AI Eye screening
Overview
Artificial Intelligence has been used to match the patient’s photo to a data base of photo’s and to grade the changes according to an international grading system. AI does not replace human oversight, and is intended to facilitate screening on a wider scale. The primary objective of AI screening is to identify ‘normal’. If the AI advises further review this does not necessarily mean there is a problem, it simply means the AI needs human intervention to make the correct diagnosis and grading.
Benefits
The ratio of ophthalmologists to patients is such that no country in the world has been able to offer an annual specialist eye exam for every diabetic patient. For every 100 screening examinations, 30 are expected to require review, of those only ten are likely to require treatment by an ophthalmologist. By using AI technology all 100 patients can be screened through the network, 30 will be reviewed by an eye care professional and 10 will be referred to an ophthalmologist. This has the advantage of improving access to a screening network whilst simultaneously improving access to specialist care for those who need treatment.
Technology
The Primary Tier Screening network makes use of non-mydriatic cameras.
The Second Tier Evaluation Clinic makes use of wide-field non-mydriatic cameras and has access to further specialised technology, such as OCT and OCT Angiography.
Screening Sites
Run under the direct control of either a Medical Doctor or an Optometrist, who are independent service providers with an interest in diabetic retinopathy screening and who collectively make up the Screening Network, whom the Retinography Team aim to support.