In vivo magnetic resonance-based virtual bone biopsy
Abstract
Metabolic bone disease fractures and treatments are associated with direct and indirect medical costs, and can cause a reduction of function, quality of life, and longevity in older patients. Both cortical and trabecular bone tissue contribute to bone strength and can be affected differently by specific diseases. Current diagnostic modalities in clinical use for bone diseases are the invasive bone biopsy and dual-energy x-ray absorptiometry (DXA), but both are unsatisfactory for longitudinal studies and treatment monitoring. Magnetic resonance imaging (MRI) has emerged as a promising modality for in vivo evaluation of trabecular bone structure. The goal of this thesis is to develop analysis methods suited to these images. Toward these goals, the virtual bone biopsy (VBB) was created from an integration of previously existing technologies to acquire trabecular and cortical bone images. Processing methods specific to these images were developed for structural analysis. Using existing dedicated MRI acquisition methods, microscopic images of the distal radial metaphysis were acquired in vivo. Applying digital topology classification to these images reveals the local topological structure type of each point in the trabecular bone network and leads to the development of parameters relating to bone loss, such as the surface-to-curve ratio and erosion index. Trabecular bone orientation is known to influence bone strength, and a topology-based method for determining local orientation and anisotropy is introduced here. A method was developed to analyze intensity profiles that transect the cortical bone to locate the periosteal and endosteal boundaries, and subsequently the structural parameters. Using synthetic, ex vivo, and in vivo data, the VBB parameters are shown to be accurate and reproducible. The VBB erosion index is a better predictor of Young's modulus than density, explaining 67% (p < 0.0001) of the variation. When applied to a study of vertebral osteoporosis (n = 79), trabecular VBB parameters from the distal radius surrogate site were better overall discriminators than DXA values between patients with vertebral deformities from those without. These results support the role of bone structure in determining strength and establish the VBB parameters as efficacious determinants of bone structural quality. ^
Subject Area
Engineering, Biomedical|Health Sciences, Medicine and Surgery
Recommended Citation
Bryon Roos Gomberg,
"In vivo magnetic resonance-based virtual bone biopsy"
(January 1, 2002).
Dissertations available from ProQuest.
Paper AAI3054945.
http://repository.upenn.edu/dissertations/AAI3054945