Objective: To explore the value of the prostate imaging reporting and data system (PI-RADS) score of prostate multi-parametric magnetic resonance imaging (mpMRI) in predicting the pathological features of PCa based on matching images and whole-mount pathology images.
Methods: This retrospective study included 318 cases of PCa treated by radical prostatectomy in our hospital from August 2016 to December 2018, with preoperative mpMRI images and complete whole-mount pathological sections. We obtained PI-RADS scores on the mpMRI lesions corresponding to the cancer lesions, evaluated the Gleason scores, pT stages, pN stages and cribriform structure, and compared them between different groups using Chi-square test or Fisher's exact test. We evaluated the efficiency of the PI-RADS score in distinguishing different pathological features by ROC curve analysis, and obtained the corresponding area under the curve (AUC) and 95% confidence interval (CI).
Results: The 318 patients averaged 69 years of age, with a median preoperative PSA level of 11.0 μg/L and a median tumor diameter of 1.8 cm. The PI-RADS score was significantly correlated with the Gleason score, pT stage, pN stage and cribriform structure (all P < 0.01), with AUCs of 0.773 (95% CI: 0.704－0.843) for distinguishing Gleason scores (3+3 vs >3+3), 0.748 (95% CI: 0.694－0.803) for distinguishing pT stages (T2 vs >T2), 0.700 (95% CI: 0.598－0.802) for distinguishing pN stages (N0 vs N1), and 0.831 (95% CI: 0.786－0.876) for distinguishing the cribriform structure (negative vs positive).
Conclusion: The preoperative PI-RADS score of mpMRI in PCa patients is significantly correlated with postoperative pathological features, and therefore can be used for risk stratification of the malignancy.
Keywords: prostate cancer; multi-parametric magnetic resonance imaging; prostate imaging reporting and data system score; Gleason score; cribriform structure.