Construction and validation of a prognostic model based on stage-associated signature genes of head and neck squamous cell carcinoma: a bioinformatics study

Background: Head and neck squamous cell carcinoma (HNSCC) is a malignancy of epithelial origin and with poor prognosis. Exploring the biomarkers and prognostic models that can contribute to early tumor detection is meaningful. A comprehensive analysis was conducted according to the stage-related signature genes of HNSCC, and a prognostic model was developed to validate their ability to predict the prognosis. Methods: The transcriptome profiles and clinical information of HNSCC patients were obtained from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) respectively. mRNA expressions of differentially expressed genes (DEGs) were analyzed in stage I–II patients and stage III–IV patients from TCGA by R packages. A protein-protein interaction (PPI) network and core-gene network map were constructed, and Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were performed to examine pathway enrichment. Kaplan-Meier, least absolute shrinkage and selection operator (LASSO), and multivariate Cox regression were applied to establish a stage-associated signature model. A Spearman analysis was conducted to examine the correlations between the characteristic genes and immune cell infiltration. Kaplan-Meier analysis and a receiver operating characteristic (ROC) curve were used to test the effectiveness of the model. Univariate multivariate Cox regression analyses were used to assess whether the risk score was an independent prognostic indicator for HNSCC. Results: In TCGA cohort, 5 genes (i.e., BRINP1, IL17A, ALB, FOXA2, and ZCCHC12) in the constructed prognostic risk model were associated with prognosis. Patients in the low-risk group had a better prognosis outcome than those in the high-risk group. The predictive power was good because all the area under the curve (AUC) of the risk score was higher than 0.6. Risk score [hazard ratio (HR) =1.985; P<0.001] was an independent risk factor for the prognosis of HNSCC. The results in the GEO cohort were consistent with those in the TCGA cohort. Conclusions: We constructed and verified a prognostic risk model of stage-related signature genes for HNSCC based on the GEO and TCGA data. Due to the good predictive accuracy of this model, the prognosis of and the tumor immune cell infiltration with patients can be estimated.