Development and validation of a survival model for thyroid carcinoma based on autophagy-associated genes

Aging (Albany NY). 2020 Oct 14;12(19):19129-19146. doi: 10.18632/aging.103715. Epub 2020 Oct 14.

Abstract

Abnormalities in autophagy-related genes (ARGs) are closely related to the occurrence and development of thyroid carcinoma (THCA). However, the effect of ARGs on the prognosis of THCA remains unclear. Here, by analyzing data from TCGA, 26 differentially expressed ARGs were screened. Cox regression and Lasso regression were utilized to analyze the prognosis of the training group, and a risk model was constructed. Our results show that low-risk patients had better overall survival (OS) than high-risk patients, and the area under the ROC curve in the training and testing groups was significant (3-year AUC, 0.735 vs 0.796; 5-year AUC, 0.821 vs 0.804). In addition, a comprehensive analysis of the 5 identified ARGs demonstrated that most of them were related to OS in THCA patients, and two of them (CX3CL1 and CDKN2A) were differentially expressed in THCA and normal thyroid tissues at the protein level. GSEA suggested that the inactivation of the cell defense system and the activation of some classical tumor signaling pathways are important driving forces for the progression of THCA. This study demonstrated that the 5 ARGs in the survival model are promising multidimensional biomarkers for the diagnosis, prognosis, and treatment of THCA.

Keywords: TCGA database; autophagy-related genes; prognosis; thyroid carcinoma.