A Systematic Review and Network Meta-analysis of Novel Androgen Receptor Inhibitors in Non-metastatic Castration-resistant Prostate Cancer

Background: Among men with high-risk non-metastatic castrate-resistant prostate cancer (nmCRPC), we used network meta-analysis to compare non-steroidal anti-androgens (NSAAs) and stratified class-level meta-analysis to identify subgroups with particular benefit from NSAAs with androgen deprivation therapy versus androgen deprivation therapy alone. Materials and Methods: We performed a systematic review of phase III parallel-group randomized controlled trials in adult men with nmCRPC. Primary outcome was metastasis-free survival (MFS). Secondary outcomes included overall survival (OS), prostate-specific antigen (PSA) progression-free survival (PFS), and rates of grade 3 to 4 adverse events (AEs). We assessed class-level effects using random effects models; effect modification owing to subgroup effects using random-effects models to pool study-level differences; and comparative outcomes between agents using fixed-effect network models in a Bayesian framework. Results: Three randomized controlled trials were identified. Pooled MFS, PSA-PFS, and OS were significantly greater with NSAA versus placebo (hazard ratio [HR], 0.32; 95% confidence interval [CI], 0.25-0.41; HR, 0.08; 95% CI, 0.05-0.13; and HR, 0.74; 95% CI, 0.61-0.90, respectively). Subgroup analysis demonstrated a greater benefit with NSAAs in men with Eastern Cooperative Oncology Group performance status 0 (HR, 0.30; 95% CI, 0.24-0.38) versus 1 (HR, 0.45; 95% CI, 0.36-0.56; P =.005), but no difference owing to PSA doubling time (P =.43) or use of osteoclast targeting therapy (P =.77). Bayesian analysis showed apalutamide and enzalutamide had a 56% and 44% likelihood of maximizing MFS, respectively, with subgroup analysis demonstrating these agents were preferred regardless of PSA doubling time and performance status. There was a 44%, 41%, and 15% likelihood that apalutamide, darolutamide and enzalutamide offered the greatest OS benefit, respectively. Grade 3 to 4 AEs were more common with NSAAs (odds ratio [OR], 1.47; 95% CI, 1.27-1.71) and there was a 61% chance that darolutamide was preferred. Conclusions: NSAAs improve survival in high-risk nmCRPC. Apalutamide and enzalutamide may result in improved oncologic outcomes. Darolutamide may result in fewer AEs. Phase IV data are needed to validate these findings.