Brief plain language background
Visual information is detected by the retina at the back of the eye, which travels to the brain. This pathway can be affected by multiple disorders and cause severe sight loss. Collectively, they are called retinal diseases.
Early diagnosis and timely treatment have been identified by those affected by retinal disease as key priority areas.
Novel computer-assisted analyses such as artificial intelligence (AI), allows for detailed interrogation of functional data and can be used to improve healthcare.
What problem/knowledge gap does it help address
Diagnosis of retinal disease is aided by measurements of retinal function, such as an electroretinogram (ERG). Interpretation is required by highly specialised clinicians in dedicated centres.
Lack of time available for analysis – or correlations not being visible even if they provide insight – can make it difficult for clinicians to use the data generated by these measurements.
The team previously showed that AI could analyse retinal function in patients with Stargardt disease with comparable accuracy to experts.
Aim of the research project
To refine the piloted AI process to interpret data from an extensive, diverse and representative ERG dataset.
Key procedures/Objectives (in laymen terms)
- Develop software to efficiently extract data from the Moorfields Eye Hospital database – over 60,000 ERGs recorded over 20 years from a rich and representative population.
- Refine and enhance the previously developed AI analyses.
- Use the computer-based procedures to interpret and examine the data for information relevant to diagnosis and clinical management.
Potential impact on people with sight loss
The AI tool developed by this project would not replace the doctor, instead enhancing their diagnostic and therapeutic ability. The tool may facilitate interpretation of ERG data outside highly specialised centres, promoting prompt and consistent evaluation and reduced waiting times. Earlier and more precise diagnosis in more patients could enable earlier treatment, potentially increasing its success.