Atomic force microscopy for the evaluation of corneal surface roughness after femtosecond laser flap creation and excimer ablation

It is well known that the femtosecond laser lamellar cut induces some degree of surface roughness. Nevertheless, as in femtosecond laser-assisted LASIK (FS-LASIK), an excimer LASIK ablation is performed, and the post-ablation stromal bed should show some degree of smoothening. We decided to compare, using atomic force microscopy (AFM), the roughness of the corneal stromal bed, after a femtosecond lasers device flap was created with or without an excimer myopic ablation. Using 6 freshly enucleated porcine eyes, we created in every eye a flap using a femtosecond laser. Additionally, in 3 eyes, an excimer laser ablation to correct-3 diopters (D) was made. AFM imaging of the remaining corneal stroma was performed. Ten different square areas of 20 μm x 20 μm at the central area of the stroma of each corneal sample were studied. The roughness parameters used were the root-mean-square deviation from a perfectly flat surface. The RMS deviation was 360 ± 120 nm in femtosecond laser only, and 110 ± 20 nm in those cases where excimer is also involved (p < 0.0001). Our results show that the roughness of the surface treated with excimer is clearly lower than in the group with no excimer ablation; thus, the application of laser excimer after a flap created by femtosecond laser seems to soften the nano-irregularities created by this technique.


Introduction
The roughness of the stromal surfaces obtained following femtosecond (FS) laser-assisted keratomileusis (LASIK) has been recognized as a clinically relevant issue, because it is associated with optical aberrations and dysphotopsia phenomena ("rainbow glare") that may compromise the patient's quality of vision in the postoperative period [1].
Vol:. (1234567890) Nowadays, it is possible to assess the characteristics of corneal tissue planes by means of three different strategies: at a near-macroscopic scale using anterior segment optical coherence tomography (OCT) [2], at a microscopic scale using scanning electron microscopy (SEM) and transmission electron microscopy (TEM) [3], and in a nano-scale (i.e., with a precision in the range of few nanometers) using atomic force microscopy (AFM) [1,4,5].
Femtosecond laser devices play an increasingly important role in modern refractive surgery and are currently preferred to mechanical microkeratomes for flap creation [6].Furthermore, FS laser-assisted refractive lenticule extraction (ReLEx) has been proposed as an alternative to laser excimer ablation for the correction of myopia [7,8].However, it is well known that the corneal surfaces after a FS cut are not smooth [5][6][7][8][9].In fact, the corneal tissue plane has some degree of roughness, mostly because the tissue-cutting laser spots are separated from each other (the so called "spot distance") [5][6][7][8][9].Recent efforts have focused on the optimal FS laser parameters that would ensure a smoother interface [9].Nevertheless, because in FS-LASIK procedures an excimer ablation is performed over the corneal stroma, the postablation stromal bed might be expected to show a decrease in roughness.
The aim of the current study was to evaluate the smoothening effect of excimer ablation on corneal tissue planes of freshly enucleated porcine eyes that underwent LASIK flap creation using a FS laser.Using AFM, the stromal bed roughness of eyes that underwent FS-LASIK flap creation alone was compared to that of eyes that had a myopic excimer ablation following the FS-LASIK flap creation.

Procedure
We designed an experimental comparative study using six freshly enucleated porcine eyes.Within 24 h from the enucleation, the same experienced surgeon (JGO) created a lamellar corneal flap using the 60 kHz IntraLase ® FS laser (Abbott Medical Optics, Inc., Santa Ana, CA, USA).The following parameters were used: flap depth 110 microns, flap diameter 9.00 mm, hinge position 90 degrees (i.e., superior), hinge angle: 50 degrees, side-cut energy 0.95 mJ, side-cut angle 135 degrees.Pocket depth and width were set at 230 µm and 0.23 µm, respectively, with 5 µm tangential and radial spot separations.Immediately after the procedure, the flap was lifted and then, repositioned using a LASIK spatula.Next, myopic excimer laser ablation to correct − 3.0 diopters (D) was performed in three of the six eyes using a Wave-Light Allegretto EyeQ 400-Hz laser (Alcon Laboratories, Inc. Fort Worth, TX, USA).The laser parameters of the ablation were: optical zone: 6.5 mm, transition zone: 0.3 mm, ablation zone: 7.10 mm, maximum central ablation: 45.5 microns.When the above procedures were completed, a 9 mm trephination of the cornea was performed with a Barron vacuum donor cornea punch (Katena Inc, Denville, NJ, USA) in all eyes.Immediately after the trephination, the flap was lifted and amputated with scissors, and the specimens were prepared for AFM analysis of the stromal bed roughness.The endothelial side of the corneal button obtained was glued with cyanoacrylate adhesive to a microscope slide and covered with 2.5% glutaraldehyde solution for 24 h before AFM measurements of the stromal bed roughness were performed.
Ten randomly selected areas from each sample were studied; thus, a total of 30 regions from each group were analyzed.

Imaging
AFM imaging was performed using a JPK NanoWizard II ® atomic force microscope coupled to a Nikon Eclipse Ti-U inverted optical microscope, in contact mode immersed in liquid, and employing Olympus OMCL-RC800PSA commercial silicon nitride cantilever tips (0.05 N/m, 18 kHz), with typical 15 nm radius at the end.Vertical accuracy of the instrument is in the order of 0.1 nm.
Surface roughness measurements were made in 10 different square areas with dimensions 20 μm x 20 μm at the central part of each corneal sample, using a 512 × 512 point resolution.The images were processed and analyzed using the JPK Data Processing software.The roughness parameter used in this study was the root-mean-square (RMS) deviation from a perfectly flat surface within the analyzed area.The RMS is used to measure the differences between values predicted by a model or an estimator, and the real values observed.It is always non-negative, and a Vol.: (0123456789) value of 0 would indicate a perfect fit to the data.In general, a lower RMS is better than a higher one.

Results
The RMS (in microns) from AFM images of every sample analyzed was calculated.The results are depicted in Table 1.
For the whole groups, the mean ± standard deviation of the RMS values was 360 ± 120 nm for the FS only group and 110 ± 20 nm for the FS and excimer group (P value = 0.0001), (Fig. 1).
Using atomic force microscopy, we obtained threedimensional and vertical deflection images of the stromal surface of the FS-LASIK flap creation without excimer ablation group (Figs. 2, 3) and of the FS-LASIK flap creation and excimer laser ablation group (Figs. 4, 5).

Discussion
Our results show that excimer laser ablation following flap creation with an FS-LASIK does smoothen the nano-scale irregularities of the corneal stroma surface induced by the FS photodisruption.This observation is clinically relevant, because the optical effects induced by a rough corneal surface after FS-LASIK surgery may adversely affect the quality of vision in the postoperative period [5][6][7][8][9].
Following its validation for ophthalmic purposes, AFM has become a method of choice for studying the roughness of various surfaces [10].In fact, AFM is considered a more attractive option than SEM because it allows the acquisition of three-dimensional images of the surface of samples and offers the possibility of quantitative and objective analysis.An additional advantage of AFM over SEM is that the preparation of specimens for the former is simpler than for the latter.For these reasons, AFM is currently considered the gold standard for the study of stromal surface roughness [5][6][7][8][9][10][11].
There are few reports on the use of AFM for the study of the stromal roughness after excimer laser ablation [12,13].However, none of them analyzed the stromal roughness before and after the ablation.Instead, these experimental studies only examined the effect of the excimer laser ablation using  different parameters for depth and energy following the mechanical removal of the epithelium [12,13].Lombardo M et al. were the first to study corneal roughness using AFM after a standard ablation using two different excimer lasers [14], as well as the effect of a smoothening technique with viscoelastic application over the corneal stroma as a masking agent in photorefractive keratectomy [4].
In a later report, the same authors published the results of AFM analysis of the corneal epithelium and the photoablated corneal stroma [1], but did not compare the roughness of the corneal stromal bed before and after the ablation.Few investigators have used AFM to examine the effect of FS lasers on corneal surface roughness.Serrao et al. studied the effect of different parameters on corneal surface roughness using the IntraLase platform for LASIK [9][10][11], as well as the surface roughness and morphology of the posterior aspect of stromal lenticules created with a femtosecond laser (ReLEx) [15].In all these studies, the amount of roughness is quite similar to the amount of roughness found in the current study (around 350 nm).Ziebarth et al. studied the quality of corneal lamellar cuts with the VisuMax platform for FS-LASIK [5].However, none of the investigators cited performed an excimer ablation after the FS laser procedure to study the excimer-induced changes on corneal surface roughness.On the other hand, although they did not perform the excimer ablation over a previously FS treated corneal surface, Serrao et al. found smoother corneal surfaces after excimer ablation 1 than after Femtosecond treatment [9] in porcine corneas, which agrees well with the results of our study.
Furthermore, the much higher standard deviation we found in the roughness values before the excimer ablation than afterward (120 vs. 20 nm, respectively), further support the fact that corneal surfaces after excimer ablation are both smoother and more homogeneous than after FS treatment.
To the best of our knowledge, the current study is the first to use AFM for the assessment of the surfacesmoothening effect induced by excimer ablation on a corneal surface previously treated by a FS laser.The fact that surface roughness decreases after the excimer ablation is clinically relevant because it indicates that in real-life surgeries, at least one of the surfaces created with after the FS laser cut is smoothened.Thus, it makes sense that the smoothening of the anterior surface of the stromal bed could translate to better quality of vision in the postoperative period.
Based on our results, we believe that in those refractive techniques in which only a FS laser is used (i.e., ReLEx procedures), as the smoothening effect of excimer ablation is not present, the visual results might be worse than in FS-LASIK, because in that procedures both surfaces are created by the FS laser [16].In fact, there seems to be some differences between FS-LASIK and ReLEx in corrected distance visual acuity 3 months after surgery [17].This study clearly showed a lower gain of Snellen lines in eyes treated with ReLEx compared to those treated with LASIK, with a statistically significant difference in the loss of 1 Snellen line and in the proportion of eyes achieving so called "supervision."Nonetheless, further studies with sufficiently long follow-up are needed to address these facts.
A possible limitation of the present study is that the animal eyes that we used were operated within approximately 24 h after enucleation.Consequently, post-mortem tissue changes may have interfered with the overall quality of the FS cut.In addition, only small areas (20 × 20 μm 2 ) of the corneal surface were studied with AFM, thus, morphologic features larger than the single area scanned do not contribute to the analysis.
The results of our study are not necessarily applicable to other commercially available femtosecond laser platforms, or even to the same platform with different settings.Nonetheless, our results clearly indicate the beneficial effect that excimer laser photoablation has on the surface roughness characteristics of tissue planes created with a FS laser.

Conclusion
Our results show that the roughness of the surface treated with excimer is clearly lower than in the group with no excimer ablation; thus, the application of laser excimer after a flap creation by femtosecond laser does soften the nano-irregularities created by this technique.

Fig. 1
Fig. 1 Mean ± standard deviation of the root-mean-square (RMS) values for the FS only and the FS and excimer group

Fig. 2 Fig. 3 Fig. 4
Fig. 2 Image of 3D atomic force microscopy (AFM) topography of the stromal surface after FS-LASIK flap creation without excimer ablation

Fig. 5
Fig. 5 Image of the AFM vertical deflection of the corneal stromal surface after FS-LASIK flap creation followed by excimer laser (3.0 D myopic ablation)

Table 1
Results