Multimodal retinal image analysis of de novo geographic atrophy development secondary to age-related macular degeneration

Abstract

Purpose : To characterize early stages of atrophy development and to determine the influence of precursor lesions on subsequent progression rates of geographic atrophy (GA) secondary to age-related macular degeneration (AMD).

Methods : Based on a retrospective database analysis, 36 GA eyes of 36 subjects (mean age 73.03 years) were identified that had been assessed before GA development. Local structural precursor lesions before GA development were evaluated by multimodal retinal imaging. Characteristic and specific features of subsequent progression of individual GA lesions were determined. The topographic progression rates of evolving GA were assessed by fundus autofluorescence imaging using the RegionFinder software and evaluated for associations with the predominant precursor phenotypes.

Results : The most common predominant local precursor lesions were large drusen (n=19), subretinal drusenoid deposits (SDD, n=8), followed by retinal pigment epithelium detachments ([PED], minimum horizontal diameter 1750 µm, n=5), and refractile deposits (n=4). For early GA lesions, the median time for progression from larger than 0.05 mm2 to larger than 0.20 mm2 and from larger than 0.05 mm2 to larger than 1.25 mm2 was 11.7 months (range 2-38) and 28.4 months (range 11-58), respectively. The predominant precursor lesion had a significant influence on the subsequent square-root GA progression rate (p<0.005), which was on average 0.64 ± 0.5 mm/year with the most rapid mean enlargement rate in eyes with SDD (1.11 ± 0.49 mm/year). A subset of these eyes exhibited the “diffuse-trickling” phenotype (3/8). Further, GA development in association with PED collapse was associated with initial rapid atrophy progression within the area, followed by slow GA enlargement outside the area of the former PED.

Conclusions : These findings underscore distinct characteristics in GA occurrence and further lesion size progression depending on the specific predominant precursor lesion along earlier stages of GA development. Analyses based on multimodal imaging may allow for identification of structural biomarkers to predict variable progressive visual loss in AMD due to atrophy development.

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