Recent Advances in Higher-order Multimodal Biomedical Imaging Agents

Abstract

Advances in biomedical imaging have spurred the development of integrated multimodal scanners, usually capable of two simultaneous imaging modes. The long-term vision of higher-order multimodality is to improve diagnostics or guidance through analysis of complementary, data-rich, co-registered images. Synergies achieved through combined modalities could enable researchers to better track diverse physiological and structural events, analyze biodistribution and treatment efficacy, and compare established and emerging modalities. Higher-order multimodal approaches stand to benefit from molecular imaging probes and in recent years, contrast agents that have hypermodal characteristics have increasingly been reported in preclinical studies. Given the chemical requirements for contrast agents representing various modalities to be integrated into a single entity, higher-order multimodal agents reported so far tend to be of nanoparticulate form. To date, the majority of reported nanoparticles have included components that are active for magnetic resonance. Herein, we review recent progress in higher-order multimodal imaging agents, which span a range of material and structural classes, that have demonstrated utility in three (or more) imaging modalities.

1. Introduction

Medical imaging is an interdisciplinary field that commands a significant proportion of all medical-related spending costs. Despite the indispensable benefits that clinical medical imaging modalities offer, X-ray computed tomography (CT), magnetic resonance imaging (MRI), ultrasound (US), positron emission tomography (PET), single-photon emission computed tomography (SPECT), and their derivatives, all have characteristic strengths and drawbacks.^{[1, 2]} A cursory overview of the characteristics of these clinical imaging modalities is presented in Figure 1. Whether it is high cost, low resolution, exposure to ionizing radiation, poor sensitivity to contrast media, or difficulty in data interpretation, emerging imaging paradigms and technologies, strive to address unmet needs.^[3] One direction this has taken is towards multimodality in medical imaging.

Figure 1.

Overview of common clinical imaging modalities which have potential for multimodal applications. MRI: magnetic resonance imaging, CT: computed tomography, PET: Positron emission tomography, SPECT: single photon emission tomography, US: ultrasound.

Multimodal imaging can be any combination of techniques resulting in a composite image from separate sources. In theory, by combining two or more methods, one can overcome limitations of each modality independently. Ideally this results in obtaining higher quality or more useful data than could be achieved otherwise. Higher-order multimodal imaging extends this concept to a greater number of imaging modalities and implies that interconnected modality synergies are exploited for the purpose of generating complementary and informative insights. In a prototypical example, Kircher et al. demonstrated this powerful paradigm by designing and testing an MRI/photoacoustic (PA)/surface enhanced Raman scattering (SERS) contrast agent to aid in brain tumor margin delineation and resection. Overcoming the sensitivity, specificity, and resolution limitations of each modality by itself, researchers were able to use a probe for pre- and intra-operative surgical planning, as well as to identify cancerous tumor margins (Figure 2).^[4] This synergistic imaging takes advantage of the strengths of each of the different imaging modes is at the core of what multimodality has to offer.

Figure 2.

Synergistic use of a triple-modality MR/Raman/PA probe (MPR). MPRs are injected intravenously into a mouse bearing an orthotopic brain tumor and accumulate in the tumor (top). The concept of proposed eventual clinical use (bottom). Detectability of MPRs by MR allows for preoperative detection and surgical planning. The probe can be detected in the tumor during surgery several days later. PA, with its relatively high resolution and deep tissue penetration, is then able to guide bulk tumor resection intraoperatively. Raman imaging, with its ultrahigh sensitivity and spatial resolution, can then be used to remove any residual microscopic tumor burden. The resected specimen can subsequently be examined using a Raman probe ex vivo to verify clear tumor margins. Adapted with permission from Kircher et al.^[4]

The rapid increase in academic research involving multimodal imaging is shown in Figure 3. The increased output in multimodal agents at the research stage is likely to be followed by development and commercialization of some of these recent advances. Several comprehensive reviews of multimodal contrast agents have been reported in recent years,^[5–9] which discuss the many classes of nanoparticles and their various therapeutic and diagnostic qualities. In this review, we describe recent examples of higher-order imaging probes, most of which are trimodal, but some of which go beyond.

Figure 3.

Growth of multimodal imaging research. The number of articles published per year containing the term “multimodal imaging” between 1995–2014 in the Google Scholar database.

2. Overview of Medical Imaging

Medical imaging techniques can be generally classified as either yielding structural or functional information. Structural imaging looks primarily at the size, shape, and morphology of objects in the body. This type of imaging can target bones, organs, and vessels and looks for physical or anatomical abnormalities in structure and can guide medical procedures. Functional imaging looks at the physiology of diverse and dynamic bioprocess at the cellular or whole-organ level and is often associated with molecular imaging. This imaging has been used in visualizing metabolism in different tissues, blood flow, detecting molecular targets, and diffusion/perfusion/biodistribution. Both structural and functional imaging modalities are important in everyday clinical applications and preclinical experiments.

2.1. Clinical modalities

CT is the result of emitting and detecting X-rays from multiple angles around an object in order to elucidate internal structures. Although it relies on ionizing radiation, CT is considered the gold standard for identifying structural abnormalities such as tumors, hemorrhages, bone trauma, vascular and heart disease, emphysema, and fibrosis, and more. It is often used in diagnosing conditions including those of the head, chest, abdomen, pelvis, heart, and upper gastrointestinal (GI) tract.^[10] While CT has had a transformative impact on medicine, due to the risks induced by the ionizing radiation on which it relies, there has been concern about the risk posed to children and by unnecessary scans.^[11]

MRI is more expensive than CT but does not involve ionizing radiation and is especially useful for visualizing soft tissues.^[12] Like CT, MRI also plays a large role in both clinical and preclinical imaging. Although its cost generally keeps it from screening applications, the benefits and potential of MRI diagnosis and treatment planning are accepted. There are many variations of MRI that can provide both structural and functional information

Ultrasound or ultrasonography (US) is one of the most commonly used types of medical imaging because it is affordable, widely accessible, transportable, real-time, and is free of ionizing radiation. Utilizing acoustic vibrations with a frequency greater than the human range of hearing, and translating the effects of tissue on these waves, forms the foundation of diagnostic sonography. It has many clinical and preclinical research applications in prenatal medicine and in diagnosis of diseases in most the major organs, with variations like Doppler echocardiography and elastography capable of discerning both structural and functional features. Drawbacks to US technology include large amounts of noise and a limited ability to penetrate deep into tissue, with resolution falling with depth. There has been work to develop endocavity transducers to image internal structures at higher frequencies and resolution.

PET and SPECT are primarily tracer imaging techniques, which require the introduction of exogenous agents. These two molecular imaging modalities rely on decay of radionuclides, usually incorporated into a bioactive molecule such as glucose. These analogues can be traced once administered in the body for minutes or hours and provide information about the activity of biological processes. The most commonly performed PET scans use this strategy to identify areas in the body exhibiting abnormally high levels of sugar metabolism, as an indication of tumor growth, for example.^[13] Although its oncologic applications are the most common, PET and SPECT are also used for imaging neural activity, myocardial function, and drug distribution.

2.2 Emerging and Preclinical Modalities

Although emerging imaging modalities, by definition, have yet to experience the same clinical success as those just described (CT, MRI, US, PET), technologies like photoacoustic imaging (PA), Raman imaging, and upconversion luminescence (UCL) imaging hold the potential for contributions in the field. There are many variations on these modalities and ways they can be implemented.

Optical and fluorescence microscopy have been used in medicine for decades as an ex vivo or in vitro tool, often for examining histological tissue samples or living cells. This is because visible light has poor tissue penetration and experiences significant scattering in biological media. Recently there has been an increase in the work done to create systems for optical and fluorescence imaging of the whole body that can be performed on a subject without any need for complicated biopsy and staining protocols.^[14] Optical techniques that can complement or replace conventional modalities are desirable because visible light radiation causes no harmful ionizing effects to the patient, and the cost of materials is cheaper compared to other highly sensitive techniques, like PET and SPECT. One benefit to optical imaging is its ability to be miniaturized to the level of a catheter or endoscope. Endoscopic imaging uses flexible tubes combined with light sources and a camera, which can be inserted into the body with minimal invasiveness. With white light as the illumination source, surgeons have been able to see inside many parts of the body without large surgical incisions and perform many essential procedures in this way. Recently, there have been significant advances in miniaturizing the hardware necessary for confocal and fluorescent imaging to the point that there are multiple commercially available confocal endomicroscope probes available. It has been shown, for example, that through using fluorescent antibodies as probes, not just the presence but also the variants of human bladder cancer can be identified.^[15]

Photoacoustic tomography (PAT) is an exciting emerging imaging modality. Based on the photoacoustic effect, this technique relies on the absorption of photons by endogenous biomolecules or exogenous contrast agents that then generate ultrasonic waves, which experience less scattering compared to light waves in tissue. With the ability to image a wide scale of sizes from organelles to organs and no use of ionizing radiation, it is not surprising this modality is trying to establish a clinical niche in breast cancer imaging, mapping sentinel lymph nodes, and brain imaging.^[16] It has been used for the functional imaging of blood flow as well as yielding structural information from the cellular level up through whole organ imaging. Because of its strength and versatility, there has been work in miniaturization of PA systems for endoscopic integration.

Raman spectroscopy is an optical technique that can be used to identify chemicals in a manner similar to infrared spectroscopy. Using physical and chemical properties on the molecular scale, Raman spectroscopy is especially powerful in distinguishing subtle differences in tissue composition or the conformational structure of pharmaceuticals and nanoparticles. In the design of biological probes, it is often desirable to track multiple targets simultaneously. To this effect, nanoparticles designed with surface enhanced Raman scattering (SERS) functionality have been synthesized with differences in their chemical structure (vibrational modes) that allows for highly multiplexed interrogation of biological systems.^[17]

3. Higher-order Multimodal Imaging

Developments in imaging acquisition and reconstruction in each independent modality have enabled the field of multimodal imaging to expand. This originally came about by using two modalities in tandem, one right after the other, but has evolved into commercial production of integrated clinical PET/CT and PET/MRI machines, and preclinical small animal imaging machines capable of many combinations of modalities. The IVIS Lumina XRMS is a single scanner manufactured by PerkinElmer capable of X-ray, FL, and Cerenkov imaging. The development of these integrated scanners has stimulated a further increase of multimodal contrast agents with diverse applications.

The first clinical implementation of multimodal imaging was combined PET-CT.^[18] The strength of this combination comes from being able obtain a greatly increased amount of information regarding structure (CT) as well as physiological function (PET) that provides information that would be unattainable with standalone modalities. This technology has virtually replaced standalone PET scanners and gained clinically acceptance at an unprecedented rate so that the clinical benefit of PET-CT over PET alone has been established.^[19] Although combined PET-MRI scanners began development around the same time as PET-CT, cost and technical challenges have delayed the widespread embracement of those machines.^[6] Figure 4 demonstrates the similarity between images acquired with PET-CT and PET-MRI.

Figure 4.

Similarity of PET/CT and PET/MRI scans, acquired on the same day in the same patient with metastasized thyroid carcinoma. (A) ¹⁸F-FDG PET acquired on PET/CT scanner. (B) Overlay of PET (in orange) and structural low-dose CT data acquired on a PET/CT scanner. (C) Overlay of PET (in orange) and MRI acquired on PET/MR scanner. (D) ¹⁸F-FDG PET acquired on PET/MR scanner. An overall similarity can be seen in pattern of suspicious lesions in A and D. Adapted with permission from Drzezga et al.^[78]

The main challenges with clinical implementation of multimodal imaging are that, with the high demands on effective communication of all parties involved, unique protocols have to be developed in each clinical setting. Making sure that all scans are performed correctly, the digital formats are appropriate, and that qualified people perform the analysis, is a highly demanding task. Then there are also technological drawbacks that are difficult to overcome without integrated scanners, such as issues from patients breathing or registering kinetics of a disease or treatment. Combined PET-CT scanners have taken tremendous steps in overcoming these challenges, but are still in active stages of improvement.

While PET-CT and PET-MRI have already been developed and implemented clinically, there are many other combined imaging techniques being explored for synergistic benefits. It is most common to see a combination of one of the more established anatomical modalities such as CT or MRI used in conjunction with more novel functional techniques like PA or UCL. Many bimodal imaging probes have been developed, encompassing virtually any combination of modalities, often with additional targeting or therapeutic applications. Although there are preclinical scanners available from different manufacturers that can combine CT or PET with FL for in vivo imaging, often multimodal imaging that incorporates emerging techniques requires scanner modification, imaging in sequence, or building new instrumentation from scratch.

4. Development of Higher-order Multimodal Contrast Agents

The development of nanoparticle (NP) based contrast agents capable of multiple modalities has increased tremendously in the past decade. Administration of a single higher-order multimodal contrast agent presents some pragmatic advantages over multiple separate administrations of different unimodal agents. Having the imaging contrast originating from a single agent also is advantageous. However, unimodal agents for various imaging modalities are already optimized and clinically established, so that there is a burden for new multimodal agents to show synergistic superiority over separate unimodal agents. But chemical design and synthesis of NPs capable of multimodality is not simple. Additionally, there could be the possibility of interference in function or storage stability between the various contrast components of integrated multimodal agents.

It may appear at first glance that a trimodal imaging probe that can simultaneously generate contrast in MRI, PET, and FL might be redundant. Although PET and FL probes both have a high sensitivity and can be used to image functional characteristics, there still is reason to combine them into a single integrated probe. Besides differences in the modalities themselves, an important benefit is being able to directly compare imaging metrics like resolution, tissue penetration, and sensitivity without introducing additional confounding factors into the system. When the anatomical resolving power of MRI is incorporated to further analyze these particles in vivo and vitro, this synergy allows for rigorous contrast agent evaluation. Many materials are compatible with higher-order imaging multimodality by incorporating additional functional moieties. For example, some platforms that are suited for multimodal imaging include upconverion NPs^{[20, 21]}, quantum dots^[22], iron oxide NPs^[23], silica NPs^[24], liposomes^[25] and lipoproteins^[26]. The following examples of multimodal agents are categorized by combined imaging modalities and serve to highlight some representative recent examples.

4.1 CT / MRI / FL

The combination of CT, MRI, and FL is one of the most explored trimodal schemes in multimodal imaging. Both CT and MRI are essential in clinical diagnostics, and FL offers a highly sensitive element capable of detecting and reporting phenomena that may not be possible with the other two. Although MRI and CT can provide similar structural information in many cases, the subtle differences between implementation, cost, safety, and target tissue to be imaged could determine which imaging technique is most appropriate. FL could then be used in a variety of ways to validate, correlate, or add new information to the observations of the other modalities at low cost and possibly on a cellular or molecular level.

Many materials have been used to impart CT, MRI, and FL imaging capabilities. Figure 5 shows some representative materials, structural schematics, and transmission electron microscopy (TEM) images of gold/silica core NPs,^[27] Fe₃O₄@dye-hybrid@Au NPs, ^{[28, 29]} and alpha-cyclodextrin (α-CD)-coated nanophosphors (DyNPs-Gd-Ir).^[30] Finding biocompatible materials that retain their respective magnetic or optical properties without compromising the functionality of the other components is a major challenge in designing multimodal imaging probes. For the gold/silica core particle, the dense core provided contrast for CT, a Gd-chelating lipid was used for MRI, a lipid-conjugated Cy5.5 provided contrast for FL, and polyethyleneglycol distearoylphosphatidylethanolamine (PEG-DSPE) was used to increase biostability (Figure 5A and 5C).^[27] This type of core-coating structure, with active lipid or polymer ligands to impart additional imaging functionality is a common design. Besides generating enhanced X-ray contrast by employing gold as the core nanoparticle, Au nanospheres have been used to decorate the surface of polymer based NPs (Figure 5B and 5D), iodinated oil has been loaded into silicon NPs,^[31] and NaDyF₄ nanorods were conjugated to polymeric NPs.^[30] Polymers like α-CD have been used to chelate Gd (Figure 5E, F, and G), which were shown to enhance both T1- and T2-weighted MRI.^[30] Iron oxide (IO) nanoparticles as the core and decorated onto the surface of other NPs have also contributed to MRI contrast in trimodal imaging probes.

Figure 5.

Exemplary NPs used for CT/MRI/FL. Schematic representations of trimodal (A) gold/silica core NP ^[27] and (B) Fe₃O₄@dye-hybrid@Au.^[28] TEM images of (C) gold/silica core NPs and (D) Fe₃O₄@dye-hybrid@Au.^[28] (E) Schematic, (F) TEM, and (G) molecular structure of alpha-CD and Ir-complex in α-CD-coated DyNPs-Gd-Ir.^[30] All figures used with permission from the indicated references.

Just like there are many components that can used to generate the trimodal signal in combined CT, MRI, and FL, there are many synthetic methods used to formulate these trimodal probes. The best methods are ones that minimize cost, risk, and time while having the highest yield, reproducibility and simplicity. A simple one-pot synthesis was described, where self-assembled amphiphilic polystyrene-b-poly(acrylic acid) with hydrophobic Fe₃O₄ nanoparticles were conjugated to Rhodamine B dye and gold nanoparticles were placed on the surface of the core.^[28] Since many of the trimodal probes currently being developed also have therapeutic or active targeting properties, the ability to load drugs or to be modified with additional targeting ligands can be useful. A trifunctional imaging probe with superparamagnetic iron oxide nanoparticles (SPIONs) as the core was designed for the additional loading of materials.^[31] After functionalization with silicon, Cy5 dye, and PEG for stability, the particles were loaded with iodinated oil to yield trimodal i-fmSiO4@SPIONs.

Phantom studies are ubiquitously performed to test the relationship between signal strength and probe concentration. This can be done in microwell plates or plastic tubes and is one of the more common ways of comparing the performance of new contrast agents to those already in use. Phantom analysis of Fe₃O₄@dye-hybrid@Au in MRI yielded an improved r₂ value more than twice that of commercial iron oxide (IO) particles such as Ferridex.^[28] With an Au concentration of 146 mg/mL, the Fe₃O₄@dye-hybrid@Au showed an equivalent contrast enhancement as 160 mg/mL of the commercial μCT agent eXIA 160. The results obtained by phantom studies sometimes can alternatively be validated or acquired while incubating samples in biologic media or in vitro. Phantom I-fmSiO4@SPION and samples in aqueous suspension were imaged as shown in in Figure 6A.^[31] Figure 6B shows that during incubation with HeLa cells, DyNP-Gd-Ir NPs were able to retain phosphorescence without any photobleaching and showed a direct relation of signal intensity with concentration of nanoparticles.^[30] Figure 6C–E shows the effect of gold/silicon core NP concentrations on MRI, FL and CT in phantoms.^[27] With the rise of endoscopic and real-time imaging aided surgery, in vitro and ex vivo techniques can also help determine the ability of imaging agents to be internalized by cells or label specific tissues. In this manner, PEGylated Fe₃O₄@dye-hybrid@Au NPs were confirmed to be internalized by normal cells by confocal microscopy, showing no toxicity.^[28]

Figure 6.

CT/MRI/FL/ imaging in phantoms. (A) T2 weighted CT/MRI/FL/ imaging (top to bottom) of phantom I-fmSiO4@SPION samples in aqueous suspension.^[31] (B) R₂ as a function of concentrations of Dy from DyNPs-Gd-Ir. Inset: T2-weighted MR images of DyNPsGd-Ir.^[30] Concentration dependence of gold/silica core NPs for signal generation in (C) MRI, (D) FL, and (E) CT.^[27] All figures used with permission from the indicated references.

PEGylated Fe₃O₄@dye-hybrid@Au NPs were also evaluated in mice with diethlynitrosamine-induced hepatocellular carcinoma. Following intravenous injection, these NPs helped to differentiate liver lesions from normal tissue.^[28] When gold/silica core NPs were administered systemically, enhanced contrast of the abdomen was achieved in MRI and CT.^[27] In contrast, 100 μL of DyNP-Gd-Ir NPs were injected at 3 mg/mL intratumorally. For these NPs there was an increase in CT contrast, improving 100% from 109 to 212 Hounsfield units. MRI negative contrast and positive contrast in the region of tumor was found to be 104% and 152% respectively, whereas the change in the surrounding tissue was negligible.^[30] Even without a cancer model, in vivo tests are essential in comparing imaging metrics between modalities, elucidating biodistribution, and pharmacokinetics. Iodinated oil-loaded i-fmSiO4@SPIONs were detected, post intravenous injection, in the liver by CT, MRI, and FL as shown in Figure 7.^[31]

Figure 7.

CT/MR/FL images of mouse livers after intravenous injection of i-fmSiO4@SPIONs at indicated times post-injection. (A) T2-weighted MRI, (B) CT, and (C) FL images. Adapted with permission from Xue et al.^[31]

4.2 CT / MRI / UCL

Although the combination of CT, MRI, and UCL is similar to CT, MRI and FL described above, the substitution of UCL for traditional FL distinguishes this combination with more emphasis on deeper optical imaging in preclinical research and development. Since the depth dependent limitations of optical imaging have prevented its implementation in deeper tissue with respect to in vivo imaging applications, the use of UCL has been an important step in overcoming this hurdle. By directly comparing the results of UCL-based optical techniques with CT and MRI, important comparative analysis can be used to push the limits of optical biomedical imaging.

UCNPs have generated recent interest for multimodal imaging and combined therapeutics. Using slightly different synthesis and formulation techniques, researchers can impart CT, MRI, and UCL functionalities.^[32–41] UCNP@SiO2-GdDTPA,^[33] upconversion sesame balls (UCS-balls),^[35] and multishell multifunctional nanoprobes (MUCNPs)^[36] have been described for this application. UCNP@SiO₂-GdDTPA nanoparticles, which incorporated Gd into a coating polymer, were recently explored.^{[33, 42]} In another system, NaY/GdF₄: Yb, Er, Tm nanocrystals were used as the core, coated with silica, and further functionalized with gold nanoparticles through adsorption. Polyethylene glycol was used to obtain biocompatible UCS-balls.^[35] Oleic acid capped NaYF₄:Yb/Tm was used as the core material for MUCNPs. An epitaxial coat method sequentially added a thick layer of NaLuF₄, a layer of NaYF₄, and a layer of NaGdF₄ followed for all the shells. Oleic acid was replaced by poly-acrylic acid (PAA) and folic acid (FA) to study performance of MUCNPs as contrast agents and tumor targeting nanoparticles.^[36]

Zhou et al. designed a mesoporous core-shell UCNP (mUCNPs) with a NaYbF₄:2%Er core and a mesoporous NaGdF₄ shell without using silicon.^{[37, 43]} Figure 8 illustrates the materials, synthesis schematic, and in vitro applications of mUCNPs. Other trimodal BaGdF₅-based UCNP contrast agents, with a convenient one-pot synthesis^[41] and which serve as drug carriers,^{[32, 44]} have also been described. For one drug carrier, BaGdF₅-based UCNPs, an active shell coating was added and doxorubicin was covalently conjugated.^[32] The other drug carrier utilized a two drug loading scheme to enhance the synergy of chemotherapeutic anticancer drugs camptothecin and doxorubicin.^[44] A NaGdF₄ nanoprobe based on chlorides of Y³⁺, Er³⁺, and Tm³⁺ was coated with tantalum oxide to form lychee-shaped upconverting nanolychees (UCNL).^[38] A method was reported for coating NaGdF₄:Yb³⁺/Er³⁺ UCNPs with polyetheleneimine (PEI, UCNP@PEI) which enables efficient gene transfection.^[39] In another example, a unique upconversion nanoparticle was developed based on Ho³⁺, which has to potential of MR, CT, and UCL imaging. These NaHoF₄ (Lipo-UCNPs) nanoparticles were synthesized by thermal decomposition.^[40] Another group synthesized a multifunctional NaLuF₄ based (Lu-UNCPs) upconversion nanophosphor.^[34] Rare-earth chlorides (Ln: Lu, Yb, Er/Tm) were used as core elements and proved to be stable for extended durations. Drug carrying BaGdF₅-based UCNP contrast agents conjugated with Dox were found to accumulate in cancer cells preferentially with their toxicity attributed to the change in pH after being endocytosed.^[32] UCNP@SiO₂-GdDTPA nanoparticles were incubated with HeLa and L02 (immortalized hepatic) cell lines for high contrast live cell imaging. In vitro CT imaging showed that UCNP@SiO₂-GdDTPA nanoparticles could enhance contrast 1.5 folds compared to the same mass concentration of Iopromide, and cells were also shown to internalize these particles.^[33] When HeLa and KB cell lines were incubated with mUCNPs, in vitro results for CT gave a value of 12.45 Hounsfield units, 2.4 folds higher than Iopromide. Folic acid conjugated mUCNPs (FA-mUCNPs) of concentration 100 μg/mL were added to HeLa, KB, and 16HBE cell lines. Almost all the cells in each case internalized FA-mUCNPs.^[37]

Figure 8.

Structure of mesoporous core shell UCNPs used for CT/MRI/UCL and drug loading. Schematic shows preparation of NPs for multimodal imaging and cancer-targeting therapy. Adapted with permission from Zhou et al.^[37]

In vitro results of CT imaging of the Lu-UCNPs suggested that they are more efficient than iodine based contrast agents. This was attributed to the fact that the absorption of lutetium is higher than iodine. After carrying out stability studies for 6 months, it was concluded that Lu-UCNPs were highly stable.^[34] Cytotoxicity studies for Lipo-UCNPs were carried out on BCEC and U87MG cells, and concentrations as high as 800 μg/mL were found to be nontoxic. Impressive relaxometric constant ratios (r₂/r₁) of 477 and 307 were found for NaYbF₄ and NaHoF₄ respectively, which are higher than commercial contrast agents like Combidex, Feridex and Resovist. Phantom CT images for Lipo-UCNPs contained 37.2 L/g, higher than commercial contrast agent iobitridol (15.8 Hounsfield unit L/g).^[40] At a concentration of 800 μg/mL of UCS-balls (100.8 μg/mL of Au), the CT clinical scanner showed a value of 250.5 HU. These results demonstrate the optimization that went into design and synthesis of UCS-balls.^[35] Trimodal CT, FL, and T1 and T2 weighted MRI images of UCNL are shown in Figure 9. MRI results yielded an r₁ and r₂ that were improve over previous reports. Trimodal in vitro imaging was also performed after incubation with RAW264.7 cells with UCNL (Figure 9E).^[38] UCNPs@PEI was tested in vitro on HeLa and L929 cells. At a concentration of 400 μg/mL, the PEI polymer was found to be toxic compared to the same concentration of UCNP@PEI after 48 hours. UCNP@PEI were internalized in HeLa cells without causing comparable cell death.^[39] Likewise, BaGdF₅:Yb/Er UCNPs were also internalized by HeLa cells.^[41]

Figure 9.

Trimodal imaging using upconverting nanolychees (UCNL). (A) CT images and the corresponding HU values of UCNLs, (B) Photos of samples under 980 excitation: from 5.2 mg/mL to 0 mg/ml (water) along the direction of the arrow. (C) T1-weighted MR images and relaxation rate R₁ and (D) T2-weighted MR images and relaxation rate R₂. (E) Trimodal imaging results in cells incubated with UCNL. From top to bottom: Fluorescent imaging, X-ray CT imaging in cell phantom (numbers indicate CT values in HU), and T1- and T2-weighted MRI. Adapted with permission from Xiao et al.^[38]

UCNP@SiO₂-GdDTPA nanoparticles were injected intravenously in Kunming nude male mice (60 μL of 1 mg/mL NPs) and luminescence was measured. UCL signal to noise ratio was 15 was confirmed by ex vivo UCL imaging. For MR imaging, 0.1 mg of the nanoparticles were injected intravenously causing the signal from the liver to increase by 22 % after 2 hours.^[33] For MR and CT in vivo studies, Lu-UCNPs were injected intradermally and readings were taken 30 minutes later (Figure 10). In the case of MR imaging, there was a 54 % increase in the signal from lymph nodes 30 minutes post injection. CT imaging followed the same trend showing an increase in the signal intensity with time. Also, the authors were able to pinpoint lymph nodes verified by 3D CT volume-rendered images.^[34] In vivo analyses of UCS-balls was performed on SD mice implanted with Walker 256 (tumor) cells. A 66.6 % increase in signal to noise ratio post intratumoral injection was observed when ~300 μg Gd/kg (corresponding to 1.1 mg Au/kg) was injected and imaged with MR. In CT, Hounsfield values of the tumor were found to increase from 41 to 102 whilst those value in soft tissue did not fluctuate. In case of FL imaging, the site of tumor emitted strong visible light when irradiated with NIR laser.^[35] In MRI of UCNL, positive and negative contrast enhancement rates of T1 and T2 signal intensities were found to be increased by 5.3% and −50.2% respectively. UCL signal was detectable 30 minutes post tail vein injection.^[38] PAA-MUCNPs exhibited typical in vivo nanoparticle behavior with slow excretion rate and accumulation in the liver and spleen. PAA-MUCNPs were intravenously injected into Kunming mice and the signals were detected in liver/spleen after 40 minutes and 2 hours for CT and MRI respectively. The authors also conducted a UCL study on FA-MUCNPs in nude mice bearing HeLa tumor. FA-MUCNPs were found to accumulate in the site of tumor after 3 hours.^[36] The efficacy of the one-pot synthesized BaGdF₅:Yb/Er nanoprobes as MRI and UCL imaging agents was evaluated in vivo in tumor bearing Kunming mice. When the mice were intravenously injected with UCNPs at 2.5 mg/mL concentration, nanoparticles accumulated in the spleen and were found to increase with time for 2 hours. CT was performed after intravenous injection with 3D volume-rendered images for different time points. Based on these images, these NPs were suggested to have the potential of diagnostic value in detecting splenic diseases.^[41] In vivo experiments using an intracranial glioblastoma mouse model in mice were performed using Lipo-UCNPs. Lipo-UCNPs (2%Ho with phospholipid-PEGylation) were injected intravenously (240 mg/kg dose) and the T2 weighted MR signal intensity was reduced by approximately 75% in the region of brain tumor 1.5 hours post injection. After successful uptake of Lipo-UCNPs in vitro, they were tested in vivo for UCL imaging. 2 hours after intravenous injection of Lipo-UCNPs, there was still a sharp enhancement in UCL images.^[40]

Figure 10.

Trimodal CT/MR/UCL imaging using Lu-UNCPs. (A) Three dimensional CT volume-rendered and (B) maximum intensity projections of coronal, (C) transversal, and (D) sagittal images of the lymph node obtained 30 min after intradermal injection of Lu-UCNPs. The red arrows indicate the locations of the lymph nodes. (E, F) T1-weighted MR images of transversal cross-section with lymph node circled in red, before and 30 min after intradermal injection. (G) In vivo, in situ, and ex vivo lymphatic UCL images of nude mice lymph nodes. UCL signals were collected at 800 nm. Adapted with permission from Zhou et al.^[34]

4.3 CT / MRI / US

In one validating example, CT and MRI were used to confirm the ability of US to locate therapeutic particles containing living cells administered to diabetic mice. A capsule-in-capsule (CIC) particle was developed that is capable of trimodal imaging (CT, MRI, and US) in addition to providing immunoprotection for transplanted cells.^[45] The CIC structure of this particle imparts enhanced acoustic properties while the CT- and MRI-active components reaffirmed the locating ability of US. Gold nanoparticles, iron oxide nanoparticles, and Ba²⁺ were loaded on to the inner capsule made of alginate. These capsules were coated with another layer of alginate crosslinked with Ba²⁺ to produce CICs. Capsules were found to be stable for at least 3 months with no leakage of nanoparticles from primary capsule to secondary capsule. Female C57BL/6 mice were used with the treatment group implanted with CICs in the peritoneal region. Spin-echo MR, gradient-echo MR, and micro-CT showed a sharp change in the contrast in the regions where CICs accumulated. US imaging could locate CICs distinctly. To assess the immunoprotection aspect of CICs, CICs containing beta-TC-6 mouse insulinoma cells were transplanted into a streptozotocin-induced mouse model of type I diabetes mellitus. Blood glucose levels were found to be normal immediately after transplanting CICs and were sustained for 75 days. CICs were extracted and analyzed ex vivo by staining, which showed the transplanted cell viability to be 55 % and the capsule remained structurally intact.

4.4 MRI / PET / FL

The benefit of complimentary information that can be achieved through combined structural and molecular techniques like CT/PET or MRI/PET is well known and evidenced by the wide clinical acceptance of these multimodal scanners. Incorporation of another highly sensitive modality like FL serves to further test this technique in a preclinical context and explore the ways it might be able to complement more expensive techniques like PET, which rely on radioactive materials.

Trimodal [¹²⁴I]HIB-Gd-liposomes were developed from radiolabeled hexadecyl-4-iodobenzoate (HIB).^{[46, 47]} In vitro analysis of liposomes in CT26 cells (murine colon cancer) showed that optical imaging had higher sensitivity and better spatial resolution when compared to PET. In case of MRI, spatial resolution was excellent but exhibited poor sensitivity. [¹²⁴I]HIB-Gd-liposomes were administered to CT26 tumor-bearing BALB/c mice and some bright FL spots were apparently located in the liver. The high tissue penetration depth of PET confirmed that the spots were originating from the liver and spleen. PET imaging also showed that the tumor was the region with second highest accumulation of the liposomes after the liver, and tumor tissue was the only region that was clearly imaged with PET after 24 hours. MR imaging proved to be a complementary imaging method by providing missing anatomical details, particularly with respect to soft tissues. Multiple tissues were identified including the liver, spleen, and tumor. PET showed contrast in the bladder although it was not detectable with FL.

Porphyrins have excellent capacity to enhance photomedicine and diagnostics.^[48–50] The potential of porphyrins as theranostic agents, when conjugated to coblock polymers-nanoporphyrins (NPs), has been demonstrated.^[51] A porphyrin-telodendrimer, PEG^5k-Por⁴-CA₄, was crosslinked with the help of reversible disulphide bond to establish a more stable crosslinked NP (CNP). Doxorubicin was encapsulated in the hydrophobic core with gadolinium (Gd(III)), and copper (⁶⁴Cu) ions were chelated in the CNPs. PDT was also carried out in vitro. Glutathione level studies were conducted on PC3m (prostate cancer cell line), MB49 (mouse bladder cancer cell line) and SKOV3 (ovarian cancer cell line) cells and showed that disulphide crosslinking was reversible, based on the intracellular dissociation of CNPs. In vivo studies were performed to evaluate the efficacy of CNPs as therapeutic (PDT, PTT) and diagnostic (FL, MRI, PET) probes. NPs chelated with Gd(III) and ⁶⁴Cu proved to be excellent imaging probes since the accumulation of these nanoparticles in the region of tumor was high.

Xie et. al., developed an IONP based multifunctional contrast agent which has a potential of carrying out PET, FL and MR imaging. ^{[52, 53]} In this work, the researchers started with coating the IONPs with dopamine. IONPs were incubated with dopamine in order to provide IONPs with a dopamine coating and then they were coated with human serum albumin (HSA) by dispersing these dopamine coated IONPs in HSA. To further functionalize the HSA-IONPs, DOTA and Cy5.5 were coupled to the surface of the NPs. ⁶⁴Cu was later chelated to DOTA for PET imaging. U87MG (human glioblastoma) cells were injected subcutaneously into the mice to induce tumor and in vitro testing was carried out with athymic nude mice bearing tumors. Phantom studies for MR imaging showed that HSA-IONP had a relaxivity (r2) of 313 mM⁻¹s⁻¹, which is higher than the commercial contrast agent Feridex. For PET/FL imaging, HSA-SPIONs were injected intravenously into the mice (10 mg Fe/kg). Post injection, mice were imaged and the tumor was clearly distinguishable at the 1 hour time point with FL. The demarcation between tumor and normal tissue became more evident with time. Tumor to muscle ratio increased from 2 at the 1 hour time point to 3 at the 18 hour time point. PET imaging showed a similar trend to FL imaging. With ratio values of 4.5 at the 1 hour time point and 8 at the 18 hour time point, PET imaging yielded superior tumor to muscle signal ratios compared to FL. This improvement was ascribed to lesser background noise in case of PET imaging. MR imaging studies were carried after intravenous administration of HSA-IONPs (10 mg Fe/kg). Results post-injection in comparison with pre-injection showed a decrease of 30 % in the region of tumor after 18 hours. Also, MR imaging showed that the nanoparticles were inhomogeneously distributed in the tumor. Ex vivo results showed accumulation of nanoparticles the tumor and liver. Histology studies further demonstrated that the NPs accumulated in tumor and were randomly distributed. Most of the particles were found outside vessels suggesting extravasation to be high. However, it was realized through ex vivo studies that there was poor correlation between nanoparticles and vasculature distribution. With Prussian blue (to visualize IONPs) and F4/80 (a macrophage marker) double staining, it was concluded that NPs in tumor were not trapped by macrophages.

Huang et. al., developed a nanoprobe based on mesoporous silica which incorporates MR, PET and optical imaging enabling noninvasive tracking of tumor metastasis.^[54] In this work, synthesis started with mixing cetyltrimethylammonium bromide (CTAB), tetraethyl orthosilicate and fluorescent dyes (FITC/ZW800) together. Later, CTAB was removed to make the structures mesoporous to obtain mesoporous silica based nanoprobes (MSNs). In the second step, Gd³⁺ ions were chelated to dye-doped MSNs through Si-Gd-DTTA to obtain dye@MSN@Gd. In the third step, MSNs were radiolabeled with ⁶⁴Cu to obtain dye@MSN@Gd@⁶⁴Cu. Gd³⁺ and ⁶⁴Cu were chelated to MSNs for MR and PET imaging, respectively. FITC-incorporated MSNs were used for in vitro studies and ZW800-embedded MSNs were used for in vivo studies. For in vivo studies, mice were subcutaneously injected with 50 μL of MSNs (2 mg/ml, 20 μCi ⁶⁴Cu). A 4T1 tumor lymph node metastasis model was used to study the efficiency of MSNs as multimodal imaging probes in vivo. First, to study the fluorescence retention in cells, FITC-MSNs were incubated with macrophages and fluorescence was visible for at least a week, showing the stability of MSNs in intracellular environments. In vivo and ex vivo results showed that MSNs were concentrated in tumor metastatic SLNs (T-SLN) and there was no signal was observed in normal contralateral SLNs (N-SLN). MR imaging showed that the MSNs were scattered along the margin of T-SLNs, which was in agreement with in vivo and ex vivo optical imaging. MR signal was visible up to 21 days after injection in T-SLN. PET imaging showed an increase in the radiotracer accumulation in T-SLNs with time, and there was no specific radiotracer signal observed in N-SLN. At 1 hour and 2 day time points T-SLN was 35 and 7 folds higher than N-SLN. Since the half-life of ⁶⁴Cu PET is 12.6 hours it is not suited for imaging with time-points over 2 days. In a clinical setup if MSNs were to be employed, PET imaging would be carried out first followed by MR imaging. Optical imaging would be performed last since it assists in surgery.

4.5 MRI / PET / UCL

With so many combinations of NP classes, therapeutic agents, and targeting moieties are being explored, finding ways to evaluate the effectiveness of each component can be challenging. Trimodal MRI, PET, and UCL can accomplish this by the MRI providing anatomical location with respect to surrounding tissues, PET can be used to quantitatively determine the particular dose located at that region, and UCL serves as a low-background optical method to determine the local targeting efficacy. NaGdF₄ upconversion nanophosphors were synthesized tagged with radiolabeled arginine-glycine-aspartic acid (RGD) in order to image the tumor angiogenesis associated with many cancers.^{[24, 55]} In vitro incubation with U87MG was imaged with UCL and the labeled NPs were confirmed to locate on the exterior of the cell. After intravenous injection in a U87MG murine tumor model with or without blocking of the RGD targeting moiety, small animal PET and MRI were performed. PET and MRI results confirmed contrast enhancement in tumor location, although there was no statistical difference between models in which the target of the RGD was blocked.

4.6 MRI / PET / PA

In one example, a multimodal imaging probe using ultrasmall (< 10 nm) melanin nanoparticles (MNP).^[56] The use of this pigment molecule serves to impart PA activity as well as provide an environment to facilitate easy incorporation of PET and MRI active components. This is an attractive alternative to more complex chemical modifications that can be used to accomplish the same goal. PEG-MNPs were functionalized by conjugating RGD, then further modified with Fe³⁺ and ⁶⁴Cu²⁺. In vitro studies were carried out with U87MG cells and in vivo studies were done on female athymic nude mice bearing tumors derived from the same cells. For MRI, in vitro results showed that nanoparticles carry a relaxivity (r1) value of 1.2 mM⁻¹s⁻¹. Cellular uptake of ⁶⁴Cu-labeled PEG-MNPs doubled after conjugating these nanoparticles to RGD peptide. For MRI, mice were imaged 4 hours after injection and showed a 30 % increase in signal intensity in the region of tumor. In vivo results for PET imaging showed that the uptake of RGD nanoparticles increased in the region of tumor to 6 % injected dose per gram of tissue after 24 hours. The nanoparticles that passed were cleared through the hepatobiliary system.

4.7 MRI / PET / Cerenkov

Because the applications and benefits of Cerenkov imaging are still actively being investigated, the ability to compare it directly with modalities like MRI and PET is useful. With a single trimodal probe active in each modality, it is possible to obtain effective comparisons. ¹²⁴I-labeled thermally cross-linked SPIONs (TCL-SPIONs) were used to identify and image SLN in mice. These particles did not exhibit any aggregation and were stable for up to half a year. Their T2 values were also higher than Feridex. The two front paws or BALB/c mice were injected with the ¹²⁴I-labeled TCL-SPIONs after the mouse breast cancer cell line 4T1 was introduced into the left shoulder ten days prior. MRI, PET, and Cerenkov imaging were then performed to determine if the lymph node closely to the injected cancer cells could be differentiated from the SLNs not located close. An indication of reduced uptake in the SLN closest to the cancer cells was noticed in all three imaging modalities consistently.^[57] When ⁶⁸Ga-SPIONs were also used to visualize SLN in these modalities, PET, MR, and Cerenkov imaging were able to clearly locate the nodes. An inhomogeneity of the SLN ⁶⁸Ga-SPIONs signal in metastatic SLNs was also observed.^{[58, 59]}

4.8 MRI / SPECT / FL

As multimodal MR scanners start to enter the market, it will be important to emphasize their versatility in imaging paradigms. Being able to design multimodal probes for MRI and SPECT may enhance the utility of these scanners. At least two trimodal MRI/SPECT/FL contrast agents have been developed; a functionalized polymeric imaging agent based on high-molecular-weight poylglycerol (HPG) with DOTA (HPGD),^[60] and SPIONs labeled with ^99mTc and Alexa dye to form ^99mTc-AF-SPIONs.^[61] HPG was functionalized with Gd³⁺, ¹¹¹In, and Alexa dye. Human vascular endothelial cells (HUVECs) were found to exhibit almost 100% viability when incubated with HPG and HPGD at the same concentration. Relaxivity (r₁) of HPGD-Gd was found to be significantly higher than commercial agents like Galbumin. Subcutaneous injection of 99mTc-AF-SPIONs in the hind paw of Wistar rats resulted in clear visualization of SLNs in MRI, SPECT, and FL after 5 hours. HPGD-Gd were alternatively injected intravenously into female C57Bl/6 mice and accumulation in the region of tumor was found to be maximal 3 days post-injection. The biodistribution of ¹¹¹In-HPGD showed considerable increase in the accumulation of NPs in the tumor region with a peak value on day three. PET and MRI were helpful for visualizing the target in whole body scans during lymph node imaging of ^99mTc-AF-SPIONs, after which FL could be used intraoperatively to highlight the target. The ability to perform this imaging during surgery then quantify the target ex vivo with either optical or radiologic techniques highlights the synergistic effects of multimodal imaging.

4.9 MRI / US / PA

US and PA, which both rely on ultrasonic detection, are naturally amenable to simultaneous detection with respect to instrumentation implementation. Microcapsules were synthesized with poly(lactic acid) loaded with IO nanoparticles (IO@PLA/GO). ^[62] MRI of nude mice after intratumorally injection with a 20 mg/kg dose showed considerable enhancement in local contrast. Pulse-inversed harmonic imaging (US) carried out on New Zealand white rabbits showed that the contrast enhancement persisted for 5 minutes, in accordance with clinical expectations. NPs placed under 44 mm of chicken breast were still detectable with PA at a concentration of 32 mg/mL. PTT applications were confirmed in vitro, showing the NPs to be have therapeutic character when irradiated.

4.10 MRI / PA / Raman

The many levels of complexity in organs and tissues can make systemic evaluation of disease and treatment challenging. When trying to analyze a complicated disease and procedure such as brain tumor surgery, multiple procedures are needed to determine the anatomical, physiological, and cellular status at the location of interest. Reducing the time, risk, and cost to patients in this situation has been addressed with trimodal MRI, PA, and Raman imaging. This imaging agent (MPR nanoparticle) was synthesized from a 60 nm Au core coated with a Raman-active molecular trans-1,2-bis(4-pyridyl)-ethylene, given a protective coating of silica, and enriched with malemide-DOTA-Gd.^[4] MPRs were subcutaneously injected into nude mice followed by MRI, PA, and Raman signal evaluations. Signal intensity was found to highly correlated to the concentration and Raman imaging was the most sensitive imaging method, followed by photoacoustic and MR imaging. For in vivo experiments, mice with orthotopic brain tumors were injected with MPRs intravenously. In vivo MRI and PA of mice implanted with eGFP⁺U87MG exhibited signal enhancement 30 minutes post injection. MRI signal was monitored until 24 hours and there was no decrease in the signal intensity compared to the 2 hours reading, whereas Gd-based imaging agents are flushed out in few minutes. In the case of Raman imaging, there was peak as soon as MPRs were injected but that was attributed to nonspecific circulation of MPRs in superficial layers. Distribution of MPRs was validated by histology. MPRs were found only in the region of the tumor and tumor boundaries, and none were seen in surrounding healthy tissue. The potential of MPRs as PA and Raman imaging in guided tumor resection was examined by craniotomy of mice with visible tumors resected. High resolution Raman images were obtained before and after resection, and finger like projections were observed around the resected area where MPRs were found to be concentrated. More histological investigation of these areas led to the conclusion that they were minute projections of tumor. This example illustrates the potential benefit that higher-order multimodal probes offer.

4.11 US / PA / FL

In order to accomplish trimodal US, PA, and FL imaging, a porphyrin-lipid based microbubble^[63–65] composed of a porphyrin-lipid together with conventional lipids was formed that encapsulated perfluoropropane gas. The porphyrin-lipid was previously used to develop liposome-like nanovesicles and has excellent potential for multimodal targeted approaches.^[66–69] Compared to unimodal microbubble made from traditional phospholipids, there was no change observed with respect to US however, PA was enhanced by 40 folds, and FL by 17 folds.^[64] Porphyrin-lipid microbubbles were injected into KB tumor bearing-xenograft mice intravenously. Contrast of the US images in case of conventional and porphyrin-lipid microbubbles was enhanced in a similar fashion. An increase in PA signal intensity was evident only for the porphyrin-lipid microbubbles. Porphyrin-lipid microbubbles enhanced signal in the tumor compared to healthy tissue illustrating potential for ex-vivo imaging either as a source for verifying drug accumulation or to estimate biodistribution.

4.12 PET / FL / UCL

FL and UCL both involve photon absorption and emission processes and are similar in implementation. In theory, the two types of imaging could provide complementary information, with UCL being better suited for deeper imaging. Because the combinations of materials, synthetic procedures, and applications of nanoparticles in imaging are virtually endless, it is important to find ways of characterizing these novel agents. Trimodal PET, FL, and UCL imaging was used to validate a range of nanoparticle synthetic techniques and applications.^[70] Alpha-cyclodextrin was used as the host material and could solubilize particles regardless of size or synthesis method (thermal decomposition, hydrothermal, solvothermal). Besides the physical characterizations the trimodal imaging was carried out in cell labeling and in vivo lymphatic imaging.

4.13 Bioluminescence

Bioluminescence, which involves enzymatic activation of light-generating chemicals is not considered an exogenous contrast mode, but rather is used to unambiguously identify genetically modified cells (such as cancer cells) in vivo. Nevertheless, there are some multimodal examples involving bioluminescence worth noting. When other modalities such as PET and CT are combined with bioluminescence, co-localization of the contrast agent with the tumor can be assessed. A trimodal imaging probe was developed to detect cancerous xenografts from different cell lines and if a melanoma cell line had metastasized in a murine model.^[71] CT and PET both registered the xenografts at the same location and combined PET/CT showed lesions that were not visible in CT alone. All three modalities registered the signal of melanoma metastases, which was confirmed with ex vivo assays. A PET, FL, and bioluminescent reporter gene has also developed for whole-body imaging.^[72] Whole-body imaging was performed after administration of D-luciferin to generate the bioluminescent signal and was successful in all three modalities.

4.14 Higher-order Multimodality

The examples described above refer to trimodal imaging agents. To date, there have also been recent examples of imaging agents that demonstrate contrast in four, five and six modalities. These hypermodal agents at this stage are more conceptual than practical, but in theory one day could simplify and streamline medical imaging. Additionally, since techniques such as MRI and CT sometimes must be avoided by patients with the presence of a metal implant or with risk of overexposure to radiation, a universal imaging agent could provide invaluable versatility. Having probes that are capable of pan-modal imaging could simplify and streamline the entire process of medical imaging. NaLuF₄-based optimized nanoparticles with the potential for tetramodal imaging (MR, CT, SPECT, and UCL imaging) were recently developed.^{[35, 73]} In vitro studies were carried out on macrophages and in vivo studies on nude mice with or without KB tumors. From the phantom studies carried out with NaLuF₄: Yb³⁺, Tm³⁺@NaGdF₄-(¹⁵³Sm) NPs for CT imaging, it was concluded that these nanoparticles at 1 mg/mL concentration show the same X-ray attenuation as 1.6 mg/mL of Iopromide. The signal-to-noise ratio for UCL imaging in vivo was found to be 21. For in vivo studies, mice were injected with 800 μL of NaLuF₄: Yb³⁺, Tm³⁺@ NaGdF₄-(¹⁵³Sm) nanoparticles (30 mg/mL) and 1 hour later high X-ray attenuation was observed in the region of the liver and spleen. Nude mice bearing KB tumors were used to probe for tumor angiogenesis. CT showed high attenuation in the region of tumor, MR detected an intense signal from the surface of tumor, SPECT imaging showed a strong signal in the region of tumor and also assisted in quantification (1.7 % of the total dose was within tumor), and UCL imaging aided in histochemical analysis of tumor angiogenesis. These results demonstrate the potential benefits of multimodal imaging probes and highlight the synergistic value of multimodal imaging data.

A magnetic-fluorescent-bioluminescent-radioisotopic particle (MFBR) was synthesized for quadruple modality MRI, PET, FL, bioluminescence.^[74] These particles were incorporated into rat glioma cell line C6 before being injected at different tissue depths in normal nude mice. Bioluminescence was detected at both intramuscular and subcutaneous locations, which was further validated by μPET. These NPs demonstrate the capacity of multimodal nanoprobes to be used for stem-cell tracking at different levels in the body via four combined imaging modalities.

Nanoparticulate platforms for pentamodal^[75] and hexamodal^[76] have also be developed for proof-of-principle studies in recent years. Rieffel et al. utilized a UCNP core, porphyrin-phospholipid (PoP), and post-synthesis chelation of a PET radiotracer to render these NPs (PoP-UCNPs) suitable for multimodal in vivo diagnostics.^[76] The signal penetration and resolution for each modality was tested as a function of tissue depth, with CT and PET imaging performing best in both categories. In vivo the PoP-UCNPs were administered to the rear footpad of BALB/c mice and imaged in each modality. The first lymph node draining the injection site was detected in PET, FL, UCL, PA, and Cerenkov imaging (Figure 11) with CT capabilities confirmed in phantom samples. This NP shows how a relatively simple agent can be used for a wide range of imaging modalities.

Figure 11.

Hexamodal lymphatic imaging using PoP-UCNPs in mice. PoP-UCNPs were injected in the rear left footpad and imaged in six modalities 1 hour post-injection. Accumulation of PoP-UCNPs in the first draining lymph node is indicated with yellow arrows. a) FL and b) UC images with the injection site cropped out of frame. c) Full anatomy PET, d) merged PET/CT, and e) Cerenkov images. f) PA images before and g) after injection show endogenous PA blood signal compared to the contrast enhancement for visualization of the previously undetected lymph node. Adapted with permission from Rieffel et al.^[76]

5. Conclusion

While advances in clinical and preclinical imaging modalities will continue to push the limits of what can be accomplished alone, higher-order multimodal imaging holds potential unite and bridge different imaging modalities. Nanoparticulate approaches are most likely required to make better use of complimentary imaging technologies to reinforce a cycle of progress. Not only have some multimodal probes shown to be competitive with existing clinical imaging agents, many also hold therapeutic capacity.^[77] Many of the multimodal contrast agents developed so far incorporate an MR active component, which reflects the clinical importance of MRI. The next steps in translating these multimodal probes from the bench to the clinic include more detailed and rigorous investigations of their in vivo pharmacokinetic and toxicity profiles. Additionally, large scale manufacturing protocols should be established that characterize the colloidal, chemical, and storage stability of the NPs. Crucially, synergistic imaging applications that address unmet needs must be identified. Although the versatility of nanoprobes in medicine and imaging has been one of the biggest stimuli towards multimodality, simplification is also a viable strategy towards clinical implementation. In experimental design this may mean sacrificing novel targeting ligands or therapeutics in favor of less complex probes based on materials that can be, or that already are, clinically approved. Higher-order multimodal probes hold great research potential, and with motivated efforts, may eventually shift biomedical imaging paradigms.