Year: 2025
Type: article
Abstract: Artificial intelligence (AI) has established itself as a transformative tool in ophthalmic practice, especially for the early diagnosis of diabetic retinopathy, glaucoma and age‑related macular degeneration (AMD). Convolutional neural networks trained on large retinal‑image databases have demonstrated accuracy comparable to—and in some scenarios surpassing—that of human specialists, enabling large‑scale screening at reduced cost and with broader reach of care. Besides expanding access in regions lacking ophthalmologists, AI reduces inter‑observer variability and streamlines routine examinations, freeing clinicians to focus on complex cases. Over recent decades, collaborative efforts among universities, technology companies and health‑care services have produced algorithms that integrate clinical data and electronic health records to generate personalised risk predictions and therapeutic recommendations. Nevertheless, algorithmic bias, the scarcity of representative datasets and regulatory hurdles still constrain widespread adoption of these solutions. We conclude that, although AI is already clinically useful in specific contexts, continuous validation, transparency and medical education are essential for the technology to realise its full potential in combating preventable blindness.
Language: English
Topic: Retinal Imaging and Analysis
Subfield: Radiology, Nuclear Medicine and Imaging
Field: Medicine
Domain: Health Sciences
Open Access status: hybrid
OpenAlex ID: W4411282225