To assess the predictive value of prostate-specific antigen density (PSAD) to detect clinically significant prostate cancer (prostate cancer grade group (GrGp) ≥2) in a series of men undergoing prostate biopsy with PSA 4-10 ng/mL. We sought to define an optimum cut-point for PSAD and assess how race and body mass index (BMI) affects PSAD performance.

Data on men (n=2162 (56% black)) with serum PSA 4-10 ng/ml undergoing prostate biopsy were analyzed. We compared area under-the-curve (AUC) between PSA and PSAD predicting clinically significant and any prostate cancer (vs. no cancer). Negative predictive values (NPVs) for PSAD cut-points ranging from 0.05 to 0.15 by every 0.01 step-off were calculated. We a priori defined the optimal cut-point for PSAD as NPV=95% and tested whether the cut-point was sensitive to BMI and race by comparing NPV across strata.

Median (IQR) PSA and PSAD were 5.6 ng/mL (4.8-7) and 0.15 ng/ml/cc (0.1-0.22), respectively. PSAD improved the performance of PSA to detect significant cancer (AUC 0.58 to 0.68) (p<0.001) and any cancer (AUC 0.55 to 0.69) (p<0.001). We identified PSAD cut-point of <0.08 having NPV=96% for GrGp ≥2 which was largely unchanged among different races and BMIs.

Regardless of race or BMI, men with a PSAD <0.08 are very unlikely to harbor GrGp ≥2 disease when PSA is 4-10 ng/mL. If validated, PSAD is a simple inexpensive and available tool that can be used to identify men who can likely forego prostate biopsies, thereby reducing over-detection and morbidity from unnecessary biopsies.