Virtual reality perimetry may produce visual field measurements comparable to Humphrey Visual Field testing while reducing testing time by more than half in patients with ocular hypertension or glaucoma, according to a retrospective study.
Researchers of Duke Eye Center compared the VisuALL virtual reality (VR) perimetry platform with Humphrey Visual Field (HVF) 24-2 testing in 79 eyes from 50 patients with ocular hypertension, glaucoma suspect, or glaucoma. They evaluated mean deviation, pattern standard deviation, testing duration, point-by-point sensitivity, reliability measures, and patient preference.
The cohort included eight eyes with ocular hypertension or glaucoma suspect, 24 eyes with mild glaucoma, 28 eyes with moderate glaucoma, and 19 eyes with severe glaucoma. The mean patient age was 72 years. All of the patients underwent HVF testing prior to VR testing using the American Virtual Algorithm Fast protocol. Seven eyes underwent testing on separate days, with a mean interval of 12 days between examinations.
Mean deviation measurements were similar between the devices at −6.8 dB with HVF testing and −6.0 dB with VR testing. Pattern standard deviation values also were comparable at 5.6 dB and 6.3 dB, respectively. Correlations between the two modalities reached 0.82 for mean deviation and 0.85 for pattern standard deviation.
However, the testing duration was shorter with VR perimetry, averaging 122 seconds compared with 253 seconds with HVF testing. The researchers also found that HVF testing time increased with the severity of glaucoma, rising from 197 seconds in patients with mild glaucoma to 317 seconds in patients with severe glaucoma. VR testing duration remained relatively stable across disease stages, ranging from 121 to 120 seconds between mild and severe glaucoma, respectively.
Point-by-point analysis did not identify systematic sensitivity differences between devices. Mean sensitivity differences ranged from −0.4 dB to +3.1 dB across the tested visual field locations, and no individual point showed a statistically significant difference between modalities. Variability between platforms increased in patients with moderate and severe glaucoma; however, the mean threshold sensitivity remained similar overall.
The researchers also assessed reliability metrics. Among 18 eyes with fixation losses greater than 20% during HVF testing, fixation loss rates were lower with VR perimetry. The researchers attributed this finding to continuous eye tracking and the device’s dynamic testing matrix. False-positive rates also were lower with VR testing among the eyes with lower fixation losses on HVF testing.
A large majority of the patients preferred the VR platform. For instance, among the 48 patients who completed both tests, 41 of them preferred VR perimetry and 7 reported no preference. None of the patients preferred HVF testing.
The researchers noted several study limitations, including the retrospective design, modest sample size, and single-center setting. Because HVF testing was always performed prior to VR testing, they could not exclude learning or fatigue effects. The researchers also cautioned that the devices used different luminance ranges and testing algorithms, limiting direct equivalence between threshold sensitivity values. In addition, they did not evaluate longitudinal monitoring or test-retest repeatability.
“VR perimetry may be used in glaucoma management, particularly in circumstances where access to a HVF analyzer is challenging either due to physical or operational constraints," wrote lead study author Peter J. Weng, of Duke Eye Center at Duke University, and colleagues. "[G]laucoma management is highly dependent on the ability to monitor disease progression. As such, future studies that evaluate how these VR perimetry track visual field changes over longer periods of time will be a critical in determining whether VR perimetry can serve as an effective adjunct to glaucoma monitoring and management.
The study authors reported no conflicts of interest and no funding.
Source: Clinical Ophthalmology
