Download “Ultrathin” DSAEK Tissue Prepared With a Low–Pulse Energy, High

BASIC INVESTIGATION
“Ultrathin” DSAEK Tissue Prepared With a Low–Pulse
Energy, High-Frequency Femtosecond Laser
Paul M. Phillips, MD,* Louis J. Phillips, OD,* Hisham A. Saad, MD,† Mark A. Terry, MD,‡§
Donna B. Stolz, PhD,¶ Christopher Stoeger, CEBT,§ Jonathan Franks, PhD,¶
and David Davis-Boozer, MPH§
Purpose: To evaluate the endothelial cell survival and stromal bed
quality when creating deep stromal cuts with a low–pulse energy,
high-frequency femtosecond laser to produce “ultrathin” tissue for
Descemet stripping automated endothelial keratoplasty.
Methods: Seventeen corneas were used for this study. Five corneas
were cut with the laser at a depth of 420 to 500 mm to produce
a tissue thickness of approximately #70 mm. Five corneas served
as an uncut comparison group. Vital dye staining and computer
digitized planimetry analysis were performed on these corneas.
The 7 remaining corneas were cut for scanning electron microscopy
evaluation.
Results: The mean central posterior stromal thickness of cut
corneas was 60.6 mm (range, 43–72 mm). Endothelial cell damage
in cut and comparison corneas was 3.92% ± 2.22% (range, 1.71%–
6.51%) and 4.15% ± 2.64% (range, 1.21%–7.01%), respectively
(P = 0.887). Low-magnification (·12) scanning electron microscopy
revealed a somewhat irregular-appearing surface with concentric
rings peripherally. Qualitative grading of higher magnification
(·50) central images resulted in an average score of 2.56 (between
smooth and rough).
Conclusions: Ultrathin tissue for Descemet stripping automated
endothelial keratoplasty can be safely prepared with minimal
endothelial cell damage using a low–pulse energy, high-frequency
femtosecond laser; however, the resulting stromal surface quality
may not be optimal with this technique.
Key Words: Descemet stripping automated endothelial keratoplasty,
Descemet stripping endothelial keratoplasty, endothelial keratoplasty, femtosecond laser, ultrathin DSAEK, endothelium
(Cornea 2012;0:1–6)
Received for publication January 31, 2012; revision received April 2, 2012;
accepted April 18, 2012.
From the *Sightline Ophthalmic Associates, Sewickley, PA; †Tanta University
Faculty of Medicine, Tanta, Egypt; ‡Devers Eye Institute, Portland, OR;
§Lions VisionGift, Portland, OR; and ¶University of Pittsburgh School of
Medicine, Center for Biologic Imaging, Pittsburgh, PA.
The authors state that they have no financial or conflicts of interests to
disclose.
Reprints: Paul M. Phillips, Sightline Ophthalmic Associates, Suite 104, 2591
Wexford-Bayne Rd, Sewickley, PA 15143 (e-mail: paulphillipsmd@
gmail.com).
Copyright © 2012 by Lippincott Williams & Wilkins
Cornea Volume 0, Number 0, Month 2012
D
escemet stripping automated endothelial keratoplasty
(DSAEK) is currently the surgical standard of care for
the treatment of endothelial failure. Although there is no
question regarding the efficacy of this surgery, which results
in the vast majority of patients with best spectacle-corrected
acuity of 20/40 or better,1–4 there has been a great deal of
debate regarding the cause for the relatively small percentage
of patients achieving 20/20 vision despite beautifully clear
postoperative corneas. One theory is that the added stromal
thickness resulting from DSAEK surgery may have some
bearing on the final visual acuity and that the use of very
thin donor tissue, also referred to as “ultrathin” tissue, may
result in better outcomes. Currently, there is no precise definition of ultrathin tissue, but most surgeons would consider
such tissue to be 100 mm or thinner.
Despite the lack of clear evidence at this time to support
the notion that thinner tissue yields better visual outcomes in
DSAEK, some surgeons argue that this ultrathin tissue may
result in both faster vision recovery and better final visual
acuity. Thinner tissue may also theoretically allow for less
crushing of tissue when using new injector systems. Many
surgeons are now attempting to produce such tissue for their
surgeries or are requesting that eye banks produce this
ultrathin tissue for their use.
With advancing femtosecond technology, allowing for
much deeper stromal cuts, it may now be possible to produce
ultrathin DSAEK tissue using this technology. One concern
regarding the use of a femtosecond laser to cut such thin
tissue is that the laser’s photodisruption process may be damaging to the endothelium when cutting close to the delicate
cells. There is some evidence that this may be the case when
cutting within around 100 mm of the endothelium using the
Intralase (Abbott Medical Optics, Abbott Park, IL) femtosecond laser.5 Additionally, although femtosecond lasers may
make smoother laser in situ keratomileusis flaps than microkeratomes when cutting at depths of 110 to 140 mm,6 some
studies have shown that when cutting deeper (300–400 mm),
the stromal surface appears to be rougher with the laser than
with the microkeratome.7,8
To date, no study has been published looking at
endothelial survival when cutting ultrathin grafts of 70-mm
thickness. Additionally, no study has looked at stromal bed
quality when using a low–pulse energy, high-frequency
(LPEHF) femtosecond laser (Ziemer Femto LDV; Ziemer
Ophthalmic Systems, Port, Switzerland). Theoretically, this
high-frequency laser (.5 MHz), which uses closely
www.corneajrnl.com |
1
Cornea Volume 0, Number 0, Month 2012
Phillips et al
overlapping spot and line separation, may lead to a smoother
bed, although at the same time, through a low pulse energy
(,100 nJ), avoid damage to the nearby endothelium. The
goal of this study, therefore, was to evaluate both endothelial
survival and stromal bed quality when using an LPEHF femtosecond laser to cut ultrathin tissue at an approximate thickness of 70 mm.
MATERIALS AND METHODS
Donor Tissue
Seventeen corneoscleral buttons unsuitable for transplantation were supplied by the Lions VisionGift. All
samples were free of gross deformities and generally clear
optically. Ten corneas were chosen to have reasonably good
endothelial cell counts (1736–3135 cells/mm2) and were
used for the endothelial cell survival analysis. Of these, 5 corneas were cut up to depths of 420 to 500 mm to produce
ultrathin tissue of approximately #70 mm, and 5 corneas
were used as uncut comparisons. An additional 7 corneas
were cut at depths of 450 to 500 mm for a goal depth of
approximately #70 mm for evaluation with scanning electron microscopy (SEM) to determine the stromal bed quality
produced by these deep cuts.
Femtosecond Laser Protocol
The endothelium of the donor button was coated with
Healon 10 (Abbott Medical Optics), and the donor button was
secured onto the Ziemer artificial anterior chamber (AAC;
Ziemer Ophthalmic Systems). The chamber was then infused
with balanced salt solution (BSS) through an intravenous line
at a bottle height of ;80 cm above the base of the anterior
chamber maintainer to achieve a moderately firm pressure by
tactile tension, and the infusion line was clamped closed. The
corneal thickness was measured with a Sonogage ultrasonic
pachymeter (Sonogage, Cleveland, OH), and the epithelium
was wiped clear of the surface with a blunt spatula and Merocel sponge (Beaver-Visitec International, Waltham, MA).
The central corneal thickness was again measured with the
pachymeter. To get precise diameter measurements, the laser
head was applied to the Ziemer AAC above the surface of the
dry cornea. The head was then lowered by turning the peripheral ring of the AAC clockwise until the central part of the
cornea touched the Intershield. The ring was turned until the
applanated surface, shown on the monitor of the laser,
reached the goal diameter of 9.5 mm. The laser head was
then removed, and Healon was applied to the cornea in
a peripheral ring of ;6 mm and BSS applied centrally. A
laser Intershield was then chosen and applied to the laser head
based on this “epi-off ” pachymetry measurement to achieve
a depth that would most likely result in the desired target goal
of a residual stromal bed of #70 mm. The laser head was then
gently reapplied to the cornea on the AAC and the laser cut
then created. The entire process from the point of application
of the laser shield to the cornea until removal after treatment
took 45 seconds. The identical procedure was followed for
uncut comparison corneas, including mounting them onto the
2
| www.corneajrnl.com
AAC and applying the laser head to the surface of the corneas
for the same 45 seconds to simulate the laser treatment, but in
these comparison cases the laser was not fired. Once the laser
procedure was performed, the interface was entered with
a Thorlakson spatula (Katena Products, Inc, Denville, NJ),
and the flap was separated and lifted from the posterior stromal bed.
In the case of the 5 cut corneas used for endothelial
analysis, the caps were then replaced, and all 10 corneoscleral
rims placed back into the Optisol storage chambers. These
10 corneas used for endothelial analysis (5 comparisons and
5 study eyes) were sent back to the eye bank where the
posterior stromal bed thickness was measured using an
Optovue optical coherence tomographer (Optovue, Inc,
Fremont, CA), before performing the endothelial cell analysis
(see following).
In the case of the 7 corneas sent for electron microscopy, the caps were removed after laser cutting, and the
corneoscleral rims were immediately placed in the fixative.
The corneoscleral rims were then placed in buffer solution
and dehydrated in ascending concentrations of ethanol. They
were treated with hexamethyldisilazane and air-dried. They
were attached to aluminum SEM stubs with colloidal graphite
and were coated with gold–palladium. They were then viewed
with a JEOL JEM-6335F transmission electron microscope
(JEOL Ltd, Tokyo, Japan). The total bed surface was
observed at ·12 magnification, whereas the central area was
observed at ·50 magnification.
Staining and Quantification of
Endothelial Damage
The tissues sent for endothelial analysis were removed
from the Optisol-GS storage chamber, rinsed with BSS, and
trephinated with an 8-mm punch. They were then stained with
0.25% trypan blue (MP Biomedicals, LLC, Solon, OH) and
0.2% alizarin red S (GFS Chemicals, Inc, Columbus, OH).
The technique used for vital dye staining was identical to that
described in a previous study.9
After vital dye staining, quantitative analysis of the
endothelial damage was performed. The graft was placed in
BSS within a clear vial and mounted on the slit lamp and
photographed at ·16 magnification. Illumination was adjusted
to avoid shadows or excessive light reflections. The image
then underwent planimetry quantitative processing using
Adobe Photoshop 7.0 software as described previously.9
The number of pixels that made up the stained area was
determined, and this was divided by the number of pixels
making up the entire endothelial area, thus giving a percentage
of the endothelial damage. The reproducibility of this technique has been previously established.9,10
Statistical Analysis
All data were analyzed using SPSS version 19.0. An
independent samples t test was used to compare cell loss
between study and comparison eyes. Power analyses were
performed using G*Power 3.
2012 Lippincott Williams & Wilkins
Cornea Volume 0, Number 0, Month 2012
Ultrathin DSAEK Tissue With Femtosecond Laser
TABLE 1. Laser Settings and Pre- and Postlaser Corneal Thickness Measurements
Donor
Cornea
1
2
3
4
5
Epi-Off Pachymetry*
(mm)
Laser Cut Depth
(mm)
Anticipated Residual Bed
(mm)
Achieved Residual Bed, OCT†
(mm)
Error (mm)
453
465
509
581
516
420
420
500
500
500
33
45
9
81
16
61
72
63
64
43
+28
+27
+54
217
+27
*Prelaser measurements performed with a pachymeter preoperatively.
†Postlaser, posterior lenticule bed thickness measurements taken with optical coherence tomography (OCT) postoperatively.
Grading of Stromal Smoothness
The SEM images of the central stromal surface of the
resulting posterior corneoscleral buttons (·50) were subjectively graded by 3 masked independent observers in a manner similar to that of Cheng et al.11 The smoothness of the
central stromal surfaces imaged at ·50 were graded on the
following scale: 1 = very smooth, 2 = smooth, 3 = rough,
and 4 = very rough.
of 3.92% ± 2.22% (range, 1.71%–6.51%) and 4.15% ± 2.64%
(range, 1.21%–7.01%), respectively (P = 0.887; Fig. 1). The
resulting effect size for the groups before removal of trephination damage (overall damage) and for the groups with
trephination damage removed was extremely small (Cohen
d = 0.095 and 0.09, respectively). As a result, a priori power
calculations using the means and standard deviations above
show that this study lacks the power needed to detect this very
small effect, which would require a sample of .1000 to gain
the appropriate statistical power.
RESULTS
Laser Cutting Accuracy
Stromal Bed Quality
The intended laser cut depths chosen ranged from 420
to 500 mm resulting in an average residual stromal bed thickness in the 5 study corneas of 60.6 mm (range, 43–72 mm).
There was an average deviation from expected target results
of 30.6 mm (range, 217 to 54 mm; Table 1).
On low-magnification inspection, some of the corneas
had noticeable concentric ring irregularities, but this pattern
was not seen uniformly (Fig. 2). In these corneas with concentric ring formation, the rings were seen peripherally,
whereas the central 2 to 4 mm seemed uninvolved. Grading
of the central stromal bed smoothness was performed at high
magnification (·50) with a subjective grading scale that
resulted in an average grading of all 5 corneas of 2.56 (range,
1.33–4.0, SD = 1.03). This grading fell between “smooth”
and “rough” on a 4-point grading scale of “very smooth” to
“very rough” (Fig. 3).
Endothelial Analysis
Endothelial cell analysis did not reveal a statistically
significant difference between the study and comparison
groups. Overall, cell damage for cut and comparison corneas
was 8.4% ± 3.6% (range, 4.88%–13.26%) and 8.7% ± 2.4%
(range, 6.05%–11.26%), respectively (P = 0.855). When
damage occurring from trephination was removed, the results
continued to demonstrate no statistically significant difference
between cut and comparison corneas, with a damage cell area
DISCUSSION
Although good prospective studies are required to
validate the theory that the use of ultrathin DSAEK grafts
FIGURE 1. Vital dye staining and
computer digitized planimetry analysis of endothelial cell damage
resulting from laser procedure
(corneas 1–5) and comparison procedure (corneas 6–10) after corneal
trephination damage was removed.
*No significant difference was found
at 3.92% (range, 1.71%–6.51%) and
4.15%
(range,
1.21%–7.01%),
respectively (P = 0.887).
2012 Lippincott Williams & Wilkins
www.corneajrnl.com |
3
Phillips et al
Cornea Volume 0, Number 0, Month 2012
FIGURE 2. SEM low-magnification
(·12) postlaser images demonstrating an example of a generally
smooth bed (cornea F) as compared
with a rougher bed with characteristic concentric ring (arrows) formation (cornea G).
can achieve better vision outcomes than standard DSAEK
grafts, surgeons have begun requesting thin DSAEK tissue from
eye banks (J. Clover B.S., C.E.B.T, Lions VisionGift, Oregon,
Personal communications, November 2011). The use of a femto-
second laser to create such tissue may prove to be a viable method
because these lasers can cut to depths of up to 500 mm. This study
was designed to provide preliminary evidence to determine
whether such deep cuts can be made effectively and safely.
FIGURE 3. SEM high magnification (·50) of central corneal images of all 7 study corneas. The stromal bed quality grading score
(average) of each image is noted in parentheses. Grading was on a scale of 1 to 4 (very smooth = 1, smooth = 2, rough = 3, and
very rough = 4).
4
| www.corneajrnl.com
2012 Lippincott Williams & Wilkins
Cornea Volume 0, Number 0, Month 2012
We believe that the laser does cut relatively accurately
at depths of 450 to 500 mm, with a deviation from expected
targeted depths of 30.6 mm. A larger study would be necessary to determine if this deviation is consistent, however.
Our study does demonstrate that the use of a low–pulse
energy (,100 nJ), high-frequency (,5 MHz) laser, to cut
deep within the stroma and close to the endothelium, does
not result in significant amounts of cell loss or death.
Although we did not find a statistically significant difference
between the cut corneas and corneas of the uncut comparison
group, as noted previously, this study was not powered to
detect the very small difference found between the groups.
It is important to note, however, that with such a small effect
size, it would take an impractically large study (.1000 corneas) to detect this difference. We feel that it is reasonable
though to consider that the 0.3 percentage points between the
groups before removing the trephination damage and the
0.23 percentage points between the groups after the trephination damage was removed are not likely to be clinically significant. This is even more evident when considering the fact
that the cell damage that did occur was variable, with no
repeatable pattern observed. Specifically, because the Ziemer
laser pulses are applied in a linear fashion, occurring from
left to right across the cornea, cell death resulting from
these pulses then would be expected to follow such a pattern
along the endothelium, which was not noted in any of the
study corneas.
Although the lack of cell death does give promise for the
possible use of this technology, the resulting stromal bed may
not be ideally smooth. In some of the study corneas, there
seemed to be a concentric ring pattern in the stromal bed,
which has been described in a previous study.12 This concentric
ring pattern is likely the result of an applanation effect of the
laser deforming the posterior stoma more than the anterior
stroma. It is unclear, however, why some tissues in this study
developed these concentric rings, whereas the others did not. It
is possible that this may be caused by different amounts of
corneal rigidity either inherent in the individual donor corneas
or, alternatively, because of differing amounts of stromal
hydration of individual corneas. Although the cause of the
concentric rings is speculation, it is also unclear whether these
concentric rings would significantly affect vision outcomes
because they mostly tended to be more pronounced in the
periphery.
The central stroma appeared more reasonably smooth,
with subjective grading of the resulting central bed falling
between smooth and rough, at an average score of 2.6. This
was similar to that found in previous studies (average score of
2.8) when cutting less deep (400 mm) within the stroma with
a 60-kHz laser.11 As noted previously, this irregular-resulting
stroma may be a result of the loosely arranged collagen deep
within the cornea.12 If loosely arranged posterior collagen
results in a rougher stromal bed, then it might also be
expected that a rougher stromal surface would occur when
using the “double-pass” microkeratome technique for the creation of ultrathin tissue, as developed by Busin.13 To our
knowledge, though, there has been no study reporting on an
electron microscopy evaluation of the stromal bed quality
created with the double-pass technique. Therefore, at this
2012 Lippincott Williams & Wilkins
Ultrathin DSAEK Tissue With Femtosecond Laser
time, a comparison cannot be made between the stromal
bed quality created with the double-pass technique and our
current technique.
Alternatively, the irregular stroma may be a result of the
increased scatter and attenuation of laser efficacy at this
greater depth in edematous corneas.12 A double cut or triple
cut with the femtosecond laser has been demonstrated to
result in a smoother stromal bed when cutting standard
DSAEK tissue at depths of 350 mm with a 40-kHz laser.14
This double-cut technique was also demonstrated in another
recent study to improve stromal bed quality with a 60- and
150-kHz laser. In this recent study, it was demonstrated that
a more ideal stromal bed quality was achieved with a doublelayer profile technique that involved first cutting at a depth of
350 mm, followed by a second cut at a depth of 150 mm. The
authors argued that this result was achieved by avoiding the
diffraction and optical aberrations occurring when cutting
deep (400–500 mm).15 Such double-cut and double-layer profile techniques were not attempted in the current study and
may improve stromal quality while potentially not increasing
endothelial damage. Further studies would be required to
validate or refute this theory.
In conclusion, the use of the LPEHF laser to create
ultrathin DSAEK tissue holds promise, as seen in our study,
that this laser can reliably create ultrathin DSAEK tissue
without causing significant damage to the delicate endothelium. Modification of energy settings and cutting techniques
may improve the final stromal bed quality, which should be
the focus of future studies.
REFERENCES
1. Bahar I, Kaiserman I, McAllum P, et al. Comparison of posterior lamellar
keratoplasty techniques to penetrating keratoplasty. Ophthalmology.
2008;115:1525–1533.
2. Chen ES, Terry MA, Shamie N, et al. Descemet-stripping automated
endothelial keratoplasty: six-month results in a prospective study of
100 eyes. Cornea. 2008;27:514–520.
3. Price MO, Baig KM, Brubaker JW, et al. Randomized, prospective
comparison of precut vs surgeon-dissected grafts for Descemet stripping automated endothelial keratoplasty. Am J Ophthalmol. 2008;
146:36–41.
4. Price MO, Price FW Jr. Descemet’s stripping with endothelial
keratoplasty: comparative outcomes with microkeratome-dissected
and manually dissected donor tissue. Ophthalmology. 2006;113:
1936–1942.
5. Sikder S, Snyder RW. Femtosecond laser preparation of donor tissue
from the endothelial side. Cornea. 2006;25:416–422.
6. Sarayba MA, Ignacio TS, Binder PS, et al. Comparative study
of stromal bed quality by using mechanical, IntraLase
femtosecond laser 15- and 30-kHz microkeratomes. Cornea. 2007;
26:446–451.
7. Jones YJ, Goins KM, Sutphin JE, et al. Comparison of the femtosecond
laser (IntraLase) versus manual microkeratome (Moria ALTK) in dissection of the donor in endothelial keratoplasty: initial study in eye bank
eyes. Cornea. 2008;27:88–93.
8. Mootha VV, Heck E, Verity SM, et al. Comparative study of Descemet
stripping automated endothelial keratoplasty donor preparation by Moria
CBm microkeratome, horizon microkeratome, and Intralase FS60.
Cornea. 2011;30:320–324.
9. Saad HA, Terry MA, Shamie N, et al. An easy and inexpensive
method for quantitative analysis of endothelial damage by using vital
dye staining and Adobe Photoshop software. Cornea. 2008;27:
818–824.
www.corneajrnl.com |
5
Phillips et al
10. Terry MA, Saad HA, Shamie N, et al. Endothelial keratoplasty: the
influence of insertion techniques and incision size on donor endothelial
survival. Cornea. 2009;28:24–31.
11. Cheng YY, Kang SJ, Grossniklaus HE, et al. Histologic evaluation of
human posterior lamellar discs for femtosecond laser Descemet’s stripping endothelial keratoplasty. Cornea. 2009;28:73–79.
12. Soong HK, Mian S, Abbasi O, et al. Femtosecond laser-assisted posterior
lamellar keratoplasty: initial studies of surgical technique in eye bank
eyes. Ophthalmology. 2005;112:44–49.
6
| www.corneajrnl.com
Cornea Volume 0, Number 0, Month 2012
13. Busin M. One-year results of the double-pass technique for ultrathin
DSAEK. Paper presented at: Cornea Society/EBAA Fall Symposium;
October 11, 2011; Orlando, FL.
14. Mehta JS, Parthasarthy A, Por YM, et al. Femtosecond laserassisted endothelial keratoplasty: a laboratory model. Cornea. 2008;
27:706–712.
15. Rousseau A, Bensalem A, Garnier V, et al. Interface quality of endothelial keratoplasty buttons obtained with optimised femtosecond laser settings. Br J Ophthalmol. 2012;96:122–127.
2012 Lippincott Williams & Wilkins