A novel intermediate endpoint for predicting overall survival in men with metastatic castration-recurrent prostate cancer (CRPC)

Abstract

5113 Background: In this proposed study, we developed and validated a novel composite clinical benefit endpoint constructed from symptoms that have intrinsic clinical importance in 800 men with CRPC who were treated with front-line chemotherapy. Methods: Data from nine multimember trials (five phase II and four randomized phase III studies) conducted by the Cancer and Leukemia Group B (CALGB) from 1992–2004 were combined. Eligible patients had progressive adenocarcinoma of the prostate during androgen ablation (despite castrate testosterone levels), an ECOG performance status of 0–2, adequate hematologic, renal and hepatic functions. The tripartite composite clinical benefit endpoint (TCCBE) had three components: one based on disease progression (whether it be PSA, bone, or soft tissue progression), and the second based on weight loss (defined as at least 10% decline from baseline) or on performance status (PS) decline (by at least one level) to capture clinical deterioration. The third component was based on pain control and opioid analgesic use (no or yes). For a person to fall in the TCCBE “yes” category, at least two of the three components had to be recorded as no. For instance, if a patient at 3 months had no progression, no change in weight and did not use opioid analgesic, then he will be classified in the “yes” group and for the purpose of this analysis was considered as someone who achieved “clinical benefit”. The sample was randomly split into 526 (67%) and 274 (33%) men in the training and testing datasets, respectively. Results: From the training dataset, the median survival times in men who had and did not have clinical benefit were 20.9 months (95% confidence interval (CI) = 18.5–22.8) and 11.1 months (95% = 8.79–12.6, p- value<0.001). In the testing dataset, the median survival times in men with and without clinical benefit were 21.7 months (95% CI= 19.1–26.1) and 8.8 months (95% CI= 7.8–11.6) and in men. The hazard ratio (HR) for men with a clinical benefit compared to men without was 0.52 (95% CI= 0.43–0.62, P<0.001). Conclusions: The TCCBE is a statistically significant intermediate endpoint for predicting overall survival. Prospective validation of this endpoint is needed. No significant financial relationships to disclose.