Ferroptosis Related Prediction Model for Hepatocellular Carcinoma Patients Sensitive to Chemotherapy Embolization Therapy Based on Bioinformatics Analysis

Abstract

Abstract Objective: The objective of this study was to develop a predictive model that can help with effective transcatheter arterial chemoembolization (TACE) in treating hepatocellular carcinoma by identifying ferroptosis-associated genes. Methods: In this study, the GSE104580 dataset from the GEO database was analyzed to identify significantly differentially expressed genes (DEGs), which were then used to identify genes associated with chemoembolization sensitivity and ferroptosis using the weighted gene co-expression network analysis (WGCNA). These genes were then used to construct a TACE treatment sensitivity prediction model using lasso regression. Immune infiltration analysis was also conducted, and a hub mRNA, hub miRNA, and hub lncRNA interaction network was established. The TCGA dataset was used to construct a prediction model which was validated by ICGC dataset. Results: Using the GSE104580 dataset, a total of 2689 DEGs were screened, resulting in the identification of 37 genes. Protein-protein interaction (PPI) network analysis was performed based on these genes, and key genes involved in predicting TACE treatment sensitivity for liver cancer were identified through GO, KEGG, and GSEA analyses. Using the lasso regression method, six hub genes were identified: GLS2, CDKN1A, GPT2, ASNS, SLC38A1, and SLC2A1. Two distinct ferroptosis patterns were identified based on these hub genes, and immune infiltration analysis was conducted to further investigate potential associations with liver cancer. Additionally, a hub mRNA, miRNA, and LncRNA interaction network was constructed using data from miRTarBase, TarBase, and Starbase databases. Utilizing a 6-gene signature, two distinct risk groups were identified. Remarkably, patients classified within the high-risk group exhibited a significant decrease in overall survival when compared to their low-risk counterparts (P < 0.001 in the TCGA cohort and P = 0.013 in the ICGC cohort). In addition, the predictive capacity of this signature was further validated by receiver operating characteristic (ROC) curve analysis. Conclusion: This study suggests that the six hub genes identified in this research could serve as important targets for improving liver cancer prognosis. Additionally, these genes can be utilized to construct effective TACE sensitive prediction models to help clinicians in treating hepatocellular carcinoma.