Related Experiment Video
Updated: Jun 27, 2025
04:23
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
Published on: April 21, 2023
1.8K
CT-Based Radiomics Models for Differentiation of Benign and Malignant Thyroid Nodules: A Multicenter Development and
Shaofan Lin1, Ming Gao2,3, Zehong Yang2,3
1Department of Radiology, Shantou Central Hospital, No. 114 Waima Rd, Shantou 515031, People's Republic of China.
AJR. American Journal of Roentgenology
|May 1, 2024
Summary
A new combined model using clinical data, CT imaging features, and multiphase CT radiomics effectively differentiates benign from malignant thyroid nodules. This validated approach shows promise for guiding patient management after CT detection.
Area of Science:
- Radiology
- Oncology
- Medical Imaging Analysis
Background:
- Computed tomography (CT) increasingly detects thyroid nodules, necessitating accurate characterization.
- Prior CT-based radiomics models showed potential but lacked external validation for thyroid nodule classification.
Purpose of the Study:
- To develop and externally validate a CT-based radiomics model for differentiating benign and malignant thyroid nodules.
Main Methods:
- Retrospective analysis of 408 thyroid nodules from 378 patients across two centers, undergoing multiphase neck CT.
- Development of conventional, radiomics (noncontrast, arterial, venous, multiphase), and combined models using feature selection.
- External validation using ROC, calibration, and decision-curve analyses.
Main Results:
- The combined model, integrating clinical, morphologic CT, and multiphase CT radiomics features, achieved an AUC of 0.923 in external validation.
- This model demonstrated high sensitivity (84.0%), specificity (94.1%), and accuracy (87.0%) in the external validation set.
- The combined model significantly outperformed conventional and individual radiomics models across all validation metrics.
Conclusions:
- A combined model incorporating clinical, morphologic CT, and multiphase CT radiomics features robustly differentiates benign and malignant thyroid nodules.
- This validated radiomics model offers a promising tool to aid clinical decision-making for thyroid nodules detected on CT.
