Prediction Of Postoperative Intraocular Lens Tilt, Decentration And Anterior Chamber Depth Using A Machine Learning Algorithm

The aim of this study was to measure and predict postoperative intraocular lens (IOL) tilt, decentration and anterior chamber depth using a machine learning algorithm.

Department of Ophthalmology, Kepler University Clinic and Johannes Kepler University, Linz, Austria

This prospective single-center study included 170 patients scheduled for bilateral cataract surgery. Assessments were performed preoperatively and 8 weeks postoperatively including optical biometry (IOLMaster 700, Carl Zeiss Meditec AG) and high-definition swept-source OCT (CASIA II, Tomey; ANTERION, Heidelberg Engineering). Additionally, autorefraction, subjective refraction and visual acuity assessments for distance vision were performed at the last follow-up. A machine learning algorithm was developed to predict postoperative IOL tilt, decentration and anterior chamber depth using preoperative data.

A total of 340 eyes of 170 patients were included in the study. At the time of submission, preliminary analysis included 64 complete datasets. The mean age was 70.16 ± 10.02 years. Axial eye length for right and left eyes was 23.17 ± 0.14 mm and 23.06 ± 0.14 mm, respectively. Mean postoperative tilt and decentration were 4.82 ± 0.81° and 0.26 ± 0.11 mm, respectively. Mean postoperative anterior chamber depth was 4.55 ± 0.21 mm.

The amount of postoperative tilt and decentration of the investigated IOL was within the physiological range. The final results regarding IOL tilt, decentration and anterior chamber depth and the performance of the machine learning-based prediction model will be presented at the congress.