Clinical and Imaging Predictors of False-Positive and False-Negative Results in Prostate Multiparametric MRI Using PI-RADS Version 2.

Abstract

Purpose To evaluate predictors of false-positive (FP) and false-negative (FN) results for prostate cancer at prostate multiparametric MRI (mpMRI) using the Prostate Imaging and Reporting Data System version 2 (PI-RADS v2). Materials and Methods This was a single-center retrospective cohort study of 2548 consecutive patients who underwent prostate mpMRI examinations (October 2016-July 2022) containing zero or one PI-RADS v2 category 3-5 lesions. Prostate mpMRI examinations were interpreted by 13 radiologists. FP results were defined as prospective PI-RADS v2 score of 3 or higher but benign or grade group 1 prostate cancer at subsequent combined targeted and systematic biopsy. FN results were defined as prospective PI-RADS v2 score 2 or lower but grade group 2 or higher prostate cancer at subsequent combined targeted and systematic biopsy. Predictors of FP and FN results were assessed by logistic regression. Results Among the 2548 patients (mean age, 65.7 years ± 7.6 [SD]; all male) analyzed, 52.0% (831 of 1597) had FP results and 15.8% (150 of 951) had FN results at mpMRI. FP results were more likely for younger patients (odds ratio [OR], 0.95/y; P < .001), smaller lesions (OR, 0.62/mm; P < .001), transition zone lesions (OR, 1.74 vs peripheral zone; P = .006), and patients with low prostate-specific antigen (PSA) density (OR, 0.55 per 0.1 ng/mL² increase; P < .001). FN results were more likely for older patients (OR, 1.03/y; P = .01) and patients with high PSA density (OR, 2.05 per 0.1 ng/mL² increase; P < .001). Conclusion PSA density and patient age independently predicted FP and FN results for detection of prostate cancer at mpMRI using PI-RADS v2. These factors are not part of the PI-RADS v2 algorithm and may inform mpMRI interpretation to improve prostate cancer diagnosis. Keywords: MR Imaging, Prostate, PI-RADS, Prostate MRI, Prostate Cancer ^©RSNA, 2025.