PREDICT-DR (v1.0)
Research type
Research Study
Full title
Deep learning for the automated prediction of diabetic retinopathy progression
IRAS ID
287236
Contact name
Timothy Jackson
Contact email
Sponsor organisation
King's College London
Duration of Study in the UK
2 years, 1 months, 0 days
Research summary
Summary of Research
Diabetic retinopathy (DR) is a disease where the light-sensing layer at the back of the eye (retina) becomes damaged by raised blood sugar levels. It affects about a third of the 422 million diabetics worldwide and is the commonest cause of vision loss in working-age adults.
Diabetic people in the UK have yearly retinal photographs taken to screen for DR. The South East London Diabetic Eye Screening Programme (SEL-DESP) screens over 106,000 patients, all of whom have retinal images and DR grading according to UK National Screening Committee (NSC) criteria at each visit for both eyes. Over 20,000 patients from this same cohort also attend diabetic clinics and have a number of clinical measurements taken during routine appointments.
Deep learning (DL), a subfield of Artificial Intelligence (AI), uses computer models that learn from large datasets to make new predictions about similar unseen data. We aim to develop DL models that predict the likelihood of DR progression over time, and the occurrence of DR that requires hospital referral. The models will also be trained to highlight high-risk features within retinal images, numerical screening data and clinical data which are associated with DR progression. Once validated, the models would allow for earlier and more accurate detection of DR progression within all UK DR screening programmes, enabling individualised, risk-based monitoring intervals and timely lifestyle or medical interventions to avoid vision loss.
Summary of Results
Diabetic retinopathy (DR) is a disease where the light-sensing layer at the back of the eye (retina) becomes damaged by raised blood sugar levels. It affects around one in three of the 463 million people with diabetes worldwide and is a leading cause of acquired vision loss in working-age adults. In this study, we developed deep learning (DL) models, a type of artificial intelligence, to predict when DR would reach a sight-threatening stage up to 3-years in the future using retinal images and participant characteristics from the southeast London DR screening service. Once prospectively evaluated, developed DL models could enable accurate, individualised risk predictions for people with diabetes who regularly attend DR screening services. This could mean fewer unnecessary visits for individuals at low-risk of DR progression, but closer monitoring and potentially earlier treatment for individuals at high-risk of DR progression, which could subsequently reduce the risk of avoidable vision loss.
REC name
East Midlands - Leicester South Research Ethics Committee
REC reference
20/EM/0250
Date of REC Opinion
7 Oct 2020
REC opinion
Favourable Opinion