Related Experiment Video
Updated: May 4, 2026
04:23
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
Published on: April 21, 2023
2.6K
A model to discriminate malignant from benign thyroid nodules using artificial neural network.
Lu-Cheng Zhu1, Yun-Liang Ye2, Wen-Hua Luo1
1Department of Radiation Oncology and Chemotherapy, The First Affiliated Hospital of Wenzhou Medical College, Wenzhou, China.
Plos One
|December 21, 2013
Summary
An artificial neural network (ANN) model using ultrasound features accurately differentiates benign and malignant thyroid nodules. This AI approach enhances diagnostic precision for thyroid cancer detection.
Area of Science:
- Medical imaging
- Artificial intelligence in medicine
- Oncology
Background:
- Thyroid nodules are common, requiring accurate differentiation between benign and malignant types.
- Ultrasound (US) is a primary imaging modality, but objective diagnostic accuracy can be improved.
Purpose of the Study:
- To develop an artificial neural network (ANN) model for differentiating benign and malignant thyroid nodules.
- To enhance the objective diagnostic accuracy of ultrasound in thyroid nodule assessment.
Main Methods:
- A dataset of 689 thyroid nodules (425 malignant, 264 benign) from 618 patients was analyzed.
- Six significant sonographic features (shape, margin, echogenicity, internal composition, calcifications, peripheral halo) were selected as inputs for the ANN model.
- A three-layer feed-forward ANN was constructed and validated.
Main Results:
- Six sonographic features were significantly associated with malignancy (p<0.001 for all).
- The ANN model achieved high diagnostic performance: training accuracy 82.3% (AUROC 0.818), validation accuracy 83.1% (AUROC 0.828).
- Sensitivity and specificity were consistently high in both training and validation cohorts.
Conclusions:
- Artificial neural network models utilizing sonographic features can effectively discriminate between benign and malignant thyroid nodules.
- The developed ANN model demonstrates high diagnostic accuracy, offering a valuable tool for objective assessment of thyroid nodules.
