CDFI In The Differential Diagnosis Of Thyroid Nodules

Objective To explore color Doppler ultrasonography in the differential diagnosis of benign and malignant thyroid nodules of the maximum diameter≤20mm. Analyzed the feature of ultrasonographic performance, then screed and identified the sensitive ultrasonographic feature of benign and malignant thyroid nodules in the multivariable stepwise regression analysis of binary logistic, in order to improve the accuracy of color Dopplar ultrasonography in the differential diagnosis of benign and malignant thyroid nodules.Methods 168 thyroid nodules, the maximum diameter of those were≤20mm, pathologically confirmed from 150 cases were retrospectively analyzed. The feature of ultrasonographic performance were valued statistically, the binary Logistic regression analysis was used in Logistic regression model, and risk degrees of all feature were ordered by OR value.Results (1). The binary Logistic regression analysis demonstrated that 8 ultrasonic features were enrolled in the Logistic equation, Logist(P)=7.534+1.583X2+2.169X3+2.146X4+2.260X5+1.824X6+2.069X7+ 1.770X8+1.702X9The coefficients were as follows:nodule with no clear boundary was 1.583,nodule with no or incomplete capsule was 2.169,nodule with a substantial hypoechoic inside was 2.146,nodule within microcalcification was 2.260,nodule missing surrounding sound halo was 1.824,nodule with posterior echo attenuation was 2.069,nodule with primary center blood supply was 1.770,nodule with aspect ratios≥1.0 1.702。(2). The risk degrees of 8 feature were ordered by OR value:nodule within microcalcification (9.581)>nodule with no or incomplete capsule (8.751)>nodule with a substantial hypoechoic inside (8.552)>nodule with posterior echo attenuation (7.919)>nodule missing surrounding sound halo (6.198)>nodule with primary center blood supply (5.873)>nodule with aspect ratios≥1.0 (5.484)>nodule with no clear boundary (4.867)。(3). The accuracy rate of the Logistic regression model was 91.1%.Conclusion Sonographic features such as nodule within microcalcification,nodule with no or incomplete capsule,nodule with a substantial hypoechoic inside,nodule with posterior echo attenuation,nodule missing surrounding sound halo,nodule with primary center blood supply,nodule with aspect ratios≥1.0, nodule with no clear boundary in the Logistic regression model may help to differentiate benign and malignant small thyroid nodules and contribute to the diagnosis of thyroid nodules.