Performance of Foundation Models vs Physicians in Textual and Multimodal Ophthalmological Questions

How do current foundation models perform when presented with offline ophthalmology textual and multimodal questions compared with experienced and nonexperienced physicians and older large language models (LLMs)?In this cross-sectional study including 7 foundation models, the latest foundation models exhibited advancements compared with older LLMs tested previously when evaluated with ophthalmology textual questions, but its current multimodal abilities remain limited, inferior to both ophthalmology trainees and experts.These results suggest that foundation models can provide useful assistance in ophthalmology when presented with textual queries, but further developments are necessary before its multimodal capabilities can be of substantial benefit in clinical practice.

Corresponding Authors: Henry Rocha, MB BChir, University of Cambridge School of Clinical Medicine, University of Cambridge, Cambridge CB2 0SP, United Kingdom (henryrocha8@gmail.com); Darren Shu Jeng Ting, MBChB, PhD, Academic Unit of Ophthalmology, Department of Inflammation and Ageing, School of Infection, Inflammation and Immunology, College of Medicine and Health, University of Birmingham, Birmingham B15 2TT, United Kingdom (d.s.j.ting@bham.ac.uk).

Link to the article in your story

We encourage you to link out to this article in your story using the link below. It includes an access token that will give free access to the article for your readers up to one year after publication. (The link will be live after the article publishes and embargo is lifted.)

Please see the article for additional information, including full author list, author contributions and affiliations, conflict of interest and financial disclosures, and funding and support.

Need more information? Contact us.