Serum prostate specific antigen as a predictor of survival in prostate cancer patients treated with second-line hormonal therapy (CALGB 9181)

Objectives: To explore whether serum prostate specific antigen (PSA) decline can be used as a prognostic factor for survival among patients with androgen-independent prostate cancer (AIPC) who are treated with a secondary hormonal therapy, megestrol acetate. Materials and Methods: One hundred forty-nine patients were randomized with equal probability to either low-dose (160 mg/day) or high-dose (640 mg/day) megestrol acetate. Patients were stratified on performance status (0-1, or 2) and on disease measurability (measurable, or evaluable disease). Results: Patients with high pretreatment PSA (≥95 ng/ml) had worse survival times than patients with low PSA (<95), with a hazard ratio of 1.6 (p = 0.003). The median survival times for patients with a ≥50% reduction in PSA at 8 weeks and those patients without a 50% reduction were 15 and 11 months, respectively (p = 0.763). A landmark analysis at 8 weeks showed that a PSA decline was significantly associated with survival. The survival time for patients whose PSA dropped below the pretreatment levels (median = 15 months) were significantly longer than for patients who did not have a decline in PSA (median = 9 months, p = 0.031). Similar significant results were observed between PSA reduction and survival when landmark analyses were performed at 12 and 16 weeks. PSA changes during the first 45 days after the initiation of therapy were significantly related to survival. The median survival time for patients whose PSA doubled during the first 45 days after treatment and for patients whose PSA did not double during the first 45 days were 10 and 14 months, respectively (p = 0.020). Conclusions: PSA decline is significantly associated with longer survival times in prostate patients treated with secondary hormonal manipulation. Longer survival time is observed among the following AIPC patients: 1) those whose pretreatment PSA were <95 ng/ml: 2) those whose PSA decline was >0% from baseline; and 3) those whose PSA did not double during the first 45 days of therapy. If these observations are generalizeable, PSA decline as a criterion of response could be used to rapidly approve new agents in patients treated with second-line hormonal therapy.

Duke Scholars

Author Susan Halabi Biostatistics & Bioinformatics, Division of Biostatistics