ABSTRACT: Neoadjuvant Chemotherapy (NAC) is not frequently used in ER-positive/HER2-negative breast cancer (BC) because around 10% patients achieve pathological complete response (pCR). Since NAC can result in cancer downstaging both in the breast and axilla and prevent a morbid surgery, thus a score to predict pCR in this population will be crucial to identify patients who can benefit from this approach. A total of 4038 patients from cohorts; GSE25066, GSE20194, Hess, GSE20181, TCGA-BRCA and METBRIC were analyzed. The score was generated by the 5 most highly expressed genes in the Hallmark E2F targets gene set amongst patients in the GSE25066 cohort with ER-positive/HER2-negative BC who achieved pCR. The area under the curve was significantly higher in the score than that for the E2F targets score. High score ER-positive/HER2-negative BCs were significantly associated with higher Nottingham pathological grade, AJCC cancer stage, MKI67 expression levels, intratumor heterogeneity, homologous recombination defects, mutation burden, neoantigen load, and infiltration of anti-cancer immune cells (CD4+, T helper type1, plasmacytoid dendritic cells, M1 macrophages). They also expressed lower abundance of stromal cells including fibroblasts, lymphatic endothelial cells, pericytes and adipocytes consistently in GSE25066, TCGA and METABRIC cohorts. All cell proliferation-related gene sets, G2M checkpoint, E2F targets, MYC targets v1 and v2, Mitotic Spindle, were strongly enriched in high score BCs consistently in 3 cohorts. The gene score was significantly associated with high pCR rate consistently in the GSE25066 (38%, P < 0.001), GSE20194 (16%, P = 0.006), and Hess cohort (23%, P = 0.037). In conclusion, the 5-gene score reflects cancer cell proliferation and immune cell infiltration, and predicts pCR after NAC in ER-positive/HER2-negative breast cancer.
