The aim of the present investigation was the evaluation of cost-effectiveness of variables used in monitoring patients with inoperable prostate cancer. Prostate-specific antigen (PSA), prostatic acid phosphatase (PAP), and radionuclide bone scan were considered. The tumor marker positivity was assessed according to dynamic criteria (> 50% increase between consecutive samples). 108 patients entered the study; 72 patients treated with a luteinizing hormone-releasing hormone analogue were followed up for periods ranging from 12 to 64 months. PSA and PAP levels were measured using immunometric assays. Both cutoff-based and dynamic, serial sample-based decision criteria were employed. With respect to a positive bone scan, PSA showed negative predictive values of 82 and 77%, respectively, using 4 and 10 ng/ml as cutoff points. Progression of the disease to the bone occurred in 20 patients: in 17 PSA was the first indicator of progression, in the other 3 PAP anticipated PSA for a very short time (3-4 months), which was not of relevance to clinical decisions. PAP is less specific and sensitive than PSA; PAP may eventually provide information on disease status in a limited percentage of patients with advanced prostate cancer treated with androgen ablation, being differently regulated with respect to PSA. No increasing PSA profile was detected in patients who responded to the therapy. From the results of the present investigation, we draw the following conclusions: (1) PSA can be used reliably as a unique tool in the follow-up of patients for the early detection of progressive disease, and (2) dynamic criteria of evaluation of serial PSA determinations probably provide more effective and earlier clinical information.