Audit of cancer yields after prostate MRI using both the PI-RADS version 2 and Likert scoring systems

Abstract

To audit diagnostic yields of the updated magnetic resonance imaging (MRI)-directed prostate cancer diagnostic service according to Prostate Imaging - Reporting and Data System (PI-RADS) version 2 and Likert assessments, comparing outcomes of the two scoring systems. Consecutive men with suspected cancer undergoing prostate MRI were included. Biopsy rates and histological diagnostic yields of all and International Society of Urological Pathology Grade Group (ISUP GG) ≥2 cancers according to PI-RADS and Likert assessment categories were documented and outcomes compared. Of 326 men (91% biopsy naive), 177 (54%) underwent transrectal (n=119) or transperineal (n=58) ultrasound-guided biopsies; 92% with negative MRI avoided immediate biopsies following multidisciplinary team (MDT) review. All cancer and ISUP GG≥2 cancer-detection rates increased with increasing suspicion scores. Prospective paired PI-RADS and Likert scoring was undertaken in 323/326 studies, with 87% concordance rate. High concordance between PI-RADS and Likert scores was observed in negative MRI (99%) and score 5 (96%). High discordance was demonstrated in the PI-RADS 4 group (32% with PI-RADS 4 categories up-risked to Likert 5). All cancer and ISUP GG≥2 cancer detection rates for MRI score ≥3 were 78% and 63%, and for MRI score ≥4 were 75% and 61%, respectively for both PI-RADS and Likert scoring systems. Most men with negative prostate MRI can avoid immediate biopsies following MDT review. Performance of PI-RADS and Likert scoring systems in clinically significant cancer detection after positive MRI is equivalent. Discordance between PI-RADS and Likert systems seems mostly confined to PI-RADS 4 categories.