Quantitative assessment of colour fundus photography in hyperopia children based on artificial intelligence

Area of Science:

• Ophthalmology
• Medical Imaging
• Artificial Intelligence in Healthcare

Background:

• Hyperopia is a common refractive error in children.
• Understanding optic nerve head and retinal vascular changes in pediatric hyperopia is crucial for early detection and management.
• Current methods for quantifying these parameters can be subjective or time-consuming.

Purpose of the Study:

• To quantitatively evaluate optic nerve head and retinal vascular parameters in children with hyperopia.
• To assess the relationship between these parameters, age, and spherical equivalent refraction (SER).
• To explore the utility of artificial intelligence (AI)-based analysis of color fundus photographs (CFP) for this evaluation.

Main Methods:

• A cross-sectional study of 324 children with hyperopia (aged 3-12 years) was conducted.
• Participants were categorized into low hyperopia (SER +0.5 D to +2.0 D) and moderate-to-high hyperopia (SER ≥ +2.0 D) groups.
• AI was used to automatically detect and quantify fundus parameters, including optic disc area and mean vessel diameter, followed by regression analysis.

Main Results:

• Children with moderate-to-high hyperopia exhibited larger superior neuroretinal rim width and greater vessel diameter compared to those with low hyperopia.
• Axial length was significantly associated with smaller superior and temporal neuroretinal rim widths and smaller vessel diameter.
• A mild inverse correlation was noted between optic disc area/vertical disc diameter and age.

Conclusions:

• AI-based CFP analysis demonstrated that children with high hyperopia have larger mean vessel diameters and smaller vertical cup-to-disc ratios.
• This study highlights AI's capability to provide objective, quantitative data on fundus parameters in pediatric hyperopia.
• AI offers a promising tool for enhanced assessment and monitoring of ocular health in children with hyperopia.