Background
Pancreatic cancer is one of the most common malignant tumors of the digestive tract, and it has a poor prognosis. Traditional methods are not effective to accurately assess the prognosis of patients with pancreatic cancer. Immunotherapy is a new promising approach for the treatment of pancreatic cancer; however, some patients do not respond well to immunotherapy, which may be related to tumor microenvironment regulation. In this study, we use gene expression database to mine important immune genes and establish a prognostic prediction model for pancreatic cancer patients. We hope to provide a feasible method to evaluate the prognosis of pancreatic cancer and provide valuable targets for pancreatic cancer immunotherapy.
Results
We used univariate COX proportional hazard regression analysis, the least absolute shrinkage and selection operator, and multivariate COX regression analysis to screen 8 genes related to prognosis from the 314 immune-related genes, and used them to construct a new clinical prediction model in the TCGA pancreatic cancer cohort. Subsequently, we evaluated the prognostic value of the model. The Kaplan–Meier cumulative curve showed that patients with low risk scores survived significantly longer than patients with high risk scores. The area under the ROC curve (AUC value) of the risk score was 0.755. The univariate COX analysis showed that the risk score was significantly related to overall survival (HR 1.406, 95% CI 1.237–1.598, P < 0.001), and multivariate analysis showed that the risk score was an independent prognostic factor (HR 1.400, 95% CI 1.287–1.522, P < 0.001). Correlation analysis found that immune genes are closely related to tumor immune microenvironment.
Conclusions
Based on the TCGA-PAAD cohort, we identified immune-related markers with independent prognostic significance, validated, and analyzed their biological functions, to provide a feasible method for the prognosis of pancreatic cancer and provide potentially valuable targets for pancreatic cancer immunotherapy.