首页 | 本学科首页   官方微博 | 高级检索  
     

基于ROC方法对门脉高压脾切后PVT早期预测的诊断研究
引用本文:冯文佳,郑璐茜,来丹飞,陈雪东,魏云海. 基于ROC方法对门脉高压脾切后PVT早期预测的诊断研究[J]. 应用数学与计算数学学报, 2014, 0(1): 109-116
作者姓名:冯文佳  郑璐茜  来丹飞  陈雪东  魏云海
作者单位:湖州师范学院理学院;湖州市中心医院
基金项目:国家社会科学基金资助项目(10BTJ001);国家自然科学基金资助项目(11171105)
摘    要:以因肝硬化行门脉高压脾切除术的患者为研究对象,观察并记录患者术后并发门脉血栓(PVT)的情况,收集所有患者的P-选择素、血栓前体蛋白(TPP)和D2聚体(D2)含量(术后第1,3,5,7,14天)的临床观察指标数据,分析各临床观察指标与并发PVT的关系.建立相应的统计模型,并利用受试者工作特征曲线(ROC曲线)方法,比较各指标在单独和联合预测PVT时的价值和意义,确定最佳的术后预测时间和阈值,进而对PVT的形成进行较准确、早期的预测,为临床防治相关患者并发PVT打下基础.

关 键 词:ROC曲线  分类树  P-选择素  D2聚体(D2)  血栓前体蛋白(TPP)

Diagnostic research on early prediction of PVT after portal hypertension operation based on ROC method
FENG Wen-jia;ZHENG Lu-xi;LAI Dan-fei;CHEN Xue-dong;WEI Yun-hai. Diagnostic research on early prediction of PVT after portal hypertension operation based on ROC method[J]. Communication on Applied Mathematics and Computation, 2014, 0(1): 109-116
Authors:FENG Wen-jia  ZHENG Lu-xi  LAI Dan-fei  CHEN Xue-dong  WEI Yun-hai
Affiliation:FENG Wen-jia;ZHENG Lu-xi;LAI Dan-fei;CHEN Xue-dong;WEI Yun-hai;School of Science,Huzhou Teachers College;Huzhou Central Hospital;
Abstract:In order to investigate the predictive factor of the portal vein thrombosis (PVT), the cirrhotic portal hypertension patients submitted to portal hypertension operation are taken as research objects. Clinical indexes of P-selection, thrombus precursor protein (TPP), .and D-dimer (1st, 3rd, 5th, 7th, and 14th days after operation) are observed and recorded, and then the relations between those clinical indexes and the concurrent PVT are analyzed with a statistical model and a receiver-operating characteristic curve (ROC curve) method. Comparing the various indexes in predicting PVT alone and in combination of value and significance, the best prediction time and postoperative threshold are determined. It plays an important role in clinical prevention of cirrhotic portal hypertension patients concurrent PVT disease.
Keywords:ROC curve  classification tree  P-selection  D-dimer (D2)  thrombusprecursor protein (TPP)
本文献已被 CNKI 维普 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号