Machine learning analysis reveals tumor heterogeneity and stromal-immune niches in breast cancer
Junjie Kuang, Guo-Fang Zhong, Linfeng Zhao, Xia Yuan, Yundong Zhou, Jun Li
Abstract
Breast cancer is a leading cause of cancer-related mortality, with tumor heterogeneity and drug resistance posing significant challenges to treatment. We integrated single-cell RNA sequencing, spatial transcriptomics, and bulk RNA-seq deconvolution to analyze BRCA samples. Our analysis identified 15 major cell clusters, including neoplastic epithelial, immune, stromal, and endothelial populations. Notably, low-grade tumors showed enriched subtypes, such as CXCR4+ fibroblasts, IGKC+ myeloid cells, and CLU+ endothelial cells, with distinct spatial localization and immune-modulatory functions. These subtypes were paradoxically linked to reduced immunotherapy responsiveness, despite their association with favorable clinical features. High-grade tumors exhibited reprogrammed intercellular communication, with expanded MDK and Galectin signaling. Bulk RNA-seq deconvolution further supported the prognostic significance of low-grade-enriched subtypes. Our findings highlight the heterogeneity of the tumor microenvironment and provide new insights into immune evasion and therapeutic resistance in breast cancer.