Identification and Validation of a Novel Six-Gene Expression Signature for Predicting Hepatocellular Carcinoma Prognosis

• Published in: Frontiers in immunology Vol. 12; p. 723271
• Main Authors: Yan, Zongcai, He, Meiling, He, Lifeng, Wei, Liuxia, Zhang, Yumei
• Format: Journal Article
• Language: English
• Published: Switzerland Frontiers Media S.A 01-12-2021
• Subjects: Adult; Aged; Aged, 80 and over; biomarkers; Carcinoma, Hepatocellular - genetics; Carcinoma, Hepatocellular - mortality; Carcinoma, Hepatocellular - physiopathology; Clinical Decision-Making; Datasets as Topic; differential gene expression analysis; Disease Progression; Female; Gene Expression Profiling; Gene Expression Regulation, Neoplastic; hepatocellular carcinoma; Humans; Immunology; Kaplan-Meier Estimate; Liver Neoplasms - genetics; Liver Neoplasms - mortality; Liver Neoplasms - physiopathology; Male; Middle Aged; Mutation; Prognosis; Proportional Hazards Models; risk score; Tumor Microenvironment; tumor microenvironment; prognosis; biomarkers; differential gene expression analysis; hepatocellular carcinoma; risk score
• ISSN: 1664-3224; 1664-3224
• DOI: 10.3389/fimmu.2021.723271

Hepatocellular carcinoma (HCC) is a highly lethal disease. Effective prognostic tools to guide clinical decision-making for HCC patients are lacking.
We aimed to establish a robust prognostic model based on differentially expressed genes (DEGs) in HCC.
Using datasets from The Cancer Genome Atlas (TCGA), the Gene Expression Omnibus (GEO), and the International Genome Consortium (ICGC), DEGs between HCC tissues and adjacent normal tissues were identified. Using TCGA dataset as the training cohort, we applied the least absolute shrinkage and selection operator (LASSO) algorithm and multivariate Cox regression analyses to identify a multi-gene expression signature. Proportional hazard assumptions and multicollinearity among covariates were evaluated while building the model. The ICGC cohort was used for validation. The Pearson test was used to evaluate the correlation between tumor mutational burden and risk score. Through single-sample gene set enrichment analysis, we investigated the role of signature genes in the HCC microenvironment.
A total of 274 DEGs were identified, and a six-DEG prognostic model was developed. Patients were stratified into low- or high-risk groups based on risk scoring by the model. Kaplan-Meier analysis revealed significant differences in overall survival and progression-free interval. Through univariate and multivariate Cox analyses, the model proved to be an independent prognostic factor compared to other clinic-pathological parameters. Time-dependent receiver operating characteristic curve analysis revealed satisfactory prediction of overall survival, but not progression-free interval. Functional enrichment analysis showed that cancer-related pathways were enriched, while immune infiltration analyses differed between the two risk groups. The risk score did not correlate with levels of PD-1, PD-L1, CTLA4, or tumor mutational burden.
We propose a six-gene expression signature that could help to determine HCC patient prognosis. These genes may serve as biomarkers in HCC and support personalized disease management.