Abstract
Purpose :
Pointwise perimetric sensitivities below ~20dB are unreliable due to excessive variability. Dynamically increasing stimulus size to increase detectability extends this dynamic range, allowing testing in more severe glaucoma. However, larger stimuli overlap multiple nerve fiber bundles so may miss smaller defects; and the bundles being stimulated differ even with perfect fixation. We propose instead using moving stimuli to increase detectability and hence increase the dynamic range of perimetry in glaucoma.
Methods :
A Size V moving stimulus was designed that travels parallel to the average nerve fiber bundle orientation at each location, with speed proportional to local magnocellular ganglion cell spacing (Fig 1). 34 locations across the visual field were tested with moving and static stimuli, using an otherwise identical Zippy Estimation by Sequential Thresholding (ZEST) seen / not seen algorithm on a clinical perimeter. 152 subjects with glaucoma / suspects were tested, of whom 52 were retested 6 months later. Resultant pointwise sensitivity estimates were compared using generalized estimating equation (GEE) regression. Test-retest variability for each stimulus was assessed by Bland-Altman plots, with limits of agreement adjusted for within-eye clustering.
Results :
The 152 subjects tested had average Mean Deviation -3.1dB from standard perimetry (interquartile range -3.7 to -0.3dB). Sensitivities using moving stimuli were 1.8dB higher; this difference increased with damage and eccentricity (both p<0.001, Fig 2A). Test-retest limits of agreement among 52 eyes were narrower for moving stimuli (-6.3 to +6.7dB, Fig 2B) than static stimuli (-12.7 to +7.9dB, Fig 2C), and test-retest variance was significantly smaller (p<0.001, F-test). 53% of subjects stated preferring the moving stimulus, vs. 24% who preferred the static stimulus.
Conclusions :
Using a moving stimulus for perimetry increases sensitivities, and hence locations stay within the dynamic range longer, allowing reliable testing at locations with more severe damage without the problems caused by increasing stimulus size. Results are more repeatable, and the test is preferred by most subjects. Work is underway to assess the normative range. Since stimulus speed is a continuous variable, a clinical test could either use moving stimuli at all sensitivities, or could increase speed proportional to loss ensuring that early detection would be unaltered.
This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.