Thirty eyes from 15 volunteers were included in this study (13 women; 2 men). The mean age was 26.8±6.8 years (range: 20 to 41 years). The mean refractive error (spherical equivalent) was −1.8±1.9 D (range, −5.00 to +1.75 D). Subjects were excluded if they had active anterior segment disease that might cause an alteration in corneal thickness, a history of ocular surgery, or a prior history of wearing contact lenses. Informed consent was obtained from each subject, and all of the subjects were treated in accordance with the tenets of the Declaration of Helsinki.

Corneal thickness was assessed using the AS-OCT Visante (Carl Zeiss Meditec Inc, Dublin, Calif.), the Orbscan II (Bausch & Lomb, Rochester, NY, version 3.12), and USP (Sonogage, Sonogage Inc., Cleveland, Ohio; calibrated by the manufacturer).

AS-OCT Visante is a non-contact imaging technique that uses infrared light to obtain high-resolution, cross-sectional images in vivo of the cornea and anterior eye (anterior chamber, iris and lens). The use of the AS-OCT Visante has been described previously.14,15 Its software can measure corneal thickness at any desired point. One single scan was performed for each eye using the scan profile protocol (selecting A-scan, which is a particular setting of the AS-OCT Visante software).

Corneal thickness was analyzed in five different corneal locations: at the center of the cornea and at the superior, inferior, nasal, and temporal regions of the tissue 4.0mm from the centre. Two different analyses were conducted for every single scan (one single scan was performed for each eye). First, an automatic analysis was conducted using the Global-Pachymetry Map, which provides the thickness of a patient's cornea automatically, and a second manual analysis of the scan was conducted by three different masked examiners who collected corneal thickness values using the Flap Tool option by placing the cursors of the caliper at the limits of the anterior and posterior corneal surfaces.

The procedure involving the use of the Orbscan II has been described previously.12,13 An experienced operator conducted the Orbscan exploration to obtain central and peripheral (4.0mm from the corneal centre at the superior, inferior, nasal and temporal areas of the tissue) corneal pachymetry (with 95% acoustic factor). One scan was performed for each eye.

Central corneal thickness (CCT) was measured using USP after analyses with the AS-OCT Visante and Orbscan II had been conducted to avoid corneal changes resulting from the direct contact of the probe with the cornea or topical anesthesia. To perform USP, the cornea was anesthetized with a drop of proparacaine hydrochloride (0.5%). The probe was applied at an angle perpendicular to the corneal surface (standard clinical conditions). The same experienced and blinded operator performed all of the USP procedures. Seven measurements were obtained for the central cornea, and the highest and lowest were discarded. The mean of the five remaining measurements was used as the final value.

All of the tests were performed between 14.00 and 20.00 hours to avoid the edematous cornea as well as sleep-induced alterations in corneal thickness. This is the time of day when the eye is most physiologically stable.1

The data were analyzed using the Statistical Package for the Social Sciences (SPSS for Windows software, version 14.0, SPSS, Inc., Chicago, Illinois, USA).

Inter-examiner agreement of the manual values of the AS-OCT Visante was determined using multiple analysis of variance (ANOVA) with Bonferroni correction. A p-value <0.05 was considered statistically significant. The manual and automatic pachymetry obtained from the AS-OCT Visante were compared for central, superior, inferior, nasal, and temporal corneal thickness using multiple analysis of variance (ANOVA) with Bonferroni correction. A p-value <0.05 was considered statistically significant.

The manual and automatic values from the AS-OCT Visante were also compared with the Orbscan II and USP using Student's paired t-test. A p-value <0.05 was considered statistically significant. Limits of agreement were calculated (mean±two standard deviations) as suggested by Bland and Altman.15 Correlations between the automatic values of the AS-OCT Visante, the Orbscan II and USP were determined by linear correlation analysis. The correlation coefficient (r2) was also calculated. A p-value <0.05 was considered statistically significant.

High inter-examiner agreement, without statistically significant differences (p=1.00 ANOVA with Bonferroni correction), was found between the AS-OCT Visante manual values obtained by each one of the three examiners in all of the corneal locations studied (Table 2).

Good agreement between the automatic and the manual values (obtained by each one of the three examiners) from the AS-OCT Visante were found for the superior, nasal, and temporal corneal locations (p>0.05 multiple ANOVA). However, statistically significant differences were found for the CCT and inferior corneal thickness between both (automatic and manual) AS-OCT Visante analyses (p<0.05 ANOVA with Bonferroni correction). The difference between the automatic and the mean manual analyses was 37±10μm for CCT and 27±11μm for inferior corneal thickness. Figure 1 and Table 3 show the difference between the manual and automatic values of the AS-OCT Visante.

Figure 1.

Agreement between manual (examiners #1, #2 and #3) and automatic values from the AS-OCT Visante pachymetry. *Statistically significant differences (p<0.05 multiple ANOVA) were found between manual and automatic values for the central and inferior corneal locations. However, non-statistically significant differences were found between the manual values of the three blinded examiners (p>0.05 multiple ANOVA with Bonferroni correction) in all of the corneal locations. Non-statistically significant differences were found (p>0.05 multiple ANOVA) between the manual and automatic values for the remaining locations.

(0.1MB).

Good linear correlation (r2>0.62) was found for all corneal locations between the AS-OCT Visante automatic values and the values from the Orbscan II and USP; there were statistically significant differences (p<0.01 on Student's paired t-test) for all corneal locations (Table 4).

The manual CCT obtained with the AS-OCT Visante was significantly different (p<0.01 Student's paired t-test) from that obtained with USP. However, no significant differences were found for the Orbscan II CCT (p>0.05 Student's paired t-test). The results of the Orbscan II peripheral pachymetry were significantly different than all of the AS-OCT Visante manual values (p<0.01 Student's paired t-test) (Table 5).

Good inter-examiner agreement was found with no statistically significant differences between the manual values of the AS-OCT Visante (p=1.00 ANOVA with Bonferroni correction) for the three masked examiners. This result is in accordance with previous reports that examined CCT 18,22–24 and are the first to describe the agreement of the manual values in the periphery of the cornea. The inter-examiner agreement obtained in this study suggests that the device is acceptable for clinical use.

The differences between the automatic and the manual values for the central24 and inferior corneal location suggest that automatic and manual analyses cannot be used interchangeably. However, the superior, nasal, and temporal areas did not show significant differences between the automatic and manual values, although the manual values were higher than the automatic values, with the exception of the temporal corneal location (all examiners) and the superior location (only the examiner #1 values). Li et al.24 noticed that the anterior corneal limit delineated by the AS-OCT Visante software was positioned slightly under the anterior corneal surface, which could lead to the underestimation of corneal thickness when measured automatically. Furthermore, Dunne et al.25 have presented improved formulas for the AS-OCT Visante software in order to get better measures for axial distances and surface curvature. We agree that one of the reasons for these differences could be related to the software, which analyzes the pachymetry automatically (Global-Pachymetry Map), as we used the same scan and it was found to have good agreement with no differences between the manual values of the three masked examiners for each one of the five corneal locations studied.

The differences in the CCT between the AS-OCT Visante (both manual and automatic analysis) and USP found in this study are consistent with previous studies in the literature,2,5,10,20,24,25 although there was no discrimination between the manual or automatic values in those reports. Automatic values for the AS-OCT Visante were significantly lower than those for USP; however, the manual values (collected by three different examiners) were significantly higher than those from USP (Tables 4 and 5).

We found a similar trend with Orbscan II pachymetry, with significant differences both in central (in accordance with the previous studies focused on CCT27) and peripheral locations (i.e., superior, inferior, nasal and temporal). The automatic values of the AS-OCT Visante were significantly lower than those taken with the Orbscan II in the central area, but higher in the peripheral locations (Table 4). Between the manual values of the AS-OCT Visante and the Orbscan II, no statistically significant differences were found for CCT (p>0.05 multiple ANOVA); however, values for peripheral pachymetry were always higher (p<0.05 multiple ANOVA) when measured with the AS-OCT Visante than with the Orbscan II (Table 5).

One of the main limitations of this study could be related with the use of both eyes of the same subject to assess agreement between the manual and the automatic analysis of AS-OCT Visante; although it has been previously reported several studies where both eyes of the same subject have been used to determine agreement between different devices.28–30 Another possible limitation is the fact that we have performed one single scan when using the Orbscan II, although it has been widely reported the high repeatability of this device.14,31 The small sample data suggest that these results should be interpreted with caution. More studies in non-healthy corneas should also be conducted to determine the agreement between the manual and the automatic analysis.