High-irradiance CXL combined with myopic LASIK: flap and residual stroma biomechanical properties studied ex-vivo

Transcription

1 A. John Kanellopoulos, MD- 2709, ASCRS San Diego 2015 Handout Collagen Crosslinking: Evolving Indications and Technology, Applications, Results, and Complication Management Senior Instructor: Gregory J. Pamel, MD Instructors: A. John Kanellopoulos, MD Doyle Stulting, MD Eric Donnenfeld, MD Soosan Jacob, MD ABSTRACT BODY: Course Description: Course will present management of cornea ectasia (keratoconus/ post-refractive surgery). Surgical treatment modalities utilized internationally will be presented, including: CXL with UVA in order to halt ectasia, combined with customised excimer laser ablation, and topography customized CXL applications. These alternatives to Intracorneal ring segment implantation, lamellar grafts, penetrating graft will be analyzed. Early and late complications will be presented and discussed. Educational Objective: Attendees will share our vast experience in managing progressive keratoconus and post-lasik ectasia in order to visually rehabilitate these patients. 1 Laservision.gr Eye Institute, Athens, Greece 2 Department of Ophthalmology, New York University Medical School, New York, New York, USA 3 Department of Cornea & Refractive Surgery, Harvard Medical School, Massachusetts Eye & Ear Infirmary, Boston, Massachusetts, USA Correspondence to Dr Anastasios John Kanellopoulos, Department of Ophthalmology, NYU Medical School, New York, NY, USA; Laservision.gr Clinical Research Eye Institute, 17 Tsocha Street, Athens 11521, Greece; ajk@brilliantvision.com Received 19 November 2014 Revised 28 January 2015 Accepted 15 February 2015 To cite: Kanellopoulos AJ, Asimellis G, Ciolino JB, et al. Br J Ophthalmol Published Online First: [please include Day Month Year] doi: /bjophthalmol High-irradiance CXL combined with myopic LASIK: flap and residual stroma biomechanical properties studied ex-vivo Anastasios John Kanellopoulos, 1,2 GeorgeAsimellis, 1 Joseph B Ciolino, 3 BorjaSalvador-Culla, 3 James Chodosh 3 ABSTRACT Background/aims To evaluate ex vivo biomechanical and enzymatic digestion resistance differences between standard myopic laser in-situ keratomileusis (LASIK) compared with LASIK+CXL, in which high-irradiance cross-linking (CXL) is added. Methods Eight human donor corneas were subjected to femtosecond-assisted myopic LASIK. Group A (n=4) served as a control group (no CXL). The corneas in LASIK+CXL group B were subjected to concurrent prophylactic high-irradiance CXL (n=4). Saline-diluted (0.10%) riboflavin was instilled on the stroma, subsequently irradiated with UV-A through the repositioned flap. The cornea stroma and flap specimens were separately subjected to transverse biaxial resistance measurements; biomechanical differences were assessed via stress and Young s shear modulus. Subsequently, the specimens were subjected to enzymatic degradation. Results For the corneal stroma specimen, stress at 10% strain was 128±11 kpa for control group A versus 293±20 kpa for the LASIK+CXL group B (relative difference Δ=+129%, p<0.05). The stress in group B was also increased at 20% strain by +68% (p<0.05). Shear modulus in group B was increased at 10% strain by +79%, and at 20% strain by +48% (both statistically significant, p<0.05). The enzymatic degradation time to dissolution was 157.5±15.0 min in group A versus ±7.5 min in group B (Δ=+18%, p=0.014). For the flaps, both biomechanical, as well as enzymatic degradation tests showed no significant differences. Conclusions LASIK+CXL appears to provide significant increase in underlying corneal stromal rigidity, up to +130%. Additionally, there is significant relevant enzymatic digestion resistance confirmatory to the above. LASIK flaps appear unaffected biomechanically by the LASIK+CXL procedure, suggesting effective CXL just under the flap. INTRODUCTION Corneal collagen cross-linking (CXL) has been clinically employed for stabilising progressive keratectasia. 1 2 This photochemical reactive process is induced by peak 365 nm ultraviolet (UV-A) radiation absorbed by riboflavin, a photosensitive vitamin B2 molecule. The procedure is broadly known as corneal cross-linking despite the fact that there are some reports suggesting that the mechanism responsible for biomechanical strengthening 34 within the stroma is related not to interlamellar cohesion increase, but to inter-fibrillar and intra-fibrillar cohesion. 5 In addition, increased Laboratory science collagen resistance against enzymatic degradation has been associated with CXL. 6 8 We have introduced an alternative CXL application, adjuvant to myopic laser in-situ keratomileusis (LASIK+CXL). The application aims to improve long-term keratometric stability 9 and to reduce regression likelihood following moderate and high myopic LASIK 10 by proactively restoring corneal biomechanical strength. 11 Riboflavin solution is briefly applied on the exposed stromal bed at the completion of the excimer ablation; the flap is then repositioned, followed by superficial UV-A irradiation To the best of our knowledge, the biomechanical and/or enzymatic degradation resistance modulations achieved via CXL application concurrent with LASIK have not been studied in human corneas. The purpose of this study is to evaluate ex-vivo biomechanical and enzymatic degradation resistance differences in such application. MATERIALS AND METHODS Eight human donor corneas were involved, obtained by the Eye Bank for Sight Restoration Inc (New York, USA), an accredited member of the Eye-Bank Association of America. The corneas were donated by eight different donors (four men, four women) of average age 62.0±9.5 (43 72) years, stored in 4 C OptiSol solution (Bausch +Lomb, Rochester, New York, USA). Surgical technique All corneas were subjected to femtosecond-laser assisted myopic treatment. The corneas were mounted on an artificial anterior chamber (Baron, Katena Products, Inc, Denville, New Jersey, USA). Flaps (120 μm thick, 8.5 mm diameter) were created with the WaveLight FS200 femtosecond laser (Alcon Surgical, Ft Worth, Texas, USA), observing standard docking, applanation and vacuum-suction procedures (figure 1A). After flap lifting (figure 1B), the WaveLight EX500 excimer laser (Alcon) was employed to create a 8.00 D myopic ablation over a 6.5 mm diameter optical zone (figure 1C). During the procedure, interferometric pachymetry embedded in the EX500 provided corneal thickness data. Isotonic saline 0.1% riboflavin solution (Vibex Rapid, Avedro Inc, Waltham, Massachusetts, USA) was instilled on the exposed stromal bed afforded by the lifted flap (figure 1C). Soaking time was 1 min (figure 1E); then excess riboflavin was wiped from the cornea surface. Special care was taken to minimise potential riboflavin soaking to the folded LASIK flap. Kanellopoulos AJ, et al. Br J Ophthalmol 2015;0:1 7. doi: /bjophthalmol

2 Laboratory science Laboratory science Table 1 Biomechanical comparative measurements between the two groups Stress Units: kpa Young s shear modulus Units: 10% 20% 10% 20% strain Group A (control) ± ± ± ±1.9 Group B (CXL) ± ± ± ±5.2 Δ 129% 68% 79% 48% p t Δ, relative (%) difference between metrics; p, Student t test p value; t, Student t test t value. Results from corneal stroma specimens. Figure 1 Schematic of the surgical procedure LASIK combined with prophylactic high-irradiance corneal cross-linking (LASIK+CXL). The corneas were then randomly formed into two groups, four in each. Control group A received no further treatment. Group B (LASIK+CXL) was subjected to cross-linking: with the flap repositioned, the cornea was UV-A irradiated at 30 mw/cm 2 for 80 s (total fluence 2.4 J/cm 2 ), employing the KXL device (Avedro) (figure 1F). The flaps were then amputated from all corneas; stroma and flap specimens were stored back to OptiSol (4 C) until testing. Biomechanical strength testing The stroma specimens were prepared by razor-blade manual dissection to approximately mm. The amputated flaps were tested without additional preparation. Transverse biaxial load-cell resistance measurements were accomplished by tangential shearforce employing the Biotester 5000 (CellScale Biomaterials Testing, Waterloo, Canada). The device records the simultaneous x-axis and y-axis displacement, applied force, and time. An integrated camera captures still, pixel images, which provide precise x-displacement and y-displacement measurements, analysed by custom software (LabJoy V.9.05). 14 The specimens were fixed (via random orientation) on a 4 5-tine rake arrangement clamped on their centre mm section. The tines, of 250 μm diameter, were spaced by 0.7 mm (figure 2). The specimens were then submerged into an isotonic saline bath, temperature controlled at 37 C, for 5 min before (for temperature stabilisation) as well as during testing (to eliminate temperature-related variability). 15 Shear rate was fixed to 4.16 μm/s. Time (s), x and y displacement (μm), and x and y force (mn) were recorded every second. Enzymatic degradation The stroma specimens were trephined into 8.5 mm diameter round buttons; no further processing on the amputated flaps. A 0.3% collagenase-a solution (active agent: Clostridium histolyticum) (Sigma-Aldrich, St Louis, Missouri, USA) was prepared via dilution in Dubecco s Phosphate Buffered Saline (DPBS, Sigma-Aldrich). Stroma and flap specimens were incubated within 1.5 ml of collagenase-a solution; the test tube racks were placed at 37 C on a plate shaker at 175 rotations per min. Figure 2 Fitting of the corneal specimens (top) and the flap (bottom) on the BioTester device. 2 Kanellopoulos AJ, et al. Br J Ophthalmol 2015;0:1 7. doi: /bjophthalmol Specimens were observed every 30 min until complete dissolution was achieved. The time to complete dissolution of the stromal and flap specimens was recorded. Data analysis Stress, a pressure metric, is the applied force (F) divided by the cross section (A) of the tested area (F/A, units=mn/ mm 2 =kpa=10 3 Pa; MPa=10 6 Pa; Pa=Pascal). Strain, expressed as percentage, is the unitless relative elongation Δx/lx (or Δy/ly, respectively), where l is the initial length. Cross-section test area was defined by the 3.5 mm sample length multiplied by the respective stroma/flap thickness, respectively. Statistical analysis and graphs of stress (expressed in kpa) in the vertical, versus strain in the horizontal axis, were constructed using the SPSS software V.21.0 (IBM Corporation, New York, USA). The exponential fitting for stress calculation was conducted at the 10% and 20% strain. Linear regression fitting was performed to calculate, by means of the slope function (gradient), the stress/ strain ratio, an expression of the shear (Young s) modulus. The shear modulus was calculated as the gradient at 10% and 20% strain. To ensure proper linear fit, we sought a minimum of 0.98 for the trend-line determination coefficient (r 2 ). Statistical significance was assessed employing Student s t test. p Values less than 0.05 were indicative of statistically significant results in this study. Results are reported in the form: average ±SD (minimum to maximum). RESULTS In group A the donor age at the time of cornea harvest was 63.3 ±13.7 years (48 72) and in group B, 60.8±4.5 years (43 69). Biomechanical strength results The distinct datasets during shear-strength measurements on stroma specimens (all stroma specimens) were on average 290 ±37 ( ), while average displacement was 1205±155 μm ( ). Mean maximum applied force was 3996 Table 2 Biomechanical comparative measurements between the two groups ±597 mn ( ). Average value of maximum strain was 30±3% (26 34%). The distinct datasets on flap specimens (all flap specimens) were on average 222±18 ( ), while average displacement was 925±74 μm ( ). Mean maximum applied force was 1505±197 mn ( ). Average value of maximum strain was 23.5±2% (21 25%). The average number for all specimens of distinct data pairs (vertical axis, stress and horizontal axis, strain) within the linear phase of analysis of the stress strain curves was 119 (95 129). The average coefficient of determination (r 2 ) of the trend-line was ( ). For the stroma specimens in control group A, average stress at 10% strain was 128±11 kpa and at 20% strain 804±31 kpa; Young s modulus at 10% strain was 3.7±2.5 MPa and at 20% strain 9.5±1.8 MPa. In the LASIK+CXL group B, stress at 10% strain was 293±20 kpa and at 20% strain 1350±46 kpa; Young s modulus at 10% strain was 6.6±2.4 MPa and at 20% strain 14.0±5.2 MPa. The relative increase of these biomechanical properties in LASIK+CXL group B in comparison to control group A was stress +129% and +68% for 10% and 20% strain, and Young s modulus +79% and +48%, respectively. All differences were statistically significant (table 1). For the flap specimens in control group A: stress at 10% strain was 158±17 kpa and at 20% strain 1566±46 kpa; Young s modulus at 10% strain was 5.9±3.9 MPa and at 20% strain 13.6±6.2 MPa. In the LASIK+CXL group B, stress at 10% strain was 182±21 kpa and at 20% strain 1534±134 kpa; Young s modulus at 10% strain was 6.5±3.2 MPa and at 20% strain 11.9±4.1 MPa. The relative changes of these biomechanical properties in LASIK+CXL group B in comparison to control group A were not statistically significant (table 2). Enzymatic digestion results Regarding the stroma specimens, the mean time to complete dissolution in control group A was 157±15 min, while in group B (LASIK+CXL) it was 186.2±7.5 min. The relative difference Stress Units: kpa Young s shear modulus Units: 10% 20% 10% 20% strain Group A (control) ± ± ± ±6.2 Group B (CXL) ± ± ± ±4.1 Δ 15% 2% 11% 12% p t Δ, relative (%) difference between metrics; p, Student t test p value; t, Student t test t value. Results from flap specimens. Kanellopoulos AJ, et al. Br J Ophthalmol 2015;0:1 7. doi: /bjophthalmol

3 Laboratory science Laboratory science Δ was +18% for group B, a statistically significant difference (p=0.014). Regarding the flaps, the mean time to complete dissolution in group A was 68.75±17.3 min, while in group B it was 80.0±8.1 min. The relative difference Δ was +16% for group B, a difference which, however, was not statistically significant (table 3). DISCUSSION Epithelium-on, in-situ CXL is reported in the peer review literature with a significantly weaker biomechanical effect in comparison to epi-off CXL. 18 This is partially attributed to reduced riboflavin permeability through the intact epithelium, mainly due to its large molecular weight. 19 This leads to insufficient and inhomogeneous riboflavin stromal diffusion and thus affects UV-A transmission to deeper layers, due to increased UV-A absorption by the superficially concentrated riboflavin. The second issue that affects epithelium-on CXL relates to UV-A absorption/filtering by the intact epithelium. 20 Some studies suggest absorption by the epithelium of about one-third, 19 leading to less energy reaching the riboflavin-saturated stroma. However, there are studies suggesting that human corneal epithelium and the underlying basement membrane absorb strongly only at wavelengths smaller than 310 nm, which compares to the peak 365 nm of the CXL-employed UV-A. The effect of refractive surgery on corneal biomechanical properties may be evaluated in-vivo by the corneal hysteresis and corneal resistance factor. These indices can be assessed by dynamic tonometry (visualisation of fast deformation of the cornea), employing the Corvis ST (Oculus Optikgeräte GmbH, Wetzlar, Germany) and the Ocular Response Analyser (Reichert, Buffalo, New York, USA). Several studies have evaluated the reduction in corneal biomechanical strength following LASIK, and following ReLEx flex and ReLEx smile procedures, suggesting similar corneal biomechanical reduction among these laser refractive procedures However, there is inconclusive evidence in the peer review literature on the specificity of these techniques in the in-vivo evaluation of the effect of corneal cross-linking. The in-situ riboflavin application investigated in this study naturally overcomes the above-mentioned restriction of riboflavin penetration through the intact epithelium, since it is applied directly on the exposed stromal bed. In addition, the ex-vivo nature of this study overcomes the potential shortcomings presented in relation to in-vivo evaluation techniques. Our team has investigated clinically in-situ (ie, intrastromal) CXL applications as a means to offer proactive cornea stabilisation in high-myopic LASIK. 31 The long-term investigation of epithelial thickness increase was investigated as an indirect measurement of overall corneal stability, supported by a difference in Table 3 Enzymatic digestion time to complete dissolution epithelial remodelling, which may indicate increased biomechanical stability in myopic LASIK combined with adjuvant CXL. In this study we conducted a human ex-vivo investigation of the biomechanical corneal strengthening occurring during in-situ CXL along myopic LASIK. We evaluated changes in the stroma (and also, for the first time reported separately, the LASIK flap) employing biaxial stress strain measurements, which may be considered superior to the one-dimensional corneal strip extensiometry 32 taking into account the nonuniform topographic distribution of the corneal strength profile We provided 10% and 20% strain values for two reasons. First, these are the ones typically reported by other researchers, 14 in an effort to provide some form of comparison. In addition, the stress strain plot had adequate linearity in these areas to enable data collection. The depth dependence of transverse shear modulus of the cornea, stronger at the anterior third, 35 indicates that tissue removal from this upper third may affect corneal rigidity the most. In our study we demonstrated that the effective CXL corneal rigidity in comparison to the non-cxl corneas that received the same treatment was of the order of +129% stress at 10% strain, by a statistically significant margin. Considering that the aim of the in-situ CXL application is to provide prophylactic corneal strengthening to counter weakening due to tissue removal, we view this very significant biomechanical effect as a very important finding. These findings are confirmatory to our previous clinical effects reported on myopic and hyperopic LASIK+CXL applications. 36 CXL maybe therefore be used as a biomechanical modulator offering refractive stability following LASIK. There was no indication that the flaps had any statistically significant biomechanical difference by any metric employed. This finding is a very important indicator that no actual cross-linking of the flap occurs during our LASIK+CXL technique. There are two main reasons that we target to avoid any cross-linking effect on the flap. First is that the flap does not contribute to the biomechanical properties of the underlying stroma Therefore, there is no benefit from a potential cross-linking of the flap. Second, and perhaps even more important, is that cross-linking such a thin (the LASIK flap consists of 50 μm of epithelium and 60 μm of stroma) stromal content may lead to undesirable stromal shrinking of the flap. Additionally, the findings in the collagen-enzymatic digestion part of this work provide corroborative evidence of the differential effects of cross-linking on the stroma and not on the flap. This is, to the best of our knowledge, the first manuscript presenting both techniques, bi-axial stress measurements and enzymatic digestion, of ex-vivo human assessment of crosslinking effects. Additional studies with larger sample size, differentiation of UV-A irradiance, as well as extending the testing period to longer time intervals following treatment are warranted to validate and further investigate the findings in this preliminary study. Competing interests Consultant/advisory positions: AJK: Alcon/WaveLight, Avedro, i-optics, Allegran; Keramed. Provenance and peer review Not commissioned; externally peer reviewed. REFERENCES 1 Wollensak G, Spoerl E, Seiler T. Riboflavin/ultraviolet-A-induced collagen crosslinking for the treatment of keratoconus. Am J Ophthalmol 2003;135: Kanellopoulos AJ, Binder PS. Collagen cross-linking (CCL) with sequential topography-guided PRK: a temporizing alternative for keratoconus to penetrating keratoplasty. Cornea 2007;26: Wollensak G, Iomdina E. Biomechanical and histological changes after corneal crosslinking with and without epithelial debridement. J Cataract Refract Surg 2009;35: Hafezi F, Kanellopoulos J, Wiltfang R, et al. Corneal collagen crosslinking with riboflavin and ultraviolet A to treat induced keratectasia after laser in situ keratomileusis. J Cataract Refract Surg 2007;33: Wollensak G, Spörl E, Mazzotta C, et al. Interlamellar cohesion after corneal crosslinking using riboflavin and ultraviolet A light. Br J Ophthalmol 2011;95: Spoerl E, Wollensak G, Seiler T. Increased resistance of crosslinked cornea against enzymatic digestion. Curr Eye Res 2004;29: Arafat SN, Robert MC, Shukla AN, et al. UV cross-linking of donor corneas confers resistance to keratolysis. Cornea 2014;33: Mazzotta C, Caporossi T, Denaro R, et al. Morphological and functional correlations in riboflavin UV A corneal collagen cross-linking for keratoconus. Acta Ophthalmol 2012;90: Kanellopoulos AJ, Asimellis G. Refractive and keratometric stability in high myopic LASIK with high-frequency femtosecond and excimer lasers. J Refract Surg 2013;29: Chayet AS, Assil KK, Montes M, et al. Regression and its mechanisms after laser in situ keratomileusis in moderate and high myopia. Ophthalmology 1998;105: Cho M, Kanellopoulos AJ. Safety and efficacy of prophylactic ultraviolet-a-induced crosslinking after high risk myopic photerefractive keratomy, Poster 5470/A441, ARVO May 3 7, 2009, Florida, USA. 12 Kanellopoulos AJ. Long-term safety and efficacy follow-up of prophylactic higher fluence collagen cross-linking in high myopic laser-assisted in situ keratomileusis. Clin Ophthalmol 2012;6: Kanellopoulos AJ, Pamel GJ. Review of current indications for combined very high fluence collagen cross-linking and laser in situ keratomileusis surgery. Indian J Ophthalmol 2013;61: Søndergaard AP, Ivarsen A, Hjortdal J. Corneal resistance to shear force after UVA-riboflavin cross-linking. Invest Ophthalmol Vis Sci 2013;54: Hatami-Marbini H, Rahimi A. Effects of bathing solution on tensile properties of the cornea. Exp Eye Res 2014;120: Boxer Wachler BS, Pinelli R, Ertan A, et al. Safety and efficacy of transepithelial crosslinking (C3-R/CXL). J Cataract Refract Surg 2010;36: Magli A, Forte R, Tortori A, et al. Epithelium-off corneal collagen cross-linking versus transepithelial cross-linking for pediatric keratoconus. Cornea 2013;32: Leccisotti A, Islam T. Transepithelial corneal collagen cross-linking in keratoconus. J Refract Surg 2010;26: Baiocchi S, Mazzotta C, Cerretani D, et al. Corneal crosslinking: riboflavin concentration in corneal stroma exposed with and without epithelium. J Cataract Refract Surg 2009;35: Podskochy A. Protective role of corneal epithelium against ultraviolet radiation damage. Acta Ophthalmol Scand 2004;82: Kolozsvári L, Nógrádi A, Hopp B, et al. UV absorbance of the human cornea in the 240- to 400-nm range. Invest Ophthalmol Vis Sci 2002;43: Ionescu AM, de la Cruz Cardona J, González-Andrades M, et al. UV absorbance of a bioengineered corneal stroma substitute in the nm range. Cornea 2010;29: Kirwan C, O Keefe M. Corneal hysteresis using the Reichert ocular response analyser: findings pre- and post-lasik and LASEK. Acta Ophthalmol 2008;86: Hamilton DR, Johnson RD, Lee N, et al. Differences in the corneal biomechanical effects of surface ablation compared with laser in situ keratomileusis using a microkeratome or femtosecond laser. J Cataract Refract Surg 2008;34: Uzbek AK, Kamburoglu G, Mahmoud AM, et al. Change in biomechanical parameters after flap creation using the intralase femtosecond laser and subsequent excimer laser ablation. Curr Eye Res 2001;36: Kamiya K, Shimizu K, Ohmoto F. Time course of corneal biomechanical parameters after laser in situ keratomileusis. Ophthalmic Res 2009;42: Pedersen IB, Bak-Nielsen S, Vestergaard AH, et al. Corneal biomechanical properties after LASIK, ReLEx flex, and ReLEx smile by Scheimpflug-based dynamic tonometry. Graefes Arch Clin Exp Ophthalmol 2014;252: Agca A, Ozgurhan EB, Demirok A, et al. Comparison of corneal hysteresis and corneal resistance factor after small incision lenticule extraction and femtosecond laser-assisted LASIK: a prospective fellow eye study. Cont Lens Anterior Eye 2014;37: Bak-Nielsen S, Pedersen IB, Ivarsen A, et al. Dynamic Scheimpflug-based assessment of keratoconus and the effects of corneal cross-linking. J Refract Surg 2014;30: Gatinel D. The mystery of collagen cross-linking when it comes to in vivo biomechanical measurements. J Refract Surg 2014;30: Kanellopoulos AJ, Asimellis G. Epithelial remodeling after femtosecond laser-assisted high myopic LASIK: comparison of stand-alone with LASIK combined with prophylactic high-fluence cross-linking. Cornea 2014;33: Elsheikh A, Anderson K. Comparative study of corneal strip extensometry and inflation tests. J R Soc Interface 2005;2: Sánchez P, Moutsouris K, Pandolfi A. Biomechanical and optical behavior of human corneas before and after photorefractive keratectomy. J Cataract Refract Surg 2014;40: Smolek MK. Interlamellar cohesive strength in the vertical meridian of human eye bank corneas. Invest Ophthalmol Vis Sci 1993;34: Petsche SJ, Chernyak D, Martiz J, et al. Depth-dependent transverse shear properties of the human corneal stroma. Invest Ophthalmol Vis Sci 2012;53: Kanellopoulos AJ, Asimellis G, Karabatsas C. Comparison of prophylactic higher fluence corneal cross-linking to control, in myopic LASIK, one year results. Clin Ophthalmol 2014;8: Reinstein DZ, Srivannaboon S, Archer TJ, et al. Probability model of the inaccuracy of residual stromal thickness prediction to reduce the risk of ectasia after LASIK part II: quantifying population risk. J Refract Surg 2006;22: Leccisotti A. Corneal ectasia after photorefractive keratectomy. Graefes Arch Clin Exp Ophthalmol 2007;245: Corneas Flaps Time Units: min Time Units: min Group A (control) ± ±17.32 Group B (CXL) ± ±8.16 Δ 18% 16% p t Δ, relative (%) difference between metrics; p, Student t test p value; t, Student t test t value. CONCLUSIONS Adjuvant intrastromal in-situ CXL combined with myopic LASIK appears in ex vivo human study to be a significant biomechanical modulator. Contributors Design and conduct of the study (AJK, GA); collection (AJK, GA, BS-C), management (AJK, JBC, JC), analysis (AJK, GA, JBC, BS-C, JC), interpretation of the data (AJK, GA, JBC, BS-C, JC); manuscript preparation (GA), manuscript review (AJK, GA, JBC, BS-C, JC), manuscript approval (AJK, GA, JBC, BS-C, JC). 4 Kanellopoulos AJ, et al. Br J Ophthalmol 2015;0:1 7. doi: /bjophthalmol Kanellopoulos AJ, et al. Br J Ophthalmol 2015;0:1 7. doi: /bjophthalmol

4 Laboratory science Laboratory science 6 Kanellopoulos AJ, et al. Br J Ophthalmol 2015;0:1 7. doi: /bjophthalmol Kanellopoulos AJ, et al. Br J Ophthalmol 2015;0:1 7. doi: /bjophthalmol

5 BASIC INVESTIGATION Kanellopoulos et al Cornea Volume 0, Number 0, Month 2015 Corneal Collagen Cross-linking Combined With Simulation of Femtosecond Laser Assisted Refractive Lens Extraction: An Ex Vivo Biomechanical Effect Evaluation Anastasios J. Kanellopoulos, MD,* Mark A. Kontos, MD, Shihao Chen, MD, OD, MSc, and George Asimellis, PhD* Purpose: To evaluate biomechanical changes induced by in situ corneal cross-linking (CXL) with stromal pocket delivered enhanced concentration riboflavin and high-fluence, high-energy UV-A irradiation. Methods: Eight human donor corneas were subjected to intrastromal lamellar corneal tissue removal of anterior 140-mm deep, 80-mm thick 5-mm diameter central stromal buttons, extracted through a 3.5-mm width tunnel, surfacing in the superior cornea periphery. Enhanced concentration riboflavin solution (0.25%) was instilled in the pocket. In study group A (CXL), superficial high-fluence UV-A irradiation was applied, whereas in control group B (no CXL), none. To comparatively assess changes in corneal rigidity, corneal specimens were subjected to transverse biaxial resistance measurements by application of a unidirectional tangential shear force. Biomechanical differences were evaluated through stress and Young shear modulus. Results: Stress at 10% strain was kpa in study group A versus kpa in control group B (relative difference D = 107%, P = 0.021). Stress at 20% strain was kpa in study group A versus kpa in control group B (D = 47%, P = 0.043). Average shear modulus in study group A at 10% strain was MPa versus MPa in control group B (D = 73%, P = 0.036). Average shear modulus in study group A at 20% strain was MPa versus MPa in group B (D = 30%, P = 0.047). Conclusions: Adjunct CXL in this ex vivo simulation refractive lens extraction procedure seems to provide significant increase in corneal rigidity, up to +107%. These findings also support our previous reported work on laser in situ keratomileusis combined with CXL. Key Words: femtosecond laser, refractive lens extraction, biomechanical simulation, in situ CXL, epithelium on, high-fluence Received for publication November 4, 2014; revision received December 15, 2014; accepted December 18, From the *Laservision.gr Clinical and Research Eye Institute, Eye Institute, Athens, Greece; New York University Medical School, New York, NY; Empire Eye Physicians, Spokane, WA; and Affiliated Eye Hospital of Wenzhou Medical University, Zhejiang, China. A. J. Kanellopoulos holds consultant/advisory positions in Alcon/WaveLight, Avedro, i-optics, and Oculus and M. A. Kontos in Abbott Medical. The remaining authors have no funding or conflicts of interest to disclose. Reprints: Anastasios J. Kanellopoulos, MD, Laservision.gr Clinical and Research Eye Institute, Eye Institute, 17 Tsocha St, Athens 11521, Greece ( ajk@brilliantvision.com). Copyright 2015 Wolters Kluwer Health, Inc. All rights reserved. CXL, high-energy CXL, higher riboflavin concentration CXL, corneal biomechanics, Young shear modulus, corneal stress strain (Cornea 2015;0:1 7) orneal collagen cross-linking (CXL) has been clinically Cused for stabilizing progressive keratectasia. 1,2 A photochemical reactive process induced by 365-nm UV-A light in the presence of riboflavin, a photosensitive molecule, results in increased intrafibrillar and interfibrillar covalent bonds 3 and stromal collagen resistance against enzymatic degradation. 4,5 The increased stromal biomechanical strength and lamellar compaction lead to corneal stabilization. 6 The original protocol Dresden technique 1 requires corneal epithelium removal (epithelium off) and anterior stromal saturation with riboflavin solution. Aiming to address the significant morbidity and postoperative pain associated with epithelial removal, 7 several alternatives have been proposed. One such alternative is to keep the epithelium in place and attempt to loosen the epithelial tight junctions to facilitate riboflavin diffusion with prolonged use of topical anesthetics, in a procedure termed epithelium-on CXL. 8 Our team has introduced alternative CXL techniques using higher fluence, shorter duration UV-A irradiation, and administration of riboflavin solution in an anterior intrastromal lamellar pocket created with a femtosecond laser. 9 A subsequent CXL alternative application we also presented (Kanellopoulos AJ Prophylactic, ultraviolet a cross-linking combined at the completion of high risk myopic LASIK cases. Subspecialty Day Paper presentation, American Academy of Ophthalmology Annual Meeting, Nov 8, 2008, Atlanta, GA) aimed to proactively restore corneal biomechanical stability 10 in myopic laser in situ keratomileusis (LASIK) corrections Riboflavin is applied on the exposed stromal bed afforded by the open LASIK flap; the flap is then repositioned, followed by UV-A irradiation of the riboflavin-soaked stroma through the flap. 14,15 Recently reported refractive error correction procedures involve intrastromal tissue dissection performed solely by a femtosecond laser. 16,17 Although the mechanism of stromal tissue reduction in these procedures is different from LASIK, the issue of potential biomechanical weakening, particularly in attempted high myopic corrections has been modeled 18 and investigated clinically. 19 The combination of CXL with such Cornea Volume 0, Number 0, Month Copyright ª 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. refractive correction application aiming to restore corneal biomechanical strength has not yet been reported. The purpose of this study was to evaluate ex vivo the biomechanical changes resulting from a simulated refractive lens extraction procedure combined with CXL. MATERIALS AND METHODS This laboratory (ex vivo) study was approved by the Laservision.gr Clinical and Research Eye Institute Ethics Committee. Eight human donor corneas were used in the study, which were obtained for research purposes from the Eye Bank for Sight Restoration Inc (New York, NY). The corneas were donated by 8 different organ donors (4 men, 4 women), whose mean age was years (range, years). Experimental Technique All corneas were subjected to creation and removal of a central lamellar intrastromal button through a customdesigned application consisting of a double intrastromal bed cut. A WaveLight FS200 femtosecond laser (Alcon Surgical, Ft Worth, TX) was used for this purpose. The corneas were mounted on an artificial chamber (Baron; Katena Products, Inc, Denville, NJ). To achieve high precision intrastromal cuts, we implemented standard docking, applanation, and vacuum suction procedures appropriate for the use of the FS200 laser along with the clear cone patient interface. 20 Subsequently, the following 2 FS200 procedures were applied: 1. Procedure A, a circular posterior lamellar bed dissection of 9-mm diameter 220-mm depth from the anterior corneal surface (Fig. 1A). To achieve this, the flap mode with no side cut was used. A 0.4-mm wide and 3.0-mm long canal was designed to serve as the channel through which the button to be created would be extracted. 2. Procedure B, a circular anterior lamellar dissection of 5-mm diameter 140-mm depth(fig.1b).toachieve this, the posterior lamellar mode was used. In addition, the circumferential side-wall cut was extended to 100 mm to ensure that the vertical dissections of the cut would transcend the posterior lamellar level. The combination of these 2 procedures resulted in the creation of an intralamellar button, 5-mm diameter, 80-mm thick, located 140 mm inside the cornea (Fig. 1C). In addition, a channel was created, enabling button extraction. The button was extracted using a Sinskey hook and toothless FIGURE 1. A, Posterior lamellar cut: left, detail from the system software report, right, schematic drawing; B, Superior lamellar cut: left, detail from the system software report, right, schematic drawing; C, Intralamellar button creation: left, detail from the system software report, right, schematic drawing. 2 Copyright 2015 Wolters Kluwer Health, Inc. All rights reserved. Copyright ª 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

6 Cornea Volume 0, Number 0, Month 2015 In Situ CXL Corneal Biomechanics Kanellopoulos et al Cornea Volume 0, Number 0, Month 2015 forceps (Katalyst Surgical, Chesterfield, MO) for the lamellar dissections (button separation) and forceps for button grip and extraction. Before and after the procedure, corneal thickness and anterior surface keratometry were evaluated through anterior segment optical coherence tomography (RtVue-100; Optovue Inc, Fremont, CA) and Scheimpflug imaging (Pentacam; Oculus Optikgeräte GmbH, Wetzlar, Germany). The corneas were then randomly assigned to 2 groups of 4 each. In study group A, a subsequent high-fluence CXL was applied. A 0.25% isotonic saline-diluted riboflavin solution (Vibex Xtra; Avedro Inc, Waltham, MA) was injected in the pocket afforded by the removed button through the channel used for its removal. Soaking time was 1minute,followedbyrinsingofexcessriboflavin from the corneal surface. UV-A illumination of 45 mw/cm 2 was applied for 2 minutes and 40 seconds (intended energy, 7.2 J/cm 2 ). The KXL-I cross-linking device (Avedro Inc) was used for this purpose. In the corneas belonging to group B, the same riboflavin solution was injected in the pocket, but no UV-A was irradiated, with the intent to serve as acontrolgroup. Data Collection and Analysis A central square block (approximately mm) was obtained from each treated corneal specimen using a disposable razor blade. A Biotester 5000 device (CellScale Biomaterials Testing, Waterloo, Canada) was used to provide biaxial resistance testing. The specimens were fixed (random orientation) on a 4 5 tine rake arrangement (Fig. 2) at their center mm section. The tines (250-mm diameter) were spaced apart by 0.7 mm. The specimens on the fixation unit were submerged into an isotonic saline bath, temperature controlled at 37 C for 5minutesbeforetesting(toallowfortemperaturestabilization), as well as during testing (to eliminate any temperature-related variability). 21 Shear rate was fixed to 4.16 mm/s. The following data, time (s), x and y size (mm), x and y displacement (mm), x and y force (mn), and x and y strain (unitless), were recorded every second. An integrated CCD camera captured still images at pixel resolution, analyzed by custom software (LabJoy v. 9.05; CellScale Biomaterials Testing). 22 Linear regression fitting using Microsoft Excel 2013 (version 15.0; Microsoft, Redmont, WA) was applied on the linear section of the loading curve to calculate, by means of the slope function (gradient), the stress strain ratio, an expression of the shear (Young) modulus E. To ensure proper linear relationship fit, a minimum value coefficient of determination (R 2 ) of the trend line of 0.98 was sought. Stress at 10% and 20% strain was recorded. Young shear modulus E was calculated as the gradient at the vicinity of 10% and 20% strain at the stress strain graph. The following definitions have been adopted: stress, a pressure metric, as the applied force F divided by cross-section A of the corneal specimen test area (F/A, units = mn/mm 2 =kpa=10 3 Pa; MPa = 10 6 Pa; where Pa is Pascal, SI unit for pressure), and strain as the unitless relative elongation (Dx/l), expressed as percentage. Crosssection of the test area was defined by 3.5-mm x (or y, respectively) width multiplied by the corneal thickness. FIGURE 2. A, Initial picture during stress strain measurement indicating the dual axis grip on the cornea specimen. B, The same cornea specimen at the completion of the test, 268 seconds later. The cornea was stretched in both directions (x and y) by mm. RESULTS In all procedures performed, the intralamellar button creation by FS200 femtosecond laser was without incident. Completion time, including corneal docking lamellar dissection by the laser was less than 1 minute in each case. The manual corneal button removal through the channel created, despite being a novel procedure, was also uncomplicated. Fig. 3 provides an example of postoperative optical coherence tomography imaging of a cornea in group A, indicating the hyperreflection line of the CXL, as well as corneal and epithelial pachymetry. Average preoperative central corneal thickness for study group A was mm versus mm for control group B (P = 0.721). Average postoperative central corneal thickness for study group A was mm versus mm for control group B (P = 0.875). Fig. 4 presents an example of axial curvature map and example of differences between postoperative and preoperative axial curvature, obtained by Scheimpflug imaging. Copyright 2015 Wolters Kluwer Health, Inc. All rights reserved. 3 Copyright ª 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. FIGURE 3. Top, optical coherence tomography derived meridional cross-section of CXL cornea, indicating a clear hyperreflection line along the cross-linked stroma. Bottom, optical coherence tomography derived corneal and epithelial pachymetry of the same cornea. The distinct data sets during each shear strength measurement were on average 273 (range, ) and the average displacement was 1129 mm (range, ). Mean maximum applied force was 2950 mn (range, ). Average value of maximum strain (relative elongation) was 27.5% (range, 24.3% 32.5%). The linear phase of analysis of the stress strain curves was sought for all corneas in both groups. The average number of distinct data pairs (y axis, stress and x axis, strain) within the linear phase in each analysis was 112 (range, ). The average coefficient of determination (R 2 ) of the trend line was (range, ). Table 1 summarizes results from the biomechanical data analysis. Stress at 10% strain was kpa in study group A versus kpa in control group B (relative difference D = 107%, P, 0.021). Stress at 20% strain was kpa in study group A versus kpa in control group B (relative difference D = 47%, P, 0.043). Average shear modulus in CXL group A at 10% strain was MPa versus MPa in control group B (relative difference D = +73%, P = 0.036). Average shear modulus in CXL group A at 20% strain was MPa versus MPa in control group B (relative difference D = +30%, P = 0.047). DISCUSSION The Dresden standard protocol CXL involves epithelial removal, superficial riboflavin instillation, and 365-nm UV-A 4 Copyright 2015 Wolters Kluwer Health, Inc. All rights reserved. Copyright ª 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

7 Cornea Volume 0, Number 0, Month 2015 In Situ CXL Corneal Biomechanics Kanellopoulos et al Cornea Volume 0, Number 0, Month 2015 FIGURE 4. Scheimpflug imaging axial curvature maps from a cornea in group A. Preoperative (left), postoperative (middle), and difference (right) maps. fluence of 5.4 J/cm, 1,23 aiming to address inherently biomechanically weakened thin corneas suffering from keratectasia. Subsequent modifications of this original technique, by maintaining intact epithelium (epithelium-on CXL), have been considered. The challenges faced in these alternative applications are related to epithelial permeability and thus insufficient/ inhomogeneous diffusion of riboflavin to the stroma, mainly because of the large molecular weight of riboflavin. 24,25 Even higher concentration (0.5%) riboflavin solutions may not facilitate improved results in comparison with the standard (0.1%) concentration. 26 The second issue is epithelial and Bowman membrane absorption/filtering of UV-A that could lead to lesser energy delivered to riboflavin-saturated stroma, as effective absorption by riboflavin at the superficial layers inhibits UV-A transmission. 27 Some studies indicate approximately onethird epithelial UV-A absorption, 24 whereas other studies suggest that human corneal epithelium and the underlying basement membrane absorb strongly only at wavelengths less than 310 nm. 28,29 The efficacy of multiple epithelium-on CXL techniques may warrant standardization and extensive investigation. Ex vivo evaluation of in situ (through a stromal pocket) CXL in porcine corneas (which do not have Bowman membrane) indicates that the biomechanical strengthening is reduced by approximately 50%, in comparison with standard TABLE 1. Biomechanical Comparative Measurements Between the 2 Groups Stress, kpa protocol CXL. 30 There is thus inconclusive evidence in the peer-reviewed literature on the aspect of efficacy of some epithelium-on CXL variations. 3,8,26,31,32 The in situ riboflavin application naturally overcomes the first of the 2 obstacles, related to riboflavin penetration through the intact epithelium and in acting as a blocking agent against UV-A propagation. This is because the epithelium and Bowman layer, that the UV-A light has to transcend, are not soaked in riboflavin. In this study, we conducted an ex vivo investigation to biomechanically address this issue, as in vivo biomechanical measurements, in our experience, have shown low specificity and sensitivity. We investigated the effect on corneal biomechanical properties by using objective biaxial stress strain measurements. This technique may be superior to corneal strip extensiometry used in earlier experiments, 33 considering the nonuniform topographic distribution of the corneal strength profile 34 due to collagen anisotropy. 35 To address the fact that corneal shear modulus naturally varies along different corneal meridians, the specimens were also randomly oriented during testing. It is known that the cornea is stronger at the anterior third. 36 This transverse depth dependence 37 suggests that tissue removal from the upper third may affect corneal rigidity the most. In our study, we demonstrated that the effective stromal rigidity increase in the cross-linked corneas was of the order of 100% at the 10% strain point, in Young s Shear Modulus, MPa 10% strain 20% strain 10% strain 20% strain D D D D Group A (CXL) Group B (control) D P D, Relative difference (%) between metrics; P, Student test P value. comparison with the non-cxl corneas that received the same treatment. This is an important finding, considering that the aim of in situ CXL was to exactly provide corneal stiffening to compensate for stromal weakening due to anterior stromal tissue removal, and thus to offer longterm corneal stability. The data in this study provide substantial ex vivo evidence that significant stromal strengthening may occur even when UV-A is projected through the intact corneal epithelium, Bowman membrane, and superficial stroma, to reach the underlying riboflavin-soaked stroma. These findings support our previous clinical investigation conducted in vivo of refractive stability in high myopic LASIK cases treated with prophylactic CXL, 12 as well as the compelling steepening refractive effect stabilization in clinical hyperopic LASIK cases treated with prophylactic CXL. Recent advances in variable fluence, customized pattern CXL may add another element to potential applications of this proposed technique: hyperopic style pattern in situ CXL may be combined to enhance cornea asphericity, overcorrection, and even address future presbyopia in myopic refractive lens extraction. 38 Myopic style pattern in situ CXL used in similar manner may address residual or regressing refractive error in high myopia refractive lens extraction. CONCLUSIONS Adjunct intrastromal cross-linking may be a valuable biomechanical modulator providing titratable increased rigidity in corneas undergoing intrastromal tissue removal such as femtosecond laser refractive intracorneal lenticular extraction. These data support our previous reported findings in LASIK combined with CXL. REFERENCES 1. Wollensak G, Spoerl E, Seiler T. Riboflavin/ultraviolet-a-induced collagen crosslinking for the treatment of keratoconus. Am J Ophthalmol. 2003;135: Hafezi F, Kanellopoulos J, Wiltfang R, et al. Corneal collagen crosslinking with riboflavin and ultraviolet A to treat induced keratectasia after laser in situ keratomileusis. J Cataract Refract Surg. 2007;33: Wollensak G, Iomdina E. Biomechanical and histological changes after corneal crosslinking with and without epithelial debridement. J Cataract Refract Surg. 2009;35: Spoerl E, Wollensak G, Seiler T. Increased resistance of crosslinked cornea against enzymatic digestion. Curr Eye Res. 2004;29: Arafat SN, Robert MC, Shukla AN, et al. UV cross-linking of donor corneas confers resistance to keratolysis. Cornea. 2014;33: Mazzotta C, Caporossi T, Denaro R, et al. Morphological and functional correlations in riboflavin UV A corneal collagen cross-linking for keratoconus. Acta Ophthalmol. 2012;90: Bakke EF, Stojanovic A, Chen X, et al. Penetration of riboflavin and postoperative pain in corneal collagen crosslinking: excimer laser superficial versus mechanical full-thickness epithelial removal. J Cataract Refract Surg. 2009;35: Stojanovic A, Chen X, Jin N, et al. Safety and efficacy of epithelium-on corneal collagen cross-linking using a multifactorial approach to achieve proper stromal riboflavin saturation. J Ophthalmol. 2012; 2012: Kanellopoulos AJ. Collagen cross-linking in early keratoconus with riboflavin in a femtosecond laser-created pocket: initial clinical results. J Refract Surg. 2009;25: Chayet AS, Assil KK, Montes M, et al. Regression and its mechanisms after laser in situ keratomileusis in moderate and high myopia. Ophthalmology. 1998;105: Kanellopoulos AJ, Asimellis G. Refractive, and keratometric stability in high myopic LASIK with high-frequency femtosecond and excimer lasers. J Refract Surg. 2013;29: Kanellopoulos AJ, Asimellis G. Epithelial remodeling after femtosecond laser-assisted high myopic LASIK: comparison of stand-alone with LASIK combined with prophylactic high-fluence cross-linking. Cornea. 2014;33: Kanellopoulos AJ, Asimellis G. Epithelial remodeling after partial topography-guided normalization and high-fluence short-duration crosslinking (Athens protocol): results up to 1 year. J Cataract Refract Surg. 2014;40: Kanellopoulos AJ. Long-term safety and efficacy follow-up of prophylactic higher fluence collagen cross-linking in high myopic laser-assisted in situ keratomileusis. Clin Ophthalmol. 2012;6: Cho M, Kanellopoulos AJ. Safety and efficacy of prophylactic ultraviolet- A-induced crosslinking after high risk myopic photorefractive keratotomy. Poster presented at: 5470/A441, ARVO; Fort Lauderdale, FL. May 3 7, 2009; FN. 16. Sekundo W, Kunert KS, Blum M. Small incision corneal refractive surgery using the small incision lenticule extraction (SMILE) procedure for the correction of myopia and myopic astigmatism: results of a 6 month prospective study. Br J Ophthalmol. 2011;95: Riau AK, Angunawela RI, Chaurasia SS, et al. Early corneal wound healing and inflammatory responses after refractive lenticule extraction (ReLEx). Invest Ophthalmol Vis Sci. 2011;52: Sinha Roy A, Dupps WJ Jr, Roberts CJ. Comparison of biomechanical effects of small-incision lenticule extraction and laser in situ keratomileusis: finite-element analysis. JCataractRefractSurg.2014;40: Vestergaard AH, Grauslund J, Ivarsen AR, et al. Central corneal sublayer pachymetry and biomechanical properties after refractive femtosecond lenticule extraction. J Refract Surg. 2014;30: Kanellopoulos AJ, Asimellis G. FS200 femtosecond laser LASIK flap digital analysis parameter evaluation: comparing two different types of patient interface applanation cones. Clin Ophthalmol. 2013;7: Hatami-Marbini H, Rahimi A. Effects of bathing solution on tensile properties of the cornea. Exp Eye Res. 2014;120: Søndergaard AP, Ivarsen A, Hjortdal J. Corneal resistance to shear force after UVA-riboflavin cross-linking. Invest Ophthalmol Vis Sci. 2013;54: Raiskup F, Spoerl E. Corneal crosslinking with riboflavin and ultraviolet A. I. Principles. Ocul Surf. 2013;11: Baiocchi S, Mazzotta C, Cerretani D, et al. Corneal crosslinking: riboflavin concentration in corneal stroma exposed with and without epithelium. J Cataract Refract Surg. 2009;35: Leccisotti A, Islam T. Transepithelial corneal collagen cross-linking in keratoconus. J Refract Surg. 2010;26: Stojanovic A, Zhou W, Utheim TP. Corneal collagen cross-linking with and without epithelial removal: a contralateral study with 0.5% hypotonic riboflavin solution. Biomed Res Int. 2014;2014: Bottós KM, Schor P, Dreyfuss JL, et al. Effect of corneal epithelium on ultraviolet-a and riboflavin absorption. Arq Bras Oftalmol. 2011;74: Kolozsvári L, Nógrádi A, Hopp B, et al. UV absorbance of the human cornea in the 240- to 400-nm range. Invest Ophthalmol Vis Sci. 2002;43: Podskochy A. Protective role of corneal epithelium against ultraviolet radiation damage. Acta Ophthalmol Scand. 2004;82: Armstrong BK, Lin MP, Ford MR, et al. Biological and biomechanical responses to traditional epithelium-off and transepithelial riboflavin-uva CXL techniques in rabbits. JRefractSurg.2013; 29: Pinelli R, Marzouky MM, El-Shawaf HI. Tensioactive-mediated transepithelial corneal crosslinking -first laboratory report. Eur Ophthalmic Rev. 2009;3: Copyright 2015 Wolters Kluwer Health, Inc. All rights reserved Copyright 2015 Wolters Kluwer Health, Inc. All rights reserved. Copyright ª 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. Copyright ª 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

8 Cornea Volume 0, Number 0, Month Kocak I, Aydin A, Kaya F, et al. Comparison of transepithelial corneal collagen crosslinking with epithelium-off crosslinking in progressive keratoconus. J Fr Ophtalmol. 2014;37: Elsheikh A, Anderson K. Comparative study of corneal strip extensometry and inflation tests. J R Soc Interf. 2005;2: Sánchez P, Moutsouris K, Pandolfi A. Biomechanical and optical behavior of human corneas before and after photorefractive keratectomy. J Cataract Refract Surg. 2014;40: Smolek MK. Interlamellar cohesive strength in the vertical meridian of human eye bank corneas. Invest Ophthalmol Vis Sci. 1993;34: In Situ CXL Corneal Biomechanics 36. Petsche SJ, Chernyak D, Martiz J, et al. Depth-dependent transverse shear properties of the human corneal stroma. Invest Ophthalmol Vis Sci. 2012;53: Dias J, Diakonis VF, Kankariya VP, et al. Anterior and posterior corneal stroma elasticity after corneal collagen crosslinking treatment. Exp Eye Res. 2013;116: Kanellopoulos AJ, Asimellis G. Hyperopic correction: clinical validation with epithelium-on and epithelium-off protocols, using variable fluence and topographically customized collagen corneal cross-linking. Clin Ophthalmol. 2014;8: Clinical Ophthalmology Open Access Full Text Article Comparison of prophylactic higher fluence corneal cross-linking to control, in myopic LASIK, one year results open access to scientific and medical research ORIGINAL R ESEARC H Journal nam Journal Des Year: 2014 Volume: 8 Running he Running he DOI: Copyright 2015 Wolters Kluwer Health, Inc. All rights reserved. 7 Copyright ª 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. Anastasios John Kanellopoulos 1,2 George Asimellis 1 Costas Karabatsas 1 1 LaserVision.gr Clinical and Research Eye Institute, Athens, Greece; 2 New York University Medical School, New York, NY, USA Correspondence: A John Kanellopoulos Laservision.gr Eye Institute, 17 Tsocha Street, Athens, , Greece Tel Fax ajk@brilliantvision.com Purpose: To compare 1-year results: safety, efficacy, refractive and keratometric stability, of femtosecond myopic laser-assisted in situ keratomileusis (LASIK) with and without concurrent prophylactic high-fluence cross-linking (CXL) (LASIK-CXL). Methods: We studied a total of 155 consecutive eyes planned for LASIK myopic correction. Group A represented 73 eyes that were treated additionally with concurrent prophylactic highfluence CXL; group B included 82 eyes subjected to the stand-alone LASIK procedure. The following parameters were evaluated preoperatively and up to 1-year postoperatively: manifest refractive spherical equivalent (MRSE), refractive astigmatism, visual acuity, corneal keratometry, and endothelial cell counts. We plotted keratometry measurements pre-operatively and its change in the early, interim and later post-operative time for the two groups, as a means of keratometric stability comparison. Results: Group A (LASIK-CXL) had an average postoperative MRSE of 0.23, 0.19, and 0.19 D for the 3-, 6-, and 12-month period, respectively, compared to D preoperatively. Flat keratometry was 37.69, 37.66, and D, compared to D preoperatively, and steep keratometry was 38.35, 38.36, and D, compared to D preoperatively. The predictability of Manifest Refraction Spherical Equivalent (MRSE) correction showed a correlation coefficient of Group B (stand-alone LASIK) had an average postoperative MRSE of 0.23, 0.20, and 0.27 D for the 3-, 6-, and 12-month period, respectively, compared with D preoperatively. Flat keratometry was 37.65, 37.89, and D, compared with D preoperatively, and steep keratometry was 38.32, 38.57, and D, compared with D preoperatively. The predictability of MRSE correction showed a correlation coefficient of The keratometric stability plots were stable for the LASIK CXL group and slightly regressing in the standard LASIK group, a novel stability evaluation metric that may escape routine acuity and refraction measurements. Conclusion: Application of prophylactic CXL concurrently with myopic LASIK surgery appears to contribute to improved refractive and keratometric stability compared to standard LASIK. The procedure appears safe and provides a new potential for LASIK correction. Keywords: myopic LASIK regression, femtosecond myopic LASIK, LASIK-CXL, LASIK- Xtra, high myopia, accelerated high-fluence collagen cross-linking Introduction Laser-assisted in situ keratomileusis (LASIK) is the most common form of refractive surgery, 1,2 offering predictable and stable refractive and visual outcomes. 3 Specifically, in correcting moderate to high myopia (equal or more than 6.00 D in the least-minus meridian), 4,5 there have been reports in the past indicating significant long-term regression. 6 8 The work by Alió et al 9 reported that 20.8% of high myopia cases required Clinical Ophthalmology 2014: Kanellopoulos et al. This work is published by Dove Medical Press Limited, and licensed under Creative Commons Attribution Non Commercial (unported, v3.0) License. The full terms of the License are available at Non-commercial uses of the work are permitted without any further permission from Dove Medical Press Limited, provided the work is properly attributed. Permissions beyond the scope of the License are administered by Dove Medical Press Limited. Information on how to request permission may be found at:

9 Kanellopoulos et al Prophylactic higher fluence corneal cross-linking in LASIK retreatment because of over- or undercorrection, or regression. Other studies have shown that the risk of regression may be between 5% and 27%. 10 Our team s experience with high myopia LASIK correction suggests a slight (0.5 D) long-term postoperative corneal steepening trend. 11 This was the motivation behind attempting to apply prophylactic in situ cross-linking (CXL) on the stromal bed, concurrently with the LASIK procedure, particularly in high-myopic eyes with thin residual stroma and in younger patients who may not yet have exhibited ectasia risk factors. 12,13 The application aims to enhance corneal rigidity and thus reduce the possibility of long-term myopic shift This study aimed to investigate potential differences in safety and efficacy, as well as in the refractive and keratometric results of myopic LASIK utilizing the Wavelight FS200 Femtosecond Laser (Alcon Laboratories Inc, Fort Worth, TX, USA) and the Wavelight EX500 Excimer Laser (Alcon Laboratories Inc) refractive surgery platforms. The study evaluated 1-year refractive and keratometric results from two groups, a LASIK-CXL and a standalone LASIK group, in which no concurrent CXL was applied. Materials and methods This prospective, observational, longitudinal study received approval by the Ethics Committee of LaserVision.gr Eye Institute and adhered to the tenets of the Declaration of Helsinki. Informed written consent was provided and documented from each subject at the time of the first study visit. The patients selected to receive the LASIK-CXL treatment were comprehensively informed of the benefits and risks involved in this procedure (which is Conformité Européene [CE]-marked in Europe). Inclusion and exclusion criteria A total of 155 consecutive patients enrolled for LASIK correction of primary myopia constituted this study. Group A (LASIK-CXL) consisted of 73 eyes in which concurrent prophylactic CXL was applied, while group B consisted of 82 eyes in which no such additional intervention was implemented (stand-alone LASIK). Only one eye was randomly selected (using randomization tables) from each patient to be included in the study, all of whom received bilateral surgery. All operations were performed by the same surgeon (AJK). The inclusion criteria comprised: no other previous ocular surgery, documented refractive stability for at least 3 years, and discontinuation of contact lens use for at least 2 weeks. Prior to intervention, a complete preoperative ophthalmologic evaluation ensured there was no present or past ocular pathology other than refractive error. The exclusion criteria comprised: systemic or ocular diseases, eyes with history of corneal dystrophy or herpetic eye disease, topographic evidence of ectatic corneal disorder, epithelial warpage from contact lens use, corneal scarring, glaucoma, severe dry eye, and collagen vascular disease. As per our clinical protocol for the last 6 years, in this study, the LASIK treatments for myopia were allocated to include LASIK-CXL if the patient had either of the following preoperative measurements: mean manifest refraction spherical equivalent (MRSE) over 5.00 D or keratometric astigmatism over 1.50 D on Scheimpflug-derived simulated keratometry, or predicted residual postoperative stromal thickness less than 330 m. No similar restrictions were applied in the inclusion criteria for the standard LASIK group B, other than applying it for myopia not exceeding D. Surgical technique In both groups, the FS200 Femtosecond Laser was employed to provide a corneal flap of 110 m thickness and 8.00 mm diameter. 17 The average pulse energy for the flap bed cut was 0.8 J, the side cut angle was 70, and the hinge position was superior. For the bed cut, the spot and line separations were 8 m. For the side cut, the spot separation was 5 m and line separation 3 m. The myopia ablation (6 mm to 6.50 mm ablation zone diameter) treatment was accomplished with the EX500 Excimer Laser. 18 Specifically for the LASIK-CXL, after the excimer laser ablation, and with the flap folded onto itself and protected with a dry Wexel sponge, one drop of Vibex Rapid (Avedro, Inc., Waltham, MA, USA), consisting of 0.10% saline-diluted riboflavin (a very slightly hypotonic solution, mixed with hydroxypropyl methylcellulose [HPMC], a dextran substitute), was placed on the exposed stromal bed afforded by the open LASIK flap and carefully spread over the bed area with an irrigating cannula for 60 seconds. It is important to avoid riboflavin immersion of the flap and its hinge for this purpose, the flap was protected, while remaining in a folded shape, as indicated in our earlier work. 14 The reason for this is to inhibit flap collagen CXL. However, a small amount of riboflavin absorption, and thus CXL, will inevitably occur as a result of osmosis during the (however short) ultraviolet A (UVA) exposure as the flap is in contact with the riboflavin-soaked stroma. One has to consider the following aspects: a riboflavin-presoaked flap will participate strongly in the UVA absorption (as it precedes the residual stroma along the illumination propagation path); however, it will not contribute any further to the corneal biomechanical stability and may negatively affect the postrefractive outcome, given that a 110 m thick flap has perhaps only a 60 m stromal (collagen) content. The application of CXL to such a thin stromal layer may lead to undesirable flap shrinking. Regarding collateral benefits, a CXL flap stromal interface might positively affect flap adherence. 19 Following stromal soaking, the flap was properly repositioned into place and the residual riboflavin-irrigated; then a UVA fluence of 30 mw/cm 2 was applied for 80 seconds (total energy 2.4 J/cm 2 ), provided by the KXL CXL system (Avedro, Inc.). The selection of the UV irradiation parameters (fluence and exposure time) was influenced by the following considerations: (a) provision of about half of the full treatment energy in comparison with the traditional CXL protocol, (b) minimization of UVA exposure in order to constrain CXL within the overlaying flap, and (c) minimization of flap dehydration and possible shrinkage. The superficial application of UVA following the in situ instillation of riboflavin was selected taking into account the following aspects: Application of CXL to the underlying stroma increases flap dehydration and potential predisposition for striae. CXL through the repositioned flap results in some riboflavin reflux in the dehydrated flap and CXL of the inner-flap collagen and the surface underlying the stroma. This may increase flap underlying stroma adherence and additionally, potentially reduce or eliminate the inadvertent space created between them (which, in postmortem standard LASIK histopathology, has been shown to be filled with amorphous deposits). CXL has well-known disinfection, if not antimicrobial activity; thus, conducting the CXL through a repositioned flap reduces the chance of flap contamination by airborne microorganisms or fomites in the operating room environment, and/or acts as an adjunct disinfectant of the LASIK procedure. Common to both groups, to avoid any risk of sameday accidental flap contact or rubbing, a bandage planar soft contact lens was then placed on the ocular surface, to be removed the following day. All patients were treated with moxifloxacin (Vigamox; Alcon Laboratories Inc) and 0.1% dexamethasone/chloramphenicol solution (Dispersadron C; Alcon Laboratories Inc) at least four times a day for 1 week. Data collection All eyes were measured for uncorrected distance visual acuity (UDVA), best (spectacle) corrected distance visual acuity (CDVA), and MRSE via manifest refraction and autorefraction measurements (Speedy-i, K Model Auto Refractometer/Keratometer; Nidek, Gamagori, Japan), visual acuity (Functional Vision Analyzer, Stereo Optical Co, Inc., Chicago, IL, USA), corneal topography, for steep and flat keratometry within the 3 mm radius, employing Placido topography (WaveLight Allegro Topolyzer Vario ; Alcon Laboratories Inc) and Scheimpflug imaging (WaveLight Oculyzer II Diagnostic Device; Alcon Laboratories Inc), and corneal thickness, employing the Oculyzer II and optical coherence tomography (OCT) (RTVue-100; Optovue, Inc, Fremont, CA, USA). 20 In addition to pachymetry, OCT was employed to provide meridional scans the hyperreflectivity lines are considered an indirect indication of CXL efficacy, 21 as demonstrated by our group s study, which showed CXL demarcation lines were evident as late as 3 years postoperatively. 22 An example of an OCT meridional image scan of a treated cornea is provided in Figure 1 and shows an increased reflectivity of the upper stromal bed. This is supportive of our argument that CXL mainly affects the underlying stroma and not the superjacent flap. The postoperative evaluation additionally included slitlamp examination, clinical evaluation of dry eye, indications of epithelial ingrowth, 23 and corneal haze. In addition, we measured endothelial cell counts preoperatively and 1-month postoperatively, employing noncontact specular microscopy (FA-3709; Konan Medical, Irvine, CA, USA). Postoperative examinations were conducted at 1 day, 1 week, 3 months, 6 months, and up to 1 year. This report presents the refractive and keratometric data analysis from the 1-year follow-up visits. Data were processed using web-based ophthalmic outcome analysis software (IBRA; Zubisoft GmbH, Oberhasli, Switzerland). 24 Descriptive statistics and analysis were performed using Minitab Figure 1 Anterior-segment optical coherency high-resolution cross-sectional (6 mm) image of an eye treated with LASIK-CXL for 2.25 D of sphere and 0.25 D of astigmatism, obtained 1-year postoperatively. Blue arrows indicate the LASIK flap, while yellow arrows indicate the stromal hyper-reflection line, which correlates with the depth of the prophylactic cross-linking effect. Abbreviations: CXL, cross-linking; LASIK, laser-assisted in situ keratomileusis Clinical Ophthalmology 2014:8 Clinical Ophthalmology 2014:8 2375

10 Kanellopoulos et al Prophylactic higher fluence corneal cross-linking in LASIK Statistical Software (Minitab Ltd., Coventry, UK) and Origin Lab 9 (OriginLab Corp, Northampton, MA, USA). P-values less than 0.05 were indicative of statistically significant results in this study. Results The 73 eyes included in group A (LASIK-CXL) belonged to 42 female and 31 male patients; of these, 36 eyes were right (OD) and 37 left (OS). The mean patient age at the time of operation was (range: 19 to 39) years. Preoperatively, the mean refractive error was sphere (range: 2.50 to 11.50) D, cylinder (range: 0.00 to 5.00) D, and the MRSE was (range: 2.50 to 11.50) D. Mean preoperative central corneal thickness was (range: 474 to 595) m and at 1-year postoperatively, was (range: 414 to 506) m. The 82 eyes in group B (stand-alone LASIK) belonged to 41 female and 41 male patients; 41 eyes were right (OD) and 41 left (OS). Mean patient age at the time of surgery was (range: 18 to 42) years. Preoperatively, mean refractive error was sphere (range: 2.50 to 9.50) D, cylinder (range: 0.00 to 4.50) D, and MRSE (range: 2.50 to 9.50) D. Mean preoperative central corneal thickness was (range: 503 to 592) m, and 1-year postoperatively, was (range: 423 to 510) m. Demographic and corneal thickness results are also reported in Table 1. No adverse events, including epithelial ingrowth, diffuse lamellar keratitis, postoperative haze, or other complications, were detected in either of the groups. Endothelial cell counts were not statistically different in either group, when compared preoperatively and 1-month postoperatively. UDVA outcome and stability The monocular UDVA outcome (Figure 2) indicates that in the LASIK-CXL group-a, 90.4% of the eyes had postoperative UDVA 20/20 (1.0 decimal) or better, and 94.5% had 20/25 (0.8 decimal) or better. In the stand-alone LASIK group, 85.4% of the eyes had postoperative UDVA better than 20/20 (1.0 decimal), and 89.0% had better than 20/25 (0.8 decimal). The differences between the two groups at the 20/20 and the 20/25 levels were statistically significant (P and P 0.037, respectively). Efficacy of CDVA The gain loss data (preoperative CDVA versus postoperative UDVA) (Figure 3) indicate that in the LASIK-CXL group, 35.6% of the eyes were unchanged, 56.2% gained one Snellen line, and 8.2% (six eyes) gained two or more Snellen lines. No eye lost any line. In the stand-alone LASIK group, 37.8% of the eyes were unchanged, 56.1% gained one Snellen line, and 4.9% (four eyes) gained two or more lines. Only 1.2% (one eye) lost one line. Refractive predictability and accuracy The predictability results are presented in the form of linear regression scatterplots (Figure 4), in which the vertical axis corresponds to the achieved MRSE, and the horizontal axis corresponds to the attempted MRSE. The data for the LASIK-CXL group had a coefficient of determination (r 2 ) of 0.979, while for the stand-alone group-b, this was The postoperative MRSE refraction results are presented, within 0.50 D intervals, in Figure 5. In the LASIK-CXL group, MRSE refraction between 0.50 and 0.00 D was achieved in 82.2% of the eyes, and in the stand-alone group, this was achieved in 81.7% (no statistically significant difference [P 0.079]). Postoperative refractive astigmatism results, within intervals of 0.50 D representing the accuracy of the cylinder correction, are illustrated in Figure 6. The LASIK-CXL group had a mean preoperative cylinder of D, while the stand-alone LASIK group had a mean preoperative Table 1 Preoperative demographics, including the planned residual stromal thickness, between the two groups Group A (LASIK-CXL): 65 eyes Age (years) Pre-op CCT ( m) Planned residual ( m) Post-op CCT ( m) Group B (stand-alone LASIK): 75 eyes Age (years) Pre-op CCT ( m) Planned residual ( m) Post-op CCT ( m) Mean Std dev Min Max Notes: CCT and planned residual stroma were reported in m. Post-op results refer to the 1-year results. Abbreviations: CCT, Central corneal thickness; CXL, cross-linking; LASIK, laser-assisted in situ keratomileusis; Max, maximum; Min, minimum; Std dev, standard deviation. A Cumulative % of eyes 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% LASIK-CXL group A 73 eyes (plano target) 12-months post-op 0.0% Post-op UDVA Pre-op CDVA 0.0% 20/ % 12.3% 90.4% 74.0% 94.5% 86.3% 97.3% 95.9% 100.0% 98.6% 20/16 20/20 20/25 20/32 20/40 Cumulative Snellen visual acuity 20/x or better Figure 2 Postoperative uncorrected distance visual acuity (blue columns) versus preoperative corrected distance visual acuity (red columns) 1-year postoperatively, in (A) the LASIK-CXL group and (B) the stand-alone LASIK group. Abbreviations: CDVA, corrected distance visual acuity; CXL, cross-linking; LASIK, laser-assisted in situ keratomileusis; UDVA, uncorrected distance visual acuity. cylinder of D. The LASIK-CXL group had, postoperatively, 90.4% of eyes with less than 0.25 D of refractive astigmatism, and mean cylinder of D. The stand-alone LASIK group had 91.5% with less than 0.25 D of refractive astigmatism, and mean cylinder of D. Refractive and keratometric stability Refractive stability was demonstrated by the MRSE correction, as followed during the 1-, 3-, 6-, and 12-month postoperative visits (Figure 7). The 1-year mean postoperative MRSE was D in the LASIK-CXL group and D in the stand-alone LASIK group. These findings indicate a reduced refractive shift in the LASIK-CXL group in comparison with the stand-alone group (P 0.063). The keratometric stability, demonstrated by the K-flat and % of eyes 60% 50% 40% 30% 20% 10% 0% 12-months post-op Lost 3 or more LASIK-CXL group A, 73 eyes (%) Stand-alone LASIK group B, 82 eyes (%) 1.2% 0.0% 0.0% 0.0% 0.0% 0.0% 37.8% 35.6% 56.2% 56.1% Figure 3 Change in corrected visual acuity, as a percentage of eyes with gain/loss in Snellen lines of corrected distance visual acuity 1-year postoperatively. Abbreviations: CDVA, corrected distance visual acuity; CXL, cross-linking; LASIK, laser-assisted in situ keratomileusis. 6.8% 3.7% 1.4% 1.2% Lost 2 Lost 1 Unchanged Gained 1Gained 2Gained 3 or more Change in Snellen lines of CDVA B Cumulative % of eyes 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Group B stand-alone LASIK 82 eyes (plano target) 12-months post-op Post-op UDVA Pre-op CDVA 0.0% 0.0% 20/ % 11.0% 85.4% 69.5% 89.0% 85.4% 95.1% 95.1% Cumulative Snellen visual acuity 20/x or better 100.0% 98.8% 20/16 20/20 20/25 20/32 20/40 K-steep average values up to the 1-year postoperative visit, is illustrated in Figure 8. The results indicate an increased keratometric stability in the LASIK-CXL group (1-year at 0.03 D in the flat and 0.05 D in the steep compared A Achieved spherical equivalent refraction (D) B Achieved spherical equivalent refraction (D) S R-Sq R-Sq (adj) Group A LASIK-CXL 73 eyes, 1-year post-op Achieved (D) = attempted (D) Regression 95% CI 95% PI % 97.9% Attempted spherical equivalent refraction (D) Group B LASIK stand-alone 82 eyes, 1-year post-op Achieved (D) = attempted (D) S R-Sq R-Sq (adj) Regression 95% CI 95% PI % 97.0% Attempted spherical equivalent refraction (D) Figure 4 Predictability of spherical equivalent correction, measured at 1-year postoperatively, showing achieved spherical equivalent (vertical axis) versus attempted spherical equivalent (horizontal axis), in (A) the LASIK-CXL group and (B) the stand-alone LASIK group. Abbreviations: CDVA, corrected distance visual acuity; CI, confidence interval; CXL, cross-linking; PI, prediction interval; S, sum of residuals; R-sq, coefficient of determination; R-sq (adj), adjusted coefficient of determination Clinical Ophthalmology 2014:8 Clinical Ophthalmology 2014:8 2377

11 % of eyes 70% 12-months post-op LASIK-CXL group A, 73 eyes 60% Stand-alone LASIK group B, 82 eyes 50% 40% 30% 20% 10% 0% 6.8% 0.0% 0.0% 9.6% 8.5% 7.3% 0.0% 1.2% 2.00 to 1.01 < to to to % 1.2% 0.0% 0.0% to >1.00 Postoperative spherical equivalent refraction (D) Figure 5 Postoperative spherical equivalent refraction for both groups, 1-year postoperatively. Abbreviations: CXL, cross-linking; LASIK, laser-assisted in situ keratomileusis. with the 1-month baseline) compared with the stand-alone LASIK group ( 0.67 D and 0.55 D, respectively), which was statistically significant (P 0.039). Discussion Improved diagnostics, ablation profiles, and laser-beam tracking refinements of the LASIK procedure, 25 and improvements attributed to femtosecond laser-assisted flap creation26 28 all have contributed to an excellent track record in myopia correction. However, refractive regression in high myopic corrections, remains a possibility. There are several possible mechanisms leading to post-lasik regression. For example, a correlation between increased epithelial thickness and myopia correction up to 1-year postoperatively was noted, in a study by Spadea et al29 in high-myopic (myopias between 8.50 and D) patients. This epithelial thickening in A B #!" # % ' #!&" ( $! % $ "!$%!"!$%!"! % $ %! %! %! %! %! %! %! Refractive astigmatism (D) "# # " ' # # # # # # # ( Refractive astigmatism (D) Group A, LASIK-CXL, 65 eyes Group B, stand-alone LASIK, 75 eyes Months following surgery 12 Figure 7 Stability of manifest spherical equivalent refraction for both groups, expressed in diopters (D), up to 1-year postoperatively. Abbreviations: CXL, cross-linking; LASIK, laser-assisted in situ keratomileusis; SD, standard deviation. Comparison of the stability results between the two groups indicates that in the stand-alone group, there was a slight positive slope in the keratometric readings, both at the flat and the steep meridian, as illustrated in Figure 8, which is suggestive of a mild progressive corneal steepening. The recorded changes correspond to 0.67 D for the flat meridian and 0.54 D for the steep meridian. The data clearly show a trend toward mild corneal steepening in the long-term postoperative period. A similar refractive shift has been reported previously by our team in LASIK corrections of high myopia with no prophylactic CXL application.11 There was no such trend of keratometric shift in the LASIK-CXL group ( 0.03 D and 0.05 respectively). Other differences between the two groups were the slightly increased stability of the MRSE (Figure 7), as well as the improved predictability (Figure 4), despite the larger range of attempted correction and increased preoperative astigmatism LASIK-CXL group, 65 eyes up to 12-months post-op (Figure 6). It is worth noting that the mean spherical error (S), as well as cylinder error (C) treated in group A (mean S 6.62 D, maximum S D; mean C 1.35 D, maximum C 5.25 D) was significantly greater than in the LASIK standalone group (mean S 5.05 D, maximum S 9.50 D; mean C 0.85 D, maximum C 3.50 D). Despite the apparently more challenging cases included in group A (LASIK-CXL) compared to group B (Standard-LASIK), the refractive results in the LASIK-CXL group were equally good and, in some cases, slightly better. One aspect that needs consideration is the possibility of refractive flattening as a result of the CXL applied. Our clinical experience, as well as the peer-review literature, suggests the continued progression of the CXL effect over time.33 We have indicated that the long-term keratometry flattening progression in the fully cross-linked corneas is of the order of 0.30 D. One has to acknowledge the following two parameters that differentiate this finding, when considering the LASIK-CXL: u The keratoconus management cases were fundamentally unstable ectatic corneas, whereas in the present work only healthy corneas were included, and u The keratoconus management cases received the full energy treatment (up to 6 J/cm2), whereas in the present work, the LASIK-CXL eyes received only a partial energy treatment (2.4 J/cm2), corresponding to less than half of the standard protocol energy. It may thus be estimated that the possibility of long-term keratometric flattening may well be restricted. Additional long-term studies are required to investigate this aspect. In view of the expressed skepticism by colleagues regarding possible regression of the refractive effect,34 sometimes, despite low preoperative risk (for example, classified as low B 47!! # % #! $ & " # " "# 1 A % of eyes % of eyes high myopia has been corroborated recently in femtosecond LASIK correction of high myopia.30 Furthermore, a comparison of stand-alone LASIK to LASIK-CXL in high myopia verified that the observed post-lasik epithelial thickening changes are significantly less prominent in LASIK-CXL cases. This difference may correlate with higher regression rates and/or may depict increased biomechanical instability in stand-alone LASIK.31 We believe that this finding is a manifestation of the same aspect, which is the chief difference between the two groups the application of preventive CXL in this group. As we have previously demonstrated,32 CXL affects epithelial thickness, leading to reduced overall thickness. In the present study, we investigated up to 1-year postoperative refractive and stability results of 155 eyes subjected to femtosecond-laser LASIK for myopia between two groups group A in which prophylactic high-fluence CXL was incorporated and group B that received stand-alone LASIK. The two groups in the study were by all other means matched: ablation zone, flap thickness, surgeon, lasers employed, and postoperative medication and treatment. Our research on this subject matter is ongoing, and we do hope to report long-term follow-up results in the future. The postoperative evaluation in the LASIK-CXL group did not indicate clinical or topographic evidence of complications in comparison with the stand-alone group. Visual rehabilitation between the two groups, as expressed by CDVA and contrast sensitivity evaluation, was at similar levels, without induction of side effects or compromise of visual safety. The refractive outcome, predictability, and stability were completely satisfactory, and in some cases superior to standard LASIK (for example, data shown in in Figure 2, subgroup of achieved visual acuity of 20/16). 3URSK\ODFWLF KLJKHU ÁXHQFH FRUQHDO FURVV OLQNLQJ LQ /$6,. 0 Time after surgery (months) Keratometric readings (D) 82.2% 81.7% 80% Keratometric readings (D) 90% Mean ' SD spherical equivalent refraction (D) Kanellopoulos et al 47 Stand-alone LASIK group, 75 eyes up to 12-months post-op Time after surgery (months) 35 Preoperative 1-month 3-months 6-months 12-months Preoperative 1-month 3-months 6-months 12-months K-flat K-flat K-steep K-steep Figure 6 Refractive astigmatism preoperatively (red columns) and 1-year postoperatively (blue columns), in (A) the LASIK-CXL group and (B) the stand-alone LASIK group. Notes: The graph shows percentage of eyes (vertical axis) versus refractive astigmatism (horizontal axis). Abbreviations: CXL, cross-linking; LASIK, laser-assisted in situ keratomileusis. Figure 8 Stability of corneal keratometry for (A) the LASIK-CXL group and (B) the stand-alone LASIK group, expressed in diopters (D), up to 1-year postoperatively. Abbreviations: CXL, cross-linking; LASIK, laser-assisted in situ keratomileusis Clinical Ophthalmology 2014:8 Clinical Ophthalmology 2014:8 2379

12 Kanellopoulos et al Prophylactic higher fluence corneal cross-linking in LASIK risk in the Ectasia Risk Score System), 35 caution has been recommended when LASIK was performed in cases of high myopia with a thin residual stromal bed. 36 Our team has performed and reported a large number of successful LASIK-CXL procedures over the last 8 years, and we view this prophylactic treatment as a pivotal biomechanical enhancement of the LASIK procedure in a young adult, in any patient under the age of 30 years with high myopia and/or astigmatism, and any patient with a difference of over 0.50 D in the amount of the astigmatism between the two eyes. In our opinion, it does not merit leniency with preoperative form-fruste keratoconus criteria. All cases studied in this work, as well as in previous reports by our group, have been screened thoroughly for any signs of tomographic cornea irregularity. Conclusion LASIK combined with a prophylactic CXL intervention appears to provide predictability as well as refractive and keratometric stability. The data reported in this study provide evidence of the safety and efficacy of this approach. The adjuvant CXL procedure adds enhanced corneal biomechanical stability. High-myopic and younger age LASIK cases, may require biomechanical re-enforcement, as means of reducing the incidence and degree of future myopic regression and/or the potential ectasia risk. Author contributions AJK design and conduct of the study, and collection and interpretation of the data. AJK performed data management, and CK and GA analyzed the data. AJK, CK and GA prepared the manuscript and AJK, CK, and GA participated in manuscript review. All authors provided manuscript approval. Disclosure AJK has held consultant or advisory positions with Alcon/ WaveLight, Allergan, Avedro, and i-optics. The other authors report no other conflicts of interest in this work. References 1. Solomon KD, Fernández de Castro LE, Sandoval HP, et al; Joint LASIK Study Task Force. LASIK world literature review: quality of life and patient satisfaction. Ophthalmology. 2009;116(4): Shortt AJ, Allan BD, Evans JR. Laser-assisted in-situ keratomileusis (LASIK) versus photorefractive keratectomy (PRK) for myopia. Cochrane Database Syst Rev. 2013;1:CD Shortt AJ, Bunce C, Allan BD. Evidence for superior efficacy and safety of LASIK over photorefractive keratectomy for correction of myopia. Ophthalmology. 2006;113(11): Liu Z, Li Y, Cheng Z, Zhou F, Jiang H, Li J. Seven-year follow-up of LASIK for moderate to severe myopia. J Refract Surg. 2008;24(9): Güell JL, Muller A. Laser in situ keratomileusis (LASIK) for myopia from 7 to 18 diopters. J Refract Surg. 1996;12(2): Oruçoğlu F, Kingham JD, Kendüşim M, Ayoğlu B, Toksu B, Göker S. Laser in situ keratomileusis application for myopia over minus 14 diopter with long-term follow-up. Int Ophthalmol. 2012;32(5): Magallanes R, Shah S, Zadok D, et al. Stability after laser in situ keratomileusis in moderately and extremely myopic eyes. J Cataract Refract Surg. 2001;27(7): Chayet AS, Assil KK, Montes M, Espinosa-Lagana M, Castellanos A, Tsioulias G. Regression and its mechanisms after laser in situ keratomileusis in moderate and high myopia. Ophthalmology. 1998;105(7): Alió JL, Muftuoglu O, Ortiz D, et al. Ten-year follow-up of laser in situ keratomileusis for myopia of up to 10 diopters. Am J Ophthalmol. 2008;145(1): Chen YI, Chien KL, Wang IJ, et al. An interval-censored model for predicting myopic regression after laser in situ keratomileusis. Invest Ophthalmol Vis Sci. 2007;48(8): Kanellopoulos AJ, Asimellis G. Refractive and keratometric stability in high myopic LASIK with high-frequency femtosecond and excimer lasers. J Refract Surg. 2013;29(12): Binder PS. Analysis of ectasia after laser in situ keratomileusis: risk factors. J Cataract Refract Surg. 2007;33(9): Randleman JB. Post-laser in-situ keratomileusis ectasia: current understanding and future directions. Curr Opin Ophthalmol. 2006; 17(4): Kanellopoulos AJ. Long-term safety and efficacy follow-up of prophylactic higher fluence collagen cross-linking in high myopic laser-assisted in situ keratomileusis. Clin Ophthalmol. 2012;6: Celik HU, Alagöz N, Yildirim Y, et al. Accelerated corneal crosslinking concurrent with laser in situ keratomileusis. J Cataract Refract Surg. 2012;38(8): Kanellopoulos AJ, Pamel GJ. Review of current indications for combined very high fluence collagen cross-linking and laser in situ keratomileusis surgery. Indian J Ophthalmol. 2013;61(8): Kanellopoulos AJ, Asimellis G. Digital analysis of flap parameter accuracy and objective assessment of opaque bubble layer in femtosecond laser-assisted LASIK: a novel technique. Clin Ophthalmol. 2013; 7: Kanellopoulos AJ, Asimellis G. Long-term bladeless LASIK outcomes with the FS200 Femtosecond and EX500 Excimer Laser workstation: the Refractive Suite. Clin Ophthalmol. 2013;7: Mi S, Dooley EP, Albon J, Boulton ME, Meek KM, Kamma-Lorger CS. Adhesion of laser in situ keratomileusis-like flaps in the cornea: Effects of crosslinking, stromal fibroblasts, and cytokine treatment. J Cataract Refract Surg. 2011;37(1): Kanellopoulos AJ, Asimellis G. Comparison of high-resolution Scheimpflug and high-frequency ultrasound biomicroscopy to anteriorsegment OCT corneal thickness measurements. Clin Ophthalmol. 2013; 7: Seiler T, Hafezi F. Corneal cross-linking-induced stromal demarcation line. Cornea. 2006;25(9): Kanellopoulos AJ, Asimellis G. Introduction of quantitative and qualitative cornea optical coherence tomography findings induced by collagen cross-linking for keratoconus: a novel effect measurement benchmark. Clin Ophthalmol. 2013;7: Henry CR, Canto AP, Galor A, Vaddavalli PK, Culbertson WW, Yoo SH. Epithelial ingrowth after LASIK: clinical characteristics, risk factors, and visual outcomes in patients requiring flap lift. J Refract Surg. 2012;28(7): Zuberbuhler B, Galloway P, Reddy A, Saldana M, Gale R. A webbased information system for management and analysis of patient data after refractive eye surgery. Comput Methods Programs Biomed. 2007;88(3): Maldonado MJ, Nieto JC, Piñero DP. Advances in technologies for laser-assisted in situ keratomileusis (LASIK) surgery. Expert Rev Med Devices. 2008;5(2): Vestergaard A, Ivarsen A, Asp S, Hjortdal JØ. Femtosecond (FS) laser vision correction procedure for moderate to high myopia: a prospective study of ReLEx( ) flex and comparison with a retrospective study of FSlaser in situ keratomileusis. Acta Ophthalmol. 2013;91(4): Kymionis GD, Kankariya VP, Plaka AD, Reinstein DZ. Femtosecond laser technology in corneal refractive surgery: a review. J Refract Surg. 2012;28(12): Kanellopoulos AJ, Asimellis G. Three-dimensional LASIK flap thickness variability: topographic central, paracentral and peripheral assessment, in flaps created by a mechanical microkeratome (M2) and two different femtosecond lasers (FS60 and FS200). Clin Ophthalmol. 2013;7: Spadea L, Fasciani R, Necozione S, Balestrazzi E. Role of the corneal epithelium in refractive changes following laser in situ keratomileusis for high myopia. J Refract Surg. 2000;16(2): Kanellopoulos AJ, Asimellis G. Longitudinal postoperative LASIK epithelial thickness profile changes in correlation with degree of myopia correction. J Refract Surg. 2014;30(3): Kanellopoulos AJ, Asimellis G. Epithelial remodeling after femtosecond laser-assisted high myopic LASIK: comparison of stand-alone with LASIK combined with prophylactic high-fluence cross-linking. Cornea. 2014;33(5): Clinical Ophthalmology Publish your work in this journal Clinical Ophthalmology is an international, peer-reviewed journal covering all subspecialties within ophthalmology. Key topics include: Optometry; Visual science; Pharmacology and drug therapy in eye diseases; Basic Sciences; Primary and Secondary eye care; Patient Safety and Quality of Care Improvements. This journal is indexed on Submit your manuscript here: Kanellopoulos AJ, Aslanides IM, Asimellis G. Correlation between epithelial thickness in normal corneas, untreated ectatic corneas, and ectatic corneas previously treated with CXL; is overall epithelial thickness a very early ectasia prognostic factor? Clin Ophthalmol. 2012;6: Kanellopoulos AJ, Asimellis G. Keratoconus management: longterm stability of topography-guided normalization combined with high-fluence CXL stabilization (the Athens Protocol). J Refract Surg. 2014;30(2): Ambrósio R, Dawson DG, Salomão M, Guerra FP, Caiado AL, Belin MW. Corneal ectasia after LASIK despite low preoperative risk: tomographic and biomechanical findings in the unoperated, stable, fellow eye. J Refract Surg. 2010;26(11): Randleman JB, Trattler WB, Stulting RD. Validation of the Ectasia Risk Score System for preoperative laser in situ keratomileusis screening. Am J Ophthalmol. 2008;145(5): Spadea L, Cantera E, Cortes M, Conocchia NE, Stewart CW. Corneal ectasia after myopic laser in situ keratomileusis: a long-term study. Clin Ophthalmol. 2012;6: PubMed Central and CAS, and is the official journal of The Society of Clinical Ophthalmology (SCO). The manuscript management system is completely online and includes a very quick and fair peer-review system, which is all easy to use. Visit testimonials.php to read real quotes from published authors Clinical Ophthalmology 2014:8 Clinical Ophthalmology 2014:8 2381

13 ARTICLE 1598 EPITHELIAL REMODELING IN KERATOCONIC EYES Epithelial remodeling after partial topography-guided normalization and high-fluence short-duration crosslinking (Athens protocol): Results up to 1 year Anastasios John Kanellopoulos, MD, George Asimellis, PhD PURPOSE: To compare epithelial remodeling in keratoconic eyes that had photorefractive keratectomy and corneal collagen crosslinking (Athens protocol) with that in untreated keratoconic eyes and healthy eyes. SETTING: Private clinical practice, Athens, Greece. DESIGN: Comparative case series. METHODS: Fourier-domain anterior segment optical coherence tomography (AS-OCT) was used to obtain in vivo 3-dimensional epithelial thickness maps and center, superior, inferior, maximum, minimum, mean, midperipheral, and variability data. RESULTS: Group A comprised 175 treated keratoconic eyes (Athens protocol); Group B, 193 untreated keratoconic eyes; and Group C, 160 healthy eyes. The 1-year mean center epithelial thickness in Group A was mm G 7.36 (SD) (range 33 to 64 mm). At the first clinical visit, it was G 6.80 mm (range 36 to 72 mm) in Group B and G 3.23 mm (range 45 to 59 mm) in Group C. The mean thickness range in Group A at 1 year was G 7.21 mm (range 6 to 34 mm). It was G mm (range 4 to 66 mm) in Group B and 6.86 G 3.33 mm (range 3 to 29 mm) in Group C. The mean topographic thickness variability in Group A at 1 year was 4.64 G 1.63 mm (range 1.6 to 8.1 mm) (P<.05). It was 5.77 G 3.39 mm (range 1.3 to 17.8 mm) in Group B and 1.59 G 0.79 mm (range 0.6 to 5.6 mm) in Group C. CONCLUSION: Anterior segment OCT indicated a thinner and more homogeneous remodeled epithelium in the keratoconic eyes treated using the Athens protocol. Financial Disclosure: Dr. Kanellopoulos is a consultant to Alcon Surgical, Inc.; Wavelight Laser Technologie AG; Avedro, Inc.; and i-optics Optikger ate GmbH. Dr. Asimellis has no financial or proprietary interest in any material or method mentioned. J Cataract Refract Surg 2014; 40: Q 2014 ASCRS and ESCRS We previously reported overall reduced corneal epithelial thickness in keratoconic eyes that were treated with (1) excimer laser debridement of the top 50 mm of the epithelium, (2) partial topographyguided excimer ablation, and (3) immediate highfluence ultraviolet-a radiation (10 mw/cm 2 ) and short-duration (10 minutes) corneal collagen crosslinking (CXL) with riboflavin in a procedure known as the Athens protocol. 1 3 Our goal was to arrest the keratectasia progression 4 and provide a less irregular anterior corneal surface. In l study, 1 which was performed using high-frequency scanning ultrasound biomicroscopy (UBM), the epithelial thickness in a group of untreated keratoconic eyes was compared with that in a group of keratoconic eyes treated using the Athens protocol. Epithelial thickness assessment has been facilitated by the development of anterior segment optical coherence tomography (AS-OCT). 5 Although there are studies of AS-OCT epithelium measurement in the peer-reviewed literature, until recently and to our knowledge, the methodology and instrumentation mainly used an on-screen caliper tool 6 ; thus, only local point thickness measurements were reported. The recent availability of in vivo, 3-dimensional (3-D) corneal epithelial mapping by AS-OCT in clinical practice 7 allows easy capture of optical images and highspeed measurements conferred by Fourier-domain signal processing. 8,9 This study used this new clinical modality to evaluate the longitudinal postoperative changes in epithelial thickness distribution as well as the epithelial layer topographic variability in a large group of keratoconic cases treated using the Athens protocol. The results in these eyes were compared with those in untreated keratoconic eyes and in healthy control eyes. PATIENTS AND METHODS This observational comparative prospective study received approval by the Ethics Committee, LaserVision.gr Institute, and adhered to the tenets of the Declaration of Helsinki. All patients provided informed written consent at the time of the first clinical visit. Exclusion criteria were systemic disease, previous corneal surgery, history of chemical injury or delayed epithelial healing, and pregnancy or lactation. Patient Enrollment and Surgical Technique Group A In Group A, eyes were treated for keratoconus with the Athens protocol. All procedures were performed by the same surgeon (A.J.K.) using an EX500 excimer laser 10 (Alcon Surgical, Inc.) with topography-guided custom partial ablation. Figure 1 shows an example of treatment planning, the distribution of the ablation depth, a preoperative axial curvature map, a postoperative axial curvature map, and the difference in axial curvature map between preoperatively and postoperatively. Immediately after surface normalization, accelerated CXL was applied using the KXL System (Avedro, Inc.). The patients were followed for up to 1 year. Group B Group B comprised eyes with keratoconus that had not received surgical treatment. Inclusion criteria were a clinical diagnosis of progressive keratoconus (confirmed by a complete ophthalmologic evaluation), minimum age 17 years, and corneal thickness of at least 300 mm. The keratoconus diagnosis was further confirmed using the Wavelight Oculyzer II (Alcon Surgical, Inc.) and the Pentacam high-resolution Scheimpflug imaging camera 11 (Oculus Optikger ate GmbH). Submitted: November 11, Final revision submitted: February 4, Accepted: February 5, From Laservision.gr Eye Institute (Kanellopoulos), Athens, Greece, and the New York University Medical School (Kanellopoulos, Asimellis), New York, New York, USA. Corresponding author: Anastasios John Kanellopoulos, MD, Laservision.gr Clinical and Research Institute, 17 Tsocha Street, Athens, Greece, Postal Code: ajk@brilliantvision. com. Figure 1. A: Preoperative tomographic (anterior corneal instantaneous [tangential]) map obtained via Scheimpflug imaging. Keratometric power reported in diopters. B: Corresponding postoperative map. C: Treatment planning showing ablation depth (in mm) and topographic distribution obtained via the refractive platform software. D: Difference between postoperative sagittal curvature map (B) and preoperative sagittal curvature map (A) (N Z nasal; T Z temporal). Group C Group C, the control group, comprised unoperated normal eyes with no current or past ocular pathology other than refractive error and no present irritation or dryeye disorder, all of which were confirmed during a complete ophthalmologic evaluation. Contact lens wearers were excluded from this group. Imaging Instrumentation The RTVue-100 Fourier-domain AS-OCT system (Optovue, Inc.), running on analysis and report software version A6 (9.0.27), was used in the study. Data output included total corneal and epithelial thickness maps corresponding to a 6.0 mm diameter area. In all cases, to avoid potential artifacts (eg, due to eyedrop instillation), OCT imaging preceded the ocular clinical examination and was performed by the same trained investigator. The settings were as follows: L-Cam lens and 8 radial meridional B-scans per acquisition consisting of 1024 A-scans each with a 5 mm axial resolution. These 8 radial meridional scans, all acquired in less than 0.5 second, were used by the system software to produce by interpolation 3-D thickness maps. Images with quality greater than 30, determined using the signal strength index parameter, were considered in the study. The signal strength index parameter measures the average signal strength across the scan. Two consecutive individual acquisitions were obtained in each case (eye) to ensure data validity; the mean value of 2 was used in this study. Data Collection and Statistical Analysis In Group A, the postoperative measurements were performed at 1 and 6 months as well as at 1 year. Imaging in Q 2014 ASCRS and ESCRS Published by Elsevier Inc /$ - see front matter J CATARACT REFRACT SURG - VOL 40, OCTOBER 2014

14 EPITHELIAL REMODELING IN KERATOCONIC EYES EPITHELIAL REMODELING IN KERATOCONIC EYES Figure 2. Comparative AS-OCT epithelial thickness (mm) 3-D maps shows an image from Group A taken 1 year postoperatively and an image from Group B (I Z inferior; IN Z inferior nasal; IT Z inferior temporal; N Z nasal; S Z superior; SN Z superior nasal; ST Z superior temporal; T Z temporal). Group B and Group C was performed during the first clinical visit. The main analysis report produced by the AS-OCT system displayed total corneal (reported as pachymetry) and epithelial 3-D thickness maps covering the 6.0 mm diameter area. Corneal pachymetry was assessed by the central corneal thickness (CCT) and minimum corneal thickness. Epithelial thickness assessment comprised the following measurements: pupil center, superior, inferior, minimum, maximum, mean, peripheral, topographic thickness variability, and epithelial thickness range. These data were collected as follows (Figure 2): Each thickness map was divided into 17 sections (2.0 mm diameter pupil center disk of mm 2 area; 8 sectors [octants] within the annulus between the 2.0 mm and 5.0 mm zones, each of 8.24 mm 2 areas; and 8 sectors [octants] within the annulus of the 5.0 to 6.0 mm zones, each of 4.32 mm 2 areas). For each of these sections, the mean epithelial thickness was displayed numerically in integer form with a minimum difference of 1 mm over the corresponding area. In this study, the reported center epithelium thickness was taken from the integer indication over the center 2.0 mm disk. The mean epithelial thickness was computed by the mean of all segments, and the peripheral epithelial thickness was computed by the mean of the thickness corresponding to 18 equispaced points along the 5.0 mm radius (data harvested by mouse-over indication over the epithelial thickness map). The superior, inferior, minimum, maximum, and topographic epithelial thickness variability (computed by the standard deviation [SD] of the 17 thickness values) were provided in tabular form by the software of the AS-OCT device (Figure 2). The thickness range was computed as follows: minimum epithelial thickness maximum epithelial thickness. Descriptive statistics, linear regression analysis to look for possible correlations, paired analysis t tests, and analysis of variance were performed using Minitab software (version , Minitab, Ltd.) and Origin Lab software (version 9, Originlab Corp.). Paired-analysis P values less than 0.05 were considered an indication of statistically significant results. RESULTS Table 1 shows the CCT, minimum corneal thickness, epithelial thicknesses, topographic thickness variability, and epithelial thickness range measured by AS-OCT in the 3 groups. Group A (Athens protocol) comprised 175 eyes, 74 of women and 101 of men. The mean patient age at the time of surgery was 26.8 years G 7.2 (SD) (range 18 to 48 years). There were 87 right eyes and 88 left eyes. The Athens protocol treatment was uneventful in all cases. Group B (untreated keratoconic) comprised 193 eyes, 92 of women and 101 of men. The mean patient age at the time of examination was 31.1 G 9.9 years (range 18.0 to 51.0 years). There were 91 right eyes and 102 left eyes. Group C (control) comprised 160 eyes, 67 of women and 93 of men. The mean patient age at the time of examination was G 9.55 years (range 18.0 to 52.0 years). There were 74 right eyes and 86 left eyes. Epithelial Thickness In Group A, the difference in the center epithelial thickness between each postoperative timepoint was statistically significant (all P!.05). The difference in the mean center epithelial thickness ( 4.31 mm; 95% confidence interval [CI], 6.31 to 2.30) between Group A 1 year after treatment and Group B at the time of examination was statistically significant (P!.05, 2-sample t test). The difference in the mean center epithelial thickness ( 4.75 mm, 95% CI, 6.59 to 2.92) between Group A 1 year after treatment and Group C at the time of examination was also statistically significant (P!.05) (Figure 3). In Group A, the difference in topographic thickness variability between each postoperative timepoint was statistically significant (all P!.05). Figure 4 shows the epithelial thickness variability and range by group. DISCUSSIONS Until recently, high-frequency UBM had been the gold standard for in vivo corneal epithelial 3-D imaging. 12 The recent, rapid development and current highspeed imaging capabilities of AS-OCT have made acquisition of in vivo 3-D pachymetry corneal maps reliable and fast Software refinement also enables clinical assessment of corneal asymmetry and focal thinning parameters for keratoconus classification. 20 In addition, the higher axial resolution, increased accuracy, and finer image-processing capabilities of the current AS-OCT imaging systems have enabled, among other things, 3-D imaging of epithelial thickness. 7 Epithelial thickness and irregularity indices (eg, center and mean epithelial thickness, epithelial thickness topographic irregularity, and thickness range) Table 1. Central corneal thickness (CCT), minimum corneal thickness, epithelial thicknesses, topographic thickness variability, and epithelial thickness range measured by the AS-OCT in the 3 groups. All units are in microns. Group CCT MinCT measured quantitatively with AS-OCT can serve as possible indicators of cornea instability, including ectasia and keratoconus. 14 In this study, we evaluated these parameters with a Fourier-domain AS-OCT system in a large group of keratoconic patients who had combined treatment of excimer laser anterior surface normalization and simultaneous high-fluence accelerated CXL. This study adds new information based on its large group of treated keratoconic eyes and its comparison with untreated keratoconic eyes and healthy eyes. In addition, our study was performed with a commercially available AS-OCT system whose use may become more widespread in clinical settings. Our findings confirm compensatory epithelial thickness changes previously described after various refractive corneal ablation procedures. 21,22,A The epithelial thickness and irregularity assessment in the Athens protocol treated Group A suggests short-term variability in corneal thickness distribution between the third month and the sixth Epithelial Thickness Center Superior Inferior Minimum Maximum Mean Mid Topographic Variability Group A (AP treated) 1 month Mean SD Max Min months Mean SD Max Min year Mean SD Max Min Group B (KCN) Mean SD Max Min Group C (healthy) Mean SD Max Min AT Z Athens Protocol; CCT Z central corneal thickness; KCN Z keratoconus, no treatment; Mid Z midperipheral; MinCT Z minimum corneal thickness Range month. 3 Specifically, in our study, the corneal and epithelial thickness distributions were characterized by large deviations that gradually became less irregular. The mean SD at the center epithelial thickness of 7.36 mm at 1 month gradually decreased to 6.80 at 3 months and to 4.57 mm at 1 year. The SD in Group B (untreated keratoconic) and in Group C (control) was 6.79 mm and 3.23 mm, respectively. In addition to fluctuating less between different eyes, the epithelial thickness in Group A progressed toward a reduced mean topographic variability and mean thickness range (from 5.43 mm and mm to 4.64 mm and mm, respectively). Both metrics were more regular than in Group B (5.77 mm and mm, respectively). These results indicate that the epithelial thickness distribution was more uniform in the Athens protocol treated group than in the untreated group of keratoconic eyes and had less overall thickness, as suggested by the reduced mean and center thickness values. J CATARACT REFRACT SURG - VOL 40, OCTOBER 2014 J CATARACT REFRACT SURG - VOL 40, OCTOBER 2014

15 EPITHELIAL REMODELING IN KERATOCONIC EYES EPITHELIAL REMODELING IN KERATOCONIC EYES Figure 3. Mean and center epithelial thicknesses in the 3 groups. Error bars correspond to the SD (KCN Z keratoconus, no treatment). The findings in the current study agree with those in our previous study 1 ; that is, although an overall thicker epithelium with large variations can be observed clinically and topographically in eyes with keratoconus, in eyes treated with CXL the variability in epithelium thickness and topographic thickness decreased by a statistically significant margin and was more uniform. We have theorized that epithelial hyperplasia in biomechanically unstable corneas (ie, increased epithelial regrowth activity) might be associated with a more elastic cornea. 1 The laboratory and clinical findings of increased corneal rigidity after CXL are widely accepted, including in studies of accelerated highfluence CXL. 26 In conclusion, we present the results in a comprehensive study of the postoperative development of corneal epithelial thickness distribution after keratoconus management using combined anterior corneal normalization by topography-guided excimer ablation and accelerated CXL. The epithelial healing processes can be monitored by AS-OCT with ease in a clinical setting, expanding the clinical application of this technology. Our findings suggest less topographic variability and overall reduced epithelial thickness distribution in keratoconus eyes treated with CXL using the Athens protocol. Figure 4. Epithelial thickness variability and range in the 3 groups. Error bars correspond to the SD (KCN Z keratoconus, no treatment). WHAT WAS KNOWN Postoperative epithelial remodeling after partial anterior surface normalization with an excimer laser and highfluence CXL, assessed with high-frequency scanning UBM, results in reduced overall epithelial thickness and topographic variability. WHAT THIS PAPER ADDS Detailed follow-up of Athens protocol treated eyes up to 1 year confirmed previous ultrasound findings of the overall thinner and smoother epithelial thickness profiles compared with the profiles of untreated keratoconic eyes. REFERENCES 1. Kanellopoulos AJ, Aslanides IM, Asimellis G. Correlation between epithelial thickness in normal corneas, untreated ectatic corneas, and ectatic corneas previously treated with CXL; is overall epithelial thickness a very early ectasia prognostic factor? Clin Ophthalmol 2012; 6: Available at: pdf. Accessed June 11, Kanellopoulos AJ. Long term results of a prospective randomized bilateral eye comparison trial of higher fluence, shorter duration ultraviolet A radiation, and riboflavin collagen cross linking for progressive keratoconus. Clin Ophthalmol 2012; 6: Available at: articles/pmc /pdf/opth pdf. Accessed June 11, Kanellopoulos AJ, Asimellis G. Keratoconus management: longterm stability of topography-guided normalization combined with high fluence CXL stabilization (the Athens Protocol). J Refract Surg 2014; 30: Kanellopoulos AJ, Asimellis G. Introduction of quantitative and qualitative cornea optical coherence tomography findings inducedbycollagencross-linking for keratoconus: anoveleffect measurement benchmark. Clin Ophthalmol 2013; 7: Available at: /pdf/opth pdf. Accessed June 11, Kanellopoulos AJ, Asimellis G. In vivo three-dimensional corneal epithelium imaging in normal eyes by anteriorsegment optical coherence tomography: a clinical reference study. Cornea 2013; 32: Francoz M, Karamoko I, Baudouin C, Labbe A. Ocular surface epithelial thickness evaluation with spectral-domain optical coherence tomography. Invest Ophthalmol Vis Sci 2011; 52: Available at: 12/9116.full.pdf. Accessed June 11, Kanellopoulos AJ, Asimellis G. In vivo 3-dimensional corneal epithelial thickness mapping as an indicator of dry eye: preliminary clinical assessment. Am J Ophthalmol 2014; 157: Chen TC, Cense B, Pierce MC, Nassif N, Park BH, Yun SH, White BR, Bouma BE, Tearney GJ, de Boer JF. Spectral domain optical coherence tomography; ultra-high speed, ultra-high resolution ophthalmic imaging. Arch Ophthalmol 2005; 123: Oh W-Y, Vakoc BJ, Shishkov M, Tearney GJ, Bouma BE. O400 khz repetition rate wavelength-swept laser and application to high-speed optical frequency domain imaging. Opt Lett 2010; 35: Available at: articles/pmc /pdf/nihms pdf.accessedjune11, Kanellopoulos AJ, Asimellis G. Long term bladeless LASIK outcomes with the FS200 femtosecond and EX500 excimer laser workstation: the Refractive Suite. Clin Ophthalmol 2013; 7: Available at: PMC /pdf/opth pdf. Accessed June 11, Kanellopoulos AJ, Asimellis G. Correlation between central corneal thickness, anterior chamber depth, and corneal keratometryasmeasured by Oculyzer II and WaveLightOB820 inpreoperative cataract surgery patients. J Refract Surg 2012; 28: Reinstein DZ, Gobbe M, Archer TJ, Silverman RH, Coleman DJ. Epithelial, stromal, and total corneal thickness in keratoconus: three-dimensional display with Artemis very-high frequency digital ultrasound. J Refract Surg 2010; 26: Available at: pdf. Accessed June 11, Potsaid B, Baumann B, Huang D, Barry S, Cable AE, Schuman JS, Duker JS, Fujimoto JG. Ultrahigh speed 1050nm swept source/fourier domain OCT retinal and anterior segment imaging at 100,000 to 400,000 axial scans per second. Opt Exp 2010; 18: Available at: nih.gov/pmc/articles/pmc /pdf/nihms pdf. Accessed June 11, Rocha KM, Perez-Straziota CE, Stulting RD, Randleman JB. SD-OCT analysis of regional epithelial thickness profiles in keratoconus, postoperative corneal ectasia, and normal eyes. J Refract Surg 2013; 29: Ge L, Yuan Y, Shen M, Tao A, Wang J, Lu F. The role of axial resolution of optical coherence tomography on the measurement of corneal and epithelial thicknesses. Invest Ophthalmol Vis Sci 2013; 54: Available at: content/54/1/746.full.pdf. Accessed June 11, Wojtkowski M, Kaluzny B, Zawadzki RJ. New directions in ophthalmic optical coherence tomography. Optom Vis Sci 2012; 89: Available at: optvissci/fulltext/2012/05000/new_directions_in_ophthalmic_ Optical_Coherence.4.aspx. Accessed June 11, Khurana RN, Li Y, Tang M, Lai MM, Huang D. High-speed optical coherence tomography of corneal opacities. Ophthalmology 2007; 114: Li Y, Shekhar R, Huang D. Corneal pachymetry mapping with high-speed optical coherence tomography. Ophthalmology 2006; 113: e2 19. Dutta D, Rao HL, Addepalli UK, Vaddavalli PK. Corneal thickness in keratoconus; comparing optical, ultrasound, and optical coherence tomography pachymetry. Ophthalmology 2013; 120: Kanellopoulos AJ, Chiridou M, Asimellis G. Optical coherence tomography-derived corneal thickness asymmetry indices: clinical reference study in 561 normal eyes. In press, J Cataract Refract Surg Gauthier CA, Holden BA, Epstein D, Tengroth B, Fagerholm P, Hamberg-Nystr om H. Role of epithelial hyperplasia in regression following photorefractive keratectomy. Br J Ophthalmol 1996; 80: Available at: 6/545.full.pdf. Accessed June 11, Erie JC, Patel SV, McLaren JW, Ramirez M, Hodge DO, MaguireLJ, BourneWM. Effect of myopiclaser in situkeratomileusison epithelialandstromalthickness; aconfocalmicroscopy study. Ophthalmology 2002; 109: Raiskup-Wolf F, Hoyer A, Spoerl E, Pillunat LE. Collagen crosslinking with riboflavin andultraviolet-alight in keratoconus: longterm results. J Cataract Refract Surg 2008; 34: Vinciguerra P, Albe E, Trazza S, Rosetta P, Vinciguerra R, Seiler T, Epstein D. Refractive, topographic, tomographic, and aberrometric analysis of keratoconic eyes undergoing corneal cross-linking. Ophthalmology 2009; 116: O Brart DPS, Kwong TQ, Patel P, McDonald RJ, O Brart NA. Long-term follow-up of riboflavin/ultraviolet A (370 nm) corneal collagen cross-linking to halt the progression of keratoconus. Br J Ophthalmol 2013; 97: Available at: bmj.com/content/97/4/433.full.pdf. Accessed June 11, Schumacher S, Oeftiger L, Mrochen M. Equivalence of biomechanical changes induced by rapid and standard corneal crosslinking, using riboflavin and ultraviolet radiation. Invest Ophthalmol Vis Sci 2011; 52: Available at: content/52/12/9048.full.pdf. Accessed June 11, 2014 OTHER CITED MATERIAL A. Reinstein DZ, Aslanides IM, Patel S, Silverman RH, Coleman DJ, Epithelial Lenticular Types of Human Cornea: Classification and Analysis of Influence on PRK, poster presented at the annual meeting of the American Academy of Ophthalmology, Atlanta, Georgia, USA, October Abstract available in: Ophthalmology 1995; 102(suppl):S156 First author: Anastasios John Kanellopoulos, MD Laservision.gr Clinical and Research Institute, Athens, Greece J CATARACT REFRACT SURG - VOL 40, OCTOBER 2014 J CATARACT REFRACT SURG - VOL 40, OCTOBER 2014

16 ORIGINAL ARTICLE The Athens Protocol/Kanellopoulos & Asimellis Keratoconus Management: Long-Term Stability of Topography-Guided Normalization Combined With High-Fluence CXL Stabilization (The Athens Protocol) Anastasios John Kanellopoulos, MD; George Asimellis, PhD ABSTRACT PURPOSE: To investigate refractive, topometric, pachymetric, and visual rehabilitation changes induced by anterior surface normalization for keratoconus by partial topography-guided excimer laser ablation in conjunction with accelerated, high-fluence cross-linking. METHODS: Two hundred thirty-one keratoconic cases subjected to the Athens Protocol procedure were studied for visual acuity, keratometry, pachymetry, and anterior surface irregularity indices up to 3 years postoperatively by Scheimpflug imaging (Oculus Optikgeräte GmbH, Wetzlar, Germany). RESULTS: Mean visual acuity changes at 3 years postoperatively were ± 0.31 (range: to +1.10) for uncorrected distance visual acuity and ± 0.21 (range: to +0.90) for corrected distance visual acuity. Mean K1 (flat meridian) keratometric values were ± 3.83 diopters (D) (range: to D) preoperatively, ± 3.97 D (range: to D) 1 month postoperatively, and ± 3.80 D (range: to D) up to 3 years postoperatively. The average Index of Surface Variance was ± (range: 17 to 208) preoperatively and ± (range: 7 to 190) up to 3 years postoperatively. The average Index of Height Decentration was ± µm (range: to µm) preoperatively and ± µm (range: to µm) up to 3 years postoperatively. Mean thinnest corneal thickness was ± µm (range: 297 to 547 µm) preoperatively, ± µm (range: 196 to 480 µm) 1 month postoperatively, and ± µm (range: 218 to 500 µm) up to 3 years postoperatively. CONCLUSIONS: The Athens Protocol to arrest keratectasia progression and improve corneal regularity demonstrates safe and effective results as a keratoconus management option. Progressive potential for long-term flattening validates using caution in the surface normalization to avoid overcorrection. [J Refract Surg. 2014;30(2):88-92.] K eratoconus is a degenerative bilateral, noninflammatory disorder characterized by ectasia, thinning, and irregular corneal topography. 1 The disorder usually has onset at puberty and often progresses until the third decade of life, may manifest asymmetrically in the two eyes of the same patient, and can present with unpredictable visual acuity, particularly in relation to corneal irregularities. 2 One of the acceptable options 3 for progressive keratoconus management is corneal collagen cross-linking (CXL) with riboflavin and ultraviolet-a. 4 To further improve the topographic and refractive outcomes, CXL can be combined with customized anterior surface normalization. 5-7 Our team has developed a procedure 8,9 we have termed the Athens Protocol, 10 involving sequentially excimer laser epithelial debridement (50 µm), partial topography-guided excimer laser stromal ablation, and high-fluence ultraviolet-a irradiation (10 mw/cm 2 ), accelerated (10, or minutes) CXL. Early results 11 and anterior segment optical coherence tomography quantitative findings 12 are indicative of the long-term stability of the procedure. Detailed studies on postoperative visual rehabilitation and anterior surface topographic changes by such combined CXL procedures are rare, particularly those reporting results longer than 1 year. This study aims to investigate safety and efficacy of the Athens Protocol procedure by analysis of longterm (3-year) refractive, topographic, pachymetric, and visual rehabilitation changes on clinical keratoconus management with the Athens Protocol in a large number of cases. PATIENTS AND METHODS This clinical study received approval by the Ethics Committee of our Institution and adhered to the tenets of the Declaration From Laservision.gr Eye Institute, Athens, Greece (AJK, GA); and New York University School of Medicine, New York, New York (AJK). Submitted: April 6, 2013; Accepted: August 21, 2013; Posted online: January 31, 2014 Dr. Kanellopoulos is a consultant for Alcon/WaveLight. The remaining author has no financial or proprietary interest in the materials presented herein. Correspondence: Anastasios John Kanellopoulos, MD, 17 Tsocha str. Athens, Greece Postal Code ajk@brilliantvision.com doi: / x of Helsinki. Written informed consent was obtained from each participant at the time of the intervention or the first clinical visit. PATIENT INCLUSION CRITERIA Two hundred thirty-one consecutive keratoconic cases subjected to the Athens Protocol procedure between 2008 and 2010 were investigated. All procedures were performed by the same surgeon (AJK) using the Alcon/WaveLight 400 Hz Eye-Q 17 or the EX500 excimer lasers. 18 Inclusion criteria were clinical diagnosis of progressive keratoconus, minimum age of 17 years, and corneal thickness of at least 300 µm. All participants completed an uneventful Athens Protocol procedure and all 231 eyes were observed for up to 3 years. Exclusion criteria were systemic disease, previous eye surgery, chemical injury or delayed epithelial healing, and pregnancy or lactation (female patients). MEASUREMENTS AND ANALYSIS Postoperative evaluation included uncorrected distance visual acuity (UDVA), manifest refraction, corrected distance visual acuity (CDVA) with this refraction, and slit-lamp biomicroscopy for clinical signs of CXL. 12 For the quantitative assessment of the induced corneal changes, postoperative evaluation was performed by a Pentacam Scheimpflug imaging device (Oculyzer II, WavLight AG, Erlangen, Germany) 19 and processed via Examination Software (Version 1.17r47). Specific anterior surface irregularity indices provided by the Scheimpflug imaging analysis were evaluated in addition to keratometric and pachymetric values. These indices are employed in grading and classification based on the Amsler and Krumeich criteria. 20,21 These are the Index of Surface Variance, an expression of anterior surface curvature irregularity, and the Index of Height Decentration, calculated with Fourier analysis of corneal height (expressed in microns) to TABLE 1 Visual Acuity Data (N = 231 Eyes) a Postoperative Value Preop 1 Month 3 Months 6 Months 12 Months 24 Months 36 Months UDVA Average SD (±) ±0.20 ±0.27 ±0.29 ±0.29 ±0.28 ±0.28 ±0.28 Gain/loss n/a CDVA Average SD (±) ±0.23 ±0.22 ±0.20 ±0.20 ±0.19 ±0.19 ±0.19 Gain/loss n/a preop = preoperative; UDVA = uncorrected distance visual acuity; SD = standard deviation; n/a = not applicable; CDVA = corrected distance visual acuity a Expressed as the difference of postoperative minus preoperative values (gain/loss). Units are decimal. quantify the degree of cone decentration. 22 Index of Surface Variance and Index of Height Decentration were computed for the 8-mm diameter zone. Descriptive and comparative statistics, analysis of variance, and linear regression were performed by Minitab version (MiniTab Ltd., Coventry, UK) and Origin Lab version 9 (OriginLab Corp., Northampton, MA). Paired analysis P values less than.05 were considered statistically significant. Visual acuity is reported decimally and keratometry in diopters (D). Results are reported as mean ± standard deviation and range as minimum to maximum. RESULTS The 231 eyes enrolled belonged to 84 female and 147 male participants. Mean participant age at the time of the operation was 30.1 ± 7.5 years (range: 17 to 57 years). All eyes were followed up to the 3-year follow-up. VISUAL ACUITY CHANGES Table 1 presents preoperative and postoperative UDVA and CDVA. UDVA increased by ± 0.31 (range: to +1.10) and CDVA by ± 0.21 (range: to +0.90). Figure 1 illustrates, in the form of box plots, visual acuity gained or lost 3 years postoperatively. These graphs (the 95% median confidence range box) indicate that 95% of the cases had at least +0.1 increase in UDVA and 95% had positive change in CDVA. KERATOMETRIC AND ANTERIOR SURFACE INDICES PROGRESSION Anterior keratometry continued to flatten over the 3-year follow-up (Figure 2). Descriptive statistics for anterior keratometry, preoperatively and postoperatively, are presented in Table A (available in the online version of this article). The values for the Index of Surface Variance and the Index of Height Decentration continued to 88 Copyright SLACK Incorporated Journal of Refractive Surgery 89

17 The Athens Protocol/Kanellopoulos & Asimellis The Athens Protocol/Kanellopoulos & Asimellis TABLE 2 Thinnest Corneal Thickness Measured by the Scheimpflug Device (N = 231 Eyes) (µm) Postoperative Value Preoperative 1 Month 3 Months 6 Months 12 Months 24 Months 36 Months Average SD ±40.02 ±53.90 ±56.41 ±53.64 ±55.70 ±57.84 ±58.21 Maximum Minimum SD = standard deviation Figure 1. Change (gain/loss) in visual acuity, expressed as the difference of postoperative minus preoperative values (expressed decimally), showing median level (indicated by (+), average symbol (x), 95% median confidence range box (red borderline boxes), and interquartile intervals range box (black borderline boxes). Figure 3. Anterior surface topometric indices Index of Surface Variance (no units) and Index of Height Decentration (units µm) as measured by the Scheimpflug imaging device (Oculyzer II, WavLight AG, Erlangen, Germany) preoperatively, 1 month postoperatively, and up to 36 months postoperatively. decrease over time (Figure 3, Table B, available in the online version of this article). PACHYMETRIC PROGRESSION The thinnest corneal decreased as a result of excimer laser ablation but then stabilized over time without additional thinning (Table 2). DISCUSSION Many reports describe the effects of CXL with or without same-session excimer laser ablation corneal normalization. 3 There is general consensus that the intervention strengthens the cornea, helps arrest the ectasia progression, and improves corneal keratometric values, refraction, and visual acuity. The key question is the long-term stability of these induced changes. For example, is the cornea inactive after Figure 2. Anterior keratometry (K1 flat and K2 steep) as measured by the Scheimpflug device (Oculyzer II, WavLight AG, Erlangen, Germany) preoperatively up to 3 years postoperatively. All units in keratometric diopters (D). the intervention and, if not, is there steepening or flattening and/or thickening or thinning? These issues are even more applicable in the case of the Athens Protocol, due to the partial corneal surface ablation. Ablating a thin, ectatic cornea may sound unorthodox. However, the goal of the topography-guided ablation is to normalize the anterior cornea and thus help improve visual rehabilitation to a step beyond that a simple CXL would provide. This study aims to address some of the above issues. The large sample and follow-up time permit sensitive analysis with confident conclusion of postoperative efficacy. We monitored visual acuity changes and for the quantitative assessment we chose to standardize on one Scheimpflug screening device and to focus on key parameters of visual acuity, keratometry, pachymetry, and anterior surface indices. 23 All of these parameters reflect changes induced by the procedure and describe postoperative progression. However, variations in the two anterior surface indices (Index of Surface Variance and Index of Height Decentration) may provide a more valid analysis than keratometry and visual function. 24 Our results indicate that the apparent disadvantage of thinning the cornea is balanced by a documented long-term rehabilitating improvement and synergy from the CXL component. VISUAL ACUITY CHANGES Based on our results, the Athens Protocol appears to result in postoperative improvement in both UDVA and CDVA. Average gain/loss in visual acuity was consistently positive, starting from the first postoperative month, with gradual and continuous improvement toward the 3-year visit. These visual rehabilitation improvements appear to be superior to those reported in cases of simple CXL treatment. 25 However, it is noted that the visual acuity presented with large variations. The standard deviation of UDVA was ±0.20 preoperatively and ±0.28 postoperatively. Likewise, the standard deviation of CDVA was ±0.23 preoperatively and ±0.20 postoperatively. We theorize that the reason for visual acuity in keratoconic cases having such large fluctuations (and often being unexpectedly good) can be attributed to a multifocal and soft (ie, adaptable) cornea, in addition to advanced neural processing in the individual visual system. However, these advantages are essentially negated with CXL treatment, which stiffens the cornea. Over time, possibly due to further topography improvement and adaptation to the partially normalized cornea, a noteworthy improvement in visual acuity is observed. KERATOMETRIC AND ANTERIOR SURFACE INDICES PROGRESSION After the 1-month visit, keratometric values are reduced. This progressive potential for long-term flattening has been clinically observed in many cases over at least 10 years. Peer-reviewed reports on this matter have been rare and only recent. 26,27 The two anterior surface indices, Index of Surface Variance and Index of Height Decentration, also demonstrated postoperative improvement. A smaller value is indication of corneal normalization (lower Index of Surface Variance, less irregular surface, lower Index of Height Decentration, cone less steep and more central). These changes are therefore suggestive of corneal topography improvement, in agreement with other smaller sample studies. 13 Such changes in Index of Surface Variance and Index of Height Decentration have been reported only recently. 28 The initial more drastic change of the Index of Height Decentration can be justified by the chief objective of surface normalization, cone centering, 6 which is noted even by the first month. The subsequent surface normalization, as also indicated by keratometric flattening, suggests further anterior surface improvement. PACHYMETRIC PROGRESSION As expected by the fact that Athens Protocol includes a partial stromal excimer ablation, there is reduction of postoperative corneal thickness, manifested by the thinnest corneal thickness. What seemed to be a surprising result is that the cornea appears to rebound, by gradually thickening, up to 3 years postoperatively. Postoperative corneal thickening after the 1-month lowest thickness baseline has also been discussed recently. 29,30 In another report, 31 the lowest thinnest corneal thickness was noted at the 3-month interval. In that study of 82 eyes treated only with CXL, the average cornea thickened by +24 µm after 1 year compared to the 3-month baseline. In our study of 212 eyes treated with the Athens Protocol procedure, the cornea thickening rate after the baseline first postoperative month was approximately half (+12 µm over the first year), in agreement with a recent publication. 29 Therefore, it is possible that stromal changes initiated by the CXL procedure are not just effective in halting ectasia, but are prompting corneal surface flattening and thickening, which appears to be longer lasting than anticipated. We note, however, that CXL alone results not just in corneal reshaping, but also in stromal density and refractive index, both possibly influencing the reported thickness by the Scheimpflug device. Therefore, true corneal thickness differences may not be accurately explored by Scheimpflug imaging due to the principle of operation (densitometry), which has been our clinical experience with abnormal density corneas (ie, corneal scars or arcus senilis corneas). Further studies of corneal thickness chances by modalities, such as anterior segment optical coherence tomography, which currently also measure epithelial thickness, 32 may be warranted. In addition, corneal biomechanical analysis and corneal volume studies may be necessary to further validate such findings. Our study indicates a significant improvement in all parameters studied. The changes induced by the procedure indicate a consistent trend toward improved visual rehabilitation, corneal flattening (validating ectasia arrest), and anterior surface improvement. The Athens Protocol procedure demonstrates impressive refractive, keratometric, and topometric results. Progressive potential for long-term flattening documented 90 Copyright SLACK Incorporated Journal of Refractive Surgery 91

18 The Athens Protocol/Kanellopoulos & Asimellis in this study suggests employment of caution in the surface normalization process to avoid overcorrection. AUTHOR CONTRIBUTIONS Study concept and design (AJK, GA); data collection (AJK, GA); analysis and interpretation of data (AJK, GA); drafting of the manuscript (GA); critical revision of the manuscript (AJK, GA); statistical expertise (GA); administrative, technical, or material support (AJK); supervision (AJK) REFERENCES 1. Gordon-Shaag A, Millodot M, Shneor E. The epidemiology and etiology of keratoconus. Int J Keratoco Ectatic Corneal Dis. 2012;1: Katsoulos C, Karageorgiadis L, Mousafeiropoulos T, Vasileiou N, Asimellis G. Customized hydrogel contact lenses for keratoconus incorporating correction for vertical coma aberration. Ophthalmic Physiol Opt. 2009;29: Chan E, Snibson GR. Current status of corneal collagen cross-linking for keratoconus: a review. Clin Exp Optom. 2013;96: Kanellopoulos AJ. Collagen cross-linking in early keratoconus with riboflavin in a femtosecond laser-created pocket: initial clinical results. J Refract Surg. 2009;25: Kanellopoulos AJ, Binder PS. Collagen cross-linking (CCL) with sequential topography-guided PRK: a temporizing alternative for keratoconus to penetrating keratoplasty. Cornea. 2007;26: Kanellopoulos AJ. Comparison of sequential vs same-day simultaneous collagen cross-linking and topography-guided PRK for treatment of keratoconus. J Refract Surg. 2009;25:S812-S Labiris G, Giarmoukakis A, Sideroudi H, Gkika M, Fanariotis M, Kozobolis V. Impact of keratoconus, cross-linking and crosslinking combined with photorefractive keratectomy on self-reported quality of life. Cornea. 2012;31: Kanellopoulos AJ, Binder PS. Management of corneal ectasia after LASIK with combined, same-day, topography-guided partial transepithelial PRK and collagen cross-linking: the athens protocol. J Refract Surg. 2011;27: Ewald M, Kanellopoulos J. Limited topography-guided surface ablation (TGSA) followed by stabilization with collagen crosslinking with UV irradiation and ribofl avin (UVACXL) for keratoconus (KC). Invest Ophthalmol Vis Sci. 2008;49:E-Abstract Kanellopoulos AJ. Long term results of a prospective randomized bilateral eye comparison trial of higher fluence, shorter duration ultraviolet A radiation, and riboflavin collagen cross linking for progressive keratoconus. Clin Ophthalmol. 2012;6: Krueger RR, Kanellopoulos AJ. Stability of simultaneous topography-guided photorefractive keratectomy and riboflavin/ UVA cross-linking for progressive keratoconus: case reports. J Refract Surg. 2010;26:S827-S Kanellopoulos AJ, Asimellis G. Introduction of quantitative and qualitative cornea optical coherence tomography findings, induced by collagen cross-linking for keratoconus; a novel effect measurement benchmark. Clin Ophthalmol. 2013;7: Greenstein SA, Fry KL, Hersh PS. Corneal topography indices after corneal collagen crosslinking for keratoconus and corneal ectasia: one-year results. J Cataract Refract Surg. 2011;37: Guedj M, Saad A, Audureau E, Gatinel D. Photorefractive keratectomy in patients with suspected keratoconus: five-year follow-up. J Cataract Refract Surg. 2013;39: Alessio G, L abbate M, Sborgia C, La Tegola MG. Photorefractive keratectomy followed by cross-linking versus cross-linking alone for management of progressive keratoconus: two-year follow-up. Am J Ophthalmol. 2013;155: Hersh PS, Greenstein SA, Fry KL. Corneal collagen crosslinking for keratoconus and corneal ectasia: one-year results. J Cataract Refract Surg. 2011;37: Kanellopoulos AJ. Topography-guided hyperopic and hyperopic astigmatism femtosecond laser-assisted LASIK: long-term experience with the 400 Hz eye-q excimer platform. Clin Ophthalmol. 2012;6: Kanellopoulos AJ, Asimellis G. Long-term bladeless LASIK outcomes with the FS200 femtosecond and EX500 Excimer Laser workstation: the Refractive Suite. Clin Ophthalmol. 2013;7: Kanellopoulos AJ, Asimellis G. Correlation between central corneal thickness, anterior chamber depth, and corneal keratometry as measured by Oculyzer II and WaveLight OB820 in preoperative cataract surgery patients. J Refract Surg. 2012;28: Krumeich JH, Daniel J, Knülle A. Live-epikeratophakia for keratoconus. J Cataract Refract Surg. 1998;24: Faria-Correia F, Ramos IC, Lopes BT, et al. Topometric and tomographic indices for the diagnosis of keratoconus. Int J Kerat Ectatic Dis. 2012;1: Kanellopoulos AJ, Asimellis G. Revisiting keratoconus diagnosis and progression classification based on evaluation of corneal asymmetry indices, derived from Scheimpflug imaging in keratoconic and suspect cases. Clin Ophthalmol. 2013;7: Markakis GA, Roberts CJ, Harris JW, Lembach RG. Comparison of topographic technologies in anterior surface mapping of keratoconus using two display algorithms and six corneal topography devices. Int J Kerat Ectatic Dis. 2012;1: Ambrósio R Jr, Caiado AL, Guerra FP, et al. Novel pachymetric parameters based on corneal tomography for diagnosing keratoconus. J Refract Surg. 2011;27: Legare ME, Iovieno A, Yeung SN, et al. Corneal collagen crosslinking using riboflavin and ultraviolet A for the treatment of mild to moderate keratoconus: 2-year follow-up. Can J Ophthalmol. 2013;48: Vinciguerra P, Albè E, Trazza S, et al. Refractive, topographic, tomographic, and aberrometric analysis of keratoconic eyes undergoing corneal cross-linking. Ophthalmology. 2009;116: Raiskup-Wolf F, Hoyer A, Spoerl E, Pillunat LE. Collagen crosslinking with riboflavin and ultraviolet-a light in keratoconus: longterm results. J Cataract Refract Surg. 2008;34: Kanellopoulos AJ, Asimellis G. Comparison of Placido disc and Scheimpflug image-derived topography-guided excimer laser surface normalization combined with higher influence CXL: The Athens Protocol, in progressive keratoconus cases. Clin Ophthalmol. 2013;7: Mencucci R, Paladini I, Virgili G, Giacomelli G, Menchini U. Corneal thickness measurements using time-domain anterior segment OCT, ultrasound, and Scheimpflug tomographer pachymetry before and after corneal cross-linking for keratoconus. J Refract Surg. 2012;28: O Brart DP, Kwong TQ, Patel P, McDonald RJ, O Brart NA. Long-term follow-up of riboflavin/ultraviolet A (370 nm) corneal collagen cross-linking to halt the progression of keratoconus. Br J Ophthalmol. 2013;97: Greenstein SA, Shah VP, Fry KL, Hersh PS. Corneal thickness changes after corneal collagen crosslinking for keratoconus and corneal ectasia: one-year results. J Cataract Refract Surg. 2011;37: Kanellopoulos AJ, Asimellis G. Anterior segment optical coherence tomography: assisted topographic corneal epithelial thickness distribution imaging of a keratoconus patient. Case Rep Ophthalmol. 2013;4: TABLE A Anterior Keratometry (K) Measured by the Scheimpflug Device (N = 231 Eyes) a Postoperative Value Preop 1 Month 3 Months 6 Months 12 Months 24 Months 36 Months K1 (flat) Average SD ±3.83 ±3.97 ±3.86 ±3.76 ±3.73 ±3.74 ±3.80 Maximum Minimum K2 (steep) Average SD ±5.14 ±5.15 ±4.99 ±4.92 ±4.91 ±4.88 ±4.91 Maximum Minimum preop = preoperative; SD = standard deviation a All units in keratometric diopters (D). TABLE B Anterior Surface Topometric Indices and Postoperative Progression (N = 231 Eyes) Postoperative Value Preoperative 1 Month 3 Months 6 Months 12 Months 24 Months 36 Months Index of Surface Variance Average SD ±43.47 ±38.73 ±38.13 ±37.79 ±38.04 ±38.28 ±38.41 Maximum Minimum Index of Height Decentration (µm) Average SD ±0.053 ±0.041 ±0.040 ±0.040 ±0.040 ±0.040 ±0.040 Maximum Minimum SD = standard deviation 92 Copyright SLACK Incorporated

19 CLINICAL SCIENCE Cornea Volume 33, Number 9, September 2014 High-Fluence CXL in Keratoprosthesis Long-Term Safety and Efficacy of High-Fluence Collagen Crosslinking of the Vehicle Cornea in Boston Keratoprosthesis Type 1 Anastasios J. Kanellopoulos, MD,* and George Asimellis, PhD* Purpose: The aim of this study was to evaluate the safety and efficacy of very high-fluence collagen crosslinking (CXL) as a means of achieving increased corneal rigidity and reduced enzymatic digestion in the vehicle cornea of Boston keratoprosthesis (KPro) type 1. Methods: Eleven consecutive patients fitted with a KPro (5 with a previous repeat cornea graft failure, 4 with ocular cicatricial pemphigoid, and 2 with chemical burn) underwent donor vehicle cornea pretreatment with very high-fluence prophylactic CXL in a 2- step procedure. First, the donor cornea was crosslinked with an intrastromal riboflavin instillation through a femtosecond laser created pocket. This was followed up with a superficial CXL treatment. On the completion of the CXL pretreatment, the cornea center was trephined with the femtosecond laser, and the KPro was fitted onto the crosslinked donor cornea. Visual acuity, corneal surface, and donor vehicle cornea stability were evaluated. Follow-up evaluations were conducted over the next 9 years with a mean of 7.5 years. Results: Mean uncorrected visual acuity improved from light perception to 20/60. One patient required a follow-up surgery, because of significant melt in the host cornea. None of the eyes developed melts and/or infection, especially on the vehicle cornea on which the KPro was fitted. Conclusions: Pretreatment with intrastromal and superficial very high-fluence CXL in conjunction with Boston type 1 KPro seems to be a safe and effective adjunctive treatment for achieving increased vehicle donor cornea rigidity. Additionally, there is an increased resistance to enzymatic degradation. This application may serve to enhance the biomechanical stability and external disease Received for publication January 27, 2014; revision received April 30, 2014; accepted May 5, Published online ahead of print July 9, From the *Laservision.gr Eye Institute, Athens, Greece; and Department of Ophthalmology, NYU Medical School, New York, NY. Presented in part at the CXL annual meeting in Dresden, Germany, in December A. J. Kanellopoulos: Alcon/WaveLight, Allegran, Avedro, i-optics; Optovue. The other author has no funding or conflicts of interest to disclose. Design and conduct of the study (A.J.K.); collection (A.J.K.), management (A.J.K.), analysis (A.J.K., G.A.), interpretation of the data (A.J.K., G.A.); manuscript preparation (G.A., A.J.K.), manuscript review (A.J.K., G.A.), manuscript approval (A.J.K., G.A.). Reprints: Anastasios J. Kanellopoulos, MD, Clinical Professor of Ophthalmology, New York University Medical School, New York, NY, Laservision.gr Eye Institute, 17 Tsocha Street, Athens , Greece ( ajk@brilliantvision.com). Copyright 2014 by Lippincott Williams & Wilkins resistance of the donor vehicle cornea in patients with advanced external disease. Key Words: prophylactic pretreatment with collagen crosslinking, Boston keratoprosthesis type 1, Dohlman keratoprosthesis, severe external disease, ocular cicatricial pemphigoid, chemical burn, repeat cornea transplantation failure (Cornea 2014;33: ) xtreme external disease has been successfully treated over Ethe last few decades with allograft cornea transplantation, relatively histocompatible limbal stem-cell transplantation, or keratoprosthesis (KPro). There are several KPro variations. Currently, the most prevalent in clinical practice are the Boston KPro, 1 4 the AlphaCor, 5 the odonto-kpro, 6 the Fyodorov KPro, 7 and the KeraClear inlay KPro. Our team has introduced the concept of accelerated, high-fluence collagen crosslinking (CXL) in post laser in situ keratomileusis (LASIK) ectasia, 8 and the use of prophylactic CXL in routine LASIK, 9 in treatment of cornea ectasia, 10 and in attempting corneal deturgescence 11 in bullous keratopathy. 12 In our 20 years of experience in using the Boston KPro, we encountered 2 significant complications: (1) melts and (2) erosion of the donor and host cornea interface. 13 When the latter occurs, there is an increase in the risk of developing infectious keratitis, which will significantly increase the risk of potential endophthalmitis 14 and may predispose prosthesis exposure and/or infection. We hypothesized that the application of a prophylactic crosslinking treatment on the donor vehicle cornea might help reduce the susceptibility of these corneas to enzymatic digestion and cornea infection. This work presents an evaluation of a longitudinal case series to study the advantages of using CXL as a prophylactic intervention adjuvant to Boston KPro surgery. MATERIALS AND METHODS This work adhered to the tenets of the Declaration of Helsinki, and the study received the approval from the Ethics Committee of our Institution. All patients were provided a written informed consent before the treatment, and a comprehensive explanation of the benefits and risks of this procedure was presented to them by the operating surgeon (A.J.K.) Cornea Volume 33, Number 9, September 2014 This study describes the outcomes from 11 different cases belonging to 11 different patients with external disease and almost total visual disability who opted to undergo Boston KPro type 1 in conjunction with prophylactic highfluence CXL in the donor vehicle cornea that would be used for their KPro surgery. All the patients had debilitating visual compromise. The most common preoperative diagnosis in this group was repeat cornea graft failure (5 patients) with at least 3 failed grafts in the past for each. The remaining patients were diagnosed with ocular cicatricial pemphigoid 15 (4 patients) and chemical corneal trauma (2 patients). Preoperatively and postoperatively, we evaluated uncorrected distance visual acuity (UDVA), best spectaclecorrected distance visual acuity (CDVA), subjective refraction, keratometry, and applied anterior-segment optical coherence tomography (OCT). 16 These examinations were performed to evaluate the integrity of the donor vehicle cornea on which the KPro was mounted, the host cornea, and the donor host interface. Macula and optic nerve OCT screening was also performed to evaluate potential glaucomatous damage and/or macular edema. Surgical Technique The donor cornea was mounted on an artificial chamber (Baron; Katena Products, Inc, Denville, NJ). A deep ( mm) lamellar central corneal pocket of an 8-mm diameter was created using a femtosecond laser (the IntraLase FS60; AMO, Abbott Park, IL, in the first 4 cases, and the WaveLight FS200; Alcon, Ft Worth, TX, in the subsequent 7 cases). This pocket was accessed from the anterior corneal surface through a 3-mm channel. After a blunt olive-tip cannula was placed into the pocket to facilitate injection, 0.1 ml of 0.1% riboflavin sodium phosphate solution was injected into the FIGURE 1. Inner donor cornea CXL through the femtosecond laser created pocket. A, FS200 femtosecond laser programming interface. B, Screen capture of the 8-mm diameter, 400-mm-deep femtosecond laser assisted pocket creation. C, Intrastromal infusion of 0.1% riboflavin solution with the olive-tip cannula. intrastromal pocket. Immediately after this, 4 minutes of 30 mw/cm 2 ultraviolet-a irradiation for a total energy of 7.2 J/cm 2 was directed onto the donor vehicle cornea. After completing this step, the corneal epithelium was scraped with a blunt crescent blade, and drops of 0.1% riboflavin sodium phosphate solution were administered every minute for a total of 10 minutes until the anterior donor cornea was briskly colored yellow. The second CXL session was also performed with a 30-mW/cm 2 fluence for 4 minutes until a total of 7.2 mj/cm 2 of energy was reached. Ultraviolet-A in all cases was administered by using the KXL crosslinking device (Avedro Inc, Waltham, MA). The central, 3-mm trephination of the donor cornea button was performed with a femtosecond laser. Before mounting the KPro, the donor cornea was trephined on a disposable Barron donor block (9.00 mm). Then, the KPro was sutured in place with 16 interrupted 10.0 nylon sutures, and the wound was deemed watertight. After the surgery, all cases were fitted with an 18-mm-diameter bandage contact lens to reduce the friction between the eyelids and the operated ocular surface. All patients were treated postoperatively with a drop of ofloxacin and 50 mg/ml of a solution of vancomycin once a day. In the later 6 years, all patients were treated with amoxifloxacin 4 times a day and with vancomycin as well. In addition, it should be noted that antifungal prophylaxis, 1 drop of natamycin (Natacyn; Alcon, Ireland), was administered weekly to all the patients. All patients were evaluated on the first postoperative day, every month for 6 months, and every 3 months thereafter to assess visual acuity and the condition of the external surface. Because the capability of assessing intraocular pressure was limited to fingertip application over the eyelids, meticulous optic nerve visualization every 3 months was used as a means to evaluate possible glaucomatous optic nerve damage. Ó 2014 Lippincott Williams & Wilkins 915

20 Kanellopoulos and Asimellis Cornea Volume 33, Number 9, September 2014 Cornea Volume 33, Number 9, September 2014 High-Fluence CXL in Keratoprosthesis RESULTS The mean age of the patients was years. Six patients were female and 5 were male. The visual acuity assessed from preoperative light perception and/or hand motion showed a 6-month postoperative improvement. The average UDVA was 20/80 (range: 20/100 20/40), and the CDVA was 20/70 (range: 20/80 20/32). These patients are still being followed up. After the first postoperative year, each patient is evaluated at least annually. During the long follow-up time that these patients have been continuously monitored (minimum 1 year, maximum 9 years), 2 of the patients required subsequent injection of intracameral and triamcinolone and bevacizumab injection (Avastin; Genentech/Roche, San Francisco, CA) because of cystoid macular edema. Additionally, 1 patient needed yttrium aluminum garnet laser intervention for a retroprosthesis inflammatory membrane that was quite dense and had resulted in a CDVA reduction from 20/60 to 20/400. After this procedure, the patient s vision returned to 20/50. Figure 1 demonstrates the procedure. Figure 1A shows the FS200 femtosecond laser programming interface. Figure 1B is a screen capture of the 8-mm diameter, 400-mm-deep femtosecond laser assisted pocket creation. Figure 1C, is a screen capture of the intrastromal infusion of 0.1% riboflavin solution with the olive-tip cannula. Figure 2 illustrates the CXL of the cornea, the anterior part of the donor cornea after epithelial debridement, and instillation of riboflavin solution. Figure 3 shows the trephination of the crosslinked vehicle donor cornea in its center to fit the KPro. Figure 4 shows results from a case 4 years postoperatively. This patient had an explosionrelated injury. The right eye had to be enucleated as a result. The left eye sustained a chemical injury, which resulted in the visual acuity being light perception and severe cicatricial changes occurring in both the conjunctiva and cornea. KPro was deemed to be the optimal treatment option for this left eye. The patient s vision improved from light perception to 20/40 UDVA in the first week postoperatively. DISCUSSION In our 20 years of experience in external disease and the use of the Boston KPro to address it, the main obstacles of prosthesis and visual rehabilitation stability that we have encountered and reported (Peralta RJ, Kanellopoulos AJ. Boston keratoprosthesis: A long-term prospective clinical study A Long- Term Prospective Clinical Study. Poster Presentation, ARVO Meeting, May 6 10, 2007, Fort Lauderdale, FL) have been intraocular pressure control, infection (whichwasattributedpartlyon cornea melt around the prosthesis), and intraocular inflammation. One of the 2 major difficulties in managing these patients has been antibiotic prophylaxis, because these eyes are especially susceptible to microbial infections, 17,18 which, after the KPro surgery, almost invariably leads to endophthalmitis in a unichamber eye with very poor prognosis The second significant postoperative management problem is donor vehicle cornea and/or host cornea melts, 23,24 especially near the graft host interface. Using antiproteolytics, such as oral tetracycline-type medications and/or topical progesterone, may be an effective alternative. 16 The decision to incorporate this adjunct prophylactic treatment in these very challenging cases was based on significant experience with CXL techniques. The long-term safety and efficacy results of this technique as noted above suggest that there may exist a very significant advantage for the long-term prognosis of the Boston KPro. This work presents 11 cases treated with adjuvant crosslinking. The relative rarity of the parameters that led to the decision for inclusion in this study leads us to believe that our cases represent 1 of the largest groups in the literature. Among the reasons for crosslinking was our observation that up to 50% of these patients suffer from cornea melting, which may increase their likelihood of developing infections. The lack of a control group is a significant flaw of this study; in our defense, the seriousness and urgency for treatment and our strong belief in the benefit of crosslinking were reasons why we did not include a control group. Questions on whether crosslinked corneas allow effective topical antibiotic penetration and/or the apoptosis of cornea keratocytes, 25 which is a side effect of CXL, and their effect on long-term microbial host defenses, are not answered in this study. However, they remain important considerations. There are recent reports indicating that crosslinked corneas do allow for effective antibiotic penetration within the corneal stroma, 26 and there is a possibility that the donor vehicle cornea may have reduced antigenicity to the host s immune mechanisms after crosslinking, because most of the keratocytes will be killed in the crosslinking process. This has been shown in basic science studies. 27,28 Because we have incorporated this novel pretreatment technique in our KPro routine, larger-scale future studies may further validate these findings. In conclusion, the very high-fluence CXL of the donor vehicle cornea as a prophylaxis against corneal melt and extreme external disease may be an efficacious adjunct that helps to reduce the susceptibility of these corneas to enzymatic digestion and cornea infection. FIGURE 2. CXL of the cornea, the anterior part of the donor cornea after epithelial debridement, and installation of riboflavin solution with very high-fluence CXL. A, The first crosslinking session of the donor cornea through intact epithelium and riboflavin solution injected in the lamellar pocket with 30 mw/cm 2 for 4 minutes. B, Scraping the donor corneal epithelium with a crescent blade before soaking the stromal surface, in preparation for the second crosslinking session. C, Soaking the deepithelialized donor cornea with riboflavin solution as a preparation for the second crosslinking session. FIGURE 3. Donor Cornea after 2.8-mm central trephination and just before the peripheral 9.5-mm trephination Ó 2014 Lippincott Williams & Wilkins FIGURE 4. Four years postoperative anterior-segment OCT imaging of the patient who had received prophylactic CXL. Top panel, meridional scan at 90, bottom panel top view of the cornea showing the scan orientation. REFERENCES 1. Cruzat A, Shukla A, Dohlman CH, et al. Wound anatomy after type 1 Boston KPro using oversized back plates. Cornea. 2013;32: Ziai S, Rootman DS, Slomovic AR, et al. Oral buccal mucous membrane allograft with a corneal lamellar graft for the repair of Boston type 1 keratoprosthesis stromal melts. Cornea. 2013;32: Ciolino JB, Belin MW, Todani A, et al; Boston Keratoprosthesis Type 1 Study Group. Retention of the Boston keratoprosthesis type 1: multicenter study results. Ophthalmology. 2013;120: Klufas MA, Colby KA. The Boston keratoprosthesis. Int Ophthalmol Clin. 2010;50: Ngakeng V, Hauck MJ, Price MO, et al. AlphaCor keratoprosthesis: a novel approach to minimize the risks of long-term postoperative complications. Cornea. 2008;27: Lim LS, Ang CL, Wong E, et al. Vitreoretinal complications and vitreoretinal surgery in osteo-odonto-keratoprosthesis surgery. Am J Ophthalmol. 2014;157: Ghaffariyeh A, Honarpisheh N, Karkhaneh A, et al. Fyodorov Zuev keratoprosthesis implantation: long-term results in patients with multiple failed corneal grafts. Graefes Arch Clin Exp Ophthalmol. 2011;249: Kanellopoulos AJ. Post-LASIK ectasia. Ophthalmology. 2007;114:1230. Ó 2014 Lippincott Williams & Wilkins 917