Value Of Ray Tracing-Based Intraocular Lens Power Calculation In Combined Cataract Surgery And Descemet Membrane Endothelial Keratoplasty

Published 2026 - 44th Congress of the ESCRS

Reference: FP22.05 | Type: Free paper | DOI: 10.82333/v4vk-kt61

Authors: Dr Klemens P. Kaiser* ¹ , Mr Jan Weber ¹ , Dr Jascha Wendelstein ² , Prof. Dr H. Burkard Dick ³ , Prof. Dr Thomas Kohnen ¹ , Dr Ingo Schmack ¹

¹Department of Ophthalmology,Goethe-University Frankfurt,Frankfurt am Main,Germany, ²Department of Ophthalmology,University Hospital Munich, Ludwig-Maximilian-University,Munich,Germany, ³Department of Ophthalmology,Knappschaftskrankenhaus Eye Clinic, University Hospital of the Ruhr University Bochum,Bochum,Germany

Purpose

To compare the predictive accuracy of ray tracing–based intraocular lens (IOL) power calculation formulas with the Barrett Universal II (BUII) formula in eyes with Fuchs endothelial corneal dystrophy (FECD) undergoing combined cataract surgery and Descemet membrane endothelial keratoplasty (triple DMEK).

Setting

Tertiary referral university hospital.

Methods

In this retrospective, single-center comparative case series, 149 eyes of 149 patients with FECD undergoing triple DMEK were included. Preoperative biometry was obtained mainly using Scheimpflug tomography (Pentacam AXL). IOL power calculations were performed using BUII, Olsen ray-tracing, and two OKULIX ray-tracing approaches (paraxial and 2.5-mm pupil). Prediction error (PE) was calculated as the difference between postoperative and predicted spherical equivalent. Trueness, precision, and accuracy were analyzed using robust statistical methods (Eyetemis). A retrospective IOLup1D adjustment was additionally evaluated.

Results

BUII demonstrated the lowest mean absolute error (0.74±0.65D) and median absolute error (0.60D), with 67.1% of eyes within ±1.00D of target refraction. Olsen ray tracing showed higher errors (mean absolute error 1.06±1.02D), while both OKULIX approaches performed similarly (mean absolute errors 0.91–1.10 D). Only OKULIX paraxial ray-tracing showed no significant systematic bias (p=0.450); all other formulas exhibited significant hyperopic shift (p<0.001). After IOLup1D adjustment, BUII and Olsen achieved comparable accuracy, with Olsen showing the highest trueness.

Conclusions

In triple DMEK eyes, ray tracing-based IOL power calculation does not provide superior accuracy compared with BUII. Vergence-based formulas combined with pragmatic adjustment strategies such as IOLup1D currently appear to offer the most reliable refractive outcomes.