Retina scans reveal your future health

A new study has found that the structure of the retina in the back of the eye can reveal a host of diseases that a person is at risk from developing. By using machine learning to marry up changes seen in retinal images with health and genetic data from tens of thousands of people who participated in the UK’s Biobank study, Mass Eye and Ear hospital ophthalmologists Maryam Zekavat and Nazlee Zebardast think they have a way of using the retina to see what a person’s future health looks like…

Nazlee - In recent years, we have come to realise that we can find out a lot more information from images of the eye than we ever thought was possible. And it's opened up this really exciting area of research. And we were fortunate enough to have access to an incredible resource, specifically a large, large study of about half a million people from the UK called the UK Biobank. And we had images of people's eyes and we were wondering, is there a link between images of the eyes? And Maryam will talk a little bit more about how those were obtained and ocular and systemic disease.

Chris - Maryam, what are these pictures you've got of people's eyes? Are you looking at the outside or are you talking about what's going on inside the eye?

Maryam - It's inside the eyes. The human retina is a multi-layered tissue, which offers a really unique window into systemic health. So here we're looking specifically at that retina, which is composed of nine different layers using OCT imaging, optical coherence tomography imaging, and that's a non-invasive imaging that uses light waves to take a cross-section picture of your retina. And so we had this OCT imaging available across 50,000 individuals in this biobank

Chris - And Nazlee, why should looking at the different layers of the retina give us an insight into a person's overall health or disease risk?

Nazlee - Some of it is actually really surprising. So for example, we didn't know before the machine learning models came out that we could tell someone's age just by looking inside their eyes, or their sex even. But the retina is a very intricate tissue. It's an extension of your brain essentially. And so it's constructed of all these different cell layers that are connected to our brain and our systemic health, or essentially your overall health. So you can look at people's eyes and when they say the eye is a window into your soul, they're not really kidding because you're seeing what is going on with someone's heart, someone's blood pressure, potentially even how they're breathing or their risk of having certain neurologic problems.

Chris - And Maryam, when you did this study, what did you actually measure and what did you marry up with what?

Maryam - So what we measured was the thickness of each of these nine layers of the retina. And we used this measurement to then look at connections to systemic health to disease and also to genetics.

Chris - Nazlee, one of the key things about any kind of disease, we always say prevention is better than cure. So the critical question I can ask you is if you do this, does it give you a window into the early stages of a disease at a time when you might be able to intervene meaningfully in a person before they actually get sick from something?

Nazlee - It's interesting that you ask that. One of the key pieces of analysis that we did in this manuscript was looking at future risk of disease. So looking at someone's retinal thickness and determining if that was linked to future disease diagnosis. And in fact, the layer thicknesses of the retina are predictive of future diseases of the heart, of the brain, of the kidneys, of the lungs, including eye diseases as well. So even though this needs to be confirmed, our study does point to the fact that we can use retinal thicknesses or retinal phenotypes to be able to predict disease before it happens and intervene.

Chris - So would the approach be then that rather than subjecting a person to a whole raft of different tests and blood samples and genetic analyses, you could plunk them down in front of you, use this technique to look at their retinal structure and then based on what you've learned about the associations between the different shapes and structures of the retina and different disease outcomes, you could make predictions about what a person might be about to develop or indeed is developing right now and then you can make obviously interventions as necessary.

Nazlee - Precisely. I think that is exactly how we envision results like this being used. As Maryam previously mentioned. Also, the OCT or these light pictures of the eye are routinely used in clinical practice in ophthalmology. And so you can imagine if someone comes in for their eye exam, then you can use this information to tell them you seem to have a high risk of MS or you have a high risk of high blood pressure, or you might have a high risk of developing diabetes, so you should go see your primary care doctor. The other way you can think about this is exactly as you said is what if we just had this as a screening tool where people would come in, get images of their retina noninvasively without having to do all these extra tests and they could know that they have risk for certain diseases that they need to get checked for or even treated for potentially.