We measured a ratio of serum prostate specific antigen (PSA) and gamma-seminoprotein concentrations (referred to as the PSA/gamma-seminoprotein ratio) and evaluated its usefulness for the diagnosis of prostate cancer. Between April 1988 and October 1992, 214 men underwent prostatic biopsy and/or transurethral resection of the prostate, and the disease was diagnosed pathologically. Of 214 patients 127 were diagnosed as having benign prostatic hyperplasia, prostatitis or a normal prostate (no cancer), while 87 had prostate cancer. Of 61 patients with a serum PSA level greater than 10 ng./ml. 50 (82.0%) had prostate cancer, compared to 31 of 84 (36.9%) with a serum PSA level of 3.0 to 10 ng./ml. Of 113 patients with a serum gamma-seminoprotein level greater than 4.0 ng./ml. 52 (46.0%) had prostate cancer. The mean plus or minus standard deviation of the PSA/gamma-seminoprotein ratio for 127 patients without cancer was 0.942 +/- 0.564, while that for 87 prostate cancer patients was 12.840 +/- 45.327 (Wilcoxon p < 0.0001). The mean plus or minus standard deviation of the PSA/-gamma-seminoprotein ratios for 37 prostate cancer patients with a PSA level of 10 ng./ml. or less and for 50 prostate cancer patients with a PSA level of more than 10 ng./ml. were 2.044 +/- 0.767 and 20.829 +/- 58.757, respectively. Even the mean PSA/gamma-seminoprotein ratio for prostate cancer patients with a PSA level of 10 ng./ml. or less was significantly greater than that for patients without cancer (Wilcoxon p < 0.0001). The sensitivities for PSA (cutoff value 3.0 ng./ml.), gamma-seminoprotein (cutoff value 4.0 ng./ml.) and PSA/gamma-seminoprotein ratio (cutoff value 1.45) were 93.1%, 59.8% and 92.0%, respectively, and the specificities were 49.6%, 52.0% and 91.3%, respectively. Of 91 patients with a PSA/gamma-seminoprotein ratio of 1.45 or more 80 (87.9%) had prostate cancer, while 116 of 123 (94.3%) with a PSA/gamma-seminoprotein ratio of less than 1.45, had no cancer. These results suggest that PSA/gamma-seminoprotein ratio yields the same sensitivity as PSA and more specificity than PSA levels, offering significant advantage over PSA in detecting prostate cancer. The mean plus or minus standard deviations of PSA/gamma-seminoprotein ratios for stages A, B, C and D prostate cancer were 1.847 +/- 0.786 (11 patients), 2.740 +/- 1.536 (30), 7.626 +/- 9.140 (12) and 27.149 +/- 70.500 (34), respectively. Stages A versus C (p = 0.0410), A versus D (p < 0.0001) and B versus D (p < 0.0001) achieved statistical significance, although stages A versus B (p = 0.0577), B versus C (p = 0.1905) and C versus D (p = 0.1128) failed. The mean plus or minus standard deviations of PSA/gamma-seminoprotein ratios for well, moderately and poorly differentiated adenocarcinomas were 2.280 +/- 1.688 (20 patients), 6.109 +/- 7.894 (36) and 27.470 +/- 73.927 (31), respectively. The mean PSA/gamma-seminoprotein ratio for each pathological grade versus those for all other pathological grades achieved statistical significance.