Purchase this article with an account.
Sarah B. Sunshine, Elvira Agron, Juliet Hartford, Frederick Ferris, Emily Y Chew; An Improved Method for Measuring Geographic Atrophy (GA) in patients with Age-related Macular Degeneration (AMD). Invest. Ophthalmol. Vis. Sci. 2016;57(12):1612a.
Download citation file:
© 2017 Association for Research in Vision and Ophthalmology.
Recent studies have utilized fundus auto-fluorescence (FAF) and color photography (CF) imaging to monitor progression of GA as an outcome variable. Grading of GA area has been done manually and using semi-automated techniques. However, these methods have led to difficulties with reproducibility and within person variance. Reduction in the variance in GA area measurements in clinical trials would increase power to assess treatment effects for AMD.
One approach to minimizing the within person variance in estimating growth rates of GA is taking into account all possible information available at the time of estimating GA area. This includes simultaneous use of FAF and CF information for grading GA area at each time point. In addition, the grader can look at previous images and future images to make the best possible estimate of where the GA border is at each study visit. Notwithstanding the potential for bias, graders make the assumption that GA did not spontaneously resolve. Registering the images leads to concentrically expanding rings to depict the progression of GA over time, allowing for both estimates of GA area over time and estimates of growth rates along specific merridians. This approach has been applied to images obtained from study participants with GA and at least 4 years follow-up at two AREDS2 clincs. To date, 48 eyes (27 participants) with 4-8 years of imaging have been preliminarily evaluated using this methodology.
Sequential GA areas were obtained from baseline and annual imaging (FAF, CF). GA area was calculated from the grading of the images using side-by-side FAF and CF images and comparison with previous and follow-up images. The slope of the growth rate was calculated using regression analysis of the square root of the area over time (mm growth per year). Variance of the growth rate was calculated using this “informed” sequential grading process and was directly compared with independent single image grading over time. Preliminary analysis suggests an important reduction in within person variance.
Using all information available on sequential visits appears to reduce the within person variance of GA growth rates. This approach will allow us to further elucidate the natural history of AMD and will allow clinical trials to more successfully characterize the effect of GA treatments.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.
This PDF is available to Subscribers Only