Free
Cornea  |   August 2014
Surface Metrology and 3-Dimensional Confocal Profiling of Femtosecond Laser and Mechanically Dissected Ultrathin Endothelial Lamellae
Author Affiliations & Notes
  • Mor M. Dickman
    Maastricht University Medical Center, University Eye Clinic Maastricht, Maastricht, The Netherlands
  • Marc P. F. H. L. van Maris
    Eindhoven University of Technology, Department of Mechanical Engineering, Multiscale Laboratory, Eindhoven, The Netherlands
  • Friso W. van Marion
    Euro Cornea Bank, Beverwijk, The Netherlands
  • Yvonne Schuchard
    Euro Cornea Bank, Beverwijk, The Netherlands
  • Petra Steijger-Vermaat
    Euro Cornea Bank, Beverwijk, The Netherlands
  • Frank J. H. M. van den Biggelaar
    Maastricht University Medical Center, University Eye Clinic Maastricht, Maastricht, The Netherlands
  • Tos T. J. M. Berendschot
    Maastricht University Medical Center, University Eye Clinic Maastricht, Maastricht, The Netherlands
  • Rudy M. M. A. Nuijts
    Maastricht University Medical Center, University Eye Clinic Maastricht, Maastricht, The Netherlands
  • Correspondence: Mor M. Dickman, Postbus 6202 AZ, Maastricht, The Netherlands;mor.dickman@mumc.nl
Investigative Ophthalmology & Visual Science August 2014, Vol.55, 5183-5190. doi:10.1167/iovs.14-14309
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Mor M. Dickman, Marc P. F. H. L. van Maris, Friso W. van Marion, Yvonne Schuchard, Petra Steijger-Vermaat, Frank J. H. M. van den Biggelaar, Tos T. J. M. Berendschot, Rudy M. M. A. Nuijts; Surface Metrology and 3-Dimensional Confocal Profiling of Femtosecond Laser and Mechanically Dissected Ultrathin Endothelial Lamellae. Invest. Ophthalmol. Vis. Sci. 2014;55(8):5183-5190. doi: 10.1167/iovs.14-14309.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose.: To determine the feasibility of confocal profiling in measuring surface roughness and obtaining 3-dimensional reconstructions of mechanically dissected and femtosecond (fs)-laser photodisrupted endothelial lamellae. To determine the predictability of single-pass dissection of ultrathin endothelial lamellae using a novel motor-driven linear microkeratome.

Methods.: Thirty (n = 30) human corneas were harvested using a motor-driven linear microkeratome (n = 20); a hand-driven rotatory microkeratome (n = 6); and a 60-kHz fs laser (n = 4). Surface roughness was measured using an optical profiler operated in confocal microscopy mode followed by environmental scanning-electron-microscopy.

Results.: Mean surface roughness for the fs laser, motor-driven linear microkeratome, and hand-driven rotatory microkeratome measured 1.90 ± 0.48 μm, 1.06 ± 0.42 μm, and 0.93 ± 0.25 μm, respectively. Femtosecond photodisrupted lamellae were significantly rougher than mechanically dissected lamellae (P < 0.001). Mean (±SD) cutting depth with the motor-driven linear microkeratome measured: 552 ± 11 μm (550-μm head); 505 ± 19 μm (550-μm head); 459 ± 19 μm (450-μm head); and 392 ± 20 μm (400-μm head).

Conclusions.: Confocal microscopy allows quantitative surface roughness analysis and 3-dimensional reconstruction of human corneal lamellae. Femtosecond-laser photodisruption at 60 kHz results in rougher surfaces compared with mechanical dissection. The motor-driven linear microkeratome allows single-pass dissection of ultrathin endothelial lamellae with a standard deviation ≤20 μm.

Introduction
Lamellar corneal dissection has become an indispensable tool in the corneal surgical armamentarium. In the field of corneal transplantation, endothelial keratoplasty has replaced the full thickness graft as the treatment of choice for endothelial dysfunction. 1 In refractive surgery, femtosecond (fs) laser technology introduced high standards of safety and predictability for flap creation in the anterior cornea. 2  
However, stromal irregularities in the donor-recipient interface may limit visual outcome after deep stromal dissection as in Descemet stripping automated endothelial keratoplasty (DSAEK). 3,4 The faster and better visual recovery following Descemet membrane endothelial keratoplasty (DMEK), in which the stromal portion of the donor is completely eliminated, supports the role of such irregularities in degrading optical quality. 5  
Recently, ultrathin (UT) DSAEK grafts (<100 μm) have been reported to achieve visual outcomes similar to DMEK with the benefits of greater ease of preparation and a lower dislocation rate, combining the advantages of both techniques. 6,7 Nevertheless, the unpredictability of graft thickness created by mechanical microkeratomes makes it difficult to harvest UT grafts without risking donor perforation. 
Surface roughness of the lamellar graft is one of the factors that may limit the postoperative quality of vision. Surface roughness measurements can help determine stromal bed quality in endothelial keratoplasty and elucidate the functional correlations between surface roughness, light scattering, optical aberrations, and graft adhesion. 
Choosing an appropriate technology for surface roughness measurement is of outmost importance. Imaging confocal microscopy enables fast, noncontact measurements of a wide variety of surfaces and materials. It does not require sample preparation and provides the best lateral resolution achievable with an optical profiler, enabling measurement of large areas with a 0.1 nm vertical resolution and a high measureable slope. 8  
In the current study, we apply confocal microscopy for measuring surface roughness and obtaining 3-dimensional reconstruction of UT endothelial lamellae, harvested by either mechanical microkeratome systems or an fs laser. 
Materials and Methods
Thirty human corneas unsuitable for transplantation were obtained from the Euro Cornea Bank (Beverwijk, The Netherlands). Corneas were only included in the study if rejected due to donor medical history or endothelial cell counts lower than 2200 cells/mm2. Mean ± SD endothelial cell density measured 1800 ± 200 cells/mm2. Corneas rejected due to stromal pathology were not included in the study. Consent for the use of tissue for research was obtained in accordance with the tenets of the declaration of Helsinki and Dutch legislation. Donor corneas were preserved according to conventional eye bank techniques. Following a short hypothermic storage (between recovery and arrival at the eye bank) corneoscleral buttons were dissected and stored in organ culture at 31°C for 20 days on average. To allow deturgescence, corneas were transferred to a transport medium consisting of modified minimal essential medium (MEM) (Biowest, France) supplemented with 6% Dextran (Sigma Aldrich, St. Louis, MO, USA) 24 hours prior to dissection. 
Posterior Corneal Lamellae Preparation
Posterior corneal lamellae were harvested using either a novel motor-driven linear microkeratome (SLc; Gebauer Medizintechnik GmbH, Neuhausen, Germany; n = 20); a hand-driven rotatory microkeratome (automated lamellar therapeutic keratoplasty [ALTK]; Moria SA, Antony, France; n = 6); or a commercial fs platform (Intralase FS60; Abbott Medical Optics, Inc., Irvine, CA, USA; n = 4). 
Corneoscleral buttons were mounted on an artificial anterior chamber (Gebauer Medizintechnik GmbH, n = 20; Moria SA, n = 6; Barron K20-2125 [Katena Products, Inc., Denville, NJ, USA], n = 4) pressurized to 65 mm Hg using an MEM infusion raised 95 cm above the chamber and clamped 15 cm from the entrance to the chamber. The epithelium was removed prior to dissection, eliminating differences among corneas induced by variation in epithelial swelling. 
Central corneal thickness was measured using Fourier domain anterior-segment optical-coherence-tomography (Casia SS-1000; Tomey Corp., Nagoya, Japan) guiding the choice of both microkeratome head and anterior side cut depth for mechanical dissection and fs photodisruption, respectively, aiming at the thinnest posterior lamellae possible without perforating the donor cornea. 
For mechanical dissection, the motor-driven linear microkeratome (Gebauer Medizintechnik GmbH) was equipped with either 400 μm (n = 5), 450 μm (n = 5), 500 μm (n = 5), or 550 μm (n = 5) heads and the hand-driven rotatory microkeratome (Moria SA) was equipped with either the 350 μm (n = 3) or the 400 μm (n = 3) heads. A new blade was used for each donor cornea. Maximal care was taken to maintain a slow movement of the hand-driven microkeratome (Moria SA), requiring 4 to 6 seconds per dissection. 
For fs photodisruption, applanation was achieved using the disposable flat interface lens of the system without need for a suction ring. The laser beam wavelength was 1053 nm, repetition rate was set to 60 kHz and a raster pattern was used. Anterior side cuts were performed using the following settings: diameter = 9.0 mm, side cut angle = 90°, side cut energy = 2.7 μJ, and laser spot and line separation 3/3 μm. Full lamellar cut was achieved using four combinations of pulse energy, spot/line separation, and depth: (2.2 μJ, 5/5 μm, 450 μm depth); (1.8 μJ, 5/5 μm, 450 μm depth); (1.5 μJ, 4/4 μm, 500 μm depth); and (1.2 μJ, 4/4 μm, 500 μm depth). The cut diameter was 9.2 mm, oversized by 0.1 mm compared with the side cut on both sides. The tightest spot and line separations available with the fs 60-kHz platform used were chosen in an attempt to create a smooth donor bed at the intended cut depth with easy dissection of the anterior from the posterior stroma. 
Surface Roughness Measurements
Following dissection, posterior lamellae were fixed in 1.25% glutaraldehyde and 0.1% formaldehyde in 0.08 M sodium cacodylate buffer (pH = 7.4). Surface roughness was quantified from surface height profiles, measured using an optical profiler (Sensofar PLu 2300; Sensofar-Tech, SL, Terrassa, Spain) operated in confocal mode with an integrated 470-nm, light-emitting diode (LED). Measurements were performed in a temperature controlled cabinet (25 ± 0.2°C) set in a temperature-controlled laboratory (21 ± 0.5°C). 
Prior to scanning, the stromal surface of the corneal buttons was gently dried with air to prevent interaction of multiple focal points at the same in-plane coordinates. To prevent sample dehydration, the step size of the z-axis was doubled to 400 nm with a z-resolution of 4 nm. The center of each lamella was first located by a single coarse measurement using a ×10 magnification confocal objective followed by surface height profile measurements using the ×50 magnification confocal objective (numerical aperture = 0.8) with a mean acquisition time of 45 seconds. Intensity threshold was set to 10% to avoid unreliable data points. To rule out the possibility of focus artifacts (interference) in the semitransparent tissue material, two measurements with different depth ranges at the same position were subtracted, showing a negligible difference. Per cornea button, surface height profile was determined in up to eight locations. At each location, an area of 254.64 × 190.90 μm2 was measured, covering in total 8 × 254.64 × 190.90 μm2, ≈ 0.4-mm2 of the central cornea, minimizing the probability of spurious measurements. 
Data analysis was performed using technical computing software (MATLAB; MathWorks, Natick, MA, USA). The center of mass algorithm was used for precise localization of metrological data. Images were leveled by fitting the measured data against a second-order polynomial and subtracting the best fit from the image, minimizing the effects of corneal curvature and slope. 
Following image leveling, the roughness root mean square (Rrms ) was calculated according to the following formula:  where is the average of Z (height) values within the given area, Zn is a specific Z value, and N is the number of data points within the given area, representing the standard deviation of the roughness data (Fig. 1).  
Figure 1
 
Surface profile of the central 254.64 × 190.90 μm2 of an ultrathin (80 μm) endothelial lamella dissected with the motor-driven linear microkeratome (Gebauer Medizintechnik GmbH) equipped with a 500-μm head. Measurement was performed with the optical profiler (Sensofar-Tech, SL) operated in confocal mode. Surface height, represented by the color map, was used to calculate surface roughness. Uniform height profiles are represented by uniform colors, corresponding with smoother surfaces.
Figure 1
 
Surface profile of the central 254.64 × 190.90 μm2 of an ultrathin (80 μm) endothelial lamella dissected with the motor-driven linear microkeratome (Gebauer Medizintechnik GmbH) equipped with a 500-μm head. Measurement was performed with the optical profiler (Sensofar-Tech, SL) operated in confocal mode. Surface height, represented by the color map, was used to calculate surface roughness. Uniform height profiles are represented by uniform colors, corresponding with smoother surfaces.
To characterize different length scales, each image was further split into nine equal areas, releveled, and Rrms was calculated again (Fig. 2). Finally, 3-dimensional surface reconstructions were obtained, allowing a more complete representation of surface topography. 
Figure 2
 
Image splitting and releveling of the surface profile shown in Figure 1. To characterize different scale lengths, each image was split in nine areas, releveled, and surface roughness was calculated again. Local height variations in adjacent areas demonstrate the importance of image splitting in reducing the influence of extreme height values on calculated surface roughness. Image releveling was performed to better account for surface slope in a curved structure as the cornea.
Figure 2
 
Image splitting and releveling of the surface profile shown in Figure 1. To characterize different scale lengths, each image was split in nine areas, releveled, and surface roughness was calculated again. Local height variations in adjacent areas demonstrate the importance of image splitting in reducing the influence of extreme height values on calculated surface roughness. Image releveling was performed to better account for surface slope in a curved structure as the cornea.
Environmental Scanning Electron Microscopy
Following confocal microscopy, all samples were examined using environmental scanning electron microscopes (ESEMs, FEI-Quanta 600F; FEI Company, Hillsboro, OR, USA) in low vacuum mode (0.6 mbar). The environmental SEM (FEI Company) was chosen because of the ability to examine samples in natural state, without applying a conductive layer (e.g., gold). Simultaneous secondary-electron images and backscattered-electron images were acquired using a large-field-detector and a backscatter-electron detector, allowing qualitative surface texture analysis over a relatively large field of view at lower magnification. 9  
Statistical Analysis
Statistical analysis was performed using statistical software (SPSS for Windows, version 20.0; SPSS, Inc., Chicago, IL, USA). To quantify the difference in surface roughness between the three dissection techniques a linear mixed model analysis was performed with roughness as a dependent variable, cornea as a grouping factor, and device as a covariate. For all statistical tests performed, statistical significance was set at 0.05. 
Results
Posterior Corneal Lamellae Preparation
Mechanical dissection was successful in all cases but one, carried out with the hand-driven rotatory microkeratome (Moria SA) equipped with a 400-μm head, resulting in donor perforation despite generous safety margins. The cutting depths of the motor-driven linear microkeratome (Gebauer Medizintechnik GmbH) using four different gap widths are shown in the Table. A representative anterior-segment optical coherence tomography (OCT) image of a UT endothelial graft harvested with the motor-driven linear microkeratome (Gebauer Medizintechnik GmbH) is shown in Figure 3. Complete fs photodisruption was achieved in all specimens and peeling the anterior cap from the posterior stroma was easily accomplished in all cases. 
Figure 3
 
Spectral domain anterior-segment OCT (SS-1000 Casia; Tomey Corp., Nagoya, Japan) image of a 72-μm thick endothelial graft harvested from a 630-μm thick donor cornea using the motor-driven linear microkeratome (Gebauer Medizintechnik GmbH) equipped with a 550-μm head.
Figure 3
 
Spectral domain anterior-segment OCT (SS-1000 Casia; Tomey Corp., Nagoya, Japan) image of a 72-μm thick endothelial graft harvested from a 630-μm thick donor cornea using the motor-driven linear microkeratome (Gebauer Medizintechnik GmbH) equipped with a 550-μm head.
Table
 
Cutting Depths of Motor-Driven Linear Microkeratome System
Table
 
Cutting Depths of Motor-Driven Linear Microkeratome System
Microkeratome Head, Nr 550-μm (5) Gap Width 500-μm (5) Gap Width 450-μm (5) Gap Width 400-μm (5) Gap Width
Mean, μm 552 505 459 392
SD, μm 11 19 19 20
Range, μm 542–567 480–530 430–495 368–495
Surface Roughness Measurements
Prior to image splitting average Rrms (±SD) measured 3.65 ± 1.43 μm (Abbott Medical Optics, Inc.), 2.87 ± 1.13 μm (Moria SA), and 2.18 ± 0.99 μm (Gebauer Medizintechnik GmbH). Surface roughness differed significantly between photodisrupted lamellae and lamellae harvested with the motor-driven linear microkeratome (Gebauer Medizintechnik GmbH; P = 0.004). There was no significant difference in surface roughness between photodisrupted lamellae and lamellae harvested with the hand-driven rotatory microkeratome (Moria SA; P = 0.116) or between lamellae harvested mechanically using the hand-driven rotatory microkeratome (Moria SA) or the motor-driven linear microkeratome (Gebauer Medizintechnik GmbH; P = 0.105; Fig. 4). 
Figure 4
 
Surface roughness distribution following single image leveling and image splitting and releveling per dissection group. Following single image leveling (blue) average Rrms (±SD) measured 3.65 ± 1.43 μm (Abbott Medical Optics, Inc.); 2.87 ± 1.13 μm (Moria SA); and 2.18 ± 0.99 μm (Gebauer Medizintechnik GmbH), differing significantly only between photodisrupted lamellae and lamellae dissected using the motor-driven linear microkeratome (Gebauer Medizintechnik GmbH; P = 0.004). Following image splitting and releveling (green), average Rrms (±SD) measured 1.90 ± 0.48 μm (Abbott Medical Optics, Inc.), 1.06 ± 0.42 μm (Moria SA) and 0.93 ± 0.25 μm (Gebauer Medizintechnik GmbH), differing significantly between photodisrupted lamellae and mechanically dissected lamellae with both mechanical microkeratome systems (P < 0.001).
Figure 4
 
Surface roughness distribution following single image leveling and image splitting and releveling per dissection group. Following single image leveling (blue) average Rrms (±SD) measured 3.65 ± 1.43 μm (Abbott Medical Optics, Inc.); 2.87 ± 1.13 μm (Moria SA); and 2.18 ± 0.99 μm (Gebauer Medizintechnik GmbH), differing significantly only between photodisrupted lamellae and lamellae dissected using the motor-driven linear microkeratome (Gebauer Medizintechnik GmbH; P = 0.004). Following image splitting and releveling (green), average Rrms (±SD) measured 1.90 ± 0.48 μm (Abbott Medical Optics, Inc.), 1.06 ± 0.42 μm (Moria SA) and 0.93 ± 0.25 μm (Gebauer Medizintechnik GmbH), differing significantly between photodisrupted lamellae and mechanically dissected lamellae with both mechanical microkeratome systems (P < 0.001).
Following image splitting and releveling, average Rrms measured 1.90 ± 0.48 μm (Abbott Medical Optics, Inc.), 1.06 ± 0.42 μm (Moria SA), and 0.93 ± 0.25 μm (Gebauer Medizintechnik GmbH). There was a significant difference in surface roughness between photodisrupted lamellae and lamellae harvested with both the hand-driven rotatory microkeratome (Moria SA) and the motor-driven linear microkeratome (Gebauer Medizintechnik GmbH; P < 0.001). No significant difference in surface roughness was found between lamellae harvested with the hand-driven rotatory microkeratome (Moria SA) and the motor-driven linear microkeratome (Gebauer Medizintechnik GmbH; P = 0.293; Fig. 4). 
No correlation was found between graft thickness and surface roughness (r = −0.41, P = 0.31) among mechanically harvested lamellae. Similarly, surface roughness did not differ among lamellae harvested using different laser settings. 
Three Dimensional Surface Reconstructions and ESEM
High resolution 3-dimensional surface reconstructions were successfully obtained from all specimens of photodisrupted and mechanically harvested lamellae (Fig. 5). Photodisrupted stromal surfaces demonstrated crater-like features of variable dimensions surrounded by cauliflower-like agglomerations most likely corresponding to cavitation bubbles surrounded by coagulated collagen fibers. Mechanically harvested stromal surfaces demonstrated relatively even surfaces with an interwoven pattern, representing collagen lamellae which became more widely spaced in thinner endothelial lamellae. Complimentary ESEM images corresponded well with the confocal reconstructions, confirming the morphological patterns observed by confocal profiling in all cases (Fig. 5). 
Figure 5
 
Representative ESEM images (upper and middle rows) and 3-dimensional surface reconstructions (lower row) demonstrating the correspondence between imaging modalities. Stucco-like collagen irregularities on the stromal bed of a photodisrupted endothelial lamella (A) Intralase FS 60 kHz using 1.8-μm pulse energy and 5-μm spot and line separation at 450-μm depth demonstrate the irregularity of the photodisrupted surface compared with mechanically dissected stromal beds, showing similar morphologies. (B) Motor-driven linear microkeratome (Gebauer Medizintechnik GmbH) equipped with a 500-μm head. (C) Hand-driven rotatory microkeratome (Moria SA) equipped with a 350-μm head.
Figure 5
 
Representative ESEM images (upper and middle rows) and 3-dimensional surface reconstructions (lower row) demonstrating the correspondence between imaging modalities. Stucco-like collagen irregularities on the stromal bed of a photodisrupted endothelial lamella (A) Intralase FS 60 kHz using 1.8-μm pulse energy and 5-μm spot and line separation at 450-μm depth demonstrate the irregularity of the photodisrupted surface compared with mechanically dissected stromal beds, showing similar morphologies. (B) Motor-driven linear microkeratome (Gebauer Medizintechnik GmbH) equipped with a 500-μm head. (C) Hand-driven rotatory microkeratome (Moria SA) equipped with a 350-μm head.
Discussion
This study demonstrated the feasibility of confocal microscopy in measuring surface roughness and obtaining 3-dimensional reconstructions of corneal lamellae. 
When considering different lamellar dissection techniques it is important to take into account the impact of anatomical differences between the anterior and posterior cornea. 10,11 Compared with the anterior stroma, the posterior stroma is composed of more heavily hydrated, less densely packed, orthogonally arranged collagen lamellae with minimal interweaving centrally. 12 These differences account for greater posterior stromal swelling during organ culture and explain the difficulties associated with dissection and manipulation of UT endothelial lamellae. The relationship between collagen lamellae arrangement and light scattering also provide a possible explanation for the different visual outcomes following transplantation of grafts with variable thicknesses. 13  
Surface roughness, defined as the standard deviation of surface elevation, has emerged as an important parameter in evaluating stromal bed quality for endothelial keratoplasty. 1420 However, standards for measuring stromal bed roughness have not yet been defined, precluding accurate comparison of dissection techniques and determination of functional correlations. Choosing the most appropriate technique depends on multiple factors, including the optical quality of the surface, the scale of desired measurements, the need for sample preparation and acceptable measurement accuracy. 21  
Previous studies on stromal bed quality analyzed SEM images, a 2-dimensional imaging modality requiring sample preparation distorting surface characteristics. 14,18,2224 However, the grayscale of SEM images reflects the amount of secondary electrons detected rather than surface height, essential for surface roughness measurement. 25  
Atomic force microscopy (AFM), a scanning probe technique, allows quantitative 3-dimensional measurements with sub-tenth–nanometer axial resolution. 26 However, AFM's lateral resolution is limited by the interaction of the probe tip in contact with the surface, making it difficult to measure large areas. Moreover, for samples with slopes of a few degrees or more, AFM loses its advantage due to simultaneous interaction of peaks and valleys in the surface with the probe tip, a phenomenon known as dilation. 21  
Imaging confocal microscopy is a well-known technology for 3-dimensional surface topography measurements. 8 Its basic principle relies on storing a sequence of confocal images taken from different z-axis planes along the depth of focus of the microscope's objective. The highest signal within the images of the sequence for each pixel relates with the height position of the topography. By locating the z-axis position of the maximum of the axial response for each pixel, the 3-dimensional surface is reconstructed. 27  
Confocal microscopy allows performing measurements under physiological conditions without destructive sample preparation. It also offers the highest lateral resolution that can be achieved by an optical instrument, allowing imaging of large areas. 27 In this study, the central ∼0.4-mm2 of the lamellae was measured, 1300 to 1600 times larger compared with previous studies using SEM-stereography and AFM. 15,16,20 Confocal objectives with high numerical apertures allow measurement of slopes with a vertical step close to 90° for optically rough surfaces. 27 The importance of sample slope is demonstrated by different results obtained prior to and following image segmentation and re-leveling (Figures 2 and 4). 
Complimentary ESEM imaging assist in guiding the choice of segmentation algorithm and confirming the accuracy of the 3-dimensional reconstructions (Fig. 5). 
Interestingly, despite macrostructural morphological differences, surface roughness of mechanically dissected lamellae was not related to graft thickness and did not differ between both microkeratome systems, emphasizing the importance of quantitative surface roughness analysis. This finding contradicted our working hypothesis, based on the anatomical differences between the anterior and posterior stroma, but is consistent with a similar dislocation rate reported following UT-DSAEK compared with standard DSAEK, 4,6 suggesting that factors other than surface roughness are responsible for the superior visual outcomes following UT-DSAEK and DMEK. These results should however be interpreted with caution due to the narrow range of graft thicknesses in our study. 
Compared with mechanical dissection, 60-kHz fs photodisruption resulted in significantly rougher stromal surfaces. While the less densely packed, isotropic arrangement of collagen lamellae in the posterior stroma predicts a lower disruption threshold compared with the anterior stroma, light scatter in edematous (donor) corneas requires considerable increase in beam energy, expected to result in rougher stromal surfaces. 28 The lack of correlation between surface roughness and laser settings in our study may be the result of a smaller sample size and limited range of laser settings in the fs laser group. Alternatively, it is possible that in the range of pulse energies used in the cutting process was governed by mechanical forces applied by expanding cavitation bubbles, irrespective of energy levels. 
These findings are consistent with previous studies reporting increased surface roughness of photodisrupted endothelial lamellae compared with mechanical dissection and inferior clinical outcomes following fs-assisted endothelial keratoplasty. 2832 They are also consistent with recent studies demonstrating elevated surface roughness of photodisrupted endothelial lamellae despite decreasing energy levels, below those used in our study. 15,20 Newer fs laser platforms allowing even tighter spot and line separations are currently available, which could improve surface roughness of photodisrupted endothelial lamellae. 33  
Several other modifications have been suggested to overcome these limitations. These include fs lasers with a liquid-filled interface that eliminate applanation induced folds and decrease laser beam defocus, OCT-guided photodisruption from the endothelial side, corneal deswelling prior to fs photodisruption and the use of higher wavelengths (in the order of 1650 nm) fs lasers. 28,29,34  
This study also showed that single-pass dissection using a motor-driven linear microkeratome system provides a reproducible alternative for harvesting UT endothelial lamellae. The major limitation to mechanical dissection of UT lamellae is graft thickness unpredictability, in particular when using heads with wider gap widths. 35 A double-pass technique has been suggested to overcome this obstacle. 36 However, it results in a high rate of donor perforation and increased endothelial damage. 37,38  
Cutting depth unpredictability of mechanical microkeratomes results in part from gradual corneal applanation during dissection. Applanation of only the fraction of the cornea to be cut next requires gradual increase in the force needed to achieve applanation. Dissection accuracy in our study was independent of microkeratome gap width and donor pachymetry. We believe this can be explained by the design of the cutting head of the SLc system, which applanates the entire cornea prior to dissection. Motor driven dissection also eliminates the influence of variable transition time on dissection accuracy 39 and an inbuilt applanation plate allows predetermination of graft diameter, ensuring maximal endothelial cell delivery. 
Despite its high reproducibility, the motor-driven linear microkeratome (Gebauer Medizintechnik GmbH) does not eliminate the meniscus shape graft thickness profile, which has been shown to result in a postoperative hyperopic shift and increased corneal higher-order aberrations. 40,41 This is due to the steep blade entry angle necessary for deep stromal dissection from the epithelial side. 
Mechanical dissection using the endothelial side as reference has the potential to eliminate this profile and should be the subject of future research. 
Our study had several limitations. Corneas used were unsuitable for transplantation, primarily because of a low endothelial cell density. However, since the lowest endothelial cell density was still above 1500 cells/mm2, we believe the endothelial cell barrier function and corneal hydration were still warranted in these discarded corneas. In addition, special care was taken to exclude corneas with evident stromal pathology that may have led to changes in stromal hydration and architecture. Although occasional intrinsic differences in donor characteristics may have influenced our measurements, we believe that random allocation of donor corneas to the dissection groups mitigated inadvertent effects. 
Organ culture, deswelling, and subsequent epithelial removal may also have influenced our measurements. Corneal swelling during organ culture occurs primarily through the posterior cornea and may reach twice the original thickness, requiring addition of Dextran to the transport medium prior to dissection. 10,42 Focal variations in stromal hydration are expected to influence mechanical dissection, increase light scatter, and interfere with photodisruption. However, Muller et al. 43 showed that changes in keratocytes and collagen fiber arrangement induced by organ culture can be fully reversed with Dextran. In line with these findings, Rossi et al. 29 recently demonstrated the protective effects of deswelling on surface texture after a corneal FS cut. The corneal epithelium is often sloughed off partly or completely during organ culture and was completely removed prior to deep stromal dissection simulating the clinical setting where it is standardly removed prior to dissection. This measure is expected to eliminate thickness differences induced by variations in epithelial swelling. 
Since the artificial anterior chamber represents an integral component of both microkeratome systems, comparing the systems using the same chamber was not possible. 
To overcome this limitation, chambers were similarly pressurized to 65 mm Hg. Statistical analysis using device as covariate was chosen to account for differences in dissection techniques, including the different chambers used. 
In conclusion, the current study provided a proof of concept for the use of confocal microscopy for quantitative topographical measurements and 3-dimensional surface reconstructions of corneal lamellae. An innovative microkeratome technology allowing reproducible single-pass mechanical dissection of UT endothelial lamellae was used to demonstrate the value of this novel surface metrology modality. 
Acknowledgments
Supported by a grant from ZonMw, The Netherlands Organization for Health Research and Development. 
Disclosure: M.M. Dickman, None; M.P.F.H.L. van Maris, None; F.W. van Marion, None; Y. Schuchard, None; P. Steijger-Vermaat, None; F.J.H.M. van den Biggelaar, None; T.T.J.M. Berendschot, None; R.M.M.A. Nuijts, Gebauer Medizintechnik GmbH (C) 
References
Eye Bank Association of America (EBAA). 2012 Eye Banking Statistical Report . Washington, DC: EBAA; 2012: 17–18.
Farjo AA Sugar A Schallhorn SC Femtosecond lasers for LASIK flap creation: a report by the American Academy of Ophthalmology. Ophthalmology . 2013; 120: e5–e20. [CrossRef] [PubMed]
Anshu A Price MO Tan DT Price FW Jr. Endothelial keratoplasty: a revolution in evolution. Surv Ophthalmol . 2012; 57: 236–252. [CrossRef] [PubMed]
Lee WB Jacobs DS Musch DC Kaufman SC Reinhart WJ Shtein RM. Descemet's stripping endothelial keratoplasty: safety and outcomes: a report by the American Academy of Ophthalmology. Ophthalmology . 2009; 116: 1818–1830. [CrossRef] [PubMed]
Price MO Price FW Jr. Descemet's membrane endothelial keratoplasty surgery: update on the evidence and hurdles to acceptance. Curr Opin Ophthalmol . 2013; 24: 329–335. [CrossRef] [PubMed]
Busin M Madi S Santorum P Scorcia V Beltz J. Ultrathin Descemet's stripping automated endothelial keratoplasty with the microkeratome double-pass technique: two-year outcomes. Ophthalmology . 2013; 120: 1186–1194. [CrossRef] [PubMed]
Neff KD Biber JM Holland EJ. Comparison of central corneal graft thickness to visual acuity outcomes in endothelial keratoplasty. Cornea . 2011; 30: 388–391. [CrossRef] [PubMed]
Artigas R Laguarta F Cadevall C. Dual-technology optical sensor head for 3D surface shape measurements on the micro and nano-scales. Proc SPIE . 2004; 5457: 166–174.
Danilatos GD. Introduction to the ESEM instrument. Microsc Res Tech . 1993; 25: 354–361. [CrossRef] [PubMed]
Cristol SM Edelhauser HF Lynn MJ. A comparison of corneal stromal edema induced from the anterior or the posterior surface. Refract Corneal Surg . 1992; 8: 224–229. [PubMed]
Jester JV Murphy CJ Winkler M Lessons in corneal structure and mechanics to guide the corneal surgeon. Ophthalmology . 2013; 120: 1715–1717. [CrossRef] [PubMed]
Meek KM Leonard DW Connon CJ Dennis S Khan S. Transparency, swelling and scarring in the corneal stroma. Eye (Lond) . 2003; 17: 927–936. [CrossRef] [PubMed]
Freund DE McCally RL Farrell RA Cristol SM L'Hernault NL Edelhauser HF. Ultrastructure in anterior and posterior stroma of perfused human and rabbit corneas. Relation to transparency. Invest Ophthalmol Vis Sci . 1995; 36: 1508–1523. [PubMed]
Cheng YY Kang SJ Grossniklaus HE Histologic evaluation of human posterior lamellar discs for femtosecond laser Descemet's stripping endothelial keratoplasty. Cornea . 2009; 28: 73–79. [CrossRef] [PubMed]
Lombardo M De Santo MP Lombardo G Surface quality of femtosecond dissected posterior human corneal stroma investigated with atomic force microscopy. Cornea . 2012; 31: 1369–1375. [CrossRef] [PubMed]
Marian A Nada O Legare F Smoothness assessment of corneal stromal surfaces. J Cataract Refract Surg . 2013; 39: 118–127. [CrossRef] [PubMed]
Monterosso C Fasolo A Caretti L Monterosso G Buratto L Bohm E. Sixty-kilohertz femtosecond laser-assisted endothelial keratoplasty: clinical results and stromal bed quality evaluation. Cornea . 2011; 30: 189–193. [CrossRef] [PubMed]
Mootha VV Heck E Verity SM Comparative study of Descemet stripping automated endothelial keratoplasty donor preparation by Moria CBm microkeratome, horizon microkeratome, and Intralase FS60. Cornea . 2011; 30: 320–324. [PubMed]
Phillips PM Phillips LJ Saad HA “Ultrathin” DSAEK tissue prepared with a low-pulse energy, high-frequency femtosecond laser. Cornea . 2013; 32: 81–86. [CrossRef] [PubMed]
Ziebarth NM Dias J Hurmeric V Quality of corneal lamellar cuts quantified using atomic force microscopy. J Cataract Refract Surg . 2013; 39: 110–117. [CrossRef] [PubMed]
Vorburger TV Rhee HG Renegar TB Song JF Zheng A. Comparison of optical and stylus methods for measurement of surface texture. Int J Adv Manuf Tech . 2007; 33: 110–118. [CrossRef]
Phillips PM Phillips LJ Saad HA “Ultrathin” DSAEK Tissue Prepared With a Low-Pulse Energy, High-Frequency Femtosecond Laser. Cornea . 2012; 32: 81–86. [CrossRef]
Hammer T Hanschke R Worner I Wilhelm FW. Evaluation of four microkeratome models: quality and reproducibility of cut edge and cut surface as determined by scanning electron microscopy. J Refract Surg . 2005; 21: 454–462. [PubMed]
Viestenz A Langenbucher A Hofmann-Rummelt C Modis L Viestenz A Seitz B. Evaluation of corneal flap dimensions and cut quality using the SKBM automated microkeratome. J Cataract Refract Surg . 2003; 29: 825–831. [CrossRef] [PubMed]
Laudate T. Imaging guidelines for scanning electron microscopy. Adv Mater Process . 2003; 161: 23–25.
Last JA Russell P Nealey PF Murphy CJ. The Applications of Atomic Force Microscopy to Vision Science. Invest Ophthalmol Vis Sci . 2010; 51: 6083–6094. [CrossRef] [PubMed]
Artigas R. Imaging confocal microscopy. In: Leach R ed. Optical Measurement of Surface Topography . Berlin: Springer-Verlag; 2011; 237–286.
Crotti C Deloison F Alahyane F Wavelength optimization in femtosecond laser corneal surgery. Invest Ophthalmol Vis Sci . 2013; 54: 3340–3349. [CrossRef] [PubMed]
Rossi M Misto R Gatto C Garimoldi P Campanelli M D'Amato Tóthová J. Protective effects of deswelling on stromal collagen denaturation after a corneal femtosecond laser cut. Invest Ophthalmol Vis Sci . 2013; 54: 4148–4157. [CrossRef] [PubMed]
Cheng YY Schouten JS Tahzib NG Efficacy and safety of femtosecond laser-assisted corneal endothelial keratoplasty: a randomized multicenter clinical trial. Transplantation . 2009; 88: 1294–1302. [CrossRef] [PubMed]
Heinzelmann S Maier P Bohringer D Auw-Hadrich C Reinhard T. Visual outcome and histological findings following femtosecond laser-assisted versus microkeratome-assisted DSAEK. Graefes Arch Clin Exp Ophthalmol . 2013; 251: 1979–1985. [CrossRef] [PubMed]
Vetter JM Butsch C Faust M Irregularity of the posterior corneal surface after curved interface femtosecond laser-assisted versus microkeratome-assisted Descemet stripping automated endothelial keratoplasty. Cornea . 2013; 32: 118–124. [CrossRef] [PubMed]
Monterosso C Galan A Bohm E Zampini A Parekh M Caretti L. Effect of 60 kHz and 150 kHz Femtosecond Lasers on Corneal Stromal Bed Surfaces: A Comparative Study. ISRN Ophthalmology . 2013; 2013: 971451. [CrossRef] [PubMed]
Talamo JH Gooding P Angeley D Optical patient interface in femtosecond laser-assisted cataract surgery: contact corneal applanation versus liquid immersion. J Cataract Refract Surg . 2013; 39: 501–510. [CrossRef] [PubMed]
Springs CL Joseph MA Odom JV Wiley LA. Predictability of donor lamellar graft diameter and thickness in an artificial anterior chamber system. Cornea . 2002; 21: 696–699. [CrossRef] [PubMed]
Busin M Patel AK Scorcia V Ponzin D. Microkeratome-assisted preparation of ultrathin grafts for Descemet stripping automated endothelial keratoplasty. Invest Ophthalmol Vis Sci . 2012; 53: 521–524. [CrossRef] [PubMed]
Sikder S Nordgren RN Neravetla SR Moshirfar M. Ultra-thin donor tissue preparation for endothelial keratoplasty with a double-pass microkeratome. Am J Ophthalmol . 2011; 152: 202–208.e202. [CrossRef] [PubMed]
Woodward MA Titus MS Shtein RM. Effect of microkeratome pass on tissue processing for Descemet stripping automated endothelial keratoplasty. Cornea . 2014; 33: 507–509. [CrossRef] [PubMed]
Bhogal MS Allan BD. Graft profile and thickness as a function of cut transition speed in Descemet-stripping automated endothelial keratoplasty. J Cataract Refract Surg . 2012; 38: 690–695. [CrossRef] [PubMed]
Dickman MM Cheng YY Berendschot TT van den Biggelaar FJ Nuijts RM. Effects of graft thickness and asymmetry on visual gain and aberrations after Descemet stripping automated endothelial keratoplasty. JAMA Ophthalmol . 2013; 131: 737–744. [CrossRef] [PubMed]
Dupps WJ Jr Qian Y Meisler DM. Multivariate model of refractive shift in Descemet-stripping automated endothelial keratoplasty. J Cataract Refract Surg . 2008; 34: 578–584. [CrossRef] [PubMed]
Van Horn DL Doughman DJ Harris JE Miller GE Lindstrom R Good RA. Ultrastructure of human organ-cultured cornea. II. Stroma and epithelium. Arch Ophthalmol . 1975; 93: 275–277. [CrossRef] [PubMed]
Muller LJ Pels E Vrensen GF. The effects of organ-culture on the density of keratocytes and collagen fibers in human corneas. Cornea . 2001; 20: 86–95. [CrossRef] [PubMed]
Figure 1
 
Surface profile of the central 254.64 × 190.90 μm2 of an ultrathin (80 μm) endothelial lamella dissected with the motor-driven linear microkeratome (Gebauer Medizintechnik GmbH) equipped with a 500-μm head. Measurement was performed with the optical profiler (Sensofar-Tech, SL) operated in confocal mode. Surface height, represented by the color map, was used to calculate surface roughness. Uniform height profiles are represented by uniform colors, corresponding with smoother surfaces.
Figure 1
 
Surface profile of the central 254.64 × 190.90 μm2 of an ultrathin (80 μm) endothelial lamella dissected with the motor-driven linear microkeratome (Gebauer Medizintechnik GmbH) equipped with a 500-μm head. Measurement was performed with the optical profiler (Sensofar-Tech, SL) operated in confocal mode. Surface height, represented by the color map, was used to calculate surface roughness. Uniform height profiles are represented by uniform colors, corresponding with smoother surfaces.
Figure 2
 
Image splitting and releveling of the surface profile shown in Figure 1. To characterize different scale lengths, each image was split in nine areas, releveled, and surface roughness was calculated again. Local height variations in adjacent areas demonstrate the importance of image splitting in reducing the influence of extreme height values on calculated surface roughness. Image releveling was performed to better account for surface slope in a curved structure as the cornea.
Figure 2
 
Image splitting and releveling of the surface profile shown in Figure 1. To characterize different scale lengths, each image was split in nine areas, releveled, and surface roughness was calculated again. Local height variations in adjacent areas demonstrate the importance of image splitting in reducing the influence of extreme height values on calculated surface roughness. Image releveling was performed to better account for surface slope in a curved structure as the cornea.
Figure 3
 
Spectral domain anterior-segment OCT (SS-1000 Casia; Tomey Corp., Nagoya, Japan) image of a 72-μm thick endothelial graft harvested from a 630-μm thick donor cornea using the motor-driven linear microkeratome (Gebauer Medizintechnik GmbH) equipped with a 550-μm head.
Figure 3
 
Spectral domain anterior-segment OCT (SS-1000 Casia; Tomey Corp., Nagoya, Japan) image of a 72-μm thick endothelial graft harvested from a 630-μm thick donor cornea using the motor-driven linear microkeratome (Gebauer Medizintechnik GmbH) equipped with a 550-μm head.
Figure 4
 
Surface roughness distribution following single image leveling and image splitting and releveling per dissection group. Following single image leveling (blue) average Rrms (±SD) measured 3.65 ± 1.43 μm (Abbott Medical Optics, Inc.); 2.87 ± 1.13 μm (Moria SA); and 2.18 ± 0.99 μm (Gebauer Medizintechnik GmbH), differing significantly only between photodisrupted lamellae and lamellae dissected using the motor-driven linear microkeratome (Gebauer Medizintechnik GmbH; P = 0.004). Following image splitting and releveling (green), average Rrms (±SD) measured 1.90 ± 0.48 μm (Abbott Medical Optics, Inc.), 1.06 ± 0.42 μm (Moria SA) and 0.93 ± 0.25 μm (Gebauer Medizintechnik GmbH), differing significantly between photodisrupted lamellae and mechanically dissected lamellae with both mechanical microkeratome systems (P < 0.001).
Figure 4
 
Surface roughness distribution following single image leveling and image splitting and releveling per dissection group. Following single image leveling (blue) average Rrms (±SD) measured 3.65 ± 1.43 μm (Abbott Medical Optics, Inc.); 2.87 ± 1.13 μm (Moria SA); and 2.18 ± 0.99 μm (Gebauer Medizintechnik GmbH), differing significantly only between photodisrupted lamellae and lamellae dissected using the motor-driven linear microkeratome (Gebauer Medizintechnik GmbH; P = 0.004). Following image splitting and releveling (green), average Rrms (±SD) measured 1.90 ± 0.48 μm (Abbott Medical Optics, Inc.), 1.06 ± 0.42 μm (Moria SA) and 0.93 ± 0.25 μm (Gebauer Medizintechnik GmbH), differing significantly between photodisrupted lamellae and mechanically dissected lamellae with both mechanical microkeratome systems (P < 0.001).
Figure 5
 
Representative ESEM images (upper and middle rows) and 3-dimensional surface reconstructions (lower row) demonstrating the correspondence between imaging modalities. Stucco-like collagen irregularities on the stromal bed of a photodisrupted endothelial lamella (A) Intralase FS 60 kHz using 1.8-μm pulse energy and 5-μm spot and line separation at 450-μm depth demonstrate the irregularity of the photodisrupted surface compared with mechanically dissected stromal beds, showing similar morphologies. (B) Motor-driven linear microkeratome (Gebauer Medizintechnik GmbH) equipped with a 500-μm head. (C) Hand-driven rotatory microkeratome (Moria SA) equipped with a 350-μm head.
Figure 5
 
Representative ESEM images (upper and middle rows) and 3-dimensional surface reconstructions (lower row) demonstrating the correspondence between imaging modalities. Stucco-like collagen irregularities on the stromal bed of a photodisrupted endothelial lamella (A) Intralase FS 60 kHz using 1.8-μm pulse energy and 5-μm spot and line separation at 450-μm depth demonstrate the irregularity of the photodisrupted surface compared with mechanically dissected stromal beds, showing similar morphologies. (B) Motor-driven linear microkeratome (Gebauer Medizintechnik GmbH) equipped with a 500-μm head. (C) Hand-driven rotatory microkeratome (Moria SA) equipped with a 350-μm head.
Table
 
Cutting Depths of Motor-Driven Linear Microkeratome System
Table
 
Cutting Depths of Motor-Driven Linear Microkeratome System
Microkeratome Head, Nr 550-μm (5) Gap Width 500-μm (5) Gap Width 450-μm (5) Gap Width 400-μm (5) Gap Width
Mean, μm 552 505 459 392
SD, μm 11 19 19 20
Range, μm 542–567 480–530 430–495 368–495
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×