Isolated prostate specific antigen (PSA) determinations in asymptomatic individuals have not demonstrated sufficient sensitivity and specificity to be useful in the routine evaluation of prostate disease. To enhance the accuracy of serum PSA we have used a quotient of serum PSA and prostate volume, which we refer to as prostate specific antigen density (PSAD). Prostate volume in this study was calculated from magnetic resonance imaging determinations of benign prostatic hypertrophy (BPH) or from the dimensions of the surgical specimen of cancer using the formula, length x width x depth x 0.5 = volume. A total of 61 patients with prostatic disease clinically confined to the prostate glands (41 with prostate cancer undergoing radical prostatectomy and 20 with BPH) was evaluated. The mean PSAD for prostate cancer was 0.581 while that for BPH was 0.044 (p less than 0.002). No patient with BPH had a PSAD of greater than 0.117 and only 1 patient had a density of 0.1 or greater. Of 34 patients with a PSAD of 0.1 or greater 33 had prostate cancer. Only 2 of the 41 prostate cancer patients and 14 of the BPH patients had a PSAD of 0.05 or less. There were 11 patients with a PSAD of greater than 0.05 and less than 0.1, including 6 with prostate cancer (1 with P0 disease) and 5 with BPH. Of the 6 prostate cancer patients 5 had a PSA of 4.0 or less and among the 5 patients with BPH 4 had a serum PSA of greater than 4.0 and 1 had a PSA of greater than 10. These results suggest that PSAD may be useful in distinguishing BPH and prostate cancer.