An artificial intelligence–powered optical coherence tomography (OCT) device for home use is reliable, easy to use, and has imaging capabilities comparable to in-office OCT scans, according to data from phase 3 trials presented Friday at the 2024 American Academy of Ophthalmology annual meeting. The Scanly Home OCT device (Notal Vision) enables physicians to monitor progression of neovascular age-related macular degeneration (nAMD) between office visits, offering a new tool for personalized care, said Eric W. Schneider, MD, who presented the findings during the late-breaking developments session of Friday’s Retina Subspecialty program.

The Scanly device captures spectral-domain OCT images in a 10°x10° (3mm x 3mm) area centered on the point of fixation. It operates in conjunction with the Notal OCT Analyzer (NOA), an AI-based program that segments and estimates the volume of hyporeflective spaces on OCT images. In May, the US Food and Drug Administration (FDA) granted de novo marketing authorization for the Scanly device, which allows retina doctors to prescribe a home monitoring program to keep tabs on disease progression between nAMD patients’ regularly scheduled office visits (Figure 1).

The FDA required 2 separate pivotal phase 3 trials to gather specific information about the effectiveness of the device, explained Dr. Schneider. For the qualitative visualization study (NCT04907409), enrolled patients were mailed a home OCT device, which they had to set up themselves with guidance from a video tutorial and help from a monitoring center if needed, and were instructed to perform daily scans for 5 weeks. Patients were also required to visit the clinic at week 1 and week 5 for an in-office OCT scan with a Zeiss Cirrus HD-OCT machine. The study’s objectives were to compare home OCT images with in-office scans for visualization of hyporeflective spaces, with a prespecified endpoint of >80% positive and negative percent agreement, as well as to evaluate the ability of nAMD patients to use the machine as prescribed.

“Reading center graders compared the home OCT with the gold standard at those 2 time points, and it showed excellent comparison,” said Dr. Schneider. “The positive percent agreement was 86% and the negative agreement was 87%, so those were above the 80% endpoint and demonstrate good reliability.” He also noted that 96.1% of enrolled patients were able to calibrate and use the device, with patients averaging 5.9 scans per week, so the adherence rate was good (Figure 2).

The purpose of the second trial (NCT05202587) was to compare the Scanly device’s AI algorithm to human graders in quantifying hyporeflective spaces, which often represent retinal fluid, explained Dr. Schneider. Patients came to the office and were scanned multiple times with 2 different Scanly devices, to ensure repeatability and reproducibility of results. They also had a standard in-office OCT scan for comparison. The coefficient of variation for the home OCT scans with AI grading was 11.1%, compared to 16.4% for in-office OCT with multiple reading center graders, Dr. Schneider reported. “Both device-grader combinations were precise, but there’s actually lower variability in the home OCT group,” he explained. “This level of repeatability should allow providers to find clinically meaningful trends in fluid with daily or near-daily imaging.”

As with any new diagnostic information, retinal physicians will have to figure out how it can be incorporated into their practice, Dr. Schneider said. Regular monitoring between visits can help doctors understand how patients are reacting to treatment and adjust the interval between intravitreal anti-VEGF injections accordingly. Because both eyes are scanned, there is also the potential for early identification of a fellow eye with dry AMD that begins neovascular conversion, he noted. “I think we’ll learn a lot about our patients and their specific disease that will allow a much more personalized approach to treatment,” he said. RP