Research Paper Volume 13, Issue 21 pp 24313—24338

Establishment of a tumor immune microenvironment-based molecular classification system of breast cancer for immunotherapy

Identified immune-related subclasses of patients with breast cancer. (A) Consensus-clustered heatmap on the TCGA-BRCA cohort (n = 1,104) based on the example genes of immune factors selected by NMF, and refine it through the multidimensional scaling random forest to define the immune class (407/1104, 36.9%, pink bar). Immune-related signatures were compared between immune and non-immune classes. Red and blue bars correspond to high and low scores of each signature, respectively. (B) Representative images of CD3, CD19, and CD163 (markers of T, B, and myeloid cells, respectively) staining in the immune and non-immune classes. Scale bar, 100 μm. Abbreviations: IFN: interferon; TLS: tertiary lymphoid structure; CYT: cytolytic activity score; NMF: non-negative matrix factorization; TCGA–BRCA: The Cancer Genome Atlas-Breast Cancer.

Figure 1. Identified immune-related subclasses of patients with breast cancer. (A) Consensus-clustered heatmap on the TCGA-BRCA cohort (n = 1,104) based on the example genes of immune factors selected by NMF, and refine it through the multidimensional scaling random forest to define the immune class (407/1104, 36.9%, pink bar). Immune-related signatures were compared between immune and non-immune classes. Red and blue bars correspond to high and low scores of each signature, respectively. (B) Representative images of CD3, CD19, and CD163 (markers of T, B, and myeloid cells, respectively) staining in the immune and non-immune classes. Scale bar, 100 μm. Abbreviations: IFN: interferon; TLS: tertiary lymphoid structure; CYT: cytolytic activity score; NMF: non-negative matrix factorization; TCGA–BRCA: The Cancer Genome Atlas-Breast Cancer.