Multiparametric MRI for Prostate Cancer Detection Using PI-RADS v2 Compared with MRI/Ultrasound Fusion- Guided Biopsy

Abstract

Objective: To determine the diagnostic performance of multiparametric magnetic resonance imaging (mp-MRI) for prostate cancer (PCa) detection compared with MRI/ultrasound fusion-guided prostate biopsy. Materials and Methods: Pre-biopsy prostate mp-MRI of 100 consecutive men were retrospectively compared with prostate biopsy obtained by MRI/ultrasound fusion guidance between June 2017 and July 2018. Two experienced radiologists assigned PI-RADS score and localization the suspicious lesions by consensus. Tumor detection rates were calculated for each PI-RADS scores, Gleason scores, and tumor location. Results: Of the 151 target lesions on mp-MRI from 100 patients, 21% (31/151) were pathologically determined as clinically significant PCa. The detection rates on targeted biopsy for PI-RADS scores of 3, 4, and 5 were 10%, 29%, and 54%, respectively. No cancer was detected in PI-RADS score of 2 lesions. Higher level of suspicion PI-RADS scores indicated higher percentages of high Gleason scores. The percentages of PCa for Gleason scores of 7, 8, and 9 were 17%, 11%, and 0% for PI-RADS 3, 50%, 44%, and 25%, for PI-RADS 4, and 33%, 44%, and 75% for PI-RADS 5. High rates at 82% (23/28) of false-positive mp-MRI findings were found for small of less than 0.5 mL peripheral zone targets and 91%(10/11) of targets in the anterior location of the transition zone. Conclusion: PI-RADS v2 showed satisfactory performance for diagnosing clinically significant PCa and provided predictive information on tumor grade. High negative predictive values for PI-RADS v2 could be used in clinical management workflow to confidently avoid prostate biopsies.