According to a new British study, artificial intelligence algorithms can predict whether a patient (even in the absence of an unhealthy lifestyle) is at risk of stroke, heart attack or death from heart disease. How? Simply by studying the images of their retina taken from an eye scan.
The thin layers of light-sensitive tissue found at the back of the eyeball can reveal a surprising amount of information on their own.
Eyes on me, please
Early studies on the subject have already found that retinal images can accurately predict a person's age and gender. And not only that: they can determine if a person is a smoker, diabetic or with other medical conditions that can be a potential cause of death.
Today another very important evaluation parameter is added.
It is possible because blood vessels in the retina are affected by blood pressure, which is also an important factor in measuring cardiovascular health risks.
The research
A team of academics from St George's University in London have conducted what they believe is the largest ever study of artificial intelligence applied to eye reading. The results, published in the British Journal of Ophthalmology, I am here.
The researchers used retinal scans from more than 70.000 patients. Data contained in two large databases, which served to train and test QUARTZ (QUAntitative Analysis of Retinal vessels Topology and siZe), an AI-based system that has given excellent results. And not very inclusive, as is still the case: “Sensitivity analyzes limited model development and validation to white ethnicity,” the authors noted.
By studying the images of the eyeballs and the medical history of each participant, QUARTZ learns more and more to associate an eye scan with coronary heart disease, heart attack, myocardial infarction and stroke.
Bottom line: These AI systems can predict, from a photo of the retina, whether someone is at risk of dying from heart disease within the next five years.
QUARTZ, Excellent results
QUARTZ's performance was compared with the Framingham Risk Score (FRS), an algorithm used to predict the chances of developing cardiovascular disease within ten years.
And he has shown that he is more accurate than he is.
An eye scan will save us (or at least forewarn us)
Researchers believe retinal scans read by AI could be used in clinical settings to assess a patient's cardiovascular health without having to take blood pressure measurements or blood tests.
AI-enabled vasculometric risk prediction is fully automated, low-cost, non-invasive. And above all it can reach a higher percentage of the population.
A fragment of the English study.
However, don't expect optometrists to start offering eye scan heart checks any time soon. There's a little something to fix before we roll out risk reduction, and it's not technological.
Diagnosis with eye scan, it is necessary to structure a system
The progress of the diagnoses applied to artificial intelligence are extraordinary, but they have consequences. To make this system widespread and keep cardiovascular risk factors under control, healthcare systems must be adapted.
Do you need to train more ophthalmologists or radiologists to interpret the results? And what happens next? To whom and how will patients be referred to receive therapeutic treatment?
These and other questions need to be answered to become part of the guidelines for the prevention of cardiovascular risk.
But before that, more training is needed for the artificial intelligence, which must be made capable of analyzing the ocular scan of all subjects, not just white ones.
