- Top of page
Purpose: To compare outcomes in wavefront–guided LASIK performed with iris recognition software versus without iris recognition software in different eyes of the same patient.
Methods: A randomised, prospective study of 104 myopic eyes of 52 patients undergoing LASIK surgery with the MEL80 excimer laser system was performed. Iris recognition software was used in one eye of each patient (study group) and not used in the other eye (control group). Higher order aberrations (HOAs), contrast sensitivity, uncorrected vision (UCV), visual acuity (VA) and corneal topography were measured and recorded pre-operatively and at one month and three months post-operatively for each eye.
Results: The mean post-operative sphere and cylinder between groups was similar, however the post-operative angles of error (AE) by refraction were significantly smaller in the study group compared to the control group both in arithmetic and absolute means (p = 0.03, p = 0.01). The mean logMAR UCV was significantly better in the study group than in the control group at one month (p = 0.01). The mean logMAR VA was significantly better in the study group than in control group at both one and three months (p = 0.01, p = 0.03). In addition, mean trefoil, total third-order aberration, total fourth-order aberration and the total scotopic root-mean-square (RMS) HOAs were significantly less in the study group than those in the control group at the third (p = 0.01, p = 0.05, p = 0.04, p = 0.02). By three months, the contrast sensitivity had recovered in both groups but the study group performed better at 2.6, 4.2 and 6.6 cpd (cycles per degree) than the control group (p = 0.01, p < 0.01, p = 0.01).
Conclusions: LASIK performed with iris recognition results in better VA, lower mean higher-order aberrations, lower refractive post-operative angles of error and better contrast sensitivity at three months post-operatively than LASIK performed without iris recognition.
Over the past two decades, LASIK surgery has become the most common surgery performed for the treatment of refractive error.1 Well over one million procedures are now performed worldwide every year.2 However, despite improvements in excimer laser equipment, software and surgical techniques, patients still occasionally complain of blurred vision and symptoms of glare after surgery.3,4 Wavefront-guided LASIK (wg-LASIK) was subsequently developed in an attempt to both decrease and minimise induction of higher-order aberrations (HOAs) after LASIK that could result in these symptoms. Despite these advances, patients still have post-operative complaints that are likely to be related to a multitude of factors that may affect the photoablation.5–9 One of these factors is cyclorotation during the procedure.10
It is well known that cyclorotation of the eye may occur when the patient moves from the sitting position during aberrometry to the supine position during laser surgery. Such cyclorotation may result in the induction of lower- or higher-order aberrations due to misalignment of the ablation.10–14 Several studies have reported a mean degree of cyclorotation ranging from two to 4.4 degrees and in some individuals the deviation was as high as six to 10 degrees.10,11,15 Unfortunately, pupil-based eye tracking systems do not detect the misalignment caused by rotational movements. To solve this problem, cyclotorsion recognition software was introduced into refractive surgery, which can identify the rotational position of anatomical features of the eye including iris structures and scleral blood vessels. In this way, iris recognition-guided LASIK surgery may give a more precise ablation than previous technologies by accounting for cyclorotation of the eye during surgery compared to during the measurement.
The current study was performed to compare visual, refractive and contrast sensitivity outcomes as well as induction of HOAs after LASIK in eyes using iris recognition software compared to outcomes in the contralateral eye of the same patient having surgery without this technology.
- Top of page
A randomised, prospective study of 122 myopic eyes of 61 patients undergoing LASIK surgery with the MEL80 excimer laser system was performed. Iris recognition software was used in one eye of each patient (study group) and not used in the other eye (control group). The eye that underwent surgery with the iris recognition software was chosen randomly for each patient. Sixty-one subjects were selected from patients with myopia and/or myopic astigmatism who presented for LASIK surgery at the eye centre of the second hospital affiliated to Zhejiang University, Hangzhou, China from 4 September 2007 to 14 December 2007. In 15 per cent of patients (nine of 61), iris recognition could not be achieved and these patients were excluded from the study. Among the remaining 52 patients, 28 of the study patients were male and 24 were female and ranged in age from 18 to 35 years (mean age 24.6 ± 4.78 years).
Routine pre-operative LASIK examinations including manifest and cycloplegic refraction with visual acuity (VA), uncorrected vision (UCV), slitlamp examination, dilated retinal examination, corneal topography, intraocular pressure, contrast sensitivity and aberrometry were performed for each patient. Inclusion criteria were a spherical equivalent of manifest refraction less than -8.00 dioptres, manifest astigmatism -3.00 D or less, VA of 6/6 or better and no history of ocular abnormalities or previous ocular surgery. Eligible patients were scheduled for bilateral wg-LASIK. The study was conducted in accordance with the tenets of the Declaration of Helsinki. Informed consent was obtained from all subjects prior to surgery.
Iris recognition was obtained pre-operatively by using the data recording system in the WASCA analyser (Carl Zeiss Meditec, Germany) that captures an iris image under both photopic and scotopic conditions. Software in the system automatically superimposes and compares the images to verify that the scotopic and photopic iris images match. This system then detects intra-operative cyclorotation ranging from -15° to 15° using the previously obtained iris images.
All surgical procedures were performed by one surgeon (YY). The LASIK procedure was performed in each eye by first creating a superiorly hinged corneal flap with the Moria-2 microkeratome (Moria, Antomy, France) with the 110 µm head. The laser ablation was then performed on an undilated pupil using the MEL80 laser system (Jena, Germany) with an ablation zone of 6.0 mm and a transition zone of 7.0 to 9.0 mm with a goal of emmetropia. As described above, the iris recognition software is an automated program in the WASCA analyser that detects the magnitude of cyclorotation by comparing the iris pattern in the sitting position during aberrometry to the iris pattern in the supine position just prior to the laser ablation and then rotationally adjusting the ablation appropriately. If iris recognition could not be achieved during surgery, the patient was excluded from the study. The active eye tracker was used to centre the ablation on the pupil centre. Patients with any intra-operative complications such as flap irregularities or post-operative complications such as recurrent epithelial defects in either eye were also excluded from the study.
The post-operative regimen included artificial tears and 0.1% fluorometholone eye-drops (Flumetholon, Japan) four times a day for 14 days.
The higher-order aberrations, contrast sensitivity, uncorrected vision (UCV), manifest refraction with VA and corneal topography were all measured at one and three months post-operatively for each eye.
The HOAs at 6.0 mm were measured under scotopic conditions using the WASCA analyser pre-operatively and at one and three months post-operatively. The wavefront aberration coefficients were determined up to the fourth order and are expressed in standard ophthalmic notation.16
Contrast sensitivity measurements were taken using the CGT-1000 analyser (Takagi Seiko, Nagano, Japan), previously described by Pérez-Carrasco and colleagues17 and Puell and associates.18 All measurements were taken at a distance of 35 cm. Patients had the appropriate refractive correction for the viewing distance while the contrast sensitivity was measured. The targets were rings of variable size with visual angles of 6.3, 4.0, 2.5, 1.6, 1.0 and 0.7 degrees. The details of the targets for calculating the visual angles and cycles per degree (cpd) are the widths of the dark rings, which are 2.9, 1.8, 1.2, 0.7, 0.5 and 0.3 mm, respectively, with corresponding cycles/degree of 1.0, 1.7, 2.6, 4.2, 6.6 and 10.4 cpd, respectively. The settings used in our study included a target viewing time of 0.2 seconds, a time between targets of two seconds and a background luminance of 10 cd/m2.
Ablation centration analysis
For each eye, we calculated the ablation decentration from the entrance pupil, as described previously.19,20 A power difference map was generated from the pre-operative and three-month post-operative topographic maps (Orbscan IIz, Bausch & Lomb, Feldkirchen, Germany) to estimate the centre of the ablation. The ablation decentration was calculated as the distance from the centre of ablation to the centre of the entrance pupil. The apparent centre of the ablation zone was indicated with a transparent grid on the computer screen. The reading of the cursor position (λ, θ) corresponding to the centre of ablation was recorded, representing the distance from the corneal vertex to the ablation centre. Next, measurements were made on the original computer image from each pre-operative examination to identify the pupil centre relative to the corneal vertex. From these data, the decentration of the ablation centre relative to the entrance pupil was obtained mathematically.
Comparisons between the study and the control group as well as comparisons between pre-operative and post-operative values within each group were performed using a paired t test (SPSS 12.0). A p value equal to or less than 0.05 was considered statistically significant. UCV and VA data were converted into logMAR units prior to analysis.
- Top of page
A pupil-based eye tracker can track the pupil in the x- and y-axes, yet it cannot detect cyclotorsion. The WASCA analyser captures images of the iris at the same time as it performs the wavefront measurement, allowing for the wavefront treatment to be precisely localised rotationally in relation to iris structures. This procedure is known as iris recognition.
The greatest benefit of iris recognition lies in the potential for reducing the effect of cyclorotation of the eye occurring when the patient changes from the sitting position during the aberrometry to the supine position during laser surgery. When ocular rotation is not compensated for by iris recognition, the laser ablation is rotated incorrectly, which theoretically may result in the induction of lower- and higher-order aberrations and decreased vision. On average, patients rotate their eye anywhere from two to 4.4 degrees when going from sitting to supine and in some individuals, the deviation can be as great as six to 10 degrees.10,11 Similarly, in our study, the mean cyclorotation was 3.8 degrees with the largest cyclorotation being 11.2 degrees. Theoretically, as the rotation of the eye can be detected and compensated for by iris recognition software, cylindrical refractive errors and higher-order aberrations should be corrected more precisely than in treatments performed without such software, resulting in better vision.
Visual acuity, refraction and AE
We found better mean logMAR UCV at one month and better mean logMAR VA at both one and three months in the study group compared to the control group. This finding implies that iris recognition software improves vision after LASIK compared to systems not employing this software. Such an improved outcome might be expected, given that the cylinder and HOA ablations are more precise when cyclorotation is taken into account.
Although the mean residual cylinder was slightly smaller in the study than the control group at three months (0.11 D versus 0.15 D) as might be expected, this difference was not statistically significant. We did find that the refraction AE was significantly smaller in the study group compared to the control group both in arithmetic and absolute means (p = 0.03, p = 0.01). Our results differ from that of Ghosh and co-workers,12 who found significantly less residual cylinder when iris recognition was used compared to when it was not. These different results are likely to be due to the differences in the mean pre-operative cylinder between the two studies. The mean pre-operative cylinders in our study are relatively small, (0.57 ± 0.41 D in the study group and 0.55 ± 0.48 D in the control group) compared to the pre-operative cylinders in Ghosh and co-workers' study12 (1.22 ± 0.93 D in the iris recognition group and 1.34 ± 1.24 D in the control group). Because of our lower mean pre-operative cylinder, it would be expected that we would be less likely to find a statistically significant difference in post-operative cylinder between groups than in Ghosh and co-workers' study.12 Given that the differences in post-operative mean residual astigmatism between iris recognition and control groups were statistically significant in the study that had higher mean pre-operative cylinder, this might imply that highly astigmatic patients might gain more benefit from iris recognition than patients with lower astigmatism. Further research is being performed by our group to confirm this hypothesis.
Our current study also showed that the study group tends to be slightly overcorrected in the mean post-operative sphere compared to the control group at three months (+0.19 versus +0.17 D). This tendency for overcorrection may have been caused by the extra time it took to establish iris recognition in the study group compared to the control group, which caused the stromal bed to dry and increase the laser ablation rate.24 This problem is likely to diminish as the surgeon becomes more familiar with the iris recognition procedure.
Our study showed that HOAs increased after LASIK regardless of whether iris recognition was used. At one month post-operatively, there was no significant difference between the two groups with respect to these increased HOAs but by three months post-operatively, aberrations in both groups had decreased slightly compared to one month for most Zernike terms and total RMS. This phenomenon is similar to the results of Oshika and colleagues,25 who showed that aberrations induced by LASIK decreased continuously over 18 months, with most of the reduction occurring in the first two months. After this initial 18-month period, the observed changes in aberration were minimal. This observed improvement in HOAs after LASIK is likely to be due to wound-healing in the early post-operative period.
While HOAs were similar between groups at one month, our study showed that three months post-operatively, the study group had a significantly lower level of trefoil, total third-order aberration, total fourth-order aberration and total RMS of HOAs than the control group, suggesting that the iris recognition software provides a more precise ablation of these higher-order aberrations through accounting for cyclorotation. There was no significant difference in the spherical aberration or coma between groups at three months. The similarity of induced spherical aberration between groups makes sense in that spherical aberration is spherically symmetric aberration and would be unaffected by cyclorotation that would be compensated for by iris recognition. On the other hand, coma could be affected by a cyclorotational misalignment. While coma was less in the study group than the control group at both one and three months, this difference was not statistically significant. The likely explanation for this is that both groups had coma induced from similar amounts of ablation decentration, as discussed below.
With respect to measurement of contrast sensitivity, low spatial frequencies (1.0 and 1.7 cpd) correspond to vision in low light situations and night vision. High spatial frequencies (6.6 and 10.4 cpd) corresponded to vision in normal lighting conditions and daytime vision. Previous studies have reported that both wavefront-guided and non-wavefront-guided LASIK cause a short-term depression in contrast sensitivity at certain spatial frequencies that recover in one to 12 months.26–28 In our study, we found that, after one month, both groups had similar slightly depressed contrast sensitivity at all spatial frequencies, however, by three months after the surgery, the study group had increased contrast sensitivity over pre-operative levels compared to the study group at the one midrange and the high spatial frequencies (2.6, 4.2 and 6.6 cpd). Because contrast sensitivity improves with time after LASIK, its early depression in both groups may be caused by the early induction of higher-order aberrations. As wound healing decreases these aberrations, contrast sensitivity improves, particularly in the iris recognition group, as described above.
Centration of refractive surgical procedures is still an area of debate among surgeons. Some surgeons prefer to centre on the entrance pupil (line of sight), others on the Purkinje image of the fixation target. Mrochen and associates29 proposed that the ideal centration point for refractive surgery is located between these two points. Like Ghosh and co-workers' study,12 we used the centre of the entrance pupil for centration of the ablation. The entrance pupil is defined as a virtual aperture that defines the area at the entrance of the system that can accept light that is, the physiologic pupil. Mrochen and associates29 found that subclinical decentration (less than one millimetre) was a major factor in increased coma-like aberrations after corneal laser surgery, even when the decentration was as little as 0.2 mm. In the present study, the decentration of the ablation centre relative to the centre of the entrance pupil was 0.37 mm and 0.44 mm in study and control groups, respectively (p = 0.052). These similar levels of decentration during the ablation are likely to explain the similar amount of coma induced in each group.
In our study, we found that most decentred ablations were decentred superiorly. Lee20 also reported that most decentrations are superior. We speculate that this tendency for superiorly decentred ablations in LASIK may be due to a systematic bias in the initial centring of the infrared tracker.
While we feel that our study provides valuable information about iris recognition in LASIK, this study was limited by two factors. We neither tracked subjective symptoms such as glare or halos, nor did we use a standardised pre-operative and post-operative patient satisfaction questionnaire. Both pieces of subjective data would have added valuable information to this comparative analysis.30
In conclusion, eyes that underwent LASIK with iris recognition had better visual outcomes and contrast sensitivity as well as fewer induced HOAs than eyes that underwent LASIK without iris recognition.