To identify the pretreatment variables that are predictive of response and the duration of response to deferred antiandrogen therapy in men with androgen-independent prostate cancer (AIPC).

A total of 375 patients receiving androgen deprivation therapy for advanced prostate cancer between 1977 and 2002 had medical records available for retrospective review. Of these 375 patients, 163 were diagnosed with AIPC. The inclusion criteria included (1) diagnosis of AIPC and (2) treatment with deferred antiandrogen therapy. AIPC was biochemically defined as two consecutive rises in the prostate-specific antigen (PSA) level during androgen deprivation therapy. The treatment response to deferred antiandrogen therapy was defined as a 50% or greater decline in the pretreatment PSA level. The prognostic value of various pretreatment parameters was determined with the appropriate statistical methods and tested with a Cox proportional hazards model.

Of the 163 patients with AIPC, 36 were treated with deferred antiandrogen therapy. Of these 36 patients, 12 (33.3%) experienced a PSA response. The median PSA failure-free survival was 9.0 months (95% confidence interval 5.2 to 12.9). The only pretreatment variable predictive of a PSA response was the PSA doubling time (PSADT). The mean PSADT in responders was 12.7 months versus 7.5 months in nonresponders (P = 0.037). Moreover, PSADT was the only statistically significant variable on univariate analysis of PSA failure-free survival in responders (hazard ratio 0.202, 95% confidence interval 0.041 to 0.990, P = 0.049). No statistically significant difference was found in cancer-specific survival between responders and nonresponders (P = 0.1501).

The PSADT predicted both the response and the duration of the response to deferred antiandrogen therapy in patients diagnosed with AIPC.