Abstract
Purpose :
The gold standard to evaluate geographic atrophy (GA) progression in clinical studies is GA lesion area growth. Perilesional irregularities and focality affect disease advancement. Retinal pigment epithelium (RPE) at risk for atrophy could be assessed by the GA lesion perimeter. This retrospective analysis determined the potential use of GA lesion perimeter for the evaluation of GA progression in clinical studies.
Methods :
Color fundus photography (CFP) images of all participants (n=107) with GA from a global, multicenter, phase 2a study of Brimonidine DDS vs sham procedure were included in this retrospective analysis. Central GA lesion perimeters and areas were measured from CFP images using ImageJ software (v1.52a). Statistical analysis was conducted using R Software (v3.6.1) under R Studio.
Results :
GA lesion area was positively correlated with GA lesion perimeter. When lesions were assessed by baseline focality, the strong association persisted for unifocal lesions (r2 = 0.75, p =8.12e-102). However, the association for multifocal lesions between GA lesion area and perimeter was not as robust (r2= 0.43, p =6.29e-36). In multifocal lesions, the correlation between area and perimeter was further weakened for GA areas greater than the median area of the data (10.94 mm2). An evaluation of the CFP images indicated that this discrepancy can be explained by the presence of foci of relatively healthy tissue in the middle of the lesion that contribute to lesion perimeter or the shift of the overall lesion shape from irregular to circular. As the GA lesion transitions from multifocal to unifocal, the lesion perimeter can decline.
Conclusions :
There is a linear correlation between GA perimeter and area primarily in unifocal lesions. For multifocal lesions, the association appears to diminish as the GA area increases. Conversion from multifocal to unifocal lesions may result in perimeter reductions over time, which contrasts with the continuous, temporal increase in GA lesion area observed with disease progression. These temporal fluctuations in GA lesion perimeter complicate interpretation of a progressing disease like GA. Therefore, these data suggest that GA lesion perimeter alone is not a reliable marker for the evaluation of GA progression, especially for larger multifocal lesions. More information regarding the GA lesion focality and morphology is needed to interpret changes.
This is a 2020 ARVO Annual Meeting abstract.