Effect of fundus tracking on perimetric test retest-variability

Abstract

Purpose : To compare test – retest variability of measurements from a fundus tracking perimeter, designed to compensate for eye movements, with and without tracking enabled.

Methods : Compass fundus perimetry (CMP; Centrevue, Padua) was performed on 33 people with glaucoma and 51 age related normals. Each participant was examined twice with tracking compensation on (Tracking ON) and twice with tracking off (Tracking OFF), randomized.
Sensitivity measurement at each visual field location was determined from a Bayesian threshold strategy (ZEST).
Test – retest variability was estimated globally using absolute difference between mean sensitivities (using Bland - Altman plots) and across all test locations (pointwise) using the mean absolute deviation (MAD).
A mixed effect model was used to examine differences between the two testing conditions and to explore the effect of fixation instability measured by Bivariate Contour Ellipse Area (BCEA) and pupil size.

Results : The 95% limits of agreement for test-retest mean sensitivity were reduced by 31% in the Tracking ON condition in glaucoma subjects and by 28% in normals (Figure 1).
MAD was significantly reduced in the Tracking ON condition both in normals (0.28 ± 0.18 dB difference, 13% reduction, p = 0.046) and glaucoma subjects (0.63 ± 0.17 dB difference, 21% reduction, p = 0.0004).
Pupil size and BCEA had no significant associations with MAD in either testing condition.

Conclusions : Fundus tracking reduces measured test – retest variability in perimetry. No association between fixation stability (BCEA) and variability could be detected in the Tracking OFF condition. Different metrics of fixation might help clarify its contribution. Pointwise MAD showed the greatest improvement, indicating a role of the tracking in obtaining repeatable localized stimulations. Fundus tracking might have a role in improving repeatability, facilitating progression detection.

This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.

 

Figure 1. Bland – Altman plots for Tracking OFF (in red) and Tracking ON (in blue) with 95% limits of agreement (shaded region) and mean difference (solid line).