Positive surgical margin during radical prostatectomy: overview of sampling methods for frozen sections and techniques for the secondary resection of the neurovascular bundles

The aim of the paper is to provide an overview of intraoperative sampling methods for frozen section (FS) analysis and of surgical techniques for a secondary neurovascular bundle (NVB) resection, as the method of surgical margin (SM) sampling and the management of a positive SM (PSM) at the nerve‐sparing (NS) area are under evaluated issues. FS analysis during radical prostatectomy (RP) can help to tailor the plane of dissection based on cancer extension and thus extend the indications for NS surgery.


Objective
The aim of the paper is to provide an overview of intraoperative sampling methods for frozen section (FS) analysis and of surgical techniques for a secondary neurovascular bundle (NVB) resection, as the method of surgical margin (SM) sampling and the management of a positive SM (PSM) at the nerve-sparing (NS) area are under evaluated issues. FS analysis during radical prostatectomy (RP) can help to tailor the plane of dissection based on cancer extension and thus extend the indications for NS surgery.

Evidence Acquisition
We performed a PubMed/Medical Literature Analysis and Retrieval System Online (MEDLINE), Web of Science, Cochrane Library, and Elton B. Stephens Co. (EBSCO)host search to include articles published in the last decade, evaluating FS analysis in the NS area and surgical attempts to convert a PSM to a negative status.

Evidence Synthesis
Overall, 19 papers met our inclusion criteria. The ways to collect samples for FS analysis included: systematic (analysing the whole posterolateral aspect of the prostate specimen, i.e., neurovascular structure-adjacent frozen-section examination [NeuroSAFE]); magnetic resonance imaging (MRI)-guided (biopsies from MRI-suspicious areas, retrieved by the surgeon in a cognitive way); and random biopsies from the soft periprostatic tissues.Techniques to address a PSM in the NS area included: full resection of the spared NVB, from its caudal to cranial aspect, often including the rectolateral part of the Denonvilliers' fascia; partial resection of the NVB, in cases where sampling attempts to localise a PSM; incremental approach, meaning a partial or full resection that extends until no prostate tissue is found in the soft periprostatic environment.

Introduction
A nerve-sparing (NS) technique during RP represents the 'gold standard' to maintain erectile function in patients with low-risk prostate cancer [1]. Given the limited reliability of nomograms and statistical models in predicting extracapsular extension (ECE) of cancer [2,3], intraoperative pathological analysis with frozen section (FS) has been suggested as an alternative. The intraoperative detection of a positive surgical margin (PSM) during a NS surgery can be useful to tailor the plane of dissection and to resect more tissue from the periprostatic environment pursuing the complete removal of all neoplastic foci [4].
Actually, surgical Margin (SM) status at the area of NS is of paramount importance, as the dissection should be performed very close to the prostate at this point (intra-or interfascial plane); furthermore, this area is particularly prone to PSM, which can result from iatrogenic intraprostatic excision into an otherwise organ-confined tumour or failure to completely excise an extraprostatic prostate cancer [5]. Consequently, the knowledge of SM status of the NS zone allows for the correction of the surgical strategy and could possibly change the course of the oncological and functional outcomes of the RP.
In recent years, several techniques have been proposed to assess SM status intraoperatively and to guide a real-time diagnosis of the excised surface [4] including: the neurovascular structure-adjacent frozen-section examination (NeuroSAFE) approach [6], augmented reality [7], and fluorescence confocal microscopy [8]; however, little is known on how to manage the finding of a PSM and how to obtain a safe radical excision of the periprostatic tissue.
The aim of the present systematic review was to provide an overview of the sampling techniques for FS analysis from the neurovascular bundle (NVB) area and the way to achieve a secondary resection of persisting neoplastic tissue.

Methodology and Reporting
This systematic review was performed in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement [9]. Two independent authors (A.E. and M.C.S.) searched the Medical Literature Analysis and Retrieval System Online (MEDLINE), Web of Science, Cochrane Library, and Elton B. Stephens Co. (EBSCO)host databases using a combination of the following keywords: 'frozen section', 'frozen tissue', 'cryosection', 'cryo-section', 'prostatectomy', 'prostate cancer', 'prostatic adenocarcinoma', 'prostatic neoplasm', and 'prostate neoplasm'. All the articles describing the use of FS during RP and reporting the management of a PSM in the NVB zone were included. Moreover, we included articles with FS analysis of the NS area (posterolateral aspect of the prostate) with or without separate assessment of the base and apex, with the exclusion of papers addressing SM only at the urethral stump, at the apex and/or at the base, without considering the NS zone. In the event of any discrepancy between the two authors, a third author helped to resolve the discrepancy. The language filter was used to include only articles published in English. Furthermore, we arbitrarily included reports published since 2010, in an effort to homogenise: (i) patient series (considering the changing trend of patients with prostate cancer undergoing surgery in the last decade, with active surveillance implementation); and (ii) RP technique, since after 2010 robot-assisted laparoscopic RP (RALP) has considerably increased.

Endpoints
FS is the most commonly used approach for assessment of the SM during RP and it is based on the use of a cryostat to induce rapid freezing of the tissue specimens (À16 to À20°C for prostatic tissues) and consequently convert the water component of the tissue to ice to allow its cutting into multiple sections. Finally, the tissue is examined under a microscope after haematoxylin and eosin staining [4]. The main aims of the present review were to assess the available sampling techniques at the region of the NVB during RP, the techniques of secondary resection, and the impact of FS on oncological and functional outcomes.

Data Extraction
The following key aspects were collected from each study and stored in a dedicated database: sample size, approach of RP (open, laparoscopic, or robot-assisted), region of FS, sampling technique, secondary resection technique, PSM rate at FS, PSM at histopathology, biochemical recurrence, and functional outcomes (erectile and continence functions).

Assessment of the Risk of Bias
The primary aim of the present review was to report the different sampling techniques of FS and not to compare between patients who underwent intraoperative FS and those who did not, thus the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) tool was used to assess the risk of bias in the included studies [10]. Two of the authors (M.C.S. and A.E.) independently performed the quality assessment of the included studies and any discrepancies were resolved.

Search Results
Overall, the search identified 523 records. All records were searched to detect and exclude duplicate results (115 studies). Subsequently, the remaining 408 articles were screened using the title and abstract to exclude all the articles irrelevant to the question of the present systematic review (340 studies), leaving 68 manuscripts for full-text review. After the screening, 49 manuscripts were excluded for various reasons detailed in (Fig. 1). Finally, 19 articles met our inclusion criteria and were included in the present systematic review.

Assessment of the Risk of Bias
The QUADAS-2 tool outcomes for all the included studies are summarised in (Fig. 2). This tool is mainly used for the assessment of four territories for any potential risk of bias: patients' selection, index test, standard reference, and time and flow. The patients' selection showed high risk of bias in nearly half of the reports, as most of the authors included patients based on surgeon's preference or the suspicion of the presence of a PSM. Furthermore, the index test showed high risk of bias in a few studies, as the authors did not report the full technique of assessment or did not define PSMs. As regards the reference standard, almost all of the studies showed unclear risk of bias, as none of the authors reported Overall Excluded (49) -Review (7) Meeting abstracts (25) Lymph nodes (2) Irrelevant (2) Apical or urethral margins (4) Ongoing trial & protocol (2) No comment on site & technique (1) Letters, Editorial, and step by step (6) - EBSCO Host (20) Cochrane Library (10)

FS Sampling and NVB Secondary Resection
A total of 19 articles reporting on NS-zone sampling techniques and NVB secondary resection were published between 2010 and 2019 [6,[11][12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28]. Table 1 summarises all the studies reporting FS sampling and secondary resection of a PSM at the NVB. According to this overview, FS assessment of the NS area could be either systematic, MRI-guided or based on random biopsies. The systematic sampling of the NVB zone includes a wedge section of the specimen at the neurovascular structure adjacent prostatic tissue, from the apex to the base on both sides (Fig. 3a); each part is dissected into 10-25 blocks of 3-4 mm, for a lateral circumferential analysis of the inked prostate surface [6,[11][12][13][14]16,18,20,[23][24][25]28]. The multiparametric MRI (mpMRI)-guided sampling consists of the marking of the area(s) of suspicious ECE at mpMRI on the NVB zone; biopsies were obtained from the suspicious areas and analysed by FS [17,27]. The random approach includes biopsies from periprostatic soft tissues retrieved in cases of suspicious ECE based on an individual surgeon's decision and/or visual suspicion (Fig. 3c) [15,19,21,22,26]. Furthermore, Jeong et al. [29], demonstrated that random biopsies can be obtained from suspicious areas on palpation of the prostatic specimen (bimanual examination); however, this manuscript was not included in the present review as it was just a step-by-step description of the technique without inclusion of any patients or results. Additionally, a single author (Obek et al. [25]) reported the systematic sampling of the entire prostate with whole circumferential assessment (Istanbul preserve).
For the management of a persistent prostate cancer at the SM the techniques employed included: the full secondary resection of the NVB, its partial resection, or a partial resection with a graded incremental retrieval of additional tissue until a negative SM was obtained. A systematic full resection of the NVB was used by 10 authors [6,11,13,14,16,18,20,25,26,28]. It consists of the identification and separation of the caudal and cranial aspects of the NVBs, separated over clips or Hem-o-Lok; the NVB is then released and excised. The resection often includes the rectolateral part of the Denonvilliers' fascia. It can be performed either before the completion of the anastomosis or afterward s, as the last step of the intervention. A partial resection of the NVBs was performed when the sampling attempted to localise a PSM, traced on the corresponding NS zone [12,17,25,27]. Hatzichristodoulou et al. [12] described in detail his method for performing a partial resection, retrieving a tangential lamella 1-2 cm in length and 3-5 mm in thickness from the point suspicious for PSM contiguity. A mere incremental approach was reported mainly from a few preliminary series, where additionally resected specimens are FS examined until no prostate tissue or carcinoma was identified [15,17,21,23,24]. Finally, two authors reported that secondary resection was performed without reporting any details about its technique [19,22]. The FS time was reported by 10/19 reports and it ranged from 17-30 min for random sampling and 30-35 min for systematic and mpMRI-guided sampling techniques; however, it should be noted that the FS time for the systematic sampling of the entire prostate was prolonged to 57 min [25].

Discussion and Clinical Implications
Indications for NS in RP are continuously evolving in accordance with the European Association of Urology guidelines [1], together with the concerns regarding surgical pre-planning based on nomograms and mpMRI [2,3,30] the role of real-time control of surgical dissection has emerged; in such a scenario, the intraoperative analysis of removed tissue is the cornerstone, i.e. FS. As a PSM is an adverse outcome of RP that is associated with worse oncological endpoints [31], knowledge of the SM status during surgery helps in tailoring the next steps of RP based on the individual aggressiveness and extension of the prostate cancer.
FS analysis represents the most consolidated way to have SM status assessed intraoperatively [4]. However, the intrinsic role of FS analysis, with subsequent surgical conversion from PSM to negative SM status, has been recently argued in an editorial from Herlemann and Meng [32]. The authors stated that, especially in the setting of high-risk diseases, the significance of a PSM seems to be overcome by other features of aggressiveness, such as advanced pT or high Gleason Score; in these cases, oncological prognosis seems to be driven mainly by the aforementioned variables rather than a PSM, either converted to a negative status or not.
Given these concerns, the aim of the present review was to report the state-of-the-art of intraoperative SM assessment, addressing separately the method to attain FS and the surgical management of a PSM in the NS zone. Herein, two comments arise.
First, we found variability in the way of sampling of the NS area. A total of 12 studies [6,[11][12][13][14]16,18,20,[23][24][25]28] described a systematic wedge approach to sample the posterolateral aspect of the specimen, the same (or similar) to the one described by Schlomm et al. [6] in 2012. To reduce time, Oxley et al. [33] suggested a slice ('Mohs') section of the posterolateral aspect of the prostate (Fig. 3b); the article has not been included as it was a 'proof-of-concept' study. Two authors described a focal sampling of the NS area guided by mpMRI; suspicious zones at imaging are manually traced by the surgeon on the specimen, in a cognitive way, and analysed with FS [17,27]. Finally, a random sampling of periprostatic soft tissue including the NVB area was performed in only five series [15,19,21,22,26], where it was mainly driven by 'surgeon's choice'. Remarkably, outcomes from these last trials were the least endorsing of FS, with the authors' concluding that 'FS analysis did not dramatically change SM status' [22], 'FS cannot be used to predict the overall final margin status' [19], 'FS do not significantly affect prognosis' [15], and 'FS is not supported during radical prostatectomy' [26]. The imprecise and subjective identification of the exact soft tissue location contiguous to the PSM, together with a focal sampling, could explain the low effectiveness of FS in these datasets. This occurrence is consistent with the recent finding from Pak et al. [34] (the paper is not included in the present review as it considers only SM from the apical and basal periprostatic tissue without including the NVB area). In the paper, they failed to find benefits of FS analysis and incremental resection in high-risk patients, concluding that tumour biology is the major determinant of disease recurrence [34]. However, given the constrained value of FS analysis of limited areas, one could argue that systematic sampling of the prostate surface would lead to a higher detection of PSM and thus improve conversion to a negative SM status.
The second comment from the present review deals with the variability we found in surgical techniques to realise the complete excision of the persisting prostatic tissue. A total of 11 studies performed a full resection of the NVB ipsilateral to a PSM [6,11,[13][14][15][16]18,20,21,26,28]; four reported the attempt of a partial resection, defining with approximation of the site of persisting tumour or resecting the NVB in the area of a PSM targeted on MRI findings [12,17,25,27]. Finally, some authors mentioned an incremental approach to achieve a negative SM: the resection of periprostatic tissue was repeated until the absence of prostate cancer or benign prostatic tissue in the specimen [15,23,24]. Whether one method is better than another, still remains questionable, given the variability of sampling technique at the very basis; meanwhile, it should be stressed that the preservation of 'as many NVB structures as possible' is crucial to improve both continence and erectile function recovery [35,36].
It should be noted that systematic sampling does not allow precise tracking of the persistent cancer inside the spared bundle; this occurrence may result in an unnecessary and untailored resection of the whole NVB.
Finally, the present review raises concerns about the great heterogeneity of the FS sampling techniques and the management approaches for PSMs, which may be responsible for the different results and the absence of high-level evidence on the actual value of FS analysis of SMs during RP.
There are some limitations that should be considered: the inclusion of open, laparoscopic and robot-assisted procedures; the great heterogeneity of the included studies that limited the performance of a meta-analysis; and the high risk of bias in the included studies. Given these limits, the present review   is the first overview of the literature reporting the state-of-the art methods for FS analysis and managing a PSM finding intraoperatively.
Future research and endpoints to be pursued should be: (i) Traceability of a PSM on the corresponding spared NVB; (ii) the implementation of NS as the main endpoint, especially in patients otherwise not suitable for NVB preservation; (iii) the assessment of long-term data from both an oncological (BRFS) and functional point of view.

Conclusion
Intraoperative control of the plane of dissection could help to extend the NS approach to patients currently not eligible for NVB preservation. The present analysis highlighted the great variability and the absence of standardisation in FS sampling methods and secondary resection of the NVB, possibly limiting the oncological benefit of FS analysis throughout the reported series. Similarly, the way to convert a PSM to a negative status is still heterogeneous, with sporadic attempts to perform a partial resection to spare as much NVB tissue as possible.
Consequently, a more standardised approach to prostatic sampling for FS analysis and a more precise localisation of the PSM on the spared NVB could be regarded as future directions to improve intraoperative tailoring of surgical dissection.