You are reading Preventing heart attacks by looking at the bottom of the eye: Google’s last frontier

Heart and cardiovascular system

Preventing heart attacks by looking at the bottom of the eye: Google’s last frontier

July 3, 2018

It can “read” a patient through the scanning of the retina, detecting blood pressure, gender and whether or not he is a smoker. The software developed by Verily, a company controlled by Google and at the forefront of research in the biomedical field, is also able to report the presence of risks to the cardiovascular system thanks to a computerized analysis of the bottom of the eye. The research, published in Nature, paves the way for a new way of making diagnoses. We talk about this topic with Maddalena Lettino, a cardiologist from Humanitas.


Verily’s research published in Nature


Recently published in the journal Nature, the new research reaffirms Google’s presence in the healthcare industry, marking the umpteenth point of conjunction between medicine and artificial intelligence.


“Cardiovascular disease is the leading cause of death globally,” Verily’s researchers say, “and a wealth of research helps us understand what can cause it: all everyday behaviors, including exercise and diet combined with genetic factors, age, ethnicity, and biological sex.

The study shows that artificial intelligence learning, applied to an image of the retinal fundus (the innermost part of the eye), can often predict and identify risk factors, including cardiovascular risk.

Related articles

Software still to be perfected


Artificial intelligence training was conducted by analyzing the retinas and medical data of more than 280,000 patients from the United States and the United Kingdom. The software has thus been programmed to autonomously search for elements capable of suggesting the presence of conditions favorable to the occurrence of health hazards. However, the method still needs to be refined: so far it has been able to correctly detect the retina of patients who had cardiovascular episodes in 70% of cases. The result is just under 72% of cases diagnosed through blood tests.


“These results make us appreciate how technological advances and the ability to process millions of data can approach the medical science always considered the prerogative of the human mind alone and how this in the future can contribute to the improvement of health and disease control – said Lettino – What has been published by Nature does not add anything to what can be done in clinical practices around the world at the moment and nothing to the medical knowledge we already have, but creates a kind of platform that the “machines” will enrich perhaps more quickly than the human mind. It is certainly interesting, it requires reflection and the development of a very complex ethical context, but it probably represents an inescapable future that could do good for humanity”.

You may also like

Do not miss our advice for your health

Sign up for the weekly Humanitas Health newsletter and get updates on prevention, nutrition, lifestyle and tips to improve your lifestyle