Front Med (Lausanne). 2025 ;12 1580583
Purpose: This study aims to generate and elucidate the latest perspectives on the application of artificial intelligence (AI) in ophthalmology using bibliometric methods. By analyzing literature from the past 5 years (2020-2024), we seek to outline the development trends of this technology, provide guidance for its future directions, and assist clinicians in adapting to these innovations.
Methods: We conducted a comprehensive search of all literature related to AI and ophthalmology in the Web of Science Core Collection (WoSCC) using bibliometric methods. The collected data were analyzed and visualized using three widely recognized bibliometric software tools: CiteSpace, VOSviewer, and the R package "Bibliometrix."
Results: A total of 21,725 documents were included from 134 countries and 7,126 institutions, consisting of 19,978 articles (91.96%) and 1,714 reviews (8.04%), with China and the United States leading the contributions. The number of publications in AI and ophthalmology has increased annually, with the University of California System, the National University of Singapore, and the University of London being the primary research institutions. Ophthalmology and Proc CVPR IEEE are the most co-cited journals and conferences in this field. These papers were authored by 87,695 individuals, with Wang Y, Liu Y, and Zhang Y the most prolific authors. Ting DSW was the most co-cited author. Major research topics include using various models to scan retinal images for diagnosing conditions such as age-related macular degeneration, diabetic retinopathy, and retinal nerve fiber layer thinning caused by glaucoma. The intersection of AI with other subfields of ophthalmology, such as in the diagnosis of ametropia, strabismus, eyelid disease, and orbital tumors, as well as in postoperative follow-up, is also rapidly developing. Key research hot spots are identified by keywords such as "deep learning," "machine learning," "convolutional neural network," "diabetic retinopathy," and "ophthalmology."
Conclusion: Our bibliometric analysis outlines the dynamic evolution and structural relationships within the AI and ophthalmology field. In contrast to previous studies, our research transcends individual domains to offer a more comprehensive insight. Notably, our analysis encompasses literature published beyond the year 2022, a pivotal year marking both the post-pandemic era and the rapid advancement of AI technologies. This temporal scope potentially fills a gap that prior bibliometric studies have not addressed. This information identifies recent research frontiers and hot spot areas, providing valuable reference points for scholars engaging in future AI and ophthalmology studies.
Keywords: artificial intelligence; deep learning; diabetic retinopathy; glaucoma; machine learning; ophthalmology