A prospective cross-sectional study of 87 Japanese glaucoma patients found that the Melbourne Rapid Fields (MRF) online perimetry system showed strong agreement with the Humphrey Field Analyzer (HFA) 30-2 SITA Fast protocol across a range of disease severity. The findings suggest that MRF’s 30-2 protocol, conducted on a standard laptop computer, can provide portable visual-field testing with reliable repeatability, but it is not interchangeable with HFA 30-2.

Across all eyes, MRF showed high concordance with HFA for mean deviation (MD) and pattern standard deviation (PSD). For MD, intra class correlation coefficient (ICC) was 0.97 with a 95% CI: 0.95–0.98; mean bias of −0.76 dB with 95% limits of agreement (LoA) of −5.82 to +4.30 dB. For PSD, ICC was 0.91 with a 95% CI: 0.86–0.94, and mean bias was +0.79 dB with LoA of −4.24 to +5.82 dB.

Subgroup analysis showed varying ICC values by severity. For MD, ICCs were 0.562 for mild glaucoma, 0.275 for moderate, and 0.947 for advanced. ICCs for PSD were 0.619 for mild glaucoma, 0.333 for moderate, and 0.694 for advanced disease. The authors noted that narrower data ranges within subgroups may mathematically reduce ICC values.

For eyes with mild glaucoma, classification using OHTS secondary criteria showed that 33 of 43 eyes were classified identically on MRF and HFA. Sensitivity was 80% and specificity was 72% for MRF. MRF also detected 3 of 4 peripheral defects that would have been missed using only 24-2 test points, and it demonstrated excellent repeatability between its first and second sessions:

• MD: ICC 0.99; mean bias +0.25 dB (LoA −2.48 to +2.99 dB)

• PSD: ICC 0.97; mean bias −0.21 dB (LoA −3.22 to +2.79 dB)

Subgroup analyses also showed small biases across severity levels.

False-positive, false-negative, and fixation loss rates were similar between devices, though MRF showed slightly higher false negatives (7.1% vs 4.3%). This difference did not reach statistical significance, however.

Test duration was also comparable. HFA was 313.9±90 seconds, while first MRF was 342.8±50.1 seconds and second MRF was 332.6±56.6 seconds. Pairwise comparison showed no significant difference between HFA and the second MRF test.

Kae Sugihara, MD, of the Department of Ophthalmology at Kurashiki Medical Center in Kurashiki, Okayama, Japan, and colleagues compared MRF with other automated perimetry systems, such as imo perimetry, the novel gaze analyzing perimeter (GAP), the PalmScan VF2000 Visual Field Analyzer, C3 field visual analyzer (CFA), the Toronto portable perimeter (TPP), and the Smart System Virtual Reality (SSVR). Imo perimetry showed correlation, but wider LoA than some alternatives. GAP showed good correlation and a similar LoA range to MRF. PalmScanVF2000 showed much wider variability compared with MRF. A report on CFA concluded that it was unreliable for matching HFA deficits. TPP had a lower LoA than MRF, and MD for SSVR was similar to HFA across severity levels.

“Although the several perimeters, including imo, GAP, TPP, SSVR, and compass, are reliable testing as well as MRF and may offer an accessible alternative to HF AVF testing, they still require a special headset or specialist testing device,” the authors wrote. “MRF does not require any special equipment except for a digital screen computer. This study is unique in that previous studies of other perimetry systems has focus on 24-2 field test, we have shown that the MRF online perimetry provides valid and repeatable assessment of the wider 30-2 visual field.”

The study included 87 eyes from 87 participants (mean age 69.6±12.0 years). Forty-three were classified as mild (MD > −6 dB), 26 moderate (−12 dB ≤ MD ≤ −6 dB), and 18 advanced (MD <−12 dB) based on HFA results. Glaucoma types included primary open-angle glaucoma (41 eyes), normal-tension glaucoma (34), primary angle-closure glaucoma (5), exfoliation glaucoma (4), secondary glaucoma (1), developmental glaucoma (1), and superior segmental optic nerve hypoplasia (1). Patients were included if they were at least 18 years or older and had BCVA of up to 0.4 logarithm of the minimum angle of resolution (logMAR). They also required previous experience with perimetry tests on HFA. Patients were excluded if they presented with any other ocular diseases or systemic conditions that could cause optic nerve lesions or visual function impairment, the authors noted. They also excluded patients who had intraocular surgery in the previous 3 months.

Participants completed both HFA 30-2 SITA Fast and 2 MRF 30-2 tests on the same day. MRF was administered on a 14-inch laptop using webcam-based gaze and viewing-distance monitoring.

Dr. Sugihara and colleagues noted limitations including absence of data on sensitivity or specificity in detecting glaucoma. The controlled environment, they wrote, may account for the strong correlations between MRF and HFA that they observed in the study. Randomization in future studies could enhance methodological rigor.

Source: Journal of Glaucoma