Teleretinal Screening and AI in Retinal Disease Detection
How Teleretinal Screening Works
Teleretinal screening begins with capturing high-quality photographs of the retina using specialized retinal cameras. These cameras can be placed in primary care offices, endocrinology clinics, community health centers, pharmacies, and other settings where patients already receive care. Many of the cameras used for screening can capture images through undilated pupils, which means that patients do not need to have their eyes dilated for the screening procedure. The imaging process is quick, typically taking only a few minutes, and is noninvasive. A technician or trained staff member positions the patient at the camera, takes photographs of each retina, and the images are then either transmitted for remote review or analyzed on-site by an AI system.
In the traditional teleretinal model, the retinal images captured at the screening site are transmitted electronically to a reading center where an eye care specialist, typically an ophthalmologist or retina specialist, reviews them. The specialist examines the images for signs of retinal disease, including hemorrhages, exudates, microaneurysms, neovascularization, drusen, and other abnormalities. Based on the review, the specialist generates a report indicating whether the images are normal, whether abnormalities are present, and whether the patient should be referred for an in-person examination. This model effectively extends the reach of retinal specialists to locations where they are not physically present, bringing screening capability to patients in underserved and rural areas.
AI-based screening systems represent an advance beyond the traditional teleretinal model by providing automated, on-site analysis of retinal images without requiring a remote specialist to review each image. These systems use deep learning algorithms that have been trained on large datasets of retinal photographs to recognize patterns associated with specific diseases. Three AI devices have received FDA clearance for autonomous diabetic retinopathy screening in the United States: LumineticsCore (formerly IDx-DR), EyeArt, and AEYE Diagnostic Screening (Retina Specialist, 2025). These systems analyze the retinal images at the point of care and provide a result within minutes, indicating whether the patient has signs of disease that warrant referral to an eye specialist or whether no referral-warranting disease was detected.
What AI Screening Can Detect
The primary application of FDA-cleared AI screening systems is the detection of diabetic retinopathy. These systems are designed to identify patients with more than mild diabetic retinopathy who should be referred for a comprehensive eye examination. In the pivotal trial of the IDx-DR system, the technology achieved a sensitivity of 87.4 percent and specificity of 89.5 percent for detecting referable diabetic retinopathy, meaning that it correctly identified the large majority of patients who needed referral while maintaining a low false-positive rate (NPJ Digital Medicine, 2018). This level of performance has been validated in real-world clinical settings, and the integration of AI screening into diabetes care workflows has been shown to increase screening rates among patients with diabetes.
While the current FDA-cleared AI systems are authorized specifically for diabetic retinopathy screening, research is actively expanding the application of AI to other retinal conditions. AI algorithms are being developed and validated for the detection of age-related macular degeneration, glaucoma, retinal vein occlusion, and other conditions visible on retinal photographs. These expanded applications have the potential to transform retinal screening from a disease-specific test into a broader assessment of retinal health that could identify multiple conditions from a single set of images. As these algorithms are validated and receive regulatory clearance, the scope of AI-based retinal screening is expected to continue growing.
While AI screening systems are valuable tools for increasing access to retinal disease detection, they have important limitations that patients should understand. Current FDA-cleared systems are designed to screen for specific conditions at specific severity thresholds and do not provide a comprehensive eye examination. They may not detect all retinal diseases, particularly conditions that do not produce the specific image features the algorithm was trained to recognize. Image quality can affect the accuracy of the analysis, and in some cases the system may produce an ungradable result that requires the patient to have a traditional eye examination. AI screening is a screening tool, not a diagnostic tool, and patients who receive a positive result require follow-up with an eye care specialist for definitive diagnosis and treatment planning.
The Patient Experience
The teleretinal or AI screening experience is designed to be quick and convenient. When you arrive at the screening site, a technician will seat you at the retinal camera and ask you to look at a target inside the camera while photographs are taken of each eye. The camera uses a bright flash of light to illuminate the retina, which may cause a brief afterimage similar to having a photograph taken with a flash. The entire imaging process typically takes less than five minutes. If the site uses an AI system, you may receive your result the same day, while teleretinal sites that use remote specialist review may provide results within a few days. No drops, injections, or contact with the eye surface are involved in the screening process, and there is no discomfort associated with the imaging.
The results from a teleretinal or AI screening are typically straightforward. You will receive one of three types of results: a result indicating that no signs of referral-warranting disease were detected, a result indicating that signs of disease were found and you should be referred to an eye specialist for further evaluation, or a result indicating that the images could not be adequately analyzed and you should have a traditional eye examination. A negative screening result does not mean that you do not need regular eye care; it means that at the time of screening, the specific disease being tested for was not detected at a referral-warranting level. You should continue to follow your recommended schedule for comprehensive eye examinations.
How Screening Fits into Retinal Care
Teleretinal screening and AI-based screening are designed to complement, not replace, comprehensive eye examinations by an eye care provider. A screening test evaluates specific aspects of retinal health and is targeted at detecting specific conditions, while a comprehensive eye examination includes assessment of visual acuity, eye pressure, anterior segment health, peripheral retinal examination, and other components that screening alone does not address. Patients who have known retinal conditions or who are under active treatment should continue their regular appointments with their retina specialist. Screening is most valuable for patients who need regular retinal assessment but face barriers to accessing specialized eye care.
One of the most significant benefits of teleretinal screening and AI is the ability to bring retinal disease detection to patients who might not otherwise receive timely screening. Many patients with diabetes, for example, do not complete the recommended annual eye examination due to barriers such as distance from an eye care provider, lack of transportation, time constraints, or lack of awareness about the importance of screening. By placing screening capability in primary care offices and community clinics where these patients already receive care, teleretinal and AI screening can identify retinal disease at earlier stages and facilitate referral before vision loss occurs. This approach is particularly valuable in rural and underserved communities where retina specialists may not be readily available.
Retinal screening technology continues to advance rapidly. AI algorithms are being developed to detect an expanding range of retinal conditions with increasing accuracy. Home-based retinal monitoring devices, including portable OCT systems, are in development and may eventually allow patients with known retinal conditions to monitor their disease between clinic visits. Integration of AI screening data with electronic health records is improving the coordination between screening and referral. As these technologies mature and are validated across diverse populations, they are expected to play an increasingly important role in the early detection and management of retinal disease, complementing the clinical expertise of retina specialists and eye care providers.
Questions and Answers
AI screening cannot replace comprehensive eye examinations. While AI screening is effective at detecting specific conditions such as diabetic retinopathy, it does not assess all aspects of eye health. A comprehensive eye examination includes evaluation of your vision, eye pressure, lens clarity, and the full peripheral retina, none of which are assessed by current AI screening systems. Think of AI screening as an additional checkpoint that helps ensure retinal diseases are detected between your regular eye examinations, particularly if you are at risk for conditions like diabetic retinopathy. You should continue to see your eye care provider at the intervals they recommend.
FDA-cleared AI screening systems have demonstrated high levels of accuracy in clinical trials and real-world validation studies. These systems have been shown to perform comparably to trained human readers in detecting referable diabetic retinopathy. However, like any screening test, AI systems are not perfect. They may occasionally miss cases of disease or produce false-positive results. The systems are designed to prioritize sensitivity, meaning they aim to identify as many patients with disease as possible, even if this means that some patients without disease are also referred for follow-up. Patients who receive a positive screening result should follow up with an eye specialist for definitive evaluation.
AI retinal screening is increasingly available at primary care offices, endocrinology clinics, diabetes care centers, and community health facilities. Your primary care physician or endocrinologist may offer AI retinal screening as part of your diabetes care or can direct you to a nearby location that provides this service. Not all healthcare facilities have AI screening capability, as the technology is still being deployed across the healthcare system. If you are unsure whether AI screening is available in your area, ask your healthcare provider about screening locations that may be accessible to you.