Vertebral osteoporosis is a major health concern both in the United States and worldwide, and is expected to affect substantially more people as the size of the aging population increases. Currently, the clinical gold standard for assessment of fracture risk for the spine is dual-energy x-ray absorptiometry (DXA). This two-dimensional scanning modality is limited in its ability to predict fracture risk, and improved methods of fracture risk prediction are therefore needed. Quantitative computed tomography (QCT), being a three-dimensional imaging modality, offers great promise at providing such improved measures, but the complexity of the underlying biomechanics of spine fractures undermines the ability of QCT alone to predict fracture risk. In particular, the vertebral body displays different strength properties for the different types of loads it encounters in vivo, such as compression vs. anterior bending. QCT scans, being descriptors of only the bone structure, cannot account for these different strengths. Recent advances in bone biomechanics and computational stress analysis techniques now enable us to produce patient-specific structural computer finite element models of an individual's vertebra directly from QCT scans in an almost entirely automated fashion. Ideally suited for clinical implementation, these """"""""voxer' finite element models can provide a fracture risk prediction that overcomes the limitations associated with DXA and QCT. Through an unique multidisciplinary team of bioengineers, clinical QCT radiological experts, and epidemiologists, we plan to implement this computational modeling technique clinically and compare its performance against DXA and QCT. In particular, we will test its ability to predict fracture risk in an ongoing NIH-funded osteoporosis prospective fracture surveillance study of almost 6000 men aged over 65, for which both DXA and QCT scans are available at baseline. To ensure that our modeling technique is optimized for successful clinical usage, we will first perform a detailed biomechanical validation of the technique as applied to cadaver vertebrae for varied loading conditions, including compression and combined compression/forward bending. We will also address the role of the posterior elements, and treat the disc condition as an uncertainty variable. Our Hypothesis is that the QCT-based finite element modeling technique, being mechanistic, is better at clinical fracture risk prediction than purely densitometric techniques such as DXA and QCT. This research will provide insight into the biomechanical mechanisms of osteoporotic spine fractures by way of our cadaver studies. It will have profound clinical impact by improving substantially the ability to predict risk of vertebral fracture in the elderly. Finally, we hope this research will instigate a paradigm change in musculoskeletal imaging in which engineering mechanistic models are integrated into medical images to provide a true functional image, in this case the """"""""biomechanical scan"""""""" of the vertebra.
Showing the most recent 10 out of 26 publications
