MRI Quantification of Cortical Bone Microstructure and Material Composition with Ultrashort Echo Time Sequences

Quantitative imaging assessment of cortical bone structure and material composition provides additional information about bone health beyond conventional radiographic assessment of bone mineral density (BMD). Large prospective studies have found that cortical bone imaging biomarkers are associated with osteoporotic fracture independent of age, sex, height, weight, or mineral density. Recent advances in solid-state 1H and 31P MRI have demonstrated feasibility of ionizing-radiation-free quantification of cortical bone microstructure, organic matrix density, and mineral content in vivo in clinically practical acquisition times. Therefore, the primary objective of this dissertation was to perform the first investigation of these cortical bone parameters in patients with diagnosed bone disease, and to explore the potential clinical translatability of these techniques. Secondary objectives include performing preclinical validation studies and developing technologies to improve clinical translatability of these protocols. First, 1H MRI biomarkers of porosity were shown to be strongly predictive of whole-bone mechanical competence in a study of human cadaveric femurs. Second, the 1H biomarkers of porosity, geometry, and collagen density, and the 31P biomarkers of mineralization, were investigated in a cohort of postmenopausal women with osteoporosis compared to age-matched healthy controls. Osteoporosis was found to be associated with elevated porosity and lower cortical thickness and mineralization, although there was no difference found in collagen density. Similarly, cortical thickness and mineralization and collagen density were positively associated with BMD while porosity was inversely associated with BMD. Third, an artificial intelligence algorithm was developed to automatically quantify porosity and thickness without the need for manual segmentation of MR images or the use of external calibration samples. Automatic quantifications of porosity and geometry were demonstrated to be highly accurate as compared to those derived from expert human segmentation. Furthermore, automated biomarkers detected significant impairments in bone quality in postmenopausal women with osteoporosis compared to age-matched postmenopausal women, and to young, healthy controls. In conclusion, this work was the first to demonstrate that MRI-derived quantification of cortical bone porosity, geometry, and mineralization detect impairments associated with bone disease in vivo.

Author dc:creator

• Jones, Brandon, Clinton

Advisors dc:contributor.advisor

• Rajapakse, Chamith, S
• Wehrli, Felix, W

• Engineering
• Medical Sciences
• Physics

Language dc:language.iso

en

Repository record dc:identifier.uri

https://repository.upenn.edu/handle/20.500.14332/60039

OAI identifier oai:identifier

oai:repository.upenn.edu:20.500.14332/60039