Are there any artificial intelligence (AI)-driven platforms to screen for ophthalmologic diseases?

From the Guidelines

Yes, there are AI-driven platforms to screen for eye diseases, such as diabetic retinopathy, age-related macular degeneration, and glaucoma. These platforms use deep learning algorithms to analyze retinal images and detect abnormalities [ 1 ]. Examples of such platforms include IDx-DR, which uses a robotic camera to take images of the retina and detect diabetic retinopathy, and EyeArt, which uses AI to detect diabetic retinopathy from retinal images. Additionally, platforms like Google's LYNA use AI to detect diabetic retinopathy and other eye diseases [ 1 ].

Key Features of AI-Driven Platforms

• Use deep learning algorithms to analyze retinal images [ 1 ]
• Can detect abnormalities such as diabetic retinopathy, age-related macular degeneration, and glaucoma [ 1 ]
• Examples of platforms include IDx-DR, EyeArt, and Google's LYNA [ 1 ]

Benefits and Limitations

• Can help identify patients who require further evaluation and treatment [ 1 ]
• Should be used in conjunction with a comprehensive eye exam by a qualified eye care professional [ 1 ]
• May have limitations in understanding discriminatory factors and generalizability to different patient groups [ 1 ]

Clinical Applications

• Can be used for population-based screening [ 1 ]
• May be useful for screening in areas where qualified eye care professionals are not readily available [ 1 ]
• Can provide timely referral for comprehensive eye examination when indicated [ 1 ]

From the Research

AI-Driven Platforms for Ophthalmologic Disease Screening

There are several artificial intelligence (AI)-driven platforms that have been developed to screen for ophthalmologic diseases, including:

• Diabetic retinopathy (DR) screening systems, such as IDx-DR, which is FDA-approved for autonomous detection of referable diabetic retinopathy 2
• Age-related macular degeneration (AMD) screening systems, which can quantify drusen and diagnose disease severity from color fundus and optical coherence tomography images 2
• Glaucoma screening systems, which can assess optic nerve head morphology in fundus photographs to detect damage 2, 3
• Combined screening systems for multiple diseases, such as the system developed by iHealthScreen Inc., which can screen for both DR and AMD 4

Performance of AI-Driven Platforms

The performance of these AI-driven platforms has been evaluated in several studies, with results showing:

• High sensitivity and specificity for DR screening, with sensitivity ranging from 97.0% to 100.0% and specificity ranging from 80.10% to 96.3% 4, 5
• High sensitivity and specificity for AMD screening, with sensitivity ranging from 86.6% to 90.91% and specificity ranging from 78.79% to 92.1% 4, 5
• High sensitivity and specificity for glaucoma screening, with sensitivity ranging from 93.26% and specificity ranging from 76.76% 5

Implementation and User Experience

The implementation and user experience of these AI-driven platforms have also been evaluated, with results showing:

• Positive feedback from clinical staff on the disease screening process, particularly praising the seamless workflow from patient registration to image transmission and obtaining the final result 5
• High satisfaction rates among users, with most finding the system easy, quick, and highly satisfactory 5
• Room for improvement in certain models, such as the AMD model, which may require enhanced training data 5