Semay Baydar

Artificial intelligence (AI) has rapidly progressed from a research concept to a clinically relevant tool, with Ophthalmology emerging as one of the specialties best suited to its adoption. This is largely due to the image-based nature of ophthalmic investigations, the increasing availability of large datasets, and growing service pressures within eye care systems (1). This article provides a brief overview of current clinical applications of AI in ophthalmology and highlights key trends.

The most established use of AI in ophthalmology is the analysis of retinal images and optical coherence tomography (OCT) scans. Deep learning algorithms, particularly convolutional neural networks, can identify subtle patterns within images and have demonstrated diagnostic performance comparable to expert clinicians (1,2). One of the earliest clinically deployed systems is IDx-DR, developed by IDx Technologies in the United States. This system became the first autonomous AI diagnostic tool approved by the US Food and Drug Administration for the detection of more-than-mild diabetic retinopathy using retinal photographs, without the need for clinician interpretation (3). Its introduction has enabled diabetic eye screening to be delivered in primary care and community settings.

Other AI platforms have since been adopted internationally. EyeArt, developed by Eyenuk Inc. (USA), is now used in diabetic retinopathy screening programmes in multiple countries, including within parts of the NHS (2). In the UK, collaboration between Moorfields Eye Hospital, University College London, and DeepMind (now Google Health) has led to the development of AI systems capable of detecting sight-threatening retinal disease on OCT imaging with high accuracy (4). These tools are intended to support clinicians by prioritising referrals and identifying patients requiring urgent review, rather than replacing Ophthalmologists.

Beyond screening, AI is increasingly being explored as a decision-support tool. Algorithms have been trained to predict disease progression, identify patients at risk of deterioration, and assist with monitoring chronic eye conditions (4). Research groups across Europe and East Asia have demonstrated promising results, with companies such as Airdoc in China investing heavily in AI-driven ophthalmic platforms (2). These developments may allow more personalised follow-up and earlier intervention.

AI also has the potential to improve efficiency within overstretched ophthalmology services. Automated image triage, virtual clinics, and remote monitoring systems are being trialled to reduce outpatient workload and waiting times (4,5). In some centres, stable patients are monitored remotely using imaging assessed by AI-assisted systems, allowing clinicians to focus attention on higher-risk cases.

Despite these advances, challenges remain. Concerns persist regarding data bias, algorithm transparency, and medico-legal responsibility (5,6). Many AI systems are trained on datasets from specific populations, which may limit generalisability. There is broad consensus that AI should augment, rather than replace, clinical judgement, with clinicians retaining responsibility for final decision-making (6).

In summary, AI is already influencing Ophthalmology through screening, diagnostic support, and service optimisation. As these technologies continue to evolve, clinicians will need a working understanding of both their benefits and limitations. For trainees and resident doctors, familiarity with AI is likely to become an essential component of modern ophthalmic practice.

References

1. Gulshan V, Peng L, Coram M, Stumpe MC, Wu D, Narayanaswamy A, et al. Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs. JAMA. 2016;316(22):2402–10.
2. Ting DSW, Cheung CY, Lim G, Tan GSW, Quang ND, Gan A, et al. Development and validation of a deep learning system for diabetic retinopathy and related eye diseases using retinal images from multiethnic populations. JAMA. 2017;318(22):2211–23.
3. Abràmoff MD, Lavin PT, Birch M, Shah N, Folk JC. Pivotal trial of an autonomous AI-based diagnostic system for detection of diabetic retinopathy in primary care offices. NPJ Digit Med. 2018;1:39.
4. De Fauw J, Ledsam JR, Romera-Paredes B, Nikolov S, Tomasev N, Blackwell S, et al. Clinically applicable deep learning for diagnosis and referral in retinal disease. Nat Med. 2018;24(9):1342–50.
5. Topol EJ. High-performance medicine: the convergence of human and artificial intelligence. Nat Med. 2019;25(1):44–56.
6. Royal College of Ophthalmologists. Artificial intelligence in ophthalmology: guidance for clinicians. London: RCOphth; 2022.