Balázs Győrffy
Abstract Background and Aim The tumor microenvironment critically influences cancer progression and patient outcome, yet the interaction between immune cell composition and prognostic gene expression patterns remains incompletely characterized. Leveraging the large-scale Kaplan–Meier Plotter (KM-plotter) breast cancer database, this study aimed to systematically identify druggable, survival-associated genes across immune-defined breast cancer subgroups and to uncover shared and context-specific biological processes. Methods Transcriptome-wide survival analysis was performed in GEO-derived breast cancer samples using Cox proportional hazards regression with an optimized cutoff approach. Tumors were stratified based on gene expression signatures for CD4+ memory T cells, CD8+ T cells, macrophages, and regulatory T cells immune cell enrichment. Genes significantly associated with relapse-free survival (false discovery rate below 1% and hazard rate over one) were identified within each subgroup, followed by gene set enrichment analysis for the cancer hallmarks and integration with druggability annotations from the Drug–Gene Interaction Database. Results The entire dataset included 4,944 breast tumor specimens. A total of 855, 682, 1,024, and 893 genes were significantly associated with poor prognosis in CD4 memory T cell–, CD8+ T cell–, macrophage-, and Treg-enriched tumors, respectively. The strongest prognostic genes included SNAPC1, KRS1, and CENPA in CD4 memory T cell–high tumors; CCNB2, CENPA, and CDC20 in CD8+ T cell–high tumors; CENPA, CCNB2, and CCNE2 in macrophage-enriched tumors; and KIF20A, AURKA, and MELK in Treg-high tumors. These top-ranked genes demonstrated strong effect sizes, with hazard ratios up to 2.54 in the Treg-high subgroup. Despite differences in immune context, hallmark analysis consistently identified genome instability (adjusted p < 1×10E-06) and reprogramming of energy metabolism (adjusted p < 1×10E-06) as the most significantly enriched processes. Druggable targets included ETFB and FKBP1A (mTOR inhibition), IFNAR2 (interferon receptor activation), RBX1 (E3 ligase inhibition), and VEGFA (anti-angiogenic blockade). Conclusions Immune cell–specific stratification reveals distinct sets of prognostic genes in breast cancer; however, these converge on common biological hallmarks dominated by proliferative and metabolic processes. Notably, all the identified top druggable targets are overexpressed in poor-prognosis tumors, supporting targeted therapeutic inhibition as a rational strategy within diverse immune microenvironments.