A PHASE III PROSPECTIVE RANDOMIZED TRIAL TO EVALUATE THE IMPACT OF AUGMENTED REALITY DURING ROBOT-ASSISTED RADICAL PROSTATECTOMY ON THE RATES OF POSTOPERATIVE SURGICAL MARGINS

Background In contemporary series, the rate of positive surgical margins (PSMs) at final pathology after robot-assisted radical prostatectomy (RARP) spans from 6.5 to 32%. Novel technologies are urgently required for implementing preservation of neuro-vascular bundles, while simultaneously decrease the rates of PSMs. 3D virtual reconstructions of 2D cross-sectional imaging have been increasingly adopted to facilitate the surgeon in better understanding the surgical anatomy. Despite initial promising results, as of today, no prospective studies were conducted to validate these technologies in daily practice. Moreover, additional comprehensive data are required to test the use of ex-vivo fluorescence confocal microscopy (FCM) as a valid alternative to intraoperative frozen section analysis (IFS) for real-time margin evaluation. Methods This is a Phase III, monocentric, prospective and randomized trial that compares augmented reality (AR) RARP with standard approach (standard RARP) for reducing the rates of PSMs. A total of 318 patients will be randomized in a 1:1 ratio to: -AR RARP: prostate 3D reconstruction according to preoperative multiparametric magnetic resonance imaging (mpMRI). The model will be projected in the surgical field using the TileProTM technology. After prostate extraction, the virtual 3D model of the prostate will be overlapped to the gland with the used of mixed reality (Hololens glasses, Microsoft, USA) and margins will be inked according to the 3D model. Intraoperative margins will be analyzed by both IFS and FCM (Vivascope 2500M-G4); -standard RARP: nerve-sparing approach according to mpMRI and mpMRI-guided standard IFS. Primary outcome is the percentage of PSMs between the two groups, with several subgroup analyses performed according to tumor and mpMRI characteristics. Secondary outcomes are: 1) the rate of nerve-sparing approaches within the two groups; 2) 3-, 6- and 12-months recovery of erectile function (IIEF-5 score) after surgery; 3) accuracy of the Vivascope 2500M-G4 vs. IFS in detecting PSMs. We reported results of the ad-interim analysis, that was planned after recruiting 159 patients. Results Overall, 80 (51%) vs. 79 (49%) patients were treated with, respectively, standard RARP vs. AR RARP. Twenty seven (33.8%) vs. 15 (19%) patients in standard RARP vs. AR RARP groups had PSMs at final pathology (p=0.01). Subgroup analyses showed a trend for lower PSMs in the AR RARP group across all tumor and mpMRI variables, with the exception of those lesions located in the apex of the prostate. This advantage was related to higher rates of margins positivity at IFS analysis in AR RARP vs standard RARP (38 vs. 20%; p=0.01). Median (IQR) surgical operative time was 202 (181-223) vs. 209 (190-239) mins for, respectively, standard RARP vs. AR RARP (p=0.17). No differences in final right (p=0.9) or left (p=0.3) nerve sparing approaches were observed between the two groups. No differences in biochemical recurrence (BCR) or erectile function recovery over time were recorded. At ex-vivo FCM, the inter-observer agreement between pathologists ranged from moderate (k=0.74) to almost perfect (k=0.90). For a highly experienced pathologist, a good balance among sensitivity (70.5%) specificity (91.8%), PPV (80.0%) and NPV (87.1%) was reached when analysis ex-vivo FCM images. Last, the agreement between ex-vivo FCM and IFS analysis ranged from moderate (k=0.62) to strong (k=0.86). Conclusions At the time of the ad interim analysis of this Phase III prospective randomized trial, we already observed lower rates of PSMs in patients treated with AR RARP vs. standard RARP. This observed advantage applied to all patient subgroups and did not compromise oncological control of the disease nor erectile function recovery. Moreover, prostate margins evaluation at ex-vivo FCM appears to be feasible and reliable, with a good agreement between readers. Nevertheless, as of today, the performance of FCM seems to be subpar when compared to the established standard of care (IFS analysis).