Download Repeatability and Reproducibility of Central Corneal Thickness

1040-5488/05/8210-0892/0 VOL. 82, NO. 10, PP. 892–899
OPTOMETRY AND VISION SCIENCE
Copyright © 2005 American Academy of Optometry
ORIGINAL ARTICLE
Repeatability and Reproducibility of Central
Corneal Thickness Measurement With
Pentacam, Orbscan, and Ultrasound
BIRGIT LACKNER, MD, GERALD SCHMIDINGER, MD, STEFAN PIEH, MD,
MARTIN A. FUNOVICS, MD and CHRISTIAN SKORPIK, MD
Departments of Ophthalmology (BL, GS, SP, CS) and Radiology (MAF), Medical University of Vienna, Vienna, Austria
ABSTRACT: Purpose. The purpose of this study was to compare central corneal thickness (CCT) measurements obtained
with a novel rotating Scheimpflug camera (Pentacam; Oculus) with scanning slit topography (Orbscan; Bausch &
Lomb), and with ultrasound pachymetry (SP-2000; Tomey). Methods. CCT in 30 healthy eyes was measured twice with
each modality by 2 independent observers in random order. The results from scanning slit topography are given both
with and without multiplication with the ⴖacoustic correction factorⴖ of 0.92. In addition, the displayed images from
the rotating Scheimpflug camera and scanning slit topography were used to calculate the signal difference-to-noise
ratios (SD/N) between cornea and background signal. Results. The mean CCT values as determined with the different
modalities (ⴞ standard deviation) were: 542 ⴞ 29 ␮m, 576 ⴞ 37 ␮m, 530 ⴞ 34 ␮m, and 552 ⴞ 32 ␮m for rotating
Scheimpflug imaging, for uncorrected and for corrected scanning slit pachymetry, and for ultrasound, respectively. The
differences between modalities (ⴞ 95% limits of agreement) were –9.8 ⴞ 31 ␮m between rotating Scheimpflug and
ultrasound, 24 ⴞ 31.2 ␮m between scanning slit and ultrasound, and 33 ⴞ 27 ␮m between scanning slit and rotating
Scheimpflug imaging. The limits of agreement for within and between observer effects were within 4.2% of the absolute
CCT values for scanning slit and ultrasound and within 2.2% for the rotating Scheimpflug imaging. The rotating
Scheimpflug camera showed similar SD/N ratios but steeper edges of the corneal surfaces in the intensity profile plots.
Conclusion. In the assessment of normal corneas, the Pentacam measured CCT values closer to ultrasound pachymetry
and with less variability compared with Orbscan. The (interobserver) reproducibility with the Pentacam was highest of
all 3 modalities. (Optom Vis Sci 2005;82:892–899)
Key Words: central corneal thickness, method comparison, optometry, Scheimpflug imaging, ultrasound pachymetry
H
igh accuracy in the assessment of central corneal thickness (CCT) is primarily desirable in 3 circumstances:
First, it is an integral part in planning keratorefractive
procedures to avoid complications such as corneal ectasia. Second,
CCT can be regarded as a correlate of the physiological condition
of the corneal endothelium and used in the diagnosis of certain
corneal diseases such as keratoconus and Fuchs dystrophy.1 Finally, the connection between decreased CCT and apparently decreased intraocular pressure readings has stimulated recent debate
about the role of CCT in the early diagnosis of glaucoma.2– 6
In the course of the development of a variety of modalities to
measure CCT, the mean thickness reported by various authors,
even in normal eyes, has not been consistent.7,8 Although handheld ultrasound-based systems offer the advantages of portability
and relative ease of use, they experience relatively higher interob-
server variability, possibly as a result of difficulties in centration
and alignment.9 –11 The need for topical anesthesia and contact of
the probe with the cornea has led to the search for noninvasive
alternative solutions without the risk of epithelial lesions or transmission of infection. Optical methods such as optical coherence
tomography12–14 or scanning slit topography systems were designed to acquire morphologic information about several structures in the anterior chamber simultaneously,15 often used in the
preoperative assessment in refractive surgery. However, several investigations showed a consistent overestimation of corneal thickness with such systems compared with ultrasound measurements,
which led to the introduction of an acoustic correction factor to
numerically reduce the measured values on such devices.16
A novel apparatus capable of modeling the anterior chamber is
the Pentacam (Oculus, Wetzlar, Germany). The system is based on
Optometry and Vision Science, Vol. 82, No. 10, October 2005
Repeatability and Reproducibility of Central Corneal Thickness Measurement—Lackner et al.
893
a rotating Scheimpflug camera, which scans and measures the
complete cornea and anterior chamber in approximately 2 seconds.
Its reproducibility and accuracy in assessing CCT compared with
traditional modalities have as yet not been established.
The purpose of the present study was to compare CCT measurements obtained with the rotating Scheimpflug camera with
scanning slit topography and with ultrasound pachymetry to provide estimates for the error between measurements with the different modalities and of the reproducibility of measurements within
and between observers.
METHODS
Thirty healthy volunteers (16 women and 14 men with a mean
age of 31.5 years [standard deviation 3.8 years]) were recruited for
the study. The study was performed in compliance with institutional and legal requirements. Informed consent was obtained in
writing from all participants. All subjects had normal eyes without
corneal abnormalities as verified by slit lamp examination. Soft
contact lens wearers were included but were required not to wear
contact lenses within 24 hours before the investigations.15,17 The
refractive error was measured with an autorefractometer (model
597; Humphrey Systems, Dublin, CA). The mean spherical equivalent was –1.44 ⫾ 1.95 D. All measurements were taken on one
randomly chosen eye per subject and at least 3 hours after awakening.6
Central corneal thickness was determined with 3 different modalities: (1) with a novel rotating Scheimpflug camera (Pentacam
70,700; Oculus, Wetzlar, Germany), (2) with scanning slit topography (Orbscan, version 4.00; Bausch & Lomb, Rochester, NY),
and (3) with ultrasound pachymetry (Pachymeter SP-2000;
Tomey, Erlangen, Germany). Each modality was carried out by 2
independent observers (BL, GS) blinded to all previous measurements of the study, who acquired 2 measurements on each eye with
each of the modalities, respectively. The order of the examiners and
of the modalities was randomly assigned to each eye; however,
ultrasound pachymetry measurements were always scheduled last
to avoid influence on the other modalities as a result of corneal
flattening.
Rotating Scheimpflug Imaging
Rotating Scheimpflug imaging was performed with the patient
seated using a chinrest and forehead strap. The patient was asked to
keep both eyes open and to fixate on a blinking fixation target. The
system uses a rotating Scheimpflug camera and a monochromatic
slit light source (blue LED at 475 nm) that rotate together around
the optical axis of the eye. Within 2 seconds, the system rotates
180° and acquires 25 images that contain 500 measurement points
on the front and back corneal surfaces to draw a true elevation map.
The software acquires the images as volume data, thus multiplanar
reprojections allow the creation of axial and tangential maps. The
thinnest value of corneal thickness was recorded as CCT.
FIGURE 1.
Representative central corneal images obtained with the Orbscan (A) and
the Pentacam (B). Differences in edge detection of the anterior and
posterior corneal surfaces are quantified in density profile plots (C and D)
through the central cornea (see white lines in A and B) obtained from the
mean of normalized density profiles on individual images. Density profiles
obtained with the Pentacam (D) consistently show steeper edges, especially on the anterior corneal surface compared with Orbscan (C).
Scanning Slit Topography
Scanning slit topography measured CCT by a noncontact
pachymeter (Orbscan; Bausch & Lomb) operating under software
version 4.00. Subjects were seated in typical position using the
chinrest; the instrument was aligned and scanned the cornea. The
Optometry and Vision Science, Vol. 82, No. 10, October 2005
894
Repeatability and Reproducibility of Central Corneal Thickness Measurement—Lackner et al.
TABLE 1.
Means and standard deviations (in ␮m) of central corneal thickness measurements obtained with 3 different modalities
in 30 normal eyes a
Mean
Standard deviation
Rotating
Scheimpflug imaging
Scanning slit topography
uncorrected data
(corrected data)
Ultrasound
pachymetry
542.0
29.3
576.0 (530.0)
36.9 (33.9)
552.0
31.7
a
The data from the scanning slit topography system are represented both with and without multiplication with the “acoustic
correction factor” of 0.92 as suggested by the manufacturer.
system software automatically detects the corneal anterior and posterior surface on the acquired images, compensates for differences
in refraction from the corneal anterior surface using a ray trace
algorithm, and calculates the corneal diameter over the whole surface. From these measurements, the thinnest value is automatically
selected as CCT. Thickness measurements were recorded both
with and without multiplication with an “acoustic factor” equal to
0.92, which had been proposed by the manufacturer and used in
previous publications to compensate for the known offset of the
Orbscan measurements compared with ultrasound measurements.
In both optical modalities, patient eye movement was constantly monitored and recorded by the respective system, and only
measurements with ⬍0.6 mm decentration between the optical
axes of the eye and the instrument were included. Furthermore, it
was ascertained that no eye in the series showed a corneal apex
farther than 0.6 mm away from the optical axis of the eye.
Ultrasound Pachymetry
CCT was measured using a handheld ultrasound pachymeter
(SP-2000; Tomey, Nagoya, Japan) calibrated by the manufacturer.
The instrument used an ultrasound velocity (acoustic index) of
1640 m/second. The cornea was anesthetized with topical 1%
oxybuprocaine. The patient was brought into a faceup position on
the examination chair and asked to fixate a target on the ceiling.
The pachymeter probe was brought in light contact with the cornea centrally and perpendicularly. CCT was recorded as the minimum of 3 individual acquisitions. This process was repeated for
each individual CCT measurement.
Image Analysis
In a subset of eyes, the displayed images from the scanning slit
topography system and the rotating Scheimpflug camera (Fig. 1A,
B) were analyzed on a standard personal computer in an 8-bit
noncompressed format. Using public domain software (ImageJ
1.32; National Institutes of Health, Bethesda, MD), a line was
drawn perpendicular to the corneal vertex, and intensity profiles
along the line were recorded. To create a mean intensity profile
throughout the cornea, the individual intensity profiles obtained
with either modality were then numerically overlaid with their
peaks aligned, and the mean intensity profiles for each modality
was plotted (Fig. 1C, D). Furthermore, circular regions of interest
(ROIs) measuring 500 ␮m in diameter were drawn in the cornea
and in the adjacent background. For each ROI, mean pixel intensity and standard deviation was calculated by the software. Signal
difference-to-noise ratios were determined by dividing the difference between cornea and background intensity (signal) by the standard deviation of the background intensity (noise).
Statistical Analysis
Central corneal thickness was measured (1) with different measurements, (2) by different observers, and (3) with different modalities. Thus, the variance of CCT consists of (1) the variance of
the CCT itself (i.e., the difference between the mean of all measurements and the mean of all measurements of that individual
eye), (2) the variance of the extra error in CCT resulting from the
observer, and (3) the variance of the error in individual measurements in a single observer.18
Variances between the measurements and between the observers
were examined using 2-factor analysis of variance (ANOVA) with
repeated measurements (RM). One factor was the number of the
measurement, the other factor the observer. Variances resulting
from the different modalities were calculated using one-factor
ANOVA with repeated measurements; the factor was the modality. For the between-modalities comparison, the mean value from
all measurements of both observers in the respective modality was
used.
To visualize the correlation between the difference between
methods and the absolute values of CCT, data were presented
graphically with Bland-Altman plots, in which for each pair of
measurements obtained with 2 modalities, the difference of the
values is plotted over the mean of the values. To compensate for
correlation between CCT difference and average, the plots were
logarithmically transformed so that CCT ratios were plotted over
average CCT.19 Limits of agreement between modalities were calculated as mean difference of the log CCT values ⫾ 1.96 standard
deviations and backconversion using the exponential function.20
The signal difference-to-noise measurements from the digitized
images of the scanning slit topography system and the rotating
Scheimpflug camera were compared with the paired t test at a level
of significance of 0.05 (all calculations: SPSS 10.0; SPSS Inc.,
Chicago, IL).
RESULTS
Device Effects
The mean values of CCT measurements and their standard
deviations are shown in Table 1. The differences between the modalities and their confidence intervals are shown in Table 2. Uncorrected scanning slit topography measured the largest mean
Optometry and Vision Science, Vol. 82, No. 10, October 2005
Repeatability and Reproducibility of Central Corneal Thickness Measurement—Lackner et al.
895
TABLE 2.
Mean differences between the modalities used to determine central corneal thickness, their standard deviations,
confidence intervals, and limits of agreement (all values in ␮m) a
Scanning slit–US pachy
Rotating Scheimpflug–US pachy
Scanning slit–rotating Scheimpflug
Corrected scanning slit–US pachy
Corrected scanning slit–rotating Scheimpflug
Mean
difference
Standard
deviation
95% confidence
interval
95% limits
of agreement
99% limits
of agreement
23.5
⫺9.8
28.0,
⫺22.6
⫺12.8
15.9
8.1
13.6
14.6
11.7
18.0, 29.0
⫺13.0, ⫺6.9
28.0, 38.0
⫺28.0, ⫺17.0
⫺17.0, ⫺8.6
⫺7.8, 55.0
⫺26.0, 6.1
6.6, 60.0
⫺51.0, 6.1
⫺36.0, 10.2
⫺18.0, 65.0
⫺31.0, 11.0
⫺1.8, 68.0
⫺60.0, 15.0
⫺43.0, 17.5
a
Note that the 95% confidence interval estimates in which the mean difference in a similar collective will be, whereas the limit of
agreement estimates in which a single observation can be expected. US, ultrasound.
CCT, while rotating Scheimpflug imaging and ultrasound measured 6% (34 ␮m) and 4.3% (24 ␮m) shorter, respectively. The
standard deviations of the respective modalities were similar. The
scanning slit topography data are given as uncorrected data and
after multiplication with the ⬙acoustic correction factor⬙ of 0.92.
The application of the “acoustic correction factor” to scanning slit
topography data led to more conservative (thinner) measurements for
CCT than with all other modalities (22 ␮m thinner than the values
acquired with ultrasound pachymetry). The correlation between scanning slit topography and ultrasound decreased with the application of
the correction factor. The signal difference-to-noise ratios as a measure
for overall image quality did not differ significantly (mean Orbscan:
31.1, mean Pentacam: 29.7, 95% confidence interval [CI] of the
difference: ⫺9.3–12.1; p ⫽ 0.738, paired t test). In contrast, the
intensity profiles through the corneal vertex obtained from the Pentacam images showed steeper edges, especially on the anterior corneal
surface, than the Orbscan images. Rotating Scheimpflug imaging
compared with ultrasound (Fig. 3D) showed both the smallest mean
difference as well as the narrowest 95% limits of agreement. The
Bland-Altman plots in the method comparisons (Fig. 3) showed a
significant correlation between average CCT and CCT difference and
have thus been transformed to logarithmic scale to provide constant
95% limits of agreement. In the observer comparison plots, no such
correlation was observed.
ities. No significant influences of either number of measurement or
observer were found on the CCT measurements. Although the 3
modalities showed similar intraobserver variabilities, rotating
Scheimpflug imaging showed the lowest interobserver variability,
and scanning slit topography and ultrasound considerably higher
values (Fig. 2). However, the 95% confidence interval of all differences did not exceed 8.5 ␮m for either modality. This equals an
observer-related reproducibility of CCT measurements within
1.6%.
DISCUSSION
The assessment of CCT with different modalities shows considerable differences between modalities. In the present collective of
60 eyes of healthy European volunteers, CCT determined with
scanning slit topography (Orbscan) was on average 20 ␮m (3.6%)
thicker than with ultrasound pachymetry, and CCT determined
with rotating Scheimpflug imaging (Pentacam) was on average 13
␮m (2.4%) thinner than with ultrasound pachymetry.
Regarding differences between modalities in CCT measurements, 2 most relevant questions need to be addressed: (1) how
reliably can we predict the differences between CCT measurements obtained with the respective modalities; and (2) what
amount of uncertainty in CCT measurement will be clinically
acceptable?
Variability In and Between Observers
Table 3 shows the mean difference and the variability, expressed
as the standard deviation of the difference and the 95% CI of the
differences within and between observers in the respective modal-
Pentacam
The Pentacam yielded the best between observer reproducibility
of all modalities (Fig. 2), and CCT values that were closer to
TABLE 3.
Mean difference, standard deviation of the difference, and 95% limits of agreement between measurements and between
observers (in ␮m) within the 3 modalities of central corneal thickness measurement
Difference between measurements
SD of difference between measurements
95% limits of agreement between measurements
Difference between observers
SD of difference between observers
95% limits of agreement between observers
Rotating Scheimpflug
imaging
Scanning slit
topography
Ultrasound
pachymetry
1.0
6.1
⫺11.0, 13.0
0.1
3.0
⫺6.0, 5.8
1.2
6.9
⫺12.0, 14.0
1.4
11.9
122.0, 25.0
0.3
3.6
⫺6.7, 7.3
2.8
9.9
⫺17.0, 22.0
SD, standard deviation.
Optometry and Vision Science, Vol. 82, No. 10, October 2005
896 Repeatability and Reproducibility of Central Corneal Thickness Measurement—Lackner et al.
ultrasound and differences showed less variability than those obtained with (corrected or uncorrected) Orbscan (Fig. 3). A potential explanation may be the different sampling modes of Orbscan
and Pentacam. The Pentacam acquires images during a 180° rotation around the optical axis so that each section runs through the
corneal vertex, whereas the Orbscan performs a translational
movement from side to side of the cornea, covering the vertex in
only a few sections.
In the assessment of image quality, the Pentacam and Orbscan
images showed similar signal difference-to-noise ratios of cornea
versus background. However, the intensity profiles showed steeper
corneal edge depiction by the Pentacam, thereby potentially contributing to a decrease in detection errors as a result of less blurred
images. This may further explain the fact that in the present study,
the Pentacam tended to give the smallest CCT values of the 3
modalities. However, the exact image processing algorithms both
within Orbscan and Pentacam are proprietary company information, and internal processing algorithms might be based on different (i.e., higher resolution or contrast) datasets.
Orbscan
FIGURE 2.
Bland-Altman plots of difference between observers over mean between
observers. (A) Ultrasound Pachymetry, (B) Scanning Slit Topography, (C)
Rotating Scheimpflug Imaging. The solid lines represent mean differences
between observers; the dotted lines represent upper and lower borders of
the 95% limit of agreement (mean difference ⫾ 1.96 * standard deviation
of the mean difference). Rotating Scheimpflug imaging shows the smallest
interobserver variations.
The differences between CCT measurements determined with
scanning slit pachymetry and ultrasound have been extensively
discussed. There seems to be a trend to higher CCT values obtained with Orbscan, at least in the European and American population. Some investigators report values close to 30 ␮m higher21,22; others found smaller differences below 10 ␮m.16,23,24
These differences are in part the result of the fact that the latter
studies multiplied the Orbscan results with the “acoustic correction factor” of 0.92, as suggested by the manufacturer, to compensate for the systematic difference between Orbscan and ultrasound
measurements. The results for corrected Orbscan values are close
to CCT measurements obtained with a calibrated confocal microscope, in which mean CCT was 39 ␮m thinner than when measured with ultrasound.7 However, as stated subsequently, any general factor applied to all measurements on a given system will not
necessarily decrease the amount of uncertainty between modalities.16 The results of the present investigation are therefore given
with and without the acoustic correction factor. The uncorrected
Orbscan data were 20 ␮m thicker than ultrasound, and the corrected Orbscan values were 26 ␮m thinner. These results are similar to a previous study comparing Orbscan and ultrasound.12
Like other optical pachymetry devices, the Orbscan requires
clear reflections on the epithelial and endothelial corneal surfaces
and homogeneous composition of the diverse optical media to
obtain precise measurements.25 As such, its use is limited to corneas that are free of sources of light scattering or distortion such as
opacities, scarring, deposits, or edema.26 There is no satisfactory
explanation why systematic measurement differences between ultrasound and slit scanning pachymetry exist, but it seems likely
that they are the result of different locations of the respective reflective interfaces in the cornea.
Ultrasound
Some authors regard ultrasound pachymetry as a “gold standard” for CCT, mainly because of its superiority to older mechan-
Optometry and Vision Science, Vol. 82, No. 10, October 2005
Repeatability and Reproducibility of Central Corneal Thickness Measurement—Lackner et al.
897
FIGURE 3.
Bland-Altman plots of difference between modalities over mean between modalities. (A) Scanning-slit topography versus ultrasound, (B) corrected
scanning slit topography (acoustic correction factor ⫽ 0.92), (C) rotating Scheimpflug imaging versus ultrasound, (D) rotating Scheimpflug imaging vs.
(uncorrected) scanning slit topography. The solid lines represent mean differences between modalities; the dotted lines represent upper and lower
borders of the 95% limits of agreement (mean difference ⫾ 1.96 * standard deviation of the mean difference). Rotating Scheimpflug imaging shows not
only the smallest systematic difference compared with ultrasound pachymetry, but also the narrowest 95% limits of agreement.
ical pachymeters (e.g., the Haag-Streit pachymeter).9 Although
ultrasound has been reported to have good intraobserver reproducibility,27 a higher degree of variation has been described between
observers.28 In the present study, the difference between ultrasound and scanning slit topography was on the lower end of the
range reported by previous studies (Gherghel,29 Chakrabarti,14
Rainer,24). The assessment of variability is complicated by the fact
that some investigators report variability as the expected ranges of
mean measurements of groups equal in size to the study collective
rather than expected changes of individual measurements.29 Difficulties with centration of the handheld ultrasound probe on the
thinnest part of the cornea may additionally contribute to a bias
toward larger measurements for CCT compared with optical
12
methods. Ultrasonic distance measurements are based on reflections of soundwaves from interfaces of tissues with different acoustic impedance. In the cornea, neither the acoustic impedances nor
the histologic locations of the interfaces are exactly known and
have to be approximated. In addition, differences in corneal water
content and stromal extracellular matrix metabolism can likely
change these properties, e.g., in the postoperative cornea.30
Reliability of the Differences
Traditionally, differences in CCT measurements between observers and modalities are given as mean difference ⫾ 95% CI. It
would, however, be incorrect to conclude that 95% of the actually
measured CCT differences will be within this interval. A more
adequate interpretation of the 95% CI is that the likelihood to
receive the observed data is equal or higher than 95% if the difference between the modalities in the whole population lies within
this interval. An estimate for the interval in which 95% of the
actual differences will be is the “limit of agreement” (mean ⫾ 1.96
* SD for normally distributed data).19,20 This interval is considerably broader than the 95% CI of the difference, and it is likely that
one in 20 patients will still have larger differences in CCT measurements.
Even if the mean CCT differences between observers or modalities are small or zero, a high standard deviation will broaden the
95% limits of agreement, and might bring the borders close to or
beyond clinically relevant levels. In such cases, it will probably be
unacceptable to have 5% of the patients outside these intervals, and
Optometry and Vision Science, Vol. 82, No. 10, October 2005
898
Repeatability and Reproducibility of Central Corneal Thickness Measurement—Lackner et al.
99% (or higher) limits of agreement will be needed (Table 3). For
these reasons, conversion equations, as suggested by some authors
and manufacturers,7,31 must remain unsatisfactory. Even if the
systematic component of the measurement differences between 2
methods is fully compensated by the application of a formula, a
high intra- or interobserver variability can still impede their comparability.
Pentacam measured CCT thinner than ultrasound. The Pentacam
values for CCT were closer to the ultrasound values than the Orbscan values; thus, we conclude that Pentacam can be regarded as
valid in determining CCT in eyes with normal cornea as the other
modalities. The validity of Pentacam measurements of CCT in
diseased corneas is currently under investigation.
Received May 11, 2005; accepted June 9, 2005.
Clinical Implications of Error Levels
Errors in determining CCT are most likely to be clinical relevant
in 2 applications: First, accurate CCT measurements are of increasing importance in the diagnosis of glaucoma. It is known that
intraocular pressure (IOP) measurements, especially when taken
by with applanation tonometry, are influenced by CCT. A 10%
increase in CCT can result in an apparent IOP increase of 3.4 mm
Hg on average and up to 10 mm Hg in patients with acute-onset
disease.8
Second, monitoring CCT in patients after refractive surgery is
of importance because postoperative eye pressure readings by applanation tonometry are lower than the preoperative measurements solely as a result of a thinner postoperative cornea rather
than a real decrease in eye pressure. False low eye pressure readings
pose the risk of delaying the diagnosis of future glaucoma in patients who undergo refractive surgery. Importantly, ultrasound assessment of CCT in patients after refractive surgery might be additionally biased because of changes in corneal acoustic impedance
resulting from differences in stromal extracellular matrix metabolism32 and carries the potential for complications in the vulnerable
postoperative phase.33
Furthermore, CCT measurement is integral part of the preoperative assessment of patients undergoing refractive surgery. In
treating patients with laser in situ keratomileusis (LASIK), it has
been suggested that a minimum of 250 ␮m of tissue remain beneath the lamellar flap after stromal ablation to ensure maintenance of corneal structural integrity.21 Given the amount of uncertainty in determining CCT, considerably more tissue would
have to be left unablated to ensure this condition. This would be an
especially important issue when treating eyes with higher myopic
refractive errors and proportionally larger ablation depths.
The results of the present study suggest that in healthy corneas,
the corrected Orbscan and the Pentacam tend to give more conservative estimates than ultrasound pachymetry. Compared with
ultrasound, the optical methods can be expected to err on the lower
side of CCT only by 11 to 15 ␮m on a 99% limit of agreement
basis, i.e., only one in 100 patients will have an optical CCT
reading that is more than 11 to 15 ␮m lower than it would have
been with ultrasound.
CONCLUSION
Of 3 investigated modalities, the Pentacam showed the lowest
interobserver variability, suggesting a high degree of reproducibility. All 3 modalities yielded similar values of repeatability. It is not
known which modality is closest to the “true” values, and all 3
modalities differ from each other with a systematic and a stochastic
component. Although (uncorrected) Orbscan yielded larger CCT
values than ultrasound in accordance with previous studies, the
REFERENCES
1. Gherghel D, Hosking SL, Mantry S, Banerjee S, Naroo SA, Shah S.
Corneal pachymetry in normal and keratoconic eyes: Orbscan II versus ultrasound. J Cataract Refract Surg 2004;30:1272–7.
2. Wolfs RC, Klaver CC, Vingerling JR, Grobbee DE, Hofman A, de
Jong PT. Distribution of central corneal thickness and its association
with intraocular pressure: the Rotterdam Study. Am J Ophthalmol
1997;123:767–72.
3. Herndon LW, Choudhri SA, Cox T, Damji KF, Shields MB, Allingham RR. Central corneal thickness in normal, glaucomatous, and
ocular hypertensive eyes. Arch Ophthalmol 1997;115:1137–41.
4. Argus WA. Ocular hypertension and central corneal thickness. Ophthalmology 1995;102:1810–2.
5. Medeiros FA, Sample PA, Zangwill LN, Bowd C, Aihara M, Weinreb
RN. Corneal thickness as a risk factor for visual field loss in patients
with preperimetric glaucomatous optic neuropathy. Am J Ophthalmol 2003;136:805–13.
6. Medeiros FA, Sample PA, Weinreb RN. Corneal thickness measurements and visual function abnormalities in ocular hypertensive patients. Am J Ophthalmol 2003;135:131–7.
7. McLaren JW, Nau CB, Erie JC, Bourne WM. Corneal thickness
measurement by confocal microscopy, ultrasound, and scanning slit
methods. Am J Ophthalmol 2004;137:1011–20.
8. Doughty MJ, Zaman ML. Human corneal thickness and its impact
on intraocular pressure measures: a review and meta-analysis approach. Surv Ophthalmol 2000;44:367–408.
9. Thornton SP. A guide to pachymeters. Ophthalmic Surg 1984;15:
993–5.
10. Javaloy J, Vidal MT, Villada JR, Artola A, Alio JL. Comparison of
four corneal pachymetry techniques in corneal refractive surgery.
J Refract Surg 2004;20:29–34.
11. Rainer G, Petternel V, Findl O, Schmetterer L, Skorpik C, Luksch A,
Drexler W. Comparison of ultrasound pachymetry and partial coherence interferometry in the measurement of central corneal thickness.
J Cataract Refract Surg 2002;28:2142–5.
12. Rainer G, Findl O, Petternel V, Kiss B, Drexler W, Skorpik C,
Georgopoulos M, Schmetterer L. Central corneal thickness measurements with partial coherence interferometry, ultrasound, and the
Orbscan system. Ophthalmology 2004;111:875–9.
13. Carkeet A, Saw SM, Gazzard G, Tang W, Tan DT. Repeatability of
IOLMaster biometry in children. Optom Vis Sci 2004;81:829–34.
14. Feng Y, Simpson TL. Comparison of human central cornea and
limbus in vivo using optical coherence tomography. Optom Vis Sci
2005;82:416–9.
15. Myrowitz EH, Kouzis AC, O’Brien TP. High interocular corneal
symmetry in average simulated keratometry, central corneal thickness, and posterior elevation. Optom Vis Sci. 2005;82:428–31.
16. Gonzalez-Meijome JM, Cervino A, Yebra-Pimentel E, Parafita MA.
Central and peripheral corneal thickness measurement with Orbscan
II and topographical ultrasound pachymetry. J Cataract Refract Surg
2003;29:125–32.
17. Moezzi AM, Fonn D, Simpson TL, Sorbara L. Contact lens-induced
Optometry and Vision Science, Vol. 82, No. 10, October 2005
Repeatability and Reproducibility of Central Corneal Thickness Measurement—Lackner et al.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
corneal swelling and surface changes measured with the Orbscan II
corneal topographer. Optom Vis Sci 2004;81:189–93.
Bland JM, Altman DG. Measuring agreement in method comparison
studies. Stat Methods Med Res 1999;8:135–60.
Bland JM, Altman DG. Comparing methods of measurement: why
plotting difference against standard method is misleading. Lancet
1995;346:1085–7.
Bland JM, Altman DG. Statistical methods for assessing agreement
between two methods of clinical measurement. Lancet 1986;1:
307–10.
Chakrabarti HS, Craig JP, Brahma A, Malik TY, McGhee CN. Comparison of corneal thickness measurements using ultrasound and
Orbscan slit-scanning topography in normal and post-LASIK eyes.
J Cataract Refract Surg 2001;27:1823–8.
Yaylali V, Kaufman SC, Thompson HW. Corneal thickness measurements with the Orbscan Topography System and ultrasonic pachymetry. J Cataract Refract Surg 1997;23:1345–50.
Radford SW, Lim R, Salmon JF. Comparison of Orbscan and ultrasound pachymetry in the measurement of central corneal thickness.
Eye 2004;18:434–6.
Giraldez Fernandez MJ, Diaz Rey A, Cervino A, Yebra-Pimentel E. A
comparison of two pachymetric systems: slit-scanning and ultrasonic.
CLAO J 2002;28:221–3.
Boscia F, La Tegola MG, Alessio G, Sborgia C. Accuracy of Orbscan
optical pachymetry in corneas with haze. J Cataract Refract Surg
2002;28:253–8.
Tam ES, Rootman DS. Comparison of central corneal thickness measurements by specular microscopy, ultrasound pachymetry, and ultrasound biomicroscopy. J Cataract Refract Surg 2003;29:1179–84.
Wheeler NC, Morantes CM, Kristensen RM, Pettit TH, Lee DA.
28.
29.
30.
31.
32.
33.
899
Reliability coefficients of three corneal pachymeters. Am J Ophthalmol 1992;113:645–51.
Bovelle R, Kaufman SC, Thompson HW, Hamano H. Corneal
thickness measurements with the Topcon SP-2000P specular microscope and an ultrasound pachymeter. Arch Ophthalmol 1999;117:
868–70.
Miglior S, Albe E, Guareschi M, Mandelli G, Gomarasca S, Orzalesi
N. Intraobserver and interobserver reproducibility in the evaluation
of ultrasonic pachymetry measurements of central corneal thickness.
Br J Ophthalmol 2004;88:174–7.
Lackner B, Pieh S, Schmidinger G, Hanselmayer G, Simader C,
Reitner A, Skorpik C. Glare and halo phenomena after laser in situ
keratomileusis. J Cataract Refract Surg 2003;29:444–50.
Suzuki S, Oshika T, Oki K, Sakabe I, Iwase A, Amano S, Araie M.
Corneal thickness measurements: scanning-slit corneal topography
and noncontact specular microscopy versus ultrasonic pachymetry.
J Cataract Refract Surg 2003;29:1313–8.
Patel S, Aslanides IM. Main causes of reduced intraocular pressure
after excimer laser photorefractive keratectomy. J Refract Surg 1996;
12:673–4.
Kawana K, Tokunaga T, Miyata K, Okamoto F, Kiuchi T, Oshika T.
Comparison of corneal thickness measurements using Orbscan II,
non-contact specular microscopy, and ultrasonic pachymetry in eyes
after laser in situ keratomileusis. Br J Ophthalmol 2004;88:466–8.
Optometry and Vision Science, Vol. 82, No. 10, October 2005
Birgit Lackner
Medical University of Vienna
Department of Ophthalmology
Währinger Gürtel 18-20
A-1090 Vienna, Austria
e-mail: [email protected]