Artificial intelligence to identify eye diseases

Dr Pearse Keane

Ground-breaking research by Moorfields Eye Hospital and DeepMind Health

Moorfields Eye Charity is delighted to report on the ground-breaking research by Moorfields Eye Hospital NHS Foundation Trust and DeepMind Health which uses artificial intelligence (AI) to identify potentially blinding eye diseases.  The latest outcomes of this truly exciting research have been published in Nature Medicine and describe how machine learning technology has been successfully trained on thousands of historic de-personalised eye scans. 

Moorfields patient, Elaine Manna, and consultant Dr Pearse Keane.

The AI system can recommend the correct referral decision for over 50 eye diseases with 94% accuracy, matching world-leading eye experts. It is hoped that the technology could revolutionise the way professionals carry out eye tests, allowing them to spot conditions earlier and prioritise patients with the most serious eye diseases before irreversible damage sets in.

Launched in 2016, this study brought together leading NHS eye health professionals and scientists from the National Institute for Health Research (NIHR) and UCL with some of the UK’s top technologists at DeepMind to investigate whether AI technology could help improve the care of patients with sight-threatening diseases, such as age-related macular degeneration (AMD) and diabetic eye disease.

Pearse Keane, consultant ophthalmologist at Moorfields Eye Hospital NHS Foundation Trust and NIHR clinician scientist at the UCL Institute of Ophthalmology said:

The number of eye scans we’re performing is growing at a pace much faster than human experts are able to interpret them. There is a risk that this may cause delays in the diagnosis and treatment of sight-threatening diseases, which can be devastating for patients.

The AI technology we’re developing is designed to prioritise patients who need to be seen and treated urgently by a doctor or eye care professional. If we can diagnose and treat eye conditions early, it gives us the best chance of saving people’s sight. With further research it could lead to greater consistency and quality of care for patients with eye problems in the future.

What are the next steps for this research?

The next step is for the research to go through clinical trials to explore how this technology might improve patient care in practice, and attain regulatory approval before it can be deployed in hospitals and other clinical settings.  

Dr Pearse Keane
Dr Pearse Keane

Taking time to speak with Moorfields Eye Charity, Pearse Keane, explained how these next steps are crucial to ensure that any new developments are properly tested, reviewed and validated.  Once these steps have been achieved and the technology is deployed, the potential impact could be revolutionary and to the benefit of large number of people.

More than 285 million people worldwide live with some form of sight loss, including more than two million people in the UK. Eye diseases remain one of the biggest causes of sight loss, and many can be prevented with early detection and treatment.

Where might technology take us and change how we interact with services?

With a view to the future, we asked Pearse to speculate how this technology might change practice in the near and distant future.  Shorter term goals might be to have ‘rapid access clinics’ where identification of cases for referral is more accurate therefore the triage - for example, urgent, non-urgent follow-up, monitor -  of patients is achieved more effectively.

Further into the future, with more research and clinical trials, it is not inconceivable that initial triaging or even treatment monitoring could be done in GP or local clinics.  This could help ease acute clinical services, allow monitoring closer to home and in timescales bespoke to the clinical needs of the individual rather than applying a standard across all patients with a similar disease.

Robert Dufton, chief executive at Moorfields Eye Charity, said:

The need for treatment for eye diseases is forecast to grow, in part because people are living longer, far beyond our ability to meet the demand using current practice. Artificial intelligence is showing the potential to transform the speed at which diseases can be diagnosed and treatments suggested, making the best use of the limited time of clinicians. AI will also help our understanding of sight loss.

Moorfields Eye Charity is proud to have funded equipment which underpins Dr Pearse Keane’s work as part of our programme of philanthropic support in pioneering researchers.

 

Related research

Any other development in areas of technology and imaging?

Another area of work for Pearse and his team is in terms of binocular OCT. This research is not related to that discussed above but is another avenue which is looking at harnessing technology to improve and advance eye care and the patient experience.  In this project, a hand held device is being tested. Binocular OCT provides extremely high-resolution images of the eye in a completely non-invasive manner. It also aims to be able to carry out a large suite of standard tests such as visual acuity and eye movements in a flexible and convenient way. This work is still in the early stages but there is great value in tackling the big questions in eye care and eye health from as many angles as possible, learning from various projects to enable the continued advancement in understanding, diagnosis and treatments.

binocular OCT
Binocular OCT

Some exciting publications have already stemmed from this work, two of which we have included below. We look forward to reporting on further updates in the future.

Use of a Binocular Optical Coherence Tomography System to Evaluate Strabismus in Primary Position. Chopra R, Mulholland PJ, Tailor VK, Anderson RS, Keane PA.
JAMA Ophthalmol. 2018 Jul 1;136(7):811-817. doi: 10.1001/jamaophthalmol.2018.1973.
 

This research asked the question; can binocular optical coherence tomography (OCT) be used to measure the size of strabismus?

This study included 15 participants with strabismus and 15 healthy volunteers. In participants with strabismus, binocular OCT imaging correctly revealed the type and direction of the deviation and had fair agreement in misalignment size with the alternating prism cover test.

These data suggest binocular anterior segment OCT imaging can provide measurements of strabismus with fair agreement to alternating prism cover test, and thus could be useful for diagnosis and monitoring.

Access the full article here

PMID: 29852032

Human Factor and Usability Testing of a Binocular Optical Coherence Tomography System. Chopra R, Mulholland PJ, Dubis AM, Anderson RS, Keane PA.
Transl Vis Sci Technol. 2017 Aug 15;6(4):16. doi: 10.1167/tvst.6.4.16. eCollection 2017 Jul.

The purpose of this study was to perform usability testing of a binocular optical coherence tomography (OCT) prototype to predict its function in a clinical setting, and to identify any potential user errors, especially in an elderly and visually impaired population.

Forty-five participants with chronic eye disease (mean age 62.7 years) and 15 healthy controls (mean age 53 years) underwent automated eye examination using the prototype. Examination included 'whole-eye' OCT, ocular motility, visual acuity measurement, perimetry, and pupillometry. Interviews were conducted to assess the subjective appeal and ease of use for this cohort of first-time users.

All participants completed the full suite of tests. Eighty-one percent of the chronic eye disease group, and 79% of healthy controls, found the prototype easier to use than common technologies, such as smartphones. Overall, 86% described the device to be appealing for use in a clinical setting. There was no statistically significant difference in the total time taken to complete the examination between participants with chronic eye disease (median 702 seconds) and healthy volunteers (median 637 seconds) (P = 0.81).

Access the full article here

PMID: 28824827 

This work is supported by Moorfields Eye Charity and the Miss Barbara Mary Wilmot Deceased Discretionary Trust.