Purchase this article with an account.
Deborah Goren, Shaban Demirel, Casie Goldman, William Swanson, Stuart Gardiner; Accuracy of sensitivities measured by perimetry at damaged locations in subjects with glaucoma. Invest. Ophthalmol. Vis. Sci. 2013;54(15):3940.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Responses from primate retinal ganglion cells (RGCs) stimulated by perimetric stimuli saturate at contrasts beyond ≈15-20dB. Increasing contrast further should not appreciably increase the response besides effects of scattered light. If this holds true in humans, sensitivities from automated perimetry (SAP) in areas of significant damage would be essentially random, and so uninformative of the true threshold. We sought to assess the relation between SAP estimates and the true sensitivities at such damaged locations.
16 subjects with moderate to advanced glaucoma were recruited. For each subject, four visual field locations were chosen. Two of these had sensitivities between 6dB and 18dB on both of their last two clinic visits (averaged to give “perimetric sensitivity”), two were positioned to aid spatial uncertainty. The method of constant stimuli (MOCS) was used to generate Frequency-of-Seeing curves at each location using an Octopus perimeter externally controlled by the Open Perimetry Interface. 35 presentations were made at each of 7 intensities centered on the perimetric sensitivity, including the brightest available stimulus (equivalent to 3.75dB on the Humphrey Field Analyzer) at the two damaged locations. 25 blank catch trials were used to measure the false positive rate. A cumulative Gaussian curve was fit to the response probabilities at each location to estimate threshold (50% detection). The significance of the correlation between MOCS and perimetric sensitivities was determined using generalized estimating equations.
Among the 32 damaged locations tested (perimetric sensitivities 9-16dB), the correlation between MOCS and perimetric sensitivities was 0.173 (p=0.101). At 17 of these locations, the brightest stimulus was detected on <50% of presentations, indicating that true sensitivity was in fact below 3.75dB. Perimetry overestimated MOCS sensitivity at most locations. Above 16dB, a significant correlation was seen (r=0.837, p<0.001).
Below around 16dB, perimetric sensitivities are only weakly informative about the true functional status. They tended to overestimate true sensitivity, by a random amount. The effective dynamic range of standard perimetry may be limited by RGC saturation, not just the maximal luminance of the perimeter.
This PDF is available to Subscribers Only