ORIGINAL RESEARCH article
Front. Immunol.
Sec. Cancer Immunity and Immunotherapy
Volume 16 - 2025 | doi: 10.3389/fimmu.2025.1663519
This article is part of the Research TopicHarnessing Macrophage Modulation: Advancing Hematologic Cancer Treatment StrategiesView all 4 articles
Machine learning-based tumor associated macrophages polarity signature predicts prognosis and treatment response in hepatocellular carcinoma
Provisionally accepted
- 1Xuanwu Hospital Capital Medical University, Beijing, China
- 2Xuancheng City Central Hospital, Xuancheng, China
- 31st Medical Center of Chinese PLA General Hospital, Beijing, China
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
Background: Tumor-associated macrophages (TAMs) shape the tumor microenvironment and drive hepatocellular carcinoma (HCC) progression. However, the prognostic significance of TAM polarity-related genes, particularly based on the CXCL9:SPP1 signature, remains unclear. Methods: We identified 372 TAM polarity-related genes in the TCGA-LIHC dataset. Prognostic candidates were selected using univariate Cox regression, bootstrap resampling, and the Boruta algorithm. Seven machine learning models were compared, and XGBoost was selected to construct a TAM polarity-related signature (TPS) consisting of 17 genes. TPS was validated in two external cohorts. Associations with clinical features, biological pathways, immune status, and drug sensitivity were explored. scRNA-seq and qRT-PCR were performed to investigate cellular expression and functional relevance. Results: TPS markedly different patients into high-and low-risk groups with significantly different survival outcomes (TCGA 1-, 3-, 5-year AUCs: 0.91, 0.89, 0.88). High-risk patients showed enrichment in glycan metabolism, DNA repair, and oncogenic pathways, whereas low-risk patients displayed elevated lipid and amino acid metabolism. Immune profiling revealed greater infiltration of immunosuppressive cells and higher expression of immune checkpoints in high-risk patients. Drug sensitivity analysis identified potential therapeutic targets and candidate compounds, including CDK1, PLK1, and statins. scRNA-seq analysis highlighted disrupted macrophage-immune interactions and identified SPP1 as a key signaling mediator. Silencing of TTC1 and G6PD suppressed HCC cell proliferation. Conclusion: We developed and validated a robust TAM polarity-related signature that effectively stratifies HCC patients by prognosis. TPS provides insights into tumor immunity, metabolism, and drug response, and may serve as a valuable tool for precision medicine in HCC.
Keywords: machine learning, Hepatocellular Carcinoma, macrophages polarity, prognosis, single-cell RNA-seq
Received: 10 Jul 2025; Accepted: 16 Oct 2025.
Copyright: © 2025 Wang, Zhang, Zheng and Lu. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence: Shichun Lu, lusc_301@163.com
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.