Download Percentage Thickness Increase and Absolute Difference from

LETTERS TO THE EDITOR
Percentage Thickness Increase and
Absolute Difference from Thinnest to
Describe Thickness Profile
To the Editor:
In the September 2009 issue of the Journal of
Refractive Surgery, Reinstein et al published a study
characterizing the normal stromal thickness profile.1
The authors should be commended for the very interesting and elegant approach that enhances the pioneering work of Mandell and Polse2 using horizontal
profile as a tool for diagnosing keratoconus. Stromal
profile has great potential in the early identification of
stromal thinning disorders, which may impact screening for refractive candidates.
Reinstein et al1 used Artemis very high-frequency
digital ultrasound scanning (ArcScan, Morrison, Colo)
across the central 10-mm corneal diameter on 110 normal eyes. The authors found a mean stromal thickness
at the thinnest point of 461.837.3 µm. For calculating
the absolute stromal thickness progression, the averages
of the thickness at 35 annular ring bands (0.06-mm
wide), centered on the thinnest point with central radii increasing in 0.1-mm increments, were calculated.
Reinstein et al then calculated the absolute difference
between the stromal thickness of each annular band
and the thinnest stroma at each radial distance of each
eye. The averages of these differences were plotted
against the distance from the thinnest point and a second order polynomial (quadratic) equation was used
to describe the averages of absolute stromal thickness
differences from the thinnest point with an excellent
fit and very high correlation (R2=0.999). Interestingly,
the absolute stromal thickness differences were independent of stromal thickness at the thinnest point.
This approach is similar to the one our group described previously for calculating the corneal thickness
spatial profile (CTSP) using the Pentacam (Oculus Optikgeräte GmbH, Wetzlar, Germany).3 In agreement with
our results on corneal thickness in normal corneas, a
tight range was observed for the gradual increase of stromal thickness from the thinnest point towards the periphery. The authors considered the absolute difference
between each annular band and the thinnest instead of
the percentage of increase. They state that the percentage increase can be misleading because the percentage
increase would be greater for a thin cornea than for a
thick cornea, which would lead to false positives in
normal thin corneas and false negatives in cases with
keratoconus and normal central thickness. However,
this is exactly the opposite of our experience using the
Pentacam software. In fact, the absolute difference in
thickness (ADT) was tested, as we first started performing thickness profile studies using the Orbscan (Bausch
& Lomb, Rochester, NY).4 The decision for proceeding
with the percentage thickness increase (PTI) was based
on statistical evidence of its superiority as a test for diagnosing keratoconus (Luz and Ambrósio, unpublished
data 2004).
For comparing the performance of the PTI and ADT
for describing thickness increase from the thinnest
point, we analyzed a database comprising both eyes of
113 normal patients (n=226) and 44 patients with keratoconus (n=88). For every radii, there was a statistically
significant difference among normals and those with
keratoconus for either the PTI and/or ADT parameters
Figure 1. Sensitivity and specificity of the
spatial annular thickness averages in relation to the thinnest for keratoconus detection in 226 normal eyes and 88 eyes with
keratoconus.
84
Copyright © SLACK Incorporated
Letters to the Editor
Figure 2. Pentacam (Oculus Optikgeräte GmbH) images of a normal thin
cornea (A and B) and bilateral keratoconus with relatively “normal central
corneal thickness” (C and D). Corneal thickness spatial profile (CTSP) and
percentage thickness increase (PTI) graphs present the mean and 95%
confidence interval (CI) limits of a normal population (dotted black lines).
The examined eye is presented as a solid red line. These graphs have the
vertical axis (Y) inverted. A) Normal thickness profiles in both eyes. Note
the graphs with PTI (lower graph) in both eyes are similar to the mean in
the normal population (blue arrows). B) Normal corneal asphericity with
low toricity in the sagittal (axial) curvature maps. C) Abnormal thickness
profile with an important escape from the 95% CI limit of the PTI line in
both eyes (red arrows). D) Typical keratoconus pattern of asymmetric
bowtie inferior steepening 48.00 D in both eyes.
(P.001, Mann-Whitney U test). The receiver operating characteristic curves for each radial distance were
calculated to provide the best cutoff value, sensitivity,
and specificity for the PTI and ADT at each radii up
to 4.6 mm (9.2 mm in diameter). Figure 1 illustrates
that both sensitivity and specificity of the PTI parameters were better than the ADT values. The best performances for PTI and ADT were found at the diameters
from 3.6 and 6.0 mm. The best performance was for the
PTI at 5.2 mm in diameter, in which a percentage increase of 17.97% from the thinnest point led to 97.3%
specificity and 95.5% sensitivity. For the ADT, 4 mm
in diameter provided the best parameter, resulting in
94.3% and 93.4% sensitivity and specificity, respectively, for a cutoff of 57 µm of difference.
The examples in Figure 2 demonstrate the Pentacam
findings of a normal thin cornea with normal topography and normal CTSP and PTI graphs in both eyes and
a case with bilateral keratoconus despite 540 µm of
central corneal thickness and abnormal thickness profile graphs. Either case could have been false positive
(thin cornea) and false negative (keratoconus) if considering the ADT instead of the PTI.
I applaud the authors for the interesting work but
urge caution when deciding not to use the percentage
increase in favor of absolute difference. In fact, differences arise from the corneal thickness and stromal
thickness and also from optical blue light Scheimpflug
and high frequency ultrasound. Thereby, the performance of the PTI and ADT approaches should be scientifically tested before deciding which one should be
used and most importantly, criticizing one or the other
may be based on false impressions.
Renato Ambrósio Jr, MD, PhD
Rio de Janeiro, Brazil
Dr Ambrósio is a consultant for Oculus Optikgeräte GmbH.
doi:10.3928/10815987X-20100121-01
REFERENCES
1. Reinstein DZ, Archer TJ, Gobbe M, Silverman RH, Coleman
DJ. Stromal thickness in the normal cornea: three-dimensional
display with Artemis very high-frequency digital ultrasound.
J Refract Surg. 2009;25:776-786.
Journal of Refractive Surgery • Vol. 26, No. 2, 2010
85
Letters to the Editor
2. Mandell RB, Polse KA. Keratoconus: spatial variation of corneal
thickness as a diagnostic test. Arch Ophthalmol. 1969;82:182-188.
3. Ambrósio R Jr, Alonso RS, Luz A, Coca Velarde LG. Corneal
thickness spatial profile and corneal-volume distribution: tomographic indices to detect keratoconus. J Cataract Refract Surg.
2006;32:1851-1859.
4. Luz A, Ursulio M, Castañeda D, Ambrósio R Jr. Corneal thickness progression from the thinnest point to the limbus: study
based on a normal and a keratoconus population to create reference values [Portuguese]. Arq Bras Oftalmol. 2006;69:579-583.
Reply:
We thank Dr Ambrósio for his letter referring to
our paper describing the stromal thickness profile in a
population of normal eyes.1 Our finding that the stromal thickness profile was independent of central stromal thickness was both surprising and interesting; it
seems remarkable that the thickness profile would be
the same despite the large range in minimum corneal
thickness from 450 to 650 µm in the population. This
result seems to show that the cornea is biologically
programmed to have a specific thickness progression
regardless of the minimum central corneal thickness.
If percentage thickness increase (PTI) is used, there
can be no argument that the PTI of a keratoconic eye
with central corneal thickness of 550 µm will be lower
(and closer to normal) than if the same eye had a central
corneal thickness of 450 µm. On the other hand, the absolute difference in thickness (ADT) will be the same for
both eyes. This means that these two eyes might be classified differently according to PTI, whereas they would
be considered identical according to ADT.
The outstanding question is whether this has an impact on screening for keratoconus.
Dr Ambrósio described a study that found PTI to be
statistically superior for diagnosing keratoconus compared to ADT. However, this result is not surprising
when comparing a normal population with a frank keratoconic population, given the corneal thinning associated with keratoconus. In a case of advanced keratoconus, the PTI is taking advantage of the thin cornea to
accentuate the difference compared with normal eyes.
However, the issue is not about identifying cases with
keratoconus that can already be diagnosed by topography or tomography. Our goal now is to improve the
sensitivity and specificity of keratoconus screening in
equivocal cases—cases in which a confident diagnosis
is not possible based on topography alone. In early or
mild cases of keratoconus, it is unlikely that the cornea will have thinned significantly, which means that
the PTI will not be able to rely on the extra diagnostic
factor of the low corneal thickness, as in cases of more
advanced keratoconus. This seems to imply that the
sensitivity and specificity of the PTI might be reduced
for early or mild keratoconus and there may be greater
opportunity for type I and type II errors.
86
The question that needs to be answered is whether
the sensitivity and specificity is higher when using PTI
or ADT for these suspect keratoconus cases, and further study is required in this area.
It is unlikely that any single parameter is going to
provide a definitive answer for keratoconus screening.
The ideal solution for keratoconus screening will be
an analysis that considers information from all available keratoconus screening techniques, including front
and back surface topography, minimum corneal thickness, coincidence of front and back surface topography
and pachymetry, pachymetric progression and profile,
thinnest pachymetry location, Ocular Response Analyzer (Reichert, Depew, NY) signal analysis (or other
biomechanical measurements), magnitude and axis
of astigmatism, keratometry, higher order aberrations,
age, fellow-eye asymmetry, ethnicity, corrected distance visual acuity, contrast sensitivity, and epithelial
thickness profiles.2,3 Alternatively, we may be fortunate to find a genetic or enzymatic test to rule keratoconus “in” or “out.”4 We acknowledge that PTI is
a step along this road in that it combines minimum
corneal thickness and pachymetric progression into
one analysis. However, it is valid to question whether
the PTI method gives too much influence to minimum
corneal thickness. Also, when we do get to the stage of
throwing everything into a single analytical pot, using
PTI would effectively mean adding minimum corneal
thickness twice.
We would like to thank Dr Ambrósio for highlighting this point for debate as the robustness of ADT
across the range of corneal thickness was certainly a
result that surprised us and one that we believe merits
further study.
Dan Z. Reinstein, MD, MA(Cantab), FRCSC, FRCOphth
Timothy J. Archer, MA(Oxon), DipCompSci(Cantab)
Marine Gobbe, MST(Optom), PhD
London, United Kingdom
Ronald H. Silverman, PhD
D. Jackson Coleman, MD
New York, New York
doi:10.3928/1081597X-20100121-02
REFERENCES
1. Reinstein DZ, Archer TJ, Gobbe M, Silverman RH, Coleman
DJ. Stromal thickness in the normal cornea: three-dimensional
display with Artemis very high-frequency digital ultrasound.
J Refract Surg. 2009;25:776-786.
2. Reinstein DZ, Archer TJ, Gobbe M. Corneal epithelial thickness profile in the diagnosis of keratoconus. J Refract Surg.
2009;25:604-610.
3. Reinstein DZ, Archer TJ, Gobbe M. Stability of LASIK in topographically suspect keratoconus confirmed non-keratoconic by
Artemis VHF digital ultrasound epithelial thickness mapping:
1-year follow-up. J Refract Surg. 2009;25:569-577.
4. Rabinowitz YS, Dong L, Wistow G. Gene expression profile
Copyright © SLACK Incorporated
Letters to the Editor
studies of human keratoconus cornea for NEIBank: a novel cornea-expressed gene and the absence of transcripts for aquaporin 5. Invest Ophthalmol Vis Sci. 2005;46:1239-1246.
Artisan Phakic IOL for the Correction
of Ametropia After Deep Anterior
Lamellar Keratoplasty
To the Editor:
We present our experience with the Artisan phakic
intraocular lens (IOL) (Ophtec BV, Groningen, The
Netherlands) for the correction of ametropia following
deep anterior lamellar keratoplasty in two patients.
Both patients were chosen due to the presence of low
regular astigmatism.
Both patients (25 and 27 years old, respectively)
were keratoconic with poor vision due to corneal scarring. Both patients underwent deep anterior lamellar
keratoplasty using the Melles technique.
In the first case, uncorrected distance visual acuity (UDVA) in the right eye 3 years after deep anterior
lamellar keratoplasty was 20/120 and corrected distance visual acuity (CDVA) was 20/30 with a refraction of 4.50 1.20 105. Uncorrected distance visual acuity in the left eye was 20/30. A 6.0/8.5-mm,
6.50-diopter (D) Artisan phakic IOL was implanted
uneventfully in the right eye. Preoperative keratometry
readings were 46.80/47.50 D. Anterior chamber depth
was 4.67 mm. Preoperative endothelial cell count was
1991 cells/mm2. One week postoperatively, UDVA had
improved to 20/30. At 6 months postoperatively, the
patient achieved UDVA of 20/20 and CDVA of 20/16
with a refraction of 0.75 1.50 91. The endothelial
cell count was 2318 cells/mm2.
In the second case, UDVA 2 years after transplantation was 20/80 and CDVA was 20/30 with a refraction of
2.00 0.25 90. Uncorrected distance visual acuity
of the left eye was 20/20. A 6.0/8.5-mm 2.50-D Artisan
phakic IOL was implanted uneventfully in the right eye.
Preoperative keratometry readings were 45.50/46.00 D.
Anterior chamber depth was 3.96 mm. Preoperative
endothelial cell count was 2721 cells/mm2. At 2 weeks
postoperatively, UDVA improved to 20/60. Ten months
postoperatively, UDVA was 20/20 and CDVA was
20/20 with a refraction of 0.75 1.00 75. The
endothelial cell count was 2327 cells/mm2.
Artisan phakic IOLs have been used with good results for the treatment of refractive errors in otherwise
healthy eyes1 and have demonstrated safety in longterm follow-up studies.2
Our patients were chosen for phakic IOL implantation, as both had good spectacle correction (indicating
Journal of Refractive Surgery • Vol. 26, No. 2, 2010
little irregular astigmatism) and intolerance to contact
lenses and spectacles.
Artisan IOLs demonstrate a safe profile regarding
endothelial cell count, with a reported loss between
8.6% and 14.5% at 5 years.3-5
In the first case, pre- and 6-month postoperative endothelial cell count was 1991 cells/mm2 and 2318 cells/mm2,
respectively. In the second case, pre- and 10-month postoperative endothelial cell count was 2721 cells/mm2 and
2327 cells/mm2, respectively.
The significant gain of CDVA after implantation
(from 20/30 to 20/16 in the first case and from 20/30
to 20/20 in the second case) was most likely due to
patient fatigue in the preoperative assessment. After
the procedure, both patients were extremely satisfied
with the result and were more motivated during final
refraction test.
To our knowledge, this is the first report of the use
of Artisan phakic IOLs for the correction of ametropia
after deep anterior lamellar keratoplasty.
Compared to wavefront-guided photorefractive keratoplasty, Artisan IOL implantation carries no risk of
postoperative haze, and it is an excellent option in settings where access to laser refractive equipment is not
available. The procedure is reversible, and it does not
compromise the structural integrity of the graft. It can
offer an excellent visual outcome in carefully selected
cases with previous deep anterior lamellar keratoplasty. Longer follow-up is needed to study the rate of endothelial cell loss in these eyes and the stability of the
refractive outcome.
P. Georgoudis, MRCOphth
M.J. Tappin, FRCOphth
Surrey, United Kingdom
doi:10.3928/1081597X-20100121-03
REFERENCES
1. Moshirfar M, Holz HA, Davis DK. Two-year follow-up of the Artisan/Verisyse iris-supported phakic intraocular lens for the correction of high myopia. J Cataract Refract Surg. 2007;33:13921397.
2. Tahzib NG, Nuijts RM, Wu WY, Budo CJ. Long-term study of
Artisan phakic intraocular lens implantation for the correction
of moderate to high myopia ten-year follow-up results. Ophthalmology. 2007;114:1133-1142.
3. Saxena R, Boekhoorn SS, Mulder PG, Noordzij B, van Rij G,
Luyten GP. Long-term follow-up of endothelial cell change after
Artisan phakic intraocular lens implantation. Ophthalmology.
2008;115:608-613.
4. Benedetti S, Casamenti V, Benedetti M. Long-term endothelial
changes in phakic eyes after Artisan intraocular lens implantation to correct myopia: five-year study. J Cataract Refract Surg.
2007;33:784-790.
5. Silva RA, Jain A, Manche EE. Prospective long-term evaluation
of the efficacy, safety, and stability of the phakic intraocular
lens for high myopia. Arch Ophthalmol. 2008;126:775-781.
87