Three tests are commonly used to diagnose prostate carcinoma to date: serum prostate specific antigen (PSA), digital rectal examination and transrectal ultrasonography. We evaluated these 3 tests in 1,001, 6-sector prostate needle biopsies to rule out prostate carcinoma. Of the biopsies 253 (25.3%) revealed prostate cancer. As a single test, PSA was superior to digital rectal examination or transrectal ultrasonography in predicting cancer in this patient population using difference of proportions tests. Receiver operating characteristic analysis also showed PSA to be the superior test. The combinations of PSA plus transrectal ultrasonography and PSA plus digital rectal examination were superior to digital rectal examination plus transrectal ultrasonography. We found cancer in 35 of 188 patients (18.6%) with intermediate PSA levels of 4.1 to 10.0 ng./ml. and normal or asymmetric nonindurated rectal examinations. Only 5 of 79 patients (6.3%) with a normal digital rectal examination and PSA level of less than 4.0 ng./ml. demonstrated carcinoma on biopsy. Of the 5 patients 4 had annual increases in PSA of 40% or greater. While hypoechoic sectors were more than twice as likely as isoechoic sectors of the prostate to contain malignancy on biopsy, nearly 37.6% of the cancers were found in isoechoic sectors. A strategy of performing biopsy of only hypoechoic sectors would have misdiagnosed 24.6% of the patients with prostate cancer. We conclude that serum PSA is the most accurate of the 3 diagnostic tests evaluated. We also recommend a systematic sextant biopsy technique.