Development of a method for in vivo mechanical characterization of articular cartilage
Metadata[+] Show full item record
[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT AUTHOR'S REQUEST.] Arthritis is one of the leading causes of disability in the United States and the second most expensive to treat according to the CDC. One of the key difficulties in diagnosing and treating arthritis, in particular osteoarthritis, is that the mechanisms for progression of the disease are poorly characterized. Mechanical engineer Joe Rexwinkle, working with Dr. Ferris Pfeiffer and the Thompson Lab for Regenerative Orthopaedics, aimed to shed some light on the links between cartilage biology and the degradation seen in osteoarthritis. The study began with obtaining cartilage samples from six patients undergoing total knee replacements and collecting information on several biomarkers with known relevance to osteoarthritis. Specifically, the concentrations of several proteins which may be determined in a standard hospital lab were analyzed. The samples were then tested to determine their mechanical properties, since the progression of osteoarthritis is always accompanied by the physical degradation of the tissue. Machine learning techniques, which are gaining increasing popularity in the field of orthopaedic research, were then used to model the relationships between these biomarkers and the mechanical state of the tissue. These models were found to be highly accurate in characterizing the mechanical state of the tissue, even when limited only to the protein concentrations that one could find in a standard hospital lab. This study has not yet produced a tool which may be used in a hospital setting, considering the low number of patients included in this study, but it does reveal promising early results in using machine learning to characterize osteoarthritis, a task which has thus far eluded the orthopaedic research community.
Access is limited to the campuses of the University of Missouri.