Construction and Validation of Prognostic Markers of Liver Cancer Based on Autophagy Genes

Abstract

Background: Liver cancer is one of the most common diseases in the world. At present, the mechanism of autophagy genes in liver cancer is not very clear. Therefore, it is meaningful to study the role and the prognostic value of autophagy genes in liver cancer. Objective: The purpose of this study is to conduct a bioinformatics analysis of autophagy genes related to primary liver cancer for establishing a prognostic model of primary liver cancer based on autophagy genes. Methods: We identified autophagy genes related to the prognosis of liver cancer through bioinformatics methods. Results: Through difference analysis, 31 differential autophagy genes were screened out and then analyzed by GO and KEGG analysis. At the same time, we built a PPI network. For optimizing the evaluation of the prognosis of liver cancer patients, we integrated multiple autophagy genes, after which a prognostic model was established. By using univariate cox regression analysis, 15 autophagy genes related to prognosis were screened out. Then we included these 15 genes into the Least Absolute Shrinkage and Selection Operator (LASSO) and performed a multi-factor cox regression analysis on the 9 selected genes for constructing a prognostic model. The risk score of each patient, who participated in the establishing of the model, was calculated based on 4 genes (BIRC5, HSP8, SQSTM1, and TMEM74). Then the patients were divided into high-risk groups and low-risk groups. In the multivariate cox regression analysis, the risk score was assessed by the independent prognostic factors (HR = 1.872, 95% CI = 1.544 - 2.196, p < 0.001). Survival analysis showed that the survival time of the low-risk group was significantly longer than that of the high-risk group. By combining clinical characteristics and autophagy genes, we constructed a nomogram for predicting the prognosis. The external dataset GSE14520 proved that the nomogram has a good prediction for individual patients with primary liver cancer. Conclusion: This study provided potential autophagy-related markers for liver cancer patients to predict their prognosis and reveal part of the molecular mechanism of liver cancer autophagy. At the same time, certain gene pathways and protein pathways related to autophagy may provide some inspiration for the development of anticancer drugs.