Advanced Biochemical MRI of the Meniscus at 3 and 7 Tesla

The principal functions of the meniscus are load transmission and shock absorption. The meniscus is made of a dense extracellular matrix composed primarily of water and collagen fibers, interspersed with cells. Glycosaminoglycans (GAG), non-collagenous proteins, and glycoproteins account for the remaining dry weight. Up to now, magnetic resonance imaging has been used for macroscopic description of meniscal tissue in order to diagnose the lesions. The recent development of the novel collagen- and GAG- sensitive MRI techniques provides the insight into connective tissues, including menisci, in non-invasive way. As a highly organized tissue, meniscus is a challenging for MRI, since at the echo time (TE) of the conventional MRI sequences, the signal from meniscus is completely or partly decayed. It is known that there is a macromolecular remodeling present in degenerative menisci. The interplay between collagen fibers (orientation and content) and water molecules is crucial for generating MRI signal. The high degree of meniscus collagen matrix anisotropy causes the splitting of T2* decay into several components. Recently developed modern MRI techniques allow for acquiring the MR signal also from the shortest T2* components which opens a variety of possibilities of using T2* as a marker for collagen fibers condition. Previously, the meniscus was predominantly investigated by qualitative MRI approaches, but recently few quantitative studies appeared in scientific community. These include mono-exponentially calculated T2*, relaxation time in the rotating frame (T) and longitudinal relaxation constant T1 with the presence of contrast agent. The goal of this project is to establish additional quantitative MRI techniques mainly focused on the meniscus component not investigated yet - GAG. The sodium imaging and gagCEST will be investigated as the potential markers for meniscus degeneration, since both of these techniques are GAG-specific. Moreover, the T2* will be deeper investigated with the emphasis of the short T2* component which sensitive to small changes in collagen mesh. All these techniques will be validated by immuno-histological assessment, polarized-light microscopy and pattern recognition algorithm. After these techniques will be established in in-vitro specimens, the group of patients with meniscal degenerations and tears will be scan as well as age-matched group of healthy volunteers. The techniques developed in this project will allow radiologists, orthopedic and trauma surgeons to diagnose, pre-operatively plan and post-operatively monitor meniscus degenerationears using advanced non-invasive MR methods.

The results of this project showed that the non-invasive assessment of biochemical composition of the meniscus and high-resolution vTE imaging, in addition to conventional morphology of the knee meniscus by magnetic resonance imaging, helps to clarify the function of the normal meniscus and improve the early diagnosis of meniscus degeneration and tears. Furthermore, we showed that MRI parameters, such as transversal relaxation constant (T2*) correlates with biochemical parameters of menisci, and thus, are able to non-invasively diagnose the status of degenerated and/or torn menisci and moreover to link it with the structural changes in the meniscal tissue. The results of this project contributed to the non-invasive assessment of meniscus using quantitative MRI. It helped to understand the meaning of the transversal relaxation constant (T2*) and its relationship to collagen fiber orientation and organization. The most important finding is the orientational dependence of the MRI signal on the orientation towards main magnetic field which has a potential significant clinical impact as the hyper-intense regions typically suggest the tissue depletion and may be mixed up with so called magic angle effect. We have also studied two glycosaminoglycan (GAG)-specific MRI methods which might have also bring valuable information on GAG content and distribution in menisci and potentially serve as a marker for early meniscus degenerations. However, due to challenging nature of these two methods we were not able conclusively demonstrate their clinical applicability. These investigations resulted in formulating new hypothesis and research questions which will be shortly translated into new project proposals. During the project, we developed two methods which were inspired by a need for advanced data evaluation. The first one expands the evaluation of the quantitative MRI images and maps - the relationship between individual pixels can bring more information than a simple average value of certain quantitative measure. We have developed a grey-level co-occurrence matrix based texture analysis (GLCM) for advanced analysis of cartilage. In the future, we plan to extend this approach to other tissues of interest for osteoarthritis, such as menisci, tendons, and ligaments. Secondly, we have developed a tool for automated evaluation of the knee cartilage using cartilage segmentation and co-registration on quantitative maps resulting in robust unsupervised knee evaluation. In the future, this tool will be applied to other parts of the knee (meniscus, ligaments) and extended by Deep Learning based tissue segmentation. The results were disseminated through number of peer-reviewed publications (11) and conference presentations (17+) and invited lectures (2). This project funds were used to financially support 7 individuals (2 PosDocs, 4 PhD students and 1 master student).