Visceral metastases on abiraterone vs. placebo: A post-hoc analysis of mode of radiographic progression in COU-AA-302.

Abstract

194 Background: Abiraterone prolongs survival in patients with prostate cancer due to its potent inhibition of androgen synthesis. We previously observed increased rates of visceral metastatic disease at the time of progression on abiraterone compared to baseline, a poor prognostic feature associated with non-AR dependent prostate cancer. We hypothesized that the rate of development of visceral disease was increased with abiraterone compared to other treatments without potent androgen signaling inhibition. In order to test this, we examined time to visceral disease among patients treated with abiraterone vs placebo on the COU-AA-302 trial. Methods: We performed a post-hoc analysis of radiographic progression data for the phase 3 study of abiraterone vs placebo, via a data sharing agreement through the Yale Open Data Access Project. Data were censored at end of study treatment. The distribution of cumulative incidence for visceral metastases was calculated by the Kaplan-Meier method and compared with log-rank testing. Multivariable cox regression analysis was performed to assess for independent association of study treatment (abiraterone vs placebo) with the development of visceral metastases. Results: Median follow-up was 12.5mo, and 84 of 1088 patients developed visceral metastases. Univariate log-rank testing showed no difference in time to visceral metastasis between groups (p = 0.97). 1- and 3-yr cumulative incidence of visceral metastases were estimated at 6.4% and 15.6% for abiraterone and 4.6% and 17.0% for placebo, respectively. Study treatment was not a significant predictor for the development of visceral metastasis (HR 0.76; 95% CI 0.47-1.21) after adjustment for baseline metastatic disease burden (soft tissue and bone) and LDH. Conclusions: Abiraterone was not associated with increased visceral metastases at progression compared to placebo. These data give more confidence to oncologists that abiraterone is not driving prostate cancer to an aggressive phenotype characterized by visceral metastases - a finding that is particularly important as abiraterone is increasingly used in early disease states.