Bold opening: Glaucoma care moves faster when testing is quick, comfortable, and reliable — but the devices delivering those benefits aren’t identical. And this is where it gets controversial: newer perimeters can outperform traditional ones in speed and patient experience, yet their measurements may diverge in key metrics. Here's a rewritten, unique presentation of the original content that preserves all essential information while clarifying concepts for beginners and keeping a professional, conversational tone.
Introduction
Glaucoma remains the leading cause of irreversible blindness worldwide, affecting an estimated 76.0 million people globally and about 4.2 million in the United States. Visual field testing is a cornerstone of comprehensive eye exams for individuals with glaucoma. Perimetry systematically measures how the field of vision functions, enabling clinicians to identify, quantify, and map patterns of vision loss to detect and monitor glaucoma. The Humphrey Field Analyzer (HFA, Carl Zeiss Meditec, Dublin, CA) is a widely used device for perimetry. However, because each eye is tested separately, the HFA can be time-consuming for both patients and clinicians. Additionally, the HFA uses a projection bowl and requires testing in a dark room, occupying a relatively large footprint.
As technology advances, new approaches aim to overcome these traditional SAP (standard automated perimetry) challenges. Innovations include virtual reality and tablet-based perimetry to address longstanding pain points and provide space-saving options. Another alternative is the TEMPO perimeter (Topcon Healthcare, Tokyo, Japan), marketed as IMOvifa outside the United States. The TEMPO is a tabletop device that allows visual field testing under ambient room lighting. The TEMPO’s 24–2 Ambient Interactive Zippy Estimated by Sequential Testing (AIZE)-Rapid protocol has shown good agreement with the HFA 24–2 Swedish Interactive Threshold Algorithm (SITA)-Fast. Perimetry using virtual reality has also emerged, with devices like the Virtual Eye ELITE (VEE; Virtual Vision, Miami, FL) being head-mounted and portable for in-home or in-office use under ambient light. Pilot data suggest reasonable agreement with HFA SAP and faster measurements, though more studies are needed. This study compares test duration and agreement of summary metrics among TEMPO, VEE, and HFA perimeters.
Methods
The study was a prospective, randomized, multi-arm, single-site trial comparing visual assessments across three perimeters. An independent IRB (Advarra IRB, approval # Pro00080959) oversaw the study, which was conducted in private practice. The study adhered to the Declaration of Helsinki, International Harmonization (ICH) guidelines, and Good Clinical Practice (GCP), and was registered in a clinical trials database (NCT06539715). Data are not available for sharing.
Inclusion criteria included adults with glaucoma and best-corrected visual acuity of 20/30 or better. Exclusions were based on intolerance to ophthalmic imaging or any ocular or systemic conditions that could affect visual field results (e.g., age-related macular degeneration, peripheral retinal disease, or severe glaucoma defined as cup-to-disc ratio > 0.8). Mild and moderate glaucoma were defined by cup-to-disc ratios of 0.6–0.7 and 0.8, respectively.
Eligible participants underwent perimetry with all three devices: HFA (model 745), TEMPO (version 20231214), and VEE (model A7510). TEMPO testing used the 24–2 AIZE-Rapid protocol, while HFA and VEE used 24–2 SITA-Fast. All devices used a Goldmann size III stimulus. Subjects were randomized to the order of device testing using pre-generated block randomization, with a five-minute rest between devices. The study took place in a single visit with informed consent and concurrent perimetry using all devices. Data collected included mean deviation (MD), pattern standard deviation (PSD), foveal threshold, false positives, false negatives, and total acquisition time for both eyes. A researcher-developed questionnaire surveyed ease of testing, comfort, and anxiety for each device on a five-point Likert scale. Total setup time was from the initial instructions to the start of visual field acquisition, while total visual field acquisition time covered testing both eyes, excluding setup and eye-switch time. There were no data exclusions based on device-specific false positives or negatives.
Primary endpoint: total acquisition time (both eyes) across devices. Secondary endpoints: MD, PSD, foveal threshold, and questionnaire responses. Exploratory endpoints included visual field index (VFI), setup time, and false positives/negatives for each device.
Statistical analysis used R (version 4.4.2). Linear mixed-effects models adjusted for device order and repeated measures within participants. Questionnaire responses were analyzed with Fisher’s Exact test, with p-values adjusted for multiple comparisons. Bland-Altman analyses evaluated differences in MD, PSD, foveal threshold, and VFI. A two-sided p-value ≤ 0.05 indicated significance. Sample size was planned for MD and PSD differences with an assumed effect size of 0.5, alpha 0.05, and 90% power, targeting 46 subjects and 54 to account for 15% dropout.
Results
Fifty-four participants completed the study, with no adverse events reported. Demographic details are summarized in Table 1. Table 2 presents bilateral acquisition times, setup times, MD, PSD, and foveal threshold across devices. Bilateral visual field acquisition time averaged 7.34 ± 2.27 minutes for HFA, 4.44 ± 0.88 minutes for TEMPO, and 7.71 ± 1.21 minutes for VEE. Differences were significant for TEMPO vs HFA and TEMPO vs VEE (p < 0.001 for both), but not significant between HFA and VEE (p = 0.34). Setup times averaged 1.92 ± 0.66 minutes for HFA, 1.69 ± 0.65 minutes for TEMPO, and 1.68 ± 0.60 minutes for VEE; TEMPO vs HFA and HFA vs VEE showed significant differences, but TEMPO vs VEE did not (p values 0.002, 0.001, and 0.98 respectively).
Mean MD values were −3.28 ± 4.19 dB (HFA), −3.16 ± 4.39 dB (TEMPO), and −4.22 ± 3.51 dB (VEE). Differences were not significant between TEMPO and HFA (p = 0.96), but were significant for TEMPO vs VEE (p = 0.01) and HFA vs VEE (p = 0.02). PSD means were 4.22 ± 3.21 dB (HFA), 4.78 ± 3.82 dB (TEMPO), and 4.51 ± 2.99 dB (VEE); differences were not significant among any pair. Foveal thresholds averaged 34.71 ± 3.63 dB (HFA), 31.40 ± 3.77 dB (TEMPO), and 27.61 ± 6.20 dB (VEE); all pairwise differences were significant (p < 0.001).
Bland-Altman analyses revealed small mean differences in MD between HFA and TEMPO (−0.12 dB) with 95% limits of agreement (LoA) −4.27 to 4.04 dB. TEMPO vs VEE showed a mean MD difference of 1.06 dB (LoA −4.36 to 6.47 dB), hinting at higher values with TEMPO on average; HFA vs VEE showed a mean MD difference of 0.94 dB (LoA −4.05 to 5.94 dB).
PSD differences were similar, with HFA vs TEMPO showing a mean difference of −0.57 dB (LoA −4.75 to 3.62 dB), TEMPO vs VEE −0.27 dB (LoA −4.04 to 4.58 dB), and HFA vs VEE −0.30 dB (LoA −4.38 to 3.79 dB).
Foveal threshold differences were more pronounced, with HFA vs TEMPO at 3.31 dB (LoA −5.28 to 11.91 dB), TEMPO vs VEE at 3.78 dB (LoA −8.48 to 16.04 dB), and HFA vs VEE at 7.09 dB (LoA −5.41 to 19.60 dB).
VFI differences were small: HFA vs TEMPO −0.36 (LoA −12.87 to 12.16%), TEMPO vs VEE −0.70 (LoA −21.10 to 19.71%), and HFA vs VEE −0.89 (LoA −18.96 to 17.19%).
Questionnaire results (Table 3) showed more participants reporting that TEMPO was easy to perform than HFA, and more reporting TEMPO and VEE as easy than HFA. Comfort followed a similar pattern, with TEMPO and VEE rated more comfortable than HFA. Anxiety scores did not differ significantly among device comparisons.
Overall, mean VFI values were 91.3% (HFA), 91.6% (TEMPO), and 92.3% (VEE), with no significant differences among devices. False positives were highest with HFA (6.2%), significantly higher than TEMPO (1.7%) and VEE (0.3%), with TEMPO higher than VEE. False negatives followed a similar pattern: highest with HFA (5.7%), significantly higher than TEMPO (0.6%) and VEE (0.2%), again with TEMPO and VEE not differing significantly.
Discussion
Advances in perimetry aim to address limitations of standard automated perimetry. This study is, to our knowledge, the first to compare HFA, TEMPO, and VEE side-by-side. A key drawback of traditional SAP is the time required to perform bilateral (two-eye) testing. TEMPO delivered significantly faster bilateral acquisitions than HFA or VEE, which is expected since TEMPO tests both eyes without breaks, whereas HFA and VEE test each eye separately. This efficiency could reduce technician time and increase clinic throughput. Prior studies from Nishida et al. reported bilateral acquisition times around 4.3–4.4 minutes for TEMPO and 6.9–7.15 minutes for HFA, aligning with our findings. Kang et al. reported similar trends, with TEMPO faster than HFA. Pilot data for VEE suggested longer times than HFA, but those studies used different protocols.
Agreement between devices was reasonable, with no clinically meaningful differences in MD or PSD. Our results align with Nishida and colleagues, who found MD differences around −0.1 to −0.2 dB and PSD differences around −0.4 to −0.6 dB between HFA and TEMPO. Data on direct TEMPO–VEE agreement are scarce; however, McLaughlin and colleagues reported reasonable agreement between HFA and VEE in a telehealth context.
A notable distinction is the VEE device’s luminance and background. VEE can reach a maximum luminance of 120 cd/m2 on a 1 cd/m2 background, whereas HFA and TEMPO reach up to 3183 cd/m2 on a 10 cd/m2 background. This difference affects Weber contrast and adaptation state, which may influence contrast detection and measurement in practice.
Questionnaire results favored TEMPO and VEE over HFA in ease and comfort, though the lack of standardized questionnaires across studies makes direct comparisons challenging. Other studies have similarly reported higher user satisfaction with newer devices.
Limitations include excluding patients with severe glaucoma, which limits generalizability to all disease stages. The study was powered for the primary endpoint (acquisition time), so conclusions about MD, PSD, FT, and VFI agreement should be interpreted cautiously. Reproducibility was not assessed, and three sequential tests could have induced fatigue, potentially biasing results despite randomization of device order.
Conclusion
The findings suggest that HFA, TEMPO, and VEE yield similar summary metrics, with TEMPO offering faster bilateral acquisition. These results support potential improvements in clinical workflow and patient experience, while underscoring the need for longitudinal monitoring and cost-effectiveness analyses in future work.
