AI system detects diabetic eye damage

Blessing Physician Services’ medical office assistant Kristy Shepherd demonstrates how to use an IDx-DR on Tuesday, Oct. 8, at Blessing Health Center. The device uses artificial intelligence to make a clinical assessment to interpret the eye images in an effort to detect diabetic blindness. | H-W Photo/Jake Shane
By Herald-Whig
Posted: Oct. 9, 2019 12:01 am

QUINCY -- Ten minutes tacked on to the end of his appointment with his doctor were all it took to check Alan Blickhan's eyes and show that there were no problems.

Blickhan, who has prediabetes, was checked with a new system called IDx-DR, which uses artificial intelligence to detect diabetic retinopathy, a condition caused by poorly controlled blood sugar that could lead to blindness.

"It doesn't take but a few minutes for them to read out and get a picture of the eye and see if there are any issues, but I had a clean bill on it," Blickhan said.

Blessing Physician Services provided a demonstration of the IDx-DR system Tuesday afternoon.

Blickhan happily underwent the test during his recent appointment.

"Eye health is really important, as my mother was legally blind from the time I was 17," he said. "Your eyes are the most important feature we have."

Blickhan's physician, Dr. Tim Beth, said not all patients visit an eye doctor each year, where diabetic retinopathy could be detected.

"This is a way for us to capture people that aren't being tested," Beth said.

The procedure, including results, takes about 10 minutes, with the patient sitting in the dark for two minutes to allow their eyes to dilate naturally. Using a robotic camera, a medical assistant takes two images of each eye, and the images are analyzed by an IDx algorithm to look for evidence of diabetic retinopathy.

If the system discovers more than mild diabetic retinopathy, patients are referred to an ophthalmologist.

The IDx system received FDA approval last year, and the clinic is one of the earliest adopters of the technology.

Beth said if diabetic retinopathy is caught early, it can be treated.

"If you wait to where the damage is done, it's going to be hard to reverse that, which is why we want to do it on a yearly basis," he said.

Dr. Irshad Siddiqui, chief health information officer for Blessing Health System, said diabetics need to have their eyes checked each year to check for symptoms of diabetic retinopathy, and only 25% of Blessing patients with diabetes were following through.

"We were missing three-fourths of our patients that had diabetes but had not been to an eye doctor," Siddiqui said. "So we thought this was a good technology to put in the primary care office to ensure that as soon as a diabetic comes from their routine diabetes care visit, they can get their eyes checked as well."

In just a couple months, 83 patients have undergone the scan, with 12 receiving positive results.

Siddiqui acknowledged the buzz around artificial intelligence in health care, and believes medical professionals need to make sure it is used in a pragmatic way.

"As the technology is growing, we have found that eye exams, skin exams, radiology, pathology, anything that has pictures in it, that's where artificial intelligence is becoming more and more useful," he said.