Correlation of hormone receptor positive HER2-negative/MammaPrint high-2 breast cancer with triple negative breast cancer: Results from gene expression data from the ISPY2 trial.

573 Background: A subset of Hormone Receptor positive (HR+), HER2 negative (HER2-) breast cancers (BCs) classified by the MammaPrint (MP) assay as High-2 (MP-H2) show clinical behavior similar to triple-negative (TN) BC. The aim of this analysis was to assess molecular similarities between the HR+/HER2-/MP-H2 (HR+/MP-H2), HR+/HER2-/MP-High-1 (HR+/MP-H1), and TN/MP-H2 (TN/MP-H2) BCs. Methods: Gene expression data used for this analysis [GEO GSE194040] was derived from pre-treatment samples of the I-SPY2 trial, which is an ongoing multicentric randomized Phase II neoadjuvant platform trial for high-risk stage II or III BC. Six hundred eighty-four pretreatment samples were included in this analysis: 274 corresponded to HR+/MP-H1, 105 to HR+/MP-H2, and 305 to TN/MP-H2, accounting for 7 experimental plus a control arm. We used principal component and distance analyses, differentially expressed genes (analysis performed for pairwise comparisons of the 3 subtypes, a linear model, log fold-change and moderated-t p-values relative to a minimum fold-change threshold of fc = 1.5), gene set enrichment analysis using KEGG and GO, and Ingenuity Pathway Analysis, as well as 32 previously defined qualifying biomarkers to study similarities in biology between the subgroups. Results: On principal component and distance analyses, results reveal close similarity between HR+/MP-H2 and TN/MP-H2 indicating shared transcriptomic features. Gene expression analysis revealed that HR+/MP-H2 tumors resemble TN/MP-H2 more than HR+/MP-H1 cancers. Differential gene expression analysis revealed that HR+/MP-H2 and TN/MP-H2 cancers presented only two differentially expressed genes, compared to HR+/MP-H1 and HR+/MP-H2, which presented 577 differentially expressed genes. Upon pathway analysis, HR+/MP-H1 cancers presented the most significant upregulation in the glucocorticoid receptor signaling, DHCR24, and HOTAIR pathways, compared to HR+/MP-H2 cancers. Furthermore, HR+/MP-H2 and TN/MP-H2 shared enrichment of DNA repair and cell cycle pathways. In addition, various biomarkers were analyzed, and among the most representative shared biomarkers for HR+/MP-H2 and TN/MP-H2 were immune features, higher proliferation, and a lower expression of ESR1 co-expressed genes as compared to HR+/MP-H1 cancers. The rates of pathologic complete response were 41% for TN/MP-H2, 31% for HR+/MP-H2, and 11% HR+/MP-H1 cancers among all included patients. Conclusions: A subset of HR+ cancers defined by MP-H2 status closely resemble TN in molecular and clinical features. The similarity of pCR rates for MP-2 observed within either HR+/HER2- and TN is supported by our findings of similarity of active biology between HR+/MP-2 and TN/MP-2. These findings have implications for understanding the biology of HR+/MP-H2 BCs and optimizing treatment strategies.