Selwyn M Prea
Professor Algis J Vingrys
PhD BSc(Optom) PGCertOcTher FARVO FAAO
Department of Optometry and Vision Sciences, The University of Melbourne
Dr George YX Kong
MBBS BMedSci PhD FRANZCO
Centre for Eye Research Australia,
Royal Victorian Eye and Ear Hospital
Ophthalmology Department of Surgery, The University of Melbourne
Visual field examination is the premier vision test in the diagnosis and management of glaucoma and other ocular and brain disorders; however, perimetric equipment is expensive and bulky, making an in-office procedure unattainable in many small practices.
Testing also requires routine clinical reviews, usually six-monthly, increasing the clinical burden of monitoring chronic eye disease. Overall, this leads to higher health-care costs and lengthy waiting lists that affect timely disease detection and patient management.
The lack of portability of bulky visual field devices also means that patients residing in aged-care centres or remote and rural areas may not receive the testing they need, compromising the detection or monitoring of their condition and delaying therapeutic intervention.
Melbourne Rapid Fields (MRF) is a suite of applications that allows visual field testing on a tablet or mobile device. MRF has been validated as a tangent perimeter and can produce reliable human visual field thresholds out to 30 degrees.1 It is robust to blur, changes in ambient light and viewing distance, although glare or reflections off the screen can have an impact on thresholds.1
Tablets are low-cost, portable and ubiquitous within the Australian population where 79 per cent of those aged 65 years or older report that they access the internet with a tablet or mobile device.2 This technology is therefore familiar to many patients who may develop chronic eye disease and it has the advantage that it can be used to self-monitor these conditions at home as well as by clinicians who need a portable perimeter.
The MRF apps are produced by a consortium that includes The University of Melbourne, The Royal Victorian Eye and Ear Hospital, glaucoma specialist and IT inventor Dr George Kong and optometrist and clinical scientist Professor Algis Vingrys.
Three apps of the MRF have been developed to target specific eye or brain conditions: MRF Glaucoma, MRF Macula and MRF Neural. (MRF Diabetes is planned for release soon). Each app has one or more near-acuity tests (Landolt C) and a visual field test grid tailored for the specific application. For example, although MRF Glaucoma and MRF Neural test the macula region, AMD is better detected and managed using the MRF Macula app which has been specifically designed for this purpose.
To provide a full suite of vision assays, each MRF app contains a specially-designed acuity test (Figure 1). As this is achieved at near (33 cm) with normal reading spectacles (if any), it might have some limitations for returning precise acuity outcomes in myopes and younger adults, although testing in Africa reports accurate and reliable outcomes compared to an Early Treatment Diabetic Retinopathy Study chart in adults.3
Figure 1. Acuity targets used in the MRF applications target three discrete channels. A high-contrast target with contour interaction for standard acuity measures in the resence of optical blur; a low contrast, low luminance target to detect lower-order contrast processing deficits (retina, nerve); and an optotype in noise that can be used to detect integrative deficiencies (brain, for example: amblyopia).
Figure 1 is an example of the optotypes available for acuity testing in the MRF. These have been designed using the principles detailed by Westheimer4 who found three discrete channels mediate acuity resolution: a high contrast size-dependent channel, which most clinicians are familiar with for detecting optical blur; a low contrast, low luminance channel that targets retinal and nerve dysfunction (note that amblyopes show little effect to this target); and an optotype shown in noise designed for higher order (brain) disorders.
Visual field testing
A 30-degree field test is achieved by having the patient fixate on the corners of the tablet. Target size has been scaled (getting bigger) to give a constant threshold (30 dB) across the visual field. The test adopts a neighbourhood logic that checks points removed from the prediction of neighbouring locations. The test patterns are described next, including an expanding field that adds extra test points in regions of change to better define the edge of a scotoma.
This software has been designed to test the peripheral visual field out to 30 degrees by 24 degrees so it is ideal for glaucoma or any other condition needing peripheral visual field evaluation. Testing commences by thresholding a region out to 15 degrees by 12 degrees (9.7 inch screen tablet) then the test goes through a phase of moving the fixation point. This movement is to the four corners of the iPad, enabling the four quadrants of the peripheral field to be assayed (see below). The test grid is either a 66 point radial grid (Figure 2) optimised for glaucoma, neuropathy or macula defects or an extended 58 point 24-2 grid (Figure 2) with four extra points added in the fovea. Testing takes about four minutes per eye for the 9.7 inch screen or three minutes per eye for the large (12.9 inch) screen where fewer refixation movements are needed.
Figure 2. Radial (blue diamonds) and extended 24-2 (red circles) test grids available in the MRF. MRF Neural does not test the far peripheral 24-2 locations (filled red circles). MRF Macula uses the grid shown in the inset (top right, green diamonds) with the box scaled to 20 degrees each side (10 degree eccentricity).
This test was designed for macula conditions such as AMD or macula oedema. It is similar to the 10-2 pattern and evaluates the macula 10 degree region with a 33 point grid. It requires about 1.5 minutes per eye for testing. An example of a macula scotoma found with this test is shown in Figure 3.
Figure 3. Example field results in two patients tested with the 33-point MRF Macula grid (left) and 66-point MRF Glaucoma grid (right). As can be seen in the left panel, the MRF Macula test added eight extra test points to the grid shown in Figure 2 and required 2 minutes 27 seconds to define the central scotoma. On the right, the MRF Glaucoma test needed 5 minutes 3 seconds to define the left superior quadrantic loss in this patient.
This test was optimised for the large screen iPad Pro (12.9 inch) and neural conditions. It presents 46 points in a reduced 24-2 grid (21 degrees by 15 degrees, Figure 2) with four extra fovea points. The test is performed using central fixation only (no need for fixation changes) and requires about 2.2 minutes per eye for testing.
An Apple iPad generation 2 or newer device with a retina screen and iOS version 8 or higher is required to run the MRF software. An iPad mini (7.9 inch) or iPad3/4/Air (9.7 inch) may be used to run MRF Macula. MRF Glaucoma and MRF Neural require an iPad3/4/Air (9.7 inch) or the larger iPad Pro (12.9 inch) screen for testing.
To perform an examination, the user is required to be seated in a dimly lit room free from reflections off the device’s screen.1 The patient’s habitual reading correction should be worn and they must take care to ensure they are positioned approximately 33 cm from the screen. This is best checked with a calibrated string (a software process using the iPad camera is under development). Patching of the fellow eye can be achieved by covering with a hand or by draping a tissue over the spectacle lens, if spectacles are worn.
When a test is commenced, voice commands guide the patient through the examination. A red viewing cross appears in the centre of the screen as a fixation target on a background of 5 cd/sqm. To respond to the stimulus, the user simply taps on the device screen to register that they have seen the spot. To avoid fingerprints on the screen and fingers from shielding a spot, a touch zone (red square) has been nominated (Figure 4).
Figure 4. A tablet showing the central fixation mark at the start of testing and the fixation pattern used during testing of peripheral locations. The response zone (red box) is shown in the lower right corner. Although the patient can respond by tapping the screen anywhere, they are encouraged to use the response zone to prevent the screen from becoming soiled and their fingers obscuring a spot location. The white spots on the screen show a schematic of the size scaling used in testing across the visual field.
Stimulus responses may also be recorded by tapping the space bar of a Bluetooth keyboard if this is available. The keyboard is a recommended method of interfacing with the software and can be activated or deactivated by selection in the test menu. As the examination progresses, the fixation cross may move to each corner of the device to test peripheral locations of the visual field (Figure 4).
Thresholding takes place using a Bayes prediction that returns one of seven different outcomes (0-30 dB) after three presentations.1 Reliability indices are assayed during the test (fixation loss, false positive, false negative) using a volley sampling method and patients who return error rates greater than 30 per cent should be considered as unreliable.
The results in Figure 3 are shown using familiar formats for the clinician: a grey scale plot, pointwise dB-plot, total deviation and pattern deviation plots with probability shading. The statistics that are calculated by the MRF are the mean defect (MD), the pattern defect (PD) and visual capacity (VC, is similar to Visual Field Index). A clinical trial has found very high concordance between the Humphrey Field Analyzer and the MRF MD indices (r = 0.91) despite the differences (test spot location, spot size, test procedure and background brightness) between these two devices.5 This means that the results of the MRF can be confidently used in the management of glaucoma patients.
Once a test has been completed, results are stored via a cloud portal where artificial intelligence analyses the data and is able to detect changes in vision and alert the managing eye-care professional if these are present. This process needs five tests to define baseline threshold and variability and any changes are confirmed before any notification is made.
We routinely ask our patients to do MRF testing every day for the first week to establish baseline.
Given the lower cost and portability of these devices, patients now have the option to become involved in the management of their chronic vision loss by undertaking visual field examinations in the comfort of their own home. We have found that weekly home-monitoring by the patient has the capacity to detect progression in one-third of the time needed by regular six-monthly in-clinic reviews (Figure 5).4 In addition, doctors practising in remote or rural locations can take advantage of the portability of the MRF and provide patients with an optimal standard of care.
Figure 5. Predicted diagnostic power for weekly (at home) and six-monthly (in clinic) reviews for patients who have fast progression (-2 dB/year) in their visual field loss and comply with the requested test frequency.6
MRF is available to download from the Apple App Store. There is an initial download fee of $38 charged by the App Store, which allows six months use of the software as part of the download package. An on-going annual fee of $120 is charged for continued (12 months) use of the software. These fees go toward further software developments and in supporting community-based programs.
MRF comes in two forms: a regular and a ‘lite’ version. The lite versions have the same capacity as the regular versions but are free of charge and limited to eight saves only; they have been designed for practitioners or patients who want to gain familiarity with the software to see if it suits their purposes.
1. Vingrys AJ, et al. Validation of a tablet as a tangent perimeter. Transl Vis Sci Technol 2016; 5: 4: 3.
2. ACMA. Digital lives of older Australians. Research Snapshot [Accessed 4/8/17]. Available from: www.acma.gov.au. 2016.
3. Bastawrous A, et al. Development and validation of a smartphone-based visual acuity test (peek acuity) for clinical practice and community-based fieldwork. JAMA Ophthalmol 2015; 133: 8: 930–937.
4. Westheimer G. Optotype recognition under degradation: comparison of size, contrast, blur, noise and contour perturbation effects. Clin Exp Optom 2016; 99: 1: 66–72.
5. Kong YX, et al. A comparison of perimetric results from a tablet perimeter and Humphrey Field Analyzer in glaucoma patients. Transl Vis Sci Technol 2016; 5: 6: 2.
6. Anderson AJ, et al. Can home monitoring allow earlier detection of rapid visual field progression in glaucoma? Ophthalmology 2017. doi: 10.1016/j.ophtha.2017.06.028. [Epub ahead of print].