ABSTRACT
INTRODUCTION: Prostate MRI detecting PI-RADsâ¯=â¯3 lesions has low diagnostic utility for prostate malignancy. Use of PSA density has been suggested to further risk-stratify these men, to potentially avoid biopsies in favor of monitoring. We evaluate the ability of PSA density (PSAd) to risk-stratify PIRADs 3 lesions across patients who underwent a prostate biopsy in a large multi-institutional collaborative. MATERIALS AND METHODS: Pennsylvania Urology Regional Collaborative (PURC) is a voluntary quality improvement collaborative of 11 academic and community urology practices in Pennsylvania and New Jersey. A retrospective analysis was performed on all patients in the PURC database that had a prostate MRI with PI-RADs 3 lesions only. PSA just before the MRI and prostate size reported on MRI were used to calculate the PSA. Clinicopathologic data were evaluated. Univariable analysis using Chi-Square and Kruskal Wallis tests and multivariable logistic regression were used to identify predictors of any PCa, and clinically significant prostate cancer (csPCa) was defined as ≥ Grade Group 2 (GG2.) RESULTS: Between May 2015 and March 2021, 349 patients with PIRADs 3 lesions only were identified and comprised the cohort of interest. Median PSA was 5.0 with a prostate volume of 58cc and a median PSA density of 0.11, 10.6% of the cohort was African American with 81.4% being Caucasian. Significant prostate cancer was detected in 70/349 (20.0%) men. Smaller prostate volume, abnormal DRE, and higher PSAd were significantly associated with clinically significant prostate cancer on univariable analysis. In men with PSAd <0.15, 31/228 (13.6%) harbored csPCa. Multivariable analysis confirmed that men with PSAd >0.15 were more likely to harbor clinically significant prostate cancer (P < 0.001). CONCLUSION: Across a large regional collaborative, patients with PIRADs 3 lesions on mpMRI were noted to have clinically significant cancer in 20% of biopsies. Using a PSA density cut-off of 0.15 may result in missing clinically significant prostate cancer in 13.6%. This information is useful for prebiopsy risk stratification and counseling.
