PET imaging of the bone
The outer layer of bones is formed of compact (cortical, dense) bone, and the inner part of spongy (cancellous, trabecular) bone with pores filled with red or yellow bone marrow.
Osseous tissue contains basically three cell types: osteoblasts, which deposit the hard extracellular matrix (ECM) of the bone; osteoclasts, which resorb the extracellular matrix; and osteocytes, surrounded by the matrix. ECM consists of hydroxyapatites and collagen.
Osteoclasts are part of the mononuclear phagocyte system (reticuloendothelial system). They are derived from differentiation and fusion of promonocytic precursors, differentiated from hematopoietic stem cells, and can be multinucleated and large in size. Like related macrophages and dendritic cells, also osteoclasts can function in acidic and hypoxic environment; acidic environment is produced by osteoclasts themselves to dissolve hydroxyapatite. Collagen is degraded with cathepsin K, secreted by osteoclasts. While osteoclasts contain mitochondria in very large number to power their high needs for energy, they can also rely on anaerobic glycolysis. Active osteoclasts can produce large amounts of ROS.
Osteoblasts originate from mesenchymal progenitors in bone marrow and periosteum; after maturation, osteoblasts produce the matrix components in abundant rough endoplasmic reticulum, and when bone matrix is formed, either undergo apoptosis or stay in the matrix as osteocytes with very long life-span. Aerobic glycolysis is active in osteoblasts in anabolic states, although they, too, contain mitochondria in large numbers. Mitochondria may produce ATP mostly from fatty acids and glutamine (Karner & Long, 2018; Kushwaha et al., 2018). Osteocytes are the main source of phosphate-regulating hormone FGF23, and sclerostin, which inhibits bone formation.
Sympathetic nerves follow arteries into and inside the bone. Osteoblasts and osteoclasts contain NMDA receptors (subtype of iGluRs in glutamatergic system), regulating bone mineralization and resorption.
Bones that are united by joints are covered with (articular) cartilage that contains chondrocytes. Normally chondrocytes maintain the cartilage by ECM remodelling. In arthritis the osteoclasts and chondrocytes disintegrate the ECM.
The high density of bones leads to strong attenuation, but it is appropriately corrected in modern PET/CT scanners. However, tracer uptake measured using PET/MRI may be underestimated compared to PET/CT because standard MRI-based attenuation correction does not accurately account for the attenuation by cortical bone (Fraum et al., 2016). CT can be used to assess bone density as Hounsfield Units.
Bone marrow constitutes 4% of the total body mass. Trabecular (spongy) bone contains the marrow adipose tissue (yellow bone marrow) and in the heads of long bones also hematopoietic tissue (red bone marrow) that produces red blood cells (RBCs), platelets, and white blood cells (Suchacki et al., 2016). At the time of birth, all of the bone marrow contains hematopoietic cells, but by the age those are replaced by adipocytes (Li et al., 2018). Erythropoietin (EPO) stimulates the production of RBCs, and EPO is mostly produced in the kidneys. Thrombopoietin (TPO) stimulates production of platelets and sustains the viability of haematopoietic stem cells. Hematopoiesis requires lot of energy, and those parts of the bone marrow have very rich vasculature, while quiescent regions can be hypoxic, and maintain undifferentiated hematopoietic stem cells. Hypoxia inhibits osteoblasts and activates osteoclasts. Erythropoietic bone marrow can be imaged with iron isotopes, or labelled transferrin. The metabolic and proliferative activity of bone marrow can be quantified with PET using tracers such as [18F]FDG, [18F]FLT, [11C]acetate, [11C]methionine, [11C]choline, and [18F]FGln (Agool et al., 2011; Zhu et al., 2019); non-invasive methods for detecting myelosuppression are clinically important for managing chemotherapy.
The bone marrow adipose tissue (MAT) accounts for about 8-10% of the total body fat mass in healthy adults, and 70% of the volume of bone marrow. The total mass of MAT is about 1.35 kg, ranging from 0.5 to 3 kg (Scheller et al., 2016). Bones accumulate large proportion of postbrandial lipoproteins in the circulation (Niemeier et al., 2008). Bone marrow fat can be measured with 1H MR spectroscopy. Regulated MAT (rMAT) is dispersed within hematopoietic cells while the constitutive MAT (cMAT) is mainly located in more distant areas (Suchacki et al., 2016). Fatty acids are mainly unsaturated in cMAT, and saturated in rMAT. MAT has endocrine functions, secreting at least adiponectin and leptin, and it also may have regulative role in bone remodelling. MAT mass is affected by temperature, and it increases in caloric restriction (although WAT decreases), osteoporosis and type 1 diabetes, generally showing negative correlation with bone density (Fazeli et al., 2013; Rendina-Ruedy and Rosen, 2017).
In acute leukaemia a hematopoietic cell line has lost the potential to differentiate. In contrast, myeloproliferative diseases (MPDs) are by cell lineage that has persisting potential for differentiation. MPD and chronic haemolytic anaemia can lead to extramedullary haematopoiesis (EMH), most often in the liver and spleen.
Bone metastases occur in red bone marrow, and can be detected with [18F]FDG PET (Blebea et al., 2007; Behzadi et al., 2018; Høilund-Carlsen et al., 2018). Granulocyte and granulocyte-macrophage colony stimulating factors (G-CSF and GM-CSF) can be used to improve chemotherapy-induced neutropenias and to reduce risk of infections; this treatment increases markedly [18F]FDG uptake in the bone marrow, which could be misinterpreted as bone marrow metastases (Sugawara et al., 1998).
Red bone marrow is often the dose-limiting organ in radioimmunotherapy. Bone-seeking 89Zr is can be used to label mAbs for immuno-PET, leading to relatively high absorbed dose. [18F]FDG and [18F]FLT PET can detect the red bone marrow, in order to spare it in radiotherapy (Wyss et al., 2016).
Bone metabolism and remodelling activity can be studied using [18F]fluoride and 99mTc-labelled bisphosphonates. Bisphosphonate-based PET tracers have also shown some promise (Wu et al., 2016). While bone density can be assessed with CT, [18F]fluoride uptake is not strongly correlated with bone density (Nawata et al., 2017). Reduced [18F]fluoride bone uptake can still been seen in osteoporotic women (Schiepers et al., 1997; Frost et al., 2004; Uchida et al., 2009; Reilly et al., 2018), and [18F]fluoride PET can be used in monitoring treatment response (Frost et al., 2011 and 2013). It should be noted that 99mTc-labelled bisphosphonates and [18F]fluoride bind to hydroxyapatite whether the hydroxyapatite is present in dead bone or hydroxyapatite implants with no new bone growth (Bernhardsson et al (2018)). Retention depends on the surface area of exposed hydroxyapatite.
Morbidly obese people have lower bone turnover than lean subjects, and it is restored after weight-loss surgery (Ivaska et al., 2017).
Paget’s disease is a benign metabolic bone disorder characterized with increased bone turnover. The disease itself, and the effect of treatment with bisphosphonates, can be studied using [18F]fluoride (Schiepers et al., 1997; Cook et al, 2002; Installé et al., 2005). [15O]H2O PET has shown that perfusion is increased in Paget’s disease, while it is decreased in old osteoporosis patients (Schiepers et al., 1997). The bone uptake of [11C]choline can be increased in Paget’s disease, and while it might cause wrong positive findings in oncological PET, it cannot be used to study Paget’s disease (Leitch et al., 2017).
Blood flow and volume
Bones are well vascularized, receiving 5.5-11% of cardiac output (Marenzana & Arnett, 2013; Tomlinson & Silva, 2013; Riddle & Clemens, 2017). Long bones get the blood from artery which is branched in the medulla to supply the medullary sinusoids and cortex; from cortex, the blood returns to the medullary sinusoids. Sinusoids drain into central sinusoid, and further into multiple veins that leave the bone. Arteries and arterioles may penetrate the bone marrow from several sites, and venous blood has several routes out of the bone as well; cortex receives blood also from small periosteal arteries; this has made the measurement of bone blood flow and metabolism difficult with traditional flow-meter and venous blood collection methods (Kane, 1968).
Bone angiogenesis is tightly coupled to bone formation processes. Two types of capillaries and fenestrated sinusoids affect the regional distribution of oxygen and other substrates and leukocyte trafficking (Stegen & Carmeliet, 2018). Impairment of the blood supply to the bone reduces growth and repair, and can cause bone loss and even necrosis.
Perfusion of the bone is usually linked to the metabolic activity, and is highly variable in different bones and bone regions (Brookes & Revell, 1998; Cook et al., 2000; Frost et al., 2009; Puri et al., 2013). Oxygen tension in bone marrow sinusoids is about half of that in arterial blood and similar to what is common in other normal tissues (Marenzana & Arnett, 2013). Ageing and cardiovascular disease causes reduced medullary bone perfusion, which may induce hypoxia, activating osteoclasts and inhibiting osteoblasts, leading to bone loss. Rat studies using microspheres have also shown marked variability of perfusion between skeletal sites, and age-related reduction of blood flow in bones (Bloomfield et al., 2002), and age reduces endothelium-dependent vasodilation in bones (Prisby et al., 2007). In a rat model of CKD, fluorescent microsphere measurement has shown increased blood flow in bone cortex and reduced blood flow in bone marrow (Aref et al., 2018).
[15O]H2O (radiowater) injection or [15O]CO2 inhalation has been used to measure blood flow in the bone (Schiepers et al., 1997; Piert et al., 1998; Kubo et al., 2001; Sörensen et al., 2003; Temmerman et al., 2008 and 2013; Heinonen et al., 2013; Raijmakers et al., 2014; Jødal et al., 2017) and bone marrow (Martiat et al., 1987; Kahn et al., 1994; Heinonen et al., 2018). Radiowater studies are often accompanied by [15O]CO inhalation study to measure blood volume. Blood volume in the bone marrow is variable in different bone sections (Iida et al., 1999).
Active osteoblasts and osteoclasts use glucose in aerobic and anaerobic glycolysis (Karner & Long, 2018). Glucose metabolism in bone marrow can be measured using [18F]FDG (Murata et al., 2006; Huovinen et al., 2014 and 2016; Derlin et al., 2015; van Vliet et al., 2016; Latva-Rasku et al., 2018; Heinonen et al., 2018). Osteoblasts express insulin receptors and IGF1Rs. In a mice study, insulin has been shown to increase bone [18F]FDG uptake (Zoch et al., 2016). Mature osteoblasts produce osteocalcin, which increases insulin synthesis and β-cell proliferation in the pancreas (Zoch et al., 2016).
Osteoblasts produce citrate, which forms ∼1% of the mass of bone mineral. Citrate is incorporated between the hydroxyapatite crystal leaflets, providing strength to the nanocomposite. Glutamine is important for the function of osteoblasts (Riddle & Clemens, 2017).
[18F]FDG can be used to study bone infections and to separate normal bone healing from bone infections (Koort et al. 2004 and 2005; Lankinen et al., 2012; Odekerken et al., 2014a and 2014b; Wenter et al., 2017; van Vliet et al., 2018). Posttraumatic and postoperative chronic osteomyelitis can be effectively localized by combined [18F]FDG and [18F]fluoride PET/CT (Christersson et al., 2018).
68Ga3+ has shown some promise in imaging bacterial bone infection in animal model (Mäkinen et al., 2005) and in human patients (Nanni et al., 2010). In porcine osteomyelitis model [18F]FDG and [11C]methionine detected infectious lesions better than [68Ga]Ga-citrate or AChE inhibitor [11C]donepezil (Jødal et al., 2017).
Chemokine receptors may provide more specific targets for imaging infection and inflammation of the bone.
Bone marrow is a common site for metastases of malignant tumours, and bone metastases are a major cause of morbidity and mortality. Bone metastases cause fractures, spinal cord compression, and severe pain. Breast, prostate, and lung cancers are frequently causing bone metastases. Bone metastases are often osteolytic, because of extensive medullary angiogenesis and activated osteoclast differentiation. Prostate cancer cells tend to induce bone formation by producing osteoblast-stimulating factors. 223Ra is a calcium-mimetic, bone-seeking α-emitter that is used to treat metastatic castration-resistant prostate cancer.
Osteosarcoma is the most common primary malignant bone tumour. Accurate initial staging and restaging after treatment is critical for survival. Glycolysis rate is increased in osteosarcoma, and [18F]FDG PET is useful in the staging, identification of metastases, and assessing treatment response in osteosarcoma, and promising also in chondrosarcoma, Ewing’s sarcoma and primary bone lymphoma (Behzadi et al., 2018).
Malignant and metastatic bone disease can be studied using [18F]FDG and [18F]fluoride and possibly other tracers (Apostolova & Brenner, 2010; Mosci & Iagaru, 2012; Beheshti & Langsteger, 2012). Accuracy of [18F]fluoride PET/CT is clearly superior to 99mTc-methylene diphosphonate (99mTc-MDP) CT for staging and restaging malignant bone diseases. Dynamic [18F]fluoride PET/CT is promising in differentiating benign from malignant bone lesions (Beheshti, 2018). [18F]Fluoride PET and [99mTc]HDP SPECT give similar SUV ratios and correlating SUVs in bone metastases of breast and prostate cancer (Arvola et al., 2019). [18F]fluoride is a feasible alternative to 99mTc-MDP also in paediatric patients (Usmani et al., 2018). [18F]FDG PET can be used for assessing treatment response in bone marrow of lymphoma patients (Goudarzi et al., 2010). PSMA tracers are useful in the detection of bone metastases in prostate cancer (Janssen et al., 2017; Sachpekidis et al., 2018; Harmon et al., 2018; Zacho et al., 2018). Integrin αvβ3 is overexpressed in active osteoclasts, which allows imaging of increased bone resorption and osteolytic bone metastases (Ocak et al., 2018).
[18F]fluoride SUVs in normal adult bones and in certain benign and malignant diseases have been reported by Win & Aparici (2014), Sabbah et al (2015) and Nawata et al (2017), and a variety of PET images have been shown by Sheth & Colletti (2012), Sarikaya et al (2017), and Woodhead et al., 2017.
Bone scintigraphy has been used to assess bone scan index (BSI), the percentage of skeletal mass affected by tumours (Imbriaco et al., 1998). Commercial software can automatically calculate BSI from 2D bone scintigram. Whole-body PET/CT allows assessment of BPIVOL, the percentage of bone volume (including bone marrow) affected by the cancer, and SUVmean inside that volume. BPISUV, calculated as
has been shown to agree well with expert rating, and has been proposed to be used for objective assessment of osseous tumour burden in prostate cancer, and treatment response evaluation and outcome prediction (Bieth et al., 2017).
- Analysis of [18F]fluoride bone scans
- Analysis of [18F]FDG
Allen MR, Burr DB: Basic and Applied Bone Biology. Academic Press, 2014.
Blake GM, Park-Holohan S-J, Cook GJR, Fogelman I. Quantitative studies of bone with the use of 18F-fluoride and 99mTc-methylene diphosphonate. Semin Nucl Med. 2001; 31(1): 28-49. doi: 10.1053/snuc.2001.18742.
Bronner F, Farach-Carson MC, Rubin J (eds.): Bone Resorption. Springer, 2005. ISBN 978-1-84628-016-0. doi: 10.1007/b136184.
Brookes M. Revell WJ. Blood Supply of Bones - Scientific Aspects. Springer, 1998. doi: 10.1007/978-1-4471-1543-4.
Burr DB, Allen MR (eds.): Basic and Applied Bone Biology. Elsevier, 2014. ISBN: 978-0-12-416015-6.
Davies AM, Sundaram M, James SLJ (eds.): Imaging of Bone Tumors and Tumor-Like Lesions - Techniques and Applications. Springer, 2009. doi: 10.1007/978-3-540-77984-1.
Grant FD, Fahey FH, Packard AB, Davis RT, Alavi A, Treves ST. Skeletal PET with 18F-fluoride: applying new technology to an old tracer. J Nucl Med. 2008; 49(1):68-78. doi: 10.2967/jnumed.106.037200.
Hsi ED (ed.): Hematopathology, 3rd ed., Elsevier, 2018. ISBN: 978-0-323-47913-4.
Krishnamurthy GT, Huebotter RJ, Tubis M, Blahd WH. Pharmaco-kinetics of current skeletal-seeking radiopharmaceuticals. Am J Roentgenol. 1976; 126(2): 293-301. doi: 10.2214/ajr.126.2.293.
Lieberman JR, Friedlaender GE (eds.): Bone Regeneration and Repair - Biology and Clinical Applications. Humana Press, 2005.
Love C, Palestro CJ. Nuclear medicine imaging of bone infections. Clin Radiol. 2016; 71(7): 632-646. doi: 10.1016/j.crad.2016.01.003.
Martiat P, Ferrant A, Cogneau M, Bol A, Michel C, Rodhain J, Michaux JL, Sokal G. Assessment of bone marrow blood flow using positron emission tomography: no relationship with bone marrow cellularity. Br J Haematol. 1987; 66(3): 307-310. doi: 10.1111/j.1365-2141.1987.00301.x-i1.
McCarthy I. The physiology of bone blood flow: a review. J Bone Joint Surg. 2006; 88-A (Suppl 3): 4-9. doi: 10.2106/JBJS.F.00890.
Palestro CJ. Radionuclide imaging of osteomyelitis. Semin Nucl Med. 2015; 45: 32-46. doi: 10.1053/j.semnuclmed.2014.07.005.
Piert M, Machulla HJ, Jahn M, Stahlschmidt A, Becker GA, Zittel TT. Coupling of porcine bone blood flow and metabolism in high-turnover bone disease measured by [15O]H2O and [18F]fluoride ion positron emission tomography. Eur J Nucl Med Mol Imaging 2002; 29(7): 907-914. doi: 10.1007/s00259-002-0797-2.
Piert M, Zittel TT, Becker GA, Jahn M, Stahlschmidt A, Maier G, Machulla HJ, Bares R. Assessment of porcine bone metabolism by dynamic [18F]fluoride ion PET: Correlation with bone histomorphometry. J Nucl Med. 2001; 42:1091-1100.
Qin L, Genant HK, Griffith JF, Leung KS (eds.): Advanced Bioimaging Technologies in Assessment of the Quality of Bone and Scaffold Materials - Techniques and Applications. Springer, 2007. ISBN 978-3-540-45456-4. doi: 10.1007/978-3-540-45456-4.
Resnick D, Kransdorf MJ: Bone and Joint Imaging, 3rd ed., Elsevier, 2005. ISBN 0-7216-0270-3.
Riddle RC, Clemens TL. Bone cell bioenergetics and skeletal energy homeostasis. Physiol Rev. 2017; 97: 667-698. doi: 10.1152/physrev.00022.2016.
Rohren EM, Macapinlac HA. Spectrum of benign bone conditions on NaF-PET. Semin Nucl Med. 2017; 47: 392-396. doi: 10.1053/j.semnuclmed.2017.02.008.
Updated at: 2019-12-09
Created at: 2015-08-18
Written by: Vesa Oikonen