Aims: To assess the efficacy of serum prostate specific antigen (PSA) in the diagnosis of prostatic cancer. To compare this tumour marker with serum acid phosphatase (ACP) in order to define the more effective diagnostic test.
Methods: Serum samples from 349 patients attending a urology department were assayed for PSA and ACP. Histological assessment of prostatic biopsy samples was used as the standard by which the diagnostic effectiveness of the tumour markers was determined.
Results: Mean serum PSA results from patients with prostatic carcinoma (159 (SEM 35) ng/mL) were significantly different to those from patients with benign prostatic hyperplasia (4.0 (0.53) ng/mL). As there was considerable overlap of results, test sensitivities and specificities were calculated for various decision points. The sensitivity and specificity of PSA at a level of 10 ng/mL were 61.2% and 93.0% respectively, while for ACP at a level of 0.8 U/L they were 47.6% and 89.9%. The areas beneath receiver-operator curves (0.81 for PSA and 0.72 for ACP) indicated that PSA gives better diagnostic information than ACP. For a stated incidence of cancer, posttest probabilities following a given PSA result have been calculated.
Conclusions: PSA is a more effective serum marker for prostatic carcinoma than ACP. We present a table so that for similar patient groups posttest probabilities for prostatic cancer can be assessed for a given serum PSA level.