Advanced analysis of leading large language models for diagnostic accuracy in retinal imaging
Current LLMs show promising but limited capabilities in ophthalmological image interpretation. While performance on common conditions like retinal detachments and age-related macular degeneration is moderately good, significant challenges remain with rare conditions, myopic pathologies and complex vascular disorders. The competitive performance between GPT-4.5 and Gemini 2.0 Pro, with each excelling in different pathology categories, suggests that leveraging their complementary strengths might offer improved diagnostic support.
Authors
- Francesco Boselli (ORCID: https://orcid.org/0009-0001-8385-5316)
- Matteo Mario Carlà (ORCID: https://orcid.org/0000-0003-2979-1638)
- Fiammetta Catania (ORCID: https://orcid.org/0000-0003-1551-8120)
- Laura De Luca (ORCID: https://orcid.org/0009-0004-3513-5826)
- G. Gambini
- Carlos Mateo (ORCID: https://orcid.org/0000-0003-0947-9575)
- Emanuele Crincoli (ORCID: https://orcid.org/0000-0001-9996-9871)
- Federico Giannuzzi
Institutions
- University of Messina (IT)
- Agostino Gemelli University Polyclinic (IT)
- Fondation Ophtalmologique Adolphe de Rothschild (FR)
- Fondation de Rothschild (FR)
- Instituto de Microcirugía Ocular (ES)
- University of the Sacred Heart (JP)
Publication Details
- Journal
- British Journal of Ophthalmology
- Published
- 2026-01-14
- DOI
- https://doi.org/10.1136/bjo-2025-327634
- Primary Topic
- Retinal Imaging and Analysis
- Type
- article
- Field-Weighted Citation Impact
- 0.00