受试者工作特征曲线对头影测量项目在拔牙判别中的应用分析

The application of ROC curve in cephalometric measurement in tooth extraction discriminant analysis

目的利用受试者工作特征曲线（receiveroperationcharacteristiccurve，ROC）曲线研究不同的头影测量指标对拔牙判别的效能，从而筛选简便的头影测量指标来进行拔牙判别。方法选取98例正畸患者，年龄在16-40岁，平均年龄27岁，对其进行常规头影测量分析。计算所有头影测量指标的ROC曲线下面积（areaundercurve，AUC）和约登指数对应的诊断阈值。统计学分析使用SPSS18．0软件包。结果上切牙至AB连线的AUC值为0．848，约登指数对应的诊断阈值为6．5mm；下切牙切缘至NB连线的AUC值为0．835，对应的诊断阈值为6n3En。结论上切牙切缘至AB连线距离，下切牙切缘至NB连线距离是良好的拔牙判定指标。

Objective ROC curve was used to evaluate the discriminant function of various cephalometric items in orthodontic extraction planning. Methods 98 orthodontic patients were chosen, aged 16 to 40 years old. The average age was 27. The area under the ROC curve of cephalometric measurement was calculated by using SPSS 18. 0 software. Results The AUC value of UIE-AB was 0. 848, and the corresponding diagnosis threshold was 6. 5 mm; The AUC value of LIE-AB was 0. 835, and the corresponding diagnosis threshold was 6 mm. Conclusions UIE-AB and LIE-NB were reliable items for orthodontic extraction planning.