Background: In Canada, osteoarthritis affects 4 million people and costs over 1.3 billion CAD annually in joint replacements. However, early detection remains a major challenge, as current clinical imaging tools cannot capture subtle tissue changes in the early stages, and the underlying mechanisms that drive disease progression are still not fully understood.
Research Objectives: This thesis investigates microstructural changes in TMM-induced OA mouse models using high-resolution X-ray microscopy (XRM). It focuses on optimizing imaging and segmentation methods to assess cartilage thickness, bone architecture, and cell morphology, with the goal of improving early OA diagnostics through detailed tissue-level insights.
Methodology: 6 Male C57BL/6 mice underwent TMM on the right knee at 8 weeks old. Two weeks later, operated and control contralateral knees were EpoFix resin embedded, harvested, and then imaged with XRM. Tissue components, including articular and calcified cartilage, subchondral bone plate, cortical and trabecular bone, and osteocytes and chondrocytes, were segmented using Attention U-Net deep learning. Cartilage thickness and cell volume changes were then quantified to assess tissue degradation.
Results: High-resolution XRM analysis revealed early osteoarthritis-induced increases in osteocyte volume and altered spatial organization in the femur. Chondrocyte sphericity was preserved, but depth-dependent shifts in cell distribution were detected. Calcified cartilage thickness increased regionally, while articular cartilage and subchondral bone plate thicknesses remained stable. Bone morphometry showed subtle femoral-specific changes in cortical and trabecular regions.
Conclusions and Future Work: High-resolution XRM enabled early detection of OA-related changes in joint morphology and cell organization, including osteocyte volume, chondrocyte distribution, and articular cartilage remodeling. Future work should explore comparative segmentation tools, regional cell density, and articular cartilage surface roughness, while expanding analysis beyond the early stages to better capture site-specific adaptations and improve OA diagnostics.