E U R O P E A N U R O L O G Y F O C U S 7 ( 2 0 21 ) 7 3 3 – 741

available at www.sciencedirect.com
journal homepage: www.europeanurology.com/eufocus

Review – Prostate Cancer

State-of-the-art Intraoperative Imaging Technologies for Prostate

Margin Assessment: A Systematic Review

Judith olde Heuvel a,b,*, Berlind J. de Wit-van der Veen a, Daphne M.V. Huizing a,
Henk G. van der Poel c, Pim J. van Leeuwen c, Patrick A. Bhairosing d,
Marcel P.M. Stokkel a, Cornelis H. Slump b
a b
Department of Nuclear Medicine, Netherlands Cancer Institute—Antoni van Leeuwenhoek, Amsterdam, The Netherlands; Technical Medical Centre,
c
University of Twente, Enschede, The Netherlands; Department of Urology, Netherlands Cancer Institute-Antoni van Leeuwenhoek, Amsterdam, The
Netherlands; d Scientific Information Service, Netherlands Cancer Institute-Antoni van Leeuwenhoek, Amsterdam, The Netherlands

Article info Abstract

Context: The main challenge in radical prostatectomy is complete excision of malignant

Associate Editor: Derya Tilki tissue, while preserving continence and erectile function. Positive surgical margins
(PSMs) occur in up to 38% of cases, are associated with tumour recurrences, and may
result in debilitating additional therapies. Despite surgical developments for prostate
cancer (PCa), no technology is yet implemented to assess surgical margins of the entire
Keywords: prostatic surface intraoperatively.
Intraoperative Objective: The aim of this systematic review is to provide an overview of novel imaging
Prostate cancer methods developed for intraoperative margin assessment in PCa surgery, which are
Margin assessment compared with standard postoperative histopathology.
Evidence acquisition: A literature search of the last 10 yr was conducted in the Scopus,
Imaging technologies PubMed, and Embase (Ovid) databases. Eligible articles had to report the PSM rate
according to their intraoperative margin assessment technology in comparison with
standard histopathology.
Evidence synthesis: The search resulted in 616 original articles, of which 11 were included
for full-text review. The main technical developments in PCa margin assessment included
optical coherence tomography, photodynamic diagnosis with 5-aminolevulinic acid,
spectroscopy, and enhanced microscopy. These techniques are described and their main
advantages, limitations, and applications in the clinical setting are discussed.
Conclusions: Several imaging methods are suggested in literature for the detection of
positive margins during PCa surgery. Despite promising qualifications of the mentioned
technologies, many struggle to find implementation in the clinic. Surgical conditions
hampering the signal, long imaging times, and comparison with histopathology are
mutual challenges. The next step towards reduction of PSMs in PCa surgery includes
evaluation of these technologies in large clinical trials.
Patient summary: In this review, new technologies are reported that can assist the
surgeon by detecting insufficient removal of all tumorous tissue during surgery, instead
of the standard postoperative histopathological assessment. Currently, it is not clear
whether these technologies improve the patient outcome directly; however, the review
shows potential future implementations.
© 2020 European Association of Urology. Published by Elsevier B.V. This is an open
access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).

* Corresponding author. Department of Nuclear Medicine, Netherlands Cancer Institute—Antoni van

Leeuwenhoek, Amsterdam, The Netherlands. Tel.: +31 205127843; Fax: +31 205122290.
E-mail address: [email protected] (J. olde Heuvel).

https://doi.org/10.1016/j.euf.2020.02.004
2405-4569/© 2020 European Association of Urology. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://
creativecommons.org/licenses/by-nc-nd/4.0/).
734 E U R O P E A N U R O L O GY F O C U S 7 ( 2 0 21 ) 7 3 3 – 741

1. Introduction efficiency, sensitivity for micrometastases, and a tumour to

nontumour distinction are mandatory without suffering
The main objective of radical prostatectomy (RP) is to ensure from changes in tissue due to surgical intervention (ie,
complete tumour resection while minimising nerve, bladder, coagulation) [5,14,15]. Finally, the technology should be user
and membranous urethra damage. Successful prostate carci- friendly, and safe for both personnel and patients, and the
noma (PCa) surgery is established by both securing good specimen should be left sufficiently intact to perform stan-
therapeutic outcome and preserving sexual function and dard diagnostic histopathological examination.
continence, the so-called trifecta [1]. However, incomplete The aim of this systematic review is to identify new
tumour resections or positive surgical margins (PSMs), technologies for intraoperative tumour margin assessment
defined as tumour on ink in histopathology, are observed for PCa and, subsequently, to evaluate the performance of
in up to 38% of cases [2–4]. The risk on a PSM is higher with a these technologies compared with standard postoperative
higher T category and biopsy Gleason score (GS). A PSM histopathology in detecting PSM status.
correlates with a shorter time to progression and an increased
rate of biochemical recurrence (BCR). Subsequently, comple- 2. Evidence acquisition
mentary therapies such as androgen deprivation therapy or
radiotherapy might be necessary for these males [5–7]. The This review was registered on PROSPERO (registration num-
presence of a PSM is associated with a high preoperative ber CRD42019124616) [16]. The literature search was per-
prostate-specific antigen levels, a high GS, pathological T formed by an information specialist (P.A.B.) in Scopus,
category, and the surgeon’s experience. These parameters PubMed, and Embase (Ovid) databases, according to the
only predict the chance of a PSM preoperatively; however, Preferred Reporting Items for Systematic Review and Meta-
the surgeon is not guided during the surgery to reduce the analysis (PRISMA) statement [17]. Different associations of
number of PSMs observed after surgery. The PSM rate is the following keywords were applied in the abstract and/or
reduced when experienced surgeons perform RP; however, title: prostate cancer; prostatectomy; margin; intraopera-
it is still present in up to 17% of the cases [8]. tive; technology; imaging (full search string is provided in
In current practice, the intraoperative frozen section (IFS) the Supplementary material). Studies published between
technique is available for margin assessment of suspected 2008 and 2019 were included. If imaging modalities were
areas. After resection, the specimen is immediately frozen not implemented within 10 yr, we presume that technologies
and stained, and areas of interest are evaluated for the pres- were not suitable enough to be implemented in the clinic.
ence of cancer cells. However, being a time-consuming pro- Letters, commentaries, editorials, case reports, reviews,
cedure with low sensitivity (42%), clinical use of IFS is con- and conference abstracts were excluded, as well as non-
troversial [9,10]. Recently, an approach with improved English manuscripts. Initially, titles and abstracts were
sensitivity, called neurovascular structure-adjacent frozen- screened by two reviewers independently (J.o.H. and D.M.
section examination (NeuroSAFE), was introduced to enable V.H.) to select publications for full-text review. Disagree-
assessment of neurovascular structures adjacent to the pros- ment was resolved by consensus or with the help of an
tate [11,12]. NeuroSAFE enables nerve-sparing surgery in a independent reviewer (B.J.d.W.v.d.V.).
larger number of patients with 93.5% sensitivity and 98.8%
specificity compared with standard histopathology. The 2.1. Selection of full-text publications
study of Schlomm et al [11] showed that patients with Neu-
roSAFE evaluation had fewer postsurgical histopathologically Articles identified based on the search strategy were assessed
confirmed PSMs than patients without NeuroSAFE (15.2% vs for relevance and scientific quality. The technologies should
21.7%); however, no difference in BCR percentage was found. compare their findings with standard histopathology. The
A disadvantage of this technique is the requirement for articles have to focus on surgical margin assessment in PCa
standby qualified pathological personnel and the absence with an aim for actual intraoperative application. Studies
of the entire prostate circumference assessment. Although were excluded if the technology was not applied in a surgical
the examination time has a duration of at least 35 min, there is setting, for example, only referred to preoperative margin
no prolonged surgical time in case of lymph node dissection prediction. Articles should at least include (1) basic descrip-
following prostate removal [13]. tion of the intraoperative technology, (2) quantitative or
There are several conditions that imaging technologies qualitative description of the margin, and (3) comparison
should adhere to, in order to be used for intraoperative of margin status with histopathological examination. Addi-
assessment in clinical practice. Ideally, margins are evalu- tional publications could be added to this review based on
ated in vivo, which obliges incorporation in minimally cross-referencing. Two independent reviewers screened full
invasive surgery tools. Other requirements for margin texts for selection (J.o.H. and D.M.V.H.).
assessment tools include fast, preferably real-time, exami-
nation times (order of minutes) without complicated sam- 2.2. Risk of bias assessment
ple preparation to minimise surgical delay. Next, the entire
surface of the prostate specimen (3–5 cm diameter) should The included publications were assessed for the risk of bias
be assessed to evaluate the total margin status and distance (RoB) by three reviewers (J.o.H., D.M.V.H., and B.J.d.W.v.d.V.)
of tumour cells to the nerves. Furthermore, high detection independently, using the criteria of the Quality Assessment
 E U R O P E A N U R O L O GY F O C U S 7 ( 2 0 21 ) 7 3 3 – 741 735

of Diagnostic Accuracy Studies (QUADAS-2) tool [18]. As scored a low RoB on all seven items, whereas four publica-
proposed by the QUADAS guidelines, four items were scored tions had fewer than five items with a low RoB and appli-
on having a low, high, or unclear RoB. Patient selection, cability concerns. All studies had an unclear RoB consider-
index test (ie, new technology), and reference standard (ie, ing reference standard, since either the reference standard
histopathology) were also assessed in terms of applicability. was performed with the knowledge of the index test or no
Initial disagreement between reviewers was resolved by reference to the histopathological protocol was present. The
discussion and consensus. RoB of the index test was unclear in some studies, due to an
incomplete description of the interpretation of the index
3. Evidence synthesis test results. All included technologies were evaluated if
requirements were met according to the previously stated
3.1. General findings criteria. Technical characteristics of all technologies are
summarised in Table 2, and an overview of advantages
An overview of the selection procedure is visualised in and limitations is provided in Table 3.
Figure 1. In total, 616 records were reviewed, of which
35 full-text analyses were performed and a total of 11 arti- 3.2. Optical coherence tomography
cles fulfilled all the selection criteria. In the following, four
consecutive sections are described: optical coherence OCT emits coherent light through tissue, and the reflected
tomography (OCT; n = 1), photodynamic diagnosis (PDD) light is used to reconstruct a cross-section of the imaged
with 5-aminolevulinic acid (ALA; n = 6), spectroscopy (n = tissue. Light reflections are accordingly reconstructed in a
2), and enhanced microscopy (n = 2). two-dimensional (2D) representation of the tissue architec-
Table 1 shows the results of the RoB analysis performed ture, a so-called tomogram [19]. OCT enables real-time
with the QUADAS-2 tool of diagnostic studies. No study evaluation of the entire prostatic circumference, but

Records idenfied through Addional records idenfied

Idenficaon

database searching through other sources

(n = 774) (n = 0)

Records aer duplicates removed

(n = 616)
Records excluded (n = 581)
No original research (n = 178)
Reviews (n = 23)
Screening

No imaging technology (n = 60)

Not on paents with PCa (n = 15)
Records screened Predicon of PSM (n = 105)
(n = 616) Diagnosis of PCa (n = 80)
Postoperave BCR risk assessment (n = 57)
Not on radical prostatectomy (n = 63)

Full-text arcles assessed Full-text arcles excluded (n = 24)

Eligibility

for eligibility No margin assessment described (n = 7)

(n = 35) Mouse and phantom experiments (n = 11)
NeuroSAFE (n = 1)
No adequate histopathological comparison
(n = 5)
Studies included in
qualitave synthesis
(n = 11)
Included

Enhanced microscopy Spectroscopy Opcal coherence Photodynamic diagnosis with

(n = 2) (n = 2) tomography (n = 1) 5-aminolevulinic acid (n = 6)

Fig. 1 – Selection workflow query, according to the PRISMA 2009 flow diagram [17]. BCR = biochemical recurrence; NeuroSAFE = neurovascular
structure-adjacent frozen-section examination; PCa = Prostate carcinoma; PRISMA = Preferred Reporting Items for Systematic Review and Meta-
analysis; PSM = Positive surgical margin.
736 E U R O P E A N U R O L O GY F O C U S 7 ( 2 0 21 ) 7 3 3 – 741

Table 1 – Risk of bias assessment of the included studies obtained with the QUADAS-2 tool [18].

QUADAS-2 = Quality Assessment of Diagnostic Accuracy Studies.

White colour indicates low risk, grey unclear risk, and black high risk.

requires careful rinsing of the specimen to remove blood are characterised by different tissue properties, regarding
clots and nonprostatic tissue and manual rotation of the nuclear size and cell density [26], enabling tissue discrimi-
specimen. The analysis can be performed within 5 min with nation. Light reflectance spectroscopy (LRS) measures the
a commercially available CE-marked portable machine. OCT intensity and spectrum of reflected or back-scattered light.
is able to assess extraprostatic extension and seminal vesi- Using LRS, 17 specimens were analysed in the study of
cle invasion, in addition to margin evaluations [19]. Using Morgan et al [27], and PSMs were observed with 86%
OCT, Dangle et al [19] observed 21 out of 100 evaluated sensitivity and 85% specificity. Lay et al [26] performed
specimens with PSMs, and a comparison with standard LRS measurements on 50 specimens, of which 197 areas
histopathology showed seven true positive and 14 false were analysed. LRS was able to detect PSMs in tumours with
positive measurements (sensitivity 70% and specificity GS  7 with 91% sensitivity and 93% specificity, in contrast
84%). to 65% and 88%, respectively, in specimens with GS 6.

3.3. PDD with 5-ALA 3.5. Confocal laser endomicroscopy

A natural amino acid 5-ALA is converted to photoactive Confocal laser endomicroscopy (CLE) is an endoscopic
protoporphyrin IX (PpIX) within the cell. This amino acid imaging tool based on standard confocal microscopy. The
shows higher accumulation in tumorous tissue compared light beam penetrates at a specific depth at a certain time;
with healthy tissue and is orally administered 3–4 h prior from the multiple 2D images obtained, a 3D reconstruction
to surgery for adequate uptake [20–22]. PpIX emits red of the specimen can be created. In the study of Lopez et al
fluorescence light after excitation with blue light, which [28], a 488 nm laser was used for prostate imaging, in
can be detected using a commercially available laparo- concurrence with a Food and Drug Administration–
scopic compatible photodynamic camera. The amino acid approved fluorophore (sodium fluorescein) intravenously
5-ALA was evaluated in different small sample size studies injected 20 min before CLE. Sensitivity and specificity were
(six to 52 patients) and compared with histopathology not mentioned in the study, and no PSM was found.
[20–25]. These studies showed sensitivities ranging from
75% to 82% and specificities between 68% and 88%. The 3.6. Structured illumination microscopy
light source can be switched from white to blue light
during surgery to excite PpIX, thereby not delaying the Structured illumination microscopy (SIM) is an optical sec-
operating time [23]. Next, the 5-ALA PDD camera has the tioning technique using wide field illumination. In the study
same view as the robotic camera, enabling real-time cancer by Wang et al [29], a video-rate SIM (VR-SIM) is used to
location visualisation in vivo. record images of the entire circumference in 24 patients.
Each tissue frame is stitched together, to get a circumferen-
3.4. Light reflectance spectroscopy tial image of the entire prostate specimen. VR-SIM was able
to detect a PSM in three of the four histologically confirmed
Spectroscopy visualises the interaction between matter and cases, which had a size of >1 mm. In one case, VR-SIM
electromagnetic radiation. Healthy and tumorous tissues detected a PSM that was not confirmed by histopathology
Table 2 – Characteristics of the included articles.

Technology Safe to use No. of In vivo/ T category PSM (%) Distinguishes signal NVB Min. FOV
and author patients ex vivo (% of patients) based on assessment invasive

OCT Dangle[19] Some heat generation 100 Ex vivo T2 85% 10 Microstructure Yes Future 2.7 mm
T3 15%
5-ALA PDD Ganzer[20] Side effects, FDA approved 24 In vivo/ex vivo T2 37.5% 33 Fluorescence of PpIX Possible Yes Camera view
T3a 33%
T3b 25%
Adam[21] Side effects, FDA approved 39 In vivo T2 42% 33 Fluorescence of PpIX Possible Yes Camera view
T3 58%
Inoue[22] Side effects, FDA approved 6 In vivo T1c 50% 0 Fluorescence of PpIX Possible Yes Camera view
T2a 17%
T2b 17%
T2c 17%
Fukuhara [23] Side effects, FDA approved 16 In vivo/ex vivo T1c 68.8% 0 Fluorescence of PpIX Possible Yes Camera view
T2a 25%

E U R O P E A N U R O L O GY F O C U S 7 ( 2 0 21 ) 7 3 3 – 741
T3a 6.25%
Fukuhara[24] Side effects, FDA approved 52 In vivo T1c 65% 2 Fluorescence of PpIX Possible Yes Camera view
T2a 17%
T2b 13%
T3a 4%
Zaak[25] Side effects, FDA approved 16 In vivo GS 2–4: 10% 13 Fluorescence of PpIX Possible Yes Camera view
5–7: 80%
8–10: 10%
LRS Lay[26] Yes 50 Ex vivo Gleason 7 88% 58 Cell density & nucleus size Not yet Future 1 mm
T3 45%
Morgan [27] Yes 17 Ex vivo Intermediate to 59 Cell density & nucleus size Future Future 1 mm
high-grade Gleason 7
CLE Lopez[28] Yes 21 In vivo T2a,b,c NS Microscopic changes Yes Yes 2.6 mm
T3a
SIM Wang[29] Yes 24 Ex vivo T2a 5% 17 Microscopic changes Yes No 1.3 mm
T2c 53%
T3a 32%
T3b 10%
Technology Commercially Resolution Sensitivity (%) Specificity (%) Time Compromising conditions Depth penetration
and author available

OCT Dangle[19] Yes 10–20 mm 70 84 1.5 s per image, <5 min No perfusion blood clots 1–2 mm
5-ALA PDD Ganzer[20] Yes NS 75 88 Real time Heat and blood NS
Adam[21] Yes NS 75 88 Real time Heat and blood NS
Inoue[22] Yes NS NS NS Real time Heat and blood NS
Fukuhara[23] Yes NS 82 68 Real time Heat and blood NS
Fukuhara[24] Yes NS 75 87 Real time Heat and blood NS
Zaak[25] Yes NS 50a 100a Real time Heat and blood NS
LRS Lay[26] No NS 91 93 No real-time data processing Blood and inflammation 2 mm
Morgan[27] No NS 86 85 NS Perfusion 2 mm
CLE Lopez[28] Yes 1 mm NS NS 10 min Blood and debris 60 mm
SIM Wang[29] No 1.3 mm 75a 93.8a 1h NS Cell layer

5-ALA = 5-aminolevulinic acid; CLE = confocal laser endomicroscopy; FDA = Food and Drug Administration; FOV = field of view; GS = Gleason score; LRS = light reflectance spectroscopy; NS = not specified; NVB = neural
vascular bundle assessment; OCT = optical coherence tomography; PDD = photodynamic diagnosis; PpIX = photoactive protoporphyrin IX; PSM = positive surgical margin; SIM = structured illumination microscopy.
a
Sensitivity and specificity were not mentioned in the article, but calculated from the obtained numbers. The PSM rate was based on histopathological findings. Safe to use indicates if there were any issues regarding the

737
patient or personnel safety that should be taken into consideration.
738 E U R O P E A N U R O L O GY F O C U S 7 ( 2 0 21 ) 7 3 3 – 741

Table 3 – Main advantages and limitations of the included technologies.

Technology Advantages Limitations

OCT [19]

- NVB assessment - Learning curve for interpretation

- Fast acquisition - Hampered by the presence of blood and nonprostatic tissue
- Small FoV

5-ALA PDD [20–25]

- Applicable in vivo using a light switch - Learning curve for interpretation

- Real time - Compromised by heat
- Preparation time
- Some side effects

LRS [26,27]

- Unambiguous result - Small FoV

- Surgical cavity assessment - No real-time data processing
- Minimisation of impairment of the signal due to
absorption by blood
- Real-time data acquisition

CLE [28]

- NVB assessment - Hampered by the presence of blood

- In vivo usage - Preparation time
- Fast acquisition - Lower signal intensity due to resolution
- 3D reconstruction of the specimen
- Micron scale resolution

SIM [29]

- NVB assessment circumferential image of the entire - Learning curve for interpretation
prostate specimen - Manual rotation to scan the circumference

5-ALA = 5-aminolevulinic acid; CLE = confocal laser endomicroscopy; 3D = three dimensional; FoV = field of view; LRS = light reflectance spectroscopy; NVB =
neural vascular bundle assessment; OCT = optical coherence tomography; PDD = photodynamic diagnosis; SIM = structured illumination microscopy.

and one PSM was missed by VR-SIM. Sensitivity and speci- the general lack of studies with a methodology to answer
ficity were not specified in the article; however, calculation our research question. Since most studies were still in the
using the published data resulted in 75% sensitivity and 94% feasibility phase, they failed to compare their results prop-
specificity. erly with standard histopathology. Second, several studies
performed their reference test (ie, histopathology) not
4. Discussion blinded for the results of the index test (ie, new technology).
This may result in a RoB regarding the sensitivity and
Intraoperative margin assessment contributes to reduction specificity, since it is possible that assessors alter histopa-
of PSMs in PCa surgery. This systematic review provides an thology results based on the technology’s suggestions. On
overview of current possibilities for intraoperative margin the contrary, marking of suspicious locations based on the
assessment, including five different technologies. All technology results enabled a direct comparison with histo-
included technologies are based on optical imaging of cel- pathology in that specific location.
lular differences between cancerous and normal tissue. Previous reviews on technologies for surgical margin
Technologies such as OCT and LRS seem to be promising, assessment in other tumour types predicted, in 2014, an
although these still face drawbacks before clinical imple- increase in the clinical use of these technologies [30–33].
mentation is possible. So far, none of these techniques affect Currently, only NeuroSAFE is used for PCa in daily practice in
clinical decisions as they are used only in research specific institutions, and one could ask why these technol-
situations. ogies are not incorporated on a more frequent basis. This
A RoB assessment of all selected studies was performed might be directly related to the strict histopathological
with the QUADAS-2 tool. A few points of consideration were definition of PSMs in PCa, which is “tumour on ink”. Hence,
extracted from this assessment. The first observation was to be as precise as histopathology, a technology should
 E U R O P E A N U R O L O GY F O C U S 7 ( 2 0 21 ) 7 3 3 – 741 739

assess margins with a depth resolution of one cell layer, intraoperative time constraints and to improve upon Neu-
which is challenging for the described intraoperative imag- roSAFE [11–13]. With future technical developments, the
ing technologies. In general, the sensitivity and specificity of aforementioned technologies should be able to reduce
the included technologies are inferior to the 93.5% sensitiv- assessment times. Preparation time needs to be considered
ity and 98.8% specificity of NeuroSAFE. However, the tech- when using fluorescent dyes with 5-ALA and CLE, as imag-
nologies included in this review are evaluated in limited ing should be performed 3–4 h and within 20 min after
sample sizes (range 6–100 patients) and are still under administration, respectively [20–22,28]. Additionally, some
development. Next to that, with the introduction of new side effects of 5-ALA have been reported [21], resulting in
techniques, a learning curve is common and could result in several contraindications for the oral use and restriction
suboptimal detection rates. Still, the results were promising from sunlight after surgery to avoid skin reactions
with specificities above 84% (except for those of Fukuhara [20,23,24].
et al [23]) and sensitivity ranging from 50% to 91%. If a PSM is detected ex vivo by an optical technology, one
Real-time tumour visualisation during PCa surgery may of the problems is to map the PSM back to the surgical cavity
improve patient outcome and survival by assisting towards to resect additional tissue. This mapping can be difficult due
a more radical resection, while preserving vital normal to changes in the surgical field, thus a limiting factor for ex
tissues. The trifecta of surgical outcome includes cancer vivo imaging. Ideally, the technology would assess margins
control, sexual function, and presence of continence. The within the surgical cavity, for example, using LRS that has
first two can be evaluated if intraoperative margin assess- the ability to measure the surgical cavity besides the excised
ment is combined with nerve-sparing surgery, for example, specimen [26,27]. Based on all previous mentioned advan-
using the OCT, CLE, and SIM techniques to assess the neural tages and drawbacks, currently no technique is optimal for
vascular bundle. Furthermore, easy incorporation in clinical intraoperative PSM detection in PCa. Therefore, the search
routine requires a user-friendly technology and unambigu- for alternative technologies persists, and is likely to end up
ous interpretation of the results. A learning curve was in the use of fluorescent or radioactive markers to enhance
required to distinguish OCT signals between tumour and signal intensities.
normal tissue due to variations in prostate anatomy. This
applies for 5-ALA as well, where training is required for the 4.1. Alternative and future technologies
subjective interpretation of 5-ALA PDD [20,21]. In contrast,
since spectroscopy technologies are based on computer- This review focused on the clinical applicability of intraop-
based algorithms, the conclusion will not primarily depend erative margin assessment, and several innovative technol-
on user interpretation [26,27]. ogies are developed for optical guidance during PCa surgery.
Most included studies, except for the 5-ALA PDD and CLE Another strategy to decrease positive margins includes
studies, are currently only performed ex vivo; yet all hold surgical experience, preoperative selection, preoperative
potential for in vivo usability. CLE images were acquired in models, and radioguided surgery [34–36]. Investigations
vivo, though image analysis and interpretation were per- in the latter area are already on-going, as well as in combi-
formed afterwards [28]. A general obstacle for in vivo usage nation with augmented and virtual reality [37,38]. Promising
is the influence of the surgical conditions on image acqui- other techniques that are still in a preclinical development
sition. For example, the 5-ALA signal is compromised by stage include fluorescence coupled to tumour-targeted
heat [21,23,24]. Thus, the use of diathermy should be probes, such as the prostate-specific membrane antigen
avoided in critical areas. To overcome this effect, the speci- (PSMA). Tumour-targeting ligands to near-infrared (NIR)
men can be prepared using a cold knife without electric fluorophores are already studied in (pre)clinical trials in
devices [23]. Besides that, the signal of multiple technolo- other cancer types. PSMA is often overexpressed by PCa
gies (5-ALA, OCT, and CLE) is hampered by the presence of cells, and the ligand can be bound to an infrared fluorescent
blood. Hence, careful rinsing of the prostate is required, dye or Cy5 dye. This dye has absorption and emission
complicating actual in vivo usage. However, research into wavelengths in the NIR and far red ranges, enabling fluo-
different wavelengths that do not overlap with the heme rescence imaging using a fluorophore [30,39,40]. PSMA can
peak may overcome this problem. This could be performed, also be coupled to 68Gallium, enabling Cerenkov lumines-
for example, using a specific wavelength range for LRS cence imaging during surgery using positron emitting prop-
(700–850 nm), which minimises impairment of the signal erties of 68Gallium. This technique has shown promising
due to absorption by blood [26,27]. Finally, OCT images can results in a preclinical setting [41].
be altered by nonprostatic tissue through tissue interfer-
ence and long periods without perfusion [19]. 5. Conclusions
A shared disadvantage of all described techniques
includes long assessment times, due to either long-lasting In conclusion, several technologies are suggested to over-
data processing or long acquisition times to assess the come the problem of postsurgery PSMs in PCa. Despite
entire prostate with a small field of view (FoV). Spectros- promising specifications of the technologies mentioned,
copy, OCT, and microscopic techniques all have an FoV of <5 many struggle to find implementation in the clinic. Surgical
mm; therefore, scanning the prostate circumference can conditions hampering the signal, long acquisition times,
take up to 1 h. Full assessment of the prostate using a and accurate comparison with histopathology are mutual
technology should be within 35 min to compete with challenges. Furthermore, large clinical trials are needed to
740 E U R O P E A N U R O L O GY F O C U S 7 ( 2 0 21 ) 7 3 3 – 741

investigate the added value of each technology in terms of [6] Silberstein J, Eastham J. Significance and management of positive
improved patient outcome and cost effectiveness, before surgical margins at the time of radical prostatectomy. Indian J Urol
incorporating margin assessment into clinical practice. 2014;30:423.
[7] Bolla M, van Poppel H, Tombal B, et al. Postoperative radiotherapy
Finally, improvements of the techniques are required to
after radical prostatectomy for high-risk prostate cancer: long-term
enhance implementation of intraoperative assessment of
results of a randomised controlled trial (EORTC trial 22911). Lancet
surgical margins.
2012;380:2018–27.
[8] Thompson JE, Egger S, Böhm M, et al. Superior biochemical recur-
Author contributions: Judith olde Heuvel had full access to all the data in rence and long-term quality-of-life outcomes are achievable with
the study and takes responsibility for the integrity of the data and the robotic radical prostatectomy after a long learning curve—updated
accuracy of the data analysis. analysis of a prospective single-surgeon cohort of 2206 consecutive
Study concept and design: olde Heuvel, Bhairosing, de Wit-van der Veen. cases. Eur Urol 2018;73:664–71.
Acquisition of data: olde Heuvel, Bhairosing, de Wit-van der Veen, [9] Goharderakhshan RZ, Sudilovsky D, Carroll LA, Grossfeld GD, Marn
Huizing. R, Carroll PR. Utility of intraoperative frozen section analysis of
Analysis and interpretation of data: olde Heuvel, de Wit-van der Veen, surgical margins in region of neurovascular bundles at radical
Huizing, Slump, van der Poel, van Leeuwen. prostatectomy. Urology 2002;59:709–14.
Drafting of the manuscript: olde Heuvel, de Wit-van der Veen, Huizing. [10] Tsuboi T, Ohori M, Kuroiwa K, et al. Is intraoperative frozen section
Critical revision of the manuscript for important intellectual content: olde analysis an efficient way to reduce positive surgical margins?
Heuvel, de Wit-van der Veen, Huizing, Bhairosing, van der Poel, van Urology 2005;66:1287–91.
Leeuwen, Stokkel, Slump. [11] Schlomm T, Tennstedt P, Huxhold C, et al. Neurovascular structure-
Statistical analysis: None. adjacent frozen-section examination (NeuroSAFE) increases nerve-
Obtaining funding: Stokkel, Slump, de Wit-van der Veen. sparing frequency and reduces positive surgical margins in open
Administrative, technical, or material support: None. and robot-assisted laparoscopic radical prostatectomy: experience
Supervision: Stokkel, Slump, de Wit-van der Veen. after 11,069 consecutive patients. Eur Urol 2012;62:333–40.
Other: None. [12] Mirmilstein G, Rai BP, Gbolahan O, et al. The neurovascular struc-
ture-adjacent frozen-section examination (NeuroSAFE) approach to
nerve sparing in robot-assisted laparoscopic radical prostatectomy
Financial disclosures: Judith olde Heuvel certifies that all conflicts of
in a British setting—a prospective observational comparative study.
interest, including specific financial interests and relationships and
BJU Int 2018;121:854–62.
affiliations relevant to the subject matter or materials discussed in the
[13] Beyer B, Schlomm T, Tennstedt P, et al. A feasible and time-efficient
manuscript (eg, employment/affiliation, grants or funding, consultan-
adaptation of NeuroSAFE for da Vinci robot-assisted radical pros-
cies, honoraria, stock ownership or options, expert testimony, royalties,
tatectomy. Eur Urol 2014;66:138–44.
or patents filed, received, or pending), are the following: None.
[14] Ackerman DA, Barry JM, Wicklund RA, Olson N, Lowe BA. Analysis of
risk factors associated with prostate cancer extension to the surgi-
Funding/Support and role of the sponsor: This research is supported by cal margin and pelvic node metastasis at radical prostatectomy. J
KWF Kankerbestrijding and Technology Foundation STW Grant number Urol 1993;150:1845–50.
15175. [15] Wieder JA, Soloway MS. Incidence, etiology, location, prevention
and treatment of positive surgical margins after radical prostatec-
tomy for prostate cancer. J Urol 1998;160:299–315.
Appendix A. Supplementary data
[16] olde Heuvel J, Huizing DM, de Wit-van der Veen BJ, et al. State-of-
the-art intra-operative imaging technologies for prostate margin
Supplementary material related to this article can be
assessment: a systematic reviewPROSPERO. 2019 CRD42019124616
found, in the online version, at doi:https://doi.org/10. 2019 http://www.crd.york.ac.uk/PROSPERO/display_record.php?
1016/j.euf.2020.02.004. ID=CRD42019124616
[17] Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items
for systematic reviews and meta-analyses: the PRISMA statement.
References
Ann Intern Med 2009;151:264–9.
[1] Bianco FJ, Scardino PT, Eastham JA. Radical prostatectomy: long- [18] Whiting PF. QUADAS-2: A revised tool for the quality assessment of
term cancer control and recovery of sexual and urinary function diagnostic accuracy studies. Ann Intern Med 2011;155:529.
(“trifecta”). Urology 2005;66:83–94. [19] Dangle PP, Shah KK, Kaffenberger B, Patel VR. The use of high
[2] Samaratunga H, Montironi R, True L, et al. International Society of resolution optical coherence tomography to evaluate robotic radical
Urological Pathology (ISUP) Consensus Conference on handling and prostatectomy specimens. Int Braz J Urol 2009;35:344–53.
staging of radical prostatectomy specimens. Working group 1: [20] Ganzer R, Blana A, Denzinger S, et al. Intraoperative photodynamic
specimen handling. Mod Pathol 2011;24:6–15. evaluation of surgical margins during endoscopic extraperitoneal
[3] Izard JP, True LD, May P, et al. Prostate cancer that is within 0.1 mm radical prostatectomy with the use of 5-aminolevulinic acid. J
of the surgical margin of a radical prostatectomy predicts greater Endourol 2009;23:1387–94.
likelihood of recurrence. Am J Surg Pathol 2014;38:333–8. [21] Adam C, Salomon G, Walther S, et al. Photodynamic diagnosis using
[4] Stephenson AJ, Wood DP, Kattan MW, et al. Location, extent and 5-aminolevulinic acid for the detection of positive surgical margins
number of positive surgical margins do not improve accuracy of during radical prostatectomy in patients with carcinoma of the
predicting prostate cancer recurrence after radical prostatectomy. J prostate: a multicentre, prospective, phase 2 trial of a diagnostic
Urol 2009;182:1357–63. procedure. Eur Urol 2009;55:1281–8.
[5] Ohori M, Wheeler TM, Kattan MW, Goto Y, Scardino PT. Prognostic [22] Inoue K, Ashida S, Fukuhara H, et al. Application of 5-aminolevulinic
significance of positive surgical margins in radical prostatectomy acid-mediated photodynamic diagnosis to robot-assisted laparo-
specimens. J Urol 1995;154:1818–24. scopic radical prostatectomy. Urology 2013;82:1175–8.
 E U R O P E A N U R O L O GY F O C U S 7 ( 2 0 21 ) 7 3 3 – 741 741

[23] Fukuhara H, Inoue K, Satake H, et al. Photodynamic diagnosis of [33] Hsu M, Gupta M, Su L-M, Liao JC. Intraoperative optical imaging and
positive margin during radical prostatectomy: preliminary experi- tissue interrogation during urologic surgery. Curr Opin Urol
ence with 5-aminolevulinic acid. Int J Urol 2011;18:585–91. 2014;345:455–9.
[24] Fukuhara H, Inoue K, Kurabayashi A, Furihata M, Shuin T. Perfor- [34] Mohareri O, Ischia J, Black PC, et al. Intraoperative registered
mance of 5-aminolevulinic-acid-based photodynamic diagnosis for transrectal ultrasound guidance for robot-assisted laparoscopic
radical prostatectomy. BMC Urol 2015;15:78. radical prostatectomy. J Urol 2015;193:302–12.
[25] Zaak D, Sroka R, Khoder W, et al. Photodynamic diagnosis of [35] Ukimura O, Aron M, Nakamoto M, et al. Three-dimensional surgical
prostate cancer using 5-aminolevulinic acid–first clinical experi- navigation model with TilePro display during robot-assisted radical
ences. Urology 2008;72:345–8. prostatectomy. J Endourol 2014;28:625–30.
[26] Lay AH, Wang X, Morgan MS, et al. Detecting positive surgical [36] Shoji S, Aron M, de Castro Abreu AL, et al. Intraoperative ultraso-
margins: utilisation of light-reflectance spectroscopy on ex vivo nography with a surgeon-manipulated microtransducer during
prostate specimens. BJU Int 2016;118:885–9. robotic radical prostatectomy. Int J Urol 2014;21:736–9.
[27] Morgan MSC, Lay AH, Wang X, et al. Light reflectance spectroscopy [37] Porpiglia F, Fiori C, Checcucci E, Amparore D, Bertolo R. Augmented
to detect positive surgical margins on prostate cancer specimens. J reality robot-assisted radical prostatectomy: preliminary experi-
Urol 2016;195:479–84. ence. Urology 2018;115:184.
[28] Lopez A, Zlatev DV, Mach KE, et al. Intraoperative optical biopsy [38] Mehralivand S, Kolagunda A, Hammerich K, et al. A multiparametric
during robotic assisted radical prostatectomy using confocal endo- magnetic resonance imaging-based virtual reality surgical naviga-
microscopy. J Urol 2016;195:1110–7. tion tool for robotic-assisted radical prostatectomy. Turk J Urol
[29] Wang M, Tulman DB, Sholl AB, et al. Gigapixel surface imaging of 2019;45:357–65.
radical prostatectomy specimens for comprehensive detection of [39] Tringale KR, Pang J, Nguyen QT. Image-guided surgery in cancer: a
cancer-positive surgical margins using structured illumination strategy to reduce incidence of positive surgical margins. Wiley
microscopy. Sci Rep 2016;6:27419. Interdiscip Rev Syst Biol Med 2018;10:1–18.
[30] Visgauss JD, Eward WC, Brigman BE. Innovations in intraoperative [40] Kularatne SA, Thomas M, Myers CH, et al. Evaluation of novel
tumor visualization. Orthop Clin North Am 2016;47:253–64. prostate-specific membrane antigen-targeted near-infrared imag-
[31] Keereweer S, Kerrebijn JDF, van Driel PBAA, et al. Optical image- ing agent for fluorescence-guided surgery for prostate cancer. Clin
guided surgery—where do we stand? Mol Imaging Biol Cancer Res 2019;25:177–87.
2011;13:199–207. [41] olde Heuvel J, de Wit-van der Veen BJ, Vyas KN, et al. Performance
[32] De Boer E, Harlaar NJ, Taruttis A, et al. Optical innovations in evaluation of Cerenkov luminescence imaging: a comparison of
68 18
surgery. Br J Surg 2015;102:56–72. Ga with F. EJNMMI Phys 2019;6:17.