You have accessJournal of UrologyProstate Cancer (V14)1 Apr 2020V14-02 COMPUTER VISION ALGORITHM ALLOWS TO PERFORM 3D AUTOMATIC AUGMENTED-REALITY ROBOT-ASSISTED RADICAL PROSTATECTOMY Francesco Porpiglia*, Enrico Checcucci, Daniele Amparore, Alberto Piana, Federico Piramide, Gabriele Volpi, Sabrina De Cillis, Matteo Manfredi, Pietro Piazzolla, Cristian Fiori, and Enrico Vezzetti Francesco Porpiglia*Francesco Porpiglia* More articles by this author , Enrico CheccucciEnrico Checcucci More articles by this author , Daniele AmparoreDaniele Amparore More articles by this author , Alberto PianaAlberto Piana More articles by this author , Federico PiramideFederico Piramide More articles by this author , Gabriele VolpiGabriele Volpi More articles by this author , Sabrina De CillisSabrina De Cillis More articles by this author , Matteo ManfrediMatteo Manfredi More articles by this author , Pietro PiazzollaPietro Piazzolla More articles by this author , Cristian FioriCristian Fiori More articles by this author , and Enrico VezzettiEnrico Vezzetti More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000000982.02AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: In this study we want to evaluate the accuracy of the cTraker, our new Automatic Augmented Reality (AR) system, to identify the tumour extracapsular extension (ECE) at the level of preserved neurovascular bundles (NVBs) during robot assisted radical prostatectomy (RARP). METHODS: From June to October 2019, 10 patients candidated to RARP with suspicious ECE at preoperative high-resolution multi-parametric magnetic resonance imaging (mpMRI, 1-mm slices) were enrolled according to dedicated protocol. The three-dimensional (3D) reconstruction was overlapped to endoscopic in-vivo anatomy and sent back to DaVinci robotic console by using Tile-Pro performing AR-RARP. In the specific case of this study, a new AR software based on computer vision algorithm was developed. This software was able to automatically identify the catheter at the level of prostatic lodge at the end of extirpative phase overlapping automatically the 3D virtual images. We placed a metallic clip at the level of suspected ECE, as indicated by the virtual images. Finally, according to the presurgical indications, the entire NVBs with suspicious ECE were removed for pathological examination. We calculated Cohen kappa coefficient (k) to define agreement between preoperative suspected ECE and pT3 stage at final pathological examination. RESULTS: The presence of ECE (κ=0.68) was confirmed in 8 cases at the final pathology (pT3). The presence of ECE at the level of metallic clip was recorded in 100% of the cases at the macroscopic assessment; then the microscopic evaluation confirmed the presence of cancer in all the cases and revealed a mean length of ECE of 4 mm. CONCLUSIONS: Our results suggest that the new evolution of our AR platform based on computer vision algorithm allows an effective Automatic AR RARP. The 3D virtual images, automatically anchored to the catheter, are able to correctly identify the location of ECE at the level of NVBs. Source of Funding: None. © 2020 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 203Issue Supplement 4April 2020Page: e1306-e1306 Advertisement Copyright & Permissions© 2020 by American Urological Association Education and Research, Inc.MetricsAuthor Information Francesco Porpiglia* More articles by this author Enrico Checcucci More articles by this author Daniele Amparore More articles by this author Alberto Piana More articles by this author Federico Piramide More articles by this author Gabriele Volpi More articles by this author Sabrina De Cillis More articles by this author Matteo Manfredi More articles by this author Pietro Piazzolla More articles by this author Cristian Fiori More articles by this author Enrico Vezzetti More articles by this author Expand All Advertisement PDF downloadLoading ...