Related Experiment Video
Updated: Jun 8, 2025
06:08
Quantitative Visualization and Detection of Skin Cancer Using Dynamic Thermal Imaging
Published on: May 5, 2011
16.8K
Artificial intelligence-enhanced infrared thermography as a diagnostic tool for thyroid malignancy detection
Panpicha Chantasartrassamee1, Boonsong Ongphiphadhanakul1, Ronnarat Suvikapakornkul2
1Division of Endocrinology and Metabolism, Department of Medicine, Faculty of Medicine Ramathibodi Hospital, Mahidol University, Bangkok, Thailand.
Annals of Medicine
|November 8, 2024
Summary
Infrared thermal imaging combined with artificial intelligence (AI) shows promise for detecting thyroid cancer. This noninvasive tool achieved 75% accuracy in distinguishing malignant from benign thyroid nodules.
Area of Science:
- Medical imaging
- Artificial intelligence in medicine
- Oncology
Background:
- Thyroid nodules are common, necessitating investigation to rule out malignancy.
- Increased intranodular vascularity in malignant tumors correlates with elevated skin surface temperatures.
- Noninvasive infrared thermography can detect these temperature changes.
Purpose of the Study:
- To develop a diagnostic tool for thyroid cancer detection.
- To evaluate the efficacy of infrared thermal images combined with an artificial intelligence (AI) algorithm for malignancy diagnosis.
Main Methods:
- Prospective cross-sectional study involving 98 participants with thyroid nodules.
- Infrared thermal images collected pre-surgery using a thermal camera.
- Machine learning model (ResNet50V2) applied to thermal images for malignancy diagnosis, validated against pathological reports.
Main Results:
- The study included 58 malignant and 40 benign thyroid nodules.
- AI-enhanced infrared thermal imaging achieved 75% accuracy and 78% sensitivity in distinguishing malignant from benign nodules.
- These results were slightly lower than AI prediction integrating ultrasound and cytology (81% accuracy, 84% sensitivity).
Conclusions:
- AI-assisted infrared thermal imaging demonstrates good performance in differentiating thyroid malignancy.
- This noninvasive imaging modality shows potential as a screening tool for adjunct evaluation of thyroid nodules.
- Further research may enhance its diagnostic capabilities for thyroid cancer.
