Lipid Metabolism-Related Gene Markers Used for Prediction Prognosis, Immune Microenvironment, and Tumor Stage of Pancreatic Cancer

Recently, more and more evidence shows that lipid metabolism disorder has been observed in tumor, which impacts tumor cell proliferation, survival, invasion, metastasis, and response to the tumor microenvironment (TME) and tumor treatment. However, hitherto there has not been sufficient research to demonstrate the role of lipid metabolism in pancreatic cancer. This study contrives to get an insight into the relationship between the characteristics of lipid metabolism and pancreatic cancer. We collected samples of patients with pancreatic cancer from the Gene Expression Omnibus (GEO), the Therapeutically Applicable Research to Generate Effective Treatments (TARGET), and the International Cancer Genome Consortium (ICGC) databases. Firstly, we implemented univariate regression analysis to get prognosis-related lipid metabolism genes screened and a construction of protein–protein interaction (PPI) network ensued. Then, contingent on our screening results, we explored the molecular subtypes mediated by lipid metabolism-related genes and the correlated TME cell infiltration. Additionally, we studied the disparately expressed genes among disparate lipid metabolism subtypes and established a scoring model of lipid metabolism-related characteristics using the least absolute shrinkage and selection operator (LASSO) regression analysis. At last, we explored the relationship between the scoring model and disease prognosis, tumor stage, tumor microenvironment, and immunotherapy. Two subtypes, C1 and C2, were identified, and lipid metabolism-related genes were studied. The result indicated that the patients with subtype C2 have a significantly lower survival rate than that of the patients with subtype C1, and we found difference in abundance of different immune-infiltrating cells. The gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses revealed the association of these differentially expressed genes with functions and pathways related to lipid metabolism. Finally, we established a scoring model of lipid metabolism-related characteristics based on the disparately expressed genes. The results show that our scoring model have a substantial effect on forecasting the prognosis of patients with pancreatic cancer. The lipid metabolism model is an important biomarker of pancreatic cancer. Using the model, the relationship between disease prognosis, molecular subtypes, TME cell infiltration characteristics, and immunotherapy in pancreatic cancer patients could be explored.