Subject-specific functional model of hard and soft tissues; skull and spine

Abstract

There is a strong demand for mechanically and morphologically accurate models of the human musculoskeletal system, particularly of the spine. Such models would have multiple applications, including surgical guides, the analysis of implant fitment and design, as well as individual strength evaluation. Current standards such as the ASTM F1717 (devised for the static and dynamic testing of implants) represent complex spine morphologies using simplified blocks of homogeneous material generally constructed from ultra-high-molecular-weight polyethylene (UHMWPE). These do not attempt to replicate morphological characteristics, and therefore do not reproduce mechanical loading properties, especially when considering the complexity of vertebral bodies and their facets. The work described in this thesis investigated the creation of a compressively accurate and validated model of a lumbar motion segment, specifically the validity of technologies such as computed tomography (CT) scanning, computer-aided scan reconstruction, rapid prototyping, digital image correlation (DIC) and finite element analysis (FEA) modelling. In particular, DIC (an optical measurement method) allowed full-field measurements of the displacements and strains. This was used to determine loading paths and magnitudes during the testing procedure. To complement this approach, FEA modelling identified the location and severity of maximum strains for subsequent comparison to the DIC and mechanical testing data. All FEA models were based on CT scan datasets of the modelled cadaveric material, and were validated against the ex vivo mechanical test measurements. The research followed a number of core stages: 1. First, the applicable technologies were tested and verified, with all channels indicating closely related data. This was achieved by the compressive loading of two types of analogue skulls, allowing the validation of DIC as a data acquisition technique in complex structures. Validation against FEA models demonstrated their potential to provide further insight into the experimental results. The initial testing identified a well-defined pathway for a sample manufacturing and preparation process, making it much easier to produce reliable analogues for subsequent experiments. ii 2. In the second stage, analogue motion segments (AMss) were created using the CT scan datasets obtained from the cadaveric porcine specimens. Motion analysis provided a better understanding of the loading paths again by using DIC as an appropriate data acquisition system. Following the creation of the AMS, different materials were considered for the creation of intervertebral discs (IVDs). The mechanically most biofidelic material was selected. 3. Finally, a sensitivity study was carried out to determine a relationship between the scanning resolution and model accuracy for both the mechanical analogue and the FEA model. The use of 3D printing was found to be an effective, efficient and economical strategy for the creation of accurate biomechanical analogues. Furthermore, DIC was a useful tool when looking at individual component strains and displacements. Finally, when considering a motion segment, the majority of the elastic loading – and thus its behaviour on the whole – was governed by the material properties of the IVD simulant. This research demonstrated a clear path towards the creation of a reliable, biofidelic motion segment, or even a partial lumbar spine analogue, that would comply in dynamic and static loading scenarios as well as conformity in compression. The capability of the techniques and the compliance and accuracy of the resulting models was confirmed by developing both analogue mechanical models and FE simulations. Given their potential advantages, it is only a matter of time before mechanical analogues and their corresponding digital models replace the outdated and inaccurate testing standards in our current medical facilities and research centres.