter, 383 Pooled Cohort Equations, 263
spatial resolution, 382 prognosis, 262
tactile feedback, 384 progression
Index 917

age- and gender-related percentiles, 276 multi-segment plaque involvement, 312, 313
cumulative proportion, 275 obstructive plaque, 310
lipid lowering and antihypertensive medications, 276 reports, 310
mean interscan time, 275 significant plaque, 310
metabolic syndrome, 276 Coronary artery subsequent to coronary intervention, 604
preventive therapies, 277–279 Coronary atherosclerotic plaques, 7
survival curves, 275, 276 AIT, 212, 213
prospective studies, 262, 263 calcified nodule, 217, 218
recommendations, 279, 280 classification, 211, 212
risk adjustment, 279 coronary calcification, 220–223
risk estimation, 263 coronary care units, 211
risk prediction/reclassification, 262, 263 fibroatheromas, 213, 214
Stable Ischemic Heart Disease, 280 fibrous and fibrocalcific plaques, 219, 220
statin therapy, 263, 264 healed ruptures, 218, 219
treatment implication by, 742 intimal xanthoma, 212, 213
Writing Group III of the American Heart Association (AHA) intraplaque hemorrhage, 219–221
Prevention Conference V, 279 luminal thrombi, 215, 217
younger individuals, 263 PIT, 212–214
Coronary artery disease (CAD), 38, 349, 472, 659, 694, plaque erosion, 217
701, 739–744, 777, 780, 849 plaque rupture, 215, 216
functional assessment, 851 TCFA, 214–216
morphologic assessment, 850 total occlusion, 219
primary prevention, 742 vulnerable plaque, 224
prognostic significance, 666, 667 Coronary blood flow (CBF), 750, 769, 777, 780
to depict and characterize, 850 Coronary Computed Tomographic Angiography for Selective Cardiac
See also Chronic chest pain Catheterization (CONSERVE) trial, 854, 855
Coronary artery disease reporting and data system (CAD-RADS), 310 Coronary computed tomography angiography (CCTA), 176–178, 181,
calcified and non calcified plaque, 180, 182, 183 229, 230, 349, 350, 659, 749, 750, 777, 783, 784, 786, 787,
CHD, 187 791, 805, 807, 811, 815, 880, 887
delayed enhancement, 186, 187 ACCF/SCCT/ACR/AHA/ASE/ASNX/NASCI/SCAI/SCMR
electrophysiology appropriate use criteria, 712
left atrial ablation, 181, 182, 184 ACCURACY trial, 684
ventricular ablation, 184 acquisition principles
extracardiac structures, 187 broad detector, 72
myocardial perfusion imaging, 186, 187 ECG-based tube current modulation, 72
stable and acute chest pain, 179, 180 ECG-padding, 71
structural heart disease high pitch single heartbeat acquisition, 71
left atrial appendage, 185, 186 iterative reconstruction (IR) techniques, 73
mitral valve interventions, 185 prospective ECG-triggering, 70, 71
TAVR, 184, 185 retrospective ECG-gating vs. ECG-synchronization, 70
Coronary artery fistula, 574 scan coverage optimization, 69, 70
Coronary artery plaque analysis triple-rule-out, 71
acute coronary syndromes, 725 x-ray tube voltage, 72, 73
CTA, 725 acute chest pain, 665, 666, 854–855
ESSCT, 730 advantages, 685–686
FFRCT, 729, 730 appropriate use, 659
histopathologic investigations, 725 in Asians, 710
low attenuation plaque, 727 in asymptomatic low-risk patients, 740
morphologic and functional characteristics, 725, 726 CAC scanning, 30
napkin-ring sign, 727, 728 CAD
non-calcified plaque, 725, 726 to depict and characterize, 850
non-invasive assessment, 725 functional assessment, 851–854
partially calcified lesions, 725, 726 morphologic assessment, 850–851
positive remodeling, 727 prognostic significance, 666, 667
quantification, 728, 729 calcium scoring, 357, 358
radiomics, 731, 732 cardiac function, 399–401
spectral assessment, 731 cardiac MRI, 30
spotty calcium, 727, 728 cardiovascular risk factors, correlation with, 740
Coronary artery protection score (CAPS), 667 CAV
Coronary artery spasm, nonatherosclerotic CAD, 297, 298 coronary artery calcification, 398
Coronary artery stenosis vs.conventional coronary angiography, 398, 399
atherosclerotic aneurysm, 310, 312 vs. intravascular ultrasound, 399, 400
atherosclerotic plaques, 310–312 CONFIRM registry, 662
CAD-RADS system, 310 age and gender, prognostic significance, 662–664
clinical scores, 311 CAD severity and left ventricular ejection fraction,
left main stenosis, 310, 311, 313 prognostic significance, 665
918 Index

Coronary computed tomography angiography (CCTA) (cont.) Coronary density difference, 754
coronary artery calcium scoring, prognostic Coronary flow measurement, CT, 750
significance, 664, 665 Coronary sinus (CS), 231
diabetic patients, prognostic significance, 665 Coronary stenosis, 763, 777, 779
ethnic differences, prognostic significance, 663, 664 Coronary stents
continual hardware and software improvements, 25 blooming artifact, 376, 377
contrast media injection protocols (see Contrast media contrast enhancement, 376
injection protocols) edge-enhancing kernel, 376, 377
CORE 64 trial, 684 high kV acquisition, 376
coronary artery stenosis, 710 iterative reconstruction techniques, 377
diabetes, asymptomatic patients, 741, 742 monoenergetic imaging, 377
diagnostic and prognostic accuracy, 26, 712 motions artifacts, 376
diagnostic estimates of, 758 spatial resolution, 377
diagnostic yield Coronary veins, 231
asymptomatic individuals with low cardiovascular risk, 740 Corrected coronary attenuation, 753–755, 760
clinical risk factors, 740 Corrected coronary opacification (CCO), 750, 792
coronary artery calcium score, 740 Cost-effectiveness analysis (CEA), 673
specific high-risk groups, 741 cardiac CT
disadvantages, 684 acute chest pain patients, 674, 675
downstream outcomes after, 668 incidental findings on, 676
Dutch trial, 684 stable chest pain, 675–677
extracardiac findings, 400, 401 health outcomes, 673
high negative predictive accuracy, 26 primer for, 673, 674
ICA, 684 Coumadin ridge, 232
image quality, 396, 397 COURAGE trials, 791
immunosuppression, 400 C-Path, 901
impact, 743 Crista terminalis, 453
indication, 357 Cross-sectional density distributions, 4
intermediate risk, 357 Cryoballoon ablation, 532, 533
iodinated intravenous contrast, 396 c-statistic, 262
low heart rate variability, 396 CT Coronary Angiography Compared to Exercise ECG
management, role of (CT-COMPARE) study, 332
coronary calcium scoring, 742–743 CT perfusion (CTP), 350
primary prevention, risk for CAD, 742 Current procedural terminology (CPT) codes, 164, 173
treatment implications by, 743 AMA, 203
vs. MPI, 685 anatomic assessment
MPS, prognostic significance, 667, 668 CHD, 205
negative predictive value, 685, 852 PHVs, 205
nonobstructive CAD, 712 pulmonary vein assessment, 205
outcome studies, 26 TAVR, 206
PET MPI, 29 calcium score, 203, 204
PCI (see Percutaneous coronary intervention) coronary CTA, 204, 205
plaque characteristics, prognostic significance, 667 Curved multiplanar reconstructions, 37
positive, 853 Curved multi-planar reformat (cMPR), 133, 137, 138
post-operative MACE, 358 Curved planar reformation (CPR), 157, 565, 781
prognostic significance, 659–662 Cystinosis, 351, 352
prognostic value, 685
clinical risk factors, 741
coronary artery calcium score, 741 D
general asymptomatic population, 741 Daqing Diabetes Prevention Study, 691
specifc high-risk groups, 742 Data acquisition system [DAS], 8
prospective acquisition protocols, 396, 397 Data per measurement system, 13
significant stenosis, 357 DECIDE-Gold trial, 805
spatial resolution, 396 Decision tree, 675
specific high-risk groups, 741 Deconvolution approach, 817, 819, 826
SPECT-MPI, 29 Deep neural networks (DL), 898–899
stenosis vs. invasive coronary angiography, 25 DEFER trial, 767
stereo CT, 83 Delayed enhancement cardiac computed tomography scan, 815
stress echocardiography, 29, 358 arterial phase scan, 839
synthesis of emerging screening criteria, 739 contrast material, 836–837
TAG, 757, 759–764 delayed image acquisition, timing of, 837
temporal resolution, 396 ECV estimation, 838
Toshiba manufactures, 80, 82 enhancement pattern, 841
utilization, 682, 683, 740 enhancing image quality, 837
usage, 684 functional recovery prediction, 839
Coronary CT Angiography Evaluation for Clinical Outcomes:An infarct size, 841
International Multicenter (CONFIRM) registry, 728 injection protocol, 837
Index 919

radiation dose, 837 split-filter, 88

transmural extent, 841 temporal coherence, 89
tube setting, 837–838 temporal resolution, 89
Delayed phase acquisition, 796 “Virtual Non-Contrast” image, 91
Denoising, 92 X-ray spectrum, 88
Density μ, 3 Dual-energy computed tomography myocardial perfusion imaging
Determination of Fractional Flow Reserve by Anatomic Computed acquisition technique and radiation dose, 794–800
Tomographic Angiography (DeFACTO) trial, 851 clinical results, 805–807
Diabetes mellitus (DM), 665, 691–692, 741, 785 cost-effectiveness, 807
CAC score, 349 data analysis, 800–805
CAD, 349 limitation, 807
cCTA, 349, 350 low-voltage and high-voltage spectra, 794
coronary CTA, 742 pharmacological stress agents, 800
CTP, 350 static, 798–799
FFR-CT, 350 technical considerations, 792–794
MI, 349 timeline of rest/stress and stress/rest protocols, 796
primary and secondary outcomes, 350 virtual non contrast dataset, 794
type 1 classification, 349 workflow, 794, 795
type 2 classification, 349 Dual-layer detectors, 88
Diagnosis of Ischemia-Causing Stenoses Obtained Via Noninvasive Dual-phase injection, 557
Fractional Flow Reserve (DISCOVER-FLOW), 851 Dual source computed tomography (DSCT) systems, 51, 73, 376, 382,
Dietary structure, China, 692, 693 471, 811, 812, 874, 877, 878
Diffuse pulmonary artery hypoplasia/stenosis, 583 advantages and disadvantages, 62
DiGeorge syndrome, 590 clinical benefits, 62
Digital Imaging and Communications in Medicine cross-scattered radiation, 63
(DICOM) files, 167, 861, 862 DE algorithms, 64
Dipyridamole, 800 high-pitch scan mode, 62
Discordant atrioventricular connections, 580 independent measurement systems, 61
Discordant ventriculoarterial connection, 580 high temporal resolution, 71
Distributed parameter model, 823 model-based/measurement based correction approaches, 63
Diverticulum of Kommerell, 591 motion artifacts, 64
D-looped, 580 temporal resolution, 62
Dobutamine, 800 z-flying focal spot, 61
Dobutamine echocardiography, 396, 835 Dual-source dual energy configurations, 88
Doppler guidewire technique, 749 Ductal patency, 595
Double-acquisition coronary CT angiography, 298 Duke prognostic coronary artery disease index, 661
Double aortic arch, 591, 592 Dynamic computed tomography myocardial perfusion imaging
Double-chambered right ventricle, 583 (CTMPI), 778, 792, 884, 885, 887
Double discordance, 589 absolute quantification of perfusion parameters, 826
Double inlet left ventricle (DILV), 597, 598 advantage of, 811, 825
Double outlet right ventricle (DORV), 588–590, 862 benefit of, 826
D shaped model, 522, 523 cardiac phase for acquisition, 812
d-TGA disadvantage of, 826
after atrial switch operation, 608 heart coverage, 812
arterial switch procedure, 611 image analysis
with mustard palliation, systemic venous baffle stenting, 609 qualitative analysis, 816
d-TGA s/p arterial switch, 608, 611 quantitative analysis, 817–823
Dual-energy acquisition protocols, 779 semiquantitative analysis, 816–817
Dual energy cardiac computed tomography, 836, 885, 886 tissue attenuation curve, 815
beam-hardening artifacts, 90 visual analysis, 816
clinical applications, 87 vs. MRI, 826
cardiac imaging, 96 with other modalities, 825
vascular imaging, 93–95 patient preparation, 813
denoising, 92 quantitative analysis, 824–825
dual-layer detectors, 88 absolute vs. relative MBF values, 824
dual spiral scans, 88 cutoff values, 824
dual-source dual energy configurations, 88 myocardial blood volume (MBV), 824–825
Fast kV-switching, 88 segment, territory and patient, 824
Hounsfield unit, 87 radiation dose, 812–813
material classification/labelling, 92 scanning protocol, 813–815
photoelectric effect, 87 semiquantitative, 825
projection-data-based base material decomposition, 90, 91 semi-quantitative analysis, 823
radiation dose, 92 vs. static, 825
scattered radiation, 89, 90 temporal sampling rate, 813
sequential/spiral acquisition mode, 88 visual analysis, 813, 823, 825
slow kV-switching, 88 Dynamic spatial reconstructor (DSR), 5
spatial resolution, 89 Dypiridamole, 800
920 Index

Dyslipidemia, 691 WATCHMAN device, 542–544

Dyslipidemias, 742 WaveCrest, 544
Endocardial cushion defects, 583
Endothelial shear stress (ESS), 730
E Epicardial adipose tissue (EAT), see Epicardial
Early Identification of Subclinical Atherosclerosis fat measurement
by Noninvasive Imaging Research (EISNER) Epicardial and pericoronany adipose tissue, 701
study, 273, 274 Epicardial coronary artery disease, 767
Ebstein’s anomaly, 586, 587 Epicardial fat measurement, 144
ECG-based dose modulation, 72 ACS, 289
ECG-based tube current modulation, 72 anatomical landmarks, 286
ECG-gated scanning techniques, 51 anatomical, embryological and biochemical diversity, 287
ECG-gated single photon emission computed automated volume quantification tools, 286
tomography (SPECT), 407 cardiovascular outcome, 290
ECG-synchronization method, 70 coronary artery progression, 290
Echocardiography, 407, 446 coronary athersolerosis
Edge-enhanced post-processing filter, 377 calcified plaques, 288
Ejection fraction, 414 clinical observations support, 288
Electroanatomic mapping, 534 community-based sample, 288
Electrocardiogram (ECG), 331 conventional risk factors, 288
Electron beam computed tomography cross-sectional study, 288
(EBCT), 6, 52, 54, 279, 874 4 CT-based scores, 288
application, 7 histopathological and immunohistochemical studies, 288
clinical situations, 7 insulin resistance, 288
commercially viable electron beam scanner, 6 intra-abdominal visceral fat, 288
coronary artery calcification, 7, 8 local peri-coronary fat volume, 289
designs, 5 mixed/non-calcified plaques, 288
e-speed, 8 non-stenotic plaques, 288
large bell-shaped X-ray tube, 6 pericoronary fat, 288
quantitated coronary artery calcium score, 7 subcutaneous fat, 288
severe left ventricular hypertrophy, 6 coronary spasm, 289
stripped down scanner, 8 cytokines, 287
ultrafast CT, 6, 7 evidence support, 291
with severe aortic stenosis, 6 free fatty acid oxidation, 287
x-ray tube/fluoroscopic unit design, 5 imaging modalities, 285
Electronic and Musical Industries, Ltd [EMI], 4 myocardial ischemia, 289
Electronic health record (EHR), 197, 198 non-atherosclerotic cardiovascular disease, 291
Electronic medical records (EMR) system, 170, 893 non-invasive imaging tools, 286
Electrophysiological applications paracrine and vasocrine mechanisms, 287
atrial fibrillation pro- and anti-inflammatory mediators, 287
AEF, 538 processing time, 286
CBA, 532 semi-automated software, 286
cryothermal injury, 541 standardized reference values, 286, 287
CT technique, 533 surrogate imaging, 285
phrenic nerve injury, 540, 541 vascular supply, 285
PNI, 540, 541 volumetric and distance measurements, 285
post-PVI interpretation, 538 Epstein-Barr virus (EBV), 304, 305
pre-PVI interpretation, 533–537 Esophageal wall thickening, 648
PVS, 538–540 Estimated glomerular filtration rate (eGFR), 126–128
RFCA, 532 Eustachian valve, 453
treatment of, 531 Exercise electrocardiography, 19
LAA occlusion Extended Brilliance Workstation (EBW), 78
CHA2Ds2-VASc score, 541 Extended Toft model, 819, 821, 823
CHA2DS2-VASc score, 541 Extensive atherosclerosis, 770
CT scanner, 545 Extra cardiac anatomy, 38
echocardiography, 548 Extracellular volume (ECV) fraction, 143, 835
extra-atrial anatomy, 547
HAS-BLED score, 541
LAA neck, 545 F
LAA ostium, 545 Facebook’s facial recognition, 893
LAmbre, 544 FACTOR-64 trial, 743
LARIAT device, 544 FAME trial, 767, 791
lobar region, 546 Fanconi syndrome, 351
postprocessing, 545 Fast alternating switching, 61
PROTECT trial, 548 Fast signal pulses, 882
surgical-based epicardial, 542 Fatty streak, 212, 213
Index 921

Fermi model, 823 Gaucher disease (GD), 350, 351

Fibroatheromas, 213, 214 GE Healthcare China, 699
Fibroelastomas, 459 Giant cell arteritis (GCA), 300
Fibroma, 463 Gillinov-Cosgrove clip, 542
Fill Hole Freeform operation, 871 Glagov effect, 774
Filtered back projection (FBP), 4, 15, 76, 82, 113–114 Glenn superior vena cava, 616
First pass arterial phase scan, 833, 834 Global left ventricular muscle mass, 7
First pass enhancement scanning, 884 Global resting flow, estimation of, 769
Flow rates, 109 Glycosaminoglycans (GAG), 351
Flow-limiting stenosis, 802 Gooseneck deformity, 583
Fontan circuit, 560, 561, 617 Gore helix septal occluder, 494
Fontan procedure, 616 Gray-level co-occurrence matrix (GLCM), 718, 719
Forward propagation phase, 894 Gray level gap length matrix (GLGLM), 720
Fossa ovalis (FO), 231 Gray-level run length matrix (GLRLM), 720
4-dimensional (4D) datasets, 131 Gray-level size zone matrix (GLSZM), 720
Fractional flow reserve (FFR), 337, 350, 358, 729, 730, 752, 760 Guidelines, cardiac CT
Framingham Risk Score, 259, 711 abnormal prior stress imaging, 252
Frank-Starling law, 424 ACCF, 245
Full width at half maximum (FWHM), 883 ACR, 245
Functional evaluation Appropriate Use Criteria and Guidelines, 245
area-length method, 411 arrhythmias without ischemic equivalent, 251
ARVD, 407 asymptomatic, 250
biplane area- length method, 411 asymptomatic without ischemic equivalent, 253
CHD, 407 CAD-RADS reporting and data system, 248, 249
cine ventriculography, 408 cardiac rehabilitation, 253
CMR, 407 coronary calcium Agatston score, 252
computed tomography ECG test stable symptoms, 252
advantages and disadvantages, 408 equipment specifications, 248
image reconstruction, 411 exercise prescription, 253
medications and contrast, 409 heart failure imaging, 250
radiation dose reduction strategies, 409 newly diagnosed heart failure, 251
scan mode, 408, 409 non-cardiac surgery, 253
diastolic dysfunction, 415 non-invasive testing, 249, 250
echocardiography, 407 normal CT/invasive angiogram/normal stress test, 253
left atrium, 419 obstructive CAD, 252
left ventricle post-revascularization symptomatic, 253
cine-ventriculography and transthoracic echocardiography, 415 prior coronary angiography, 252
function, 407 prior testing without intervening revascularization, 251
global systolic dysfunction, 415 qualifications and responsibilities of
heart failure, 415, 416 personnel, 246–247
ischemic heart disease, 415 risk assessment, 250
normal regional LV functional parameters, 415 SCCT, 245
normal right ventricular values, 417 sequential/follow-up testing, 251, 252
quantification of, 412 stable ischemic heart disease, 250
retrospective ECG gating, 410 stress imaging study/non-obstructive CAD, 252, 253
nuclear medicine techniques, 407 syncope without ischemic equivalent, 251
qualitative, 414 unknown functional capacity, 253
quantitative parameters, 414 worsening symptoms, 253
right ventricle
ARVD, 417
congenital heart disorders, 417 H
dysfunction, 418 Hagen-Poiseuille equation, 432
function, 407 Half-scan reconstruction (HALF), 838
manual, automated or semi-automated techniques, 416 Health Information Portability and Accountability
pulmonary embolism, 417 Act (HIPAA), 167
pulmonary hypertension, 417 Healthcare system, 39
quantification of, 416 Heart coverage, 812
volumes, 418 Heart ischemia, 128
Simpson’s method, 411 Heart motion, variation of, 57
stroke volume, 414 Heart transplantation
threshold-based segmentation, 411–413 cardiovascular anatomy
heterotopic transplantation, 392, 394
orthotopic transplantation, 391–393
G CAV (see Cardiac allograft vasculopathy)
Gamma emitting monovalent cations, 19 CCTA (see Coronary CT angiography)
Garbage in, garbage out, 899 HeartFlow algorithm, 769, 770, 772, 774
922 Index

Heart-lung axis IBM’s AI software (Watson), 899

atherosclerosis and respiratory disease, 627 Idiopathic pulmonary fibrosis, 628
cardiovascular disease and COPD, 628–630 Image data display and analysis
cardiovascular disease and lung cancer screening, 631 automated segmentation, 133, 134
cardiovascular disease and respiratory diseases, 630 cardiac planes, 133, 135
cardiac CT, extra-cardiac findings in lungs, 631 cMPR, 133, 137, 138
CT imaging methods, to assess, 623–625 coronary tree, 134, 139
pulmonary hypertension and heart, 625 MAR, 133
acute pulmonary embolism, 625, 626 MINIPs, 135, 139
causes of, 627 MIPs, 135
chronic thromboembolic pulmonary hypertension, 626 projection techniques, 134
lung disease, 626, 627 semi-automated/manual segmentation, 133, 134
shared aetiology and pathogenesis, 625 “simple” MPR image, 133
HeartMate II, 498 sMPR, 133
HeartWare pump, 497 thick-MPRs, 133, 136
Heel effect, 59 3D VRT, 131–133
Heinz Nixdorf Recall Study (HNR), 263 Image noise, 70
Hemangioma, 461 Image quantification
Hemodynamic significance, 767 calcium scoring, 136
Hepatic cyst, 655 chamber volumes and ejection fraction, 141–143
Heterotaxy syndromes, 597, 599 coronary stenosis quantification/fractional flow reserve, 136, 140
Hiatal hernia, 457 CT data, 144
High-definition (HD) scan modes, 878 dual-energy CT/perfusion imaging, 143–145
High pitch single heartbeat acquisition, 71 myocardium, regional function and epicardial fat, 142–144
High resolution coronary imaging, 605 plaque composition analysis, 140, 141
High-risk plaque (HRP) features SPECT/PET, 144
adverse outcomes, 314 Image reconstruction
atherothrombosis, 315 back projection, 118, 120
CT appearance, 314, 315 calibration phantoms, 117
in-stent restenosis, 315, 316 convolution kernel, 119, 120
low-attenuation plaque, 314 ECG editing, 118, 119
myocardium, 315, 316 iterative reconstruction, 121–123
napkin-ring sign, 314 raw data, 117
perivascular fat stranding, 315, 316 Z-interpolation, 117, 118
positive remodeling, 314 Imatron scanners, 6–8
spotty calcium, 314 Immunoglobulin G4-related disease (IgG4-related
High-sensitivity C-reactive protein (hs-CRP), 260, 297 disease), 302–304
High-sensitivity troponin (hs-troponin) assays, 339, 340 Impella, 497
Histoplasmosis, 629 Implantable loop recorder (ILR), 494
Holmes heart, 597 Implantable pacemakers, 492
Homogentisate 1,2-dioxygenase (HGO), 351 Impulse response function (IRF), 817–818, 822
Hounsfield units (HU), 87, 133, 309, 715 Incidental pulmonary hamartoma, 653
Hybrid acquisition techniques, 71 Incomplete LAA exclusion, 542
Hybrid iterative reconstruction, 121, 123 Incremental cost-effectiveness ratio (ICER), 674
Hyperattenuating myocardial pattern, 795 Indeterminate left hilar lymph node enlargement, 653
Hypercholesterolemia, 691 Inflammation, 625
Hyperemia, simulation of, 769 inPrint, 862
Hyperlipidemia, 691 In-stent restenosis (ISR), 375, 376
Hypertension, 689–690 diagnostic accuracy, 378
Hypertriglyceridemia, 691 guidelines, 378
Hypo-attenuated myocardium, 800 lumen visibility, 378
Hypoplastic left heart syndrome (HLHS), 595 non-invasive assessment, 378
sensitivity and specificity, 378
stent evaluation, 378
I Integrated discrimination improvement index (IDI), 763
IAC accreditation program Intensity profile, 59
coronary calcium scoring, 42 Interarterial septal abnormalities, 639, 640
criteria and guidelines, 43 Interatrial septal aneurysm, 458
daily and periodic QC tests, 43 Interleukin 6 (Il-6), 287
interpretive quality, 43 International Chapel Hill Consensus Conference
preventive maintenance service, 43 (CHCC), 299–300
quality improvement program, 42 International Society for Heart and Lung
radiation safety, 43 Transplantation (ISHLT), 391
scanning requirements, 42 Interpreting studies, 38
selecting or de-selecting coronary calcification, 42 Interrupted aortic arch (IAA), 586, 587
technologist requirements, 43 Intimal xanthoma, 212, 213
Index 923

Intra-aortic balloon pump (IABP), 496 Left internal mammary arteries (LIMA), 381, 383
Intra-cardiac blood flow simulation, 774, 775 Left lateral ridge, 453
Intraplaque hemorrhage, 219–221 Left main coronary artery, 309, 310
Intravascular ultrasound (IVUS), 299, 350, 395 Left superior pulmonary vein (LSPV), 532
Intravenous (IV) heparin, 331 Left superior vena cava (LSVC), 597
Intravenous immunoglobulin (IVIG), 301 Left upper lobe pulmonary vein (LUPV), 186
Invasive coronary angiography (ICA), 382, 384, Left ventricle (LV), 231
676, 684, 787, 805, 834, 849, 851, 854, 855, 884 dysfunction, 829
Invasive fractional flow reserve (FFR), 767, 771 thrombus, 636
Iodine concentration, 792, 800, 817 Left ventricular apical thin-point (LVATP), 231
Iodine delivery rate (IDR), 110 Left ventricular assist devices (LVADs), 497
Iodine maps, 884 Left ventricular ejection fraction (LVEF), 426, 665
Iodine volume, 109 Left ventricular end-diastolic volume index (LVEDVI), 426
Ischemic cardiomyopathy, 830 Left ventricular muscle mass index (LVMMI), 426
Ischemic heart disease (IHD), 239, 427, 829 Left ventricular outflow tract (LVOT), 472, 504–505, 508, 519
Ischemic stroke, 635 Left ventricular stroke volume index (LVSVI), 426
cardioembolic sources, 636 Length of stay (LOS), 332
aortic atheroma, 643 Lesion length ratio (LLR), 759
cardiac thrombus, 636, 637 LIMA, see Left internal mammary arteries
cardiac tumors, 638, 640 Lipoma, 461
interarterial septal abnormalities, 639, 640 Lipomatous hypertrophy of interatrial septum (LHIS), 456
SEC, 638, 639 Lipomatous metaplasia, 831
valvular calcification, 642 Local resting flow, estimation of, 769
valvular vegetation, 642 Low cardiovascular risk, asymptomatic individuals with, 740
cardioembolic stroke, 635 Low plaque attenuation, 314
clinical importance, 635 Low-density lipoprotein (LDL), 691
Isotropic 3-dimensional (3D) datasets, 131 Low-density lipoprotein cholesterol (LDL-C), 742
Iterative methods, 4 Lower heart rate
Iterative reconstruction (IR) technique, 14–17, 73, 121–123 beta-blockers, 103, 104
IT infrastructure calcium channel blockers, 104
acquisition plane, 195, 197 cardio-selective β-adrenergic antagonist, 103
clinical management, 198, 199 ivabradine, 104
computed tomography, 195–197 metoprolol, 103
EHR, 197–199 Lower tube voltage (kVp), 113
x-ray films, 195, 196 Luminal myofibroblastic proliferation (LMP), 301
Luminal stenosis, 851
Luminal thrombi (Th), 215, 217
J Lung disease, 626, 627
Joint Commission Accreditation Program Lung inflammation, 625
scanner requirements, 43 Lymphoma, 467
technologists requirements, 43 Lysosomal cystine transporter (CTNS), 351
Jungner classic screening criteria, 739

M
K Machine learning (ML), 198, 770, 879, 887, 889
Kaplan-Meier curve, 663, 667 adapting to, 903–904
Kawasaki disease (KD), 300–303 advantage, 899
Klippel-Trenaunay-Weber syndrome, 630 back propagation phase, 896
Kommerell’s diverticulum, 591, 593 coronary artery calcium detection, 898
definition, 894
development and training, 898
L forward propagation phase, 894
LAA pseudothrombus, 537 importance in medicine, 893
Lambert-Beer law, 3, 4 in medicine
Late enhancement computed tomography (LE-CT), 830, 832, 834, accuracy, 900
835, 837–839, 842 early detection, 900, 901
Late gadolinium enhancement-magnetic resonance ICU, 900
imaging (LGE-­MRI), 829, 830, 833–835, 838 mortality prediction, 900
Leadless pacemaker, 494 outcome prediction, 901
Left anterior descending (LAD) artery, 176, 227, precision medicine, 901, 902
309, 310, 565, 749, 750, 802, 804, 816, 853 prognosis prediction, 901
Left atrial (LA) ablation, 181, 182, 184 screening, 900
Left atrial appendage (LAA), 185, 186, 546 Watson, IBM’s AI software, 899, 900
Left atrial ejection fraction (LAEF), 429 in radiology
Left atrial systolic volume index (LASVI), 426 abnormalities detection, 902
Left circumflex coronary artery (LCx), 176, 227, 565, 804 accuracy, 902
924 Index

Machine learning (ML) (cont.) TMVR implantation, 519

CACS, 902–904 VSD, 527
cardiac datasets, 902 Minimally invasive direct coronary artery
cardiac echocardiography, 902 bypass (MIDCAB), 384
cardiac imaging algorithms, 902 Minimum intensity projections (MinIP), 839
challenges, 902 Minkowski functionals, 722
efficiency, 902 Minkowski–Bouligand dimension, 722
image segmentation, 902 MitraclipTM device, 491, 519
myocardial perfusion imaging, 902 Mitral annular calcification (MAC), 456, 476, 520, 642
pattern recognition, 902 Mitral annulus, 642
prognosis and outcome prediction, 903 Mitral Clip, 489, 491
SPECT, 902 Mitral regurgitation (MR), 489
limitations, 893, 899 Mitral stenosis (MS), 476
neural units, 894 Mitral valve annuloplasty, 491
supervised learning, 896, 898 Mitral valve prolapse (MVP), 232, 476
unsupervised learning, 898, 899 Model fabrication, 861
workflow, 894 Model-based iterative reconstruction, 121, 123
Magnetic resonance imaging (MRI) Moderate stenosis, 770
cardiac masses, 23 Monte Carlo simulation, 675
cardiovascular disease, 23 Motion artifacts, 64
CHT, 579 Motion correction approach, 875, 876
interpretation of, 23 Mucopolysaccharidoses (MPS)
left and right ventricular structure and function, 23 diagnostic performance, 351
pharmacologic stress, 23 GAG, 351
soft tissue differentiation, 23 left sided cardiac valves, 351
valvular heart disease, 23 subtypes I and II, 351
Major adverse cardiovascular events (MACE), 290, 332, transthoracic two-dimensional echocardiography, 351
430, 841, 887 Multi detector computed tomography (MDCT), 8, 9, 140, 195, 503,
Markov model, 675 520, 531, 684, 873, 878
Materialise software, 862 acquisition time, 382
Mathematical modeling technique, 674 ADC electronics and data transmission, 54
Maximum intensity projection (MIP), 472, 565, 839 aortic cross-clamping, 381
in coronal plane, 598 coronary calcium scanning, 381
in sagittal plane, 598 dual source acquisition, 382
MDCT coronary angiogram, 572 ECG-dyssynchrony, 382
Mediastinum, 358, 457, 464, 539, 540, 647, 648, 652–653 ECG-signal, 56, 57
Medical 3D printing, 859, 861 fan-beam CT design, 54
Medical turf wars, 37 key requirements, 54
Mesotholioma, 464 LIMA, 381
Metal artifacts, 96 minimally invasive techniques, 381
Metastasis, 464 motion artifact, 382
Microvascular obstruction (MO), 839 myocardial revascularization, 381
Mid-aortic syndrome, 606 off-pump surgery, 381
Minimally invasive cardiac procedures patency and long-term survival, 381
ASD, 527 patency vs. occlusion, 382
mitral valve postoperative planning, 384–386
ACC/AHA guidelines, 519 preoperative planning, 382–385
access point, 526 quantum and electronic choice, 54
anatomy, 521 radiographic assessment, 381
annular area, 521–523 reoperative surgery, 381, 386–388
CardiAQTM, 520 16-slice CT detector, 55, 56
EVEREST I trial, 520 slip-ring designs, 54
EVEREST II trial, 520 solid-state scintillation detectors, 54
fluoroscopic angulation, 524, 525 state-of-the art X-ray tubes, 54
landing zone characteristics, 523, 524 temporal resolution, 381, 382
LVOT obstruction, 524 3rd generation DSCT, 382
MDCT, 520 whole body scan field of view, 54
Mitraclip device, 520 z-flying focal spot, 56, 57
pre-procedural CT, 520 Multi-Energy CT, see Spectral CT imaging
STS risk predication model, 519 Multi ethnic study of atherosclerosis (MESA) study, 136, 743
TendyneTM, 520 Multilevel baffle obstruction, 609
TiaraTM, 520 Multi-planar reformation (MPR), 57, 133, 472, 565, 839
TwelveTM, 520 Multiple gated acquisition (MUGA), 408
pulmonary valve replacement, 526 Multiple gray-level size zone matrix (MGLSZM), 720
pulmonary veins isolation, 527, 529 Multiple perfusion defects, 626
TAVR, 519 Multi-segment reconstruction, 80
Index 925

Multi-slice computed tomography (MSCT) scanners, 830 Neural network, 894, 898, 902
Multisource interior tomography system, 874, 875 Neural units, 894
Multivessel disease, 803 Nikaidoh procedure, 610
Muscular defects, 581 Nodes, 894
Mustard palliation, 609 Nonalcoholic fatty liver disease (NAFLD), 337
Myocardial blood flow (MBF), 811–813, 816, 817, Nonatherosclerotic coronary artery disease
819, 822–824, 826, 885 CAEBV, 304, 305
Myocardial blood volume (MBV), 815, 824 coronary artery dissection, 298, 299
Myocardial bridging (MB), 572, 764 coronary artery spasm, 297, 298
Myocardial delayed enhancement techniques, 23 IgG4-related disease, 302–304
Myocardial infarction (MI), 341, 349, 637, 638, 802, 831 KD, 300–303
Myocardial ischemia, 297, 767, 777, 784, 786, 787, 792, 813, 826 TAK, 299–301
Myocardial mass, 414 Non-cardiac computed tomography, 831
Myocardial perfusion defects (MPD), 800 Non-cardiac surgery, pre-operative risk assessment
Myocardial perfusion imaging (MPI), 324, 681, 682 ACC and AHA, 355
add-on codes, 681 algorithm for, 355, 356
vs. CCTA, 685 appropriate use criteria, 356
in machine learning, 902 beta-blockade, 358
negative results, 684 CCTA, 357, 358
preauthorization, 682 clinical challenges, 358, 359
primary CPT-4 code, 681 clinical risk factors, 355
Myocardial perfusion reserve (MPR), 824 complications, 355
Myocardial perfusion scintigraphy (MPS), 667, 668, 850, 854 FFRCT, 358
Myocardial scar calcification, 831–832 functional assessment, 355, 356
Myocardial viability assessment, CT, 843 medical management, 358
contrast enhanced cardiac computed tomography, 829–830, 841 non-flow limiting stenosis, 358
dual-energy CT, delayed imaging, 836 RCRI, 355
extracellular volume fraction, 835 recommendations, 356
first pass arterial phase scan, 833, 834 statins, 358
late contrast enhancement, 834, 835 Non-contrast calcium scoring scan, 70
post-percutaneous coronary intervention Non-contrast computed tomography, 831
delayed imaging, 835, 836 Noninvasive testing (NIT), 849, 850
wall thinning, 832 Nonspecific left adrenal nodule, 655
delayed enhancement cardiac CT scan Non-viable myocardium, 829
arterial phase scan, 839 No-reflow phenomenon, 839
contrast material, 836 Normal pulmonary vein variations, 535, 536
delayed image acquisition, timing of, 837 Nuclear cardiology
ECV extimation, 838 combined assessment of clinical and imaging data, 20
enhancement pattern, 841 components, 21
enhancing image quality, 837 ECG-gated SPECT imaging, 20
functional recovery prediction, 839 first pass or multiple-gated equilibrium technique, 20
infarct size, 841 gamma emitting monovalent cations, 19
injection protocol, 837 incremental information, 21
radiation dose, 837 PET imaging, 20
transmural extent, 841 planar imaging methods, 20
tube setting, 837 semi-quantitative and quantitative scoring systems, 20
non-contrast/non-cardiac CT, 831–832 SPECT-MPI, 20
pros and cons, 830–831 test abnormality, 20
revascularization, 829 thallium-201, 19
Myocardial volume reserve (MVR), 824 universal scoring system, 20
Myocardial wall thickness, 414 NXT trial, 770, 772, 773, 854
Myocardium, 446–448
Myxoma, 458, 638
O
Obesity, 692
N BMI, 352
Napkin-ring sign, 314 CAC, 352
National Cardiovascular Data Registry, 684 comorbidities, 352
National Cardiovascular Data Registry's CathPCI Registry, 849 development of, 352
National Institute for Health and Care Excellence (NICE), 673 diagnostic accuracy, 352
National Survey on Chinese Students’ Physical Fitness EAT, 352
and Health, 692 1st obtuse marginal branch (OM1), 176
Navier-Stokes equations, 769 Odds ratio (OR), 357
Negative predictive value (NPV), 382, 636, 850, 855 Old myocardial infarction (OMI), 829
Neighborhood gray-tone difference matrix (NGTDM), 720 Optical coherence tomography (OCT), 299
Net reclassification index (NRI), 262 Optimal medical therapy (OMT), 205, 327
926 Index

Outsourcing, 145 pericardial effusions, 443, 444

Overweight, 692 pericardial tamponade, 445
Oxidized low-density lipoproteins (ox-LDL), 288 pericarditis, 445
Pericardial effusions, 443
Pericardial tamponade, 445
P Perimembranous defects, 582
Paracardiac fat, 285 Periodic acid Schiff (PAS), 213
Paraganglioma, 462 Peripheral pulmonary embolism, 652
Paravalvular aortic regurgitation (PAR), 511 Peripheral pulmonic stenosis, 583
Partial cor-triatriatum, 595 Phantom measurements, 41
Partial dataset misalignment artifacts, 784 Pharmacological interventions
Partially anomalous pulmonary venous connection lower heart rate
(PAPVC), 592, 593 beta-blockers, 103, 104
Passive ejection fraction (pEF), 430 calcium channel blockers, 104
Passive emptying volume (pEV), 430 cardio-selective β-adrenergic antagonist, 103
Patent ductus arteriosus (PDA), 499, 582, 583 ivabradine, 104
Patent foramen ovale (PFO), 527, 639, 640 metoprolol, 103
Pathobiologic Determinants of Atherosclerosis in Youth stress perfusion evaluation
(PDAY) study, 212 adenosine, 105
Pathologic intimal thickening (PIT), 212–214 dipyridamole, 105
Patient positioning, 154–156 dobutamine, 106
Patient selection regadenoson, 106
acute chest pain, 237, 238 vasodilatation, 104, 105
advantages, 235–237 Phase-contrast computed tomography, 880, 881
appropriateness criteria, 235, 236 Photoelectric effect, 87
chronic chest pain, 238 Photon counting detectors, 878, 880
disadvantages, 235 Phrenic nerve (PN) injury, 184
echocardiography Physical activity, 692
cardiac devices and cardiac surgery, 240 Planar imaging methods, 20
cardiac embolic source, 239 Plaque calcification
CHD, 238, 239 CAC progression, 309
great vessel anomalies/abnormalities, 238, 239 dual-energy scanning, 309
new systolic/diastolic heart failure, 239 high-risk feature, 309
pericardium, 239 LAD, 309, 310
regional wall motion abnormalities, 239 left main coronary artery, 309, 310
valve disease, 239 RCA, 309, 310
MRI society guidelines, 309
biventricular function, 240 volume averaging and beam hardening, 309
cardiac mass, 241 volume calcium subtraction, 309
CHD, 240 Plaque composition analysis, 140, 141
great vessel anomalies, 240 Plaque erosion, 217
ischemic heart disease, 241 Plaque rupture, 215, 216
pericardial disease, 241 Plaque vulnerability, 774
valve disease, 240 PLATFORM trial, 772
nuclear cardiac imaging, 241, 242 Pleura, CTA, 654
risk stratification, 237 Poiseuille-based index, 763
Patlak method, 822–823, 826 Polyarteritis nodosa (PAN), 300
PDA Clip, 499 Polysplenia syndrome, 599
Percent atheroma volume (PAV), 278 Positive predictive value (PPV), 636, 851
Percutaneous coronary intervention (PCI), 313, 327, 375, Positive remodeling, 314
381, 796, 805 Positron emission tomography (PET), 407, 811, 835
coronary anomalies, 368 Post-percutaneous coronary intervention (PCI) delayed
CTO, 366–370 imaging, 835, 836
flow-limiting plaque, 361, 362 Pre-test probability, 678
ischemia-guided intervention, 363, 364 Primary care provider (PCP), 275
peri-procedural guidance, 369–372 Primum defects, 580
plaque assessment, 363–365 Prinzmetal’s angina, 321, 322
stent size and length, 361–363 Private medical-grade cloud, 197
SYNTAX III REVOLUTION trial, 371 Progression of Early Atherosclerosis Study, 30
time-saving multidimensional information, 371 Projection-data-based base material decomposition, 90, 91
vulnerable plaque, 364–366 Prospective Army Coronary Calcium (PACC) Project, 274
Percutaneous valve endocarditis, 612 Prospective Longitudinal Trial of FFRCT Outcome and Resource
Pericardial cyst, 456, 457 Impacts (PLATFORM) trial, 851
Pericardial disease Prospective Multicenter Imaging Study for
additional morphologic abnormalities, 446 Evaluation of Chest Pain (PROMISE)
anatomy, 443 trial, 28, 204, 326, 850, 851, 854
constrictive pericarditis, 445, 446 Prosthetic heart valves (PHV), 205
Index 927

CCTA technology, 479, 483 Agatston score, 717

degenerative lesions, 481 cut-off values, 718
dehiscence, 481 definition, 716
paravalvular abscess, 481 distribution shape, 716
vs. periprosthetic fibrosis, 483 histogram analysis, 716
pseudoaneurysm, 481 interquartile range, 716
renal dysfunction, 483 kurtosis, 716
ruputured chordae, 481 mean, 716
stuck valve, 483 median, 716
SVD, 482 moments, 716
TTE and TEE, 479 percentiles, 716
PROTECTION I study, 700 Shannon entropy, 717
Pseudoaneurysms, 457 skewness, 716
Pseudo-thrombus, Fontan circuit, 617 standardized/normalized central moments, 716
Pulmonary atresia, 584, 596 variance, 716
with intact ventricular septum, 595 voxel values, 716, 718
Pulmonary disease, 701 radiomics-based patient analysis, 715
Pulmonary embolism, 95 shape-based metrics
Pulmonary hypertension, 628, 630 coronary geometry, 722
heart-lung axis, 625 fractal dimension, 722
acute pulmonary embolism, 625, 626 MACE rates, 722
causes of, 627 Minkowski functionals, 722
chronic thromboembolic pulmonary hypertension, 626 1-dimensional (1D) metrics, 721
lung disease, 626, 627 2-dimensional (2D) metrics, 721, 722
Pulmonary nodule, 653 3-dimensional (3D) metrics, 721, 722
Pulmonary stenosis, 583 texture-based metrics
Pulmonary valve leaflets, 617 GLGLM, 720
Pulmonary valve placement, 612 GLRLM, 720
Pulmonary vascular resistance (PLVR), 431 GLSZM, 720
Pulmonary vein mapping protocol, 171 laws' texture energy measures, 721
Pulmonary vein stenosis (PVS), 538, 540 MGLSZM, 720
Pulmonary veins, 232–234 NGTDM, 720
plaque composition, 718
second-order statistics, 718–720
Q texture analysis, 718
Qualified medical physicist (QMP), 41, 42 tumor heterogeneity, 721
Quality assurance (QA) program, 170 transform-based metrics
Quality improvement (QI) program, 42 definition, 722
Quality-adjusted life-year (QALY), 673 Fourier and Gabor transforms, 722, 723
Quantitated coronary artery calcium score, 7 wavelet transforms, 723
Quantitative coronary angiography (QCA), 350 Radionuclide ventriculography (RVG), 407
RadSite Accreditation Program
scanner requirements, 43, 44
R technologists requirements, 44
Radiation dose, 794, 812, 837, 878 Receiver operating characteristic (ROC)
cancer risk, 17 curve, 261, 752, 761–764
dose reduction techniques, 17 Reconstruction technologies, 113
filtered back projection, 15 Redo PFO closure, 495
iterative reconstruction, 14–17 Regadenosone, 780, 787–788, 800
prospective ECG-gating Region of interest (ROI), 314, 862, 864
with a sequential (or “step-and-shoot”) data Regional myocardial perfusion, 7
acquisition, 12, 13 Regional wall motion abnormalities (RWMA), 666
with a spiral data acquisition and high-pitch, 13, 14 Regional/segmental EF, 414
retrospective ECG-gating techniques, 12 Relative myocardial blood flow values, 824
scan length, 15 Relative value units (RVU’s), 38
tube current modulation, 14 Remodeling index (RI), 667
voltage adaptation, 14 Rényi dimensions, 722
Radiation dosimetry measurements, 42 Repaired CHD, cardiovascular CT
Radiofrequency catheter ablation, 532 aortic arch imaging, 606, 607
Radiologists, 38 d-TGA s/p arterial switch, 608, 611
Radiology benefits management companies (RBMs), 682 indications, 618
Radiomics intervention, coronary artery imaging after, 604–606
classification, 716 risk of, 603
coronary artery plaque analysis, 731, 732 single ventricle heart disease, 612–616, 618
datamining techniques, 715 situations in, 604
definition, 715 Tetralogy of Fallot, 611, 612
HU values, 715 transposition complexes, 607
intensity-based metrics use of, 604
928 Index

Reporting, cardiac CT Semi-quantitative and quantitative scoring systems, 20

assure quality and consistency, 173 Sensitive detector technology, 849
CAD-RADS (see Coronary artery disease reporting and data Sensitivity analysis, 678
system (CAD-RADS)) Sensitivity and specificity, 851
CCTA, 176–178, 181 Septal defect patch, 867, 869, 870, 872
coronary artery calcium scoring, 175 Sequential segmental analysis, 579, 580
image acquisition, 174 Service lines, 38
image quality, 174, 175 Shannon entropy, 717
image reconstruction, 174 Short-axis (SAX), 234
impressions and management recommendations Siemens cFFR application, 771
calcium scoring, 189 Simpson’s method, 407, 411
comprehensive report, 188, 189 Single bolus technique, 837
emergency department, 190 Single lead pacemaker, 492
indications, 173, 174 Single photon emission computed tomography
medical record documentation, billing, and compliance, 173 (SPECT) imaging, 20, 279, 811, 812, 815,
noncoronary cardiac findings, 178, 179 816, 823–825, 831, 834, 835, 854, 902
Respiratory disease, 623 Single tilting disc valve, 487
cardiovascular disease and COPD, 628–630 Single ventricle heart disease, 612, 613
cardiovascular disease and lung cancer screening, 631 anatomic findings
cardiovascular disease and respiratory diseases, 630 after 3rd stage palliation, 615
Rest transluminal attenuation gradient, 763 stage 1 and 2, 613
Resting coronary blood flow, 769 functional cardiovascular CT imaging, 616, 618
Rest-stress approach, 780 scan acquisition considerations and recommendations, 615, 616
Revascularization, 774, 829 scan modifications, 616
Reverse attenuation gradient (RAG) sign, 763, 764 scan recommendations, 613, 615
Reverse attenuation sign, 764 Single-energy computed tomography myocardial
Reversible ischemia, 802 perfusion imaging, 805, 807
Revised Cardiac Risk Index (RCRI), 355 Single-energy technique, 792
Rhabdomyoma, 463 Single-shot acquisition, 779
Right atrial stroke volume index (RASVI), 431 Single-source CT systems (SSCT), 73
Right atrium (RA), 231 Single-source helical, 88
Right coronary artery (RCA), 176, 227, 309, 310 Single-source rapid-switching, see Fast KV-switching
Right internal mammary artery (RIMA), 384 Single-source sequential Dual Energy CT scanners, see Slow
Right pulmonary artery occlusion, 629 KV-switching Dual Energy CT scanners
Right ventricular ejection fraction (RVEF), 431 Single-source twin-beam, see Split-filter Dual Energy CT scanners
Right ventricular MMI (RVMMI), 431 Single-tube multi-detector CT (MDCT) scanners, 812
Right ventricular outflow tract (RVOT), 604 Sinoatrial nodal artery (SNA), 227, 566
Right ventricular output index (RVCI), 431 Sinogram, 117, 118
right ventricular stroke volume index (RVSVI), 431 Sinus venosus defects, 580
Right ventricular-dependent coronary circulation, 595 Situs ambiguous, 580
Risk Or Benefit IN Screening for Cardiovascular Disease Situs inversus, 579
(ROBINSCA) trial, 275 Situs solitus, 579
Röentgen, Wilhelm Conrad, 3 64-slice CT systems, 51
Royal Australian and New Zealand College of Radiologists Slip-ring technology, 8
(RANZCR), 44 Slow kV-switching Dual Energy CT scanners, 88
Rule Out Myocardial Infarction using Computer Assisted Smaller scan field of view (SFOV), 154
Tomography (ROMICAT) trial, 204, 666 Smoking, 690–691
Snap Shot Freeze (SSF), 76
Society of Cardiovascular Computed Tomography
S (SCCT), 25, 176, 229, 245, 310, 666
Saddle pulmonary embolism, 627 Society of Thoracic Surgeons (STS) database, 387, 519
St. Francis Heart Study, 277 Soft cardiac plaque composition, 96
Saline chaser, 110 Solid part creation operation, 866
Sarcomas, 464 Spatial resolution, 89
Scan field of view (SFOV), 873 Speckle tracking echocardiography, 902
Scattered radiation, 89, 90 Spectral computed tomography, 97, 792, 882
SCI-EXPANDED, 701 Split-filter dual energy CT scanners, 88
Scimitar syndrome, 593 Spontaneous coronary artery dissection (SCAD), 298, 299
SCOT-HEART trial, 743, 850 Spontaneous echo contrast (SEC), 638, 639
Scottish Aortic Stenosis and Lipid Lowering Trial:Impact on Spotty calcium, 177, 314
Regression (SALTIRE), 277 Stable angina, 781, 785, 850
Scottish Computed Tomography of the HEART Trial Stable chest pain, 204, 675–677
(SCOT-HEART) trial, 28 Stable microvascular angina, 322, 323
Scottish system, 28 Stair-step or slab registration artifacts, 57, 59
Secundum atrial septal defect, 580, 581 Standard optimal care (SOC), 339
Semi-automated/manual segmentation, 133, 134 Standard Tessellation Language (STL) file format, 861, 862
Index 929

Standard 3D posterior left atrial models, 534 Temporal coherence, 89

Stanford type B intramural hematoma, 648 Temporal resolution, 89
Static computed tomography myocardial perfusion Tesla’s self-driving cars, 893
imaging, 825 Test abnormality, 20
Static computed tomography perfusion, 792, 796 Tetralogy of Fallot (ToF), 583–585
dual-energy acquisition protocols, 779 cardiovascular CT, repaired CHD, 611, 612
flow-limiting stenosis, 779 pulmonary artery hypoplasia in, 612
optimal time of image acquisition, 779 surgical intervention for, 614
qualitative analysis of, 783 Thick multi-planar reformat (thick-MPR), 133, 136
scan protocols, 780 Thin-cap fibroatheroma, 211
optimal RR phase, 782 Thin fibrous cap atheroma (TCFA), 214–216
optimal scan delay, 782 Thoracic aortic coarctation, 606
rest and stress acquisitions, sequences of, 780 Thoracic skeleton, 654
single-shot acquisition, 779 3D laboratory, cardiovascular CT
time-attenuation-curves (TAC), 779 community needs and priorities, 161, 163
STEMI, see ST-segment elevation MI compliance, 164, 167, 168
Stent fracture, 375 CT post-processing, 161, 162
Step-and-shoot acquisition, 149 documentation and archive of results, 170
Step-and-shoot mode, 70 imaging workflow map, 169
Stewart-Hamilton equation, 427, 431 patient centricity, 168
Stress echocardiography (SE), 23, 682, 683 proper communication channels, 168, 169
Stress myocardial perfusion, 801 protocolled images/reports, 169
Stress perfusion evaluation pulmonary vein mapping protocol, 171
adenosine, 105 quality and efficiency, 163–166
dipyridamole, 105 quality assurance, 170
dobutamine, 106 referrals enhancement, 168
regadenoson, 106 research and education, 168
Stress radionuclide myocardial perfusion imaging, 396 software upgrades, 169, 170
Stress-rest approach, 780–782, 786, 788 surgical planning and guidance programs, 169, 170
Stress transluminal attenuation gradient, 763 training and quality control, 170
Stretched MPR (sMPR), 133 workflows, 161, 162
Stroke, 636 3D printing
Structural bioprosthetic valve degeneration (SVD), 482 computer-aided design (CAD), 861
ST-segment elevation MI (STEMI), 331, 774 creation and printing guide
Subarterial defects, 582 Materialise software, 862
Subarterial-type ventricular septal defect, 590 model preparation, 862
Subclinical coronary atherosclerosis, 740 STL format, 862
Subcutaneous ICD, 493 data acquisition, 859–860
Subendocardial curvilinear calcium density, 832 heart and great vessels, segmentation of
Subtotal occlusion (STO), 764 built-in import wizard, 869
Subvalvar obstruction, 583 Contrast Walls, 868
Sudden cardiac arrest (SCA), 339 Curve > Close Curve operation, 871
Summed difference score (SDS), 786 Edit Part menu, 865, 867
Summed rest score (SRS), 784 export model dialogue, 869
Summed stress score (SSS), 786 Fill Hole Freeform operation, 871
Superior sinus venosus defect, 581 heart simulation, 870
Superior vena cava (SVC), 231 Hounsfield unit thresholds, 862
Surgical aortic valve replacement (SAVR), 503 intraluminal contrast, 865, 867
Swan-Ganz catheter, 498 part creation, 865
Symptomatic patients, 739 printer-specific software solutions, 872
Systemic Coronary Risk Evaluation (SCORE) models, 259 reate Part menu, 865
Systemic venous abnormalities, 597 region growing algorithm, 865
Systemic venous baffle stenting, 609 region growing semi-automated algorithm, 862
Systemic ventricle, 417 ROI, 862
Systolic percentage wall thickening, 414 septal defect patch, 867, 869, 870, 872
smoothed contrast model, 867
smoothing operation, 866
T three-dimensional object, 871
TAC, see Time attenuation curves three-dimensional patch, 872
Takayasu arteritis (TAK), 299–301 thresholding settings and resultant ROI, 864
Takotsubo cardiomyopathy, 416 model fabrication, 861
TandemHeart, 497 segmentation of relevant anatomy, 860–861
Targeted spatial frequency filtration (TSFF), 838 workflow, 859
Taussig-Bing anomaly, 590 3 dimensional reconstructions, 38
TDG, see Transluminal Diameter Gradient 3D trans-esophageal echocardiography (TEE), 507
Teamwork, 39 3D volume rendering techniques (VRT), 37, 131–133, 472
930 Index

3-matic, 862 contrast attenuation-time curves, 749

Thrombosis in myocardial infarction (TIMI) score, 331, 753 CT-FFR, 761–763
Thrombus, 454, 455, 617 CT perfusion, 763
TiaraTM, 520 CTO, 764
Time attenuation curves (TAC), 750, 779, 792, 817 diagnostic estimates of, 758
Time-density curve (TDC), 757 distinct coronary contrast opacification gradients, 749
Tissue attenuation curve, 815, 817, 818 measurement, 751, 759
Toshiba Medical Systems, 769 predicted probability of, 749, 761
Total cholesterol (TC), 691 RAG sign, 764
Total coronary blood flow, 769 rest/stress, 763
Total Endoscopic Coronary Artery Bypass (TECAB), 384 reverse attenuation gradient sign, 763
Total iodine load (TIL), 111 reverse attenuation sign, 764
Total triglyceride (TG), 691 STO, 764
Totally anomalous pulmonary venous connection (TAPVC), 591 and TDG, 755–757
Tracer kinetic models, 819–823 techniques for, 750
Traditional filtered-backprojection, 879 Transluminal diameter gradient (TDG), 751, 755, 756
Traditional iterative image reconstruction approaches, 880 Transmitral valve implantation (TMVI), 185
Tranesophageal echocardiography (TEE), 407 Transmural perfusion ratio (TPR), 783
Transaortic valve replacement, 52, 184, 185 Transposition complexes, 607
Transcatheter aortic valve implantation (TAVI), 387, 472 Transposition of the great arteries (D-TGA), 587, 588
Transcatheter aortic valve replacement (TAVR), Transthoracic echocardiography (TTE), 451, 476
199, 206, 489, 490, 519, 701 Tribalism, 37
acquisition protocols, 505, 506 Tricuspid annular plane systolic excursion (TAPSE), 431
angiography, 511 Tricuspid annuloplasty, 491
aortic annulus Tricuspid atresia, 597
aortic root orientation, 509 Tricuspid valve, 595
area derived virtual diameters, 508 Triphasic injection, 558
BAV, 511 Triphenyltetrazolin-chloride (TTC) staining, 834
calcification burdens, 509 Triple-rule-out (TRO) CT, 334, 666
coronary ostia location, 509 Trisucpid annulus, 595
MDCT perimeter, 508 Truncal valve, 590
aortic valve cusps, 505 Truncus arteriosus, 590, 591
aortic valve replacement, 504 Tube current modulation, 14
aortic valve stenosis, 503 Tumor necrosis factor (TNF) alpha, 287
contrast media volume, 506 Turf war, 37
coronary arteries, 512, 513 Turner’s syndrome, 584
coronary occlusion, 511 TwelveTM, 520
echocardiography, 503 Twin beam dual-energy, 794
eight transcatheter aortic valve replacement, 505 Two-compartment model, 822–823
fluoroscopic projection angle, 509 2D transthoracic echocardiography (TTE), 507
LVOT, 505 “Two-step” stress test, 19
MDCT, 503, 505
post outcomes, 511
post TAVI pseudo-aneurysm, 512, 513 U
pre-procedural report, 506 Ultrafast CT, 6
radiation exposure, 506 Unbalanced atrioventricular septal defect, 599
SAVR, 503, 505 Universal scoring system, 20
sinotubular junction, 504 Unprotected coronary territories (UCT), 667
transapical approach, 503 Unqual Treatment: Confronting Racial and Ethnic
transfemoral approach, 503, 511 Disparities in Health Care, 707
valve in valve, 512 Unstable angina, 804
vascular access, 510 Upper abdominal organs, 654
ventriculoaortic junction, 504 Upslope method, 817
Transcatheter mitral valve replacement (TMVR), 519 U.S. Department of Health and Human Services
Transcatheter valves, 604, 612 Secretary’s Task Force, 707
Transesophageal echocardiography (TEE), 451, 635
Transient asymptomatic atrial-ventricular
block, 800 V
Transluminal attenuation flow encoding Valsalva maneuver, 639
(TAFE), 750, 753 Value-based health care, 39
Transluminal attenuation gradient (TAG), 791 Valvar pulmonary stenosis, 584
in animal model, 752 Valvular calcification, 642
cardiac output and vessel diameters, 757 Valvular disease, 701
CCO, 760 Valvular vegetation, 642
chronic total occlusion, 764 Vasa vasorum, 220, 221
clinical evidences of, 757–760 Vascular access complications, 849
Index 931

Vascular imaging, 93–95 Web-client based systems, 145, 146

Vascular rings, 590, 591 Wide cone cardiac axial (WCCA), 75
Vasculitis of large vessels (LVV), 300 Williams syndrome, 584
Vasculitis of medium vessels (MVV), 300 Wilson criteria, 739
Vasodilatation, 104, 105 Women's ischemia syndrome evaluation (WISE) study, 323
Vasodilative effects, 800 Workflow optimization
Vasospastic angina, 321, 322 acquisition strategy
Vegetations, 458 average heart rate, 151
Venous two-phase injection protocol, 558 cardiac rhythm, 151
Ventricular dyssynchrony, 774 CaSc, 151, 153–155
Ventricular function assessment, 610 CCTA, 151–153
Ventricular septal defect (VSD), 527, 581, 582, 861 detector collimation, 150
Ventricular tachycardia, 800 end-systole and end-diastole phases, 150
Virtual monoenergetic imaging, 94, 96 hardware based system, 150
“Virtual Non-Contrast” image, 91 helical scanning, 149
Visceral heterotaxy, 599 image reconstruction, 150
Viscosity, 110 indication-driven, 151
VNC imaging, 96 single heart beat volume, 150
Volume CT approach, 471 step-and-shoot acquisition, 149
Volume high definition (VHD), 75 breath-hold training, 156–159
Volume perfusion computed tomography (VPCT), 817, 822 CT systems, 153
Volume rendering technique (VRT), 781 ECG, 156, 157
Vulnerable plaque, 211, 224 patient positioning, 154–156
Worsening symptoms, 253

W
Wall attenuation, 414 X
Wall motion/shortening, 414 X-ray tube voltage, 5, 72, 73
WATCHMAN device, 495, 542
Watchman LAAO, 543
Watson, IBM’s AI software, 899, 900, 902 Z
Wavefront phenomenon, 839 Z-interpolation, 117, 118