Construction and validation of prognostic risk score model based on autophagy-related genes for head and neck squamous cell carcinoma

Abstract

Objective To construct and validate a prognostic risk-scoring model based on autophagy-related genes for head and neck squamous cell carcinoma (HNSCC). Methods All HNSCC transcriptome expression data (RNA sequencing, RNA-seq) and clinical information downloaded from the Cancer Genome Atlas (TCGA) database, and differentially expressed genes were screened. These differentially expressed genes of HNSCC were intersected with autophagy related genes (ARGs) retrieved from the GeneCards database to obtain differentially expressed ARGs. After integrating clinical information, prognostic ARGs were obtained by prognostic analysis, and then enrichment analysis was performed. The least absolute shrinkage and selection operator (LASSO) regression and Cox regression model were used to construct a risk scoring model for predicting the prognosis and survival of HNSCC. The receiver operating characteristic (ROC) curve was drawn, the area under the curve (AUC) and the best cut-off value were calculated, and the patients were divided into the high- and low-risk score groups with the best cut-off value. Kaplan-Meier survival curve was drawn to assess the predictive performance of the model. The clinical information was integrated with the risk score, and the independent prognostic value of the risk score was evaluated by Cox regression analysis. Results The prognostic risk score model of HNSCC was constructed based on the 9 ARGs significantly related to prognosis were obtained by LASSO regression and Cox regression analysis through the prognostic analysis for differentially expressed ARGs which screened 20 ARGs related to prognosis. The survival time of the low-risk score group was better than that of the high risk score group, and the survival time of the two groups was significantly different (P < 0.001), according to ROC curve and Kaplan-Meier survival curve. The model showed good prediction performance in both the training set (the maximum AUC, 0.69) and the external validation set (the maximum AUC, 0.822). Cox regression analysis showed that the risk score was significantly correlated with the prognosis of HNSCC patients (P < 0.001), indicating that the risk score had independent prognostic value for HNSCC. Conclusion The HNSCC risk scoring model composed of 9 ARGs can effectively predict the prognosis of patients with HNSCC.

Keywords

• autophagy-related genes
• head and neck squamous cell carcinoma
• cancer genome map database
• risk score