A global review of publicly available datasets for ophthalmological imaging: barriers to access, usability, and generalisability

AI Generated Summary*

A review catalogued publicly available ophthalmic imaging datasets found through MEDLINE, Google, and Google Dataset Search. Of 140 datasets, 94 permitted download, 27 nominally open ones were unobtainable, and 19 required licensing, approval, or payment. Those 94 held 507,724 images and 125 videos from at least 122,364 patients, mostly from Europe, Asia, or North America. Fundus photography featured in 57%, OCT or OCT angiography in 19%. Healthy eyes (62%), diabetic eye disease (37%), glaucoma (20%), and macular degeneration (16%) dominated, while cataract, trachoma, and refractive error barely featured. Only 20% reported age, sex, or ethnicity; 15% gave inclusion and exclusion criteria. No known datasets existed for 172 countries, 45% of the global population, which the authors call data poverty.

*This summary was generated by AI and is published unedited. Oku does not alter these summaries. It may contain errors or omissions and is provided for general informational purposes only. Oku does not guarantee its accuracy, completeness, or reliability. For authoritative information, please refer to the original, peer-reviewed article.