Nov 29 2022

Innovations in Eye Care: How Technology Is Transforming Ophthalmology

Artificial intelligence applications are helping eye doctors personalize care for patients with conditions such as AMD and diabetic retinopathy.

Innovations in eye care have the potential to improve patient outcomes and deliver new treatment options.

Artificial intelligence applications in ophthalmology are gaining traction as AI algorithms help physicians predict risk for a wide range of ophthalmic conditions such as diabetic retinopathy and age-related macular degeneration (AMD).

AI systems can analyze and interpret complex eye scans and help detect, monitor and treat AMD and diabetic retinopathy.

In addition, telehealth platforms and remote patient monitoring are gaining momentum. During the height of the COVID-19 pandemic, ophthalmic practices used telehealth and remote vision monitoring to help patients get the care they needed when in-person visits were largely unavailable.

“These technologies have the potential to fundamentally transform eye care,” says Dr. Konstantinos Balaskas, director of the Moorfields Ophthalmic Reading Centre and Clinical AI Lab at Moorfields Eye Hospital in London.

Click the banner below for access to exclusive HealthTech content and a customized experience.

How AI Applications Can Revolutionize Eye Care

AI software in ophthalmology holds great potential in helping physicians automate the diagnostic process, according to Dr. T.Y. Alvin Liu, assistant professor of ophthalmology at the Johns Hopkins Medicine Wilmer Eye Institute and director of the Wilmer Precision Ophthalmology Center of Excellence.

“AI enables the screening of ophthalmic diseases at scale and more efficiently,” Liu says.

Screening of diabetic retinopathy, which involves damage to the retina’s blood vessels, is a key use case for AI, especially considering the exponential increase in the incidences of diabetes, Liu notes.

Primary care doctors could use AI to perform diabetic retinopathy screening, Liu says. Companies that make FDA-approved AI diabetic retinopathy tools include Digital Diagnostics and Eyenuk.

“Once you get the picture taken on the back of the eye by a fundus camera, the AI automatically makes a recommendation and detection of whether you have diabetic retinopathy, and that is all done autonomously, without human intervention,” Liu says. “So, it is happening. It’s not science fiction. The question is whether the deployment of these devices actually improves the metrics that we care about.”

Liu is exploring whether AI-based screening improves the compliance rate for annual diabetic retinopathy screening at Johns Hopkins Medicine.

“We just got our preliminary data back showing that in clinics with AI-based screening, the compliance rate improved, and this improvement was, statistically, significantly better than the primary care sites without AI screening,” Liu says.

Dr. T.Y. Alvin Liu
AI enables the screening of ophthalmic diseases at scale and more efficiently.”

Dr. T.Y. Alvin Liu Assistant Professor of Ophthalmology, Johns Hopkins Medicine Wilmer Eye Institute

A cloud provider works with an eye care clinic to manage a cloud server, perform data analysis and then provide test results, Liu explains.

The AI screening allows ophthalmologists to focus on screening patients with positive disease, and after the AI program analyzes an image, it can return a result within about a minute, he says.

At Moorfields Eye Hospital in London, Balaskas developed and validated an AI model that can detect and quantify a type of AMD called geographic atrophy, a chronic progressive degeneration of the macula.

“OCT scans play a crucial role in the management of AMD,” Balaskas says. “Detailed interpretation and measurements of the AMD type called geographic atrophy on an OCT scan by a human expert can take up to 43 minutes. The same task takes only 2 seconds for the AI system.”

He notes that this type of AMD was previously untreatable. Now, therapeutics for AMD exist; the first one is soon to receive an FDA license for use.

“It will be indispensable to have automation to monitor these patients, and to monitor their response to treatment,” Balaskas says.

EXPLORE: Practical applications for artificial intelligence in diagnostic intelligence.

How Ophthalmology Benefits from Telehealth and Virtual Care

Ophthalmologists face different challenges with telehealth compared with medical specialties that are well suited to virtual care, such as psychiatry. The lack of reliable and high-quality imaging limits telehealth implementation in ophthalmology, according to Liu.

“The problem with ophthalmology is we are very driven by images,” Liu says. “And the only way to get over it is to really push the development of at-home ophthalmic imaging devices.”

Although patients can’t take a picture of the back of their eye by themselves, during the COVID-19 pandemic they would send a selfie of the eye to doctors to triage and decide which patients need to be seen in person, says Dr. Bennie H. Jeng, director of the Scheie Eye Institute at Penn Medicine and a spokesperson for the American Academy of Ophthalmology.

In addition, patients would go to pressure checking stations during the pandemic and get screened for glaucoma, Jeng says. A physician or technician would check the eye pressure and then contact an ophthalmologist if the result was abnormal. They could then determine if the patient needed a change in medication or another intervention. At some institutions, ophthalmologists use secure platforms such as Doximity to connect with patients.

Click the banner below to learn more about virtual care implementation.

Telehealth holds more potential for examining the front of the eye for triaging, compared with the back of the eye when working directly with patients, according to Jeng, because patients generally do not have the ability to image the back of the eye. Doctors can see if eyes are red or if there’s opacity on the cornea, he says. If patients have images taken of the back of their eye at a doctor’s office, then these images can be interpreted by an ophthalmologist.

“We’re able to screen out things that don’t look malignant,” Jeng says. He predicts that telehealth will increase in use as the technology becomes more advanced and less expensive.

Moorfields in London has pioneered the concept of diagnostic imaging hubs, where patients with conditions like AMD, diabetic retinopathy or glaucoma can have tests and scans but not a live consultation with a physician. Doctors review the scans within 24 hours and decide on appropriate steps for the patient. Treatment can include a consultation, further monitoring in the hub or medical treatment.

“The result is a better experience of care for our patients who avoid unnecessary long consultations and increased capacity of the healthcare team to provide the right care to the right patients at the right time,” Balaskas says.

However, scanning equipment must become smaller and more affordable for home use and for doctors to monitor eye scans remotely, he adds.

DISCOVER: How to achieve an end-to-end virtual care solution.

Advances in Optical Coherence Tomography for Eye Care

Optical coherence tomography is a widely used type of ophthalmic imaging that can be thought of as a light-based ultrasound. It lets doctors image tissue at a high-resolution noninvasively. Home-based OCT machines have been undergoing FDA approval and could revolutionize tele-ophthalmology, Liu says.

Although home OCT technologies are promising, they have yet to gain traction due to difficulty in scaling such expensive devices for personal use.

The goal is for every patient to have an inexpensive, compact, miniaturized OCT device at home, according to Balaskas.

“We are not quite there yet from the technology side because it’s still not possible to sufficiently reduce OCT component parts in size and cost,” Balaskas says. He predicts that in the “medium term” they are likely to become more affordable to allow scaling across an entire patient population.

Tracking the Future of Eye Care Technology

The benefit of AI going forward will be in predictive analytics and “oculomics,” according to Balaskas. Oculomics involves the association of ophthalmic biomarkers with systemic health and disease.

AI will help determine which patients may have a positive or negative response to a certain eye treatment and will be able to make predictions based on a patient’s history, images and clinical genetics.

“Multimodal AI will be able to make predictions based on a patient’s history, lifestyle, clinical information, various medical images and scans, and genetics,” Balaskas says. “Lab tests and even remote monitoring sensors will allow ‘the gift of time’ for the all-important patient-clinician interaction.”

Despite the growth of AI, the jobs of eye doctors should be safe because they will be needed to interpret the images, just as a cardiologist is needed to review an EKG, according to Jeng.

“I don’t think we’re ever going to be replaced,” Jeng says. “But do I think that it’s going to make our job easier? Yes. Do I think that it might make it so we’re less reliant on physicians? It certainly might.”

RELATED: Understand key concerns for adoption of medical devices using AI.

peakSTOCK/Getty Images

aaa 1

Register