Li Kang, Song Yi, Qin Ling, Li Ang, Jiang Sanjie, Ren Lei, Zang Chaoran, Sun Jianping, Zhao Yan, Zhang Yonghong
Biomedical Information Center, Beijing You'An Hospital, Capital Medical University, Beijing, China.
Experimental Center, Beijing Friendship Hospital, Capital Medical University, Beijing, China.
Front Oncol. 2021 Oct 20;11:756326. doi: 10.3389/fonc.2021.756326. eCollection 2021.
Aberrant methylation of CpG sites served as an epigenetic marker for building diagnostic, prognostic, and recurrence models for hepatocellular carcinoma (HCC).
Using Illumina 450K and EPIC Beadchip, we identified 34 CpG sites in peripheral blood mononuclear cell (PBMC) DNA that were differentially methylated in early HCC versus HBV-related liver diseases (HBVLD). We employed multiplex bisulfite sequencing (MBS) based on next-generation sequencing (NGS) to measure methylation of 34 CpG sites in PBMC DNA from 654 patients that were divided into a training set ( = 442) and a test set ( = 212). Using the training set, we selected and built a six-CpG-scorer (namely, cg14171514, cg07721852, cg05166871, cg18087306, cg05213896, and cg18772205), applying least absolute shrinkage and selection operator (LASSO) regression. We performed multivariable analyses of four candidate risk predictors (namely, six-CpG-scorer, age, sex, and AFP level), using 20 times imputation of missing data, non-linearly transformed, and backwards feature selection with logistic regression. The final model's regression coefficients were calculated according to "Rubin's Rules". The diagnostic accuracy of the model was internally validated with a 10,000 bootstrap validation dataset and then applied to the test set for validation.
The area under the receiver operating characteristic curve (AUROC) of the model was 0.81 (95% CI, 0.77-0.85) and it showed good calibration and decision curve analysis. Using enhanced bootstrap validation, adjusted C-statistics and adjusted Brier score were 0.809 and 0.199, respectively. The model also showed an AUROC value of 0.84 (95% CI 0.79-0.88) of diagnosis for early HCC in the test set.
Our model based on the six-CpG-scorer was a reliable diagnosis tool for early HCC from HBVLD. The usage of the MBS method can realize large-scale detection of CpG sites in clinical diagnosis of early HCC and benefit the majority of patients.
CpG 位点的异常甲基化作为一种表观遗传标记，用于构建肝细胞癌（HCC）的诊断、预后和复发模型。
我们使用 Illumina 450K 和 EPIC Beadchip，在 654 例患者外周血单个核细胞（PBMC）DNA 中鉴定出 34 个 CpG 位点，这些位点在早期 HCC 与乙肝相关肝病（HBVLD）中存在差异甲基化。我们采用基于下一代测序（NGS）的多重亚硫酸氢盐测序（MBS）来检测 654 例患者 PBMC DNA 中 34 个 CpG 位点的甲基化情况，这些患者被分为训练集（n = 442）和测试集（n = 212）。利用训练集，我们应用最小绝对收缩和选择算子（LASSO）回归选择并构建了一个六 CpG 评分器（即 cg14171514、cg07721852、cg05166871、cg18087306、cg05213896 和 cg18772205）。我们对四个候选风险预测因子（即六 CpG 评分器、年龄、性别和 AFP 水平）进行多变量分析，采用 20 次缺失数据插补、非线性变换以及逻辑回归的向后特征选择。最终模型的回归系数根据“鲁宾法则”计算。该模型的诊断准确性通过 10000 次自助验证数据集进行内部验证，然后应用于测试集进行验证。
该模型的受试者操作特征曲线下面积（AUROC）为 0.81（95%CI，0.77 - 0.85），并且显示出良好的校准和决策曲线分析。使用增强自助验证，调整后的 C 统计量和调整后的布里尔评分分别为 0.809 和 0.199。该模型在测试集中对早期 HCC 的诊断 AUROC 值也为 0.84（95%CI 0.79 - 0.88）。
我们基于六 CpG 评分器的模型是一种用于从 HBVLD 中诊断早期 HCC 的可靠工具。MBS 方法的使用能够在早期 HCC 的临床诊断中实现对 CpG 位点的大规模检测，使大多数患者受益。
