Producing micro-finite element models from real-time clinical CT scanners: calibration, validation and material mapping strategies

Yanni Cai^1*Peter Zioupos²Nicholas Marquez-Grant¹Basil Budair³Sarah Junaid⁴

• ¹Cranfield University, Cranfield, United Kingdom
• ²University of Hull, Hull, United Kingdom
• ³Royal Orthopaedic Hospital NHS Foundation Trust, Birmingham, United Kingdom
• ⁴Aston University, Birmingham, United Kingdom

Finite element models from living anatomical structures to produce patient-specific models offer improved diagnosis, precision pre-op planning for surgeries, and reliable biofidelic stress loading analysis. These models require the use of clinical scanners that are safe to use in-vivo but offer relatively lower resolution than in-vitro micro-CT ones. To capitalise on the clinical advantages, this route offers certain technical challenges which must be ironed out to derive a reliable validated route from scanning to in-silico modelling. In the present study, sheep vertebrae were used to create biofidelic phantoms for scanning by using one of the latest technology high-resolution (300 micron) clinical standing scanners (HiRise, Curvebeam). Geometric information was used to produce FEA models (Abaqus/CAE), which were then validated under compression loading in the lab. The main challenges had to do first with reading and converting the scan data from voxels to material property assignment for each FE element, which was performed by using a number of different conversion equations from the literature, and second to a lesser degree with the minor challenges of seeking convergence and refining the boundary conditions. The fit between the model and the experimental results was best for two equations from the literature, while others were less reliable. The selection of the most suitable and universally applicable material conversion equation is significant because it can streamline the route to produce scanner to computer patient-specific models, and make these widely available and ultimately more easily immediately obtainable post-scans. Some known clinical examples highlight the potential use of this methodology for situations where loading and unloading configurations are equally challenging for modelling (i.e. standing CT scans of feet), and we discuss the importance of our approach for such examples. Unlike previous studies using micro-CT or non-clinical setups, this work validates a real-time, weight-bearing CT-based workflow for biomechanically consistent finite element modelling.