Precursors and Development of Geographic Atrophy with Autofluorescence Imaging

Accepted Manuscript Precursors and Development of Geographic Atrophy with Autofluorescence Imaging: Age-Related Eye Disease Study 2 Report No. 18 Ian C. Holmen, MD, Bryce Aul, MD, Jeong W. Pak, PhD, Ralph Moeller Trane, MS, Barbara Blodi, MD, Michael Klein, MD, Traci Clemons, PhD, Emily Chew, MD, Amitha Domalpally, MD, the AREDS2 Research Group PII:

S2468-6530(18)30672-9

DOI:

https://doi.org/10.1016/j.oret.2019.04.011

Reference:

ORET 520

To appear in:

Ophthalmology Retina

Received Date: 21 November 2018 Revised Date:

8 April 2019

Accepted Date: 9 April 2019

Please cite this article as: Holmen I.C., Aul B., Pak J.W., Trane R.M., Blodi B., Klein M., Clemons T., Chew E., Domalpally A. & the AREDS2 Research Group, Precursors and Development of Geographic Atrophy with Autofluorescence Imaging: Age-Related Eye Disease Study 2 Report No. 18, Ophthalmology Retina (2019), doi: https://doi.org/10.1016/j.oret.2019.04.011. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Precursors and Development of Geographic Atrophy with Autofluorescence Imaging: Age-Related

2

Eye Disease Study 2 Report No. 18

3

Ian C. Holmen MD1, Bryce Aul MD1, Jeong W. Pak PhD1, Ralph Moeller Trane MS1, Barbara Blodi

4

MD1, Michael Klein MD2, Traci Clemons PhD3, Emily Chew MD4, Amitha Domalpally MD1 and the

5

AREDS2 Research Group5

6

1

7

2

8

3

9 10

4

11

5

RI PT

1

Fundus Photographic Reading Center, The University of Wisconsin, Madison, WI, USA

The Emmes Corporation, Rockville, MD, USA

M AN U

Division of Epidemiology and Clinical Applications, National Eye Institute, National Institutes of Health, Bethesda, MD, USA Appendix for the listing of all the members of the AREDS2 Research Group

12

Funding: This study was supported by the intramural program funds and contract HHS-N-260-200500007-C and ABD contract N01-EY-5-00007 from the National Eye Institute (NEI), National Institutes of Health (NIH), Department of Health and Human Services, Bethesda, Maryland, and in part by an unrestricted grant from Research to Prevent Blindness to the Department of Ophthalmology and Visual Sciences, University of Wisconsin, Madison (IH,JWP,RMT, BB,AD)

18

TE D

13 14 15 16 17

SC

Casey Eye Institute, Oregon Health and Science University, Portland, OR, USA

Corresponding Author:

20

Amitha Domalpally, MD

21

Suite 205, 310 N Midvale Blvd

22

Madison, WI 53717

23

Phone 608 2631088

24

[email protected]

AC C

EP

19

25 1

ACCEPTED MANUSCRIPT

Abstract (311)

27

Purpose: To describe the sequence of events leading to development of geographic atrophy (GA) in age-

28

related macular degeneration (AMD) with fundus autofluorescence (FAF) imaging.

29

Design: Post hoc analysis of FAF images from the Age-Related Eye Disease Study 2.

30

Participant: FAF images of 120 eyes (109 patients) with incident GA and at least 2 years of preceding

31

FAF images.

32

Methods: Images of incident GA were stacked and aligned over FAF images of preceding annual visits.

33

The regions of retina that developed into incident GA were assessed on prior years FAF images. These

34

regions, defined as precursor lesions were classified into Minimal ChangeAF, Predominant HypoAF

35

(decreased AF), Predominant HyperAF (increased AF), and MixedAF. The natural progression in

36

precursor lesions leading to GA formation and their associations with incident GA size and GA

37

enlargement rate was evaluated.

38

Results: Incident GA had a mean area of 1.00 mm2 (range 0.15 - 8.22 mm2) and an enlargement rate of

39

0.97 mm2/year (SD 1.66). Predominant HypoAF was the most common precursor lesion, increasing from

40

42% to 81% over 3 years prior to onset of GA. Almost 30% of eyes had Minimal ChangeAF 3 years prior

41

to GA. Among the other precursors, 70% progressed to Predominant HypoAF before developing GA. The

42

type of precursor lesions was not associated with incident GA area. GA evolving from Minimal

43

ChangeAF precursor lesions was associated with faster GA enlargement rates compared to other

44

precursor lesion classes.

45

Conclusion: Using image registration we identified changes in AF mages prior to onset of GA.

46

Decreased autofluorescence was the most common change although minimal changes were also seen in a

47

third of the images. Incident GA that arises from predominantly normal AF is associated with faster

48

enlargement rates compared to GA arising from abnormal AF. Faster GA enlargement rates were also

49

associated with incident GA size, area of surround abnormal autofluorescence, and presence of reticular

50

pseudodrusen.

52 53

SC

M AN U

TE D

EP

AC C

51

RI PT

26

54 55 56 2

ACCEPTED MANUSCRIPT

57 Introduction

59

Geographic atrophy (GA) is a late stage phenotype of age-related macular degeneration (AMD) that

60

causes dense scotomatas, reading and facial recognition challenges, and poor visual acuity when affecting

61

the fovea.1 In the US, nearly 1.5% of the population over the age of 40 suffer from late AMD, including

62

GA.2 No effective treatments are currently available to restore visual loss from GA. In addition, to date,

63

all phase 3 clinical trials of therapies to slow GA enlargement have not been successful.3,4 Given the

64

extensive research on the natural progression of GA using fundus autofluorescence (FAF), the European

65

Medicines Agency and the U.S. Food and Drug Administration accept FAF imaging to assess the

66

progression of established GA as a clinically relevant endpoint.5-12 However, there is very little research

67

on the natural evolution from early/intermediate stage AMD to GA using FAF imaging. Evaluating this

68

evolution with FAF can give important insight into monitoring the progression of AMD to GA,

69

understanding the pathophysiology underlying its development, and developing effective treatments for

70

the prevention of GA formation.

71

The natural evolution of early stage AMD to late stage GA without neovascularization has been well

72

described on color fundus photographs (CFP). Klein et al. studied the 5-year follow-up data from Age-

73

Related Eye Disease Study and showed the evolution of dry AMD in four general stages: (1) early large

74

drusen, (2) hyperpigmentation, (3) drusen regression with depigmentation, and (4) geographic atrophy.13

75

Wu et al. has since identified precursor lesions to drusen-associated disruption and subsequent atrophy on

76

spectral-domain optical coherence tomography (SD-OCT), which they termed nascent GA.14 In a subset

77

of eyes with FAF imaging, Wu et al. associated nascent GA with mixed autofluorescence (both hypo

78

autofluorescence and hyper autofluorescence) on FAF on a cross-sectional analysis.15 Thiele et al. later

79

associated drusen-associated GA with hyper autofluorescence on FAF, and identified 2 other pathways to

80

GA formation with distinct FAF signals.16

81

Hyper autofluorescence signals on FAF are associated with increased accumulation of lipofuscin (the

82

intrinsic fluorophore in retinal pigment epithelial (RPE) cells), vertical stacking of fluorophore containing

83

cells, and cellular fragments, while hypo autofluorescence is mainly associated with RPE loss.17-19

84

Monitoring the changes in autofluorescence signal prior to onset of GA can provide information on the

85

development of GA. To-date studies of GA precursor lesions on FAF have been limited to smaller case

86

series so further in-depth analysis may yield more insight into the pathogenesis of GA.

AC C

EP

TE D

M AN U

SC

RI PT

58

3

ACCEPTED MANUSCRIPT

The Age-Related Eye Disease Study 2 (AREDS 2) was a multicenter phase III randomized controlled

88

clinical trial to study the impact of nutritional supplements on the progression of AMD.20 The primary

89

outcome of the clinical trial was development of late AMD defined as central GA or neovascular AMD.

90

The purpose of the current report is to describe the sequence of progression leading to incident GA

91

associated with AMD with FAF in AREDS2 participants and identify associations between FAF

92

precursor lesions and incident GA enlargement and development.

93

Methods

94

Study Design

95

The AREDS2 study design has been described previously.20 The AREDS2 FAF ancillary study involved

96

obtaining annual FAF images from a subset of participating AREDS2 sites with available cameras.

97

Fundus autofluorescence images were obtained from Heidelberg Retina Angiograph (Heidelberg

98

Engineering, Heidelberg, Germany) instruments by certified photographers. Methods pertaining to

99

evaluation of FAF have been described in detail in AREDS2 Report Number 11.7 The research was

M AN U

SC

RI PT

87

conducted under the Declaration of Helsinki and complied with the Health Portability and Accessibility

101

Act. Institutional review board approval was obtained at each clinical site and written informed consent

102

for the research was obtained from all study participants.

103

For the purpose of this study, eyes with at least 2 sequential annual visits with FAF imaging preceding

104

incident GA formation were included for analysis. Eyes with neovascular AMD, poor image quality,

105

confounding peripapillary atrophy, and/or poor field definition were excluded. Poor image quality was

106

subjectively determined by certified graders and was due to moving shadows, uneven illumination, or

107

significant differences in photo bleaching between visits. Changes in quality between visits prohibited

108

direct comparison of registered images. Given that evaluating central macular integrity in the presence of

109

GA by scanning laser ophthalmoscopy FAF images alone is challenging, eyes with lesions affecting only

110

the central macula were also excluded.21

111

Evaluation of Incident Visit

112

Geographic atrophy (GA) was classified as a well-defined, homogenously black area, with a minimum

113

size of I-2 (diameter 430 microns, area 0.15 mm2), on FAF. The first visit at which GA was identified on

114

FAF, without any GA observed on previous visits, was considered the “incident visit.” Incident visit FAF

115

images were evaluated for GA area, distance from GA to central macula, central macular involvement,

116

presence of halo (hyper autofluorescence), perilesional abnormal area, GA focality, and presence of

117

reticular pseudodrusen. Measurements of area and distance were made using software planimetry tools.

AC C

EP

TE D

100

4

ACCEPTED MANUSCRIPT

The central macula was assumed to be involved if the hypo AF border of the GA merged with the

119

darkness of the central macula and there was no clear region demarcating the two. Halo was defined as

120

presence of hyper AF surrounding at least 10% of the perimeter of GA. Abnormal perilesional area was

121

defined as a mix of hypo and hyper AF surrounding and contiguous to the area of GA of any size. GA

122

focality was graded as multifocal if there was more than one area of GA meeting the criteria described

123

above. Reticular pseudodrusen was ill-defined background network of hypo AF and hyper AF ribbons in

124

a reticulated pattern with a minimum size of 0.5 disc area. Fundus autofluorescence evaluations were

125

independently performed by two trained and certified graders.

126

Characteristics of the CFP from the corresponding incident visit were also recorded per AREDS2 grading

127

methodology and included area of drusen and pigment changes along with AREDS severity scale for

128

AMD.7

129

Defining Precursor Lesions

130

Incident visit images were stacked and aligned over preceding annual FAF images from the same eye

131

using vascular markings in Adobe Photoshop CC 2018 (Adobe Systems, San Jose, CA, USA). This

132

allowed for pixel-to-pixel correlation between images of sequential visits. Eyes were excluded if major

133

rescaling of sequential FAF images was necessary for proper registration. Planimetry tools were used to

134

define the region of interest (ROI) around the incident GA; the ROI remained in the same position in

135

preceding FAF images. In cases of multifocality, only the largest GA lesion was selected. Measurements

136

of halo, GA area, and perilesional area were only done on the largest GA lesion in cases of multifocality.

137

Once the ROI was drawn on a transparent layer, the incident visit images were removed from the stack so

138

the ROI annotation was visible over the preceding visit. The area within the ROI at preceding visits was

139

defined as a precursor lesion with visits identified as Year -1 (1 year before incident GA), -2, -3, etc. The

140

ROI of each FAF image preceding the incident visit was graded for percentage of normal

141

(isofluorescence) or abnormal (hyper AF and/or hypo AF). Images were subjectively and independently

142

graded by 2 certified graders at the University of Wisconsin Fundus Photograph Reading Center. In cases

143

in which the precursor lesion included the center macula, center macular hypo AF was included in the

144

precursor lesion. Figure 1 demonstrates the stacking and alignment of FAF images, placement ROI, and

145

progression of precursor lesions prior to incident visit.

146

All precursor lesions were assigned to one of four classes represented in Figure 2. The precursor lesion

147

was classified as Minimal ChangeAF if normal AF (isofluorescence) was greater than 50% of the ROI.

148

When abnormal AF was greater than or equal to 50% of the ROI, the precursor lesion was classified

149

based on the relative percentage of decreased (hypoAF) or increased autofluorescence (hyperAF) as

AC C

EP

TE D

M AN U

SC

RI PT

118

5

ACCEPTED MANUSCRIPT

Predominant HypoAF, Predominant HyperAF, or MixedAF. MixedAF was defined as subjectively equal

151

percentage of decreased and increased AF.

152

Evaluation of GA Enlargement

153

The five-year longitudinal follow-up in AREDS2 study allowed the identification of precursor lesions,

154

along with analysis of GA enlargement for 1-3 years following the incident visits. The GA enlargement

155

was assessed for eyes with at least 1 follow-up visit after the development of GA. Area of GA was

156

measured on FAF at the final follow-up study visit and GA enlargement per year was calculated.

157

Statistical Analysis

158

Frequency distributions and summary statistics were analyzed. A multivariable model was used with

159

generalized estimating equations (GEE) to analyze correlation between precursor lesion classes and

160

incident visit GA characteristics with GA enlargement rate. Square root transformation was used for

161

analysis of incident GA area and rate of enlargement. All analysis was performed using R software

162

(version 3.5.1, R Foundation for Statistical Computing, Vienna, Austria)

163

Results

164

The AREDS2 ancillary FAF study included 5048 eyes from 2524 participants with follow-up varying

165

from 2 to 6 years. For the purpose of this project, 226 eyes (192 participants) with incident GA and at

166

least two preceding annual visits with FAF imaging were included. Images were excluded for poor image

167

quality (n = 37), confounding peripapillary atrophy (n = 14), confounding central macular pigmentation

168

(n = 34), and poor image registration (n = 21). One-hundred and twenty eyes (109 participants) fitting

169

these criteria were included. Of these, 69 eyes (60 participants) had 3 complete years of FAF images prior

170

to the incident visit.

171

Characteristics of GA at the incident visit (n=120) are presented in Table 1. The mean area of incident

172

visit GA on FAF was 1.0 mm2 (SD 1.2) (0.39 disc area (DA), SD 0.47) with a range from 0.15 mm2 to

173

8.22 mm2 (0.06 - 3.23DA). The median enlargement rate of incident GA was 0.41 mm2/year (IQR 0.09 –

174

0.95) (0.16 DA/yr, IQR 0.04 – 0.25) with 1-3 years of follow-up. Mean proximity of the incident GA to

175

central macula was 581 µm (SD 664), with 45% of eyes having contiguous central macula involvement.

176

The mean abnormal perilesional AF was 12.64 mm2 (SD 10.21) (4.97 DA, SD 4.02). Halo was seen in

177

43% and reticular pseudodrusen in 38%. Twenty-five percent of incident GA lesions were multifocal, and

178

none amalgamated during follow-up. Evaluation of the corresponding CFP showed GA in 49% of eyes at

179

the corresponding visit. Other features of AMD from the corresponding CFP at the incident visit included

AC C

EP

TE D

M AN U

SC

RI PT

150

6

ACCEPTED MANUSCRIPT

large drusen in all eyes, 96% with increased pigment, and 76% with depigmentation. The sub-sample of

181

eyes (n=69) with three years of complete FAF imaging prior to the incident visit had similar

182

characteristics to the overall sample (Table S1, Supplemental Data).

183

The percentage of eyes with precursor lesions characterized by abnormal AF increased gradually through

184

time. At 3 years prior to the incident visit, 29% of eyes had precursor lesions with Minimal ChangeAF,

185

while 71% had abnormal AF. At 1 year prior to the incident visit, the percentage with Minimal ChangeAF

186

decreased to 9%, while that with abnormal AF increased to 91%. The percentage of eyes with

187

Predominant HypoAF increased from 42% (-3 years) to 81% (-1 year) (Figure 3). In contrast, the

188

percentage of eyes with Predominant HyperAF decreased from 23% (-3 years) to 7% (-1 year). The

189

distribution of precursor lesions at each year prior to incident visit did not change when only selecting

190

eyes that also had GA present on CFP.

191

Predominant HypoAF was the most common precursor lesion at all preceding visits, with all other

192

precursor classes converting to Predominant HypoAF over time. Once classified as Predominant

193

HypoAF, 96% remained classified as Predominant HypoAF in subsequent years prior to developing GA.

194

Seventy percent of Minimal ChangeAF and Predominant HyperAF lesions also progressed to

195

Predominant HypoAF in the years prior to developing GA (Figure 4).

196

There was no clear association between precursor lesion class and area of incident GA (p = 0.785).

197

However, in the subgroup with 3 year preceding visits, smaller GA (areas < 0.5 mm2) was associated with

198

Predominant HyperAF precursor lesions compared to larger GA (area > 0.5 mm2) (36% vs 9%, p =

199

0.025).

200

A multivariable model evaluated the correlation between precursor class and enlargement rate of GA

201

evaluated yearly for 1-3 years after incident visit (n = 120) (Table 2, Figure S1, Supplemental Data).

202

Incident GA evolving from Minimal ChangeAF lesions had faster enlargement rates when compared to

203

incident GA evolving from any abnormal precursor lesion. GA evolving from Predominant HyperAF,

204

Predominant HypoAF, and MixedAF precursor lesions grew slower than GA that evolved from Minimal

205

ChangeAF lesions, respectively (p = 0.017, p = 0.012, and p = 0.024). A second model was evaluated

206

with additional incident visit characteristics including presence of reticular pseudodrusen, abnormal

207

perilesional area, presence of halo, and focality. When adding these variables, incident GA evolving from

208

Minimal ChangeAF precursor lesions only had significantly faster enlargement rate in comparison to GA

209

evolving from MixedAF precursor lesions (p = 0.011). All abnormal precursor lesions combined into a

210

single class showed significantly slower enlargement rate compared to Minimal ChangeAF lesions with

211

all incident visit variables included (p = 0.04) (Figure S2, Supplemental Data). When only selecting eyes

AC C

EP

TE D

M AN U

SC

RI PT

180

7

ACCEPTED MANUSCRIPT

that also had GA present on CFP at incident visit (n = 59), all abnormal precursor lesions enlarged

213

significantly slower than Minimal ChangeAF lesions (Supplemental Table S2). Presence of abnormal

214

perilesional area and presence of reticular pseudodrusen were found to significantly associate with faster

215

GA enlargement (p = 0.002 and p = 0.032, respectively). Larger incident GA area significantly associated

216

with faster GA enlargement rate in both models (p<0.001). The results remained similar with the square

217

root transformation of GA area and enlargement rate.

218

Discussion

219

The natural history of GA development and the precursor lesions that evolve into GA in dry AMD have

220

been studied using color photographs. Aside from smaller studies that have used FAF alongside

221

multimodal imaging, this is the first large study to define the precursor lesions of GA on FAF imaging,

222

observe their evolution to GA and their associations with GA enlargement. These data from AREDS2

223

imaging show that GA can evolve from a variety of changes in FAF including Predominant HypoAF,

224

Predominant HyperAF, MixedAF, and Minimal ChangeAF. Among all the precursor lesions,

225

Predominant HypoAF was the most common, accounting for 42% to 81% of precursor lesions from 3

226

years to 1 year prior to developing GA respectively. On FAF images, Predominant HypoAF appears to be

227

the final step before GA is visible, for the majority of eyes. Precursor lesion classification was not

228

associated with the area of incident GA but was associated with enlargement rate.

229

The characteristics of incident GA on FAF have not been well studied. The incident GA in the AREDS2

230

dataset was significantly smaller in area, with a mean area of 1.0 mm2, than the previously described

231

mean area of incident GA and baseline GA in clinical trials, which ranged from 4.6 to 8.85 mm2.5-11,22

232

Incident GA on FAF tends to be surrounded by a broader area of abnormal AF that we termed perilesional

233

abnormal AF, and 43% had hyper AF halo around its border. In our sample, the majority of the lesions

234

were in close proximity to the center of the macula, with a mean distance to the fovea of 581 µm, and

235

45% had central macular involvement. Just under half of incident GA on FAF were visible on CFP, which

236

is consistent with previous studies. A previous AREDS2 report showed that GA is often visible earlier on

237

FAF than it is on CFP, with only 42.9% of GA visible on FAF also being present on CFP at baseline,

238

which increased to over 75% after 2 years.7 We found that the majority of incident GA on FAF

239

correspond to drusen and pigment changes on CFP. Previous studies have shown associations between

240

FAF patterns and area of baseline GA.6,11 We did not observe a consistent association between precursor

241

lesion class and size of incident GA.

242

Klein et al. originally described the natural history of dry AMD on CFP as progressing from large drusen

243

to increased pigmentation and depigmentation before the development of GA.13 Recently, Wu et al.

AC C

EP

TE D

M AN U

SC

RI PT

212

8

ACCEPTED MANUSCRIPT

identified precursors to drusen-associated GA on OCT, which they termed nascent GA, and associated

245

these changes to AF abnormalities on FAF.14,15 They found that most nascent GA associated with mixed

246

hyperAF and hypoAF progressed to mostly hypoAF once GA was definite on OCT.15 Thiele et al.

247

associated varying hyperAF, irregular AF, and mottled AF precursor signals depending on the

248

pathogenesis of the GA.16 We observe trends similar to those suggested by Wu et al. and Thiele et al.

249

Overall, each successive year leading to GA had an increasing number of precursor lesions classified as

250

Predominant HypoAF. Once an eye was classified as Predominant HypoAF, it nearly always remained

251

the same at each successive year until GA was identified (Figure 4). Moreover, precursor lesions

252

classified as Minimal ChangeAF or Predominant HyperAF primarily progressed to Predominant HypoAF

253

before developing into GA. Early observational studies of GA formation using FAF noted that incident

254

GA was often preceded by hyper AF, and Thiele et al. noted that all drusen-associate atrophy was

255

preceded by hyper AF signals.16,23 However, we observe that incident GA is not necessarily preceded by

256

hyper AF, which may be due to variations in GA pathogenesis as suggested by Thiele et al.16 While

257

Thiele et al. also performed yearly FAF imaging, the long interval between images and the small sample

258

size may explain the difference in findings. Our data suggest that significant RPE loss may be present

259

before clear GA formation on FAF, and that large, visible accumulations of lipofuscin may not be

260

necessary for GA development. In fact, 29% of eyes have Minimal ChangeAF (>50% of ROI is normal)

261

three years prior to developing GA and 9% continue to remain in the same group.

262

Eyes with Minimal ChangeAF seem to develop into GA with faster enlargement rates compared to GA

263

evolving from other types of precursors. The enlargement rate of incident GA in this dataset is 0.97

264

mm2/year. Other factors such as baseline GA area, extent of abnormal autofluorescence, and reticular

265

pseudodrusen are also significantly associated with enlargement rate. The area of baseline GA has

266

consistently been associated with GA enlargement rate, with larger baseline lesions having faster

267

enlargement rates.5,6,11,24 Square root transformation of area has been used to reduce the effect of baseline

268

area on enlargement.5,25 We found that area of incident GA area is also associated with GA enlargement

269

rate regardless of other incident FAF features and square root transformation. Previous research has

270

shown an association between the overall increase in the amount of hyper AF and faster enlargement

271

rates.26 We found that halo, which was defined as hyper AF contiguous to incident GA perimeter, was not

272

associated with incident GA enlargement rate. However, larger abnormal perilesional area, which

273

consisted of both increased hyper AF and hypo AF surrounding the region of incident GA, correlated with

274

faster enlargement rates of incident GA. Reticular pseudodrusen has been associated with spatial

275

progression of GA and with faster GA enlargement rates in some cases.27-29 These findings suggest that

276

diffuse outer retinal changes in the form of contiguous abnormal AF and reticular pseudodrusen have an

AC C

EP

TE D

M AN U

SC

RI PT

244

9

ACCEPTED MANUSCRIPT

effect on GA enlargement. Lastly, multifocal lesions have been correlated with faster enlargement rates

278

in various studies.5,10 We did not see an association between faster enlargement and multifocal lesions in

279

eyes with incident GA in our complete-sample multivariable model. This may be due to the fact that our

280

incident GA lesions are significantly smaller than GA reported in other studies and our model has a

281

smaller sample size.5-11,22

282

Association between FAF precursor lesions and GA enlargement rates showed that GA evolving from

283

Minimal ChangeAF 2 years prior to incident visit had a faster enlargement rate than GA developing from

284

other precursor lesions. In other words, GA developing after mostly normal AF may be at higher risk of

285

faster enlargement. Precursor lesions showing minimal change 2 years before GA formation appear to

286

precipitously develop into definitive GA and then continue to grow at a faster rate than other GA lesions.

287

When including multiple incident GA characteristics in the model, only MixedAF remained associated

288

with a decreased enlargement rate while Predominant HyperAF and Predominant HypoAF trended

289

towards significance. When all abnormal precursor lesions were combined, they were associated with

290

slower enlargement rates compared to Minimal ChangeAF. If this association is validated through future

291

studies it may suggest predominantly abnormal AF in precursor lesions provides a protective effect in GA

292

enlargement rate early in GA progression, or that precursor lesions that show minimal change prior to

293

incident GA may be more of an acute process that develop quickly into incident GA and have faster

294

subsequent enlargement.

295

There are several limitations to this AREDS2 report. Annual FAF imaging creates large gaps in our

296

understanding of the evolution of precursor lesions. Previous studies have used three-monthly intervals

297

for imaging.14,15 The higher imaging frequency may allow for increased detection of changes immediately

298

prior to development of atrophy that we may have missed with annual imaging, and consequently the

299

exact sequence of events can only be estimated. Future studies will also need larger sample sizes in order

300

to confirm the association between precursor lesions and GA enlargement rates. The absence of OCT

301

images and histopathology to correlate AF changes with changes in retinal layers is also a limitation. This

302

study is unable to contribute to our understanding of retinal anatomical changes that cause the visible

303

changes on FAF imaging is also unable to differentiate between central macular hypo AF and GA

304

boundaries, which limited our ability to assess GA at the macula. Lastly, our definition of GA was limited

305

to FAF imaging with a relatively large size of 0.15mm2. GA is commonly defined on CFP and by smaller

306

size on inclusion criteria. Changing our definition of GA to another imaging modality such as CFP or

307

changing the size inclusion criteria may change our results. Additionally, changing specific percentage

308

definitions of precursor categories may change our results. Future research should take our definitions

309

into account when making comparisons to our results. Nevertheless, the robust AREDS2 dataset allowed

AC C

EP

TE D

M AN U

SC

RI PT

277

10

ACCEPTED MANUSCRIPT

for an analysis of evolution of incident GA and precursor lesions on FAF. The use of registration and

311

stacking assisted in reproducibly defining precursor lesions and their natural progression to GA. While

312

there is still much left to understand in the evolution of AMD to GA on FAF, continued research may

313

provide pharmaceutical development and clinicians with valuable information in the future treatment and

314

prevention of GA.

RI PT

310

315

TE D

M AN U

SC

1. Danis RP, Lavine JA, Domalpally A. Geographic atrophy in patients with advanced dry agerelated macular degeneration: current challenges and future prospects. Clin Ophthalmol. 2015;9:2159-2174. 2. Freidman DS, O’Colmain BJ, Muñoz B, Tomany SC, McCarty C, de Jong PT, Nemesure B, Mitchell P, Kempen J; The Eye Diseases Prevalence Research Group. Prevalence of age-related macular degeneration in the United States. Arch Ophthalmol 2004;122:564-72. 3. Yehoshua Z, de Amorim Garcia Filho CA, Nunes RP, Gregori G, Penha FM, Moshfeghi AA, Zhang K, Sadda S, Feuer W, Rosenfeld PJ. Systemic complement inhibition with eculizumab for geographic atrophy in age-related macular degeneration: the COMPLETE study. Ophthalmology. 2014;121:693-701. 4. Holz FG, Sadda SR, Busbee B, Chew EY, Mitchell P, Tufail A, Brittain C, Ferrara D, Gray S, Honigberg L, Martin J, Tong B, Ehrlich JS, Bressler NM; Chroma and Spectri Study Investigators. Efficacy and Safety of Lampalizumab for Geographic Atrophy Due to Age-Related Macular Degeneration: Chroma and Spectri Phase 3 Randomized Clinical Trials. JAMA Ophthalmol. 2018;136(6):666-677. 5. Keenan T, Agrón E, Domalpally A, Clemons T, Van Asten F, Wong W, Danis R, Sadda S, Rosenfeld P, Klein M, Ratnapriya R, Swaroop A, Ferris FL 3rd, Chew EY; AREDS2 Research Group. Progression of geographic atrophy in age-related macular degeneration: AREDS2 Report 16. Ophthalmology. 2018 Jul 27 pii: S0161-6420(18)30875-3. 6. Holz FG, Bindewald-Wittich A, Fleckenstein M, Dreyhaupt J, Scholl HP, Schmitz-Valckenber S; FAM-Study Group. Progression of geographic atrophy and impact of fundus autofluorescence patterns in age-related macular degeneration. Am J Ophthalmol. 2007;143:463-472. 7. Domalpally A, Danis R, Agrón E, Blodi BA, Clemons T, Chew E; AREDS2 Research Group. Evaluation of geographic atrophy from color photographs and fundus autofluorescnce images: Age-Related Eye Disease Study 2 Report Number 11. Ophthalmology. 2016;123:2401-2407. 8. Jaffe GJ, Schmitz-Valckenberg S, Boyer D, Heier J, Wolf-Schnurrbusch U, Staurenghi G, Schmidt-Erfurth U, Holz FG. Randomized trial to evaluate tandospirone in geographic atrophy secondary to age-related macular degeneration: the GATE Study. Am J Ophthalmol. 2015;160:1226-34. 9. Yaspan BL, Williams DF, Holz FG, Regillo CD, Li Z, Dressen A, van Lookeren Campagne M, Le KN, Graham RR, Beres T, Bhangale TR, Honigberg LA, Smith A, Henry EC, Ho C, Strauss EC, MAHALO Study Investigators. Targeting factor D of the alternative compliment pathway

EP

317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348

References

AC C

316

11

ACCEPTED MANUSCRIPT

14.

15.

16.

17.

18.

RI PT

SC

13.

M AN U

12.

TE D

11.

EP

10.

reduces geographic atrophy progression secondary to age-related macular degeneration. Sci Transl Med. 2017;9:395. Schmitz-Valckenberg S, Sahel JA, Danis R, Fleckenstein M, Jaffe GJ, Wolf S, Pruente C, Holz FG. Natural history of geographic atrophy progression secondary to age-related macular degeneration (Geographic Atrophy Progression Study). Ophthalmology. 2016;123:361-368. Biarnés M, Arias L, Alonso J, Garcia M, Hijano M, Rodríguez A, Serrano A, Badal J, Muhtaseb H, Verdaguer P, Monés J. Increased Fundus Autofluorescence and Progression of Geographic Atrophy Secondary to Age-Related Macular Degeneration: The GAIN Study. Am J Ophthlamol. 2015;160:345-353. Holz FG, Sadda SR, Staurenghi G, Lindner M, Bird AC, Blodi BA, Bottoni F, Chakravarthy U, Chew EY, Csaky K, Curcio CA, Danis R, Fleckenstein M, Freund KB, Grunwald J, Guymer R, Hoyng CB, Jaffe GJ, Liakopoulos S, Monés J, Oishi A, Pauleikhoff D, Rosenfeld PJ, Sarraf D, Spaide RF, Tadayoni R, Tufail A, Wolf S, Schmitz-Valckenberg S; CAM group. Imaging Protocols in Clinical Studies in Advanced Age-Related Macular Degeneration: Recommendations from Classification of Atrophy Consensus Meetings. Ophthalmology. 2017;124(4):464-78. Klein ML, Ferris FL 3rd, Armstrong J, Hwang TS, Chew EY, Bressler SB, Chandra SR; AREDS Research Group. Retinal precursors and the development of geographic atrophy in age-related macular degeneration. Ophthalmology. 2008;115:1026-31. Wu Z, Luu CD, Ayton LN, Goh JK, Lucci LM, Hubbard WC, Hageman JL, Hageman GS, Guymer RH. Optical coherence tomography-defined changes preceding the development of drusen-asssociated atrophy in age-related macular degeneration. Ophthalmology. 2014;121:241522. Wu Z, Luu C, Ayton LN, Goh JK, Lucci LM, Hubbard WC, Hageman JL, Hageman GS, Guymer RH. Fundus autofluorescence characteristics of nascent geographic atrophy in age-related macular degeneration. Invest Ophthalmol Vis Sci. 2015;56:1546-1552. Thiele S, Pfau M, Larsen PP, Fleckenstein M, Holz FG, Schmitz-Valckenberg S. Multimodal Imaging Patterns for Development of Central Atrophy Secondary to Age-Related Macular Degeneration. Invest Ophthalmol Vis Sci. 2018;59:AMD1-AMD11. Sayegh RG, Simader C, Scheschy U, Montuoro A, Kiss C, Sacu S, Kreil DP, Prunte C, SchmidtErfurth U. A systematic comparison of spectral-domain optical coherence tomography and fundus autofluorescence in patients with geographic atrophy. Ophthalmology. 2011;118:1844-1851. Delori FC, Dorey CK, Staurenghi G, Arend O, Goger DG, Weiter JJ. In vivo fluorescence of the ocular fundus exhibits retinal pigment epithelium lipofuscin characteristics. Invest Ophthalmol Vis Sci. 1995;36:718-729. Rudolf M, Vogt SD, Curcio CA, Huisingh C, McGwin G Jr, Wagner A, Grisanti S, Read RW. Histologic basis of variations in retinal pigment epithelium autofluorescence in eyes with geographic atrophy. Ophthalmology. 2013;120:821-828. AREDS2 Research Group, Chew EY, Clemons T, SanGiovanni JP, Danis R, Domalpally A, McBee W, Sperduto R, Ferris FL. The Age-Related Eye Disease Study 2 (AREDS2): study design and baseline characteristics (AREDS2 report number 1). Ophthalmology. 2012;119:22829. Lindner M, Böker A, Mauschitz MM, Göbel AP, Fimmers R, Brinkmann CK, Schmitz-

AC C

349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392

19.

20.

21.

Valckenberg S, Schmid M, Holz FG, Fleckenstein M, Fundus Autolfluorescence in AgeRelated Macular Degeneration Study Group. Directional Kinetics of Geographic Atrophy 12

ACCEPTED MANUSCRIPT

26.

27.

28.

29.

RI PT

SC

25.

M AN U

24.

TE D

23.

EP

422

22.

Progression in Age-Related Macular Degeneration with Foveal Sparing. Ophthalmology. 2015;122:1356-65. Joachim N, Mitchell P, Kifley A, Rochtchina E, Hong T, Wang JJ. Incidence and progression of geographic atrophy: observations from a population-based cohort. Opthalmology. 2013;120:20422050. Holz FG, Bellman C, Staudt S, Schütt F, Völcker HE. Fundus autofluorescence and development of geographic atrophy in age-related macular degeneration. Invest Ophthalmol Vis Sci. 2001;42:1051-56. Feuer WJ, Yehoshua Z, Gregori G, Penha FM, Chew EY, Ferris FL, Clemons TE, Lindblad AS, Rosenfeld PJ. Square root transformation of geographic atrophy area measurements to eliminate dependence of growth rates on baseline lesions measurements: a reanalysis of age-related eye disease study report no. 26. JAMA Ophthalmol. 2013;131:110-1. Lindblad AS, Lloyd PC, Clemons TE, Gensler GR, Ferris FL 3rd, Klein ML, Armstrong JR; AgeRelated Eye Disease Study Research Group. Change in area of geographic atrophy in the AgeRelated Eye Disease Study: AREDS report number 26. Arch Ophthalmol.2009;127:1168-74. Schmitz-Valckenber S, Bindewald-Wittich A, Dolar-Szczasny J, Dreyhaupt J, Wolf S, Scholl HP, Holz FG; Fundus Autofluorescence in Age-Related Macular Degeneration Study Group. Correlation between the area of increased autofluorescence surrounding geographic atrophy and disease progression in patients with AMD. Invest Ophthalmol Vis Sci. 2006;47:2648-2654. Finger RP, Chong E, McGuinness MB, Robman LD, Aung KZ, Giles G, Baird PN, Guymer RH. Reticular Pseudodrusen and Their Association with Age-Related Macular Degeneration: The Melbourne Collaborative Cohort Study. Ophthalmology. 2016;123:599-608. Zhou Q, Daniel E, Maguire MG, Grunwald JE, Martin ER, Martin DF, Ying GS; Comparison of Age-Related Macular Degeneration Treatments Trials Research Group. Pseudodrusen and Incidence of Late Age-Related Macular Degeneration in Fellow Eyes in the Comparison of AgeRelated Macular Degeneration Treatments Trials. Ophthalmology.2016;123:1530-1540. Marsiglia M, Boddu S, Bearelly S, Xu L, Breaux BE Jr, Freund KB, Yannuzzi LA, Smith RT. Association between geographic atrophy progression and reticular pseudodrusen in eyes with dry age-related macular degeneration. Invest Ophthalmol Vis Sci. 2013;54:7362-9.

AC C

393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421

13

ACCEPTED MANUSCRIPT

Figure Legends

RI PT

Figure 1. Alignment and grading of precursor GA on autofluorescence imaging. Stacks A and B show color fundus photographs and aligned FAF images at incident visit and preceding precursor lesions prior to GA development on FAF. Stack A demonstrates the development of a unifocal precursor lesion from 3 years prior to incident visit. Stack B demonstrates development of a multifocal lesion with central macular involvement.

SC

Figure 2. Precursor lesion classification. Precursor lesions were classified into (a) Minimal ChangeAF – more than 50% normal (isofluorescence), (b) Predominant HypoAF – ≥50% abnormal AF with majority decreased AF, (c) Predominant HyperAF - ≥50% abnormal AF with majority increased AF, (d) MixedAF - abnormal AF with decreased and increased AF being equal.

M AN U

Figure 3. The composition of precursor lesions by class in the years prior to GA at incident visit (n = 69). Figure 4. Trends in progression GA precursor lesions on FAF from 3 years prior to incident visit to 1 year prior to incident visit (n = 69). The arrows represent the percentage of each class that progressed from one classification to another the following year. The majority of Predominant HypoAF precursor lesions remained Predominant HypoAF from one year to the next. Minimal ChangeAF and Predominant HyperAF primarily progressed to Predominant HypoAF or remained the same. MixedAF remained too small to distinguish consistent patterns.

TE D

Supplemental Figure S1. Predicted enlargement rate of incident visit GA based on the square root of GA area at incident visit and its precursor lesion class at 2 years prior to GA formation. Incident visit GA with precursor lesions classified as Minimal ChangeAF (being greater than 50% normal) had significantly faster enlargement rates than incident visit GA with precursor classified as predominantly abnormal. Sqrt = square root.

AC C

EP

Supplemental Figure S2. Predicted enlargement rate of incident visit GA based on the square root of GA area at incident visit and its precursor lesion class at 2 years prior to GA formation. All abnormal precursor lesions were grouped in comparison to Minimal ChangeAF precursors (being greater than 50% normal) (p = 0.04). Sqrt = square root.

ACCEPTED MANUSCRIPT

Table 1. Characteristics of Geographic Atrophy at Incident Visit (n=120) Mean (SD) 1.00 (SD 1.23) 581 (664) 12.64 (10.21) N (%) 54 (45%)

Central Macula Involvement GA Focality Unifocal Multifocal Halo Present Reticular Pseudodrusen

SC

90 (75%) 30 (25%) 51 (43%) 46 (38%)

RI PT

Fundus Autofluorescence Area of Incident GA (mm2) Distance of GA to Fovea (µm) Abnormal Perilesional Area (mm2)

Color Fundus Photograph

59 (49%) 120 (100%) 1 (1%) 115 (96%) 91 (76%)

AC C

EP

TE D

M AN U

GA Drusen Drusenoid PED Increased Pigment Depigmentation

ACCEPTED MANUSCRIPT

Table 2. Association of Incident Geographic Atrophy Enlargement Rate with FAF Characteristics (n = 120) Model 2: Including Incident Visit Characteristics β (95% CI) P Value

Abnormal Precursor Lesion Class Reference Standard: Minimal ChangeAF

0.019 0.012 0.042

0.246 (0.072, 0.419)

0.005

-1.058 (-2.193, 0.077) -1.026 (-2.122, 0.071) -1.48 (-2.614, -0.345)

SC

Incident Visit Characteristics Area of Incident GA (sqrt) Reticular Pseudodrusen Present Area Perilesional Abnormal AF (sqrt) Halo Present Focality

-0.333 (-0.611, -0.055) -0.338 (-0.603, -0.073) -0.282 (-0.554, -0.011)

M AN U

Predominant HyperAF Precursor Predominant HypoAF Precursor MixedAF Precursor

RI PT

Model 1: Excluding Incident Visit Characteristics β (95% CI) P Value

1.196 (0.529, 1.864) 0.601 (0.025, 1.188) 0.204 (0.027, 0.381) 0.514 (-0.155, 1.183) -0.577 (-1.41, 0.255)

0.067 0.068 0.011 <0.001 0.041 0.024 0.132 0.173

AC C

EP

TE D

*Abnormal precursor lesions were compared to Minimal ChangeAF precursor lesions at 2 years prior to incident visit. FAF = fundus autofluorescence. AF = autofluorescence. Sqrt = square root.

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Precis: Precursor lesions for development of geographic atrophy were assessed from autofluorescence images showing that areas of autofluorescence changes, both increase and decrease, can evolve into atrophy. Atrophy developing from minimal changes in autofluorescence is associated with faster growth rate.