Integrative Bulk and Single-Nucleus Analyses Nominate COL5A2 as a CAF/ECM-Associated Marker Associated with PDAC Progression

Background/Objectives: Pancreatic ductal adenocarcinoma (PDAC) is characterized by an extensive desmoplastic microenvironment; however, reproducible stromal-associated biomarkers linked to disease progression remain limited. This study therefore aimed to identify and validate a biologically relevant stromal/extracellular matrix (ECM)-associated candidate biomarker for PDAC. Methods: Three GEO bulk transcriptomic PDAC cohorts (GSE15471, GSE16515, and GSE62452) were integrated for differential expression, functional enrichment, protein–protein interaction, and hub-gene analyses. Candidates identified as a promising biomarker were further evaluated using the following: public proteomic and survival resources; head-to-head receiver operating characteristic (ROC) comparisons against COL1A1, COL3A1, and COL5A1; a progression cohort (GSE43288); and single-nucleus RNA sequencing data (GSE202051). Results: Among 206 shared differentially expressed genes, COL5A2 was the only consensus hub retained across multiple network-ranking methods. COL5A2 protein expression was found to be elevated in tumor tissue and associated with worse overall and disease-free survival. In ROC analyses, COL5A2 exhibited stable tumor-versus-non-tumor discrimination across GSE15471, GSE16515, and GSE62452 (AUC = 0.932, 0.760, and 0.782, respectively) and significantly outperformed COL3A1 in two cohorts. In GSE43288, COL5A2 expression increased along the normal–pancreatic intraepithelial neoplasia–PDAC axis and remained positively associated with ECM and cancer-associated fibroblast (CAF) signature scores after adjustment for disease group. Reanalysis of GSE202051 restricted to the original 18 untreated PDAC specimens revealed that COL5A2 expression was concentrated in fibroblast-lineage compartments, with CAFs accounting for the largest overall contribution and myCAFs demonstrating the strongest per-specimen expression enrichment. Conclusions: COL5A2 is a reproducible stromal/ECM-associated candidate biomarker linked to PDAC progression, with predominant expression in fibroblast/CAF compartments.