The EyeStar AI system screens patients for diabetic retinopathy and macular edema, conditions that lead to vision loss and blindness if left untreated. VisionQuest brought the EyeStar AI system to Clínicas del Azúcar in 2016, and since then the software has identified over 4,000 patients with sight-threatening disease who would not otherwise have been referred for care by a specialist in a timely manner.
The screening procedure at Clínicas del Azúcar is straightforward and quick: A patient arrives at the clinic and has their vital measurements recorded. A nurse or technician takes retinal photographs (without needing to dilate the patient’s pupils) using a tabletop or handheld camera and uploads them to the cloud, where the user-friendly EyeStar system applies a deep-learning classifier to determine whether the patient needs to be referred to an ophthalmologist. EyeStar’s analysis takes less than a minute and allows physicians at the clinics to manage their patients with efficiency and care and to refer them to a specialist when needed.
VisionQuest’s quality assurance program shows that EyeStar can provide a result in 98 percent of cases due to the software’s resilience to variations in image quality. EyeStar was developed with funding from the National Eye Institute. The system uses state-of-the-art deep-learning algorithms trained on hundreds of thousands of images from VisionQuest’s proprietary dataset to achieve over 90 percent sensitivity in the detection of diabetic retinopathy. Providers refer only patients who are found to have severe disease for dilated-eye exams and possible treatment, which makes the best use of scarce eye-care resources in Mexico. EyeStar thus has a major positive impact on clinical efficiency and health-care access, as well as improving outcomes for patients who might otherwise be reluctant to return for further care.
Read the full press release here.