Related Experiment Video
Updated: Jan 18, 2026
04:23
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
Published on: April 21, 2023
2.3K
Comparative Analysis of Diagnostic Performance Between Elastography and AI-Based S-Detect for Thyroid Nodule
Jee-Yeun Park1,2, Sung-Hee Yang2
1Department of Radiological Science, Jangpalpal Internal Medicine Clinic, 369, Haeundae-ro, Haeundae-gu, Busan 48062, Republic of Korea.
Diagnostics (Basel, Switzerland)
|September 13, 2025
Summary
Elastography and radiologist assessments are more accurate than AI-based S-detect for distinguishing benign from malignant thyroid nodules. Radiologist interpretation showed the highest diagnostic accuracy, followed by elastography, with S-detect performing lowest.
Area of Science:
- Medical Imaging
- Oncology
- Diagnostic Technology
Background:
- Elastography is a non-invasive imaging technique assessing tissue stiffness and elasticity.
- Thyroid nodules require accurate differentiation between benign and malignant types for effective management.
- Deep learning-based computer-aided diagnosis (DL-CAD) software like S-detect offers potential for nodule analysis.
Purpose of the Study:
- To evaluate the diagnostic performance and clinical utility of elastography and S-detect in distinguishing benign from malignant thyroid nodules.
- To compare the accuracy of elastography, S-detect, and radiologist interpretation using fine needle aspiration cytology (FNAC) as the gold standard.
Main Methods:
- A retrospective study of 159 patients who underwent thyroid ultrasonography, elastography, S-detect analysis, and FNAC.
- Malignancy status was confirmed by FNAC findings.
- Diagnostic performance metrics including accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were assessed for each method.
Main Results:
- Radiologist interpretation achieved the highest diagnostic accuracy (AUC 89%), with high sensitivity (98.28%) and NPV (98.8%).
- Elastography (elasticity contrast index) showed good performance with 85% accuracy, 87.93% sensitivity, and 81.19% specificity.
- S-detect demonstrated the lowest accuracy (78%), with a sensitivity of 87.93% but lower specificity (68.32%) and PPV (61.4%).
Conclusions:
- Radiologist interpretation and elastography are valuable tools for diagnosing thyroid nodules, outperforming AI-based S-detect in this study.
- While S-detect shows potential, further development and validation are needed to improve its diagnostic accuracy for thyroid nodules.
- The study highlights the comparative diagnostic utility of these methods in clinical practice, despite limitations of single-center design and sample size.
