ORIGINAL RESEARCH article
Front. Endocrinol.
Sec. Thyroid Endocrinology
Volume 16 - 2025 | doi: 10.3389/fendo.2025.1552479
This article is part of the Research TopicAdvances in Management of Aggressive Thyroid Cancer: Medullary and Advanced Thyroid CancerView all 6 articles
A Machine Learning-Based Model for Predicting Recurrence in Intermediate-and High-Risk Differentiated Thyroid Cancer: Insights from a Retrospective Single-Center Study of 2388 Patients
Provisionally accepted- 1Department of Breast and Thyroid Surgery, Wuhan Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China
- 2Department of Radiology, Wuhan Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei Province, China
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
Current guidelines provide a recognized yet broad framework for stratifying recurrence risk in differentiated thyroid cancer (DTC) patients. More precise tools are needed for intermediate-and high-risk groups. This study aims to identify recurrence-associated risk factors and develop a machine learning-based predictive model.Methods: In this retrospective analysis, 2,388 DTC patients were randomly assigned to a training group (1,910 cases) and a validation group (478 cases). Predictive factors were identified using univariate and multivariate analyses. Six machine learning models were trained and validated, with performance evaluated through accuracy, area under the curve, and clinical utility via decision curve analysis.Results: Independent risk factors for recurrence included intraglandular dissemination, total tumor size, bilateral cervical lymph node involvement, and Hashimoto's thyroiditis, while normal/elevated TSH and multifocal nodules were protective. The random forest model demonstrated the best performance (training accuracy: 0.801; validation accuracy: 0.808). A random forest-based online calculator was developed to facilitate individualized risk assessment in clinical settings.Conclusions: The random forest model effectively predicts DTC recurrence, offering a practical tool for individualized risk assessment and aiding clinical decision-making.
Keywords: differentiated thyroid cancer (DTC), cancer recurrence, predictive models, machine learning, Risk factors, random forest
Received: 28 Dec 2024; Accepted: 21 May 2025.
Copyright: © 2025 Li, Tang, Ren, Tian, zhang, Wang, Liu and Ming. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence: Jie Ming, Department of Breast and Thyroid Surgery, Wuhan Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.