Abstract
Purpose :
For the past 15 years, adaptive optics scanning light ophthalmoscopy (AOSLO) has allowed noninvasive, routine observation of the human cone mosaic and enabled quantification of cone density versus retinal eccentricity in both normal and disease retina. Despite the potential for AOSLO to quantify disease progression at the cellular level, there remain few longitudinal studies investigating changes in cone density as a measure of disease progression. Here, we undertake a prospective, longitudinal study to investigate the variability of cone density measurements in normal subjects over two years.
Methods :
The cone mosaics of fourteen eyes of nine subjects ages 18 to 54 with no known pathology were imaged along the temporal meridian at baseline and two years using confocal AOSLO. Images from each study visit at five retinal locations (190, 350, 500, 900, and 1500 µm from the fovea) were manually aligned and square regions of interest (ROIs) 85 µm per side were cropped from each aligned image, resulting in 140 ROIs. Semi-automated custom software was used to identify cones in each ROI (graded blind and in random order) and cone density was extracted from each image. Cone density at baseline and at 2 years was compared for each retinal location. The inter-eye correlation was accounted for by using generalized estimating equations.
Results :
Mean (standard deviation) baseline cone densities were 87,300 (14,000), 61,800 (9,400), 46,800 (7,500), 29,300 (3,200), and 17,600 (2,600) cones/mm2 at 190, 350, 500, 900, and 1,500 µm, respectively. The mean difference (95% confidence limit for the agreement) between 2 year and baseline cone densities were 200 (-14,600 – 15,100), 300 (-5,700 – 6,300), 1,300 (-2,400 – 4,900), 900 (-1,600 – 3,300), and 1,300 (-3,900 – 6,600) cones/mm2 at 190, 350, 500, 900, and 1,500 µm, respectively. These changes were not statistically significant (p-values: 0.92, 0.83, 0.58, 0.42, and 0.09).
Conclusions :
While a measurable change in cone photoreceptor density did not occur over two years in our subjects, high variability of changes in cone density measurements between two time points was observed, despite careful image alignment between time points. This inter-visit variability must be considered when planning prospective longitudinal clinical trials using changes in cone density as an outcome measure for assessing retinal disease progression.
This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.