Huang Shenjun, Tu Tingting
Department of Oncology, Nantong Tumur Hospital (Affiliated Tumur Hospital of Nantong University), Nantong, China.
Department of Radiation Oncology, Lianyungang Second People's Hospital (Lianyungang Tumur Hospital), Lianyungang, China.
Front Immunol. 2025 Jan 7;15:1534723. doi: 10.3389/fimmu.2024.1534723. eCollection 2024.
Hepatocellular carcinoma (LIHC) poses a significant health challenge worldwide, primarily due to late-stage diagnosis and the limited effectiveness of current therapies. Cancer stem cells are known to play a role in tumor development, metastasis, and resistance to treatment. A thorough understanding of genes associated with stem cells is crucial for improving the diagnostic precision of LIHC and for the advancement of effective immunotherapy approaches.
This research combines single-cell RNA sequencing with machine learning techniques to identify vital stem cell-associated genes that could act as prognostic biomarkers and therapeutic targets for LIHC. We analyzed various datasets, applying negative matrix factorization alongside machine learning algorithms to reveal gene expression patterns and construct diagnostic models. The XGBoost algorithm was specifically utilized to identify key regulatory genes related to stem cells in LIHC, and the expression levels and prognostic significance of these genes were validated experimentally.
Our single-cell analysis identified 16 differential prognostic genes associated with liver cancer stem cells. Cluster analysis and diagnostic models constructed using various machine learning techniques confirmed the significance of these 16 genes in the diagnosis and immunotherapy of LIHC. Notably, the XGBoost algorithm identified S100A10 as the stem cell-related gene most relevant to the prognosis of LIHC patients. Experimental validation further supports S100A10 as a potential prognostic marker for this cancer type. Additionally, S100A10 shows a positive correlation with the stem cell marker POU5F1.
The results of this study highlight S100A10 as an essential predictor for liver cancer diagnosis and treatment response, particularly regarding immunotherapy. This research offers valuable insights into the molecular mechanisms underlying LIHC and suggests S100A10 as a promising target for enhancing treatment outcomes in liver cancer patients.
肝细胞癌（LIHC）在全球范围内构成了重大的健康挑战，主要原因是诊断较晚以及当前治疗方法的有效性有限。已知癌症干细胞在肿瘤发展、转移和治疗耐药性中发挥作用。全面了解与干细胞相关的基因对于提高LIHC的诊断准确性以及推进有效的免疫治疗方法至关重要。
本研究将单细胞RNA测序与机器学习技术相结合，以识别可能作为LIHC预后生物标志物和治疗靶点的重要干细胞相关基因。我们分析了各种数据集，应用非负矩阵分解以及机器学习算法来揭示基因表达模式并构建诊断模型。特别利用XGBoost算法来识别与LIHC中干细胞相关的关键调控基因，并通过实验验证这些基因的表达水平和预后意义。
我们的单细胞分析确定了16个与肝癌干细胞相关的差异预后基因。使用各种机器学习技术构建的聚类分析和诊断模型证实了这16个基因在LIHC诊断和免疫治疗中的重要性。值得注意的是，XGBoost算法确定S100A10是与LIHC患者预后最相关的干细胞相关基因。实验验证进一步支持S100A10作为这种癌症类型的潜在预后标志物。此外，S100A10与干细胞标志物POU5F1呈正相关。
本研究结果突出了S100A10作为肝癌诊断和治疗反应（特别是在免疫治疗方面）的重要预测指标。这项研究为LIHC的分子机制提供了有价值的见解，并表明S100A10是改善肝癌患者治疗结果的一个有前景的靶点。
