Foveal vs. Total Geographic Atrophy as a Predictor of Visual Acuity in AMD

Abstract

Purpose : To understand the relationship between total geographic atrophy (GA), foveal GA and visual acuity (VA) in age-related macular degeneration (AMD), and to explore potentially more sensitive anatomical predictors of VA in GA.

Methods : Retrospective review of blue autofluorescence (AF) fundus images, optical coherence tomography (OCT, Heidelberg Engineering) and Snellen VA of patients with GA due to AMD. The foveal area, 1.77 mm², was defined by a 1.5 mm diameter circle centered on the fovea centralis. Using the Heidelberg built-in software, two graders independently measured the total GA area and the GA area within the fovea (Fig. 1a) and correlated it with VA. The average between the two graders was computed, except when two graders disagreed by more than 20%, in which case a third grader was used for adjudication. In addition, we graded for the presence of foveolar sparing, GA covering the entire foveal area, subretinal drusenoid deposits (SDD) and multifocal GA.

Results : 18 patients (mean age, 82.1±6.6 [70-96] years, 83% female) with GA due to AMD were assessed for data spanning n=39 eyes spread across up to 3 visits. Mean VA at time of imaging was 0.86 (Snellen 20/145) ±0.6 [0.2-3.0] logMAR, mean total GA 9.56±4.78 [3.96-20.04] mm², mean percentage GA within the fovea was ~79±24 [23-100] %. About 97% of assessed eyes had SDD and ~54% had multifocal GA.
Percentage GA within the fovea was better associated with VA (R²=0.3) than total GA area (R²<0.01) (Figs. 1b, 1c) showing an exponential decline in logMAR VA with increasing percentage of foveal GA: on average, VA remained better than 20/60 for GA lesions covering less than ~50% of the fovea, while a sharp decrease in visual acuity was observed for larger percentages. Eyes with GA covering the entire foveal area (~10%) were associated with poor (<20/200) VA, while prediction of VA in patients with foveolar sparing (~8%) was inconclusive, possibly due to the low occurence of this feature in our cohort.

Conclusions : Our findings suggest that the percentage GA within the fovea may be more sensitive in predicting the degree of VA impairment than the total GA area or foveolar involvement. Further validation is needed in larger cohorts.

This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.

 

Figure 1. (a) Representative blue autofluorescence fundus image outlining geographic atrophy (GA) within the fovea. (b) Scatterplot of logMAR visual acuity (VA) vs. percent foveal GA. (c) Scatterplot of logMAR VA vs. total GA area in mm2.