Screening Prognostic Risk Markers For Head And Neck Squamous Cell Carcinoma Based On The Expression Profile Of Epithelial-mesenchymal Transition Genes

Objective:Head and neck squamous cell cancer（HNSCC）is one of the most common types of cancer worldwide.There have been many reports suggesting that biomarkers explored via database mining plays a critical role in predicting HNSCC prognosis.However,a single biomarker for prognostic analysis is not adequate.Hence,there is growing evidence indicating that gene signature could be a better choice for HNSCC prognosis.Methods:1.Download mRNA data and clinicopathological parameters related to head and neck squamous cell carcinoma in TCGA（Human tumor Genome Database,The Cancer Genome Atlas）).2.By（GSEA）,analysis,we found that the gene set HNSCC involved in epithelial-mesenchymal transformation（EMT）is closely related to the occurrence and development.The gene set was further screened under the condition of normalized P value（P<0.05）.3.To screen the differentially expressed genes in cancer and paracancerous tissues.4.Univariate and multivariate Cox regression models were used to screen the relationship between m RNAs and prognosis,and molecular tags based on mRNAs expression were constructed to predict the prognosis of patients.5.K-M curve predicts the ability of molecular tags to predict the prognosis of patients with head and neck squamous cell carcinoma.6.The patients were randomly divided into two groups:the training set and the test set,and the efficacy of mRNAs molecular tags in predicting the prognosis of patients was tested respectively.7.Univariate and multivariate Cox regression analysis was used to analyze the correlation between molecular tags and clinicopathological parameters.8.Functional annotation and pathway enrichment analysis of all differentially expressed genes in each group were performed by DAVID online analysis tool.Results:1.A total of 494 HNSCC patients were obtained from TCGA with clinical features and expression data sets of 57,072 genes.The EMT gene set obtained by GSEA gene set analysis contains 197 mRNAs for further analysis.First,univariate Cox regression analysis was used to evaluate the relationship between the expression levels of 197 differentially expressed mRNAs（DEMis）and patient overall survival（OS）screened in the previous study,and 12 mRNAs were found to be significantly correlated with OS（P<0.05）.2.A stepwise multivariate Cox regression analysis was performed,and the previous12 mRNAs were selected to establish a prediction model.The prediction model is defined as a linear combination of four-mRNA expression levels weighted by its relative coefficients in a multivariate Cox regression test.Survival risk score（SRS）=(-0.23734×Exp_WIPF1)+(0.19805×Exp_PPIB)+(0.09904×Exp_BASP1)+(0.15977×Exp_PLOD2).Risk stratification suggests that four-mRNA tags perform well in predicting overall survival in patients with head and neck squamous cell carcinoma.For the 494 patients in our study,they were divided into a high-risk group（n=247）or a low-risk group（n=247）based on the median risk score.The four-mRNA-based Tekaplan-Meier overall survival curves were significantly different（log-rank p=0<0.001）,and the high-risk prognosis was poor.Conclusion:In this study,we used TCGA head and neck squamous cell carcinoma data to analyze using single-factor and multi-factor COX proportional hazard regression models.We identified a four-mRNA marker combination,namely,WIPF1,PPIB,BASP1,and PLOD2.Composed collection.The prognosis of head and neck squamous cell carcinoma patients with the high-risk score group and the low-risk score group of this mRNA combination is significantly different,indicating that the four-mRNA marker combination model has good sensitivity and specificity in predicting the survival risk of patients with head and neck squamous cell carcinoma.A reliable biomarker for predicting prognosis in patients with head and neck squamous cell carcinoma.