Reproducibility of Cornea Measurements in Anterior Segment OCT Images of Normal Eyes and Eyes with...

Reproducibility of Cornea Measurements in AnteriorSegment OCT Images of Normal Eyes and Eyes withBullous Keratopathy Analyzed with the ZhongshanAssessment Program

Joy B. Chan,1,2 Leonard H. Yuen,1,2 Elaine H. Huang,1,2 Hla M. Htoon,1 Mingguang He,3

Tin Aung,1,2,4 Donald T. Tan,1,2,4 and Jodhbir S. Mehta1,2,4,5

PURPOSE. To determine the interobserver and intraobservermeasurement reproducibility of cornea parameters of bothnormal eyes and eyes with bullous keratopathy (BK) obtainedwith the Zhongshan Assessment Program (ZAP) on anteriorsegment optical coherence tomography (AS-OCT) images.

METHODS. A comparative study was carried out on 24 healthyvolunteers and 25 subjects with BK. AS-OCT images wereindependently analyzed by two examiners. Parameters exam-ined: anterior chamber depth (ACD), central corneal thickness(CCT), posterior corneal curvature (PCC), and posterior cor-neal arc length (PCAL). Interobserver and intraobserver repro-ducibility of these parameters was calculated in terms of limitsof agreement (mean of differences � 1.96SD of differences).

RESULTS. In the normal group, both horizontal and vertical ACDwere successfully measured in 23 of 24 (96%) images. Themean bias for two measurements by two different observersranged from 0.003 to 0.117 mm for ACD, PCC, and PCALmeasurements and from 0.013 to 2.25 �m for CCT measure-ments, and there were no differences between the two observ-ers (P � 0.05). Mean bias for two measurements by the samegrader ranged from 0.005 to 0.327 mm for ACD, PCC, andPCAL measurements and 1.46 to 2.53 �m for CCT measure-ments. There was no difference between the two observations(P � 0.05). Similar results were found in the BK group.

CONCLUSIONS. There was high inter- and intraobserver reproduc-ibility for normal and pathologic corneas using the ZAP soft-ware. The ZAP software may serve as a new investigatory toolfor accurately evaluating the anterior segment and cornealparameters for corneal procedures. (Invest Ophthalmol VisSci. 2011;52:8884–8890) DOI:10.1167/iovs.10-6411

Anterior segment optical coherence tomography (AS-OCT,Visante; Carl Zeiss Meditec, Dublin, CA) is a rapid, non-

contact modality for imaging the anterior segment.1–4 It iscapable of real-time imaging of a cross-section of the entirecornea, both angles as well as the anterior lens surface, and itsease of acquisition makes it ideal for quantitative measurementof corneal parameters.1

Quantitative measurements of corneal and anterior segmentparameters have become increasingly important in the era ofmodern anterior segment surgery. Accurate measurements ofthe internal cornea diameter would, for example, enable cor-neal surgeons to accurately size endothelial grafts for endothe-lial keratoplasty procedures, such as Descemet’s strippingautomated endothelial keratoplasty (DSAEK).5 Precise mea-surements of the internal corneal width would allow forbetter sizing of anterior chamber phakic intraocular lenses(IOLs).6 The follow-up of other novel corneal refractive pro-cedures (e.g., implantation of a corneal stromal inlay for thecorrection of presbyopia; AcuFocus, Inc., Irvine, CA)7 is de-pendent on accurate imaging of the implant and precise mea-surement of the depth of implantation.

The Zhongshan Assessment Program (ZAP; Guangzhou,China) is a customized algorithm software developed to obtainquantitative measurements of anterior segment parametersfrom AS-OCT images. Both its clinical and research utility andits measurement reproducibility have been documented formeasurements of the anterior chamber angle.8 Modification ofthe software allows quantitative measurements of corneal pa-rameters such as central corneal thickness (CCT), posteriorcorneal curvature (PCC), and posterior corneal arc length(PCAL), a novel parameter defined as the distance betweensclera spurs.9 This distance gives a good approximation of theinternal corneal width, a parameter that has not previouslybeen accurately and reproducibly measured. Quantitative mea-surements are objective and repeatable, and thus are useful forthe diagnosis and subsequent follow-up of anterior segmentconditions. However, the accuracy of quantitative measure-ments is influenced by the reproducibility of the instrumentused. The aim of this study was to evaluate the reproducibilityof corneal measurements in AS-OCT images of both normalcorneas and corneas with bullous keratopathy (BK) using theZAP software. This would allow validation of the software forfuture studies.

METHODS

Subjects

Two groups of subjects participated in this prospective study. The firstgroup comprised 24 normal volunteers and the second group com-

From the 1Singapore Eye Research Institute, Singpore; the 2Singa-pore National Eye Center, Singapore; the 3Zhongshan OphthalmicCenter, Guangzhou, China; the 4Department of Ophthalmology, YongLoo Lin School of Medicine, National University of Singapore, Singa-pore; and 5Duke National University of Singapore Graduate MedicalSchool, Singapore.

Supported in part by National Research Foundation FundedTranslational and Clinical Research Programme Grant NMRC/TCR/002-SERI/2008.

Submitted for publication August 15, 2010; revised February 10and July 5, 2011; accepted July 17, 2011.

Disclosure: J.B. Chan, None; L.H. Yuen, None; E.H. Huang,None; H.M. Htoon, None; M. He, None; T. Aung, None; D.T. Tan,None; J.S. Mehta, None

Corresponding author: Jodhbir S. Mehta, Singapore National EyeCentre, 11 Third Hospital Avenue, Singapore 168751;[email protected].

Cornea

Investigative Ophthalmology & Visual Science, November 2011, Vol. 52, No. 128884 Copyright 2011 The Association for Research in Vision and Ophthalmology, Inc.

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/Journals/IOVS/933459/ on 09/18/2015

prised 25 subjects with BK. The study had the approval of the hospi-tal’s Ethics Committee and was conducted according to the tenets ofthe Declaration of Helsinki. Written informed consent was obtainedfrom study subjects after the nature and intent of the study had beenfully explained to them. The exclusion criteria for the normal groupwere any ocular abnormalities, history of eye disease, prior refractivesurgery, and contact lens wear. In the BK group, eyes with mild tomoderate corneal clouding affecting visual acuity and requiring DSAEKwere included. The authors chose this condition because it is commonand mild to moderate corneal clouding will have only a minimal effecton AS-OCT image acquisition (see Figs. 2 and 3 in the following text).This disease is the primary indication in which PCAL will be mostuseful to measure graft size from the posterior border. The inclusioncriteria are all patients with mild to moderate corneal edema affectingvisual acuity requiring DSAEK.

All scanning was performed in the undilated left eye in the normalgroup. In the BK group, scanning was done in the pathologic eye. Thiswas the right eye in 10 subjects and the left eye in the remaining 15subjects. During scanning, subjects were asked to fix on the externalfixation lighting the contralateral eye. The external fixation light waspositioned so that the subject was looking straight ahead.

Scanning

Performance of the AS-OCT (Visante; Carl Zeiss Meditec, Dublin, CA)has been described previously.10 Study subjects underwent slit-lampexamination, in a sitting position by a single examiner (JC), beforeimaging. AS-OCT images of the anterior chamber were obtained underdark conditions using the standard anterior segment single-scan proto-col. To obtain the best-quality image, the examiner adjusted the satu-ration and noise and optimized the polarization for each scan duringthe examination.

Patients were instructed to keep their eyes wide open duringscanning and, when necessary, the lids were gently held apart (withcare not to exert pressure on the globe) to ensure that the lids did notblock the 10-mm-diameter corneal mapping. The operator adjusted thesystem to position the vertex at the center of the AS-OCT image and tomaximize the vertex reflection. All scanning was performed by thesame operator (JC). All measurements were taken between 10:00 AMand 4:00 PM (to minimize any effect of diurnal variation on cornealthickness).11–13

In the control group, both vertical and horizontal measurementswere taken and compared to validate the software, despite the verticalmeasurement being harder to obtain. In the BK group, measurementswere taken only in the horizontal meridians because having alreadyvalidated the controls in both meridians, we wanted to validate onlythe consistency of horizontal scans that were part of the study. Verticalscans were more difficult to obtain, for AS-OCT imaging and for scleralspur identification, due to eyelid blocking, and thus not used in thediseased group.

Image Processing and Analysis

The AS-OCT scans were processed using in-built dewarping softwarethat adjusted for image distortion resulting from corneal optical prop-erties. Scleral spur location was the only observer input requiredduring image assessment. This was defined by a prominent innerextension of the sclera at its thickest part,14–16 and often appeared asan inward protrusion or change in the curvature of the inner surface ofthe angle wall.

The Zhongshan Assessment Program (ZAP, Guangzhou, China)software automatically extracted the 300 � 600 8-bit grayscale (inten-sities from 0 to 255) image portion of the output file and performednoise and contrast conditioning.8 A binary copy of the image was thenproduced where pixels were either 1’s (tissue) or 0’s (open space)depending on whether they were brighter or darker than a calculatedthreshold value. Algorithms defined the borders of the corneal epithe-lium and endothelium and the anterior surface of the iris. The algo-rithms used basic edge arguments (5 consecutive 0’s above, and 5consecutive 1’s below indicated an anterior surface point) to describethe borders. The corneal border data were fitted with polynomiccurves and a line-smoothing algorithm, explicitly defined by the edgefinding algorithms, used the derivative data to repair step-like portionsof the border. The radius (in mm) of the curve on the central one thirdof the posterior cornea was calculated as PCC.

The parameters obtained from image analysis included PCC (inmm), CCT (in �m), anterior chamber depth (ACD, in mm), and PCAL(in mm). PCC is defined as the radius of curvature of the posteriorsurface of the cornea. CCT was measured from the anterior to theposterior surface of the cornea, taken at the center of the AS-OCTimage. ACD was measured from the posterior surface of the cornea tothe anterior surface of the lens, also taken at the center of the AS-OCTimage. PCAL is a novel parameter, defined as the arc distance betweentwo scleral spurs along the corneal endothelium (Fig. 1).

Repeatability of Image Analysis

Two ophthalmologists (JC and EH) independently carried out theimage analysis for the full set of images of the first group (normalsubjects) using the ZAP software. One week later, one of the examin-ers (JC) repeated the analysis for the same set of images and she wasmasked to the results of initial analysis.

For the BK group, two ophthalmologists (JC and LY) carried out theimage analysis for the full set of images of the second group using theZAP software. One week later, one of the examiners (JC) repeatedthe analysis for the same set of images and again she was masked to theresults of initial analysis. Examples of images from the BK group areshown in Figures 2 and 3.

Statistical Methods

Bland Altman analysis was performed to analyze inter- and intraob-server agreement (MedCalc, Version 9.6.4.0; MedCalc Software, Mari-

FIGURE 1. An AS-OCT image analyzed by ZAP software.

FIGURE 2. AS-OCT image of bullous keratopathy showing good pen-etrance and minimal effect of corneal edema on scan.

IOVS, November 2011, Vol. 52, No. 12 Reproducibility Cornea AS-OCT Zhongshan Assessment Program 8885


akerke, Belgium). Interobserver and intraobserver reproducibility ofthe above parameters was calculated in terms of limits of agreement(LOA; mean of differences � 1.96SD of differences). Paired t-tests wereused for the differences between observer measurements. Statisticalanalysis was performed with data analysis/spreadsheet software (Mi-crosoft Excel; Microsoft Corp., Redmond, WA).

RESULTS

Analysis in the Normal Group

The normal group of study participants consisted of 24 volunteers(10 males and 14 females), ranging in age from 21 to 55 years (meanage, 30.48 years). The ethnicity of normal subjects was distributed asfollows: 12 Chinese, 7 Indian, 3 Malay, and 2 Burmese.

Intraobserver Reproducibility. In both horizontal and ver-tical meridians, the mean of differences (bias) between two re-peated measurements by the same grader was small, and LOAvalues were correspondingly small. In the horizontal meridian,intraobserver bias was 0.196 mm for PCAL (P � 0.189), 0.005 mmfor ACD (P � 0.941), and 0.055 mm for PCC (P � 0.613) measure-ments, respectively, and was 2.533 �m for CCT (P � 0.788) mea-surements. In the vertical meridian, intraobserver bias was 0.327 mmfor PCAL (P � 0.082), 0.015 mm for ACD (P � 0.649), and 0.006 mmfor PCC (P � 0.946) measurements, and was 1.463 �m for CCT (P �0.874) measurements (Table 1 and Fig. 3).

Interobserver Reproducibility. Comparing parameter forparameter, the mean of differences between two graders waslower for all parameters except horizontal ACD and horizontalPCC. However, the LOA values in the interobserver agreementwere similar to those in the intraobserver agreement for bothhorizontal and vertical scans. Interobserver bias in horizontalmeasurements was 0.025 mm for PCAL (P � 0.857), 0.009 mmfor ACD (P � 0.901), 0.086 mm for PCC (P � 0.446) measure-ments, and was 0.013 �m for CCT (P � 0.999) measurements.Interobserver bias in vertical measurements was 0.117 mm forPCAL (P � 0.498), 0.003 mm for ACD (P � 0.967), 0.017 mmfor PCC (P � 0.843) measurements, and was 2.250 �m for CCT(P � 804) measurements.

ANALYSIS IN THE BK GROUP

The BK study group consisted of 25 patients with BK of varyingcauses. There were 9 males and 16 females, ranging in age from56 to 79 years. The ethnicity of subjects with BK was distrib-uted as follows: 17 Chinese, 5 Indonesian, 2 Vietnamese, and 1Eurasian. All parameters were measured in the horizontal me-ridian only.

Intraobserver Reproducibility. As with the normalgroup, the mean of differences and limits of agreement forrepeated measurements by the same grader were small. Agree-

FIGURE 3. Bland Altman plots of corneal measurements of normal corneas. Intraobserver agreement, top left: ACD measurements by one grader(repeated); top right: CCT measurements by one grader (repeated); bottom left: PCC measurements by one grader (repeated); bottom right: PCALmeasurements by one grader (repeated).

8886 Chan et al. IOVS, November 2011, Vol. 52, No. 12


TA

BLE

1.

Inte

rob

serv

eran

dIn

trao

bse

rver

Var

iab

ility

of

Par

amet

ers

of

No

rmal

Co

rnea

sA

nal

yzed

wit

hZ

AP

Soft

war

e

Par

amet

erM

ean

(SD

)O

bse

rver

1M

ean

(SD

)O

bse

rver

2P

(Tw

o-s

ided

)M

ean

Bia

s(9

5%

CI)

Lim

its

of

Agr

eem

ent

Low

erLi

mit

Up

per

Lim

it

Inte

rob

serv

erva

riab

ility

Ho

rizo

nta

lP

CA

L12

.928

(0.4

47)

12.9

03(0

.483

)0.

857

�0.

025

(�0.

145

to0.

095)

�0.

581

(�0.

789

to�

0.37

3)0.

532

(0.3

24to

0.74

0)H

ori

zon

tal

CC

T57

3.00

0(3

3.24

0)57

3.01

3(3

2.65

9)0.

999

0.01

3(�

2.86

5to

2.89

0)�

13.3

46(�

18.3

34to

�8.

358)

13.3

71(8

.383

to18

.359

)H

ori

zon

tal

AC

D3.

161

(0.2

40)

3.16

2(0

.238

)0.

901

0.00

9(�

0.00

9to

0.02

6)�

0.07

2(�

0.10

2to

�0.

042)

0.08

9(0

.059

to0.

118)

Ho

rizo

nta

lP

CC

6.44

0(0

.389

)6.

527

(0.3

88)

0.44

60.

086

(�0.

036

to0.

209)

�0.

481

(�0.

693

to�

0.26

9)0.

654

(0.4

42to

0.86

6)V

erti

cal

PC

AL

13.2

98(0

.599

)13

.182

(0.5

84)

0.49

8�

0.11

7(�

0.21

2to

�0.

021)

�0.

559

(�0.

725

to�

0.39

4)0.

326

(0.1

61to

0.49

1)V

erti

cal

CC

T56

9.48

0(3

1.33

1)56

9.72

9(3

1.06

8)0.

804

2.25

0(�

0.26

2to

4.76

2)�

9.40

8(�

13.7

61to

�5.

055)

13.9

08(9

.555

to18

.261

)V

erti

cal

AC

D3.

149

(0.2

36)

3.14

6(0

.241

)0.

967

�0.

003

(�0.

006

to�

0.00

0)�

0.01

5(�

0.02

0to

�0.

011)

0.01

0(0

.005

to0.

014)

Ver

tica

lP

CC

6.41

0(0

.277

)6.

428

(0.3

17)

0.84

30.

017

(�0.

038

to0.

072)

�0.

236

(�0.

331

to�

0.14

2)0.

271

(0.1

76to

0.36

5)

Par

amet

erFir

stA

nal

ysi

sM

ean

(SD

)Se

con

dA

nal

ysi

sM

ean

(SD

)P

(Tw

o-s

ided

)M

ean

Bia

s(9

5%

CI)

Lim

its

of

Agr

eem

ent

Low

erLi

mit

Up

per

Lim

it

Intr

aob

serv

erva

riab

ility

Ho

rizo

nta

lP

CA

L12

.927

8(0

.447

)12

.732

(0.5

62)

0.18

90.

196

(0.0

37to

0.35

5)�

0.54

2(�

0.81

7to

�0.

266)

0.93

3(0

.658

to1.

208)

Ho

rizo

nta

lC

CT

573.

000

(33.

240)

570.

467

(31.

493)

0.78

82.

533

(�2.

075

to7.

142)

�18

.858

(�26

.847

to�

10.8

70)

23.9

25(1

5.93

7to

31.9

13)

Ho

rizo

nta

lA

CD

3.16

1(0

.240

)3.

166

(0.2

36)

0.94

1�

0.00

5(�

0.03

6to

0.02

6)�

0.14

7(�

0.20

0to

�0.

094)

0.13

7(�

0.08

4to

0.19

0)H

ori

zon

tal

PC

C6.

440

(0.3

89)

6.38

6(0

.353

)0.

613

0.05

5(�

0.05

5to

0.16

5)�

0.45

6(�

0.64

7to

�0.

265)

0.56

5(0

.375

to0.

756)

Ver

tica

lP

CA

L13

.298

(0.5

99)

12.9

70(0

.673

)0.

082

0.32

7(0

.165

to0.

489)

�0.

424

(�0.

705

to�

0.14

4)1.

079

(0.7

98to

1.36

0)V

erti

cal

CC

T56

9.48

0(3

1.33

1)56

8.94

2(3

2.00

7)0.

874

�1.

463

(�7.

704

to4.

779)

�30

.433

(�41

.251

to�

19.6

15)

27.5

08(1

6.69

0to

38.3

26)

Ver

tica

lA

CD

3.14

9(0

.236

)3.

118

(0.2

24)

0.64

90.

015

(�0.

018

to0.

049)

�0.

137

(�0.

195

to�

0.07

9)0.

168

(0.1

09to

0.22

6)V

erti

cal

PC

C6.

410

(0.2

77)

6.41

7(3

53.0

00)

0.94

6�

0.00

6(�

0.10

6to

0.09

3)�

0.46

7(�

0.63

9to

�0.

295)

0.45

5(0

.283

to0.

627)

CI,

con

fid

ence

inte

rval

.



ment for intraobserver measurements was equally strong for allparameters measured. The bias was 0.06 mm for PCAL (P �0.104), 0.037 mm for ACD (P � 0.275), and 0.059 mm for PCC(P � 0.065) measurements, and was 0.371 �m for CCT (P �0.767) measurements (Table 2).

Interobserver Reproducibility. When horizontal mea-surements were performed by two graders, the mean of differ-ences was smaller than that of repeated horizontal measure-ments by the same grader. The interobserver bias was 0.0017mm for PCAL (P � 0.761), 0.00046 mm for ACD (P � 0.741),and 0.043 mm for PCC (P � 0.311) measurements, and was1.57 �m for CCT (P � 0.315) measurements. The limits ofagreement for measurements by two graders was correspond-ingly small. Agreement for interobserver measurements wasequally strong for all parameters measured (Table 2 and Fig. 4).

DISCUSSION

In this study, we found that the ZAP software was a highlyreproducible tool for quantitative analysis of corneal parame-ters measured on AS-OCT. Mean of differences between mea-surements by the same grader was found to be small forparameters in both the normal and the BK groups. Means ofdifferences between measurements by different observerswere also found to be small in both groups, and limits ofagreement were high for both interobserver and intraobservervariation. Interestingly, P values were marginally significant fortwo parameters measured by the same observer. These param-eters were vertical PCAL in the normal group (P � 0.082) andhorizontal PCC in the BK group (P � 0.065). These sameparameters had high P values for measurements by differentobservers. There does not appear to be a reason for the low Pvalues; moreover, all P values were found to be �0.05. Becausethe reproducibility of AS-OCT has been previously documen-ted,9,17–21 these observations validate the reproducibility andthus the clinical utility of the ZAP software, given that itdemonstrates that differences between users of the softwaretend to be small; thus valid comparisons between analyses bydifferent users can be made. In addition, this study found thatthe ZAP software showed the repeatability of measurements inboth normal and diseased corneas, thus validating its use in aclinical setting to diagnose and follow-up cornea pathology.

The cornea parameters measured in this study included theanterior chamber depth (ACD), posterior cornea curvature(PCC), central cornea thickness (CCT), and posterior cornealarc length (PCAL). The PCAL is a novel cornea parameter thathas not previously been accurately and reproducibly measured,and to our knowledge, the ZAP software is the only imageprocessing software able to do so. Although the other threeparameters (ACD, PCC, and CCT) can be manually measuredon AS-OCT scans using the in-built measurement tool, imageprocessing using the ZAP software uses algorithms to identifythe borders of the cornea endothelium, epithelium, and ante-rior surface of the iris. This avoids potential inaccuracies thatcan arise from manual placement of the AS-OCT measurementtool on the AS-OCT image, which may be done in an arbitraryfashion. In addition the software gives new values not able tobe calculated with current “in-built ” software.

The ability to reproducibly measure cornea parameters hasa myriad of clinical and research applications. In particular, theadvent of lamellar and refractive cornea procedures necessi-tates accurate measurement of the CCT and PCAL. The PCAL isa novel parameter defined as the posterior corneal borderdistance between scleral spurs. This gives a good estimation ofthe limits of the diameter of the corneal endothelium. Itspotential utility is particularly relevant to endothelial kerato-plasty (EK) procedures (such as DSAEK and Descemet’s mem- T

AB

LE2

.In

tero

bse

rver

and

Intr

aob

serv

erV

aria

bili

tyo

fH

ori

zon

tal

Par

amet

ers

of

Co

rnea

sw

ith

Bu

llou

sK

erat

op

ath

yA

nal

yzed

wit

hZ

AP

Soft

war

e

Par

amet

erM

ean

(SD

)O

bse

rver

1M

ean

(SD

)O

bse

rver

2P

(Tw

o-s

ided

)M

ean

Bia

s(9

5%

CI)

Lim

its

of

Agr

eem

ent

Low

erLi

mit

Up

per

Lim

it

Inte

rob

serv

erva

riab

ility

Ho

rizo

nta

lP

CA

L12

.87

(0.6

4)12

.88

(0.6

4)0.

761

�0.

0017

(�0.

129

to0.

095)

�0.

537

(�0.

73to

�0.

34)

0.50

3(0

.309

to0.

697)

Ho

rizo

nta

lC

CT

774.

64(1

30)

773.

06(1

26.6

8)0.

315

1.57

(�0.

755

to3.

91)

�9.

24(�

13.3

to�

5.2)

12.3

9(8

.35

to16

.43)

Ho

rizo

nta

lA

CD

4.19

(1.9

5)3.

162

(0.2

38)

0.74

10.

0004

6(�

0.00

3to

0.00

24)

�0.

014

(�0.

02to

�0.

009)

0.01

3(0

.008

to0.

017)

Ho

rizo

nta

lP

CC

6.86

(0.7

3)6.

527

(0.3

88)

0.31

1�

0.04

3(�

0.12

9to

0.04

3)�

0.44

1(�

0.59

to�

0.29

)0.

355

(0.2

07to

0.50

4)

Par

amet

erFir

stA

nal

ysi

sM

ean

(SD

)Se

con

dA

nal

ysi

sM

ean

(SD

)P

(Tw

o-s

ided

)M

ean

Bia

s(9

5%

CI)

Lim

its

of

Agr

eem

ent

Low

erLi

mit

Up

per

Lim

it

Intr

aob

serv

erva

riab

ility

Ho

rizo

nta

lP

CA

L12

.87

(0.6

4)12

.92

(0.6

5)0.

104

�0.

06(�

0.13

4to

0.01

3)�

0.40

3(�

0.53

to�

0.28

)0.

283

(0.1

548

to0.

410)

Ho

rizo

nta

lC

CT

774.

64(1

30)

775.

01(1

27.8

9)0.

767

�0.

371

(�2.

928

to2.

187)

�12

.24

(�16

.7to

�7.

81)

11.5

0(7

.07

to15

.93)

Ho

rizo

nta

lA

CD

4.19

(1.9

5)4.

15(1

.92)

0.27

50.

037

(�0.

031

to0.

104)

�0.

278

(�0.

395

to�

0.16

)0.

351

(0.2

33to

0.46

8)H

ori

zon

tal

PC

C6.

86(0

.73)

6.92

(0.7

9)0.

065

�0.

059

(�0.

123

to0.

004)

�0.

355

(�0.

47to

�0.

24)

0.23

6(0

.126

to0.

347)

CI,

con

fid

ence

inte

rval

.



brane endothelial keratoplasty), where accurate preoperativemeasurement of the PCAL in a recipient can guide an appro-priate choice of donor graft diameter. Currently, most surgeons“guesstimate ” the appropriate EK size graft from measure-ments of the anterior corneal surface,22 similar to what isconventionally done for penetrating keratoplasty (PK). Theadvantage of an EK procedure is that the surgeon can trans-plant much larger grafts, and thus more endothelium than thatfor conventional PK. In our center, the mean size of the EKgraft is 8.75 mm (range, 7–10 mm), mode 9 mm,23 whereasthat for PK is 8 mm (range, 6.5–11.5 mm), mode 7.5–7.9 mm.23

The additional 1-mm increase in the graft area adds a 26%increase in the amount of endothelial cells transplanted.24

Thus, optimal graft sizing can have a direct impact on theamount of endothelial cells transplanted. The choice for hori-zontal scans was used because the horizontal meridian is larg-est and thus most appropriate for DSAEK cases.

AS-OCT can be used to image patients with EK postopera-tively, and recent further modification of the ZAP software canbe used to calculate the diameter of the implanted endothelialgraft, providing useful clinical parameters for subsequent fol-low-up of these patients. Further work is under way to validatethe DSAEK graft measurements from AS-OCT images using amodification of the ZAP software in a comparison with intra-operative graft trephine diameter.

There are a number of other anterior segment proceduresfor which the ZAP software has potential utility, both in pre-

operative assessment and in postoperative follow-up. For ex-ample, in anterior chamber phakic IOL implantation, measure-ment of the PCAL and ACD allows the surgeon to accuratelyselect an appropriate size implant preoperatively. Postopera-tive measurement of these parameters is useful in determiningif the implant has been placed in an optimal position; subse-quent follow-up measurements can determine whether theimplant shifts or subluxes with time. The other strength of thisstudy is the high limits of agreement between observers andthose within one observer, suggesting the high repeatability ofthe software.

The ZAP software has some limitations. First, image pro-cessing was dependent on the manual identification of thescleral spur as a measurement reference point. The difficulty ofthis practice has been well recognized.2,3,10,16 In a cross-sec-tional study of 502 eyes, Sakata and colleagues24 found thedetectability of the sclera spur on AS-OCT images to be 72%overall, and was less detectable on AS-OCT images obtained inthe superior (64%) and inferior (67%) quadrants. To overcomethis limitation, the ZAP software has an in-built magnificationwindow that can be moved to any position on the scan. Wefound the placement of this window over the scleral spurs tobe highly useful in determining their exact position. In addi-tion, during vertical scan acquisition, the subject’s eyelids weregently held apart to ensure that the lids did not obscure thescleral spurs. This increased the detectability of the scleral spurin our study, and we were able to detect the scleral spur in all

FIGURE 4. Bland Altman plots of corneal measurements of BK corneas. Interobserver agreement, top left: ACD measurements by two graders, topright: CCT measurements by two graders; bottom left: PCC measurements by two graders; bottom right: PCAL measurements by two graders.



AS-OCT images analyzed. This is particularly important if one isplanning on using this software to estimate the graft size sincethe major limitation will be the vertical graft size.

The ZAP software is also limited in that it is unable toprocess high-resolution spectral-domain (SD) OCT images. Thisis due to the limited 6-mm scan width of the SD-OCT, which isunable to image more than one angle concurrently in oneimage. Thus, the ZAP software is able to process only time-domain OCT images, which have a lower resolution but alarger scan width than can encompass both angles in the sameimage.

Finally, the limitation of this study lies in the small samplesize, and that analysis was done in a variety of bullous kera-topathies that were not stratified. In conclusion, the Zhong-shan Assessment Program offers accurate measurement of cor-nea parameters from AS-OCT images with good inter- andintraobserver reproducibility, for scans in both the horizontaland vertical meridians. The program is relatively easy to run,with manual identification of the sclera spur as the only ob-server input required. The data obtained from the scans allowdetailed quantitative analysis of the anterior segment, which isuseful for the preoperative planning of many anterior segmentprocedures.

References

1. Nolan WP, Aung T, Machin D, et al. Detection of narrow anglesand established angle closure in Chinese residents of Singapore:potential screening tests. Am J Ophthalmol. 2006;141:896–901.

2. Radhakrishnan S, Goldsmith J, Huang D, et al. Comparison ofoptical coherence tomography and ultrasound biomicroscopy fordetection of narrow anterior chamber angles. Arch Ophthalmol.2005;123:1052–1059.

3. Radhakrishnan S, Huang D, Smith SD. Optical coherence tomog-raphy imaging of the anterior chamber angle. Ophthalmol ClinNorth Am. 2005;18:vi, 375–381.

4. Huang D, Li Y, Radhakrishnan S. Optical coherence tomography ofthe anterior segment of the eye (Review). Ophthalmol Clin NorthAm. 2004;17:1–6.

5. Tan DT, Janardhanan P, Zhou H, et al. Penetrating keratoplasty inAsian eyes: the Singapore corneal transplant study. Ophthalmol-ogy. 2008;115:975–982.

6. Werner L, Izak AM, Pandey SK, Apple DJ, Trivedi RH, SchmidbauerJM. Correlation between different measurements within the eyerelative to phakic intraocular lens implantation. J Cataract RefractSurg. 2004;30:1982–1988.

7. Yilmaz OF, Bayraktar S, Agca A, Yilmaz B, McDonald MB, van dePol C. Intracorneal inlay for the surgical correction of presbyopia?J Cataract Refract Surg. 2008;34:1921–1927.

8. Console J, Sakata LM, Aung T, Friedman DS, He M. Quantitativeanalysis of anterior segment optical coherence tomography

images: the Zhongshan angle assessment program. Br J Ophthal-mol. 2008;92:1612–1616.

9. Yuen LH, He M, Aung T, Htoon HM, Tan DT, Mehta JS. Biometryof the cornea and anterior chamber in Chinese eyes: an anteriorsegment optical coherence tomography study. Invest OphthalmolVis Sci. 2010;51:3433–3440.

10. Su DH, Friedman DS, See JL, et al. Degree of angle closure andextent of peripheral anterior synechiae: an anterior segment OCTstudy. Br J Ophthalmol. 2008;92:103–107.

11. Lattimore MR, Kaupp S, Schallhorn S, Lewis R. Orbscanpachymetry: implications of a repeated measures and diurnal vari-ation analysis. Ophthalmology. 1999;106:977–981.

12. Harper CL, Boulton ME, Bennett D, et al. Diurnal variations inhuman corneal thickness. Br J Ophthalmol. 1996;80:1068–1072.

13. Feng Y, Varikooty J, Simpson TL. Diurnal variation of corneal andcorneal epithelial thickness measured using optical coherencetomography. Cornea. 2001;20:480–483.

14. Pavlin CJ, Foster FS. Ultrasound Biomicroscopy of the Eye. NewYork: Springer; 1995.

15. Pavlin CJ. Practical applications of ultrasound biomicroscopy. CanJ Ophthalmol. 1995;30:225–229.

16. Pavlin CJ, Harasiewicz K, Foster FS. Ultrasound biomicroscopy ofanterior segment structures in normal and glaucomatous eyes.Am J Ophthalmol. 1992;113:381–389.

17. Muscat S, McKay N, Parks S, Kemp E, Keating D. Repeatability andreproducibility of corneal thickness measurements by optical co-herence tomography. Invest Ophthalmol Vis Sci. 2002;43:1791–1795.

18. Muller M, Dahmen G, Porksen E, et al. Anterior chamber anglemeasurement with optical coherence tomography: intraobserverand interobserver variability. J Cataract Refract Surg. 2006;32:1803–1808.

19. Li H, Leung CK, Cheung CY, et al. Repeatability and reproducibil-ity of anterior chamber angle measurement with anterior segmentoptical coherence tomography. Br J Ophthalmol. 2007;91:1490–1492.

20. Radhakrishnan S, See J, Smith SD, et al. Reproducibility of anteriorchamber angle measurements obtained with anterior segmentoptical coherence tomography. Invest Ophthalmol Vis Sci. 2007;48:3683–3688.

21. Kim DY, Sung KR, Kang SY, et al. Characteristics and reproduc-ibility of anterior chamber angle assessment by anterior-segmentoptical coherence tomography. Acta Ophthalmol. 2011;89:435–441.

22. Tan DT, Mehta JS. Special DSEK techniques for Asian eyes. In:Price MO, Price FW, eds. DSEK: What You Need to Know AboutEndothelial Keratoplasty. Thoroughfare, NJ: SLACK Publishers;2009:97–108.

23. Price MO, Price FW. Endothelial cell loss after Descemet strippingwith endothelial keratoplasty: influencing factors and 2-year trend.Ophthalmology. 2008;115:857–865.

24. Sakata L, Lavanya R, Friedman D, et al. Assessment of the scleralspur in anterior segment optical coherence tomography images.Arch Ophthalmol. 2008;125:181–185.



Reproducibility of Cornea Measurements in Anterior Segment OCT Images of Normal Eyes and Eyes with...

Documents

Transcript of Reproducibility of Cornea Measurements in Anterior Segment OCT Images of Normal Eyes and Eyes with...