A total of 64 eyes were enrolled, comprising 24 from normal participants and 40 from patients with dry eye. Mean age of participants was 26.7 ± 2.4 years, with 17 female patients. Demographic characteristics are summarized in Table 1.
Table 1
Total | 64 eyes |
Age (years) | 26.7 ± 2.4 |
Sex (female (number, %)) | 17, 26.5 |
TMH (㎛) | 308.57 ± 77.79 |
BUT (sec) | 5.28 ± 3.59 |
Variation of anterior axial curvature (D) | 26.99 ± 17.59 |
Variation of posterior axial curvature (D) | 2.61 ± 3.96 |
Variation of pachymetry (㎛) | 2.76 ± 5.62 |
BUT, Break-up time ; TMH, tear meniscus height
When dividing the participants into normal and dry eye groups based on a BUT of 5 seconds, there were 24 eyes in the normal group and 40 eyes in the dry eye group, with average BUT of 9.1 ± 2.8, and 3.0 ± 1.1, respectively (p < 0.001). The average TMH was 222.9 ± 73.9 for the normal group and 195.7 ± 79.9 for the dry eye group, which was not statistically significant (p = 0.181). Both the RUV and ICC of anterior axial curvature, posterior axial curvature, and corneal thickness showed no significant difference with p > 0.05. (Table 2, Fig. 2).
Table 2
Results of ratio of unacceptable variation and intraclass correlation in normal and dry eyes when groups with a BUT of less than 5 seconds were classified as having dry eye.
| Normal | Dry Eye (BUT < 5sec) | P value |
Total patients | 24 | 40 | |
BUT (sec) | 9.1 ± 2.8 | 3.0 ± 1.1 | < 0.001 |
TMH (㎛) | 222.9 ± 73.9 | 195.7 ± 79.9 | 0.181 |
Ratio of Unacceptable variation(%) | Anterior Axial Curvature | 6.9 ± 7.9 | 7.9 ± 7.8 | 0.634 |
Posterior Axial Curvature | 0.7 ± 2.0 | 0.5 ± 1.6 | 0.616 |
Pachymetry | 2.7 ± 6.6 | 1.6 ± 2.2 | 0.321 |
Intraclass correlation coefficient | Anterior Axial Curvature | 0.97 ± 0.04 | 0.98 ± 0.02 | 0.396 |
Posterior Axial Curvature | 0.90 ± 0.17 | 0.94 ± 0.10 | 0.355 |
Pachymetry | 0.99 ± 0.02 | 0.99 ± 0.003 | 0.171 |
BUT, Break-up time ; TMH, tear meniscus height
BUT, Break-up time
When classified based on a TMH threshold of 150㎛, there were 29 eyes in the normal group and 35 eyes in the dry eye group, with average BUT of 5.39 ± 2.9, and 5.18 ± 3.9, respectively, showing no significant difference. However, the TMH was 275.1 ± 59.3 for normal group and 148.5 ± 31.4 for the dry eye group, indicating a statistically significant difference (p < 0.001). Even in this classification, both RUV and ICC of anterior axial curvature, posterior axial curvature, and corneal thickness showed no statistically significant differences, with p > 0.05 (Table 3, Fig. 3).
Table 3
Results of ratio of unacceptable variation and intraclass correlation in normal and dry eyes when groups with a TMH of less than 150㎛ were classified as having dry eye.
| Normal | Dry Eye (THM < 150㎛) | P value |
Total patients | 29 | 35 | |
BUT (sec) | 5.39 ± 2.9 | 5.18 ± 3.9 | 0.814 |
TMH (㎛) | 275.1 ± 59.3 | 148.5 ± 31.4 | < 0.001 |
Ratio of Unacceptable variation(%) | Anterior Axial Curvature | 7.1 ± 7.4 | 7.9 ± 8.2 | 0.688 |
Posterior Axial Curvature | 1.0 ± 2.3 | 0.2 ± 0.9 | 0.616 |
Pachymetry | 2.1 ± 2.9 | 1.9 ± 5.4 | 0.102 |
Intraclass correlation coefficient | Anterior Axial Curvature | 0.98 ± 0.02 | 0.98 ± 0.03 | 0.852 |
Posterior Axial Curvature | 0.90 ± 0.15 | 0.94 ± 0.11 | 0.177 |
Pachymetry | 0.99 ± 0.02 | 0.99 ± 0.01 | 0.447 |
BUT, Break-up time ; TMH, tear meniscus height
TMH, tear meniscus height
The repeatability map could be divided into forms with no significant variability, forms where the anterior refractive values showed variability of more than 0.5D but the posterior refractive values and corneal thickness were consistent, forms where values other than the anterior refractive values showed significant variability, and forms where values other than the posterior refractive values showed significant variability (Fig. 4).
RUV, Ratio of unacceptable variation
When plotting the angles and positions from the corneal center of the points where the variability was more than 0.5D or 25um, it was observed that the anterior axial curvature showed a lot of variability even within 3mm of the center. In the case of posterior axial curvature and corneal thickness, the central area was stable, but variations could be seen in the peripheral areas of the upper and lower cornea. Additionally, it can be observed that overall, the upper cornea exhibits greater variability compared to the lower cornea (Fig. 5).
This study found high repeatability in measurements of anterior and posterior axial curvature, and corneal thickness obtained through repeated Galilei anterior segment camera assessments, regardless of corneal location. Variations in tear film parameters such as BUT and TMH did not affect repeatability. Although RUV was low within the central 3 mm of the cornea for posterior axial curvature and corneal thickness, higher variability was noted in anterior axial curvature within the same region.
It is known that Galilei system showed better repeatability than other devices (Orbscan II or Pentacam).6–9 In previous studies, repeatability is typically measured for metrics such as flat keratometry, steep keratometry, or central corneal thickness, which show results localized to the center of the cornea. In this study, using MATLAB, the entire cornea was divided into 18,002 points, and the RUV at each topographically and geographically specific point was measured to assess the repeatability across the whole cornea. While the analysis of the entire cornea showed results that were somewhat consistent with those previously limited to the center in terms of repeatability, it was observed that the RUV was significantly present in the center portion in the case of anterior axial curvature.
Although variability in anterior axial curvature is anticipated to be closely related to changes in the tear film, this study did not observe any significant differences among groups divided based on BUT and TMH. Previous study has shown that the pattern of tear film disruption varies with BUT, where linear and dimple breaks occur in mild dry eyes with BUT of 2–3 seconds, and random breaks occur when BUT exceeds 5 seconds.10 Given the high ratio of RUV observed within 3 mm of the center of the cornea for anterior axial curvature in this study, it could be inferred that this variability may be due to tear film disruption in the central area.
The upper cornea exhibits greater variability compared to the lower cornea. It can be understood that the differences in the position of the upper and lower eyelids, as well as variations in the tear film's volume and quality in their vicinity, may influence the outcomes. According to the study of ocular surface height analysis11, there was an observed increase in tear film height at the superior cornea after blinking. This increase is likely due to the swift upward motion of tears following a blink suggesting a localized thickening of the tear film in this area. It is proposed that immediately after a blink, a reduced thickness of the lipid layer at the superior cornea creates higher surface tension, driving the tear film upwards and thickens the tear film in the upper portion of the cornea.12 The changes in the tear film of the upper cornea following a blink may potentially contribute to greater variability during measurements. It has been observed that the period 1 to 4 seconds after a blink offers the highest reliability and minimal influence from the tear film11, suggesting that considering this timing during examinations could be a strategy to reduce errors.
This study proceeded by randomly selecting one eye from each participant for research. Previous studies7,13,14 have included both eyes of participants, which, while potentially increasing the sample size, introduces a fundamental issue of inter-eye correlation, potentially leading to slight differences in the results.
Previous studies15–18 assessing repeatability of Scheimpflug analyzer, typically involved young and healthy participants, which might not fully align with real-world settings. In our study, we focused on a young patient cohort with an average age of 26.7 years and found favorable repeatability outcomes. It is speculated that repeatability might decrease in older patients due to difficulties with eye fixation and lower cooperation levels. However, according to the findings by Kim et al.19, examining patients aged 56 ± 18 years (ranging from 22 to 87 years) revealed repeatability comparable to that of younger participants, indicating that the Galilei demonstrates strong performance in clinical settings.
Our study has some limitations. First, this study did not analyze variations according to the tear film break-up pattern and excluded patients with severe dry eye of grade 2 or above. Second, TMH was manually measured and it might produce some errors. Also, variability between sessions was not examined, and it is anticipated that this factor would also add to the variability observed in the measurements. However, the comprehensive analysis of the entire cornea using MATLAB on 18,002 points represents a strength of this research, as there have been no prior studies analyzing changes across the whole cornea.
In conclusion, this study demonstrated high repeatability of corneal topography measurements obtained through three repeated measurements using the Galilei, with no significant differences observed based on the BUT and TMH. However, caution should be made when interpreting measurements of anterior axial curvature, especially within 3 mm of the center, due to potential changes exceeding 0.5D. Future research should explore whether variability differs according to the pattern of the tear film disruption, as well as among patients with corneal diseases other than dry eye syndrome, indicating the need for further studies in this area.