Osteoarthritis is a degenerative disease affecting entire joints and leading to pain, stiffness, and loss of mobility. It affects around 13% of the Canadian population and commonly presents in the knee. Traditionally, osteoarthritis has been visualized using radiography because it is the most accessible imaging method and can detect bone alterations, but this method is unable to show changes to the articular cartilage and meniscus, which have been shown to play an important role in the disease process. Quantitative magnetic resonance imaging (qMRI) is able to provide images of the soft tissue within the knee joint as well as numerical values representative of the state of the tissue health. One particular qMRI technique is quantitative magnetization transfer (qMT), and it allows for the determination of the properties of the bound pool within tissues (macromolecules such as proteoglycan and collagen) that has resonance too short to be captured with conventional MRI. Because qMT probes the properties of the hydrogen bound to macromolecules, it is expected to be more sensitive to the changes in composition of a tissue associated with osteoarthritis. The primary objective of this research is to establish a relationship between qMT parameters (f, k, T_2b relaxation time, T_2f relaxation time, and T_1f relaxation time) and the biochemical, histological, and mechanical properties of human articular cartilage and meniscus, and a secondary objective is to compare in vivo to ex situ qMT parameters. Two separate studies were conducted using differing populations in order to accomplish these objectives.

The first study assessed six human cadaver knees with no history of injury or illness in order to validate the methods and gain a baseline of values to be expected in a healthy population. Intact cadaver knees were imaged using qMT MRI techniques and qMT parameters extracted. Subsequent to imaging, core samples were taken from each meniscus and digested and assayed to determine the liquid, collagen, and proteoglycan contents. Menisci were dissected into pieces for histology and scored using an established histological scoring system customized to the meniscus. Pearson product moment and Spearman’s rho correlation coefficients were calculated for the biochemistry and histology results respectively compared to the qMT parameters to determine if any of the imaging metrics were predictive of the biochemical content or histological score. Results of this study showed several significant correlations between the qMT parameters and tissue properties. Some of these key findings included correlations in the collective samples where increasing liquid content was associated with decreasing bound pool fraction (r=-0.248, p<0.5);​ increasing collagen per dry mass showed increasing T_1f (r=0.413, p<0.01) and T_2f (r=0.510, p<0.01);​ and an increase in total histology score was related to a decrease in T_1f (ρ=-0.232, p<0.05), T_2f (ρ=-0.277, p<0.01), and T_2b (ρ=-0.207, p<0.05). In the medial side samples, key correlations were observed between increasing collagen per dry mass and increasing T_1f (r=0.477, p<0.01), T_2f (r=0.585, p<0.01), and T_2b (r=0.415, p<0.05);​ and increasing histology score and decreasing T_1f (ρ=-0.232, p<0.05), T_2f (ρ=-0.277, p<0.01), and T_2b (ρ=-0.207, p<0.05). In the lateral side samples, key correlations were between increasing liquid content and decreasing f (r=-0.380, p<0.05) and increasing sulfated glycosaminoglycan (sGAG) per wet mass was associated with increases in f (r=0.391, p<0.05) and kf (r=0.404, p<0.05).

The second study focused on an end-stage osteoarthritis population by assessing total knee arthroplasty patients. The aim of this study was to explore the relationships between qMT parameters and tissue properties in damaged tissue. Two patients were scanned using the qMT MRI protocol prior to their surgery, and the excised tissues were scanned post-operatively using the same sequence. From these samples, seven separate articular cartilage and meniscus surfaces (both medial and lateral) were assessed. After imaging, the surfaces underwent mechanical indentation testing and the instantaneous modulus, elastic fit mean squared error, and tissue thicknesses were determined. Core samples were then removed from the surfaces for biochemical and histological analysis. Biochemistry protocols were the same as utilized in the cadaver study, and histology preparation was the same as well with different scoring methods used depending on the tissue type (articular cartilage versus meniscus). Pearson and Spearman correlation coefficients were once again determined in order to assess correlations between the qMT parameters and the tissue properties. A Wilcoxon signed rank test was performed to assess differences between in vivo and ex situ qMT results. The key results of this study showed significant correlations in the in vivo cartilage between increasing instantaneous modulus and decreasing T_1f (r=-0.221, p<0.05) and T_2f (r=-0.233, p<0.05) in the lateral side samples;​ increasing liquid content and T_1f (r=0.836, p<0.05) in the lateral samples;​ and histology score and f in the combined samples (ρ=0.670, p<0.05) and medial samples (ρ=1.000, p<0.01). In the ex situ cartilage, significant correlations were found between increasing histology score and decreasing T_2b (ρ=-0.896, p<0.01) in the lateral samples. In the lateral menisci samples in vivo, key correlations were found between increasing liquid content and decreasing kf (r=-0.890, p<0.05);​ increasing sGAG/dry mass and increasing T_2b (r=0.869, p<0.05);​ and increasing collagen/wet mass and increasing kf (r=0.820, p<0.05). In the lateral ex situ menisci, a negative correlation was observed between instantaneous modulus and T_2f (r=-0.563, p<0.05). In the global surface analysis (combining all cartilage and meniscus surfaces), key correlations were between increasing liquid content and increasing T_1f (r=0.926, p<0.01) and T_2f (r=0.864, p<0.05);​ increasing sGAG/dry mass and increasing T_1f (r=826 p<0.05) and T_2f (r=0.964, p<0.01);​ increasing collagen/dry mass and decreasing T_1f (r=-0.780, p<0.05);​ and increasing histology score and increasing T_2f (ρ=0.893, p<0.01). Significant decreases in T1obs, T_1f, T_2f and T_2b were also found from in vivo to ex situ scanning environments.

The findings in the correlation analysis of this project show the potential of qMT MRI imaging as a valuable modality for determining the structure, function, and composition of osteoarthritic articular cartilage and meniscus. It has been shown that ex situ qMT parameters are not the same as in vivo but steps have been made in a direction towards quantifying the relationships between the differing environments. Possible uses of this technique lie in early diagnosis of OA, monitoring of disease progression, and evaluation of potential treatments.