Can At-Home Vision Tests Detect Active Neovascular AMD?

Three at-home vision-monitoring tools failed to detect worsening neovascular age-related macular degeneration (AMD), a diagnostic-test accuracy study reported.

The paper-based KeepSight Journal from the International Macular and Retinal Foundation, Genentech's MyVisionTrack app, and Visumetrics' MultiBit app failed to identify active disease among 259 patients under treatment for neovascular AMD, with an estimated area under the receiver operating characteristic (ROC) curve less than 0.6 for all three, according to researchers led by Ruth Hogg, PhD, of Queen's University Belfast in Ireland.

"In this context, I would not be confident on relying on any of these tests," Hogg told MedPage Today. The paper in JAMA Ophthalmology reiterated that message, saying that "implementing any of these evaluated tests, with ophthalmologists only reviewing test positives, would mean most active lesions were missed, risking unnecessary sight loss."

Neovascular AMD is the most common cause of visual loss in the elderly population. It can lead to a central blind spot and legal blindness, although treatment can stop worsening of vision and even improve it in some cases.

Ideally, Home Monitoring Could Detect Worsening AMD

According to Hogg, worsening advanced AMD -- also known as wet AMD -- can be tricky to detect. "Long-term wet AMD can mean distortion is always present. Reliably picking up changes in the amount of distortion is much more difficult than picking up new-onset distortion."

She explained that the impetus for the study came from its funder, the U.K. National Institute for Health Research, which analyzes ways to improve the British healthcare system, Hogg said. "Monitoring of wet [advanced] AMD patients is causing huge pressure on service delivery, so this was considered a research priority," she said.

Ideally, she said, home-monitoring vision tools would allow patients to track their condition at home and "then only require an appointment when the home monitoring created an alert that a change had occurred. An alternative would be less frequent appointments with the reassurance that an appointment could be sooner if an alert was generated."

However, "the available literature on the tests we evaluated was really not very supportive," co-author Barnaby Reeves, PhD, of the University of Bristol in England, told MedPage Today. "There was no evidence about the performance of the tests in relation to treatment decisions for individual patients."

According to the study, "none of the tests had adequate diagnostic test accuracy for the primary outcome to enable patients to monitor their vision at home. This conclusion was unaltered by the sensitivity analyses."

Kevin Blinder, MD, of the Retina Institute in St. Louis, agreed in a commentary accompanying the study that the tools examined "should not be used or relied on for monitoring patients at this point."

Makers of Vision-Monitoring Tools Object to Study Findings

Siddarth Rathi, MD, MBA, global medical director for Remote Vision Monitoring at Roche, of which Genentech is part of, told MedPage Today that the study "provided no evidence regarding the MyVisionTrack application when used as intended."

He said the prescription-only app, an FDA-registered medical device, "allows patients to test their vision function at home through a shape discrimination hyperacuity test that can accurately identify significant changes in vision function. The web portal allows physicians and their staff to manage patients and view their test results."

This technology "helps patients and their ophthalmologists by gathering additional clinical data between visits," he added, and "is not intended to diagnose retinal diseases, which is the responsibility of the prescribing eye-care professional."

With regard to the MultiBit app, Lars Frisén, MD, PhD, of Sweden-based Visumetrics, told MedPage Today that the study misrepresents his product. "Remarkably, the report does not provide scientific references for MultiBit or its PC-based predecessors, all of which were published in peer-reviewed journals of good standing."

He highlighted a 2015 pilot study and added that "MultiBit cannot provide a diagnosis but offers a means for sensitive self-monitoring over time of the function of the retinal macula."

The International Macular and Retinal Foundation, developer of the KeepSight Journal, didn't respond to a request for comment about the study.

Study Details

The MONARCH study included 259 patients with AMD (312 eyes) over age 50 who were recruited from six U.K. hospital eye clinics from 2018 to 2020. Enrollment required neovascular AMD treatment for 6 to 42 months prior to baseline and a visual acuity no worse than 6/60 on Snellen testing. Participants mean age was 75 years, 58.6% were women, and 69% were white. Among them, 60.1% had active neovascular AMD, 37.2% had inactive disease, 2.6% had uncertain status, and 0.1% had no lesion.

The patients were taught how to use the tests and told to test weekly with all three tools. Median home-monitoring testing frequency was three times per month. The results of the tests were compared against in-hospital ophthalmologist examinations.

"Among the home-monitoring summary scores, only the KeepSight Journal worse-vision summary score showed a statistically significant association with lesion activity," the researchers reported. Patients who reported worse vision with that tool had an odds ratio of 3.48 (95% CI 1.09-11.13, P=0.04) for an active lesion.

An alternative definition of diagnostic accuracy -- change in disease classification from inactive to active between pairs of monitoring visits -- yielded similar results, "though imprecisely estimated due to the small number of pairs of visits documenting lesion reactivation," the researchers noted. "We specified this alternative reference because it represented how home-monitoring could be implemented, i.e., discharge from hospital follow-up when a patient has inactive disease and recall only when home-monitoring test results diagnose reactivation."

Limitations included smaller than planned sample size, with less than half the target number of monitoring visits and withdrawal of 33% of recruited patients. Nonetheless, the researchers noted narrow 95% confidence intervals that consistently overlapped those for no test, "ruling out the [possibility] that any test provided adequate accuracy for diagnosing active neovascular AMD to enable patients to be monitored at home."

Also, they added, "The study had no control over monitoring visits, and participants are likely to have reported their subjective visual experience to their consultants, which may have influenced the reference standard."

While the findings didn't support clinical roll-out, they "do support rigorous evaluation of such technologies in the settings and populations for which they are intended before widespread implementation," Hogg and colleagues concluded.

Comment