Application Of SsGSEA In Bladder Cancer And Establishment Of Prognostic Model

Objective:Bladder cancer（BLCA）is an aggressive malignant tumor with high recurrence rate.The steady increase of its incidence makes BLCA one of the most common urogenital tract cancers in the world.The prognosis of BLCA is usually poor when it progresses to the middle or late stage.Studies have found that there are many factors related to the prognosis of BLCA,among which age and grade are the most important factors affecting the overall survival of patients^[1].At present,there is still no consensus on the relationship between the prognosis of BLCA patients and immune-related genes.Immune-related genes are widely used in BLCA and are closely related to the treatment and prognosis of BLCA.Immune-related genes related to the prognosis of BLCA can be obtained through single sample gene set enrichment analysis（ss GSEA）.Therefore,it is of great clinical significance to study the relationship between immune-related genes and BLCA and establish a prognostic model.Methods:Through bioinformatics,BLCA transcriptome data,clinical data and tumor mutation data were downloaded from tumor Genome Atlas（TCGA）database,and the BLCA transcriptome data were preprocessed,and then the data were enriched and analyzed by ss GSEA to obtain immune typing files.The relationship between tumor microenvironment,immune cell infiltration and HLA genes was evaluated.Then the differential genes were obtained by GSEA4.1.0 enrichment analysis,immune-related genes were obtained by IMMPORT database,and the differential immune-related genes were screened out by R software combined with expression data files and typing results files,and visualized.The intersection of immune-related genes with differences between TCGA transcriptome data and typing was selected,and the prognostic immune-related genes were screened by R software combined with survival data files.Based on immune-related genes associated with prognosis,the protein interaction network（PPI）was constructed using interactive gene retrieval tool（STRING）,and the key modules and core genes of PPI network were screened by Cytocape3.8.2 software.Then univariate and multivariate Cox regression analysis was performed to identify the independent prognostic immune-related genes in BLCA,and the prognostic model was constructed.In order to verify the effectiveness of the model,all patients were randomly divided into the training group（Train group）and the verification group（Test group）by1:1,and the risk score was calculated.According to the median risk score,the Train group and Test group were divided into high risk group and low risk group respectively.Then the validity and reliability of the model are verified from the survival curve,ROC curve,calibration curve and risk curve.After the reliability of the model was verified,R software was used to further evaluate the relationship between immune-related genes and immune cells,tumor mutational burden（TMB）and high and low risk groups,and evaluate whether risk score was an independent prognostic factor of BLCA.Finally,based on immune-related genes with independent prognosis of BLCA,the rograms were constructed to predict the overall survival rate of BLCA.Results:Complete gene expression data of 433 patients were obtained from TCGA database,including 19 normal bladder tissue samples,414 BLCA tissue samples,complete clinical information of 412 patients,4 groups of tumor mutation data.High and low immune group files were obtained by ss GSEA enrichment analysis,and ion pathway enrichment analysis of differential genes was conducted by GSEA software to explore the possible key ion pathway of BLCA occurrence.IMMPORT database was used to download immune-related genes,and R software was used to screen a total of476 different immune-related genes.After combining the gene with BLCA survival data,32 immune-related genes were obtained by univariate Cox regression analysis.PPI and its sub-network were constructed,and 5 core regulatory genes including STAT1,STAT5A,GATA3,PPARG and CEBPB were obtained.Then 32 immune-related genes were analyzed by multivariate Cox regression analysis,and the best prediction models of 11 immune-related genes were obtained,including PDGFC,GNLY,AGER,NRP2,PPARG,GMFG,CCL17,SRC,S100A10,IL9R and CTSE.Multiple indicators were verified based on the prediction model.In the ROC curve,the area under curve（AUC）of Train group in 1,3 and 5 years were 0.737,0.773 and 0.761,respectively,showing good verification ability.In the Test group,the AUC of 1 year,3 years and 5 years were0.691,0.637 and 0.679,respectively,which also verified the predictive value of the model.Similarly,survival curve,risk curve and other aspects also confirmed the prediction ability of the model.Then,correlation analysis showed that immune-related genes were differentially expressed in some immune cells and had significant correlation,which may play an important role in immune cells.In terms of the influence of high and low TMB and high and low risk group on the overall survival rate（OS）of BLCA,h-TMB-low risk group showed the best survival rate.Risk score was further confirmed as an independent prognostic factor for BLCA.Finally,a rograph was constructed based on the risk scores of 11 immune-related genes,which could better predict the 1-year,3-year and 5-year overall survival rates of BLCA patients.Conclusions:In this study,5 core regulatory genes were obtained by bioinformatics method through the analysis of BLCA related data in TCGA database,which may become potential therapeutic targets.In addition,11 immune-related genes related to BLCA prognosis,including PDGFC,GNLY,AGER,NRP2,PPARG,GMFG,CCL17,SRC,S100A10,IL9R and CTSE,were obtained,and multiple indicators verified the validity of the model and established the best prediction model.This study may contribute to the search for potential biomarkers of BLCA and a more accurate assessment of the prognosis of BLCA patients.