Predicting the risk of stroke using retinal imaging
- Dr Mervyn Thomas
- Jatinder Minhas
- Frank Proudlock
- Irene Gottlob
- Tom Robinson
- Professor Zhang
Studying the neurosensory retina and vasculature offers a unique opportunity to non-invasively visualise and quantify vascular and central nervous system health. Retinal microvasculature changes have been identified as independent predictors for hypertension, diabetes, heart disease and stroke. Moreover, retinal microvascular changes can precede clinical manifestation of end organ damage.
In this project, the ACF will utilise UKBIOBANK and local annotated retinal imaging datasets for the purposes of developing a machine learning model to predict the risk of developing a stroke. This model will be compared against previously published and validated stroke prediction scores. Developing novel risk prediction systems and surrogate biomarkers from retinal vascular morphometric parameters has significant predictive value and provides a window for aggressive risk reduction. The project will utilise the research computing facility, including the £2M High-Performance Computing cluster for “big data” analysis and will be linked with the local HDRUK network.
The ACF will be hosted by the Ulverscroft Eye Unit (UEU), Cerebral Haemodynamics in Ageing and Stoke Medicine (CHiASM) group and Computing at Leicester. The supervisory team brings together experts in the field: Dr Thomas (Ophthalmology), Dr Minhas (Stroke Medicine), Dr Proudlock (Ophthalmology), Professor Gottlob (Ophthalmology, FARVO), Professor Zhang (Machine Learning, FIET) and Professor Robinson (Stroke Medicine, NIHR Senior Investigator). The supervisory team have an established record with publications in high-impact journals in the fields of ophthalmology, stroke medicine and machine learning. UEU and CHiASM have successfully hosted NIHR trainees with excellent career progression attaining external fellowship funding and subsequently lectureship and senior lectureship posts.