rss
Br J Ophthalmol doi:10.1136/bjophthalmol-2012-301713
  • Clinical science

Errors in neuroretinal rim measurement by Cirrus high-definition optical coherence tomography in myopic eyes

  1. Yong Ho Sohn1
  1. 1Department of Ophthalmology, Konyang University, Kim's Eye Hospital, Myung-Gok Eye Research Institute, Seoul, Republic of Korea
  2. 2Department of Ophthalmology, Korea University College of Medicine, Seoul, Republic of Korea
  3. 3Department of Ophthalmology, Konyang University College of Medicine, Daejeon, Republic of Korea
  1. Correspondence to Dr Yong Ho Sohn, Department of Ophthalmology, Konyang University, Kim's Eye Hospital, Myung-Gok Eye Research Institute, #156 Youngdeungpo-dong 4ga, Youngdeungpo-gu, Seoul 150-034, Republic of Korea; yhsohn{at}kimeye.com
  • Accepted 30 July 2012
  • Published Online First 31 August 2012

Abstract

Background/aims To investigate the prevalence of, and factors associated with, errors in neuroretinal rim measurement by Cirrus high-definition (HD) spectral-domain optical coherence tomography (OCT) in myopic eyes.

Methods Neuroretinal rim thicknesses of 255 myopic eyes were measured by Cirrus HD-OCT. The prevalence of, and factors associated with, optic disc margin detection error and cup margin detection error were assessed by analysing 72 cross-sectional optic nerve head (ONH) images obtained at 5° intervals for each eye.

Results Among the 255 eyes, 45 (17.6%) had neuroretinal rim measurement errors; 29 (11.4%) had optic disc margin detection errors at the temporal (16 eyes), superior (11 eyes), and inferior (2 eyes) quadrants; 19 (7.5%) showed cup margin detection errors at the nasal (17 eyes) and temporal (2 eyes) quadrants; and 3 (1.2%) had both disc and cup margin detection errors. Errors in detection of temporal optic disc margin were associated with presence of parapapillary atrophy (PPA), higher myopia, and greater axial length (AL) (p<0.001). Cup margin detection errors were associated with vitreous opacities attached to the ONH surface or acute cup slope angles (p<0.001).

Conclusions Errors in neuroretinal rim measurement by Cirrus HD-OCT were found in myopic eyes, especially in eyes with PPA, higher myopia, greater AL, vitreous opacity or acute cup slope angle. These findings should be considered when interpreting neuroretinal rim thickness measured by Cirrus HD-OCT.

Introduction

Optical coherence tomography (OCT) is widely used for the assessment of the neuroretinal rim, peripapillary retinal nerve fibre layer (RNFL), and macular retinal thickness in eyes with glaucoma.1–7 Neuroretinal rim thickness, as determined by the recently developed spectral-domain OCT has been reported to have excellent reproducibility, glaucoma diagnostic ability and low rate of incorrect optic disc margin detection.4–7 Optic nerve head (ONH) analysis algorithms with Cirrus high-definition (HD)-OCT automatically identify the borders of optic discs and cups, and neuroretinal rim thickness is defined as the distance between the optic disc margin and cup margin.4 Therefore, the presence of errors in the detection of the optic disc or cup margin can cause errors in neuroretinal rim thickness measurement.

In previous studies using Cirrus HD-OCT, the prevalence of optic disc margin detection errors ranged from 0.5%8 to 2.6%,6 and the prevalence of cup margin detection errors was 1.1%.8 In these studies, eyes with optic disc and cup margin detection errors were excluded from analysis, and the factors associated with optic disc and cup margin detection errors were not evaluated. We believe that assessment of factors associated with optic disc and cup margin detection errors may provide valuable clues for the proper interpretation of OCT results and contribute to the improvement of OCT ONH analysis algorithms.

Myopia is an important risk factor for development of glaucoma.9 ,10 In myopic eyes, various optic disc changes, such as optic disc tilt, rotation, nasal elevation, temporal flattening and parapapillary atrophy (PPA) can develop,11–13 and these changes may affect neuroretinal rim measurements obtained using OCT. To date, little is known about errors in neuroretinal rim thickness measured by OCT in myopic eyes.

This study was performed to investigate the prevalence of, and factors associated with, errors in detection of optic disc and cup margin by Cirrus HD-OCT in healthy myopic eyes.

Methods

Participants

The study protocol was approved by the Institutional Review Board of the Armed Forces Capital Hospital, Seongnam, Republic of Korea. All participants provided informed consent before enrolment, and all procedures conformed to the guidelines of the Declaration of Helsinki. Healthy young myopic subjects were recruited from our previous study population.12 Each subject underwent a full ophthalmic examination, including an assessment of visual acuity, refractive error expressed as the spherical equivalent (SE), axial length (AL) measurement by A-scan (HiScan; Optikon, Rome, Italy), ONH evaluation and fundus examination. Eyes with best-corrected visual acuity of 20/20 or better, SE≤0 dioptre (D), normal intraocular pressure (<21 mm Hg), normal anterior segment by slit-lamp biomicroscopic examination, ONH without glaucomatous changes (ie, increased cup-disc ratio, narrowing of neuroretinal rim), and no RNFL defect on red-free fundus photography were enrolled in this study.

OCT measurement

A 200×200 cube optic disc scan was obtained using Cirrus HD-OCT (Carl Zeiss Meditec, Dublin, California, USA) after pupil dilation with instillation of 2.5% phenylephrine hydrochloride eye drops (Mydfrin; Alcon, Fort Worth, Texas, USA) three times at 5 min intervals as described previously.12 For image acquisition, scanning laser images were focused after subjects were seated and properly positioned. Using the iris and fundus viewports, alignment was properly adjusted with the ONH in the centre of the scan. Once the ONH was centred on the live scanning laser image, a 6×6 mm square of data was captured. Images without a prominent involuntary saccade during the scan and a signal strength of >8 were included.

The built-in algorithms of the Cirrus HD-OCT software (V.5.1.0.96) are capable of automatically identifying the optic disc margins at the termination of Bruch's membrane (figure 1). In this process, the rim thickness around the entire circumference of the optic disc is automatically determined by measuring the thickness of the neuroretinal tissue in the optic nerve as it turns to exit through the opening in Bruch's membrane (figure 1).4 The presence of errors in detection of optic disc and cup margin was assessed by analysing 72 cross-sectional ONH images (obtained at 5° intervals from 5° to 360°) from each eye. Error in detection of the optic disc margin was defined as a failed positioning of the optic disc margin at the level of Bruch's membrane (figure 1B,C); error in detection of the cup margin was defined as positioning of the cup margin at the vitreous opacity attached to the ONH surface (figure 1D), or at the point inside the cup (figure 1E,F). Neuroretinal rim measurement error was defined as either optic disc margin detection error or cup margin detection error. The locations of optic disc and cup margin detection errors were classified as originating in the superior, nasal, inferior or temporal quadrants.

Figure 1

Examples of cross-sectional optic nerve head (ONH) images obtained by Cirrus high-definition (HD) optical coherence tomography (OCT) and optic disc photographs in myopic eyes. Cirrus HD-OCT automatically identifies optic disc margins (black circles) and cup margins (red circles); neuroretinal rim thickness is defined as the distance between the optic disc margin and the cup margin. (A) An eye without errors in detection of optic disc and cup margin; (B) An eye with an error in detection of optic disc margin at the superior quadrant (white arrow); (C) an eye with β-zone parapapillary atrophy (PPA) and sloped temporal optic disc margin configuration (yellow arrow) showing an error in detection of optic disc margin at the temporal quadrant (white arrow); (D) an eye with vitreous opacity attached to the ONH surface showing a cup margin detection error (white arrow); (E, F) eyes with acute cup slope angles (cup slope angle (θ),≤90°) and PPAs showing cup margin detection errors (white arrows). An error in detection of temporal disc margin was found in an eye with a stepped temporal optic disc margin configuration (yellow arrow) (E). This figure is only reproduced in colour in the online version.

To investigate the factors associated with optic disc and cup margin detection errors, the presence of β-zone PPA, the presence of vitreous opacity attached to the ONH surface, and the presence of acute cup slope angles (a cup slope angle of <90°, figure 1F) were evaluated. In a previous study, optic disc margin configuration was classified into several types, according to the configuration of scleral bed and retinal pigment epithelium (RPE)-Bruch's membrane complex within the optic disc margin area.14 In the present study, to identify the effect of optic disc margin configuration on optic disc margin detection error, scleral bed and RPE-Bruch's membrane complex configurations were classified as flat (non-oblique configuration, figure 1A), sloped (oblique configuration, figure 1C), or stepped (two flat areas connected by a short, steep, slope (figure 1E)).14

Statistical analyses

The distribution of all variables was examined for normality using the Kolmogorov-Smirnov 1 sample test. The associations between refractive error, AL, the presence of myopic PPA, vitreous opacity attached to the ONH surface, acute cup slope angle, and the presence of optic disc and cup margin detection errors were evaluated by Fisher's exact test. A p value of <0.05 was considered statistically significant. Statistical analyses were performed using SPSS V.12.0 (SPSS, Chicago, Illinois, USA).

Results

Two hundred and fifty-five eyes from 255 subjects were enrolled in this study. The mean±SD age, spherical refractive error, cylindrical refractive error, and AL were 21.03±1.45 years (range, 19–25), −2.74±2.28 D (range, −10.50–0.00), −0.86±0.80 D (range, −3.50–0.00), and 24.96±1.25 mm (range, 22.79–28.13), respectively. Among the 255 eyes, 45 eyes (17.6%) had neuroretinal rim measurement errors; 29 eyes (11.4%), optic disc margin detection errors; 19 eyes (7.5%), cup margin detection errors; and 3 eyes (1.2%), both disc and cup margin detection errors. The number of sections among the 72 sections for each eye where optic disc margin and cup margin detection errors were detected was 14.0±2.9 (range, 7–19) and 6.7±3.8 (range, 3–17), respectively.

Optic disc margin detection errors were found at the temporal (16 eyes), superior (11 eyes), and inferior (2 eyes) quadrants. Temporal optic disc margin detection errors were associated with the presence of PPA, higher myopia and greater AL (p<0.001); all eyes with temporal optic disc margin detection errors had PPA. Seventy-five of the 255 eyes (29.4%) had PPA, among them, 16 eyes (16 / 75, 21.3%) had optic disc margin detection errors. The presence of PPA and sloped or stepped temporal optic disc margin configurations were correlated with higher myopia and greater AL (p<0.001). Errors in detection of temporal optic disc margin were associated with slopped or stepped temporal optic disc margin configurations (p<0.01, table 1). Errors in detection of superior and inferior optic disc margins were not significantly associated with the presence of PPA, degree of myopia or AL (p>0.05).

Table 1

Distribution of eyes with or without errors in detection of temporal optic disc margin according to the temporal optic disc margin configurations

Errors in detection of cup margin were located in the nasal (17 eyes) and temporal (2 eyes) quadrants. Errors in detection of cup margin were associated with vitreous opacities attached to the ONH surface and acute cup slope angles (p<0.001). Among the 17 eyes with errors in detection of nasal cup margin, 8 eyes (47.1%) had vitreous opacities on the nasal side of the ONH surface, and 9 eyes (52.9%) had acute nasal cup slope angles. Among the two eyes with errors in detection of temporal cup margin, one eye had a vitreous opacity on the temporal side of the ONH surface, and the other eye had an acute temporal cup slope angle. Twenty-one of the 255 eyes (8.2%) had vitreous opacities on the surface of the ONH, and among them, 9 eyes (42.9%) showed cup margin detection errors. Acute cup slope angles were found in 10 of 255 eyes (3.9%), and all eyes (100%) with acute cup slope angles had cup margin detection errors.

Discussion

In the present study, errors in measurement of the neuroretinal rim by Cirrus HD-OCT were found in 17.6% of myopic eyes, especially in eyes with PPA, higher myopia, greater AL, vitreous opacity or acute cup slope angle.

In a previous study, Sung et al6 reported that when errors in detection of the optic disc margin were evaluated by horizontal and vertical ONH tomograms of Cirrus HD-OCT, only 2.6% of eyes (21 from a total of 796 eyes) had incorrect determination of the disc margin. In another study, Cheung et al8 reported that the prevalence of errors in optic disc and cup margin detection were 0.5% and 1.1% (3 and 6 from a total of 565 eyes), respectively. In previous studies, the mean SE was −0.74 D6 and −0.93 D,8 whereas, the mean SE of the present study was −3.17 D. In the previous study,6 only horizontal and vertical ONH tomograms were assessed, whereas, 72 cross-sectional images from each eye were inspected in the present study. These differences may contribute to the higher prevalence of errors in detection of the optic disc margin in the current study compared with previous studies.6 ,8

When the locations of errors in optic disc and cup margin detection were classified into four quadrants, the errors were found at the temporal (16 eyes), superior (11 eyes), and inferior (2 eyes) quadrants. Errors in detection of the temporal optic disc margin were associated with presence of PPA, especially in eyes with sloped or stepped temporal optic disc margin configurations. These results imply that structural changes in the temporal optic disc margin caused by myopia may result in errors in detection of the temporal optic disc margin. During the development and progression of myopia, the temporal aspect of the ONH moves back and becomes stretched.11 During this process, mechanical stress on the temporal ONH area may cause PPA and sloped or stepped temporal optic disc margin configurations. Compared with eyes with flat optic disc margin configurations, eyes with sloped or stepped optic disc margin configurations may have more errors in optic disc margin detection. The significant association between refractive error or AL, and the presence of errors in temporal optic disc margin detection may be explained by a greater possibility of the presence of PPA in eyes with higher myopia or greater AL.

It has been hypothesised that at least two different mechanisms (mechanical stretch associated with myopia and atrophic change associated with age-related change or glaucoma) may underlie PPA development.11 In the present study, only young myopic eyes with PPA were enrolled. Given that PPA is not only associated with myopic change, but also correlated with age-related change or glaucoma,15 ,16 further studies may be required to investigate the prevalence of and factors associated with errors in neuroretinal rim measurement by OCT in aged eyes and eyes with glaucoma.

Unlike errors in detection of temporal optic disc margin, errors in detection of superior optic disc margin were not associated with refractive error and AL. The cause of errors in detection of superior optic disc margin is not clear. In the present study, sloped or stepped optic disc margin configurations were not found in the superior and inferior quadrants; therefore, errors in detection of superior optic disc margin may not be associated with sloped or stepped optic disc margin configurations. Given that major retinal vessels are located at the superior and inferior areas of the ONH, vascular shadows may play a role in the development of superior or inferior optic disc margin detection errors.17 An interesting finding of the current study was that the prevalence of errors in optic disc margin detection was less in the inferior quadrant (0.8%) than in the superior quadrant (4.3%). This finding may be associated with differences in optic disc margin structure,18 retinal vascular characteristics (for instance, course, bifurcation, crossover and crowdedness), or OCT scan angle19 between superior and inferior quadrants. To verify these hypotheses, further studies are required.

In eyes with errors in optic disc margin detection, optic disc margins were misidentified as points located higher than the termination levels of the original Bruch's membrane, as shown in figure 1B,C, and E. As a result, errors in detection of optic disc margin can cause neuroretinal rim thinning. For instance, in the case presented in figure 2, although disc photographs showed no glaucomatous changes in the ONH, errors in detection of the superior optic disc margin at the superior quadrant (figure 2B) caused neuroretinal rim thinning at the superior quadrant (figure 2D).

Figure 2

An example of an error in detection of optic disc margin in a myopic eye. (A) Disc photograph revealed no glaucomatous changes; (B) cross-sectional image of the optic nerve head (ONH) demonstrated an error in detection of optic disc margin at the superior quadrant (white arrow); (C) en face ONH image revealed an error in detection of the optic disc margin at the superior quadrant (black arrow); (D) neuroretinal rim thickness curve showed neuroretinal rim thinning at the superior quadrant (black arrow) as a consequence of the error in detection of the optic disc margin. This figure is only reproduced in colour in the online version.

Errors in detection of cup margin were found in the nasal and temporal quadrants, and these errors were associated with vitreous opacities attached to the ONH surface or acute cup slope angles. With the currently available imaging technologies, it is difficult to discriminate between neural and non-neural (connective tissues or vitreous tissues) elements at the ONH.20 Therefore, if a vitreous opacity attached to the ONH surface has similar reflectivity as that of the neuroretinal rim, it can be misidentified as a cup margin. An eye with an acute cup slope angle tends to have abrupt changes in the ONH surface structure; thus, an eye with an acute cup slope angle may have a higher chance of cup margin detection error than an eye with an obtuse cup slope angle that has gradual changes in the ONH surface.

In the present study, only young subjects without ocular disorders, and only images with signal strength of >8 were included. However, in the real world, OCT is used in elderly patients who often have media opacity, and cooperate poorly during the image acquisition. In those eyes, the error rate in neuroretinal rim measurement may be different from that found in this study. Further studies regarding the effect of scan quality on the errors in neuroretinal rim measurement by OCT may be needed.

In conclusion, errors in neuroretinal rim measurement by Cirrus HD-OCT were found in myopic eyes, especially in eyes with PPA, higher myopia, greater AL, vitreous opacity or acute cup slope angle. These findings should be considered when interpreting neuroretinal rim thickness measured by Cirrus HD-OCT.

Footnotes

  • Contributors (1) Study design, acquisition of data, and analysis and interpretation of data: YHH, YYK, YHS; (2) drafting the article or revising it critically for important intellectual content: YHH, YYK, SJ, JHN, HKK, YHS; and (3) final approval of the version to be published: YHH, YYK, YHS.

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval This study was performed with approval of the Institutional Review Board of the Armed Forces Capital Hospital, Republic of Korea. All procedures conformed to the Declaration of Helsinki.

  • Provenance and peer review Not commissioned; externally peer reviewed.

References

Register for free content


Free sample
This recent issue is free to all users to allow everyone the opportunity to see the full scope and typical content of BJO.
View free sample issue >>

Don't forget to sign up for content alerts so you keep up to date with all the articles as they are published.