@article{JGO116920,
author = {Wei Li and Hang Jiang and Jian Duan and Jinlan He and Liping Zhao and Guoping Zhong and Chenghu Fan},
title = {Mining and experimental validation of machine learning-based immune-related diagnostic biomarkers for hepatocellular carcinoma},
journal = {Journal of Gastrointestinal Oncology},
volume = {17},
number = {3},
year = {2026},
keywords = {},
abstract = {Background: Hepatocellular carcinoma (HCC) is a highly malignant and aggressive tumor. Immune-related genes (IRGs) expression correlates closely with the HCC immune microenvironment, and this study aims to identify immune-related diagnostic markers in HCC.Methods: HCC-related datasets and IRGs were obtained from public databases (TCGA, ICGC, TISIDB, and InnateDB). Differential expression analysis screened differentially expressed genes (DEGs) between HCC and control samples, while weighted gene co-expression network analysis identified key module genes correlated with immune scoring systems. Candidate genes were obtained via the intersection of DEGs, IRGs, and key module genes. Hub genes were then determined through protein-protein interaction analysis, followed by correlation and survival analyses. Moreover, three machine learning algorithms identified diagnostic biomarkers, which were evaluated via a logit model and receiver operating characteristic curve analysis for HCC diagnostic efficacy. Finally, biomarker expression was validated in clinical samples.Results: After identifying 8,800 DEGs and 2,337 key module genes, we intersected them with IRGs to obtain 87 candidate genes. Thereafter, 33 hub genes were obtained. The hub genes showed a notable positive correlation, suggesting their potential involvement in regulating common biological processes. Additionally, six hub genes (e.g., FCGR2A, TNFRSF4) exhibited expression correlation with HCC patients’ survival. Three diagnostic biomarkers (FCGR2A, MARCO, TNFRSF4) were further identified via machine learning algorithms; logit model analysis confirmed their significant diagnostic utility for HCC. Reverse transcription quantitative polymerase chain reaction (RT-qPCR) validation showed markedly increased expression of FCGR2A and TNFRSF4, consistent with dataset trends.Conclusions: This study identified three immune-related diagnostic biomarkers for HCC, which may provide novel insights into HCC prognostic management.},
issn = {2219-679X}, url = {https://jgo.amegroups.org/article/view/116920}
}