Correlation between overall response rate and progression-free survival/overall survival in comparative trials involving targeted therapies in molecularly enriched populations.

Abstract

3588 Background: Randomized trials involving agents targeting oncogene addicted tumors have greatly increased over the past decade. Whether clinical response rates can predict or correlate with efficacy measures such as progression-free survival (PFS) or overall survival (OS) has not been established in molecularly enriched patient populations. In this meta-analysis, we investigated whether improvements in objective response rate (ORR) in comparative trials using targeted agents could serve as a potential surrogate endpoint for improvements in PFS or OS in populations with oncogene addicted cancer. Methods: CT.gov and MEDLINE databases were queried (using commercial text mining software I2E) for randomized, phase 3 clinical trials based on the following prospectively defined criteria: (1) use of agents targeting EGFR activating mutations (erlotinib, gefitinib, afatinib, dacomitinib, osimertinib), ALK and ROS1 rearrangements (crizotinib, ceritinib, alectinib), BRAF V600E or V600K mutations (dabrafenib), and BCR-ABL fusion protein (imatinib, dasatinib, nilotinib, ponatinib); (2) must include molecularly enriched trial populations (biomarker subgroup data included if available); (3) control arms should not include targeted agents directed towards those molecularly enriched populations. ORR, OS, and PFS data were manually extracted from the relevant studies and correlative analyses (weighted Pearson correlation) were performed. Results: 61 trials were identified with 15 ultimately meeting the prespecified criteria. ORR effect size (both the ORR difference and log odds ratio) and the log PFS hazard ratio were strongly correlated (-0.78, p-value = 0.0007). No significant correlation was found between ORR and OS. Conclusions: In our analyses, a strong correlation between ORR and PFS was found in randomized clinical trials investigating agents targeting oncogene-driven cancers. Establishing a correlation between ORR and OS was limited, most probably due to confounding factors such as treatment cross-over following progression, number of subsequent therapies and long post-progression survival in this setting. These findings further warrant the use of ORR as a surrogate for PFS in biomarker-driven studies.