You are viewing the site in preview mode

Skip to main content
Download PDF

Assessing the interchangeability of keratometry measurements from four biometric devices in intraocular lens power calculations: insights into the predictive accuracy of five modern IOL formulas

BMC Ophthalmology volume 25, Article number: 236 (2025) Cite this article

Abstract

Background

Achieving accurate intraocular lens (IOL) power calculation is crucial for successful refractive outcomes in cataract surgery. This study aimed to evaluate the interchangeability of keratometry (K) values obtained from four biometric devices (IOLMaster 700, CASIA2, Pentacam, and iTrace) and assess the predictive accuracy of five modern IOL calculation formulas (Barrett Universal II, Cooke K6, EVO 2.0, Kane, and PEARL-DGS) when using K values from these different devices.

Methods

This prospective study included K values obtained from four biometric devices for use in five IOL power calculation formulas. Predictive accuracy was assessed using multiple statistical parameters, including standard deviation (SD), mean absolute error (MAE), median absolute error (MedAE) and root mean square absolute error (RMSAE). The interchangeability of devices was evaluated by comparing predictive outcomes across devices and formulas, with statistical analyses focusing on consistency and agreement.

Results

Predictive accuracy across the five IOL formulas was stable and showed no statistically significant differences when using keratometry measurements from the same biometric device. However, significant variability was noted when comparing K values from different devices using the same formula. The SS-OCT-based devices (IOLMaster 700 and CASIA2) showed higher consistency in predictive accuracy compared to Scheimpflug-based Pentacam and ray-tracing-based iTrace. Despite this inter-device variability, all five IOL formulas showed overall robust performance across different devices.

Conclusions

Our findings indicate that keratometry measurements from different biometric devices are not fully interchangeable. SS-OCT-based devices (IOLMaster 700 and CASIA2) provided superior consistency in refractive prediction accuracy. Therefore, clinicians should carefully select biometric device-formula combinations based on the specific measurement principles and desired refractive outcomes. Further research involving larger sample sizes, additional IOL types and biometric devices, as well as assessment of surgeon-related factors, is warranted to optimize refractive accuracy in cataract surgery.

Peer Review reports

Background

Achieving precise refractive outcomes in cataract surgery is critical, particularly with the increasing adoption of premium intraocular lenses (IOLs) designed to reduce spectacle dependence [1]. Accurate corneal power measurements are essential for selecting the optimal IOL and ensuring consistent postoperative refractive results [2]. However, errors in keratometry (K) values have been reported to contribute to up to 22% of the total error in IOL power calculation [3, 4]. This highlights the necessity of precise and reliable K measurements, as even small inaccuracies can lead to significant postoperative refractive deviations. Despite advances in biometry technologies, including swept-source optical coherence tomography (SS-OCT), Scheimpflug imaging, and ray-tracing aberrometry, variability persists among devices, raising concerns about the interchangeability of their measurements and their potential impact on refractive outcomes [5,6,7].

In our previous study, we compared the performance of four commonly used biometric devices: IOLMaster 700, CASIA2, Pentacam, and iTrace. While strong correlations were observed in anterior corneal curvature measurements, significant discrepancies were noted in posterior and total corneal power. These differences were attributed to variations in measurement principles and the optical zones analyzed. The findings underscored the necessity of caution when interpreting measurements across devices and highlighted the lack of standardization in device outputs [5]. However, the study did not address how device-derived measurements influence refractive prediction accuracy when integrated into contemporary IOL power calculation formulas.

Building on these findings, the present study investigates the interchangeability of K values derived from four biometric devices and their integration into five modern IOL power calculation formulas: Barrett Universal II, Cooke K6, EVO 2.0, Kane, and PEARL-DGS. Given the critical need to minimize postoperative refractive errors [8], this prospective study evaluates refractive prediction errors (PE) derived from these biometric devices by applying their corneal power measurements (K values) in the five formulas. Predictive accuracy is assessed using metrics such as standard deviation (SD), mean absolute error (MAE), median absolute error (MedAE), root mean square absolute error (RMSAE) and the percentage of eyes achieving refraction within specific error ranges (± 0.25D, ± 0.50D, ± 0.75D, and ± 1.00D). By focusing on the interchangeability of K values across devices and the consistency of these formulas, this study aims to provide valuable insights for optimizing preoperative planning and improving refractive outcomes in cataract surgery.

Methods

Study design

This prospective, observational study was conducted between April and September 2023, adhering to the tenets of the Declaration of Helsinki. Ethical approval was obtained from the institutional ethics committee (No.2023KY-04), and written informed consent was acquired from all participants prior to enrollment.

Study population

The study included 194 eyes from 194 patients scheduled for phacoemulsification cataract surgery. Inclusion criteria included: age ≥ 18 years, corneal astigmatism < 3 diopters, successful preoperative evaluation with all four devices, and a minimum postoperative follow-up of four weeks. Patients with coexisting ocular conditions, intraoperative complications, or poor compliance were excluded.

Biometric measurements

Preoperatively, all eyes underwent biometry with four devices: IOLMaster 700 (Carl Zeiss Meditec AG, Jena, Germany; software version 1.90.38.02), CASIA2 (Tomey Corporation, Nagoya, Japan; software version 50.5B.C1), Pentacam (OCULUS, Wetzlar, Germany; software version 1.21r25), and iTrace (Tracey Technologies Corp., Houston, TX, USA; software version 6.1.0). IOLMaster 700 is based on swept-source optical coherence tomography (SS-OCT), which enables precise axial length and keratometry measurements by scanning across different depths of the eye. CASIA2 is also an SS-OCT-based device, providing high-resolution anterior segment imaging and measuring corneal power through an anterior segment tomography approach. Pentacam utilizes Scheimpflug imaging, capturing a rotating series of slit images to analyze both anterior and posterior corneal surfaces for total corneal power assessment. iTrace is a ray-tracing aberrometer and corneal topographer, which measures corneal wavefront aberrations and computes corneal power based on multiple ray-path analyses. Measurements were conducted by experienced ophthalmic technicians, and strict calibration protocols were followed for each device.

IOL power calculation

Refractive predictions were calculated using five IOL power formulas: Barrett Universal II (available at https://calc.apacrs.org/barrett_universal2105/, accessed on August 15, 2024), Cooke K6 (available at https://cookeformula.com/, accessed on August 17, 2024), EVO 2.0 (available at https://www.evoiolcalculator.com/, accessed on August 17, 2024), Kane (available at https://www.iolformula.com/, accessed on August 17, 2024), and PEARL-DGS (available at https://iolsolver.com/, accessed on August 20, 2024). Corneal power measurements (K values) from each device were used as inputs. For the actual IOL implantation, the Barrett Universal II formula was applied to determine the final IOL power selection. The implanted intraocular lens was the Alcon AcrySof IQ SN60WF (Alcon Laboratories, Inc., Fort Worth, TX, USA), with an optimized A-constant of 118.97, obtained from the IOLCon website (https://www.iolcon.org). Prediction error (PE) was calculated as the difference between the actual postoperative spherical equivalent (SEQ) and the predicted SEQ. The study design was based on recommendations for intraocular lens power calculation protocols [9, 10]. Additionally, the impact of testing distance on IOL power calculation [11, 12] was considered to minimize potential measurement errors.

Statistical analysis

Data were analyzed using Excel 2016 (Microsoft Corp., Redmond, WA, USA), MedCalc version 15.2.2 (MedCalc Software Ltd., Ostend, Belgium), and R software version 4.4.1 (R Foundation for Statistical Computing, Vienna, Austria). Normality was assessed using the Shapiro-Wilk test. Normally distributed variables are presented as mean ± SD, while non-normally distributed variables are reported as median and interquartile range (IQR). Differences were considered statistically significant when the P value was less than 0.05.

Additionally, the SEQ_PE wrap-up function from the set of 28 R functions (Rallfun-v43.txt, https://osf.io/xhe8u/, accessed September 7, 2024) was used for spherical equivalent prediction error analysis. The statistical parameters used for the analysis included SD, MAE, MedAE, RMSAE, and the percentage of eyes within specific prediction error intervals, aligning with recent studies emphasizing the importance of RMSAE in evaluating IOL power calculation accuracy [13,14,15,16].

Pearson correlation coefficients (r), intraclass correlation coefficients (ICC), and 95% confidence intervals (CI) were calculated using a two-way mixed-effects model to assess agreement between devices. Agreement between device-formula combinations was evaluated with Bland-Altman plots. G*Power 3.1 was used to calculate the sample size (n = 180), based on an effect size of 0.25, a power of 0.80, and α = 0.05. Graphical representations were generated using GraphPad Prism version 9.4.1 (GraphPad Software, San Diego, CA, USA).

Results

Baseline characteristics

The study involved 194 cataract surgery patients, with a mean age of 67.4 ± 8.6 years (range: 39–83). The cohort included 42.3% males and 57.7% females, with an approximately equal distribution of right (51.0%) and left eyes (49.0%). The mean axial length was 23.72 ± 1.32 mm, and the mean implanted IOL power was 20.51 ± 2.56 diopters.

Keratometry measurements were obtained using four different devices, which showed comparable results. Full demographic and biometric details, including keratometry values, lens thickness, corneal diameter, central corneal thickness, and anterior chamber depth, are summarized in Table 1.

Table 1 Demographic and biometric characteristics of cataract patients included in the study
Full size table

Visual and refractive outcomes

Figure 1 summarizes the postoperative visual and refractive outcomes. All patients achieved uncorrected distance visual acuity (UDVA) of 20/400 or better, consistent with preoperative surgical planning, including cases of targeted myopia (Fig. 1a). Corrected distance visual acuity.

Fig. 1
figure 1

Standard cataract surgery outcome analysis. (a) Postoperative uncorrected distance visual acuity (UDVA) and corrected distance visual acuity (CDVA). (b) Histogram of the lines of difference between postoperative UDVA and CDVA. (c) Postoperative spherical equivalent (SEQ) outcomes. (d) Residual refractive cylinder outcomes

Full size image

(CDVA) outcomes were more indicative of visual potential, with all patients achieving CDVA of 20/40 or better, a study inclusion criterion.

A comparison between postoperative UDVA and CDVA showed that 39% of eyes achieved UDVA equal to or better than CDVA, with 70% of eyes within 1 line of CDVA (Fig. 1b). This suggests reasonable alignment between uncorrected and corrected visual outcomes, despite intentional under-correction in some cases.

Refractive predictability was moderate, with 60% of eyes achieving postoperative SEQ within ± 0.50 D of the target, and 76% within ± 1.00 D (Fig. 1c). The predictability within ± 0.50 D was reduced due to the preoperative goal of retaining myopia in some patients.

Postoperative refractive cylinder analysis revealed that 37% of eyes had residual cylinder ≤ 0.50 D, while 65% had residual cylinder ≤ 1.00 D (Fig. 1d). Although these results are acceptable, they indicate potential areas for improving residual astigmatism control.

Intra-device comparisons

For intra-device comparisons, all five IOL power calculation formulas (Barrett Universal II, Cooke K6, EVO 2.0, Kane, and PEARL-DGS) generally performed similarly across the four biometric devices (IOLMaster 700, CASIA2, Pentacam, and iTrace). However, differences were observed in key statistical parameters, including SD, MAE, MedAE, and RMSAE across devices.

On the IOLMaster 700, Cooke K6 demonstrated the lowest SD (0.399 D), MAE (0.305 D), MedAE (0.233 D), and RMSAE (0.398 D) (Table 2; Fig. 2). indicating the highest predictive accuracy among the formulas. PEARL-DGS followed closely with SD (0.401 D), MAE (0.309 D), MedAE (0.243 D), and RMSAE (0.400 D). Barrett Universal II, while slightly higher, still performed well with SD (0.411 D), MAE (0.312 D), MedAE (0.250 D), and RMSAE (0.410 D). Cooke K6 had the highest percentage of eyes within ± 0.25 D (52.85%), while PEARL-DGS exhibited the best accuracy within ± 0.50 D (81.44%). Within ± 1.00 D, PEARL-DGS performed the best, reaching 98.45% (Table 2; Fig. 3). However, statistical analysis showed no significant differences in SD, MAE, MedAE, and RMSAE among different formulas when using IOLMaster 700 (P > 0.05).

Fig. 2
figure 2

Comparative analysis of biometric devices and formulas. (a) Bar chart comparing Kf, Ks, and Km values across four devices. (b) Bar chart showing standard deviation (SD) values for five formulas across four devices. (c). Bar chart comparing mean absolute error (MAE) for five formulas across four devices. (d) Bar chart illustrating median absolute error (MedAE) for five formulas across four devices. (e) Violin plot showing absolute prediction error (APE) distributions for each device-formula combination. Note: Statistically significant difference at *p-values < 0.05; **p-value < 0.01

Full size image
Table 2 Refractive outcomes obtained by five formulas on four different machines
Full size table

On CASIA2, Cooke K6 again had the lowest SD (0.420 D), MAE (0.319 D), MedAE (0.252 D), and RMSAE (0.419 D), supporting its strong predictive performance. EVO 2.0 (SD 0.428 D, MAE 0.331 D, MedAE 0.264 D, RMSAE 0.427 D) and PEARL-DGS (SD 0.429 D, MAE 0.327 D, MedAE 0.276 D, RMSAE 0.428 D) followed closely. The highest percentage of eyes within ± 0.25 D was observed with Cooke K6 (48.19%), which also achieved the highest accuracy within ± 0.50 D (83.42%), followed by Barrett Universal II (81.96%) and PEARL-DGS (81.44%). The highest accuracy within ± 1.00 D was observed for Barrett Universal II and Kane (96.91%) (Table 2; Fig. 3). However, statistical comparisons showed no significant differences in SD, MAE, MedAE, and RMSAE among different formulas on CASIA2 (P > 0.05).

Fig. 3
figure 3

Refractive error prediction accuracy across devices and formulas description. (a) Stacked bar graph showing percentage distribution of eyes within refractive error ranges for five formulas across four devices. (b) Stacked bar graph comparing refractive error ranges for each device across the five formulas

Full size image

For Pentacam, Cooke K6 once again exhibited the lowest SD (0.456 D), MAE (0.354 D), MedAE (0.274 D), and RMSAE (0.455 D), maintaining its predictive reliability. Kane (SD 0.459 D, MAE 0.360 D, MedAE 0.298 D, RMSAE 0.458 D) and PEARL-DGS (SD 0.462 D, MAE 0.362 D, MedAE 0.295 D, RMSAE 0.460 D) displayed slightly higher values (Table 2; Fig. 2). The highest accuracy within ± 0.25 D was recorded for Cooke K6 (46.11%), which also achieved the highest accuracy within ± 0.50 D (75.65%), followed by EVO 2.0 (74.09%) and Barrett Universal II (73.71%). Kane demonstrated the highest accuracy within ± 1.00 D (96.39%) (Fig. 3). However, no significant differences in SD, MAE, MedAE, and RMSAE were observed among different formulas when using Pentacam (P > 0.05).

iTrace exhibited greater variability compared to other devices. Barrett Universal II and EVO 2.0 exhibited the lowest SD (0.478 D), followed by Kane (0.488 D) and PEARL-DGS (0.489 D). The lowest MAE was observed for EVO 2.0 (0.354 D), while Kane had the highest MAE (0.359 D). In terms of MedAE, Cooke K6 had the lowest value (0.265 D), while PEARL-DGS had the highest MedAE (0.288 D). The lowest RMSAE was observed for Barrett Universal II and EVO 2.0 (0.477 D), whereas Cooke K6 showed the highest RMSAE (0.491 D) (Table 2; Fig. 2). Regarding refractive prediction accuracy, Cooke K6 and EVO 2.0 achieved the highest percentage of eyes within ± 0.25 D (47.15%), while Kane outperformed other formulas within ± 0.50 D (78.35%). Kane and PEARL-DGS demonstrated the highest accuracy within ± 1.00 D, reaching 95.36% (Table 2; Fig. 3). However, statistical analysis revealed no significant differences in SD, MAE, MedAE, and RMSAE among different formulas when using iTrace (P > 0.05).

Overall, Cooke K6 consistently achieved the most favorable predictive metrics across all devices, particularly on IOLMaster 700 and CASIA2, while greater variability was noted for Pentacam and iTrace. The choice of biometric device significantly influenced the predictive accuracy of each IOL formula.

Inter-device comparisons using the same formula

Significant differences in SD (P = 0.017), RMSAE (P = 0.024) and MAE (P = 0.012) were observed between IOLMaster 700 and Pentacam when using the Barrett Universal II formula. IOLMaster 700 demonstrated superior prediction accuracy but showed no significant differences compared to CASIA2 or iTrace (P > 0.05) (Table 2; Fig. 2).

For Cooke K6, significant differences in SD were observed between IOLMaster 700 and both Pentacam (P = 0.010) and iTrace (P = 0.017). IOLMaster 700 also showed significantly lower MAE compared to Pentacam (P = 0.006) and iTrace (P = 0.030), indicating superior predictive performance. Additionally, IOLMaster 700 exhibited a significantly lower RMSAE compared to Pentacam (P = 0.006), further reinforcing its predictive accuracy (Table 2; Fig. 2).

For EVO 2.0, IOLMaster 700 had a significantly lower SD than Pentacam (P = 0.047) and lower MAE (P = 0.012), Similarly, IOLMaster 700 also demonstrated a significantly lower RMSAE compared to Pentacam (P = 0.012), further supporting its superior predictive performance. No significant differences were found in other comparisons (P > 0.05) (Table 2; Fig. 2).

The Kane formula showed significant differences in ± 0.50 D accuracy between IOLMaster 700 and Pentacam (P = 0.048), with IOLMaster 700 performing better. Additionally, IOLMaster 700 exhibited a significantly lower RMSAE compared to Pentacam (P = 0.048). No significant differences were found in other comparisons (Table 2; Fig. 2).

For PEARL-DGS, significant differences were observed between IOLMaster 700 and both Pentacam (P = 0.008) and iTrace (P = 0.046), with IOLMaster 700 demonstrating lower MAE. Additionally, IOLMaster 700 exhibited a significantly lower RMSAE compared to Pentacam (P = 0.012). Refractive prediction differences were also noted between IOLMaster 700 and Pentacam in the ± 0.25 D (P = 0.046) and ± 0.50 D (P = 0.028) ranges, and between CASIA2 and Pentacam in the ± 0.50 D range (P = 0.036) (Table 2; Fig. 2).

Table 3 Agreement and correlation of prediction errors among five formulas across different devices
Full size table

Bland-altman analysis

Bland-Altman analysis demonstrated the narrowest limits of agreement (LoA) between IOLMaster 700 and CASIA2, indicating strong agreement. Wider LoA were observed for iTrace, especially in comparison with IOLMaster 700 and CASIA2, indicating greater measurement variability (Table 3; Fig. 4). Intraclass correlation coefficients (ICC) also supported the superior consistency of IOLMaster 700, particularly compared with CASIA2.

Fig. 4
figure 4

Bland-Altman comparison of refractive prediction errors between biometric devices. (a) IOLMaster 700 vs. CASIA2, with subplots a1-a5 showing comparisons for the five formulas. (b) IOLMaster 700 vs. Pentacam, with subplots b1-b5 for each formula. (c) IOLMaster 700 vs. iTrace, subplots c1-c5 showing comparisons across formulas. (d) CASIA2 vs. Pentacam, subplots d1-d5 for each formula. (e) CASIA2 vs. iTrace, subplots e1-e5 for formula comparisons. (f) Pentacam vs. iTrace, subplots f1-f5 showing comparisons for each formula. Each plot illustrates the agreement and variation between the devices, with 95% limits of agreement and mean differences highlighted. BII, Barrett Universal II; K6, Cooke K6; EVO, emmetropia verifying optical; DGS, Pearl-DGS

Full size image

Pearson correlation coefficients

Pearson correlation coefficients (r) were calculated to evaluate the correlation between devices using the same formula. Across all formulas, the highest correlations were found between IOLMaster 700 and CASIA2, ranging from r = 0.898 to r = 0.924 (P < 0.001), indicating strong agreement (Table 3). iTrace showed the lowest correlation with IOLMaster 700 (r = 0.753 to r = 0.778, P < 0.001), indicating greater variability. EVO 2.0 and Kane demonstrated the highest correlation between IOLMaster 700 and CASIA2, with r = 0.923 and r = 0.924, respectively. Pentacam showed slightly lower but substantial correlations with IOLMaster 700 across all formulas (ranging from r = 0.819 to r = 0.851).

Discussion

This study aimed to evaluate whether K values obtained from four different biometric devices (IOLMaster 700, CASIA2, Pentacam, and iTrace) are interchangeable in modern intraocular lens (IOL) power calculation formulas, while also analyzing the predictive accuracy of five IOL formulas (Barrett Universal II, Cooke K6, EVO 2.0, Kane, and PEARL-DGS) using K values from different devices. Overall, although there were differences in predictive outcomes across various statistical parameters when using the same device’s K values with different IOL formulas, these differences were not statistically significant, indicating consistent predictive performance among the formulas. However, when analyzing K values from different devices using the same calculation formula, the SS-OCT-based devices (IOLMaster 700 and CASIA2) showed higher consistency in predictive accuracy, while the Scheimpflug and iTrace devices demonstrated greater variability in certain predictive parameters. This suggests that devices with different measurement principles may vary in predictive accuracy, necessitating careful device selection based on specific performance characteristics.

Our findings align with previous studies on the consistency between different biometric devices [17, 18]. The differences between Placido-Scheimpflug and OCT technologies may partially be attributed to their differing measurement principles and the optical zones they analyze. While measurements of anterior and posterior corneal curvature showed good correlation across devices, significant differences were observed in overall corneal power, which may influence the accuracy of IOL power calculations [19, 20]. Further analysis of the five IOL calculation formulas indicated that they maintained good predictive stability and consistency across different K values.

Although statistical analyses revealed no significant differences in predictive accuracy among the five modern intraocular lens calculation formulas when using keratometry (K) values from the same biometric device, it is essential to emphasize the clinical implications of these findings. Small yet consistent performance variations, such as those observed with the Cooke K6 and Barrett Universal II formulas on swept-source OCT-based devices (IOLMaster 700 and CASIA2), could clinically translate into better refractive outcomes and higher patient satisfaction, especially in patients opting for premium multifocal or toric IOLs. Therefore, even minor numerical improvements in refractive prediction metrics, despite lacking statistical significance, remain clinically important and should guide surgeons toward optimal biometric device and formula selection, particularly in refractive cataract surgery where achieving minimal residual refractive error is paramount.

Clinically, the observed differences in keratometry measurements among the four biometric devices have practical implications for refractive outcomes. Specifically, the superior consistency demonstrated by the SS-OCT-based devices (IOLMaster 700 and CASIA2) makes them preferable for routine clinical use, particularly when employing advanced IOL formulas like Barrett Universal II and Cooke K6. Conversely, the higher variability noted with the Scheimpflug-based Pentacam and ray-tracing iTrace devices may result in greater unpredictability in refractive outcomes, potentially diminishing patient visual satisfaction, especially in premium IOL or refractive lens exchange scenarios. Clinicians should therefore critically consider device selection, utilizing Pentacam and iTrace primarily as supplementary diagnostic tools for complex corneas or cases with prior refractive surgeries, where detailed posterior corneal curvature or aberration analyses may provide valuable supplementary data.

Specifically, IOLMaster 700 and CASIA2 exhibited the most stable performance, with more consistent predictive accuracy compared to the other devices, indicating greater refractive prediction accuracy and high interchangeability between these two devices [21,22,23,24]. Pentacam and iTrace, while showing greater variability in corneal power measurements, can still be used as supplementary tools under certain conditions [25,26,27,28]. However, caution is advised when interpreting their results for direct IOL power calculations.

The choice of IOL power calculation formula was critical in determining refractive accuracy [6, 29,30,31]. The study also demonstrated that, while differences between different IOL formulas using the same device’s K values were not statistically significant, analyzing the same IOL formula with K values from different devices showed statistically significant differences in some parameters. This finding is consistent with previous studies [6] comparing the predictive accuracy of IOL power using two SS-OCT-based biometric devices, both of which demonstrated high predictive accuracy without significant differences. Furthermore, a comprehensive analysis of the overall performance of each formula revealed that all five IOL formulas (Barrett Universal II, Cooke K6, EVO 2.0, Kane, and PEARL-DGS) showed stable predictive accuracy across different devices, further underscoring their robustness and applicability in clinical practice. By comparing multiple parameters (such as SD, MAE, RMSAE and MedAE), the study found no significant differences in the predictive results when using the same device’s K values. However, differences were observed when K values from different devices were used, suggesting that careful selection of devices and formulas is crucial to optimize refractive outcomes.

Another significant finding is that, while no statistically significant differences were found in predictive accuracy between different IOL formulas using the same device’s K values, differences in biological parameters were observed when analyzing the same IOL formula with K values from different devices. This indicates that, in clinical practice, leveraging the specific strengths of individual devices can allow for more targeted selection of suitable device-formula combinations to optimize refractive prediction accuracy and minimize postoperative refractive errors. This is crucial for clinical surgical planning, enabling ophthalmologists to make more precise decisions based on device characteristics, ultimately improving patient outcomes and satisfaction [32, 33], ensuring precise IOL power calculation is essential for optimizing visual outcomes and patient satisfaction [34].

This study has several limitations. First, the sample size was relatively small, which, although sufficient for initial analysis, may limit the generalizability of the results. Additionally, the study population was drawn from a single center, which may not be representative of a broader population. Furthermore, the study only evaluated a specific type of IOL, excluding other types that might affect the generalizability of the findings. The selection of IOL calculation formulas was also limited to five modern formulas (Barrett Universal II, Cooke K6, EVO 2.0, Kane, and PEARL-DGS), and other formulas were not considered.

Future research should expand the sample size and include a broader range of IOL types and biometric devices to enhance the generalizability of the findings. Multi-center studies are recommended to further validate the interchangeability of different devices and calculation formulas across diverse populations. Additionally, future studies should focus on the impact of surgeon-related factors on the accuracy of IOL power calculations to comprehensively optimize refractive prediction precision and consistency.

Although our study confirmed the superior predictive accuracy and interchangeability of SS-OCT-based devices (IOLMaster 700 and CASIA2), for eyes with complex corneal morphologies, combining multiple devices may yield additional diagnostic benefits. Recognizing the study’s limitations—such as a relatively small sample size, single-center design, and evaluation limited to one IOL type—these findings should be validated further. Despite these limitations, our results provide clear guidance for clinicians to carefully select appropriate device-formula combinations to achieve optimal refractive outcomes in clinical practice.

Conclusions

In conclusion, keratometry values obtained from different biometric devices are not completely interchangeable. The SS-OCT-based devices (IOLMaster 700 and CASIA2) exhibited higher consistency and predictive accuracy compared with the Scheimpflug-based Pentacam and the ray-tracing-based iTrace. These differences hold clinical significance, particularly in premium intraocular lens implantation or cases where precise refractive outcomes are essential. Careful selection of biometric devices and corresponding IOL calculation formulas is recommended to optimize refractive outcomes and enhance patient satisfaction. Future studies should broaden the scope by including larger sample sizes, diverse IOL types, additional biometric devices, and assessing surgeon-related variables, thereby contributing to improved standardization and accuracy of IOL power predictions in clinical practice.

Data availability

The datasets generated and analyzed during the current study, including raw keratometry values and predictive error data, are available from the corresponding author upon reasonable request.

Abbreviations

IOL:

Intraocular lens

SD:

Standard deviation

MAE:

Mean absolute error

MedAE:

Median absolute error

SS-OCT:

Swept-source optical coherence tomography

CDVA:

Corrected distance visual acuity

UDVA:

Uncorrected distance visual acuity

CCT:

Central corneal thickness

ACD:

Anterior chamber depth

LT:

Lens thickness

WTW:

White-to-white distance

SimK:

Simulated keratometry

SEQ:

Spherical equivalent

MPE:

Mean prediction error

MIN:

Minimum

MAX:

Maximum

IQR:

Interquartile range

ICC:

Intraclass correlation coefficients

CI:

Confidence intervals

LoA:

Limits of agreement

References

  1. Narang R, Agarwal A. Refractive cataract surgery. Curr Opin Ophthalmol. 2024;35(1):23–7.

    Article  PubMed  Google Scholar 

  2. Savini G, Negishi K, Hoffer KJ, Schiano Lomoriello D. Refractive outcomes of intraocular lens power calculation using different corneal power measurements with a new optical biometer. J Cataract Refract Surg. 2018;44(6):701–08.

    Article  PubMed  Google Scholar 

  3. Stopyra W. The exactness of intraocular Lens power calculation formulas for short eyes and correlation between method accuracy and eyeball axial length. Cesk Slov Oftalmol. 2022;78(5):236–40.

    CAS  PubMed  Google Scholar 

  4. Stopyra W. Analysis of accuracy of twelve intraocular lens power calculation formulas for eyes with axial hyperopia. Saudi J Ophthalmol. 2023;37(2):125–30.

    Article  PubMed  PubMed Central  Google Scholar 

  5. Ma S, Gao R, Sun J, Yang J, Wen K, Chen X, Zhao F, Xu X, Tian F. Comparison of two swept-source optical coherence tomography devices, a Scheimpflug camera system and a ray-tracing aberrometer in the measurement of corneal power in patients with cataract. Graefes Arch Clin Exp Ophthalmol. 2024;262(5):1567–78.

    Article  PubMed  Google Scholar 

  6. Sorkin NIR, Zadok R, Savini G, Kan-Tor Y, Benjamini Y, Levinger E, Bardugo J, Abulafia ADI. Comparison of intraocular Lens power prediction accuracy between 2 Swept-Source optical coherence tomography biometry devices. Am J Ophthalmol. 2024;265(156– 64.

  7. Wagner FM, Hoffmann P, Preussner PR. Accuracy comparison of tomography devices for ray tracing-based intraocular lens calculation. J Cataract Refract Surg. 2024;50(2):110–15.

    Article  PubMed  Google Scholar 

  8. Kieval JZ, Al-Hashimi S, Davidson RS, Hamilton DR, Jackson MA, LaBorwit S, Patterson LE, Stonecipher KG, Donaldson K. Prevention and management of refractive prediction errors following cataract surgery. J Cataract Refract Surg. 2020;46(8):1189–97.

    Article  PubMed  Google Scholar 

  9. Hoffer KJ, Aramberri J, Haigis W, Olsen T, Savini G, Shammas HJ, Bentow S. Protocols for studies of intraocular lens formula accuracy. Am J Ophthalmol. 2015;160(3):403–e051.

    Article  PubMed  Google Scholar 

  10. Hoffer KJ, Savini G. Update on intraocular Lens power calculation study protocols: the better way to design and report clinical trials. Ophthalmology. 2021;128(11):e115–20.

    Article  PubMed  Google Scholar 

  11. Simpson MJ, Charman WN. The effect of testing distance on intraocular lens power calculation. J Refract Surg. 2014;30(11):726.

    Article  PubMed  Google Scholar 

  12. Hill WE, Abulafia A, Wang L, Koch DD. Pursuing perfection in IOL calculations. II. measurement foibles: measurement errors, validation criteria, IOL constants, and lane length. J Cataract Refract Surg. 2017;43(7):869–70.

    Article  PubMed  Google Scholar 

  13. Wang L, Holladay JT, Wilcox RR, Koch DD. Wrap-up R functions for statistical analysis of spherical equivalent and astigmatic prediction errors. J Cataract Refract Surg. 2023;49(8):890–91.

    Article  PubMed  Google Scholar 

  14. Holladay JT, Wilcox RR, Koch DD, Wang L. Review and recommendations for univariate statistical analysis of spherical equivalent prediction error for IOL power calculations. J Cataract Refract Surg. 2021;47(1):65–77.

    Article  PubMed  Google Scholar 

  15. Stopyra W, Voytsekhivskyy O, Grzybowski A. Accuracy of 20 intraocular Lens power calculation formulas in Medium-Long eyes. Ophthalmol Ther. 2024;13(7):1893–907.

    Article  PubMed  PubMed Central  Google Scholar 

  16. Stopyra W, Voytsekhivskyy O, Grzybowski A. Accuracy of 7 artificial Intelligence-Based intraocular Lens power calculation formulas in extremely long Caucasian eyes. Am J Ophthalmol. 2025;271(337– 46.

  17. Montes-Mico R, Pastor-Pascual F, Ruiz-Mesa R, Tana-Rivero P. Ocular biometry with swept-source optical coherence tomography. J Cataract Refract Surg. 2021;47(6):802–14.

    Article  PubMed  Google Scholar 

  18. Borgia A, Raimondi R, Sorrentino T, Santoru F, Buzzi M, Borgia V, Scorcia V, Giannaccare G. Swept-Source optical coherence Tomography-Based biometry: A comprehensive overview. Photonics. 2022;9(12).

  19. Shammas HJ, Shammas MC, Jivrajka RV, Cooke DL, Potvin R. Effects on IOL power calculation and expected clinical outcomes of axial length measurements based on multiple vs single refractive indices. Clin Ophthalmol. 2020;14:1511–19.

    Article  PubMed  PubMed Central  Google Scholar 

  20. Hessler P, Kunzel P, Degle S. Comparison of three different devices for the evaluation of axial length, refractive error, and keratometry. Optom Vis Sci. 2023;100(8):557–63.

    Article  PubMed  PubMed Central  Google Scholar 

  21. Aksoy M, Asena L, Gungor SG, Kucukoduk A, Akman A. Comparison of refractive outcomes using Scheimpflug holladay equivalent keratometry or iolmaster 700 keratometry for IOL power calculation. Int Ophthalmol. 2021;41(6):2205–12.

    Article  PubMed  Google Scholar 

  22. Choi JY, Choi A, Kwon H, Jeon S. Accuracy of IOL power calculation formulas for quadrifocal acrysof IQ PanOptix TFNT implantation in patients with previous corneal refractive surgery: comparison of SS-OCT-Based biometers. J Refract Surg. 2021;37(12):836–41.

    Article  PubMed  Google Scholar 

  23. Oh R, Oh JY, Choi HJ, Kim MK, Yoon CH. Comparison of ocular biometric measurements in patients with cataract using three swept-source optical coherence tomography devices. BMC Ophthalmol. 2021;21(1):62.

    Article  PubMed  PubMed Central  Google Scholar 

  24. Ruan X, Yang G, Xia Z, Zhang J, Gu X, Tan Y, Liu Z, Luo L. Agreement of anterior segment parameter measurements with CASIA 2 and iolmaster 700. Front Med (Lausanne). 2022;9:777443.

    Article  PubMed  Google Scholar 

  25. Muzyka-Wozniak M, Oleszko A, Grzybowski A. Measurements of anterior and posterior corneal curvatures with OCT and Scheimpflug biometers in patients with low total corneal astigmatism. J Clin Med. 2022;11(23).

  26. Lu AQ, Poulsen A, Cui D, Seeger C, Lehman E, Scott IU, Pantanelli SM. Repeatability and comparability of keratometry measurements obtained with swept-source optical coherence and combined dual Scheimpflug-Placido disk-based tomography. J Cataract Refract Surg. 2020;46(12):1637–43.

    Article  PubMed  Google Scholar 

  27. Llorens-Quintana C, Lee DJ, Pavlatos E, Chamberlain W, Huang D, Li Y. Measuring corneal astigmatism using OCT in keratoconus. J Cataract Refract Surg. 2022;48(11):1285–91.

    Article  PubMed  PubMed Central  Google Scholar 

  28. Shetty R, Trivedi D, Ranade R, Arun S, Khamar P, Kundu G. Repeatability and agreement of wavefront aberrations of a new hybrid topographer and aberrometer in healthy eyes. J Cataract Refract Surg. 2022;48(4):408–16.

    Article  PubMed  Google Scholar 

  29. Stopyra W, Cooke DL, Grzybowski A. A review of intraocular Lens power calculation formulas based on artificial intelligence. J Clin Med. 2024;13(2).

  30. Savini G, Hoffer KJ, Kohnen T. IOL power formula classifications. J Cataract Refract Surg. 2024;50(2):105–07.

    Article  PubMed  Google Scholar 

  31. Mo ER, Chen Z, Feng KE, Zhu Z, Xu J, Zhu C, Chang P, Li JIN, Zhao Y-E. Accuracy of modern intraocular Lens formulas in highly myopic eyes implanted with Plate-Haptic intraocular lenses. Am J Ophthalmol. 2024;265(105– 16.

  32. Schallhorn JM, Pantanelli SM, Lin CC, Al-Mohtaseb ZN, Steigleman WA 3rd, Santhiago MR, Olsen TW, Kim SJ, Waite AM, Rose-Nussbaumer JR. Multifocal and accommodating intraocular lenses for the treatment of presbyopia: A report by the American academy of ophthalmology. Ophthalmology. 2021;128(10):1469–82.

    Article  PubMed  Google Scholar 

  33. Blehm C, Potvin R. Reported patient satisfaction and spectacle independence following bilateral implantation of the PanOptix((R)) trifocal intraocular Lens. Clin Ophthalmol. 2021;15:2907–12.

    Article  PubMed  PubMed Central  Google Scholar 

  34. Kane JX, Chang DF. Intraocular Lens power formulas, biometry, and intraoperative aberrometry: A review. Ophthalmology. 2021;128(11):e94–114.

    Article  PubMed  Google Scholar 

Download references

Acknowledgements

The authors thank the staff at Tianjin Medical University Eye Hospital for their support during data collection.

Funding

This study was supported by Tianjin Key Medical Discipline (Specialty) Construction Project (No. TJYXZDXK-037 A); Weifang Science and Technology Bureau Project (No.2020YX065).

Author information

Authors and Affiliations

  1. Tianjin Key Laboratory of Retinal Functions and Diseases, Tianjin Branch of National Clinical Research Center for Ocular Disease, Eye Institute, School of Optometry, Tianjin Medical University Eye Hospital, Tianjin, China

    Shan Ma, Jing Sun, Jun Yang, Kai Wen, Xiteng Chen, Fangyu Zhao & Fang Tian

  2. Zhengda Guangming Eye Group, Weifang, China

    Shan Ma & Xuequan Sun

  3. Binzhou Polytechnic, Binzhou, China

    Cheng Li

Authors
  1. Shan Ma
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Cheng Li
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Jing Sun
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Jun Yang
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Kai Wen
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Xiteng Chen
    View author publications

    You can also search for this author inPubMed Google Scholar

  7. Fangyu Zhao
    View author publications

    You can also search for this author inPubMed Google Scholar

  8. Xuequan Sun
    View author publications

    You can also search for this author inPubMed Google Scholar

  9. Fang Tian
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

S.M. performed the experiment and data analyses and was a major contributor to writing the manuscript; C.L. contributed significantly to analysis and manuscript preparation; J.S. helped with analysis through constructive discussions; J.Y. contributed to the conception of the study; K.W., X.C., F.Z., X.S., and F.T. contributed to data collection; X.S. and F.T. contributed to manuscript revisions. All authors read and approved the final manuscript.

Corresponding authors

Correspondence to Xuequan Sun or Fang Tian.

Ethics declarations

Ethics approval and consent to participate

This prospective study was approved by the Institutional Review Board of the Tianjin Medical University Eye Hospital, Tianjin, China (No.2023KY-04). All procedures adhered to the tenets of the Declaration of Helsinki. All participants provided written informed consent before cataract surgery for the use of their clinical data.

Consent for publication

Not Applicable.

Competing interests

The authors declare no competing interests.

Additional information

Declarations.

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Ma, S., Li, C., Sun, J. et al. Assessing the interchangeability of keratometry measurements from four biometric devices in intraocular lens power calculations: insights into the predictive accuracy of five modern IOL formulas. BMC Ophthalmol 25, 236 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12886-025-04067-y

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12886-025-04067-y

Keywords

BMC Ophthalmology

ISSN: 1471-2415

Contact us