Introduction
Corneal refractive surgeries, such as laser-assisted in situ keratomileusis (LASIK) and keratorefractive lenticule extraction (KLEx), have become increasingly popular as the prevalence of myopia has increased globally [
1]. In response to this rise, many advancements in excimer laser platforms have been made to make these procedures more accurate, effective, and safer for patients [
2]. WaveLight® Contoura Vision topography-guided LASIK (TG-LASIK), approved by the US Food and Drug Administration (FDA) in 2016, utilizes preoperative corneal topography measurements in its excimer laser platform to create a customized ablation to maintain the aspheric corneal shape and minimize surgically induced corneal irregularities [
3,
4].
KLEx has been a viable alternative to traditional femtosecond-assisted LASIK, as it boasts a “flapless” approach and avoids flap-related complications [
5]. Although studies have demonstrated that KLEx and LASIK provide patients with similar postoperative visual outcomes [
6,
7], KLEx may also potentially induce less iatrogenic dry eye in the early postoperative period and may better maintain the cornea’s biomechanical properties [
8,
9]. However, these findings have been controversial among other studies.
Several studies have compared visual and refractive outcomes and higher-order aberrations (HOAs) of TG-LASIK and KLEx [
4,
10‐
13]. One study also assessed epithelial remodeling between these two procedures [
14]; however, to the best of our knowledge, no studies have evaluated corneal biomechanics or extensive corneal topographic parameters between these two procedures. In this prospective, contralateral study, we aim to assess and compare visual and refractive outcomes, HOAs, various corneal tomographic parameters such as corneal and epithelial thickness, and corneal biomechanics of TG-LASIK and KLEx.
Methods
This is a single-center randomized, prospective contralateral controlled trial that was conducted at the Hoopes Vision Research Center in Draper, Utah (NCT05611294). This study was approved by the Salus Institutional Review Board (IRB) (IORG00005674) and adhered to the tenets of the Declaration of Helsinki. Informed consent was provided and documented in written form from each patient prior to the start of intervention.
Thirty-four patients (68 total eyes) with myopia (− 7.5 D to − 0.75 D) with or without astigmatism (− 2.75 D to 0 D) were recruited for participation in this study. To be eligible for inclusion, all patients met the following criteria: age of 22–50 at the time of consent, diagnosed with myopia or myopia with astigmatism with a preoperative spherical equivalent (SEQ) greater than or equal to − 2.00 and smaller than or equal to − 9.00 D, a preoperative spherical component of greater than or equal to − 2.00 and less than or equal to − 8.00 D, a refractive cylinder of greater than or equal to − 3.00 D, corrected distance visual acuity (CDVA) of 20/20 or better in each eye, a stable refraction defined as a change in SEQ no greater than 0.50 D comparing the screening visit manifest refraction to previous refractions over a 1-year period prior to surgery. Patients meeting any of the following criteria for either eye were excluded from the study: clinically significant dry eye, irregular astigmatism, keratoconus, abnormal corneal topography, a history of corneal dystrophies or guttata, herpetic keratitis or active disease, prior refractive surgery, glaucoma or glaucoma suspect, uncontrolled diabetes, unstable hypertension, unstable autoimmune disease, or female patients who were pregnant, breastfeeding, or intending to become pregnant at any time during the study, as determined by verbal inquiry. Subjects were then randomized to receive TG-LASIK in one eye and small incision lenticule extraction (SMILE®) in the contralateral eye via selection of sealed envelopes containing cards indicating one of the two choices for right eye treatment. Card selection was done at the time of surgery scheduling.
All patients received a complete preoperative ophthalmic evaluation which included uncorrected distance visual acuity (UDVA), CDVA, best corrected 25% and 5% contrast sensitivity testing under mesopic conditions (without glare), intraocular pressure (IOP), slit lamp examination, dilated fundus examination, manifest refraction, and keratometry. Corneal topographic mapping and Zernike analysis for HOAs (total HOAs, spherical aberration (SA), horizontal/vertical coma, horizontal/oblique trefoil) were performed with a Pentacam HR (Oculus, Arlington, WA, USA). Corneal pachymetry was obtained with the Avanti Widefield optical coherence tomography (OCT) system (Optovue, Fremont, CA, USA) and corneal biomechanics were measured with the Ocular Response Analyzer® G3 (ORA) (Reichert Technologies, Depew, NY, USA). Patients were also provided with an adaptation of the Quality of Vision (QoV) [
15] questionnaire to assess the following qualities both preoperatively and postoperatively: glare, disabling halos/rings/starbursts, dry-eye issues, and eye preference.
Patients were evaluated postoperatively at 1 day, 1 week, 1 month, and 3 months. Visual acuity was measured at each visit. Refractive outcomes and corneal tomography were measured at 1 week, 1 month, and 3 months. Contrast sensitivity, corneal biomechanics, and patient-reported outcomes (PROs) were measured at 1 and 3 months. HOAs were evaluated only at 3 months.
Surgical Interventions
Both TG-LASIK and SMILE were performed sequentially on each subject on the same day by a single surgeon (MM) with 26 and 8 years of experience in performing LASIK and SMILE, respectively. Prior to starting TG-LASIK, multiple topographical measurements with the WaveLight® VARIO Topolyzer (Alcon Laboratories, Inc., Fort Worth, TX, USA) were utilized for corneal mapping for customized ablation treatment. At least four of these measurements were chosen on the basis of mean deviation. VARIO results were uploaded into the Phorcides Analytic Engine (PAE) (Phorcides LLC, St. Paul, MN, USA) for further precision and customization of the ablation pattern to account for corneal asymmetry. The updated VARIO results were then relayed to the WaveLight® EX500 Excimer Laser (Alcon Laboratories Inc., Fort Worth, TX, USA) for stromal ablation. The FS200 femtosecond laser system was used to create a superiorly hinged corneal flap with an 8.9-mm diameter and 100 µm thickness. The following laser settings were used: 0.72 µJ laser-bed energy, 7.0 µm bed spot/line, 0.72 µJ side-cut energy, and 5.0/3.0 µm side-cut spot/line.
For SMILE, our previously established nomogram from Datagraph-med® version 5.80 (Datagraph, Chapel Hill, NC, USA) based on 1-year outcomes was used for target spherical and cylindrical correction. Patients with a cylinder of ≤ 0.50 D underwent myopic SMILE based on SEQ, whereas those with a cylinder of ≥ 0.75 D received myopic astigmatic SMILE based on nomogramic data. The VisuMax 500 kHz femtosecond laser system (Carl Zeiss Meditec AG, Jena, Germany) was used with pulse durations of 220–580 fs and with the following laser parameters: 3.9 mm superior corneal incision, small cone size, 7.5 mm cap diameter, 120 µm cap thickness, 6.0–6.5 mm lenticule diameter, superior incision position, 90° side-cut angle, 50° incision angle, laser-bed energy of 145 nJ, and 3 µm, 2.5 µm, 3 µm, and 2 µm spot separations for the lenticule, lenticule side-cut, cap, and incision side-cut, respectively.
After completion of both procedures, patients received one drop of moxifloxacin 0.5% ophthalmic solution eye drops in both eyes. The postoperative drop regimen included moxifloxacin four times daily for 1 week in both eyes, prednisolone acetate 1% ophthalmic suspension drops four times daily for 1 week in both eyes, and twice daily for an additional 2 weeks in the SMILE eye. Preservative-free artificial tears were used as needed.
Statistical Analysis
Statistical analysis was conducted using Excel (Microsoft Corporation, Redmond, WA, USA) and Statistical Package for the Social Sciences (SPSS) version 27.0.0.0 (IBM, Armonk, NY, USA). Variable distributions were evaluated for normality with Kolmogorov–Smirnov and Shapiro–Wilk testing. Non-normalized variables were treated with transformations for use in statistical testing. Homogeneity of variances were tested with Levene’s test. A significance level of above 0.05 was used to judge acceptable normality and homogeneity. Multicollinearity analysis was performed to observe the variance inflation factor, of which no variable surpassed a threshold of five, signifying no excessive correlations. Chi-square testing was performed for comparisons between categorical variables. Analysis of variance (ANOVA) with post hoc Tukey testing was conducted on preoperative variables between the two surgery groups to evaluate for differences. Two-way analysis of covariance (ANCOVA) testing with pairwise comparisons was performed in multivariate fashion on all postoperative variables, with repeated measures versions used when necessary. Factors utilized were surgery type and time. All ANCOVA models incorporated covariates of age and sex to compensate for their effects on postoperative outcomes. ANCOVA models had general power above 0.99 and powers above 0.8 for each reported variable observation. Bonferroni corrections were used in estimated marginal mean calculation with pairwise comparison in ANCOVA testing to decrease type I errors. Standard refractive surgery graphs were created, and vector analysis was performed using mEYEstro (MathWorks Inc, Natick, MA, USA) [
16] and astigMATIC software (McGill University, Montreal QC, Canada) [
17], respectively. Cylinder is depicted in these materials with negative notation converted to positive notation. A threshold of 0.05 was used to define statistical significance for reported observations.
Discussion
In our study, TG-LASIK showed faster visual recovery at 1 day and 1 week compared to SMILE, but both were similarly effective from 1 to 3 months. Yang et al. found TG-LASIK significantly more effective than SMILE at 1 day, with 100% of eyes achieving 20/20 or better (compared to 71% of SMILE) [
10]. However, Kanellopoulos showed TG-LASIK significantly more effective at 3 months, with 9.1% gaining two or more lines of UDVA (compared to 4.1% of SMILE) [
13] (Table
7). In an earlier retrospective study Moshirfar et al. reported that at 12 months, 10% of TG-LASIK eyes lost one or more lines of UDVA (compared to 19% of SMILE) [
12]. The differences in the rate of visual recovery may be due to higher energy levels used in SMILE, as lower energy settings have been associated with improved UDVA on postoperative day 1 [
18]. In our study, both groups had excellent safety profiles, with no significant differences. Kang and Zhang et al. similarly found no safety differences between TG-LASIK and SMILE, with over 90% of eyes in both groups experiencing no change or improvement in lines of CDVA 6 months postoperatively [
4,
11].
Table 7
Studies comparing outcomes of TG-LASIK and SMILE
Kanellopoulos (2017) Prospective contralateral study | 22, 22 | 22 | 29.5 | 3 M | – | – | – | ≥ 20/20: 86.4% vs. 68.2% (p < 0.002) ≥ 20/16: 59.1% vs. 31.8% (p < 0.002) | Gain of 1 line: 64% vs. 36% *Gain of 2 lines: 9.1% vs. 4.5% (*p < 0.002) No change: 27% vs. 59% | – | 95.5% vs. 77.3% (p < 0.002) |
Moshirfar et al. (2019) Retrospective non-contralateral comparative study | 249, 357 | 212, 357 | TG-LASIK: 34 ± 9.3 SMILE: 33.1 ± 7.2 | 12 M | 1 line: 27% vs. 22% 2 lines: 10% vs. 1% (p = 0.57) | 57% vs. 74% | 1 line: 2% vs. 2% | ≥ 20/20: 93% vs. 89% (p > 0.05) ≥ 20/16: 65% vs. 59% (p > 0.05) | loss of 1 + lines: 10% vs. 19% (p = 0.009) | – | 95% vs. 95% (p > 0.05) |
Yang et al. (2021) Retrospective non-contralateral comparative study | 23, 26 | 23, 26 | TG-LASIK: 23.7 ± 5.18 SMILE: 23.4 ± 4.09 | 3 M | – | – | – | ≥ 20/13: 30% vs. 11% (p < 0.05) | – | – | – |
Kang et al. (2022) Retrospective non-contralateral comparative study | 45, 45 | 45, 45 | TG-LASIK: 24.9 ± 4.86 SMILE: 25.9 ± 5.34 | 6 M | No change/gain lines: 97.8% vs. 93.3 (p < 0.05) | – | ≥ 20/20: 97.8% vs. 93.3% ≥ 20/16: 80% vs. 68.9% | – | 44% vs. 44% (p > 0.05) | 87% vs. 89% (p > 0.05) |
Zhang et al. (2022) Prospective case-series study | 81, 102 | 81, 102 | TG-LASIK: 29.5 ± 6.7 SMILE: 28.2 ± 6.1 | 6 M | 1 line: 14% vs. 10% 2 lines: 17% vs. 12% (p > 0.05) | 65% vs. 72% | 1 line: 4% vs. 7% | – | – | 54% vs. 55% (p > 0.05) | 95% vs. 92% (p > 0.05) |
Current study (2024) Prospective contralateral study | 34, 34 | 34 | 33.7 ± 6.0 | 3 M | 1 line: 59% vs. 35% 2+ lines: 3% vs. 6% | 35% vs. 56% | 1 line: 3% vs. 3% 2+ lines: 0% vs. 0% | ≥ 20/20: 97% vs. 97% ≥ 20/16: 91% vs. 79% (p > 0.05) | Gain 1 line: 35% vs. 18% No change: 56% vs. 53% Loss of 1 line: 6% vs. 26% Loss of 2 lines: 3% vs. 3% | 44% vs. 56% (p > 0.05) | 100% vs. 94% (p > 0.05) |
In our study, both platforms demonstrated excellent and comparable predictability, with over 40% of eyes in each group achieving a SEQ within ± 0.13 D of plano. Our findings are comparable to other studies, as Kang et al. showed that 44% of TG-LASIK and SMILE eyes were within ± 0.13 D of target, whereas Zhang et al. reported that 54% of TG-LASIK eyes were within ± 0.13 D of target (compared to 55% of SMILE) [
4,
11].
Despite original presumptions that TG-LASIK would reduce total HOAs, our study showed that both groups had a significant increase in total HOAs (27% increase for TG-LASIK vs. 53% increase for SMILE) from the preoperative period to 3 months, but the TG-LASIK group significantly induced less total HOAs, vertical coma, and oblique trefoil than SMILE. Conversely, Yang et al. demonstrated that both groups did not show any significant increases in total HOAs from preoperative to 3 months, but similarly demonstrated that the TG-LASIK group exhibited significantly lower induced total HOAs compared to the SMILE group [
10]. Similar to Kang et al., our study showed that SMILE induced significantly more vertical coma than TG-LASIK [
11]. TG-LASIK’s customized ablation treatment reduces susceptibility to errors caused by pupil centroid shifts due to changes in pupil size, improving correction of peripheral corneal irregularities [
19]. Additionally, SMILE has been shown to cause more vertical decentration than TG-LASIK, specifically around the area of lenticule extraction, potentially contributing to the higher amount of vertical coma [
20,
21]. However, newer SMILE platforms such as the VisuMax 800 (Carl Zeiss Meditec AG, Jena, Germany), which was recently approved by the FDA, will offer cyclotorsion and centration control, potentially addressing the induced vertical coma seen with older platforms [
22].
Epithelial remodeling after traditional LASIK and SMILE has been well documented. Although controversial, the current proposed mechanism is that changes in corneal curvature from stromal ablation or lenticule extraction may incite corneal epithelial cell migration from the limbus [
23,
24]. Changes in epithelial thickness are thought to contribute to myopic regression in patients after corneal refractive surgery [
23]. Rocha et al. observed significant central epithelial thickening after LASIK at 1 and 3 months postoperatively [
25], whereas Luft et al. observed an increase in CET to 63.0 ± 3.6 µm after SMILE 6 months postoperatively [
26]. Our study showed that both TG-LASIK and SMILE groups had significant and comparable increases in amount and variation (max − min CET; SD CET) in epithelial remodeling during the first week and remained relatively stable throughout the rest of the postoperative period. Kanellopoulos similarly demonstrated that CET increased comparably for both groups but noted stability from 3 to 24 months and a significantly greater change in thickness variation (max − min CET) for the TG-LASIK group from 1 to 12 months [
14]. Interestingly, our study showed that the SMILE group had a significantly thicker nasal CET than the TG-LASIK group at the 7.0-mm zonal area at 3 months. One study demonstrated that SMILE treatments have a tendency to be decentered nasally due to their lack of cyclotorsion control, resulting in more corneal tissue being removed in that area [
21]. This could potentially explain the compensatory epithelial hyperplasia and remodeling observed in our SMILE group in the nasal 7.0 mm zonal area.
Corneal biomechanics is a relatively new area of focus, as studies have found that these properties may play a role in understanding myopic regression after corneal refractive surgery [
27,
28]. Prior studies have shown that SMILE provides better biomechanical outcomes than traditional LASIK, likely due to less disturbance of the corneal surface with its flapless approach [
27,
28]; however, to the best of our knowledge, no studies have compared TG-LASIK and SMILE with regards to these parameters. Our study showed that although both CH and CRF decreased in both groups postoperatively, there were no significant intergroup differences at any point in time. These results demonstrate that TG-LASIK is not biomechanically inferior to the flapless procedure, SMILE.
The utilization of a contralateral study design enables patients to serve as their own controls, eliminating potential biases arising from differences in healing rates, socioeconomic factors, or environmental influences between subjects. Concerning limitations, we realize that our sample size only consisted of 34 patients (68 eyes) and may limit the generalizability and validity of our outcomes; however, our statistical power for all variables was over 0.8. Another limitation is that a one-dimensional diagnostic tool (ORA) was used to measure CH and CRF, only two of the many biomechanical properties. The lack of comprehensive biomechanical data may limit the validity of our results. Although our study group was only followed up to 3 months postoperatively, it is generally known that visual outcomes stabilize around 3 months after LASIK or SMILE [
10,
13]; however, as seen in Kanellopoulos’ study [
14], prolonged follow-up of these patients could reveal variations in certain parameters, such as epithelial thickness.