THE 500 EYES LENS POWER TRIAL: INTRODUCING MACHINE LEARNING FOR INTRAOCULAR LENS POWER CALCULATION

To train regression models for intraocular lens (IOL) power calculations using several machine learning approaches in order to minimize the refractive error after cataract surgery.

Department of Ophthalmology and Optometry, Johannes Kepler University, Linz, Austria | Department of Ophthalmology, Medical University Graz, Austria

This two-center study included patients that underwent cataract surgery between July 2021 and January 2022. All eyes received a  swept-source OCT biometry (IOL Master 700) followed by implantation of a monofocal hydrophobic open loop haptic IOL. Study examination was performed minimum 4 weeks and maximum 24 months after surgery and included optical biometry, autorefraction as well as subjective refraction. Main parameter for the machine learning approached was the spherical equivalent deriving from subjective refraction. Next to classical random forest plots, Bayesian statistics and support vector machine learning was used to predict the mentioned primary endpoint parameter. 

In total, 500 eyes of 500 patients were included in this trial. After constant optimisation for different IOL power calculation formulae the mean absolute refractive error (MAE) was found to be between 0.39 D and 0.52 D. No significant improvement was found when using classical random forest plots (MAE: 0.41 D) or support vector machine learning (MAE: 0.43 D), however, a slight improvement was found when using Bayesian statistics (MAE: 0.32 D). A novel concept for refractive surprise detection will be presented at the ESCRS meeting.

Bayesian statistics can improve the post-operative refractive outcome and decrease the number of refractive surprises. However, not all cases of refractive surprises can be avoided by this method.