Prostate specific antigen: A prognostic marker of survival in good prognosis metastatic prostate cancer? (EORTC 30892)

Purpose: We study the value of PSA response and PSA progression as prognostic factors for survival in good prognosis metastatic prostate cancer. Methods: Data from 257 patients treated with Flutamide or Cyproterone acetate within the EORTC GU Group protocol 30892 have been used and analysis by Cox models. Results: A PSA response defined as a decrease to less than or equal to1 ng/ml and to between 1 and 10 ng/ml was associated with a hazard ratio of 0.30 and 0.61 for overall survival, respectively, as compared to the non-responders (PSA > 10 ng/ml). Five definitions of PSA progression were considered: (1) a confirmed or (2) a repeated doubling of the PSA over nadir and unconfirmed (3) 100%, (4) 50% and (5) 20% increase of the PSA over nadir, each to a value >4 ng/ml. Definition (5) was the most sensitive with sensitivity 76.20% and specificity 32.08%. With this definition, 70.0% of the patients had a PSA progression, which occurred in median 1.98 years before death. Conclusions: For good prognosis metastatic prostate cancer patients under anti-androgen treatment, PSA response at 6 months with cut-off levels of less than or equal to1 ng/ml and less than or equal to10 ng/ml is prognostic for survival. A 20% increase over nadir to a value >4 ng/ml is prognostic for a poor survival with a 76.20% sensitivity. In this study, confirmation of the increase by a second observation did not seem necessary. Genuine surrogacy is not established in this study. (C) 2003 Elsevier Science B.V. All rights reserved.