Histopathologic correlation of PI-RADS V.2 lesions on 3T multiparametric prostate MRI.

Abstract

10 Background: Prostate Imaging Reporting and Data System (PIRADS) Version 2 was recently issued by The American College of Radiology. Each score, from 1 to 5, corresponds to the likelihood of clinically significant cancer. We aim to evaluate the correlation between a higher PIRADS score and the presence of high grade tumor on histopathology. Methods: 300 consecutive 3T Multiparametric Prostate MRI (MP-MRI) were identified from March 26, 2012, to June 29, 2015. We reviewed patient charts and a total of 102 patients underwent either a transrectal ultrasound guided (TRUS) biopsy or a total prostatectomy after the MP-MRI. Thirty-four patients were excluded due to prior radiation or cryoablation to the prostate, leaving 68 treatment-naive patients (44-84 years). MP-MRIs were re-interpreted utilizing the new PIRADS V.2 scoring system and correlated with pathology findings. Statistical comparisons were made using R (v.3.1.2) and Fisher’s exact test. Results: High grade (Gleason 8-10) tumors were found only in MP-MRI with a PIRADS score of 5 and a higher percentage of tumor is found in MP-MRI with PIRADS scores of 4 and 5 (80%) when compared to PIRADS scores of 2 and 3 (32%). Low grade (Gleason 6-7) tumors were also found in both PIRADS 4 and 5 categories. There was a significant relationship between a higher PIRADS score and a finding of cancer on pathology (p = 0.0003, OR=0.118, 95% CI 0.03-0.41). Conclusions: The PIRADS scoring system is a useful tool in suggesting the likelihood of clinically significant cancer, with high grade tumors found only in MP-MRI with a score of PIRADS 5. However, MP-MRI and PIRADS criteria alone are not sufficient for determining clinically significant prostate cancer since both high grade and low grade tumors were found in PIRADS 5 and 4 categories. Correlation with other clinical parameters and biopsy is required for confirmation. A larger scale study would be helpful for further delineation. [Table: see text]