Adoption of prognostic molecular assays for prostate cancer requires evidence of robust performance in different racial groups. Retrospective analysis was conducted to assess the performance of the Oncotype DX® Genomic Prostate Score® test in African American and Caucasian American men with surgically treated prostate cancer. We compared the assay results (scale 0-100) and the 4 gene group scores in biopsy specimens from 201 African American and 1,144 Caucasian American men with clinically localized prostate cancer in 6 cohorts. Adverse pathology was defined as high grade (primary Gleason pattern 4 or any pattern 5) and/or nonorgan-confined disease (≥pT3). Binary logistic regression models were used for adverse pathology. Biochemical recurrence was defined as 2 successive prostate specific antigen levels >0.2 ng/ml or initiation of salvage therapy after radical prostatectomy. Cox proportional hazards models evaluated the association of the assay result or racial group with time to biochemical recurrence. Each cohort had different clinical risk distributions and percentages of African Americans, although median and interquartile ranges of the assay results and gene group scores were similar between both racial groups. In a multivariable model with the assay and pathological/clinical features including race, the assay was significantly associated with adverse pathology (p ≤0.004) and biochemical recurrence (p <0.001). Race was not a significant predictor of either end point. The assay is similarly predictive of outcomes in African American and Caucasian American patients, and improves risk stratification in men with newly diagnosed prostate cancer from both racial groups.