Construction Of A Predictive Model For The Prognosis Of Skin Melanoma Based On The Immune Microenvironment

Objective:Bioinformatic analysis was used to screen the immune microenvironment-related genes associated with the prognosis of SKCM and construct a new risk score model to provide a theoretical basis for predicting the prognosis of SKCM patients and investigating new immunotherapeutic targets.Methods:We obtained SKCM transcriptomic data from The Cancer Genome Atlas(TCGA).The immune score-related genes were obtained by ESTIMATE algorithm,singlesample gene set enrichment analysis(ss GSEA),Kaplan-Meier(K-M)curve,weighted gene co-expression network analysis(WGCNA),and differential expression analysis were obtained for immune score-associated genes.Univariate Cox regression analysis and lasso operator(LASSO)were used to identify prognostic genes and establish risk profiles associated with immune scores.The receiver operating characteristic(ROC)curve further evaluated the reliability and sensitivity of the model.The GSE65904 dataset of the Gene Expression Omnibus(GEO)database was used as a validation set to verify the prognostic value of this risk scoring model.Cox regression analyses were used to identify independent prognostic factors for SKCM.Finally,the Tumor Immune Dysfunction and Exclusion(TIDE)algorithm was used to predict patient response to immunotherapy.Results:This study obtained 468 SKCM samples from the TCGA database for analysis.ESTIMATE analysis showed that immune score and stromal score were different in T stage and Stage(p < 0.01);ss GSEA and Pearson correlation analyses showed that eosinophils were positively correlated with T stage(p < 0.05),CD56 bright NK cells,CD56 dim NK cells,and T helper cells 17 were negatively correlated with T stage(p <0.05),and T helper cells 17 were positively correlated with N stage(p < 0.05).WGCNA analysis used immune and matrix scores as traits to construct a co-expression network,and K-M analysis showed that immune score was correlated with overall survival(OS)of patients(p < 0.001).Therefore,WGCNA selected the yellow module most correlated with immune score(cor = 0.87,p < 0.05)for subsequent analysis,and the yellow module had 299 key genes.The SKCM samples were divided into high and low immune scores according to the median immune score,and 633 differentially expressed genes(DEGs)were obtained by screening with |log2 fold change(FC)| > 1 and p < 0.05 as the threshold.There were 173 overlapping genes between the 633 DEGs and the 299 yellow module genes in WGCNA.The TCGA database samples with complete survival information of SKCM(n = 458)were randomly divided into training set(n = 321)and test set(n = 137)in the ratio of 7 : 3.Univariate and LASSO Cox analysis of 173 overlapping genes in the training set revealed that HLA-DRB1,XCL2,CCR1,HLADQB2 and DERL3 genes could construct a new risk model for predicting the prognosis of SKCM patients.K-M analysis showed that the overall survival(OS)of the high-risk group was significantly lower than that of the low-risk group(p < 0.001),and the areas under the ROC curve(AUC)at 1,3,and 5 years were 0.722,0.634,and 0.627,respectively.The K-M curves of the test set and the external GEO validation set(GSE65904,n =210)were consistent with the training set,and could effectively distinguish the survival difference between the high and low groups(p < 0.05)and the 1,3,and 5-year AUC values were all greater than 0.6.Univariate and multivariate Cox regression found that risk score,N stage,and T stage were independent prognostic factors in patients with SKCM,which constructed a nomogram model with C index = 0.705.In addition,the low-risk group had lower TIDE scores and higher immune checkpoint expression compared with the high-risk group.Conclusions:A novel prognostic prediction model constructed by HLA-DRB1,XCL2,CCR1,HLA-DQB2 and DERL3 genes based on bioinformatic analysis of the immune microenvironment of SKCM patients.The model had good predictive power for 1,3,and 5-year overall survival in SKCM patients.