Identification of a Novel Epithelial-to-mesenchymal-related Gene Signature in Predicting Survival of Patients with Hepatocellular Carcinoma

Abstract

Background: Epithelial-mesenchymal transformation (EMT) promotes cancer metastasis, including hepatocellular carcinoma. Therefore, EMT-related gene signature was explored.

Objective: The present study was designed to develop an EMT-related gene signature for predicting the prognosis of patients with hepatocellular carcinoma..

Methods: An integrated gene expression analysis based on tumor data of the patients with hepatocellular carcinoma from The Cancer Genome Atlas (TCGA), HCCDB18, and GSE14520 dataset was conducted. An EMT-related gene signature was constructed by the least absolute shrinkage and selection operator (LASSO) and COX regression analysis of univariate and multivariate survival.

Results: A 3-EMT gene signature was developed and validated based on gene expression profiles of hepatocellular carcinoma from three microarray platforms. Patients with a high-risk score had significantly worse overall survival (OS) than those with low-risk scores. The EMT-related gene signature showed a high performance in accurately predicting prognosis and examining the clinical characteristics and immune score analysis. Univariate and multivariate Cox regression analyses confirmed that the EMT-related gene signature was an independent prognostic factor for predicting survival in hepatocellular carcinoma patients. Compared with the existing models, our EMTrelated gene signature reached a higher area under the curve (AUC).

Conclusion: Our findings provide novel insight into understanding EMT and help identify hepatocellular carcinoma patients with poor prognosis.

Keywords: Hepatocellular carcinoma; bioinformatics; epithelial-mesenchymal transformation; liver cancer; prognosis; signature.