BACKGROUND: Earlier time-to-event (TTE) endpoints than overall survival (OS) are needed to accelerate drug development and regulatory approval in metastatic castration-resistant prostate cancer (mCRPC). In the phase 3 VISION trial, [177Lu]Lu-PSMA-617 (lutetium (177Lu) vipivotide tetraxetan) plus SoC prolonged PCWG3-defined radiographic progression-free survival (rPFS) and OS, and delayed time to first symptomatic skeletal event (SSE) or death in patients with mCRPC. This post hoc analysis aimed to estimate correlations between these TTE endpoints and provide further evidence for rPFS as an early endpoint for regulatory approval, in the context of radioligand therapy (RLT).
METHODS: Because censoring prevented observation of events, three Spearman’s rank correlation methods adapted to TTE endpoints were used to estimate the strength of association between rPFS, OS and time to first SSE. Correlation estimations were used for both VISION protocol data (as is) and imputed data. For the latter, imputation methods were used for rPFS and OS to adjust for informative censoring; for time to first SSE, patients were censored at 30 days after first treatment change or death (Table).
RESULTS: The three Spearman’s rank correlation methods provided similar results. The correlation between rPFS and OS was moderate-to-strong using both protocol and imputed data, with estimates of ∼0.7 or higher. Stronger associations between rPFS/OS and SSE were observed for protocol vs. imputed data; however, SSE data mainly contained information on OS for the protocol data (data not shown). Estimates (95% confidence interval). Imputed data were used for the correlation between rPFS/OS and SSE.aEarly censored data imputed; bProtocol data; cPatients censored at 30 days after first treatment change or death.
CONCLUSIONS: This analysis, the first for an RLT, shows a moderate-to-strong correlation between rPFS and OS, indicating that rPFS, as defined by PCWG3, may be relevant as an early endpoint for regulatory approval and clinical trial design in this context. Correlations between TTE endpoints and PRO scores will complement this analysis.
