Secondary Logo

Journal Logo

Article

Effect of corneal thickness on the agreement between ultrasound and Orbscan II pachymetry

Hashemi, Hassan MD; Roshani, Maryam; Mehravaran, Shiva MD; Parsafar, Hiva MD; Yazdani, Kamran MD, MPH

Author Information
Journal of Cataract & Refractive Surgery: October 2007 - Volume 33 - Issue 10 - p 1694-1700
doi: 10.1016/j.jcrs.2007.05.036
  • Free

Abstract

The popularity of refractive surgery has increased the importance of accurate corneal thickness measurements. Preoperative assessment of central corneal thickness (CCT) is a requisite to ensure patient eligibility in terms of a sufficient corneal thickness for a safe correction and an acceptable residual thickness to reduce the risk for iatrogenic ectasia.1,2 Pachymeters are generally based on ultrasound (US) or optical technologies.3 The most commonly used devices are US pachymeters, which are also recognized as the gold standard for performing CCT measurements.4–6 However, in this technique, a probe must come in contact with the corneal surface. This requires local anesthesia and is associated with the risk for infection.7,8 Other disadvantages of the technique are the probability of misalignment, variations in placing the probe in repeated measurements, and corneal indentation by the probe.9,10 Also, US measurements are affected by the fluctuations in corneal tissue hydration. For these reasons, use of noncontact methods with an acceptable accuracy has been suggested.11

The Orbscan topography system (Bausch & Lomb) uses a computerized slit-scanning method to image the anterior and posterior corneal surfaces and a full corneal pachymetry map in a noncontact acquisition that takes fewer than 2 seconds.12–14 Although highly repeatable,2,15–17 the validity of its corneal thickness readings compared with US measurements is a matter of debate.

Results in initial comparative studies showed that the Orbscan overestimated CCT up to approximately 30 μm5,11,14,18; thus, the use of acoustic equivalent correction factors was suggested to make the Orbscan readings more comparable to US readings. The Orbscan II software allows users to apply such an acoustic factor to corneal thickness measurements displayed by the system, and the value recommended by the manufacturer is 0.92.

González-Méijome et al.19 measured and compared the CCT and peripheral corneal thickness at equal distances from the center. They found that a different equation was required for each measured point. Other studies suggest using a different correction factor for corneas that are thin postoperatively8,20 and performing further studies of the validity of the Orbscan corneal thickness measurements in thin and thick corneas.21

Our experience and published reports hint that the Orbscan II tends to underestimate the CCT in thinner corneas, whether normal, keratoconic,22,23 or after laser refractive surgery,24 and generates overestimated readings in thicker corneas. This study was designed to determine the effect of corneal thickness on the agreement between the Orbscan II and US measurements in unoperated eyes. We also tried to determine whether any correction factor could provide acceptable validity for Orbscan II measurements, especially in corneas thinner or thicker than normal.

PATIENTS AND METHODS

Prospectively, in a stratified sampling approach, 177 patients admitted to Noor Vision Correction Center were enrolled in 1 of 3 groups based on corneal thickness: less than 500 μm (thin corneas), 500 to 600 μm (medium corneas), and more than 600 μm (thick corneas). Exclusion criteria were a history of ocular surgery, history of ocular trauma causing apparent damage to the corneal surface, very poor cooperation, corneal haze, keratoconus, and use of a soft contact lens in the past 3 days or a hard contact lens in the past 3 weeks.

Central Corneal Thickness Measurement

The central thickness was measured in all corneas at least 3 hours after waking, first with the noncontact Orbscan II and then with the US method. The same technician performed measurements in each patient on the same day, and readings in the right eyes were recorded.

With the Orbscan II, after proper positioning, patients were instructed to focus on the red blinking light in front of them. The slits were adjusted, and an acquisition was made. During acquisition, 40 slits were projected and scanned over the cornea, each providing 240 data points from the anterior and posterior corneal surfaces. The corneal thickness was computed for every corneal point from the elevation difference between the 2 surfaces. In this study, only central readings were collected without applying an acoustic equivalent correction factor.

For US measurements, the cornea was anesthetized with tetracaine 0.5% eyedrops, and the sterilized probe was placed on the cornea as perpendicularly as possible. Five consecutive measurements were generated, the mean of which was used in the analysis.

Statistical Analysis

The Kolmogorov-Smirnov test was used to confirm the normal distribution of data sets (P>.05). Mean interdevice differences were determined through paired t tests in each group and in all eyes. The Pearson correlation coefficient and regression coefficient between the mean Orbscan II and US readings and their differences were calculated. The interdevice agreement was determined by the method described by Bland and Altman,25 differences between paired measurements were plotted against their means, and 95% limits of agreement (LoA) were calculated. Intraclass correlation coefficients (ICC) were used to assess the level of agreement as described by Fleiss and Cohen.26 To interpret the degree of agreement, the guidelines provided by Landis and Koch27 (Table 1) were used.

Table 1
Table 1:
Interpretation of ICC into level of agreement.

Regression analysis was used to determine regression equations: Ultrasound CCT = (Orbscan II CCT × slope) + intercept. The equation obtained in each group was applied to that group's Orbscan II data as the acoustic equivalent correction factor to derive corrected readings, after which the interdevice difference and agreement analyses were repeated. Results are expressed as mean ± standard deviation (SD) based on a significance level of <.05.

RESULTS

One hundred seventy-seven right eyes were examined in 66 men (37%) and 111 women (63%). The mean age of the patients was 29.7 ± 8.2 years (range 18 to 54 years). Table 2 shows the number of eyes, male-to-female ratio, and age range of patients in each corneal thickness group. The Pearson correlation coefficient between mean Orbscan II and US readings and their differences was 0.414 (P<.001) (Figure 1).

Table 2
Table 2:
Details of study participants grouped by CCT.
Figure 1
Figure 1:
Regression plot of interdevice differences in measuring the CCT against their mean in all eyes (P value for β coefficient (0.186)<.001).

Interdevice Differences

Table 3 shows the mean CCT readings and the interdevice differences. In the less than 500 μm group, the Orbscan II underestimated the CCT in 19 of 31 eyes but the mean difference (−1.6 ± 15.8 μm) was not statistically significant (P = .580). The Orbscan II overestimated the CCT in 105 of 115 eyes in the 500 to 600 μm group, and the mean difference (21.3 ± 20.2 μm) was statistically significant (P = .0001). In the more than 600 μm group, the Orbscan II overestimated the readings in 28 of 31 eyes (P<.0001), with a mean difference of 27.2 ± 20.9 μm. Overall, there was a mean overestimation of 18.0 ± 21.7 μm by the Orbscan II, a difference that was statistically significant (P<.0001).

Table 3
Table 3:
Central corneal thickness measurements with US and Orbscan II without correction and interdevice differences in the 3 groups and the total sample.

Linear Correlation Between Data Sets

Table 4 shows the equations derived from the regression analyses. After these equations were applied, the mean difference between the data sets was near zero in all groups (P>.05). These differences are shown in Table 4.

Table 4
Table 4:
Equations derived from linear regression analysis for each group, corrected Orbscan II readings with corresponding equations, and the interdevice differences after applying the equation.

Interdevice Agreement

Table 5 shows the ICC values in each group and in all eyes. There was a substantial interdevice agreement in the thin and medium cornea groups. Measurements in the thick cornea group, however, showed moderate agreement between the 2 systems.

Table 5
Table 5:
Intraclass correlation coefficients and the 95% confidence intervals for US and Orbscan II CCT measurements in each group.

Bland-Altman plots for each group before and after applying each correction equation (Table 4) are shown in Figures 2 and 3, respectively. Identical scales are used for easier comparisons. Correcting the Orbscan II readings with these equations reduced data distribution and the width of 95% LoA. This change was more apparent in the less than 500 μm group. The upper and lower 95% LoA before and after correcting the Orbscan II readings are shown in Table 6.

Figure 2
Figure 2:
Scatterplots of the agreement between Orbscan II and US CCT measurements in the thin group (A), medium group (B), thick group (C), and all eyes (D). The middle line denotes the mean differences, and the lines above and below indicate the 95% LoA.
Figure 2
Figure 2:
(continued)
Figure 2
Figure 2:
(continued)
Figure 2
Figure 2:
(continued)
Figure 3
Figure 3:
Scatterplots of the agreement between corrected Orbscan II and US CCT measurements in the thin group (A), medium group (B), thick group (C), and all eyes (D). The Orbscan data in each group is corrected with its specific acoustic equivalent correction factor (Table 4). The middle line denotes the mean differences, and the side lines indicate the 95% LoA.
Figure 3
Figure 3:
(continued)
Figure 3
Figure 3:
(continued)
Figure 3
Figure 3:
(continued)
Table 6
Table 6:
The 95% LoA for US and Orbscan II CCT measurements in each group, before and after applying acoustic equivalent correction factors (Table 4).

DISCUSSION

Studies of the earlier versions of the Orbscan report overestimation of corneal thickness compared with US measurements,5,11,14,18 just as we found a mean overestimation of 18 μm in the total group, for which the following equation was computed: (0.77 × Orbscan II) + 110.70. Based on such findings, the Orbscan II and acoustic equivalent correction factor were introduced and the value recommended by the manufacturer was 0.92. Most subsequent studies found underestimations by this system when the correction factor was used4,21,28,29 and concluded that the interdevice agreement is not acceptable and the devices should not be used interchangeably.17,19,28,30–34 However, some based their judgment on mean differences and correlation coefficients,14,35,36 reported no significant interdevice differences,6,31,35 and accepted the validity of the Orbscan corneal thickness measurements.14,21,35,36 In addition, several acoustic equivalent correction factors for improving the validity of the Orbscan pachymetry readings have been suggested.9,14,31,32

In this study, we attempted to investigate this matter further and assess the effect of corneal thickness on interdevice differences. Although it has been mentioned19,28 (H. Hashemi, MD, S. Mehravaran, MD, “Central Corneal Thickness Measurement with Pentacam, Orbscan II, and Ultrasound Before and After Laser Refractive Surgery for Myopia,” presented at the XXIV Congress of the European Society of Cataract & Refractive Surgeons, London, United Kingdom, September 2006), to our knowledge, there are no peer-reviewed published reports concerning such an effect. For this purpose, we divided our sample into groups of different CCT ranges and found that the mean interdevice difference was not significant in the thin group but increased in thicker corneas. Naturally, correcting measurements in the thin group with an equation proper for the thick group would lead to great underestimations. This may, to some extent, explain the Orbscan II underestimations after ablative procedures, as reported in the above-mentioned presentation.

In addition, we determined the interdevice agreement in each group by computing the ICC and 95% LoA; results were indicative of lower agreement in measurements of thicker corneas. However, the range of data can affect the ICC. In this study, the ICC for the total may have been overrated because of the wide range in the total sample population, while in the group of thick corneas, the low ICC may be attributable to the small range of values.

To compare levels of agreement expressed in 95% LoA, it would be easier to consider the width of these limits. For example, according to the information in Table 6, the width of the 95% LoA was 62.2 μm (−32.7 to 29.5) in thin corneas, 79.1 μm (−18.2 to 60.9) in medium corneas, and 82.1 μm (−13.8 to 68.3) in thick corneas. This means the agreement is worst with thick corneas and the difference between the Orbscan II CCT measurements and their US pairs may range from an underestimation of 13.8 μm to an overestimation of 68.3 μm in 95% of cases. The remaining 5% may even have differences beyond this range. Such differences are not acceptable because a difference as little as 10 μm can lead to a different treatment plan.

In the next step, the correction factor for each range of thickness was calculated and applied to the Orbscan data and the analyses were repeated with corrected data. As expected, the mean difference in all 3 groups changed to near zero and the levels of agreement improved. To be more precise, the width of the 95% LoA decreased as much as 25 μm in thin corneas, 9 μm in medium corneas, and 17 μm in thick corneas, leaving widths of 37 μm, 70 μm, and 65 μm, respectively. One could reason that the agreement did not acceptably improve. However, a practical advantage of using correction factors is that the ranges of expected differences lie symmetrically along the zero axis. For instance, if the CCT reading is 550 μm and we use the bottom equation in Table 4, there is a 95% chance that the acoustic equivalent is 550 ± 35 μm because the 95% LoA for this correction factor are −35 μm and 35 μm.

Overestimations and underestimations were seen not only in the total sample but also in each separate group. However, rates varied in different groups; the Orbscan II underestimated the CCT in 19.8% of the total 177 corneas, 61.0% in the thin group, 11.0% in the medium group, and 9.7% in the thick group. Applying a correction factor evens out the distribution (ie, evens rates of overestimation and underestimation) rather than reduces the variability. This means a simple linear transformation is not sufficient to achieve interchangeable data.

Further studies with more samples in the higher and lower ranges of corneal thickness are needed to investigate this issue in more detail, as are studies of the accuracy of US pachymeters. In the meantime, these 2 devices should not be used interchangeably and when the accuracy of the measurement is vital or a reading is borderline for proceeding with a particular procedure, it is advisable to recheck the corneal thickness with a US pachymeter.

REFERENCES

1. Marsich MM, Bullimore MA. The repeatability of corneal thickness measures. Cornea. 2000;9:792-795.
2. Bechmann M, Thiel MJ, Neubauer AS, et al. Central corneal thickness measurement with a retinal optical coherence tomography device versus standard ultrasonic pachymetry. Cornea. 2001;20:50-54.
3. Javaloy J, Vidal MT, Villada JR, et al. Comparison of four corneal pachymetry techniques in corneal refractive surgery. J Refract Surg. 2004;20:29-34.
4. Iskander NG, Anderson Penno E, Peters NT, et al. Accuracy of Orbscan pachymetry measurements and DHG ultrasound pachymetry in primary laser in situ keratomileusis and LASIK enhancement procedures. J Cataract Refract Surg. 2001;27:681-685.
5. Chakrabarti HS, Craig JP, Brahma A, et al. Comparison of corneal thickness measurements using ultrasound and Orbscan slit-scanning topography in normal and post-LASIK eyes. J Cataract Refract Surg. 2001;27:1823-1828.
6. Gherghel D, Hosking SL, Mantry S, et al. Corneal pachymetry in normal and keratoconic eyes: Orbscan II versus ultrasound. J Cataract Refract Surg. 2004;30:1272-1277.
7. Drexler W, Baumgartner A, Findl O, et al. Submicrometer precision biometry of the anterior segment of the human eye. Invest Ophthalmol Vis Sci. 1997;38:1304-1313.
8. Kawana K, Tokunaga T, Miyata K, et al. Comparison of corneal thickness measurements using Orbscan II, non-contact specular microscopy, and ultrasonic pachymetry in eyes after laser in situ keratomileusis. Br J Ophthalmol. 2004;88:466-468.
9. Solomon OD. Corneal indentation during ultrasonic pachometry. Cornea. 1999;18:214-215.
10. Tam ES, Rootman DS. Comparison of central corneal thickness measurements by specular microscopy, ultrasound pachymetry, and ultrasound biomicroscopy. J Cataract Refract Surg. 2003;29:1179-1184.
11. Yaylali V, Kaufman SC, Thompson HW. Corneal thickness measurements with the Orbscan Topography System and ultrasonic pachymetry. J Cataract Refract Surg. 1997;23:1345-1350.
12. Nissen J, Hjortdal J.Ø, Ehlers N, et al. A clinical comparison of optical and ultrasonic pachometry. Acta Ophthalmol (Copenh). 1991;69:659-663.
13. Samapunphong S, Azar D. Placido- and elevation-based corneal topography; a review. Ophthalmol Clin North Am. 1998;11(3):311-329.
14. Liu Z, Huang AJ, Pflugfelder SC. Evaluation of corneal thickness and topography in normal eyes using the Orbscan corneal topography system. Br J Ophthalmol. 1999;83:774-778.
15. Lackner B, Schmidinger G, Pieh S, et al. Repeatability and reproducibility of central corneal thickness measurement with Pentacam, Orbscan, and ultrasound. Optom Vis Sci. 2005;82:892-899.
16. Fam H-B, Lim K-L, Reinstein DZ. Orbscan global pachymetry; analysis of repeated measures. Optom Vis Sci. 2005;82:1047-1053.
17. Jonuscheit S, Doughty MJ. Regional repeatability measures of corneal thickness: Orbscan II and ultrasound. Optom Vis Sci. 2007;84:52-58.
18. Módis L Jr, Langenbucher A, Seitz B. Corneal thickness measurements with contact and noncontact specular microscopic and ultrasonic pachymetry. Am J Ophthalmol. 2001;132:517-521.
19. González-Méijome JM, Cerviño A, Yebra-Pimentel E, Parafita MA. Central and peripheral corneal thickness measurement with Orbscan II and topographical ultrasound pachymetry. J Cataract Refract Surg. 2003;29:125-132.
20. Cheng ACK, Tang E, Mohamed S, Lam DSC. Correction factor in Orbscan II in the assessment of corneal pachymetry. Cornea. 2006;25:1158-1161.
21. Rainer G, Petternel V, Findl O, et al. Comparison of ultrasound pachymetry and partial coherence interferometry in the measurement of central corneal thickness. J Cataract Refract Surg. 2002;28:2142-2145.
22. Kawana K, Miyata K, Tokunaga T, et al. Central corneal thickness measurements using Orbscan II scanning slit topography, noncontact specular microscopy, and ultrasonic pachymetry in eyes with keratoconus. Cornea. 2005;24:967-971.
23. Haque S, Simpson T, Jones L. Corneal and epithelial thickness in keratoconus: a comparison of ultrasonic pachymetry, Orbscan II, and optical coherence tomography. J Refract Surg. 2006;22:486-493.
24. Cheng ACK, Rao SK, Tang E, Lam DSC. Pachymetry assessment with Orbscan II in postoperative patients with myopic LASIK. J Refract Surg. 2006;22:363-366.
25. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;1:307-310.
26. Fleiss JL, Cohen J. The equivalence of weighted kappa and the intraclass correlation coefficient as measures of reliability. Educ Psychol Meas. 1973;33:613-619.
27. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977;33:159-174.
28. Giraldez Fernandez MJ, Diaz Rey A, Cerviño A, Yebra-Pimentel E. A comparison of two pachymetric systems: slit-scanning and ultrasonic. CLAO J. 2002;28:221-223.
29. Prisant O, Calderon N, Chastang P, et al. Reliability of pachymetric measurements using Orbscan after excimer refractive surgery. Ophthalmology. 2003;110:511-515.
30. Cho P, Cheung SW. Central and peripheral corneal thickness measured with the TOPCON specular microscope SP-2000P. Curr Eye Res. 2000;21:799-807.
31. Suzuki S, Oshika T, Oki K, et al. Corneal thickness measurements: scanning-slit corneal topography and noncontact specular microscopy versus ultrasonic pachymetry. J Cataract Refract Surg. 2003;29:1313-1318.
32. Hashemi H, Yazdani K, Mehravaran S, Fotouhi A. Central corneal thickness measurement with Orbscan II and an ultrasound pachymeter. J Iran Soc Ophthalmol. 2004;4(17):1-9.
33. Fishman GR, Pons ME, Seedor JA, et al. Assessment of central corneal thickness using optical coherence tomography. J Cataract Refract Surg. 2005;31:707-711.
34. Basmak H, Sahin A, Yildirim N. The reliability of central corneal thickness measurements by ultrasound and by Orbscan system in schoolchildren. Curr Eye Res. 2006;31:569-575.
35. Fakhry MA, Artola A, Belda JI, et al. Comparison of corneal pachymetry using ultrasound and Orbscan II. J Cataract Refract Surg. 2002;28:248-252.
36. Wong A.C.-M, Wong C-C, Yuen N.S.-Y, Hui S-P. Correlational study of central corneal thickness measurements on Hong Kong Chinese using optical coherence tomography, Orbscan and ultrasound pachymetry. Eye. 2002;16:715-721.
© 2007 by Lippincott Williams & Wilkins, Inc.