Quantitative Magnetic Resonance Fluorine Imaging: Today and tomorrow

Abstract

Fluorine (¹⁹F) is a promising moiety for quantitative magnetic resonance imaging (MRI). It possesses comparable MR sensitivity to proton (¹H) but exhibits no tissue background signal, allowing specific and selective assessment of the administrated ¹⁹F-containing compounds in vivo. Additionally, the MR spectra of ¹⁹F-containing compounds exhibited a wide range of chemical shifts (> 200 ppm). Therefore, both MR parameters (e.g. spin-lattice relaxation rate R₁) and the absolute quantity of molecule can be determined with ¹⁹F MRI for unbiased assessment of tissue physiology and pathology. This article reviews quantitative ¹⁹F MRI applications for mapping tumor oxygenation, assessing molecular expression in vascular diseases, and tracking labeled stem cells.

Magnetic resonance imaging (MRI) is a powerful medical imaging modality with excellent sof-tissue contrast. MRI also provides a number of quantitative measures for unbiased assessment of tissue physiology and pathology. Quantitative MRI typically utilizes two approaches: 1) quantification of the intrinsic MR parameters (e.g., spin-lattice relaxation rate R₁) of imaging moieties through fitting detected MR signal to a mathematical model; 2) determination of the absolute quantity of imaging moieties based on the registered MR signal intensity. While the first category of quantitative MRI can be achieved with any nucleus exhibiting a magnetic resonance effect, the second category is primarily restricted to those nuclei with no background tissue signal, such as fluorine (¹⁹F).

The feasibility of ¹⁹F MRI was first demonstrated by Holland and colleagues [6] four years after the development of ¹H MRI [7]. ¹⁹F has 100% natural abundance, a spin of ½, and a gyromagnetic ratio of 40.08 MHz/T (slightly lower than the 42.58 MHz/T of ¹H), resulting in 83% of the sensitivity of ¹H [8, 9]. Additionally, the chemical shift of ¹⁹F is sensitive to the molecular environment of its nucleus because of the seven outer-shell electrons of ¹⁹F atom (as compared to only one electron of ¹H). The ¹⁹F spectroscopic signature manifests a range of > 200 ppm [10, 11] (Fig. 1A), which permits unambiguous identification of distinctive ¹⁹F-containing compounds with ¹⁹F MRI. Finally, in contrast to the prominent ¹H signal from mobile water in biological tissue, only trace amounts of ¹⁹F (< 10⁻⁶ M) are present in tissue, i.e., these being immobilized in solid phase in the teeth and bones [12]. For practical purposes, no background ¹⁹F MR signal exists in vivo. Thus, ¹⁹F is a preferred probe for quantitative MR applications because of its high sensitivity, unique spectroscopic signature, and no tissue background. This review will focus on studies involving ¹⁹F imaging. The role of ¹⁹F in quantitative MR spectroscopy is beyond the scope of this review and comprehensive reviews that summarize current knowledge on this topic have been published [13, 14].

Figure 1.

(A) Representative ¹⁹F spectrum of PFPE and PFOB nanoparticles shows the chemical shift of ¹⁹F signatures. The single PFPE peak and five discernible PFOB peaks are easily detected and individually resolved. (B) Chemical structure of PFPE shows its twenty ¹⁹F atoms. (C) Schematic of a PFPE nanoparticle functionalized with homing ligands in the outer phospholipid monolayer (shown in green). The PFPE nanoparticle provides ¹H MR contrast by its surface payload of ~90,000 Gd³⁺ (shown in gold) and ¹⁹F MR contrast by ~100M ¹⁹F in its core. (From Morawski et al. [1] and Southworth et al. [2].)

PERFLUOROCARBON AS A USEFUL COMPOUND FOR QUANTITATIVE ¹⁹F MRI

Perfluorocarbons (PFCs) are a group of ¹⁹F-containing compounds derived from hydrocarbons by complete substitution of ¹H with ¹⁹F [15] (Fig. 1B). PFCs are non toxic and biologically stable. After in vivo administration, PFCs are not metabolized by the tissue but cleared by circulation and then in large part vaporized to the air through respiration. Because of their high payload of ¹⁹F atoms, PFCs are the most frequently used compounds for ¹⁹F MRI [16, 17].

For biological applications, PFCs are typically emulsified into a nanoparticle form (nominal size ranging from < 100 nm to several hundreds of nm [18]) to overcome their hydrophobic and lipophobic constraints on preparation (Fig. 1C). A typical PFC nanoparticle emulsion formulated in our lab contains 40% (v/v) perfluorooctylbromide (PFOB), 2% (w/v) safflower oil, 2% (w/v) of surfactant commixture, 1.7% (w/v) glycerin, and water balance [19]. Each formulated PFC nanoparticle comprises a liquid PFC core encapsulated by a lipid monolayer, resulting in a high concentration of ¹⁹F atoms (~ 100 M) for ¹⁹F MRI at both clinical (e.g., 1.5T) and research (e.g., 9.4T) field strengths [20, 21]. Intravenously administrated PFC nanoparticles do not leak out of intact vasculature because of the relatively large particle size. Instead, PFC nanoparticles are removed from the blood stream primarily by the reticuloendothelial system and macrophage endocytosis that eventually expelled out of the lung through gaseous exchange. Blood half-life of PFC nanoparticles varies with particle size [18] with a typical value of 2–12 hours [22, 23].

Plain PFC nanoparticles can be functionalized with various agents for molecular imaging of vascular disease [24]. Multiple copies of binding ligands can be covalently or noncovalently linked to the particle surface for targeted binding to molecular epitopes in diseased regions, such as cancer angiogenesis, atherosclerotic plaques, and renal inflammation [25]. The particle surface also can be complexed with other imaging contrast agents (e.g., fluorescent lipids) for multi-modality imaging [26]. Finally, the lipid membrane of PFC nanoparticles can dissolve lipophilic drugs (e.g., fumagillin) for targeted drug delivery [27]. Overall, functionalized PFC nanoparticles enabled a wide variety ¹⁹F MRI applications.

APPLICATIONS

Assessing regional tumor hypoxia by ¹⁹F MRI pO₂ mapping

Tumor hypoxia is a major therapeutic target in oncology [28–31]. Regional hypoxia is common in solid tumor because of the poorly organized tumor vasculature and the high oxygen demand of proliferative tumor cells. Hypoxic tumor cells are more resistant to radiotherapy and chemotherapy than well-oxygenated tumor cells [32]. Thus, non-invasive assessment of tumor hypoxia is critical for advancing tumor diagnosis and therapy [33, 34].

Quantitative ¹⁹F MRI of PFCs is the most extensively explored MRI method for in vivo mapping tumor oxygenation [15]. Although ¹H MRI methods, such as blood oxygen level dependent (BOLD) MRI that utilizes T₂* difference between paramagnetic deoxyhemoglobin and diamagnetic oxyhemoglobin [35], can achieve high spatial and temporal resolution to map tissue oxygenation without the need of exogenous agents, the detected change in blood oxygenation does not necessarily reflect tissue hypoxia. In contrast, PFCs can carry a high payload of O₂ and possess a fast gas exchange rate with surrounding tissue through free diffusion [15]. Because of the paramagnetic effect of O₂ [36], the partial pressure of dissolved O₂ (pO₂) in PFCs is linearly correlated with the ¹⁹F₁ R of PFCs at a given temperature. Thus, MR determined ¹⁹F₁ R of PFCs provides a non-invasive measure of tissue oxygenation when temperature is tightly controlled [37]. For certain PFCs exhibiting minimal sensitivity to physiological temperature variations within [30 – 42°C], such as hexafluorobenzene (HFB) and perfluoropolyether (PFPE), tissue oxygenation can be derived directly using the quantified ¹⁹F R₁ and a priori calibrated R₁ - pO₂ curve [38]. The precision of the ¹⁹F MRI method can reach 1–3 mmHg in hypoxic region [39]. It has been shown that tumor pO₂ measured by ¹⁹F MRI is comparable to that measured with fine electrode [40], fiberoptics [41], and near-infrared spectroscopy [42, 43].

Both PFCs and PFC nanoparticles have been used for assessing tumor oxygenation with ¹⁹F MRI. PFC nanoparticles administrated systematically can be sequestered in tumors, which is a method entailing minimal tissue damage [44, 45]. However, the major limitation of this method is that blood delivered PFC nanoparticles are primarily concentrated in well-vascularized tumor area instead of the poorly perfused hypoxic area, resulting in erroneous overestimation of tumor pO₂ [46]. To avoid this problem, PFCs have been directly injected into different tumor regions to achieve a comprehensive spatial measure of pO₂ [23, 39, 47]. Alternatively, PFCs can be enclosed in gas-permeable alginate capsules and implanted together with tumor cells to produce a stable pO₂ readout [48] that can persist for up to 2 years [49].

Fast ¹⁹F MRI techniques, such as fluorocarbon relaxometry using echo planar imaging for dynamic oxygen mapping (FREDOM), have been developed for dynamic mapping of tumor pO₂ [50, 51]. Using these techniques, heterogeneous distributions of pO₂ was observed both within tumors and between large and small tumors [23]. The measured tumor pO₂ exhibited a strong correlation with tumor size [41, 52, 53] except in a few reports [54]. Interestingly, oxygen inhalation only transiently elevated pO₂ in some but not all types of tumors [4, 39, 43](Fig. 2). It was demonstrated that successful elevation of tumor pO₂ during radiation therapy correlates with delayed tumor development [4], thus validating the usefulness of tumor pO₂ mapping for prediction of therapeutic response.

Figure 2.

FREDOM determined pO₂ maps of two representative AT1 tumors in rats. The pO₂ value was calculated pixel-by-pixel based on the quantified ¹⁹F R₁ and a priori calibrated ¹⁹F R₁ - pO₂ curve of HFB. (A and D) Composite ¹⁹F (displayed in color) and ¹H (displayed in grayscale) MR images show HFB distribution in a large tumor (A, 3.6 cm³) and a small tumor (D, 1 cm³). (B and E) Baseline pO₂ maps show higher pO₂ in the small tumor when both animals were breathing air. Mean pO₂ of large and small tumors were 0.1 ± 1.8 torr and 25.4 ± 1.1 torr, respectively. (C and F) Tumor pO₂ maps of same animals obtained at 24 minutes after oxygen breathing, mean pO₂ of large and small tumors were 8.1 ± 4.5 torr and 90.6 ± 3.9 torr, respectively. Both values were significantly higher than that of baseline (p < 0.01). (From Bourke et al. [4].)

In summary, ¹⁹F MRI of PFC facilitated pO₂ mapping has been demonstrated in animal models at clinically achievable field strength (< 7T). Future translation of this technique to human will improve tumor prognosis and help predict the therapeutic response in patients.

Molecular ¹⁹F MRI with site-targeted PFC nanoparticles

The unique capability of ¹⁹F MRI to directly determine the absolute quantity of ¹⁹F atoms has been largely unexploited until very recently, primarily due to lack of a method for non-invasively delivering PFC nanoparticles to region of interest. Fortunately, functionalized PFC nanoparticles that are specifically targeted to vascular diseases and cancers have been employed experimentally for over a decade [24]. In the first stage, these particles carried paramagnetic agents for ¹H MRI, and the binding of particles to region of interest was determined by the local proton signal enhancement relative to the native tissue contrast [26]. Thus, it is reasonable to expect that targeted delivery of functionalized PFC nanoparticles will also result in regional accumulation of ¹⁹F atoms above minimally required concentration for ¹⁹F MRI.

Neubauer et al. first reported a quantitative ¹⁹F MRI assessment of functionalized PFC nanoparticles [1]. Using fibrin-targeted PFC nanoparticles, the results showed that the bound particles on the surface of fibrin clots provided enough ¹⁹F atoms for ¹⁹F MRI at 4.7T field strength. Additionally, a linear relationship between the quantity of bound ¹⁹F atoms and the measured ¹⁹F MR signal was confirmed for functionalized PFC nanoparticles (Fig. 3A). By applying an external reference standard of ¹⁹F-containing compound that possesses different spectroscopic signature (Fig. 3B), the bound PFC nanoparticles on fibrin clot surface was quantitatively mapped with ¹⁹F MRI using the following equation:

$$\#\text{of}\text{bound}{}^{19}\mathrm{F}=\#\text{of}{}^{19}\mathrm{F}\text{in}\text{standard}\times (\text{signal}\text{of}\text{bound}{}^{19}\mathrm{F}/\text{signal}\text{of}{}^{19}\mathrm{F}\text{standard})\times k$$

where k is a constant calibrated a priori, representing the ratio between ¹⁹F MR signal of PFC nanoparticles and ¹⁹F MR signal of reference standard containing equal number of ¹⁹F atoms (Fig. 3C).

Figure 3.

(A) A representative ¹⁹F spectrum of PFPE nanoparticle emulsion (−90 ppm) and trichlorofluormethane reference standard (0 ppm) acquired at 4.7 T. (B) The calibration curve for PFPE nanoparticle emulsion shows a linear relationship between the quantity of PFC nanoparticles and ¹⁹F signal intensity. (C) Left: An optical image of a human carotid endarterectomy sample shows moderate luminal narrowing and several atherosclerotic lesions. Middle: A ¹⁹F projection image acquired through the entire thickness of carotid artery sample shows high ¹⁹F signal along the lumen because of the binding of nanoparticles to fibrin. Right: The calculated concentration map of bound nanoparticles in the carotid sample based on ¹⁹F signal intensity in each voxel and the calibrated standard curve in (B). (From Morawski et al. [1].)

The later work of Caruthers et al. from the same group showed that the binding of functionalized PFOB and PFPE nanoparticles to biological specimen could be simultaneously or selectively assessed with ¹⁹F MRI at 1.5T field strength, validating the feasibility of this ¹⁹F MRI method for phenotypic characterization of pathological biosignatures in clinical settings [20]. Neubauer et al. further showed that incorporation of paramagnetic gadolinium (Gd³⁺) on PFC nanoparticles increased ¹⁹F₁ R by four fold, resulting in up to 125% higher ¹⁹F signal intensity at 1.5T field strength [55]. However, direct mixing Gd³⁺ (up to 250 mM) in PFC nanoparticle emulsion has no detectable effect on ¹⁹F R₁, suggesting Gd³⁺ must be in very close proximity to the PFC core of nanoparticles to affect ¹⁹F R₁.

Recently, the potential for using quantitative ¹⁹F MRI of functionalized PFC nanoparticles to delineate vascular diseases was assessed in ex vivo studies. In a rabbit model of atherosclerotic valve angiogenesis, Waters et al. reported that valve leaflets of rabbits treated with α_vβ₃-integrin targeted PFC nanoparticles exhibited ~3 times higher ¹⁹F signal than valve leaflets treated with untargeted nanoparticles, suggesting the specific binding α_vβ₃-integrin targeted PFC nanoparticles to valve angiogenesis [56]. Southworth et al. produced further evidence for molecular ¹⁹F MRI of functionalized PFC nanoparticles in the kidneys of ApoE-knockout mice, an animal model of renal inflammation. In this case, ApoE-knockout mice treated with vascular cell adhesion molecular 1 (VCAM-1) targeted PFC nanoparticles exhibited substantial higher kidney ¹⁹F signal than ApoE-knockout mice treated with non-targeted PFC nanoparticles and control C57BL/6 mice treated with VCAM-1 targeted or non-targeted PFC nanoparticles, reflecting up regulated VCAM-1 expression in the kidneys of ApoE-knockout mice [2].

However, in vivo quantitative ¹⁹F MRI of site-targeted PFC nanoparticles is not straightforward because of the background signal from unbound circulating nanoparticles in the blood pool. A potential solution of this problem that utilized diffusion weighted ¹⁹F MRI technique has been recently reported. This technique specifically detects ¹⁹F signal from bound nanoparticles by applying a pair of diffusion sensitizing gradients to selectively suppress ¹⁹F signal from unbound (flowing) PFC nanoparticles [5] (Fig. 4).

Figure 4.

Diffusion weighted ¹⁹F signal in the ear of K14-HPV16 mice (open symbols), an animal model of squamous cell cancer with dysplastic lesions developed in the ear epidermis, and in the ear of control C57BL/6 mice (filled symbols). All animals were intravenously injected with of α_vβ₃-integrin targeted PFC nanoparticles before MRI. (A) Results acquired with modest b-values (i.e., an index of diffusion weighting) shows complete decay of ¹⁹F signal in control mouse ears when b-value > 1500 s/mm². In contrast, a large fraction of ¹⁹F signal persisted in the ears of K14-HPV16 mice at all b-values, reflecting the specific binding of targeted nanoparticles to the ear neovasculature of K14-HPV16 mice. (B) Diffusion weighted ¹⁹F signal in the ears of K14-HPV16 mice persisted even when b-values > 10,000 s/mm². (From Waters et al. [5].)

In summary, molecular ¹⁹F MRI of functionalized PFC nanoparticles has been demonstrated at both research (e.g., 11.7T) and clinical (e.g., 1.5T) field strengths. The site-targeted PFC nanoparticles are currently used in preclinical research. Clinical trials are expected within a few years.

In vivo stem cell trafficking using ¹⁹F MRI of intracellularly labeled PFC nanoparticles

Regenerative therapy using stem cells offers great promises for treating many types of diseases. The development of quantitative imaging methods to localize therapeutic stem cells in vivo is critical for noninvasive assessment of stem cell therapy [57]. Because of the excellent soft tissue contrast and readily available imaging contrast agents, ¹H MRI has served as a primary method for stem cell trafficking [58, 59]. Typically, in vitro cultured stem cells are incubated with ¹H MRI contrast agents, such as super-paramagnetic iron oxide or paramagnetic Gd-DTPA [60, 61], which results in uptake by endocytosis entailing the cells themselves to serve as imaging agents. Internalization of these contrast agents by stem cells generally requires transfection techniques or other adjunctive mechanical methods such as electroporation [62–64]. After in vivo administration, the labeled stem cells can be detected by ¹H MRI based on the “negative” (i.e. dark) or “positive” (i.e. bright) contrast effects of the labeling agents. These ¹H MRI techniques can detect as few as a single cell under certain circumstances [65].

It was recently shown that ¹⁹F MRI can be used as an alternative method for quantitative trafficking of stem cells in vivo [66]. The seminal work of Ahrens et al. demonstrated that PFPE nanoparticles could be effectively internalized by dendritic stem cells with the help of a cationic transfection agent [67]. The intracellular labeling of PFPE nanoparticles was proved to be a biologically safe and stable. After local injection or systematic delivery, the biodistribution of labeled cells could be specifically detected by ¹⁹F MRI at 11.7T. Subsequent work by our lab showed that PFC nanoparticles could be readily internalized by stem/progenitor cells without the need of transfection agent [3]. Using this technique, stem/progenitor cells can be labeled with multiple types of PFC (i.e., PFOB and PFPE) nanoparticles that can be rapidly detected by ¹⁹F MRI at both 1.5T and 11.7T field strengths (Fig. 5). In a recent study, Ruiz-Cabello et al. showed that PFPE nanoparticles with cationic surface charge could also be effectively internalized by neural stem cells without the need of transfection agents [21]. After in vivo administration into the mouse brain, the labeled cells remained viable and exhibited a constant19F signal that can be detected at 9.4T for up to one week, reflecting the stability and non-toxicity of this labeling technique.

Figure 5.

Localization of PFC nanoparticles labeled cells in mice using ¹⁹F MRI. (A) ¹⁹F MRI trafficking of stem/progenitor cells labeled with either PFOB (green) or PFPE (red) nanoparticles. Labeled cells were locally injected into the skeletal muscle of mouse thigh before MRI. (B–D) At 11.7T field strength, ¹⁹F spectral discrimination permits respective imaging of ~1×10⁶ PFOB-loaded cells (B) and PFPE-loaded cells (C). The composite ¹⁹F (displayed in color) and ¹H (displayed in grayscale) image (D) reveals the location of PFOB labeled cells in the left leg and PFPE labeled cells in the right leg (dashed line indicates 3×3 cm² field of view for ¹⁹F images). (E) Similarly, a ¹⁹F image acquired at 1.5T field strength shows ¹⁹F signal from ~4×10⁶ PFPE nanoparticles labeled cells. (F) The composite ¹⁹F and ¹H image shows the location of PFPE nanoparticles labeled cells in a mouse thigh. Overall, the absence of background signal in ¹⁹F images (B, C, and E) enables unambiguous localization of PFC-containing cells at both 1.5 T and 11.7 T field strength. (From Partlow et al. [3].)

In vivo ¹⁹F MRI of the migration of PFPE nanoparticles labeled T cells has been reported by Ahrens and colleagues. The work of Srinivas et al. and Janjic et al. respectively demonstrated the homing of intraperitoneally injected T cells to the pancreas of diabetic mice and to the lymph nodes of BALB/c mice at 48 hours after cell injection [68, 69]. In a recent study by Srinivas et al., longitudinal¹⁹F MRI over 21 days delineated a dynamic accumulation of ovalbumin-specific T cells in the lymph node proximal to, but not distal to, the inflammatory region induced by focal chicken ovalbumin inoculation [70].

Despite the difference in the stem cell lines and labeling techniques, previous reports have delineated several common approaches for stem cell labeling and trafficking with ⁹F MRI: (1) intracellular labeling was confirmed by high resolution microscopy showing PFC nanoparticles within the cell cytosol; (2) the lack of cellular toxicity of internalized PFC nanoparticles was validated by the comparable viability, proliferation, and function between labeled cells and control cells; (3) the stability of cell labeling was reflected by the detectable ¹⁹F MR signal for up to 21 days post labeling; (4) after in vivo administration, labeled cells was specifically detected by ¹⁹F MRI based on their positive signal with no background; (5) the biodistribution of cells was visualized by overlaying ¹⁹F image of labeled cells on the ¹H anatomical image acquired at the same location; (6) the local concentration of labeled cells was determined by quantitative ¹⁹F MRI. The primary limitation of this cell trafficking technique is its relative low sensitivity, i.e., a minimum of 2000 labeled cells are needed for detection with ¹⁹F MRI [3].

In summary, in vivo stem cell trafficking using ¹⁹F MRI of intracellularly labeled PFC nanoparticles is still restricted to preclinical research. Because of the low ¹⁹F signal intensity from individual labeled cell, MR experiments were mostly performed at high field strength (i.e., 9.4T and 11.7T) to improve the sensitivity of ¹⁹F MRI to labeled cells.

Other quantitative ¹⁹F MRI applications

Recently, ¹⁹F MRI of PFC nanoparticles has been used to detect cardiac and cerebral ischemia [71]. It was shown that intravenously administrated PFC nanoparticles were actively internalized by circulating monocytes/macrophages. After both acute cardiac and cerebral ischemia, the progressive accumulation of these immunocompetent cells in corresponding inflammatory areas was delineated on the composite ¹⁹F/¹H image by overlaying the detected positive ¹⁹F signal on anatomical ¹H image. The fast imaging time (~ 20 minutes) and high spatial resolution (~ 0.5 × 0.5 × 2 mm³) of ¹⁹F MRI suggested that this method could be applied to assess inflammatory diseases in general. However, the contracting viable myocardium may suffer motion induced ¹⁹F signal decay relative to the ischemic non-contracting myocardium and thus complicating MRI data interpretation.

Compounds other than PFCs have also been used for quantitative ¹⁹F MRI. For example, Higuchi et al. showed that ¹⁹F MRI using a ¹⁹F-containing amyloidophilic probe is more reliable than T₁ and T₂ weighted ¹H MRI to detect amyloid β plaques in a mouse model of Alzheimer’s disease [72]. Additionally, Porcari showed ¹⁹F MRI delineated the biodistribution and pharmacokinetics of ¹⁹F-labeled boronophenylalanine, a carrier for Boron neutron capture therapy (BNCT), in rat glioma tumor [73, 74]. The results suggested that the ability to dynamically assess the concentration of boronophenylalanine in tumor may improve the outcome of BNCT by providing an estimation of optimal time for neutron irradiation [75]. Finally, fluorinated gases have been used in some studies to image and measure gas diffusion in lungs [76–78].

Conclusion

¹⁹F MRI is a promising quantitative medical imaging method for assessing the structure, function, and molecular display of various diseases. The ability to specifically detect administrated ¹⁹F-containing compounds without tissue background is a unique advantage over ¹H MRI. Although ¹⁹F MRI was usually performed at relatively low spatial (millimeters) and temporal (minutes to hours) resolution because of low concentration of ¹⁹F atoms in region of interest, the improved MR sensitivity at higher field-strength (e.g., 7T) may overcome this limitation for translational application. Additionally, the development of functionalized PFC nanoparticles that can specifically accumulate in region of interest will also enhance the sensitivity ¹⁹F MRI. The major limitations of ¹⁹F MRI are the difficulties of approval for clinical use of ¹⁹F-containing agents and the requirement of additional hardware for registering the ¹⁹F nucleus on existing clinical ¹H MR scanners.

In summary, quantitative ¹⁹F MRI has been applied to assess tumor pO₂ for disease prognosis and optimization of therapy, to image vascular diseases with site-targeted PFC nanoparticles, and to monitor stem cell migration and proliferation. Because of the long blood half-life of PFC nanoparticles, novel imaging methods that can effectively suppress ¹⁹F signal from circulating PFC nanoparticles need to be developed to enable specific and dynamic assessment of PFC accumulation in region of interest. The continuous improvement in MR hardware and software will eventually permit widespread molecular imaging of ¹⁹F contrast agents on all imaging platforms.