Bone is a highly vascularized and metabolically active tissue that undergoes continuous remodeling. Bone remodeling in the joint serves to adjust bone architecture to meet changing mechanical loads on the joint, to repair microdamages in the bone matrix, as well as in response to hormonal signaling, trauma, or bone pathology (1,2). While bone remodeling is important to the normal homeostasis of bone function, increased bone remodeling has been linked to several musculoskeletal disorders including osteoarthritis and osteoporosis.

Magnetic Resonance Imaging (MRI) plays a pivotal role in the clinical evaluation of the musculoskeletal system, including muscle, tendon and ligaments, bone marrow, and more recently bone. While bone is ordinarily invisible in MR images as a result of the extremely short life-times (T2) of the protons (tightly bound water and collagen), its structure can be inferred from the signal of the surrounding soft tissues, i.e. bone marrow. In addition to high-resolution morphologic information, quantitative MR methods have been developed to study bone composition and structure to assess fracture risk. However, these methods have yet to be applied to cortical-like subchondral bone in joints. Further, metabolic information regarding bone stress and remodeling cannot be directly measured using this modality. Finally, as structural changes occur at a later stage in disease, these degenerative changes may not be reversible or at a point when other tissues are affected. Detection of metabolic changes maybe an important marker of early disease and evaluation of efficacy of disease modifying treatments.

18F-NaF is a long recognized bone-seeking agent that is able to probe bony remodeling(3-5). Early investigations demonstrated that the concentrations of 18F ions are 10 times higher in areas of regenerating bone, compared with areas of normal bone(3). Uptake of 18F-Fluoride indicates osteoblastic activity by identifying reactive changes in the affected bone(6,7). Standardized Uptake Values (SUV), the tissue activity concentration normalized by the fraction of the injected dose/unit weight, provides a measure of PET radiotracer uptake. New hybrid PET-MRI systems offer to combine the benefits of MRI high-resolution morphologic imaging with functional information from PET for simultaneous, sensitive and quantitative assessment of subchondral bone in OA. This presents an opportunity to not only assess bony activity, but to also to characterize established quantitative and qualitative MRI metrics of bone health. Finally, PET-MR systems offer an opportunity to relate metabolic, biochemical and structural properties of bone to degenerative changes in adjacent tissues.

18F-NaF was first recognized as a bone-seeking agent in 1962(3) and has been approved for PET imaging by the food and drug administration (FDA) since 1972. The mechanism of skeletal uptake of 18F-NaF is based on ion exchange. Bone tissue is continuously renewing itself through remodeling, and this remodeling occurs at the bone surface. 18F ions exchange with hydroxyl ions (-OH) on the surface of the hydroxyapatite to form fluoroapatite. This exchange occurs at a rapid rate; however, the actual incorporation of 18F ions into the crystalline matrix of bone may take days or weeks. Uptake of 18F-NaF is a function of osseous blood flow and reflects bone remodeling. Abnormal areas of increased 18F-NaF uptake depicted on PET images are due to processes that increase exposure of the surface of bone and provide a higher availability of binding sites, such as osteolytic and osteoblastic processes.

Several unique characteristics of 18F-NaF make it a desirable radiotracer for imaging of bone. 18F-NaF has minimal binding to serum proteins, which allows a rapid single-pass extraction and fast clearance from the soft tissues. This high bone uptake and faster soft-tissue clearance lead to a higher quality of images with high bone-to-background ratio and allow for shorter imaging times. 18F-NaF PET imaging can be performed less than 1 hour after injection.

Osteoarthritis (OA): OA is a chronic degenerative disease affecting all tissues in the joint. MRI has long been utilized to non-invasively study and understand many of the complex disease processes involved in OA(8). MRI provides excellent high-resolution morphologic information of joint tissue. However, structural degenerative changes observed on MRI are likely at a late stage in the disease process when tissue loss has already occurred and treatments are unlikely to be effective. While quantitative MRI techniques, such as T2 or T1ρ relaxation times, is able to provide information regarding of tissue microstructure in OA, it is mostly studied in soft tissue such as cartilage (9-11). The role of bone injury or remodeling is often neglected or studied in the context of qualitative assessment of osteophytes and marrow signal changes(12).

18F-NaF PET relates information regarding osteoblast activity in subchondral bone which can be correlated with high-resolution quantitative MR methods of other tissues to study the pathogenesis of OA. Initial results have shown significantly different levels of 18F-NaF uptake in various subchondral bone pathology (bone marrow lesions, osteophytes, sclerosis) identified on MRI (13). Furthermore, high 18F-NaF uptake in subchondral bone did not always correspond to structural damage detected on MRI. Subchondral bone is a region that is associated with the development of pain as well as cartilage degeneration, and early 18F-NaF PET/MR data suggests that metabolic abnormalities in the bone may occur prior to structural changes are seen on MRI. Additionally, another PET tracer, fluorodeoxyglucose (18F-FDG) can provide commentary information about inflammatory processes occurring in the joint (14). This allows PET/MR to simultaneously assess several early metabolic and biochemical marker of knee OA progression across all tissues in the joint. This may provide new insights into OA pathogenesis and lead to new treatment targets to arrest the onset and progression of OA.

Metabolic Bone Disorders: Quantitative dynamic 18F-NaF PET has also been used to study low-turnover bone diseases, including osteoporosis, and to assess age-related changes in pre- and postmenopausal women(15,16). Bone metabolism and fluoride binding to bone mineral were found to be significantly reduced in osteoporosis, whereas at the same time biochemical markers of bone turnover (bone-specific alkaline phosphatase) were increased. This somewhat surprising result highlights the importance of regional measurements of bone turnover (from imaging) to improve the understanding of metabolic bone diseases. MR, in addition to obtaining high-resolution morphological images, is able to characterize bone quality through measures of cortical bone water and bone marrow composition(17). Together, PET-MR offers an opportunity to longitudinally study multiple facets of disease and therapeutic efficacy.

Radiotracer Dose: MR knee and hip protocols (20-60 minutes) are considerably longer than the data collection time in one patient bed position in clinical PET-CT (3-4 minutes). As current PET-MR hybrid systems allow all of the MR scan time to be used to collect PET data, the injected dose of radiotracer can be significantly reduced. In initial studies, acquiring PET data for the duration of the 30-minute MRI protocol, doses as low as 2.5 mCi (1/4 of clinical dose) of 18F-Fluoride was used while maintaining a high SNR in PET SUV maps.

PET Attenuation Correction (AC): In order to obtain accurate PET images, emission data recorded during a PET scan must be corrected for tissue and hardware attenuation. This is performed during reconstruction using an attenuation map (µ-map). Differences in PET attenuation between tissue fat and water as well as attenuation are accounted for using fat-water separated images and known AC coefficients. However, attenuation correction for MRI Coils is a new challenge that came about with the development of simultaneous PET/MR imaging systems(18,19). AC of these hardware components is required as their presence leads to considerable attenuation of the PET signal(20). Rigid and stationary MR hardware components, such as the patient table, are corrected for by integrating computed tomography (CT)-based attenuation templates of these parts at a fixed position in the μ-map used for AC in PET image reconstruction(21). However flexible radiofrequency (RF) coils, which have considerably reduced attenuation of PET photons, are currently disregarded in MR based AC (MRAC) since their position and individual geometry are unknown in patient scans. Additionally, cortical bone, which is an important area of study in osteoporosis, has higher attenuation of PET photons than soft tissue, and is also unaccounted for in MRAC due to its lack of signal on MR images. Investigations are ongoing to characterize and correct for MR hardware and cortical bone attenuation.