Abstract
In this thesis, we present automated methods that quantify information from medical images; information that is intended to assist and enable clinicians gain a better understanding of the underlying pathology. The first part of the
thesis presents methods that analyse the articular cartilage, segmented from MR images of the knee. The cartilage tissue is considered to be a key determinant in the onset of Osteoarthritis (OA), a degenerative joint disease, with no known cure. The primary obstacle has been the dependence on radiography as the ‘gold standard’ for detecting the manifestation of cartilage changes. This is an indirect assessment, since the cartilage is not visible on xrays. We propose Cartilage Homogeneity, quantified from MR images, as a marker for detection of the early biochemical alterations in the articular cartilage. We show that homogeneity provides accuracy, sensitivity, and
information beyond that of traditional morphometric measures. The thesis also proposes a fully automatic and generic statistical framework for identifying biologically interpretable regions of difference (ROD) between two groups of biological objects, attributed by anatomical differences or changes relating to pathology, without a priori knowledge about the location, extent, or topology of the ROD. Based on quantifications from both morphometric and textural based imaging markers, our method has identified the most pathological regions in the articular cartilage. The remaining part of the thesis presents methods based on diffusion tensor imaging, a technique widely used for analysis of the white matter of the central nervous system in the living human brain. An inherent drawback of the traditional diffusion tensor model is its limited ability to provide detailed information about multi-directional fiber architecture within a voxel. This
leads to erroneous fiber tractography results in locations where fiber bundles cross each other. We present a novel tractography technique, which successfully traces through regions of crossing fibers. Detection of crossing
white matter pathways can improve neurosurgical visualization of functionally relevant white matter areas. We also present preliminary results of analysing the meshwork of the collagen fibers in the articular cartilage by high-resolution diffusion tensor imaging.
thesis presents methods that analyse the articular cartilage, segmented from MR images of the knee. The cartilage tissue is considered to be a key determinant in the onset of Osteoarthritis (OA), a degenerative joint disease, with no known cure. The primary obstacle has been the dependence on radiography as the ‘gold standard’ for detecting the manifestation of cartilage changes. This is an indirect assessment, since the cartilage is not visible on xrays. We propose Cartilage Homogeneity, quantified from MR images, as a marker for detection of the early biochemical alterations in the articular cartilage. We show that homogeneity provides accuracy, sensitivity, and
information beyond that of traditional morphometric measures. The thesis also proposes a fully automatic and generic statistical framework for identifying biologically interpretable regions of difference (ROD) between two groups of biological objects, attributed by anatomical differences or changes relating to pathology, without a priori knowledge about the location, extent, or topology of the ROD. Based on quantifications from both morphometric and textural based imaging markers, our method has identified the most pathological regions in the articular cartilage. The remaining part of the thesis presents methods based on diffusion tensor imaging, a technique widely used for analysis of the white matter of the central nervous system in the living human brain. An inherent drawback of the traditional diffusion tensor model is its limited ability to provide detailed information about multi-directional fiber architecture within a voxel. This
leads to erroneous fiber tractography results in locations where fiber bundles cross each other. We present a novel tractography technique, which successfully traces through regions of crossing fibers. Detection of crossing
white matter pathways can improve neurosurgical visualization of functionally relevant white matter areas. We also present preliminary results of analysing the meshwork of the collagen fibers in the articular cartilage by high-resolution diffusion tensor imaging.
| Original language | English |
|---|
| Place of Publication | København |
|---|---|
| Publisher | Department of Computer Science, University of Copenhagen |
| Number of pages | 161 |
| Publication status | Published - 2008 |
