基于贝叶斯网络建立风险预测模型用于VEGF受体抑制剂心脏毒性评估

项目名称： 基于贝叶斯网络建立风险预测模型用于VEGF受体抑制剂心脏毒性评估

项目编号： No.81503145

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 医药、卫生

项目作者： 何华

作者单位： 中国药科大学

项目金额： 17.9万元

中文摘要： 随着分子靶向抗肿瘤药物的广泛使用，肿瘤病人的生存率以及生存期有了显著提高，但以VEGF受体抑制剂为代表的部分靶向抗肿瘤药物的心脏毒性也引起广泛的关注和重视。由于VEGF对于维持心血管系统的正常具有重要意义，特异性抑制其受体将不可避免地损伤心血管系统，甚至引起心力衰竭的发生。因此，在临床应用该类药物时需对其心脏毒性进行早期的预测评估，并予以及时的干预，从而达到控制其心脏毒性的目的。近年来兴起的多标记物法被认为是疾病预测的一种较好手段，而进一步将该方法与数学模型相结合有助于提高对疾病风险预测的准确性。本研究拟以VEGF受体抑制剂舒尼替尼和索拉非尼作为研究对象，基于多个常用于心力衰竭预测和诊断的生物标记物构建贝叶斯网络，用于建立预测上述药物心脏毒性的贝叶斯网络模型，希望通过本研究为药物心脏毒性的预测和早期诊断提供一种新的思路与方法，同时这一方法对于其他疾病或药物毒性的风险预测也具有一定借鉴意义。

中文关键词： 心脏毒性；贝叶斯网络；多标记物法；风险预测模型；VEGF受体抑制剂

英文摘要： The use of molecularly targeted agents has markedly improved the treatment outcomes in cancer patients. However, the molecularly targeted agents such as VEGF receptor inhibitors bring both tremendous oncologic success and worrisome cardiovascular toxicities. As the inhibition of VEGF receptor affects not only the tumor tissue but also cardiovascular system, the cardiotoxicity of VEGF receptor inhibitor is inevitable. Therefore, it is extremely important to predict cardiotoxicity in patients with these molecularly targeted agents. Multimarker approach has been considered as a useful method for the prediction and early diagnosis of disease. The combination of multimarker approach and mathematical model might improve the accuracy of prediction and diagnosis. As a result, the present project proposes a bayesian network model to predict the cardiotoxicity of two VEGF receptor inhibitors, sunitinib and sorafinib. As heart failure is one of the main serious side effect of these two anti-cancer drugs, the bayesian network is developed based on the multiple biomarkers of heart failure. The present study proposes a new idea and strategy for prediction the cardiotoxicity of anticancer drugs and it is also a paradigm for other disease or toxicity prediction and evaluation.

英文关键词： Cardiotoxicity;Bayesian network;Multimarker approach;Risk prediction model;VEGF receptor inhibitor