Several previous clinical research reports have suggested that LIKE has a certain potential in correcting moderate to high hyperopia[5–8, 10]. Our previous study (unpublic data) and Brar et al.'s[11] study have found that LIKE is more advantageous in terms of stability after correcting hyperopia. In order to further understand the reasons, this study aims to explore the characteristics of corneal morphological changes after LIKE. In this study, the data of cases corrected by HLASIK in the same department and the same period have been reviewed as the historical control group to better reflect the characteristics after the LIKE procedure. We found that LIKE will be superior in preserving EOZ than HLASIK after correcting moderate to high hyperopia.
Previously, the method commonly used to analyze the EOZ diameter was the average figured by manual multi-radial measurement[9, 12]. This method may be limited by low measurement efficiency and poor repeatability. To address this technical problem, our team found that ImageJ, an open-source software developed by public health organizations, can efficiently analyze the corneal topography. In this way, the EOZ area can be automatically intercepted by the software's color discrimination tool, which can reduce the deviation caused by manual positioning. In addition, the software comes with the functions of graphical analysis of the best-fitting circle and positioning of the centration coordinates of the graph, which can efficiently and quickly fit the corresponding graph to obtain measurement information. The EOZ diameter parameter is helpful information that allows us to compare pupil diameter or POZ diameter intuitively. However, this parameter is just single-dimensional, so we further collected the EOZ area as a two-dimensional indicator to compare the differences in the optical zone between the two groups. The software can directly measure the actual area of the postoperative EOZ instead of the fitting outcomes, which helps us understand the area difference between the actual EOZ and POZ, thereby verifying the further comparison results. Although this semi-automated technology cannot process images in batches, its method is not limited to the researcher's programming skills[13, 14], allowing researchers to be widely used to quantitatively analyze the EOZ area or decentration after various refractive surgery.
This study found that the POZ of LIKE was significantly smaller than the settings of HLASIK in both diameter and area. Interestingly, the EOZ after LIKE was significantly larger than that after HLASIK in both diameter and area. Previous studies have reported that the diameter of corneal flap created by the VisuMax femtosecond laser platform is 0.47mm greater than that of intended POZ diameter[15]. Liu et al.[16] measured the diameter of the lenticule removed from SMILE, which was placed on the surface of the model eye (keratometry = 42D). They found that the chord length of lenticular diameter was similar to the POZ diameter (6.53 ± 0.07mm vs. 6.5mm). We also further verified the lenticular diameter derived from myopic SMILE procedure by the anterior segment optical coherence tomography (AS-OCT) (Avanti Optovue, Inc). It was measured by laying the lenticule flat on the surface of a Petri dish. A coordinate was attached to the rear surface of the Petri dish. Under the microscope observationke, the lenticular centration was moved to the origin of the coordinate (Supplementary Figure. 1st step). Then "crossline" mode was used to align the crossline from AS-OCT with the coordinate's origin, and the horizontal and vertical diameters of lenticule can be measured, and an average of 3 times measurements were counted for each lenticule (Supplementary Figure. 2nd step). This measurement is about the arc length of lenticular diameter; the chord length of lenticular diameter can be calculated from the preoperative keratometry (43.35 ± 1.22D). The result is shown in Supplementary Table 1. We found that the chord length of the lenticular diameter was slightly shorter than that of the POZ (6.44 ± 0.06mm vs. 6.5mm). This outcome may be caused by the fracture of the stromal collagen fibres when the lenticule is severed, which makes the fibre shrink. Therefore, we conclude that the diameter of the lenticule implanted in LIKE may be slightly shorter than the POZ diameter in myopic SMILE without significantly difference. The POZ diameter of extracted lenticule from the donor can be directly used as the target amount of LIKE to compare with the EOZ diameter after LIKE. The result has shown that the EOZ diameter after LIKE was 5.20 ± 0.28mm, accounting for 80% of the target. It is similar to Liu et al.'s[17] report: the EOZ diameter is 5.33 ± 0.28mm, accounting for 82% of the target.
In the HLASIK group reviewed in this study, the EOZ diameter was significantly smaller than the POZ diameter (4.61 ± 0.20mm vs. 6.56 ± 0.20mm). The postoperative EOZ diameter accounts for about 70% of the POZ diameter, and previous studies have found that the EOZ diameter after HLASIK accounts for about 65–72% of the target (supplementary table 2)[4, 9]. Roesler et al.[4] found that the higher hyperopia corrected by HLASIK, the smaller EOZ diameter achieved after surgery. The LIKE procedure was more advantageous in preserving the EOZ in correcting hyperopia than HLASIK. Previously, researchers believed that femtosecond laser is different from excimer laser in essence, and the refractive lenticule made by femtosecond laser will not be impacted by laser fluence projection and reflection, like excimer laser[9, 18]. The outcomes after LIKE corrected hyperopia (EOZ diameter: 5.20 ~ 5.33mm, accounting for 80 ~ 82% of the target) were similar to those after SMILE correcting hyperopia (EOZ diameter: 5.03mm, accounting for 79% of the target)[9]. It can be speculated that the larger EOZ diameter obtained by LIKE procedure is probably not related to the method in changing corneal curvature (additive or subtractive procedure), but may be related to the type of laser. However, the main advantage of LIKE for correcting hyperopia is as a type of corneal "additive" surgery. The procedure is relatively reversible to a certain extent, giving the opportunity to restore a relatively natural corneal shape after removing the implanted lenticule; it can still work even further excimer laser enhancement would be performed for reshaping lenticule thickness after the LIKE procedure. Besides, when cataract surgery is needed in the future, less challenge would be met in calculating the power of the intraocular lens (IOL), because the negative induction of SA can be erased to a large extent after removing the reversible lenticule, the accuracy of IOL implantation would be better than HLASIK[19].
Another concern about the LIKE may be the optical decentration induced by the manual manipulation during the procedure. In the EOZ centration scatter plot (Fig. 3) in this study, it can be seen that the distribution in centration of the EOZ after LIKE dispersed slightly. Nevertheless, there was no statistical difference between the decentration after LIKE and eye-tracking HLASIK. Moreover, there also was no statistical difference in the corneal coma changes before and after the two procedures (Supplementary table 3). After LIKE procedure, cumulatively 95% of the eyes had centration offset of optical zone within 0.5 mm. The decentration in LIKE of this study was similar to those of Liu et al.[16] (0.27 ± 0.14mm vs. 0.27 ± 0.08mm). We found that the centration offset in LIKE group did not threaten the correction's safety. However, this result may depend on the technique of the experienced surgeon. This may also be related to the large EOZ obtained after LIKE, which can tolerate some extent decentration in the optical zone[20]. Even so, we still need to explore ways to center the lenticule to improve the accuracy of EOZ effectively.
In this study, we further analyzed the changes in HOAs of the cornea with a range of 6mm diameter after the two surgeries. As previously reported, the higher hyperopia corrected by corneal refractive surgery, the more negative SA would be induced[21]. However, the induction of SA after correcting moderate to high hyperopia in the LIKE group was similar to outcomes from the HLASIK group in correcting moderate to low hyperopia (0.126µm vs. 0.140µm, P = 0.65). We further compared the induction of SA by weighing the refraction between the two groups. The change after LIKE is slightly lower than that of HLASIK (0.018µm vs. 0.040µm), but there is still no statistical difference between both (P = 0.13). We speculated that LIKE would obtain a more natural corneal shape after hyperopic correction than HLASIK procedure, because of the less induction of SA after LIKE procedure, as Ganesh et al.[22] previously reported. However, the HOAs change in this study could not be compared with other research on hyperopic refractive surgery, which may be related to different measuring equipment or calculation methods.
There are some limitations in this retrospective study, the two groups in this study could not be matched in the refractive correction, because the indication of HLASIK in correcting high hyperopia does not meet the current excimer technical requirement and ethical principle. More sample sizes of clinical trials should be considered for the LIKE procedure to better evaluate the correction effect. However, this preliminary study inspires that the LIKE procedure can be a comparable surgical option for treating moderate to high hyperopia. And we also propose a new semi-automatical analytical method for postoperative measurement in EOZ, which can be relatively efficient, reliable, and simple.