Scheimpflug tomography has become a widely employed technique in the early diagnosis of corneal ectasia, because it can provide accurate data of corneal elevations for the entire anterior segment.1 The elevation data are obtained by comparing the reconstruction of the anterior or posterior corneal surface to the best-fit sphere (BFS). The typical dimension of BFS is 8 mm in diameter and the curvature of BFS increases with expansion of its dimension.1 In addition to the elevation maps, the Scheimpflug system also provides several standardized indices for ectasia detection including the Belin/Ambrósio Enhanced Ectasia display (BAD-D).
As reported previously, Chinese patients have a smaller corneal diameter (corneal diameter) than White patients.2 The distributions of corneal diameter in Chinese patients as measured by Scheimpflug tomography were 9% to 13% for a corneal diameter of 11 mm or less, 75% to 78% for a corneal diameter of 11 to 12 mm, and 9% to 16% for a corneal diameter of 12 mm or greater.3 For eyes with relatively small (eg, ≤ 11 mm) or large (eg, ≥ 12 mm) corneas, the standard dimension of 8 mm in diameter may be inadequate for the evaluation of corneal ectasia. To date, a limited number of studies have concentrated on the relationship of corneal diameter and the Pentacam (Oculus Optikgeräte) ectasia detection indices, and their sample size was small.4 To address this issue, we set out in this prospective study to investigate the Pentacam Scheimpflug tomography findings in a group of Chinese patients with different corneal diameters.
Patients and Methods
Patients
Participants of this prospective cross-sectional study were consecutively recruited from those visiting the Ophthalmology Department of Eye and ENT Hospital of Fudan University (Shanghai, China) for correction of myopia and myopic astigmatism between June and December 2019. Inclusion criteria were as follows: patients with unremarkable slit-lamp examination and normal tomography (ABCD classification system, stage 0: ARC [anterior radius of curvature for a 3-mm zone centered on the thinnest point] > 7.25 mm [< 46.5 diopters [D]], posterior radius of curvature [PRC] for a 3-mm zone centered on the thinnest point > 5.9 mm, thinnest pachymetry [TP] > 490 µm, best documented visual acuity ≥ 1.0, and absence of corneal scarring) in both eyes. Those cases with any pathological ocular conditions or relevant systemic diseases were excluded. Wearing soft contact lenses was discontinued for at least 7 days prior to the examination, and rigid or hybrid contact lenses were discontinued for a minimum period of 3 weeks. The horizontal corneal diameter was measured for each patient (Pentacam). The study participants were grouped according to their corneal diameter.
This study was approved by the Ethics Committee of the Eye and ENT Hospital of Fudan University (Shanghai, China), and was conducted in compliance with tenets of the Declaration of Helsinki. Written informed consent was obtained from all participants prior to inclusion in this study.
Corneal Tomography
All patients underwent a comprehensive ophthalmic examination, including the slit-lamp examination, objective and subjective refractions, and Pentacam HR examination. Pentacam imaging was performed by the same experienced examiner (LN) for all participants with three measurements averaged for each individual. Only scans registered as “OK” by the Examination Quality Specification of the instrument were saved and analyzed.
The following corneal descriptors were obtained by the Pentacam software: flat central power (Kf), steep central power (Ks), mean central power [Km = (Kf + Ks)/2], maximum corneal power (Kmax); corneal astigmatism (Ka = Ks − Kf); anterior radius of curvature for a 3-mm zone centered on the thinnest point (ARC); posterior radius of curvature for a 3-mm zone centered on the thinnest point PRC; the BFS for the anterior cornea (BFSa), and the BFS for the posterior cornea (BFSp); TP; front and back corneal elevations at TP (FE and BE); pachymetric progression index (PPI, minimum, average, and maximum); Ambrósio’s maximum relational thickness index (ARTmax); normalized indices: deviation of normality of the front elevation (Df), deviation of normality of the back elevation (Db), deviation of normality of pachymetric progression (Dp), deviation of normality of corneal thinnest point (Dt), deviation of normality of relational thickness (Da), and overall deviation of normality (BAD-D). The analyzing dimensions for these indices are all 8 mm in diameter.
PPI represents the change in corneal thickness from TP to periphery and can be calculated over all 360 degrees of the cornea. The average of these meridians is represented as PPIavg, whereas the meridian with maximal pachymetric increase is PPImax, and minimal pachymetric increase is PPImin. On request, the manufacturer informed that the deviation-based indices can be classified by the software as normal (< 1.6 standard deviation [SD] from the population mean, shown in white), suspicious (> 1.6 SD and < 2.6 SD, highlighted in yellow), and pathologic (> 2.6 SD, highlighted in red), according to data reported by Ambrósio et al,5 and this scheme was followed throughout the current study. FE was classified as normal (< 5.01), suspicious (≥ 5.01 and < 7.14), and pathologic (≥ 7.14), and BE (for myopia only) was classified as normal (< 11.77), suspicious (≥ 11.77 and < 16.42), and pathologic (≥ 16.42). For PPI and ARTmax, each observation was categorized as normal or abnormal according to the cutoff values reported by the manufacturer or Feng et al6 and Ambrósio et al5: PPImin (abnormal: ≥ 0.79); PPIavg (abnormal: ≥ 1.15); PPImax (abnormal: ≥ 1.44); and ARTmax (abnormal: ≤ 313). Vector analysis was performed to decompose Ka into two cross cylinder components, J0 and J45. J0 and J45 were defined as follows: J0 = − (Ka/2) cos (2 α); J45 = − (Ka/2) sin (2α), where Ka is the magnitude of cylinder power and α represents the axis in radians. J0 denotes the horizontal/vertical component of astigmatism, and J45 is astigmatism with angles of 45° and 135°.7
Statistical Analysis
Statistical analysis was performed using SPSS 13.0 software (IBM Corporation). Comparisons between three corneal diameter–based groups were done using one-way analysis of variance. Linear regression analyses were undertaken for indices. The percentages of abnormality between groups were compared by Pearson’s chi-square test. A P value of less than .05 was considered statistically significant.
Results
The right eyes of 305 men (47.4%) and 338 women (52.6%) were included. The mean age was 26.6 ± 6.6 years (range: 18 to 50 years). The number of each corneal diameter group was 11 mm or less: n = 206; 11 to 12 mm: n = 219; and 12 mm or greater: n = 218. The mean values and ranges of the spherical and cylindrical errors at the spectacle plane for each group are shown in Table 1. There were no significant differences in the distributions of age, sphere, and cylinder between each group.
 |
Table 1: Demographic, Topographic, and Pentacam Corneal Descriptors in 3 Groups and Regression Analyses for the Pentacam Descriptors With Respect to Corneal Diameter |
Corneal Indices
Mean values for every Pentacam corneal descriptor and the results of regression analyses with respect to corneal diameter are summarized in Table 1. The mean values of corneal powers (Kf, Ks, Km, and Kmax) and TP were minimum in eyes with a corneal diameter of 12 mm or greater (analysis of variance, P < .05). The corneal powers and TP were negatively correlated with corneal diameter (linear regression analysis, P < .001). Both the ARC and PRC were positively correlated with corneal diameter. The scatter plots for the distributions of Km and TP with respect to corneal diameter are illustrated in Figure A (available in the online version of this article). The corneal astigmatism (Ka) was positively correlated with corneal diameter (linear regression analysis, P < .001, Figure A). Both the difference between anterior BFS and ARC (BFSa-ARC) and the difference between posterior BFS and PRC (BFSp-PRC) were negatively correlated with corneal diameter (linear regression analysis, P < .001). Both FE and BE were negatively correlated with corneal diameter (linear regression analysis, P < .001). The scatter plots for the distributions of FE and BE are shown in Figure B (available in the online version of this article). The results of classifications (normal, suspect, and abnormal) for FE and BE are summarized in Table 2. The differences in the classifications between corneal diameter groups were statistically significant for BE (Pearson’s chi-square test = 59.89, P < .001), but not for FE (Pearson’s chi-squared test = 4.61, P = .10).
 |
Figure A. Scatter plots of the distribution of mean central corneal power (Km), the thinnest pachymetry (TP) and the corneal astigmatism (Ka) with respect to corneal diameter (CD). D = diopters |
 |
Figure B. Scatter plots of the distributions of front elevation and back elevation at the thinnest point (TP) with respect to corneal diameter (CD). |
 |
Table 2: Results of Classification for Front and Back Elevations and Pentacam Keratoconus Indices |
The normalized indices were negatively correlated with corneal diameter except for Dt. Additionally, Df, Db, Dp, Da, and D were all maximum in eyes with a corneal diameter of 11 mm or less, whereas Dt was maximum in eyes with a corneal diameter of 12 mm or greater (Table 1). The scatter plots for the distributions of all normalized indices are depicted in Figure C (available in the online version of this article). Among the six indices, corneal diameter had the greatest influence on Db (R2 = 0.47, P < .001) and BAD-D (R2 = 0.32, P < .001). The results of classifications for each index are shown in Table 2. The differences in the classifications between corneal diameter groups were statistically significant except for Dt. In the 11 mm or less group, 62 eyes (30.1%) had normal values for every D index. In the 11 to 12 mm group, 141 eyes (64.4%) had normal values for every D index. In the 12 mm or greater group, 190 eyes (87.2%) had normal values for every D index (Pearson’s chi-square test = 146.60, P < .001).
![Scatter plots of the distributions of the 6 normalized indices (deviation of normality of the front elevation [Df], deviation of normality of the back elevation [Db], deviation of normality of pachymetric progression [Dp], deviation of normality of corneal thinnest point [Dt], deviation of normality of relational thickness [Da], and Belin/Ambrósio Enhanced Ectasia display [BAD-D]) with respect to corneal diameter (CD).](https://m4.healio.com/~/media/journals/jrs/2020/10_october/10_3928_1081597x_20200730_02/fig3.jpg) |
Figure C. Scatter plots of the distributions of the 6 normalized indices (deviation of normality of the front elevation [Df], deviation of normality of the back elevation [Db], deviation of normality of pachymetric progression [Dp], deviation of normality of corneal thinnest point [Dt], deviation of normality of relational thickness [Da], and Belin/Ambrósio Enhanced Ectasia display [BAD-D]) with respect to corneal diameter (CD). |
All three PPI indices (minimum, average, and maximum) were maximum in eyes with a corneal diameter of 11 mm or less (Table 1), and were negatively correlated with corneal diameter (linear regression analysis, P < .001). ARTmax was the lowest in eyes with a corneal diameter of 11 mm or less, and was positively correlated with corneal diameter (linear regression analysis, P < .001). The scatter plots for the distributions of three PPIs and ARTmax are displayed in Figure D (available in the online version of this article). Among the four indices, corneal diameter had the greatest influence on PPImin (R2 = 0.164, P < .001), followed by PPIavg (R2 = 0.158, P < .001). The results of classifications according to each index are summarized in Table 2. The differences in the results of classifications were statistically significant (Pearson’s chi-square test, P < .05) for all indices. In the 11 mm or less group, 69 eyes (33.5%) had normal values for every PPI. In the 11 to 12 mm group, 106 eyes (48.4%) had normal values for every PPI. In the 12 mm or greater group, 164 eyes (75.2%) had normal values for each PPI (Pearson’s chi-square test, P < .001).
 |
Figure D. Scatter plots of the distributions of pachymetric progression indices minimum (PPImin), maximum (PPImax), and average (PPIavg), and maximum Ambrósio’s relational thickness (ARTmax) with respect to corneal diameter (CD). |
Overall, the numbers and percentages of eyes with normal values for all 10 indices (4 PPIs plus 6.00 Ds) were 39 (18.9%), 89 (40.6%), and 155 (71.1%) for the three groups, respectively (Pearson’s chi-square test, P < .001).
Discussion
Pentacam Scheimpflug tomography provides several indices specifically designed to diagnose ectatic changes. In the current study, we explored the performance of these descriptors in a sample of tomographically normal patients with different horizontal corneal diameters to indicate whether corneal diameter has an influence on these indices and whether it is adequate to evaluate eyes with different corneal diameters with the same analytical zone. The results showed that the following indices, such as corneal powers, TP, FE and BE, PPI (minimum, average, and maximum), Df, Db, Dp, Da, and D, were negatively correlated with corneal diameter. Among the ectatic indices, corneal diameter had the greatest influence on Db, followed by BAD-D and BE. The rates of being flagged as suspect or abnormal were significantly higher in eyes with a corneal diameter of 11 mm or less, especially in some indices (32.5% for Db, 43.7% for Dp, 50.5% for BAD-D, and 42.2% for PPIavg). The BAD indices might have overestimated the risk of ectasia in small corneas. Corneal diameter should therefore be incorporated as an additional variable of BAD display.
To date, a limited number of studies assessed the effect of corneal diameter on anterior segment parameters. In the current research, we found that the corneal powers were negatively correlated with corneal diameter. This is consistent with the findings of Jiang et al.8 We also noted that TP was negatively correlated with corneal diameter, whereas Dt was positively correlated with corneal diameter, indicating that larger cornea could be thinner and weaker. Montard et al9 demonstrated that corneal diameter was negatively correlated with cornea hysteresis and corneal resistance factor, in which both parameters are used in quantitative assessment of the biomechanical properties of cornea. Moreover, a negative correlation was observed between the cornea hysteresis and corneal resistance factor and spur-to-spur distance of ciliary body as reported by Shah et al.10 However, further studies are required to clarify the relationship between corneal size and its biomechanics.
The mean values of FE and BE obtained in the current study were in agreement with others.11–14 We also found that both FE and BE were negatively correlated with corneal diameter. FE or BE represents the maximum elevation in a zone above the standardized reference shape, which is typically a BFS or best-fit toric ellipsoid. Due to its asphericity, the cornea can flatten gradually toward the periphery.15 The curvature may become flat earlier in corneas with a small diameter. As a result, a BFS may have an increased radius of curvature (flatter power) in eyes with a small corneal diameter. It was proved by the increased BFSa-ARC and BFSp-PRC in the 11 mm or less group. On the other hand, although both FE and BE were negatively correlated with corneal diameter, the R2 values were markedly higher for BE than FE. In the 11 mm or less group, the rate of abnormal BE was 20.9% and the rate of abnormal FE was 2.9%. It seems that corneal diameter had a greater influence on BE than on FE. One possible explanation is that PRC is smaller than the ARC, which means the posterior cornea was steeper than the anterior cornea. As a consequence, PRC may become flat more quickly than ARC toward the limbus, especially in small corneas. Moreover, the increased PPI in the 11 mm or less group means that small corneas may have a higher rate of thickening toward the periphery, which also suggests that the posterior cornea may become flat more quickly than the anterior cornea, especially in small corneas. Consequently, corneal diameter might have a greater impact on BE.
In contrast to FE and BE, focalized corneal thinning seems to typify several of the earliest changes in keratoconic eyes. The Pentacam software provides several metrics designed to detect these ectatic signs: pachymetric progression and Ambrósio’s relational thickness indices. In the current study, we found that the PPI indices were negatively correlated with corneal diameter. Because PPIs represent the rate of change in corneal thickness from the thinnest point to the periphery, eyes with a smaller corneal diameter may have a higher rate of change because the overall distance between the thinnest point and periphery is less. Among the four indices, corneal diameter had the greatest influence on PPImin, followed by PPIavg. In a study by Boyd et al,4 corneal diameter had the greatest impact on PPIavg. The positive correlation between ARTmax and corneal diameter also indicated that corneal thickness changes faster in smaller corneas.
The BAD display uses both elevation data and pachymetric data to screen ectatic changes.16,17 It comprises nine different indices, in which the final D (BAD-D) is calculated based on regression analysis of those nine indices, and has been shown in multiple studies to have the highest accuracy in detecting both clinical keratoconus and pre-keratoconus.18,19 Our study noted that five normalized indices (Df, Db, Dp, Da, and BAD-D) were negatively correlated with corneal diameter, and corneal diameter had the greatest influence on Db and final D. Our result about final D was consistent with that of Boyd et al.4 Because BAD-D is one of the best Pentacam indices in identifying both definitive and pre-keratoconus,18,19 it may become more appropriate if corneal diameter can be incorporated as one of its valuables in the future.
It is noteworthy that a sizable number of cases are flagged as suspect or abnormal by the Pentacam analysis, who are otherwise considered to be normal by clinical examinations (ABCD classification, stage 0). The rates of suspect or abnormal cases were significantly higher in eyes with a corneal diameter of 11 mm or less, especially for some indices (32.5% for Db, 43.7% for Dp, 50.5% for final D, and 42.2% for PPIavg). It has been reported that no significant difference was found between normal and keratoconic eyes with respect to the horizontal corneal diameter.20,21 The significant differences in the rate of abnormality between different corneal diameter–based groups should be taken into consideration during the screening of refractive surgical candidates, and additional testing should be applied to these patients (eg, corneal biomechanical testing and epithelial thickness analysis) and can be used with Scheimpflug tomography.22–24
Besides the BAD parameters, several systems have also been used in detecting early ectasia such as the Screening Corneal Objective Risk of Ectasia Analyzer (SCORE, a software linked to the Orbscan topography system; Bausch & Lomb).25–27 The SCORE Analyzer combines 12 Placido and tomographic indices in a weighted fashion to classify corneas as suspicious for keratoconus or normal. These variables include the TP, the difference between the central pachymetry and TP, thinnest point decentration, the difference between inferior and superior keratometry, posterior elevation of the thinnest point, the 3-mm irregularity, and data derived from the pachymetry thinning rate. Chan et al26 investigated the efficacy of the SCORE Analyzer in detecting forme fruste keratoconus in Asian eyes and concluded that no discriminant function adjustments for this system are required for the Asian group of patients. However, the corneal diameter was not incorporated as a variable in the SCORE Analyzer either. Considering the influence of corneal diameter on the posterior elevation and the pachymetry progression rate that was proved in our study, it is meaningful to study the influence of corneal diameter on the other keratoconus detection systems.
In the current study, Pentacam HR was used to measure the corneal diameter. Pentacam measures the horizontal corneal diameter with an iris camera optic that can recognize iris landmarks and determine the pupil location and shape. It has been found to have good agreement with other devices (eg, Orbscan) and has been widely used to determine the size of phakic posterior chamber intraocular lenses.28,29 The main limitation of this study is that the corneal diameter was measured by only one type of instrument. Because all study participants were measured by the same device, the results could be used for comparison. Another limitation of the study is that the inclusion criteria were based mainly on the ABCD classification system, which has yet to be adopted in a widespread fashion. According to the recently published Global Consensus on Keratoconus and Ectatic Diseases30 there is currently no clinically adequate classification system for keratoconus. The most widely used Amsler-Krumeich system fails to recognize any changes other than on the anterior corneal surface. The ABCD classification system incorporates anterior (A) and posterior (B) radius of curvature, thinnest corneal pachymetry (C), and best corrected distance visual acuity (D), as well as an indication of corneal scarring. Several studies have validated its clinical value in early screening. However, this system still fails to incorporate several key indices for detection of pre-keratoconus, such as index of vertical asymmetry, and the vertical thinnest point decentration and may misclassify some suspect cases as normal individuals. Further study using a more comprehensive classification system might make the results more convincing. Moreover, the current study found that the rates of suspect or abnormal cases were significantly lower in eyes with a corneal diameter of 12 mm or greater. Could BAD indices underestimate the risk of ectasia in larger cornea? Because all cases were labeled as normal in the current study, further research is needed to investigate the issue in a population of suspect keratoconus with different corneal diameters.
This study suggests that corneal diameter has an influence on the BAD parameters, especially Db, BE, and BAD-D. Corneal diameter should be incorporated as an additional variable in BAD analysis. For eyes with relatively small or large corneas, the standard analytical dimension of 8 mm may be inadequate. The analytical dimensions should be individualized for eyes with individual corneal diameter. Further study is required to set up a rule for choosing an appropriate analytical dimension for each corneal diameter.
