Vault Height Is a Key Predictive Factor for Anterior Segment Measurement Error by IOLMaster 700 in Eyes With Phakic Intraocular Lens

Abstract

Purpose: To identify risk factors of ocular anterior segment measurement error by the IOLMaster 700 in eyes implanted with an implantable Collamer lens (ICL).

Methods: In total, 152 patients with clear lens (152 eyes, group 1) and another 32 cataract patients (57 eyes, group 2) who underwent ICL implantation were included, and the presence of measurement error by the IOLMaster 700 was determined based on B-scan images. The risk factors for measurement error were evaluated by logistic regression, and the optimal threshold was determined using receiver operating characteristic analysis.

Results: The ICL was misidentified as the anterior surface of the crystalline lens in 51.97% of eyes (79/152) in group 1 and 80.70% of eyes (46/57) in group 2. For every 100-µm decrease in the vault height, a 3.57- and 5.78-fold increase in the risk of measurement error was observed in group 1 and group 2, respectively. We identified an optimal threshold of the vault height at 389.47 µm for predicting biometric measurement error in eyes implanted with ICL, which showed an area under the curve of 0.93 (95% confidence interval, 0.90–0.97), a sensitivity of 0.87, and a specificity of 0.86.

Conclusions: Patients with ICL implantation, particularly those with a vault height less than 389.47 µm, are at a greater risk of anterior segment biometric measurement error by the IOLMaster 700.

Translational Relevance: The threshold of vault height can help to identify high-risk patients and further optimize biometric measurement.

Introduction

Phakic intraocular lens (PIOL) achieved widespread popularity (more than two million implantations in 2022) for the correction of high refractive error due to many advantages, including rapid vision rehabilitation, preservation of accommodation ability, and reversibility.^1–3 However, studies showed that PIOL, especially posterior chamber PIOL (PC-PIOL), may accelerate the development of cataracts.^4–6 At 10 years after surgery, it has been reported that 17% to 18.3% of eyes required PC-PIOL explantation combined with cataract surgery.^7,8 

Correct ocular biometry measurement is vital for intraocular lens (IOL) power calculation in modern cataract surgery, but the presence of a PIOL can affect biometry measurement by ultrasound A-scan or partial coherence interferometry (PCI)–based biometers.⁹ Previous studies also observed that the IOLMaster 700 (Carl Zeiss Meditec AG, Jena, Germany), based on anterior segment swept-source optical coherence tomography, could misidentify the PC-PIOL as the anterior surface of the crystalline lens in 62.5% to 75.0% of eyes, resulting in an underestimation of anterior chamber depth (ACD) and overestimation of lens thickness (LT).^10–12 This notable and frequent measurement error will affect the doctors’ evaluation on the structure of the anterior segment, as well as the IOL power calculation, particularly new generation formulas that incorporate ACD and LT. Identification of patients with a higher risk of biometric measurement error could better guide clinical assessment, especially for those planning for cataract surgery. 

The present study aimed to investigate risk factors and further predict ocular anterior segment measurement error by the IOLMaster 700 for eyes with PC-PIOL implantation. 

Methods

The retrospective case series study was conducted at Zhongshan Ophthalmic Center, Sun Yat-Sen University, Guangzhou, China, with the consent of the Institutional Review Board/Ethics Committee of Zhongshan Ophthalmic Center (2022KYPJ095). All procedures were performed in accordance with the Declaration of Helsinki. 

Patients who underwent PC-PIOL (Visian Implantable Collamer Lens [ICL]; Staar Surgical, Monrovia, CA, USA) implantation with clear lens (group 1) and cataract (group 2) were included in this study. Patients with any evidence of the following conditions were excluded: (1) unavailable IOLMaster 700 measurement data; (2) unacceptable quality of IOLMaster 700 measurement, indicated by a yellow or red light; and (3) ocular comorbidities or history of ocular surgery, which could affect biometry measurement, including keratopathy, pars plana vitrectomy, and ocular trauma. 

The medical records of included patients were retrospectively reviewed, and the following data were collected: age, gender, history of diseases and surgeries, preoperative subjective refraction, the logarithm of the minimum angle resolution (logMAR) best-corrected visual acuity (BCVA), and biometry measurements by the IOLMaster 700 (including axial length [AL], keratometry, ACD, LT, and horizontal cornea diameter [CD]). 

The B-scan images obtained by the IOLMaster 700 were reexamined by the same ophthalmologist (FC) to evaluate if the ACD and LT measurements were correct. The green lines on the image represented structures automatically identified by the IOLMaster 700, including the anterior and posterior surface of the cornea and the crystalline lens (Fig. 1A). The anterior surface of PC-PIOL was misidentified as the anterior surface of the crystalline lens in some cases (Fig. 1B), for whom the ACD, LT, and vault height were remeasured manually by the same ophthalmologist (FC) using the ImageJ software (http://rsb.info.nih.gov/ij; National Institutes of Health, Bethesda, MD, USA), following methods described in our previous study.¹⁰ The vault height of the PIOL is the distance between the posterior surface of the PIOL and the anterior surface of the crystalline lens. 

Figure 1.

 

Correct (A) and wrong (B) anterior segment measurements by the IOLMaster 700. Green lines indicate structures automatically detected by the IOLMaster 700. Red lines show the vault height, measured from the posterior surface of the phakic intraocular lens to the anterior surface of the crystalline lens.

The quantitative data were displayed by mean ± standard deviation (SD) based on the Shapiro–Wilk test for normality of data, while the qualitative data were represented by number (percent). Independent t-test (for continuous variables) and χ² test (for categorical variables) were used to compare the participant characteristics between eyes with correct measurements and those with measurement error. Multiple logistic regression models were used to identify risk factors for measurement error. Factors with P < 0.05 in the univariable models were included in the multiple logistic regression model. The receiver operating characteristic curves (ROCs) were performed to determine the optimal threshold, and the area under the curve (AUC) was used to evaluate the efficacy of risk factors in predictively detecting measurement error. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were also calculated accordingly. All statistical analyses were performed with commercially available software (Stata version 16.0; StataCorp, College Station, TX, USA; R version 4.1.2, R Foundation, Vienna, Austria). A P value of <0.05 was considered statistically significant. 

Results

In total, group 1 included 152 patients (152 eyes, 129 females) with clear lens and ICL (version v4c), and group 2 included 32 cataract patients (57 eyes, 16 females) with prior ICL implantation (23 eyes with version v4c and 34 eyes with version v4). Table 1 lists the demographic and clinical characteristics of the included patients. Compared to patients in group 1, patients in group 2 were statistically older (43.77 ± 9.10 vs. 28.42 ± 5.11 years), had worse BCVA (0.76 ± 0.61 vs. 0.02 ± 0.10 logMAR), longer AL (30.96 ± 2.89 vs. 26.62 ± 1.34 mm), thicker LT (4.11 ± 0.40 vs. 3.83 ± 0.25 mm), and a smaller vault height (316.78 ± 185.74 vs. 420.01 ± 229.26 µm). 

Table 1.

 

Demographic and Clinical Characteristics of Participants

 

The rate of ocular biometric measurement error (i.e., underestimation of ACD and overestimation of LT) was 51.97% (79/152) with a mean error of 0.50 ± 0.17 mm in group 1 and 80.70% (46/57) with a mean error of 0.48 ± 0.17 mm in group 2, respectively. Table 2 exhibits the patient characteristics and ocular biometry between eyes with correct and incorrect biometric measurements. In group 1, eyes with measurement error were older and had a shorter AL, shallower ACD, thicker LT, smaller CD, and pupil diameter compared to eyes with correct measurements (all P < 0.01). The vault height was significantly lower in eyes with measurement error (265.83 ± 139.94 µm) than those with accurate measurements (586.87 ± 186.29 µm, P < 0.001). In group 2, only the distribution of vault height was statistically different between eyes with correct (591.61 ± 158.28 µm) and incorrect (251.06 ± 120.36 µm) measurements, whereas no differences were observed in other parameters. 

Table 2.

 

Characteristics Between Eyes With Incorrect and Correct Biometric Measurements

 

Logistic regression (Table 3) demonstrated that the only significant risk factor for measurement error was low vault height (odds ratio [OR] for per 100-µm decrease in vault heigh, group 1: 3.57, 95% confidence interval [CI], 2.33–5.56; group 2: 5.78, 95% CI, 2.10–15.90). Figure 2 shows the ROC curve in group 1, group 2, and the total population. In group 1, a cutoff of 389.37 µm for vault height distinguished eyes with correct and wrong measurements with a sensitivity of 0.84, a specificity of 0.85, a PPV of 0.86, and an NPV of 0.83. In group 2, the optimal cutoff value for vault height was 389.47 µm, which achieved an AUC of 0.96 (95% CI, 0.91–1.00), a sensitivity of 0.94, a specificity of 0.91, a PPV of 0.98, and an NPV of 0.77. The threshold of vault height that achieved the highest sensitivity and specificity was 389.47 µm in the total population. Figure 3 shows the distribution of vault height in eyes with correct and incorrect measurements, and the red line indicates the optimal threshold of vault height in each group. Cases of incorrect measurements with different vault heights are shown in the Supplemental Figure. 

Table 3.

 

Univariate and Multivariate Logistic Regression of the Risk Factors of Incorrect Anterior Segment Measurements by the IOLMaster 700

 

Figure 2.

 

The ROC curve of vault height in group 1 (A), group 2 (B), and the total population (C). The threshold of vault height that shows the highest overall sensitivity and specificity is displayed. sens, sensitivity; spec, specificity.

Figure 3.

 

Distribution of vault height in eyes with correct and incorrect measurements in group 1 (A) and group 2 (B). The rain cloud plots show the distribution of vault height (jittered raw data, a boxplot with the median, and the probability density). The red line indicates the threshold of vault height with the highest overall sensitivity and specificity.

Discussion

The present study demonstrated that after PIOL implantation, 51.97% of eyes with clear lens and 80.70% of eyes with cataracts would encounter an average 0.50-mm underestimation of ACD and overestimation of LT measured by the IOLMaster 700. The major risk factor for measurement error was low vault height, with the risk of measurement error increasing by 3.57- to 5.78-fold for every 100-µm decrease in vault height. A cutoff vault height value at 389.47 µm could effectively predict measurement error in eyes with PIOL implantation regardless of lens status with high sensitivity and specificity. 

It is known that the presence of PIOL could lead to a noteworthy ACD and LT measurement error by ultrasound A-scan and optical biometry. One previous study reported that compared with preoperative measurements, the ACD assessed by applanation ultrasonography was 1.07 to 1.31 mm shorter after two types of iris-fixed PIOL implantation, whereas a negligible difference (0.05 to 0.08 mm) was detected in the IOLMaster based on PCI technology.⁹ For eyes with PC-PIOL implantation, Amro et al.¹³ observed an average 0.27-mm difference between the preoperative and postoperative ACD measurements by the IOLMaster 500. To be noted, the PCI-based IOLMaster measured ACD using the lateral slit illumination approach, which could induce greater intersubject variability and could not indicate whether the PIOL is incorrectly identified as the anterior surface of the crystalline lens. In contrast, the IOLMaster 700 used a wavelength of 1055 nm with a total of 2000 A-scans taken per second. Subsequently, multiple adjacent A-scans were combined to form a B-scan image, allowing the user to assess ACD on the visual axis with high consistency and visually verify what anatomic structures are being measured.¹⁴ In this study, we detected an average 0.50-mm underestimation of ACD and overestimation of LT in clear lens eyes with PC-PIOL, which was consistent with a previous study (0.51 mm).¹¹ 

The risk of measurement error increased with decreasing vault height in our study, suggesting that the proximity of PIOL to the natural lens is the primary cause of measurement errors. Most cases of misidentification occurred in eyes with a vault height less than 389.47 µm (AUC, 0.93). According to the literature, the ideal range of vault height postoperatively is between 250 and 750 µm.¹⁵ Insufficient (<250 µm) or excessive (>750 µm) vault height may result in complications such as cataract or glaucoma. It was worth noting that the vault height of PIOL would reduce over time.^8,16–18 The annually thickening of the crystalline lens (24 µm per year) and ciliary muscle, as well as the erosion of the PC-PIOL into the sulcus, is implicated in the aforementioned process.^19–21 Choi et al.⁵ reported a mean vault height of 562.4 ± 175.9 µm in eyes with PC-PIOL implantation at 6 months postoperatively, decreasing to 352.9 ± 171.8 µm at 10 years. Based on the threshold in this study, it could be assumed that patients with PC-PIOL implantation longer than 10 years were at higher risk of ocular biometric measurement error. 

The underlying cause of this measurement error is the strong reflection signal induced by the implanted ICL. A relatively strong reflection peak is obtained at the medium interface. When multiple reflection peaks are close together, it is difficult for the machine to determine which is the true position of the optical medium. For example, PCI-based optical biometry could generate a double peak in eyes with macular disease, thus affecting axial length measurement.²² Similarly, the closer the ICL is to the crystalline lens (corresponding to a lower vault height), the easier it is for the program of the optical device to misidentify the ICL as the anterior surface of the crystalline lens. Our results suggest that the program of the IOLMaster 700 designed for eyes with PIOL needs to be further optimized with reference to the vault height threshold in the clinical setting. 

Several factors (including age, biometric parameters, and pupil size) were associated with the risk of measurement error in univariate regression but were no longer significant in the multiple regression models. This could be explained in part by the correlation between these variables and vault height. First, age was negatively correlated with the vault height due to the age-related decline in vault height.^8,16–18 Second, the ocular biometric characteristics (such as CD, AL, keratometry, and ACD) are crucial determinants of the PIOL position in the eye and thus have been used to predict the vault height.^23–26 Previous studies also revealed that vault height could vary with pupil size. Specifically, vault height decreased with pupil constriction, while the effect of cycloplegic medications on vault height differed between individuals.^27–29 Overall, the low vault height was the crucial risk factor for measurement error, whereas other factors were involved indirectly through their interaction with vault height. 

Regarding the impact of this ACD measurement error on IOL calculation, the findings of previous studies were inconsistent. Ouchi¹¹ reported the estimated IOL power calculated by the Haigis and Barrett Universal II formula was significantly lower using the ocular parameters measured by the IOLMaster 700 after PC-PIOL implantation, whereas other studies suggested that the PIOL did not affect the IOL power calculation.^10,13 It was worth mentioning that only highly myopic eyes, which were less sensitive to ACD measurement errors,³⁰ were involved in these studies. However, a growing amount of PC-PIOL has been used to correct high levels of hyperopia as well as low to moderate myopia.^31,32 The ACD measurement error could not be ignored for these individuals considering their relatively short AL and low vault height.^25,30 Additionally, myopia prediction error frequently happens in eyes with shallow ACD,^33,34 and thus surgeons might choose a hyperopic target for these patients. Misidentification of this measurement error that underestimated the ACD might result in a hyperopic prediction error in cataract patients with PIOL implantation. We found that more than 80% of cataract patients had incorrect ACD and LT measurements with a mean error of 0.48 ± 0.17 mm. Given the high incidence and the impact on IOL calculation and target refractive design, these errors need to be identified and corrected. 

Some limitations of this study should be addressed. This was a single-center retrospective study that included a relatively limited sample size, especially for the cataract cases, and exclusively Chinese patients. Moreover, the ocular biometric data were retrospectively evaluated and only assessed using the IOLMaster 700. The repeatability and reproducibility of the IOLMaster 700 in eyes implanted with ICL need to be investigated in prospective studies in the future. In addition, the rate and extent of measurement error in biometry measurement based on other techniques (e.g., Lenstar LS 900; Haag-Streit AG, Koenig, Switzerland) and their optimal cutoff value remain unknown. Nevertheless, the study provided important information that patients with a vault height less than 389.47 µm should be treated with caution regarding ACD and LT measurement by the IOLMaster 700. 

In summary, a low vault height was the major risk factor for anterior segment measurement error by the IOLMaster 700 in eyes with PIOL implantation. We also provided a threshold vault height value at 389.47 µm to help predict measurement error in clinical practice for both clear lens and cataract patients with proven high sensitivity and specificity. The results also provided preliminary guidance for optimizing the IOLMaster 700 in eyes with ICL implantation. 

Acknowledgments

Supported by the National Natural Science Foundation of China (No. 82070940, 82070941, 82101171) and the Construction Project of High-Level Hospitals in Guangdong Province (No. 303020102). The funding organization had no role in the design or conduct of this research. 

Disclosure: J. Zhang, None; F. Chen, None; X. Han, None; X. Qiu, None; Z. Liu, None; X. Chen, None; G. Jin, None; B. Qu, None; H. Yao, None; Y. Ye, None; K. Yu, None; X. Tan, None; L. Luo, None