Eyenuk, Inc. - Artificial Intelligence Eye Screening

Tech Brief

The EyeArt System is a clinical game-changer in the diagnosis of DR as it eliminates the need for an eye care specialist to conduct DR screening, thereby vastly increasing access to screening and saving the vision of potentially millions of people across the globe from preventable loss due to DR. EyeArt consistently outperforms competitors in head-to-head studies and is currently in use by over 100 organizations around the world improving health equity in previously hard-to-reach populations. Moreover, EyeArt allows eyecare specialists to spend less time on screening and provide more availability to handle serious cases. That’s because only patients who have a positive DR screening through EyeArt are referred for care, ensuring that ophthalmologists’ time and expertise is optimized for assessing and treating the most complex cases (ophthalmologists strongly prefer this approach since they receive more compensation for treatment than for diagnosis). In recognition of Eyenuk’s success, on June 8 Juniper Research announced that Eyenuk had earned its Future Digital Health Innovation Award for the “Best Digital Diagnostics Solution.” The awards are presented to companies that deliver imaginative, innovative products or services that have the potential to disrupt their ecosystems and provide significant benefits to their target audience.

Problem Tech Solves

EyeArt fills a substantial gap in the care for people with diabetes by enabling them to be screened for diabetic retinopathy (DR) – the leading cause of blindness in adults aged 20-74 – during a regular primary care office visit rather than requiring a separate visit to an eye-care specialist. DR affects 1 in 3 people with diabetes. All people with diabetes are at risk of developing DR, and the longer a person has had diabetes, the greater the risk that they will develop DR. Since the vision loss from DR can typically be prevented, the American Diabetes Association and the American Academy of Ophthalmology have established guidelines recommending all people with diabetes have an annual exam to enable early detection of DR. However, only about 60% of Americans with diabetes receive their annual eye exam and preventable blindness occurs in thousands of people every year. Currently, a standard diabetic eye exam for DR screening typically requires an eye care specialist to dilate a patient’s eyes and examine the back of the eyes, which is expensive, requires long waits for appointments and keeps specialists from focusing on patients needing treatment (vs. diagnosis alone). The EyeArt system addresses these problems by enabling non-specialists in the primary care setting to identify patients requiring referral to an eye care specialist, using fully autonomous, real-time screening of DR during a patient’s regular office visit, which increases accessibility and health equity, improves efficiency and accuracy and reduces costs.

Validation

The EyeArt AI System has been prospectively validated in a pivotal clinical trial with 942 participants at 15 U.S. clinics. The EyeArt system was shown to achieve high accuracy (mtmDR sensitivity: 96%, specificity: 88% and vtDR sensitivity: 97%, specificity: 90%) against the rigorous Early Treatment Diabetic Retinopathy Study reference standard determined by standardized, adjudicated grading by experts at the Wisconsin Reading Center. It successfully graded more than 97% of the eyes scored manually, with most not requiring dilation. The strong real-world clinical performance was the basis for the EyeArt system’s FDA clearance. Also, the EyeArt System has been validated in the largest AI in healthcare study on >100,000 patient visits and by the UK National Health Service (NHS) in the largest prospective AI study on >30,000 patients. A head-to-head study of seven AI DR screening systems on 23,724 veterans at the Veterans Affairs Puget Sound Health and Atlanta Health Care Systems concluded that the EyeArt system (Algorithm G) was the only algorithm that was statistically indistinguishable from the standard of care and enabled the highest amount of cost savings. The comprehensive review of AI technologies was issued by the UK National Screening Committee, which provides advice about all aspects of population screening to the NHS and the ministers of the four UK countries. In the study, the EyeArt System was confirmed to have “sufficient high quality evidence” as “safe and cost-effective,” and is the only diabetic eye screening AI technology ready for live clinical implementation in the United Kingdom.