Angiogenesis 2025: AI-driven home OCT insights with Anat Loewenstein, MD

Anat Loewenstein, MD, shared key takeaways from a cross-sectional pivotal home OCT trial that compared the performance of AI-enabled algorithms with human expert graders in neovascular age-related macular degeneration during her presentation at the 22nd Annual Angiogenesis, Exudation, and Degeneration Meeting, held virtually on February 8, 2025. The presentation emphasized the utility of home OCT in capturing regular retinal images by patients, alongside the challenge of managing large volumes of data generated.

The study¹ focused on two key measures: repeatability (the consistency of results from repeated scans) and agreement between AI algorithms and expert human graders. The results showed that home OCT demonstrated higher repeatability than in-office OCT, with a coefficient of variation of 11.1% for home OCT compared with 16.4% for in-office OCT. The study also revealed excellent agreement between AI segmentation of hyperreflective areas and human graders, as evidenced by a Dice similarity score similar to inter-grader variability.

In terms of clinical implications, the study suggested that AI-enhanced home OCT can provide valuable insights into disease dynamics, including volume estimation of hyperreflective spaces and projections of retinal volume changes over time. A Monte Carlo simulation further supported the reliability of AI in detecting variations in disease progression, with only a slight delay in notifications, within a two-day window.

Overall, the study demonstrated that combining home OCT with AI provides a high level of precision and reliability, with clinical impact sufficient for the management of neovascular AMD. The data from this pivotal trial supports its potential in enhancing patient care and disease monitoring for macular degeneration.

Reference

1. Schneider EW, Heier JS, Holekamp NM, et al. Pivotal trial toward effectiveness of self-administered OCT in neovascular age-related macular degeneration. Report 2-Artificial Intelligence Analytics. Ophthalmol Sci. 2024;5(2):100662. Published November 26, 2024. doi:10.1016/j.xops.2024.100662