Thirty prostatic cancer patients were evaluated for staging purposes with both CT (18-sec scan speed) and certain commonly used 'conventional' diagnostic tests, namely: radionuclide bone scan with correlative plain films, the prostatic fraction of the serum acid phosphatase, excretory urogram, and chest radiograph. All patients included in the study had histopathologic proof of diagnosis. CT correctly identified extracapsular prostatic cancer spread locally or in pelvic lymph nodes in 14% of patients with completely negative conventional studies. Sensitivity of extracapsular tumor spread detection increased from 41% to 59% by adding CT to the conventional studies. CT confirmed the presence of tumor spread and localized it in 43% of patients with positive conventional studies. Conventional studies were positive when tumor spread was present in 32% of patients with negative CT. CT reduces understaging when conventional tests are negative, localizes and confirms tumor spread when conventional tests are positive, but cannot demonstrate tumor spread in some patients whose conventional tests are positive and who are subsequently shown to have tumor spread histopathologically.