A MyCAF-Based Signature for Prognosis, Immune Landscape, and Therapeutic Response in Pancreatic Cancer

Pancreatic ductal adenocarcinoma (PDAC) remains a highly aggressive cancer with a poor prognosis, driven in part by a complex tumor microenvironment where myofibroblastic cancer-associated fibroblasts (myCAFs) are key players. This study aimed to develop and validate a robust myCAF-related gene signature for prognostic stratification and therapeutic guidance in PAAD. Clinical and transcriptomic data from the TCGA-PAAD cohort were processed, and differentially expressed genes (DEGs) were identified, including 4222 up-regulated and 827 down-regulated genes. By intersecting DEGs with a myCAF gene set, we constructed a four-gene prognostic model (TOP2A, MKI67, COL7A1, MMP1) using LASSO-Cox regression. The model significantly stratified patients into high- and low-risk groups with distinct overall survival outcomes in the TCGA training set (p < 0.0001), a finding which was robustly validated in the independent ICGC-PACA-AU cohort (p = 0.0046). The high-risk group exhibited a significantly suppressed tumor immune microenvironment, characterized by reduced infiltration of critical T cells, including gamma delta T cells, CD8 + T cells, and resting memory CD4 + T cells, as well as monocytes. Crucially, drug sensitivity analysis revealed that the high-risk group demonstrated reduced sensitivity to several targeted therapies, including Afatinib, Dasatinib, Lapatinib, and Trametinib. In conclusion, we have established and validated a novel four-gene prognostic signature based on myCAF biology. This model not only serves as an independent prognostic indicator but also provides insights into the immune landscape and potential therapeutic vulnerabilities, offering a valuable framework for personalized treatment strategies in pancreatic cancer.