Purpose: Aim of this study was to compare the diagnostic performance of PI-RADS version 1 (v1) and version 2 (v2) in the detection of prostate cancer (PCa).
Methods: Multiparametric MRIs (mpMRI) of 50 consecutive patients with biopsy proven PCa, which had originally been evaluated according to PIRADS v1, were now retrospectively re-evaluated, comparing PI-RADS v1 and v2. MpMRI data were evaluated in comparison with histopathological whole-mount step-section slides. MRI examinations included T2-weighted, diffusion-weighted, and dynamic contrast-enhanced MRI.
Results: Overall PI-RADS v1 showed a significantly larger discriminative ability of tumor detection: PI-RADS v1 AUC 0.96 (95 % CI 0.94-0.98) and v2 AUC 0.90 (95 % CI 0.86-0.94). For peripheral zone lesions, PI-RADS v1 showed a significantly larger ability of PCa discrimination: v1 AUC 0.97 (95 % CI 0.95-0.99) and v2 AUC 0.92 (95 % CI 0.88-0.96). For transition zone lesions, PI-RADS v1 showed more discrimination: v1 AUC 0.96 (95 % CI 0.92-1.00) and v2 0.90 (95 % CI 0.83-0.97), but the difference was not significant. PI-RADS v2 resulted in significantly more false negative results (3 % in v1, 14 % in v2) and a comparable number of true positive results (82 % in v1, 80 % in v2).
Conclusion: PI-RADS v2 uses a simplified approach, but shows a lower diagnostic accuracy. This could lead to a higher rate of false negative results with the risk of missing tumors within low PI-RADS score levels. Therefore, its use cannot be recommended unconditionally, and further improvement should be considered.
Keywords: Multiparametric MRI; PI-RADS; Prostate cancer.