Examination of the cornea by very highfrequency

Transcription

1 Epithelial and Stromal Changes Induced by Intacs Examined by Three-dimensional Very Highfrequency Digital Ultrasound Dan Z. Reinstein, MD, MA(Cantab), FRCSC; Sabong Srivannaboon, MD; Simon P. Holland, MB, MRCP, FRCS, FRCSC ABSTRACT PURPOSE: To examine epithelial and stromal layers by three-dimensional very high-frequency (VHF) digital ultrasound scanning before and after implantation of Intacs (intracorneal ring segments [ICRS]). METHODS: Three-dimensional scanning was performed in five eyes before and 3 months after Intacs insertion. Digital signal processing techniques provided high-resolution B-scan imaging and I-scan traces for high-precision (1-µm) threedimensional pachymetry. Thickness maps of individual corneal layers were constructed of the epithelium, stroma, and full cornea before and after surgery. Difference maps for epithelium and stroma were produced to examine anatomical changes in the thickness profile induced in each layer and correlate these to refractive changes. RESULTS: B-scan examination revealed stromal and epithelial anatomy anterior and adjacent to the Intac. Ring depth could be measured topographically. There was stromal lamellar displacement by the ring segments that produced a concave From Weill Medical College of Cornell University, New York, NY (Reinstein), Centre Hospitalier National d'ophtalmologie des Quinze Vingts, Paris, France (Reinstein), and the University of British Columbia, Department of Ophthalmology, Vancouver, BC, Canada (Reinstein, Srivannaboon, Holland). Certain aspects of the ultrasound technology described in this article are protected by U.S. and international patents. Patents are administered by the Cornell Research Foundation, Ithaca, New York. Supported in part by NIH grant EY01212, the Dyson Foundation, the St. Giles Foundation, and Research to Prevent Blindness, Inc. Presented in part at the American Society of Cataract and Refractive Surgery Annual Meeting, Boston, MA, April 13, 1999 and at the Association for Research in Vision and Ophthalmology Annual Meeting, May 17, Dr. Reinstein has a proprietary interest in the ultrasound technology employed in this study. Drs. Srivannaboon and Holland have no proprietary interest. Preparation in partial fulfillment of the requirements for the doctoral thesis, University of Cambridge, for Dr. Reinstein. The authors thank Nadina Bukowski for technical assistance. Correspondence: Dan Z Reinstein, MD, MA(Cantab), FRCSC, Department of Ophthalmology, Weill Medical College of Cornell University, 1300 York Ave., Rm A-855, New York, NY Fax: ; DanReinstein@compuserve.com Received: November 13, 2000 Accepted: February 16, 2001 anterior stromal groove within an annulus central to the ring. Epithelial filling of this concavity was shown in three dimensions in such a way as to produce orthogonally asymmetrical flattening of the corneal surface, thus potentially accounting for induced astigmatism. Mapping of the central stroma demonstrated thickening, potentially also accounting for astigmatic changes ascribable to orthogonal asymmetry. CONCLUSIONS: VHF digital ultrasound scanning provided imaging and three-dimensional thickness mapping of corneal layers, enabling anatomical evaluation of the changes induced in the cornea by Intacs. [J Refract Surg 2001;17: ] Examination of the cornea by very highfrequency (VHF) ultrasound is a relatively new experimental application in refractive surgery. In 1993, we reported the first confirmed measurement of the epithelium of the cornea in vivo using VHF ultrasound, and demonstrated that acoustic interfaces were located spatially at the epithelial surface and the interface between epithelial cells and the surface of Bowman s layer. 1 By using digital signal processing techniques (I-scan), reproducibility has been improved to 0.61 µm for epithelial and 0.74 µm for corneal thickness measurements. 2,3 By combining multiple scan planes, it has been possible to produce three-dimensional epithelial thickness mapping 4 as well as that of other distinct layers, such as the flap within the cornea. 5 Application of this biometric and imaging technology has yielded useful information for the study of corneal surgery. 3 It has been possible to characterize central epithelial anatomy and demonstrate that the power of the epithelium may not be constant from eye to eye (Reinstein DZ, Aslanides IM, Patel S, et al. Epithelial lenticular types of human cornea: classification and analysis of influence on PRK. Ophthalmology 1995:102(suppl):S156). The shape of Bowman s layer 6, measurement of anterior 310 Journal of Refractive Surgery Volume 17 May/June 2001

2 corneal scars for planning therapeutic keratectomy 7,8, objective quantification of corneal haze after photorefractive keratectomy (PRK) 9, and measurement of the depth of radial keratotomy incisions 10 have been studied. Analysis of epithelial and stromal changes after lamellar corneal surgery has demonstrated that significant epithelial changes occur after uncomplicated laser in situ keratomileusis (LASIK) and has shown the masking of stromal surface irregularities producing optical complications. 5 In April 1999, the United States Food and Drug Administration approved the intracorneal ring segments (ICRS; Intacs, KeraVision Inc., Fremont, CA), for mild nearsightedness (-1.00 to diopters [D]), with mild astigmatism (1.00 D or less). Approval was based on a review of safety and effectiveness data on 450 eyes. After 12 months, 97% of eyes were corrected to 20/40 or better and 74% were corrected to 20/20 or better. Thirty-nine (8.6%) patients chose to have their Intacs removed because of side effects or because they were unhappy with the corrected vision. Of those, 19 (4.2%) reported symptoms such as glare, halos, and problems with night vision. Approximately 4% of eyes were found to have more than 1.00 D of surgically induced astigmatism. No study to date has reported on the layered anatomical changes induced within the cornea after insertion of Intacs. We present a pilot study to examine the layered anatomical changes demonstrated by VHF ultrasound three-dimensional scanning. PATIENTS AND METHODS Patient Preparation One eye in each of five consecutive patients undergoing insertion of Intacs was included. Patients were scanned before and a median of 3 months after insertion of Intacs for the treatment of myopia. This study conformed to the tenets of the Declaration of Helsinki and volunteers were scanned after their fully informed consent. Our scanning system and patient set-up has been described. 3 Briefly, patients were placed in the supine position and scanned using a standard ophthalmic immersion technique. A plum-bob and fixation target above the eye not being scanned provided vertical alignment of both visual axes as well as fixation of the eye during the scan sequence. Scanning System and Procedure A very high-frequency (50 MHz) broadband transducer (Panametrics, Inc., Waltham, MA) was controlled in a reverse-arc motion to follow the corneal contour and enable an 8- to 10-mm wide corneal B-scan to be acquired in a single sweep. 11 Details of the scanning procedure have been described. 3 Briefly, three-dimensional scan sets consist of meridional scans at 45 intervals. Each scan sweep is performed in approximately 0.5 seconds. Data Analysis The radio frequency ultrasound data digitized and stored were subsequently processed to B-scan images for visualization and I-scan traces for biometry 2, using a speed-of-sound constant for the cornea of 1640 m/s. 12 The thickness of each corneal layer was then derived from the distance between surfaces in the radial direction (perpendicular to the back surface of the cornea). 13 The depth of the ring segment within each B-scan was determined for topographic analysis of ring depth. A simple linear polar/radial interpolation function was used to interpolate between scan meridians for three-dimensional plotting. For plotting of thickness mapping, we imported the above data-matrix into Deltagraph v.4.5 for Macintosh (SPSS Inc., Richmond, CA). Surface fill plots X,Y, Z were employed to display thickness data on a color scale. Color scales were adjusted subjectively to delineate variability in thickness profile of the corneal layer in question. RESULTS Clinical Refractive Outcome The Table summarizes clinical refractive results, which appear within the expected range of published outcomes. 14 Before surgery, all eyes had a magnitude of refractive astigmatism of less than or equal to 0.50 D of cylinder. Vector analysis using the standard angle-doubling method demonstrated an average (SD) surgically induced astigmatism of 1.26 (0.44) D. VHF Ultrasound Two-dimensional B-scan Examination Figure 1 shows a horizontal corneal B-scan through the visual axis 3 months after insertion of Intacs in the left eye of patient 1. The epithelial layer is clearly resolved from the stromal layer over the entire 9.5-mm diameter B-scan image. The Journal of Refractive Surgery Volume 17 May/June

3 Table Clinical and Refractive Data* Including Vector Analysis for Five Eyes Patient Eye Age Preoperative Refraction (D) Postoperative Refraction (D) Vector Analysis No. (yr) Sphere Cylinder Axis SE Sphere Cylinder Axis SE Surgically Axis Induced (deg) (deg) Induced of Error On-axis Astig From Cylinder (D) Original (D) Axis 1 OS OD OS OD OS Mean SD ±1.05 ±0.21 ±1.09 ±0.41 ±0.31 ±0.29 ±0.44 ±20.15 ±0.67 Range *Mean, standard deviation, and range are calculated for each parameter. There was induced astigmatism in all eyes. Spherical equivalent Positive equals clockwise Figure 1. VHF ultrasound B-scan horizontal section of a cornea after insertion of Intacs (above). A photograph demonstrating the positioning of the Intacs relative to the pupil shows the scan plane (red dotted line). There is stromal bulging anterior and posterior to the Intacs, with epithelial compensation for the stromal surface changes induced. Three-dimensional thickness maps (color scale in µm) derived from multiple meridional scanning of the cornea are shown (bottom row) representing on a color scale in microns the epithelial thickness (bottom left), stromal layer thickness (bottom, middle), and full corneal thickness (bottom, right). Epithelial profile changes are consistent with the induction of cylinder in this eye. X,Y grid is 1 mm. 312 Journal of Refractive Surgery Volume 17 May/June 2001

4 Figure 2. Photograph of Intacs demonstrates inferior decentration when compared to the two black circles that are concentric with the limbus (left). Three-dimensional thickness maps (color scale in µm) derived from multiple meridional scanning of the cornea are shown representing the epithelial thickness before surgery (center left), and 3 months after surgery (center right). A topographic subtraction map between epithelial layers before and after surgery is shown (right). Epithelial difference mapping demonstrates asymmetric changes that are consistent with the induction of refractive cylinder. X,Y grid is 1 mm. polymethylmethacrylate (PMMA) Intacs are visualized in cross-section within the peripheral stroma. Changes in epithelial thickness are apparent over and around the Intacs. The posterior surface of the cornea behind the Intacs is not visualized due to the acoustic shadowing produced by the intervening PMMA. By extrapolation of the corneal back surface adjacent to the zones of the ring, it is possible to estimate the percentage depth achieved on ring insertion. Gross inspection of the stromal surface (Bowman s layer) shows forward arching of stromal tissue by the space occupying effect of the Intacs. The posterior surface is similarly arched inward. There is relative epithelial thinning over the Intacs. The stromal surface within the zone just central to the forward arch produced by the ring is concave before regaining its convex shape more centrally. This concavity in Bowman s surface is effectively filled in by epithelial thickening. The epithelium thins again in a centripetal fashion. Thus, a visible portion of the contour changes induced by the Intacs is modified by epithelial compensation. VHF Ultrasound Three-dimensional Pachymetry Figure 1 shows the epithelial, stromal, and full corneal thickness profiles constructed from the multiple meridional B-scans in patient 1, three months after Intacs insertion. There was an on-axis surgically induced astigmatism of DC at approximately 13 (Table). The epithelial thickness within the central 4-mm zone is approximately 45 µm. The epithelium is thinned to 21 µm over the Intacs. There is annular thickening central to the Intacs, maximal (80 µm) along the 2 to 8 o clock meridian and minimal (50 µm) overlying the gaps between PMMA segments along the 11 to 5 o clock meridian. There is epithelial compensation over the stromal surface curvature that will produce relatively more corneal surface flattening along the 2 to 8 o clock meridian relative to the perpendicular axis. The induced astigmatism may be, at least in part, due to the difference in epithelial thickening between the inner annular zone where there were gaps between the rings and the epithelial changes within the zone of ring placement. Figure 2 shows a photograph of the right eye of patient 4 with inferior decentration of the rings. Vector analysis showed an on-axis surgically induced astigmatism of DC at 170. An epithelial change map calculated by subtraction of the preoperative from postoperative epithelial profile shows similar annular epithelial thickening central to the ring segments. There is, again, differential epithelial compensation in the axis of the gap between segments relative to the perpendicular axis through the ring segments. This epithelial compensation respects ring localization rather than the corneal center and again will cause increased flattening in the horizontal relative to the vertical axis. Figure 3 demonstrates layer-by-layer thickness mapping of the epithelium, stroma, and full cornea, before and after Intacs insertion in the left eye of patient 3. There was an on-axis surgically induced astigmatism of DC at approximately 90. Temporal grouping is in columns and anatomical grouping is in rows. The preoperative epithelial, stromal, and full corneal thickness maps appear in the first column; corresponding postoperative maps appear in column two. The third column shows calculated maps for epithelial and stromal layer thickness change. The bottom right-hand video image shows the positioning of the Intacs. The epithelial thickness profile before surgery shows a relatively homogeneous profile 50 to 52 µm within the central zone, thickening inferonasally to approximately Journal of Refractive Surgery Volume 17 May/June

5 Figure 3. Photograph of Intacs in the left cornea of a 42-year-old woman demonstrates relative ring positioning with respect to the pupil (bottom right). Three-dimensional thickness maps (color scale in µm) derived from multiple meridional scanning of the cornea are shown. Column 1 depicts preoperative layers: epithelium (top), stroma (middle), and full cornea (bottom). Column 2 depicts corneal layers after surgery: epithelium (top), stroma (middle), and full cornea (bottom). Topographic subtraction maps between corneal layers before and after surgery are shown in column 3: epithelial change (top), and stromal layer change (middle). Figure 3 demonstrates relatively symmetrical annular epithelial changes, but asymmetric central stromal thickness changes consistent with the induction of refractive cylinder. X,Y grid is 1 mm. 54 µm. After surgery, central epithelial thickness decreased from 46 to 48 µm. The epithelial difference map shows annular epithelial thickening of approximately 30 µm just within the ring zone and epithelial thinning up to 5 µm centrally within the cornea. In this eye, the annulus of epithelial thickening bridges the gaps between segments, appears circumferentially homogeneous, and would not explain induced astigmatism. Comparing the stromal thickness maps from before and after surgery (respectively, the first and second maps within the second row), there was stromal thickening within the central 6-mm zone induced by surgery, despite the lack of direct surgical intervention within the central cornea. The stromal difference map (column 3, second row) shows that the changes in stromal 314 Journal of Refractive Surgery Volume 17 May/June 2001

6 Figure 4. Right eye of a 46-year-old woman (patient 4) after implantation of 30-µm-thick Intacs. Photograph of Intacs demonstrates relative ring positioning with respect to the pupil (bottom right). Three-dimensional thickness maps derived from multiple meridional scanning of the cornea are shown represented on a color scale in microns; X,Y grid is 1 mm. Column 1 depicts preoperative layers: epithelium (top), stroma (middle), and full cornea (bottom). Column 2 depicts the corneal layers after surgery: epithelium (top), stroma (middle), and full cornea (bottom). Topographic subtraction maps between corneal layers before and after surgery are shown in column 3: epithelial change (top) and stromal layer change (middle). Despite well-centered Intacs insertion, there was induced astigmatism in this eye. Both epithelial and stromal asymmetrical changes could explain the induction of cylinder. thickness were not symmetrical; along the vertical meridian there was stromal thickening of approximately 25 to 30 µm inferiorly with 10 to 15 µm thickening superiorly. Along the horizontal meridian there was 25 to 20 µm thickening temporally and 20 to 25 µm nasally. There was relatively more thickening in the horizontal meridian (and therefore more flattening along the horizontal meridian). In contrast to the previous cases, this case demonstrates that changes in the stromal layer alone may also account for induction of astigmatism. Figure 4 shows the VHF ultrasound analysis of the right cornea of a 46-year-old woman (patient 4), which demonstrated surgically induced astigmatism Journal of Refractive Surgery Volume 17 May/June

7 316 Journal of Refractive Surgery Volume 17 May/June 2001 Figure 5. Thickness mapping before and 3 months after Intacs insertion. The layout of maps is similar to those of Figures 3 and 4, with the addition of column 4, depicting ring depth within the stroma alone (first row) and within the cornea from the surface of the epithelium (second row). The depth of the ring at the time of insertion (excluding epithelial changes) is obtained by topographic addition of the epithelium before surgery (first column, first row) to the ring depth within the stroma 3 months after surgery (fourth column, first row). There appears to be a progressive increase in Intacs depth from the site of insertion. X,Y grid is 1 mm. VHF Ultrasound in Intracorneal Ring Segment (Intacs) Surgery/Reinstein et al

8 of D at axis 35. The epithelium thickened more in the horizontal than the vertical meridian. The stromal change map (second row, third column) demonstrates thickening throughout the central cornea, with more thickening along the meridian at approximately 55. This stromal layer change is consistent with the D of cylinder at 55 induced by surgery. This eye shows a combined epithelial and stroma contribution to induced astigmatism. Figure 5 shows VHF ultrasound of the left cornea of a 42-year-old woman with surgically induced astigmatism of D at 126. The video-capture photograph shows inferonasal ring decentration with respect to the pupil. There is cumulatively more epithelial thickening in the meridian at 145 (superonasally and inferotemporally) than in the opposite axis with noticeably less thickening inferonasally. This is consistent with differential flattening along the 145 meridian, as evidenced by the patient s postoperative refraction of plano x 145. The stromal thickness maps (second row) demonstrate asymmetric changes induced by surgery; the temporal side of the corneal stroma appears to have thickened by approximately 20 to 25 µm, and the nasal side by approximately 10 to 15 µm. The region overlying the entry point for insertion of the rings shows an area of stromal thinning of approximately 10 µm. Depth of Ring Segment Figure 5 also depicts the depth of the temporal intracorneal ring (third column) within the stromal layer (top row), measured from the surface of the epithelium (second row) and measured assuming no changes in the epithelial layer (third row). Because of epithelial changes induced by surgery, to find the original depth of insertion, it is necessary to add the stromal depth after surgery to the epithelial thickness preoperatively. This temporally displaced topographic addition of the epithelial and stromal layers enables determination of the surgical accuracy of insertion. Ring depth within the stromal layer (top row) shows a depth superiorly of approximately 380 µm, decreasing to approximately 300 µm inferotemporally. Examination of ring depth from the corneal (epithelial) surface reveals a depth superiorly of 430 µm, decreasing to approximately 370 µm inferotemporally. Not including epithelial changes induced by surgery, the depths superiorly and inferotemporally at the time of insertion were 440 µm and 385 µm, respectively. Although the ring was inserted within a particular stromal lamellar plane from the site of insertion superonasally, the depth of the ring from the surface of the cornea increased as the ring was advanced further along its lamellar channel. Another possibility is that the stromal component of the cornea above or below the ring has changed in thickness due to compression of the stromal lamellae by the space occupying effect of the ring. DISCUSSION We have demonstrated that examination of the cornea before and after insertion of Intacs by VHF three-dimensional ultrasound has potential use in understanding of the mechanism of refractive effects. The epithelium appears to possess the ability to smooth changes in the stromal surface in LASIK 3,5 and PRK. 4 After Intacs insertion, large and refractively significant epithelial profile thickness changes may occur as a result of compensation for stromal surface shape changes over and around the region of the inserted rings. There appear to be changes in the stromal component of the cornea within the central corneal zone, despite no direct surgical intervention within the central cornea. The central stroma may thicken after insertion of Intacs. Although the central epithelium thins (probably in a compensatory fashion), this still results in an effective increase in corneal central thickness. Histological animal studies have documented only localized changes around the ring. 15,16 Whether the central corneal thickness observed in this study is due to biomechanical changes within the central cornea due perhaps to compression/rarefaction of stromal lamellar packing, or a change in endothelial cell function perhaps in the region posterior to the inserted rings (and thus hidden from direct visualization for counting or morphographic study) remains to be determined. VHF ultrasound technology is capable of determining the depth of insertion of Intacs. It appears that the intralamellar dissection that occurs during insertion may not always be parallel to the surface of the cornea, implying that corneal lamellar packing density may vary from superior to inferior within the same annular radial region of the cornea. As ring depth is known to correlate with the magnitude of refractive effect induced by the Intacs, perhaps this variable can lead to asymmetrical changes in refraction and induction of ocular aberrations such as coma. Aberrometry (Mrochen M, Kaemmerer M, Mierdel P, Krinke HE, Seiler T. Principles of Journal of Refractive Surgery Volume 17 May/June

9 Tscherning aberrometry. J Refract Surg 2000;16:S570-S571) or spatially resolved refractometry (Burns SA. The spatially resolved refractometer. J Refract Surg 2000;16:S566-S569), coupled with corneal surface calculated aberrometry 17, may explain some of the subjective visual complaints that can lead to explantation of the Intacs. VHF ultrasound provides high-resolution anatomical visualization and high-precision biometry for the assessment of Intacs surgery. VHF ultrasound may elucidate the mechanisms of Intacsinduced astigmatism and can measure central corneal thickness changes, separating stromal and epithelial components. This data may be useful in improving mathematical predictive modeling to improve refractive accuracy of Intacs. REFERENCES 1. Reinstein DZ, Silverman RH, Coleman DJ. High-frequency ultrasound measurement of the thickness of the corneal epithelium. Refract Corneal Surg 1993;9: Reinstein DZ, Silverman RH, Rondeau MJ, Coleman DJ. Epithelial and corneal thickness measurements by high-frequency ultrasound digital signal processing. Ophthalmology 1994;101: Reinstein DZ, Silverman RH, Raevsky T, Simoni GJ, Lloyd HO, Najafi DJ, Rondeau MJ, Coleman DJ. A new arc-scanning very high-frequency ultrasound system for 3D pachymetric mapping of corneal epithelium, lamellar flap and residual stromal layer in laser in situ keratomileusis. J Refract Surg 2000;16: Reinstein DZ, Silverman RH, Trokel SL, Coleman DJ. Corneal pachymetric topography. Ophthalmology 1994; 101: Reinstein DZ, Silverman RH, Sutton HF, Coleman DJ. Very high-frequency ultrasound corneal analysis identifies anatomic correlates of optical complications of lamellar refractive surgery: anatomic diagnosis in lamellar surgery. Ophthalmology 1999;106: Patel S, Reinstein DZ, Silverman RH, Coleman DJ. The shape of Bowman's layer in the human cornea. J Refract Surg 1998;14: Reinstein DZ, Silverman RH, Trokel SL, Allemann N, Coleman DJ. High-frequency ultrasound digital signal processing for biometry of the cornea in planning phototherapeutic keratectomy [letter] [published erratum; Arch Ophthalmol 1993;111:926]. Arch Ophthalmol 1993;111: Reinstein DZ, Aslanides IM, Silverman RH, Asbell PA, Coleman DJ. High-frequency ultrasound corneal pachymetry in the assessment of corneal scars for therapeutic planning. Clao J 1994;20: Allemann N, Chamon W, Silverman RH, Azar DT, Reinstein DZ, Stark WJ, Coleman DJ. High-frequency ultrasound quantitative analyses of corneal scarring following excimer laser keratectomy. Arch Ophthalmol 1993;111: Lazzaro DR, Aslanides IM, Belmont SC, Silverman RH, Reinstein DZ, Muller JW, Lloyd HO, Coleman DJ. High frequency ultrasound evaluation of radial keratotomy incisions. J Cataract Refract Surg 1995;21: Silverman RH, Reinstein DZ, Raevsky T, Coleman DJ. Improved system for sonographic imaging and biometry of the cornea. J Ultrasound Med 1997;16: Coleman DJ, Woods S, Rondeau MJ, Silverman RH. Ophthalmic ultrasonography. Radiol Clin North Am 1992;30: Segall M, Reinstein DZ, Johnson NF. Computer aided analysis and visualization of high-frequency ultrasound scanning of the human cornea. IEEE Computer Graphics Applications 1999;19: Schanzlin DJ. Studies of intrastromal corneal ring segments for the correction of low to moderate myopic refractive errors. Trans Am Ophthalmol Soc 1999;97: Quantock AJ, Kincaid MC, Schanzlin DJ. Stromal healing following explantation of an ICR (intrastromal corneal ring) from a nonfunctional human eye. Arch Ophthalmol 1995;113: D'Hermies F, Hartmann C, von Ey F, Holzkamper C, Renard G, Pouliquen Y. Biocompatibility of a refractive intracorneal PMMA ring. Fortschr Ophthalmol 1991;88: Oshika T, Klyce SD, Applegate RA, Howland HC, El Danasoury MA. Comparison of corneal wavefront aberrations after photorefractive keratectomy and laser in situ keratomileusis. Am J Ophthalmol 1999;127: Journal of Refractive Surgery Volume 17 May/June 2001