Abstract
Purpose :
Virtual reality (VR) headsets are an attractive emerging option for conducting visual field tests. As opposed to traditional tests, VR systems offer an alternative that potentially decreases human effort, time, and cost while offering accurate and repeatable results. Here, we validate the application of a VR system that uses only eye telemetry data to provide visual field analysis comparable to traditional tests that utilize a hand clicker.
Methods :
We prototyped and conducted a series of visual field exams using VS 2020 (Twenty/Twenty Therapeutics, LLC), an eye tracking-based ophthalmic software device running on a VR headset that presents a series of time and response-dependent stimuli while capturing eye telemetry data. Stimuli are presented in a 24-2 central pattern within the headset field of view following a proprietary strategy relying only on the user’s eye response to the presented stimuli without additional user input. The responses are collected to construct sensitivity maps which are later normalized considering an internal age-normative transformation to create total deviation maps. The average of the sensitivity map is used to generate a mean sensitivity value (MS) and the total deviation map is weighted to generate a mean deviation value (MD). Visual field exams from the same subjects were also collected from a Humphrey Visual Field Analyser (HFA; Carl Zeiss Meditec, Inc.) using a 24-2 central pattern and SITA Fast algorithm in order to compare the VS 2020 MS and MD results to those produced by a traditional test. Repeated tests in both the VS 2020 and HFA were also collected to analyze the repeatability of the exams.
Results :
Data collected from 34 eyes of 34 subjects (ages 21-84, 6 normal, 28 abnormal) in routine clinical care were analyzed. Analysis of Pearson correlation between VS 2020 and HFA values yielded values of 0.87 for MS (p<0.001) and 0.86 for MD (p<0.001). Bland-Altman analysis of the repeated eye exams yielded standard deviation values of 2.6 dB and 2.8 dB for MS and MD respectively for VS 2020 data and of 1.0 dB and 1.0 dB for MS and MD respectively for HFA data.
Conclusions :
This work validates the use of a VR system to conduct prototype visual field exams relying exclusively on eye telemetry data. The system performs reliable measurements with significant correlation and high repeatability when compared to traditional tests.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.