Comparison of Corneal Densitometry and Visual Quality after Small Incision Lenticule Extraction (SMILE) and Laser Epithelial Keratomileusis (LASEK): One-Year Comparative Study

Purpose To investigate changes in corneal densitometry (CD) and visual quality following small incision lenticule extraction (SMILE) and laser epithelial keratomileusis (LASEK) in patients with mild-to-moderate myopia. Methods A retrospective analysis was performed on 24 and 25 patients (46 eyes each) who underwent SMILE and LASEK, respectively, for mild-to-moderate myopia. The visual quality and CD values were recorded. Using the Pentacam Scheimpflug system, CD values were collected in three concentric optical zones at the depths of the anterior, central, and posterior layers. Efficacy, safety, predictability, corneal wavefront aberrations, and QoV scores were measured to evaluate visual quality. A correlation analysis was performed between changes in CD and clinical characteristics. Results There were no statistical differences in efficacy and safety indices between the two groups. At 3 months postoperatively, a pronounced reduction in several zones was observed in the LASEK group (p < 0.05), whereas no obvious change was observed in the SMILE group. There were obvious changes in the CD values in several zones in the SMILE and LASEK groups (p < 0.05) after 1 year. The magnitude of the CD changes in the anterior and central corneal layers was smaller in the SMILE group than in the LASEK group (all p < 0.05). Lower HOAs, spherical aberration, and horizontal comas of the anterior and whole corneal surfaces were observed in the SMILE group. QoV scores were similar between the two groups. Conclusion CD decreased in the SMILE and LASEK groups after 1 year; there was a smaller reduction in SMILE than in LASEK. SMILE and LASEK did not differ significantly in terms of safety and effectiveness in correcting mild-to-moderate myopia.


Introduction
Myopia is the most frequent cause of distance impairment and a global public health concern, with an increasing annual prevalence. It is estimated that globally, by 2050, 4758 and 938 million people will have myopia and high myopia, respectively, which would constitute 49.8% and 9.8% of the world population, respectively [1]. Corneal refractive surgery remains the mainstay of ocular refractive correction as the cornea provides three-fourth of ocular refractive power. Corneal refractive surgery has been widely performed to correct myopia and is recognized as an effective, safe, and predictable strategy [2].
Small incision lenticule extraction (SMILE) is an advanced technology of corneal refractive surgery without flap creation, mainly depending on a 3-4 mm side cut for lenticule extraction [3]. Laser epithelial keratomileusis (LASEK) mainly utilizes an excimer laser to correct myopia. Both procedures are safe and effective [4,5].
Corneal transparency is indispensable for achieving optimal vision. It can be measured via corneal light backscatter and is commonly displayed as corneal densitometry (CD) values [6]. The CD values can be investigated via several medical devices, such as in vivo confocal microscopy [7], Oculus Pentacam [8], and optical coherence tomography [9]. The Pentacam Scheimpflug system (Oculus GmbH, Wetzlar, Germany) was used for scanning the cornea and recording CD values. This is a noninvasive and convenient approach for evaluating the backscatter profile of the anterior, central, and posterior layers of the entire cornea. This method has been used to evaluate the progression of various ocular surface diseases, including corneal dystrophies [10] and keratoconjunctivitis [11], and corneal clarity after refractive corneal surgery [12,13]. Previous studies reported that the CD values increased quickly in the first 24 h after SMILE [14]. Meanwhile, Wei et al. showed that the CD values decreased significantly after SMILE or femtosecond laser-assisted in situ keratomileusis (FS-LASIK) at 5 years, with fewer changes found after SMILE [15]. No studies have compared CD between SMILE and LASEK after a 1-year follow-up. Therefore, we used the Pentacam Scheimpflug system to evaluate CD changes in SMILE and LASEK at baseline, 3 months, and 1 year postoperatively. Our study provided a different evaluating index to compare these two types of refractive surgery and examine the relationship with clinical characteristics.

Inclusion and Exclusion
Criteria. The main inclusion criteria were (1) age ≥ 18 years with stable refraction for at least 2 years, (2) spherical equivalent ðSEÞ < −6:00 D, (3) corrected distance visual acuity of 20/20 or better, and (4) discontinuation for 2 weeks and 6 months of soft contact and orthokeratology lens use, respectively, before examination. The exclusion criteria were a history of ocular disease, surgery, or systemic diseases.
2.3. Surgical Techniques. All the surgical procedures were performed by a single surgeon (Dr. Zou) with extensive experience. The VisuMax femtosecond laser system (Carl Zeiss Meditec, Jena, Germany) settings were as follows: 500 kHz repetition rate, 150 nJ pulse energy, 120 μm intended cap thickness, and 6.7-6.8 mm optical zone. An incision with a circumferential width of 2-3 mm was made for lenticule dissection and extraction.
LASEK was performed using 20% alcohol solution to create corneal epithelial flaps. The epithelial flaps were lifted using a crescent corneal spatula. The MEL 90 excimer laser platform (Carl Zeiss Meditec AG, Jena, Germany) was used to complete corneal stromal tissue ablation at a repetition rate of 500 Hz.
Postoperatively, topical 0.1% fluorometholone was administered four times daily, tapering off once a week and once a month in the SMILE and LASEK groups, respectively. Topical levofloxacin was administered four times daily for 10 days, and topical sodium hyaluronate was administered four times daily for 3-6 months in both groups.

Corneal Densitometry
Analysis. Data on CD values were collected using the Pentacam HR and expressed in grayscale units (GSUs). The values could quantify corneal clarity by monitoring transient haze-like reactions [16]. Four annular zones of the cornea (0-2 mm, 2-6 mm, 6-10 mm, and 10-12 mm) displayed CD values [17]. According to the anatomical corneal layers based on depth, there were four corneal layers, including a superficial anterior layer of 120 μm, a posterior layer of the innermost cornea of 60 μm, central layers of subtraction of the anterior and posterior layer thickness from the total layer thickness, and a total layer [18]. The zone of peripheral 10-12 mm was excluded due to the lowest degree of repeatability and reproducibility of this zone [19].
2.6. Quality of Vision (QoV) Questionnaire. The QoV questionnaire was used to measure the subjective visual quality at 1 year postoperatively [21]. There were 10 symptoms scaled by frequency, severity, and bothersome nature, including halos, glare, hazy, starbursts, blurred, double vision, visual fluctuation, distortion, and difficulty in focusing and judging distance or depth perception. The scores for all of the above analyses ranged from 0 to 3 according to the degree. The questionnaires were completed telephonically.

Statistical Analysis.
Each eye with mild-to-moderate myopia was selected for the statistical analysis. All data were analyzed using SPSS statistical software (version 26.0, SPSS Inc., Chicago, IL, USA). Continuous variables were presented as mean ± standard deviation (SD), while categorical variables were recorded as counts. The independent t-test was used for normally distributed data, and the Mann-Whitney U test was used for nonnormally distributed data (except for sex and eye, which were compared using the chi-square test). Analysis of variance was performed in different rounds to evaluate the standardized differences in the preoperative and postoperative visits at different time points. The associations between the changes in CD values 2 BioMed Research International and the type of surgery, UDVA, age, SE, endothelial cells, MCT, and LT/AD were assessed using Spearman's correlation analysis. p was set at < 0.05.

Study Population.
Among the 49 patients, 46 eyes of 24 patients who underwent SMILE and 46 eyes of 25 patients who underwent LASEK were evaluated at 3 months and 1 year postoperatively. No significant preoperative differences were observed in most clinical characteristics between the two groups (all p > 0:05). The MCT and LT/AD were significantly different between the groups before surgery (p < 0:001) ( Table 1). All surgical procedures were uneventful, and no high intraocular pressure or any other visionthreatening complications were observed after 1 year.  Table 2). At 3 months after LASEK, the CD values at the 0-2 mm and 2-6 mm (both p < 0:001) and 6-10 mm zones (p = 0:006) of the anterior layer, 0-2 mm zone (p = 0:015) of the central layer, and 0-2 mm and 2-6 mm zone (both p < 0:001) of the total layer were reduced. There was no difference in the CD values of the three zones of the     Table 2).

Efficacy and
Comparing the changes in CD between the groups, the CD values in the LASEK group showed a more obvious reduction in the three zones of the anterior layer (all p < 0:001), the 0-2 mm zone of the central layer (p = 0:032), and the total layer (all p < 0:001) at the 3-month follow-up. Similarly, a greater reduction in CD was observed in the LASEK group than in the SMILE group in the three zones of the anterior layer (all p < 0:001), central layer (p = 0:021 for 0-2 mm, p < 0:001 for 2-6 mm, and p = 0:027 for 6-10 mm), and total layer (p < 0:001 for 0-2 mm and 2-6 mm and p = 0:002 for 6-10 mm) after 1 year. However, no significant changes in the CD values were observed in the posterior layer between the groups (p > 0:05).

Corneal Wavefront Aberrations.
Three months postoperatively, HOAs, spherical aberration, and vertical and horizontal coma of the anterior and whole corneal surfaces increased in both groups (all p < 0:05). At 1 year postoperatively, there was a significant increase in all values of anterior and whole corneal aberrations, except trefoil 30°in the SMILE group and trefoil 0°in the LASEK group (all p < 0:05). Lower HOAs, spherical aberration, and horizontal comas of the anterior and whole corneal surfaces were observed in the SMILE group than in the LASEK group at 3 months and 1 year postoperatively (Table 3, all p < 0:05, respectively).

QoV Scores.
There was no significant difference in the frequency, severity, and bothersome score between the groups at 1 year postoperatively (p = 0:72, p = 0:56, and p = 0:79, respectively) ( Figure 2 and Table 4). The three most prevalent reported visual symptoms in the SMILE and LASEK groups were glare (48% versus 50%), vision fluctuation (48% versus 50%), and halos (33% versus 35%). However, most patients considered the severity of their visual symptoms to be mild or nonexistent and responded that it bothered them a little or not at all.

Discussion
In this study, we investigated the changes in CD and visual quality after SMILE and LASEK for mild-to-moderate myopia. A greater reduction in CD was observed in the LASEK group than in the SMILE group in the three zones of all layers, except for the posterior layer, after 1 year. Our findings resonated partly with those of Lazaridis et al. who reported that the CD values recorded by the Scheimpflug showed no difference between baseline and 3 months after SMILE [13]. Rozema et al. indicated that a significant decrease was observed in the anterior zone of corneal backscatter in LASEK surgery at 6 months postoperatively [22]. Furthermore, Litwak et al. also reported good corneal clarity after LASEK [23]. However, Shajari et al. reported that there was no significant difference in CD between LASIK and SMILE at the short-or long-term follow-up [12].
Boote et al. reported that fibril packing over the corneal surface was nonuniform and fibril matrix appeared more compact in the prepupillary cornea [24]. Ni et al. reported that the corneal anterior layer showed higher densitometry than the central and posterior layers [19]. Hence, we hypothesized that the decrease in CD was possibly due to the ablation of the anterior layer. Our results were consistent with those of a previous study performed using confocal microscopy, which suggested that the increased light scatter was correlated with activated keratocytes and mainly occurred in the anterior one-third of the corneal stroma [25]. At the early stage after surgery, corneal transparency increased in response to inflammation, inducing a longer time for the recovery of visual outcomes. Topical steroid eye drops are commonly used to alleviate the inflammatory reactions. Thus, owing to the control of inflammation and ablation of stromal collagen fibrils, a decrease in CD was observed after both SMILE and LASEK 1 year postoperatively. The femtosecond laser in SMILE procedure was applied to create lenticule, accompanied with plasma, shockwave, and cavitation bubble, while the excimer laser in LASEK producer was applied to ablation, followed with the damage of the organic molecular bonds of cornea [26,27]. Considering this situation, investigating the significance of regional CD in these two surgeries was meaningful.
There was no significant difference in the efficacy or safety index between SMILE and LASEK at 1 year postoperatively, suggesting that LASEK was as safe and effective as SMILE for the treatment of mild-to-moderate myopia. Our findings partly confirmed those of Yu et al. who reported no difference in visual outcomes after SMILE and LASEK for low-to-moderate myopia at 3 months postoperatively [28]. Our previously published studies also proved the safety and stability of LASEK for the correction of mild, moderate, and high myopia [29,30]. However, SMILE became a superior choice over LASEK in terms of lesser intensity of postoperative pain. Exposure to corneal nerve endings and the release of inflammatory mediators lead to intense postoperative pain after LASEK. The pain was relieved by complete corneal reepithelialization [31]. Moreover, considering patients with thinner corneal thickness who required refractive surgery, LASEK had the advantage of less ablation depth than the lenticule thickness of SMILE for correcting the same SE. Hence, the two surgical procedures have their respective advantages and disadvantages for myopic correction, with almost consistent long-term visual outcomes.
With the development of refractive surgery, increasing attention has been paid to the postoperative visual quality. HOAs induced by corneal refractive surgeries can cause symptoms such as glare, halos, and starbursts [32]. Sekundo et al. found that 54 eyes after SMILE showed a slight increase in HOAs, coma, and spherical aberrations from 3 to 12 months [33]. Previous studies reported that the changes in corneal HOAs and spherical aberration were significantly higher after LASEK than after SMILE from 3 months to 3 years [34]. Our results were also consistent with those observed at 3 months and 1 year in previous studies [5,28]. QoV is an important indicator for assessing postoperative visual quality and patient satisfaction. Yu et al. indicated that the scores of glare and halos in the QoV questionnaire were lower in the SMILE group than in the LASEK group at the 3-month follow-up [28]. In the present study, there was a slightly lower induction rate of HOAs, spherical aberrations, and horizontal coma in the SMILE group than in the LASEK group after 1 year. Additionally, there was no significant difference in the subjective visual symptoms between the groups postoperatively. The reduction of CD values might influence the actual visual function. A previous study reported that there was no relationship between the changes of CD values with the visual function such as corneal wavefront aberrations [15]. And in our study, we found that the efficacy index, safety index, and QoV scores had no  [13]. In our study, all CD values, except those at the posterior corneal layers, were significantly associated with the type of surgery, MCT, and LT/AD. Since the lenticule thickness in SMILE procedure and the ablation depth in LASEK procedure were applied at the anterior elastic layer and stroma of cornea, the correlation of CD and LT/AD was mainly at the region of anterior layer. This might be associated with the volume of extracted or ablated stromal corneal tissues, inducing the decrease of keratocyte density, so that influenced the backscattered light from residual corneal tissue. Our study has some limitations. First, the sample size of 46 eyes in each group was relatively small, which increased the risk of a type 2 statistical error (false-negative results because of the underpowered study arms). Therefore, our results were therefore conservative in terms of the differences between the SMILE and LASEK procedures. Second, we mainly analyzed the data at 3-month and 1-year followup. Therefore, our study was not equipped to make inferences on longer-term changes in the CD values and visual outcomes between the SMILE and LASEK procedures. Third, the Scheimpflug device showed greater reflection at the interfaces between layers with different refractive indices owing to noncontact examination.
In conclusion, CD decreased in both the SMILE and LASEK groups at the 1-year follow-up; however, the reduc-tion in CD values in the SMILE group was smaller than that in the LASEK group. Nevertheless, SMILE and LASEK are equally safe, effective, and predictable refractive surgeries for correcting mild-to-moderate myopia. Long-term observations and mechanisms of CD changes after SMILE and LASEK should be employed in further research.

Data Availability
The data used to support the findings of this study are included within the article.

Ethical Approval
All procedures were performed in accordance with the principles of the Declaration of Helsinki. Observational retrospective study was approved by the Ethical Board Committee of the Shanghai Tenth People's Hospital.