JUL 25, 2022
The Notal Vision Home OCT (NVHO) system was evaluated for its reliability to daily detect fluid in patients with neovascular AMD, increasing the data available upon which to base better treatment decisions. The results present the success rate of patient use of the device and the agreement of scans evaluated by the device's deep learning algorithm vs human evaluators.
This was a prospective, observational longitudinal study of 15 patients with neovascular AMD undergoing anti-VEGF treatments in at least 1 eye at 2 centers. Patients were instructed to perform daily home self-imaging using the Notal Vision Home OCT (NVHO) for 3 months. The macular 3 x 3 mm2 cube scans were uploaded automatically to the Notal Health Cloud. The images were evaluated by both the Notal OCT Analyzer (NOA), an automated deep learning-based algorithm, and human expert graders. Clinically trained staff at a remote monitoring center were available for assistance throughout the study period to support participants who encountered device or operational issues. Patients were excluded if they had poorly dilated pupils and/or significant media opacity that interfered with OCT imaging.
A total of 39 eyes were included in the study, of which 5 (17%) had intermediate AMD and 24 (83%) had neovascular AMD. Each patient completed a mean of 5.7 ± 0.9 scans per week. There were 82 ± 24 successful scans in 86 ± 25 scan attempts per study eye, with a mean success rate of 95% ± 7%. A total of 317 NVHO scans were both manually graded by the investigators and automatically graded using NOA. There was agreement between the NOA and retinal specialists on the fluid status in 262 (83%) scans and disagreement for 55 (17%) scans, all of which involved subretinal fluid. Of the scans with disagreement, 48 (15%) had fluid detected by the retina specialist but not by the NOA. The remaining 7 (2%) scans had fluid detected using the NOA but not by the retina specialists, with 4 exhibiting subretinal fluid and 3 with intraretinal fluid.
The study has a small sample size and is of relatively short duration.
This is an early study of home OCT, demonstrating the potential of self-imaging and an automated deep leaning algorithm to detect fluid on OCTs. Further larger and longer-term studies are needed to better assess the clinical utility of the device.