To study the usefulness of the complexed-to-total (C:T) prostate-specific antigen (PSA) ratio in the early detection of prostate cancer in patients with a total PSA value <4.0 ng/ml. Total PSA and PSA complexed to alpha(1)-antichymotrypsin were measured in plasma from 193 men with benign prostatic hyperplasia (BPH) and 34 with prostate cancer. The diagnosis was confirmed in 28 BPH and 16 prostate cancer patients by biopsy and in 165 BPH and 18 prostate cancer patients by histological study following transurethral prostatectomy or open prostatectomy. The area under the receiver operating characteristic (ROC) curve was significantly greater for the C:T PSA ratio (0.908) than for total PSA (0.692) (p<0.001). Using a cut-off point of 0.83 for the C:T PSA ratio and regardless of the digital rectal examination (DRE) finding, 20 of the 34 prostate cancer patients would have been given a correct diagnosis (59% sensitivity) and in only 8 of the 193 BPH patients would a biopsy have been necessary (96% specificity). With a cut-off of 0.79, the sensitivity increased to 85% with a specificity of 92%. When the analysis was restricted to the 44 patients with abnormal DRE, the area under the ROC curve for the C:T PSA ratio was 0.919, and a cut-off point of 0. 78 gave a sensitivity of 87% and a specificity of 93%. Using a cut-off of 0.63, all prostate cancers were detected (100% sensitivity) and 54% of the negative biopsies would have been eliminated. For the 183 patients diagnosed following surgery, a cut-off of 0.82 gave a sensitivity of 72% and a specificity of 94%. Our results show that the C:T PSA ratio significantly improves the clinical utility of the PSA assay for detecting prostate cancer in patients with total PSA < 4 ng/ml, increasing the sensitivity without significantly increasing the number of biopsies.