Representational / Getty picture.
The analysis, published in the journal Nature Biomedical Engineering, revealed that deep learning applied to a retinal fundus picture, a photograph that includes the blood vessels of the eye, can forecast risk factors for heart ailments — by blood pressure to smoking status.
The algorithm which the researchers generated can even help predict the occurrence of a future major cardiovascular event on par with current steps, said Michael McConnell, Head of Cardiovascular Health Innovations in Verily in website post.
Cardiovascular disease is the main cause of death globally and researchers know that lifestyle factors such as diet and exercise in combination with genetic factors, age, ethnicity, and sex all contribute to it.
However, they do not exactly understand how these factors add up in a specific person, and so in some patients it will become necessary to perform complex tests, like coronary calcium CT tests, to help better stratify an individual’s risk for getting a heart attack or a stroke, also such other cardiovascular events.
Retina picture. Google
In this study, using deep learning algorithms trained on data in 284,335 patients, the investigators could predict cardiovascular risk factors from retinal images with amazingly large accuracy for patients from two separate datasets of 12,026 and 999 patients.
The algorithm could differentiate the retinal images of a smoker out of that of a non-smoker 71 percent of the time, the study found.
“In addition, while doctors can normally distinguish between the retinal images of patients with acute hypertension and regular patients, our algorithm could go further to forecast the systolic blood pressure in 11 mmHg normally for patients overall, such as those with and without hypertension,” research co-author Lily Peng, Product Manager, Google Brain Team, stated.
“Among the fascinating aspects of this research is the creation of ‘attention maps’ to reveal which aspects of the retina contributed most to the algorithm, thus providing a window into the ‘black box’ often related to machine learning,” McConnell, who’s also a co-author of this study, said.
This may give clinicians greater confidence in the algorithm, and potentially provide new insights to sociological attributes not previously associated with cardiovascular risk factors or future risk, McConnell said.
The findings indicate that a simple retinal image may one day help comprehend the wellbeing of a patient’s blood vessels, crucial to cardiovascular health.
“This is promising, but early research — more work has to be done in order to develop and affirm these findings on larger patient cohorts earlier this may arrive in a clinical setting,” McConnell added.