The number of individuals experiencing blindness is increasing, but there is hope in the form of AI to combat this issue.

 The field of ophthalmology has experienced a significant increase in clinical appointments within the UK's National Health Service (NHS) since 2017. Eye-related issues now account for nearly 10 percent of all NHS outpatient appointments, totaling close to 10 million appointments annually. This represents a more than one-third rise over the past five years.

While diabetic eye disease is the leading cause of blindness among individuals aged 18 to 65, the aging population has led to a higher prevalence of age-related macular degeneration (AMD), which is the most common cause of blindness. A recent study revealed that approximately 25.3 percent of individuals aged 60 and above in Europe display early signs of AMD. Additionally, in the UK, around 200 people per day develop a severe form of AMD known as wet AMD, resulting in blindness due to bleeding at the back of the eye.

The growing number of patients seeking ophthalmic care has posed challenges for ophthalmologists, leading to delays in diagnosis and treatment, ultimately contributing to cases of avoidable blindness. However, technology, particularly optical coherence tomography (OCT) devices, has emerged as a potential solution. These advanced scanners, now available in optometry practices across the country, can capture highly detailed images of the retina in a noninvasive manner.

The deployment of OCT scanners presents its own set of challenges, as community optometrists may lack the expertise to interpret OCT scans accurately. Consequently, many patients are unnecessarily referred to eye hospitals, exacerbating the burden on the healthcare system.

Artificial intelligence (AI) can play a crucial role in addressing these challenges by bringing world-class expertise to the community level. A collaborative effort between Moorfields Eye Hospital and DeepMind showcased in a 2018 Nature paper demonstrated that an AI system could effectively analyze OCT scans and detect over 50 retinal diseases, performing on par with expert ophthalmologists.

Since then, efforts have focused on clinically validating the AI system by training it on diverse datasets to ensure its applicability to patients of all ethnicities and clinical settings. Once validated, the AI system can be implemented on a large scale within the community. The algorithm will enable the identification and prioritization of patients with the most severe conditions, allowing for timely treatment at hospitals. This approach will alleviate the burden of chronic diseases like AMD.

The advancement of medical AI in ophthalmology can be likened to Thomas Edison's invention of the electric light bulb. Just as Edison recognized the need for a network of innovations to usher in the electrical age, ophthalmology AI requires a similar infrastructure. This includes connecting optometry practices with OCT machines to the cloud, initiating national transformation programs in the NHS, and establishing pathways and payment systems to facilitate patient transfer from the community to hospitals. Once these innovations converge, the AI network will be ready for comprehensive deployment, revolutionizing the field of ophthalmology.