On the anisotropy of trabeculae in vivo : computationally and morphometry aware

Abstract

Bones and their inner trabecular tissue transmit mechanical loads as part of the skeleton and continuously adapt to forces. Medically indicated implant systems significantly increase the bone’s stiffness, which induces the remodeling of the trabeculae. Aseptic loosening of the bone-implant system is a recurring and, in part, preventable outcome.

Resembling the patient’s tissue stiffness by the implant mitigates this phenomenon. Computed tomography imaging is available to capture the bone’s state a priori of the implantation. However, the spatial resolution in the regime of multiple tens of millimeters depicts the morphology only in a limited manner.

High-resolution microfocus computed tomography is radiologically prohibitive for use in vivo but enables the numerical evaluation of anisotropic stiffness tensors ex vivo. The direct discretization of voxels within cubic subvolumes into finite elements allows for direct mechanics analyses. Therefore, fully triclinic material characterizations in vivo are available as the ground truth.

This report describes a new approach to computing the tensor fields based on cubic subvolumes to quantify the patient-specific linear elasticity, e.g., for designing patient-specific implants. For the first time, anisotropic effective and their enhanced approach, effective numerical stiffnesses based on clinical computed tomography in vivo are available a priori of the implantation.

Direct mechanical evaluations require considerable computing capacities. As an exemplary engineering application, hardware characteristics and a priori available morphological parameters of trabeculae enable energy and runtime savings for simulations. A new approach to documenting the history of datasets, MeRaDat, extends the current state of the art for future, domain-specific research on high-performance computers.