MARS Spectral CT Technology for Orthopaedic Applications

Abstract

This thesis investigates MARS spectral CT technology for preclinical orthopaedic applications. Current CT methods for orthopaedic imaging have limited success due to their to inability to quantify tissue types and contrast media, and in reducing metal artefacts arising from orthopaedic implants. Spectral CT has the potential to quantify contrast media for cartilage evaluation, to quantify strontium in additive-manufactured scaffolds, and to reduce CT image artefacts. This thesis is presented in two parts, with the first half focussing on MARS image processing techniques, and the second half focussing on preclinical studies on cartilage imaging and metal implant imaging. A projection-space image denoising algorithm for MARS multi-energy data was developed during my research. Image noise in the acquired data corrupts the information required for multienergy CT reconstruction, material discrimination, and image analysis. The benefits of this denoising technique were demonstrated on data from preclinical studies such as metal implant imaging, atheroma imaging, and contrast molecular imaging. Image segmentation methods using component analysis techniques were developed for a hybrid-CT setup, and were also used for material characterisation in an additive-manufactured polymer scaffold used in regenerative medicine. Quantitative cartilage imaging using spectral CT was demonstrated using bovine cartilage and osteoarthritic human cartilage samples. A contrast media used as an inverse marker of gylcosaminoglycans (GAG) in cartilage enabled cartilage-bone differentiation and material quantification. The results were compared with cartilage histology stained for GAG. Metal artefact reduction through high-energy acquisition using MARS was demonstrated. Samples made from titanium and cobalt-chromium alloys in the form of phantoms, small implants and 3D-printed porous scaffolds were imaged. Metal artefacts were quantified in the multienergy data using appropriate CT image metrics. In conclusion, this thesis has developed imaging strategies and image processing methods using MARS small-animal modalities to demonstrate the potential of spectral CT for quantitative cartilage imaging, for reducing metal artefacts in orthopaedic implants, and for non-destructive imaging of additive-manufactured scaffolds.