Can computer-aided diagnosis assist in the identification of prostate cancer on prostate MRI? A multi-center, multi-reader investigation

Authors

Sonia Gaur, National Cancer Institute (NCI)
Nathan Lay, National Institutes of Health (NIH)
Stephanie A. Harmon, National Cancer Institute (NCI)
Sreya Doddakashi, National Cancer Institute (NCI)
Sherif Mehralivand, National Cancer Institute (NCI)
Burak Argun, Acibadem Mehmet Ali Aydinlar Universitesi
Tristan Barrett, University of Cambridge
Sandra Bednarova, Università degli Studi di Udine
Rossanno Girometti, Università degli Studi di Udine
Ercan Karaarslan, Acibadem Mehmet Ali Aydinlar Universitesi
Ali Riza Kural, Acibadem Mehmet Ali Aydinlar Universitesi
Aytekin Oto, The University of Chicago
Andrei S. Purysko, Cleveland Clinic Foundation
Tatjana Antic, The University of Chicago
Cristina Magi-Galluzzi, Cleveland Clinic Foundation
Yesim Saglican, Acibadem Mehmet Ali Aydinlar Universitesi
Stefano Sioletic, Università degli Studi di Udine
Anne Y. Warren, University of Cambridge
Leonardo Bittencourt, Universidade Federal Fluminense
Jurgen J. Fütterer, Radboud University Nijmegen
Rajan T. Gupta, Duke University
Ismail Kabakus, Hacettepe Üniversitesi
Yan Mee Law, Singapore General Hospital
Daniel J. Margolis, Cornell University
Haytham Shebel, Mansoura University
Antonio C. Westphalen, University of California, San Francisco
Bradford J. Wood, National Institutes of Health (NIH)
Peter A. Pinto, National Cancer Institute (NCI)
Joanna H. Shih, National Cancer Institute (NCI)
Peter L. Choyke, National Cancer Institute (NCI)
Ronald M. Summers, National Institutes of Health (NIH)
Baris Turkbey, National Cancer Institute (NCI)

Document Type

Journal Article

Publication Date

9-1-2018

Journal

Oncotarget

Volume

9

Issue

73

DOI

10.18632/oncotarget.26100

Keywords

Computer-aided diagnosis; Multiparametric MRI; PI-RADSv2; Prostate cancer; Tumor detection

Abstract

© Gaur et al. For prostate cancer detection on prostate multiparametric MRI (mpMRI), the Prostate Imaging-Reporting and Data System version 2 (PI-RADSv2) and computeraided diagnosis (CAD) systems aim to widely improve standardization across radiologists and centers. Our goal was to evaluate CAD assistance in prostate cancer detection compared with conventional mpMRI interpretation in a diverse dataset acquired from five institutions tested by nine readers of varying experience levels, in total representing 14 globally spread institutions. Index lesion sensitivities of mpMRI-alone were 79% (whole prostate (WP)), 84% (peripheral zone (PZ)), 71% (transition zone (TZ)), similar to CAD at 76% (WP, p=0.39), 77% (PZ, p=0.07), 79% (TZ, p=0.15). Greatest CAD benefit was in TZ for moderately-experienced readers at PI-RADSv2 <3 (84% vs mpMRI-alone 67%, p=0.055). Detection agreement was unchanged but CAD-assisted read times improved (4.6 vs 3.4 minutes, p<0.001). At PI-RADSv2 . 3, CAD improved patientlevel specificity (72%) compared to mpMRI-alone (45%, p<0.001). PI-RADSv2 and CAD-assisted mpMRI interpretations have similar sensitivities across multiple sites and readers while CAD has potential to improve specificity and moderately-experienced radiologistsf detection of more difficult tumors in the center of the gland. The multi-institutional evidence provided is essential to future prostate MRI and CAD development.

This document is currently not available here.

Share

COinS