Identification of Core Genes and Prognostic Models of Laryngeal Cancer by Autophagy Related Biomarkers

Abstract

Background: The aims of our article were to identify the core genes of the autophagy-related genes (ARGs) which abnormally expressed in laryngeal cancer (LC) and constructed a risk prognostic models with these genes. Methods: In this study, we identified genes with abnormally expressed in LC, and they were mainly involved in some cancer-related gene ontology (GO) terms and Kyoto Encyclopedia of Genes and Genomes pathways (KEGG). Statistical analysis were conducted to identify the powerful independent prognostic factors associated with clinical factors and survival. Results: A total of 35 DEGs were identified in our research. The risk prediction model was constructed with three potential prognostic genes (VEGFA, SPNS1 and CCL2) of autophagy by lasso regression analysis that can successfully predict the prognosis in LC. We applied ROC curve to evaluate the effectiveness of the risk prognostic model, and found that AUC was 0.693 below the curve. Risk prediction model was only related to survival status (P < 0.01), and was not related to clinicopathological factors. Furthermore, the genes (VEGFA and CCL2) were considered as core genes not only because they were the highly connected genes but also they were the composed genes of risk prognostic model. Conclusions: Taken together, ARGs were considered as important roles in the progression of LC and the prognostic model can help to identification of new targets to guide the diagnosis and therapy.