Predicting Intraocular Lens Tilt Using OCT Measurements and Partial Least Squares Regression Modelling

Overview

During cataract surgery an artificial intraocular lens (IOL) is implanted in the eye to replace the lens and to correct the refraction for distance vision. Misalignment of IOLs can cause severe loss of visual quality. Different types of misalignment are known. Tilt, one type of IOL misalignment is thought to play a negative role for the optical performance in eyes with IOL designs, especially, if they have aspheric, toric, or multifocal optics. Various methods to measure IOL misalignments have been described. Studies assessing the IOL position have used subjective grading methods at the slit lamp examination or a Scheimpflug camera to assess IOL decentration and tilt. The subjective grading at the slitlamp may display considerable variability between examiners. This method is more qualitative than quantitative and does not allow fine resolution when reporting IOL tilt. The fact that the patient has no standardized target to focus on makes the method even less reliable. Scanning methods such as Scheimpflug photos require a very well dilated pupil exceeding 6mm to assess the IOL position. Additionally, it can be difficult to identify the anatomical structures of the eye that need to be used as points of reference. Scheimpflug camera images have been used for assessing IOL tilt previously, but erroneous results, often due to corneal magnification, have diminished their widespread use. Another possibility to assess tilt is the use of Purkinje reflexes. The light reflections of Purkinje images at ocular surfaces to evaluate ocular alignment have recently been utilized. Since light is reflected at all interfaces of media with a difference in refractive index, these reflections, called Purkinje images, can be used to assess tilt and decentration of IOLs. Two different clinically applicable Purkinjemeter system provide the measurement of IOL decentration and tilt. The main problem with Purkinje meters is accessibility, as there are only a few prototypes available worldwide. The most recently developed method for tilt quantification is the use of optical coherence tomography (OCT). This method has several advantages compared to the previous methods: OCT based devices are available in most clinics, the resolution of modern OCT devices is high, and the measurements are reproducible. Aim of this study is to measure tilt with two modern OCT based devices and one Scheimpflug camera and to predict the post-operative tilt using partial least squares regression. This method was developed by Wold and introduced to ophthalmology previously.

During cataract surgery an artificial intraocular lens (IOL) is implanted in the eye to replace the lens and to correct the refraction for distance vision. Misalignment of IOLs can cause severe loss of visual quality. Different types of misalignment are known. Tilt, one type of IOL misalignment is thought to play a negative role for the optical performance in eyes with IOL designs, especially, if they have aspheric, toric, or multifocal optics. For example, in the case of aspheric IOLs, it appears that even slight amount of tilt may result not only in the loss of the effect of reducing spherical aberrations but in more severe cases even a worsening of the optical quality compared to spherical IOLs. In the case of toric IOLs, tilt introduces higher order aberrations potentially mimicking astigmatism. In the case of multifocal IOLs, tilt increases higher order aberrations, which leads to decreased visual quality. Various methods to measure IOL misalignments have been described. Studies assessing the IOL position have used subjective grading methods at the slit lamp examination or a Scheimpflug camera to assess IOL decentration and tilt. The subjective grading at the slitlamp may display considerable variability between examiners. This method is more qualitative than quantitative and does not allow fine resolution when reporting IOL tilt. The fact that the patient has no standardized target to focus on makes the method even less reliable. Scanning methods such as Scheimpflug photos require a very well dilated pupil exceeding 6mm to assess the IOL position. Additionally, it can be difficult to identify the anatomical structures of the eye that need to be used as points of reference. Scheimpflug camera images have been used for assessing IOL tilt previously, but erroneous results, often due to corneal magnification, have diminished their widespread use. Another possibility to assess tilt is the use of Purkinje reflexes. The light reflections of Purkinje images at ocular surfaces to evaluate ocular alignment have recently been utilized. Since light is reflected at all interfaces of media with a difference in refractive index, these reflections, called Purkinje images, can be used to assess tilt and decentration of IOLs. Two different clinically applicable Purkinjemeter system provide the measurement of IOL decentration and tilt. The main problem with Purkinje meters is accessibility, as there are only a few prototypes available worldwide. The most recently developed method for tilt quantification is the use of optical coherence tomography. This method has several advantages compared to the previous methods: OCT based devices are available in most clinics, the resolution of modern OCT devices is high, and the measurements are reproducible. Due to the fact that tilt has an influence on visual quality, the prediction of the post-operative tilt could improve IOL power calculation significantly, especially for toric IOLs. In the literature, the prediction of tilt was shown to be acceptable for the orientation of tilt, but not for the amount of tilt. Furthermore, there is disagreement concerning the fact, if the amount of tilt increases or decreases after cataract surgery. Aim of this study is to measure tilt with two modern OCT based devices and one Scheimpflug camera and to predict the post-operative tilt using partial least squares regression. This method was developed by Wold and introduced to ophthalmology previously.