Abstract
Purpose::
Shifting from film to digital imaging, we have extended the Age-Related Eye Disease Study (AREDS) classification of geographic atrophy (GA) by using digital measurement tools, and by recording more features of GA morphology.
Methods::
GA is the atrophic form of age-related macular degeneration (AMD) causing significant vision loss in the elderly. The AREDS classification (Report 6, Am J Ophthalmol 2001; 2001;132:668-81), similar to the international classification (Surv Ophthalmol 1995;39:367-74), describes GA with a categorical scale indicating absence/presence and extent (six steps defined by graduated circles), and proximity to the macular center (by concentric region of the AREDS grid). We developed an extension with digital tools: planimetry of GA area and measurement of its macular proximity. Annual images of 123 AREDS participants with GA were evaluated longitudinally by four graders using this new extension (data on change in GA presented in a separate poster). Slides were digitized, then displayed and evaluated using Topcon IMAGEnet. To test reproducibility, 12 eyes (10%) with multiple visits (N=82) were selected for masked replicate measurement by a different grader.
Results::
Comparison of GA data between the original and replicate gradings yielded the following intra-class correlation coefficients: GA area, 0.99; macular proximity, 0.72. For AREDS 2, the AREDS 1 classification is being updated to use digital measurement. As the poster will illustrate, it is also being extended further to record the degree of atrophy observed (mild, moderate and severe) and the shape of the GA lesion according to the criteria proposed by Sunness and coworkers (Ophthalmology 1999; 106:1768-1779): small, multifocal, horseshoe, ring, solid, and indeterminate.
Conclusions::
To improve evaluation of GA from retinal photos, the AREDS classification has been extended for AREDS 2 by adding digital measurement and by recording more features of GA morphology. Computerized measurement compares favorably to manual estimation, producing data that are continuous (rather than categorical) and highly reproducible.
Clinical Trial::
www.clinicaltrials.gov 1
Keywords: age-related macular degeneration • clinical (human) or epidemiologic studies: natural history • clinical (human) or epidemiologic studies: outcomes/complications