Development of Radiomics-Based Risk Prediction Models for Stages of Hashimoto's Thyroiditis Using Ultrasound, Clinical, and Laboratory Factors

Jia-Hui Chen; Kai Kang; Xin-Yue Wang; Jian-Ning Chi; Xue-Meng Gao; Yong Xin Li; Ying Huang

doi:10.1016/j.ultrasmedbio.2025.05.025

Development of Radiomics-Based Risk Prediction Models for Stages of Hashimoto's Thyroiditis Using Ultrasound, Clinical, and Laboratory Factors

Ultrasound Med Biol. 2025 Jun 21:S0301-5629(25)00180-2. doi: 10.1016/j.ultrasmedbio.2025.05.025. Online ahead of print.

Authors

Jia-Hui Chen¹, Kai Kang¹, Xin-Yue Wang¹, Jian-Ning Chi², Xue-Meng Gao¹, Yong Xin Li³, Ying Huang⁴

Affiliations

¹ Department of Ultrasound, Shengjing Hospital of China Medical University, Shenyang, China.
² Faculty of Robot Science and Engineering, and Key Laboratory of Intelligent Computing in Medical Image of Ministry of Education, Northeastern University, Shenyang, China.
³ Department of Research and Development, Yizhun Medical AI Co. Ltd., Beijing, China.
⁴ Department of Ultrasound, Shengjing Hospital of China Medical University, Shenyang, China. Electronic address: huangying712@163.com.

PMID: 40545378
DOI: 10.1016/j.ultrasmedbio.2025.05.025

Abstract

Objectives: To develop a radiomics risk-predictive model for differentiating the different stages of Hashimoto's thyroiditis (HT).

Methods: Data from patients with HT who underwent definitive surgical pathology between January 2018 and December 2023 were retrospectively collected and categorized into early HT (HT patients with simple positive antibodies or simultaneously accompanied by elevated thyroid hormones) and late HT (HT patients with positive antibodies and beginning to present subclinical hypothyroidism or developing hypothyroidism). Ultrasound images and five clinical and 12 laboratory indicators were obtained. Six classifiers were used to construct radiomics models. The gradient boosting decision tree (GBDT) classifier was used to screen for the best features to explore the main risk factors for differentiating early HT. The performance of each model was evaluated by receiver operating characteristic (ROC) curve. The model was validated using one internal and two external test cohorts.

Results: A total of 785 patients were enrolled. Extreme gradient boosting (XGBOOST) showed best performance in the training cohort, with an AUC of 0.999 (0.998, 1), and AUC values of 0.993 (0.98, 1), 0.947 (0.866, 1), and 0.98 (0.939, 1), respectively, in the internal test, first external, and second external cohorts. Ultrasound radiomic features contributed to 78.6% (11/14) of the model. The first-order feature of traverse section of thyroid ultrasound image, texture feature gray-level run length matrix (GLRLM) of longitudinal section of thyroid ultrasound image and free thyroxine showed the greatest contributions in the model.

Conclusion: Our study developed and tested a risk-predictive model that effectively differentiated HT stages to more precisely and actively manage patients with HT at an earlier stage.

Keywords: Artificial intelligence; Clinical indicators; Early diagnosis; Hashimoto's thyroiditis; Machine learning; Radiomics; Risk prediction; Ultrasound imaging.