Aspects on reference values for tumor markers in human prostatic carcinoma.

The efficiency of the tumor markers prostatic acid phosphatase (PAP), prostate-specific antigen (PSA), neopterin, and osteocalcin was tested with regard to their ability to predict cancer death within 2 years plus survival beyond 2 years in a series of patients with newly diagnosed prostate cancer. For all markers, an elevated level suggested a tumor with a worse prognosis. Moreover, the extent to which the level was increased carried additional information. The prognostic efficiency was routinely improved by selecting cutoff levels higher than the standards suggested by the radioimmunoassay (RIA) kit manufacturers. Seventy-four percent of the patients with elevated levels of neopterin were still alive after 2 years when 8 nmol/L was selected as the upper normal value compared to only 43% at 12 nmol/L. At a cut-off value of 3 micrograms/L for osteocalcin, 79% of the patients with elevated levels were still alive after 2 years compared with only 20% when 7 micrograms/L was selected. Such adjustments to higher cutoff levels could be made without increasing the number of "false-negatives." The efficiency of PAP to predict short-term prognosis was poor at the standard cutoff level of 1.9 microgram/L. Not until 20 micrograms/L was selected did the efficiency exceed 80%. PSA was highly sensitive but little specific at any of the cutoff levels tested with regard to ability to indicate prognosis.