Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


case1152214735.pdf (17.77 MB)

ETD Abstract Container

Abstract Header

CHARACTERIZATION OF ATHEROSCLEROSIS WITH MAGNETIC RESONANCE IMAGING, CHALLENGES AND VALIDATION

Salvado, Olivier
http://rave.ohiolink.edu/etdc/view?acc_num=case1152214735

Abstract Details

2006, Doctor of Philosophy, Case Western Reserve University, Biomedical Engineering.
The long-term goal of this research is the characterization of atherosclerosis lesions in vivo using Magnetic Resonance Imaging (MRI). Our particular interest is the computer-aided diagnosis and staging of the disease, especially the detection of vulnerable plaques. Multi-contrast MRI scans are used (e.g. PDW, T1W, and T2W) to assess the presence and morphology of vessel lesions in human arteries. The first chapter of this thesis reviews the existing techniques for atherosclerosis imaging with emphasis on MRI. The challenges facing intensity-based computer methods are then investigated. The subsequent four chapters address specifically the four main issues that have been identified: 1) Intensity inhomogeneity, which is present when surface coils or intravascular microcoils induce shading artifacts across the data. Artifacts are modeled as a smooth multiplicative bias field with a cubic spline. The cubic spline is optimized to minimize the entropy of the corrected image. Extension to intravascular axial images is presented. 2) Partial volume effect; despite recent improvements in MRI technologies, image resolution is still a limiting factor owing to the small size of the vessel wall and its lesions. A new approach is investigated to address this issue. Under the assumption that homogenous tissues are well separated by an edge, a reverse diffusion model is implemented. Extension to reduce noise is also introduced by using a diffusion coefficient function of the image gradient. 3) Misregistration; Patient and physiological motions induce misregistration between the datasets with different contrast, requiring methods with sub-voxel accuracy to re-align them. The limits of registration methods are investigated and a new constant variance filter is described to improve intensity-based automatic registration accuracy and robustness. This filter removes local maxima and global bias in similarity measures. 4) The last issue investigated is validation; a new method is presented to produce three-dimensional ground truth of ex vivo vessel specimens. This method uses a novel Cryoimaging system where frozen specimens are serially cut and imaged using bright field and fluorescence microscopy. The last chapter discusses potential applications to other diseases and image modalities, future works, and concludes about the perspectives of atherosclerosis imaging with MRI.
David Wilson (Advisor)
231 p.
Engineering, Biomedical
Atheroclerosis; Magnetic Resonance Imaging; Intensity inhomegeneity; Partial volume effect; Noise; Anisotropic diffusion; Vulnerable plaque; Cryoimaging; Segmetnation; Tissue classification; Registration

Recommended Citations

Citations

  • Salvado, O. (2006). CHARACTERIZATION OF ATHEROSCLEROSIS WITH MAGNETIC RESONANCE IMAGING, CHALLENGES AND VALIDATION [Doctoral dissertation, Case Western Reserve University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=case1152214735

    APA Style (7th edition)

  • Salvado, Olivier. CHARACTERIZATION OF ATHEROSCLEROSIS WITH MAGNETIC RESONANCE IMAGING, CHALLENGES AND VALIDATION. 2006. Case Western Reserve University, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=case1152214735.

    MLA Style (8th edition)

  • Salvado, Olivier. "CHARACTERIZATION OF ATHEROSCLEROSIS WITH MAGNETIC RESONANCE IMAGING, CHALLENGES AND VALIDATION." Doctoral dissertation, Case Western Reserve University, 2006. http://rave.ohiolink.edu/etdc/view?acc_num=case1152214735

    Chicago Manual of Style (17th edition)

case1152214735
553
© 2006, all rights reserved.
This open access ETD is published by Case Western Reserve University School of Graduate Studies and OhioLINK.