How to Improve the Refractive Predictability of SMILE

Fig. 15.1
Plot of the spherical equivalent prediction error (D) by the difference in central corneal thickness between the expected (VisuMax read-out lenticule thickness) and the observed change measured by optical low-coherence tomography

The mechanism of femtosecond laser cutting may affect the refractive precision of the SMILE procedure [2]. At the laser focus point, the laser energy increases above a critical level, and plasma formation and cavitation take place. The actual point of focus is associated with some degree of imprecision that is determined by lateral and axial beam scanning imprecision. However, the axial point of interaction between laser pulse and corneal tissue may fluctuate due to non-linear effects within a certain range that is approximately given by the Rayleigh length (zR) (Fig. 15.2).


Fig. 15.2
The actual point of focus of the femtosecond laser is associated with some degree of imprecision and is characterised by the Rayleigh length (z R). The Rayleigh length is dependent on the wavelength (λ) and the beam waist (w 0). (Courtesy: Dirk Mühlhoff, Carl Zeiss Meditec AG, Jena, Germany)

The VisuMax® femtosecond laser operates with a wavelength of 1,043 nm (λ) and a beam waist (w0) of 1 μm. The beam waist is dependent on a number of optical factors related to the femtosecond laser system. From the known parameters, the Rayleigh length can be calculated to be about 3 μm. This shows that the axial position of individual points of interaction is stable within a range of about 3 μm. The VisuMax technology fulfils the applicative requirements if it is taken into account that the cut surface is defined by millions of adjacent points of interactions, and the cap will function as a sort of low-pass filter to level out variations in the cut surfaces.

Sometimes the Rayleigh length is discussed in relation to the central profile thickness for a 1-diopter correction that is approximately 13 μm for a 6-mm diameter optical zone. However, from the previous considerations, this is misleading, because the relevant parameter to observe is the overall change in corneal curvature and not the central profile thickness. Further optimisation of the femtosecond laser with respect to wavelength and beam waist is discussed today. If the wavelength could be reduced to one-third and/or the beam waist could be lowered further, the Rayleigh length would be reduced. However, it is questionable whether such technological modifications would actually be useful and are at all possible.

15.5 Biological Variations

15.5.1 Epithelium

From basic studies of excimer laser refractive procedures, it is known that compensatory wound healing takes place after surgery. In photorefractive surgery, stromal remodelling and new synthesis of collagen, which is laid down on the bared stroma, can affect the refractive reproducibility. After PRK and LASIK for myopia, considerable thickening of the central epithelium takes place, and this will also reduce the net refractive effect of the procedures. Up to a 10 μm increase in central epithelium has been documented by confocal microscopy over time after correction for approximately 7 diopters of myopia [3]. This is acknowledged as a result of the normal beneficial smoothing effect of the epithelium to reduce abrupt changes in the stromal surface shape. After SMILE, central epithelial thickening have also been documented by optical coherence tomography, but the thickening amounted to only 6 μm on average with considerable variation between eyes (range −2 to 13 μm) [12]. Using very high-frequency ultrasound, epithelial thickening of 15 μm has, however, been observed after SMILE [8]. A 6-μm increase in epithelial thickness corresponds to a regression of approximately 0.4 dioptres. Some of the variation in epithelial thickness changes is caused by measurement inaccuracy. However, individual differences between patients and eyes concerning epithelial hyperplasia undoubtedly are a major contributor to variations in refractive predictability after SMILE. Compared with LASIK, the effective optical zone is larger after SMILE, and this possibly explains that compensatory epithelial thickening is less after SMILE.

15.5.2 Biomechanics

In SMILE, the majority of the anterior lamellae are preserved compared to the anatomical situations after LASIK, and as such, the SMILE-operated cornea is more biomechanically intact compared with the post-LASIK cornea. Mathematical simulations have confirmed this ([5], Chap. 13). The effect of individual variations in corneal biomechanics and susceptibility to development of corneal ectasia, in theory, should be less in SMILE compared with LASIK-treated eyes. Today, corneal ectasia development after SMILE has not been published although more than 100,000 procedures have been performed worldwide. Overall, it is fair to expect that individual variations in corneal biomechanical properties are less prone to affect refractive predictability after SMILE compared with LASIK.

15.5.3 Empirical Studies of Factors Affecting the Refractive Predictability of SMILE

In 2012, we published a paper on predictors for the outcome of SMILE [1]. Three hundred thirty-five patients with myopia up to 10 diopters (spherical equivalent refraction) and astigmatism up to 2 diopters were treated with small-incision lenticule extraction in both eyes (670 eyes) and followed for 3 months. The preoperative spherical equivalent averaged −7.19 D ± 1.30 D. In eyes with emmetropia as target refraction, 84.0 % obtained an uncorrected distance visual acuity ≤0.10 (logMAR) at 3 months. Mean corrected distance visual acuity improved from −0.03 to −0.05 (logMAR) (p < 0.01). 2.4 % (16 eyes) lost two or more lines of corrected distance visual acuity. The achieved refraction was 0.25 ± 0.44 D less than attempted after 3 months, and 80.1 % (537 eyes) and 94.2 % (631 eyes) were within ±0.5 and ±1.0 diopters of attempted, respectively. Multiple linear regression analyses revealed that spherical equivalent undercorrection was predicted by increasing patient age (0.1 D per decade; p < 0.01) and steeper corneal curvature (0.04 D per D; p < 0.01). The safety and efficacy of the procedure were minimally affected by age, gender, and simultaneous cylinder correction. In conclusion, the findings of an undercorrection of 0.25 diopters and small effects of patient age and corneal curvature suggest that the standard nomogram for SMILE needs only minor adjustments. However, this conclusion is related to a standard cup thickness (up to 130 μm) only. As stated in Chap. 12, thicker caps (or deeper lenticules) might need some adjustments in refractive data entered into the laser.

Here, we now extend the study to include more than 1,500 eyes treated for myopia with SMILE at Department of Ophthalmology, Aarhus University Hospital. The cohort of patients is the same as described by Ivarsen et al. [4]. After a thorough preoperative evaluation by trained optometrists, patients were thoroughly informed about keratorefractive surgery, including known side effects and complications.

From January 2011 to March 2013, 1,800 eyes of 922 consecutive patients were operated with SMILE at the Department of Ophthalmology, Aarhus University Hospital, Denmark. Patients were seen the day after surgery and again after 3 months; however, only 808 patients (1,574 eyes) attended the 3-month follow-up visit. The average preoperative patient characteristics are given in Table 15.1.

Table 15.1
Pre-operative demographics of 1,800 eyes treated with SMILE for myopia

All eyes (n = 1,800)

Gender (male/female)


Age (yrs)

38 ± 8 (19–59)

Sphere (D)

−6.79 ± 1.99 (−14.25 to +1.75)

Cylinder (D)

−0.93 ± 0.90 (−5.75 to 0.00)

Spherical equivalent refraction (D)

−7.25 ± 1.84 (−14.50 to −0.25)

Keratometry (D)

43.27 ± 1.47 (38.93–48.35)

Central corneal thickness (µm)

535 ± 27.7 (473–634)

Intraocular pressure (mmHg)

15.2 ± 2.8 (7–24)

Values are given as mean ± standard deviation, with range in parentheses

Patients underwent SMILE treatment by the VisuMax® femtosecond laser. At the 12-o’clock position, a 30–60° incision was created for lenticule extraction. Lenticule diameter ranged from 6.0 to 7.0 mm, and the lenticule side cut was 15 μm in all cases. Cap diameter varied from 7.3 to 7.7 mm, and the intended cap thickness was 110–130 μm. All cases were operated with one of two different laser energy settings. Setting 1 used a laser cut energy index of 25–27 (~130 nJ) and a spot spacing of 2.5–3.0 μm. In setting 2, a laser cut energy index of 34 (~170 nJ) and a spot spacing of 4.5 μm were used. Laser setting 1 was used for surgery in 656 eyes, and setting 2 was used in 1,144 procedures. After the laser treatment, a blunt spatula was used to break any remaining tissue bridges, allowing the lenticule to be removed with a pair of forceps. After removal of the lenticule, the stromal pocket was flushed with saline, and patients received 1 drop of chloramphenicol and 1 drop of diclofenac.

Only gold members can continue reading. Log In or Register to continue

May 26, 2017 | Posted by in OPHTHALMOLOGY | Comments Off on How to Improve the Refractive Predictability of SMILE
Premium Wordpress Themes by UFO Themes