Application of artificial intelligence in the diagnosis and treatment of thyroid cancer
- 1Department of Thyroid Surgery, The First Affiliated Hospital of Zhengzhou University
Abstract
Thyroid cancer is one of the most common endocrine malignancies, and its rising incidence has increased the demand for more accurate and efficient diagnostic and management strategies. Current clinical management mainly relies on ultrasonography, fine-needle aspiration (FNA), surgery, radioactive iodine therapy, thyroid-stimulating hormone suppression, and systemic treatment in selected advanced cases. However, substantial interobserver variability, indeterminate cytology, and heterogeneity in tumor behavior still complicate diagnosis and risk stratification. In thyroid cancer, AI(Artificial intelligence) -assisted models have been applied to ultrasound, CT, MRI, cytology, histopathology, and genomic data, with promising performance in distinguishing benign from malignant nodules, predicting lymph node metastasis and key mutations, and supporting individualized management. This review critically synthesizes evidence across these applications, focusing on study design, validation strategies, generalizability, and clinical relevance. It also examines explainable AI, emerging foundation models, and regulatory and implementation challenges. Although many models report high diagnostic or predictive performance, these results alone do not establish improvements in clinical decision-making or patient outcomes. Prospective clinical validation, cost-effectiveness assessment, and evaluation of integration into clinical workflows are needed to determine whether AI provides meaningful benefits over standard care.
Related
- Artificial Intelligence for Thyroid Nodule Diagnosis: From Multimodal Integration to Clinical Utility.
- Artificial Intelligence-Guided Extracellular Vesicle Platforms for Diagnosis and Treatment of Chronic Kidney Disease: A Scoping Review and Future Directions
- Artificial intelligence-assisted bronchoscopy for the diagnosis and treatment of pulmonary diseases: Advances and clinical practice
- The application progress of artificial intelligence in the diagnosis of Alzheimer’s disease
- Artificial intelligence in hepatitis: Advances in diagnosis, treatment optimization, and prognostic prediction
BibTeX
@article{su2026application,
title = {{Application of artificial intelligence in the diagnosis and treatment of thyroid cancer}},
author = {Zhen Su and Shuo Wang and Xintao Wang and Wenjing Yao and Yanshu Li and Hao Wang and Jie Wang and Meng Jia},
journal = {Frontiers in Endocrinology},
year = {2026},
month = nov,
doi = {10.3389/fendo.2026.1911045},
url = {https://www.frontiersin.org/articles/10.3389/fendo.2026.1911045/full},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Application of artificial intelligence in the diagnosis and treatment of thyroid cancer
AU - Zhen Su
AU - Shuo Wang
AU - Xintao Wang
AU - Wenjing Yao
AU - Yanshu Li
AU - Hao Wang
AU - Jie Wang
AU - Meng Jia
PY - 2026
DA - 2026/11/04
JO - Frontiers in Endocrinology
DO - 10.3389/fendo.2026.1911045
UR - https://www.frontiersin.org/articles/10.3389/fendo.2026.1911045/full
AB - Thyroid cancer is one of the most common endocrine malignancies, and its rising incidence has increased the demand for more accurate and efficient diagnostic and management strategies. Current clinical management mainly relies on ultrasonography, fine-needle aspiration (FNA), surgery, radioactive iodine therapy, thyroid-stimulating hormone suppression, and systemic treatment in selected advanced cases. However, substantial interobserver variability, indeterminate cytology, and heterogeneity in tumor behavior still complicate diagnosis and risk stratification. In thyroid cancer, AI(Artificial intelligence) -assisted models have been applied to ultrasound, CT, MRI, cytology, histopathology, and genomic data, with promising performance in distinguishing benign from malignant nodules, predicting lymph node metastasis and key mutations, and supporting individualized management. This review critically synthesizes evidence across these applications, focusing on study design, validation strategies, generalizability, and clinical relevance. It also examines explainable AI, emerging foundation models, and regulatory and implementation challenges. Although many models report high diagnostic or predictive performance, these results alone do not establish improvements in clinical decision-making or patient outcomes. Prospective clinical validation, cost-effectiveness assessment, and evaluation of integration into clinical workflows are needed to determine whether AI provides meaningful benefits over standard care.
SN - 1664-2392
ER -