Multimodal artificial intelligence in ophthalmology: Applications, challenges, and future directions.

Area of Science:

• Ophthalmology
• Artificial Intelligence
• Medical Informatics

Background:

• Multimodal artificial intelligence (AI) integrates diverse data types, offering enhanced analytical capabilities.
• AI's rapid development presents significant opportunities within specialized medical fields like ophthalmology.

Purpose of the Study:

• To systematically review and evaluate the applications, technical features, and clinical value of multimodal AI in ophthalmology.
• To assess the performance improvements of multimodal AI systems compared to unimodal approaches.

Main Methods:

• Systematic review adhering to PRISMA guidelines.
• Literature search across PubMed, Web of Science, Scopus, and Google Scholar (2018-2025).
• Inclusion of 10 studies for final analysis.

Main Results:

• Key applications include glaucoma, age-related macular degeneration, corneal diseases, emergency triage, and diagnostic chatbots.
• Multimodal AI systems demonstrated superior performance over unimodal systems.
• Improvements noted: 4-5% in Area Under the Curve and 2-7% in accuracy.

Conclusions:

• Multimodal AI offers comprehensive and accurate diagnostic information in ophthalmology.
• Future directions include clinical validation, advanced fusion techniques, interpretability, and model optimization.
• Promoting clinical translation of multimodal AI is crucial for advancing eye care.