Abstract
Purpose :
The Radius is a lightweight virtual reality headset that performs automated perimetry. Self-administered visual field (VF) tests such as the Radius have the potential to significantly improve the ability to monitor glaucoma progression by enabling more frequent measurements of VFs in a patient’s home. In this preliminary non-randomized comparative case series, we compare estimated sensitivities of the Radius to those of the Humphrey Field Analyzer (HFA) on the 24-2 test pattern for 27 glaucoma and glaucoma suspect eyes. The Radius presents the Goldmann III stimulus on a 10 cd/m2 background, the same as the HFA.
Methods :
24-2 VFs were measured on 27 eyes from 19 patients with glaucoma or glaucoma suspect status at the North Bay Eye Associates using both the HFA (SITA-Standard algorithm) and the Radius (RATA Standard algorithm). Distributions of estimated sensitivities in the 15-40 dB range — the Radius has a dynamic range of 15-40 dB — were compared across all 54 test locations using the Wilcoxon rank sum test with Bonferroni correction (). Deming regression was used to estimate the slope of the best-fitting line between estimated sensitivities. Overall test times were also recorded
Results :
Across 54 test locations, the smallest p-value for the Wilcoxon rank sum test was p = 0.0026, which is non-significant given the Bonferroni corrected . Figure 1 shows that the slope of the Deming regression line in the 15-40 dB range was 0.62 (the size of the circles in Figure 1 is proportional to the number of data points at that location). The relation is linear, but the slope of 0.62 demonstrates reduced dynamic range for the Radius within the HFA 15-40 dB range. Average test times for the Radius were less than for the HFA (see Figure 2): 5 minutes 36 seconds for the Radius and 6 minutes 28 seconds for the HFA.
Conclusions :
This preliminary study demonstrates statistical non-inferiority of estimated sensitivities from the Radius compared to those of the HFA for the 24-2 test pattern within the 15-40 dB range.
This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.