Teams of researchers have shown this week that we could detect Alzheimer’s disease and the risk of heart attack by photographing the retina. The details of Géraldine Zamansky, journalist at the Magazine de la Santé on France 5, and every Saturday on franceinfo.
franceinfo: What is Géraldine concretely about?
Geraldine Zamansky: The first study comes from the Faculty of Medicine of Hong-Kong and its principal author, Carol Cheung, explained to me how they had created an artificial intelligence, capable of detecting Alzheimer’s disease from photographs of the retina. It was already known that this pathology was associated with deformations of this membrane which lines the back of the eye. So they provided thousands of photos of retinas of people with Alzheimer’s to their algorithm. Result: he manages to make the diagnosis in more than 80% of cases.
Would that be much simpler than current diagnostic methods?
Yes it’s sure ! Today, it is generally necessary to combine memory tests, difficult samples and sometimes even a scanner. Whereas there, if you went to an ophthalmologist recently, you already know the type of machine needed, a retinograph. In a few minutes, your retina is photographed. It would then suffice to show the pictures to the artificial intelligence of Carol Cheung, to have a diagnosis on the health of your retina of course, and also that of your brain. But the researcher told me that they need to confirm their results with people in the early stages of the disease.
And from the same photos, another team is able to identify a risk of heart attack or stroke?
Yes, it’s the team of Alicja Rudnicka, from St George’s University in London. Their starting point is that the very small vessels of the retina reflect the state of the other vessels of the body. She summed it up simply: if they are too narrow and winding, for example, that’s a bad sign. So they too created an artificial intelligence, from photos of the retinas of more than 88,000 volunteers, followed for years, in the incredible British Biobank. They added known risk factors such as smoking or hypertension.
And it works. Their system identifies those most at risk of a heart attack or stroke, the leading causes of death linked to poor artery health. As it is a public search, their algorithm is free. So Alicja Rudnicka hopes that it won’t take too long to create a secure digital link between the retinograph screening network and the general practitioners who will take care of patients classified as “red alert”.
>>> Studies
– The research of the team of Alicja Rudnicka, from St George’s University in London, dyears the British Journal of Ophthalmology.
– The article published in The Lancet concerning the work of the Chinese Carol Cheung