In vivo confocal microscopy revealed a hyperintense signal consistent with mild interface scarring.
A, Donor/host interface after penetrating keratoplasty. Light microscopy with Ki-67 staining (original magnification x200) did not show significant cellular proliferation close to the donor/host interface even if we noticed that the density of keratocytes was higher in the posterior stroma than in the anterior stroma. Ki-67 staining was also negative in case 1. B, Light microscopy with SMA staining (original magnification x200) did not reveal significant myofibroblastic contraction that could have explained the stuccolike aspect found in electron microscopy. The SMA staining was also negative in case 1.
Donor/host interface after penetrating keratoplasty in case 1 (A) and case 2 (B). Light microscopy (hematoxylin-eosin, original magnification x100) showed mild interface scarring in both cases. The arrowheads denote the interface between the anterior and posterior lamellae. In case 1 (A), the asterisk shows that the scarring was mainly located in the posterior stroma (residual stromal bed after femtosecond laser–assisted lamellar keratoplasty). In case 2 (B), the interface was smooth with mild scarring located in the anterior lamellae. The trephination edge was regular with no epithelial ingrowth (B).
Electron microscopy of donor posterior surface (A, original magnification [Magn] ×305) and recipient anterior surface (B, original Magn ×328) showed a mild stuccolike appearance in both the donor and recipient corneal lamellar surfaces. Acc. V indicates acceleration voltage; Det, detector; Exp, exposure; GSE, gaseous secondary electron; and WD, working distance.
Banitt M, Cabot F, Hussain R, Dubovy S, Yoo SH. In Vivo Effects of Femtosecond Laser–Assisted Keratoplasty. JAMA Ophthalmol. 2014;132(11):1355-1358. doi:10.1001/jamaophthalmol.2014.2389
The femtosecond laser is reported to cut lamellar surfaces with varying degrees of smoothness depending on the depth of the cut, with deeper cuts leaving less smooth surfaces. We attempted to evaluate the smoothness of the deeper lamellar surface as cut by the femtosecond laser after allowing 3 months of in vivo healing.
Two patients underwent penetrating keratoplasty 3 months after inadequate visual rehabilitation following femtosecond laser–assisted sutureless anterior lamellar keratoplasty for the treatment of anterior stromal scars. In vivo confocal microscopy that was performed before penetrating keratoplasty demonstrated an acellular zone with a hyperintense signal consistent with a mild interface opacification. Light microscopy in one patient demonstrated scarring limited primarily to the posterior stroma; in the other patient, the interface was smooth with mild scarring of the anterior lamellae. When studied with electron microscopy, the cut surfaces revealed a smooth to very mild stuccolike appearance that was smoother than anticipated.
Conclusions and Relevance
After 3 months of in vivo healing, the lamellar interface produced by the femtosecond laser, as imaged by electron microscopy, appeared to be nearly smooth with minimal roughness to the cut surfaces. We attribute this to the effects of in vivo healing and remodeling.
By causing photodisruption only where it is focused, the femtosecond laser does not damage adjacent structures. This property makes it ideally suited for use in biologic tissue, including the cornea. Therefore, the femtosecond laser has been used to create lamellar dissections, shaped host and donor incisions for penetrating keratoplasty, and flaps for laser in situ keratomileusis, as well as to assist with cataract extraction.1
Earlier reports2,3 have indicated that the femtosecond laser creates very smooth lamellar cuts in the superficial cornea. Deeper stromal cuts have resulted in a stuccolike appearance within the interface.2- 4 A prior study5 has described the microscopic findings of the lamellar and trephined surfaces of a rabbit cornea 3 months after the interface was created, allowing for the natural healing response. However, only the trephined edges of the rabbit cornea (not the interface) were studied with an electron microscope. To our knowledge, femtosecond laser electron microscopy findings after in vivo healing in a human eye have not previously been documented. In this report, we describe the use of light and electron microscopy, as well as in vivo confocal microscopy, to characterize the lamellar interfaces of 2 corneas that underwent femtosecond laser–assisted lamellar keratoplasty followed by penetrating keratoplasty.
A man in his 50s who wore contact lenses presented 3 months after an episode of bacterial keratitis in his right eye. The keratitis had been treated with topical antibiotics, and the patient developed corneal thinning and an anterior stromal scar. In an attempt to rehabilitate the eye, sutureless femtosecond laser anterior lamellar keratoplasty (with a target residual stromal bed thickness of 250 μm) was performed (IntraLase femtosecond laser system; Abbott Medical Optics Inc) with a lamellar energy setting of 2.20 mJ and spot and line separations of 7 and 7 µm. Because of residual scarring and irregular astigmatism, only partial rehabilitation was achieved. In vivo confocal microscopy (ConfoScan 3; Nidek Co Ltd) was performed postoperatively. The patient elected to undergo penetrating keratoplasty 3 months after the initial procedure.
Intraoperatively, the corneal specimen containing the lamellar interface was cut in half, preserved in formalin, 10%, dehydrated in alcohol and xylol, and embedded in paraffin. One-half of the specimen was sectioned, stained with hematoxylin-eosin, and examined with a light microscope. The other half was rinsed in buffered saline, fixed with glutaraldehyde, 2.5%, dehydrated in an ascending-strength series of ethanol (to 100%), and infiltrated with 3 changes of hydroxymethyldisalazine. The hydroxymethyldisalazine was pipetted off, and the samples were allowed to air dry overnight. The specimens used for scanning electron microscopy were manually separated to reveal each donor and recipient interface. The tissue was then mounted on aluminum stubs, dried, and sputter-coated with a thin layer of gold. The specimens were then examined with a scanning electron microscope.
A man in his 60s had a history of photorefractive keratectomy for 4.00 diopters of myopia and pigment dispersion syndrome without glaucoma. His vision had been 20/20 for 10 years until he developed cataracts and underwent uncomplicated cataract extraction in both eyes. Several years later, he developed fungal keratitis in his left eye. After treatment, corneal topography revealed irregularity, a central corneal thickness of 533 µm, and an opacity that was approximately 230 µm deep as determined by time-domain optical coherence tomography (Visante; Carl Zeiss Meditec).
The patient underwent a femtosecond laser–assisted sutureless anterior lamellar keratoplasty (with a target residual stromal bed thickness of 300 µm). The procedure was performed using a lamellar energy setting of 1.30 mJ and spot and line separations of 5 and 5 µm (IntraLase femtosecond laser system). In an attempt to fully ablate the scar, a decision was made intraoperatively to perform an additional 10-µm ablation to the lamellar bed (phototherapeutic keratectomy). The anterior lamellar donor button and residual posterior host cornea specimens were sent to the laboratory for analysis. Both specimens were prepared in the manner discussed above.
Six weeks later, both significant irregular astigmatism and residual scarring deep to the lamellar interface remained. Best-corrected visual acuity with rigid gas-permeable lenses was 20/80, and the patient elected to undergo penetrating keratoplasty 3 months later. Perioperatively, the specimens (anterior lamellar donor button and residual posterior host cornea removed during the operation) were sent to the laboratory for analysis.
Like the Nd:YAG laser, the femtosecond laser uses ultrashort pulses to deliver cavitation bubbles and small shock waves to cause photodisruption. Lamellar dissections achieved with the femtosecond laser have produced cut corneal surfaces with appearances that vary depending on corneal depth.1- 3 A smooth surface results when cuts are made in the superficial cornea, such as when laser in situ keratomileusis is performed.2 Cuts made in the deeper stroma can produce surfaces that appear to be minimally changed and stuccolike.1,2 However, Soong and Malta1 described rough interfaces in patients who underwent femtosecond laser–assisted penetrating keratoplasty. Laser scatter and attenuation are postulated to contribute to the greater surface irregularity seen with deep corneal dissections.3 Earlier reports1- 3 described the contour of the interface immediately postoperatively, but changes that occur in vivo have not been reported.
Our patients demonstrated an interface surface smoother than anticipated based on the findings of in vitro studies1- 3 after undergoing deep stromal cuts. We attribute this to 3 months of in vivo healing. We postulate that the surface was smoother than anticipated owing to the contact between the 2 cut surfaces and remodeling of the lamellar surface by keratocytes. In patients who had undergone femtosecond laser–assisted keratoplasty, Shtein et al6described an initial increase in keratocyte activation, which is thought to represent a transition to a repair phenotype.7,8 In case 1 described in the present report, in vivo confocal microscopy at approximately 230 µm from the epithelial surface that was thought to best represent the interface, performed 3 months after the femtosecond laser–assisted lamellar keratoplasty, revealed an acellular zone with a hyperintense signal consistent with a mild interface opacification (Figure 1). Some microdots were noted, but no needlelike opacities or keratocytes were found near the interface, which suggests the lack of an ongoing repair process as well as minimal inflammation. We picked a level that appeared to best represent the interface and could not determine whether this was the host or donor side. The depth was approximately 230 µm from the epithelial surface. Light microscopy with Ki-67 staining in samples from both patients did not reveal significant cellular proliferation at the interface, which supports our findings of a lack of an ongoing repair process and minimal inflammation at the interface (Figure 2A).
Light microscopy (hematoxylin-eosin staining) showed mild interface scarring in both cases. In case 1 (Figure 3A), scarring was largely limited to the posterior stroma or residual stromal bed. In case 2 (Figure 3B), the interface was smooth with mild scarring of the anterior lamellae. In both cases, the posterior surface of the anterior lamellae and the anterior surface of the posterior lamellae were regular, with fibrosis present above and below the interface.
Electron microscopy revealed a smooth to very mild stuccolike appearance of both the donor and recipient corneal lamellar surfaces (Figure 4). Smooth-muscle actin staining was performed to determine whether myofibroblastic contraction could explain the stuccolike aspect noted on electron microscopy. However, the staining was negative in both cases as shown in Figure 2B for case 2.
Mian and colleagues5 reported seeing well-integrated lamellar interfaces and smooth trephination edges in vivo 3 months after performing femtosecond laser–assisted posterior lamellar keratoplasty in rabbits. Our findings regarding the trephination edges and the lamellar interface in human eyes corroborate their results.
In conclusion, the deep stromal surfaces 3 months after femotosecond laser–assisted keratoplasty in 2 patients were smoother than expected. We attribute this finding to the effects of in vivo healing and remodeling. Further studies with larger sample sizes are needed to confirm these results.
Corresponding Author: Michael Banitt, MD, Department of Ophthalmology, Bascom Palmer Eye Institute, 900 NW 17th St, Miami, FL 33136.
Submitted for Publication: October 3, 2013; final revision received March 30, 2014; accepted April 8, 2014.
Published Online: July 31, 2014. doi:10.1001/jamaophthalmol.2014.2389.
Author Contributions: Dr Banitt had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Banitt, Yoo.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Banitt, Cabot, Hussain.
Critical revision of the manuscript for important intellectual content: Banitt, Cabot, Dubovy, Yoo.
Statistical analysis: Banitt, Dubovy.
Administrative, technical, or material support: Banitt, Cabot Hussain, Dubovy.
Study supervision: Cabot, Dubovy, Yoo.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was supported by National Institutes of Health center grant P30-EY014801, a Research to Prevent Blindness unrestricted grant, and US Department of Defense grant W81XWH-09-1-0675.
Role of the Sponsor: The sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.