< Back to previous page

Publication

Cone-beam Computed Tomography is a Fast and Promising Technique for Microstructural Imaging in Clinical Practice

Book - Dissertation

Due to the rising life expectancy, bone and joint diseases (e.g., osteoporosis, osteoarthritis and rheumatoid arthritis) have become an important socioeconomic burden. For these diseases, the importance of assessing the bone's microarchitecture make-up in clinical practice has been emphasized in basic science. However, it remains challenging to assess it in clinical practice. High-resolution in vivo imaging became possible with the advent of a highresolution peripheral computed tomography (HR-pQCT) scanner. Two systems are currently on the market, XTremeCT-I and XTremeCT-II (Scanco Medical AG, Switzerland) which provide a voxel size of down to 82µm and down to 60.7µm, respectively. So far, a breakthrough of the scanner for widespread clinical applications is still lacking. The two main disadvantages of the scanner are the slow scanning time (2 à 3min. for a stack of 0.9 à 1cm), which makes scanning of a large volume of interest challenging in vivo, and the dedicated nature of the scanner which does not allow routine clinical use for standard musculoskeletal diagnostics. A promising alternative is high-resolution cone-beam computed tomography (CBCT), which is already the gold standard in many dental and maxillofacial applications. The top high-resolution CBCT scanners on the market, e.g. CBCT Newtom 5G (Cefla, Italy), feature a fast scanning time (18 à 31s.), a large field of view (12x12x8cm3) and a low radiation dosage, in addition to a high resolution (voxel size down to 75µm). Yet, CBCT is impaired by the presence of image artefacts that reduce image contrast, leading to it being currently used for qualitative evaluation only. The overarching aim of this PhD is to determine whether CBCT can be enhanced by means of artefact correction algorithms and advanced segmentation techniques in order to be used to visualize and quantify bone microstructure and to quantify bone mechanical parameters in clinical practice. To attain this main aim, four sub aims were formulated and worked out in this thesis. The first sub aim was to identify CBCT artefacts and to enhance CBCT images. In literature, scattering is addressed as the main detrimental factor in CBCT imaging and beam hardening is another often mentioned important artefact. To address these artefacts, a Monte Carlo simulation, as well as a C++ program was developed. In contrast to what is often assumed in literature, our simulations demonstrated that scattering is limited when scanning a wrist with a normal gantry and that the impact of beam hardening, rather than scattering, is more pronounced in CBCT images. The development of a beam hardening correction technique, which took into account the high bone content of extremities, was able to enhance the images significantly. Next to beam hardening correction, an in-house reconstruction and projection processing program was developed which enhanced the images already significantly, compared to the standard reconstruction of the scanner. The second sub aim of this thesis was to evaluate the accuracy of CBCT images in quantifying bone microstructural parameters. To enable quantification of the bone microstructural parameters, the images had to be segmented and a trabecular volume of interest had to be selected. For the segmentation, an adaptive segmentation technique was proposed in this thesis. This adaptive segmentation technique was a key element to enhance accuracy of the quantified bone microstructural parameters, because it enabled segmentation of not completely homogeneous images. To select the trabecular volume of interest automatically, the technique of Buie et al. was extended. Two ex vivo validations were performed in this thesis, one on 19 trapezia and another on 19 distal radii. Both validations demonstrated that our enhanced CBCT images were able to quantify bone microstructural parameters with high accuracy. The third sub aim was to evaluate the accuracy of CBCT images in quantifying bone mechanical parameters. Thus far, simple-flat surface boundary conditions have been mostly applied on sections of a bone, such as a section of the distal radius. To apply boundary conditions on entire bones or multiple bones, more advanced boundary conditions are needed. Therefore, a software technique was developed to apply more general loading conditions. Afterwards, the accuracy of CBCT to quantify bone mechanical parameters was tested on the previously mentioned 19 trapezia and 19 distal radii. Both validations demonstrated that the enhanced CBCT images had an adequate accuracy even when quantifying bone mechanical parameters. The last sub aim was to compare the accuracy of CBCT to HRpQCT, the current standard for in vivo high-resolution scanning of extremities. Although CBCT is visually less sharp, it provides a very similar, albeit slightly lower, accuracy in quantifying bone microstructural and mechanical parameters when compared with XTremeCT-II, the newest generation HR-pQCT scanner. In conclusion, it can be stated that our enhanced CBCT images are able to quantify bone microstructural and mechanical parameters with high accuracy. Hence, high-resolution CBCT, which features fast scanning of large FOV at high resolution and low radiation dosage, is a promising scanner for high-resolution imaging in clinical practice.
Publication year:2019
Accessibility:Closed