Predicting the Postoperative Intraocular Lens Position Using Continuous Intraoperative Optical Coherence Tomography Measurements

Abstract

Purpose.: The aim of this study was to assess, if measuring the position of the lens capsule intraoperatively with a continuous intraoperative optical coherence tomography (OCT) device could be useful to improve the prediction of the intraocular lens (IOL) position.

Methods.: This prospective study included patients who were scheduled for cataract surgery. A prototype operating microscope with an integrated continuous OCT device was used to measure the anterior and posterior lens capsule position at different time points during cataract surgery. In all cases, a capsular tension ring (CTR) was used to tauten the lens capsule. Partial coherence interferometry was used to measure anterior chamber depth (ACD) immediately before, and 1 hour and 3 months postoperatively. Partial least squares regression (PLSR) was used to assess the influence of different pre- and intraoperatively measured parameters.

Results.: In total, 70 eyes of 70 patients were included. Mean axial eye length was 23.6 mm (range, 20.6 mm–30.8 mm), mean used IOL power was 22.2 diopters (D; range, 6.0 D–31.5 D). PLSR showed that the anterior lens capsule measured after removing the crystalline lens and after implanting a CTR was a significantly better predictor for the postoperative ACD compared with preoperative ACD measurements.

Conclusions.: The main problem of IOL power calculation, the prediction of the IOL position after surgery, could possibly be reduced by using intraoperative lens capsule measurements instead of preoperative ACD measurements. (ClinicalTrials.gov number, NCT01867541.)

Introduction

Postoperative emmetropia is the major determining factor for patient satisfaction after cataract surgery ¹ —and especially critical for functioning of modern intraocular lens (IOL) designs, such as aspherical, multifocal, and toric IOLs. However, after cataract surgery, approximately 9% ^2,3 to 20% ⁴ of all patients need a refractive correction of more than 1 diopter (D). Approximately 50% of this postoperative refractive error is the imprecise prediction of the postoperative IOL position, or postoperative anterior chamber depth (ACD). ⁵ Different IOL power formulae were developed to overcome this problem using the so-called effective lens position (ELP), which can be interpreted as a “fudge” factor used to attempt to optimize the formulae for empirical data. However, it does not directly correlate with the anatomical IOL position. Retzlaff et al., ⁶ Hoffer, ⁷ and Holladay et al. ⁸ used axial eye length (AL) and corneal power (K) for estimation of ELP and subsequent IOL power calculation. Haigis ⁹ used AL and preoperative ACD for estimation of ELP. Olsen ¹⁰ developed a thick lens formula using AL, ACD, crystalline lens thickness (LT), corneal radius (CR), and preoperative refraction in a prediction algorithm for ELP. Preussner et al. ¹¹ and Norrby ¹² developed ray tracing programs for IOL power calculation purposes. However, these methods cannot predict with certainty either the IOL position immediately after surgery, or the ACD shift within the first months after surgery. 

The aim of this study was to assess whether measuring the position of the lens capsule intraoperatively, after removing the crystalline lens, could be useful to improve the prediction of the IOL position. For this purpose, a prototype of an anterior segment time-domain optical coherence tomography (OCT) device connected to an operating microscope was used. 

Patients and Methods

This prospective study included patients, who were scheduled for cataract surgery. Thirty short eyes (emmetropic IOL power, >23.0 D); thirty “normal” eyes (emmetropic IOL power, 18.0–23.0 D); and 10 long eyes (emmetropic IOL power, <18.0D) were included. Exclusion criteria were preceding ocular surgery or trauma, pseudoexfoliation syndrome, intraoperative floppy iris syndrome, and a preoperative visual acuity of less than 0.05 Snellen. All the research and measurements followed the tenets of the Declaration of Helsinki and were approved by the local ethics committee. Written informed consent was obtained from all patients in the study. 

Preoperatively, the eye to be operated was examined at the slit-lamp. Optical biometry was performed using commercial optical devices (IOLMaster 500 and ACMaster; Carl Zeiss Meditec AG, Jena, Germany); the scan was performed to assess AL, preoperative ACD, and lens thickness (LT). 

During surgery, continuous intraoperative anterior segment OCT (AS_CI OCT) imaging was performed and digitally recorded. Screenshots of the videos were taken in after surgery. 

Surgery was performed under topical anesthesia in all cases by one experienced surgeon (OF). The first AS_CI OCT screenshot was taken at the beginning of surgery (“phakic,” Fig. 1). A self-sealing clear corneal incision with a 2.8-mm single-beveled steel blade was performed. The incision was followed by the injection of an ophthalmic viscoelastic device (OVD), capsulorhexis, phacoemulsification, and irrigation/aspiration of cortical material. The OVD was removed carefully and a second AS_CI OCT screenshot was taken with the infusion hand piece of the bimanual I/A set placed into the paracentesis and the bottle height adjusted to control the pressure in the anterior chamber (as explained below). Then, after instillation of the OVD, a capsular tension ring (CTR) with a diameter of 11 mm (Morcher, Germany) was implanted; and after complete OVD removal, a third AS_CI OCT screenshot was taken again under continuous irrigation (“CTR,” Fig. 1). In all cases, a hydrophobic acrylic IOL with an offset haptic design (ZCB00) was implanted into the capsule bag after OVD instillation using a commercial injector (Emerald AR; Abbott Medical Optics, Santa Ana, CA). Following the implantation of the IOL, the OVD was aspirated thoroughly using a bimanual I/A set. Care was taken to completely remove OVD from behind the IOL by slightly displacing and tilting it and reaching behind the optic with the aspiration cannula. At this time point, the fourth AS_CI OCT screenshot was taken (“IOL,” Fig. 1). Postoperative therapy within the first month was ketorolac gtt (Acular; Allergan, Irvine, CA) three times daily for 1 month. 

Figure 1

 

Intraoperative AS_CI OCT measurements at different time points: anterior segment at the beginning of surgery (phakic), after implanting a CTR (CTR) and after implantation of an IOL (IOL). *anterior lens capsule. #center of the anterior surface of the IOL.

Outside the operating theater, partial coherence interferometry (PCI) measurements (PCI meter; ACMaster; Carl Zeiss Meditec AG) of the anterior chamber depth were performed preoperatively, and 1 hour and 3 months postoperatively. Additional examinations at 3 months postoperatively were autorefraction in IOL mode (Topcon RM 8800; Topcon Corporation, Tokyo, Japan) and subjective refraction using the cross-cylinder method. 

Intraoperative AS_CI OCT Setup and Analysis

Requirements for AS_CI OCT measurements of the lens capsule are high resolution and deep penetration depth. These parameters are defined by the wavelength and the bandwidth: a shorter wavelength and a wider bandwidth, result in a higher resolution, but a reduced penetration depth due to scatter. ^13,14 For this study, a time domain OCT with a wavelength of 1310 nm, an axial resolution of 18 μm, and an acquisition time of 0.5 seconds (Visante; Carl Zeiss Meditec AG) was used. This AS OCT was shown to be highly reproducible for measurements of the ACD and of the IOL/crystalline lens. ^15,16 Until recently, the AS OCT was only available as a standalone device to be used pre- and postoperatively. In this project, a prototype of a combination of an AS_CI OCT and an operating microscope (OPMI 200; Carl Zeiss Meditec AG) was used to allow continuous measurements of the crystalline lens, the lens capsule itself, and furthermore, the position of the IOL intraoperatively. To assure imaging at the center of the cornea, a cross-hair was introduced into the eyepiece of the operating microscope so that the surgeon knew the exact location of the OCT scan axis during surgery. During surgery, the surgeon guided the patient to look straight into the operating microscope light, and then centered the cross-hair, on the corneal apex and the Purkinje I reflex. For further analysis, the AS_CI OCT measurements were recorded together with the synchronized “2D view” video that represented the surgeon's view without the cross-hair. Special care was taken to align the AS_CI OCT scan with the continuously recorded 2D view before the trial started and alignment was checked at regular time points during the study. All screenshots were imported into a graphics editing program (Photoshop CS4; Adobe Systems, Mountain View, CA), with the purpose of determining distances in pixels. These distances were converted into mm using a reference distance and afterward, a correction for the refractive index was performed. For horizontal distances, we used the known diameter of an IOL in five different cases and for depth distances, we used the central corneal thickness measured with PCI technology before surgery in another five patients). For all intraoperative measurements the distance between the endothelium of the cornea and the anterior and the posterior lens capsule were analyzed. Therefore, the cornea itself was not included in the measurements. For analysis, the distance between the center (in the x/y axis) of the cornea and the center of the posterior lens capsule was used. For the anterior lens capsule, no direct measurement of the center was possible; therefore, a line was drawn from one edge of the rhexis to the other edge. The osculation point of this line and a line perpendicular to the center of the cornea was taken for further analysis. 

One of the main challenges of intraoperative ACD measurements was to measure at a normal intraocular pressure (IOP). Therefore, the irrigating handpiece was held into one of the paracenteses, and the height of the infusion bottle was adapted, as followed: We attempted to standardize the intraocular pressure (IOP) at 2666 Pa (20 mm Hg). Due to the conversion factor of 1.36 (from mm Hg to cmH₂O) the height of the infusion bottle should be 27.2 cm above the patient's eye. Before starting the trial we measured the IOP of seven patients intraoperatively, increasing the height of the infusion bottle in 5-cm steps and measuring the IOP with a Schiötz tonometer (Riester, Tuttlingen, Germany) intraoperatively. For exact measurements we measured the height of the patient's eye and the actual height of the infusion bottle on the phaco-machine. During the study, the height of the operation table was measured during surgery to assure a standardized setting concerning IOP during the measurements. Another challenge was to ensure a taut and ‘straight' planar posterior capsule. Therefore, we introduced a CTR in all patients to cause a tauter capsule. 

Partial Coherence Interferometry (PCI) Measurements

As shown by Findl et al., ¹⁷ PCI technology showed an accuracy of 3 μm for measurements of ACD in pseudophakic eyes. Each peak of a PCI scan gives the optical path length (OPL) between two surfaces with a different refractive index (RI), but not directly anatomical distances. ¹⁸ To overcome this problem, each single PCI scan was analyzed separately and each peak in each scan was marked manually. Additionally, the mode n = 1 was used to analyze each OPL with an RI of 1. After detecting each peak in all scans manually, all measured peaks were entered into a spreadsheet file (Excel; Microsoft Corporation, Redmond, WA). Afterwards, all peaks within 0.1 mm were grouped and allocated regarding their anatomical structure (anterior surface of the cornea, posterior surface of the cornea, anterior surface of the lens, and posterior surface of the lens). Finally, OPL was converted to ACD using the appropriate group refractive index for aqueous at 850 nm (1.3454). ¹⁸  

Statistical Analysis

For statistical analysis, spreadsheet software (Excel; Microsoft Corporation) for Mac (Apple Inc., Cupertino, CA) with commercial plug-ins (StatPlus:mac version 5.8.3.8; AnalystSoft, Vancouver, BC, Canada; and XLSTAT; Addinsoft, Inc., Brooklyn, NY) was used. For missing data, observations were excluded from analysis. Descriptive data are always shown as mean, SD, and range. 

For statistical modeling, partial least squares regression (PLS) was performed with data analysis software (XLSTAT; Addinsoft, Inc.). Advantages of PLS regression are explained in the discussion section; here we only want to introduce the outcome variables of PLS regression. Cumulated Q ² gives the contribution of a component (or some components) to the predictive quality of a model. Variable importance for projection (VIP) measures the importance of an explanatory variable to explain the dependent variable (more precisely: not the dependent variables, but the t-scores that contain compressed explanatory variables). More relevant for clinicians are the suggested thresholds of the VIPs ¹⁹ : VIP of >0.8 to <1.0 means that the explanatory variable moderately influences the model and values of 1.0, or more means that it highly influences the regression model. 

To evaluate the regression model, a bootstrap method ²⁰ was used to estimate the weighting of each explanatory variable. This method avoids the bias of the good fit of the model for the data it has been derived from and can be seen as simulating the real-world situation: all data from all patients except one are used to create a PLSR model. Afterward it is observed, if the PLSR model can be used to explain this one patient. Afterward this process is repeated for each patient in the data set. The result of this bootstrapping method is then shown in standardized coefficients (beta coefficients) plots. For interpretation purposes: the larger the absolute value of a coefficient, the larger the weight of the variable and if the confidence interval (whiskers) includes 0, the weighting of the variable is not significant. Scatter plots were used to show the correlation of the observed and the predicted dependent variables. 

Results

In total, 70 patients (60% female and 40% male) were included in this study. Descriptive data are depicted in Table 1. After removing the crystalline lens, AS_C OCT measurements of the anterior lens capsule were possible in all cases, and measurements of the posterior lens capsule in 93% of all cases. In total, 12,600 PCI scans were performed, but in 7% of all patients, no clearly identifiable peaks could be detected. 

Table 1. 

 

Demographic Data of the Study Population

	Mean (SD)	Range
Age, y	72.3 (11.1)	41–100
AL, mm	23.6 (1.9)	20.6–30.8
Mean K, D, RI: 1.332	43.3 (1.6)	39.6–47.2
Pre-OP LT, mm	4.2 (0.6)	3.2–7.5
Pre-OP ACD, mm	2.8 (0.5)	1.9–4.0
IOL power, D	22.2 (4.9)	6.0–31.5

Descriptive Analysis of Intraoperative Measurements

ACD measurements with the AS_CI OCT at the beginning of surgery (“phakic”) correlated well with preoperative ACD measurements performed with PCI (r ² = 0.9, P < 0.001; Fig. 2) and all except 10 measurements were within the limits of agreement (1.96×SD). Intraoperative AS_CI OCT measurements resulted in larger distances compared with PCI measurements. The conversion factor was found to be 1.13 × distance_(PCI) = distance_(OCT). For the regression model, the original values were used and no correction factor was introduced in the formula. The distances between the endothelium of the cornea and the anterior and posterior lens capsules are shown in Table 2. Intraoperatively, IOL position was measured significantly deeper, compared with 1 hour postoperatively (mean difference 0.6 mm, SD: 0.7 mm; P < 0.001 ANOVA). 

Figure 2. 

 

Correlation of preoperatively measured ACD with partial coherence interferometry (preoperative ACD [PCI]) and intraoperatively measured ACD with AS_CI OCT scans (intraoperative ACD “phakic” [AS_CI OCT]).

Table 2. 

 

Distances Between the Endothelium of the Cornea and the Anterior and Posterior Lens Capsules in μm at 4 Different Time Points

	Anterior Capsule	Posterior Capsule
Beginning of surgery	3.2 (SD: 0.6; 1.7–4.5)	8.5 (SD: 1.2; 7.3–10.0)
After phaco	4.8 (SD: 0.6; 3.2–6.0)	6.5 (SD: 1.6; 3.2–9.7)
With CTR	5.1 (SD: 0.6; 3.6–6.8)	6.5 (SD: 1.2; 3.7–10.3)
With IOL	5.5 (SD: 0.7; 4.2–8.0)	6.5 (SD: 0.7; 5.0–8.7)

Which Intraoperative Measurement Is the Best Predictor of the Postoperative ACD?

For intraoperative measurements (“aphakic” and “CTR”) the anterior (aphakic_a; CTR_a) and posterior lens capsule (aphakic_p; CTR_p) was included in a PLS model to predict the postoperative ACD. As shown in Figure 3 (left), CTR_a showed the highest VIP and, furthermore, it performed best concerning standardized coefficients (Fig. 3, right). Although the aphakic measurement of the anterior lens capsule (without CTR) showed a good prediction, it was eliminated from further analysis and only CTR_a was used. This can be justified, because in a model with only aphakic_a and CTR_a, the aphakic_a measurement showed a VIP below the 0.8 level. Additionally, it should be mentioned that measurements of the posterior lens capsule were poor predictors for the postoperative ACD (Fig. 3, left). 

Figure 3. 

 

Influence of intraoperative measurements (explanatory variables) on the postoperative ACD (dependent variable). VIP plot (left) and standardized coefficients (right) shows the intraoperative measurements of the anterior and posterior lens capsule. Standardized coefficients show the weighting of the explanatory variables (the larger the blue box, the higher the weighting). Additionally, the whiskers show the 95% confidence interval estimated using bootstrap methods.

Prediction of the 1 Hour Postoperative IOL Position

Firstly, prediction power of the postoperative ACD was compared between the preoperatively measured ACD and the CTR_a. VIP of preoperative ACD measurements and CTR_a was 1.15 (95% CI < 0.01) and 0.83 (95% CI < 0.01), respectively (Fig. 4). 

Figure 4. 

 

Variable importance for projection (VIPs; left) and standardized coefficients (right) of preoperativley measured ACD and CTR_a on the ACD measured 1 hour after surgery.

For ACD prediction purposes, AL, CTR_a, ACD, and LT were included. VIPs and standardized coefficients are shown in Figure 5. VIP for lens thickness was poor and the 95% confidence interval of the bootstrapping validation included 0, meaning that the influence of this explanatory variable was not significant. Additionally, the 95% confidence interval of the preoperatively measured ACD included zero. Therefore, lens thickness and preoperatively measured ACD were not included in the model. Cumulated Q ² for AL plus CTR_a and only CTR_a was 0.10, respectively. 

Figure 5

 

Variable importance for projection (VIPs; left) and standardized coefficients (right) of different explanatory variables on the ACD measured 1 hour after surgery.

Prediction of the ACD shift in the first 3 months after surgery was generally poor. For none of the PLS regression models cumulated Q ² was higher than 0.06. Only lens thickness showed to have a moderate predictive power (VIP: 1.41), but Q ² was 0.03. 

Prediction of the 3-Month Postoperative IOL Position

For ACD prediction AL, CTR_a, ACD, and LT were included. As shown in Figure 6, LT was a poor predictor and was not included in the model (VIP: 0.35), whereas AL, CTR_a, and preoperatively measured ACD showed significant VIPs of 1.18, 1.13, and 1.09, respectively. Standard coefficients for AL, CTR_a, preoperatively measured ACD and LT were 0.39 (0.15–0.64); 0.35 (0.10–0.61); 0.22 (0.01–0.42); and 0.26 (0.03–0.50), respectively (Fig. 5, right). Due to the fact that the lower 95% CI bound of preoperatively measured ACD was almost 0, the weighting of this explanatory variable was low. 

Figure 6. 

 

Variable importance for projection (VIPs; left) and standardized coefficients (right) of different explanatory variables on the ACD measured 3 months after surgery.

The PLS regression formula (not including LT) was:  

 

 

Discussion

To our knowledge, this is the first study dealing with postoperative ACD prediction using intraoperative biometrical measurements of the capsule positions. OCT measurements for IOL power calculation were used in the past, but only for preoperative measurements, and results were shown to be similar to those obtained with the optical device (Carl Zeiss Meditec AG) when using the SRK/T IOL power formula. ²¹ Concerning intraoperative measurements, two different technologies were used for IOL power calculation recently: intraoperatively performed autorefraction showed to implicate several problems, such as alignment problems, because the patients couldn't see the fixation target and the use of OVDs made measurements impossible. ²² The second more recently introduced technology is intraoperatively measured wavefront aberrometry. However, this technology was found to be inaccurate for this purpose due to several problems, such as residual viscoelastics in the anterior chamber and tear film problems. ²³ Additionally, these methods have the same problem as conventional IOL power calculation formulae: they use fudge factors to predict the postoperative IOL position as well as the postoperative ACD shift. 

This study showed that continuous intraoperative measurements of the lens capsule after removing the lens were a powerful tool to predict the postoperative IOL position. In this study, best intraoperative prediction factor was the anterior lens capsule after implanting a CTR (CTR_a), followed by the anterior lens capsule measurement after removing the crystalline lens without a CTR. The posterior lens capsule was a poor predictor at all measurement time points. The reason could be that the posterior lens capsule was to a higher extent influenced by the variable hydration of the vitreous than the anterior lens capsule. Additionally, a significant difference between the ACD measured at the end of surgery and the ACD 1 hour after surgery was found. This observation is similar to findings by Erickson. ²⁴  

In our model including preoperatively measured ACD together with CTR_a showed a better postoperative ACD prediction compared with using only preoperatively measured ACD, or only CTR_a.. However, comparing both explanatory variables, CTR_a was found to be a better predictor than preoperatively measured ACD. Concerning the postoperative IOL position 1 hour after surgery, axial eye length and CTR_a were excellent predictors, preoperatively measured ACD was a good predictor, and lens thickness was a poor predictor. These findings are similar to findings by Sverker Norrby (unpublished observations, 2010), but contrary to observations by Olsen et al. ¹⁰ Lens thickness had a slightly higher, but still low influence, on the IOL position at the 3 months follow-up in this study, compared with the measurement performed immediately after surgery. Additionally, lens thickness had a higher correlation to the ACD shift within the first 3 months after surgery compared with all other factors (axial eye length, preoperative ACD, CTR_a), but this relation was not significant and a large scatter between observations was found. Our model suggests that lens thickness should not be included in a general prediction model, but it could possibly be useful to detect patients with a very unusual lens thickness and, therefore, a potential for a larger ACD shift in the first months after surgery. However, our sample size was too small to define cutoff values for this purpose. 

Comparing the impact of CTR_a and preoperatively measured ACD, it should be mentioned that CTR_a measurements were much more powerful for predicting the IOL position 1 hour after surgery, but the difference in prediction power between CTR_a and preoperatively measured ACD was smaller, at 3 months' follow-up. This means that the position of the anterior lens capsule of the aphakic eye corresponds very well with the IOL position immediately after surgery and this is the advantage of the used setup. However, in some patients, where a relevant ACD shift was observed within the first weeks after surgery, this advantage is reduced. This would also explain the reduced advantage of the CTR_a measurement compared with the preoperative ACD measurement at 3 months' follow-up. However, CTR_a measurements were superior to preoperatively measured ACD measurements at all time points. 

In this study, we avoided first generation modeling techniques, such as multiple linear regression models that were used in the past to predict the postoperative ACD for several reasons. One main problem of multiple regression modeling is “overfitting.” This means, the more explanatory variables are used in an ACD prediction model, the better it will appear (R ² will increase), although this will not be the case for the subjects outside the sample the regression was derived from. This is especially a problem if the sample size is very large, because then almost every explanatory variable in the prediction model will be significant and included in the model. Another problem with multiple linear regression is that linear independence of explanatory variables is required. However, it is well known that there is a linear relationship between some variables, such as preoperatively measured ACD and axial eye length. Another disadvantage is that one of the assumptions of multiple linear regression is that all measured values have to be completely reliably determined, which is often not the case in the clinical setting, especially due to missing data, or inaccurate measurements. Structural equation models, such as principal component analysis (PCR) and PLS regression, can deal with the overfit and the collinearity problem, as well as the “reliability of all measurements” problem, at least to a higher extend compared to conventional multiple linear regression. ²⁵ Additionally, PLS does not only explain the conditional distribution of the dependent variable given the explanatory variable, but uses both as random variables, connected by a latent variable and a smaller sample size is needed compared with other regression methods. Furthermore, PLS is a variance based structural equation model that not only includes a structural part to explain influences of variables to each other, but also performs a weighting of variables. This weighting has two advantages: first, it is more accurate, because it is unlikely that all variables have the same weighting; and second, it takes care of variables that are less reliable. ²⁶ PLS can cope with a large number of explanatory variables, because all explanatory variables are compressed into so-called t-scores. These t-scores then replace the original explanatory variables-matrix in a subsequent regression step. ²⁷ The prediction quality is given by cumulated Q ² that is lower, or equal R ², and not depending on the number of explanatory variables. 

To summarize, in the future it could be possible to use the intraoperatively measured position of the anterior lens capsule after removing the crystalline lens to better predict the postoperative IOL position. This would be a large step forward in reducing the number of outliers and the variability of refractive outcomes after routine cataract surgery. However, the entire process of intraoperative measurements is still time consuming, and currently all steps have to be performed manually. To be used routinely, it would be necessary to semi-automate this method and to integrate IOL power calculations based on these intraoperative measurements. 

One problem that has to be tackled is the variable hydration of the vitreous. This problem can be minimized, but not eliminated by measuring the anterior and not the posterior lens capsule and by using a CTR. One option to completely correct for this error could be to measure not only the anterior segment, but also the posterior segment to detect changes of the vitreous. For this purpose, it could be sufficient to use an A-scan for the posterior segment additionally to the B-scan for the anterior segment, if the assumption is met that the patient's eye is aligned. Another aspect that should be addressed is that 3D OCT imaging with Fourier domain OCT devices would have several advantages, such as improving the alignment of the measurement, improving the resolution, and reducing the acquisition time. Furthermore, with this method tilt, an IOL could be taken into account, as shown by Ortiz et al. ²⁸ Another option would be to use swept source OCT imaging, preferably using ultralong range full eye imaging, as shown by Grulkowski et al. ²⁹  

What was known: 

•  
  Prediction of the postoperative IOL position is the main source of refractive error, especially in short eyes;
•  
  Different IOL formulae try to overcome this problem by using preoperatively measured parameters, such as anterior chamber depth, or lens thickness.

What this paper adds: 

•  
  Intraoperative optical coherence tomography measurements of the anterior capsule are a better predictor of the postoperative IOL position compared with preoperatively measured factors;
•  
  Best prediction of the postoperative IOL position was observed when using measurements of the anterior lens capsule after implanting a capsular tension ring.

Acknowledgments

Special thanks go to Sverker Norrby, PhD, for his advice and helpful suggestions. 

Presented at the annual meeting of the Association for Research in Vision and Ophthalmology, Fort Lauderdale, Florida, May 2012. 

Supported by the “Jubilaeumsfond” of the Austrian National Bank (project number 14052). The authors alone are responsible for the content and writing of the paper. 

Disclosure: N. Hirnschall, None; S. Amir-Asgari, None; S. Maedel, None; O. Findl, None