Wu Di, Pan Yun, Zheng Xueyong
Department of General Surgery, Sir Run-Run Shaw Hospital, School of Medicine, Zhejiang University, Hangzhou, 310016, China.
Department of Emergency, Sir Run-Run Shaw Hospital, School of Medicine, Zhejiang University, Hangzhou, 310016, China.
J Cancer. 2021 Jan 30;12(7):1884-1893. doi: 10.7150/jca.52089. eCollection 2021.
Though various hub genes for HCC have been identified in decades, the limited sample size, inconsistent bioinformatic analysis methods and lacking evaluation in validation cohorts would make the results less reliable, novel biomarkers and risk model for HCC prognosis are still urgently desired. The Robust Rank Aggression method was applied to integrate 12 HCC microarray datasets to screen for robustly and stably differentially expressed candidates. The Least Absolute Shrinkage and Selection Operator regression and multivariate Cox regression analysis were performed to construct a six hub genes-based prognostic model, which was further verified in matched tumor and non-tumor hepatic samples and two independent validation cohorts. Six hub genes for HCC were identified including CD163, EHHADH, KIAA0101, SLC16A2, SPP1 and THBS4. The risk score according to hub genes-based prognostic model could be an independent predictive factor for HCC. Quantitative real-time polymerase chain reaction results showed significant difference in expression level between tumor and non-tumor hepatic tissues. Prognostic value of risk model has been verified in TCGA-HCC and GSE76240 datasets. Biological function analysis revealed these hub genes were closely associated with tumorigenesis processes. A novel six hub genes predictive risk model for HCC has been established based on multiple datasets analyses, providing novel features for the prediction of HCC patients' outcome.
尽管数十年来已鉴定出多种肝癌的枢纽基因，但样本量有限、生物信息分析方法不一致以及缺乏在验证队列中的评估，会使结果的可靠性降低，因此仍迫切需要用于肝癌预后的新型生物标志物和风险模型。应用稳健秩聚合方法整合12个肝癌微阵列数据集，以筛选出稳健且稳定差异表达的候选基因。进行最小绝对收缩和选择算子回归以及多变量Cox回归分析，以构建基于六个枢纽基因的预后模型，并在匹配的肿瘤和非肿瘤肝脏样本以及两个独立验证队列中进一步验证。鉴定出六个肝癌枢纽基因，包括CD163、EHHADH、KIAA0101、SLC16A2、SPP1和THBS4。基于枢纽基因的预后模型得出的风险评分可能是肝癌的独立预测因子。定量实时聚合酶链反应结果显示肿瘤和非肿瘤肝脏组织之间的表达水平存在显著差异。风险模型的预后价值已在TCGA-HCC和GSE76240数据集中得到验证。生物学功能分析表明，这些枢纽基因与肿瘤发生过程密切相关。基于多个数据集分析建立了一种新型的六个枢纽基因预测肝癌风险模型，为预测肝癌患者的预后提供了新特征。
