A systematic review and meta-analysis of individual patient data on the impact of the BIM deletion polymorphism on treatment outcomes in epidermal growth factor receptor mutant lung cancer

Abstract

Background: A germline deletion in the BIM (BCL2L11) gene has been shown to impair the apoptotic response to tyrosine kinase inhibitors (TKIs) in vitro but its association with poor outcomes in TKI-treated non-small cell lung cancer (NSCLC)patients remains unclear. We conducted a systematic review and meta-analysis on both aggregate and individual patient data to address this issue. Results: In an aggregate data meta-analysis (n = 1429), the BIM deletion was associated with inferior PFS (HR = 1.51, 95%CI = 1.06-2.13, P = 0.02). Using individual patient data (n = 1200), we found a significant interaction between the deletion and ethnicity. Amongst non-Koreans, the deletion was an independent predictor of shorter PFS (Chinese: HR = 1.607, 95%CI = 1.251-2.065, P = 0.0002; Japanese: HR = 2.636, 95%CI = 1.603-4.335, P = 0.0001), and OS (HR = 1.457, 95% CI = 1.063-1.997, P = 0.019). In Kaplan-Meier analyses, the BIM deletion was associated with shorter survival in non-Koreans (PFS: 8.0 months v 11.1 months, P < 0.0005; OS: 25.7 v 30.0 months, P = 0.042). In Koreans, the BIM deletion was not predictive of PFS or OS. Materials and Methods: 10 published and 3 unpublished studies that reported survival outcomes in NSCLC patients stratified according to BIM deletion were identified from PubMed and Embase. Summary risk estimates were calculated from aggregate patient data using a random-effects model. For individual patient data, Kaplan-Meier analyses were supported by multivariate Cox regression to estimate hazard ratios (HRs) for PFS and OS. Conclusions: In selected populations, the BIM deletion is a significant predictor of shorter PFS and OS on EGFR-TKIs. Further studies to determine its effect on response to other BIM-dependent therapeutic agents are needed, so that alternative treatment strategies may be devised.