filo/DigitalVision Vectors via G

How AI Can Help Diagnose and Streamline Care for Diabetic Retinopathy

AI technology may be able to facilitate early diagnosis and streamlined care for diabetic retinopathy in specialized, primary care, and mobile settings.

According to the National Eye Institute, a subset of the NIH, “diabetic retinopathy is an eye condition that can cause vision loss and blindness in people who have diabetes. It affects blood vessels in the retina (the light-sensitive layer of tissue in the back of your eye).” Earlier diagnosis can improve patient outcomes and reduce the risk of total blindness. To facilitate early diagnosis and streamline care in specialized, primary care, and mobile settings, providers can utilize AI in the diagnosis and treatment process.

The current recommendation of the NIH for patients at risk for diabetic retinopathy is to have an annual comprehensive dilated eye exam. Unfortunately, many patients forgo these yearly exams. Those who go for an annual check are often met with long wait times, expensive visits, and misdiagnosed or undiagnosed conditions. Eyenuk developed an FDA-cleared AI technology, EyeArt, to help reduce some barriers associated with a diagnosis.

LifeScienceIntelligence sat down with Kaushal Solanki, PhD, founder, and CEO of Eyenuk, to discuss the technology and its implications. Solanki’s personal experience with ophthalmic conditions inspired him to look for additional solutions to these problems.

The CDC states that one-third of Americans with diabetes also have diabetic retinopathy. “Anybody with diabetes is vulnerable to blindness that progresses without any symptoms. This is over half a billion people worldwide — over 35,000,000 in the United States alone — who have diabetes and are vulnerable to vision loss,” stated Solanki.

An Overview of the Technology

“To solve this problem, my team and I developed EyeArt first for the autonomous detection of diabetic retinopathy. The AI is just a software program that can process images of the retina. Trained experts and eye care professionals must grade these same photos. When EyeArt comes in, it can just process these images within seconds. This study results show that the performance, sensitivity, and accuracy of EyeArt in detecting people who have the disease are much higher than that of a traditional exam by an ophthalmologist, “explained Solanki.

Solanki shares that the AI technology is meant to work in multiple ways. There are three main ways that this AI functions to facilitate improved patient care and provider workflow: spellcheck AI, autonomous AI, and superpower AI. “Eyenuk's AI offerings fall within these three frameworks,” asserted Solanki.

Spell Check AI

“One way could be just AI working in the background, like a spell check. As physicians do their regular work, AI can work in the background to point out errors that need adjudication. That’s what I call spell check AI, and that's how it can help physicians make fewer errors and do their work more accurately,” he stated.

Solanki shared that eye screening programs have fundus cameras that image eyes, and human experts often make conclusions based on collected scans.

“The spell check AI can check things in the background, and that offering is called an eye screen,” added Solanki.

Autonomous AI

The second way that AI can be a physician tool is through autonomous AI. Solanki describes this AI as giving doctors the ability to scale their work. Rather than requiring physicians to be present at every screening. The AI technology conducts the screening, reducing the need for physicians and allowing more patients to be seen daily.

“Eyenuk’s flagship AI system — the autonomous AI — is called EyeArt and does not need any doctor to make the diagnosis. EyeArt is FDA-cleared for anybody with a high school diploma to operate,” noted Solanki.

Superpower AI

“The third is superpower AI. This AI can do things doctors cannot do today, including looking at multiple visits and the smallest changes. AI can indicate how fast the disease is progressing, and that gives providers a superpower, so to speak,” added Solanki.

He explained that this kind of AI could look at the retina between multiple visits and create prognostic biomarkers to use while monitoring disease progression. While the Eyenuk technology has not yet become a superpower AI, the company is working toward this technology.

“Eyenuk is working on this superpower AI with funding from the National Eye Institute,” revealed Solanki.

Impacts on Healthcare

Solanki told LifeScienceIntelligence that this technology will have a plethora of healthcare implications.

Access to Care and Provider Workflow

According to the Mayo Clinic, Black, Hispanic, and Native American populations are at a higher risk of diabetic retinopathy. Unfortunately, these marginalized communities often have limited access to care. Solanki shared that more patients would have access to screenings because of this AI technology's easy and quick design.

“I believe the biggest impact will be expanding the number of patients who get eye screens. This technology is designed to scale beyond ophthalmology offices to the front lines of care, to primary care offices, to senior living, and even to your neighborhood pharmacies,” he stressed. “That will enable eye screening in the front lines of care, getting more patients to meet their annual screening requirements, reducing vision loss.”

Solanki focused on the efforts to deploy the technology in primary care settings. “The team is doing a lot of work deploying this technology in primary care offices where any nurse or technician can do this procedure within five minutes,” he stated.

Solanki explained that, when diabetic patients enter their primary care provider’s office for their annual check-up, they get their vitals taken. In an ideal world where Eyenuk’s technology is available in every primary care office, vision screenings will be a part of those initial steps. At that point, the provider will have the vision report before seeing the patient.

The technology will either tell providers that the patient’s vision is fine and should be rescreened in a year or that they should be referred to a specialist.

“That certainly goes a long way in preserving vision for these patients and reaching more patients while the workflow impact is minimal at the primary care offices,” he claimed.

Diagnosis and Treatment

In addition to improving access to care and workflow changes, LifeScienceIntelligence asked Solanki how the technology will change the diagnosis and treatment of diabetic retinopathy.

“There are two ways this happens,” he responded. “One is getting this screening done within seconds. The AI report returns in 10 seconds, and the procedure can happen in five minutes without needing dilation.” This speedy procedure offers additional convenience to the patient and the provider.

In addition to timely diagnosis, Solanki emphasized that this technology provides a more accurate and specific diagnosis than an ophthalmologist alone. “The sensitivity of EyeArt is much higher than dilated ophthalmoscopy. Even after a dilated exam, the disease is missed. EyeArt is not going to miss that disease,” he asserted. 

Cost Efficacy

Beyond the patient and provider benefits, Solanki shares that this technology will be cost-effective. He shares that ophthalmologists’ time and billing codes are expensive. Instead of immediately sending patients to an ophthalmologist, primary care providers can conduct these screenings and only refer patients needing care. This workflow change will save money for patients and payers.

Additionally, a timely diagnosis will allow for early interventions that can prevent more expensive treatments down the line.

“From a financial perspective, that gives them additional reimbursement and revenues for doing this. There is a new CPT code with a national payment that pays them handsomely for doing this procedure,” said Solanki. “Payers are happy because they will not pay for expensive treatments for too many patients. At the same time, more people stay healthy. It's a win-win for everybody, including employers or governments who are paying for healthcare.”

Looking Ahead

Solanki and his team shared that Eyenuk is currently being used as a mobile solution by practice in Virginia.

“There is an eye care group in Virginia doing community events and sending vans to senior living and other similar facilities to bring patients into their facilities. Temple University is also one of our major users deploying these camera systems with EyeArt in their primary care offices,” he revealed.

“This project is all driven centrally by their ophthalmology department and public health, where the idea is to funnel only those patients who need attention into the busy ophthalmology practices. Eyenuk also has users who are endocrinologists or diabetes specialists who see many patients and don't want or like the idea of not knowing what's going on with their eyes.”

As this technology continues to develop and expand, healthcare professionals may consider incorporating it into their practices as a form of chronic disease management and prevention. Further studies continue to be done on this technology and its ability to streamline care. 

Editors Note: This article was edited to reflect that the technology is FDA-cleared, not FDA-approved.

Next Steps

Dig Deeper on Biotechnology in medicine

xtelligent Healthtech Analytics
xtelligent Healthcare Payers
xtelligent Health IT and EHR
xtelligent Healthtech Security
Close