Objectives. In multivariable analysis, post-therapy change in prostate-specific antigen (PSA) was shown to be the most significant factor predictive of survival in patients with androgen-independent prostate cancer. To refine the model, we studied the patterns of change in acid phosphatase, alkaline phosphatase, and lactate dehydrogenase after treatment.

Methods. One hundred seven patients with androgen-independent prostate cancer treated on seven different protocols in Memorial Sloan-Kettering Cancer Center were evaluated. For tumor-specific (acid phosphatase and PSA) and nontumor-specifc (alkaline phosphatase and lactate dehydrogenase) enzymes, a minimum 50% or 80% decrease from baseline documented on three separate occasions a minimum of 6 weeks apart was required to categorize a patient as having a decline.

Results. Nineteen patients (18%) had either a 50% decline in acid phosphatase or PSA, of whom 13 (68%) had a decline of both markers. Six (32%) patients showed discordance between the two parameters. Declines in PSA level typically preceded declines in acid phosphatase levels. The median survival of patients showing declines in both markers exceeded that of patients showing declines in PSA alone by 1 year. Although baseline measurements of alkaline phosphatase or lactate dehydrogenase did add additional prognostic information, post-therapy changes did not.

Conclusions. Post-therapy declines in PSA and acid phosphatase represent reproducible endpoints for clinical trials in androgen-independent disease. The requirement of a repeated and parallel decline in both markers may improve the results observed by monitoring declines in PSA alone. Monitoring the two parameters may allow the development of models that can be used as surrogate endpoints for response and survival in a disease in which reproducible measurements of response are lacking.