Content area
Purpose
To evaluate wound strength for patient safety during transport and endothelial viability when partial and complete femtosecond laser-assisted keratoplasty (FLAK) incisions are made in cadaveric corneas.
Methods
19 human corneoscleral rims were divided into six groups, mounted on an anterior chamber maintainer and cut with a femtosecond laser programmed to the following patterns: 'zigzag' (A), 'mushroom' (B) and 'top hat' (C) in both full (1) and partial (2) thicknesses. The pressure required to produce leakage from the corneal incision was then measured. Eight additional corneas were cut with the 'zigzag' pattern: four full and four partial thickness, prepared and transported per standard eye bank protocol, and analysed for endothelial cell loss with trypan blue staining and digital image analysis.
Results
Mean leakage pressure in mm Hg for group A1 was 110 (SD 94); group A2, 1180 (SD 468); group B1, 978 (SD 445); group B2, 987 (SD 576); group C1, 710 (SD 474); group C2, 1290 (SD 231). There was a significant difference in leakage pressure between groups A1 and A2 (p=0.05), groups A1 and B1 (p=0.05), and groups A1 and C1 (p=0.05). Mean percentage endothelial damage after full-thickness cuts was 8.40 (SD 2.34) and 5.30 (SD 1.33) in partial-thickness cuts (p=0.11).
Conclusions
Partial thickness zigzag, top hat and mushroom-style partial FLAK incisions left an intact tissue wall with high resistance to rupture, whereas full-thickness cuts were more variable. Laser trephination and eye bank handling protocol for donor FLAK buttons leads to moderate peripheral endothelial cell loss in tissue with both complete and partial cuts.
Full text
Introduction
Penetrating keratoplasty has been the mainstay treatment of corneal disease for the past 100 years. 1 It has undergone a number of major changes over that time, including the development of the suction trephine, 2 a variety of suture strategies aimed at reducing astigmatism, 3 and combination procedures aimed at reducing the impact of future operations on the health of the donor corneal button. 4 Ever since Barraquer's landmark description of a 'Penetrating Keratoplasty in Two Planes', 5 the promise of a full-thickness graft without the inherent instability of a vertical edge to vertical edge interface has been waiting to be realised. There has been recent re-examination of these techniques by Busin and Affra, 6 7 and with the development of femtosecond laser-assisted keratoplasty (FLAK) techniques, 8 donor corneal button and host cornea can be cut in a variety of wound patterns to achieve a more stable fit. 9-12 The most common incisions patterns reported are zigzag, top hat and mushroom. 8 13 14
A practical consideration of this surgical technique has been that the femtosecond laser and the operating suite are often in different locations requiring that patients be moved from one to the other in order to complete the keratoplasty operation. Host cornea stability during this transportation is desired to avoid theoretical complications of premature wound dehiscence, including possible subsequent prolapse of intraocular tissue. It has been demonstrated that leaving some of the host tissue uncut by the laser adds strength and stability to the host cornea in the 'top hat' configuration. 15 In this report, McAllum and colleagues did not perform a direct comparison between different shapes of FLAK wounds with and without uncut tissue 'bridges'. Additionally, they did not study the strength of the most common tissue bridge used in large clinical reports: a posterior stromal bridge at the most posterior aspect of the laser incision. 13 16 Instead, they tested the resilience of tissues bridges left in the mid and anterior stroma. The purpose of the posterior tissue bridge design is to create an uninterrupted laser incision to allow for a smooth and well-matched graft-host interface in the anterior 85-90% of the cornea.
Evaluations of endothelial cell loss after FLAK wounds in porcine eyes have shown that gross endothelial cell damage can be observed in both partial thickness (100 µ posterior tissue bridge intact) and full-thickness cuts. 17 At this time, however, there are no quantitative evaluations of the effect of FLAK wounds on endothelial cell death in human tissue.
In this study, we performed a preliminary evaluation of the resistance to burst pressure across a variety of wound patterns, with full thickness and partial thickness (70 μ posterior tissue bridge left intact). Additionally we have used digital image analysis software to quantify the endothelial cell loss after both full-thickness and partial-thickness cuts in experimental eye bank corneas.
Materials and methods
Corneas
Nineteen human cadaveric corneoscleral rims from donors aged 47-72 years (Lions Visiongift, Portland, Oregon, USA), all within 1 week of harvest and preservation in Optisol GS (Bausch & Lomb Surgical, Irvine, California, USA) were divided into five groups of three and one group of four. Corneas with peripheral or central opacities were excluded from this study. Corneal thickness maps using OCT (RTVUE, Optovue, Fremont, California, USA) were generated with a single reliable scan to determine the thinnest pachymetry at 8.0-9.0 mm diameters centred on the cornea. Thicknesses were confirmed once corneas were mounted on the artificial chamber maintainer (Moria, Inc., Doylestown, Pennsylvania, USA). Pressure was maintained during applanation and laser incision by closure of three-way stopcock on the outflow tube.
Laser trephination
The corneas were cut by a 150 kHz Femtosecond Laser (IntraLase iFS Laser, Abbot Medical Optics, Santa Ana, California, USA) into the following six wound patterns as illustrated in figure 1 : (A1) full-thickness zigzag; (A2) partial-thickness (70 μ posterior tissue bridge left intact) zigzag; (B1) full-thickness mushroom; (B2) partial-thickness mushroom; (C1) full-thickness top hat; and (C2) partial-thickness top hat. The cuts were made to create an 8.5 mm (external diameter) corneal button. The full-thickness cuts were assured by setting the femtosecond laser to maximal cut depth of 1200 μ (exceeding all corneal thicknesses) for groups A1, B1 and C1 and were always confirmed by air bubbles in the anterior chamber. Posterior side-cut depth for groups A2, B2 and C2 were based on OCT pachymetry measurements done on the corneas at 8.0-9.0 mm diameters and ranged from 488 to 530 µm in order to leave a 70 µm tissue bridge. The power setting on the laser was 1.9 mJ in all groups for all cuts. All groups shared the following laser settings for the lamellar cut: lamellar cut depth (300 µm), tangential spot separation (2 µm) and radial spot separation (2 µm). The spiral start differed between the groups with the zigzag and top hat configurations in, and the mushroom out. The short diameter for the top hat and mushroom configurations was 6.9 mm while the zigzag was 7.4 mm. The anterior side cuts of mushroom and zigzag shared the same diameter (8.5 mm), posterior depth (330 µm), spot separation (5 µm) and layer separation (5 µm). The posterior side cuts of all groups shared the same spot separation (3 µm), layer separation (2 µm) and the top hat posterior side cut had a diameter of 8.5 mm The zigzag cut had anterior and posterior side-cut angles of 30° while the top hat and mushroom configurations had anterior and posterior side-cut angles of 90°. During laser applanation and trephination, no leaks were observed.
Figure 1Schematic demonstrating femtosecond laser cut patterns/dimensions/angles for zigzag full-thickness (A1) and partial-thickness (A2) cuts, mushroom full-thickness (B1) and partial-thickness (B2) cuts, and top hat full-thickness (C1) and partial-thickness (C2) cuts as described in 'Materials and methods' section.
Measurement of wound strength
Following laser trephination, the corneas were mounted on a Moria anterior chamber maintainer (Moria Surgical, Antony, France), which was then situated under a microscope (Zeiss Stemi 2000 dissecting scope, Carl Zeiss Microscopy, Thornwood, New York, USA) for direct visualisation. A normal saline solution infusion system was connected by using a three-way connector to the artificial anterior chamber and to two digital manometers (Heise-PTE1 HQS-30212, Stratford, CT (high-pressure ranges > 400 mm Hg) and Ohmic Instruments PPM-30, Easton, MD (low-pressure ranges 0-500 mm Hg)) with polypropylene tubing in parallel. Each cornea was then manually pressurised using a 25 cc syringe with slow infusion to the point of 'burst' or 'ooze', defined as the point where fluid leak was observed rapidly or slowly, respectively, by direct visualisation from the FLAK wound ( figure 2 demonstrates experimental set-up). The pressure measured by the manometers prior to wound failure was recorded as the 'burst' or 'ooze' pressure for each cornea.
Figure 2Photograph of the anterior chamber maintainer set-up with dual manometers and microscope, used to measure wound dehiscence after femtosecond laser trephination.
Four additional corneoscleral buttons prepared for endothelial cell death analysis were cut in the full-thickness zigzag pattern detailed earlier, and four were cut in the partial-thickness zigzag pattern with 70 μ residual posterior stromal bridge. These rims were then placed in Optisol GS, and shipped both to and from a remote site over a 24 h period to subject the tissue to standard eye bank handling protocol. Once returned, the rims were stained using trypan blue 0.25% (Sigma, St Louis, Missouri, USA) for 2 min to highlight the devitalised tissue. Tissues were rinsed briefly with balanced saline solution. Digital photographs were obtained using a Pro-Microscan Camera Model 310 (Oplenic Optronics Co., Ltd, Hangzhou, CN). These images were then cropped to only include a 9 mm diameter ring image centred on the ∼8.5 mm corneal FLAK wound. This circular image was then loaded into the FIJI image analysis software (open access, free software; website Fiji.sc), converted to 16-bit gray-scale images with fixed image widths (1000x666 pixels), and then a trainable segmentation process was employed to distinguish vital from devitalised tissue and create a binary image (see 'Results' section and figure 5 ). These images were then analysed using Photoshop C3 (Adobe Software, 2011) to determine total pixel counts, the pixel count of the 1.5 mm in width ring aligned over the FLAK wound and the 7.5 mm diameter central portion of the tissue. Percentage areas of cell death were then calculated with these values by dividing the number of dark pixels (Trypan stained) over the total number of pixels.
Statistics
The results were evaluated using the Mann-Whitney U test. p Values of less than 0.05 were considered statistically significant.
Results
Table 1 details the pressure required for each cornea to 'burst'. Burst refers to a frank dehiscence of the cut tissue. Ooze is a slow egress of fluid without a notable break at the cut. The designation of 'no burst' was used when the polypropylene tubing detached from its coupling prior to wound dehiscence, suggesting that wound strength exceeded maximum rim or tube coupling connection strength. Figure 3 represents this data graphically.
Table 1
Zigzag (A) |
|||||||
Wound type | Partial thickness (A2) | Full thickness (A1) |
|||||
Specimen ID | 1 | 2 | 3 | 1 | 2 | 3 | - |
Burst pressure (mm Hg) | 648 | 1391 | 1515 | 23 | 98 | 210 | - |
Notes | Ooze | No burst | No burst | Burst | Burst | Ooze | - |
Mean (SD) | 1180 (468) | 110 (94) |
|||||
Specimen ID | 1 | 2 | 3 | 1 | 2 | 3 | - |
Burst pressure (mm Hg) | 1463 | 408 | 1339 | 523 | 998 | 1412 | - |
Notes | Burst | Burst | No burst | Burst | Burst | No burst | - |
Mean (SD) | 1290 (231) | 710 (474) |
|||||
Specimen ID | 1 | 2 | 3 | 1 | 2 | 3 | 4 |
Burst pressure (mm Hg) | 1091 | 1546 | 1246 | 1215 | 140 | 640 | 275 |
Notes | Burst | No burst | No burst | Burst | Burst | Ooze | Burst |
Mean (SD) | 987 (576) | 978 (445) |
Figure 3Mean burst pressure for zigzag, mushroom and top hat configurations. Error bars represent SD.
Mean leakage pressure ( table 1 ) in mm Hg for the full-thickness zigzag group was 110 (SD 94); 1180 (SD 468) for the partial-thickness zigzag group; 978 (SD 445) for the full-thickness top hat group; 987 (SD 576) for the partial-thickness top hat group; 710 (SD 474) for the full-thickness mushroom group; and 1290 (SD 231) for the partial-thickness mushroom group. There was a borderline statistically significant difference in leakage pressure between groups A1 and A2 (p=0.05), between groups A1 and B1 (p=0.05) and between groups A1 and C1 (p=0.05).
Table 2 details the endothelial cell loss after both full-thickness and partial-thickness zigzag FLAK wounds. Figure 4 represents this data graphically. Mean percentage endothelial cell damage after full-thickness cuts was 8.40 (SD 2.34) and 5.30 (SD 1.33) in partial-thickness cuts. Mean percentage endothelial cell damage in the 1.5 mm ring centred over the laser cut was 17.78 (SD 8.25) in the full-thickness cuts compared with 9.17 (SD 2.23) (p=0.11) in partial-thickness cuts. Mean percentage endothelial cell damage in the 7.5 mm inner diameter centre circle was 3.22 (SD 1.05) in the full-thickness cuts compared with 3.43 (SD 0.95) in partial-thickness cuts (p=0.88). Mean percentage endothelial cell damage in the entire full-thickness cuts was 8.40 (SD 2.34) and 5.30 (SD 1.33) in partial-thickness cuts (p=0.11). A representative sample of the stained tissue and digital rendering of cell death in corneal buttons is shown in figure 5 .
Table 2
Full thickness |
|||
Tissue | Cell death, % |
||
Inner circle | Outer ring | Total |
|
1 | 4.35 | 10.04 | 6.45 |
2 | 2.99 | 26.56 | 11.21 |
3 | 3.62 | 11.48 | 6.50 |
4 | 1.91 | 23.04 | 9.44 |
Mean (SD) | 3.22 (1.03) | 17.8 (8.25) | 8.40 (2.34) |
1 | 2.51 | 7.96 | 4.31 |
2 | 3.01 | 6.68 | 4.17 |
3 | 3.46 | 11.46 | 5.72 |
4 | 4.74 | 10.58 | 7.00 |
Mean (SD) | 3.43 (0.95) | 9.17 (2.23) | 5.30 (1.33) |
p Value (two-tailed) | 0.78 | 0.11 | 0.11 |
Figure 4Endothelial cell death as a percentage of total endothelial cells, within the centre of the corner (inner circle), the 1.5 mm ring of tissue centred on the femtosecond laser-assisted keratoplasty posterior side-cut incision (outer circle) and the overall cell death (total). Error bars represent SD.
Figure 5(A) Full-thickness corneal button stained with trypan blue, demonstrating peripheral and central endothelial cell damage. (B) Binary image produced by FIJI software demonstrating quantifiable regions of cell death over the entire graft and (C) isolated on the outer rim of the graft to highlight zones where laser may have contributed to cell death.
Discussion
This study demonstrates that a variety of full-thickness and partial-thickness FLAK wound patterns with a posterior 70 μ bridge of tissue provide impressive resistance to rupture under previously reported physiological and pathological intracameral pressure ranges. 18
Multiple theoretical concerns of dehiscence during patient transport exist. These include patient valsalva from cough, sneezing or vomiting, posterior pressure from retrobulbar block or placement of a Honan balloon. Thus, a tissue bridge is a safety factor used by many surgeons to prevent possible complications, including the worst of which could be an expulsive suprachoriodal haemorrhage. Wound configuration and optimisation makes the posterior stroma the most ideal place to leave a tissue bridge. Incision continuity and smooth graft-host interface are probably more important in the anterior portion of the wound as they may contribute more to the overall fit of graft and host and anterior surface contour. Indeed, two previous larger-scale clinical reports have used posterior stromal bridges during laser trephination. 13 19 A more recent report looked at 123 femtosecond laser keratoplasties done with full-thickness trephinations, but the laser was in the operating room and so no patient transport was required. 20
We have shown that 70 μ posterior stromal bridges impart a significant resistance to rupture in zigzag, top hat and mushroom patterns requiring upwards of 900 mm Hg of intracameral pressure to dehisce. Full-thickness incisions in all categories demonstrated a surprising amount of resistance to intracameral burst pressure as well. In fact, there was no significant difference in full-thickness and partial-thickness cut burst pressures in mushroom and top hat incisions. This is probably a function of wound configuration and remnant collagen bridges not broken by the laser, but requires further investigation. Interestingly, the zigzag full-thickness incision did not have the same level of burst resistance as full-thickness top hat and mushroom incisions. In our preliminary study, this mean was influenced by one full-thickness zigzag tissue that opened at a pressure of 23 mm ( table 1 ). Thus, we cannot say with certainty whether full-thickness trephination is safe if significant patient transport or movement is anticipated. It is possible that the angled side cut lowered the strength of the full-thickness incision in zigzag tissues compared with the vertical side cuts, which may have increased remnant collagen bridges when cut at the same energy. Additional variables not controlled in this study may have contributed to some of the variations seen in wound strength. These include variation in corneal thickness and tissue age, which could alter light scatter properties of tissue and biomechanical strength due to age-associated cross-linking, respectively.
A recent study demonstrated that profiled top hat incisions with short horizontal lamellar cuts provide a posterior graft underlay on the host cornea that contributes significantly better to wound strength than straight incisions in a variety of suture patterns. 21 While not directly comparable, it is interesting that in our study the full-thickness mushroom profile had higher burst pressure than the full-thickness top hat, which seems geometrically paradoxical. This trend was not statistically significant but may reflect higher collagen density in the anterior stroma and larger circumference of the mushroom anterior side cut. Theoretically, more residual collagen bridges would exist in the mushroom when cut at the same energy as the top hat anterior side cut, which has a smaller anterior side-cut circumference.
Data from this study must be interpreted with caution. Our sample number in each group is small, creating uncertainty in statistical analysis with outlier effects, which will require a larger follow-up study to verify the patterns observed. Previous studies have used similar small numbers as a proof of principle initial study 15 22 and reflect the availability and cost of research tissue. A straight-cut control would be an interesting comparison, but various cut patterns assessed are adequate internal controls in our study. Additionally, the mechanism tested was only wound dehiscence due to intracameral pressurisation in an artificial anterior chamber model. While this model has been commonly accepted in assessing these types of wounds, 9 11 15 22 further studies examining global external pressures or point source pressures on the corneal surface and on whole globes could shed further light on incision strength. These mechanisms may be relevant for patient transport between laser suite and operating room.
We were also able to demonstrate in this study that the endothelial cell death from full-thickness and partial-thickness FLAK wounds is moderate given current eye bank handling protocols in donor tissue. We focused on zigzag cuts due to limited research tissue availability and because zigzag is the most common of the three cuts used at this and other centres. 8 13 14 Typically endothelial cell counts are estimated either by specular microscopy 23 or by vital or devitalised tissue staining, 24 both of which have limitations. We used trypan blue staining coupled with the trainable segmentation function within FIJI image analysis software to create both reproducible and accurate binary images for calculation of percentage endothelial cell loss. FIJI has been widely used among other disciplines in cell sorting and quantification; 25 however, to our knowledge, this is the first time using Fiji's trainable segmentation towards corneal endothelial cell death analysis. While trypan staining may not provide the power to detect cell loss at an individual resolution, we have another manuscript in submission demonstrating a clear protocol for its use with vital dyes and its potential advantages over other reported methods for cell death analysis (Jardine G, et al ).
At our centre we typically receive full-thickness precut tissue transported in specially designed cradle from SightLife (Seattle Eye Bank) for our FLAK surgeries. Endothelial attenuation at the edges and centre of the graft under these transport conditions prior to this study have not been characterised. We noted that cell death is concentrated in the zone where laser trephination occurred even if the laser did not penetrate the endothelium. Interestingly, we noted some minor central cell death as well, which likely reflects the additional tissue manipulation required for precut tissue (3.2%, table 2 ). The converse situation of this is what happens to partially trephined tissue of the host. The analysis of endothelial cell death from the effect of femtosecond laser alone is important as this technology is being increasingly used in procedures that conserve the host's endothelium, such as deep anterior lamellar keratoplasty (DALK). We did not detect a significant difference in the endothelial cell loss in these partial-thickness cut tissues, but our analysis revealed a 5.30% (SD 1.30) cell death suggesting that the femtosecond laser causes modest cell death even if the endothelial plane is never penetrated.
In conclusion, we have demonstrated the posterior stromal bridge in femtosecond penetrating keratoplasty incisions imparts significant resistance to burst pressure and more consistently than do full-thickness incisions. This provides further evidence of a theoretical safety factor when transporting patients posttrephination between laser suite and operating room to complete the surgery. Additionally, we have demonstrated that surprisingly both full-thickness and partial-thickness femtosecond incisions in eye bank donor tissue cause modest endothelial cell death, concentrated in the zone of posterior laser trephination, which may have implications for its use in DALK.
Contributors: KRK made substantial contributions to data acquisition, analysis and interpretation, drafting the article and revising it critically for publication. MAP, JH, AP and BI made substantial contributions to conception and design of experiments and data acquisition. WC (corresponding author) planned the study, made substantial contributions to conception and design, acquisition of data, analysis and interpretation, revising the article critically for important intellectual content, and final approval of the version to be published.
Funding: Material support in the form of research tissue and laser interfaces was provided by Lions VisionGift, Portland, OR, and SightLife, Seattle, WA.
Competing interests: None.
Provenance and peer review: Not commissioned; externally peer reviewed.
Data sharing statement: All data from this investigation are available for sharing; however, there are no unpublished data. The corresponding author (WC) should be contacted for further inquiries regarding the data herein.
1 Zirm E. Eine erfolgreiche totale Keratoplastik. Archiv Ophthalmol 1906; 64: 580-93.
2 Hessburg PC Barron M. A disposable corneal trephine. Ophthalmic Surg 1980; 11: 730-3.
3 Boruchoff SA Jensen AD Dohlman CH. Comparison of suturing techniques in keratoplasty for keratoconus. Ann Ophthalmol 1975; 7: 433-6.
4 Hunkeler JD Hyde LL. The triple procedure: combined penetrating keratoplasty, cataract extraction, and lens implantation. Am Intraocular Implant Soc J 1979; 5: 222-4.
5 Barraquer J Rutllan J. eds. Microsurgery of the Cornea: an Atlas and Textbook. Barcelona: Ediciones Scriba, 1984.
6 Busin M. A new lamellar wound configuration for penetrating keratoplasty surgery. Arch Ophthalmol 2003; 121: 260-5.
7 Busin M Arffa RC. Microkeratome-assisted mushroom keratoplasty with minimal endothelial replacement. Am J Ophthalmol 2005; 140: 138-40.
8 Steinert RF Ignacio TS Sarayba MA. Top-hat shaped penetrating keratoplasty using the femtosecond laser. Am J Ophthalmol 2007; 143: 689-91.
9 Bahar I Kaiserman I McAllum P. Femtosecond laser-assisted penetrating keratoplasty: stability evaluation of different wound configurations. Cornea 2008; 2: 209-11.
10 Ignacio TS Nguyen TB Chuck RS. Top hat wound configuration for penetrating keratoplasty using the femtosecond laser: a laboratory model. Cornea 2006; 25: 336-40.
11 Malta JB Soong HK Shtein R. Femtosecond laser-assisted keratoplasty: laboratory studies in eye bank eyes. Curr Eye Res 2009; 34: 18-25.
12 Kook D Derhartunian V Bug R. Top-hat shaped corneal trephination for penetrating keratoplasty using the femtosecond laser: a histomorphological study. Cornea 2009; 29: 795-800.
13 Farid M Steinert RF Gaster RN. Comparison of penetrating keratoplasty performed with a femtosecond laser zig-zag incision versus conventional blade trephination. Ophthalmology 2009; 116: 1638-43.
14 Buratto L Böhm E. The use of the femtosecond laser in penetrating keratoplasty. Am J Ophthalmol 2007; 143: 737-42.
15 McAllum P Kaiserman I Bahar I. Femtosecond laser top hat penetrating keratoplasty: wound burst pressures of incomplete cuts. Arch Ophthalmol 2008; 6: 822-5.
16 Chamberlain WD Rush S Mathers WD. Comparison of femtosecond laser-assisted keratoplasty versus conventional penetrating keratoplasty. Ophthalmology 2011; 118: 486-91.
17 Kim JH Choi SK Lee D. The comparison of femtosecond laser-assisted penetrating keratoplasty with conventional surgery in terms of endothelial safety: ex vivo study using porcine eyes. Cornea 2009; 28: 812-16.
18 Brody S Erb C Veit R. Intraocular pressure changes: the influence of psychological stress and the Valsalva maneuver. Biol Psychol 1999; 51: 43-57.
19 Price MO Jordan CS. Safety of incomplete incision patterns in femtosecond laser-assisted penetrating keratoplasty. J Cataract Refract Surg 2008; 34: 2099-103.
20 Birnbaum F Wiggermann A Maier PC. Clinical results of 123 femtosecond laser-assisted penetrating keratoplasties. Graefes Arch Clin Exp Ophthalmol 2013; 251: 95-103.
21 Maier P Böhringer D Birnbaum F. Improved wound stability of top-hat profiled femtosecond laser-assisted penetrating keratoplasty in vitro. Cornea 2012; 31: 963-6.
22 Gilmer W McLeod SD Jeng BH. Laboratory model of bursting pressures of femtosecond laser-assisted penetrating keratoplasty wounds using novel pattern designs. Br J Ophthalmol 2012; 96: 1273-4.
23 Terry MA Shamie N Straiko MD. Endothelial keratoplasty: the relationship between donor tissue storage time and donor endothelial survival. Ophthalmology 2011; 118: 36-40.
24 Gain P Thuret G Chiquet C. Value of two mortality assessments techniques for organ cultured corneal endothelium: trypan blue versus TUNEL technique. Br J Ophthalmol 2002; 86: 306-10.
25 Schindelin J Arganda-Carreras I Frise E. Fiji: an open-source platform for biological-image analysis. Nature Methods 2012; 9: 676-82.
Copyright: 2014 Published by the BMJ Publishing Group Limited. For permission to use (where not already granted under a licence) please go to http://group.bmj.com/group/rights-licensing/permissions