Breast cancer molecular subclassification and estrogen receptor expression to predict efficacy of adjuvant anthracyclines-based chemotherapy: a biomarker study from two randomized trials

Background: The purpose of this study was to determine the predictive value of breast cancer molecular subclassification regarding the benefit of adjuvant anthracycline-based chemotherapy. Patients and methods: Tumor samples from 823 patients included in two randomized trials that compared an anthracycline-based chemotherapy with no treatment were used to construct a tissue array. Estrogen receptor (ER), Her2, epidermal growth factor receptor, cytokeratine 5/6 expressions were determined by immunohistochemistry (IHC). The potential predictive factors of treatment effect on disease-free survival (DFS) were assessed by interaction tests and multivariate analysis. Results: Sixty-four (8%), 98 (12%), 109 (14%) and 527 (66%) patients presented a Her2+/ER-, basal-like, Her2-/ER-/nonbasal and luminal-like breast cancer. ER expression, when assessed by IHC, was an independent predictive factor for the benefit of chemotherapy on DFS (test for interaction, P = 0.0015). The molecular subclassification significantly predicted the efficacy of chemotherapy (test for interaction, P = 0.01), but had no significant added value (P = 0.32) as compared to the ER by treatment interaction. Adjuvant chemotherapy was associated with an adjusted hazard ratio for relapse or death of 0.42 [95% confidence interval (Cl): 0.17-1.05], 0.54 (95% CI: 0.27-1.08), 0.35 (95% CI: 0.18-0.68), 1.07 (95% CI: 0.81-1.41) for patients with Her2+/ER-, basal-like, Her2-/ER-/nonbasal and luminal-like tumors, respectively. Conclusion: The breast cancer molecular subclassification was predictive for chemotherapy, efficacy in adjuvant setting, but did not provide significant additional information to ER.