Retinal vessel morphology as a predictor of cardiovascular risk and disease outcome in later life
Approved Research ID: 522
Approval date: July 1st 2013
Lay summaryThis is an application for use of the UK Biobank data by the population Health research Centre at St George?s University of London in collaboration with Kingston University. The focus of this proposal is to quantify the shape and size of retinal vessels from the digital photographs of the retina. Recent studies have also shown that retinal vessel width and tortuosity are associated with cardiovascular risk factors from an early age, suggesting that these measures may be early physio-markers of vascular health. We propose to evaluate the distribution and determinants of these indicators of vascular health and their relationship to the future onset of both fatal and non-fatal disease. A central purpose of Biobank is to identify predictors of future disease, with a view to prevention, early detection, and selective clinical management of high risk groups. Our proposal falls within this remit. Drs Rudnicka, Owen and Barman (members of UK Biobank Eyes and Vision Consortium) and have worked together on retinal images analysis of a large dataset of child retinal images. Owen and Rudnicka are investigators on an HTA funded project examining performance and cost effectiveness for the NHS of automated diabetic retinal imaging software. The focus of this proposal is to: i. Fully automate identification and quantification of vessels in retinal images, including identification of arterioles and venules ii. Examine cross-sectional determinants of retinal vessel size and shape iii. Investigate the association between retinal vessel size and shape with cardiovascular disease precursors, and future onset of both fatal and non-fatal disease. This research can be used to understand factors that influence the size and shape of retinal vessels and their role in prediction of cardiovascular risk / events. This project requires access to data only (not the biological samples) but we wish to access data on biochemical markers and health outcomes, as and when they become available.
- Automated arteriole and venule classification using deep learning for retinal images from the UK Biobank cohort
- Automated retinal image quality assessment on the UK Biobank dataset for epidemiological studies
- Retinal Vascular Tortuosity and Diameter Associations with Adiposity and Components of Body Composition