Skip to main content

Using AI Algorithms to Meet Growing Need for Glaucoma Detection

As glaucoma cases rise amid an impending shortage of eyecare specialists, a research initiative looks at artificial intelligence algorithms as a potential diagnostic tool. The results could help develop methods for detecting preventable vision loss, especially in underserved populations.

In the next decade, researchers expect there to be a rapid rise in the number of patients who have or appear at risk for glaucoma. This is paired with an impending shortage of eyecare specialists.

As the field looks for ways to address this growing divide, artificial intelligence (AI) may provide the answer for improving access to eyecare services – especially to underserved communities.

Benjamin Xu, MD, PhD, an ophthalmologist, glaucoma specialist and clinician-scientist at the USC Roski Eye Institute, part of Keck Medicine of USC, has embarked on a research initiative to use AI to automate the detection and referral of Los Angeles County (LAC) Department of Health Services (DHS) patients with glaucoma.

Supported by a two-year grant from the Southern California Clinical and Translation Science Institute (SC-CTSI) and LAC DHS, as well as generous support from the Carlson Family Foundation, Xu’s lab will work on developing the preliminary AI algorithms to identify eyes with or at risk for glaucoma using fundus photographs and optical coherence tomography (OCT) for implementation in LAC DHS glaucoma screening clinics.

“Within the next decade or so, it’s projected there will be a shortage of around 6,000 ophthalmologists,” Xu said. “AI is urgently needed to address this impending eyecare access crisis. There will be many patients who will develop glaucoma, who could permanently lose their vision, but we won’t have the healthcare capacity to detect them and deliver sight-saving treatment.”

Localized approach to AI research

Currently, Xu’s team is taking a localized approach to AI implementation, using data from LAC patient populations to develop algorithms that serve those same populations.

“At present, our AI algorithm is developed using data from LAC DHS, where the majority of patients are Latino,” Xu said. “While our algorithms may work more effectively for local patients, the methods and algorithms we’re trying to develop are meant to validate an eyecare delivery model that can be broadly adopted worldwide.”

The team eventually plans to establish a nationwide consortium that shares data and works together to better understand whether localized AI algorithms can be applied in diverse and independent populations.

Current challenges lie not in developing AI algorithms, but in implementing them—and doing so in a way that is equitable across patient populations.

“AI implementation requires merging laboratory research with real-world clinical practice,” Xu said. “We are leveraging data from tens of thousands of under-served patients and developing AI algorithms specifically tailored to improve glaucoma care among similar patient communities."

Using AI to streamline diagnostic and referral processes

According to Xu, a key focus for the study will be demonstrating that AI algorithms provide a high level of performance that simulates a well-trained eyecare provider.

“OCT provides high-quality images that are effective for identifying patients at risk for glaucoma,” Xu said. “I ventured into the field of AI to automate the analysis of these images so that OCT can be used more conveniently by physicians.”

Xu also said that AI has the potential to completely revolutionize and transform the way that eyecare is provided—and not just for glaucoma. AI can be designed to augment diagnosis and treatment of other common eye diseases like macular degeneration and diabetic retinopathy, helping to facilitate the referral process to specialists.

“We're not suggesting that the machine do all of the work,” Xu said. “We see AI as a tool to help human providers streamline care so that high-suspicion cases are moved to the top of the pile for manual evaluation.”