A multiomic ferroptosis-associated prognostic signature incorporating epigenetic and transcriptional biomarkers for hepatocellular carcinoma.

Background: Ferroptosis was reported to have tremendous promise in the treatment and prognosis of hepatocellular carcinoma (HCC). Here, we identified a novel ferroptosis-related prognostic signature incorporating epigenetic and transcriptional biomarkers could help predicting survival of patients with HCC. Methods: We employed multi-omics and clinical data from The Cancer Genome Atlas (TCGA) database to identify the ferroptosis-associated methylation CpG sites associated with HCC survival using sure independence screening (SIS). Then we utilized Kaplan-Meier curves to evaluate the prognostic significance of gene expression and DNA methylation. Receiver operating characteristic (ROC) curve was used predicting the 3- and 5-year survival. Mediation analysis of ferroptosis-related methylation and transcriptional score was performed. Results: We firstly identified 114 significant CpG sites under the criteria of false discovery rate (FDR) <0.05 in training set. Then we screened out 5 candidate CpG sites in validation set for multivariate screening and stepwise regression. We found that the high-risk group had significantly shorter survival time than the low-risk group in the prognostic signature combined with epigenetic and transcriptional scores (HR =2.72 95% CI: 2.01-3.68, P=8.75E-11). And the predictive model involving clinical information, gene expression, and methylation data performed best for 3-year survival prediction (AUC =0.672) and 5-year survival prediction (AUC =0.742). Conclusions: Our results suggested a signature combining clinical information, ferroptosis-related gene expression, and methylation presented a superior ability for prognostic prediction in HCC, which may bring us novel tool and targets in the treatment of HCC.