Diagnostic performance of prostate multiparametric magnetic resonance imaging in African-American men.

Abstract

To compare test performance of multiparametric magnetic resonance imaging (mpMRI) for detection of prostate cancer (PCa) in African-American men (AAM) and white men (WM) using the Prostate Imaging Reporting and Data System (PIRADS) in unmatched groups as well as a cohort matched for clinical factors.We examined our database of men who underwent prostate mpMRI prior to biopsy between October 2014 and June 2017 (n=601; 60 AAM, 541 WM). Test performance was defined using PIRADS classification 4 or 5 considered test positive and Gleason grade group (GGG) 2 or greater from any biopsy core considered outcome positive. A subset analysis was performed using a propensity score caliper matching algorithm to match AAM to WM in a 1:2 ratio using the variables age, PSA, and PSA density.No significant differences in test performance were found with similar sensitivity (86.7% vs 83.6, p=1.00), specificity (45.9% vs 49.1%, p=0.71), positive predictive value (50.0% vs 46.9%), and negative predictive value (85.0% vs 84.8%, p=1.00) for AAM and WM. Similar results were noted in our matched comparison. The rate of upgrading between targeted and systematic biopsy cores did not statistically differ between AAM and WM in both unmatched (12.2% vs. 15.8%, p=0.66) and matched (12.2% vs 12.8%, p=0.92) comparisons.Our findings provide supporting evidence that AAM have similar outcomes to WM in PCa detection using mpMRI. We suggest that mpMRI should not be withheld or offered preferentially on the basis of race when used for the detection of PCa.