应用ROC曲线评价彩色多普勒血流显像对肾动脉狭窄的诊断价值

Evaluation of color Doppler flow imaging in diagnosis of renal artery stenosis by receiver operating characteristic curves

目的 用受试者工作特征（receiver operating characteristic，ROC）曲线评价彩色多普勒血流显像（CDFI）诊断肾动脉狭窄（RAS）的价值。方法 选择同时行肾动脉CDFI及数字减影血管造影（DSA）检查的RAS患者56例，以DSA检查结果为“金标准”，绘制CDFI诊断RAS的正确率及RAS狭窄程度的ROC曲线并计算曲线下面积（AUC）。根据经过培训的检查者从事彩色多普勒超声诊断工作的年资及所使用的仪器分组，比较各组间的诊断符合率。结果CDFI诊断RAS正确率ROC曲线的AUc达0．756±0．048，用CDFI诊断RAS具有统计学意义（P〈0．001）；CDFI诊断RAS狭窄程度ROC曲线的AUC仅为0．512±0．065，即用CDFI诊断RAS狭窄程度无统计学意义（P〉0．05）。不同年资超声医师对RAS的诊断符合率差异无统计学意义（P〉0．05），不同仪器组对RAS的诊断符合率差异无统计学意义（P〉O．05）。结论 在规范化检查的条件下，CDFI可作为诊断RAS的有效筛选手段，为临床提供重要的诊断信息，但对RAS程度的判断则价值有限。

Objective To evaluate the value of color Doppler flow imaging （CDFI） in diagnosis of renal artery stenosis （RAS） by receiver operating characteristic （ROC） curves. Methods Fifty-six patients were performed ultrasound and digital subtraction angiography （DSA）. ROC curve of CDFI diagnosing RAS and RAS＇ degree were drawn depending on DSA, which was regard as ＂gold standard＂, and the areas under curves were calculated. The diagnose accordance rate was compared among examiners with different past services or among different equipments. Results The results of ROC diagnosing RAS by CDFI demonstrated that the area under curves reached 0. 756 ± 0. 048, there was significant statistical difference （ P 〈 0.001 ）. The results of ROC diagnosing the degree of RAS by CDFI demonstrated that the area under curves only reached 0. 512 ± 0. 065 ,there was no significant statistical difference （ 19 〉0. 05）. There was no significant different about the diagnosis of RAS among the ultrasound doctors with different past services or among different equipments（ P〉0. 05）. Conclusions In standardized checking condition,CDFI is an effectual inspecting method to evaluate RAS, but it is limited to value the degree of RAS.