Applications of Large Language Models in Glaucoma: A Scoping Review

Area of Science:

• Ophthalmology
• Artificial Intelligence
• Medical Informatics

Background:

• Large language models (LLMs) and vision-language models (VLMs) are increasingly used in ophthalmology for patient education, diagnosis, and surgical support.
• Their capabilities in processing medical information make them valuable for glaucoma care.

Purpose of the Study:

• To conduct a scoping review of LLMs and VLMs in glaucoma.
• To summarize their applications, performance, and limitations to inform future research and clinical use.

Main Methods:

• Systematic literature search across major databases (PubMed, Scopus, Web of Science, arXiv, IEEE Xplore) from 2014 to July 2025.
• Inclusion of studies on LLMs/VLMs in glaucoma for diagnosis, education, or surgical support.
• Screening and review by two independent reviewers following PRISMA-ScR guidelines.

Main Results:

• 27 studies met inclusion criteria from 316 identified records.
• Applications were categorized into patient education (11 studies), diagnosis/risk prediction (10 studies), and surgical management (6 studies).

Conclusions:

• LLMs are currently best suited as assistive tools in glaucoma care, not autonomous systems.
• Potential is high for patient communication and text-based decision support, but accuracy and multimodal integration need improvement.
• Future work should focus on domain-specific LLMs, enhanced multimodal fusion, patient readability, and ethical frameworks.