An integrated breast cancer transcriptomic atlas defines clinical subtypes and informs cellular origin

To address this, we adapted single cell RNA sequencing analytical strategies to harmonize 7,612 bulk breast cancer transcriptomes from 22 datasets, generating a high resolution atlas that resolves both known and previously unrecognized malignant transcriptional states. This framework identified existing intrinsic subtypes, microenvironmental programs, and newly identified stem like, basal-luminal progenitor (BLP), mucinous, and neuroendocrine differentiated (NED) subtypes, each displaying distinct clinical behaviors.

The NED subtype exhibited luminal features, neuroendocrine marker expression, and associated with chemo-endocrine therapy poor prognosis. The HER2E-kinase inhibitor sensitive (HER2EKIS) subtype identified patients who achieved high pCR rate from HER2-targeted kinase inhibitors. Transcriptomic positioning of BLP tumors suggests a bipotent or myoepithelial progenitor association, limited responsiveness to chemotherapy, and potential benefit to VEGF and receptor tyrosine kinase inhibitors.