I-SCREEN Methodology Paper Published in Eye (Nature)

The I-SCREEN consortium has published its methodology paper in Eye, a peer-reviewed journal from the Nature portfolio. The paper describes the full design and clinical framework of the project, marking a significant milestone in the consortium's four-year mission to transform how age-related macular degeneration (AMD) is detected across Europe.

Why early detection matters

AMD is one of the leading causes of irreversible vision loss, affecting roughly 20% of older adults and more than 15 million people across Europe. Because early AMD develops without affecting central vision, most patients remain completely unaware of the disease until significant damage has already occurred. This late presentation severely limits treatment options.

What I-SCREEN does differently

Rather than waiting for patients to reach specialist clinics, I-SCREEN brings screening directly into community optometry practices. Local opticians acquire high-resolution OCT scans, which are then reviewed remotely by board-certified ophthalmologists via telemedicine. AI-driven analysis supports detection and risk prediction throughout the process.

The shared-care network now spans 6 European countries, 28 community optometry practices, and 7 ophthalmology clinics, operating across three interconnected clinical studies: PYRENEES, SUDETES, and APENNINES.

Early results

The PYRENEES screening study has already imaged over 2,300 individuals across community practices. Approximately 20% were found to have signs of early AMD or non-vision-affecting geographic atrophy that had previously gone undetected. These individuals have been referred for clinical follow-up and two-year monitoring.

What comes next

Data collected across all three studies will inform the development of AI-based predictive models capable of estimating an individual's risk of progression to advanced AMD. The goal is a fully automated, scalable screening tool deployable in community eye care settings across Europe.

The full paper is available here.