Value of a column-line diagram model incorporating BRAFV600E gene and enhanced CT in the differential diagnosis of thyroid nodules categorized as TI-RADS 3 and above
Objective To construct a thyroid imaging reporting and data system (TI-RADS) benign-malignant prediction model for thyroid nodules categorized as TI-RADS 3 and above, incorporating both BRAFV600E gene mutation status and enhanced CT features, and assess its diagnostic efficacy. Methods A retrospective analysis of data from 251 patients with TI-RADS 3 and above thyroid nodules admitted to Chaohu Hospital of Anhui Medical University from October 2022 to February 2024 were conducted. Ultrasound-guided fine-needle aspiration cytology and postoperative pathology served as the "gold standard", with 177 nodules classified as benign and 74 as malignant. The LASSO regression method was employed for variable and predictor selection, leading to the establishment of a prediction model. Results LASSO regression identified four variables for inclusion in the prediction model: age, BRAFV600E gene mutation status, presence of blurred borders on enhanced CT, and discontinuity of the nodule envelope. A prediction model for BRAFV600E gene mutation status in enhanced CT was developed based on these variables and subsequently validated. The AUC for the combined prediction model was 0.816, surpassing that of the enhanced CT prediction model alone (AUC=0.755) with statistical significance (P< 0.05). The joint prediction model demonstrated a sensitivity of 88.7％, specificity of 63.5％, and accuracy of 81.7％, with a Hosmer-Lemeshow fit test yielding P=0.4564. The net reclassification index compared to the enhanced CT prediction model alone was 0.308 (0.151-0.465) (P<0.001), and the integrated discrimination improvement index was 0.114 (0.060-0.167) (P<0.001). Decision curve analysis and calibration curves confirmed the high predictive performance of the combined prediction model. Conclusion The column-line diagram model combining BRAFV600E gene mutation status with enhanced CT features demonstrates significant diagnostic value in distinguishing between benign and malignant nodules categorized as TI-RADS 3 and above.
thyroid noduleBRAFV600E geneenhanced CTcolumn line drawing modelvalidationnet reclassification index
国家科技期刊平台
NSTL
万方数据
