Leveraging Artificial Intelligence for Diabetic Retinopathy Screening and Management: History and Current Advances

Area of Science:

• Ophthalmology
• Medical Imaging
• Artificial Intelligence

Background:

• Diabetic retinopathy (DR) screening is a global challenge due to limited human resources.
• Digital health innovations, particularly AI, show significant promise for DR screening.
• Artificial intelligence (AI)-based deep learning (DL) algorithms are effective for diagnosing referable DR (RDR).

Purpose of the Study:

• To evaluate the application of AI in DR screening.
• To review currently available automated DR detection algorithms.

Main Methods:

• A narrative review of articles published up to May 15th, 2024.
• Searches conducted in PubMed, Medline, Embase, Scopus, and Google Scholar.
• Keywords included diabetic retinopathy, AI, deep learning, automated detection, and retinal imaging.

Main Results:

• AI has advanced digital health for DR diagnosis.
• Key AI/DL concepts and validated algorithms for DR detection were identified.
• IDx-DR, EyeArt, and AEYE-DS are USFDA-approved automated DR screening algorithms.

Conclusions:

• AI revolutionizes DR screening through early automated detection.
• Advances in AI and retinal imaging facilitate early diagnosis and better patient outcomes.