A literature-guided integrative transcriptomic analysis framework for prioritizing angiogenesis-associated genes across cancers.

Angiogenesis is a fundamental process in cancer progression, but computational prioritization of angiogenesis-associated genes often relies on predefined gene sets or canonical signaling pathways, limiting the identification of non-canonical and context-dependent regulatory programs. To address this limitation, we developed a literature-guided integrative transcriptomic framework to prioritize angiogenesis-associated candidates across heterogeneous cancer types. The framework integrates bibliometric trend profiling, pan-cancer gene expression analysis, survival association assessment, tumor microenvironment-related transcriptional characterization, and functional enrichment analysis. Rather than prespecifying angiogenic regulators, candidate genes were prioritized through convergence across these complementary evidence layers. Across diverse cancer cohorts, the framework consistently identified cadherin 2 (CDH2) as a representative non-canonical angiogenesis-associated signal. CDH2-associated transcriptional programs were enriched in extracellular matrix organization, cell-cell adhesion, and structural remodeling processes, and tumor microenvironment analyses indicated preferential alignment with stromal and vascular-associated features, particularly cancer-associated fibroblast-related signatures, rather than immune-dominant profiles. In independent external glioma cohorts, CDH2 showed a non-uniform and context-dependent relationship with vascular endothelial growth factor A (VEGFA)-associated transcriptional patterns, while CDH2-high tumors reproducibly exhibited higher extracellular matrix remodeling, adhesion/focal adhesion, epithelial-mesenchymal transition, and angiogenesis-related signature scores. In an independent Chinese Glioma Genome Atlas cohort, higher CDH2 expression was also associated with worse overall survival in a lower-grade glioma-preferred subset. Collectively, this study demonstrates the utility of a knowledge-guided integrative transcriptomic framework for prioritizing angiogenesis-associated genes beyond canonical regulators and highlights structure-oriented, non-canonical angiogenesis-related programs as reproducible biological signals.