Prostate imaging reporting and data system version 2.1 as a predictor of clinically significant and aggressive prostate cancer: a radical prostatectomy-validated study.

BACKGROUND

Prostate cancer incidence continues to rise globally, emphasizing the need for improved diagnostic and risk-stratification tools. This study aimed to evaluate the performance of Prostate Imaging Reporting and Data System (PI-RADS) version 2.1 in detecting clinically significant prostate cancer (csPCa), identifying high-risk patients, and predicting adverse pathological outcomes using radical prostatectomy (RP) pathology as the reference standard.

METHODS

This retrospective study included 190 men who underwent preoperative multiparametric prostate magnetic resonance imaging (mpMRI) followed by RP. PI-RADS scores were independently assigned by two fellowship-trained radiologists and correlated with RP pathology.

RESULTS

Higher PI-RADS scores were strongly associated with higher radical prostatectomy Gleason scores (rpGSs), with each one-category increase yielding 8-10-fold higher odds of elevated rpGS (P<0.001). Associations remained significant after adjustment for age, prostate-specific antigen (PSA), and biopsy Gleason scores (bxGSs) [adjusted odds ratio (OR) 5.54-8.19, P<0.001]. Increasing PI-RADS categories were also significantly linked to extraprostatic extension, seminal vesicle invasion, lymphovascular invasion, perineural invasion, and pathological T3 (pT3) disease (P<0.001-0.042). Each category increase corresponded to a threefold increase in odds of pT3 disease [OR 3.02, 95% confidence interval (CI): 1.26-8.28, P=0.02]. Among diagnostic thresholds, PI-RADS ≥3 had the highest sensitivity but lower specificity, whereas PI-RADS ≥4 provided the best balance (accuracy 0.78-0.81; specificity 0.53-0.58). PI-RADS 5 showed the strongest association with aggressive disease features, including grade group ≥3 and pT3 status (accuracy 0.67-0.74). Incorporating MRI improved risk-prediction performance, increasing the concordance index (C-index) from 0.864 (biopsy-only) to 0.883-0.894 and raising the area under the curve (AUC) for csPCa detection from 0.88 to 0.91-0.92. Interobserver agreement was near-perfect (weighted κ=0.96).

CONCLUSIONS

PI-RADS v2.1 is a robust predictor of clinically significant and aggressive prostate cancer, demonstrating excellent reliability and incremental value beyond clinical parameters and biopsy. PI-RADS ≥4 represents an optimal threshold for csPCa detection, while PI-RADS 5 identifies patients at highest risk for adverse pathology. MRI-integrated assessment should inform risk stratification, active surveillance selection, and surgical planning.