Related Experiment Video
Updated: Jul 5, 2025
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
The BONSAI (Brain and Optic Nerve Study with Artificial Intelligence) deep learning system can accurately identify
Mung Yan Lin1, Raymond P Najjar2, Zhiqun Tang3
1Department of Ophthalmology, Emory University School of Medicine, Atlanta, Georgia.
Insights
A deep-learning system (DLS) accurately identifies pediatric papilledema from ocular fundus photographs. This AI tool aids in diagnosing optic disk abnormalities in children, improving early detection of serious neurologic conditions.
Area of Science:
- Ophthalmology
- Artificial Intelligence
- Pediatric Neurology
Background:
- Pediatric papilledema often indicates severe neurologic disorders and can be challenging to diagnose in children.
- Ocular fundus photography offers a feasible method for data acquisition in pediatric patients, even in non-specialist settings.
Purpose of the Study:
- To evaluate an improved deep-learning system (DLS), previously validated in adults, for its accuracy in identifying papilledema and other optic disk abnormalities in children.
- To assess the DLS's capability to serve as a diagnostic aid for optic nerve head assessment in the pediatric population.
Main Methods:
- The DLS was tested on mydriatic fundus photographs from a multiethnic pediatric cohort (<17 years) across three international centers.
- Performance metrics including multiclass accuracy, area under the curve (AUC), sensitivity, and specificity were calculated, comparing DLS performance against expert neuro-ophthalmologists.
Main Results:
- The DLS achieved an overall multiclass accuracy of 89.6% on 898 fundus photographs from 447 pediatric patients.
- The system demonstrated high accuracy in distinguishing normal from abnormal optic disks (AUC 0.99) and specifically detecting papilledema (sensitivity 98.0%, specificity 94.1%).
Conclusions:
- The DLS reliably differentiates papilledema from normal optic disks and other abnormalities in children.
- The findings suggest the DLS can be a valuable diagnostic aid for assessing optic nerve head appearance in pediatric patients.
Background:
Pediatric papilledema often reflects an underlying severe neurologic disorder and may be difficult to appreciate, especially in young children. Ocular fundus photographs are easy to obtain even in young children and in nonophthalmology settings. The aim of our study was to ascertain whether an improved deep-learning system (DLS), previously validated in adults, can accurately identify papilledema and other optic disk abnormalities in children.
Methods:
The DLS was tested on mydriatic fundus photographs obtained in a multiethnic pediatric population (<17 years) from three centers (Atlanta-USA; Bucharest-Romania; Singapore). The DLS's multiclass classification accuracy (ie, normal optic disk, papilledema, disks with other abnormality) was calculated, and the DLS's performance to specifically detect papilledema and normal disks was evaluated in a one-vs-rest strategy using the AUC, sensitivity and specificity, with reference to expert neuro-ophthalmologists.
Results:
External testing was performed on 898 fundus photographs: 447 patients; mean age, 10.33 (231 patients ≤10 years of age; 216, 11-16 years); 558 normal disks, 254 papilledema, 86 other disk abnormalities. Overall multiclass accuracy of the DLS was 89.6% (range, 87.8%-91.6%). The DLS successfully distinguished "normal" from "abnormal" optic disks (AUC 0.99 [0.98-0.99]; sensitivity, 87.3% [84.9%-89.8%]; specificity, 98.5% [97.6%-99.6%]), and "papilledema" from "normal and other" (AUC 0.99 [0.98-1.0]; sensitivity, 98.0% [96.8%-99.4%]; specificity, 94.1% (92.4%-95.9%)].
Conclusions:
Our DLS reliably distinguished papilledema from normal optic disks and other disk abnormalities in children, suggesting it could be utilized as a diagnostic aid for the assessment of optic nerve head appearance in the pediatric age group.
