Chapter 45

Fundus Autofluorescence Imaging

in Age-Related Macular Degeneration

and Geographic Atrophy

Srilaxmi Bearelly and Scott W. Cousins

Abstract The traditional method for documenting and quantifying geographic atrophy (GA) is color photography. This method has been shown to be reproducible in several clinical trials, including the Age-related Eye Disease Study (AREDS) and the natural progression of GA studies by Sunness et al. (AREDS No. 6, Am J Ophthalmol 132(5):668–681, 2001; Sunness et al., Invest Ophthalmol Vis Sci 40(8):1761–1769, 1999). Nevertheless, it can be difficult to distinguish between dead/nonfunctioning retinal pigment epithelium (RPE), living but depigmented RPE (RPE often release melanin granules upon injury), and yellowish coloration caused by large drusen or calcified regressed drusen. Two imaging technologies that seem promising are fundus autofluorescence (FAF) and spectral domain (high resolution) optical coherence tomography (SDOCT). Here we provide an overview of FAF imaging in the setting of age-related macular degeneration (AMD) and GA.

45.1 Background

Approximately 10 million Americans and 60 million individuals worldwide exhibit some form of age-related macular degeneration (AMD). Most affected patients manifest early AMD, characterized by sub retinal pigment epithelium (RPE) deposits called drusen. However, almost 20% of AMD patients progress to one of two late forms of AMD: geographic atrophy (GA) and neovascular AMD, both of which are associated with severe vision loss. The prevalence of GA and neovascular AMD are similar. Approximately 1.2 million individuals manifest neovascular AMD in at least one eye, but approximately 973,000 individuals exhibit geographic atrophy in at least one eye (Friedman et al. 2004). These prevalence rates are likely to double

S. Bearelly (B)

The Duke Center for Macular Diseases, Duke Eye Center, Durham, NC, USA e-mail: beare002@mc.duke.edu

Medicine and Biology 664, DOI 10.1007/978-1-4419-1399-9_45,C Springer Science+Business Media, LLC 2010

by the age 2030. Since poor vision may result from either form of late AMD, novel approaches to management and treatment of GA are important.

The traditional method for documenting and quantifying GA is with color fundus photographs (CFP). This method has been shown to be reproducible in several clinical trials, including the Age-related Eye Disease Study (AREDS) and the natural progression of GA studies by Sunness et al. (AREDS No. 6 2001; Sunness et al. 1999). Nevertheless, it can be difficult to distinguish between dead/nonfunctioning RPE, living but depigmented RPE (RPE often release melanin granules upon injury), and yellowish coloration caused by large drusen or calcified regressed drusen. In addition, the predictive sensitivity of color photos is poor, and is capable of identifying only 5–7% of eyes that progress to late stages of AMD in 5 years (Klein et al. 1997). An imaging technology that seems promising in measuring and predicting GA is fundus autofluorescence (FAF) imaging (Fig. 45.1).

If FAF imaging is more reproducible and accurate as compared with the current gold standard, color fundus photos, this could potentially have a major impact on how we measure and follow GA in future prospective trials. Imaging could help stratify slow and fast progressors, and thus enable smaller trials with shorter duration and enhanced power. Without the development of methods for stratification of risk, trials for treatment of early AMD and GA will remain prohibitively large and expensive, making it difficult to test efficacy of novel therapeutic approaches.

45.2 Fundus Autofluorescence Overview

FAF is a photographic technique that measures emitted fluorescent light from the retina after excitation with 488 nm light. Emission is detected above 500 nm with a barrier filter. Although standard fundus cameras equipped with appropriate barrier filters can detect autofluorescence, the Heidelberg Retinal Angiograph, which is a confocal scanning laser ophthalmoscope (cSLO) equipped with excitation 488 nm solid-state laser and Heidelberg image analysis software, can register and average multiple FAF images. Typically from 9 to 15 single images are averaged in order to amplify the FAF signal. The source of FAF is not completely understood. Delori et al. demonstrated with spectrophotometric investigations that lipofuscin granules in the RPE monolayer contain the dominant fluorophores responsible for FAF imaging (Delori et al. 2001, 1995). Other potential fluorescent structures include photoreceptors (rhodopsin), and autofluorescence of vitreous and lens. The confocal nature of the HRA, however, limits the autofluorescence from other planes of the eye, and maximizes the signal from the plane of interest.

Figure 45.1 demonstrates the appearance by FAF and grayscale CFP with (a) normal macula, (b) multiple, large coalesced drusen consistent with AMD, and (c) geographic atrophy. The normal macula has a decrease in the FAF signal concentrically in the macula, which corresponds with the increased lutein and zeaxanthin

concentration. Drusen themselves do not typically correspond exactly to the autofluorescence in AMD. Different patterns of FAF have been described in early dry AMD, as described below.

A discrete area of GA is noted superotemporal to the fovea in image C of Fig. 45.1. This area of RPE death is easily demarcated on FAF as black (or absent autofluorescence). While CFP does show the demarcated area of GA with visibility of the underlying choroidal structures, the extent of GA especially inferiorly becomes obscured by depigmentation. This depigmentation of the RPE inferior to the fovea could be mistaken for dead RPE, and an overestimation of GA is possible. Such a patient may show no difference in progression by CFP, when in fact there is (ß error). The FAF image also shows an area of focal increase in autofluorescence around part of the perimeter of GA (arrows) or ‘rim area focal hyperautofluorescence (RAFH)’ indicative of increased lipfuscin load in these RPE cells.

45.3 FAF Findings in Early AMD with Drusen Only

Analysis of FAF patterns in the setting of non-neovascular AMD (NNVAMD) performed by Einbock et al. suggests that risk of progression may be stratified by pattern type. The International Fundus Autofluorescence Group has identified eight patterns: minimal change, focal increase, lace-like, reticular, speckled, patchy, linear and plaque-like patterns. While this is a qualitative approach to interpretation, patchy FAF may predict a higher risk of progression to neovascular changes, focal or plaque-like FAF in the macula may result in progression to geographic atrophy (Einbock et al. 2005). This heterogeneity of patterns may reflect underlying differences in cell kinetics and metabolism. Hyperpigmentation, seen clinically as clumping of pigment in the posterior pole, typically exhibits a higher level of FAF which is thought to be secondary to a higher level of RPE lipofuscin (Lois et al. 2002). The lack of correspondence between distribution of drusen and FAF supports the conclusion that drusen and autofluorescence represent independent measures of aging in the posterior pole. These FAF findings may exceed funduscopically visible alterations, suggesting that changes at the level of the RPE precede the occurrence of any visible lesions.

45.4 FAF Findings in Late AMD with Geographic Atrophy

It has been theorized that increased FAF may precede the enlargement of preexisting atrophy and the development of new atrophy over time (Holz et al. 2001). This supports the observation that excess lipofuscin in the RPE may be key in the progression of GA. In addition, the areas of increased FAF outside of GA may be associated with variable loss of retinal sensitivity as measured by fundus perimetry, thereby suggesting a functional correlate of excessive lipofuscin accumulation