Background and Aim. With regard to patients with intermediate-stage, irresectable hepatocellular carcinoma (HCC), transcatheter arterial chemoembolization (TACE) is the mainstay of treatment. There is an urgent clinical requirement to identify reliable biomarkers to predict the response of HCC patients to TACE treatment. We aimed to identify a gene signature for predicting TACE response in HCC patients based on bioinformatics analysis. Methods. We downloaded the gene expression profile GSE104580 based on 147 tumor samples from 81 responders to TACE and 66 nonresponders from the Gene Expression Omnibus (GEO) database. Then, we randomly divided the 147 tumor samples into a training set () and a validation set () and screened differentially expressed genes (DEGs) in the training set. Gene Ontology (GO) term and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed to annotate functions of the DEGs. The DEGs were mapped into the STRING website for constructing protein-protein interaction (PPI). The predictive value of the candidate genes by receiver-operating characteristic (ROC) curves was further verified in the validation set. Results. We totally found 158 DEGs (92 upregulated genes and 66 downregulated genes) in the training set. The GO enrichment analysis revealed that DEGs were significantly enriched in metabolic and catabolic processes, such as drug metabolic process, fatty acid metabolic process, and small molecule catabolic process. The KEGG pathway analysis revealed that the DEGs were mainly concentrated in drug metabolism-cytochrome P450, metabolism of xenobiotics by cytochrome P450, and chemical carcinogenesis. We identified 6 candidate genes (CXCL8, AFP, CYP1A1, MMP9, CYP3A4, and SERPINC1) based on the PPI network of the DEGs, which had high predictive value in HCC response to TACE with an area under the curve (AUC) value of 0.875 and 0.897 for the training set and validation set, respectively. Conclusion. We identified a six-gene signature which might be biomarkers for predicting HCC response to TACE by a comprehensive bioinformatics analysis. However, the actual functions of these genes required verification.

中文翻译：

---

通过生物信息学分析确定肝细胞癌中 TACE 的六基因特征预测反应

背景和目标。对于中期、不可切除的肝细胞癌 (HCC) 患者，经导管动脉化疗栓塞 (TACE) 是主要治疗方法。临床迫切需要确定可靠的生物标志物来预测 HCC 患者对 TACE 治疗的反应。我们旨在基于生物信息学分析确定预测 HCC 患者 TACE 反应的基因特征。方法。我们下载了基于 147 个肿瘤样本的基因表达谱 GSE104580，这些样本来自对 TACE 的 81 名响应者和来自基因表达综合 (GEO) 数据库的 66 名无响应者。然后，我们将 147 个肿瘤样本随机分成一个训练集（)和验证集 ()并筛选训练集中的差异表达基因 (DEG)。进行基因本体论 (GO) 术语和京都基因和基因组百科全书 (KEGG) 通路富集分析以注释 DEG 的功能。DEG 被映射到 STRING 网站，用于构建蛋白质 - 蛋白质相互作用（PPI）。在验证集中进一步验证了受试者操作特征（ROC）曲线对候选基因的预测值。结果. 我们在训练集中总共发现了 158 个 DEG（92 个上调基因和 66 个下调基因）。GO富集分析显示DEGs在代谢和分解代谢过程中显着富集，如药物代谢过程、脂肪酸代谢过程和小分子分解代谢过程。KEGG通路分析表明，DEGs主要集中在药物代谢-细胞色素P450、细胞色素P450对异生物质的代谢和化学致癌作用中。我们基于 DEG 的 PPI 网络确定了 6 个候选基因（CXCL8、AFP、CYP1A1、MMP9、CYP3A4 和 SERPINC1），它们对 HCC 对 TACE 的反应具有很高的预测价值，曲线下面积 (AUC) 值为 0.875训练集和验证集分别为 0.897。结论. 我们通过综合生物信息学分析确定了一个六基因特征，它可能是预测 HCC 对 TACE 反应的生物标志物。然而，这些基因的实际功能需要验证。