Related Experiment Video
Updated: Jun 23, 2026
Heterotopic and Orthotopic Tracheal Transplantation in Mice used as Models to Study the Development of Obliterative Airway Disease
Published on: January 20, 2010
Building and validating an artificial intelligence model to identify tracheobronchopathia osteochondroplastica by
Chongxiang Chen1, Fei Tang2, Felix J F Herth3
1State Key Laboratory of Respiratory Disease, National Clinical Research Center for Respiratory Disease, Guangzhou Institute of Respiratory Health, The First Affiliated Hospital of Guangzhou Medical University, Guangzhou, Guangdong Province, China.
An artificial intelligence model can now differentiate tracheobronchopathia osteochondroplastica (TO) from other multinodular airway diseases using bronchoscopic images. This AI tool aids physicians in diagnosing this rare condition, improving patient care.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Pulmonology
Background:
- Tracheobronchopathia osteochondroplastica (TO) is a rare airway disease.
- Young doctors may struggle to identify TO from bronchoscopic findings due to its rarity.
Purpose of the Study:
- To develop an artificial intelligence (AI) model for differentiating TO from other multinodular airway diseases using bronchoscopic images.
- To assist healthcare professionals in the accurate diagnosis of TO.
Main Methods:
- A deep learning model using convolutional neural networks (CNNs) was trained on 2183 bronchoscopic images of multinodular lesions (including TO, amyloidosis, tumors, and inflammation) and 1733 images without airway lesions.
- The model, EfficientNet, was validated using bronchoscopic images from 21 TO patients.
- Data were collected from January 2010 to October 2022.
Main Results:
- The AI model achieved 98.9% accuracy in identifying multinodular lesions.
- The model demonstrated 89.2% accuracy in detecting TO from multinodular lesions.
- External validation showed 89.8% accuracy, with all 21 TO cases correctly diagnosed.
Conclusions:
- An AI model was successfully developed to differentiate TO from other multinodular airway diseases based on bronchoscopic images.
- This AI tool can significantly aid physicians, especially those in primary hospitals, in identifying rare airway diseases like TO.
