We have all heard that the eyes are the window to the soul, but they may also be the window to the heart.
Specifically, the blood vessels in the retina, which can act as indicators to heart disease, say researchers at the University of Leeds in England.
The study was published this week in the journal Nature Machine Intelligence.
Scientists developed an artificial intelligence (AI) system that analyses scans of eyes taken during standard optometrist appointments. From the scans, the AI can tell whether a person has any heart problems.
“Cardiovascular diseases, including heart attacks, are the leading cause of early death worldwide and the second-largest killer in the U.K.,” Alex Frangi, a chair in computational medicine at the University of Leeds, said in a statement. “This causes chronic ill-health and misery worldwide.”
Frangi adds that the technique opens-up the possibility of revolutionizing the screening of cardiac disease given that retinal scans are comparatively cheap and routinely used.
“As a result of automated screening, patients who are at high risk of becoming ill could be referred for specialist cardiac services,” he said.
Predicting a heart attack with 80 per cent accuracy
AI was between 70 and 80 per cent accurate at detecting a heart attack within the next year, the researchers say.
To maximize the amount of information available, the U.K. Biobank, a biomedical database with in-depth genetic and health information from more than 500,000 people across the U.K., provided data for the study. From there, more than 5,000 retinal and cardiac scans were analyzed by AI.
The machine was then able to detect patterns between the eyes and hearts, and was even able to estimate the efficiency of the left ventricle — one of four chambers in the heart which helps pump blood. An enlarged left ventricle was recognized as a clue to increased risk of heart disease.
“The AI system is an excellent tool for unravelling the complex patterns that exist in nature, and that is what we have found — the intricate pattern of changes in the retina linked to changes in the heart,” Sven Plein, professor of cardiovascular imaging at the University of Leeds said.
This is not the first time AI was used to detect medical conditions.
Machine learning can identify 74 diseases
A 2020 study published in the Orphanet Journal of Rare Diseases says machine learning is able to identify as many as 74 different diseases. Researchers hope that future work will be facilitated by the machine learning they used to help further identify patients.
According to the federal government, the most common form of cardiovascular disease is ischemic heart disease — a condition in which the coronary arteries, which supply blood to the heart itself, become blocked. This results in a lack of oxygenated blood in the heart and can cause chest pain or shortness of breath.
Every year, about 158,7000 people aged 20 or older in Canada are diagnosed with heart disease, the federal government says. Of those, roughly 63,200 adults have a heart attack for the first time every year.
The Heart and Stroke Foundation says 90 per cent of Canadians have at least one risk factor for a stroke or vascular cognitive impairment. Additionally, one person in Canada dies every five minutes from a heart condition or stroke. The foundation says that there are 13 per cent more deaths from heart conditions or stroke than all forms of cancer combined.