Abstract
EGFR and cMET cross-talk is involved in breast cancer (BC) progression and resistance to different targeted therapies, however little is known about the co-expression patterns of EGFR and cMET or its prognostic significance in BC. Protein levels of EGFR, cMET and their phosphorylated proteins were measured in 825 BC samples using reverse phase protein array (RPPA). Given unimodal distribution of proteins, the median was selected as a cut-off after sensitivity analyses. Kaplan-Meier survival curves were used to estimate relapse-free (RFS) and overall survival (OS). Cox-proportional hazards models were utilized to determine associations between EGFR and cMET with outcomes. Mean age was 58 years with 457 (55%) hormone receptor (HR) positive, 211 (26%) triple-negative (TN) and 148 (18%) HER2 positive tumors (HER2+). HER2+ was associated with higher EGFR expression and phosphorylation, compared to HR and TN (p<0.05). High EGFR expression was associated with higher phosphorylated-cMET (p-cMET) but not cMET (ANOVA p-cMET p < 0.001; cMET p = 0.34). The same association was found with high phosphorylated-EGFR (p-EGFR) group at Tyr992 and Tyr1068 (both p < 0.001). High expressions in either of two p-EGFRs were linked with higher cMET as well (all p<0.001). For the TN subtype, high expression in EGFR and p-EGFR at Tyr992 but not at Tyr1068 was associated with higher p-cMET (p<0.00, p = 0.012, p = 0.4 respectively). Only high expression in p-EGFR at Tyr992 was linked with higher expression of cMET (p = 0.02). In contrast, among HER2 subtype, high expression in p-EGFR at Tyr1068 but not at Tyr992 was associated with higher cMET and p-cMET (cMET p = 0.023;p-cMET p<0.001). Four subgroups of patients defined by dichotomized EGFR/p-EGFR and cMET/p-cMET level demonstrated no significant differences in survival. In multivariate analyses, neither cMET nor EGFR expression/activation was found to be an independent prognostic factor in survival outcome.