From FDA clearances to peer review: The published science behind AEYE-DS
At AEYE Health, our mission is simple: eliminate preventable blindness by bringing autonomous AI diagnostic screening to every point of care. Today, we are proud to share that the clinical trials data behind our groundbreaking FDA approvals has officially been published in Frontiers in Digital Health.
Spanning 1,210 patients across 13 clinical sites, the published data from three pivotal trials (AEYE-1, AEYE-2, and AEYE-3) details the rigorous data that supported our FDA clearances. The accepted gold standard was established by comparing AEYE-DS results against a multi-expert reading center using a comprehensive multi-model data for ground truth (including dilated exam, OCT and anterior imaging). Every FDA endpoint was cleared with margin, using images captured entirely by novice operators - clinic staff with no prior ocular imaging experience.
Key metrics at a glance:
Only portable solution: AEYE-DS is the only FDA-cleared autonomous AI for a portable handheld camera. The exam goes where the patient is, in the room, at a satellite site, or in a mobile clinic, so no patient is left behind. Clinical sites can mix both portable and tabletop cameras when using AEYE-DS.
Single-Image & Non-Dilated Efficiency: As the only FDA-cleared solution requiring just one non-dilated image per eye, AEYE-DS significantly reduces screening complexity to under a minute, saving time, and improving patient comfort.
1 minute screening: the entire screening takes 1 minute, from capture to instant result, fitting seamlessly into primary care workflows.
Groundbreaking Diagnostic Accuracy: Demonstrated outstanding sensitivity (92%-93% for the portable solution, 93% for tabletop) and specificity (89%-94% for the portable solution, 91.4% for tabletop) for detecting more-than-mild diabetic retinopathy (mtmDR).
100% Sensitivity for Severe Cases: Achieved a 100% sensitivity rate for ETDRS ≥ 43, including all severe diabetic retinopathy cases across all 3 clinical trials, ensuring critical high-risk cases are never missed.
>99% Imageability: Successfully analyzed >99% of patient images across all three pivotal trials, delivering instant diagnostic results and virtually eliminating ungradable image rejections.
99% Intra-Operator Repeatability: Proven reliable when operated by general clinic staff rather than eye care specialists, ensuring effortless 1-minute screening integration into daily workflows.
The only FDA-cleared autonomous AI for portable imaging and only 1-image-per-eye technology
Workflow speed and hardware flexibility are vital for busy primary care practices and rural health clinics. The study validates equivalent diagnostic performance across both portable handheld and tabletop cameras, allowing clinics to select hardware based on room layout and patient volume.
Metric | Handheld camera (Aurora) | Desktop camera (Topcon NW400) |
Form factor | Portable & lightweight | Fully robotic tabletop |
Sensitivity | 92% – 93% 100.0% sensitivity for severe cases (ETDRS ≥ 43) | 93% 100.0% sensitivity for severe cases (ETDRS ≥ 43) |
Specificity | 89% – 94% | 91.4% |
Imageability rate | >99% | >99% |
Images requires | Just 1 image per eye | Just 1 image per eye |
By validating the market's only portable handheld autonomous AI solution, AEYE Health gives healthcare organizations the flexibility to select the hardware that best fits their room layouts, clinic mobility needs, and patient volume. Combined with superior usability with just 1 image per eye, no dilation, and 1-minute screening anywhere, clinics can effortlessly close diabetic eye care gaps without adding workflow overhead.
Solving the ungradable image bottleneck
Diabetic retinopathy remains the leading cause of blindness among working-age adults. Timely diagnosis and treatment can prevent more than 95% of vision loss cases, yet over 60% of patients miss their annual eye exams due to specialist shortages and referral hurdles.
Tele-retinal and remote AI screening were built to fix that, yet they moved the bottleneck instead. When a meaningful share of captured images come back ungradable, and there’s no way to complete the procedure and notify the patient on the spot, the clinic is back where it started, chasing the patient for a repeat image or the specialist visit it was trying to avoid.
Our pivotal trial data proves that AEYE-DS solves this bottleneck. By achieving >99% imageability, AEYE-DS ensures that virtually every patient who gets screened leaves their primary care visit with an instant, actionable diagnostic result on the spot.
Transforming care gap closure into a standard of care
With zero physician overread required, seamless EHR integration, and established reimbursement pathways via CPT code 92229, AEYE-DS turns diabetic eye screening from a frequently missed referral into a frictionless, financially sustainable standard of care.You can read the full paper in Frontiers in Digital Health here. And if you’re ready to eliminate care gaps in your practice, you’re welcome to speak with our product experts.
Resources
Dvey-Aharon Z, et al. "Autonomous AI-driven point-of-care screening for diabetic retinopathy compared to reading center multi-expert clinical review: results from three prospective controlled pivotal validation studies with AEYE-DS in over 1,200 patients." Frontiers in Digital Health, vol. 8, 2026, art. 1907804. Open access.
