Professor Pearse Keane on using eye health to detect dementia

Professor Pearse Keane explains how using artificial intelligence (AI) to investigate the eye's retinal tissue can help diagnose dementia.

What is your primary field of research?

I’m a consultant ophthalmologist at Moorfields Eye Hospital and a Professor of Artificial Medical Intelligence at UCL Institute of Ophthalmology. I am also Director of the INSIGHT Eye & Oculomics Health Data Research Hub at Moorfields, which is the world’s largest bioresource of ophthalmic imaging data.

Although I’m not a computer scientist or engineer, I lead a multidisciplinary research group which aims to develop and apply artificial intelligence in healthcare, both in the UK and globally, using ophthalmology as an exemplar. We cover the full spectrum of research in clinical AI, going “from idea to algorithm” and “from code to clinic”.

What can detailed images of the inside of the eye tell us about dementia at large?

For more than 100 years, doctors have known that the eye can be used as a window to the rest of the body. This makes sense because the retinal blood vessels are the only microvasculature that can be directly visualised in living human subjects.

And similarly, the retinal tissue is the only part of the central nervous system that can be directly seen without removing part of the skull. It’s an approachable part of the brain! From a number of studies, we have evidence showing that signs of degeneration in the retinal tissue may be seen in patients with dementia.

How have developments in artificial intelligence (AI) and machine learning affected your field of research and how did you become involved in the use of these tools?

Our ability to use the eye as a window to systemic health has been supercharged in recent years by the combination of: 1) big data, 2) the latest advances in artificial intelligence, and 3) advances in retinal imaging such as optical coherence tomography.

To describe this emerging field of study, our research group coined the term “oculomics”. AI can allow us to find new information hidden within retinal images that is not currently visible to human experts.

Have there been any big advancements in AI and ‘oculomics’ in the past year?

Over the past year, the research team I lead across the UCL Institute of Ophthalmology and Moorfields has made some exciting advances in Oculomics. One of the most important has been the launch of the Global RETFound initiative, a research consortium that is building the first globally representative medical AI foundation model.

The initiative is one of the largest medical AI collaborations ever undertaken, involving more than 100 study groups across 65 countries spanning all continents (except Antarctica). Together we are producing one of the most geographically and ethnically diverse ophthalmic imaging datasets, comprising 100 million eye images.

Global RETFound builds on the success of RETFound, the first foundation model for retinal and systemic disease detection. Published in Nature in 2023, RETFound was developed by researchers at Moorfields Eye Hospital and UCL Institute of Ophthalmology, using 1.6 million retinal images curated by the INSIGHT Health Data Research Hub at Moorfields.

Like RETFound, Global RETFound could help improve diagnosis of some of the most debilitating eye diseases, including diabetic retinopathy and glaucoma, and predict systemic diseases such as Parkinson’s, stroke and heart failure.

You co-lead a research study with Dr Siegfried Wagner called AlzEye, which aims to use eye scans to detect neurodegenerative diseases, what progress has been made in the field of dementia detection?

The AlzEye study is a research study involving UCL and Moorfields Eye Hospital. In it, we have linked ophthalmic imaging data from Moorfields with a national NHS database called Hospital Episode Statistics (HES).

This means that we can identify patients that have had retinal scans done at Moorfields and who have subsequently developed a range of cardiovascular and neurodegenerative diseases.

The AlzEye project is one of the breakthroughs that has helped to forge the field of oculomics, linking more than 24 million retinal images from over 350,000 people attending Moorfields Eye Hospital with corresponding NHS Hospital Episode Statistics (HES) covering myocardial infarction, all-cause stroke and all-cause dementia.

We used the AlzEye dataset to produce a landmark study, using AI to detect Parkinson’s Disease in retinal biomarkers up to seven years before diagnosis. Led by Siegfried Wagner, the study was published in Neurology, and covered by over 400 international news sources, including the BBC.

In 2025, Siegfried coordinated an update to AlzEye. In agreement with NHS England, the HES data was updated, expanding the original linkage from 2018 to 2024, including cause of death certificates. AlzEye has ethical approval for neurodegeneration and dementia studies, providing exceptional opportunities for research into systemic diseases using images of the eye.

We are currently exploring opportunities to link with routinely collected brain imaging at our partner hospitals. This could help us unlock even more information about systemic changes that are detectable from eye images.

From 2026 we are embarking on a genomics linkage project with the INSIGHT Health Data Research Hub at Moorfields in partnership with UCL, funded by the Medical Research Council and NIHR. This will allow us to link with large-scale retinal imaging data and genetic data from UK Biobank and NIHR BioResource to explore shared genetic architecture between retinal traits and conditions such as dementia.

What is the potential of this work for public health?

If we can find strong signs of systemic disease from eye examinations, it could have huge public health potential.

For example, we know that only a small percentage of people in their 40s and 50s attend their GPs for the NHS Health Check. By contrast, however, the majority of these people will attend their local optometrists for regular eye tests. If we can use such tests to pick up the earliest signs of systemic diseases, it would be transformative!

Image: Optical Coherence Tomography (OCT) scan of the retina at the back of the eye, with the layers of the retina highlighted in colour.

Read more about Oculomics on the INSIGHT website.