Objective To screen key prognosis-related immune genes(KIGs) in colon cancer(CC) and construct immune-related prognostic model.

Methods The gene expression data and clinical data of CC patients were downloaded from the Cancer Genome Atlas (TCGA), differential expressed genes (DEGs) were screened based on immune typing, and the intersection of DEGs and immune genes of ImmPort database was used to obtain differentially expressed immune genes (DEIGs). Univariate Cox analysis and Lasso-Cox analysis were used to screen KIGs to establish immune-related prognostic models. Next, the predictive performance of the prognostic model was verified. The correlations of risk score with clinical features, immune infiltrating cells, tumor mutation burden, and microsatellite instability were analyzed. Independent prognostic factors of CC were screened by univariate and multivariate Cox analysis, and a nomogram related to overall survival rate (OS) was constructed to predict the overall survival rate of CC patients.

Results A total of 1, 439 DEGs were screened out based on immune typing, and 379 DEIGs were obtained by the intersection of DEGs and immune genes of ImmPort database. A total of 12 KIGs were screened out by univariate Cox analysis and Lasso-Cox analysis, and an immune-related prognostic model was constructed. There was a significant difference in OS between high-risk score patients(high-risk group) and low-risk score patients (low-risk group)(P < 0.05). The area under the receiver operating characteristic curve (AUC) for 1-, 3- and 5-year were 0.735, 0.725 and 0.699, respectively, which showed that the model had a good predictive effect. There were significant differences in risk scores of patients with different disease stages and different TNM stages (P < 0.05); KIGs was correlated with numerous immune infiltrating cells (P < 0.05); micro-satellite instability was greater in the high-risk group than in the low-risk group. Univariate and multivariate Cox regression analysis showed that risk score was an independent prognostic factor for CC (P < 0.001), and the further constructed nomogram had good predictive power and accuracy for the survival status of CC patients.

Conclusion An immune-related prognostic model based on 12 KIGs is constructed in the study, and a nomogram is developed to predict overall survival in CC patients, which could help physicians make individualized treatment decisions.