Yue Zhong, Min Feng, Rawaz D Tawfeeq, Jun Cheng
Artificial intelligence and deep learning research in ophthalmology has increased significantly, with imaging-based studies contributing the highest output. Publications on LLM-based applications have also increased during recent years. This bibliometric analysis summarizes the distribution of publications, research themes, and methodological trends in ophthalmic artificial intelligence and provides a basis for future clinical and research applications.
BACKGROUND: To examine global research activity in the application of artificial intelligence, large language models, machine learning, and deep learning in ophthalmic diseases from 2015 to 2025.
METHODS: The study retrieved relevant publications from the Web of Science Core Collection. Bibliometric indicators were analysed using VOSviewer, CiteSpace, and Bibliometrix. Annual publication numbers, citation distribution, countries, institutions, journals, co-cited journals, author productivity, co-authorship networks, keyword frequency, burst terms, and thematic clusters were assessed.
RESULTS: A total of 1,997 articles were included in this analysis based on the inclusion criteria. Annual publication numbers increased after 2018 and remained high from 2020 to 2025. The United States, China, and the United Kingdom showed the highest research output.The University of California System, the University of London and University College London were leading institutions in terms of publication output. The most common keywords included deep learning, optical coherence tomography, diabetic retinopathy, and macular degeneration. Citation burst analysis identified influential articles on diabetic retinopathy detection, optical coherence tomography biomarker analysis, and residual network models. Visual cluster analysis identified themes related to automated retinal imaging, diabetic retinopathy screening, fluid quantification models, multimodal image analysis, and recent studies using large language models.
CONCLUSION: Artificial intelligence and deep learning research in ophthalmology has increased significantly, with imaging-based studies contributing the highest output. Publications on LLM-based applications have also increased during recent years. This bibliometric analysis summarizes the distribution of publications, research themes, and methodological trends in ophthalmic artificial intelligence and provides a basis for future clinical and research applications.