The transepithelial single-step photorefractive keratectomy (TransPRK) procedure has rapidly gained popularity during the past decade due to successful logistical implementation by iVIS (CTEN),1 SCHWIND (SmartSurfACE),2 and, more recently, WaveLight One-Touch PRK (Streamlight)3 and Bausch & Lomb No-Touch PRK.4 The aim of single-step laser epithelial removal and stromal ablation is to remove the epithelial layer in a precise and uniform manner, all while allowing little room for unequal stromal hydration. This approach treats refractive errors by superimposing a defined epithelial thickness profile on a refractive ablation profile. Additionally, the diameter of epithelial removal matches the ablation zone (combination of optical and transition zones), thus decreasing the wound surface and speeding up the healing process. TransPRK also maximizes correspondence between the measured corneal topography and the ablation profile.
This technology has witnessed several modifications, but the core principle consists of applying an epithelial thickness ablation profile that resembles the general population epithelial mapping. By using very high-frequency digital ultrasonography, Reinstein et al.5 found that the central epithelial thickness (CET) in normal corneas averaged 53 µm centrally and 59 µm at the 6-mm periphery. Several studies revealed similar, albeit not identical, data using optical coherence tomography (OCT) and confocal microscopy.6,7
Epithelial thickness mapping, initially performed by very high-frequency ultrasonic imaging, has recently become a standard part of corneal imaging thanks to the development of high-resolution OCT systems.8 Although most devices display the measurement from the surface normal to the Bowman's membrane surface (Figure A, available in the online version of this article), thickness can be measured in multiple ways, each of which may have a clinical application.9 In refractive surgery, thickness measured by the normal-to-the-surface tangent (NTST) has its origins in radial keratotomy, where the diamond blade should be perpendicular to the corneal surface. Here the clinical application and method of measurement coincide.
Comparison of the normal-to-the-surface-tangent (NTST) vs parallel vertical sections (PVS) methods for determining/defining epithelial thickness. The thickness measured along the PVS reads thicker than measured along the NTST. AES = anterior epithelial surface; PES = posterior epithelial surface (anterior stromal surface)
Other methods of measuring thickness may be better suited to different clinical situations. Parallel vertical sections (PVS) (Figure A) emphasize vertical distance throughout the epithelium as opposed to incidental angle. In industry, thickness determined by parallel vertical measurements may be better suited for ablative or tissue removal devices, which move along the object's surface and whose laser beam is always almost vertical regardless of the object's surface architecture.
A variant of the PVS method would be thickness measured from a fixed point (Figure B, available in the online version of this article). Epithelial thickness will vary depending on the distance of the fixed point to the cornea. When the single point method is set to infinity, the rays behave as if they were parallel and the two maps (single fixed point and PVS) are equivalent. This scenario applies to most current lasers; the latter are not translational as they rotate or angle their beam, but at a virtually infinite working distance (ie, the distance is much larger than the epithelium thickness and the corneal anterior radius of curvature).
Illustrative demonstration of thickness measured along a line relative to a fixed point. The figure depicts eight different reference points (RefP) from above or below the corneal surface. RefP −∞ and ∞, correspond to parallel vertical sections (PVS); whereas RefP 8 mm resembles normal-to-the-surface-tangent (NTST) for normal corneas. AES = anterior epithelial surface; PES = posterior epithelial surface (anterior stromal surface)
Considering that the epithelial profile is subjected to variations in normal and especially corneas that had previous surgery, the aim of this theoretical analysis was to quantify the effect of different reference points, in particular NTST and PVS, on the read-out of epithelial thickness mapping, and to predict the refractive implications when used interchangeably in planning single-step TransPRK.
Patients and Methods
Modeling of Corneal Epithelial Profiles
We have generated corneal epithelial profiles based on three single parameters (central thickness, peripheral thickness, and peripheral thickness diameter), assuming a radial parabolic distribution of the cross-section of the profile. This simple model resulted in the following equations:
where CET is the central epithelial thickness, PET the peripheral epithelial thickness, PD the peripheral epithelial thickness diameter, and r the radial distance.
Modeling Read-outs of Corneal Epithelial Profiles
Numeric values coming for devices capable of measuring the epithelial thickness are usually reported in the form of averages of local finite areas (discs, annulus, and sections). We have assumed the reported epithelial profiles are based on two averages: one over a central disc of a given dimension and another over an annulus confined by known inner and outer diameters. This simple model resulted in the following equation:
where epithelial thickness is the average of the epithelial thickness within the region from inner diameter “a” to outer diameter “b” defined by the integrals. So that:
Having two averages, one over a more central annulus of a given dimension and another more peripheral annulus, each confined by known inner and outer diameters, one can solve the equation to get the local CET and PET for a particular PD, under the assumption of a radial parabolic distribution of the cross-section of the profile. This leads to:
where “A” and “B” are the diameters defining the first (more central) annulus and “C” and “D” defining the second (more peripheral) annulus, which for a central disc of a given dimension reduces to:
Modeling PVS of Corneal Epithelial Profiles
We have generated the simplest corneal surfaces based on the Baker's equation for aspheric shapes. This simple refractive model resulted in the following equation:
where z is the corneal sagitta at radial location r, R0 is the apical radius of curvature of the cornea, and Q is the asphericity quotient.
For the anterior corneal surface (which is equivalent to the anterior epithelial surface), the parameters R0 and Q are input values. For the anterior stromal surface (equivalent to the posterior epithelial surface), the parameters R0 and Q are calculated as follows.
First, the semi-axes of the ellipsoid have been calculated for the anterior corneal surface:
Using Equation 1, the local epithelial thickness at the radial position TransversalSemiAxisACS has been calculated (transversal epithelial thickness [TET]). With this, the semi-axes of the ellipsoid have been calculated for the anterior stromal surface:
With those, R0 and Q have been calculated for the anterior stromal surface:
and the PVS of corneal epithelial profiles are simply:
Quantification of the Refractive Power of the Epithelial Lens
For our simulations we have considered nEpithelium = 1.401 and nAir = 1.
For our simulations, we considered the average read-outs as the average of the 3-mm central disc (central epithelium) and the average of the 3- to 6-mm diameter annulus (peripheral epithelium). Additionally, we performed the simulations for a 7-mm peripheral diameter in the reconstructions. We have covered the keratometry values from 30.00 to 50.00 diopters (D), along with asphericities from −1.00 to +1.00 D, and the CET from 45 to 65 µm, along with differences from center to periphery in the average read-out from −10 to +10 µm.
Figure 1 shows four common cases corresponding to a normal cornea, after myopic laser keratorefractive surgery, after hyperopic laser keratorefractive surgery, and a keratoconus case. Epithelium measured along PVS becomes thicker toward the periphery. For all four examples, the average values (read-outs) show lower values than the local reconstructed epithelial profiles (NTST), and these are lower than the PVS values. Even for cases that apparently have a thinner peripheral epithelium compared to CET, the PVS show a thicker peripheral epithelium compared to CET.
Four common cases corresponding to (A) a normal cornea, (B) after myopic laser keratorefractive surgery, (C) after hyperopic laser keratorefractive surgery, and (D) a keratoconus case. Epithelium measured along parallel vertical sections (PVS) becomes thicker toward the periphery. For all four examples, the average values (read-outs) show lower values than the local reconstructed epithelial profiles (normal-to-the-surface-tangent [NTST]), and these are lower than the PVS values. Even for cases that apparently have a thinner peripheral epithelium compared to central epithelial thickness (CET), the PVS show a thicker peripheral epithelium (PET) compared to CET.
Effects of Geometry on the Epithelial Thickness Profile
Figure 2 summarizes the effects of keratometry, asphericity, CET, and difference central to periphery on the epithelial thickness profile measured along PVS. The steeper the cornea and the thicker the CET, the faster the rate of increase in epithelial thickness as we go toward the periphery, whereas corneal asphericity seems to have a much less important role. The central-to-periphery difference of the epithelial read-out has a relevant impact on the epithelial thickness profile measured along PVS.
Effects of (A) keratometry, (B) asphericity, (C) central epithelial thickness (CET), and (D) difference central to periphery on the epithelial thickness profile measured along parallel vertical sections (PVS) (at a fixed distance of 3.5 mm). The steeper the cornea and the thicker the CET, the faster the rate of increase in epithelial thickness toward the periphery, whereas corneal asphericity seems to have a much less important role. The central-to-periphery difference of the epithelial read-out has a relevant impact on the epithelial thickness profile measured along PVS. PET = peripheral epithelial thickness; D = diopters; NTST = normal-to-the-surface-tangent
Quantification of the Induced Refractive Error
Additional refractive errors are induced whenever the actual center-to-periphery difference in the corneal epithelial profile deviates from the difference in the applied epithelial ablation profile. The amount of additional refractive errors when considering the NTST read-outs instead of the PVS is a factor of the rate of peripheral growth, as shown in Figure 3. The error is restricted to 0.40 to 0.80 D, except for the cases in which a large difference from center to periphery is already observed in the read-outs. Because the PVS method provides steeper gradients than the NTST method, a hyperopic shift is induced that would lead to an overcorrection in myopia and an undercorrection in hyperopia.
Additional refractive errors are induced whenever the actual center-to-periphery difference in the corneal epithelial profile deviates from the difference in the applied epithelial ablation profile. The amount of additional refractive errors when considering the normal-to-the-surface-tangent (NTST) read-outs instead of the parallel vertical sections (PVS) is a factor of the rate of peripheral growth. The error is restricted to 0.40 to 0.80 diopters (D), except for the cases in which a large difference from center to periphery is already observed in the read-outs. Because the PVS method provides steeper gradients than the NTST method, a hyperopic shift is induced. This would lead to an overcorrection in myopia and an undercorrection in hyperopia. CET = central epithelial thickness; PET = peripheral epithelial thickness
A simulation for changing the reference point from NTST to PVS quantifying its effect on the read-out of epithelial thickness mapping has been presented. The simulation includes a simple modeling of corneal epithelial profiles and allows the analytical quantification of the refractive implications in terms of induced refractive error. The simple model uses a radial parabolic distribution for the cross-section of the profile, because a parabolic profile is the simplest derivable radial profile providing a well-defined center, and it approximates the profiles for corneal epithelium reported in the literature.10,11 Measurement differences based on different reference points have been discussed qualitatively in an editorial.9 In this study, we focused on epithelial thickness and quantified, rather than merely qualifying, the effects through a simple model.
The induced refractive error is directly related to the center-to-periphery epithelial thickness gradient: a hyperopic shift develops if the PVS epithelial profile is steeper than the NTST profile. A constant NTST epithelial profile would suggest applying a flat-depth phototherapeutic keratectomy (PTK) for laser epithelium removal, with equal depths at all ablation positions. This induces a hyperopic shift due to three reasons. First, because the central epithelium is thinner than the peripheral epithelium by approximately 10 µm in the PSV model, a flat-depth PTK would lead to a stromal ablation that is approximately 10 µm too thin/shallow peripherally.12,13 Second, removing corneal tissue leads to a marginally shorter axial length, inducing minimal hyperopia. This effect is minor and can be regarded as less than 0.25 D per 100 µm of tissue ablated.14 For a 55-µm central ablation, a +0.10 D hyperopic shift is expected. Third, in line with the radial compensation principle, peripheral loss of efficiency will result in less tissue ablated, enlarging the gap between central and peripheral stromal ablation. The loss of efficiency depends heavily on the laser system used, but it averages 25%, resulting in approximately +0.80 D of hyperopic shift.15,16 The summation of all of the three aforementioned effects leads to approximately +1.50 D of hyperopic shift. Thus, even for a constant NTST measured epithelium, the clinical experience would prevent using a constant PTK depth for the removal of the epithelial layer.
The refractive error induced by applying an NTST-derived epithelial profile instead of a PVS model in transepithelial laser ablation would be amplified in hyperopic eyes after laser ablation and minimized in myopic eyes after laser ablation because the normal center-to-periphery epithelial thickness gradient is minimized and often reversed after myopic correction and magnified after hyperopic correction.17,18 The epithelial gradient change, center-to-periphery, is even more pronounced the larger the amount of treated errors, whether myopic or hyperopic.19 An alternative risk for shallow corrections is that the peripheral epithelium may have not been completely removed (because PVS readings are much thicker than NTST readings toward the periphery).
The amplification of the expected refractive error derived from applying an NTST-derived epithelial profile instead of a PVS model is also true in keratoconic eyes, leading to a diameter ablated on the stroma smaller than planned. Haque et al.20 found that CET measured by OCT over four meridians was thinnest in keratoconic eyes (44 ± 7 µm), followed by normal eyes wearing rigid gas permeable contact lenses (50 ± 4 µm), and then normal eyes without contact lenses (54 ± 2 µm). Similar values for keratoconic eyes were reported by Reinstein et al.21,22
Another possible source of refractive error in TransPRK is the mismatch between the epithelial thickness profile and the epithelial ablation profile. First, the cornea being flatter nasally, the thickness of the epithelial layer may not be equally distributed from the nasal to temporal hemi-cornea, in contradistinction to the epithelial ablation.
Second, in corneas with appreciable toricity (leading to corneal astigmatism), the epithelial layer may also exhibit a different toricity than the underlying Bowman's membrane. A clinically significant increase in the amount of anterior astigmatism has been reported when moving from the anterior epithelial (air/tear film) interface to the Bowman's layer surface in an in vivo study using slit-scanning topography.23 This partial compensation demonstrates that there are azimuthal differences in the epithelial thickness distribution, because the epithelial thickness profile along the steepest meridian may differ from the thickness profile along the flattest meridian. The application of a rotationally symmetrical transepithelial profile on a toric corneal surface with underlying epithelial toricity may induce lenticules of wasted tissue (or variations in the achieved optical zone) with an elliptical perimeter. All of these uncertainties may lower the precision of astigmatic correction and limit any benefit provided by TransPRK ablations. Although this does not represent a major issue for the spherical component, marginal amounts of coma (decentered epithelium thinning compared to the visual axis) and astigmatism (different epithelial profile in all four cardinal directions) may be induced. The mean value of the error would be approximately 0.25 D worse in astigmatism correction with TransPRK for the normal population, but can peak to up to 0.63 D.
Third, the center of the laser ablation typically does not coincide with that of the thinnest epithelial spot, which could be located anywhere in the vicinity of the geometric center of the cornea. Depending on the surgeon's habits and the laser platform used, the excimer ablation is usually performed on the entrance pupil center, the corneal vertex, or in between these two. In eyes with large angle alpha and/or kappa, the discrepancy between the ablation center and the point of minimal epithelial thickness may even be larger. This would lead to either excessive, unpredictable, and asymmetrical stromal ablation or optical zone perimeter reduction, when delivering any TransPRK preset ablation. Similarly, the chances of such misalignment may be high in re-treatments or eyes with keratoconus. This limitation could be bypassed by creating an epithelial ablation profile with an offset on the thinnest epithelial point, while keeping the stromal ablation centered on target, being the corneal vertex, the pupil centroid, or in between; this would be especially useful in keratoconic eyes. The advent of high-resolution OCT technology has allowed proper representation of the epithelial layer, which may ultimately lead to customized epithelial ablations, circumventing such problems.
We have used an extrapolated TET at the edge of the transversal axis of the generating ellipsoid for the corneal surface. This may represent estimating the epithelium thickness using Equation 1 for radial distances exceeding 8 mm. Considering the geometry of the cornea (whose typical radius of curvature is 7.81 mm), this may be interpreted as a “stretched” extrapolation of the epithelium thickness, because the epithelium is measured up to an approximately 3.5-mm radius. Further, the intersection point of the transversal semi-axis (of the ellipsoidal cornea model) is not present on the real cornea. One could directly determine the stromal surface asphericity just within the region covered by the average epithelium thickness data (ie, the region where epithelium thickness is actually measured). Newer systems (eg, Optovue Inc., Irvine, CA, and MS-39; Costruzione Strumenti Oftalmici, Florence, Italy) report epithelium up to a 9-mm diameter (and can measure even wider areas). Some studies reported epithelial thicknesses at the limbal junction.10 The diameter of the limbus is classically considered to be 11.7 mm horizontally. Consejo et al.24 found 12 to 13 mm to be a better estimate of the limbus diameter, and in a recent review it was closer to 13 to 14 mm (extending the radial distance beyond 6.5 mm).25 The applied extrapolation for TET may go beyond these. Extrapolating to the edge of the transversal axis provides mathematical simplicity (determining radius and asphericity based on perpendicular axes) but does not change the determined values (RASS and QASS).
A mathematical advantage of NTST compared to PVS is that NTST thickness measurements are independent of tilt or gaze rotations, whereas PVS readings require coaxial alignment defining the vertical.
Read-outs from OCT or equivalent devices for measuring epithelial thickness using the NTST method can be directly used for the central point (or a small area around the center) in PVS or near PVS platforms such as excimer lasers. In that region, the difference between average read-outs and NTST local readings is minor and at the center NTST and PVS actually coincide. The wider the area or the annular region considered, the larger the differences between average read-outs and NTST, and between NTST and PVS. Thus, the risk for hyperopic shifts is maximal for wide peripheral areas.
The recent clinical availability of accurate and detailed corneal epithelial mapping has the potential to further refine TransPRK results and safety. However, unless a direct link between high-resolution OCT devices and excimer laser platforms is established, surgeons need to be cautious in manually transferring the epithelial measurement results into their ablation profiles. We have outlined the possible refractive implications using a simplified theoretical model, while providing a simplistic formula to convert NTST epithelial data read-out to PVS values that can be used for excimer laser ablation planning. Additionally, establishing a separate offset for the thinnest point in the epithelial ablation profile that matches the location of the measured corneal epithelial thinnest point is essential, especially in eyes with high aberration such as those with previous surgery or keratoconus.
- Chen X, Stojanovic A, Zhou W, Utheim TP, Stojanovic F, Wang Q. Transepithelial, topography-guided ablation in the treatment of visual disturbances in LASIK flap or interface complications. J Refract Surg. 2012;28(2):120–126. doi:10.3928/1081597X-20110926-01 [CrossRef]
- Aslanides IM, Kymionis GD. Trans advanced surface laser ablation (TransPRK) outcomes using SmartPulseTechnology. Cont Lens Anterior Eye. 2017;40(1):42–46. doi:10.1016/j.clae.2016.11.004 [CrossRef]
- Cummings AB, Joseph DA. Trans-epithelial PRK outcomes with a new one-touch continuous PTK/PRK procedure using a 500Hz excimer laser. Presented at the ASCRS/ASOA annual meeting. ; May 3–7, 2019. ; San Diego, California. .
- Braddon M. Bausch + Lomb releases new software on TECHNOLAS® TENEO™ 317 Model 2 providing surgeons with improved usability. EyeNews. September10, 2019.
- Reinstein DZ, Archer TJ, Gobbe M, Silverman RH, Coleman DJ. Epithelial thickness in the normal cornea: three-dimensional display with Artemis very high-frequency digital ultrasound. J Refract Surg. 2008;24(6):571–581. doi:10.3928/1081597X-20080601-05 [CrossRef]
- Feng Y, Simpson TL. Comparison of human central cornea and limbus in vivo using optical coherence tomography. Optom Vis Sci. 2005;82(5):416–419. doi:10.1097/01.OPX.0000162649.97059.72 [CrossRef]
- Patel SV, McLaren JW, Hodge DO, Bourne WM. Confocal microscopy in vivo in corneas of long-term contact lens wearers. Invest Ophthalmol Vis Sci. 2002;43(4):995–1003.
- Vega-Estrada A, Mimouni M, Espla E, Alió Del Barrio J, Alio JL. Corneal epithelial thickness intrasubject repeatability and its relation with visual limitation in keratoconus. Am J Ophthalmol. 2019;200:255–262. doi:10.1016/j.ajo.2019.01.015 [CrossRef]
- Villavicencio O, Belin MW, Ambrósio R Jr, Steinmueller A. Corneal pachymetry: new ways to look at an old measurement. J Cataract Refract Surg. 2014;40(5):695–701. doi:10.1016/j.jcrs.2014.04.001 [CrossRef]
- Feng Y, Simpson TL. Corneal, limbal, and conjunctival epithelial thickness from optical coherence tomography. Optom Vis Sci. 2008;85(9):E880–E883. doi:10.1097/OPX.0b013e318185272d [CrossRef]
- Wu S, Tao A, Jiang H, Xu Z, Perez V, Wang J. Vertical and horizontal corneal epithelial thickness profile using ultra-high resolution and long scan depth optical coherence tomography. PLoS One. 2014;9(5):e97962. doi:10.1371/journal.pone.0097962 [CrossRef]
- Camellin M, Arba Mosquera S. Simultaneous aspheric wavefront-guided transepithelial photorefractive keratectomy and phototherapeutic keratectomy to correct aberrations and refractive errors after corneal surgery. J Cataract Refract Surg. 2010;36(7):1173–1180. doi:10.1016/j.jcrs.2010.01.024 [CrossRef]
- Salah-Mabed I, Saad A, Gatinel D. Topography of the corneal epithelium and Bowman layer in low to moderately myopic eyes. J Cataract Refract Surg. 2016;42(8):1190–1197. doi:10.1016/j.jcrs.2016.05.009 [CrossRef]
- Arba-Mosquera S, de Ortueta D. Analysis of optimized profiles for ‘aberration-free’ refractive surgery. Ophthalmic Physiol Opt. 2009;29(5):535–548. doi:10.1111/j.1475-1313.2009.00670.x [CrossRef]
- Arba-Mosquera S, de Ortueta D. Geometrical analysis of the loss of ablation efficiency at non-normal incidence. Opt Express. 2008;16(6):3877–3895. doi:10.1364/OE.16.003877 [CrossRef]
- de Ortueta D, Arba Mosquera S, Häcker C. Theoretical considerations on the hyperopic shift effect observed when treating negative cylinder in laser refractive surgery. J Emmetropia. 2010;1:23–28.
- Reinstein DZ, Srivannaboon S, Gobbe M, et al. Epithelial thickness profile changes induced by myopic LASIK as measured by Artemis very high-frequency digital ultrasound. J Refract Surg. 2009;25(5):444–450. doi:10.3928/1081597X-20090422-07 [CrossRef]
- Ivarsen A, Fledelius W, Hjortdal JØ. Three-year changes in epithelial and stromal thickness after PRK or LASIK for high myopia. Invest Ophthalmol Vis Sci. 2009;50(5):2061–2066. doi:10.1167/iovs.08-2853 [CrossRef]
- Reinstein DZ, Archer TJ, Gobbe M, Silverman RH, Coleman DJ. Epithelial thickness after hyperopic LASIK: three-dimensional display with Artemis very high-frequency digital ultrasound. J Refract Surg. 2010;26(8):555–564.
- Haque S, Jones L, Simpson T. Thickness mapping of the cornea and epithelium using optical coherence tomography. Optom Vis Sci. 2008;85(10):E963–E976. doi:10.1097/OPX.0b013e318188892c [CrossRef]
- Reinstein DZ, Archer TJ, Gobbe M. Corneal epithelial thickness profile in the diagnosis of keratoconus. J Refract Surg. 2009;25(7):604–610. doi:10.3928/1081597X-20090610-06 [CrossRef]
- 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(4):259–271. doi:10.3928/1081597X-20100218-01 [CrossRef]
- Gatinel D, Racine L, Hoang-Xuan T. Contribution of the corneal epithelium to anterior corneal topography in patients having myopic photorefractive keratectomy. J Cataract Refract Surg. 2007;33(11):1860–1865. doi:10.1016/j.jcrs.2007.06.041 [CrossRef]
- Consejo A, Llorens-Quintana C, Radhakrishnan H, Iskander DR. Mean shape of the human limbus. J Cataract Refract Surg. 2017;43(5):667–672. doi:10.1016/j.jcrs.2017.02.027 [CrossRef]
- Bergmanson JP, Martinez JG. Size does matter: what is the corneolimbal diameter?Clin Exp Optom. 2017;100(5):522–528. doi:10.1111/cxo.12583 [CrossRef]