A meta-analysis of the efficacy of epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor treatment in non-small-cell lung cancer patients.

Abstract

e20604 Background: Non-small cell lung cancer (NSCLC) accounts for approximately 85% of lung cancer cases. Tyrosine kinase inhibitors (TKI) targeting epidermal growth factor receptor (EGFR) mutations have been the driving force for the treatment of NSCLC in recent years. An important clinical question that arises within oncology drug development is whether we can use early clinical endpoints from phase II or even phase I studies to predict phase III overall survival (OS) outcome for a new treatment. Methods: The Pubmed database was searched for randomized clinical trials from 2010 to 2015. Only phase II or phase III trials using EGFR TKIs (gefitinib, erlotinib, afatinib, dacomitinib, and icotinib) for patients with advanced NSCLC were included. The trials were categorized by line of treatment (1st line, 2nd line or above) and by patient populations (unselected or molecularly selected (EGFR mutation)). Median progression free survival (PFS) and median OS, PFS hazard ratio (HR) and OS HR were analyzed using Spearman’s rank order correlation. Results: In total, 38 trials (68 arms) were identified for unselected and molecularly selected patient population. Out of those, 62 arms were on EGFR TKI treatment. The median number of patients in each arm was 71 (range: 18- 617). Of 38 trials, 30 were unselected, and 8 were molecularly selected trials. The correlation between median PFS and median OS or between PFS HR and OS HR was generally low for most of the trials. There was no significant difference in OS between monotherapy and combination (median OS: 9.2 vs 9.1 months). Strongest correlation between median PFS and OS was observed from the molecularly selected population (r2= 0.626). The median OS was higher in molecularly selected patients than those in unselected (16.7 vs. 8.7 months). Conclusions: There is no significant correlation between median PFS and OS for most of trials except for studies in EGFR molecularly selected patients, indicating PFS could be a reasonable predictor of OS in molecularly selected patients for TKI therapies.