Agreement of Total Keratometry and Posterior Keratometry Among IOLMaster 700, CASIA2, and Pentacam

Abstract

Purpose: The purpose of this study was to compare total keratometry (TK) and posterior keratometry (PK) obtained by two swept-source optical biometers (IOLMaster 700 and CASIA2) and one Scheimpflug-based topography (Pentacam AXL).

Methods: The TK and PK in cataract surgery candidates obtained by IOLMaster 700, CASIA2, and Pentacam AXL were compared. Intraclass correlation coefficients (ICCs), limit of agreement, and Bland-Altman plots were used to assess the agreement.

Results: One hundred two patients with a mean age of 68.21 ± 8.70 years were included. There were significant differences among IOLMaster 700, CASIA2, and Pentacam AXL in the mean TK (TK_m) (44.23 ± 1.59 diopters [D] vs. 43.25 ± 1.53 D vs. 43.94 ± 1.68 D; all P < 0.001), mean PK (PK_m; −5.90 ± 0.24 D vs. −6.25 ± 0.25 D vs. −6.37 ± 0.26 D; all P < 0.001) and TK-J0 (−0.34 ± 0.65 D vs. −0.23 ± 0.53 D vs. −0.12 ± 0.62 D; all P < 0.001). We also observed significant differences in PK-J45 between IOLMaster 700 and Pentacam AXL as well as between CASIA2 and Pentacam AXL (both P < 0.001). There was a good agreement in TK_m, TK-J0, TK-J45, and PK-J0 (ICC = 0.887, 0.880, 0.751, and 0.807, respectively), a moderate agreement in PK-J45 (ICC = 0.626), and a poor agreement in PK_m (ICC = 0.498) among these 3 biometers.

Conclusions: TK, PK, and the corresponding astigmatism obtained by IOLMaster 700, CASIA2, and Pentacam AXL showed significant differences, and could not be used interchangeably.

Translational Relevance: Our study may help to guide preoperative keratometry measurement for intraocular lens (IOL) power calculation and astigmatism evaluation for patients with cataract.

Introduction

Accurate measurement of keratometry is a critical step for intraocular lens (IOL) power calculation and astigmatism evaluation in cataract surgery.¹ The simulated keratometry (SimK), which is calculated based on the anterior corneal curvature and a standard keratometric index (usually 1.3375), has been widely used for IOL power calculation.^2,3 The use of the standard keratometric index is based on the assumption that there is a fixed linear relationship between the anterior and posterior corneal curvatures. However, the anterior/posterior corneal curvature ratio varies in different eyes.⁴ Therefore, ignoring the true posterior keratometry (PK) and corneal thickness could result in inaccurate estimations of total keratometry (TK) and total corneal astigmatism (TCA).¹ Compared to SimK, TK achieves higher prediction accuracy in patients with prior corneal refractive surgery or keratoconus, and showed comparable or better refractive results for patients with routine cataract.^5–7 In addition, for toric IOL power calculation, ignoring posterior corneal astigmatism (PCA) may lead to overcorrection in eyes having with-the-rule astigmatism and under-correction in eyes having against-the-rule astigmatism.^1,8,9 

TK can be calculated based on the anterior and posterior corneal curvature and true corneal thickness using the thick lens formula.⁵ With advances in technology, many biometers are available for PK measurement, including the IOLMaster 700 (Carl Zeiss Meditec AG, Jena, Germany), CASIA2 (Tomey Corporation, Nagoya, Japan), Pentacam AXL (Oculus, Wetzlar, Germany), Anterion (Heidelberg Engineering, Heidelberg, Germany), and Orbscan (Bausch & Lomb, Rochester, NY). The measuring principles of different biometers are varied. The IOLMaster 700, one of the most widely used optical biometers for IOL power calculation, measures the anterior corneal radius using a telecentric keratometry, and simultaneously measures the posterior corneal radius and central corneal thickness using the swept-source optical coherence tomography (SS-OCT) technology.^3,10 Whereas the CASIA2, as a newly introduced three-dimensional anterior segment OCT (AS-OCT), measures keratometry with a faster scanning speed (50,000 axial scans/s), wider scanning range (16 × 16 × 13 mm), and higher imaging resolution.^4,11 The Pentacam AXL is another widely used device for keratometry assessment in clinical practice, which uses a 360-degree rotating Scheimpflug camera to measure both the anterior and posterior corneal surfaces, and then the total corneal refractive power (TCRP) is calculated using ray tracing and Snell's law of refraction.^3,10,12 Differences in measurement principles, areas, and image processing methods of the above three biometers may result in discrepancies in the corneal measurement values. The agreement of keratometry measurements obtained by different biometers is inconsistent in previous studies. Some studies have reported good agreement in SimK and TK measurements among varied biometers, whereas significant differences in TCA and PK were observed in other studies.^{3,10,11,13,14} 

Given the vital role of TK and PK in IOL power calculation and astigmatism evaluation for patients with cataract, especially for those with prior refractive surgery, keratoconus, and significant corneal astigmatism, the agreement and interchangeability of TK and PK by different biometers need further investigation. Therefore, in this study, we aimed to comprehensively assess and compare the agreement of TK, PK, and corresponding astigmatism measurements obtained by IOLMaster 700, CASIA2, and Pentacam AXL. 

Methods

This cross-sectional study was performed at the Zhongshan Ophthalmic Center (ZOC) of Sun Yat-sen University, Guangzhou, China. All procedures of this study were in accordance with the tenets of the Declaration of Helsinki and were conducted with the approval of the Institutional Review Board/Ethics Committee of ZOC (2019KYPJ033). All participants were given detailed explanations and provided written informed consents before entering the study. 

Participants

Patients scheduled for cataract surgery at ZOC from November 2021 to April 2022 were recruited for this study. The exclusion criteria included: (1) with dry eye, pterygium, keratoconus, or other ocular diseases; (2) history of contact lenses wearing in the past 3 months; and (3) history of ocular surgery. 

Instruments and Measurements

Before cataract surgery, all participants underwent a full anterior and posterior segment slit-lamp examination, followed by ocular biometric measurements using IOLMaster 700, CASIA2, and Pentacam AXL in a dark room. All these 3 biometers were calibrated on a daily basis, and related measurements for the same patient were performed within 30 minutes per standardized protocol. Artificial tear drops were not administered before or during the examination, and the examiners would assist in separating the eyelid of the participant when necessary. Only high-quality measurements were included in the final analysis. For any case with a low-quality warning by the device, the measurement was repeated until a high-quality image was obtained, or until a maximum number of three attempts had been made. The following parameters were obtained from these three biometers, including flat simulated keratometry (SimK_f), steep simulated keratometry (SimK_s), flat total keratometry (TK_f), steep total keratometry (TK_s), flat posterior keratometry (PK_f), and steep posterior keratometry (PK_s). Regarding TK, the real keratometry of CASIA2 and total corneal refractive power (TCRP) of Pentacam AXL were analyzed. 

Astigmatism Calculation

In order to compare corneal astigmatism, J0 and J45 vectors were calculated using the following formulas, respectively¹⁵:  

\begin{eqnarray*}\begin{array}{@{}l@{}} {\rm{J0}} = - {\rm{C/2xcos2a}}\\ {\rm{J45}} = - {\rm{C/2xsin2a}}{\rm{.}} \end{array}\end{eqnarray*}

 

In both equations, positive values of J0 represent with-the-rule astigmatism, negative values represent against-the-rule astigmatism, and J45 corresponds to oblique astigmatism.¹⁶ Cylinder C is equal to steep keratometry minus flat keratometry and the axis is the degree of steep keratometry. 

Statistical Analysis

Statistical analysis was performed with the SPSS software (version 22.0; IBM Corp.). The normality of distribution was checked by the Shapiro-Wilk test. Repeated-measures analysis of variance (rANOVA) was used to assess differences in the keratometry measurements obtained by the three biometers. If the rANOVA showed significant differences, post hoc analysis using the Bonferroni test was further performed for pairwise comparisons. Besides, we used the Bland-Altman diagram and intraclass correlation coefficients (ICCs) to evaluate the agreement among biometers. A single-measurement, absolute-agreement, two-way random-effects model was used to estimate the ICCs. The mean difference, 95% confidence interval (CI), and 95% limits of agreement (LoA; calculated as mean difference ± 1.96 SD) were used to illustrate the agreement and variability of the keratometry measurements. A P value less than 0.05 was considered statistically significant. 

Results

Demographic and Clinical Characteristics of Participants

A total of 102 patients (41 men and 61 women, 102 eyes) were included in the final analysis, with a mean age of 68.21 ± 8.70 years. Detailed demographic and clinical characteristics of the participants are shown in Table 1. The mean axial length (AL) was 24.26 ± 2.05 mm, with 11 eyes (10.78%) having AL longer than 26 mm. 

Table 1.

 

Demographic and Clinical Characteristics of Participants.

 

Agreement of Keratometry

The IOLMaster 700 had the highest mean total keratometry (TK_m; 44.23 ± 1.59 diopters [D]), followed by Pentacam AXL (43.94 ± 1.68 D) and CASIA2 (43.25 ± 1.53 D). As for mean posterior keratometry (PK_m), Pentacam AXL showed the highest value (-6.37 ± 0.26 D), followed by CASIA2 (-6.25 ± 0.25 D) and IOLMaster 700 (-5.90 ± 0.24 D). In terms of mean SimK (SimK_m), CASIA2 showed the highest value (44.32 ± 1.57 D), followed by IOLMaster 700 (44.18 ± 1.59 D) and Pentacam AXL (44.14 ± 1.63 D). Statistically significant differences were observed for SimK_m, TK_m, and PK_m obtained by these three biometers (all P < 0.001). In addition, statistical differences were observed in TK_m and PK_m among all pairwise comparisons for the three biometers (all P < 0.001; Table 2). 

Table 2.

 

Keratometry and Astigmatism Obtained by IOLMaster 700, CASIA 2, and Pentacam AXL

 

There were good agreements in TK_m (ICC = 0.887) and poor agreements in PK_m (ICC = 0.498) among these three biometers. The mean differences for TK_m ranged from 0.29 to 0.98 D, with a 95% LoA of –1.50 to 1.42 D. The largest mean difference was observed between IOLMaster 700 and CASIA2 (0.98 D). Regarding PK_m, the mean differences ranged from 0.13 to 0.47 D, with a 95% LoA of 0.01 to 0.62 D. The largest mean difference (0.47 D) was found between IOLMaster 700 and Pentacam AXL (Figs. 1, 2, Supplementary Table S1). 

Figure 1.

 

Bland-Altman plots of TK among IOLMaster 700, CASIA2 , and Pentacam AXL. (A, D, G) IOLMaster 700 versus CASIA2; (B, E, H) IOLMaster 700 versus Pentacam AXL; (C, F, I) CASIA2 versus Pentacam AXL. TK_f, flat total keratometry; TK_s, steep total keratometry; TK_m, mean total keratometry.

Figure 2.

 

Bland-Altman plots of PK among IOLMaster 700, CASIA2 , and Pentacam AXL. (A, D, G) IOLMaster 700 versus CASIA2; (B, E, H) IOLMaster 700 versus Pentacam AXL; (C, F, I) CASIA2 versus Pentacam AXL. PK_f, flat posterior keratometry; PK_s, steep posterior keratometry; PK_m, mean posterior keratometry.

Agreement of Astigmatism

As for corneal astigmatism, statistically significant differences were found for SimK-J0, TK-J0, and PK-J45 among these 3 biometers (all P < 0.001). In contrast, no significant differences were identified for SimK-J45, TK-J45, and PK-J0 (P = 0.569, P = 0.218, and P = 0.750, respectively). Additionally, TK-J0 was statistically different in all pairwise comparisons for the three biometers (all P < 0.001). 

There were good agreements in TK-J0 (ICC = 0.880), TK-J45 (ICC = 0.751), and PK-J0 (ICC = 0.807), and moderate agreements in PK-J45 (ICC = 0.626) among these 3 biometers. The mean differences for TK-J0 ranged from 0.11 to 0.22 D, with a 95% LoA of −0.74 to 0.40 D. The largest difference for TK-J0 was observed when comparing IOLMaster 700 and Pentacam AXL. The mean differences for PK-J45 were smaller, ranging from 0 to 0.04 D, with a 95% LoA of −0.07 to 0.08 D (Fig. 3, Supplementary Table S2). 

Figure 3.

 

Bland-Altman plots of TK-J0, TK-J45, PK-J0 , and PK-J45 among IOLMaster 700, CASIA2, and Pentacam AXL. (A, D, G, J) IOLMaster 700 versus CASIA2; (B, E, H, K) IOLMaster 700 versus Pentacam AXL; (C, F, I, L) CASIA2 versus Pentacam AXL. TK-J0, 0 vector of total keratometry; TK-J45, J45 vector of total keratometry; PK-J0, J0 vector of posterior keratometry; PK-J45, J45 vector of posterior keratometry.

Discussion

The current study identified significant differences in TK and PK measurements by the three biometers. It has been reported that the difference between SimK and TK can be as large as 0.50 D, and even larger in eyes with prior corneal refractive surgery or keratoconus.¹⁷ TK achieved higher refractive prediction accuracy for IOL power calculation than SimK.^7,8,9,18 In our study, the IOLMaster 700 showed the steepest TK_m and CASIA2 yielded the flattest TK_m. Based on the premise that 1.00 D of keratometry measurement error could result in an IOL power change of 1.10 to 1.30 D, the observed TK_m differences (0.29–0.98 D) among the three biometers could lead to clinically significant differences (0.32–1.27 D) in IOL power calculation.¹⁹ Similarly, Oh R et al. reported that TK_m measured by IOLMaster 700 was 1.05 D steeper than CASIA2.²⁰ On the contrary, Shajari M et al. reported TK_m measured by Pentacam AXL was 0.29 D steeper than IOLMaster 700. The discrepancy may be explained by that they used true net power (TNP) instead of TCRP measured by Pentacam AXL.³ In view of the considerable difference in TK measurement by the three biometers, they could not be used interchangeably in clinical practice. 

Currently, PK can be directly measured by several biometers, which further improves the accuracy of IOL power calculation.²¹ In our study, Pentacam AXL showed the steepest PK_m, followed by CASIA2 and IOLMaster 700. Specifically, PK_m obtained by IOLMaster 700 was 0.47 D flatter than Pentacam AXL. Similarly, Kose B. found that IOLMaster 700 exhibited a 0.51 D flatter PK_m than Pentacam HR in myopic eyes.¹⁴ Zhang T et al. also found that PK_m measured by CASIA2 was flatter than Pentacam AXL by 0.16 D in healthy volunteers aged 18 to 36 years. As a corollary, the agreement of PK among these three biometers was low and these devices cannot be used interchangeably. 

A good repeatability and reproducibility of IOLMaster 700, CASIA2, and Pentacam AXL has been proved in many previous studies.^11,13,22,23 Discrepancies in TK and PK measurements among these three devices are largely attributable to the different measuring technologies, image processing methods, and corneal zone analyzed.^5,11,12 IOLMaster 700 measures the anterior surface at 1.5 mm, 2.5 mm, and 3.2 mm diameters using 18 projected reference points (equivalent to 9 meridians) and maps the posterior surface using 6 meridional SS-OCT scans.²⁴ Data from all diameters and points are analyzed for consistency and correlation quality to maximize accuracy, and data from the 2.5 mm diameter are used for IOL power calculation.⁵ The CASIA2 performs 16 consecutive meridional scans centered on the corneal vertex, and only keratometry at the 3.0 mm diameter, the default value of the machine, was used for analysis.^4,11 Pentacam AXL scans 25 meridians to measure the corneal curvature using a 360-degree rotating Scheimpflug camera,^12,24,25 and TCRP at 4.0 mm ring was analyzed in our study. Previous studies demonstrated that a larger measurement diameter rendered higher corneal irregularity,^26,27 so the measurement meridians or analyzed zone could have a significant impact on TK, PK, and corresponding astigmatism. The axial resolution of CASIA2, IOLMaster 700, and Pentacam AXL is ≤10 µm, 22 µm, and 50 µm, respectively,^24,28,29 which could be another possible reason for the difference in TK and PK measurements among devices. 

TCA based on PCA is more accurate than the SimK-based keratometric astigmatism for toric IOL power calculation regardless of measurement methods.^1,9,30–32 Our study observed significant differences in TK-J0 but not in TK-J45 among these three biometers. Similarly, Wang L. et al. found clinically significant differences in TCA between the IOLMaster 700 and Pentacam AXL.¹⁰ In addition, Shajari M. et al. reported poor agreement of TK and TCRP in the astigmatism vectors between IOLMaster 700 and Pentacam AXL.³ As for PCA, statistically significant differences were observed for PK-J45 between IOLMaster 700 and Pentacam AXL as well as between CASIA2 and Pentacam AXL. Wang L. et al. reported that IOLMaster 700 obtained a lower PCA compared with Pentacam AXL.¹⁰ Kose B. et al. also reported significant differences in both PK-J0 and PK-J45 between IOLMaster 700 and Pentacam AXL in myopic eyes.¹⁴ These results collectively imply poor agreements of both the magnitude and axial position of astigmatism among these three biometers, suggesting non-interchangeability in clinical practice. 

There are several limitations in this study. First, the sample size of this study was relatively small. Future studies with larger sample sizes are needed. Second, this study was a single-center study of Chinese patients scheduled for cataract surgery, the agreement and interchangeability of the three biometers in other populations, especially those with prior corneal refractive surgery or keratoconus need further assessment. Third, which keratometry measurements could yield more accurate results for toric IOL power calculation is unknown and warrants further investigation. In addition, participants with a history of dry eye were excluded based on self-report medical history, but objective tests for dry eye disease, such as tear break-up time or Schirmer's test, were not performed. Future studies on the effect of dry eye on keratometric measurements are warranted. 

In conclusion, compared with IOLMaster 700, Pentacam AXL and CASIA2 showed a flatter TK_m and a steeper PK_m. TCA and PCA were different among these three devices both in magnitude and axial position. Therefore, TK, PK, and the corresponding astigmatism measured by IOLMaster 700, CASIA2, and Pentacam AXL cannot be used interchangeably. 

Acknowledgments

Supported by the National Natural Science Foundation of China (Nos. 82070940 and 82070941). The funding organization had no role in the design or conduct of this research. 

Disclosure: A. Jin, None; X. Han, None; J. Zhang, None; X. Qiu, None; Y. Zhang, None; B. Qu, None; X. Tan, None; L. Luo, None