Multi-dimensional structure and subtype-specific heterogeneity of angiogenesis in breast cancer

Background: The prognostic relevance of angiogenesis in breast cancer remains unclear when modeled as a coordinated gene program rather than individual gene effects. This study aimed to characterize the multidimensional structure of angiogenesis and assess the prognostic relevance of its latent components for overall survival in breast cancer. Methods: Using RNA sequencing data from 1,096 patients with The Cancer Genome Atlas Breast Invasive Carcinoma (TCGA-BRCA) from the United States, we summarized 36 angiogenesis-related genes via principal component analysis (PCA) applied to standardized log-counts per million values. The first three components (Z1−Z3), retained based on variance explained, scree plot inspection, and interpretable gene loading patterns, were evaluated for association with overall survival using Kaplan-Meier and Cox proportional hazards models adjusted for age and molecular subtype. Results: The first three components captured a multi-dimensional structure of angiogenesis activity. None was significantly associated with overall survival in multivariable Cox models (Z1: HR = 1.016, 95% CI: 0.957–1.078, p = 0.603; Z2: HR = 1.049, 95% CI: 0.954–1.153, p = 0.319; Z3: HR = 1.007, 95% CI: 0.918–1.105, p = 0.878), and Kaplan–Meier analyses showed no significant survival differences across median-based groups. Functional enrichment analyses indicated that the retained components represented distinct biological dimensions of angiogenesis. Subtype-specific analyses showed no statistically significant associations within individual subtypes and no strong evidence of effect modification. However, marked differences in latent component distributions across molecular subtypes revealed substantial subtype-specific heterogeneity. Collectively, these findings indicate limited independent prognostic relevance for overall survival despite substantial biological heterogeneity and subtype-specific variation. Conclusions: Angiogenesis is a multi-dimensional and biologically heterogeneous process. No latent angiogenesis component demonstrated independent prognostic value for overall survival, indicating that angiogenesis is insufficient as a standalone prognostic biomarker. Nevertheless, subtype-specific variation across latent dimensions highlights its relevance for understanding biological heterogeneity in breast cancer.