"Predicting intraocular lens tilt using a machine learning concept".

OBJECTIVE: Aim of this study was to use a combination of partial least squares regression and a machine learning approach to predict IOL tilt using pre-operative biometry data.

SETTING: Patients scheduled for cataract surgery at the Kepler University Clinic Linz.

DESIGN: Prospective single center study.

METHODS: Optical coherence tomography, autorefraction and subjective refraction was performed at baseline and 8 weeks after cataract surgery. In analysis I only one eye per patient was included and a tilt prediction model was generated. In analysis II a pair-wise comparison between right and left eyes was performed.

RESULTS: In analysis I 50 eyes of 50 patients were analysed. Difference in amount, orientation and vector from pre- to post-operative lens tilt was -0.13°, 2.14° and 1.20° respectively. A high predictive power (variable importance for projection) for post-operative tilt prediction was found for pre-operative tilt (VIP=2.2), pupil decentration (VIP=1.5), lens thickness (VIP=1.1), axial eye length (VIP=0.9) and pre-operative lens decentration (VIP=0.8). These variables were applied to a machine learning algorithm resulting in an out of bag score of 0.92°. In analysis II 76 eyes of 38 patients were included. The difference of pre- to post-operative IOL tilt of right and left eyes of the same individuum was statistically relevant.

CONCLUSION: Post-operative IOL tilt showed excellent predictability using pre-operative biometry data and a combination of partial least squares regression and a machine learning algorithm. Pre-operative lens tilt, pupil decentration, lens thickness, axial eye length and pre-operative lens decentration were found to be the most relevant parameters for this prediction model.