Comparison of a New Head Mount Virtual Reality Perimeter (C3 Field Analyzer) With Automated Field Analyzer in Neuro-Ophthalmic Disorders

Automated perimetry in neurologically disabled patients is a challenge. We have devised a patient-friendly virtual reality perimeter, the C3 field analyzer (CFA). We aim to assess the utility of this as a visual field-testing device in neuro-ophthalmic patients for screening and monitoring.

Original Contribution Section Editors: Clare Fraser, MD Susan Mollan, MD Comparison of a New Head Mount Virtual Reality Perimeter (C3 Field Analyzer) With Automated Field Analyzer in Neuro-Ophthalmic Disorders Annamalai Odayappan, DNB, Priya Sivakumar, MS, Sandal Kotawala, BTech, Ramalakshmi Raman, MSc, Sivagami Nachiappan, DO, Arulmozhivarman Pachiyappan, PhD, Rengaraj Venkatesh, DNB Background: Automated perimetry in neurologically disabled patients is a challenge. We have devised a patientfriendly virtual reality perimeter, the C3 field analyzer (CFA). We aim to assess the utility of this as a visual field-testing device in neuro-ophthalmic patients for screening and monitoring. Methods: Neuro-ophthalmic patients and controls were selected to participate in the study between September and December 2018. They randomly underwent either the CFA or automated field analyzer (HFA) first followed by the other in an undilated state. The CFA results were compared with the HFA, and the correlation of the pattern of the field defect was assessed by an independent masked physician. Results: In total, 59 eyes of 33 neuro-ophthalmic patients (cases) and another 95 normal individuals (controls) were enrolled. CFA was found to have greater proportion of reliable fields (81.4%) than HFA (59.3%) (P = 0.009). There were less false negatives (P , 0.001) and more false positives in CFA (P , 0.001) among neuro-ophthalmic patients compared with controls. Among neuroophthalmology patients, the number of fixation losses was greater with CFA (P , 0.001), whereas false negatives were greater in HFA (P , 0.001). On assessing the pattern of the field defects, we found that there was almost 70% correlation of CFA with HFA. Moreover, in classical neurological fields such as hemianopia, the correlation was 87.5%. Conclusions: The CFA seems to correlate well with HFA in classic neurological fields such as hemianopias and may Glaucoma Services (AO), Aravind Eye Hospital, Pondicherry, India; Neuro-Ophthalmology Services (PS), Aravind Eye Hospital, Pondicherry, India; Alfaleus Technology Private Limited (SK), Jaipur, Rajasthan, India; Department of Biostatistics (RR), Aravind Eye Hospital, Madurai, Tamil Nadu, India; Indira Gandhi Govt. General Hospital and Post Graduate Institute (SN), Pondicherry, India; School of Electrical Engineering (PA), Vellore Institute of Technology, Vellore, Tamil Nadu, India; and Aravind Eye Hospital (RV), Pondicherry, India. S. Kotawala is the inventor of the CFA device. However, he had no role in the conduct of the study or the analysis of the results. Address correspondence to Annamalai Odayappan, DNB, C/o Aravind Eye Hospital, Cuddalore main road, Thavalakuppam, Pondicherry—605007, India. E-mail: [e-mail] 232 serve as an alternative in patients unable to perform a standard automated perimetry. Further developments are currently underway to incorporate threshold testing. Journal of Neuro-Ophthalmology 2023;43:232–236 doi: 10.1097/WNO.0000000000001714 © 2022 by North American Neuro-Ophthalmology Society A comprehensive neuro-ophthalmology examination is incomplete without visual field evaluation. The pattern of the visual field defects helps in clinical localization of the site of lesion along the visual pathway, guiding in appropriate imaging and thereby timely intervention. Apart from its diagnostic utility, visual field examination is also useful to monitor resolution or recurrence and planning rehabilitation strategies (1). Currently, the standard method of visual field examination is the automated static perimetry. However, the automated static perimetry has its limitations. Patients need good cognition to do an automated perimetry testing. They need to maintain their posture for the duration of the testing that might be difficult in patients with neurological problems, resulting in unreliable fields and futile testing. In addition, it is expensive and time consuming as well. A poor field assessment may result in the patients being subjected to needless defensive investigations such as MRI of the brain. Sometimes, on the contrary, appropriate investigations may be missed. In some disabled patients, confrontation testing is helpful but it has low sensitivity for detecting visual field loss associated with parasellar tumors or compressive optic neuropathies. The reported sensitivity in detecting anterior and posterior visual field defects by confrontation testing is around 50% when compared with automated static perimetry (2). Odayappan et al: J Neuro-Ophthalmol 2023; 43: 232-236 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution These factors suggest that there is a need for a low-cost, user-friendly method to evaluate the visual field as accurately as the automated static perimetry. We are attempting to bridge this gap using the C3 field analyzer (CFA) (Alfaleus Technology Private Limited, Jaipur, Rajasthan, India), which is a virtual reality–based perimeter. We aimed to evaluate the validity of the CFA device as a screening and diagnostic tool for neuro-ophthalmic disorders by measuring the sensitivity, specificity, and correlation of CFA screening algorithm in comparison with the standard automated field analyzer (HFA) (Carl Zeiss Meditec Inc, Dublin, CA). METHODS Institutional review board approval was obtained at the Aravind Eye Hospital, Pondicherry, for this study. This research adhered to the tenets of the Declaration of Helsinki. This is a prospective comparison study. Patients were enrolled from the neuro-ophthalmology clinic (cases), and age-matched controls were selected from the general outpatient department at the Aravind Eye Hospital, Pondicherry, between September and December 2018. Cases refers to patients who were diagnosed with neuroophthalmic defects based on visual acuity, comprehensive anterior, posterior segment examination, neurological examination, HFA, and neurological imaging when needed. We included patients who had any neurological pattern of field defect. Age-matched controls underwent visual acuity testing, comprehensive anterior, posterior segment examination, neurological examination, and confrontational testing of visual fields and were diagnosed to be normal. Classification of patients as having neuro-ophthalmological diseases or controls was confirmed by the treating physician. We enrolled 63 eyes of 35 patients in the study. Two neuro-ophthalmic patients (4 eyes) had erroneously performed testing with a 10–2 program on HFA and were, therefore, excluded. Finally, 59 eyes of 33 patients were included for the analysis. Another 95 individuals who were found to be normal without any field defects were selected as controls. The common inclusion criteria included age between 18 and 65 years, having a best-corrected visual acuity (BCVA) better than or equal to 6/12 and able to provide informed written consent. All participants underwent visual field testing using the CFA and HFA. Automated static perimetry was performed using the HFA II-i (Carl Zeiss Meditec Jena, Germany), where a 30-2 threshold test with a Swedish Interactive Threshold Algorithm Standard (SITA-Standard) strategy was used. The CFA is a virtual reality (VR)-based perimeter (Fig. 1). The system comprises 3 subsystems—a VR headset, a mobile phone, and a response button. All 3 subsystems are interconnected through Bluetooth. The user can wear the VR headset, and the operator can control the test through the CFA application on the mobile phone. Using Odayappan et al: J Neuro-Ophthalmol 2023; 43: 232-236 the mobile application, the operator can choose parameters, such as field of view, test strategy, and eye to be tested. Test can be started, paused/resumed, and aborted from the mobile application at any time. Based on input parameters of the operator, a test is started in the VR headset. Only the chosen eye is presented with stimulus at a time. Since the display screen for each eye is independent, eye patching is not required. The padding on the device provides a dark environment inside the VR headset; therefore, a dark room is not required for performing the test. To correct for refractive errors, a set of lenses are made available; however, only spherical lenses are available, and for cylindrical power, a spherical equivalent is used. The VR headset screen is calibrated to a semi-white (shade of white) background to give a luminance of 10 cd/m2, which is equivalent to 31.5 apostilbs. White stimulus can be presented to the user with varying intensity; however, for the study, we used a fixed suprathreshold stimuli brightness preset at 60 cd/m2 using an HTC Instrument LX- 101A Light Meter Luxmeter (HTC, Taoyuan City, Taiwan) approximating an 18 dB stimulus of HFA. A 30-2 suprathreshold test was performed. As the test begins, the user can see a fixation spot in the center. The user is instructed to keep looking at the center and press the response button on seeing a blinking dot (stimulus). The stimulus is presented for a duration of 200 milliseconds, and a time interval of 1,300 milliseconds is available for the user to respond after which the next stimulus is presented, and in this pattern, the test progresses. Before starting the test, the device shows 10 points near the potential blind spot location, and based on the response, the blind spot is mapped. During the test, up to 3% of stimuli are projected at the blind spot, and if the user still clicks on it, it is counted as a fixation loss. In addition to FIG. 1. C3 field analyzer (CFA). 233 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution this, another 3% of stimuli during the test are displayed after a delay 1,200 milliseconds. If the user still clicks when no stimulus is visible on the screen, it is counted as a false positive response. Each stimulus position was presented twice during a test. If for any given point the user has clicked once and not clicked other time, they are considered false negatives. The operator can track the test progress on the mobile phone on which a graphical representation of areas where stimulus is being presented, and the reliability parameters such as fixation losses and false positives are updated in real time. Based on this information, the operator can reinstruct the user or abort and restart the test if reliability parameters are out of limit. The test result is available and stored in the software on the smartphone which can be exported or printed directly. Patients were shown an instructional video on the use of the CFA. A short demo test was conducted to ensure understanding. All participants were randomized arbitrarily to undergo either CFA or HFA first followed by the other in an undilated state. Both the eyes were included in the study depending on the eligibility. A mandatory 5-minute gap was given between testing of the 2 eyes and between the 2 visual field testing. Individuals with unreliable fields were given a short 10-minute break, and testing was repeated. The results of the CFA report were compared with the HFA reports for all participants. The correlation of the pattern of the field defect was assessed by an independent consultant who was not aware of the study protocol. Statistical Analysis Mean ± SD and frequency (%) were used to describe the summary data. The chi‐squared test was used to assess the association between the pattern in CFA and pattern in HFA. Two sample t tests were used to find the mean difference of fixation losses, false positives, or false negatives between cases and controls. P value ,0.05 were considered as statistically significant. All the statistical analyses were performed by STATA 14.0 (TX). RESULTS We included 59 eyes of 33 neuro-ophthalmology patients in the analysis. Among them, 24 eyes had hemianopic field defects by HFA. The other field defects present were altitudinal field defect, arcuate scotomas, blind spot enlargement, peripheral constriction of visual fields, and central and paracentral scotomas. The average mean deviation (MD) and pattern SD (PSD) of the neuroophthalmic patients was 215.91 and 11.32 dB, respectively. The mean (±SD) visual field index (VFI) was 56.1% (±25.5%). The mean refractory correction required was +2.03 D. The mean age (49 ± 14.7 vs 49.8 ± 9.2 years, P = 0.678) and gender distribution (67.8% vs 54% males, P = 0.090) were similar between the neuro-ophthalmology cases and controls. The BCVA (0.13 ± 0.3 vs 0.14 ± 0.2, 234 TABLE 1. Reliability in the standard HFA Reliability Pattern Not Reliable n (%) Reliable n (%) Total Hemianopia Others Total 4 (16.7) 20 (57.1) 24 (40.7) 20 (83.3) 15 (42.9) 35 (59.3) P* 24 0.002 35 59 *Chi-squared test. HFA, automated field analyzer. P = 0.804) was also similar between the neuroophthalmology cases and controls. Reliability Poor reliability was defined by the presence of either fixation loss, false positive, or false negative over 33%. Increased fixation loss was the most common cause of poor reliability in both groups (54.2% in HFA vs 81.8% in CFA among all unreliable fields). We found that we had significantly greater proportion of reliable fields with CFA (81.4%) than with HFA (59.3%) (P = 0.009) (Tables 1 and 2). Moreover, this percentage was better in patients with hemianopic fields where again CFA was found to have a greater proportion of reliable fields than HFA (95.7% vs 83.3%) but this difference was not statistically significant (P = 0.171). The reliability indices of the neuro-ophthalmology cases were also compared with the controls, and we found that there were significantly higher fixation losses among neuroophthalmology patients with both CFA and HFA (P # 0.001 for both). There were less false negatives (P , 0.001) and more false positives with CFA in neuroophthalmology patients (P , 0.001) compared with controls (Table 3). Among neuro-ophthalmology patients, the number of fixation losses was greater with CFA than HFA (P , 0.001), whereas false negatives were greater in HFA than CFA (P , 0.001). There was no significant difference in the number of false positives between the 2 groups (P = 0.061). Correlation On assessment of the similarity between the patterns of the field defects between the 2 instruments, we found that there TABLE 2. Reliability in the CFA Reliability Pattern Not Reliable n (%) Reliable n (%) Total Hemianopia Others Total 1 (4.3) 10 (27.8) 11 (18.6) 22 (95.7) 26 (72.2) 48 (81.4) P* 23 0.024 36 59 *Chi-squared test. CFA, C3 field analyzer. Odayappan et al: J Neuro-Ophthalmol 2023; 43: 232-236 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution TABLE 3. Comparison of the reliability indices for CFA and HFA in controls and neuro-ophthalmology cases Controls (n = 95 Patients = 190 Eyes) Average Fixation loss (%) False negatives (%) False positives (%) CFA HFA 6.74 ± 12 7.17 ± 7.5 1.47 ± 4.1 0.09 ± 0.2 4.43 ± 6.8 3.31 ± 3.7 Neuro-Ophthalmology Cases (P)* CFA (n = 48 Eyes) HFA (n = 35 Eyes) 13.1 ± 6.2 (P = 0.001) 0.6 ± 2.4 (P , 0.001) 7.92 ± 17.3 (P = 0.001) 5.8 ± 6.5 (P , 0.001) 7.0 ± 7.8 (P = 0.069) 2.3 ± 2.9 (P = 0.148) *Two sample t test. CFA, C3 field analyzer; HFA, automated field analyzer. *Chi-squared test. CFA, C3 field analyzer; HFA, automated field analyzer. them from fixing properly. Moreover, higher fixation loss in CFA could be due to the lack of orientation to the new VR devices among the general population. It could be that because the entire external environment is blocked, the participants start to search for light leading to fixation losses. The HFA being more time consuming may lead to greater patient fatigue and increased false negatives. CFA is much faster which could be one of the reasons for having very low false negatives. Newer perimeters are either tablet based such as the Melbourne Rapid Fields, Eyecatcher, and Visual Fields Easy (4) or laptop screen based such as the Moorfields Motion Displacement Test (5) or can involve the use of VR headsets which includes the C3 Field Analyzer, Oculus Quest virtual reality headset (6), GearVision (7), VisuALL (Olleyes, Inc, Summit, NJ) (8), Virtual Eye (9), and imo (10). Stapelfeldt et al compared the Oculus Quest VR headset with the Octopus 900 in glaucoma patients and found that there was high mean defect correlation between the 2 devices (6). Pradhan et al compared the GearVision with the HFA and found good agreement of threshold sensitivities between the 2 devices (7). Tsapakis et al (11) have devised a home-based visual field test using a PC monitor or VR glasses and suggested that it is useful for glaucoma screening. The Toronto Portable Perimeter (VEM Medical Technologies) is another perimeter with similar MD, PSD, VFI, and test duration compared with the HFA (12). Apart from these, various other approaches of visual field testing have been tried such as the flicker perimetry (13), high-pass resolution or ring perimetry (14), microperimetry (15), rarebit perimetry (16), and the edge perimetry (17). Patients with neurological disorders often find performing an automated perimetry difficult due to various reasons such as inability to maintain posture, poor comprehension, accessing a health care facility to perform the test, and longer duration of the testing. This portable perimeter, the CFA, overcomes these disadvantages with fairly good correlation and, thus, can be a useful tool to assess the neuro-ophthalmic status and aid in monitoring the disease. The strength lies in the prospective nature of this comparison with the standard automated perimetry. This CFA device was the prototype device, and current research includes the development of threshold testing and other programs. The limitations are that we had used a Odayappan et al: J Neuro-Ophthalmol 2023; 43: 232-236 235 was almost 70% correlation of CFA with HFA. However, in classical neurological fields such as hemianopia, the correlation was 87.5% (Table 4). The average time taken (mean ± SD) to complete the CFA test was 4 minutes 1 ± 38 seconds, whereas HFA took 8 minutes 2 sec ± 2 minutes 1 seconds. As expected, the suprathreshold CFA test was faster than the threshold test of HFA (P , 0.001). CONCLUSIONS Our study evaluated the role of the CFA device using suprathreshold stimulus in visual field testing among neuroophthalmology patients in an Indian population. We found that there was good correlation between the CFA and HFA. Recently, there is significant interest in the development of VR-based perimeters with the advantages being not needing to maintain a fixed position and allowing head movement. Moreover, it could be performed by bedridden patients. The entire device is compact enough so that it can be carried to outreach camps or vision centers. A constant power supply or Internet is also not required. In a similar study in glaucoma patients with the same device, it was found that fixation losses and false negatives were more common with CFA and false positives were more common in HFA (3). By contrast, we found that in neuroophthalmology patients, fixation losses were greater with CFA; however, false negatives were greater in HFA. We found greater fixation losses in neuro-ophthalmology patients with both CFA and HFA compared with controls. This is most likely a result of involvement of the fixation in large number of neuro-ophthalmic diseases which prevents TABLE 4. Correlation of the pattern of field defects on HFA with CFA in neuro-ophthalmology patients CFA Correlation Pattern on HFA No Yes Total P* Hemianopia Others Total 3 (12.5) 15 (42.9) 18 (30.5) 21 (87.5) 20 (57.1) 41 (69.5) 24 (100) 35 (100) 59 (100) 0.013 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution suprathreshold stimulus for testing with CFA, and comparing a screening procedure with a threshold test may not be ideal. There is a possibility that a few relative scotomatous areas may have been missed. However, we do not see this as a major barrier in neurological field testing because the correlation was over 85%. In nonhemianopic fields, the CFA had just over a 50% correlation, but nonetheless, CFA could be used as a screening tool for neurological diseases. In conclusion, we find that the CFA can be an ideal alternative for visual field estimation in neuro-ophthalmic patients. It may be more beneficial in bed-ridden patients and those who are unable to maintain the posture for long time. Further device upgrades are currently being developed to better approximate or supersede the gold standard automated perimetry. STATEMENT OF AUTHORSHIP Conception and design: P. Sivakumar, S. Kotawala, P. Arulmozhivarman, R. Venkatesh. Acquisition of data: P. Sivakumar. Analysis and interpretation of data: A. Odayappan, R. Raman, S. Nachiappan. Drafting the manuscript: P. Sivakumar, A. Odayappan, S. Nachiappan. Revising the manuscript for intellectual content: S. Kotawala, R. Raman, P. Arulmozhivarman, R. Venkatesh. Final approval of the completed manuscript: A. Odayappan, P. Sivakumar, S. Kotawala, R. Raman, S. Nachiappan, P. Arulmozhivarman, R. Venkatesh. ACKNOWLEDGMENTS The authors thank Dr. Keerthi Gayam, the independent masked consultant, who analyzed the pattern of the various visual field defects. REFERENCES 1. Kedar S, Ghate D, Corbett JJ. Visual fields in neuroophthalmology. Indian J Ophthalmol. 2011;59:103–109. 2. Johnson LN, Baloh FG. The accuracy of confrontation visual field test in comparison with automated perimetry. J Natl Med Assoc. 1991;83:895–898. 3. Mees L, Upadhyaya S, Kumar P, Kotawala S, Haran S, Rajasekar S, Friedman DS, Venkatesh R. Validation of a headmounted virtual reality visual field screening device. J Glaucoma. 2020;29:86–91. 236 4. Prager AJ, Kang JM, Tanna AP. Advances in perimetry for glaucoma. Curr Opin Ophthalmol. 2021;32:92–97. 5. Ong EL, Zheng Y, Aung T, Tan L, Cheng CY, Wong TY, How A. Performance of the Moorfields motion displacement test for identifying eyes with glaucoma. Ophthalmology. 2014;121:88– 92. 6. Stapelfeldt J, Kucur SS, Huber N, Höhn R, Sznitman R. Virtual reality-based and conventional visual field examination comparison in healthy and glaucoma patients. Transl Vis Sci Technol. 2021;10:10. 7. Pradhan ZS, Sircar T, Agrawal H, Rao HL, Bopardikar A, Devi S, Tiwari VN. Comparison of the performance of a novel, smartphone-based, head-mounted perimeter (GearVision) with the Humphrey field analyzer. J Glaucoma. 2021;30:e146– e152. 8. Montelongo M, Gonzalez A, Morgenstern F, Donahue SP, Groth SL. A virtual reality-based automated perimeter, device, and pilot study. Transl Vis Sci Technol. 2021;10:20. 9. Wroblewski D, Francis BA, Sadun A, Vakili G, Chopra V. Testing of visual field with virtual reality goggles in manual and visual grasp modes. Biomed Res Int. 2014;2014:206082. 10. Matsumoto C, Yamao S, Nomoto H, Takada S, Okuyama S, Kimura S, Yamanaka K, Aihara M, Shimomura Y. Visual field testing with head-mounted perimeter ’imo. PLoS One. 2016;11:e0161974. 11. Tsapakis S, Papaconstantinou D, Diagourtas A, Kandarakis S, Droutsas K, Andreanos K, Brouzas D. Home-based visual field test for glaucoma screening comparison with Humphrey perimeter. Clin Ophthalmol. 2018;12:2597–2606. 12. Ahmed Y, Pereira A, Bowden S, Shi RB, Li Y, Ahmed IIK, Arshinoff SA. Multicenter comparison of the Toronto portable perimeter with the Humphrey field analyzer: a pilot study. Ophthalmol Glaucoma. 2022;5:146–159. 13. Nomoto H, Matsumoto C, Takada S, Hashimoto S, Arimura E, Okuyama S, Shimomura Y. Detectability of glaucomatous changes using SAP, FDT, flicker perimetry, and OCT. J Glaucoma. 2009;18:165–171. 14. Martin LM, Lindblom B, Gedda UK. Concordance between results of optic disc tomography and high-pass resolution perimetry in glaucoma. J Glaucoma. 2000;9:28–33. 15. Lima VC, Prata TS, De Moraes CG, Kim J, Seiple W, Rosen RB, Liebmann JM, Ritch R. A comparison between microperimetry and standard achromatic perimetry of the central visual field in eyes with glaucomatous paracentral visual-field defects. Br J Ophthalmol. 2010;94:64–67. 16. Steven Houston SK, Weber ED, Koga SF, Newman SA. Rarebit perimetry for bedside testing: comparison with standard automated perimetry. J Neuroophthalmol. 2010;30:243–247. 17. Horn FK, Kremers J, Mardin CY, Jünemann AG, Adler W, Tornow RP. Flicker-defined form perimetry in glaucoma patients. Graefes Arch Clin Exp Ophthalmol. 2015;253:447– 455. Odayappan et al: J Neuro-Ophthalmol 2023; 43: 232-236 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited.