Surrogacy analysis of prostate-specific antigen (PSA) decline for improved overall survival (OS) with enzalutamide (ENZ) in AFFIRM.

Abstract

266 Background: ENZ significantly increased OS for men with metastatic castration-resistant prostate cancer (mCRPC) vs. placebo and significantly decreased PSA levels. However, the prognostic significance and role of PSA falls as a surrogate for OS are not established. Methods: Men in the AFFIRM trial were grouped by maximal unconfirmed PSA decline during the 1st 90 days of treatment in a post-hoc analysis. Each PSA decline criterion was assessed for surrogacy for OS by the proportion of treatment effect (PTE)-explained and Prentice criteria. We also assessed the association of PSA decline with OS, progression-free survival (PFS), and pain response. Results: ENZ improved OS (hazard ratio 0.63, p < 0.001) and was associated with higher rates of PSA declines when compared to placebo (odds ratio > 19.0, p < 0.001). Greater declines in PSA were associated with longer OS, PSA PFS, radiographic PFS, and higher pain response when compared with no PSA decline or PSA increase (table). All decline measures from baseline were highly prognostic for OS and several ( > 0%, ≥ 30%, ≥ 50% declines) explained a proportion of the treatment effect (PTE 1.07–1.29, 95% CI lower bounds > 0.63), in which treatment was no longer significant after adjustment for the decline measures (p > 0.20). Full surrogacy was not demonstrated. Conclusions: In AFFIRM, > 0, ≥30%, and ≥50% PSA declines within 90 days of treatment fulfilled Prentice surrogacy criteria 1–3. Prentice 4, equivalency of survival adjusting for PSA decline outcomes, could not be demonstrated. PSA declines are associated with longer PFS and improved pain response. External prospective validation is needed. Clinical trial information: NCT00974311. [Table: see text]