In response to the COVID-19 situation, OCLI is following enhanced safety protocols consistent with guidelines put forth by the CDC and the World Health Organization (WHO) to protect our patients and colleagues. Click here to learn about how we are keeping you safe.

Detecting Diabetic Retinopathy with Artificial Intelligence

December 7, 2017

World Diabetes Day is November 14 but the entire month of November is dedicated to increasing diabetes awareness. One-third of the 17 million people in America that suffer from diabetes aren’t even aware they have the disease. A dilated comprehensive eye exam can detect and diagnose diabetes by recognizing early stages of diabetic retinopathy. Diabetic retinopathy is a condition affecting the eyes that develops as a complication from diabetes. Anyone with Type I or Type II diabetes can develop this complication and if left untreated, it can lead to blindness.

Diabetes is often linked to vision loss because even if diabetes is caught and treated early on, it can still cause cataracts to form, blocking light and decreasing vision. Diabetes is the leading cause of vision loss in 20-74 year olds. Early signs of diabetic retinopathy that can be detected by your ophthalmologist include leaking blood vessels, retinal swelling known as macular edema, and deposits on the retina.

Until recently, diabetic retinopathy could only be detected by trained doctors and specialists who scan and interpret pictures of the retina. Unfortunately there is currently a shortage of qualified ophthalmologists trained to diagnose diabetic retinopathy but there may soon be a new test available for use in diagnosing diabetes. Doctors in New Zealand were able to detect diabetic retinopathy using deep neural networks, a type of artificial intelligence (AI). The deep neural network was able to detect referable diabetic retinopathy from photographs. 1,685 diabetic retinal photographs were chosen from two different databases and assessed by the deep neural network software to determine the diabetic retinopathy grade of each photo.

The artificial intelligence software was found to be accurate, achieving specificity in the range of 79.7-90% and sensitivity ranging from 84.6-96%. Researchers are hoping the deep neural networks can be integrated into community screening once they can successfully detect both diabetic retinopathy and diabetic macular edema.

Google Brain (Google’s research team) conducted a similar study using 12,000 images. The images were evaluated by Google Brain’s neural network algorithm as well as a panel of ophthalmologists. Google Brain’s algorithm was considered to be on par with the diagnoses of the experienced ophthalmologists. As AI becomes more accurate, its impact on the healthcare industry becomes more significant. This research has the potential to assist eye hospitals in evaluating more patients in less time and prioritizing them based on their diagnosis.

Microsoft is also working on artificial intelligence programs to assist doctors with eye care and have partnered with the LV Prasad Eye Institute to launch MINE – Microsoft Intelligent Network for Eyecare. Artificial intelligence is significantly impacting the healthcare industry by filling a need for AI-assisted screening and diagnosis.

Whether or not you have been diagnosed with diabetic retinopathy, it’s important to have your eye health checked regularly. Let National Diabetes Month be a reminder for you to take care of yourself. If you’re concerned you may have diabetes or feel like your eyesight isn’t what it used to be, contact OCLI today to schedule your next eye exam. 

Back to our blog

Services offered at OCLI

Our world-class team of professionals at OCLI can help you with the latest treatment options for you.

Request an appointment

* required