Blog Hero

Google algorithms allow detection for heart disease and diabetes through retinal screening

Book Appointment

A study published Feb 19th, in Nature and Biomedical Engineering, researches reported that they used retinal images from 284,335 people to determine signs of cardiovascular disease. Using Google algorithms, they then created models to predict 70 percent of the time which patients were likely to have a heart attack in 5 years.  “The clinical utility of such features still requires further study”, researchers said.

Diabetic retinopathy is the leading cause of blindness in the United States caused by diabetes.  The blood sugar causes blood vessels in the back of the eye (retina) to breakdown. This results in leaky blood vessels that can cause vision loss overtime.  By teaching software to distinguish stages of diabetic retinopathy, artificial intelligence had sensitivities of 87 percent to 95 percent.

Currently, retinal imaging is offered at our office during a comprehensive eye exam.  Cardiovascular diseases such as diabetes, high blood pressure, and high cholesterol are often picked up during these retinal exams.   With the help of aritifical intelligence eye doctors may be able to catch ife threatening conditions earlier than with current screening methods. This is an exciting time in eye care and we look forward to integrating artificial intelligence algorithms to our retinal imaging as they become available.   This technology can provide better communication and coordination of care between primary care physicians, eye doctors, and patients leading to better health care.


Written by Dr. Michael Duong

Dr. Duong received his bachelor’s degree in biology with a minor in chemistry at San Jose State University. He then attended Pennsylvania College of Optometry and received his Doctor of Optometry in 2009.
instagram facebook facebook2 pinterest twitter google-plus google linkedin2 yelp youtube phone location calendar share2 link star-full star star-half chevron-right chevron-left chevron-down chevron-up envelope fax