【作者】 张一休； 李建初； 欧阳云淑； 孟华； 张波； 戴晴； 姜玉新；

【Author】 Zhang Yixiu;Li Jianchu;Ouyang Yunshu;Department of Ultrasound,Peking Union Medical College Hospital,Chinese Academy of Medical Science;

【机构】 中国医学科学院北京协和医院超声科；

【摘要】 目的筛选甲状腺微小癌的独立恶性征象,评价灰阶超声对其诊断价值。方法回顾性分析手术病理证实的145个甲状腺微小结节(良性结节73个,微小癌72个)的灰阶超声表现,通过Logistic回归分析筛选独立诊断指标,建立Logistic模型,并进一步评价其预报能力。结果最后进入Logistic模型的5个特征分别为:钙化类型、回声水平、周边声晕、内部回声和结节形态。Logistic模型的ROC曲线下面积为0.929。如果以预报概率P=0.5作为阈值,则预测准确率高达86.2%。结论微小钙化、极低回声、不规则声晕、实性、形态不规整是甲状腺微小癌的独立特征指标。灰阶超声特征的Logistic回归模型有助于甲状腺微小癌的鉴别诊断。更多还原

【Abstract】 Objective To evaluate malignant gray-scale sonographic features of thyroid microcarcinoma(TMC)and to assess the diagnostic accuracy of sonographic findings of subcentimeter thyroid nodules.Methods The sonographic characteristics of 72 TMC and 73 benign thyroid nodules(≤1 cm) were retrospectively reviewed.A Logistic model for predicting TMC was obtained,and assessment of the predicting value was performed.Results 5 ultrasonic features were finally entering the Logistic model,they were calcification type,echogenicity,halo sign,internal solid/cystic component and configuration.The area under curve was 0.929.If apply P=0.5 as cutoff,the diagnostic accuracy was 87.6%.Conclusions Microalcification,marked hypoechoic,irregular halo sign,solid lesion and irregularshape were main characteristics of TMC.The Logistic model can efficiently differentiate malignant subcentimeter thyroid nodules from benign ones.更多还原