Logistic regression analysis of ultrasound diagnosis of thyroid follicular adenoma and follicular carcinoma

This study discussed the application of Logistic regression model in differential diagnosis of benign and malignant thyroid follicular tumors by high-frequency ultrasound. A total of 90 cases of pathologically confirmed thyroid follicular tumors were selected from January 2010 to November 2022, including 73 cases of thyroid follicular adenoma (FA) and 17 cases of thyroid follicular carcinoma (FTC). Analysis measures included age, sex, tumor location, size, aspect ratio, shape, boundary, halo, internal real part echo uniformity, calcification, cystic degeneration, and blood flow. The Logistic regression model was established through univariate and multifactor regression analysis, and the ROC curve of the model was plotted and the area under the curve was calculated. The results showed that there were 5 variables in Logistic regression analysis into the regression model, and the area under the ROC curve of the model was 0.903. The Logistic regression model of thyroid follicular tumors based on high-frequency ultrasound characteristics established in this study has good diagnostic effect and important clinical value in distinguishing benign and malignant thyroid follicular tumors.