AI screening could prevent vision loss disease, study finds

Imagine walking into a supermarket, train station, or shopping mall and having your eyes checked for glaucoma within seconds—no appointment needed. Thanks to a new AI-powered system called AI-GS (AI-based Glaucoma Screening), this futuristic idea could soon become a reality.

The Silent Threat of Glaucoma

Glaucoma is a leading cause of irreversible blindness, both in Japan and worldwide. The disease progresses quietly, gradually narrowing a person’s field of vision.

Because the loss of vision starts at the edges, most people do not notice the symptoms until significant and permanent damage has already occurred. This delay in diagnosis is a major problem, especially in areas where access to eye specialists is limited.

The traditional method for detecting glaucoma requires a visit to an ophthalmologist and specialized testing, making widespread screening difficult. Many cases go undiagnosed until it is too late to prevent severe vision loss.

A Game-Changing AI Screening System

To tackle this problem, researchers at Tohoku University, led by Professor Toru Nakazawa and Associate Professor Parmanand Sharma, have developed the AI-GS network. This system can quickly and accurately analyze multiple key indicators of glaucoma using images of the back of the eye, where glaucoma-related damage occurs.

Tested on a dataset of 8,000 retinal images, AI-GS achieved an impressive 93.52% sensitivity at 95% specificity—matching the accuracy of expert ophthalmologists. Unlike many traditional AI models, this system is particularly skilled at detecting early-stage glaucoma, even when visible damage is minimal.

Overcoming the “Black Box” Problem

One major challenge in AI-driven healthcare is transparency. Many AI systems operate as “black boxes,” meaning doctors and patients cannot see how they reach their conclusions.

AI-GS solves this issue by providing numerical values for each diagnostic feature, allowing ophthalmologists to understand and verify its decisions. This transparency increases trust and makes it easier to integrate into clinical practice.

Small, Fast, and Accessible Anywhere

Another key feature of AI-GS is its portability. The system is only 110 MB in size, requires minimal computing power, and delivers a diagnosis in under a second. This means it can run on mobile devices, allowing glaucoma screenings to be conducted in public places such as shopping centers and train stations.

“This technology brings expert-level glaucoma screening to your pocket,” says Associate Professor Sharma. “Because of its portability, it can be used anywhere—even in remote areas where eye specialists are scarce.”

Professor Nakazawa emphasizes the system’s potential to prevent blindness on a massive scale. “By enabling early detection, we can prevent unnecessary vision loss for millions of people worldwide,” he says.

The Future of Glaucoma Screening

With its high accuracy, transparent decision-making, and lightweight design, AI-GS represents a breakthrough in AI-driven eye care. If widely implemented, this system could revolutionize glaucoma screening by making it easily accessible outside of hospitals and clinics.

Millions of people, especially those in underserved communities, could benefit from earlier diagnosis and treatment, reducing the risk of blindness worldwide.

The research findings can be found in npj Digital Medicine.