The initial growth rate of geographic atrophy lesions appears to strongly predict future growth rate

Abstract

Purpose : To assess the prognostic potential of the initial growth rate of geographic atrophy (GA) on future growth rate and its utility in clinical trial design (Sunness, J.S., et al. Ophthalmology. 2007).

Methods : A retrospective analysis was performed using the biannual fundus autofluorescence (FAF) and monthly spectral domain optical coherence tomography (OCT) image data from patients in the sham study arms of the MAHALO (NCT01229215) study. Initial annualized growth rates were obtained in two ways: (1) based on the reading center (DIRC) assessment of GA lesion areas at baseline and month 6 on FAF, and (2) based on GA lesion areas derived from OCT at baseline and month 3. These initial rates were compared to the growth rate over the subsequent 12 months, using the month 6 and 18 FAF assessments. To derive the GA lesion area on OCT, en face choroidal images were generated, and a semi-automated algorithm was used to identify GA following affine registration of the month 3 image to baseline (Figure 1A to 1D). Patients with OCTs with significant motion artifacts were excluded from the analysis.

Results : Twenty sham eyes had FAF reads and gradable OCTs. At baseline, GA lesion area from FAF and OCT showed a Pearson correlation coefficient (R) of 0.99. The correlation between the initial 6-month, annualized FAF growth rate and the subsequent 12-month growth rate was 0.84. The correlation between the initial 3-month, annualized OCT growth rate and the month 6 to 18 FAF growth rate was 0.89. Splitting the data by the median initial growth rate results in subgroups with lower growth rate coefficient of variation (CV) (Figure 1E).

Conclusions : The initial growth rate of GA lesions strongly predicts future growth rate in the sham patients of MAHALO, suggesting that the precision of treatment effect estimates for GA would improve when pre-treatment growth rates are included as baseline covariates in study analyses. Furthermore, subgroups based on initial growth rate may have a lower growth rate CV than the original population, which may impact statistical power.

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

 

Figure 1. En face choroidal images from a patient at (A) baseline and (B) month 3 with corresponding GA segmentation (C, D). (E) Plot of annualized initial growth rate versus annualized month 6 to 18 growth rate depicting subgroups (blue, red) with lower CV than the original group.