Construction Of A Risk Model For Colorectal Cancer Based On Immune Signatures For Prognostic Analysis

Colorectal cancer(CRC)is one of the most common cancers in china.There are many people die for colorectal cancer every year,and the deaths is still increasing.With the development of early immune tumor markers and immunotherapy technology,Some achievements have been made in its diagnosis and treatment,but the improvement of the OS rate is not obvious,and the prognosis is still poor.The main reason is that the molecular characteristics and overall immune landscape of colorectal cancer tumors are insufficiently studied,the selection of treatment methods needs to be optimized,and lack of effective biomarkers for prognostic judgment.Immunophenotyping can not only fully display the molecular characteristics of tumors and the overall immune landscape,but also can select more appropriate immunotherapy methods for patients with different levels of immunity,and prognostic markers based on immune genes can also more accurately predict the treatment effect of patients.Here,we download and organize the RNA-FPKM data and clinical characteristic data of colorectal cancer(CRC)samples from two public databases,TCGA and GEO,and enrich score the immune-related items of CRC patients through the enrichment method of ssGSEA.Based on immune enrichment results,CRC patients were divided into Immunity-H and Immunity-L,and next step we explored the heterogeneity between the two subtypes,verifying their differences in tumor microenvironment,immune cell infiltration,overall survival outcomes,and HLA and Check point differences in gene expression levels,analyze biological functions and metabolic mechanisms between subtypes,and provide reference for immunotherapy.In addition,based on the immune landscape of CRC,we established a risk score model based on immune signatures,which can effectively predict patient survival outcomes and response to immunotherapy,and this signature is correlated with clinical traits,immune cell infiltration,immune checkpoint genes.There is a significant correlation between the expression of,and the final comprehensive risk score and clinical characteristics can be used to timely and effectively predict the patient’s treatment effect through the nomogram,and predict the patient’s 1-year,3-year,and 5-year survival rates.Guided by a combined analysis of immune-infiltrating subtypes and immune-related prognostic features in CRC patients,insights into CRC typing and precise immunotherapy are provided.In this paper,the ssGSEA method was used to quantify the immune-related items in the samples,and the TCGA samples were effectively divided into 278 Immunity-H and 201 Immunity-L immune subtypes by using the HC algorithm,and the t-distributedstochastic neighbor embedding(t-SNE)analysis to estimate the clustering ability of the model.In order to verify the feasibility of the grouping strategy,we performed ESTIMATE analysis and CIBERSORT algorithm analysis to fully demonstrate the tumor microenvironment status and immune cell infiltration status of the two subgroups.The main purpose of our sample typing is to provide reference for clinical treatment,so we performed statistical analysis on the expression of HLA family and immune checkpoint genes in the two subtypes,and found the expression of immune checkpoint genes in different immune subtypes There was a significant difference(p<0.001),which provided an important reference for us to select appropriate therapeutic targets when performing immunotherapy on patients with different levels of immunity.In addition,we used the Kaplan-Meier(K-M)method for survival analysis to verify the effect of different degrees of immunophenotyping on survival.Through the enrichment of GO terms and KEGG pathways of GSEA,functional pathways and metabolic pathways that were significantly enriched in Immunity-H and Immunity-L groups were found.After fully demonstrating the effectiveness of immune typing,we wanted to find genes with significantly up expression between high and low immune subtypes.After screening,649 differential genes(DEGs)were obtained,and the differential genes were intersected with the immune genes downloaded by Imm Port.Differential immune genes(DEIRGs)were obtained,252 differentially expressed immune-related genes(DEIRGs)were identified,and GO and KEGG were performed on them to observe the functional characteristics and enrichment pathways of these genes.Based on the significantly differentially expressed immune genes screened between high and low immune subtypes,univariate COX regression and multivariate correlation analysis were performed on them,and 4(CX3CL1,CCL22,CHGB,FABP4)were screened out with the survival of CRC patients.Significantly related genes,based on these genes and their correlation with survival,get the risk score formula: Risk score=CX3CL1*0.35485+CCL22*(-0.74145)+CHGB*0.28134+FABP4*0.22572,build a risk score model,and use GEO data as independent verification.According to the scoring model,the risk score of the CRC patients studied was calculated,and the research samples of TCGA(Train cohort)and GEO(Test cohort)were divided into High risk and Low risk according to their median values,and the survival status and survival curve analysis of the Train cohort and the Test cohort were performed to evaluate the risk scoring model,and the ROC curve was drawn to verify the accuracy and validity of the model.The correlation between risk scores and clinical traits,tumor-infiltrating cells,and immune checkpoint genes was further analyzed,as well as the independent prognostic ability of the risk score model through univariate and multivariate analysis,and finally a nomogram was constructed by combining the risk score model and clinical traits.It is used to more accurately predict the survival rate of patients,and the prognostic ability of each factor is compared through the ROC curve,which provides an important reference for the prognosis analysis of clinical immunotherapy effects.