利用ROC曲线对门脉血栓进行确诊预测

The Prediction of the Portal Vein Thrombosis by Using the ROC Curve

为了对门脉血栓(PVT)进行较为精确的早期预测,减少实际治疗过程中的盲目性,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PVT的情况,收集所有患者的P-选择素、血栓前体蛋白(TPP)、D-二聚体和血小板含量(PLT)(术后1、3、5、7、14天)的临床观测指标数据,利用受试者工作曲线(ROC曲线),根据患者术后形成血栓的时间结合不同的时间间隔,对上述4个指标的预测价值进行比较.研究结果表明:血小板对PVT几乎没有预测价值,而P-选择素的预测效果最好,血栓前体蛋白和D-二聚体也有较好的预测效果.同时也发现术后第3天的测量数据在预测门脉血栓方面具有较高价值.通过本文研究可知,在实际治疗过程中必须重点关注患者术后P选择素含量的变化,同时还要结合指标TPP和D2的含量变化,只有结合这三种指标进行诊断,才能使诊断更具意义.

The goal of this paper is to make the early prediction of the portal vein thrombosis （PVT） be more accurate so as to reduce the blindness in the process of actual medical treatment. In this paper, we take the cirrhotic portal hypertension patients submitted to portal hypertension operation as research objects. The indica- tors of clinical observation data of P - selection, thrombus precursor （TPP）, D- dimer and platelet （1, 3, 5,7, 14 days after operation） are observed and recorded. Then we take the time of thrombosis of every patient after operation into account and use the receiver - operating characteristic curve （ROC curve） combined with different interval to compare the predictive value of the above four indexes. The research results show that： platelet （PLT） almost has no predictive value to PVT, but the P - selection has the best predictive value, and thrombus precursor protein and D- dimer also have good forecast effect. At the same time, we find that the measurement data of the third day after surgery is of high value in terms of predicting PVT. Through this study we can know that in actual medical treatment doctors must pay more attention to the changes in levels of P- selection and at the same time doctors should observe the changes of TPP and D2. Considering the changes of the P - selection, TPP and D2 can make the diagnosis be more meaningful.