Quantitative HER2 protein levels predict outcome in fluorescence in situ hybridization‐positive patients with metastatic breast cancer treated with trastuzumab

Abstract

Only a portion of breast cancer patients currently selected for trastuzumab therapy respond. Using a novel assay (HERmark) to quantify total human epidermal growth factor receptor 2 (HER2) expression, the authors examined outcomes in 102 trastuzumab-treated metastatic breast cancer patients previously assessed as immunohistochemistry (IHC) 3+ by local but not central IHC, or fluorescence in situ hybridization (FISH) positive, and then retested by central FISH. Of 102 MBC patients previously scored as IHC 3+ or 2+/FISH-positive and treated with trastuzumab-containing regimens, 98 had both central FISH and HER2 total expression values. Sixty-six of 76 central FISH-positive patients (87%) had high HER2 total expression levels (concordant positive), and 19 of 22 central FISH-negative patients (86%) were HER2 total expression low (concordant negative). Fourteen percent (3 of 22) of central FISH-negative patients were HER2 total expression high (discordant HER2 total expression high), and 13% (10 of 76) of central FISH-positive patients were HER2 total expression low (discordant HER2 total expression low). The concordant positive group had a significantly longer time to progression (TTP, median = 11.3 months) compared with the concordant negative group (median TTP, 4.5 months; hazard ratio [HR] = 0.42, P < .001), and also compared with the discordant HER2 total expression low group (median TTP, 3.7 months; HR = 0.43, P = .01). The discordant HER2 total expression low group behaved similarly compared with concordant negatives (HR = 1, P = .99). In analyses restricted to central FISH-positive patients only (n = 77), Cox proportional hazards multivariate regression identified HER2 total expression as an independent predictor of TTP (HR = 0.29, P = .0015) and overall survival (HR = 0.19, P < .001). A subset of patients with HER2 gene amplification by FISH express low levels of HER2 protein and have reduced response to trastuzumab-containing therapy, similar to FISH-negative patients. This cohort represents a training dataset, and the observed relationships and derived cutoffs require validation in an independent cohort of trastuzumab-treated metastatic breast cancer patients.