- Email: [email protected]
Laser in situ keratomileusis flap measurements: Comparison between observers and between spectral-domain and time-domain anterior segment optical coherence tomography
ARTICLE
Laser in situ keratomileusis flap measurements: Comparison between observers and between spectral-domain and time-domain anterior segment optical coherence tomography Reece C. Hall, FRANZCO, Farook K. Mohamed, D Opt, Hla M. Htoon, PhD, Donald T. Tan, FRCOphth, Jodhbir S. Mehta, FRCS(Ed)
PURPOSE: To evaluate the between-observer (interobserver) and between-instrument (intraobserver) variability in flap thickness measurements after laser in situ keratomileusis (LASIK) using spectral-domain and time-domain anterior segment optical coherence tomography (AS-OCT). SETTING: Singapore National Eye Centre. DESIGN: Evaluation of diagnostic test or technology. METHODS: Two independent masked observers measured flap thickness 1 month after LASIK using spectral-domain (RTVue) or time-domain (Visante) AS-OCT. The measurements were taken at central (0.0 mm), 1.5 mm, and C1.5 mm locations. Measurements were repeated to assess between-instrument variability. RESULTS: There was no statistically significant difference in mean flap thickness between the 2 observers at 1.5 mm, 0.0 mm, and C1.5 mm on spectral-domain AS-OCT and at 1.5 mm and C1.5 mm on time-domain AS-OCT (P < .01). There was a statistically significant difference between the 2 observers in the central (0.0 mm) values on time-domain AS-OCT (PZ.0008). There was stronger interobserver correlation for spectral-domain AS-OCT at 1.5 mm (r Z 0.82), 0.0 mm (r Z 0.88), and C1.5 mm (r Z 0.88) than for time-domain AS-OCT (r Z 0.73, r Z 0.62, and r Z 0.79, respectively). There was no statistically significant difference in between-instrument measurements. There was stronger between-instrument correlation with spectral-domain AS-OCT than with time-domain AS-OCT at all locations. The mean standard deviation (measure of instrument repeatability) for spectral-domain AS-OCT was 4.19 mm. CONCLUSION: Spectral-domain AS-OCT had closer agreement in between-observer and betweeninstrument measurements than time-domain AS-OCT and provided more consistent measurements of post-LASIK flap thickness. Financial Disclosure: No author has a financial or proprietary interest in any material or method mentioned. J Cataract Refract Surg 2011; 37:544–551 Q 2011 ASCRS and ESCRS
Laser in situ keratomileusis (LASIK) is the most commonly performed refractive procedure to correct ametropia and is considered safe and effective.1 Accurate measurement of corneal thickness is important in planning treatment. Knowledge of flap thickness and residual stromal bed thickness is important in reducing the risk for keratectasia, especially in cases requiring enhancement.2,3 Postoperative measurements after LASIK are useful in assessing the accuracy 544
Q 2011 ASCRS and ESCRS Published by Elsevier Inc.
and repeatability of a LASIK flap created with a microkeratome or a femtosecond laser.4–7 It can also help in calibrating a femtosecond laser during installation of new systems. Ultrasound (US) pachymetry is the most commonly used method for measuring corneal thickness preoperatively and after the LASIK flap has been cut. However, this method requires intraoperative contact with the eye and cannot measure flap thickness after 0886-3350/$ - see front matter doi:10.1016/j.jcrs.2010.10.037
LASIK FLAP MEASUREMENT WITH AS-OCT
LASIK. Recent studies show that intraoperative measurements with US pachymetry after LASIK flap creation with a femtosecond laser are different from postoperative optical coherence tomography (OCT) measurements.A This may be due to a variation in stromal hydration during measurements. Anterior segment OCT (AS-OCT) is a recently introduced noncontact method that enables postoperative measurement of the LASIK flap and the remaining stromal bed.8 The aim of this study was to evaluate the between-observer and between-instrument variability in post-LASIK flap thickness measurements with a Fourier/spectral-domain AS-OCT system and a time-domain AS-OCT system. PATIENTS AND METHODS This prospective study evaluated patients who had femtosecond-assisted LASIK between November 2008 and January 2009 at Singapore National Eye Centre. The study was performed according to the tenets of the Declaration of Helsinki, and patients gave informed consent after receiving a full explanation of the nature and intent of the study. The local ethics committee approved the study. To be included in the study, patients had to be older than 21 years with myopia higher than 1.50 diopters (D) and less than 15.00 D. Exclusion criteria were a history of eye disease and ocular abnormalities found on examination that would normally exclude the patient from being suitable for LASIK.
545
Optical Coherence Tomography Systems Two AS-OCT systems were used in this study. Both are approved by the U.S. Food and Drug Administration. The Visante (Carl Zeiss Meditec AG) is a time-domain acquisition system. Its performance has been described.9 The system uses a 1310 nm superluminescent diode source and operates up to a speed of 8 frames per second (2000 A-scans per second). The transverse resolution is 60 mm and the axial resolution, 10 to 20 mm. It can scan the anterior segment up to a width of 13.0 mm. The reference mirror moves 1 cycle for each axial scan; this mechanical movement limits the speed of image acquisition. The system uses widefield scanning optics (16.0 mm) and a deep axial scan range (8.0 mm) to image the entire anterior chamber in a single frame. After acquisition, the scanned images are processed by customized dewarping software, which compensates for the index of refraction transition at the air–tear interface and the different group indices of air, cornea, and aqueous to correct the image’s physical dimensions.10 The RTVue (Optovue, Inc.) is a newer AS-OCT system that uses Fourier/spectral-domain acquisition. It has an 830 nm light source and performs and analyzes 26 000 axial scans per second to achieve an image resolution of 5 mm. It can scan the anterior segment up to a width of 6.0 mm. In the system, the reference mirror is kept stationary and spectral interferograms obtained between the sample and reference reflections are transformed to axial scans. There are no moving parts, which allows rapid acquisition of images. The reflections from all layers are detected simultaneously. After acquisition, the scanned images are similarly modified to provide dispersion compensation (dewarping, segmentation, and measurements) before image processing.
Scanning Technique Surgical Technique One of 2 surgeons performed all LASIK procedures. The LASIK flap was created with a femtosecond laser (VisuMax, Carl Zeiss Meditec AG) and the corneal ablation, with an excimer laser (Technolas T127z100, Bausch & Lomb). The intended flap thickness was 120 mm in all eyes.
Submitted: November 23, 2009. Final revision submitted: August 31, 2010. Accepted: October 5, 2010. From Singapore National Eye Centre (Hall, Mohamed, Tan, Mehta), Singapore Eye Research Institute (Mohamed, Htoon, Tan, Mehta), Yong Loo Lin School of Medicine (Tan), National University of Singapore, and the Department of Clinical Sciences (Mehta), Duke-NUS Graduate Medical School, Singapore. Funded by the National Research Foundation, Funded Translational & Clinical Research Programme Grant (NMRC/TCR/002-SERI/ 2008-TCR 621/41/2008), Singapore. Presented at XXVII Congress of the European Society of Cataract & Refractive Surgeons, Barcelona, Spain, September 2009. Corresponding author: Jodhbir S. Mehta, FRCS(Ed), Singapore National Eye Centre, 11 Third Hospital Avenue, Singapore 168751. E-mail: [email protected].
Before scanning, an examiner (F.M.) performed a slitlamp examination with the patient seated. The same examiner performed all scanning with the patient’s eye undilated 1 month postoperatively. During scanning, patients were asked to fixate on the external fixation light, which was positioned so the patient was looking straight ahead. The AS-OCT images of the cornea were obtained in a dark room using a standard anterior segment single-scan protocol. To obtain the best quality image, the examiner adjusted the saturation and noise and optimized the polarization for each scan during the examination. The definition of reproducibility used in this study was based on definitions adopted by the International Organization for Standardization.11,12 Patients were instructed to keep their eyes wide open during scanning; when necessary, the lids were gently held apart (with care not to exert pressure on the globe) to ensure that the lids did not block the image. Cross-sectional scans were displayed continuously on the integrated video monitor. The examiner adjusted the system to position the vertex at the center of the AS-OCT image and to maximize vertex reflection. All measurements were taken between 10 AM and 4 PM to minimize the effect of diurnal variation on corneal thickness.13,14 More than 1 horizontal scan from each AS-OCT system was performed. The best-quality scan was used for measurements. Acceptable scans were selected as soon as they appeared. Images were judged to be of adequate quality based on the absence of artifacts caused by motion or the eyelid margins. The final images from each system were an averaged image of consecutive OCT frames. This was done to reduce speckle noise and increase the signal-to-noise ratio.
J CATARACT REFRACT SURG - VOL 37, MARCH 2011
546
LASIK FLAP MEASUREMENT WITH AS-OCT
Statistical Analysis The recorded measurements were entered into a database, and statistical analysis was performed using SPSS for Windows software (version 17.0, SPSS, Inc.). A 2-tailed paired t test was used for the difference between measurements. A P value less than 0.01 was considered statistically significant. The Pearson correlation coefficient (r) was used to assess the correlation between measurements. A correlation was considered very strong if the r value was more than 0.8 and moderately strong if it was 0.6 to 0.8.15 BlandAltman plots were used to determine agreements between measurements from each AS-OCT system. MedCalc software (version 9.3, MedCalc Software bvba) was used to analyze Bland-Altman plots; 95% limits of agreement (LoA) were considered valid for Bland-Altman plots.16
Figure 1. Anterior segment-OCT images of the same patient showing flap thickness measurements from the spectral-domain (top) and time-domain (bottom) systems.
Image Analysis Two trained, independent, masked observers (F.M., R.H.) reviewed both sets of AS-OCT scans. Using calipers provided by the software of each AS-OCT system, they measured the flap thickness at the reflective line on each horizontal scan. Measurements of scans from both systems were taken at the center (0.0 mm) and at C1.5 mm and 1.5 mm on either side of the center. Figure 1 shows examples of the scans. After flap measurement, the calipers were deleted from the scan and the data recorded on a spreadsheet. One week later, each observer repeated the flap thickness measurements, taking 2 measurements on the spectral-domain scans and 2 on the time-domain scans. Each observer was masked to the measurement of the other observer and from previous measurements. A distance of 1.5 mm was chosen because both AS-OCT systems can measure that distance. One observer (F.M.) performed a substudy of instrument repeatability using 5 consecutive measurements on 10 patient scans taken with the spectral-domain system. The mean standard deviation (SD) of these measurements was used as a measure of instrument repeatability for that system.
RESULTS Twenty patients (9 men, 11 women) with a mean age of 32.5 years (range 23 to 54 years) participated in this study. There were no intraoperative or postoperative complications, and no complication occurred as a result of scanning. Overall, the scanning quality was considered to be high. The difference in resolution between the 2 scans was notable, and both observers agreed that the spectral-domain images were better and made identifying the flap interface for measurement easier. Between Observers Table 1 shows the mean flap thickness measured by observer 1 and observer 2 with both AS-OCT systems. There was no statistically significant difference in the between-observer spectral-domain measurements. The between-observer difference at the central (0.0 mm) time-domain measurement was statistically significant (PZ.0008). There was no statistically significant difference in the between-observer spectral-domain reading
Table 1. Between-observer mean flap thickness measured with the 2 AS-OCT systems. Mean Thickness (mm) G SD System/Position Spectral domain 1.5 mm 0.0 mm C1.5 mm Time domain 1.5 mm 0.0 mm C1.5 mm
Observer 1
Observer 2
131.9 G 10.1 133.2 G 12.8 133.6 G 12.1
133.0 G 10.9 133.0 G 12.1 133.8 G 11.7
134.5 G 9.8 138.2 G 10.6 134.8 G 10.9
129.4 G 9.9 127.5 G 9.5 130.2 G 10.3
Bias (95% CI) †
r Value
LoA Mean
Upper Limit
Lower Limit
.650 .668 .933
0.828 0.888 0.880
1.1 (3.1 to 0.9) 1.2 (0.7 to 3.1) 0.2 (2.1 to 1.7)
11.1 (7.7 to 14.6) 12.9 (9.6 to 16.1) 11.3 (8.0 to 14.5)
13.3 (16.7 to 9.8) 10.4 (13.7 to 7.2) 11.7 (15.0 to 8.5)
.023† .0008x .055
0.729 0.620 0.789
5.1 (2.8 to 7.5) 10.7 (7.9 to 13.6) 4.6 (2.4 to 6.8)
19.4 (15.4 to 23.4) 28.1 (23.2 to 33.0) 18.2 (14.4 to 22.0)
9.1 (13.1 to 5.1) 6.6 (11.5 to 1.8) 9.0 (12.8 to 5.1)
P Value*
CI Z confidence interval; LoA Z limits of agreement *Paired t test † Based on observer 1 observer 2 values z Statistically significant (P!.05) x Statistically significant (P!.01)
J CATARACT REFRACT SURG - VOL 37, MARCH 2011
547
LASIK FLAP MEASUREMENT WITH AS-OCT
at any point. The spectral-domain measurements had a lower systemic bias than the time-domain measurements for between-observer comparisons. The difference in between-observer spectral-domain measurements was within the a priori tolerance level (1 to 2 mm). There was a larger variation in the betweenobserver time-domain measurements. There was very strong between-observer correlation for spectral-domain measurements and a moderately strong correlation for time-domain measurements at all locations.
was no statistically significant between-instrument difference in spectral-domain measurements, and there was a very strong correlation for observer 1 and observer 2. There was no statistically significant difference in between-instrument time-domain measurements; there was a very strong correlation for observer 1 and a moderately strong correlation for observer 2. There was no statistically significant difference in between-instrument measurements for either AS-OCT system. The mean SD as a measure of instrument repeatability for spectral-domain system was 4.19 mm.
Between Instruments Table 2 shows the spectral-domain and timedomain mean flap thickness and t test for each observer point. Observer 2 had a consistently higher reading for time-domain readings than for spectraldomain readings. In contrast, observer 1 had a higher reading for spectral-domain than for time-domain. However, there were no statistically significant differences between the 2 methods. The Bland-Altman plots (Figure 2) showed agreement between spectral-domain and time-domain measurements at 1.5 mm and C1.5 mm for both observers. For observer 1, the time-domain reading was slightly greater than (mean 1.2 mm) the spectral-domain reading at C1.5 mm. At 1.5 mm, the time-domain reading was slightly greater than (mean 2.6 mm) the spectral-domain reading. For observer 2, the time-domain reading was slightly lower than (mean 3.6 mm) the spectraldomain reading at C1.5 mm. At 1.5 mm, the timedomain reading was slightly lower than (mean 3.6 mm) the spectral-domain reading. There was no relationship between discrepancy and level of measurement. Table 3 shows the statistical analysis of betweeninstrument measurements for each observer. There
DISCUSSION We evaluated the between-observer and betweeninstrument variability in measurement of post-LASIK flap thickness using the RTVue spectral-domain AS-OCT system and the Visante time-domain AS-OCT system. Most differences in measurements between observers or between instruments were not statistically signficant. However, we found a statistically significant difference in between-observer measurements at the center (0.0 mm) position with the time-domain AS-OCT system (PZ.0008). The Pearson correlation coefficients were very strong for the spectral-domain system and moderately strong for the time-domain system for between-observer and between-instrument measurements. This suggests that the LASIK flap measurements from the spectraldomain system’s images are more reproducible. The observers had some difficulty using the Visante calipers during measurements. When the caliper was moved, it had a tendency to jump at a greater distance (O15 mm) than the desired point selected by the observers. Our mean flap thickness and SD results are comparable to those in other studies using this
Table 2. Mean flap thickness measured by the 2 AS-OCT systems at each observer position. Mean Thickness (mm) G SD
Bias (95% CI)
Observer/Position Spectral Domain Time Domain P Value* r Value Observer 1 1.5 mm 0.0 mm C1.5 mm Observer 2 1.5 mm 0.0 mm C1.5 mm
LoA Mean†
Lower Limit
Upper Limit
131.9 G 10.1 133.2 G 12.8 133.6 G 12.1
134.5 G 9.8 138.2 G 10.6 134.8 G 10.9
.123 .034z .380
0.508 0.387 0.603
2.6 (5.8 to 0.6) 22.1 (27.6 to 16.6) 16.8 (11.4 to 22.3) 5.0 (9.2 to 0.8) 30.7 (38.0 to 23.5) 20.7 (13.5 to 27.9) 1.2 (4.5 to 2.1) 21.5 (27.2 to 15.8) 19.0 (13.4 to 24.7)
133.0 G 10.9 132.0 G 12.1 133.8 G 11.7
129.4 G 9.9 127.5 G 9.9 130.2 G 10.3
.044z .022z .114
0.642 0.540 0.599
3.6 (0.7 to 6.4) 13.8 (18.7 to 8.9) 20.9 (16.1 to 25.8) 4.5 (1.1 to 7.9) 16.3 (22.1 to 10.4) 25.3 (19.4 to 31.1) 3.6 (0.4 to 6.8) 15.9 (21.4 to 10.4) 23.1 (17.6 to 28.6)
CI Z confidence interval; LoA Z limits of agreement *Paired t test † Based on Fourier-domain time-domain values z Statistically significant (P!.05)
J CATARACT REFRACT SURG - VOL 37, MARCH 2011
548
LASIK FLAP MEASUREMENT WITH AS-OCT
Figure 2. Bland-Altman plots for observer 1 and observer 2 comparing spectral-domain and time-domain measurements at 1.5 mm and C1.5 mm. The horizontal lines represent the mean and 95% LoA.
device.6,8,17 Data provided by the manufacturer of the Visante system for between-observer flap measurement at C1.0 mm and 1.0 mm show an SD of 18.0 mm and 20.2 mm, respectively.B The SD of the mean bias between observers that we found (7.3 mm at 1.5 mm, 8.9 mm at 0.0 mm, 6.9 mm at 1.5 mm) was smaller than in other studies using Visante and prototype machines with a 1310 nm laser.8,18 To our knowledge, there are no published data with which we can compare our RTVue flap measurements after LASIK. The only statistically significant difference in between-observer measurements was at 0.0 mm when a P value of less than 0.01 was used. If a P value of less than 0.05 were used, the between-observer measurement at 1.5 mm with the time-domain system would also be statistically significantly different between observers. Cheng et al.8 used a P value of less than 0.01 in their study to compare measurements,
but the manufacturer of the Visante system used a P value of less than 0.05. Had Cheng et al.8 also used a significance value of less than 0.05, in addition to finding a significant difference between US pachymetry measurements and OCT, they would have found a statistically significant difference in betweenobserver central measurements. The RTVue system has a speed of 26 kHz and an axial resolution of 5 mm. This is 13-times faster with 3 times more resolution than the Visante system. The RTVue system uses a shorter wavelength (830 nm) and broader bandwidth, producing better transverse and axial resolutions. The higher speed also results in very-high-definition images in a small fraction of a second.15 We found the higher resolution of RTVue scans made it less challenging to discern the flap interface than on the Visante scans. The lower resolution of Visante scans made determining the interface difficult,
J CATARACT REFRACT SURG - VOL 37, MARCH 2011
LASIK FLAP MEASUREMENT WITH AS-OCT
Table 3. Statistical analysis of between-instrument measurements for each observer. Observer 1 Position Spectral domain 1.5 mm 0.0 mm C1.5 mm Time domain 1.5 mm 0.0 mm C1.5 mm
Observer 2
P Value*
r Value
P Value*
r Value
.853 .940 .733
0.926 0.943 0.937
.680 .675 .486
0.936 0.907 0.935
.086 .070 .103
0.828 0.841 0.867
.125 .045† .040†
0.772 0.751 0.763
*t test † Statistically significant (P!.05)
and we found that this became statistically significant at the center 0.0 mm point. This can be explained by the strong specular reflection seen on the scan at this point, making it difficult to see the flap interface in some cases. In myopic treatments, the thinnest point is the center 0.0 mm, making it the most important location for obtaining a residual stromal thickness measurement when planning retreatment. Studies by the Visante manufacturerB used 1.0 mm left and right of center (0.0 mm). We recommend taking Visante measurements at 1.5 mm and C1.5 mm left and right of center. This is not as important with the RTVue system, and measurements taken at 0.0 mm with this device are more reproducible than with the Visante system. At present, the RTVue system is limited in the small scan range in tissue (approximately 6.0 mm wide) and the scan depth of 2.3 mm. It therefore cannot measure as wide as the Visante device, nor can it scan the angle recess, iris root, or ciliary body.19,C This would be an issue when determining the thinnest point in hyperopic treatments when a 7.0 mm ablation zone is used. Newer software upgrades allow scanning up to 8.0 mm in length. Other methods of measuring flap thickness include the Artemis 3-dimensional very-high-frequency (VHF) digital US device (ArcScan, Inc.), which has a repeatability of 1.14 mm for LASIK flap thickness.20 This is better than published data on the repeatability for the Visante system (5.0 mm to 17.6 mm)21 and our measurement of repeatability for the RTVue system (4.19 mm). However, measurements with the Artemis device require direct contact with the eye via a waterbath. A study comparing US pachymetry and Visante AS-OCT8 found that AS-OCT overestimated flap thickness. Another study of total corneal thickness22 found that AS-OCT significantly underestimated corneal thickness. However, intraoperative US pachymetry can be highly variable.23 This is likely because of
549
the different amounts of stromal hydration at the time of measurement, and this is difficult to standardize. The flap is thickest immediately postoperatively as a result of flap edema, and the thickness decreases over the following 5 days.23 A main advantage of AS-OCT is that it is noninvasive. Factors reported to influence AS-OCT measurements are flap edema, epithelial thickness, and hyperplasia.18,24–26 Changes in epithelial thickness are permanent and will be included in any postoperative flap thickness measurements. The Artemis VHF US system avoids this by adding the postoperative stromal thickness to the preoperative epithelium thickness.17 With time, the flap interface also becomes more difficult to differentiate on AS-OCT; early measurements (ie, those at 1 month) are more defined than those at 1 year. In our study, the Pearson correlation coefficients were very strong (O0.8) for between-observer measurements with the RTVue system and moderately strong (0.6 to 0.8) with the Visante system. The between-instrument correlations were very strong (O0.8) for both observers with the RTVue system. With the Visante system, the between-instrument correlation was very strong (O0.8) for observer 1 and moderately strong for observer 2. Cheng et al.8 found similar moderately strong between-observer correlations and very strong between-instrument correlations with measurements taken with the Visante system. Our Visante results are equivalent to those in other published studies. We found the between-observer measurements were better with the RTVue system than with the Visante system. To our knowledge, there are no other published data with which to compare our RTVue results. The variability in between-observer measurements with the 2 AS-OCT system is not uncommon because in many practices, more than 1 person performs the scans and measurements. The stronger correlations of RTVue between-observer measurements would make it the preferred system in a practice in which more than 1 person is responsible for scanning and measuring LASIK flaps. Bland-Altman plots showed agreement between Visante and RTVue measurements at points 1.5 mm and C1.5 mm for both observers. Both machines performed well, with a slight variation in under and over measurements between the observers. There were no consistent deviations between the 2 machines. Limitations of our study include that the examiner who performed the scans was also an observer, which may have introduced bias. However, this was minimized by performing the measurements and the scanning on different days. We did not compare our results with those of US pachymetry during laser flap creation, which has been the most common method of
J CATARACT REFRACT SURG - VOL 37, MARCH 2011
550
LASIK FLAP MEASUREMENT WITH AS-OCT
measuring flap thickness. However, US pachymetry after femtosecond laser flap creation is reported to have highly variable results.A In addition, surgeons performed the LASIK procedures, which added to the variability in the thickness of flaps. However, our study aimed to compare measurements of the 2 AS-OCT systems, not the consistency of LASIK flap thickness created by the VisuMax femtosecond laser. Knowledge of flap thickness is important to achieve reproducible and repeatable flap creation with a femtosecond laser or microkeratome. It is also important to know the residual cornea stromal thickness when planning LASIK retreatments. With AS-OCT, these measurements can be taken postoperatively in a noncontact manner without lifting the flap. Flap thickness measurements by AS-OCT include epithelial thickness changes; thus, this method slightly overestimates the flap thickness and the measurement does not equal the flap thickness at the time of creation. In this study, we evaluated the between-observer and between-instrument variability in LASIK flap measurements with the RTVue and Visante AS-OCT systems. The RTVue system had closer agreement in between-observer and between-instrument measurements than the Visante system and provides more consistent estimates of flap thickness after LASIK.
16.
REFERENCES
17.
1. Shortt AJ, Bunce C, Allan BDS. Evidence for superior efficacy and safety of LASIK over photorefractive keratectomy for correction of myopia. Ophthalmology 2006; 113:1897–1908 2. Reinstein DZ, Srivannaboon S, Archer TJ, Silverman RH, Sutton H, Coleman DJ. Probability model of the inaccuracy of residual stromal thickness prediction to reduce the risk of ectasia after LASIK. Part I: quantifying individual risk. J Refract Surg 2006; 22:851–860 3. Reinstein DZ, Srivannaboon S, Archer TJ, Silverman RH, Sutton H, Coleman DJ. Probability model of the inaccuracy of residual stromal thickness prediction to reduce the risk of ectasia after LASIK. Part II: quantifying population risk. J Refract Surg 2006; 22:861–870 4. Solomon KD, Donnenfeld E, Sandoval HP, Al Sarraf O, Kasper TJ, Holzer MP, Slate EH, Vroman DT. Flap Thickness Study Group; Flap thickness accuracy: comparison of 6 microkeratome models. J Cataract Refract Surg 2004; 30:964–977 5. Miranda D, Smith SD, Krueger RR. Comparison of flap thickness reproducibility using microkeratomes with a second motor for advancement. Ophthalmology 2003; 110:1931–1934 6. Li Y, Netto MV, Shekhar R, Krueger RR, Huang D. A longitudinal study of LASIK flap and stromal thickness with high-speed optical coherence tomography. Ophthalmology 2007; 114:1124–1132 7. Stahl JE, Durrie DS, Schwendeman FJ, Boghossian AJ. Anterior segment OCT analysis of thin IntraLase femtosecond flaps. J Refract Surg 2007; 23:555–558 8. Cheng ACK, Ho T, Lau S, Wong AL, Leung C, Lam DSC. Measurement of LASIK flap thickness with anterior segment optical coherence tomography. J Refract Surg 2008; 24:879–884 9. Su DHW, Friedman DS, See JLS Chew PTK, Chan Y-H, Nolan WP, Smith SD, Huang D, Zheng C, Li Y, Foster PJ,
10.
11.
12.
13.
14.
15.
18.
19.
20.
21.
22.
23.
24.
Aug T. Degree of angle closure and extent of peripheral anterior synechiae: an anterior segment OCT study. Br J Ophthalmol 2008; 92:103–107 Console JW, Sakata LM, Aung T, Friedman DS, He M. Quantitative analysis of anterior segment optical coherence tomography images: the Zhongshan Angle Assessment Program. Br J Ophthalmol 2008; 92:1612–1616 International Organization for Standardization. Accuracy (Trueness and Precision) of Measurement Methods and Results. Part 1. General Principles and Definitions. Geneva, Switzerland, ISO, 1994; (ISO 5725-1) International Organization for Standardization. Accuracy (Trueness and Precision) of Measurement Methods and Results. Part 2. Basic Methods for the Determination of Repeatability and Reproducibility of a Standard Measurement Method. Geneva, Switzerland, ISO, 1994; (ISO 5725-2) Lattimore MR Jr, Kaupp S, Schallhorn S, Lewis R IV. Orbscan pachymetry; implications of a repeated measures and diurnal variation analysis. Ophthalmology 1999; 106:977–981 Harper CL, Boulton ME, Bennett D, Marcyniuk B, JarvisEvans JH, Tullo AB, Ridgway AE. Diurnal variations in human corneal thickness. Br J Ophthalmol 1996; 80:1068–1072. erratum 1997; 81:175 (the upper part of Figure 2, C on page 1070 should read “28-year-old woman” not “20-year-old woman.” Article available at: http://www.ncbi.nlm.nih.gov/pmc/articles/ PMC505705/pdf/brjopthal00012-0038.pdf. Accessed October 29, 2010 Chan YH. Biostatistics 104: correlational analysis. Singapore Med J 2003; 44:614–619. Available at: http://www.sma.org.sg/ smj/4412/4412bs1.pdf. Accessed October 29, 2010 Bunce C. Correlation, agreement, and Bland-Altman analysis: statistical analysis of method comparison studies. Am J Ophthalmol 2009; 148:4–6 von Jagow B, Kohnen T. Corneal architecture of femtosecond laser and microkeratome flaps imaged by anterior segment optical coherence tomography. J Cataract Refract Surg 2009; 35:35–41 Wang J, Thomas J, Cox I, Rollins A. Noncontact measurement of central corneal epithelial and flap thickness after laser in situ keratomileusis. Invest Ophthalmol Vis Sci 2004; 45:1812– 1816. Available at: http://www.iovs.org/cgi/reprint/45/6/1812. Accessed October 29, 2010 Ramos JLB, Li Y, Huang D. Clinical and research applications of anterior segment optical coherence tomography – a review. Clin Exp Ophthalmol 2009; 37:81–89 Reinstein DZ, Couch DG, Archer T. Direct residual stromal thickness measurement for assessing suitability for LASIK enhancement by Artemis 3D very high-frequency digital ultrasound arc scanning. J Cataract Refract Surg 2006; 32:1884–1888 Reinstein DZ, Archer TJ, Gobbe M, Johnson N. Accuracy and reproducibility of Artemis central flap thickness and visual outcomes of LASIK with the Carl Zeiss Meditec VisuMax femtosecond laser and MEL 80 excimer laser platforms. J Refract Surg 2010; 26:107–119 Cheng ACK, Rao SK, Lau S, Leung CKS, Lam DSC. Central corneal thickness measurements by ultrasound, Orbscan II, and Visante OCT after LASIK for myopia. J Refract Surg 2008; 24:361–365 Nagy ZZ, Resch M, Suveges I. Ultrasound evaluation of flap thickness, ablation depth, and corneal edema after laser in situ keratomileusis. J Refract Surg 2004; 20:279–281 Erie JC, Patel SV, McLaren JW, Ramirez M, Hodge DO, Maguire LJ, Bourne WM. Effect of myopic laser in situ keratomileusis on epithelial and stromal thickness; a confocal microscopy study. Ophthalmology 2002; 109:1447–1452
J CATARACT REFRACT SURG - VOL 37, MARCH 2011
551
LASIK FLAP MEASUREMENT WITH AS-OCT
25. Spadea L, Fasciani R, Necozione S, Balestrazzi E. Role of the corneal epithelium in refractive changes following laser in situ keratomileusis for high myopia. J Refract Surg 2000; 16: 133–139 26. Feng Y, Varikooty J, Simpson TL. Diurnal variation of corneal and corneal epithelial thickness measured using optical coherence tomography. Cornea 2001; 20:480–483
C. Bower BA, Zhao M, Chu A, Zawadzki RJ, Toth C, Izatt JA. Rapid volumetric imaging of the human retina in vivo using a low-cost, spectral domain optical coherence tomography system. IOVS 2005; 46: E-Abstract 1050 Abstract available at: http:// abstracts.iovs.org/cgi/content/abstract/46/5/1050. Accessed October 29, 2010
OTHER CITED MATERIAL A. Murakami Y, Manche EE, “Intraoperative Subtraction Pachymetry Versus Postoperative Anterior Segment OCT Imaging of Femtosecond and Mechanical Keratome LASIK Flaps,” presented at the ASCRS Symposium on Cataract, IOL and Refractive Surgery, San Francisco, California, USA, April 2009 B. Visante Operation Manual. Jena, Germany, Carl Zeiss Meditec AG, 2006; appendix D-6:158
J CATARACT REFRACT SURG - VOL 37, MARCH 2011
First author: Reece C. Hall, FRANZCO Singapore National Eye Centre, Singapore
Laser in situ keratomileusis flap measurements: Comparison between observers and between spectral-domain and time-domain anterior segment optical coherence tomography Reece C. Hall, FRANZCO, Farook K. Mohamed, D Opt, Hla M. Htoon, PhD, Donald T. Tan, FRCOphth, Jodhbir S. Mehta, FRCS(Ed)
PURPOSE: To evaluate the between-observer (interobserver) and between-instrument (intraobserver) variability in flap thickness measurements after laser in situ keratomileusis (LASIK) using spectral-domain and time-domain anterior segment optical coherence tomography (AS-OCT). SETTING: Singapore National Eye Centre. DESIGN: Evaluation of diagnostic test or technology. METHODS: Two independent masked observers measured flap thickness 1 month after LASIK using spectral-domain (RTVue) or time-domain (Visante) AS-OCT. The measurements were taken at central (0.0 mm), 1.5 mm, and C1.5 mm locations. Measurements were repeated to assess between-instrument variability. RESULTS: There was no statistically significant difference in mean flap thickness between the 2 observers at 1.5 mm, 0.0 mm, and C1.5 mm on spectral-domain AS-OCT and at 1.5 mm and C1.5 mm on time-domain AS-OCT (P < .01). There was a statistically significant difference between the 2 observers in the central (0.0 mm) values on time-domain AS-OCT (PZ.0008). There was stronger interobserver correlation for spectral-domain AS-OCT at 1.5 mm (r Z 0.82), 0.0 mm (r Z 0.88), and C1.5 mm (r Z 0.88) than for time-domain AS-OCT (r Z 0.73, r Z 0.62, and r Z 0.79, respectively). There was no statistically significant difference in between-instrument measurements. There was stronger between-instrument correlation with spectral-domain AS-OCT than with time-domain AS-OCT at all locations. The mean standard deviation (measure of instrument repeatability) for spectral-domain AS-OCT was 4.19 mm. CONCLUSION: Spectral-domain AS-OCT had closer agreement in between-observer and betweeninstrument measurements than time-domain AS-OCT and provided more consistent measurements of post-LASIK flap thickness. Financial Disclosure: No author has a financial or proprietary interest in any material or method mentioned. J Cataract Refract Surg 2011; 37:544–551 Q 2011 ASCRS and ESCRS
Laser in situ keratomileusis (LASIK) is the most commonly performed refractive procedure to correct ametropia and is considered safe and effective.1 Accurate measurement of corneal thickness is important in planning treatment. Knowledge of flap thickness and residual stromal bed thickness is important in reducing the risk for keratectasia, especially in cases requiring enhancement.2,3 Postoperative measurements after LASIK are useful in assessing the accuracy 544
Q 2011 ASCRS and ESCRS Published by Elsevier Inc.
and repeatability of a LASIK flap created with a microkeratome or a femtosecond laser.4–7 It can also help in calibrating a femtosecond laser during installation of new systems. Ultrasound (US) pachymetry is the most commonly used method for measuring corneal thickness preoperatively and after the LASIK flap has been cut. However, this method requires intraoperative contact with the eye and cannot measure flap thickness after 0886-3350/$ - see front matter doi:10.1016/j.jcrs.2010.10.037
LASIK FLAP MEASUREMENT WITH AS-OCT
LASIK. Recent studies show that intraoperative measurements with US pachymetry after LASIK flap creation with a femtosecond laser are different from postoperative optical coherence tomography (OCT) measurements.A This may be due to a variation in stromal hydration during measurements. Anterior segment OCT (AS-OCT) is a recently introduced noncontact method that enables postoperative measurement of the LASIK flap and the remaining stromal bed.8 The aim of this study was to evaluate the between-observer and between-instrument variability in post-LASIK flap thickness measurements with a Fourier/spectral-domain AS-OCT system and a time-domain AS-OCT system. PATIENTS AND METHODS This prospective study evaluated patients who had femtosecond-assisted LASIK between November 2008 and January 2009 at Singapore National Eye Centre. The study was performed according to the tenets of the Declaration of Helsinki, and patients gave informed consent after receiving a full explanation of the nature and intent of the study. The local ethics committee approved the study. To be included in the study, patients had to be older than 21 years with myopia higher than 1.50 diopters (D) and less than 15.00 D. Exclusion criteria were a history of eye disease and ocular abnormalities found on examination that would normally exclude the patient from being suitable for LASIK.
545
Optical Coherence Tomography Systems Two AS-OCT systems were used in this study. Both are approved by the U.S. Food and Drug Administration. The Visante (Carl Zeiss Meditec AG) is a time-domain acquisition system. Its performance has been described.9 The system uses a 1310 nm superluminescent diode source and operates up to a speed of 8 frames per second (2000 A-scans per second). The transverse resolution is 60 mm and the axial resolution, 10 to 20 mm. It can scan the anterior segment up to a width of 13.0 mm. The reference mirror moves 1 cycle for each axial scan; this mechanical movement limits the speed of image acquisition. The system uses widefield scanning optics (16.0 mm) and a deep axial scan range (8.0 mm) to image the entire anterior chamber in a single frame. After acquisition, the scanned images are processed by customized dewarping software, which compensates for the index of refraction transition at the air–tear interface and the different group indices of air, cornea, and aqueous to correct the image’s physical dimensions.10 The RTVue (Optovue, Inc.) is a newer AS-OCT system that uses Fourier/spectral-domain acquisition. It has an 830 nm light source and performs and analyzes 26 000 axial scans per second to achieve an image resolution of 5 mm. It can scan the anterior segment up to a width of 6.0 mm. In the system, the reference mirror is kept stationary and spectral interferograms obtained between the sample and reference reflections are transformed to axial scans. There are no moving parts, which allows rapid acquisition of images. The reflections from all layers are detected simultaneously. After acquisition, the scanned images are similarly modified to provide dispersion compensation (dewarping, segmentation, and measurements) before image processing.
Scanning Technique Surgical Technique One of 2 surgeons performed all LASIK procedures. The LASIK flap was created with a femtosecond laser (VisuMax, Carl Zeiss Meditec AG) and the corneal ablation, with an excimer laser (Technolas T127z100, Bausch & Lomb). The intended flap thickness was 120 mm in all eyes.
Submitted: November 23, 2009. Final revision submitted: August 31, 2010. Accepted: October 5, 2010. From Singapore National Eye Centre (Hall, Mohamed, Tan, Mehta), Singapore Eye Research Institute (Mohamed, Htoon, Tan, Mehta), Yong Loo Lin School of Medicine (Tan), National University of Singapore, and the Department of Clinical Sciences (Mehta), Duke-NUS Graduate Medical School, Singapore. Funded by the National Research Foundation, Funded Translational & Clinical Research Programme Grant (NMRC/TCR/002-SERI/ 2008-TCR 621/41/2008), Singapore. Presented at XXVII Congress of the European Society of Cataract & Refractive Surgeons, Barcelona, Spain, September 2009. Corresponding author: Jodhbir S. Mehta, FRCS(Ed), Singapore National Eye Centre, 11 Third Hospital Avenue, Singapore 168751. E-mail: [email protected].
Before scanning, an examiner (F.M.) performed a slitlamp examination with the patient seated. The same examiner performed all scanning with the patient’s eye undilated 1 month postoperatively. During scanning, patients were asked to fixate on the external fixation light, which was positioned so the patient was looking straight ahead. The AS-OCT images of the cornea were obtained in a dark room using a standard anterior segment single-scan protocol. To obtain the best quality image, the examiner adjusted the saturation and noise and optimized the polarization for each scan during the examination. The definition of reproducibility used in this study was based on definitions adopted by the International Organization for Standardization.11,12 Patients were instructed to keep their eyes wide open during scanning; when necessary, the lids were gently held apart (with care not to exert pressure on the globe) to ensure that the lids did not block the image. Cross-sectional scans were displayed continuously on the integrated video monitor. The examiner adjusted the system to position the vertex at the center of the AS-OCT image and to maximize vertex reflection. All measurements were taken between 10 AM and 4 PM to minimize the effect of diurnal variation on corneal thickness.13,14 More than 1 horizontal scan from each AS-OCT system was performed. The best-quality scan was used for measurements. Acceptable scans were selected as soon as they appeared. Images were judged to be of adequate quality based on the absence of artifacts caused by motion or the eyelid margins. The final images from each system were an averaged image of consecutive OCT frames. This was done to reduce speckle noise and increase the signal-to-noise ratio.
J CATARACT REFRACT SURG - VOL 37, MARCH 2011
546
LASIK FLAP MEASUREMENT WITH AS-OCT
Statistical Analysis The recorded measurements were entered into a database, and statistical analysis was performed using SPSS for Windows software (version 17.0, SPSS, Inc.). A 2-tailed paired t test was used for the difference between measurements. A P value less than 0.01 was considered statistically significant. The Pearson correlation coefficient (r) was used to assess the correlation between measurements. A correlation was considered very strong if the r value was more than 0.8 and moderately strong if it was 0.6 to 0.8.15 BlandAltman plots were used to determine agreements between measurements from each AS-OCT system. MedCalc software (version 9.3, MedCalc Software bvba) was used to analyze Bland-Altman plots; 95% limits of agreement (LoA) were considered valid for Bland-Altman plots.16
Figure 1. Anterior segment-OCT images of the same patient showing flap thickness measurements from the spectral-domain (top) and time-domain (bottom) systems.
Image Analysis Two trained, independent, masked observers (F.M., R.H.) reviewed both sets of AS-OCT scans. Using calipers provided by the software of each AS-OCT system, they measured the flap thickness at the reflective line on each horizontal scan. Measurements of scans from both systems were taken at the center (0.0 mm) and at C1.5 mm and 1.5 mm on either side of the center. Figure 1 shows examples of the scans. After flap measurement, the calipers were deleted from the scan and the data recorded on a spreadsheet. One week later, each observer repeated the flap thickness measurements, taking 2 measurements on the spectral-domain scans and 2 on the time-domain scans. Each observer was masked to the measurement of the other observer and from previous measurements. A distance of 1.5 mm was chosen because both AS-OCT systems can measure that distance. One observer (F.M.) performed a substudy of instrument repeatability using 5 consecutive measurements on 10 patient scans taken with the spectral-domain system. The mean standard deviation (SD) of these measurements was used as a measure of instrument repeatability for that system.
RESULTS Twenty patients (9 men, 11 women) with a mean age of 32.5 years (range 23 to 54 years) participated in this study. There were no intraoperative or postoperative complications, and no complication occurred as a result of scanning. Overall, the scanning quality was considered to be high. The difference in resolution between the 2 scans was notable, and both observers agreed that the spectral-domain images were better and made identifying the flap interface for measurement easier. Between Observers Table 1 shows the mean flap thickness measured by observer 1 and observer 2 with both AS-OCT systems. There was no statistically significant difference in the between-observer spectral-domain measurements. The between-observer difference at the central (0.0 mm) time-domain measurement was statistically significant (PZ.0008). There was no statistically significant difference in the between-observer spectral-domain reading
Table 1. Between-observer mean flap thickness measured with the 2 AS-OCT systems. Mean Thickness (mm) G SD System/Position Spectral domain 1.5 mm 0.0 mm C1.5 mm Time domain 1.5 mm 0.0 mm C1.5 mm
Observer 1
Observer 2
131.9 G 10.1 133.2 G 12.8 133.6 G 12.1
133.0 G 10.9 133.0 G 12.1 133.8 G 11.7
134.5 G 9.8 138.2 G 10.6 134.8 G 10.9
129.4 G 9.9 127.5 G 9.5 130.2 G 10.3
Bias (95% CI) †
r Value
LoA Mean
Upper Limit
Lower Limit
.650 .668 .933
0.828 0.888 0.880
1.1 (3.1 to 0.9) 1.2 (0.7 to 3.1) 0.2 (2.1 to 1.7)
11.1 (7.7 to 14.6) 12.9 (9.6 to 16.1) 11.3 (8.0 to 14.5)
13.3 (16.7 to 9.8) 10.4 (13.7 to 7.2) 11.7 (15.0 to 8.5)
.023† .0008x .055
0.729 0.620 0.789
5.1 (2.8 to 7.5) 10.7 (7.9 to 13.6) 4.6 (2.4 to 6.8)
19.4 (15.4 to 23.4) 28.1 (23.2 to 33.0) 18.2 (14.4 to 22.0)
9.1 (13.1 to 5.1) 6.6 (11.5 to 1.8) 9.0 (12.8 to 5.1)
P Value*
CI Z confidence interval; LoA Z limits of agreement *Paired t test † Based on observer 1 observer 2 values z Statistically significant (P!.05) x Statistically significant (P!.01)
J CATARACT REFRACT SURG - VOL 37, MARCH 2011
547
LASIK FLAP MEASUREMENT WITH AS-OCT
at any point. The spectral-domain measurements had a lower systemic bias than the time-domain measurements for between-observer comparisons. The difference in between-observer spectral-domain measurements was within the a priori tolerance level (1 to 2 mm). There was a larger variation in the betweenobserver time-domain measurements. There was very strong between-observer correlation for spectral-domain measurements and a moderately strong correlation for time-domain measurements at all locations.
was no statistically significant between-instrument difference in spectral-domain measurements, and there was a very strong correlation for observer 1 and observer 2. There was no statistically significant difference in between-instrument time-domain measurements; there was a very strong correlation for observer 1 and a moderately strong correlation for observer 2. There was no statistically significant difference in between-instrument measurements for either AS-OCT system. The mean SD as a measure of instrument repeatability for spectral-domain system was 4.19 mm.
Between Instruments Table 2 shows the spectral-domain and timedomain mean flap thickness and t test for each observer point. Observer 2 had a consistently higher reading for time-domain readings than for spectraldomain readings. In contrast, observer 1 had a higher reading for spectral-domain than for time-domain. However, there were no statistically significant differences between the 2 methods. The Bland-Altman plots (Figure 2) showed agreement between spectral-domain and time-domain measurements at 1.5 mm and C1.5 mm for both observers. For observer 1, the time-domain reading was slightly greater than (mean 1.2 mm) the spectral-domain reading at C1.5 mm. At 1.5 mm, the time-domain reading was slightly greater than (mean 2.6 mm) the spectral-domain reading. For observer 2, the time-domain reading was slightly lower than (mean 3.6 mm) the spectraldomain reading at C1.5 mm. At 1.5 mm, the timedomain reading was slightly lower than (mean 3.6 mm) the spectral-domain reading. There was no relationship between discrepancy and level of measurement. Table 3 shows the statistical analysis of betweeninstrument measurements for each observer. There
DISCUSSION We evaluated the between-observer and betweeninstrument variability in measurement of post-LASIK flap thickness using the RTVue spectral-domain AS-OCT system and the Visante time-domain AS-OCT system. Most differences in measurements between observers or between instruments were not statistically signficant. However, we found a statistically significant difference in between-observer measurements at the center (0.0 mm) position with the time-domain AS-OCT system (PZ.0008). The Pearson correlation coefficients were very strong for the spectral-domain system and moderately strong for the time-domain system for between-observer and between-instrument measurements. This suggests that the LASIK flap measurements from the spectraldomain system’s images are more reproducible. The observers had some difficulty using the Visante calipers during measurements. When the caliper was moved, it had a tendency to jump at a greater distance (O15 mm) than the desired point selected by the observers. Our mean flap thickness and SD results are comparable to those in other studies using this
Table 2. Mean flap thickness measured by the 2 AS-OCT systems at each observer position. Mean Thickness (mm) G SD
Bias (95% CI)
Observer/Position Spectral Domain Time Domain P Value* r Value Observer 1 1.5 mm 0.0 mm C1.5 mm Observer 2 1.5 mm 0.0 mm C1.5 mm
LoA Mean†
Lower Limit
Upper Limit
131.9 G 10.1 133.2 G 12.8 133.6 G 12.1
134.5 G 9.8 138.2 G 10.6 134.8 G 10.9
.123 .034z .380
0.508 0.387 0.603
2.6 (5.8 to 0.6) 22.1 (27.6 to 16.6) 16.8 (11.4 to 22.3) 5.0 (9.2 to 0.8) 30.7 (38.0 to 23.5) 20.7 (13.5 to 27.9) 1.2 (4.5 to 2.1) 21.5 (27.2 to 15.8) 19.0 (13.4 to 24.7)
133.0 G 10.9 132.0 G 12.1 133.8 G 11.7
129.4 G 9.9 127.5 G 9.9 130.2 G 10.3
.044z .022z .114
0.642 0.540 0.599
3.6 (0.7 to 6.4) 13.8 (18.7 to 8.9) 20.9 (16.1 to 25.8) 4.5 (1.1 to 7.9) 16.3 (22.1 to 10.4) 25.3 (19.4 to 31.1) 3.6 (0.4 to 6.8) 15.9 (21.4 to 10.4) 23.1 (17.6 to 28.6)
CI Z confidence interval; LoA Z limits of agreement *Paired t test † Based on Fourier-domain time-domain values z Statistically significant (P!.05)
J CATARACT REFRACT SURG - VOL 37, MARCH 2011
548
LASIK FLAP MEASUREMENT WITH AS-OCT
Figure 2. Bland-Altman plots for observer 1 and observer 2 comparing spectral-domain and time-domain measurements at 1.5 mm and C1.5 mm. The horizontal lines represent the mean and 95% LoA.
device.6,8,17 Data provided by the manufacturer of the Visante system for between-observer flap measurement at C1.0 mm and 1.0 mm show an SD of 18.0 mm and 20.2 mm, respectively.B The SD of the mean bias between observers that we found (7.3 mm at 1.5 mm, 8.9 mm at 0.0 mm, 6.9 mm at 1.5 mm) was smaller than in other studies using Visante and prototype machines with a 1310 nm laser.8,18 To our knowledge, there are no published data with which we can compare our RTVue flap measurements after LASIK. The only statistically significant difference in between-observer measurements was at 0.0 mm when a P value of less than 0.01 was used. If a P value of less than 0.05 were used, the between-observer measurement at 1.5 mm with the time-domain system would also be statistically significantly different between observers. Cheng et al.8 used a P value of less than 0.01 in their study to compare measurements,
but the manufacturer of the Visante system used a P value of less than 0.05. Had Cheng et al.8 also used a significance value of less than 0.05, in addition to finding a significant difference between US pachymetry measurements and OCT, they would have found a statistically significant difference in betweenobserver central measurements. The RTVue system has a speed of 26 kHz and an axial resolution of 5 mm. This is 13-times faster with 3 times more resolution than the Visante system. The RTVue system uses a shorter wavelength (830 nm) and broader bandwidth, producing better transverse and axial resolutions. The higher speed also results in very-high-definition images in a small fraction of a second.15 We found the higher resolution of RTVue scans made it less challenging to discern the flap interface than on the Visante scans. The lower resolution of Visante scans made determining the interface difficult,
J CATARACT REFRACT SURG - VOL 37, MARCH 2011
LASIK FLAP MEASUREMENT WITH AS-OCT
Table 3. Statistical analysis of between-instrument measurements for each observer. Observer 1 Position Spectral domain 1.5 mm 0.0 mm C1.5 mm Time domain 1.5 mm 0.0 mm C1.5 mm
Observer 2
P Value*
r Value
P Value*
r Value
.853 .940 .733
0.926 0.943 0.937
.680 .675 .486
0.936 0.907 0.935
.086 .070 .103
0.828 0.841 0.867
.125 .045† .040†
0.772 0.751 0.763
*t test † Statistically significant (P!.05)
and we found that this became statistically significant at the center 0.0 mm point. This can be explained by the strong specular reflection seen on the scan at this point, making it difficult to see the flap interface in some cases. In myopic treatments, the thinnest point is the center 0.0 mm, making it the most important location for obtaining a residual stromal thickness measurement when planning retreatment. Studies by the Visante manufacturerB used 1.0 mm left and right of center (0.0 mm). We recommend taking Visante measurements at 1.5 mm and C1.5 mm left and right of center. This is not as important with the RTVue system, and measurements taken at 0.0 mm with this device are more reproducible than with the Visante system. At present, the RTVue system is limited in the small scan range in tissue (approximately 6.0 mm wide) and the scan depth of 2.3 mm. It therefore cannot measure as wide as the Visante device, nor can it scan the angle recess, iris root, or ciliary body.19,C This would be an issue when determining the thinnest point in hyperopic treatments when a 7.0 mm ablation zone is used. Newer software upgrades allow scanning up to 8.0 mm in length. Other methods of measuring flap thickness include the Artemis 3-dimensional very-high-frequency (VHF) digital US device (ArcScan, Inc.), which has a repeatability of 1.14 mm for LASIK flap thickness.20 This is better than published data on the repeatability for the Visante system (5.0 mm to 17.6 mm)21 and our measurement of repeatability for the RTVue system (4.19 mm). However, measurements with the Artemis device require direct contact with the eye via a waterbath. A study comparing US pachymetry and Visante AS-OCT8 found that AS-OCT overestimated flap thickness. Another study of total corneal thickness22 found that AS-OCT significantly underestimated corneal thickness. However, intraoperative US pachymetry can be highly variable.23 This is likely because of
549
the different amounts of stromal hydration at the time of measurement, and this is difficult to standardize. The flap is thickest immediately postoperatively as a result of flap edema, and the thickness decreases over the following 5 days.23 A main advantage of AS-OCT is that it is noninvasive. Factors reported to influence AS-OCT measurements are flap edema, epithelial thickness, and hyperplasia.18,24–26 Changes in epithelial thickness are permanent and will be included in any postoperative flap thickness measurements. The Artemis VHF US system avoids this by adding the postoperative stromal thickness to the preoperative epithelium thickness.17 With time, the flap interface also becomes more difficult to differentiate on AS-OCT; early measurements (ie, those at 1 month) are more defined than those at 1 year. In our study, the Pearson correlation coefficients were very strong (O0.8) for between-observer measurements with the RTVue system and moderately strong (0.6 to 0.8) with the Visante system. The between-instrument correlations were very strong (O0.8) for both observers with the RTVue system. With the Visante system, the between-instrument correlation was very strong (O0.8) for observer 1 and moderately strong for observer 2. Cheng et al.8 found similar moderately strong between-observer correlations and very strong between-instrument correlations with measurements taken with the Visante system. Our Visante results are equivalent to those in other published studies. We found the between-observer measurements were better with the RTVue system than with the Visante system. To our knowledge, there are no other published data with which to compare our RTVue results. The variability in between-observer measurements with the 2 AS-OCT system is not uncommon because in many practices, more than 1 person performs the scans and measurements. The stronger correlations of RTVue between-observer measurements would make it the preferred system in a practice in which more than 1 person is responsible for scanning and measuring LASIK flaps. Bland-Altman plots showed agreement between Visante and RTVue measurements at points 1.5 mm and C1.5 mm for both observers. Both machines performed well, with a slight variation in under and over measurements between the observers. There were no consistent deviations between the 2 machines. Limitations of our study include that the examiner who performed the scans was also an observer, which may have introduced bias. However, this was minimized by performing the measurements and the scanning on different days. We did not compare our results with those of US pachymetry during laser flap creation, which has been the most common method of
J CATARACT REFRACT SURG - VOL 37, MARCH 2011
550
LASIK FLAP MEASUREMENT WITH AS-OCT
measuring flap thickness. However, US pachymetry after femtosecond laser flap creation is reported to have highly variable results.A In addition, surgeons performed the LASIK procedures, which added to the variability in the thickness of flaps. However, our study aimed to compare measurements of the 2 AS-OCT systems, not the consistency of LASIK flap thickness created by the VisuMax femtosecond laser. Knowledge of flap thickness is important to achieve reproducible and repeatable flap creation with a femtosecond laser or microkeratome. It is also important to know the residual cornea stromal thickness when planning LASIK retreatments. With AS-OCT, these measurements can be taken postoperatively in a noncontact manner without lifting the flap. Flap thickness measurements by AS-OCT include epithelial thickness changes; thus, this method slightly overestimates the flap thickness and the measurement does not equal the flap thickness at the time of creation. In this study, we evaluated the between-observer and between-instrument variability in LASIK flap measurements with the RTVue and Visante AS-OCT systems. The RTVue system had closer agreement in between-observer and between-instrument measurements than the Visante system and provides more consistent estimates of flap thickness after LASIK.
16.
REFERENCES
17.
1. Shortt AJ, Bunce C, Allan BDS. Evidence for superior efficacy and safety of LASIK over photorefractive keratectomy for correction of myopia. Ophthalmology 2006; 113:1897–1908 2. Reinstein DZ, Srivannaboon S, Archer TJ, Silverman RH, Sutton H, Coleman DJ. Probability model of the inaccuracy of residual stromal thickness prediction to reduce the risk of ectasia after LASIK. Part I: quantifying individual risk. J Refract Surg 2006; 22:851–860 3. Reinstein DZ, Srivannaboon S, Archer TJ, Silverman RH, Sutton H, Coleman DJ. Probability model of the inaccuracy of residual stromal thickness prediction to reduce the risk of ectasia after LASIK. Part II: quantifying population risk. J Refract Surg 2006; 22:861–870 4. Solomon KD, Donnenfeld E, Sandoval HP, Al Sarraf O, Kasper TJ, Holzer MP, Slate EH, Vroman DT. Flap Thickness Study Group; Flap thickness accuracy: comparison of 6 microkeratome models. J Cataract Refract Surg 2004; 30:964–977 5. Miranda D, Smith SD, Krueger RR. Comparison of flap thickness reproducibility using microkeratomes with a second motor for advancement. Ophthalmology 2003; 110:1931–1934 6. Li Y, Netto MV, Shekhar R, Krueger RR, Huang D. A longitudinal study of LASIK flap and stromal thickness with high-speed optical coherence tomography. Ophthalmology 2007; 114:1124–1132 7. Stahl JE, Durrie DS, Schwendeman FJ, Boghossian AJ. Anterior segment OCT analysis of thin IntraLase femtosecond flaps. J Refract Surg 2007; 23:555–558 8. Cheng ACK, Ho T, Lau S, Wong AL, Leung C, Lam DSC. Measurement of LASIK flap thickness with anterior segment optical coherence tomography. J Refract Surg 2008; 24:879–884 9. Su DHW, Friedman DS, See JLS Chew PTK, Chan Y-H, Nolan WP, Smith SD, Huang D, Zheng C, Li Y, Foster PJ,
10.
11.
12.
13.
14.
15.
18.
19.
20.
21.
22.
23.
24.
Aug T. Degree of angle closure and extent of peripheral anterior synechiae: an anterior segment OCT study. Br J Ophthalmol 2008; 92:103–107 Console JW, Sakata LM, Aung T, Friedman DS, He M. Quantitative analysis of anterior segment optical coherence tomography images: the Zhongshan Angle Assessment Program. Br J Ophthalmol 2008; 92:1612–1616 International Organization for Standardization. Accuracy (Trueness and Precision) of Measurement Methods and Results. Part 1. General Principles and Definitions. Geneva, Switzerland, ISO, 1994; (ISO 5725-1) International Organization for Standardization. Accuracy (Trueness and Precision) of Measurement Methods and Results. Part 2. Basic Methods for the Determination of Repeatability and Reproducibility of a Standard Measurement Method. Geneva, Switzerland, ISO, 1994; (ISO 5725-2) Lattimore MR Jr, Kaupp S, Schallhorn S, Lewis R IV. Orbscan pachymetry; implications of a repeated measures and diurnal variation analysis. Ophthalmology 1999; 106:977–981 Harper CL, Boulton ME, Bennett D, Marcyniuk B, JarvisEvans JH, Tullo AB, Ridgway AE. Diurnal variations in human corneal thickness. Br J Ophthalmol 1996; 80:1068–1072. erratum 1997; 81:175 (the upper part of Figure 2, C on page 1070 should read “28-year-old woman” not “20-year-old woman.” Article available at: http://www.ncbi.nlm.nih.gov/pmc/articles/ PMC505705/pdf/brjopthal00012-0038.pdf. Accessed October 29, 2010 Chan YH. Biostatistics 104: correlational analysis. Singapore Med J 2003; 44:614–619. Available at: http://www.sma.org.sg/ smj/4412/4412bs1.pdf. Accessed October 29, 2010 Bunce C. Correlation, agreement, and Bland-Altman analysis: statistical analysis of method comparison studies. Am J Ophthalmol 2009; 148:4–6 von Jagow B, Kohnen T. Corneal architecture of femtosecond laser and microkeratome flaps imaged by anterior segment optical coherence tomography. J Cataract Refract Surg 2009; 35:35–41 Wang J, Thomas J, Cox I, Rollins A. Noncontact measurement of central corneal epithelial and flap thickness after laser in situ keratomileusis. Invest Ophthalmol Vis Sci 2004; 45:1812– 1816. Available at: http://www.iovs.org/cgi/reprint/45/6/1812. Accessed October 29, 2010 Ramos JLB, Li Y, Huang D. Clinical and research applications of anterior segment optical coherence tomography – a review. Clin Exp Ophthalmol 2009; 37:81–89 Reinstein DZ, Couch DG, Archer T. Direct residual stromal thickness measurement for assessing suitability for LASIK enhancement by Artemis 3D very high-frequency digital ultrasound arc scanning. J Cataract Refract Surg 2006; 32:1884–1888 Reinstein DZ, Archer TJ, Gobbe M, Johnson N. Accuracy and reproducibility of Artemis central flap thickness and visual outcomes of LASIK with the Carl Zeiss Meditec VisuMax femtosecond laser and MEL 80 excimer laser platforms. J Refract Surg 2010; 26:107–119 Cheng ACK, Rao SK, Lau S, Leung CKS, Lam DSC. Central corneal thickness measurements by ultrasound, Orbscan II, and Visante OCT after LASIK for myopia. J Refract Surg 2008; 24:361–365 Nagy ZZ, Resch M, Suveges I. Ultrasound evaluation of flap thickness, ablation depth, and corneal edema after laser in situ keratomileusis. J Refract Surg 2004; 20:279–281 Erie JC, Patel SV, McLaren JW, Ramirez M, Hodge DO, Maguire LJ, Bourne WM. Effect of myopic laser in situ keratomileusis on epithelial and stromal thickness; a confocal microscopy study. Ophthalmology 2002; 109:1447–1452
J CATARACT REFRACT SURG - VOL 37, MARCH 2011
551
LASIK FLAP MEASUREMENT WITH AS-OCT
25. Spadea L, Fasciani R, Necozione S, Balestrazzi E. Role of the corneal epithelium in refractive changes following laser in situ keratomileusis for high myopia. J Refract Surg 2000; 16: 133–139 26. Feng Y, Varikooty J, Simpson TL. Diurnal variation of corneal and corneal epithelial thickness measured using optical coherence tomography. Cornea 2001; 20:480–483
C. Bower BA, Zhao M, Chu A, Zawadzki RJ, Toth C, Izatt JA. Rapid volumetric imaging of the human retina in vivo using a low-cost, spectral domain optical coherence tomography system. IOVS 2005; 46: E-Abstract 1050 Abstract available at: http:// abstracts.iovs.org/cgi/content/abstract/46/5/1050. Accessed October 29, 2010
OTHER CITED MATERIAL A. Murakami Y, Manche EE, “Intraoperative Subtraction Pachymetry Versus Postoperative Anterior Segment OCT Imaging of Femtosecond and Mechanical Keratome LASIK Flaps,” presented at the ASCRS Symposium on Cataract, IOL and Refractive Surgery, San Francisco, California, USA, April 2009 B. Visante Operation Manual. Jena, Germany, Carl Zeiss Meditec AG, 2006; appendix D-6:158
J CATARACT REFRACT SURG - VOL 37, MARCH 2011
First author: Reece C. Hall, FRANZCO Singapore National Eye Centre, Singapore