- AI programme can spot warning signs of heart problems.
- It measures the width of small arteries and veins behind eyes.
- It’s being called QUantitative Analysis of Retinal vessels Topology and SiZe (QUARTZ).
A recent study has suggested that scanning a person’s eyes might actually hint toward their risk of early death. Scientists have created an AI programme that has the capacity to spot warning signs of heart problems.
It measures the width of the small arteries and veins behind the eyes which supposedly have crucial information about the health of the heart.
If there are any issues in blood circulation, blood vessels in the retina swell which can even lead to loss of vision.
A study was conducted on 88,000 people in Britain. The AI was able to predict 54% of cardiac arrests and strokes and 58% of deaths related to heart diseases.
The test, doctors said was non-invasive and could be utilised during any eye checkups.
In America, heart diseases are the leading cause of death. They take a life every 34 seconds which is nearly 700,000 yearly.
While there are no guidelines on how exactly one can screen for these illnesses, the Centers for Disease Control and Prevention says that all individuals that are over the age of 20 should get screened for high cholesterol every five years.
Regular checkups for blood pressure and blood glucose are also recommended.
The method that is now being explored is called retinal vasculature imaging. It takes pictures of the back of the eyes so that doctors can see the blood vessels there. The technique has been helping doctors diagnose glaucoma and other ophthalmologic problems.
Researchers developed software which can then read thousands of scans and report the thickness of the vessels across the retina. It is called Quantitative Analysis of Retinal vessels Topology and Size (QUARTZ) and was revealed in a study in the British Journal of Ophthalmology.
Once scientists have the results, they can compare them to other important information related to the patient like their age, smoking status and general medical history.
Their algorithm accurately reported 58% of early deaths. It predicted 454 strokes and heart attacks and 839 were actually recorded.
The system was tested in Europe as well where it worked as well as in Britain.
