Related Experiment Video
Updated: Jun 12, 2026
07:11
Assessing Early Stage Open-Angle Glaucoma in Patients by Isolated-Check Visual Evoked Potential
Published on: May 25, 2020
ProptoView: AI-based digital exophthalmometry using multi-view facial images in a multinational validation study
Chaoyu Lei1,2, Sifan Song3,4, Jingyuan Fan5
1State Key Laboratory of Eye Health, Department of Ophthalmology, Shanghai Ninth People's Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, 200011, China.
Journal of Translational Medicine
|June 11, 2026
Summary
ProptoView, an AI tool, accurately measures proptosis using images, offering a cost-effective alternative to CT scans. This digital exophthalmometer enhances clinical workflows and supports telemedicine for thyroid eye disease management.
Area of Science:
- Ophthalmology
- Medical Imaging
- Artificial Intelligence
Background:
- Accurate proptosis measurement is crucial for managing thyroid eye disease (TED) and other orbital conditions.
- Current methods like the Hertel exophthalmometer lack precision, while CT scans are costly and involve radiation exposure.
Purpose of the Study:
- To develop and validate ProptoView, an AI-based digital exophthalmometer for accurate, non-contact proptosis measurement.
- To compare the accuracy of ProptoView against traditional methods like the Hertel exophthalmometer and CT scans.
Main Methods:
- Developed ProptoView using a deep learning approach with 5676 images from 763 TED patients.
- Trained and validated three AI models: single-view, multi-view, and dynamic input.
- Externally validated ProptoView using 644 images from 324 patients across multiple countries and hospitals.
Main Results:
- The dynamic input AI model achieved the highest accuracy with an intraclass correlation coefficient (ICC) of 0.901, surpassing the Hertel exophthalmometer.
- The multi-view model (ICC=0.890) also outperformed the Hertel exophthalmometer (ICC=0.871).
- ProptoView demonstrated robust external validation (ICC=0.845) and reduced misclassification rates compared to the Hertel exophthalmometer.
Conclusions:
- ProptoView offers an accurate, non-contact, and cost-effective solution for proptosis measurement.
- The AI tool has the potential to streamline clinical workflows and facilitate telemedicine applications in ophthalmology.
- ProptoView represents a significant advancement in the objective assessment of orbital conditions.