Effect of fat-contamination and fat-suppression on T2 quantitation of knee articular cartilage in vivo

• Main Author: Paakkari, P. (Petri)
• Format: Dissertation
• Language: English
• Published: University of Oulu 2018
• Subjects: Physics
• Online Access: http://urn.fi/URN:NBN:fi:oulu-201805312261; http://nbn-resolving.de/urn:nbn:fi:oulu-201805312261

This study aims to investigate the effect of fat contamination and fat suppression (FS) on in vivo T2 mapping of knee articular cartilage. T2 mapping is a technique used in magnetic resonance imaging (MRI) and it has advantages over radiography techniques, such as no ionizing radiation and T2 maps give information about cartilage content and structure. Knee articular cartilage has high water concentration and near the cartilage, there is fat tissue, i.e. bone marrow and subcutaneous fat. Fat causes chemical shift artifact and can contaminate T2 values of T2 maps. To minimize chemical shift artifacts, FS is needed to suppress the fat signal. Four volunteers were imaged on a 3 Tesla MRI clinical scanner and T2 values were calculated in several regions of tibiofemoral cartilage using a multi-echo spin-echo sequence (MSME) with and without FS (NoFS). Two frequency encoding directions, superior-inferior (SI) and inferior-superior (IS), were used to evaluate the effects of FS on chemical shift artifact. Sagittal slices of the knee were manually segmented to 18 region of interests (ROIs) and then T2 maps were calculated with MATLAB based scripts. The repeatability of segmentation was evaluated, which was achieved by segmenting same slices multiple times. Differences between NoFS T2-SI and NoFS T2-IS ranged widely and in 11 out of 18 ROIs the differences were high, suggesting that the chemical shift artifact affects most of the cartilage regions. Particularly affected were the superficial and deep femoral cartilage and deep tibia. Differences between FS T2-SI and FS T2-IS values were low to moderate in most of the ROIs, with the highest differences observed in deep parts of anterior and central femur. Comparing NoFS T2 and FS T2 values with the same frequency encoding direction, differences were on average higher using SI direction than IS direction, and the deep ROIs were the most affected. Repeatability of segmentation was great in eight ROIs without FS and six ROIs with FS and the average differences were low for both NoFS and FS T2 mapping. For NoFS T2 mapping, segmentation repeatability was high in deep tibial cartilage. On the contrary, FS T2 mapping showed that the repeatability of the segmentation was low to moderate in superficial parts of posterior femur and deep parts of posterior tibia. The use of FS improved slightly repeatability of the cartilage segmentation in several regions and reduced the chemical shift artifacts. However, the regional heterogeneity in FS sequence introduced further uncertainties in T2 measurements.