Visual field testing in a portable virtual reality environment
Z HOT TOPIC Z
Author Disclosure Block: R.B. Shi: None. Y.M. Buys: None. T.L. Fee: None. Y.J. Mahsood: None. G.E. Trope: None. M. Eizenman: None.
Abstract Body:
Purpose: Standard automated perimetry is widely used for detection and monitoring of visual field (VF) defects such as in glaucoma. Current devices like the Humphrey Field Analyzer (HFA) are expensive, have limited portability and require a testing facility with technical support. Using a standard smartphone (Huawei Nexus 6P Google 32GB Model H1511 android smartphone), virtual reality viewer (VR, google cardboard second generation) and Bluetooth controller (clicker) we have created a head-mounted mobile perimeter (MP). In this study we compared the performance of the MP to the HFA.
Study Design: Prospective cross-over study.
Methods: 19 consenting subjects from the Toronto Western Hospital glaucoma unit, with varying degrees of glaucomatous VF loss were enrolled. Each subject underwent a HFA 24-2 SITA standard followed by the MP. In addition seven subjects underwent a second MP to test for inter-test variability. For each subject one eye was tested.
Results: The mean age of the subjects was 64.9±9.1 years. Test duration was longer with MP (8.6±1.0 mins) compared to HFA (5.7±1.1 mins, p<0.001). The reliability indices were similar for false positives and negatives however there were significantly more fixation losses with the MP (22% versus 7.2%, p=0.01). We found a good correlation between the HFA and MP for threshold (R2=0.64), and excellent agreement for MD (R2=0.87), PSD (R2=0.89) and VFI (R2=0.92). Overall there was a lower sensitivity with MP and higher variance at the lower thresholds. In the 7 subjects who repeated the MP there was no significant difference in test duration or reliability indices. There was a good agreement between the tests for threshold (R2=0.61) and excellent agreement for MD (R2=0.98), PSD (R2=0.96) and VFI (R2=0.97).
Due to the limited dynamic range of the cell-phone’s display, we used larger Goldmann size targets to test VF thresholds that were below 10dB. This resulted in larger variations between the HFA and MP for such thresholds. The performance of the MP could be improved by using smartphone displays with larger dynamic range and reducing the scattering from the optical lenses of the virtual reality system.
Conclusions: MP shows promise as a complimentary device to HFA. The portability and affordability of this device could make this a valuable tool for home monitoring and screening purposes. Further development of smartphone based mobile perimeters could lead to highly cost efficient VF tests especially for subjects who are unable to visit a VF lab.
