Longitudinal assessment of plasma androgen receptor copy number predicts overall survival in subsequent treatment lines in castration-resistant prostate cancer: analysis from a prospective trial

Background: Baseline plasma androgen-receptor copy number (AR-CN) is a promising biomarker for metastatic castration-resistant prostate cancer (mCRPC) outcome and treatment response; however, the role of its longitudinal testing is unproven. We aimed to evaluate the prognostic role of AR-CN assessed before subsequent treatment lines in mCRPC patients. Methods: A subgroup analysis of a prospective multicenter biomarker trial (IRSTB030) was carried out. Plasma AR-CN status (classified as normal or gain, cut-off value = 2) was assessed with digital PCR before each treatment line. Results: Forty mCRPC patients receiving sequentially docetaxel, cabazitaxel and an AR signaling inhibitor (abiraterone or enzalutamide) were analyzed. At multivariate analysis, at each assessment overall survival (OS) was independently correlated with AR-CN status [first line: hazard ratio (HR) 4.1 [95% confidence interval (CI) 1.6-10.5]; second line: HR 2.4 (95% CI 1.1-5.3); third line: HR 2.1 (95% CI 1.0-4.3)] and median prostate-specific antigen [first line: HR 4.4 (95% CI 1.8-10.9); second line: HR 3.4 (95% CI 1.6-7.2); third line: HR 2.5 (95% CI 1.2-5.6)]. In the three subsequent assessments, AR-CN status changed from normal to gain in 15 (38%) patients. These patients had longer OS (47 months) compared with patients presenting AR-CN gain from first assessment (36 months), but shorter than those maintaining normal AR-CN (69 months) (P = 0.003). Conclusions: Plasma AR-CN correlates with survival not only at baseline (before first treatment), but also in the assessments before the following lines. Interestingly, AR-CN status may change from normal to gain across subsequent treatments in a significant number of cases, identifying a group of patients with intermediate outcomes. Longitudinal assessment of AR-CN status could represent a promising method to capture mCRPC intrinsic heterogeneity and to improve clinical management.