常规灰阶超声和超声造影诊断甲状腺结节模型的建立

Modeling gray-scale and contrast-enhanced ultrasound in the diagnosis of thyroid lesions

目的探寻常规灰阶超声和超声造影单独及联合诊断甲状腺良恶性结节的典则判别模型，并探讨其诊断价值。方法综合分析经手术病理证实的211个甲状腺结节的超声表现，对常规灰阶超声显示的形状（x1）、生长方向（X2）、内部回声（X3）、周边暗环（X4）和微钙化（X5）5项指标和超声造影显示的结节周边（x6）及内部（X7）开始增强时间、结节周边（X8）及内部（X9）峰值强度、环状增强（x10）和增强均-度（x11）6项指标进行判别分析，分别拟合常规灰阶超声、超声造影及两者联合诊断甲状腺结节的典则判别函数，并获得三者诊断甲状腺良恶性结节的诊断正确率。结果常规灰阶超声诊断甲状腺良恶性结节的典则判别函数为：g1（X）=0．715X．＋0．276X，＋1．028X，＋1．197X。＋0．923X，-2．202，诊断正确率为86．3％。超声造影诊断甲状腺结节的典则判别函数为：g，（x）=-0．392X。＋0．541X。-0．117X。＋0．562X。＋1．173X，。＋2．200X。，-1．956，诊断正确率为89．1％。常规灰阶超声和超声造影联合诊断甲状腺良恶性结节的典则判别函数为：g，（x）=0．418X1＋0．173X2＋O．626X3＋0．558X4＋0．183X5—0．476X6＋0．474X7—0．071X8＋0．399X。＋O．985X，。＋1．639X，。-2．530，诊断正确率为91．0％。结论常规灰阶超声和超声造影有助于甲状腺良恶性结节的鉴别诊断，两者联合使用时诊断价值最高。

Objective To investigate the discriminant analysis model of gray-scale ultrasound （ GSUS）, contrast-enhanced ultrasound （ CEUS ） and the combination of them in the differential diagnosis of benign and malignant thyroid lesions and the diagnostic values. Methods Ultrasound images of 211 thyroid lesions confirmed by pathology were synthetically reviewed by scoring 5 GSUS indicators including shape （ Xl ） , orientation （ X2 ） , interior eehogenicity （ X3 ） , halo sign （ X4 ） , and microcaleification （ X5 ） , as well as 6 CEUS indicators including relative arrival time of mierobubbles in the periphery （ X6 ） and interior （ X7 ） , peak periphery （ Xs ） and interior （ X9 ） echogenicity, peripheral ring-enhancement （X^0） , homogeneity of enhancement （XH ）. The diagnostic models with their values of GSUS, CEUS and the combination of them were explored by discriminant analysis. Results The diseriminant analysis function of GSUS in the diagnosis of thyroid benign and malignant lesions was g~ （X）= 0.715X1 ＋0. 276X2 ＋ 1. 028X3 ＋ 1. 197X4 ＋0. 923X5 -2. 202 with the diagnostic value 86.3% , the discriminant analysis function of CEUS was g2（X） = -0. 392X6 ＋0. 541X7 -0.117X8 ＋0. 562X9 ＋ 1. 173X10 ＋2. 200Xll - 1. 956 with the diagnostic value 89.1% , and the discriminant analysis function of the combination of GSUS and CEUS was g3 （ X ） = 0.418X1＋0. 173X2 ＋0. 626X3 ＋0. 558X4 ＋0. 183X5 -0. 476X6 ＋0. 474X7 -0. 071Xs ＋0. 399X9 ＋0. 985X^0 ＋ 1. 639X, -2. 530 with the diagnostic value 91.0%. Conclusions GSUS and CEUS were valuable in the differential diagnosis of benign and malignant thyroid lesions, and the combination of GSUS and CEUS was most valuable.