This narrative review examined AI-based clinical decision support across four retinal disease settings: diabetic retinopathy (DR) screening and referral, diabetic macular edema (DME) treatment assessment, neovascular age-related macular degeneration (nAMD) retreatment and longitudinal monitoring, and inherited retinal disease (IRD) diagnosis, genotype-phenotype support, progression modeling, and trial enrichment.
Evidence is strongest for DR screening (supported by prospective and real-world validation), while treatment-oriented AI applications in DME and nAMD, and multimodal diagnostic/prognostic tools in IRDs, remain inconsistently validated and lack robust clinical utility data.
As a narrative review, it does not perform systematic search or meta-analysis, limiting reproducibility; the review does not quantify pooled performance metrics across included studies; translational gaps between model performance and real-world clinical utility are acknowledged but not resolved.
AI-assisted DR screening is the most clinically ready application and can support referral decisions with real-world evidence; clinicians should be cautious about adopting AI tools for DME/nAMD treatment decisions or IRD diagnosis outside of validated, prospectively evaluated workflows.
Explore related topics