Abstract
Purpose :
Flash color fundus photography has been the historical gold standard for assessment of geographic atrophy (GA), but has largely been supplanted by fundus autofluorescence due to its superior contrast and clarity for assessing the borders of atrophic regions. More recently, confocal color imaging has become available, which offers potentially better contrast. The purpose of this study is to compare the precision of these various methods for assessing the area of atrophic lesions.
Methods :
Twenty six eyes from 16 patients with GA secondary to age-related macular degeneration (AMD) were prospectively recruited in this IRB-approved study (written, informed consent) and imaged with fundus autofluorescence (FAF; Heidelberg Spectralis), confocal white-light color (CenterVue Eidon), and flash white-light color (Topcon Triton). Images were exported and reviewed by two masked, certified reading center graders who manually segmented the areas of GA. Images for individual modalities were reviewed separately without reference to other modalities. The mean difference in area between graders was computed and the intraclass correlation coefficients (ICC) between the two graders for each modality were analyzed using SPSS 23.0.
Results :
Among the 26 eyes, images were gradable for area in all three modalities for 21 cases, with a percent gradability of 88%, 96%, 81% for FAF, confocal color, and flash color, respectively. The mean area of the atrophic lesions in this cohort was 6.78 ± 6.17 mm2 (range: 0.49-21.53 mm2). The mean difference between graders was 0.50 mm2 for FAF, 0.03 mm2 for confocal color, and 0.81 mm2 for flash color. Paired T test with Bonferroni correction (α=0.05/3=0.0167) shown that the inter-grader differences are not significant between either two modalities However, the ICCs between graders for FAF are significantly higher than flash color fundus imaging (FAF, ICC 0.995, 95% CI 0.989-0.998; confocal color, 0.983, 0.958-0.993, and flash color, 0.953, 0.883-0.981).
Conclusions :
Confocal color imaging yields a high-level of gradable images and appears to facilitate precise and reproducible measurements of GA lesions, though FAF imaging still yielded the best agreement among graders. The feasibility of confocal color imaging as an outcome tool for GA studies warrants further investigation.
This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.