Identification of ferroptosis and autophagy-related diagnostic markers in hepatocellular carcinoma based on machine learning

Abstract

Abstract Background The hepatocellular carcinoma (HCC) is a malignant tumor of the hepatic domain that is one of the most prevalent malignancies of the digestive system worldwide. Hepatocellular carcinoma, ferroptosis, and autophagy are closely related. The understanding of these relationships enables researchers to uncover the pathogenesis of liver cancer and develop new treatment strategies. Methods Four genes most related to the prognosis of LIHC were screened, and a risk score model was established. The model classifies HCC patients from different datasets and platforms into low-risk and high-risk groups. Multivariate Cox regression analysis confirmed that the model was an independent prognostic factor for HCC. Results Our results showed significant differences in gene expression profiles in HCC samples compared to normal tissues, with 2596 genes significantly up-regulated and 3036 genes significantly down-regulated. In addition, our results showed that the proportion of Macrophages M0, NK cells activated and T cells CD4 memory resting was significantly increased in HCC samples compared to normal tissues. In addition, our analysis showed that a combination of four ferroptosis and autophagy related genes including CDKN2A, HSPA5, GABARAPL1, and MAPK3 could serve as the optimal markers for the diagnosis of HCC. The combined analysis of these four genes accurately diagnosed the occurrence of HCC. Conclusions In this study, we investigated new biomarkers associated with hepatocellular carcinoma (HCC) based on ferroptosis and autophagy. We aimed to establish a risk scoring model to predict the prognosis of HCC, with the goal of providing appropriate treatment strategies for HCC patients.