Integrated Multiomics Analysis Identifies PHLDA1+ Fibroblasts as Prognostic Biomarkers and Mediators of Biological Functions in Pancreatic Cancer

Background: Pancreatic cancer (PC) is marked by extensive heterogeneity, posing significant challenges to effective treatment. The tumor microenvironment (TME), particularly cancer-associated fibroblasts (CAFs), plays a critical role in driving PC progression. However, the prognostic and functional contributions of distinct CAF subtypes remain inadequately understood. Here, we introduce a novel 7-gene risk model that not only robustly stratifies PC patients but also unveils the unique role of PHLDA1 as a key mediator in tumor-stroma crosstalk.Methods: By integrating single-cell RNA sequencing (scRNA-seq), spatial transcriptomics, and bulk RNA sequencing data, we comprehensively characterized the heterogeneity of CAFs in PC. We identified five CAF subtypes and focused on matrix CAFs (mCAFs), which were strongly associated with poor prognosis. A 7-gene mCAF-associated risk model was constructed using advanced machine learning algorithms, and the biological significance of PHLDA1 was validated through co-culture experiments and pan-cancer analyses.Results: Our multiomics analysis revealed that the novel 7-gene model (comprising USP36, KLF5, MT2A, KDM6B, PHLDA1, REL, and DDIT4) accurately predicts patient survival, immunotherapy response, and TME status. Notably, PHLDA1 was uniquely overexpressed in CAFs and correlated with the activation of key protumorigenic pathways, including EMT, KRAS, and TGF-β, underscoring its central role in modulating the crosstalk between CAFs and malignant ductal cells.Pan-cancer analysis further supported PHLDA1's prognostic and immunomodulatory significance across multiple tumor types.Our study presents a novel 7-gene prognostic model that significantly enhances risk stratification in PC and identifies PHLDA1+ CAFs as promising prognostic biomarkers and therapeutic targets. These findings provide new insights into the TME of PC and open avenues for personalized treatment strategies.