One of the key responsibilities for optometrists in primary care is the early detection of macular disease that may be the start of a choroidal neovascular incident (CNV). If promptly referred, patients may be assessed, usually by angiography, and, where appropriate, some treatment may be undertaken. The eye care pathway proposals for AMD management hinge on this decision.
At present there are a number of ways of detecting a CNV incident. As there is serous fluid under the retinal pigment epithelium in the macular area, there is a resultant distortion of vision (often reported as a 'kinking' of straight edges) and occasionally a hypermetropic shift. Patients may also notice changes in image size. The distortion is typically assessed using an Amsler grid which, combined with the symptoms and a view of the macula showing any RPE elevation, would be the basis for urgent referral. As with much in optometry, the scope for inaccuracy is significant. Concerns include:
Too much reliance on patient symptoms for initial presentation
Amsler charts suffer from inacuracy due to poor fixation, crowding and the 'completion phenomenon' (the tendency for the perceptual system to 'fill in the gaps' when viewing a regular grid pattern)
Fluorescein angiography is both expensive and induces adverse side-effects in many patients.
Zeiss has recently introduced the Preferential Hyperacuity Perimeter (PHP) which is designed to detect CNV incidents in those at risk of them occurring.
Preferential Hyperacuity Perimetry
Hyperacuity (or Vernier acuity) is the ability to resolve minute differences between two or more stimuli which may be as small as three to six minutes of arc difference in retinal subtense. The PreView PHP uses this sensitivity by presenting on a screen a sequence of linear test patterns. A normal retina will perceive these line targets as straight, whereas eyes with a diseased retinal pigment epithelium, extensive drusen or early CNV will perceive the image as having shifted from its actual position. This image relocation is the reason for metamorphopsia as detected by an Amsler grid.
The PHP presents a small line target to a patient (corrected for reading) on a screen for 160ms. The patient is encouraged to touch the screen with a stylus wherever they feel they saw a distorted line. A dot appears where the patient touched the screen and this dot then moves towards the central fixation dot. Once the dots fuse, a new line flashes up on the screen at a different location. In all, around 100 line targets are presented which takes around five minutes per eye.
The data is presented by means of a thermal printout at the end of each test. Figure 1 shows the plot for the eye of a 76-year-old with established atrophic AMD who had demonstrated no visual symptoms. The machine concludes that, though there is a central field defect detected, this was not consistent with CNV. In other words, the defect was a static one and the image shift was not as expected with an exudative change.
Just as with many other automated field screener, there is also some indication of the reliability of the plot. The low rates of false positives and negatives indicate good reliability. Figure 2, on the other hand, suggests a level of hyperacuity reduction indicative of a CNV episode that would merit immediate referral for assessment and possible treatment.
Who might use this?
Obviously, if one was able to use this on elderly patients on a daily basis, the detection rate for CNV would be massively improved. This is, however, not a feasible option. Instead, the manufacturers recommend regular (six-monthly) screening of those patients already at risk of a CNV incident and trials have indicated that sufferers of established ('intermediate') atrophic AMD are more likely to have a CNV episode predicted by this approach.
optician will be reporting on hospital experience with the instrument in a forthcoming issue.
Bill Harvey describes a new instrument designed to allow earlier detection of CNV than is possible with standard optometric techniques