PD56-06 AUTOMATED DETECTION OF PROSTATE CANCER LESIONS BASED ON TARGETED BIOPSY ON MULTIPARAMETRIC MAGNETIC RESONANCE IMAGING USING DEEP LEARNING

Sherif Mehralivand,Dong Yang,Daguang Xu,Ziyue Xu,Holger Roth,Stephanie Harmon,Nathan Lay,Maria Merino,Bradford Wood,Peter Choyke,Peter Pinto,Baris Turkbey

DOI: https://doi.org/10.1097/ju.0000000000002090.06

2021-01-01

Abstract:You have accessJournal of UrologyProstate Cancer: Detection & Screening VI (PD56)1 Sep 2021PD56-06 AUTOMATED DETECTION OF PROSTATE CANCER LESIONS BASED ON TARGETED BIOPSY ON MULTIPARAMETRIC MAGNETIC RESONANCE IMAGING USING DEEP LEARNING Sherif Mehralivand, Dong Yang, Daguang Xu, Ziyue Xu, Holger Roth, Stephanie Harmon, Nathan Lay, Maria Merino, Bradford Wood, Peter Choyke, Peter Pinto, and Baris Turkbey Sherif MehralivandSherif Mehralivand More articles by this author , Dong YangDong Yang More articles by this author , Daguang XuDaguang Xu More articles by this author , Ziyue XuZiyue Xu More articles by this author , Holger RothHolger Roth More articles by this author , Stephanie HarmonStephanie Harmon More articles by this author , Nathan LayNathan Lay More articles by this author , Maria MerinoMaria Merino More articles by this author , Bradford WoodBradford Wood More articles by this author , Peter ChoykePeter Choyke More articles by this author , Peter PintoPeter Pinto More articles by this author , and Baris TurkbeyBaris Turkbey More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000002090.06AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: Targeted biopsy of prostate lesions on multiparametric magnetic resonance imaging (MRI) improves the detection of clinically significant prostate cancer. However, access to highly specialized genitourinary radiologists is still limited outside of larger academic centers. With the advent of Deep Learning (DL), more and more automated algorithms achieve human or superhuman performance for medical imaging tasks. These tools can assist less-experienced readers to achieve expert-level performance. In this study, we sought to develop and test a DL-based automated detection tool for prostate cancer lesions using targeted-biopsy confirmed prostate cancer histopathology as training labels. METHODS: A total of 525 patients were included in this study. All lesions which histopathologically proven prostate cancer after MRI-guided targeted biopsy were contoured by a single genitourinary radiologist with > 14 years of experience and > 1500 reads per year leading to a total of 864 lesions. The algorithm is based on two cascaded 3D U-Net structures for 1. Segmentation of the whole prostate and 2. Detection and segmentation of prostate cancer lesions. Detection performance was assessed by calculating lesion-based sensitivity, positive predictive value and false-discovery rate. RESULTS: The 525 patient scans were subdivided into a training (n = 368), validation (n = 79) and test set (n = 78). Within the test set, lesion-based sensitivity was 49/75 (65.3%) and positive predictive value was 49/91 (53.8%). This came at the cost of a false discovery rate of 42/91 (46.2%). Figure 1 demonstrates a patient with a prostate lesion in the left base peripheral zone which was biopsy-confirmed as Gleason grade group 5 on targeted biopsy. The lesion was correctly detected and segmented by the algorithm but a benign prostatic hyperplasia lesion on the opposite site was also detected constituting a false positive lesion call. CONCLUSIONS: We developed a DL-based automated detection algorithm for biopsy-confirmed prostate cancer lesions on multiparametric MRI. The system needs to be compared against human performance in a prospective multi-reader study to determine its additional value to human-only reading. Source of Funding: This project is based on a Cooperative Research and Development Agreement (CRADA) between the National Institutes of Health and the NVIDIA Corporation. This project has been funded in whole or in part with federal funds from the National Cancer Institute, National Institutes of Health, under Contract No. HHSN261200800001E. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government. © 2021 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 206Issue Supplement 3September 2021Page: e1005-e1005 Advertisement Copyright & Permissions© 2021 by American Urological Association Education and Research, Inc.MetricsAuthor Information Sherif Mehralivand More articles by this author Dong Yang More articles by this author Daguang Xu More articles by this author Ziyue Xu More articles by this author Holger Roth More articles by this author Stephanie Harmon More articles by this author Nathan Lay More articles by this author Maria Merino More articles by this author Bradford Wood More articles by this author Peter Choyke More articles by this author Peter Pinto More articles by this author Baris Turkbey More articles by this author Expand All Advertisement PDF downloadLoading ...