Predictors for Growth of Geographic Atrophy over 12 Months

Abstract

Purpose : To evaluate the relationship between characteristics of geographic atrophy (GA) lesions at baseline and Month 6, on the growth rate of atrophy over 12 months.

Methods : A total of 82 eyes of 82 subjects with geographic atrophy (GA) and 12 months of follow-up data at the Doheny Image Reading Center were included in this retrospective analysis. All subjects had undergone blue light fundus autofluorescence (FAF) imaging at baseline, Month 6, and Month 12. To be included in this analysis, the total GA lesion area was required to be between 2.54 and 17.78 mm², and in multifocal cases, at least one of the lesions was required to be larger than 1.27 mm². Areas of definite decreased autofluorescence (at all visits) corresponding to GA were manually segmented using planimetric reading center software. Lesion location (subfoveal vs nonfoveal), number, morphology (multifocal, horseshoe, ring, single), perimeter, and circularity index were measured at baseline and at month 6. Development of new/distinct lesions, as well as merger of two or more lesions at Month 6, were also assessed. These lesion characteristics at baseline and month 6 were correlated with the month 12 GA growth rate by linear regression analysis.

Results : Of 82 eyes included in this analysis, 50 were multifocal. Baseline GA area and annual growth rate of GA were 8.06 ± 4.06 mm² and 2.02 ± 1.13 mm² respectively. The average number of atrophic lesions at baseline was 3.8 ± 3.7 and this decreased to 3.5 ± 3.5 at month 12 (p = 0.58). The average perimeter of the lesions at baseline was 20.61 ± 10.90 mm and this increased to 20.74 ± 10.42 mm at month 6, but the difference was not significant (p = 0.94). In addition, the increase in CI from 0.34 ± 0.21 at baseline to 0.37 ± 0.21 at month 6 was also not statistically significant (p = 0.41). Figure 1 illustrates the Descriptive Regression Coefficients for GA area growth at month 12. Factors associated with a faster enlargement of GA are shown in Figure 2.

Conclusions : Although many baseline lesion characteristics can influence the enlargement rate of GA, their predictive value is small compared to the prior growth rate. This suggests that a “run-in” period prior to randomization may be useful for standardizing treatment arms in interventional trials.

This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.

 

Table 1: Descriptive Regression Coefficients for month 12 enlargement of GA

 

Table 2: Multivariate linear regression analysis for GA enlargement at Month 12