Ligands of EGFR and the insulin-like growth factor family as serum biomarkers for response to EGFR-inhibitors in patients with advanced NSCLC

Abstract

8100 Background: To determine whether concentrations of transforming growth factor α (TGFa), amphiregulin (ARG), both ligands of the epidermal growth factor receptor (EGFR), insulin-like growth factor 1(IGF1) or IGF-binding protein 3 (IGFBP3) measured in serum of advanced non-small cell lung cancer (NSCLC) patients are predictive of EGFR-inhibitors (EGFRi) response. Methods: We assessed serum levels of marker candidates using ELISA (TGA and ARG) and chemiluminescent (IGF1 and IGFBP3) assays. Sixty-one advanced NSCLC patients treated with EGFRi (gefitinib or erlotinib, >14 days) were matched for gender, age and histology to a control group of 63 EGFRi-untreated advanced NSCLC patients. We dichotomized marker levels at the 20th, 50th or 80th percentile and evaluated whether the effect of EGFRi treatment on overall survival (OS) differed by marker level based on multivariate proportional hazards regression with an interaction term. We adjusted for gender, smoking, stage, histology and prior chemotherapy. Results: While 6-months OS did not appreciably differ between 50 EGFRi treated and 50 control patients whose TGFa levels were below the 80th percentile (39% vs. 54%, multivariate HR 0.78, 95% CI 0.45–1.34, p=0.359), it was substantially worse for 11 EGFRi treated patients compared with 13 control patients whose TGFa levels were high (9% vs. 42%, multivariate HR 2.38, 95% CI 0.87–6.52, p=0.092). The difference of EGFRi effects by TGFa level was statistically significant (interaction p=0.033). There was no evidence that EGFRi treatment effects differed by levels of ARG, IGF1 and IGFBP3. Patients with high concentrations of IGFBP3 (above the median) had significantly longer OS than patients with low IGFBP3 concentrations, independent of treatment (HR 0.46, 95% CI 0.29–0.73, p=0.001). Conclusions: This is the first study in NSCLC patients of Caucasian origin in which serum concentrations of TGFa are predictive for EGFR-inhibitor response, suggesting this is a potential predictive marker for EGFRi treatment. Furthermore, we coincidentally found that levels of IGFBP3 are predictive for overall survival, indicating this might be a prognostic factor in advanced NSCLC patients. [Table: see text]