Osteoarthritis (OA) is a degenerative joint disease that affects 12% of the adult population of the United States and costs the health care system $45-80 billion per year in direct and indirect costs[20, 112]. Currently, there is no cure for OA. It is managed with conservative treatment in the early stages of the disease and necessitates a joint replacement in advanced stages. One of the challenges of developing effective treatments for OA is that it cannot be detected early enough to intervene before irreversible soft tissue damage has occurred using the most prevalent detection method, a standing radiograph. In this dissertation, apparent diffusivity of a contrast agent measured using computed tomography arthrography (CTa) scans is introduced as a potential biomarker for early-stage cartilage degeneration. This could expand access to early detection methods to patients for whom a magnetic resonance imaging (MRI) exam is inaccessible due to cost, availability in their area, or MRI-incompatible medical devices.

While in theory, apparent diffusivity will be elevated in cartilage with early signs of degradation compared to healthy cartilage, it is not a perfectly quantitative measurement. This is because apparent diffusivity is measured by fitting the Fickian diffusion model to contrast agent concentration in cartilage, a method that assumes that only diffusion is taking place. Cartilage responds to osmotic changes in its environment, which means that interstitial fluid flow is also occurring, which can accelerate or slow down solute transport. The first aim of this dissertation (Chapter 3) seeks to answer the question of how quantitative apparent diffusivity is by measuring it in enzymatically degraded cartilage explants and healthy, unaltered ones in the presence of different contrast agent concentrations and boundary conditions. In this study, it was shown that degradation detection using apparent diffusivity was robust to a wide range of contrast agent concentrations from 6-50% by volume. It was also shown that bulk apparent diffusivity is different in absorption and desorption. Simulations were used to predict the effects of imaging protocol variation on diffusivity measurements with a large range of resolutions and scan times. This allowed a range of scan parameters to be examined without the confounding factors introduced by using multiple imaging systems. The results of these simulations showed a decrease in apparent diffusivity as voxel size increased, a decrease in apparent diffusivity as total scan time increased, and a decrease in apparent diffusivity as scan frequency increased. These studies could provide a framework for future comparison of apparent diffusivity measurements across institutions.

Aims 2, 3, and 4 of this dissertation are steps towards translating this method for eventual in vivo use. In Aim 2 (Chapter 4), two different methods for making 3D voxel-wise apparent diffusivity measurements are presented. They are used to analyze a bovine explant dataset consisting of enzymatically degraded and unaltered cartilage samples to confirm that voxel-wise measurements show increases in apparent diffusivity in damaged tissue, like the bulk measurements did in Aim 1. Aim 3 (Chapter 5) details a method for identifying the cartilage surface in areas where it is in contact with cartilage or menisci, a step towards overcoming the challenge of accurate cartilage segmentation in CTa datasets. In Aim 4 (Chapter 6), simulations are used to study the effects of contrast agent solutions on cartilage mechanical properties. Solute size, bath ionic strength, and bath osmolarity are altered individually, which is not possible in the laboratory, and their effect on indentation force is quantified. It was found that increased bath ionic strength and osmolarity generally corresponded to decreased compressive stiffness. Increased solute size had mixed effects on transient compressive stiffness that depended on the other properties of the test solution. This work increases the utility of apparent diffusivity measurements by enabling its measurement in 3D and providing context for benchtop confined compression experiments.

The work presented in this dissertation could enable CTa-based detection of cartilage degeneration, expand access to early OA detection, and greatly increase treatment options in conjunction with the development of new interventions.