One of the potential advantages of small incision lenticule extraction (SMILE)1–4 over LASIK is that the potential errors associated with excimer laser ablation are minimized or avoided, such as stromal hydration,5 laser fluence,6–8 and other environmental factors.9 Therefore, it would be expected for there to be close agreement between the predicted and achieved refractive lenticule extraction lenticule depth, whereas there is usually a difference between the predicted and achieved ablation depth for excimer lasers due to the factors described above and the difficulty of verifying the ablation rate in vivo, which has been previously reported.10,11
The Artemis VHF digital ultrasound (ArcScan, Inc., Morrison, CO) technology is capable of measuring individual layers within the cornea in three dimensions across the central 10-mm diameter.12,13 The repeatability of single point thickness measurements using the Artemis VHF digital ultrasound arc-scanning system has been shown to be 1.68 μm for cornea and 1.78 μm for stroma.13 Measurement of the achieved excimer laser ablation depth by comparing Artemis stromal thickness data before and after LASIK has been previously described.10–13 Including the high repeatability, this method brings two main advantages. First, variable epithelial changes known to occur after LASIK14,15 are excluded because the stromal thickness is measured directly. Second, Artemis scans are centered on the corneal vertex so the same fixed location can be found more confidently before and after surgery, improving the potential error of finding the same measurement location compared with using a single-point handheld ultrasound pachymeter. Also, this method avoids potential error due to variability in tissue hydration that can affect intraoperative hand-held ultrasound pachymetry.5,16
The aim of the current study was to compare the readout lenticule thickness for myopic SMILE treatments using the VisuMax femtosecond laser with the Artemis measuring maximum stromal change.
Patients and Methods
This study was a retrospective noncomparative cases series of patients who underwent a SMILE procedure at the London Vision Clinic, London, United Kingdom, between May 2011 and December 2012. A complete ocular examination was performed to screen for corneal abnormalities and determine patient candidacy for refractive surgery. Eyes with ocular pathology such as keratoconus, corneal scar, corneal dystrophy, and previous ocular surgery were excluded. Our routine preoperative assessment, which has been described previously,15 including an Artemis VHF digital ultrasound scan, was performed.11–13 Informed consent and permission to use their data for analysis and publication were obtained from all patients.
All SMILE treatments were performed by a single surgeon (DZR) using the VisuMax 500-kHz femtosecond laser, which has been described in detail previously.1–4,17,18 Details of the geometry and software set-up of the SMILE lenticule, cap, and small incision have also been described previously.1–4 In all eyes, two access incisions were created: a 2-mm incision superonasally and a 3-mm incision superotemporally. The optical zone diameter was between 5.75 and 7.00 mm. A personalized nomogram adjustment was implemented after the first patient (2 eyes) was treated and used for all subsequent eyes.
During the SMILE procedure, at the moment of contact between the disposable curved contact glass and the cornea, a meniscus tear film appears, at which point the patient is able to see the fixation target clearly because the vergence of the system is focused according to the patient’s refraction. In this way, the patient essentially auto-centrates the system and hence the center of the lenticule on the corneal vertex, which closely approximates the visual axis. The femtosecond ablation was performed as described previously.1–4 Total suction time was approximately 35 seconds, independent of refractive error treated.
The 3-mm superotemporal incision was opened and the upper and lower edges of the lenticule were delineated so that the tissue planes were well defined. If it was not possible to delineate the edge of both interfaces, the second 2-mm superonasal incision was opened to find the edge(s) that had yet to be delineated. The upper interface was separated first using a standard lamellar corneal surgical technique of waving the instrument back and forth. The lower layer was then dissected in a similar fashion. Once both layers had been separated, the lenticule was removed from the cornea using a pair of retinal microforceps.
The patient was then taken to the slit lamp and fluorescein was instilled and the full distention of the cap centrally was achieved by a dry microspear sponge to ensure that any redundant cap (due to mismatch of cap vs bed length) was not left in the central cornea, but rather redistributed to the periphery.
Patients were instructed to wear plastic shields for 7 nights. Tobramycin and dexamethasone (Tobradex; Alcon Laboratories, Inc., Fort Worth, TX), and ofloxacin (Exocin; Allergan Ltd., Marlow, UK) were prescribed four times daily for the first week. Patients were followed up at 1 day and 1 and 3 months. Artemis VHF digital ultrasound was performed at the 3-month postoperative visit.
Artemis Lenticule Thickness Calculation
The method of calculating the lenticule thickness using the Artemis VHF digital ultrasound arc-scanner is identical to that of calculating excimer laser ablation depth, which has been described previously.10–12 Briefly, a single Artemis examination was performed before and at least 3 months after surgery. The lenticule depth was calculated by subtracting the postoperative Artemis stromal thickness data from the Artemis stromal thickness data obtained before surgery. A myopic SMILE lenticule is thicker in the center and this value is reported on the VisuMax readout. Because the SMILE procedures were all centered on the corneal vertex and the Artemis scans were also centered on the corneal vertex, the maximum ablation would theoretically be located at the Artemis (0, 0) coordinate. However, the Artemis measured lenticule thickness was taken to be the maximum difference in stromal thickness within the central 1-mm zone to account for any minor lenticule decentration or Artemis scan misalignment.
Obtaining the postoperative Artemis measurement at least 3 months after surgery ensures that postoperative edema had resolved and isolating the stromal thickness excludes any epithelial thickness changes. There might still be inaccuracies in the lenticule thickness value because biomechanical changes in the stroma were not considered. A summary of all potential errors associated with this method of measuring stromal thickness change have been described in detail previously.11
The difference between the VisuMax readout lenticule thickness and the Artemis measured stromal change was calculated. The mean, standard deviation, and minimum and maximum differences were found. A Student’s paired t test and linear regression analysis were performed and the coefficient of determination (R2) was calculated to investigate the correlation between the VisuMax readout and Artemis measured stromal change.
Finally, data from a previous publication on the accuracy of excimer laser ablation depth of wavefront-optimized ablations with the MEL80 excimer laser (Carl Zeiss Meditec)11 was reanalyzed to include only those eyes with ablation depth in the range found in the current SMILE population to compare the scatter in the data between treatment types according to the R2 of the linear regression analysis. The R2 were statistically compared using an online calculator,19 which converts each correlation coefficient into a z-score using Fisher’s r-to-z transformation and compares the z-scores while also taking sample size into account.
Microsoft Excel 2010 (Microsoft Corporation, Redmond, WA) was used for data entry and statistical analysis. A P value of .05 was considered statistically significant.
The study included 70 eyes of 37 patients. The contralateral eye in 3 patients underwent LASIK, whereas 1 patient only had one eye treated. The mean age of the population was 34.8 ± 11.1 years (range: 21 to 61 years). The mean maximum myopic meridian treated was −7.81 ± 2.33 diopters (D) (range: −2.25 to −12.50 D). The mean cylinder treated was 0.57 ± 0.42 D (range: 0.00 to 1.50 D). The mean average simulated keratometric value was 43.59 ± 1.53 D (range: 40.56 to 46.61 D). The mean central corneal thickness with handheld ultrasound was 532 ± 32 μm (range: 468 to 591 μm). The mean intraocular pressure was 16 ± 3 mm Hg (range: 9 to 22 mm Hg).
The programmed optical zone diameter was 5.75 mm in 6 eyes (9%), 6.00 mm in 22 eyes (31%), 6.25 mm in 12 eyes (17%), 6.50 mm in 16 eyes (23%), 6.75 mm in 3 eyes (4%), and 7.00 mm in 11 eyes (16%). The programmed cap thickness was 100 μm in 2 eyes (3%), 110 μm in 16 eyes (23%), 120 μm in 13 eyes (19%), 130 μm in 31 eyes (44%), and 140 μm in 8 eyes (12%). The programmed mean minimum thickness at the edge of the lenticule was 9 ± 4 μm (range: 3 to 25 μm).
Lenticule Depth: Maximum Change in Stromal Thickness
The mean VisuMax readout lenticule thickness was 129 ± 22 μm (range: 82 to 175 μm). The mean Artemis measured stromal change was 121 ± 21 μm (range: 82 to 167 μm). On average, the VisuMax readout lenticule depth was 8.2 ± 8.0 μm (range: −8 to +29 μm) thicker than the Artemis measured stromal change (P < .001). Figure 1 shows the linear regression analysis comparing the Artemis measured stromal change with the VisuMax readout lenticule thickness.
Scatter plot of Artemis (ArcScan, Inc., Morrison, CO) measured stromal thickness change (blue data points) and corneal thickness change (green data points) plotted against VisuMax (Carl Zeiss Meditec, Jena, Germany) readout lenticule thickness for myopic small incision lenticule extraction treatments. The regression equations and coefficients of determination (R2) are displayed. The red line represents equality between the measured stromal or corneal thickness change and the readout lenticule thickness. Points plotted above the red line mean that the VisuMax readout lenticule thickness overestimated the achieved stromal or corneal thickness change and points plotted below the red line mean that the VisuMax readout lenticule thickness underestimated the achieved stromal or corneal thickness change.
Comparison With Excimer Laser Ablation
Figure 2 shows the linear regression analysis comparing the Artemis measured stromal change with the MEL80 readout ablation depth for all eyes with an ablation depth between 80 and 180 μm for wavefront-optimized ablations (107 eyes), obtained from a previous study on MEL80 ablation depth.11 The R2 was 0.738 for the LASIK group, which was lower than the R2 of 0.868 for the SMILE group (P = .015).
Scatter plot of Artemis (ArcScan, Inc., Morrison, CO) measured stromal thickness change plotted against MEL80 excimer laser (Carl Zeiss Meditec, Jena, Germany) readout ablation depth for myopic LASIK treatments using wavefront-optimized ablations with the MEL80 excimer laser. The regression equations and coefficients of determination (R2) are displayed. The red line represents equality between the measured stromal change and the readout lenticule thickness or ablation depth.
Change in Central Epithelial Thickness
On average, central epithelial thickness was 15.0 ± 5.2 μm (range: 5 to 30 μm) thicker after the procedure. These epithelial thickness changes meant that the VisuMax readout lenticule depth was 23.2 ± 10.9 μm (range: +5 to +49 μm) thicker than the Artemis measured corneal thickness change, as shown in Figure 1.
Predictability of Spherical Equivalent Refraction
The 3-month spherical equivalent refraction was used to calculate the predictability of the refractive correction. The mean ± standard deviation predictability achieved (for the eyes using the personalized nomogram) was −0.04 ± 0.52 D (range: −1.13 to +1.50 D). The mean ± standard deviation predictability based on an analysis using the laser data entry would have been −0.78 ± 0.53 D (range: −2.01 to +0.82 D).
Direct measurement of the change in stromal thickness using VHF digital ultrasound found that the VisuMax lenticule thickness readout was 8 μm thicker on average than the achieved stromal thickness change. The accuracy of lenticule thickness in SMILE was found to be comparable, if not superior, to that of excimer laser ablation given the lower variability in the data; the R2 of 0.868 for the SMILE population was better than the 0.738 in the LASIK group. Because the same methodology of measuring the stromal change by VHF digital ultrasound was used in both studies, the scatter due to measurement error can be assumed to be identical. Therefore, the difference in variability can be attributed to the type of procedure.
The systematic difference of 8 μm could be due to one of three reasons (or a combination of these): (1) an error in the VisuMax cutting accuracy for one of the two layers; (2) an error with the stromal change measurement by Artemis VHF digital ultrasound; or (3) evidence for a biomechanical change in the stroma. The VisuMax has been shown to have high flap thickness predictability when creating LASIK flaps,20–24 and we have previously reported the reproducibility of cap thickness in SMILE to be 4.4-μm centrally and less than 6 μm in the majority of locations.4 Also, the femtosecond cutting process cannot be affected by outside factors, unlike the environmental,9 corneal hydration,5 and laser fluence factors6–8 that affect an excimer laser ablation. For there to be an error in the lenticule thickness due to VisuMax cutting accuracy, there would have to be an error in only one of the interfaces. However, in our previous study, we showed that the cap thickness was accurate with a central accuracy of −0.7 μm.4 Therefore, if the lenticule thickness difference was due to the VisuMax cutting accuracy, the error must have been in the lower interface of the lenticule. However, in our previous study, the accuracy was found to be similar for cap thicknesses between 80 and 140 μm,4 which provides evidence that the accuracy of the VisuMax does not vary with depth (although this needs to be confirmed for depths at which the lower interface of the lenticule is created).
With any measurement, there are always associated measurement errors. The method of using Artemis VHF digital ultrasound to measure the three-dimensional stromal thickness profile before and after surgery eliminates the error induced by epithelial thickness changes that would be included with any full corneal thickness change method.25 In addition, measurements by Artemis VHF digital ultrasound do not introduce further variable pachymetric error due to inclusion of the natural tear film, which is the case for epithelial measurements by current modern Fourier-domain optical coherence tomography scanning technology. The Artemis also has a high repeatability for corneal and stromal thickness measurements, so this source of error was minimized.13 In any event, errors such as these would be randomly distributed and would be likely to average out rather than result in a systematic error; measurement errors probably explain the majority of the scatter in the data. Another source of error is the alignment between the two scans. In contrast to the other sources of measurement error, alignment error could be expected to be more likely to occur in one direction. Because the corneal thickness is thinnest centrally and radially thicker toward the periphery, the lenticule is centered close to the thinnest point on the cornea in most cases unless the corneal thickness is significantly decentered from the corneal vertex. Therefore, any misalignment in the postoperative scan will mean that the thinnest point of the postoperative scan will not be aligned with the thinnest point of the preoperative scan. This means that in the majority of cases, an alignment error will tend to underestimate the change in stromal thickness, which was observed in this population.
However, it is unlikely that these alignment errors could explain a systematic difference of 8 μm because the pachymetric progression of the central stroma is relatively gradual.26 Therefore, the current study seems to provide evidence for some central stromal expansion caused by biomechanical changes occurring after SMILE. One possible mechanism could be that the lamellae severed by the lenticule in between the residual bed and the cap might be recoiling and causing expansion of the stroma because they are no longer under tension, similar to the known peripheral stromal expansion after LASIK.12,27 We intend to investigate the changes to peripheral stromal thickness in a future study. This expansion might be keeping the bottom lamellae of the cap slightly apart from the top lamellae of the residual bed (Figure 3). It seems unlikely that there would be any reason for the stroma in the residual bed or the cap to be expanding because they are still under tension. For example, the high accuracy of cap thickness that we have previously reported4 provides evidence for biomechanical stability within the cap.
Diagram of one possible biomechanical explanation for the systematic 8-μm difference between Artemis stromal thickness change and VisuMax lenticule thickness readout. It might be that the lamellae severed by the lenticule in between the residual bed and the cap (red lines) expand slightly given that they are no longer under tension. This expansion might be keeping the bottom lamellae of the cap slightly further apart from the top lamellae of the residual bed compared to the lamellae within the cap and the residual bed (represented by the green arrow).
It is almost inevitable that there will be biomechanical changes after any corneal surgical procedure, so it is not surprising that there was a difference between the theoretical and achieved lenticule thickness. The fact that the difference was only 8 μm, of which a proportion can be explained by measurement error, implies that there is actually little biomechanical change after SMILE, which might be expected given that only the stroma required is removed and that the strongest anterior stroma28 and Bowman’s layer29 remain intact. Also, the theoretical increased biomechanical stability of SMILE compared to LASIK30 may have contributed to the lower degree of scatter in the SMILE population compared with the LASIK population. Further study is required to verify this finding and investigate the source of the biomechanical changes.
There was a mean undercorrection of −0.78 D when comparing the outcome that would have been achieved using the laser data entry directly as the refraction. This may wholly or partly be explained by the difference in achieved lenticule thickness, particularly because −0.78 D corresponds to a 9-μm depth. However, to confirm whether the thickness difference was contributing a refractive effect would require a comparison with the whole diameter of the lenticule, but only the central lenticule thickness was available. Also, the refractive undercorrection might be partly due to epithelial thickness changes, which may not be fully accounted for by the lenticule profile design or the internal treatment nomograms.
The lowest readout lenticule thickness for any eye included in the curent study was 82 μm, so the accuracy of lenticule thickness for lenticules thinner than this could not be verified. However, it is reasonable to assume that the result would be similar. Because SMILE simply involves the creation of two intrastromal interfaces, the potential error is independent of the lenticule thickness and hence the refractive error being corrected.
The slope of the linear regression line was almost exactly 1.0, which indicates that the difference between intended and achieved lenticule thickness was similar for both low and high myopia. However, this does not take into account other variables such as optical zone diameter, cap thickness, corneal thickness, or intraocular pressure. This would be an interesting area to study further, but we were not able to answer this question with the current population because of the uneven distribution of the different variables. For example, the majority of the eyes with a larger optical zone were for lower myopia. Therefore, a multivariate regression analysis would be extrapolating into a region with few or no data points (ie, high myopia with large optical zone). This means that any results would be unreliable because of the skewed population bias.
The central epithelial thickness was found to become thicker after SMILE to compensate for the stromal tissue that has been removed, in the same way that it does after myopic LASIK.14,15 Analysis of the lenticule thickness using the change in corneal thickness therefore found a significantly different result, which demonstrates that change in corneal thickness is an unreliable method for measuring stromal tissue removal.
The VisuMax readout agreed closely with the achieved lenticule thickness, although there appears to be evidence for some biomechanical changes after SMILE. The accuracy of VisuMax lenticule creation is comparable to or slightly better than the accuracy of excimer laser ablation. Given all of the above discussion, it is impressive that the lenticule thickness predicted mathematically was so close to that which was eventually measured in the biological system using the most accurate methods available today.
- 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;95:335–339. doi:10.1136/bjo.2009.174284 [CrossRef]
- Shah R, Shah S, Sengupta S. Results of small incision lenticule extraction: all-in-one femtosecond laser refractive surgery. J Cataract Refract Surg. 2011;37:127–137. doi:10.1016/j.jcrs.2010.07.033 [CrossRef]
- Hjortdal JØ, Vestergaard AH, Ivarsen A, Ragunathan S, Asp S. Predictors for the outcome of small-incision lenticule extraction for myopia. J Refract Surg. 2012;28:865–871. doi:10.3928/1081597X-20121115-01 [CrossRef]
- Reinstein DZ, Archer TJ, Gobbe M. Accuracy and reproducibility of cap thickness in small incision lenticule extraction (SMILE). J Refract Surg. 2013;29:810–815. doi:10.3928/1081597X-20131023-02 [CrossRef]
- Dougherty PJ, Wellish KL, Maloney RK. Excimer laser ablation rate and corneal hydration. Am J Ophthalmol. 1994;118:169–176.
- Seiler T, Kriegerowski M, Schnoy N, Bende T. Ablation rate of human corneal epithelium and Bowman’s layer with the excimer laser (193 nm). Refract Corneal Surg. 1990;6:99–102.
- Mrochen M, Seiler T. Influence of corneal curvature on calculation of ablation patterns used in photorefractive laser surgery. J Refract Surg. 2001;17:S584–S587.
- Arba-Mosquera S, de Ortueta D. Geometrical analysis of the loss of ablation efficiency at non-normal incidence. Opt Express. 2008;16:3877–3895. doi:10.1364/OE.16.003877 [CrossRef]
- Schena E, Silvestri S, Franzesi GT, Cupo G, Carito P, Ghinelli E. Theoretical model and design of a device to reduce the influence of environmental factors on refractive surgery outcomes. Conf Proc IEEE Eng Med Biol Soc. 2006;1:343–346. doi:10.1109/IEMBS.2006.260184 [CrossRef]
- Reinstein DZ, Srivannaboon S, Archer TJ, Silverman RH, Sutton H, Coleman DJ. Probability model of the inaccuracy of residual stromal thickness prediction to reduce the risk of ectasia after LASIK: Part I. Quantifying individual risk. J Refract Surg. 2006;22:851–860.
- Reinstein DZ, Archer TJ, Gobbe M. Corneal ablation depth readout of the MEL80 excimer laser compared to Artemis three-dimensional very high-frequency digital ultrasound stromal measurements. J Refract Surg. 2010;26:949–959. doi:10.3928/1081597X-20100114-02 [CrossRef]
- Reinstein DZ, Silverman RH, Raevsky T, et al. Arc-scanning very high-frequency digital ultrasound for 3D pachymetric mapping of the corneal epithelium and stroma in laser in situ keratomileusis. J Refract Surg. 2000;16:414–430.
- Reinstein DZ, Archer TJ, Gobbe M, Silverman RH, Coleman DJ. Repeatability of layered corneal pachymetry with the Artemis very high-frequency digital ultrasound arc-scanner. J Refract Surg. 2010;26:646–659. doi:10.3928/1081597X-20091105-01 [CrossRef]
- 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;25:444–450. doi:10.3928/1081597X-20090422-07 [CrossRef]
- Reinstein DZ, Archer TJ, Gobbe M. Change in epithelial thickness profile 24 hours and longitudinally for 1 year after myopic LASIK: three-dimensional display with Artemis very high-frequency digital ultrasound. J Refract Surg. 2012;28:195–201. doi:10.3928/1081597X-20120127-02 [CrossRef]
- Silverman RH, Patel MS, Gal O, et al. Effect of corneal hydration on ultrasound velocity and backscatter. J Ultrasound Med. 2009;35:839–846. doi:10.1016/j.ultrasmedbio.2008.11.009 [CrossRef]
- Sekundo W, Kunert K, Russmann C, et al. First efficacy and safety study of femtosecond lenticule extraction for the correction of myopia: six-month results. J Cataract Refract Surg. 2008;34:1513–1520. doi:10.1016/j.jcrs.2008.05.033 [CrossRef]
- Reinstein DZ, Archer TJ, Gobbe M, Johnson N. Accuracy and reproducibility of Artemis central flap thickness and visual outcomes of LASIK with the Carl Zeiss Meditec VisuMax femtosecond laser and MEL 80 excimer laser platforms. J Refract Surg. 2010;26:107–119. doi:10.3928/1081597X-20100121-06 [CrossRef]
- Preacher KJ. Calculation for the test of the difference between two independent correlation coefficients; 2002. Available at http://www.quantpsy.org/corrtest/corrtest.htm. Accessed July 10, 2013.
- Ju WK, Lee JH, Chung TY, Chung ES. Reproducibility of LASIK flap thickness using the zeiss femtosecond laser measured postoperatively by optical coherence tomography. J Refract Surg. 2011;27:106–110. doi:10.3928/1081597X-20100428-04 [CrossRef]
- Yu ZQ, Xu Y, Yao PJ, Qin B, Zhou XT, Chu RY. Analysis of flap thickness by anterior segment optical coherence tomography in different flap preparation styles of excimer laser surgery [article in Chinese]. Zhonghua Yan Ke Za Zhi. 2010;46:203–208.
- Yao P, Xu Y, Zhou X. Comparison of the predictability, uniformity and stability of a laser in situ keratomileusis corneal flap created with a VisuMax femtosecond laser or a Moria microkeratome. J Int Med Res. 2011;39:748–758. doi:10.1177/147323001103900306 [CrossRef]
- Ahn H, Kim JK, Kim CK, et al. Comparison of laser in situ keratomileusis flaps created by 3 femtosecond lasers and a microkeratome. J Cataract Refract Surg. 2011;37:349–357. doi:10.1016/j.jcrs.2010.08.042 [CrossRef]
- Issa A, Al Hassany U. Femtosecond laser flap parameters and visual outcomes in laser in situ keratomileusis. J Cataract Refract Surg. 2011;37:665–674. doi:10.1016/j.jcrs.2010.10.049 [CrossRef]
- Reinstein DZ, Gobbe M, Archer TJ. Inaccuracies in reporting the accuracy of flap creating devices. J Refract Surg. 2011;27:784–785. doi:10.3928/1081597X-20110908-01 [CrossRef]
- Reinstein DZ, Archer TJ, Gobbe M, Silverman R, Coleman DJ. Stromal thickness in the normal cornea: three-dimensional display with Artemis very high-frequency digital ultrasound. J Refract Surg. 2009;25:776–786. doi:10.3928/1081597X-20090813-04 [CrossRef]
- Roberts C. The cornea is not a piece of plastic. J Refract Surg. 2000;16:407–413.
- Randleman JB, Dawson DG, Grossniklaus HE, McCarey BE, Edelhauser HF. Depth-dependent cohesive tensile strength in human donor corneas: implications for refractive surgery. J Refract Surg. 2008;24:S85–S89.
- Seiler T, Matallana M, Sendler S, Bende T. Does Bowman’s layer determine the biomechanical properties of the cornea?Refract Corneal Surg. 1992;8:139–142.
- Reinstein DZ, Archer TJ, Randleman JB. Mathematical model to compare the relative tensile strength of the cornea after PRK, LASIK, and small incision lenticule extraction. J Refract Surg. 2013;29:454–460. doi:10.3928/1081597X-20130617-03 [CrossRef]