Current Eye Research, 2015; 40(5): 496–500 ! Informa Healthcare USA, Inc. ISSN: 0271-3683 print / 1460-2202 online DOI: 10.3109/02713683.2014.930155
ORIGINAL ARTICLE
Reproducibility of Angle Metrics Using the Time-Domain Anterior Segment Optical Coherence Tomography: Intra-Observer and Inter-Observer Variability Curr Eye Res Downloaded from informahealthcare.com by Nyu Medical Center on 06/19/15 For personal use only.
Jyotsna Maram1, Xiaojing Pan1,2, Srinivas Sadda1,3, Brian Francis1,3, Kenneth Marion1 and Vikas Chopra1,3 1
Doheny Image Reading Center, Doheny Eye Institute, Los Angeles, CA, USA, 2The Affiliated Hospital of Medical College, Qingdao University, Shandong, P.R. China, and 3Department of Ophthalmology, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA
ABSTRACT Purpose: To evaluate the reproducibility of anterior chamber angle measurements obtained by the Zeiss Visante anterior segment optical coherence tomography (AS-OCT). Methods: Twenty eyes from 20 normal subjects with open anterior chamber angles were studied. The anterior chamber angle was imaged using the Visante AS-OCT. The angle-opening distance (AOD 500, AOD 750), trabercular iris space area (TISA 500, TISA 750) and scleral spur angle (SS angle) at the inferior angle location were measured. All the subjects underwent imaging in a darkened room (1 foot candles measured at the eye). Images were graded in a masked fashion by certified Doheny Image Reading Center graders. For intra-grader reproducibility assessments, images were re-graded by the same grader 1 week later after random sorting of images. For inter-grader assessments, a second masked grader independently reviewed the images. Intraclass correlation coefficients (ICC) were used to assess reproducibility. Results: Inferior angle measurements of AOD (500, 750), TISA (500, 750) and SS angle for 20 normal eyes were calculated. The intra-observer ICC calculations showed excellent reproducibility for all measurements (AOD 500 = 0.95, AOD 750 = 0.97, TISA 500 = 0.93, TISA 750 = 0.94, SS = 0.96; p50.001 for all). The inter-observer ICC calculations showed lower reproducibility for all measurements (AOD 500 = 0.71, p50.001; AOD 750 = 0.82, p50.001; TISA 500 = 0.49, p = 0.08; TISA 750 = 0.61, p = 0.02; SS = 0.75). Conclusion: Determination of anterior chamber angle measurements was possible with the time-domain AS-OCT, but only modest inter-observer reproducibility was found even among experienced graders. Keywords: Anterior chamber angle, anterior-segment OCT, glaucoma, metrics, optical coherence tomography, reproducibility
INTRODUCTION be complicated by variables such as the lighting conditions and the amount of indentation by the physician.1 Recently, anterior segment optical coherence tomography (AS-OCT) has been suggested as a potential high-resolution, non-contact approach for evaluation and quantification of the anterior chamber angle morphology.2–4
Historically, the gold standard for examining and assessing the morphology of the anterior chamber angle has been indentation gonioscopy. This method, however, is subjective, requires significant illumination and requires contact with the ocular surface. As such, precise analysis of the angle geometry may
Received 23 July 2013; revised 19 May 2014; accepted 28 May 2014; published online 20 June 2014 Correspondence: Vikas Chopra, Doheny Image Reading Center, Doheny Eye Institute, MD, 1450 San Pablo Street DEI 3615, Los Angeles, CA 90033, USA. Tel: +323 442 6393. Fax: +323 442 6760. E-mail: [email protected]
496
Curr Eye Res Downloaded from informahealthcare.com by Nyu Medical Center on 06/19/15 For personal use only.
Reproducibility of Angle Metrics on Visante ASOCT Conventional OCT instruments for posterior segment imaging may be adapted for evaluation of the anterior segment and the iridocorneal angle (ICA). In addition, these retinal devices typically provide only short B-scan lengths, precluding visualization of the inferior and superior (or nasal and temporal) angle in the same B-scan. To address these limitations, purpose-built AS-OCT devices have been developed, the Visante AS-OCT (Carl Zeiss Meditec, Dublin, CA) being the most well-established. The Visante OCT uses a 1310-nm superluminscent light-emitting diode as the light source, offering excellent visualization of the angle recess. As with posterior segment OCT devices, investigators have utilized AS-OCT for quantitative analyses. A variety of ICA parameters such as angle-opening distance (AOD), trabecular iris space area (TISA) and the scleral spur angle (SS angle) have been described as potential metrics to define the angle geometry and estimate the region available for filtration. These AS-OCT metrics have been largely based on identification of the scleral spur, since Schwalbe’s line is usually not well-seen on AS-OCT B-scans. Since the exact span of the trabecular meshwork (TM) anterior to the spur in a given patient is not known, investigators have suggested standardized distances of 500 or 750 microns anterior to the spur as proxies for use in measurement of the TISA and AOD. Because of increasing interest in these parameters, the Visante OCT includes caliper and calibration tools to facilitate calculation of angle metrics by the operator. The interest in these metrics has been further enhanced by direct TM surgical procedures (e.g. goniotomy) and angle drainage devices (e.g. iStent), and OCT-derived angle metrics are now being included in glaucoma clinical trials.5,25 For these metrics to be valid clinical tools, however, the reproducibility must be established. Previous studies have shown good repeatability and reproducibility for measurement of AOD, TISA, ARA (angle recess area) and TIA (trabecular-iris angle) with anterior segment OCT.4,6 Li et al.4 reported a coefficient of variation of the angle width for intra-grader reproducibility to be 7% and intergrader reproducibility to be 10% using the Visante OCT. Studies have also reported significant variability in the identification of scleral spur and angle metrics using AS-OCT (between 70% and 78.9%).7,8 Other studies have reported an increase in variability when more than one observer identifies the angle metrics and scleral spur using a custom built software.2,9,10 The purpose of the present study was to investigate whether reading center-trained OCT graders could achieve acceptable inter- and intra-observer reproducibility in ICC measurements using the Visante caliper and calculation tools. !
2015 Informa Healthcare USA, Inc.
497
MATERIALS AND METHODS Recruitment of Subjects Subjects were recruited from the department of ophthalmology at the University Of Southern California Keck School Of Medicine. All the study participants were healthy volunteers with no history of any intraocular surgery. The study was conducted in accordance with the ethical standards stated in the 1964 Declaration of Helsinki and was approved by University of Southern California Institutional Review Board. Written informed consent was obtained from the study subjects. All subjects underwent a complete ophthalmic examination including visual acuity, slit lamp biomicroscopy, intraocular pressure measurements and fundus examination. In a darkened room, gonioscopy was performed using a Zeiss 4-mirror indirect gonioscopy lens by an experienced, fellowship-trained glaucoma specialist (V.C. or B.F.). The inferior angle was graded using the modified Shaffer classification. To be eligible, a participant needed to have an open angle which was defined as having a Shaffer grade equal to or greater than three during dark-room gonioscopy.
OCT Image Acquisition After consent was obtained, all subjects underwent non-mydriatic anterior segment OCT imaging of both eyes using a Visante OCT (Carl Zeiss Meditec, Dublin, CA). Scans were obtained in a darkened room, with lighting standardized to 1 cd/m2 at the imaging plane, and confirmed with a light meter (model no. 840021, Sper Scientific Ltd, Scottsdale, AZ). A single vertical line scan was obtained including both the inferior and superior angles in the same B-scan. Scans were immediately inspected by the operator (X.P. and J.M.) after acquisition, and were repeated if both angles were not fully included in the scan. One eye was randomly selected as the study eye, and only the scans from this eye were included in the analyses.
OCT Grading Procedures Images were exported from the device and graded in a masked fashion by certified Doheny Image Reading Center anterior segment OCT graders, using a prespecified standardized grading protocol. Grading was performed using the semi-automated built-in analysis software on a Visante OCT device within the reading center. The axial resolution of the Visante OCT is 18 mm with a scan width of 16 mm 6 mm. The number of pixels in each scan is 1121 557 pixels. In order to compute angle metrics using the review software, the grader was required to first identify the
Curr Eye Res Downloaded from informahealthcare.com by Nyu Medical Center on 06/19/15 For personal use only.
498 J. Maram et al. scleral spur. Metrics were only computed for the inferior angle, and included the angle-opening distance (AOD 500, AOD 750), trabecular iris space area (TISA 500, TISA 750) and scleral spur angle (SS angle). The AOD 500 was calculated as the distance from the trabecular meshwork to the iris at a position 500 mm anterior to the scleral spur, with the AOD 750 calculated at a position 750 mm anterior to the scleral spur (Figure 1). Radhakrishnan et al.2 have suggested that the TISA may be a more robust method for angle quantification, and the TISA was calculated by also drawing a line from the scleral spur perpendicular to the plane of the inner scleral wall to the opposing iris. The TISA space is bounded superiorly by the inner corneoscleral wall, and inferiorly by the iris surface. In our study, since all images were inspected immediately after acquisition by the operator to make sure they had adequate image quality, no case was deemed to be ungradable at the time of grading. For intra-grader reproducibility assessments, images were re-graded by the same grader 1 week later after random sorting of images to maintain masking. For inter-grader assessments, a second masked grader independently also reviewed the images, and computed the same angle metrics.
FIGURE 1 Semi-automated quantitative determination of the anterior chamber angle parameters assessed by built-in analysis software of anterior-segment Visante OCT. Abbreviations: SS angle, scleral spur angle; AOD, angle opening distance; TISA, trabecular iris space area.
The average of the two measurements was used to compute the inter-observer ICC.
Statistics Statistical analysis was performed with commercial software (SPSS version 11.0; SPSS, Chicago, IL). The absolute mean difference, percentage error and standard deviation between grading’s were calculated. Intraclass correlation coefficients (ICC) were also calculated to evaluate reproducibility.
RESULTS Inferior angle measurements of AOD (500, 750), TISA (500,750) and SS angle for 20 normal eyes are shown in Tables 1 and 2. The mean absolute difference between graders for SS angle was 4.36 ± 4.21 (14.1) [mean ± SD (Max)] ( ) and the mean percentage error between graders was 9.29 ± 9.33 (34.51). The mean absolute difference between the graders (Inter-grader) for AOD 500 was 0.1 ± 0.1 (0.32) (mm) and the mean percentage error was 15.53 ± 15.27 (49.34). Similarly, the mean absolute difference between graders for TISA 500 was 0.05 ± 0.06 (0.24) (mm2) and the mean percentage error was 21.18 ± 22.23 (79.1). The mean absolute difference within the intragrader for SS angle was 2.28 ± 1.98 (6.8) ( ) and the mean percentage error was 5.02 ± 4.5 (14.76). For AOD 500, the mean absolute difference was 0.05 ± 0.04 (0.13) (mm) and the percentage error was 8.35 ± 6.83 (23.65). For TISA 500, the mean absolute difference was 0.03 ± 0.02 (0.07) (mm2) and the percentage error was 11.2 ± 8.53 (29.44). The intra-observer measurements showed excellent reproducibility with ICC’s of 0.95 and 0.97 (AOD 500 and 750), 0.93 and 0.94 (TISA 500 and 750) (p50.001), respectively. The inter-observer measurements showed lower reproducibility with ICC values of 0.71 and 0.82 (AOD 500 and 750) (p50.001), 0.49 and 0.61 (TISA 500 and 750) (p = 0.08, p = 0.02), respectively. The ICC for intra-observer and inter-observer reproducibility of the SS angle was 0.96 and 0.75 (p50.001), respectively.
TABLE 1 Intra-observer reproducibility of angle metrics with the Visante OCT. Inter-observer reproducibility
1st time grading Mean ± SD (Max)
2nd time grading Mean ± SD (Max)
ABS difference Mean ± SD (Max)
% Error Mean ± SD (Max)
SS angle ( ) AOD500 (mm) AOD750 (mm) TISA500 (mm2) TISA750 (mm2)
46.67 ± 7.02 0.55 ± 0.13 0.77 ± 0.18 0.19 ± 0.05 0.36 ± 0.09
46.76 ± 7.81 0.55 ± 0.15 0.76 ± 0.20 0.19 ± 0.05 0.35 ± 0.09
2.28 ± 1.98 0.05 ± 0.04 0.06 ± 0.04 0.03 ± 0.02 0.04 ± 0.03
5.02 ± 4.5 (14.76) 8.35 ± 6.83 (23.65) 6.7 ± 4.41 (16.84) 11.2 ± 8.53 (29.44) 8.98 ± 6.55 (24.51)
(56.4) (0.76) (1.08) (0.28) (0.51)
(57.9) (0.80) (1.20) (0.27) (0.51)
(6.8) (0.13) (0.14) (0.07) (0.09)
Inter-class correlation coefficient ICC
95% CI
0.96 0.95 0.97 0.93 0.94
0.88–0.98 0.87–0.98 0.93–0.98 0.81–0.97 0.86–0.97
SS angle, scleral spur angle; AOD, angle opening distance; TISA, trabecular iris space area; SD, standard deviation. Current Eye Research
Reproducibility of Angle Metrics on Visante ASOCT
499
TABLE 2 Inter-observer reproducibility of angle metrics with the Visante OCT. Inter-observer reproducibility (n = 20 eyes)
Grader 1 Mean ± SD (Max)
Grader 2 Mean ± SD (Max)
ABS Difference Mean ± SD (Max)
% Error Mean ± SD (Max)
SS angle ( ) AOD500 (mm) AOD750 (mm) TISA500 (mm2) TISA750 (mm2)
46.67 ± 7.02 0.55 ± 0.13 0.77 ± 0.18 0.19 ± 0.05 0.36 ± 0.09
46.72 ± 6.8 (63.9) 0.55 ± 0.16 (1.02) 0.77 ± 0.2 (1.26) 0.2 ± 0.08 (0.44) 0.36 ± 0.12 (0.72)
4.36 ± 4.21 (14.1) 0.1 ± 0.1 (0.32) 0.11 ± 0.11 (0.36) 0.05 ± 0.06 (0.24) 0.07 ± 0.08 (0.3)
9.29 ± 9.33 (34.51) 15.53 ± 15.27 (49.34) 12.96 ± 11.6 (40.65) 21.18 ± 22.23 (79.1) 17.77 ± 16.85 (62.31)
(56.4) (0.76) (1.08) (0.28) (0.51)
Inter-class correlation coefficient ICC
95% CI
0.75 0.71 0.82 0.49 0.61
0.38–0.90 0.27–0.89 0.54–0.93 0.30–0.80 0.02–0.85
SS angle, scleral spur angle; AOD, angle opening distance; TISA, trabecular iris space area; SD, standard deviation.
Curr Eye Res Downloaded from informahealthcare.com by Nyu Medical Center on 06/19/15 For personal use only.
DISCUSSION In this study, we were able to demonstrate that established metrics for characterizing anterior chamber morphology could be computed with excellent intra-grader and moderate inter-grader reproducibility using a standardized grading protocol. Although gonioscopy has historically been considered the gold standard for anterior chamber angle assessment, the requirement for contact and the need for an illuminating light are potential confounders for accurate and reproducible angle assessment.1 Ultrasound biomicroscopy has also proven to be a useful method for visualization of the anterior chamber angle and iridocorneoscleral junction, but it also required contact which could potentially artificially widen the angle and alter the results.11,12,13 The reproducibility of this technique has also been thought to be poor due to difficulties in obtaining the scan in the same location and with a consistent alignment.14,15 In contrast, the non-contact and dim infrared illumination of the OCT offers potential advantages for accurate depiction of angle morphology. In addition, the rapid scanning capability of anterior segment OCT opens the possibility for dynamic studies of the angle. With advancements in OCT imaging technologies, TD-OCT has given way to ultrahigh resolution OCT (UHROCT) and spectral domain OCT (SD-OCT) which have further improved on the image quality of TD-OCT. Currently available commercial SD-OCT systems provide a high scanning speed almost 50 times faster data acquisition than the TD-OCT systems. SD-OCT also provides superior image quality and reduced motion artifacts compared with TD-OCT. Furthermore, the SD-OCT system also has a moderate improvement in axial resolution (5 mm).16–18 There are various reports in the literature some have shown good repeatability and reproducibility for measurement of AOD, TISA, ARA (angle recess area) and TIA (trabecular-iris angle) with anterior segment OCT while others have shown poor reproducibility of angle metrics. Muller et al.6, for example, reported low inter- and intra-observer reproducibility in their !
2015 Informa Healthcare USA, Inc.
study of anterior chamber angles assessed from slit lamp OCT images. Other studies have also shown low intra- and inter-observer variability in angle measurements derived from AS-OCT.6,19,20 In our study, however, the intra-grader reproducibility for all parameters was excellent with ICC’s of 0.93–0.97. Inter-grader reproducibility, however, was not as good, with ICC’s of 0.49–0.82. There was a statistical significance difference for AOD (500, 750), TISA (500,750) and SS angle for Intra-grader assessment (p50.001). Inter-grader assessment of the angle also showed statistical significance difference for AOD (500, 750) and SS angle (p50.001), however no statistical significance difference was observed for TISA (500,750) (p = 0.08, p = 0.02). This could possibly be due to reasons such as the iris contour issues that might have affected the graders during the grading procedures and also identification of scleral spur could be difficult, especially in marked grading fashion. However, the fact that each grader individually was able to achieve excellent intra-grader reproducibility opens the possibility that with further training and standardization of imaging and grading procedures, better inter-grader reproducibility could be achieved. In this study the largest discrepancy in measurements between graders revealed that the principal cause of the discrepancy was slight differences in the placement of the position of the scleral spur. Using the current Visante built-in angle tool, the location of scleral spur has to be determined by hand which my introduce subjectivity in the analysis of an AS-OCT images. Studies investigating the visibility of the scleral spur with AS-OCT showed visualization between 70% and 75% which is similar to what we found in our study.21–23 Tian et al. developed an automatic algorithm that could assess the anterior chamber angle by detecting the Schwalbe’s line, AOD, TISA automatically in a very fast and accurate manner. Such new technologies could be employed to reduce the subjective error associated with manual or semi-automatic measurement of angle parameters when more than one observer identifies the scleral spur.24
500 J. Maram et al. In summary, we observed that using a standardized grading protocol, trained reading center graders could achieve excellent intra-grader and moderate inter-grader reproducibility for the computation of angle metrics using Visante anterior segment OCT. With the development of anterior chamber angle instrumentation and drainage devices, reproducible assessment of the anterior chamber angle with OCT is likely to become of increasing importance.
DECLARATION OF INTEREST
Curr Eye Res Downloaded from informahealthcare.com by Nyu Medical Center on 06/19/15 For personal use only.
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the article.
REFERENCES 1. Friedman DS, He M. Anterior chamber angle assessment techniques. Surv Ophthamol 2008;53:250–273. 2. Radhakrishnan R, See J, Smith SD, Nolan WP, Ce Z, Friedman DS, et al. Reproducibility of anterior chamber angle measurements obtained with anterior segment optical coherence tomography. IOVS 2007;48:3683–3688. 3. Leung CK, Chan WM, Ko CY, Chui SI, Woo J, Tsang MK, Tse RK. Visualization of anterior angle dynamics using optical coherence tomography. Ophthalmology 2005;112: 980–984. 4. Li H, Leung CK, Cheung CY, Wong L, Pang CP, Weinreb RN, Lam DS. Repeatability and reproducibility of anterior chamber angle measurement with anterior segment optical coherence tomography. Br J Ophthalmol 2007;91: 1490–1491. 5. Memarzadeh F, Li Y, Chopra V, Varma R, Francis BA, Huang D. Anterior segment optical coherence tomography for imaging the anterior chamber following laser peripheral iridotomy. Am J Ophthalmol 2007;143:877–879. 6. Mu¨ller M, Dahmen G, Po¨rksen E, Geerling G, Laqua H, Ziegler A, Hoerauf H. Anterior chamber angle measurements with optical coherence tomography: intraobserver and interobserver variability. J Cataract Refract Surg 2006; 32:1803–1808. 7. Sakata LM, Lavanya R, Friedman D, Aung HT, Seah SK. Assessment of the scleral spur in anterior segment optical coherence tomography images. Arch Ophthalmol 2008;126: 181–185. 8. Wong HT, Lim MC, Sakata LM, Aung HT, Amerasinghe N, Friedman DS, Aung T. High-definition optical coherence tomography imaging of the irido corneal angle of the eye. Arch Ophthalmol 2009;127:256–260. 9. Console JW, Sakata LM, Aung T, Friedman DS, He M. Quantitative analysis of anterior segment optical coherence tomography image: the Zhongshan Angle Assessment Program. Br J Ophthalmol 2008;92:1612–1616.
10. Leung CK, Cheung CY, Li H, Dorairaj S, Yiu CK, Wong AL, et al. Dynamic analysis of dark-light changes of the anterior chamber angle with anterior segment OCT. Invest Ophthalmol Vis Sci 2007;48:4116–4122. 11. Ishikawa H, Inazumi K, Liebmann JM, Ritch R. Inadvertent corneal indentation can cause artifactitious widening of the iridocorneal angle on ultrasound biomicroscopy. Ophthalmic Surg Lasers 2000;31:342–345. 12. Woo EK, Pavlin CJ, Slomovic A, Taback N, Buys YM. Ultrasound biomicroscopic quantitative analysis of lightdark changes associated with pupillary block. Am J Ophthalmol 1999;127:43–47. 13. Ishikawa H, Esaki K, Liebmann JM, Uji Y, Ritch R. Ultrasound biomicroscopy dark room provocative testing: a quantitative method for estimating anterior segment chamber angle width. Jpn J Ophthalmol 1999;43:526–534. 14. Tello C, Liebmann J, Potash SD, Cohen H, Ritch R. Measurement of ultrasound biomicroscopy images: intraobserver and interobserver reliability. Invest Ophthalmol Vis Sci 1994;35:3549–3552. 15. Urbak SF, Pedersen JK, Thorsen TT. Ultrasound biomicroscopy. II. Intraobserver and interobserver reproducibility of measurements. Acta Ophthalmol Scand 1998;76:546–549. 16. Wojtkowski M, Srinivasan V, Fujimoto JG, Ko T, Schuman JS, Kowalczyk A, Duker JS. Three-dimensional retinal imaging with high-speed ultrahigh-resolution optical coherence tomography. Ophthalmology 2005;112: 1734–1746. 17. Jiao S, Knighton R, Huang X, Gregori G, Puliafito C. Simultaneous acquisition of sectional and fundus ophthalmic images with spectral-domain optical coherence tomography. Opt Express 2005;13:444452. 18. Spaide RF, Koizumi H, Pozzoni MC. Enhanced depth imaging spectral-domain optical coherence tomography. Am J Ophthalmol 2008;146:496–500. 19. Dawczynski J, Koenigsdoerffer E, Augsten R, Strobel J. Anterior optical coherence tomography: a non-contact technique for anterior chamber evaluation. Graefes Arch Clin Exp Ophthalmol 2007;245:423–425. 20. Cumba RJ, Radhakrishnan S, Bell NP, Nagi KS, Chuang AZ, Lin SC, et al. Reproducibility of scleral spur identification and angle measurements using fourier domain anterior segment optical coherence tomography. J Ophthalmol 2012;2012:487309. Epub 2012 Nov 1. 21. Kim DY, Sung KR, Kang SY, Cho JW, Lee KS, Park SB, et al. Characteristics and reproducibility of anterior chamber angle assessment by anterior-segment optical coherence tomography. Acta Ophthalmol 2011;89:435–441. 22. Pekmezci M, Porco TC, Lin SC. Anterior segment optical coherence tomography as a screening tool for the assessment of the anterior segment angle. Ophthalmic Surg Lasers Imaging 2009;40:389–398. 23. Tian J, Marziliano P, Baskaran M, Wong HT, Aung T. Automatic anterior chamber angle assessment for HDOCT images. IEEE Trans Biomed Eng 2011;58:3242–3249. 24. Memarzadeh F, Tang M, Li Y, Chopra V, Francis BA, Huang D. Optical coherence tomography assessment of angle anatomy changes after cataract surgery. Am J Ophthalmol 2007;144:464–465.
Current Eye Research
ORIGINAL ARTICLE
Reproducibility of Angle Metrics Using the Time-Domain Anterior Segment Optical Coherence Tomography: Intra-Observer and Inter-Observer Variability Curr Eye Res Downloaded from informahealthcare.com by Nyu Medical Center on 06/19/15 For personal use only.
Jyotsna Maram1, Xiaojing Pan1,2, Srinivas Sadda1,3, Brian Francis1,3, Kenneth Marion1 and Vikas Chopra1,3 1
Doheny Image Reading Center, Doheny Eye Institute, Los Angeles, CA, USA, 2The Affiliated Hospital of Medical College, Qingdao University, Shandong, P.R. China, and 3Department of Ophthalmology, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA
ABSTRACT Purpose: To evaluate the reproducibility of anterior chamber angle measurements obtained by the Zeiss Visante anterior segment optical coherence tomography (AS-OCT). Methods: Twenty eyes from 20 normal subjects with open anterior chamber angles were studied. The anterior chamber angle was imaged using the Visante AS-OCT. The angle-opening distance (AOD 500, AOD 750), trabercular iris space area (TISA 500, TISA 750) and scleral spur angle (SS angle) at the inferior angle location were measured. All the subjects underwent imaging in a darkened room (1 foot candles measured at the eye). Images were graded in a masked fashion by certified Doheny Image Reading Center graders. For intra-grader reproducibility assessments, images were re-graded by the same grader 1 week later after random sorting of images. For inter-grader assessments, a second masked grader independently reviewed the images. Intraclass correlation coefficients (ICC) were used to assess reproducibility. Results: Inferior angle measurements of AOD (500, 750), TISA (500, 750) and SS angle for 20 normal eyes were calculated. The intra-observer ICC calculations showed excellent reproducibility for all measurements (AOD 500 = 0.95, AOD 750 = 0.97, TISA 500 = 0.93, TISA 750 = 0.94, SS = 0.96; p50.001 for all). The inter-observer ICC calculations showed lower reproducibility for all measurements (AOD 500 = 0.71, p50.001; AOD 750 = 0.82, p50.001; TISA 500 = 0.49, p = 0.08; TISA 750 = 0.61, p = 0.02; SS = 0.75). Conclusion: Determination of anterior chamber angle measurements was possible with the time-domain AS-OCT, but only modest inter-observer reproducibility was found even among experienced graders. Keywords: Anterior chamber angle, anterior-segment OCT, glaucoma, metrics, optical coherence tomography, reproducibility
INTRODUCTION be complicated by variables such as the lighting conditions and the amount of indentation by the physician.1 Recently, anterior segment optical coherence tomography (AS-OCT) has been suggested as a potential high-resolution, non-contact approach for evaluation and quantification of the anterior chamber angle morphology.2–4
Historically, the gold standard for examining and assessing the morphology of the anterior chamber angle has been indentation gonioscopy. This method, however, is subjective, requires significant illumination and requires contact with the ocular surface. As such, precise analysis of the angle geometry may
Received 23 July 2013; revised 19 May 2014; accepted 28 May 2014; published online 20 June 2014 Correspondence: Vikas Chopra, Doheny Image Reading Center, Doheny Eye Institute, MD, 1450 San Pablo Street DEI 3615, Los Angeles, CA 90033, USA. Tel: +323 442 6393. Fax: +323 442 6760. E-mail: [email protected]
496
Curr Eye Res Downloaded from informahealthcare.com by Nyu Medical Center on 06/19/15 For personal use only.
Reproducibility of Angle Metrics on Visante ASOCT Conventional OCT instruments for posterior segment imaging may be adapted for evaluation of the anterior segment and the iridocorneal angle (ICA). In addition, these retinal devices typically provide only short B-scan lengths, precluding visualization of the inferior and superior (or nasal and temporal) angle in the same B-scan. To address these limitations, purpose-built AS-OCT devices have been developed, the Visante AS-OCT (Carl Zeiss Meditec, Dublin, CA) being the most well-established. The Visante OCT uses a 1310-nm superluminscent light-emitting diode as the light source, offering excellent visualization of the angle recess. As with posterior segment OCT devices, investigators have utilized AS-OCT for quantitative analyses. A variety of ICA parameters such as angle-opening distance (AOD), trabecular iris space area (TISA) and the scleral spur angle (SS angle) have been described as potential metrics to define the angle geometry and estimate the region available for filtration. These AS-OCT metrics have been largely based on identification of the scleral spur, since Schwalbe’s line is usually not well-seen on AS-OCT B-scans. Since the exact span of the trabecular meshwork (TM) anterior to the spur in a given patient is not known, investigators have suggested standardized distances of 500 or 750 microns anterior to the spur as proxies for use in measurement of the TISA and AOD. Because of increasing interest in these parameters, the Visante OCT includes caliper and calibration tools to facilitate calculation of angle metrics by the operator. The interest in these metrics has been further enhanced by direct TM surgical procedures (e.g. goniotomy) and angle drainage devices (e.g. iStent), and OCT-derived angle metrics are now being included in glaucoma clinical trials.5,25 For these metrics to be valid clinical tools, however, the reproducibility must be established. Previous studies have shown good repeatability and reproducibility for measurement of AOD, TISA, ARA (angle recess area) and TIA (trabecular-iris angle) with anterior segment OCT.4,6 Li et al.4 reported a coefficient of variation of the angle width for intra-grader reproducibility to be 7% and intergrader reproducibility to be 10% using the Visante OCT. Studies have also reported significant variability in the identification of scleral spur and angle metrics using AS-OCT (between 70% and 78.9%).7,8 Other studies have reported an increase in variability when more than one observer identifies the angle metrics and scleral spur using a custom built software.2,9,10 The purpose of the present study was to investigate whether reading center-trained OCT graders could achieve acceptable inter- and intra-observer reproducibility in ICC measurements using the Visante caliper and calculation tools. !
2015 Informa Healthcare USA, Inc.
497
MATERIALS AND METHODS Recruitment of Subjects Subjects were recruited from the department of ophthalmology at the University Of Southern California Keck School Of Medicine. All the study participants were healthy volunteers with no history of any intraocular surgery. The study was conducted in accordance with the ethical standards stated in the 1964 Declaration of Helsinki and was approved by University of Southern California Institutional Review Board. Written informed consent was obtained from the study subjects. All subjects underwent a complete ophthalmic examination including visual acuity, slit lamp biomicroscopy, intraocular pressure measurements and fundus examination. In a darkened room, gonioscopy was performed using a Zeiss 4-mirror indirect gonioscopy lens by an experienced, fellowship-trained glaucoma specialist (V.C. or B.F.). The inferior angle was graded using the modified Shaffer classification. To be eligible, a participant needed to have an open angle which was defined as having a Shaffer grade equal to or greater than three during dark-room gonioscopy.
OCT Image Acquisition After consent was obtained, all subjects underwent non-mydriatic anterior segment OCT imaging of both eyes using a Visante OCT (Carl Zeiss Meditec, Dublin, CA). Scans were obtained in a darkened room, with lighting standardized to 1 cd/m2 at the imaging plane, and confirmed with a light meter (model no. 840021, Sper Scientific Ltd, Scottsdale, AZ). A single vertical line scan was obtained including both the inferior and superior angles in the same B-scan. Scans were immediately inspected by the operator (X.P. and J.M.) after acquisition, and were repeated if both angles were not fully included in the scan. One eye was randomly selected as the study eye, and only the scans from this eye were included in the analyses.
OCT Grading Procedures Images were exported from the device and graded in a masked fashion by certified Doheny Image Reading Center anterior segment OCT graders, using a prespecified standardized grading protocol. Grading was performed using the semi-automated built-in analysis software on a Visante OCT device within the reading center. The axial resolution of the Visante OCT is 18 mm with a scan width of 16 mm 6 mm. The number of pixels in each scan is 1121 557 pixels. In order to compute angle metrics using the review software, the grader was required to first identify the
Curr Eye Res Downloaded from informahealthcare.com by Nyu Medical Center on 06/19/15 For personal use only.
498 J. Maram et al. scleral spur. Metrics were only computed for the inferior angle, and included the angle-opening distance (AOD 500, AOD 750), trabecular iris space area (TISA 500, TISA 750) and scleral spur angle (SS angle). The AOD 500 was calculated as the distance from the trabecular meshwork to the iris at a position 500 mm anterior to the scleral spur, with the AOD 750 calculated at a position 750 mm anterior to the scleral spur (Figure 1). Radhakrishnan et al.2 have suggested that the TISA may be a more robust method for angle quantification, and the TISA was calculated by also drawing a line from the scleral spur perpendicular to the plane of the inner scleral wall to the opposing iris. The TISA space is bounded superiorly by the inner corneoscleral wall, and inferiorly by the iris surface. In our study, since all images were inspected immediately after acquisition by the operator to make sure they had adequate image quality, no case was deemed to be ungradable at the time of grading. For intra-grader reproducibility assessments, images were re-graded by the same grader 1 week later after random sorting of images to maintain masking. For inter-grader assessments, a second masked grader independently also reviewed the images, and computed the same angle metrics.
FIGURE 1 Semi-automated quantitative determination of the anterior chamber angle parameters assessed by built-in analysis software of anterior-segment Visante OCT. Abbreviations: SS angle, scleral spur angle; AOD, angle opening distance; TISA, trabecular iris space area.
The average of the two measurements was used to compute the inter-observer ICC.
Statistics Statistical analysis was performed with commercial software (SPSS version 11.0; SPSS, Chicago, IL). The absolute mean difference, percentage error and standard deviation between grading’s were calculated. Intraclass correlation coefficients (ICC) were also calculated to evaluate reproducibility.
RESULTS Inferior angle measurements of AOD (500, 750), TISA (500,750) and SS angle for 20 normal eyes are shown in Tables 1 and 2. The mean absolute difference between graders for SS angle was 4.36 ± 4.21 (14.1) [mean ± SD (Max)] ( ) and the mean percentage error between graders was 9.29 ± 9.33 (34.51). The mean absolute difference between the graders (Inter-grader) for AOD 500 was 0.1 ± 0.1 (0.32) (mm) and the mean percentage error was 15.53 ± 15.27 (49.34). Similarly, the mean absolute difference between graders for TISA 500 was 0.05 ± 0.06 (0.24) (mm2) and the mean percentage error was 21.18 ± 22.23 (79.1). The mean absolute difference within the intragrader for SS angle was 2.28 ± 1.98 (6.8) ( ) and the mean percentage error was 5.02 ± 4.5 (14.76). For AOD 500, the mean absolute difference was 0.05 ± 0.04 (0.13) (mm) and the percentage error was 8.35 ± 6.83 (23.65). For TISA 500, the mean absolute difference was 0.03 ± 0.02 (0.07) (mm2) and the percentage error was 11.2 ± 8.53 (29.44). The intra-observer measurements showed excellent reproducibility with ICC’s of 0.95 and 0.97 (AOD 500 and 750), 0.93 and 0.94 (TISA 500 and 750) (p50.001), respectively. The inter-observer measurements showed lower reproducibility with ICC values of 0.71 and 0.82 (AOD 500 and 750) (p50.001), 0.49 and 0.61 (TISA 500 and 750) (p = 0.08, p = 0.02), respectively. The ICC for intra-observer and inter-observer reproducibility of the SS angle was 0.96 and 0.75 (p50.001), respectively.
TABLE 1 Intra-observer reproducibility of angle metrics with the Visante OCT. Inter-observer reproducibility
1st time grading Mean ± SD (Max)
2nd time grading Mean ± SD (Max)
ABS difference Mean ± SD (Max)
% Error Mean ± SD (Max)
SS angle ( ) AOD500 (mm) AOD750 (mm) TISA500 (mm2) TISA750 (mm2)
46.67 ± 7.02 0.55 ± 0.13 0.77 ± 0.18 0.19 ± 0.05 0.36 ± 0.09
46.76 ± 7.81 0.55 ± 0.15 0.76 ± 0.20 0.19 ± 0.05 0.35 ± 0.09
2.28 ± 1.98 0.05 ± 0.04 0.06 ± 0.04 0.03 ± 0.02 0.04 ± 0.03
5.02 ± 4.5 (14.76) 8.35 ± 6.83 (23.65) 6.7 ± 4.41 (16.84) 11.2 ± 8.53 (29.44) 8.98 ± 6.55 (24.51)
(56.4) (0.76) (1.08) (0.28) (0.51)
(57.9) (0.80) (1.20) (0.27) (0.51)
(6.8) (0.13) (0.14) (0.07) (0.09)
Inter-class correlation coefficient ICC
95% CI
0.96 0.95 0.97 0.93 0.94
0.88–0.98 0.87–0.98 0.93–0.98 0.81–0.97 0.86–0.97
SS angle, scleral spur angle; AOD, angle opening distance; TISA, trabecular iris space area; SD, standard deviation. Current Eye Research
Reproducibility of Angle Metrics on Visante ASOCT
499
TABLE 2 Inter-observer reproducibility of angle metrics with the Visante OCT. Inter-observer reproducibility (n = 20 eyes)
Grader 1 Mean ± SD (Max)
Grader 2 Mean ± SD (Max)
ABS Difference Mean ± SD (Max)
% Error Mean ± SD (Max)
SS angle ( ) AOD500 (mm) AOD750 (mm) TISA500 (mm2) TISA750 (mm2)
46.67 ± 7.02 0.55 ± 0.13 0.77 ± 0.18 0.19 ± 0.05 0.36 ± 0.09
46.72 ± 6.8 (63.9) 0.55 ± 0.16 (1.02) 0.77 ± 0.2 (1.26) 0.2 ± 0.08 (0.44) 0.36 ± 0.12 (0.72)
4.36 ± 4.21 (14.1) 0.1 ± 0.1 (0.32) 0.11 ± 0.11 (0.36) 0.05 ± 0.06 (0.24) 0.07 ± 0.08 (0.3)
9.29 ± 9.33 (34.51) 15.53 ± 15.27 (49.34) 12.96 ± 11.6 (40.65) 21.18 ± 22.23 (79.1) 17.77 ± 16.85 (62.31)
(56.4) (0.76) (1.08) (0.28) (0.51)
Inter-class correlation coefficient ICC
95% CI
0.75 0.71 0.82 0.49 0.61
0.38–0.90 0.27–0.89 0.54–0.93 0.30–0.80 0.02–0.85
SS angle, scleral spur angle; AOD, angle opening distance; TISA, trabecular iris space area; SD, standard deviation.
Curr Eye Res Downloaded from informahealthcare.com by Nyu Medical Center on 06/19/15 For personal use only.
DISCUSSION In this study, we were able to demonstrate that established metrics for characterizing anterior chamber morphology could be computed with excellent intra-grader and moderate inter-grader reproducibility using a standardized grading protocol. Although gonioscopy has historically been considered the gold standard for anterior chamber angle assessment, the requirement for contact and the need for an illuminating light are potential confounders for accurate and reproducible angle assessment.1 Ultrasound biomicroscopy has also proven to be a useful method for visualization of the anterior chamber angle and iridocorneoscleral junction, but it also required contact which could potentially artificially widen the angle and alter the results.11,12,13 The reproducibility of this technique has also been thought to be poor due to difficulties in obtaining the scan in the same location and with a consistent alignment.14,15 In contrast, the non-contact and dim infrared illumination of the OCT offers potential advantages for accurate depiction of angle morphology. In addition, the rapid scanning capability of anterior segment OCT opens the possibility for dynamic studies of the angle. With advancements in OCT imaging technologies, TD-OCT has given way to ultrahigh resolution OCT (UHROCT) and spectral domain OCT (SD-OCT) which have further improved on the image quality of TD-OCT. Currently available commercial SD-OCT systems provide a high scanning speed almost 50 times faster data acquisition than the TD-OCT systems. SD-OCT also provides superior image quality and reduced motion artifacts compared with TD-OCT. Furthermore, the SD-OCT system also has a moderate improvement in axial resolution (5 mm).16–18 There are various reports in the literature some have shown good repeatability and reproducibility for measurement of AOD, TISA, ARA (angle recess area) and TIA (trabecular-iris angle) with anterior segment OCT while others have shown poor reproducibility of angle metrics. Muller et al.6, for example, reported low inter- and intra-observer reproducibility in their !
2015 Informa Healthcare USA, Inc.
study of anterior chamber angles assessed from slit lamp OCT images. Other studies have also shown low intra- and inter-observer variability in angle measurements derived from AS-OCT.6,19,20 In our study, however, the intra-grader reproducibility for all parameters was excellent with ICC’s of 0.93–0.97. Inter-grader reproducibility, however, was not as good, with ICC’s of 0.49–0.82. There was a statistical significance difference for AOD (500, 750), TISA (500,750) and SS angle for Intra-grader assessment (p50.001). Inter-grader assessment of the angle also showed statistical significance difference for AOD (500, 750) and SS angle (p50.001), however no statistical significance difference was observed for TISA (500,750) (p = 0.08, p = 0.02). This could possibly be due to reasons such as the iris contour issues that might have affected the graders during the grading procedures and also identification of scleral spur could be difficult, especially in marked grading fashion. However, the fact that each grader individually was able to achieve excellent intra-grader reproducibility opens the possibility that with further training and standardization of imaging and grading procedures, better inter-grader reproducibility could be achieved. In this study the largest discrepancy in measurements between graders revealed that the principal cause of the discrepancy was slight differences in the placement of the position of the scleral spur. Using the current Visante built-in angle tool, the location of scleral spur has to be determined by hand which my introduce subjectivity in the analysis of an AS-OCT images. Studies investigating the visibility of the scleral spur with AS-OCT showed visualization between 70% and 75% which is similar to what we found in our study.21–23 Tian et al. developed an automatic algorithm that could assess the anterior chamber angle by detecting the Schwalbe’s line, AOD, TISA automatically in a very fast and accurate manner. Such new technologies could be employed to reduce the subjective error associated with manual or semi-automatic measurement of angle parameters when more than one observer identifies the scleral spur.24
500 J. Maram et al. In summary, we observed that using a standardized grading protocol, trained reading center graders could achieve excellent intra-grader and moderate inter-grader reproducibility for the computation of angle metrics using Visante anterior segment OCT. With the development of anterior chamber angle instrumentation and drainage devices, reproducible assessment of the anterior chamber angle with OCT is likely to become of increasing importance.
DECLARATION OF INTEREST
Curr Eye Res Downloaded from informahealthcare.com by Nyu Medical Center on 06/19/15 For personal use only.
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the article.
REFERENCES 1. Friedman DS, He M. Anterior chamber angle assessment techniques. Surv Ophthamol 2008;53:250–273. 2. Radhakrishnan R, See J, Smith SD, Nolan WP, Ce Z, Friedman DS, et al. Reproducibility of anterior chamber angle measurements obtained with anterior segment optical coherence tomography. IOVS 2007;48:3683–3688. 3. Leung CK, Chan WM, Ko CY, Chui SI, Woo J, Tsang MK, Tse RK. Visualization of anterior angle dynamics using optical coherence tomography. Ophthalmology 2005;112: 980–984. 4. Li H, Leung CK, Cheung CY, Wong L, Pang CP, Weinreb RN, Lam DS. Repeatability and reproducibility of anterior chamber angle measurement with anterior segment optical coherence tomography. Br J Ophthalmol 2007;91: 1490–1491. 5. Memarzadeh F, Li Y, Chopra V, Varma R, Francis BA, Huang D. Anterior segment optical coherence tomography for imaging the anterior chamber following laser peripheral iridotomy. Am J Ophthalmol 2007;143:877–879. 6. Mu¨ller M, Dahmen G, Po¨rksen E, Geerling G, Laqua H, Ziegler A, Hoerauf H. Anterior chamber angle measurements with optical coherence tomography: intraobserver and interobserver variability. J Cataract Refract Surg 2006; 32:1803–1808. 7. Sakata LM, Lavanya R, Friedman D, Aung HT, Seah SK. Assessment of the scleral spur in anterior segment optical coherence tomography images. Arch Ophthalmol 2008;126: 181–185. 8. Wong HT, Lim MC, Sakata LM, Aung HT, Amerasinghe N, Friedman DS, Aung T. High-definition optical coherence tomography imaging of the irido corneal angle of the eye. Arch Ophthalmol 2009;127:256–260. 9. Console JW, Sakata LM, Aung T, Friedman DS, He M. Quantitative analysis of anterior segment optical coherence tomography image: the Zhongshan Angle Assessment Program. Br J Ophthalmol 2008;92:1612–1616.
10. Leung CK, Cheung CY, Li H, Dorairaj S, Yiu CK, Wong AL, et al. Dynamic analysis of dark-light changes of the anterior chamber angle with anterior segment OCT. Invest Ophthalmol Vis Sci 2007;48:4116–4122. 11. Ishikawa H, Inazumi K, Liebmann JM, Ritch R. Inadvertent corneal indentation can cause artifactitious widening of the iridocorneal angle on ultrasound biomicroscopy. Ophthalmic Surg Lasers 2000;31:342–345. 12. Woo EK, Pavlin CJ, Slomovic A, Taback N, Buys YM. Ultrasound biomicroscopic quantitative analysis of lightdark changes associated with pupillary block. Am J Ophthalmol 1999;127:43–47. 13. Ishikawa H, Esaki K, Liebmann JM, Uji Y, Ritch R. Ultrasound biomicroscopy dark room provocative testing: a quantitative method for estimating anterior segment chamber angle width. Jpn J Ophthalmol 1999;43:526–534. 14. Tello C, Liebmann J, Potash SD, Cohen H, Ritch R. Measurement of ultrasound biomicroscopy images: intraobserver and interobserver reliability. Invest Ophthalmol Vis Sci 1994;35:3549–3552. 15. Urbak SF, Pedersen JK, Thorsen TT. Ultrasound biomicroscopy. II. Intraobserver and interobserver reproducibility of measurements. Acta Ophthalmol Scand 1998;76:546–549. 16. Wojtkowski M, Srinivasan V, Fujimoto JG, Ko T, Schuman JS, Kowalczyk A, Duker JS. Three-dimensional retinal imaging with high-speed ultrahigh-resolution optical coherence tomography. Ophthalmology 2005;112: 1734–1746. 17. Jiao S, Knighton R, Huang X, Gregori G, Puliafito C. Simultaneous acquisition of sectional and fundus ophthalmic images with spectral-domain optical coherence tomography. Opt Express 2005;13:444452. 18. Spaide RF, Koizumi H, Pozzoni MC. Enhanced depth imaging spectral-domain optical coherence tomography. Am J Ophthalmol 2008;146:496–500. 19. Dawczynski J, Koenigsdoerffer E, Augsten R, Strobel J. Anterior optical coherence tomography: a non-contact technique for anterior chamber evaluation. Graefes Arch Clin Exp Ophthalmol 2007;245:423–425. 20. Cumba RJ, Radhakrishnan S, Bell NP, Nagi KS, Chuang AZ, Lin SC, et al. Reproducibility of scleral spur identification and angle measurements using fourier domain anterior segment optical coherence tomography. J Ophthalmol 2012;2012:487309. Epub 2012 Nov 1. 21. Kim DY, Sung KR, Kang SY, Cho JW, Lee KS, Park SB, et al. Characteristics and reproducibility of anterior chamber angle assessment by anterior-segment optical coherence tomography. Acta Ophthalmol 2011;89:435–441. 22. Pekmezci M, Porco TC, Lin SC. Anterior segment optical coherence tomography as a screening tool for the assessment of the anterior segment angle. Ophthalmic Surg Lasers Imaging 2009;40:389–398. 23. Tian J, Marziliano P, Baskaran M, Wong HT, Aung T. Automatic anterior chamber angle assessment for HDOCT images. IEEE Trans Biomed Eng 2011;58:3242–3249. 24. Memarzadeh F, Tang M, Li Y, Chopra V, Francis BA, Huang D. Optical coherence tomography assessment of angle anatomy changes after cataract surgery. Am J Ophthalmol 2007;144:464–465.
Current Eye Research
