基于DNA甲基化和miRNA两种不同芯片数据的肾癌预后模型的比较与整合

我们前期通过microRNA（miRNA）芯片建立肿瘤miRNA预后模型表明，基于生物芯片数据的分子预后模型可弥补临床预后参数的不足，提高肿瘤预后预测准确度。随着芯片种类的增加（如DNA甲基化芯片等），肿瘤预后研究面临如何比较与整合基于不同芯片源的预后模型的新问题。本项目针对肾癌同一研究人群通过基于DNA甲基化芯片和miRNA芯片筛选差异甲基化位点和差异表达miRNA，分别应用焦磷酸测序和qPCR定量检测肾癌标本差异甲基化位点和差异表达miRNA，结合患者预后信息分别建立DNA甲基化和miRNA两种预后模型，通过国际多中心验证来评估比较两种预后模型的准确度。进一步整合两种预后模型和临床预后参数建立列线图，并通过肿瘤基因图谱（TCGA）数据库中的肾癌数据验证该列线图的预测准确度，评估基于两种不同芯片数据的复合预后模型能否显著提高肾癌预后预测准确度，更好地满足肾癌个体化治疗的需要。

We previously established a miRNA prognosis model through genome-wide microarray analysis, which shows that molecular prognosis models based on biological microarray data can make up for the insufficient of clinical parameters in the prognosis of the tumor, improve the accuracy of prediction. With the increase of biological microarray types (such as DNA methylation microarray), tumor prognosis research faces how to compare and integrate new prognosis models based on different microarray sources. According to the same study population of renal cell carcinoma（RCC）, we will use DNA methylation microarray and miRNA microarray to screen differential methylation CpG sites and differential expression miRNAs in this project, respectively. Differential methylation CpG sites and differential expression miRNAs will be detected by using pyrosequencing and qPCR with RCC specimens, respectively. We will then combine the data with patients' prognostic information in the training set to establish the DNA methylation and miRNA expression prognostic models. We will then use international multicentre dataset to validate the two models and compare the accuracy of the two kinds of prognosis model. Furthermore, we will integrate the two prognostic models and clinical prognostic parameters to establish a nomogram, and validate the prediction accuracy of the nomogram through the renal cell carcinoma data in the Cancer Genome Atlas (TCGA) database. We hope the nomogram will significantly improve the predictive accuracy for RCC prognosis, and will facilitate patient counseling and individualize management of patients with RCC.