Abstract
This thesis aims to evaluate the capabilities of MARS spectral photon counting computed tomography for diagnosis and follow-up of bone and soft tissue changes. The imaging modality provides material-specific maps with high spatial resolution and reduces the beam hardening effect. Based on these attributes, we hypothesized that we could track bone changes, differentiate newly formed bone based on treatment, and differentiate soft tissue in the presence of beam hardening.
Prior to using the scanner for a clinical purpose, its accuracy and how different parameters affect the image quality need to be evaluated. A monthly assessment of the image quality parameters was performed to evaluate the scanner’s reliability and stability. Various parameters including translation
of the camera, bias voltage, and voxel size were assessed for their effect on the image quality. The results showed a consistent outcome for spatial resolution, signal-to-noise ratio, and material identification across monthly scans. Furthermore, across all MARS scanners, the spectral response of the specific material exhibited the same spectral trend. In addition, it has been shown that when the bias voltage increases to 800 keV or when the camera translation increases so the gap between chips is covered, the material identification will increase. During the assessment of the images, it was found that all MARS scanners have the ring artefact, and it is more pronounced when larger FOVs are used when the detectors are defective, or when the gap between the chips is not accounted for by the camera translation and
alignment.
SPPCT allows you to use smaller isotropic voxels in reconstruction which
results in a large volume of images to analyse. Analysing bone tissue features will therefore be time-consuming and subject to user error in the case of manual segmentation. To efficiently measure bone features, a semiautomatic method needed to be devised. The active contour method and weka segmentation tool embedded in ImageJ were used to semi-automatically segment bone compartments. Results from both developed methods and manual measurements showed good correlations for all the bone parameters.
The possibility of tracking bone changes using this imaging modality was assessed by two studies. In the first study, bones were decalcified to track calcium loss. Results acquired from MARS were compared with atomic absorption spectroscopy. We found that both methods showed similar patterns of calcium loss during the process. In the second study, after surgically inducing bilateral damage to sheep tibiae, bone healing was monitored using plain radiographs, iDXA, clinical single- and dual-energy CT, and MARS
SPCCT. The results of this study showed that MARS SPCCT can visualise
bone healing and quantify bone mineral density and content in the damaged areas. MARS SPCCT was also used to visualize and quantify strontium as a biomarker for new bone formation. The results demonstrated that MARS SPCCT can distinguish strontium from calcium at higher concentrations (strontium greater than 8 mg/ml and calcium greater than 50 mg/cm3).
These results were verified with Laser ablation inductively coupled plasma mass spectrometry. Additionally, bone-induced beam hardening on the brain was evaluated. Clinical single- and dual-energy CT images were compared with MARS SPCCT images. As a result, the effect of beam hardening using SPCCT
was lesser than that of the clinical CT. Comparing MARS SPCCT with clinical CT operated in single and dual-energy modes, the grey-white matter contrast was higher in the presence of subdural grid electrodes and skull.
In conclusion, this thesis demonstrated the potential of SPCCT to investigate the changes of bone tissue mineral density and structure, as well as to assess soft tissue differentiation in the presence of beam hardening.