A Validated Prognostic Index Predicting Response to Dexamethasone and Diethylstilbestrol in Castrate-Resistant Prostate Cancer

BACKGROUND: The objective of this study was to identify subgroups of patients with castrate-resistant prostate cancer (CRPC) who had a prolonged response to combined dexamethasone and diethylstilbestrol (DS) therapy by constructing a prognostic index. METHODS: Multivariate and cutoff point analyses with bootstrapping were performed 1 month after commencing DS therapy, and data were validated using an independent external dataset. RESULTS: The median overall survival for 145 patients was 18.3 months (95% confidence interval [CI], 15.4-23.3 months). Only 2 factors were significant on multivariate analysis: a prostate-specific antigen (PSA) level below the median at the start of DS therapy and a decline >50% in the PSA level after 1 month on DS therapy. Subsequent cutoff point analyses revealed that a PSA level <88 ng/mL at the start of DS therapy and a reduction >53.1% in PSA after the first month of treatment were the most significant factors. These 2 factors were used to construct a prognostic index, which was validated successfully by the external dataset (median overall survival, 18.3 months [95% CI, 16.4-28.0 months]; concordance, 72% [95% CI, 68%-76%]). The prognostic index identified 3 prognostic groups: The 2-year overall survival rate for these 3 groups was 68% (95% CI, 57%-98%) for the good prognostic group, 40% (95% CI, 31%-52%) for the intermediate prognostic group, and 12% (95% CI, 5%-25%) for the poor prognostic group. CONCLUSIONS: The easy-to-use prognostic index that the authors developed was able to identify a subgroup of patients with CRPC who had prolonged survival only 1 month after starting DS therapy. Cancer 2010;116;3595-602. (C) 2010 American Cancer Society