AI in Glaucoma: What's Working Now, What's Coming Next, and Where the Real Opportunities Lie

This audio article is from VisualFieldTest.com.

Read the full article here: https://visualfieldtest.com/en/ai-in-glaucoma-what-s-working-now-what-s-coming-next-and-where-the-real-opportunities-lie

Test your visual field online: https://visualfieldtest.com

Support the show so new episodes keep coming: https://www.buzzsprout.com/2563091/support

Excerpt:

Introduction Glaucoma is a group of eye conditions that damage the optic nerve and can lead to irreversible blindness. Often called “the silent thief of sight,” glaucoma affects millions worldwide. In fact, an estimated 111.8 million people will have glaucoma by 2040 (). Early detection and treatment are critical because vision loss cannot be fully recovered. This is where artificial intelligence (AI) is making inroads: by analyzing eye images and test data, AI can help screen, diagnose, and monitor glaucoma more efficiently. In this article we explore how AI is being applied today in glaucoma care – citing real-world tools and studies – and examine emerging opportunities, especially in vision restoration research. We focus on proven results (e.g. sensitivity and specificity of AI tools) and on concrete future applications, providing practical guidance for patients and researchers alike.AI in Current Glaucoma Screening and Diagnosis Smartphone and Fundus Image Analysis One major use of AI today is automated analysis of fundus photographs (images of the retina) to screen for glaucoma. Research teams have paired portable fundus cameras or smartphone attachments with AI classifiers to flag glaucomatous optic discs. For example, a recent prospective study in India tested an offline AI model embedded on a smartphone fundus camera (Medios AI-Glaucoma on Remidio’s FOP NM-10 device). This system detected patients needing referral (“referable glaucoma”) with about 94% sensitivity and 86% specificity compared to a full clinical workup (). In numbers, the AI correctly identified 93.7% of true glaucoma cases and correctly excluded 85.6% of non-glaucoma cases (). Such high accuracy shows that smartphone-based AI screening can reliably find patients with glaucoma changes in their optic discs. Another study used a similar AI-camera setup across all severities of glaucoma. It found the AI achieved 91.4% sensitivity and 94.1% specificity for detecting glaucoma or suspect cases (). Performance was slightly lower for very early disease (about 87% sensitivity) and highest for advanced cases (96% sensitivity) (). These results come from outpatient clinics and show that AI tools can match specialist exams in flagging suspicious eyes. They also highlight that AI often errs on the side of caution by flagging mild or suspect cases; in one study most false-positives were eyes labeled “disc suspect” by specialists (). This conservative approach helps avoid missing true disease at the cost of some extra referrals. Commercial and research groups are already developing such systems. For instance, the Medios AI-Glaucoma system (Remidio, India/ Singapore) integrates on a smartphone fundus camera and has shown the results above () (). Other AI platforms (e.g. BegIA) use smartphone images to estimate cup-to-disc ratios or even analyze facial images for eye abnormalities (). In one clinical evaluation, a smartphone app reported an area under curve (AUC) of 0.966 for glaucoma detection, with 95.4% sensitivity and 87.3% specificity (). Telemedicine and Remote Screening AI-enabled apps are also used in telemedicine for glaucoma. For example, the iPredict cloud platform runs AI on uploaded fundus images. In a real-w