An accurate assessment of abdominal aortic aneurysm (AAA) progression is essential to its clinical management. Currently, the maximum diameter of AAA at diagnosis is considered as the primary indicator of rupture risk. However, it is not optimal as rupture can happen at any size. Several patient-specific factors may also influence AAA rupture risk. Given the clinical variability in aneurysm progression, additional prognostic markers are desirable to enhance patient-specific risk stratification. Radiomics is an image processing technique that extracts quantitative and high-dimensional features from medical images. While it has emerged as a novel approach for solving diagnosis in oncology, its application in cardiovascular diseases is still limited.

This study set out with an aim to determine the feasibility of radiomics in identifying AAA with a fast growth rate (>0.3cm/year) using CT images. An automatic AAA segmentation algorithm was developed in our pipeline. Based on the radiomics features of an 84 CT dataset, supervised classification models were implemented with two feature selection algorithms and two classifiers in a machine-learning framework. An AUC of 0.80 was achieved and the predictive power was proved through comparisons to the maximum diameter and conventional risk factors. Further multivariate analysis suggested that a radiomics-based classification model could be used as an independent, yet strong predictor for fast AAA growth rate.

Background: 4D flow (4DF) magnetic resonance imaging (MRI) offers a promising way to evaluate blood flow patterns in intracranial aneurysms, although long scan times present a major limitation to broad clinical implementation. Compressed sensing (CS), an accelerated imaging technique using strategically undersampled data for data reconstruction, offers a possible solution to reduce scan times. The aim of this study was to understand the effects and limitations of varying compressed sensing acceleration factors, R, at different resolutions in in vitro 4D flow acquisitions.

Methods: This study employed a phantom depicting a saccular aneurysm. Experiment 1 evaluated the reliability of 4D flow with varying levels of compressed sensing acceleration factors (R=7.6, 12.8, and 16.6). Experiment 2 assessed the effects of varying resolutions (0.5, 1.0, 1.5, and 2.0 mm) with a compressed sensing R=12.8. Qualitative analysis included a visual assessment of velocity vectors and streamlines. Quantitative analysis compared the velocity components, peak velocity, flow rate, and wall shear stress in each experiment. All studies were post-processed using a clinically-geared software as well as with an In-House engineering pipeline, with the purpose of understanding the advantages and disadvantages of each approach and validating any results.

Results: The addition of compressed sensing reduced scan times to approximately 4-7 minutes. The In-House processing pipeline is superior to the clinical software in visualizing of velocity; visual analysis showed velocity overestimations in the R=16.8 streamlines, indicating the limits of compressed sensing to be 7.6 < R < 12.8. As expected, comparison of velocity components reflects a decrease in linear regression slopes and correlations as acceleration factors increased (m > 0.90 for all acceleration factors except R=16.6, and all r > 0.9). Experiment 2 highlights partial voluming effect at R=12.8: as resolutions decrease, velocities along the wall are less reliable than velocities furthest from the wall, which retain high slope and correlation values (both above 0.9 at 1.0 mm and 1.5 mm resolutions). High variability peak velocity, flow rate, and wall shear stress in both pipelines point to the need for a reliable way to post-process 4D flow.

Conclusion: This study showed reliable velocity data can be obtained from 4D flow studies acquired with compressed sensing lower to moderate acceleration factors at higher resolutions. With clinically-acceptable scan times, the focus now shifts towards establishing a robust and validated workflow for 4D flow studies before clinical implementation can truly be feasible.

Purpose

Methods

Results

Conclusions

Introduction: Information from MR blood flow quantification can be used to evaluate patient health, inform surgical decisions and provide boundary conditions for numerical simulations. Thus, it is important to determine the accuracy and precision of the flow measurement and its contributing factors. In this study, we investigated the reproducibility of the measurement by comparing flow results from 2D MRV, 4D MRV and CFD. Also, we investigated the sensitivity of the measurements at different positions in the scanner, since these could be affected by gradient field imperfections.

Methods: A phantom which was an exact replica of a patient intracranial aneurysm was set up with a flow system where flow was driven by a gravity-fed pressure head to provide constant flow. A contrast-enhanced MRA image was obtained at 0.5 mm isotropic resolution to determine the geometry. 4D MRV within the phantom was measured under 3 conditions: cine-MR at the isocenter (0 cm offset) for 5 time points, continuous measurement at the isocenter, and continuous measurement 10cm offset from the isocenter. CFD simulation was performed on commercial software COMSOL. Image post-processing was done using in-house Python tools. Qualitative comparisons were made using Paraview while quantitative comparisons were assessed by correlation and Bland-Altman plots.

Results: Imaging at 10 cm from isocenter was found to adequately visualize secondary flows such as jets in the aneurysm. Total flow through the inlets obtained from the 2D- and 4D- MRV acquisitions were 4.03 ± 0.07 mL/s and 3.65 ± 0.12 mL/s, which were significantly larger than the directly measured flow of 3.1 mL/s. The overall velocity field from the CFD results underestimate the measurements from 4D MRV, suggesting they provided inlet boundary conditions are too low. The greatest differences between the CFD and 4D MRV results appear at the vessel walls.

Conclusion: We developed a pipeline for evaluating flow measured under different conditions and modalities. A combination of noise and partial volume effects compromise velocity measurements, particularly in voxels at the vessel edge. Additionally, CFD can provide precise measurements at the wall, but the accuracy of the results depends highly on the image-derived geometric and flow boundary conditions.

Atherosclerotic plaque rupture at the carotid bifurcation is a major cause of stroke. While plaques vulnerable to rupture progress under significant influence from the local biochemical environment, and often experience a chronic or acute inflammatory process, mechanical forces are also of importance. Finite element studies have been conducted on models of diseased vessels to elucidate the effects of lesion characteristics on the stresses within plaque materials. Ultimately, it is hoped that patient-specific biomechanical analyses may serve as a robust clinical tool to assess the rupture potential for any particular lesion, allowing better stratification of patients into the most appropriate treatment plans.

The relationship between various mechanical descriptors such as stresses or strains and rupture vulnerability is incompletely known, however, and the patient-specific utility of biomechanical analyses is thus unclear. Progress on this front has been impeded by several distinct challenges. First, data on in vivo plaque rupture, under normal physiologic conditions, is unfortunately sparse. Second, methods for building highly realistic patient-specific finite element models of diseased vessels are lacking, and the simplifications common in the literature may render current analyses inadequate. Third, the time and computational resource demands of realistic analyses including fluid-structure interaction prohibit large-scale investigation.

In this dissertation, a strategy for accurate and efficient patient-specific modeling of the atherosclerotic carotid bifurcation is developed. We present surface-based methods for rapid yet detailed geometric discretization of image-derived vessel features, allowing for highly resolved stress calculations. A two-stage solution method is also introduced to compliment the unstructured meshes representing different tissues, allowing sizeable reductions in the time and computing resources needed for arterial fluid-structure interaction simulations.

Using these methods, we present a fluid-structure interaction analysis of a patient for whom pre- and post-rupture imaging data is available. The effects of image imprecision on the calculated stress fields are characterized to further elucidate challenges of image-based modeling. We find that plaque rupture location and extent, derived from post-rupture imaging data, correspond well to a region of elevation in first principal stress within the fibrous plaque layer of the lesion.

The pathology of many cardiovascular diseases is believed to be heavily tied to local hemodynamics. A positive feedback loop of remodeling and unhealthy flow conditions can lead to blood vessel abnormalities, such as aneurysms, stenoses, and plaques. Doctors use medical imaging techniques to visualize vascular disease within patients. However, while treatments based on these images for some diseases have been well-established, there is room for improvement. Many treatment decisions are based on simplified criteria and incomplete data and our understanding behind the pathophysiology of many of these diseases remains woefully incomplete. A stronger grasp of the relationship between disease and hemodynamics can help clinicians plan better treatment outcomes or help prevent some complications from occurring.

4D-Magnetic Resonance (MR) Imaging can be a powerful tool for quantifying the flow field within the cardiovasculature, but remains difficult to integrate for clinical use due to the involving process required to convert noisy magnetic resonance image data from a set of symbols to an easily digestible image. Treatment decisions are based on standardized practices, but no standardization for the processing and interpretation of 4D-MR images exist. Computational Fluid Dynamics (CFD) can simulate patient hemodynamics with much finer temporal and spatial resolution than 4D-MR. However, computational methods must still use boundary conditions obtained from imaging data, which means CFD results can only be as accurate as the images they are based upon. Additionally, the logistics of performing CFD, which require extensive training and time, is a strong barrier against its adoption in a clinical setting.

This thesis presents a model for using in vivo (live) and simulated flow data as complementary methods to explore potential mechanisms of pathophysiology and aid the treatment of cardiovascular diseases. We developed pipelines for processing 4D-MR images and for building patient-specific CFD simulations. We first attempted to measure the experimental error in 4D-Flow MRI within a cerebral aneurysm phantom and explored how that error might propagate into CFD results based on those MR measurements. We used those results to predict how this sensitivity would affect the treatment of aneurysms based on their hemodynamics. We then used both 4D-MR and CFD methods to characterize a rarely explored vascular territory, the cerebral venous outflow tract (CVOT). 4D-Flow MR was used to categorize geometries and flows in the CVOT. CFD was used to investigate a potential association between flow and symptoms in a subcategory of internal jugular vein geometries.

Objectives

Background

Methods

Results

Conclusions

Aim

Methods

Results

Conclusion