受试者工作特征曲线在血液病毒筛检试验中的应用

Receiver operating characteristic curve applied in screening for transfusion-transmitted virus

目的通过应用案例,将ROC曲线法引入血站实验室病毒筛检试验中诊断阈值的判定、及对诊断方案的评价。方法采用2种厂家生产的ELISA诊断试剂对34例阴性、22例阳性样本中的抗-HCV进行测定。基于吸光值,绘制ROC曲线对诊断试验临界值进行判定、对2组诊断试验准确性进行比较评价。结果应用厂家A试剂开展HCV诊断试验的最佳临界值为2.459,此时灵敏度与特异度综合精度达到最大,为0.73、0.97。ROC曲线下面积为0.9082,表明该诊断试验准确性较高。厂家B组试验价值优于厂家A组。结论由ROC曲线法判定的诊断阈值能够使诊断试验灵敏度、特异度达到最佳优化,诊断试验的Youden指数达到最高。ROC曲线法能够客观、一致地评价诊断试验的准确性、比较不同诊断试验的诊断价值。

Objective To promote ROC curve and to evaluate the decided thresholds and diagnostic performance in TTVs screening in blood station. Methods Enzyme-linked immunosorbent assay was used to detect antibody for hepatitis C virus （anti-HCV） in 56 human plasma （ positive 34, negative 22）. The cutoff value and accuracy between two diagnostic tests via the ROC curve were determined. Results The best cut point of manufacturer A was 2.459, with values of sensitivity and specificity 0.73 and 0.97, respectively. The area under ROC curve was 0. 9082, which suggested a higher level of diagnostic performance. The test accuracy of manufacturer B was superior to A. Conclusion An optimal statistical index of sensitivity, specificity and Youden value could be calculated by ROC curve. ROC curve leads to more objective results and can evaluate the accuracy and compare between two diagnostic tests in a more reliable way than other indices.