RetinaUWF - AI Detection of Diabetic Retinopathy in Ultra-Wide-Field Retinal Images

Lead Participant: RetinaScan Ltd

Abstract

"Diabetic Retinopathy (DR) is a common complication of diabetes mellitus, which affects around half of 430m diabetics worldwide (WHO). It is a major cause of blindness (\>7% UK blindness) but can be easily ameliorated if detected and treated early. Hence the importance of annual screening when images of the retina are taken and reviewed by qualified graders for symptomatic features. Diabetic Eye Screening Programmes (DESP) are run in the UK and other countries, however they are labour intensive, slow and expensive (annual NHS screening cost of \>£100m).

Automated Retinal Image Analysis Systems (ARIASs) are emerging as a powerful tool for improving DR screening efficiencies and clinical outcomes. However, current systems only work with conventional narrow-field fundus cameras, limiting imaging to approximately 20% of the retina surface and leaving 80% out of view. This restricts their ability to detect some of the most harmful eye disease, such as diabetic retinopathy and ocular tumours, which occur frequently in the far periphery of the eye. Ophthalmic best practice is driving adoption of a new generation of Ultra-Wide-Field cameras, capable of imaging 85% of the retina. However, while UWF images are larger and more costly to grade manually (up to 5x), no ARIAS exists capable of reliably analysing UWF images for disease. This creates a significant, unmet and growing need which this project responds to.

In this project, RetinaScan Ltd (RSL) partners with the leading NHS-DESP at Gloucester and Surrey University (CVSSP) to meet this challenge by developing the world's first AI competent to analyse UWF retinal images for DR. We will further innovate our successful A-CNN deep-learning architecture to achieve high detection performance and robustness to UWF distortions such as blurring and occlusions.

The key outcomes of the R&D will be: (i) a fully capable prototype suite of trained UWF-Augmented-CNN technology; (ii) prototype validation via NHS user trials; and a detailed business plan for commercialisation as a cloud-based service for markets globally.

Significant economic benefits will be realised by healthcare providers globally who deliver retinopathy screening services by UWF imaging with ARIAS."

Lead Participant

Project Cost

Grant Offer

RetinaScan Ltd, Guildford £538,443 £ 376,909
 

Participant

University of Surrey, United Kingdom £218,626 £ 218,626
Gloucestershire Hospitals NHS Foundation Trust £241,625 £ 241,625

Publications

10 25 50