Hepatocellular carcinoma (HCC) remains highly fatal due to late diagnosis and scarce early biomarkers. Multi‑omics and liquid biopsy advances improve early detection, prognostication, and monitoring. Gene signatures derived from genomic, transcriptomic, and immune profiling predict prognosis and response to immune checkpoint inhibitors. Circulating tumor DNA enables non‑invasive real‑time tracking. However, clinical validation, cross‑platform reproducibility, and ethical issues (data privacy, consent, standardization) hinder routine use. Integrating multi‑omics with liquid biopsy offers a path to personalized therapy, but success requires multidisciplinary collaboration to overcome these barriers.
Introduction
HCC is the third leading cause of cancer death. Risk factors include hepatitis B/C, alcohol, aflatoxin, obesity, and metabolic dysfunction. Curative options are limited to early stages; advanced cases rely on multi‑kinase inhibitors (toxicity, resistance) or immunotherapy (variable response). Tumour heterogeneity—genomic, epigenetic, and microenvironmental—drives therapy resistance and recurrence. This review synthesises gene‑signature and multi‑omics strategies to improve prognosis and immunotherapy selection, emphasising validation and clinical feasibility.
Molecular Basis of HCC
HCC exhibits marked genetic heterogeneity. Frequent alterations include TP53 (~50%), TERT promoter (~60%), and Wnt/β‑catenin (~40%) mutations, influencing aggressiveness and immune evasion. Epigenetic changes (methylation, non‑coding RNAs) and proteomic/metabolomic alterations contribute to tumor progression and provide potential biomarkers.
Prognostic Gene Signatures
Many multigene expression panels (e.g., 8‑gene, 9‑gene, 4‑gene) have been validated across cohorts, often outperforming conventional staging (BCLC, TNM). They stratify patients by survival and recurrence risk, guiding transplant prioritisation, adjuvant therapy, and surveillance. Public databases (TCGA, GEO) enable cross‑platform validation, though formalin‑fixed samples pose RNA quality challenges.
Tumour Microenvironment and Immune Classification
The TME critically influences outcomes. Transcriptomic signatures quantify immune infiltration: high CD8A/GZMB indicates cytotoxic T‑cell activity and better prognosis; M2 macrophage markers correlate with immunosuppression. HCC can be classified as immune‑"hot" (inflamed, high IFN‑γ signature, responsive to ICIs) or "cold" (immune‑excluded, Wnt‑active, resistant). Stromal and cytokine signatures (e.g., VEGFA, IL6) further inform tumour‑stroma crosstalk and metastasis.
Gene Signatures for Immunotherapy Guidance
ICI response is unpredictable; gene expression signatures improve patient selection. IFN‑γ‑related transcripts (CXCL9/10, IDO1, STAT1) correlate with anti‑PD‑1/PD‑L1 benefit. Clinical trial analyses (CheckMate 040, atezolizumab/bevacizumab) show that multi‑gene immune scores outperform single biomarkers like PD‑L1 IHC. For cold tumours, priming strategies—oncolytic viruses, radiotherapy—can convert them to hot and enhance ICI efficacy.
Multi‑Omics Integration and Non‑Coding RNAs
Integrating genomics, transcriptomics, epigenomics, proteomics, and metabolomics yields more robust and stable signatures than single‑layer approaches. Non‑coding RNAs (circPRDM4, HOTAIR, MALAT1) regulate PD‑L1 and immune evasion, contributing to immunotherapy resistance and offering potential therapeutic targets.
Full text
https://www.xiahepublishing.com/1555-3884/GE-2025-00073
The study was recently published in the Gene Expression .
Gene Expression (GE) is an open-access journal. It was launched in 1991 by Chicago Medical School Press, and transferred to Cognizant Communication Corporation in 1994. From August 2022, GE is published by Xia & He Publishing Inc.
GE publishes peer-reviewed and high-quality original articles, reviews, editorials, commentaries, and opinions on its primary research topics including cell biology, molecular biology, genes, and genetics, especially on the cellular and molecular mechanisms of human diseases.
GE has been indexed in Medline (1991-2021), Scopus, Biological Abstracts, Biosis Previews, ProQuest, etc.
