Improving toric iol calculations
3D quantitative OCT and ray tracing show promise for achieving better outcomes
Cataract surgeons in the future will be using three-dimensional (3D) quantitative optical coherence tomography (OCT) imaging to obtain preoperative biometry. The information it provides should improve traditional IOL power formulas and feed into custom-eye ray-tracing models to customise IOL selection and the best rotation for toric IOLs, said Susana Marcos PhD, at the XXXV Congress of the ESCRS in Lisbon, Portugal.
“With a single scan, the 3D quantitative OCT can determine topography, keratometry, and astigmatism of the anterior and posterior cornea and a better estimate of the effective lens position (ELP). It can also provide data on the thickness, anterior surface topography, posterior surface topography, tilt, and decentration for the crystalline lens, along with pachymetry, pupillometry, and anterior chamber depth data,” said Dr Marcos, Director, Visual Optics and Biophotonics Laboratory, Consejo Superior de Investigaciones, Científicas, Madrid, Spain.
Dr Marcos demonstrated the promise of 3D quantitative OCT by reviewing findings from research that has been conducted with an experimental anterior segment spectral domain system developed by her laboratory. The custom platform uses algorithms with full distortion correction and automatic analysis tools to quantify the biometric parameters.
She discussed a study showing how the system could be used to characterise changes in corneal aberrations that are caused by the surgical incision.
“This is something we can incorporate in the model for better accuracy,” she said.
She also reviewed an in vivo study showing how the use of the 3D imaging technique to obtain patient-specific data on the full crystalline lens shape preoperatively improved prediction of the postoperative IOL position when compared to existing methods that are considered state-of-the-art.
“It is well accepted that a more accurate estimation of the ELP will result in improved IOL power calculations, and therefore a more accurate correction of refractive error,” she said, noting that a novel OCT-based method that allowed full quantification of the crystalline lens 3D shape was associated with improvement of up to 0.6D compared with the SRK/T formula or the Olsen Constant optimised. Dr Marcos also showed how having an anatomically correct full model of the lens translates into achieving better results with toric IOL implantation.
“When custom model eyes are built using patient-specific anatomical data and there is a better estimate of where the lens is going to fit, you can use these as a platform to perform ray-tracing calculations, and for example, rotate the toric lens to determine which orientation produces an optimised wavefront. Applying this approach, there is no longer any need to rely on regression or theoretical formulas or to determine the axis during surgery,” she said.
RAY TRACING
Paul-Rolf Preussner MD, PhD, Professor, University Eye Hospital, Mainz, Germany, discussed the application of ray tracing to IOL formulas. Ray tracing is used by the commercially available Okulix software.
Input data for ray-tracing IOL power calculation can be obtained using devices that give anterior and posterior cornea surface measurements. Currently, the Okulix software can interface with all Tomey devices, the Oculus Pentacam and the Haag-Streit Lenstar, and it will be compatible with other diagnostic imaging devices soon.
Dr Preussner noted that the ray-tracing prediction algorithm is the same for toric and non-toric IOLs because there should be no difference in lens position between the two types of optics, and he recommended using anterior and posterior cornea surface data for the calculation.
To illustrate the performance of the software, Dr Preussner reviewed findings from a study that analysed data from 78 eyes of 56 patients implanted with different toric IOL models. Using the ray-tracing software to predict the residual refraction, Dr Preussner and colleagues showed that the cylindrical prediction error was lower when the software used preoperative keratometry, topography and tomography than if it used any of the three measurements alone. Among eyes that achieved VA ≥20/20, the median absolute cylindrical vector prediction error was only 0.38D using the combined data.
A study including five “problem eyes” with a history of corneal surgery was undertaken to investigate the accuracy limits of the measuring process in such cases. All eyes were measured three times with each of three different devices (Tomey TMS-5, Ziemer Galilei G6 and Oculus Pentacam). The raw tomography data were processed by the ray-tracing software independent of the device’s software, and the vector average of the total corneal astigmatism was calculated for each of the nine measurements for each eye.
The results for two eyes showed good agreement across all measurements. For two other eyes, there was reasonably good agreement but greater intra- and inter-device data spread. In the remaining eye, there were large differences in outcomes between devices and comparing the repeated measurements taken with a single device.
“I think most surgeons do not have different devices, but they still want to know what kind of accuracy to expect in an individual case. My recommendation is to do repeated measurements, calculate the outcome, and use the data to inform the patient about the possibility that the outcome may not be ideal,” Dr Preussner said.
Susana Marcos: susana@io.cfmac.cscic.es
“With a single scan, the 3D quantitative OCT can determine topography, keratometry, and astigmatism of the anterior and posterior cornea and a better estimate of the effective lens position (ELP). It can also provide data on the thickness, anterior surface topography, posterior surface topography, tilt, and decentration for the crystalline lens, along with pachymetry, pupillometry, and anterior chamber depth data,” said Dr Marcos, Director, Visual Optics and Biophotonics Laboratory, Consejo Superior de Investigaciones, Científicas, Madrid, Spain.
Dr Marcos demonstrated the promise of 3D quantitative OCT by reviewing findings from research that has been conducted with an experimental anterior segment spectral domain system developed by her laboratory. The custom platform uses algorithms with full distortion correction and automatic analysis tools to quantify the biometric parameters.
She discussed a study showing how the system could be used to characterise changes in corneal aberrations that are caused by the surgical incision.
“This is something we can incorporate in the model for better accuracy,” she said.
She also reviewed an in vivo study showing how the use of the 3D imaging technique to obtain patient-specific data on the full crystalline lens shape preoperatively improved prediction of the postoperative IOL position when compared to existing methods that are considered state-of-the-art.
“It is well accepted that a more accurate estimation of the ELP will result in improved IOL power calculations, and therefore a more accurate correction of refractive error,” she said, noting that a novel OCT-based method that allowed full quantification of the crystalline lens 3D shape was associated with improvement of up to 0.6D compared with the SRK/T formula or the Olsen Constant optimised. Dr Marcos also showed how having an anatomically correct full model of the lens translates into achieving better results with toric IOL implantation.
“When custom model eyes are built using patient-specific anatomical data and there is a better estimate of where the lens is going to fit, you can use these as a platform to perform ray-tracing calculations, and for example, rotate the toric lens to determine which orientation produces an optimised wavefront. Applying this approach, there is no longer any need to rely on regression or theoretical formulas or to determine the axis during surgery,” she said.
RAY TRACING
Paul-Rolf Preussner MD, PhD, Professor, University Eye Hospital, Mainz, Germany, discussed the application of ray tracing to IOL formulas. Ray tracing is used by the commercially available Okulix software.
Input data for ray-tracing IOL power calculation can be obtained using devices that give anterior and posterior cornea surface measurements. Currently, the Okulix software can interface with all Tomey devices, the Oculus Pentacam and the Haag-Streit Lenstar, and it will be compatible with other diagnostic imaging devices soon.
Dr Preussner noted that the ray-tracing prediction algorithm is the same for toric and non-toric IOLs because there should be no difference in lens position between the two types of optics, and he recommended using anterior and posterior cornea surface data for the calculation.
To illustrate the performance of the software, Dr Preussner reviewed findings from a study that analysed data from 78 eyes of 56 patients implanted with different toric IOL models. Using the ray-tracing software to predict the residual refraction, Dr Preussner and colleagues showed that the cylindrical prediction error was lower when the software used preoperative keratometry, topography and tomography than if it used any of the three measurements alone. Among eyes that achieved VA ≥20/20, the median absolute cylindrical vector prediction error was only 0.38D using the combined data.
A study including five “problem eyes” with a history of corneal surgery was undertaken to investigate the accuracy limits of the measuring process in such cases. All eyes were measured three times with each of three different devices (Tomey TMS-5, Ziemer Galilei G6 and Oculus Pentacam). The raw tomography data were processed by the ray-tracing software independent of the device’s software, and the vector average of the total corneal astigmatism was calculated for each of the nine measurements for each eye.
The results for two eyes showed good agreement across all measurements. For two other eyes, there was reasonably good agreement but greater intra- and inter-device data spread. In the remaining eye, there were large differences in outcomes between devices and comparing the repeated measurements taken with a single device.
“I think most surgeons do not have different devices, but they still want to know what kind of accuracy to expect in an individual case. My recommendation is to do repeated measurements, calculate the outcome, and use the data to inform the patient about the possibility that the outcome may not be ideal,” Dr Preussner said.
Susana Marcos: susana@io.cfmac.cscic.es
Authors
Cheryl Guttman Krader
Published
Tuesday, January 29, 2019
Category