Abstract
OBJECTIVES:: To develop and validate a logistic regression model and point-based scoring system for predicting ≥Grade Group 3 clinically significant prostate cancer using a combination of multiparametric MRI findings and patient risk factors.
METHODS:: This Institutional Review Board-approved, retrospective cohort study was conducted1/1/2022–12/31/2022 with data analysis 7/1/2023–6/30/2025. Males undergoing prostate multiparametric MRI during the study period at a multi-institutional health system with prostate-specific antigen ≥4ng/mL and prostate biopsy and/or radical prostatectomy within 6 months post-multiparametric MRI were included in the study. A separate derivation cohort included 960 men who underwent multiparametric MRI from 2015–2019. A logistic regression and point-based scoring system for predicting high-grade clinically significant prostate cancer (≥Grade Group 3) was developed using predictors including prostate-specific antigen density (PSAD), highest PI-RADS score from multiparametric MRI, extraprostatic extension, and age groups (e.g., 65-<70). Discrimination was assessed using area under the curve.
RESULTS:: 1245 patients met inclusion criteria; 83% were White; 86% were ≥60 years of age. 83% had a focal lesion with PI-RADS score of ≥3. Based on the new point-based scoring system for predicting high-grade (≥Grade Group 3) prostate cancer, 28% of patients had a cumulative score of 0–7, with an estimated clinically significant prostate cancer risk of 9%. The area under the curve was 0.77 for both the logistic regression model and the point-based system.
CONCLUSIONS:: In a multi-institutional health system, age, as well as prostate-specific antigen density, highest PI-RADS score from multiparametric MRI, and extraprostatic extension significantly predicted high-grade clinically significant prostate cancer in a logistic regression and point-based scoring system.