Accuracy and Interobserver Agreement of the Correlation Between Prostate Imaging Reporting and Data System Version 2.1 and International Society of Urological Pathology Scores

Introduction: This research aims to evaluate accuracy and interobserver agreement on the correlation between the Prostate Imaging Reporting and Data System version 2.1 (PI-RADS v2.1) and the International Society of Urological Pathology (ISUP) scores. Methods: We examined patients who underwent prostate multiparametric magnetic resonance imaging (MpMRI) prior to transrectal ultrasound–guided cognitive fusion biopsy between April and December 2019. MpMRI examinations were evaluated by two radiologists according to PI-RADS v2.1. Interobserver agreement was recorded and the final PI-RADS category was decided by consensus. The correlation of cognitive fusion biopsy results with PI-RADS v2.1 score was evaluated. Lesions with Gleason score ≥7 were considered to be clinically significant prostate cancer. Results: A total of 84 patients with 106 lesions were included in the study. The rates of prostate cancer in the PI-RADS groups 1, 2, 3, 4, and 5 were 0%, 0%, 22.2%, 56%, and 94.45%, respectively. There was a positive correlation with an area under the curve value of 0.814 between the PI-RADS v2.1 and the ISUP score. Using PI-RADS ≥3 as the cut-off value in the peripheral zone (PZ) and the whole gland, the negative predictive value for malignancy was 100%. For PI-RADS ≥4, it was 76.47% for PZ and 80.65% for the whole gland. Without applying cut-off values, the interobserver agreement for PI-RADS score was κ = 0.562. Conclusion: Our data support the notion that PI-RADS v2.1 facilitates the evaluation of MpMRI and improves interobserver agreement.