Back

Integration of mechanical testing, in vivo optical coherence elastography and personalized finite element modeling to predict geometrical outcomes of corneal cross-linking

Frigelli, M.; Lohmuller, R.; Ariza Gracia, M. A.; Schlunck, G.; Lang, S. J.; Torres-Netto, E. A.; Hafezi, F.; Buchler, P.; Kling, S.

2025-09-06 bioengineering
10.1101/2025.09.02.673622 bioRxiv
Show abstract

PurposeCorneal cross-linking (CXL) induces both mechanical and geometrical changes in the cornea, which are typically overlooked in pre-operative planning. We propose a patient-specific finite element model (FEM) to predict the topographic alterations resulting from CXL. MethodsTo calibrate the model, we performed nanoindentation and ex vivo OCE inflation tests before and after CXL on five human donor corneas. Nanoindentation results tuned the visco-hyperelastic parameters, while ex vivo OCE axial strains were used for validation. Personalized corneal models were generated from the topographies of three keratoconus patients, with regional stiffness in the affected areas adjusted based on axial strain measured by in vivo pressure-modulated OCE. Simulated CXL outcomes were then compared to 6-month clinical results. ResultsCXL induces a 16-fold increase in the fiber-related mechanical parameters and reduces the viscoelasticity time constant by three. In vivo OCE measurements showed an average mechanical weakening of 57% in the KC regions. When compared to the clinical topography at the 6-month follow-up, the CXL-induced curvature changes predicted by the model were -1.5 D vs. -1.76 D, -1.65 D vs. - 1.91 D, and -1.76 D vs. -1.57 D, for the three patients, respectively. ConclusionBy combining FEM with in vivo corneal mechanical characterization, patient-specific topographic changes can be predicted, which can be used to improve the planning of CXL treatments.

Published in Annals of Biomedical Engineering (predicted rank #19) · training set

Matching journals

The top 8 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.