Use of Scheimpflug corneal anterior-posterior imaging in ray-tracing intraocular lens power calculation

Abstract

Purpose: To examine improvement with the use of Scheimpflug imaging of the anterior and posterior corneal surfaces in the accuracy of ray-tracing intraocular lens (IOL) power calculation for normal cataractous eyes. Methods: Prospective case series comprised 136 eyes of 136 consecutive patients who had undergone cataract surgeries. Scheimpflug imaging of the cornea was included with routine preoperative examinations. Postoperative refractions were predicted using three methodologies; ray-tracing calculation using Scheimpflug imaging and Placido topography, ray-tracing calculations using Placido topography, and the SRK/T formula using autokeratometry. Prediction errors from the manifest refraction at 1 month postoperatively were compared between the methods. Influence of the posterior corneal curvature was also evaluated. Results: Mean prediction errors were 0.008, −0.103 and −0.042 D, respectively without significant difference between the three methods (p = 0.23). The prediction errors were significantly correlated with the posterior corneal curvature when the Scheimpflug imaging was not used (p < 0.03). Conclusion: Use of Scheimpflug imaging in ray-tracing IOL power calculation was as accurate as the other calculations in normal cases, showing no bias in the posterior corneal curvature, as is the case with the other calculations. Use of a multifocal intraocular lens (IOL) or a toric IOL necessitates an accurate IOL power calculation, because the patient anticipates good postoperative visual acuity without spectacles. Theoretically, involving the posterior corneal curvature (PCC) and central corneal thickness (CCT) would improve the accuracy (Olsen 2007) in the power calculation. However, significant improvement was not reported in the previous papers (Savini et al. 2009, 2011; Shammas et al. 2009). Equivalent keratometry (Eq-K) and true net power (TNP) obtained from a Scheimpflug imaging of the anterior and posterior corneal surfaces with the Pentacam (Oculus, Wetzlar, Germany) did not improve the accuracy for unoperated cataractous eyes (Shammas et al. 2009), whereas they were effective for eyes after myopic laser in situ keratomileusis (LASIK) (Holladay et al. 2009; Kim et al. 2009). Combination of dual Scheimpflug photography and Placido topography in the GALILEI system (Ziemer Ophthalmic Systems, Port, Switzerland) did not improve the accuracy in normal eyes (Savini et al. 2011). Utilizing ray tracing is a better way to describe the pseudophakic eye optics (Olsen 2007). An IOL power calculation, OKULIX (Tomey, Nagoya, Japan), performed ray tracing through a pseudophakic eye used in a TMS topography system (Tomey) (Preußner et al. 2002, 2005). For the ray-trace calculation, physical curvatures, thickness and refractive index of the particular IOL design, that were not usually divulged by the IOL manufacturer, were required. The accuracy of OKULIX was comparable with the SRK/T formula in normal cataractous eyes (Preußner et al. 2005, 2008), although it significantly improved in postcorneal surgery cases (Rabsilber et al. 2007). The ray-tracing calculation enables incorporate additional measurements such as PCC and CCT. The purpose of this study was to evaluate improvement with the use of a Scheimpflug imaging of the anterior and posterior corneal surfaces in the ray-tracing IOL power calculation in normal cataractous eyes. This prospective case series comprised 230 eyes of 160 patients who underwent cataract surgeries at Miyata Eye Hospital between December 2009 and December 2010. For evaluation, 136 eyes of 136 patients who obtained best corrected distance visual acuity (BCVA) of 0.15 logMAR (20/30 in Snellen) or better at 1 month postoperatively and topography measurement with a good fixation and corneal surface quality were selected. This study followed the tenets of the Declaration of Helsinki and was approved by the institutional review board of Miyata Eye Hospital. Demographic data of the subjects are shown in Table 1. Before surgery, axial length (AL) and autokeratometry (Auto-K) were measured with a partial coherent pachymetry, IOLMaster version 5 (Carl Zeiss Meditec, Jena, Germany), and an autokeratometer ARK-730A (Nidek, Aichi, Japan), respectively. As routine preoperative examinations, a slit-lamp examination and measurements of intraocular pressure were performed to rule out abnormality. Corneal topography was measured with TMS-5 version 6B (Tomey) that incorporates a Placido topography for measurement of the anterior corneal surface and a rotating Scheimpflug photography for 3-dimensional imaging of the anterior and posterior corneal surfaces. Distortion in the Scheimpflug image due to the camera configuration was geometrically corrected for the anterior surface and then pachymetry mapping data were generated using ray tracing (Richards et al. 1988). The Placido topography was merged with the anterior surface profile in the Scheimpflug imaging to compensate for poor quality or missing portions in the Placido topography. The Scheimpflug image had a distortion in corneal peripheral area due to the strong refraction on the corneal surface, as shown in Fig. 1 (Khurana et al. 2007; Nakagawa et al. 2011). The distortion was corrected by a comparison with the height reference that was measured with an anterior segment optical coherence topographer (AS-OCT) for restoring the corneal asphericity. Scheimpflug image of anterior segment cornea. Edge of the anterior cornea was sharp around the corneal centre and dull on the peripheral due to refraction on the cornea surface. All cataract surgeries were conducted by a surgeon (Kazunori Miyata) with cataract removals by phacoemulsification through 2.4-mm scleral incisions and implantations of IOL NY-60 (HOYA, Tokyo, Japan) in the capsular bags. IOL power was determined with the SRK/T formula using the AL and Auto-K measurements and the A-constant of 118.6. BCVA and manifest refraction spherical equivalent (MRSE) were measured at 1 month postoperatively. Postoperative refractions for the implanted IOL were predicted using three methods: ray-tracing calculation using the Scheimpflug imaging and Placido topography (RayTrace-S), ray-tracing calculation using the Placido topography (RayTrace-P), and the SRK/T formula using Auto-K (SRK/T). Ray-tracing calculation was performed with OKULIX version 8.45, which conducted a single numerical ray tracing from the retina to the IOL and cornea (Preußner et al. 2002, 2005). Corneal curvatures of the anterior and posterior surfaces were calculated from topography data in areas of 6 mm diameter. In the RayTrace-P, PCC was obtained with the anterior curvature multiplied by 0.83 and CCT was assumed as 550 μm. The IOL position was predicted with the AL measurement with the regression equation (Preußner et al. 2002, 2005), because the postoperative IOL position was significantly correlated with the AL (Preußner et al. 2008). Curvatures and thickness of the IOL was retrieved from a database within the software. Predicted postoperative refraction for a best focus was calculated for a default pupil diameter (2.5 mm). Prediction errors between the postoperative MRSE and the predicted postoperative refraction, and its absolute values were calculated using three methods. The prediction errors were compared by the one-way anova and ad-hoc Bonferroni’s test. The mean and median of the absolute error was obtained with examinations of anova and Kruskal–Wallis test, respectively. Percentages of the prediction errors within ± 0.5 D and ± 1.0 D were also obtained and examined by the chi-square test for trend. Actual measurement of the PCC was used in the RayTrace-S only, while the others did not include the PCC measurements. Influence of the PCC on the prediction error was evaluated by linear regression analysis. There have been frequently observed eyes with small corneas or tight eye lids in elder Asian patients. In the ray-tracing calculation, topography data within 6 mm diameter were required for a corneal curvature calculation. When the measured area was <6 mm diameter, the calculated curvature was less accurate. Measurement of the posterior corneal surface was more restricted due to image configuration of the Scheimpflug camera. Diameters where complete topography data were available were measured for 86 of the 136 eyes. Results are expressed by mean ± SD (standard deviation). p < 0.05 was considered as statistically significant. The postoperative uncorrected visual acuity (UCVA) and BCVA were 0.137 ± 0.286 and −0.090 ± 0.102 logMAR, respectively, and the MRSE was −0.458 ± 1.006 D (range, −5.75 to 1.75 D). IOL power prediction errors of the SRK/T and the ray-tracing calculations are shown in Table 2. There was no significant difference (p = 0.23) between the three methods. In the absolute errors, the ray-tracing calculations were lower than SRK/T in the mean and median, although there was no significant difference (p = 0.19). Percentage of refraction errors within ± 1.0D and ± 0.5D are also shown in Table 2 with no significant difference (p = 0.88 and 0.98). Influence of the PCC is shown in Fig. 2. The measured PCC was 6.46 ± 0.33 mm (range, 5.83–8.07). The prediction errors of RayTrace-P varied with the PCC; however, the correlation was weak (p < 0.0001, R2 = 0.114). There was no association in RayTrace-S (p = 0.33, R2 = 0.007). Effect of posterior corneal curvature on refraction error with ray-tracing calculation using Placido topography (RayTrace-P, •) and using Placido and Scheimpflug topography (RayTrace-S, ○). There was a significant correlation in RayTrace-P (solid line), while there was no association in RayTrace-S (dotted line). Diameters of the measured areas in the Placido topography and the posterior Scheimpflug imaging were 6.36 ± 0.91 mm and 5.06 ± 1.04 mm, respectively. Figure 3 shows the distribution and cumulative percentage of each corneal surface. Areas with 6 mm diameter or larger were 73% in the anterior corneal surface, but only 15% in the posterior. Distribution and cumulative percentage of measured area in the anterior surface measured with Placido topography (black) and the posterior surface measured with Scheimpflug imaging (while); bars and the left axis denote distribution of eyes, and solid line and the right axis denote the cumulative percentage. In comparing the three IOL power calculation methods, there was no significant difference in the accuracy. Results in the current study were comparable with the study by Shammas et al. (2009). However, ray-tracing calculation using Placido and Scheimpflug topography (RayTrace-S) was not dependent upon PCC. This finding demonstrates that the Scheimpflug measurement of anterior and posterior corneal surfaces was effectively incorporated in the IOL power calculation using ray-tracing methodology. The ray-tracing calculation using Placido topography (RayTrace-P) was significantly associated with PCC. This finding points to the fact that the variation in PCC contributed to the prediction error; however, the effect of PCC variation was sufficiently small, because most refraction power in the cornea was on the anterior corneal surface. Posterior corneal abnormality would be found in the eye after high-myopia LASIK with a forward shift of the posterior cornea (Kamiya & Oshika 2003; Miyata et al. 2004). It was anticipated that ray-tracing calculation using measurement of the both corneal curvatures would allow more accurate IOL power calculation in abnormal cornea cases. Scheimpflug imaging (Dubbelman et al. 2002; Read et al. 2009) and AS-OCT (Radhakrishnan et al. 2001; Chen et al. 2012; Nakagawa et al. 2011) are currently available technologies to measure 3-dimensional corneal images. In the Scheimpflug imaging, a narrow slit beam illuminates the cornea, and a sectioned image of the cornea is captured. The captured image is distorted due to the camera configuration being tilted by 45° to the optical axis of the slit beam. Distortion on the anterior surface is geometrically corrected, and correction of the posterior surface requires ray tracing to compensate the outer optical components (Richards et al. 1988). Another distortion is due to difficulty in the surface detection due to strong refraction in the peripheral, where the edge of the corneal surface had degraded to dullness (Khurana et al. 2007; Nakagawa et al. 2011). This distortion makes the anterior corneal profile flatter than the actual. For reconstructing the anterior profile, differences from AS-OCT measurements were corrected in the current study. Assessment by Nakagawa et al. (2011) demonstrated that AS-OCT topography was more consistent and stable in comparison with rotating Scheimpflug topography. With proper corrections, accurate posterior corneal curvature could be obtained for ray-tracing calculation. Another problem in the Scheimpflug imaging was a limitation in the measured area. Image observation from a 45-degree orientation is occasionally blocked with the lid, supercilium or corneal arcus (Moss et al. 2000; Sugar et al. 2005). Furthermore, the posterior corneal measurement in Scheimpflug imaging was more limited: only 15% of eyes could measure 6-mm-diameter area (Fig. 3). The ray tracing was performed at 2.5 mm pupil diameter, so that impact of the reduced measurement area on accuracy in the PCC and asphericity calculations was not critical. The use of AS-OCT topography would be more effective to avoid this issue (Minami et al. 2012). In conclusion, the use of the Scheimpflug measurement of anterior and posterior corneal surfaces was as accurate as the IOL power calculations using measurement of the anterior corneal surface only in normal cataractous eyes. Whereas it was least associated with PCC, which demonstrated that incorporating additional measurements to ray-tracing calculation would allow an accurate IOL power calculation for eyes after corneal refractive surgery. No author has a financial or proprietary interest in any material or method mentioned.