Three-dimensional T1 quantification techniques for assessment of cartilage quality using dGEMRIC

Osteoarthritis (OA) is a common chronic disease and one of the major global causes for functional disabilities. The disease is characterized by loss and degradation of cartilage, commonly affecting the knees and hips. Delayed gadolinium-enhanced MRI of cartilage (dGEMRIC) is a previously presented method for identification of early OA using magnetic resonance imaging (MRI). By quantifying the T1 parameter in the cartilage after distribution of a contrast agent, a measure of the glycosaminoglycan (GAG) content is retrieved, which in turn is known to be decreased at early stages of OA. In this thesis a series of three-dimensional (3D) T1 quantification methods have been developed and evaluated for use in dGEMRIC. Traditionally, such sequences have not been widely adopted for this purpose and in this work it has been shown that the main obstacles are related to B1 variations within the volume. As part of the work a 3D Look-Locker (3D-LL) T1 quantification pulse sequence have been created. In addition, new methods were developed for correction of B1 inhomogeneities and slab profile flip angle variations in the 3D-LL data. From in vivo and phantom measurements the methods were shown to be reliable, with T1 results that agreed very well to gold standard two-dimensional inversion recovery (2D-IR). A 3D variable flip angle (3D-VFA) T1 quantification sequence in combination with a B1 mapping sequence was also investigated. The results with and without B1 correction was studied in vivo and in phantoms. It was concluded that 3D-VFA should always be used with B1 correction, especially at higher field strengths In addition, two dedicated clinical studies were set up, to both verify the usability of the developed methods and to explore new dGEMRIC applications. In one of these studies, the repeatability of successive dGEMRIC measurements for each of the T1 quantification methods was investigated for a group of subjects at risk of developing OA. 2D-IR and 3D-LL performed equally well, while 3D-VFA (without B1 correction) was inferior. Repeatability was shown to be similar to previously reported results in healthy subjects. The other of these studies was a time-response study, using 3D-LL, to explore the feasibility of performing dGEMRIC in the meniscus. It was concluded that the temporal contrast uptake in the meniscus follows that of the articular cartilage and that differences can be seen between vascular and avascular parts of the meniscus. The overall conclusion of this work is that 3D T1 quantification in dGEMRIC is feasible and should allow for both new and improved means of diagnostics