Cone spacing measures in visually normal eyes imaged at baseline and 12 months

Abstract

Purpose: To determine the inter-visit and inter-grader variability of Adaptive Optics Scanning Laser Ophthalmoscopy (AOSLO) derived cone spacing measures in normal eyes monitored longitudinally over 12 months. Analysis of cone spacing measures in normal eyes will establish the variability of cone spacing as an outcome measure for use in clinical studies of photoreceptor change utilizing AOSLO.

Methods: Eight visually normal patients underwent AOSLO imaging at 3 visits; 2 baseline visits were separated by no more than 1 month, and one visit occurred 12 months after the baseline visits. Cone spacing was measured in each image by two independent graders. Cone spacing measures were compared between visits and graders, and then correlated with standard macular measures including visual acuity, foveal sensitivity, Goldmann perimetry, and spectral-domain optical coherence tomography (SD-OCT)-derived outer segment layer thickness.

Results: Cone spacing was measured in 424 regions within 3.5 degrees of the fovea. Agreement between graders was strong (ICC = 0.87, 95% confidence interval (CI) 0.81-0.90); mean difference between graders was 0.07 arcminutes (95% CI 0.05-0.08). Median absolute difference between cone spacing measures at each ROI between the 2 baseline visits was 0.047 arcminutes (95% CI=0.04-0.059). Cone spacing at 12 months decreased by -0.0039 arcminutes (95% CI: -0.01 to 0.0023), consistent with no change over time in this normal population. Similarly, there were no significant changes in any clinical measures of retinal degeneration.

Conclusions: A high degree of inter-visit and inter-grader agreement of AOSLO-derived cone spacing measures was observed in visually normal eyes examined twice at baseline and at 12 months. The reproducibility of cone spacing measures in normal eyes lends support to the future use of AOSLO in clinical trials studying photoreceptor change over time.