Directional Kinetics of Geographic Atrophy Progression in Age-Related Macular Degeneration with Foveal Sparing Moritz Lindner, MD,1 Alexander Böker, MD,1 Matthias M. Mauschitz, MD,1 Arno P. Göbel, MD,1 Rolf Fimmers, PhD,2 Christian K. Brinkmann, MD, FEBO,1 Steffen Schmitz-Valckenberg, MD, FEBO,1 Matthias Schmid, PhD,2 Frank G. Holz, MD, FEBO,1 Monika Fleckenstein, MD,1 for the Fundus Autofluorescence in Age-Related Macular Degeneration Study Group* Purpose: To describe the directional kinetics of the spread of geographic atrophy (GA) spread in eyes with age-related macular degeneration and foveal sparing. Design: Prospective, noninterventional natural history study: Fundus Autofluorescence Imaging in Age-Related Macular Degeneration (FAM; clinicaltrials.gov identifier, NCT00393692). Subjects: Participants of the FAM study exhibiting foveal sparing of GA. Methods: Eyes were examined longitudinally with fundus autofluorescence (FAF; excitation wavelength, 488 nm; emission wavelength, >500 nm) and near infrared (NIR) reflectance imaging (Spectralis HRAþOCT or HRA2; Heidelberg Engineering, Heidelberg, Germany). Areas of foveal sparing and GA were measured by 2 independent readers using a semiautomated software tool that allows for combined NIR reflectance and FAF image grading (RegionFinder; Heidelberg Engineering). A linear mixed effect model was used to model GA kinetics over time. Main Outcome Measure: Change of GA lesion size over time (central vs. peripheral progression). Results: A total of 47 eyes of 36 patients (mean age, 73.87.5 years) met the inclusion criteria. Mean follow-up time was 25.216.9 months (range, 5.9e74.6 months). Interreader agreement for measurements of GA and foveal-sparing size were 0.995 and 0.946, respectively. Mean area progression of GA toward the periphery was 2.270.22 mm2/year and 0.250.03 mm2/year toward the center. Analysis of square rootetransformed data revealed a 2.8-fold faster atrophy progression toward the periphery than toward the fovea. Faster atrophy progression toward the fovea correlated with faster progression toward the periphery in presence of marked interindividual differences. Conclusions: The results demonstrate a significantly faster centrifugal than centripetal GA spread in eyes with GA and foveal sparing. Although the underlying pathomechanisms for differential GA progression remain unknown, local factors may be operative that protect the foveal retinaeretinal pigment epithelial complex. Quantification of directional spread characteristics and modeling may be useful in the design of interventional clinical trials aiming to prolong foveal survival in eyes with GA. Ophthalmology 2015;122:1356-1365 ª 2015 by the American Academy of Ophthalmology. *Supplemental material is available at www.aaojournal.org.

Geographic atrophy (GA) represents the late stage of dry age-related macular degeneration (AMD).1e5 Geographic atrophy is present in 3.5% of people older than 75 years6,7 and becomes the predominant type of AMD in the population 85 years of age and older.8 In industrial countries, late-stage neovascular or dry AMD is the leading cause of legal blindness in the elderly.9,10 Although various pathways have been proposed to be involved in the atrophic disease phenotype, the exact underlying pathophysiologic mechanisms are incompletely understood. Typically, patches of GA initially occur in the parafoveal retina. With spread over time, multifocal atrophic areas may coalesce, and new atrophic areas may occur. In advanced stages, GA areas may form a ring surrounding the intact and still functioning fovea. On clinical examination, the fovea

1356

 2015 by the American Academy of Ophthalmology Published by Elsevier Inc.

may remain uninvolved by the atrophic process until late in the course of the disease, a phenomenon referred to as foveal sparing.4,11,12 Geographic atrophy areas are associated with a corresponding absolute scotoma. Thus, the foveal-sparing pattern of disease evolution corresponds with progressive visual impairment that is characterized initially by reading difficulties resulting from parafoveal scotomata while the central visual acuity is still preserved.4,13e17 When the fovea finally becomes involved, a dramatic loss in central visual acuity occurs. However, visual impairments such as decrease in reading speed may start much earlier, putatively correlating with the size of the spared fovea.16,17 With the advent of confocal scanning laser ophthalmoscopy fundus autofluorescence (FAF) imaging it is possible to identify and quantify GA areas readily.18,19 Because of the http://dx.doi.org/10.1016/j.ophtha.2015.03.027 ISSN 0161-6420/15

Lindner et al



Progression of Foveal-Sparing in GA

absence of fluorophores in the retinal pigment epithelium (RPE), atrophic areas in GA eyes are associated with a severely reduced signal, resulting in high contrast between the atrophic and nonatrophic retina. Image analysis software has been developed and validated to quantify the size of atrophic areas and the spread over time.20,21 Because of luteal pigment, blue-light FAF (excitation, 488 nm) intensities typically are decreased in the fovea. Although atrophic patches exhibit an even lower FAF intensity than the central macula, judgment on foveal involvement based on FAF images only can be challenging (Figs 1 and 2). Additional use of near infrared (NIR) reflectance confocal scanning laser ophthalmoscopy imaging, spectral-domain optical coherence tomography imaging, or both have proven helpful in determining foveal involvement (Fig 1).21,22 A novel imaging analysis tool incorporated in the RegionFinder software (Heidelberg Engineering, Heidelberg, Germany) allows for combined NIR reflectance and FAF image grading. Using the region growth algorithm, the area of foveal sparing is outlined semiautomatically in the NIR reflectance image. Subsequently, the constraint delineating the foveal island is registered automatically to the corresponding FAF images to quantify GA areas outside the residual foveal island in the usual fashion.21 Recent studies have addressed the natural history of GA in detail, and several agents that may affect GA progression are in preclinical or clinical development (reviewed in Holz et al23). Patients in whom the fovea is affected by the atrophic process still would benefit because eccentric fixation abilities may be preserved and the enlargement of the scotoma size may be reduced. In patients with foveal sparing, preservation of both the fovea and parafoveal retina, even for a limited period, would have an enormous impact on the quality of life. Certain visual tasks necessary for activities of daily living, including recognizing faces and reading, are dependent on the fovea and parafoveal retina. If these areas could be preserved, patients would be able to maintain independent living, and quality of life, for longer. For the design of interventional trials aiming to preserve the foveal island in eyes with GA, the availability of natural history data is essential. Furthermore, a reliable technique that assesses GA progression toward the fovea is crucial. Therefore, the aim of this study was to quantify the directional kinetics of GA with foveal sparing using a semiautomated software tool that allows for combined NIR reflectance and FAF image grading.

Methods Patients Patients were recruited from the Fundus Autofluorescence Imaging in Age-Related Macular Degeneration (FAM) study (clinicaltrials.gov identifier, NCT00393692). This noninterventional, prospective natural history study followed the tenets of the Declaration of Helsinki and was approved by the institutional review boards of the participating centers. Informed consent was obtained from each participant after explanation of the study’s nature and possible consequences of participation. Patients were included in the current analysis with (1) unilateral or bilateral GA resulting from AMD with clear ocular media that

allowed for good-quality FAF and NIR reflectance imaging, (2) foveal sparing of GA in at least 1 eye, and (3) serial examinations of at least 6 months.

Definition of Geographic Atrophy Resulting from Age-Related Macular Degeneration Geographic atrophy was defined funduscopically as 1 or more well-defined, usually more or less circular patches of partial or complete depigmentation of the RPE, typically with exposure of underlying large choroidal blood vessels.24 In the FAM study, GA resulting from AMD was defined further as a sharply demarcated lesion with clearly reduced FAF of an extent of 0.05 mm2 or more (approximately 178 mm in diameter) that does not correspond to exudative retinal changes (e.g., bleeding, exudates, fibrous scar) in an eye with funduscopically visible soft drusen, retinal pigment abnormalities consistent with AMD, or both.25

Definition of Foveal Sparing in Geographic Atrophy In the current analysis, eyes were defined as having foveal sparing in GA if the residual foveal island was more than 270 surrounded by well-demarcated areas of GA. An eye was classified as having complete foveal sparing if GA surrounded the fovea in ring configuration. If there were bridges between the residual foveal island and the surrounding retina, an eye was classified as having incomplete foveal sparing. The definition and classification of foveal sparing were based on FAF and NIR reflectance images. Functional data were not included in this definition. If the 2 eyes of a patient met the inclusion criteria, both eyes were included in the analysis. Exclusion criteria included the presence of other retinal diseases such as diabetic retinopathy, present or past exudative AMD in the study eye, and retinal dystrophy, as well as a history of laser photocoagulation or retinal surgery. Furthermore, patients in whom the area of atrophy exceeded the 30 30 confocal scanning laser ophthalmoscopy images were excluded. Best-corrected visual acuity was determined with Early Treatment Diabetic Retinopathy Study charts on a quasilogarithmic ordinal scale. Before fundus examination, the pupil of the study eye was dilated with 1% tropicamide eye drops.

Image Acquisition, Processing, and Grading Fundus autofluorescence and NIR reflectance images were acquired using HRA 2 or Spectralis (Heidelberg Engineering). Fundus autofluorescence images were acquired with an excitation wavelength of 488 nm and an emission spectrum of 500 to 700 nm using the high-speed mode. Near infrared reflectance images were obtained at a wavelength of 820 nm. The field of view was set to 30 30 with a minimum resolution of 512512 pixels and was centered on the fovea. Single FAF images were aligned automatically and averaged to maximize the signal-to-noise ratio using the manufacturer’s software. Measurement of atrophy area and foveal sparing area were performed using the RegionFinder software (Heidelberg Engineering).21 The applied version (2.5.5.0) includes a newly implemented feature that automatically registers FAF to corresponding NIR reflectance images and allows easily toggling from one to the other modality (Fig 2). The size of the residual foveal island was determined semiautomatically in the NIR reflectance image. The reader manually sets a seeding point inside the foveal island to start the automatic region identification algorithm that also is used for atrophic area segmentation in FAF images. This algorithm is based on the

1357

Ophthalmology Volume 122, Number 7, July 2015

Figure 1. Foveal sparing in geographic atrophy visualized by different imaging methods: (A) fundus autofluorescence imaging (excitation wavelength, 488 nm), (B) near-infrared reflectance imaging, and (C) spectral-domain optical coherence tomography (vertical scan).

increase of the NIR reflectance signal at the site of atrophy compared with the low NIR reflectance signal of the island itself (Fig 2A). In case of an incomplete foveal sparing, a constraint was placed by manually drawing a line at the most narrow place (the bridge) of the incomplete part of the residual foveal island at baseline. This line also was copied to all follow-up images. The area size of the residual foveal island as determined in the NIR reflectance image was documented and blocked (using the “region to block” tool) for further analysis by the region identification algorithm, that is, for the next step when toggling to the corresponding FAF image for quantification of atrophic areas as previously described (Fig 2B).21 The final value that was analyzed further for both the size of the residual island and the total size of atrophy was defined as the average of the 2 measurements between both readers, provided that the 2 measurements did not differ by more than 0.15 mm2, respectively. If the difference exceeded 0.15 mm2, a senior reader additionally performed the measurement (M.M.M or A.P.G.). This latter measurement along with the closer of the 2 reader measurements then were averaged and taken as the final value. Area progression was assessed separately into 2 directions. The first was atrophy spread toward the fovea, that is, the centripetal progression of GA. This value was obtained by calculating the difference in area measurements of the foveal island at different time points. The second was atrophy spread toward the periphery, that is, the centrifugal progression of GA. This value was obtained by calculating the difference of atrophy area between different time points and subtracting the change in the area of the spared fovea that occurred during the same interval.

Statistical Analysis Data were compiled in a spreadsheet application and analyzed using SPSS software version 22 (IBM SPSS Statistics, Chicago, IL) and SAS software version 9.2 (SAS Inc, Cary, NC). Concordance between both eyes of a patient and interobserver agreement were assessed using the interclass correlation coefficient (ICC). Furthermore, the degree of interobserver agreement was assessed according to the descriptive method of Bland and Altman.26 For each patient, the differences between measurements (i.e., reader 1 [M.F.] and 2 [M.L.]) were calculated. A difference of 0 indicated perfect agreement. The limits of agreement were defined as the mean difference 2 standard deviations (observed standard deviation of the difference between the 2 measurements per patient). The ICC and Bland-Altman statistics were calculated on basis of the values obtained by each of the 2 first readers. To quantify central and peripheral progression rates within the patient collective, a linear mixed-effects model was used as described previously.27,28 The 2-level random effects model used here separates eye-specific and patient-specific effects and accounts

1358

for dependencies between measurements that were obtained from the same patient, eye, or both. A detailed description of the modeling process is given elsewhere.28 Comparison of progression kinetics of different areas (like centrifugal and centripetal GA progression) requires knowledge about the nature of the factors promoting this progression. Because these factors are not known for GA progression, we aimed to achieve comparability between progression toward the fovea and toward periphery by additionally modeling the square roots of GA and foveal-sparing areas. Comparing square rootetransformed areas is reasonable when contemplating foveal-sparing GA as a donut-like structure bordered by 2 circles of distinct radii and supposing that the enlargement of the atrophy area is proportional to the circumference of these circles, as recent studies suggest.29,30 If the enlargement was proportional to the circumference, square root transformation would enable comparison of centrifugal and centripetal progression rates.

Results A total of 47 eyes from 36 patients (7 men and 29 women; mean age, 73.87.5 years) with foveal sparing in GA were included in the analysis. This represents 23% of patients included in the FAM study with unilateral or bilateral GA resulting from AMD with clear ocular media that allows for good-quality FAF and NIR reflectance imaging and serial examinations of at least 6 months (n ¼ 158 patients). In eyes with foveal sparing, mean best-corrected visual acuity at baseline was 0.340.29 logarithm of the minimum angle of resolution units (range, 0.1 to 1.5 logarithm of the minimum angle of resolution units). Mean follow-up time was 25.216.9 months (range, 5.9e74.6 months). According to the classification of FAF phenotypes in GA,31 44 eyes were classified as diffuse. Of those, the diffuse-fine granular (18 eyes) and the diffuse-trickling (15 eyes) phenotypes were the most prevalent groups. At baseline, there were 7 eyes with complete (i.e., the atrophic area entirely surrounded the fovea in a ring configuration) and 40 eyes with incomplete foveal sparing (i.e., with bridges between the residual foveal island and the peripheral retina). Over the observational period, in 21 eyes, such bridges disappeared and a complete sparing developed. At last presentation, foveal sparing was still present and measurable in 39 eyes. Reasons for loss of follow-up of the other eyes were development of neovascularization (n ¼ 1), complete disappearance of foveal sparing (n ¼ 3), or inability to delineate clearly the residual foveal island because of insufficient image quality (n ¼ 4). The NIR reflectance and FAF images of all 47 eyes were registered successfully and processed at baseline using the RegionFinder software version 2.5.5.0. Fourteen follow-up visits were not included into the analysis either because of insufficient

Lindner et al



Progression of Foveal-Sparing in GA

Figure 2. Measurement of geographic atrophy and foveal-sparing size in the right eye of a representative patient using the RegionFinder software (Heidelberg Engineering, Heidelberg, Germany). A and B, Screenshots from the program’s interface obtained while performing the analysis. The unmodified fundus autofluorescence (FAF) image is shown on the left, whereas either (A) the corresponding near infrared (NIR) reflectance image or (B) the FAF image in process is displayed on the right. A, First, the NIR reflectance image mode is selected (highlighted with the red rectangle at top), and the area of foveal sparing is outlined and measured semiautomatically (red outlines). B, After applying the region to block function, creating a constraint outlining the fovea, the FAF image mode is selected. The generated constraint is still displayed in the FAF image (red outlines). The atrophy area can now be measured semiautomatically (white outlines indicate the borders of the large geographic atrophy lesion, the yellow outline indicates a small satellite lesion, and the red constraints at the superotemporal lesion border is set manually to block the lesion growth algorithm at a small vessel).

1359

Ophthalmology Volume 122, Number 7, July 2015 FAF or NIR reflectance image quality for grading, because of failure of registration of NIR reflectance and FAF images, or because the alignment of serial images resulted in a shift of the GA border out of the image frame. This resulted in a total of 159 visits (mean, 3.41.5 visits per patient; range, 2e7 visits) with quantitative values on foveal-sparing size and geographic atrophy size, respectively. At 8 of these visits (6 eyes), the residual foveal island was delineated directly in the FAF images because the NIR reflectance images did not add superior information for determination of the central atrophic border. The interreader agreement (ICC) was 0.995 for atrophy size and 0.946 for foveal-sparing size. The Bland-Altman plot for GA measurements (Fig 3A) displayed no systematic difference between the values obtained by the 2 readers. For measurements of foveal sparing, the Bland-Altman plot showed a slight tendency of reader 1 to measure the spared fovea larger than reader 2. This difference and the variance between the 2 readers increased with increasing size of foveal sparing (Fig 3B). Atrophy size and sparing size could be determined with high interreader agreement in eyes with complete as well as with incomplete foveal sparing: the ICC was 0.911 for foveal-sparing area and 0.998 for atrophy area in eyes with an atrophy entirely surrounding the spared fovea. In eyes with a nonatrophic bridge connecting the fovea with the peripheral retina, the ICC was 0.943 for foveal-sparing size and 0.991 for atrophy size. Area measurements differed by more than 0.15 mm2 in 95 visits for GA size and in 62 visits for foveal-sparing size, respectively. In these occasions, senior grading and arbitration was performed. The size of the atrophic area at baseline was 8.844.75 mm2, whereas the size of the residual foveal island was 1.480.88 mm2. Overall, atrophy progression, as determined by the linear mixed model, was 2.520.24 mm2/year. Centrifugal progression of GA was 2.270.22 mm2/year. In contrast, centripetal progression was 0.250.03 mm2/year (Fig 4). Square rootetransformation of the data revealed a centrifugal progression of 0.3190.024 mm/year, whereas centripetal progression was 0.1160.014 mm/year. Hence, on average there was a 2.8-fold faster atrophy progression toward the periphery than toward the fovea based on the square rootetransformed data. Overall, centrifugal and centripetal GA progression correlated with r ¼ 0.35 (P < 0.05; slope, 0.085; Fig 5). Although most eyes showed progression kinetics closely around the mean values given above, single eyes exhibited extreme values. Square root centrifugal progression ranged from 0.146 mm/year to 0.807 mm/year and square root centripetal progression ranged from 0.073 mm/year to 0.301 mm/year. Figure 6 gives 3 examples with ratios of centrifugal versus centripetal progression of 2.7, 1.5, and 5, respectively (Fig 6AeC, respectively). To assess if there was a difference in progression kinetics between GA with complete versus GA with incomplete foveal sparing, the variable complete sparing (yes vs. no) was added into the mixed model. Analysis revealed no significant differences in rates of progression toward the periphery (P ¼ 0.11) and toward the fovea (P ¼ 0.06) regarding area measurements. After square roote transformation, peripheral progression was significantly faster by 0.070.02 mm/year (P < 0.0001) in eyes with incomplete foveal sparing versus complete foveal sparing. However, progression toward the fovea was not significantly different for square rootetransformed data (P ¼ 0.31). Subsequently, agreement in atrophy progression between the 2 eyes of a single patient was assessed. Eleven patients with bilateral foveal-sparing GA could be included into this subanalysis. For square rootetransformed centrifugal progression, the ICC was 0.82 (95% confidence interval, 0.61e0.93), and for centripetal progression, the ICC was 0.48 (95% confidence interval, 0.15e0.83).

1360

Discussion The qualitative observation of the phenomenon of foveal sparing implicates different directional expansion rates of atrophic areas.4,11,17,32 However, the differential kinetics of atrophy progression toward the periphery as compared with the central retina in eyes with foveal sparing have not been quantified previously. By applying an image analysis software tool that allows for quantification of GA in eyes with foveal sparing and by modeling progression kinetics using a linear mixed model, this analysis revealed that, after square rootetransformation, GA progression toward the periphery is on average 2.8- times more quickly compared with progression toward the fovea in the presence of marked interindividual variation. The relevance of this locally distinct progression kinetics for a patient becomes evident if one imagines how long an average spared fovea would survive in their absence: mean size of the spared fovea at baseline was 1.480.88 mm2, which is only little more than half of the annual atrophy progression rate (2.520.24 mm2/year). In case of symmetrical progression kinetics toward the periphery and the fovea, the initially spared fovea would be lost after a short period. However, because of the differential progression kinetics, survival of the foveal island is markedly prolonged. Local differences in GA progression kinetics seem not to represent a phenomenon limited to eyes with GA and foveal sparing. The present work is in line with previous studies demonstrating local differences in progression of GA areas: a preferential vertical progression of GA areas was described first by Schatz and McDonald.3 Additionally, a more rapid progression of atrophy areas located within the inner zone (i.e., 600e1200 mm parafoveally) of an Early Treatment Diabetic Retinopathy Study grid has been reported.33 More recently, local GA border characteristics associated with fast GA progression were identified.34 The mean difference between fastest and slowest progressing GA border within the same eye was as high as 0.12 mm/year.27 In agreement with the present data, these results indicate that local factors may exist that determine the distinct directional GA progression kinetics. The mechanisms responsible for these differential progression kinetics are unclear. Regarding foveal sparing, histologic observations35e37 led to the hypothesis of a preferential vulnerability of the rod system, whereas the cone system seems more resistant to factors promoting GA.38e40 An alternative explanatory approach posits that the unique choroidal blood supply to the fovea that may exhibit a local protective effect.33,41,42 Furthermore, the ratio of rods per RPE cells is unfavorably high in the parafoveal retina. This could lead to an early decompensation of the metabolic function of parafoveal RPE cells, and thereby promote foveal-sparing shaped atrophy development.43,44 Finally, it was hypothesized that the macula pigment, which has its physiologic peak at the foveola, exhibits a protective effect against atrophy progression. This idea is supported by the observation of decreased values of macula pigment in eyes at high risk of developing AMD.45 In this context, the results for agreement of atrophy progression between the

Lindner et al



Progression of Foveal-Sparing in GA

Figure 3. Bland-Altman plots for (A) atrophy size and (B) foveal sparing size. The upper diagram illustrates that there is no systematic difference between the values obtained by the 2 readers for atrophy measurement. As shown in (B), reader 1 had a tendency to measure the spared fovea larger than reader 2. This difference increased with increasing size of foveal sparing. Also, the variance between the 2 readers increased with increasing size of foveal sparing. Black points represent visits where the spared fovea entirely surrounded the fovea; gray points indicate an incomplete foveal sparing.

2 eyes of a single patient presented in this work are of particular interest. Agreement for progression toward the fovea was lower than for peripheral progression. This may point toward a local rather than systemic reason for foveal-sparing maintenance, for example, by anatomic variations in foveal blood supply. Of note, case numbers for this subanalysis were low and they may be explained at least in part by coincidence or the lower accuracy of determining foveal sparing. Among the patients observed, 39 of the 47 eyes still had a foveal sparing at last visit (>2 years mean follow-up). This observation is in agreement with an earlier study by Sunness and et al17 reporting that 80% of eyes with a horseshoe or ring configuration at baseline still had a foveal-sparing configuration at 4 years. The kinetic data presented here may explain this long survival of the preserved foveal island: with a baseline size of 1.48 mm2 and a progression rate toward the fovea of 0.25 mm2/year, it would take almost 6 years for the average foveal island to be lost completely. Yet, it remains to be investigated whether foveal-sparing regression becomes slower with decreasing size of the spared fovea (e.g., because of a maximum abundance of a protective factor at the foveola). This could explain even longer times of foveal survival. The current study revealed marked interindividual differences for both centrifugal and centripetal GA progression among different eyes. High variability in the GA progression rate between patients has been described, whereby various prognostic markers have been identified that also may influence centripetal progression kinetics. This includes GA size at baseline,46,47 the previous progression rate,47 the GA progression rate of the fellow eye,24,47,48 as well as the diagnosis of the fellow eye.25 More recently, reticular pseudodrusen49 and distinct spectral-domain optical coherence tomography findings, including a splitting of the RPEeBruch membrane complex, were described as risk factors for GA progression.34,50e52 Furthermore, GA configuration seems to possess a prognostic value46,53,54 as well as distinct patterns and the extent of increased FAF surrounding atrophy.31,46,55

Interestingly, overall GA progression rate observed in this patient cohort with foveal sparing was faster than that reported for the entire FAM collective.28,31 A possible explanation is the relatively high percentage of eyes with diffuse FAF patterns in the current analysis. Among those, the diffuse trickling phenotype that has been reported to show extremely fast GA progression was relatively frequent.31,56 Of note, the phenomenon of foveal sparing is not exclusive for AMD, but may occur with other retinal diseases that lead to outer retinal atrophy. These include toxic retinopathies as well as monogenetic macular and retinal dystrophies.43,57,58 Herein, the most striking example of prolonged structural and functional foveal sparing is seen in mitochondrial retinal dystrophy.43

Figure 4. A, Box-and-whisker plot illustrating distinct square rootetransformed centrifugal and centripetal atrophy progression, estimated separately for each eye. Mean peripheral progression occurs 2.8times more quickly than centrally directed progression, that is, centripetal progression. B, Data as presented in (A), but without square root transformation, given for illustration.

1361

Ophthalmology Volume 122, Number 7, July 2015

Figure 5. Scatterplot showing the square rootetransformed data for centrifugal and centripetal atrophy progression plotted against each other. The linear fit illustrates the correlation between square root peripheral atrophy progression kinetics and square root foveal sparing regression kinetics (r ¼ 0.35; P < 0.05).

Identification and quantification of the spared fovea in eyes with GA may be challenging. Fundus autofluorescence imaging with an excitation wavelength of 488 nm may lead to an overinterpretation of the size of atrophic patches and

foveal involvement,59 because there is a decreased signal resulting from the absorption of luteal pigment. Greenlight FAF (excitation wavelength, 514 nm) and spectraldomain optical coherence tomography imaging seem to be superior in the identification of the spared fovea when compared with blue-light FAF (488 nm) alone.22,59 Combined NIR reflectance and 488-nm FAF grading that was performed in the current analysis seems to be a timeeffective and feasible approach for identification and quantification of the spared fovea and the surrounding GA. However, the interreader agreement for GA size was better (ICC, 0.995) as compared with foveal-sparing measurements (ICC, 0.946). Although there was no systematic difference between the values obtained by the 2 readers for atrophy measurement, the difference and the variance between the 2 readers increased with the increasing size of foveal sparing. These results may reflect the complicated nature of foveal-sparing quantification in general. In particular, they may reflect difficulties in achieving agreement in the measurement of foveal-sparing areas that had a bridge toward the peripheral retina. Readers were asked to draw a line at the narrowest part of the bridge at baseline to confine the spared fovea from the peripheral macula. This rather subjective task may represent a source of interreader disagreement. However, in the present analysis, ICC between readers was also high in eyes where these lines were set. The decision to keep the line at the same position in all serial images of an eye was made to reduce differences in measurements of a single reader because of changes in the location of this constraint during follow-up. However, this strategy also may be source of nonprecise measurements of

Figure 6. Images showing the variability of atrophy progression and survival of the residual foveal island among the eyes evaluated: (upper row) fundus autofluorescence and (lower row) near infrared images (left column) at baseline and (right column) at latest follow-up. A, Eye with 2.7 years of follow-up. Square root progression toward the periphery was 0.35 mm/year and progression toward the fovea was 0.129 mm/year (ratio, 2.77). Both values being in close proximity to the mean values from the total patient collective analyzed here. During follow-up, the size of the spared fovea decreased from 2.36 to 1.40 mm2. B, Eye with 3.9 years of follow-up. Square root centrifugal progression was 0.329 mm/year. In this eye, centripetal progression clearly was above the average with 0.222 mm/year (ratio, 1.5). The size of the spared fovea decreased from 1.31 to 0.07 mm2. C, Eye with 3.3 years of follow-up exemplifying extremely slow centripetal progression as present in some eyes. Square root centrifugal progression was 0.160 mm/year and centripetal progression was 0.032 mm/year (ratio, 5). The size of the spared fovea decreased from 0.47 to 0.32 mm2.

1362

Lindner et al



Progression of Foveal-Sparing in GA

progression kinetics because the possibility remains that during follow-up, the bridge truncates at another location. In the present study, an interreader agreement of less than 0.15 mm2 was demanded, or a senior grading was performed. This limit of less than 0.15 mm2 was set in agreement with the standard operating procedures of the GRADE (GRAding of Digital fundus Examinations) Reading Center, Bonn, Germany, and resulted in an apparently high frequency of senior gradings. Yet, it reflects the strictness of agreement limits, rather than an extensive interreader disagreement, as the high values obtained for ICC and the Bland-Altman plots (Fig 3) for atrophy and foveal sparing area illustrate. Of note, a disagreement of 0.15 mm2 represents only 2% of the total area of GA at baseline, but 10% of the baseline fovealsparing area. By accepting this higher relative cutoff for senior grading, the accuracy of foveal-sparing measurement is probably lower than the accuracy of atrophy measurement. This represents a possible limitation of the present study. In conclusion, the current analysis quantifies differential progression kinetics within eyes with GA and foveal sparing showing a 2.8-fold faster centrifugal GA progression as compared with centripetal progression. Local, yet unidentified, factors seem to drive differential disease progression. These results contribute further to the understanding of the natural disease process of GA. With regard to interventional trials aiming to preserve the fovea in patients with GA, the analysis strategy applied here that combines FAF and NIR reflectance imaging seems to be reliable for assessment of the residual foveal island. Furthermore, the presented natural history data may be used for study design, particularly for sample size calculations.

References 1. Gass JD. Drusen and disciform macular detachment and degeneration. Arch Ophthalmol 1973;90:206–17. 2. Blair CJ. Geographic atrophy of the retinal pigment epithelium. A manifestation of senile macular degeneration. Arch Ophthalmol 1975;93:19–25. 3. Schatz H, McDonald HR. Atrophic macular degeneration. Rate of spread of geographic atrophy and visual loss. Ophthalmology 1989;96:1541–51. 4. Sunness JS. The natural history of geographic atrophy, the advanced atrophic form of age-related macular degeneration. Mol Vis 1999;5:25–30. 5. Lim LS, Mitchell P, Seddon JM, et al. Age-related macular degeneration. Lancet 2012;379:1728–38. 6. Klein R, Klein BE, Franke T. The relationship of cardiovascular disease and its risk factors to age-related maculopathy. The Beaver Dam Eye Study. Ophthalmology 1993;100: 406–14. 7. Vingerling JR, Dielemans I, Hofman A, et al. The prevalence of age-related maculopathy in the Rotterdam Study. Ophthalmology 1995;102:205–10. 8. Klein R, Klein BE, Knudtson MD, et al. Fifteen-year cumulative incidence of age-related macular degeneration: the Beaver Dam Eye Study. Ophthalmology 2007;114:253–62. 9. Congdon N, O’Colmain B, Klaver CC, et al. Causes and prevalence of visual impairment among adults in the United States. Arch Ophthalmol 2004;122:477–85.

10. Resnikoff S, Pascolini D, Etya’ale D, et al. Global data on visual impairment in the year 2002. Bull World Health Organ 2004;82:844–51. 11. Sunness JS, Gonzalez-Baron J, Applegate CA, et al. Enlargement of atrophy and visual acuity loss in the geographic atrophy form of age-related macular degeneration. Ophthalmology 1999;106:1768–79. 12. Sarks JP, Sarks SH, Killingsworth MC. Evolution of geographic atrophy of the retinal pigment epithelium. Eye (Lond) 1988;2(Pt 5):552–77. 13. Sunness JS, Schuchard RA, Shen N, et al. Landmark-driven fundus perimetry using the scanning laser ophthalmoscope. Invest Ophthalmol Vis Sci 1995;36:1863–74. 14. Schmitz-Valckenberg S, Bultmann S, Dreyhaupt J, et al. Fundus autofluorescence and fundus perimetry in the junctional zone of geographic atrophy in patients with age-related macular degeneration. Invest Ophthalmol Vis Sci 2004;45: 4470–6. 15. Forte R, Querques G, Querques L, et al. Multimodal evaluation of foveal sparing in patients with geographic atrophy due to age-related macular degeneration. Retina 2013;33:482–9. 16. Sunness JS, Rubin GS, Applegate CA, et al. Visual function abnormalities and prognosis in eyes with age-related geographic atrophy of the macula and good visual acuity. Ophthalmology 1997;104:1677–91. 17. Sunness JS, Rubin GS, Zuckerbrod A, Applegate CA. Fovealsparing scotomas in advanced dry age-related macular degeneration. J Vis Impair Blind 2008;102:600–10. 18. Delori FC, Dorey CK, Staurenghi G, et al. In vivo fluorescence of the ocular fundus exhibits retinal pigment epithelium lipofuscin characteristics. Invest Ophthalmol Vis Sci 1995;36: 718–29. 19. von Ruckmann A, Fitzke FW, Bird AC. Distribution of fundus autofluorescence with a scanning laser ophthalmoscope. Br J Ophthalmol 1995;79:407–12. 20. Deckert A, Schmitz-Valckenberg S, Jorzik J, et al. Automated analysis of digital fundus autofluorescence images of geographic atrophy in advanced age-related macular degeneration using confocal scanning laser ophthalmoscopy (cSLO). BMC Ophthalmol 2005;5:8–16. 21. Schmitz-Valckenberg S, Brinkmann CK, Alten F, et al. Semiautomated image processing method for identification and quantification of geographic atrophy in age-related macular degeneration. Invest Ophthalmol Vis Sci 2011;52:7640–6. 22. Sayegh RG, Simader C, Scheschy U, et al. A systematic comparison of spectral-domain optical coherence tomography and fundus autofluorescence in patients with geographic atrophy. Ophthalmology 2011;118:1844–51. 23. Holz FG, Strauss EC, Schmitz-Valckenberg S, van Lookeren Campagne M. Geographic atrophy: clinical features and potential therapeutic approaches. Ophthalmology 2014;121:1079–91. 24. Lindblad AS, Lloyd PC, Clemons TE, et al. Change in area of geographic atrophy in the Age-Related Eye Disease Study: AREDS report number 26. Arch Ophthalmol 2009;127: 1168–74. 25. Fleckenstein M, Schmitz-Valckenberg S, Adrion C, et al. Progression of age-related geographic atrophy: role of the fellow eye. Invest Ophthalmol Vis Sci 2011;52:6552–7. 26. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–10. 27. Fleckenstein M, Schmitz-Valckenberg S, Adrion C, et al. Tracking progression with spectral-domain optical coherence tomography in geographic atrophy caused by age-related macular degeneration. Invest Ophthalmol Vis Sci 2010;51:3846–52.

1363

Ophthalmology Volume 122, Number 7, July 2015 28. Dreyhaupt J, Mansmann U, Pritsch M, et al. Modelling the natural history of geographic atrophy in patients with age-related macular degeneration. Ophthalmic Epidemiol 2005;12:353–62. 29. Feuer WJ, Yehoshua Z, Gregori G, et al. Square root transformation of geographic atrophy area measurements to eliminate dependence of growth rates on baseline lesion measurements: a reanalysis of age-related eye disease study report no. 26. JAMA Ophthalmol 2013;131:110–1. 30. Yehoshua Z, Rosenfeld PJ, Gregori G, et al. Progression of geographic atrophy in age-related macular degeneration imaged with spectral domain optical coherence tomography. Ophthalmology 2011;118:679–86. 31. Holz FG, Bindewald-Wittich A, Fleckenstein M, et al. Progression of geographic atrophy and impact of fundus autofluorescence patterns in age-related macular degeneration. Am J Ophthalmol 2007;143:463–72. 32. Schmitz-Valckenberg S, Fleckenstein M, Helb HM, et al. In vivo imaging of foveal sparing in geographic atrophy secondary to age-related macular degeneration. Invest Ophthalmol Vis Sci 2009;50:3915–21. 33. Mauschitz MM, Fonseca S, Chang P, et al. Topography of geographic atrophy in age-related macular degeneration. Invest Ophthalmol Vis Sci 2012;53:4932–9. 34. Moussa K, Lee JY, Stinnett SS, Jaffe GJ. Spectral domain optical coherence tomography-determined morphologic predictors of age-related macular degeneration-associated geographic atrophy progression. Retina 2013;33:1590–9. 35. Curcio CA, Millican CL, Allen KA, Kalina RE. Aging of the human photoreceptor mosaic: evidence for selective vulnerability of rods in central retina. Invest Ophthalmol Vis Sci 1993;34:3278–96. 36. Jackson GR, Owsley C, Curcio CA. Photoreceptor degeneration and dysfunction in aging and age-related maculopathy. Ageing Res Rev 2002;1:381–96. 37. Curcio CA. Photoreceptor topography in ageing and agerelated maculopathy. Eye (Lond) 2001;15(Pt 3):376–83. 38. Chen C, Wu L, Wu D, et al. The local cone and rod system function in early age-related macular degeneration. Doc Ophthalmol 2004;109:1–8. 39. Owsley C, Jackson GR, Cideciyan AV, et al. Psychophysical evidence for rod vulnerability in age-related macular degeneration. Invest Ophthalmol Vis Sci 2000;41:267–73. 40. Wang JS, Kefalov VJ. The cone-specific visual cycle. Prog Retin Eye Res 2011;30:115–28. 41. Chuang EL, Sharp DM, Fitzke FW, et al. Retinal dysfunction in central serous retinopathy. Eye (Lond) 1987;1(Pt 1):120–5. 42. Xu W, Grunwald JE, Metelitsina TI, et al. Association of risk factors for choroidal neovascularization in age-related macular degeneration with decreased foveolar choroidal circulation. Am J Ophthalmol 2010;150:40–47e2. 43. de Laat P, Smeitink JA, Janssen MC, et al. Mitochondrial retinal dystrophy associated with the m.3243A/G mutation. Ophthalmology 2013;120:2684–96. 44. Snodderly DM, Sandstrom MM, Leung IY, et al. Retinal pigment epithelial cell distribution in central retina of rhesus monkeys. Invest Ophthalmol Vis Sci 2002;43:2815–8. 45. Beatty S, Murray IJ, Henson DB, et al. Macular pigment and risk for age-related macular degeneration in subjects from a

1364

46.

47.

48.

49.

50.

51.

52.

53. 54.

55.

56.

57. 58. 59.

Northern European population. Invest Ophthalmol Vis Sci 2001;42:439–46. Holz FG, Schmitz-Valckenberg S, Fleckenstein M, et al. Lesion characteristics and progression in the Natural History of Geographic Atrophy (GAP) study. ARVO Meeting Abstracts 2010;51:94. Sunness JS, Margalit E, Srikumaran D, et al. The long-term natural history of geographic atrophy from age-related macular degeneration: enlargement of atrophy and implications for interventional clinical trials. Ophthalmology 2007;114:271–7. Fleckenstein M, Adrion C, Schmitz-Valckenberg S, et al. Concordance of disease progression in bilateral geographic atrophy due to AMD. Invest Ophthalmol Vis Sci 2010;51: 637–42. Marsiglia M, Boddu S, Bearelly S, et al. Association between geographic atrophy progression and reticular pseudodrusen in eyes with dry age-related macular degeneration. Invest Ophthalmol Vis Sci 2013;54:7362–9. Schmitz-Valckenberg S, Fleckenstein M, Gobel AP, et al. Optical coherence tomography and autofluorescence findings in areas with geographic atrophy due to age-related macular degeneration. Invest Ophthalmol Vis Sci 2011;52:1–6. Nunes RP, Gregori G, Yehoshua Z, et al. Predicting the progression of geographic atrophy in age-related macular degeneration with SD-OCT en face imaging of the outer retina. Ophthalmic Surg Lasers Imaging Retina 2013;44:344–59. Hariri A, Nittala MG, Sadda SR. Outer retinal tubulation as a predictor of the enlargement amount of geographic atrophy in age-related macular degeneration. Ophthalmology 2015;122: 407–13. Klein R, Meuer SM, Knudtson MD, Klein BE. The epidemiology of progression of pure geographic atrophy: the Beaver Dam Eye Study. Am J Ophthalmol 2008;146:692–9. Danis R, Domalpally A, Corkery E, et al. Index of non circularity as a predictor for progression in eyes with geographic atrophy (GA). ARVO Meeting Abstracts 2013;54: 6311. Schmitz-Valckenberg S, Bindewald-Wittich A, DolarSzczasny J, et al. Correlation between the area of increased autofluorescence surrounding geographic atrophy and disease progression in patients with AMD. Invest Ophthalmol Vis Sci 2006;47:2648–54. Fleckenstein M, Schmitz-Valckenberg S, Martens C, et al. Fundus autofluorescence and spectral-domain optical coherence tomography characteristics in a rapidly progressing form of geographic atrophy. Invest Ophthalmol Vis Sci 2011;52: 3761–6. Saksens NT, Fleckenstein M, Schmitz-Valckenberg S, et al. Macular dystrophies mimicking age-related macular degeneration. Prog Retin Eye Res 2013;39:23–57. van Huet RA, Bax NM, Westeneng-Van Haaften SC, et al. Foveal sparing in Stargardt disease. Invest Ophthalmol Vis Sci 2014;55:7467–78. Wolf-Schnurrbusch UE, Wittwer VV, Ghanem R, et al. Bluelight versus green-light autofluorescence: lesion size of areas of geographic atrophy. Invest Ophthalmol Vis Sci 2011;52: 9497–502.

Lindner et al



Progression of Foveal-Sparing in GA

Footnotes and Financial Disclosures Originally received: December 3, 2014. Final revision: March 21, 2015. Accepted: March 23, 2015. Available online: May 9, 2015. 1

Manuscript no. 2014-1943.

Department of Ophthalmology, University of Bonn, Bonn, Germany.

2

Institute for Medical Biometry, Informatics and Epidemiology, University of Bonn, Bonn, Germany. *Principal investigators in the Fundus Autofluorescence in Age-Related Macular Degeneration Study Group are listed in the Appendix. Financial Disclosure(s): The author(s) have made the following disclosure(s): M.L.: Financial support - Genentech, Inc (San Francisco, CA; Grant provided during conduct of this study); Nonfinancial support - Heidelberg Engineering (Heidelberg, Germany); Carl Zeiss Meditec (Jena, Germany); Optos (Dunfermline, Scotland, UK); Alimera Sciences (Alpharetta, GA; Financed meeting participation); Novartis (Basel, Switzerland; Financed meeting participation). A.B.: Financial support - Genentech, Inc (Grant provided during conduct of this study); Nonfinancial support - Heidelberg Engineering; Carl Zeiss Meditec; Optos.

M.M.M.: Financial support - Genentech, Inc (Grant provided during conduct of this study); Nonfinancial support - Heidelberg Engineering; Carl Zeiss Meditec; Optos. A.P.G.: Financial support - Genentech, Inc (Grant provided during conduct of this study); Nonfinancial support - Heidelberg Engineering; Carl Zeiss Meditec; Optos. C.K.B.: Financial support - Genentech, Inc (Grant provided during conduct of this study); Nonfinancial support - Heidelberg Engineering; Carl Zeiss Meditec; Optos. S.S.-V.: Financial support - Genentech, Inc (Grant provided during conduct of this study); Novartis (Grants and personal fees outside this study); Bayer Healthcare (Leverkusen, Germany; Grants and personal fees outside this study); Acucela (Seattle, Washington; Grants outside this study); Alcon (Huenenberg, Switzerland; Grants outside this study); Allergan (Parsippany, NJ; Grants outside this study); Optos (Grants and personal fees outside this study); Heidelberg Engineering (Grants and personal fees outside this study).

and personal fees outside this study); Alcon (Grants and personal fees outside this study); Allergan (Grants and personal fees outside this study); Optos (Grants and personal fees outside this study); Boehringer Ingelheim (Ingelheim, Germany; Personal fees outside this study); Heidelberg Engineering (Grants and personal fees outside this study). M.F.: Financial support - Novartis (Grants or personal fees outside this study); Allergan (Grants or personal fees outside this study); Bayer (Grants or personal fees outside this study); Formycon (Martinsried, Germany; Grants or personal fees outside this study); Genentech, Inc (Grant provided during conduct of this study); Heidelberg Engineering (Grants or personal fees outside this study); Optos (Grants or personal fees outside this study); Roche (Basel, Switzerland; Grants or personal fees outside this study); Alimera (Grants or personal fees outside this study); Nonfinancial supportCarl Zeiss Meditec; Patent - US20140303013 A1 pending. Supported in part by the German Research Foundation (DFG, Bonn, Germany; grant nos.: Ho1926/3-1 and FL 658/4-1); and Genentech, Inc. The sponsor or funding organization had no role in the design or conduct of this research. Author Contributions: Conception and design: Lindner, Fimmers, Schmitz-Valckenberg, Holz, Fleckenstein Analysis and interpretation: Lindner, Fimmers, Schmid, Fleckenstein Data collection: Fleckenstein

Lindner,

Böker,

Mauschitz,

Göbel,

Brinkmann,

Obtained funding: not applicable. Overall responsibility: Lindner, Schmitz-Valckenberg, Holz, Fleckenstein Abbreviations and Acronyms: AMD ¼ age-related macular degeneration; fluorescence Imaging in Age-Related FAF ¼ fundus autofluorescence; GA ICC ¼ interclass correlation coefficient; RPE ¼ retinal pigment epithelium.

FAM ¼ Fundus AutoMacular Degeneration; ¼ geographic atrophy; NIR ¼ near infrared;

Correspondence: Monika Fleckenstein, MD, Department of Ophthalmology, University of Bonn, Ernst-Abbe-Str. 2, 53127 Bonn, Germany. E-mail: Monika. [email protected].

F.G.H.: Financial support - Genentech, Inc (Grant provided during conduct of this study); Novartis (Grants and personal fees outside this study); Bayer Healthcare (Grants and personal fees outside this study); Acucela (Grants

1365