论文部分内容阅读
目的根据甲状腺微小结节的三维能量多普勒定量参数、弹性值及超声特征建立Logistic回归模型,得到对其良恶性具有鉴别诊断意义的变量。方法对215例247个甲状腺微小结节进行分析和检验,采用Logistic回归得出预测甲状腺微小结节良恶性公式。结果 5个变量最终进入回归方程,该模型对甲状腺恶性微小结节预报的敏感度、特异度和准确度分别为71.01%、94.94%和88.26%。结论结节边界模糊、纵横比≥1、微钙化、后方声影衰减和弹性值较高5个特征,对超声鉴别诊断甲状腺微小结节的良恶性具有较高价值。
Objective To establish a Logistic regression model based on three-dimensional energy Doppler quantitative parameters, elastic values and ultrasonic features of small thyroid nodules, and to obtain the variables that have the significance of differential diagnosis between benign and malignant. Methods A total of 215 patients with 247 thyroid nodules were analyzed and tested. Logistic regression was used to predict benign and malignant thyroid nodules. Results Five variables finally entered the regression equation. The sensitivity, specificity and accuracy of the model for the prediction of malignant microthyroid nodules were 71.01%, 94.94% and 88.26%, respectively. Conclusion Nodular boundary fuzzy, aspect ratio ≥ 1, micro-calcification, attenuation of the back of the sound and high elasticity value of five characteristics of the differential diagnosis of benign and malignant thyroid nodules have high value.