Abstract
BACKGROUND: Dysregulation of lipid metabolism is implicated in the progression of hepatocellular carcinoma (HCC). We therefore investigated the molecular characteristics of lipid-metabolism-related genes in HCC. METHODS: Multi-dimensional bioinformatics analysis was conducted to comprehensively identify the lipid metabolism-related genes (LMRGs) from public databases, as well as the clinical information, immune features, and biological characteristics of HCC. The IMGR were then used to classify HCC into molecular phenotypes. Six lipid metabolism-related genes sets with the potential to predict the prognosis of HCC patients were identified. RESULTS: A total of 770 HCC patients with complete clinical information and corresponding 776 LMRGs were downloaded from 3 databases. Univariate cox and non-negative matrix factorization analyses were used to classify HCC patients into 2 clusters. This molecular classification was associated with overall survival, clinical characteristics, and immune cells. The biological function of the differentially expressed LMRGs in the 2 clusters showed the genes associated with tumor-related metabolism pathways. A combination of multivariate/univariate cox regression and least absolute shrinkage and selection operator analyses were conducted to build a 6 LMRGs signature (6-IS) to predict the prognosis of HCC. The 6-IS signature was found to be an independent prognostic factor. Performance of the 6-IS prognostic signature was verified in a validation set and compared with an external data set. Results revealed the 6-IS signature could effectively predict the prognosis of patients with HCC. CONCLUSIONS: This study provides new insights into the role of LMRG in the pathogenesis of HCC and presents a novel prognostic signature 6-IS monitoring the clinical course of HCC.