Cataract surgery is one of the most frequently performed and highly successful medical procedures worldwide. In this procedure, an intraocular lens (IOL) replaces the opaque crystalline lens. Ocular biometry is essential for optimizing surgical outcomes and achieving high patient satisfaction. Over the past two decades, the evolution of premium IOLs, including toric, multifocal, and extended depth of focus (EDOF) lenses, has further emphasized the importance of obtaining highly accurate biometric measurements for refractive predictability and optimal visual rehabilitation.^(1,2)

Swept-source optical coherence tomography (SS-OCT) has emerged as a cutting-edge technology for ocular biometry, offering high-resolution imaging, superior penetration in dense cataracts, and increased speed compared to older optical coherence biometry (OCB) technologies.^(3,4) The Eyestar900 (Haag-Streit, Switzerland) is a fully automated SS-OCT-based biometer capable of anterior segment tomography, corneal topography, and axial length measurements using a 1,060 nm-wavelength and a scan speed of 30 kHz. It employs dual-zone keratometry (K) based on 32 infrared light-emitting diode sources (850 nm) and allows three-dimensional analysis of anterior segment structures, enhancing the accuracy of biometric measurements.^(5,6)

The IOLMaster700 (Carl Zeiss Meditec, Germany) is another SS-OCT-based biometer widely used in clinical practice, featuring a scanning rate of 2000 scans per second and a wavelength range of 1,035 to1,095 nm. It employs telecentric K that measures in three optical zones (1.5 mm, 2.5 mm, and 3.2 mm) and utilizes a refractive index of 1.3375 for keratometric calculations.^(7,8) The IOLMaster700 has been extensively validated in the literature and is often considered a reference device for biometric comparisons.^(9,10)

Several studies have evaluated the agreement of SS-OCT biometers, revealing minor yet statistically significant differences in K, axial length, and anterior segment parameters across different devices.^(11,12) However, discrepancies in anterior chamber depth (ACD) measurements have been reported as a potential limitation when considering the interchangeable use of these devices for IOL power calculation and phakic IOL sizing.^(13,14) Given that ACD variations can influence the prediction of postoperative effective lens position (ELP), even small differences in measurement techniques may have clinical implications, particularly for eyes with extreme axial lengths or previous refractive surgery.^(15,16)

Although previous research has assessed the agreement between various SS-OCT-based biometers, studies comparing the Eyestar900 and IOLMaster700 remain limited, particularly in specific populations. Given the anatomical and demographic variations that may influence biometric measurements, regional studies are necessary to validate the clinical applicability of these devices across different patient groups.^(17,18) The present study aims to evaluate the agreement of ocular biometric parameters between the Eyestar900 and the IOLMaster700 in a Brazilian cataract population. This study seeks to determine whether these devices can be used interchangeably in routine clinical practice by assessing the statistical and clinical relevance of measurement discrepancies.

This study included 43 eyes from 33 patients who underwent optical biometry with both devices. The biometric parameters obtained from each device are summarized in table 1. All parameters demonstrated high precision and excellent agreement between measurements, except for ACD, which showed greater variability.

Accurate ocular biometry is essential for optimal IOL power calculation and for achieving predictable refractive outcomes in cataract surgery. This study compared the Eyestar900 and the IOLMaster700, both SS-OCT biometers, to evaluate their agreement in key biometric parameters and their potential interchangeability in clinical practice.

The results demonstrated excellent agreement between the devices for most parameters, including AL, WTW, K, and LT. However, ACD showed lower concordance, which requires careful interpretation. The observed mean differences for these parameters were statistically significant but remained within clinically acceptable limits, confirming previous findings that swept-source optical coherence tomography (SS-OCT) biometers provide highly consistent measurements across different platforms. The strong correlation coefficients obtained further support the reliability of both devices for routine biometric assessments.^(1–3)

AL measurements exhibited nearly perfect agreement, with an ICC of 1.000 and a mean difference of 0.08 mm. These results align with previous studies comparing SS-OCT-based biometers, which have shown similarly high levels of reproducibility.⁽⁴⁾ Given that a 0.1 mm error in AL measurement corresponds to approximately 0.27 diopters of IOL power error, the minor discrepancy between the two devices is unlikely to impact clinical decision-making significantly. Furthermore, the tight limits of agreement in the Bland-Altman plot reinforce the interchangeability of these devices for AL measurement. Other studies have also demonstrated high repeatability in AL measurements between these devices, with differences that are not clinically significant.^(5,6)

Anterior K also demonstrated excellent agreement, with a mean difference of 0.03 D and an ICC of 0.990. Sorkin et al. ⁽¹⁶⁾ found that both devices exhibited high consistency in anterior K measurements, supporting their use in IOL calculations with the Barrett Universal II formula. Other investigations have reported similar findings, with limits of agreement within ± 0.5 diopters, indicating that these devices can be used interchangeably in clinical practice. The agreement in K measurements is critical for Toric IOL calculations, as even small variations can influence refractive outcomes. A study by Asena et al. further confirmed that the IOLMaster700 showed strong agreement with other keratometric devices, reinforcing its reliability.^(7–9)

The largest discrepancy between devices was observed in ACD, with a mean difference of 0.48 mm and an ICC of −0.081. The Bland-Altman analysis revealed wider limits of agreement, suggesting significant measurement variability. These findings are consistent with previous reports showing that ACD measurements can differ between biometers, likely due to differences in image acquisition methods and reference points for measurement.^(10,11) Given that a 0.5 mm-error in ACD can correspond to a refractive error of approximately 0.3-0.5 D in IOL power calculation, especially in premium lenses (e.g., EDOF, trifocal), this discrepancy is clinically relevant and should not be overlooked.

To mitigate the impact of ACD variability, two strategies may be considered in clinical practice: standardizing ACD measurements using a single device within the same institution and applying conversion equations derived from inter-device differences when measurements are obtained from different biometers. Additionally, cross-validation of IOL calculations using modern formulas such as Barrett Universal II can help minimize the influence of single-parameter variability, as these formulas incorporate multiple biometric values simultaneously. Recent evidence suggests that omitting parameters such as ACD, LT, or WTW in the Barrett Universal II formula results in minimal and clinically insignificant changes in predictive accuracy in eyes with normal biometry, provided that results are cross-checked with other formulas.⁽²⁷⁾ Moreover, a modified version of the Barrett Universal II incorporating total corneal power and virtual AL demonstrated improved refractive prediction accuracy in intermediate biometric profiles.^(28–31)

Lens thickness measurements showed excellent agreement between the two biometers, with a high intraclass correlation coefficient. Previous studies have also demonstrated that differences in LT measurements between the Eyestar900 and the IOLMaster700 are statistically significant but clinically negligible. Galzignato et al. reported that the agreement in measuring LT was strong, with small mean differences that do not impact IOL power calculations. These findings suggest that both devices can be used interchangeably for this parameter in routine clinical practice.^(12,13)

Regarding IOL power prediction, both the Eyestar900 and IOLMaster700 have been shown to yield comparable accuracy when using modern formulas such as Barrett Universal II. A recent study evaluating IOL power prediction accuracy between SS-OCT-based biometers found no statistically significant differences in spherical equivalent prediction error when comparing devices, with agreement in power selection within ± 0.5 diopters in most cases. Given the high degree of agreement in AL and K measurements, both devices are likely to provide similar predictive accuracy when used with contemporary IOL formulas. However, studies have noted that adjustments to lens factor calculations may be necessary when incorporating posterior corneal curvature into IOL power calculations.^{(2,3,14–26)}

Despite the overall high level of agreement observed, some biases were identified. The main concern was related to ACD, which showed greater variability between devices. This suggests that while AL, K, WTW, and LT can be used interchangeably, ACD should be interpreted with caution, particularly in cases where this parameter is critical for effective lens position estimation or phakic IOL sizing.^(2,3,9)

This study has limitations. Its retrospective design may introduce selection bias and non-standardized measurement conditions. Additionally, the relatively small sample size (n = 43) may reduce statistical power and limit generalizability, particularly for eyes with extreme AL (e.g., high myopia or hyperopia). Future prospective studies with larger and more diverse populations are recommended to determine whether inter-device differences, particularly in ACD, can influence IOL selection and postoperative refractive outcomes.^{(1,4,10,27–31)}

In conclusion, the Eyestar900 and IOLMaster700 demonstrated excellent agreement in most biometric parameters, supporting their interchangeability for routine clinical applications. However, the variability observed in ACD measurements highlights the need for caution when using this parameter for IOL power selection or phakic lens planning. Given the increasing reliance on SS-OCT biometers for preoperative planning, further research is warranted to refine calibration methods, develop conversion models, and optimize IOL prediction accuracy across different platforms.