Development of a metastasis-related genes prognostic model for colorectal cancer

Abstract BackgroundThe colorectal tumor is a malignant tumor, the most important cause of death from colorectal cancer is metastasis. The primary objective of this study was to build a prognosis model by analyzing the differential genes between metastatic and non-metastatic CRC; Create a corresponding nomogram for clinical risk assessment of prognostic indicators to facilitate interpretation.MethodsIn this study, the main data on colorectal tumors are obtained from TCGA data set. Then, we used univariate COX regression analysis & Lasso regression analysis to find and filter characteristic variables. We construct the Cox proportional hazards regression model and verified the stability of the model by GEO external data set, the nomograms based on MR-risk score were obtained for clinical use. Finally, we explored the underlying mechanism of MR-risk score and colorectal cancer prognosis by using immune infiltration and immune checkpoint analysis.ResultsIn this study, the prognostic model of colorectal cancer patients was constructed based on the colorectal metastasis-related genes, and the correlation between MR-risk core and the prognosis of colorectal cancer patients was obtained. In TCGA database, AUC of training set was 0.72, and the AUC of the testing set was 0.76, at the same time, AUC of GEO external data set was 0.68, which proved that treatment and prognosis of patients could be effectively determined. Finally, we found that the immune cells infiltrated in the tissues of patients with high MR-risk score were mostly in immune static or inactivated state compared with those of patients with low MR-risk score. ConclusionsIn summary, MR-risk score has a direct correlation with the prognosis of colorectal cancer, It is useful for predicting prognosis and immunotherapy of colorectal cancer.