引用本文:黄耀渠,赵晓梅,秦红卫,周守国,伍琼慧.颞骨测量值诊断梅尼埃病的ROC曲线分析[J].中国临床新医学,2009,2(11):1137-1139.
【打印本页】   【下载PDF全文】   查看/发表评论  【EndNote】   【RefMan】   【BibTex】
←前一篇|后一篇→ 过刊浏览    高级检索
本文已被:浏览 1353次   下载 989 本文二维码信息
码上扫一扫!
分享到: 微信 更多
颞骨测量值诊断梅尼埃病的ROC曲线分析
黄耀渠,赵晓梅,秦红卫,周守国,伍琼慧
528000 广东,佛山市中医院MRI室
摘要:
[摘要] 目的 利用ROC曲线选择颞骨测量值诊断梅尼埃病(MD)的截断点, 探讨其在诊断MD中的应用价值。方法 利用22例MD患者和20例正常志愿者MR水成像图像的半规管至颞骨岩部后缘的最短距离(PP值)、总骨脚至颞骨岩部后缘的最短距离(CP值)及前庭至颞骨岩部后缘的最短距离(VP值)数据绘制ROC曲线。结果 PP值、CP值及VP值诊断MD的ROC曲线下面积分别为0.814、0.837及0.796,最佳诊断截断点分别为2.035 mm、1.995 mm及4.730 mm,其敏感性分别为72.7%、70.5%及65.9%,特异性分别为85.0%、92.5%及90.0%。结论 PP值、CP值及VP值对诊断MD有较好的诊断价值,根据ROC曲线可以选择最佳诊断截断点。
关键词:  梅尼埃病  颞骨  接收者工作特征曲线
DOI:10.3969/j.issn.1674-3806.2009.11.07
分类号:R 764
基金项目:广东省佛山市卫生局医学科研立项课题(No.2009110)
Measurement values of the shape of temporal bone in diagnosis of Meniere′s disease using analysis of ROC curve
HUANG Yao-qu, ZHAO Xiao-mei, QIN Hong-wei, et al.
Room of MRI, Foshan Hospital of TCM, Guangdong 528000, China
Abstract:
[Abstract] Objective To use ROC curve to choose the cut-off of PP value, CP value and VP value, and investigate the value of them in the diagnosis of Meniere′s disease(MD). Methods The MR hydrography images of 22 MD patients and 20 normal volunteers were analyzed retrospectly, the minimum distance between the posterior semicircular canal and the posterior petrous surface (PP value), between the common crus and the posterior petrous surface (CP value), and between the vestibule and the posterior petrous surface (VP value) were measured in axial plane, and ROC curve was established by them.Results The area under ROC curve of PP value, CP value and VP value reached 0.814, 0.837 and 0.796, and the best diagnostic cut-off was 2.035 mm, 1.995 mm and 4.730 mm respectively, the sensitivity of them was 72.7%, 70.5% and 65.9%respectively, and specificity was 85.0%, 92.5% and 90.0%.Conclusion PP value, CP value and VP value have medium value in diagnosis of MD, useing ROC curve can choose the best cut-off of them.
Key words:  Meniere′s disease  Temporal bone  Receiver operator characteristic curve