Current research interest in noncontact laser photothermal keratoplasty has evolved because of its noninvasiveness, versatility in correcting refractive errors, and possible clinical applications.12 In our previous study, the correlation between the corneal topographic change and the selection of laser treatment parameters was established.2
The purpose of this study is to investigate the corneal tissue response to different beam geometry, a range of laser radiant exposures, and number of pulses.
MATERIALS AND METHODS
The noncontact laser photothermal keratoplasty laser system described in our previous paper1 is used in this study. In this system, the laser energy emitting from the holmium:YAG laser (wavelength = 2.10 µm; pulse duration = 250 µm, repetition rate = 5 Hz) is delivered by a 300-micrometer fiber optic cable and is coupled into a beam reshaping device, which is mounted on an ophthalmic slitlamp. The beam reshaping device employs either a conical axicon lens or an eight-facet polyprismatic lens, which splits the beam into a ring pattern or eight spots on a ring. The diameter of the ring can be adjusted from 2 to 8 mm by vertically translating the beam reshaping device. The width of the projected annulus ring was 150 µm and the size of the projected spots was 300 µm in diameter when measured on the cornea. A mask can be placed into the beam reshaping device so that only part of the ring pattern or one of the eight laser spots is allowed to emit from the system. The radiant exposure level of the laser treatment can be varied from 1 to 20 J/cm2 by adjusting the laser output energy. The number of pulses emitting from the system can be preselected and electronically controlled. The system is activated by a foot-pedal.
Figure 1: Distribution of laser treatment spots and their corresponding parameters.
Figure 2: Clinical appearance of the owl monkey cornea immediately after a 5-millimeter annular ring pattern laser photothermal keratoplasty treatment (8 J/cmp 2, 25 pulses). Deformation of the cornea was visible. A strong corneal haze was observed at the laser-treated site, indicating the corneal collagen lost its organized structure.
Annular Treatment - Monkey Model. The laser photothermal keratoplasty annulus treatment pattern was studied on an owl monkey. After a deep level of general anesthesia was achieved using an intramuscular injection of ketamine/xylazine, a vacuum ring was applied to the monkey's limbus and was then attached to the laser system for a stable fixation.,1 A 5-millimeter annulus ring pattern was projected onto the monkey's cornea centered at the pupil. The radiant exposure of the laser treatment was 8 J/cmp 2. Twenty-five consecutive laser pulses were delivered to the cornea at 1 Hz. The laser photothermal keratoplasty treatment was performed bilaterally. Clinical examination and photographic documentation were performed immediately following laser treatment. The animal was killed by lethal injection. The right eye was used to estimate endothelial cell damage by vital stain and the left eye was preserved following steps described below for histological study.
Figure 3: Scanning electron micro-photographs of the owl monkey cornea after laser photothermal keratoplasty treatment with annular pattern (8 J/cmp 2, 25 pulses), (A) Overall cross-section view. (B) Cross-section view from endothelial side. The laser-induced contraction of the corneal collagen fibers causing an Increase in tissue volume at the laser treatment site. (C) Scanning electron microscopy (SEM) over epithelial surface. The epithelium was overheated by the laser treatment and shrunken into a convoluted and tortuous form, (D) SEM over endothelial surface. Local endothelial damages were observed underneath the laser-treated area.
Spot Treatment - Cat and Rabbit Model. The laser photothermal keratoplasty spot treatment pattern was studied and compared on cat and rabbit models because of their different corneal thicknesses. General anesthesia with ketamine and xylazine was administered to the rabbits and cats. After the animals reached a satisfactory level of anesthesia, the corneas were divided into eight parts by an eight-incision radial keratotomy marker. A single laser spot of one pulse or five consecutive pulses at 5 Hz of the same radiant exposure were delivered in the middle area of each part at 1.5 mm, and 2.5 mm from the center. Eight different radiant exposures, ranging from 5.00 J/cmp 2 to 18.01 J/cmp 2, were studied. The treatment pattern and parameters are summarized in Figure 1. After the laser treatment, the animals were killed by lethal injection and the eyes prepared for histological study.
The eyes were processed for light, scanning, and transmission electron microscopy. To prevent inadvertent mechanical deformation of the cornea, we injected 2.5% glutaraldehyde buffered in 0.1 M sodium cacodylate into the anterior chamber of the globes. This also provided rapid and uniform fixation of the cornea. The globes were then immersed into this fixative. After 2 hours of fixation, the corneas were excised at the limbus and post-fixed in 2% osmium tetroxide buffered in 0. 1 M sodium cacodylate for light, scanning, and transmission electron microscopy.
Volume of Tissue Shrinkage - Cadaver Eye Model
This part of the experiment was designed to study collagen shrinkage induced by laser treatment at a range of radiant exposure levels of a single pulse or of five consecutive pulses delivered at 5 Hz. Five human cadaver eyes were used. Eyes showing evidence of previous ocular surgery or significant corneal pathology were excluded from the study. The corneas were deepithelialized. Their thicknesses and intraocular pressures were maintained at a physiological level (central thickness = 450 ± 20 µm, IOP = 20 mm Hg).
Figure 4: Light microscopy of the owl monkey cornea treated by laser photothermal keratoplasty annular pattern (8 J/cmp 2, 25 pulses). The effect of the laser-induced collagen shrinkage reaches the full thickness of the cornea. The corneal collagen loses its lamella structure beneath the laser-treatment site. The contraction of the collagen fibers caused a local increase in corneal thickness at the laser-treated sites.
Figure 5: Transmission electron microscopy of keratocytes: (A) Within the laser photothermal keratoplasty treatment zone: The keratocyte is thermally denatured and shrunken; and (B) outside the treatment zone: The keratocyte maintains its normal morphology.
Prior to the laser treatment, a drop of 1% dextran B solution was topically applied to simulate the effect of the tear film. After the dextran drop reached stable distribution on the cornea, the eye was treated following the same laser treatment protocol and tissue fixation as described above. However, each laser spot was studied independently. They were separately embedded in epoxy resin (Epon) for thin sectioning (1 µm). For each specimen, a series of histological sections were made through the lesion with a step size of 20 µm and stained with toluidine blue. The sectioning process continued to the mid point of the laser spot. After mounting, the shrinkage zone boundary of each section in the series was traced manually on scientific graph paper for reconstruction of the affected tissue dimensions.
In Vivo Response- Spot Treatment on Cat Model
The purpose of this experiment was to study the healing response after the laser photothermal keratoplasty procedure. The cat model was selected because its corneal thickness is similar to that of a human. Five cats were treated by the laser photothermal keratoplasty procedure with an eight-spot treatment pattern. The eight spots were distributed equally on a 3-millimeter ring centered at each cat's pupil. A single pulse at 15.6 J/cmp 2 was used for the treatment. The clinical course of the laser photothermal keratoplasty-treated cats was followed over a 3-month period. Then, all cats were killed by lethal injections and the eyes were preserved for histological study using the above-described fixation process. The laser spots were sectioned for light microscopy.
Morphological Study of Laser Photothermal Keratoplasty
Annular Treatment - Monkey Model. In the owl monkey experiment, laser-induced collagen shrinkage, stress lines, and corneal surface deformation were visible after five pulses. After 25 pulses were delivered, a dense corneal haze was observed at the laser-treated site. The clinical appearances of the corneal surface during and immediately after laser photothermal keratoplasty treatment with the annular pattern are shown in Figure 2. The vital stain analysis revealed an annular-shaped zone of damaged endothelial cells beneath the laser treatment. This was further confirmed by scanning electron microscopy. Figures 3A and B show a cross-sectional view of the laser photothermal keratoplasty-treated cornea. The laser-induced contraction of the corneal collagen fibers caused the increased tissue volume seen at the laser-treated site. The tissue bulged up on both sides of the cornea and created a rim. Scanning electron microscopy (SEM) of the epithelial surface (Fig 3C) showed that epithelial cells were shrunken into a convoluted and tortuous form, resembling spaghetti. Damage to the endothelial cells was also demonstrated by SEM (Fig 3D). Light microscopy showed a conical shrinkage area with the laser effect reaching the endothelium for the laser treatment: 8 J/cm2, 25 pulses (Fig 4). Transmission electron microscopy on the owl monkey cornea showed that laser photothermal keratoplasty treatment thermally denatured the keratocytes within the laser treatment area (Fig 5A), whereas the keratocytes outside the treatment zone retained their normal morphology (Fig 5B).
Figure 6 Scanning electron microscopic comparison of a single-pulse (A.B) versus a five-pulse (C1D) treatment at 1 8.01 J/cmp 2 on cat model (top row = epithelium, bottom row = endothelium): Single-pulse treatment creates minimal epithelial surface disruption and no endothelial damage, whereas the five-pulse treatment creates the epithelial surface disruption and localized endothelial damage. (Spot size - 350 µm.)
Spot Treatment - Cat and Rabbit Model. In the cat and rabbit studies, tissue reaction to the laser spots was not observed for either a single pulse or for five pulses until the radiant exposure of the laser treatment was increased to 8.16 J/cmp 2. The stress lines induced by the collagen shrinkage appeared at a radiant exposure of 12.38 J/cmp 2 for the one-pulse treatment, and 10.26 J/cmp 2 for the five-pulse treatment. Clinical examination with the slit-lamp microscope immediately following the laser treatment revealed conical corneal lesions for all treatments with radiant exposures above 13.32 J/cmp 2. Significant light scattering was observed at all lesions produced by the five-pulse treatment. They appeared white and cloudy. However, all of the single-pulse lesions appeared transparent within the cone. The laser treatment changed their relative refractive index compared to the surrounding tissues, but the collagen still remained organized and no increase in light scattering was noted in the cone. Figures 6 and 7 illustrate the SEM comparison of a single-pulse treatment and five-pulse treatment at 18.01 J/cmp 2 on the cat and rabbit model, respectively. A singlepulse treatment created minimal surface disruption to the cat cornea and no endothelium damage (Fig 6), whereas endothelial damage was observed on the rabbit cornea (Fig 7). The five-pulse treatment created significant epithelium damage (although the surfaces remained attached) and endothelial damage to both cat and rabbit corneas. In all cases, collagen shrinkage was evident by tissue elevation at the laser treatment sites.
Figure 7: Scanning electron microscopic comparison of a single-pulse (A,B) versus a five-pulse (C1D) treatment at 18.01 J/cmp 2 on rabbit model: Both the single- and five-pulse treatments damage the epithelial (top row) and endothelial (bottom row) cells. This is due to the 360-micrometer penetration depth of the 2.10-micrometer Ho: YAG laser radiation which is similar to the rabbit's corneal thickness (320 to 350 µm). As in the cat (and human), the corneal thickness is >500 µm, and endothelial damage does not occur with a single laser pulse treatment at the 2.10-micrometer wavelength. (Spot size = 350 µm).
Figure 8: Typical appearance of toluidine blue-stained histological sections made across the laser photothermal keratoplasty lesions, with spot treatment patterns.
Volume of Tissue Shrinkage- Cadaver Eye Model
No laser photothermal keratoplasty lesion could be detected on the samples with a laser radiant exposure below 10.26 J/cmp 2. Figure 8 shows a typical example of histological sections of the laser photothermal keratoplasty treatment spots. The dark, conical portion of the cornea identifies the collagen shrinkage area in crosssection. The cross-sections for each laser lesion at radiant exposures of 13.4 J/cmp 2 or above were traced to provide an accurate estimation of the volume and depth of the shrinkage cone (Fig 9). Histological cross-sections showed that effects of the five-pulse treatment extended to the endothelial layer at the radiant exposure of 13.4 J/cmp 2, but no single-pulse treatment reached the endothelium for any radiant exposure level studied. The volume of shrinkage, however, did increase proportionally with higher laser radiant exposures for the single-pulse treatment. For the five-pulse treatments, the total volume of the shrunken collagen reached a maximal point at 15.43 J/cmp 2. Additional energy deposited in the same treatment area, such as five pulses at 18.01 J/cmp 2, caused essentially no increase in the volume of the shrunken collagen.
In Vivo Response- Spot Treatment on Cat Model
The mechanical deformation of the cat cornea to the eight-spot laser treatment (15.6 J/cmp 2, one pulse) was instantaneous when the laser pulse reached its surface. Laser-induced collagen shrinkage created stress lines around each laser spot and formed an octagonal deformation area centered around the pupil (Fig 10, top). Epithelial defects were observed at the center of each spot. The apex of the collagen-shrinkage cone extended about twothirds depth into the cornea and appeared transparent as observed above. The laser-induced stress lines and the collagen-shrinkage cone remained present and transparent after 3 months of healing (Fig 10, bottom). The histological sections across these lesions showed a denser keratocyte concentration at the laser-treated sites, a potential indication of scar formation (Fig 11).
Although thermokeratoplasty has been attempted and improved from cautery to microwave over a century,4"16 the lack of control of heat generation, extensive damage to the corneal tissue, and inadequate delivery systems hampered efforts to establish the efficacy of the procedure. Our study demonstrated that the amount of collagen shrinkage, and its location and geometrical shape, can be accurately and precisely controlled by an Ho:YAG laser coupled to a noncontact optical delivery system. It confirmed the results from an earlier, similar experiment as reported by Moreira and colleagues.17 Endothelial damage to the rabbit cornea was observed when the radiant exposure of the laser treatment was 18 J/cmp 2. Multiple laser energy deposition at this radiant exposure level caused the saturation of tissue shrinkage in volume and excessive thermal injury to the keratocytes and underlying structure.
The objective of laser photothermal keratoplasty is to predictably contract stromal collagen with minimal amounts of energy, time, and tissue involved for the refractive correction. Although both annulus and spot treatment patterns can be used for the laser photothermal keratoplasty procedure, there are several advantages of using a spot pattern rather than an annulus pattern. A spot treatment pattern requires less tissue for the same refractive change and causes less damage to epithelium and endothelium than the annulus pattern. It provides more flexibility in the treatment pattern. Additional corneal tissue can be treated in the same diameter, making additive refractive effect by rotating the polyprism. With spot treatment, the upper epithelium layer was always affected by the laser photothermal keratoplasty procedure. However, scanning electron micrographs of the cat and rabbit corneas showed that in spite of treatments with increasing energy doses (up to 18 J/cmp 2, five pulses), the basal epithelium was not ablated and covered the treated surface, allowing it to serve as a natural barrier to infection. The use of a contact lens during laser exposure has been suggested as a method of minimizing epithelial damage.
Figure 9: Contour maps of shrinkage cones as traced from histological slides for one pulse (left) and five pulses (right) at 13.4 J/cmp 2, 15.43 J/cmp 2, and 18.01 J/cmp 2. No collagen shrinking effect from a single-pulse treatment reaches the endothelial surface for any radiant exposure level studied; the volume of the shrunken collagen increases proportionally with higher laser radiant exposure. The effect of a five-pulse treatment reaches the endothelial surface at 13.4 J/cmp 2, the volume of the shrunken collagen reaches a maximum at 1 5.43 J/cmp 2, and an 18.01 J/cmp 2 dose causes essentially no increase in the volume of the shrunken collagen.
In the spot-treatment pattern, the energy distributed in each spot can be independently adjusted for uniform distribution. In contrast, the annulus pattern is easily subject to alignment error and laser beam nonuniformity, and may cause uncontrollable astigmatic error.
Our experiment and our previous studies2 both confirmed that a set of eight spots with one pulse at 15.6 J cmp 2 is sufficient in generating a refractive effect. However, one pulse at 18.01 J/cmp 2 provides the most predictable refractive correction without damaging the endothelial cells in the cat model and in human cadaver eyes. The use of a long sequence of pulses (eg, 5 or 20) caused excessive epithelial and endothelial damage by heat accumulation and conduction. Multiple-pulse treatments lengthen the procedure, and may be affected by saccadic ocular movements which might off-set the centration. The tissue shrinkage study also indicated that the laser-induced collagen shrinkage volume reaches a saturation point at 15.43 J/cm2, five pulses. Depositing more energy onto the same area causes no significant increase in shrinkage but does increase tissue damage.
Figure 10: Clinical appearance of an eight-spot laser photothermal keratoplasty treatment (15.6 J/cm2, one pulse, treatment zone = 3 mm, spot size = 350 µ?t?, pulse duration = 250 µe, radiant exposure = 15.6 J/cm2) on the cat" model immediately postoperative (top): The laser induced collagen shrinkage creates stress lines around each laser spot and forms an octagonal deformation area centered around the pupil, the collagen shrinkage cones extend into the cornea about two-thirds of its thickness and appear transparent; and 3-months postoperatively (bottom): the octagonal stress lines remain present and the collagen shrinkage cones remain transparent.
Figure 11: Light microscopy of laser photothermal keratoplasty treatment spot (15.6 J/cm2, one pulse) on cat cornea 3 months postoperatively. The higher keratocyte population density in the laser treated area might indicate formation of a scar.
The goal of laser photothermal keratoplasty is to produce a controlled stromal collagen shrinkage and cellular responses in the creation of a predictable and stable refractive change. When a laser photothermal keratoplasty eight-spot pattern at one pulse, 15.6 J/cm2 was applied to the cat cornea, both cellular and interstitial alternations were observed histologically. However, no changes in either the appearance of the octagonal stress lines in the anterior cornea or in the increased index of refraction in the treated cone under the laser spot were observed with the slit-lamp microscope over the 3-month follow-up period. This is an indication of biomechanical stability in the laser photothermal keratoplasty effect. It might be due to the observed cellular response - an increase in the keratocyte population density at the treated sites - which may, in turn, indicate the slow formation of a scar. Further studies are necessary to accurately quantify the laser photothermal keratoplasty biological, biomechanical, and optical effects. Clinical trials are needed to assess the long-term stability of laser photothermal keratoplasty.
1. Parel JM, Een QS, Simon G. Noncontact laser photothermal keratoplasty I: laser beam delivery system. J Refract Corneal Surg. 1994;10:511-518.
2. Simon G, Ren QS, Parel JM. Noncontact laser photothermal keratoplasty II: refractive effects, treatment parameters, and therapeutic applications. J Refract Corneal Surg. 1994;10:519-528.
3. Ren QS1 Simon G, Legeais JM, Parel JM, Culbertson W, Shen JH, Takesue Y, Savoldelli M. UV solid state laser 213 nm) photo-refractive keratectomy: in vivo study. Ophthalmology. In press.
4. Lans LJ. Experimentelle Untersuchungen über Entstehung von Astigmatismus durch nicht-perforirende comeawunden. GraefesArch Ophthalmol. 1889;44:117-152.
5. Gasset A, Shaw E, Kaufman H, Itoi M, Sakimoti T. Thermokeratoplasty. Transactions of the American Academy of Ophthalmology and Otolaryngology. 1973;77:441-454.
6. Aquavella J. Thermokeratoplasty. Ophthalmic Surg. 1974;4:39-48.
7. Keates R, Dingle J. Thermokeratoplasty for keratoconus. Ophthalmic Surg. 1975;6:89-92.
8. Aquavella J, Smith R, Shaw E. Alterations in corneal morphology following thermokeratoplasty. Arch Ophthalmol. 1976;94:2082-2085.
9. Fogle J, Kenyon K, Stark W. Damage to epithelial basement membrane by thermokeratoplasty. Am J Ophthalmol. 1977;83:392-401.
10. Caster AI. The Fyodorov technique of hyperopia correction by thermal coagulation: a preliminary report. Journal of Refractive Surgery. 1988;14:105-108.
11. Feldman ST, Ellis W, Frucht-Pery J, Chayet A, Brown SI. Regression of effect following radial thermokeratoplasty in humans. J Refract Corneal Surg. 1989;15:288-291.
12. Neuman AC, Sanders D, SaIz J. Radial thermokeratoplasty for hyperopia, J Refract Corneal Surg. 1989;5:50-54.
13. Neuman AC, Fyodorov S, Sanders DR. Radial thermokeratoplasty for the correction of hyperopia. J Refract Corneal Surg. 1990;6:404-412.
14. Neuman AC, Sanders D, Reanan M, DeLuca M. Hyperopic thermokeratoplasty: clinical evaluation. J Cataract Refract Surg. 1991;17:830-838.
15. Rowsey JJ, Doss JD. Preliminary report of Los Alamos keratoplasty techniques. Ophthalmology. 1981;88:755-760,
16. Rowsey JJ, Gaylor JR, Dahlstrom R. Los Alamos keratoplasty techniques. Contact Intraocular Lens Medical Journal. 1980;6:1-12.
17. Moreira H, Campos M, Sawusch MR, McDonnell JM, Sand B, McDonnell PJ. Holmium laser thermokeratoplasty. Ophthalmology. 1993;100:752-761.