Eyenuk has received US Federal Drug Administration (FDA) clearance to market EyeArt, an autonomous artificial intelligence (AI) system to detect diabetic retinopathy.
The EyeArt system is designed for use by healthcare providers to detect more than mild diabetic retinopathy (mtmDR) and vision-threatening diabetic retinopathy (vtDR) in adults diagnosed with diabetes who haven’t already been diagnosed with more than mild DR.
EyeArt’s FDA clearance was based on analysis of EyeArt’s performance in a pre-registered, prospective, multicentre, pivotal clinical trial involving 655 participants (of which 380 were enrolled at seven primary care sites and 275 were enrolled at four ophthalmology sites). Participants underwent two-field retinal photography and dilated four-widefield stereo retinal photography for the reference standard, which was determined by expert graders at the Fundus Photography Reading Centre (FPRC), according to the Early Treatment for DR Study severity (ETDRS) scale.
The analysis demonstrated that EyeArt achieved 96% sensitivity and 88% specificity for detecting mtmDR; 92% sensitivity and 94% specificity for detecting vtDR; and all eyes with ETDRS level 43 or higher were correctly identified as mtmDR positive.
The study also showed that operators can be trained to obtain medical-grade retinal images. Even at primary care centres, where most operators did not have prior ophthalmic imaging experience, EyeArt had an imageability of 97% (ie. disease detection results were provided for 97% of eyes that received a reading centre grading), with the majority of participants’ eyes (90%) receiving disease detection results without the need for dilation.
