常规MRI联合体素内不相干运动成像鉴别甲状腺良恶性结节的价值  

Value of conventional MRI and intravoxel incoherent motion imaging for differentiating benign and malignant thyroid nodules

作　　者：周静仪 周智凤 张家雄[1] 李俊[1] 李恒国[3] ZHOU Jingyi;ZHOU Zhifeng;ZHANG Jiaxiong;LI Jun;LI Hengguo(Department of Radiology,Foshan Hospital of Traditional Chinese Medicine,Foshan,Guangdong Province 528000,China;Neuropsychiatric Imaging Center,Shenzhen Kangning Hospital,Shenzhen,Guangdong Province 518000,China;Imaging Center,the First Affiliated Hospital of Jinan University,Guangzhou 510630,China)

机构地区：[1]佛山市中医院放射科,广东佛山528000 [2]深圳市康宁医院精神影像中心,广东深圳518000 [3]暨南大学附属第一医院影像中心,广东广州510630

出　　处：《实用放射学杂志》2021年第10期1594-1599,共6页Journal of Practical Radiology

基　　金：佛山市重点专科培育项目(Fspy3G2015019).

摘　　要：目的探讨常规MRI联合体素内不相干运动(IVIM)成像鉴别甲状腺良恶性结节的诊断价值.方法收集经手术病理证实的95例甲状腺结节患者155个结节的术前常规MRI及IVIM成像资料.观察甲状腺结节最大径、形态、边界、钙化或出血、囊变、有无淋巴结肿大等特征.84例患者89个结节可以测量IVIM参数值[纯扩散系数(D)值、假性扩散系数(D^(∗))值、灌注分数(f)值和表观扩散系数(ADC)值]并重建伪彩图.根据单因素分析得出有统计学意义的指标,采用似然比前进法对89个结节建立二分类Logistic回归模型并进行预报.结果单因素分析显示结节最大径、形态、边界、囊变及淋巴结肿大、ADC值及D值在鉴别诊断良恶性结节的差异有统计学意义(P<00.1).二分类Logistic回归模型的曲线下面积(AUC)为09.45,敏感度为879.%,特异度为91.1%,准确率为89.9%.结论以边界、囊变及D值建立的Logistic回归模型的诊断效能得到体现.常规MRI联合IVIM成像可以有效地用于临床评估甲状腺结节的良恶性.Objective To investigate the value of conventional MRI combined with intravoxel incoherent motion(IVIM)imaging in differential diagnosis of benign and malignant thyroid nodules.Methods Totally 95 patients with 155 thyroid nodules diagnosed by pathology were enrolled in the study.All patients underwent conventional MRI and IVIM scan before surgery.The MRI features of the maximal diameter,shape,boundary,calcification or hemorrhage,cystic change and lymph node enlargement were observed.The IVIM parameters[D,D^(*),f and apparent diffusion coefficient(ADC)values]of 84 patients with 89 thyroid nodules in normal gland and thyroid nodules group were measured and the pseudo-color images were reconstructed.According to the univariate analysis,the statistical indicators were obtained.The binary Logistic regression model of 89 nodules was established and predicted by likelihood ratio progression method.Results Univariate analysis showed that there were significant differences in the maximum diameter,shape,boundary,cystic change,lymph node enlargement,ADC value and D value in the differential diagnosis of benign and malignant nodules(P<0.01).The area under the curve(AUC),sensitivity,specificity and accuracy of the binary Logistic regression model were 0.945,87.9%,91.1% and 89.9%respectively.Conclusion The diagnostic efficiency of Logistic regression model based on boundary,cystic change and D value is demonstrated.Conventional MRI combined with IVIM imaging can effectively evaluate the benign and malignant thyroid nodules.

关 键 词：甲状腺结节 体素内不相干运动 磁共振成像 LOGISTIC回归分析 

分 类 号：R581[医药卫生—内分泌] R445.2[医药卫生—内科学]