T 2 Mapping Refined Finite Element Modeling to Predict Knee Osteoarthritis Progression.

This paper presents a novel method for informing cartilage material properties in finite element models from T ₂ relaxometry. In the developed pipeline, T ₂ relaxation values are mapped to elements in subject-specific finite element models of the cartilage and menisci. The Young's modulus for each element within the cartilage is directly calculated from its corresponding T ₂ relaxation voxel value. Our model was tested on a single subject (Subject ID 9932809, Kellgren-Lawrence grade 2) from the Osteoarthritis Initiative dataset at baseline imaging. For comparison, an identical finite element model was built with homogeneous material properties. Kinematics of the stance phase of a standard gait cycle were used as model constraints. Simulation results were compared qualitatively to the MRI Osteoarthritis Knee Score (MOAKS) from the same baseline timepoint. Our T ₂ -refined material model showed higher maximum shear strain in regions with moderate cartilage loss as compared to the homogeneous material model, and the homogeneous model showed higher maximum principal stress and maximum shear strain in regions with no cartilage loss. These results show that a homogeneous material model likely underestimates tissue strains in regions with cartilage damage while overestimating strains in regions with healthy cartilage. This preliminary study demonstrates that T ₂ -refined material properties are more appropriate than assumptions of homogeneity in predictive models of cartilage damage.Clinical relevance- The proposed pipeline demonstrates a computationally efficient way to improve the subject-specificity of finite element models used for evaluation of osteoarthritis.