Second-generation keratorefractive lenticule extraction: will there be a third?

It has been over 20 years since the idea of an all-in-one femtosecond laser refractive procedure was first theorized. However, early results were of insufficient quality in their scientific and refractive evaluation to further continue this work in a methodological approach. After extensive research by Carl Zeiss Meditec AG, a series of preclinical studies on porcine eyes, rabbit eyes, and in amblyopic patients with poor vision were conducted by Sekundo et al.1 The procedure performed was called femtosecond lenticule extraction (FLEx), in which a refractive lenticule was created, under a LASIK (laser in situ keratomileusis)-style flap where the interface creating the anterior surface of the lenticule simultaneously created the posterior surface of the flap. The flap was then lifted and the lenticule peeled from the stromal surface.1 Following the successful phase 2 studies, a series of eyes, with good visual potential, were treated.2 In the initial series, there was variation in the flap thickness, with no nomogram available, and the geometry of the lenticule was calculated mathematically. Despite this, 98.1% of eyes were within 1 diopter (D) of the refractive target, which far exceeded expectations.2 After this initial success, Shah et al. were instrumental in developing the procedure further.3 By gaining a vast amount of clinical experience with this technique, they were able to optimize the laser-firing sequence to improve the early visual acuity results. The creation of a lenticule by a femtosecond laser, using an Archimedes spiral pattern, can be done in 4 ways. Shah et al. showed that there was delayed visual recovery in patients who underwent the initial laser-firing sequence and theorized this was due to irregularities in the interface. Riau et al., in a series of experiments on just 8 rabbits, proved this theory and showed that there was collagen derangement in the central area of the lenticule, due to disruption of the collagen fibrils.4 By altering the firing sequence, this derangement could be improved. This was a pivotal time for refractive lenticule procedures and opened the way for the widespread dissemination of FLEx and further refinement in the form of small-incision lenticule extraction (SMILE). The first SMILE procedure was performed with 2 opposite side-cut incisions to extract the lenticule, by Sekundo et al., and the encouraging results led the way for further refinement of the procedure, performed with a single incision.5 Although the early results were good, there was a delay in visual recovery and a large variation in reported outcomes. Further optimization of energy parameters, spot spacing, a new 500 kHz laser, and new instrumentations have all helped to improve this early visual recovery. Over the years, SMILE has gained immense popularity among the refractive community, and it is incredible that last year Carl Zeiss Meditec celebrated the 10 millionth procedure performed. This is a remarkable achievement when you consider the impressive outcomes of modern-day LASIK. However, although the results have been good, there is a wide variation of results reported in the literature. This is partially due to the limitations of this first-generation procedure. Instead of examining individual surgeon series, which are often biased, one may look at meta-analyses as the highest level of evidence. However, we must also be cautious to the effect of large sample sizes from a few centers and countries, skewing the results and conclusions.6 Lessons on the use of antiresorptive agents for patients at risk of bone fractures had a detrimental effect on the conclusion of many meta-analyses. Despite this, if conducted properly, they can provide an immense amount of information. Focusing on those that have been published in the past few years to minimize bias from earlier publications, we can learn of certain trends. Tian et al. compared the clinical outcomes and higher-order aberrations (HOAs) of wavefront-guided LASIK (WFG-LASIK) vs SMILE.7 They analyzed 12 studies and concluded that there was no significant difference in clinical outcomes and induced HOAs between both SMILE and WFG-LASIK. Both procedures were found to be safe, effective, and predictable, but SMILE was found to induce more vertical coma and trefoil aberrations, whereas WFG-LASIK induced more spherical aberrations. Similarly, Yao et al. looked at the outcomes of SMILE and WFG-LASIK in myopic patients and patients with myopic astigmatism.8 Focusing only on randomized clinical trials, one can argue that this is a study of a very high level of evidence. Of the 11 studies, overall there were no significant differences between the 2 groups with regard to the final mean refractive spherical equivalent, safety, efficacy, postoperative mean logMAR uncorrected distance visual acuity (UDVA), postoperative astigmatism, and postoperative HOA. However, the spherical aberration was significantly less in the SMILE group, and the proportion of eyes within ± 0.5 D was significantly higher in the WFG-LASIK group, with P values of 0.04. Focusing on astigmatic corrections, Song et al. published a review of 17 studies. Similar to the above, there was no difference in visual acuity, refraction, or total HOAs between the 2 groups.9 Again, spherical aberration was found to be less in the SMILE group. With respect to astigmatism alone, the only significant difference was found in the correction index in favor of WFG-LASIK. The correction index was more apparent in the low-to-moderate astigmatism correction where SMILE exhibited a smaller correction index and a larger difference vector.9 In high myopia cases, Fu et al. found 12 relevant studies. There were no significant differences in the postoperative UDVA and postoperative mean refractive spherical equivalent. However, the SMILE group had better postoperative corrected distance visual acuity. Similar to the findings from other meta-analyses, the postoperative spherical aberration was lower in the SMILE group; in addition, the postoperative total HOAs were also lower.10 An interpretation of the refractive result variability can be attributed to variation in surgeon-controlled centration and torsion control. It is amazing that the results are so close compared with LASIK, considering we are relying on surgeon control for alignment. The importance of eye trackers during excimer laser ablation has been recognized for over 20 years. However, with SMILE, the eye position is fixed by the suction device during the lenticule creation. Individual surgeons have shown that they can achieve centration results equivalent to modern-day excimer lasers, using manual marking on the cornea.11 However, larger series have shown the link between induction of coma in SMILE and greater decentration.12 With the introduction of a second-generation lenticule extraction procedure, this was clearly an issue that can be improved upon. The success of any surgical procedure will always be dependent on reproducibility of outcomes among multiple surgeons of varying standards. Other requests for a second-generation procedure in addition to centration and cyclotorsional control would be the ability to perform hyperopic treatment, that is, an extended treatment range, as well as the ability to reduce the energy further to improve early visual outcomes, reducing the risk of suction loss during lenticule creation and the possibility of topography-guided treatment. In the past 4 years, industry has taken note and responded to the success of SMILE, with several lenticule procedures now available from different manufacturers. These include SMILE Pro (VisuMax 800, Carl Zeiss Meditec AG), SmartSight (ATOS, Schwind eye-tech-solutions GmbH & Co. KG), corneal lenticular extraction for advanced refractive correction (CLEAR) (Z8, Ziemer Ophthalmic Systems AG), and smooth incision lenticular keratomileusis (SILK) (ELITA, Johnson & Johnson Vision). From a user perspective, competition does breed innovation and faster development, but it may also lead to many false claims that are not scientifically based. It is the role of researchers and clinicians to affirm or dismiss these beliefs and not just take them for granted. There are differences between these lenticule procedures, and it is important for clinicians to know and understand these differences. However, from a patient's perspective, it can still be confusing. Similarly, there are differences between wavefront-optimized, topography-guided, and WFG-LASIK, but patients know they are all essentially LASIK procedures. In view of this dilemma, our professional community, in a joint editorial statement, has decided to use the terminology of keratorefractive lenticule extraction (KLEx) as the unifying name for all these procedures.13 This can be seen to be discouraging from an industry point of view, but from a patient point of view, it becomes easier to understand. Our role as refractive practitioners, including industry, and doctors is to grow the refractive market so that many more people can benefit from this quality-of-life changing technology. Of the list of requests above, for a second-generation procedure, some have already been met with these new machines. Reported suction loss rates are lower on all second-generation machines, either by having a much faster procedure (eg, SMILE Pro) or by having a higher machine-controlled suction (eg, SmartSight, SILK, CLEAR). The key point above, about the refractive outcomes being dependent on centration and torsional control, has also been partly met. Currently, all second-generation machines allow the lenticule to be adjusted after application of suction (eg, CLEAR, SILK), hence taking away any patient dependency, or provide a guidance system to help the surgeon center the lenticule more accurately (eg, SMILE Pro, SmartSight). Certainly, early published results seem to show these second-generation procedures to be superior to first-generation procedures, but there are still few publications in the literature because of time for regulatory approval of these technologies. However, currently, the centration and torsion is semiautomated on 3 of the devices, but full automated centration and torsion is coming soon on all systems and is already available on the ATOS. With hyperopic correction soon to be released on the VISUMAX 800 (Carl Zeiss Meditec AG), the extended range is now present, with the other companies following suit in their pipeline development. The question remains if topography-guided KLEx is possible for asymmetric corneas to further extend the treatment range. First, the laser and diagnostic device will have to be linked, which is currently in development. Ideally, the laser profile would have to be applied to the cornea in a situation where the cornea is not under any additional stress. All the current laser systems distort the cornea, either weakly using a curved interface or strongly using an applanation scheme. It is possible to do this without adding stress, which has already been proven in preclinical studies.14 By using a liquid system, one could truly apply a femtosecond corneal incision without any corneal distortion. The question remains whether the laser can be applied in an accurate profile to change its symmetry during photo disruption. This would require software enhancement to achieve. The laser would need to scan every pulse laid down to change its depth as and when needed. In addition, the arc length of both the anterior and posterior surfaces of the lenticule would need to be equal to allow the lamellae above and below to be similarly stressed for transmission of the customized shape to the anterior surface of the cornea. This would also be necessary for long-term stability, especially with changes in intraocular pressure. Another option would be to try to achieve this with even more precision with an attosecond pulse. The latter requires more research and investment. It is likely that current second-generation systems, when they reach their full potential, with full linkage to diagnostic devices, will be adequate for most KLEx surgical cases. However, if we want to push the boundaries in reducing energy and achieving topography-guided corrections, a third generation will be required.