Purchase this article with an account.
Lukas Mees, Swati Upadhyaya, Pavan Kumar, Sandal Kotowala, Shankar Haran, Shruthi R, David S Friedman, Rengaraj Venkatesh; Validation of a head mounted virtual reality visual field screening device. Invest. Ophthalmol. Vis. Sci. 2019;60(9):2482. doi: https://doi.org/.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Testing the visual field is a key part in the diagnosis and management of glaucoma. The current gold standard, the Humphrey visual field analyser (HFA), is large, expensive and can be uncomfortable for some patients. The current study investigated the C3 fields analyser (CFA), a virtual reality head mounted visual field testing device, as a possible replacement for glaucoma screening and eventually glaucoma monitoring.
The CFA presented stimuli in the same 54 positions as the HFA 24-2 SITA Standard test using a suprathreshold algorithm approximating an 18dB deficit for 157 patients at the Aravind Eye Hospital, Pondicherry, India. Patients were diagnosed clinically, and then designated as control or glaucoma through consensus by blinded researchers using HFA, optical coherence tomography and fundus photos.
The number of stimuli missed on the CFA correlated with HFA mean deviation (r = 0.62, P <0.001), and with pattern standard deviation (r = 0.36, P <0.001). The area under the receiving operator characteristic curve (AROC) was 0.77 ± 0.06 for mild glaucoma (HFA mean deviation ≥ -6 dB) and 0.86 ± 0.04 for moderate-advanced glaucoma (HFA mean deviation < -6 dB). Each individual missed stimulus on the CFA had an average positive predictive value of 0.76 for identifying glaucoma. On an analysis of each individual stimulus point, the CFA correctly identified 38% identical deficits as the HFA and 91% identical correct responses.
Despite the fact that the CFA did not reliably identify the same points as abnormal as the HFA, it was effective at identifying glaucoma subjects as shown by the robust AROC and positive predictive value. Further refinements to the device will be required to improve on point by point testing performance.
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.
This PDF is available to Subscribers Only