Abstract
OBJECTIVES: To evaluate the recommendations for multiparametric prostate MRI (mp-MRI) interpretation introduced in the recently updated Prostate Imaging Reporting and Data System version 2 (PI-RADSv2), and investigate the impact of pathologic tumour volume on prostate cancer (PCa) detectability on mpMRI. METHODS: This was an institutional review board (IRB)-approved, retrospective study of 150 PCa patients who underwent mp-MRI before prostatectomy; 169 tumours ≥0.5-mL (any Gleason Score [GS]) and 37 tumours <0.5-mL (GS ≥4+3) identified on whole-mount pathology maps were located on mp-MRI consisting of T2-weighted imaging (T2WI), diffusion-weighted (DW)-MRI, and dynamic contrast-enhanced (DCE)-MRI. Corresponding PI-RADSv2 scores were assigned on each sequence and combined as recommended by PI-RADSv2. We calculated the proportion of PCa foci on whole-mount pathology correctly identified with PI-RADSv2 (dichotomized scores 1-3 vs. 4-5), stratified by pathologic tumour volume. RESULTS: PI-RADSv2 allowed correct identification of 118/125 (94 %; 95 %CI: 90-99 %) peripheral zone (PZ) and 42/44 (95 %; 95 %CI: 89-100 %) transition zone (TZ) tumours ≥0.5 mL, but only 7/27 (26 %; 95 %CI: 10-42 %) PZ and 2/10 (20 %; 95 %CI: 0-52 %) TZ tumours with a GS ≥4+3, but <0.5 mL. DCE-MRI aided detection of 4/125 PZ tumours ≥0.5 mL and 0/27 PZ tumours <0.5 mL. CONCLUSIONS: PI-RADSv2 correctly identified 94-95 % of PCa foci ≥0.5 mL, but was limited for the assessment of GS ≥4+3 tumours ≤0.5 mL. DCE-MRI offered limited added value to T2WI+DW-MRI. KEY POINTS: • PI-RADSv2 correctly identified 95 % of PCa foci ≥0.5 mL • PI-RADSv2 was limited for the assessment of GS ≥4+3 tumours ≤0.5 mL • DCE-MRI offered limited added value to T2WI+DW-MRI.