Presentation Authors: Sohrab Afshari Mirak*, Amirhossein Mohammadian Bajgiran, Melina Hosseiny, Sepideh Shakeri, Afshin Azadikhah, Anthony Sisk, Robert Reiter, Steven Raman, LOS ANGELES, CA
Introduction: To investigate the performance of PI-RADSv2 & a quantitative PI-RADSv1 based scoring system in 3T multiparametric-MRI (3TmpMRI) for prostate cancer (PCa) detection, grading & staging with whole-mount histopathology (WMHP) correlation.
Methods: In this HIPAA-compliant, IRB-approved retrospective study, the cohort included 569 men with PCa, who underwent 3T mpMRI within 6 months prior to robotic assisted radical prostatectomy (RALP), at a single tertiary care center between December 2009 and October 2017. The performance of a quantitative PI-RADSv1 based protocol (method 1) was compared to qualitative PI-RADSv2 protocol (method 2) evaluating sensitivity, PPV, grading and staging. In method 1, lesions were assigned an assessment grade based on the formula (T2-weighted imaging (T2WI) + 2Â· apparent diffusion coefficient (ADC) + dynamic contrast-enhanced (DCE) âˆ’ 0.25Â· transition zone (TZ))/4, where T2WI, ADC, and DCE were the 1â€“5 assessment, and TZ was the binary assessment of the TZ (1 if the lesion was in the TZ and 0 if not)(table 1). Method 2 was the standard protocol for PIRADSv2 scoring system with T2WI diffusion-weighted imaging (DWI) as the main sequences for TZ and peripheral zone (PZ) lesions. For both methods, score 3 was considered as low suspicion and scores 4 and 5 were considered as high suspicion for PCa. All 3T mpMRI and histopathology lesions (true positive, false positive and false negative) were included and evaluated using STATA version 12.
Results: 1353 PCa lesions in 569 patients were included in the study. Methods 1 and 2 had similar sensitivity (44.3% vs 45%) and PPV (81.2% and 81.5%), respectively. The AUROC for PCa grading (Gleason score 6 vs >6) for methods 1 was slightly but non-significantly higher than method 2 (0.67 vs 0.64)(roccomp command, p = 0.25). However, both methods showed similar AUROC for PCa staging (
Conclusions: In this study the quantitative PI-RADSv1 and qualitative PI-RADSv2 based methods had similar performance for PCa detection, grading and staging.
Source of Funding: This work was supported by funds from the Integrated Diagnostics Program, Department of Radiological Sciences & Department of Pathology and Laboratory Medicine, David Geffen School of Medicine at UCLA.