Macular Choroidal Thickness and Volume in Eyes With Angioid Streaks Measured by Swept Source OCT

A

UAJ

Vk

Macular Choroidal Thickness and Volume in Eyes WithAngioid Streaks Measured by Swept Source Optical

Coherence Tomography

ABDALLAH A. ELLABBAN, AKITAKA TSUJIKAWA, AKIKO MATSUMOTO, KEN OGINO,MASANORI HANGAI, SOTARO OOTO, KENJI YAMASHIRO, MASAHIRO AKIBA, AND

NAGAHISA YOSHIMURA

piaBdoabc

dnt

● PURPOSE: To study the mean choroidal thickness andvolume of the macula in eyes with angioid streaks usingswept source optical coherence tomography (OCT) inthe 1050-nm wavelength range.● DESIGN: Prospective case series.● METHODS: The macular area of 39 eyes of 23 patientswith angioid streaks and of 20 normal eyes of 20 matchedcontrols (Group 1) was studied with a swept source OCTprototype system. Eyes with angioid streaks were classi-fied into 1 of 4 groups: those without choroidal neovas-cularization (CNV) (Group 2); those with CNV thathad no history of treatment (Group 3); those with CNVthat had previously received only anti–vascular endothe-lial growth factor treatments (Group 4); and those withCNV that had previously received photodynamic therapy(Group 5). Using a raster scan protocol with 512 � 128A-scans, we produced a macular choroidal thickness map(6 � 6 mm2).● RESULTS: There were no significant differences in age,axial length, or refractive error among the 5 groups.Mean choroidal thickness of the macula in Group 2(218.9 � 46.8 �m) was as great as that in Group 1(218.8 � 69.2 �m). However, the macular choroidalthickness in Group 3 (119.7 � 49.2 �m), Group 4(140.1 � 64.9 �m), and Group 5 (144.0 � 52.6 �m)was significantly less than that of Group 1 (P < .05).There were no statistical differences between Groups 3through 5. In each group, the choroid of the nasalquadrant was significantly thinner compared to that inother quadrants (P < .05).● CONCLUSIONS: The choroid in eyes with angioidstreaks without CNV was as thick as that in normalcontrols, but was significantly thinner in eyes withangioid streaks that had developed CNV. (Am J Oph-thalmol 2012;153:1133–1143. © 2012 by Elsevier Inc.All rights reserved.)

Supplemental Material available at AJO.com.ccepted for publication Dec 29, 2011.From the Department of Ophthalmology and Visual Sciences, Kyotoniversity Graduate School of Medicine, Kyoto, Japan (A.A.E., A.T.,.M., K.O., M.H., S.O., K.Y., N.Y.); and Topcon Corporation, Tokyo,

apan (A.M., M.A.).Inquiries to Akitaka Tsujikawa, Department of Ophthalmology and

isual Sciences, Kyoto University Graduate School of Medicine, Sakyo-
u, Kyoto 606-8507, Japan; e-mail: [email protected]
© 2012 BY ELSEVIER INC. A0002-9394/$36.00doi:10.1016/j.ajo.2011.12.013

P SEUDOXANTHOMA ELASTICUM IS A RARE MULTISYS-

tem disorder associated with a mutation in theABCC6 gene.1 Pseudoxanthoma elasticum causes

rogressive fragmentation and calcification of elastic fibersn connective tissue, resulting in pathologic changes thatre most pronounced in the dermis, blood vessels, andruch membrane.2 Characteristic fundus changes of pseu-oxanthoma elasticum are peau d’orange, areas of chori-retinal atrophy, and angioid streaks.3,4 Angioid streaksre irregular crack-like dehiscences in the Bruch mem-rane that often allow the ingrowth of choroidal neovas-ularization (CNV) through Bruch membrane defects.3,4

CNV secondary to angioid streaks occurs in 72% to 86% ofeyes, is bilateral in more than 70% of cases, and oftenoccurs at a younger age than does the CNV of age-relatedmacular degeneration (AMD). Because CNV secondary toangioid streaks is often refractory to treatment, it leads inmany cases to poor vision if the CNV involves the maculararea.5–9

Recently, the choroid, which is a structure with one ofthe highest blood flows in the body,10 has been reported tobe involved in the pathogenesis of various ocular diseases,including AMD,11 polypoidal choroidal vasculopathy,12

central serous chorioretinopathy,13 Vogt-Koyanagi-Haradadisease,14 and myopic chorioretinal degeneration.15 Inangioid streaks, although the Bruch membrane is thoughtto be most affected, previous reports have suggested thatthe choroid is also involved in the pathogenesis of CNV.3,4

Today, optical coherence tomography (OCT) is used tomeasure retinal thickness and to detect morphologicchanges in various retinal diseases. However, commerciallyavailable OCT equipment visualizes the entire choroidonly in eyes with high myopia because of low penetrationand high backscattering at the level of the retinal pigmentepithelium (RPE).16,17 So, to date, little in vivo informa-tion is available on the choroidal changes associated withangioid streaks.

Since Spaide and associates16 introduced enhancedepth imaging OCT based on spectral-domain OCT tech-ology, many investigators have studied the choroidalhickness in eyes with various pathologies.11,14,15,18–25

Although normal choroidal thickness in the macula varies
with region and can vary even more in pathologic condi-
LL RIGHTS RESERVED. 1133

http://AJO.com

mailto:[email protected]

tormute

ba0wOctN

wAu(d6

tions, in most of these studies the choroidal thickness wasmeasured at the foveal center or, sometimes, at severalmeasurement points.11,14,15,18–25 This is because enhanceddepth imaging requires the averaging of 50 to 100 B-scansof an identical location of interest,16 hindering us fromhigh-density scanning. With the advances in OCT tech-nology, the recent generation has used swept source lasertechnology, which has the advantage of a high-speedscanning rate and low-sensitivity roll-off vs depth com-pared to spectral-domain OCT.26–34 With the use of alight source at the 1-�m wavelength region, swept sourceOCT allows for a high-contrast, high-penetration image ofthe entire choroid. Based on these advantages, sweptsource OCT at a longer wavelength allows us to obtain a3-dimensional (3D) high-contrast image of the choroid.

In the study described herein, we scanned the maculararea of eyes with angioid streaks associated with pseudo-xanthoma elasticum using swept source OCT at 1050 nmwith a 3D raster scan protocol, and produced a choroidalthickness map of the macular area. By applying the gridused by the Early Treatment Diabetic Retinopathy Study(ETDRS) to this map, the mean choroidal thickness andvolume in each sector was measured and compared withthose of normal subjects.

PATIENTS AND METHODS

THIS PROSPECTIVE STUDY CONSISTED OF 39 EYES OF 23

patients with angioid streaks secondary to pseudoxan-thoma elasticum. The macular area of these 39 eyes wasexamined with a swept source OCT prototype system atKyoto University Hospital between September 2010 andNovember 2011. The diagnosis of pseudoxanthoma elasti-cum was based on systemic manifestations, fundus exami-nation, and fluorescein angiography,35 and the diagnosiswas confirmed by characteristic histopathologic abnormal-ities in a skin biopsy.

All subjects underwent a thorough ocular examination,including autorefractometry (ARK1; Nidek, Gamagori, Ja-pan), best-corrected visual acuity measurement with a 5-me-ter Landolt chart, axial length measurement using ocularbiometry (IOLMaster; Carl Zeiss Meditec, Jena, Germany),slit-lamp examination, intraocular pressure measurement,dilated color fundus photography (TRC50LX; TopconCorp, Tokyo, Japan), and prototype swept source OCTexamination. All eyes with angioid streaks underwentsimultaneous fluorescein angiography and indocyaninegreen angiography using Spectralis HRA�OCT (Heidel-berg Engineering, Heidelberg, Germany). Eyes with poorimages attributable to opaque media (eg, cataracts orvitreous hemorrhage) or to unstable fixation were excludedfrom the study.

Eligible eyes with angioid streaks were classified into 4groups based on the presence of CNV and the previous
treatment of the CNV: eyes with angioid streaks without
AMERICAN JOURNAL OF1134

CNV (Group 2), eyes with CNV secondary to angioidstreaks that had no history of treatments (Group 3), eyeswith CNV secondary to angioid streaks that had previouslyreceived only anti–vascular endothelial growth factor(VEGF) treatments with no history of photodynamictherapy (PDT) (Group 4), and eyes with CNV secondaryto angioid streaks that did have a history of PDT (Group5). Data obtained from the 4 groups with angioid streakswere compared with data obtained from 20 healthy eyes(20 subjects) of controls, matched for age, axial length,and refractive error (Group 1).36 Exclusion criteria forhese 20 controls included history or evidence of chori-retinal or vitreoretinal diseases, including AMD, diabeticetinopathy, central serous chorioretinopathy, epiretinalembrane, and macular dystrophy, or a history of intraoc-

lar surgery. Subjects with systemic diseases or conditionshat might affect retinal or choroidal thickness were alsoxcluded.

● SWEPT SOURCE OCT SYSTEM AND SCAN PROTOCOLS:

The prototype swept source OCT system (Topcon Corp)used in the current study has been reported previously.36 Inrief, this swept source OCT system uses a light source ofwavelength-sweeping laser centered at 1050 nm with 10000-Hz repetition rate. Axial and transverse resolutionsere 8 �m and 20 �m in tissue, respectively. Swept sourceCT imaging operated at the 1-�m wavelength region was

onducted at �1 mW on the cornea, which is well belowhe safe retinal exposure limit established by the Americanational Standards Institute.Swept source OCT examinations of the eligible subjects

ere performed by trained examiners after pupil dilation.3D imaging data set was acquired on each subject by

sing a raster scan protocol of 512 (horizontal) � 128vertical) A-scans per data set (total 65 536 axial scans/ata set) in 0.8 seconds. Each 3D raster scan covered a 6 �-mm2 area centered on the fovea, which was confirmed by

internal fixation and by a fundus camera built into theswept source OCT system. To reduce speckle noise, eachimage was enhanced by weighted moving average of 3consecutive original images.

In each subject, 50 averaged horizontal and vertical scanimages in 12-mm transverse scan length were obtained aswell. The 50 single images, where each image consisted of1024 A-scans, were registered and averaged by software tocreate an averaged single image. The vertical scan wascentered on the fovea while the horizontal scan wascentered on the midpoint between the fovea and opticdisc.

● CHOROIDAL THICKNESS AND VOLUME MEASURE-

MENT PROTOCOL: The choroidal thickness was measuredas the distance between the outer border of the RPE andthe chorioscleral border. In areas where the RPE wasdefective, the Bruch membrane was considered as an inner
border of the choroid. In each image of the 3D data set,
OPHTHALMOLOGY JUNE 2012

lines of both RPE and the chorioscleral border weredetermined manually by a trained observer in a masked

FIGURE 1. Images of a healthy subject obtained with swept source ohorizontal OCT image of 12 mm in length was obtained by averaging 5limiting membrane to the chorioscleral border is seen clearly beyon(Middle and Right) In a raster scan protocol, 128 consecutive imagesimage was enhanced by weighted moving average of 3 consecutiveprotocol without weighted moving average. (Right) Single OCT ima

TABLE 1. Characteristics of Eyes and Mean ChoroidalCoherenc

Group 1

Normal Eyes

Group

AS Without

Number of eyes 20 6

Sex (male/female) 10/10 4/2

Age (years) 67.4 � 13.2 66.0 � 6

Axial length (mm) 23.5 � 1.3 23.3 � 1

Refractive error (diopters) �0.51 � 2.96 0.01 � 1

Visual acuity (logMAR) 0.01 � 0.21 �0.14 � 0

Foveal retinal thickness (�m) 204.4 � 35.5 194.0 � 1

Foveal choroidal thickness (�m) 254.8 � 76.0 244.7 � 5

Mean choroidal thickness within a

circle of 1.0-mm diameter (�m)

238.7 � 75.0 239.0 � 4



233.0 � 72.4 232.7 � 4



218.8 � 69.2 218.9 � 4

Choroidal volume within a circle of

1.0-mm diameter (mm3)

0.187 � 0.059 0.188 � 0



1.646 � 0.512 1.644 � 0



6.183 � 1.957 6.185 � 1

CNV (active/inactive) — —

Treatment

Number of anti-VEGF treatments 0 0

Number of PDT treatments 0 0

AS � angioid streaks; CNV � choroidal neovascularization; logM

therapy; VEGF � vascular endothelial growth factor.aP � .01, compared with Group 1.bP � .05, compared with Group 1.

fashion. Automated built-in calibration software deter-

MACULAR CHOROIDAL THICKNVOL. 153, NO. 6

mined the distance between these 2 lines. The measure-ment points per image consisted of 512 points with an

l coherence tomography (OCT) at 1050 nm. (Left) A multi-averagedages, which consisted of 1024 A-scans. Fine structure from the innervascular arcade. Regional changes in choroidal thickness are seen.

sisting of 512 A-scans, were obtained. To reduce speckle noise, eachal images. (Middle) Single OCT image acquired with a raster scanuired with a raster scan protocol with weighted moving average.

kness and Volume Obtained by Swept Source Opticalography

Group 3

AS With CNV and

No History of

Treatment

Group 4

AS With CNV and a

History of Only

Anti-VEGF Treatments

Group 5

AS With CNV and a

History of PDT P Value

7 11 15

4/3 8/3 8/7

65.1 � 7.1 65.6 � 7.8 66.7 � 9.9 .960

23.6 � 1.1 23.9 � 1.5 23.4 � 1.8 .949

0.11 � 1.93 �1.38 � 2.50 �2.20 � 1.90 .134

1.02 � 0.32a 0.45 � 0.61a 0.80 � 0.45a �.001

260.3 � 119.5 209.9 � 97.8 154.4 � 64.3 .027

142.0 � 67.9a 144.0 � 79.8a 124.8 � 54.2a �.001

117.4 � 55.9a 144.1 � 69.8a 130.5 � 52.8a �.001

116.5 � 52.4a 139.0 � 68.9a 137.7 � 57.6a �.001

119.7 � 49.2a 140.1 � 64.9b 144.0 � 52.6b �.001

0.092 � 0.044a 0.113 � 0.056a 0.102 � 0.041a �.001

0.823 � 0.370a 0.982 � 0.487a 0.973 � 0.407a �.001

3.381 � 1.389a 3.959 � 1.853b 4.070 � 1.487b �.001

4/3 5/6 1/14

0 3.4 � 2.1 3.1 � 3.9

0 0 2.9 � 1.9

� logarithm of minimal angle of resolution; PDT � photodynamic

ptica0 im

d the, conoriginge acq

Thice Tom

2

CNV

.8

.0

.02

.05

2.6

1.0

5.4

9.7

6.8

.036

.351

.323

AR

interval of �12 �m. From the 128 images of each 3D data

ESS IN ANGIOID STREAKS 1135

ct

set, a choroidal thickness map of 6 � 6-mm2 area wasreated. After the choroidal thickness map was obtained,he ETDRS grid was applied to the map.37

We compared data on the mean choroidal thickness and

FIGURE 2. Multi-averaged horizontal section through the fovefundus photograph (Right) in a healthy eye and in eyes withstructures and the chorioscleral border are clearly seen. The chwith angioid streaks without choroidal neovascularization (CNrow) An eye with CNV secondary to angioid streaks thatsubstantially reduced in the entire macular area. (Fourth row)received only anti–vascular endothelial growth factor treatmeChoroidal thickness is substantially reduced in the entire macuside (arrow). (Bottom row) An eye with CNV secondary toChoroidal thickness is reduced not only in the PDT-treated m

volumes of the ETDRS grid between the 5 study groups


and also compared these values within each group. Usingthe averaged OCT images, we also performed measure-ment of retinal thickness and of choroidal thickness at thecenter of the fovea. Retinal thickness was defined as the

ained by swept source optical coherence tomography (Left) andioid streaks. (Top row) A healthy eye (Group 1). Choroidald seems to be thin around the optic disc. (Second row) An eye(Group 2). The choroidal thickness is well preserved. (Thirdno history of treatments (Group 3). Choroidal thickness isye with CNV secondary to angioid streaks that had previouslyith no history of photodynamic therapy (PDT) (Group 4).

ea. Regional thickening of the choroid is seen on the temporalid streaks that was treated previously with PDT (Group 5).r area but also beyond the vascular arcade.

a obtangoroiV)

hadAn ents wlar arangioacula

distance between the vitreoretinal interface and RPE.


1

tddv

● STATISTICAL ANALYSIS: Statistical analysis was per-formed using SPSS statistical software (version 16; SPSSInc, Chicago, Illinois, USA). All values are presented asmean � standard deviation. The data were analyzed using

FIGURE 3. Choroidal thickness map of an eye with angioid strFundus photograph of the right eye of a 68-year-old man; visua window defect. (Middle) A multi-averaged horizontal imageshows choroidal structures and the chorioscleral border. ThChoroidal thickness map of 6 � 6 mm2 of the macular area. ChIn addition, there is regional thickening in the outer superior arthe ETDRS grid to the map, mean choroidal thickness was ob

-way analysis of variance with Tukey’s post hoc analysis s


o compare mean choroidal thickness and volume atifferent regions. To compare the thickness and volumeata in each group, 2-way analysis of variance was used. Palues less than .05 were considered to be statistically

that had no choroidal neovascularization (Group 2). (Top left)ity was 20/20. (Top right) Fluorescein angiogram shows only

ined with swept source optical coherence tomography clearlyroid appears to have a physiologic thickness. (Bottom left)

d in the nasal quadrant is thinner than that of other quadrants.d in the outer inferotemporal area. (Bottom right) By applyingd for each sector.

eaksal acuobta

e chooroiea antaine

ignificant.


wo2

mm

RESULTS

IN THE CURRENT STUDY, 39 EYES (23 PATIENTS, 14 MEN AND

9 women) with angioid streaks secondary to pseudoxan-

FIGURE 4. Choroidal thickness map of an eye with choroidal nFundus photograph of the left eye of an 80-year-old woman;intravitreal injection of bevacizumab and 3 injections of ranibichoroidal neovascularization. (Middle) A multi-averaged horizonshows the thin choroid with type 1 choroidal neovascularizat

acular area. Choroidal thickness is reduced in the entire macean choroidal thickness was obtained for each area.

thoma elasticum, ranging in age from 57 to 80 years (mean: c


65.4 � 7.4 years), were examined. Mean refractive erroras �1.19 � 2.15 diopters (range: �5.5 to �1.75 di-pters). Mean axial length was 23.83 � 1.39 mm (range:2.15 to 26.54 mm). Eyes with angioid streaks were

scularization (CNV) associated with angioid streaks. (Top left)al acuity was 20/30. This eye was previously treated with 1b (Group 4). (Top right) Fluorescein angiogram shows occultmage obtained with swept source optical coherence tomography(Bottom left) Choroidal thickness map of 6 � 6 mm2 of thearea. (Bottom right) By applying the ETDRS grid to the map,

eovavisu

zumatal i

ion.ular

lassified into 4 groups: those with angioid streaks but with


amC(ttbtwb

otaattvcCtaoobww3s(

mt

admdc(ad

1

(1

3crG

no CNV (Group 2, n � 6); eyes with CNV secondary toangioid streaks with no history of treatments for theirCNV (Group 3, n � 7); eyes with CNV secondary tongioid streaks with a history of only anti-VEGF treat-ents for their CNV (Group 4, n � 11); and eyes withNV secondary to angioid streaks with a history of PDT

Group 5, n � 15). Table 1 shows the ocular characteris-ics of each group. All eyes in Group 4 received anti-VEGFreatments for CNV, with the mean number of injectionseing 3.4 � 2.1. Eyes in Group 5 received 2.9 � 1.9 PDTreatments and 3.1 � 3.9 anti-VEGF treatments. Thereere no differences in age, axial length, or refractive erroretween the 5 groups.

● CHOROIDAL STRUCTURE: The swept source OCT systemperated at 1050-nm wavelength range enables visualiza-ion of the clear structure of the posterior pole as well asllowing deeper penetration into the choroid. Multi-veraged scans of 12 mm in length revealed structures ofhe retina and choroid, which allowed precise identifica-ion in all eyes of the chorioscleral border beyond theascular arcade (Figure 1, Left). Figure 2 shows typicalases of each group. In eyes with angioid streaks withoutNV (Group 2), the choroid was well preserved and was as

hick as that of healthy controls (Group 1). Multi-aver-ged scans of the current OCT system often show a layerf medium-diameter blood vessels (Sattler’s layer) and anutermost layer of the choroid consisting of larger-diameterlood vessels (Haller’s layer). However, the choroid in eyesith CNV secondary to angioid streaks was remarkably thinith a disorganized configuration of choroidal vessels (Groups–5), regardless of a history of treatment. In eyes with CNVecondary to angioid streaks that had a history of PDT

TABLE 2. Choroidal Volume in Each Sector of Early TreatmeOptical Coher

Sector

Group 1

Normal Eyes

Group 2

AS Without CNV

Central area (P valuea) 0.187 � 0.059 0.188 � 0.036 (1.000)

Inner superior (P valuea) 0.375 � 0.114 0.402 � 0.081 (.947)

Inner nasal (P valuea) 0.354 � 0.118 0.331 � 0.084 (.990)

Inner inferior (P valuea) 0.368 � 0.113 0.352 � 0.097 (.989)

Inner temporal (P valuea) 0.369 � 0.119 0.371 � 0.073 (1.000)

Outer superior (P valuea) 1.301 � 0.379 1.361 � 0.322 (.984)

Outer nasal (P valuea) 0.939 � 0.375 0.826 � 0.218 (.878)

Outer inferior (P valuea) 1.111 � 0.394 1.173 � 0.414 (.983)

Outer temporal (P valuea) 1.185 � 0.378 1.197 � 0.231 (1.000)

AS � angioid streaks; CNV � choroidal neovascularization; VEGaP value, compared with values in Group 1 (mean choroidal volu

Tukey’s post hoc analysis).

Group 5), choroidal thickness was reduced not only in the i


acular area, where the PDT was performed, but also beyondhe vascular arcade.

● MEAN CHOROIDAL THICKNESS AND VOLUME: Usingraster scan protocol with 512 � 128 A-scans, 3D imagingata of the 6 � 6-mm2 area were acquired. By weightedoving average, each image had quality sufficient to allow

elineation of both the outer border of the RPE and thehorioscleral border, even in eyes with retinal pathologyFigure 1, Middle and Right; Supplemental Videos 1 and 2,vailable at AJO.com). Based on 128 images of the 3Data set, a choroidal thickness map of the 6 � 6-mm2 area

centered on the fovea was created for each eye. In thecurrent study, the mean choroidal thickness in the macula(within the outer circle of the ETDRS grid) was 218.8 �69.2 �m in Group 1, 218.9 � 46.8 �m in Group 2, 119.7 �49.2 �m in Group 3, 140.1 � 64.9 �m in Group 4, and44.0 � 52.6 �m in Group 5. Mean macular choroidal

thickness in Group 2 was as great as that in Group 1(Figure 3). However, mean macular choroidal thickness inGroups 3 through 5 was significantly less than that inGroup 1 (P � .05, respectively) (Figure 4). There were nostatistical differences between Groups 3 and 5.

● CHOROIDAL VOLUME: Macular choroidal volumewithin the outer circle of the ETDRS grid) was 6.183 �.957 mm3 in Group 1, 6.185 � 1.323 mm3 in Group 2,

3.381 � 1.389 mm3 in Group 3, 3.959 � 1.853 mm3 inGroup 4, and 4.070 � 1.487 mm3 in Group 5. In Groups

through 5, choroidal volume of the macula was signifi-antly reduced compared with that of Group 1 (P � .05,espectively). There were no statistical differences betweenroups 3 through 5. Table 2 shows the choroidal volume

abetic Retinopathy Study Grid Obtained With Swept SourceTomography

ean Choroidal Volume (mm3)

Group 3

S With CNV and No

istory of Treatments

Group 4

AS With CNV and a

History of Only Anti-VEGF

Treatments

Group 5

AS With CNV and a History

of Photodynamic Therapy

92 � 0.043 (.001) 0.113 � 0.056 (.002) 0.102 � 0.041 (�.001)

86 � 0.080 (.001) 0.232 � 0.119 (.005) 0.229 � 0.098 (.001)

50 � 0.615 (�.001) 0.203 � 0.117 (.004) 0.164 � 0.080 (�.001)

82 � 0.104 (.003) 0.207 � 0.108 (.002) 0.228 � 0.113 (.003)

12 � 0.092 (.007) 0.226 � 0.097 (.003) 0.249 � 0.090 (.006)

53 � 0.296 (.007) 0.887 � 0.422 (.018) 0.873 � 0.289 (.006)

84 � 0.192 (.002) 0.554 � 0.384 (.014) 0.473 � 0.220 (.001)

59 � 0.300 (.046) 0.713 � 0.324 (.029) 0.805 � 0.347 (.085)

70 � 0.286 (.033) 0.806 � 0.278 (.021) 0.946 � 0.280 (.139)

ascular endothelial growth factor.

n each area were compared using 1-way analysis of variance with

nt Dience

M

A

H

0.0

0.1

0.1

0.1

0.2

0.7

0.3

0.6

0.7

F � v

mes i

n each sector of the ETDRS grid. In most sectors,


http://AJO.com

2at(wos

etbcurwstfsdIs

choroidal volume in Groups 3 through 5 was significantlyless than that of Group 1, respectively.

● REGIONAL CHOROIDAL THICKNESS: Table 3 shows themean choroidal thickness of each area of the ETDRS grid in thefive groups. Mean choroidal thickness within the central areawas 238.7 � 75.0 �m in Group 1, 239.0 � 45.4 �m in Group, 117.4 � 55.9 �m in Group 3, 144.1 � 69.8 �m in Group 4,nd 130.5 � 52.8 �m in Group 5. However, in these choroidalhickness maps, there were regional irregularities in most eyesFigures 3 and 4). In all groups, the choroid in the nasal quadrantas thinner than in other quadrants; the mean thickness of theuter nasal sector was significantly less than that of the otherectors (P � .05, respectively).

DISCUSSION

ENHANCED DEPTH IMAGING OCT IS COUPLED USUALLY

TABLE 3. Mean Choroidal Thickness in Each Sector of theSwept Source Optica

Inner S

Group 1

Normal eyes (P valuea) 235.5 � 7

Group 2

AS without CNV (P valuea) 256.2 � 5

Group 3

AS with CNV and no history of treatments (P valuea) 118.4 � 5

Group 4

AS with CNV and a history of only anti-VEGF

treatments (P valuea)

148.1 � 7

Group 5

AS with CNV and a history of photodynamic therapy

(P valuea)

145.7 � 6

Outer

Group 1

Normal eyes (P valuea) 245.6 � 7

Group 2

AS without CNV (P valuea) 256.9 � 6

Group 3

AS with CNV and no history of treatments (P valuea) 142.1 � 5

Group 4

AS with CNV and a history of only anti-VEGF

treatments (P valuea)

167.5 � 7

Group 5

AS with CNV and a history of photodynamic therapy

(P valuea)

164.9 � 5

AS � angioid streaks; CNV � choroidal neovascularization; VEGaP value, compared with values in nasal sector (mean choroidal t

with Tukey’s post hoc analysis.)

with multiple averaging in order to produce high-contrast h


images with low speckle noise.16 In the studies that haveused the enhanced depth imaging technique, it was difficultto provide as many sections as does 3D imaging and thechoroidal thickness was measured only at the foveal center or,in some cases, at several measurement points.20,21,38–42 How-ver, because the choroid often shows focal thickening orhinning, measurement at a single point may be influencedy such irregularity or even by focal indistinctness of thehorioscleral border.11,18,43 In the current study, the mac-lar area of eyes with angioid streaks was examined with aaster scan protocol using a prototype swept source OCT,hich operated at the 1-�m wavelength region. Swept

ource OCT at this longer wavelength has higher penetra-ion and lower scattering at the RPE, and thus allows forull-depth imaging of the choroid.36 The tunable laserource in the swept source OCT has a lower signal decay vsepth than does the existing spectral-domain OCT system.n addition, high-speed scanning coupled with the highensitivity of the swept source OCT allowed us to obtain

Treatment Diabetic Retinopathy Study Grid Obtained Witherence Tomography

Mean Choroidal Thickness of Inner Ring Sectors (�m)

r Inner Nasal Inner Inferior Inner Temporal

06) 220.4 � 75.0 234.4 � 72.2 (.058) 235.5 � 75.4 (.043)

12) 210.9 � 53.6 224.0 � 61.9 (.724) 236.5 � 46.6 (.217)

19) 95.7 � 39.2 116.0 � 66.4 (.186) 135.3 � 58.5 (.003)

67) 129.4 � 74.9 132.1 � 69.0 (.983) 144.0 � 62.3 (.202)

.001) 104.9 � 51.2 145.1 � 72.2 (�.001) 158.7 � 57.6 (�.001)

Mean Choroidal Thickness of Outer Ring Sectors (�m)

r Outer Nasal Outer Inferior Outer Temporal

01) 177.2 � 70.9 209.6 � 74.5 (�.001) 223.6 � 73.1 (�.001)

01) 155.9 � 41.1 221.4 � 75.0 (.031) 222.8 � 40.2 (.027)

.001) 72.5 � 36.3 124.4 � 56.6 (�.001) 145.3 � 54.0 (�.001)

.001) 104.6 � 72.6 134.6 � 61.2 (.022) 155.1 � 51.5 (�.001)

.001) 89.4 � 41.6 152.0 � 65.2 (�.001) 178.3 � 53.0 (�.001)

ascular endothelial growth factor.

ess in each sector was compared using 2-way analysis of variance

Earlyl Coh

uperio

5.4 (.0

1.6 (.0

0.9 (.1

5.9 (.0

2.5 (�

Superio

1.6 (.0

0.8 (.0

5.9 (�

9.7 (�

4.5 (�

F � v

hickn

ighly reproducible measurements of the choroidal thick-


aRcbclTscs

eCCadcHrdcetlItaa

hs

iiearrcm(AOstttwet

obcqnudzaCietosetrwct

ness using a 3D raster scan protocol. Averaging of thick-ness values obtained at multiple measurement points in the3D data set contributed to lessening of the error indetermining choroidal thickness in the macular area.26,36

In the current study, the choroid of eyes with angioidstreaks without CNV was as thick as that in normal controls.Multi-averaged scans of the current OCT system often showa layer of medium-diameter blood vessels (Sattler’s layer) andan outermost layer of the choroid consisting of larger-diameter blood vessels (Haller’s layer).34 However, the cho-roid was significantly thinner in eyes with angioid streaksthat had developed CNV. Because the configuration ofchoroidal vessels is disorganized, it is difficult to determinewhich layer was primarily degenerated. Furthermore, be-cause this study was performed on a cross-sectional basis,the cause of this finding is uncertain, but several hypoth-eses can be proposed.

First, eyes with angioid streaks may show progressivethinning of the choroid over time. This thinning becomesmost evident at a later stage of the disease, at which timemany eyes also show secondary CNV. Second, a decreasein choroidal thickness may actually contribute to thedevelopment of CNV. It is possible that eyes with angioidstreaks have a wide range of choroidal thicknesses, andthat eyes with a thin choroid develop CNV, while eyeswith a more preserved choroid rarely do. Previous histo-pathologic reports4,44,45 have shown that angioid streaksre accompanied by atrophic degeneration of the overlyingPE and by focal breaks or even an absence of the underlyinghoriocapillaris. These histopathologic changes, which mighte attributable primarily to thickening of the choroidalapillary walls or, more probably, obliteration of theirumens, may be associated with development of CNV.hird, the atrophy of RPE and choriocapillaris that is often

een around the CNV may be associated with the macularhoroidal thinning in eyes with CNV secondary to angioidtreaks.

In the current study, the choroidal thinning was seen inyes with CNV, independent of a history of treatments forNV. However, it is possible that previous treatments forNV are also involved in the choroidal thinning. Maruko

nd associates20,42 reported recently that subfoveal choroi-al thickness is decreased after PDT in eyes with polypoidalhoroidal vasculopathy or central serous chorioretinopathy.owever, it has also been reported that the subfoveal cho-

oidal thickness is increased during the active phase of theseiseases,11,39 so it is possible that PDT does not make thehoroid pathologically thin by its direct photoreactiveffect on the choroidal vessels, but, rather, only reduceshe increased subfoveal choroidal thickness to physiologicevels by the reduction of the choroidal hyperpermeability.n our patients treated with PDT (Group 5), choroidalhickness was significantly less, not only in the macularrea, where the PDT had been previously performed, but
lso beyond the vascular arcade. The choroid in eyes that

ad developed CNV associated with angioid streaks wasignificantly thinner than that of normal eyes.

Anti-VEGF treatments for CNV may also be involvedn the thinner choroid in Groups 4 and 5. In a computer-zed search of PubMed, however, we could find no refer-nce to a reduction of choroidal thickness attributable tonti-VEGF treatment. Recently, Koizumi and associateseported that in eyes with exudative AMD treated withanibizumab, the mean subfoveal choroidal thickness de-reased from 244 �m at the baseline to 226 �m at 3onths, 229 �m at 6 months, and 226 �m at 12 months

P � .01) (Poster presentation at 2011 Annual Meeting ofmerican Academy of Ophthalmology, Orlando, Florida,ctober 22–25, 2011). Although the authors found a

tatistically significant decrease in the foveal choroidalhickness, the mean reduction was only 18 �m. Based onhis evidence, the effect of either PDT or anti-VEGFreatments on the choroidal thickness in our patientsould be limited. Previous treatments for CNV do notxplain sufficiently the decrease in macular choroidalhickness in eyes with CNV secondary to angioid streaks.

The current results have shown the asymmetric naturef macular choroidal thickness not only in normal controlsut in eyes with angioid streaks. In our subjects, thehoroid in the nasal quadrant was thinner than in otheruadrants; this thinning was most evident in the outerasal sector. The reason for nasal choroidal thinning isnknown, but it may be related to the vascular bedistribution of the choroid and the choroidal watershedone, which is frequently observed between the maculand optic disc.46 In eyes with angioid streaks, secondaryNV occurs usually adjacent to angioid streaks, sometimes

n the peripapillary region.3,4 Angioid streaks per se may bessential to the development of CNV. In addition, thishinning of nasal choroid might be related to the frequentccurrence of CNV secondary to angioid streaks at thisite. However, while eyes with high myopia often show anxtremely thin choroid, especially in the nasal quadrant ofhe macula, myopic CNV is seen usually around the fovea,arely in the peripapillary area. In addition, all such eyesith a thin choroid do not show myopic CNV. Recently,horoidal thinning has been reported in eyes with exuda-ive AMD,11 but so far, the clinical significance of choroi-

dal thinning in the development of CNV has beeninsufficiently elucidated.47

One of the major limitations of the current study is thesmall sample size in each study group. Unexpectedly, themacular choroidal thickness in Group 3 was slightlythinner than that in Groups 4 and 5, although thedifference was not statistically significant (P � .956 andP � .903, respectively). This difference could be explainedby the variability of the macular choroidal thickness ineligible eyes, because the sample size of Group 3 was only7. Because of this limitation, we could not preciselyestimate the effect of previous treatments for CNV on the
choroidal thickness. Another limitation of the current

eTs

study is that the line of RPE and the chorioscleral borderwere determined manually. The raster scan protocol, withover 60 000 measurement points, would have minimizedany error in measurement made by the observers, sosoftware to determine these lines automatically is essentialto further standardize this technique.

In spite of these limitations, our study did show thatswept source OCT at 1 �m allows us to comprehensivelyvaluate the macular choroid in eyes with angioid streaks.he choroid in eyes without CNV secondary to angioid

treaks was as thick as that in normal controls, although

840–845.

1

1

1

1

1

1

1

1

2

2

2

2


the choroid was significantly thinner in eyes with angioidstreaks that had developed CNV. In addition, the currentstudy confirmed the asymmetric nature of the choroid inthe macular area in eyes with angioid streaks. However, asthis was a cross-sectional study, various interpretations ofour findings are possible and it is difficult to determinewhether choroidal thinning is a cause or a consequence ofthe development of CNV. Additional longitudinal pro-spective studies are needed to elucidate the role played bythe choroid in the development of CNV in eyes with
angioid streaks secondary to pseudoxanthoma elasticum.
ALL AUTHORS HAVE COMPLETED AND SUBMITTED THE ICMJE FORM FOR DISCLOSURE OF POTENTIAL CONFLICTS OFInterest. Publication of this article was supported in part by the Japan Society for the Promotion of Science (JSPS), Tokyo, Japan (Grant-in-Aid forScientific Research, no. 21592256), and by the Japan National Society for the Prevention of Blindness, Tokyo, Japan. The authors disclose no financialconflicts of interest. Involved in conception and design of the study (A.A.E., A.T., A.M., N.Y.); analysis and interpretation (A.A.E., A.T., A.M., K.O.,M.H., S.O., K.Y., M.A., N.Y.); writing of the article (A.A.E., A.T.); critical revision of the article (A.M., K.O., M.H., S.O., K.Y., M.A., N.Y.); finalapproval of the article (A.A.E., A.T., A.M., K.O., M.H., S.O., K.Y., M.A., N.Y.); and data collection (A.A.E., A.T., A.M., M.H., S.O., K.Y.). TheInstitutional Review Board and Ethics Committee of Kyoto University approved this study, which adhered to the tenets of the Declaration of Helsinki.Written informed consent for research participation was obtained from each subject before examination.

REFERENCES

1. Le Saux O, Urban Z, Tschuch C, et al. Mutations in a geneencoding an ABC transporter cause pseudoxanthoma elasti-cum. Nat Genet 2000;25(2):223–227.

2. Finger RP, Charbel Issa P, Ladewig MS, et al. Pseudoxan-thoma elasticum: genetics, clinical manifestations and ther-apeutic approaches. Surv Ophthalmol 2009;54(2):272–285.

3. Clarkson JG, Altman RD. Angioid streaks. Surv Ophthalmol1982;26(5):235–246.

4. Dreyer R, Green WR. The pathology of angioid streaks: astudy of twenty-one cases. Trans Pa Acad OphthalmolOtolaryngol 1978;31(2):158–167.

5. Mimoun G, Tilleul J, Leys A, Coscas G, Soubrane G, SouiedEH. Intravitreal ranibizumab for choroidal neovasculariza-tion in angioid streaks. Am J Ophthalmol 2010;150(5):692–700.

6. Myung JS, Bhatnagar P, Spaide RF, et al. Long-term out-comes of intravitreal antivascular endothelial growth factortherapy for the management of choroidal neovascularizationin pseudoxanthoma elasticum. Retina 2010;30(5):748–755.

7. Connor PJ Jr, Juergens JL, Perry HO, Hollenhorst RW,Edwards JE. Pseudoxanthoma elasticum and angioid streaks.A review of 106 cases. Am J Med 1961;30:537–543.

8. Shaikh S, Ruby AJ, Williams GA. Photodynamic therapyusing verteporfin for choroidal neovascularization in angioidstreaks. Am J Ophthalmol 2003;135(1):1–6.

9. Mennel S, Schmidt JC, Meyer CH. Therapeutic strategies inchoroidal neovascularizations secondary to angioid streaks.Am J Ophthalmol 2003;136(3):580–582.

10. Linsenmeier RA, Padnick-Silver L. Metabolic dependenceof photoreceptors on the choroid in the normal anddetached retina. Invest Ophthalmol Vis Sci 2000;41(10):3117–3123.

11. Chung SE, Kang SW, Lee JH, Kim YT. Choroidal thicknessin polypoidal choroidal vasculopathy and exudative age-related macular degeneration. Ophthalmology 2011;118(5):

2. Gomi F, Tano Y. Polypoidal choroidal vasculopathy andtreatments. Curr Opin Ophthalmol 2008;19(3):208–212.

3. Spaide RF, Hall L, Haas A, et al. Indocyanine greenvideoangiography of older patients with central serous cho-rioretinopathy. Retina 1996;16(3):203–213.

4. Fong AH, Li KK, Wong D. Choroidal evaluation usingenhanced depth imaging spectral-domain optical coherencetomography in Vogt-Koyanagi-Harada disease. Retina 2011;31(3):502–509.

5. Fujiwara T, Imamura Y, Margolis R, Slakter JS, Spaide RF.Enhanced depth imaging optical coherence tomography ofthe choroid in highly myopic eyes. Am J Ophthalmol2009;148(3):445–450.

6. Spaide RF, Koizumi H, Pozzoni MC. Enhanced depth imag-ing spectral-domain optical coherence tomography. Am JOphthalmol 2008;146(4):496–500.

7. van Velthoven ME, Faber DJ, Verbraak FD, van LeeuwenTG, de Smet MD. Recent developments in optical coher-ence tomography for imaging the retina. Prog Retin Eye Res2007;26(1):57–77.

8. Koizumi H, Yamagishi T, Yamazaki T, Kawasaki R, Kinoshita S.Subfoveal choroidal thickness in typical age-related maculardegeneration and polypoidal choroidal vasculopathy. GraefesArch Clin Exp Ophthalmol 2011;249(8):1123–1128.

9. Margolis R, Spaide RF. A pilot study of enhanced depthimaging optical coherence tomography of the choroid innormal eyes. Am J Ophthalmol 2009;147(5):811–815.

0. Maruko I, Iida T, Sugano Y, et al. Subfoveal choroidalthickness after treatment of Vogt-Koyanagi-Harada disease.Retina 2011;31(3):510–517.

1. Reibaldi M, Boscia F, Avitabile T, et al. Enhanced depthimaging optical coherence tomography of the choroid inidiopathic macular hole: A cross-sectional prospective study.Am J Ophthalmol 2011;151(1):112–117.

2. Spaide RF. Age-related choroidal atrophy. Am J Ophthalmol2009;147(5):801–810.

3. Yeoh J, Rahman W, Chen F, et al. Choroidal imaging in
inherited retinal disease using the technique of enhanced

3

3

3

3

4

4

4

4

4

4

4

4

depth imaging optical coherence tomography. Graefes ArchClin Exp Ophthalmol 2010;248(12):1719–1728.

24. Blackburn J, McGwin G Jr. Enhanced depth imaging opticalcoherence tomography of the choroid in idiopathic macularhole. Am J Ophthalmol 2011;151(3):560–561.

25. Imamura Y, Iida T, Maruko I, Zweifel SA, Spaide RF.Enhanced depth imaging optical coherence tomography ofthe sclera in dome-shaped macula. Am J Ophthalmol 2011;151(2):297–302.

26. Agawa T, Miura M, Ikuno Y, et al. Choroidal thicknessmeasurement in healthy Japanese subjects by three-dimen-sional high-penetration optical coherence tomography.Graefes Arch Clin Exp Ophthalmol 2011;249(10):1485–1492.

27. Ikuno Y, Kawaguchi K, Nouchi T, Yasuno Y. Choroidalthickness in healthy Japanese subjects. Invest OphthalmolVis Sci 2010;51(4):2173–2176.

28. Ikuno Y, Maruko I, Yasuno Y, et al. Reproducibility of retinaland choroidal thickness measurements in enhanced depthimaging and high-penetration optical coherence tomogra-phy. Invest Ophthalmol Vis Sci 2011;52(8):5536–5540.

29. de Bruin DM, Burnes DL, Loewenstein J, et al. In vivothree-dimensional imaging of neovascular age-related maculardegeneration using optical frequency domain imaging at 1050nm. Invest Ophthalmol Vis Sci 2008;49(10):4545–4552.

30. Yasuno Y, Hong Y, Makita S, et al. In vivo high-contrastimaging of deep posterior eye by 1-micron swept sourceoptical coherence tomography and scattering optical coher-ence angiography. Opt Express 2007;15(10):6121–6139.

31. Esmaeelpour M, Povazay B, Hermann B, et al. Mappingchoroidal and retinal thickness variation in type 2 diabetesusing three-dimensional 1060-nm optical coherence tomog-raphy. Invest Ophthalmol Vis Sci 2011;52(8):5311–5316.

32. Esmaeelpour M, Povazay B, Hermann B, et al. Three-dimensional 1060-nm OCT: choroidal thickness maps innormal subjects and improved posterior segment visualiza-tion in cataract patients. Invest Ophthalmol Vis Sci 2010;51(10):5260–5266.

33. Wood A, Binns A, Margrain T, et al. Retinal and choroidalthickness in early age-related macular degeneration. Am JOphthalmol 2011;152(6):1030–1038.

34. Povazay B, Hermann B, Hofer B, et al. Wide-field opticalcoherence tomography of the choroid in vivo. Invest Oph-thalmol Vis Sci 2009;50(4):1856–1863.

35. Shields JA, Federman JL, Tomer TL, Annesley WH Jr.Angioid streaks. I. Ophthalmoscopic variations and diagnos-
tic problems. Br J Ophthalmol 1975;59(5):257–266.

6. Hirata M, Tsujikawa A, Matsumoto A, et al. Macularchoroidal thickness and volume in normal subjects measuredby swept-source optical coherence tomography. Invest Oph-thalmol Vis Sci 2011;52(8):4971–4978.

7. Early Treatment Diabetic Retinopathy Study research group.Photocoagulation for diabetic macular edema. Early Treat-ment Diabetic Retinopathy Study report number 1. ArchOphthalmol 1985;103(12):1796–1806.

8. Spaide RF. Enhanced depth imaging optical coherencetomography of retinal pigment epithelial detachment inage-related macular degeneration. Am J Ophthalmol 2009;147(4):644–652.

9. Imamura Y, Fujiwara T, Margolis R, Spaide RF. Enhanceddepth imaging optical coherence tomography of the choroidin central serous chorioretinopathy. Retina 2009;29(10):1469–1473.

0. Maruko I, Iida T, Sugano Y, Ojima A, Ogasawara M, SpaideRF. Subfoveal choroidal thickness after treatment of centralserous chorioretinopathy. Ophthalmology 2010;117(9):1792–1799.

1. Vance SK, Imamura Y, Freund KB. The effects of sildenafilcitrate on choroidal thickness as determined by enhanceddepth imaging optical coherence tomography. Retina 2011;31(2):332–335.

2. Maruko I, Iida T, Sugano Y, Saito M, Sekiryu T. Subfovealretinal and choroidal thickness after verteporfin photody-namic therapy for polypoidal choroidal vasculopathy. Am JOphthalmol 2011;151(4):594–603.

3. Rahman W, Chen FK, Yeoh J, Patel P, Tufail A, Da Cruz L.Repeatability of manual subfoveal choroidal thickness mea-surements in healthy subjects using the technique of en-hanced depth imaging optical coherence tomography. InvestOphthalmol Vis Sci 2011;52(5):2267–2271.

4. Gass JDM, Clarkson JG. Angioid streaks and disciformmacular detachment in Pagets disease (osteitis deformans).Am J Ophthalmol 1973;75(4):576–586.

5. Walker ER, Frederickson RG, Mayes MD. The mineraliza-tion of elastic fibers and alterations of extracellular matrix inpseudoxanthoma elasticum. Ultrastructure, immunocyto-chemistry, and X-ray analysis. Arch Dermatol 1989;125(1):70–76.

6. Hayreh SS. In vivo choroidal circulation and its watershedzones. Eye (Lond) 1990;4(Pt 2):273–289.

7. Nickla DL, Wallman J. The multifunctional choroid. Prog
Retin Eye Res 2010;29(2):144–168.

Biosketch

Abdallah A. Ellabban, MD, graduated from Suez Canal University (Egypt) and obtained his medical degree in 2003. Hecompleted his residency program and obtained a masters degree in ophthalmology from Suez Canal University in 2007.Dr Abdallah is now in a PhD program in the Department of Ophthalmology and Visual Sciences at Kyoto University(Japan) under the supervision of Professor Nagahisa Yoshimura.

MACULAR CHOROIDAL THICKNESS IN ANGIOID STREAKSVOL. 153, NO. 6 1143.e1

Health & Medicine

Macular Choroidal Thickness and Volume in Eyes With Angioid Streaks Measured by Swept Source OCT