A predictive atlas of disease onset from retinal fundus photographs: a modelling study using data from population-based cohorts - PubMed
7 hours ago
- #population health
- #disease prediction
- #retinal imaging
- Early detection of high-risk individuals is crucial for healthcare systems.
- Retinal fundus photographs are non-invasive, low-cost, and scalable.
- Study assessed retinal images' predictive potential for 752 diseases in 61,256 UK Biobank participants.
- Improved predictive performance for 306 diseases, 280 outside ophthalmology.
- No improvement in cardiovascular disease prediction compared to existing scores.
- Predictive improvements linked to vascular patterns, eye color, and lens morphology.
- Genetic findings revealed associations between retinal risk phenotypes and complex disorders.
- Retinal images show potential for broader disease screening but require rigorous benchmarking.
- Study highlights limitations and opportunities for cost-efficient screening strategies.