Summary
This retrospective study aimed to develop ultrasound radiomics models for predicting central lymph node metastasis (CLNM) in patients with papillary thyroid carcinoma (PTC). The study analyzed ultrasound images of PTC and used deep learning techniques to extract multimodal ultrasound image features. Three different classifiers were employed for classification: adaptive boosting (AB), linear discriminant analysis (LDA), and support vector machine (SVM).
The multimodal SVM model demonstrated the best predictive performance for CLNM, with high diagnostic accuracy. The findings suggest that these ultrasound radiomics models can be valuable for predicting CLNM and aiding in treatment decisions for PTC patients.
