Testing optic nerve head pallor for associations with dementia risk factors
Approved Research ID: 95450
Approval date: April 17th 2023
Background: Dementia is a growing challenge. Mass screening would help a) target healthcare to those who are likely to benefit most, and b) identify suitable candidates for clinical trials. However, MRI brain-scans are expensive, and cognitive tests lack sensitivity. Retinal fundus imaging is a cost-effective, non-invasive, and widely available technology that may offer a solution, owing to the homology between the retina and the brain.
Rationale and aims: The optic disc is a prominent structure in the retina and is the point at which all the retinal ganglion cells, which are responsible for vision, converge to form the optic nerve, which enters the brain. A pale optic disc indicates irreversible damage to the ganglion cells, and is a marker for RNFL thickness (a 3D scan of the retina). It is usually identified through examination with a slit-lamp or via visual assessment of a photograph. However, this requires expertise and time, and does not lend well to research. In previous work, we developed software to automatically quantify the paleness of the optic disc and found strong associations between pallor and RNFL thickness, which itself is related to dementia. In the UK Biobank, we aim to explore the relationship between optic disc pallor and dementia-related variables, including diagnosis, time to diagnosis, risk factors, inflammation, and cognition.
Public health impact: Determining the association between optic disc pallor and dementia-related variables in a large sample would represent an important step in making the tool available to healthcare practitioners. In Scotland, for example, the NHS fund full eye health examinations every two years. Embedding this technology within high street opticians could aid practitioners in making referrals and allow for the monitoring of individuals over several years. This can also be done retrospectively.
Project duration: This project will form part of the applicants PhD research and will be complete no later than January 2025. However, we anticipate publications and a release of derived data much sooner.