Molecular Mechanisms of Bone ¹⁸F-NaF Deposition

PREPARATION OF ¹⁸F-NaF

¹⁸F-fluoride is produced by 11-MeV proton irradiation of ¹⁸O-water in a tantalum target body (10) using a cyclotron. The irradiated aqueous solution containing ¹⁸F-fluoride is diluted with sterile water (5 mL) and passed through a cation exchange (H⁺ form) cartridge. The eluent from the cation exchange cartridge is passed through an anion exchange (HCO₃⁻ form) cartridge to trap the ¹⁸F-fluoride. The anion exchange cartridge is then flushed with 10 mL of sterile water, and the ¹⁸F fluoride is then eluted with 10 mL of sterile normal saline and is passed through a sterile filter into a sterile multidose vial. The quality control tests for the ¹⁸F-NaF are conducted according to the method of the U.S. Pharmacopeia (11).

KINETICS OF ¹⁸F-NaF

The rate-limiting step of ¹⁸F-NaF bone uptake is blood flow (12), and almost all ¹⁸F-NaF delivered is retained by bone after a single pass of blood (13). The initial ¹⁸F-NaF distribution therefore represents blood flow that varies among different bones (14). Uptake of tracer by the bone marrow is negligible (15). Around 30% of the injected ¹⁸F-NaF is present in erythrocytes, but this fact does not impede ion exchange in bone because ¹⁸F-NaF is freely diffusible across membranes (16). ¹⁸F-NaF is rapidly cleared from plasma and excreted by the kidneys. One hour after injection, only 10% of ¹⁸F-NaF remains in plasma, and the 5-h integrated plasma concentration was only 1.1%–2.6% of the dose per liter. The 5-h cumulative urine excretion was 7.4%–24.8% in 8 patients (17).

For bone deposition, ¹⁸F-ions need to “pass from plasma through the extracellular fluid space into the shell of bound water surrounding each crystal, onto the crystal surface and in the interior of the crystal,” as described by Blau et al. (12). After chemisorption onto hydroxyapatite, ¹⁸F exchanges rapidly for OH on the surface of the hydroxyapatite matrix (Ca₁₀(PO₄)₆OH₂) to form fluoroapatite (Ca₁₀(PO₄)₆F₂) (18). The incorporation of ¹⁸F into bone matrix as fluoroapatite is slow (days to weeks) (12).

Bone surface encompasses 300 m²/g of bone tissue for a total of 3 × 10⁶ m² for the whole body (19). ¹⁸F-NaF uptake and retention in bone depends on the area of the “exposed” bone surface, which is larger in a variety of benign and malignant bone disorders. The relationship between osteoblastic and osteoclastic activity determines the incorporation of ¹⁸F-NaF into the bone matrix (20). Negligible plasma protein binding, rapid blood and renal clearance (21), and high bone uptake early after injection result in high target-to-background ratios, which in turn permit whole-body imaging as early as 45–60 min after tracer injection.

TRACER KINETIC MODELS

Quantitative bone imaging is potentially important because several metabolic bone diseases are subtle at onset, diffuse in nature, and difficult to detect with other techniques, including blood- or urine-derived biomarkers or other imaging techniques (22). For instance, measurements of bone density are of unknown prognostic value for fracture risk assessments and unproven for monitoring disease during treatment (23). Quantitative measurements of bone blood flow and metabolism can be derived from dynamic PET with ¹⁸F-NaF and appropriate tracer kinetic models. Such models can provide quantitative measurements in patients with a variety of bone diseases and might greatly aid drug response assessments.

The model that best describes fluoride incorporation into bone is a simple 3-compartment model (18) consisting of a vascular, an extravascular, and a bone compartment. The model describes regional rather than whole-body pharmacokinetics (24,25) (Fig. 2) and is similar to the standard in vivo model for ¹⁸F-FDG (26). This and other models have been described previously (27,28).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

¹⁸F-NaF kinetic model with 3 compartments: vascular (1), extravascular (2), and bone (3). Plasma clearance of ¹⁸F-NaF is measured in left ventricle or aorta with PET scanner or from arterial blood draw; k₁–k₄ are rate constants; k₁ and k₂ represent forward and reverse transport from plasma, and k₃ and k₄ represent uptake and release from bone. If extraction fraction equals 1, then k₁ represents local bone blood flow. Influx rate K_i = k₁ · k₃ · (k₂ + k₃)⁻¹ is related to Ca²⁺ influx and bone apposition rate and, presumably, represents bone remodeling rate. Ki is determined by both bone blood flow and bone turnover. Therefore, measurements have to be interpreted in context of each individual study. Respective concentrations of fluoride are denoted as C_p (in plasma), C_e (in extravascular space), and C_b (in bone compartment). RBC = red blood cells

ANIMAL EXPERIMENTAL STUDIES

PET has been used to measure bone blood flow (29), first-pass bone extraction fraction (30), whole-body biodistribution (21,31), tracer kinetics (24,25), and the pharmacokinetics (17) of ¹⁸F-NaF. These and other studies were reviewed extensively by Blake et al. (8).

The normal kinetics and biodistribution of ¹⁸F-NaF have also been exploited to study physiologic processes. For instance, assessing renal function is of critical importance in preclinical drug development. Traditionally, this requires blood and urine sampling, which is difficult in small animals. Because renal fluoride clearance is related to the glomerular filtration rate, the pharmacokinetics of ¹⁸F-NaF were used to assess renal function in rats (32). The ¹⁸F-NaF–based measurements correlated well with traditional measurements of renal function derived from imaging (⁹⁹Tc-mercaptoacetyltriglycine) and from serum markers, suggesting that this quantitative method could be used for measuring renal function preclinically and clinically. However, given the existence of other accurate tests, this application is less likely to be adopted clinically in the near future.

Several animal experimental studies have used quantitative approaches to describe alterations in bone blood flow and metabolism in a variety of diseases.

Murine Tumor Models

¹⁸F-NaF PET bone imaging protocols have been optimized in mice. The coefficient of variation from 4 consecutive measurements as an index of reproducibility of ¹⁸F-NaF bone uptake was less than 15% when bone counts per pixel per minute were normalized to a region of interest encompassing the whole skeleton (35). The excellent reproducibility supports the use of ¹⁸F-NaF small-animal PET to study a variety of bone diseases.

Small-animal PET or PET/CT has been used to image relevant models of metastatic prostate cancer and to monitor natural disease progression. For instance, the formation of osteolytic, osteoblastic, and mixed bone metastases from prostate cancer was studied with small-animal ¹⁸F-NaF and ¹⁸F-FDG PET/CT scans, radiographs, and histology or histomorphometry (36). ¹⁸F-NaF and ¹⁸F-FDG PET/CT accurately identified osteoblastic lesions and their progression (Fig. 3). Mixed blastic–lytic lesions were well visualized only with ¹⁸F-FDG, whereas ¹⁸F-NaF uptake was low or absent. In pure lytic lesions, ¹⁸F-FDG uptake increased over time. In contrast, ¹⁸F-NaF uptake was initially high but decreased to background levels after 6 wk, suggesting early increases in tumor vascularization resulting in flow-dependent accumulation of the probe.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Radiographs (left), ¹⁸F-NaF PET/CT scans (middle), and photomicrographs of histologic specimen (right). PET/CT images reveal osteoblastic lesion earlier (4 wk) than radiography (arrows denote bone lesions). Increasing ¹⁸F-NaF uptake over time corresponds to increased bone formation seen on histology (asterisks). (Reprinted from (36).)

Another group injected human Ewing sarcoma cells subcutaneously or intravenously into immune-deficient mice (37). Small-animal PET with ¹⁸F-FDG and ¹⁸F-NaF was then used to search for soft-tissue, lung, and bone metastases. Only intravenous tumor cell injection reproduced the human metastatic pattern (lung, bone, and soft tissue). In contrast to the study by Hsu et al. (36), Ewing sarcoma bone lesions exhibited decreased ¹⁸F-NaF uptake relative to normal bone. This difference is most likely explained by the use of a different tumor model and the different timing of PET scans after tumor cell injection.

Thus, relevant animal models of primary and metastatic bone lesions that mimic the biologic behavior of these tumors in humans were generated. These models can be used to evaluate novel bone imaging agents or to monitor the effects of therapeutic interventions on the course of primary or secondary bone malignancies.

QUANTITATIVE HUMAN STUDIES

The long-term precision and reproducibility of quantitative bone metabolism measurements with ¹⁸F-NaF PET was established by Frost et al. (38) in patients with osteopenia or osteoporosis. All underwent PET scans of the lumbar spine at baseline and after 6 mo. Measurement precision and reproducibility were compared with biomarkers of bone formation. The coefficient of variation was comparable among PET-based measurements and nonimaging biomarkers. In general, the precision of PET-based measurements ranged from 10% to 15%.

Hawkins et al. (18) measured an average influx rate of 0.036 ± 0.006 mL/min/mL in healthy bones (18). Schiepers et al. (28) reported a blood flow to lumbar vertebrae of 0.058 mL/min/mL and influx of 0.022 mL/min/mL for osteoporosis. For Paget disease, the flow rate was 0.205 mL/min/mL (or 4 times higher than flow to normal bone) and the influx was 0.114 mL/min/mL. Using the same model (Fig. 2), Cook et al. (27) also found an almost 4-fold increase in Paget-affected bones.

Messa et al. (39) measured bone flow and fluoride influx in patients with renal osteodystrophy and primary hyperparathyroidism and found significantly increased influx rates that decreased by 30%–40% after parathyroidectomy and medical therapy. Excellent correlations between influx rates and serum alkaline phosphatase and parathyroid hormone levels, as well as histomorphometric indices of bone formation rate, were reported.

The blood flow to the femoral head was estimated by Schiepers et al. (40) in 5 patients with unilateral hip trauma. A flow ratio of greater than 2 between the abnormal and normal femoral head was necessary for a successful outcome with conservative treatment. A minimum flow of 0.04 mL/min/mL was measured in 1 patient whose affected femoral head healed conservatively.

Using the model shown in Figure 2 (18,28), Frost et al. studied bone flow and turnover in women with treated and untreated metabolic bone disease (14). Bone perfusion was highly variable between the lumbar spine and the humerus, an observation that was not attainable by biomarkers of global bone turnover or by biopsy.