Addressing the expected survival benefit for clinical trial design in metastatic castration-resistant prostate cancer: Sensitivity analysis of randomized trials