Nanomaterial-based contrast agents

Abstract

Medical imaging, which empowers the detection of physiological and pathological processes within living subjects, has a vital role in both preclinical and clinical diagnostics. Contrast agents are often needed to accompany anatomical data with functional information or to provide phenotyping of the disease in question. Many newly emerging contrast agents are based on nanomaterials as their high payloads, unique physicochemical properties, improved sensitivity and multimodality capacity are highly desired for many advanced forms of bioimaging techniques and applications. Here, we review the developments in the field of nanomaterial-based contrast agents. We outline important nanomaterial design considerations and discuss the effect on their physicochemical attributes, contrast properties and biological behaviour. We also describe commonly used approaches for formulating, functionalizing and characterizing these nanomaterials. Key applications are highlighted by categorizing nanomaterials on the basis of their X-ray, magnetic, nuclear, optical and/or photoacoustic contrast properties. Finally, we offer our perspectives on current challenges and emerging research topics as well as expectations for future advancements in the field.

Introduction

Contrast agents can be used in imaging procedures to enhance the detectability of underlying tissue anatomy or to derive information about certain pathological processes. These agents have traditionally been based on small molecules such as gadolinium chelates or iodinated small molecules, but their use has been restricted to a relatively small number of imaging applications, such as blood pool or urinary tract imaging. Therefore, much of the research in recent decades has been devoted to improving current designs and developing alternative contrast agents, such as nanomaterial-based formulations with novel compositions and structures. Nano-based contrast agents have performed well in a number of advanced biomedical applications, such as cell tracking, intraoperative tumour detection and implant monitoring^1,2.

There are several criteria to consider when designing nanomaterial-based contrast agents for specific bioimaging applications³ (Fig. 1). First, nano-based contrast agents are typically in the 1–400 nm size range, but their size can be tuned to exhibit very different behaviours in vivo. Sub-5-nm nanoparticles (NPs) tend to rapidly clear via kidney filtration to prevent long-term retention in the mononuclear phagocyte system (MPS) organs^4–6, whereas larger NPs may remain in the body longer. NPs can accumulate in some diseased tissues via processes such as the enhanced permeation and retention (EPR) effect if long-circulating, receptor-specific uptake if targeted or opsonization by plasma proteins and others^7–10. Size may also affect the contrast properties of an NP platform; for instance, the fluorescence emission wavelength of quantum dots (QDs) generally increases with increasing core diameter¹¹. Nano-based contrast agents are mostly spherical in shape, but can also be formed as rods, stars or other shapes, which can determine their imaging utility and avidity in target binding. For example, gold nanostars have a high surface-to-core ratio for improved ligand conjugation, and the tips of their points exhibit enhanced Raman spectra for effective surface-enhanced Raman spectroscopy (SERS)-based imaging¹². Next, materials that can generate or enhance contrast, such as metals, fluorescent probes and radionuclides, can be incorporated into the core, integrated in the shell or attached to the surface or coating of the NP. Notably, NPs carry more contrast payloads per entity compared with conventional small molecules, resulting in image contrast of higher intensity¹³. Contrast agents can be encapsulated in or bound to larger carrier structures, such as liposomes, dendrimers, polymers and silica, to facilitate desired circulation half-lives for specific applications, biodegradability and colloidal stability¹⁴. NP surfaces can be easily modified with coating ligands, such as neutrally charged polyethylene glycol (PEG), that yield varying pharmacokinetic and biodistribution profiles¹⁵. The surface can also be functionalized with drugs or nucleic acids for therapeutic effects, or targeting moieties, such as antibodies and peptides, for targeted imaging.

Fig. 1 |. Design principles of nanomaterial-based contrast agents for various imaging modalities and biomedical applications.

The physicochemical properties of nano-based contrast agents are generally described by their core and hydrodynamic diameter, chemical composition, shape, surface chemistry and functionalization. These agents are designed and optimized for major bioimaging modalities that produce contrast based on X-rays, radioactive decay, magnetism, optical photons and acoustics. These agents, in conjunction with the appropriate imaging techniques, are applied to provide structural and functional information of the disease in question. HDL, high-density lipoprotein; NP, nanoparticle; PEG, polyethylene glycol.

Nano-based contrast agents can be used for many major medical imaging modalities^16–18, as summarized in Table 1 and Fig. 1. Computed tomography (CT) generates image contrast on the basis of differential mass-dependent absorption or scattering of ionizing X-ray radiation. Although non-contrast enhanced CT is effective for imaging electron-dense tissues, such as bones and calcifications, and non-dense tissues, such as lungs, distinguishing soft tissues without the use of exogenous contrast agents is difficult. The earliest reports of nano-based CT contrast agents were based on emulsions or liposomes containing clinically approved iodinated molecules^19–21. Subsequent formulations typically involve high atomic number (Z) elements such as cerium, gold and bismuth^22,23. Contrast-enhanced mammography, or dual-energy mammography (DEM), is a clinical tool that improves lesion detection in dense breasts. Breast imaging via DEM uses lower X-ray energy spectra compared with CT; therefore, lower Z elements, such as silver and molybdenum, would be suited for development into DEM-specific nano-based contrast agents^24,25.

Table 1 |.

Features of commonly used medical imaging modalities and relevant nano-based contrast agents

Imaging modality	Spatial resolution	Temporal resolution	Sensitivity	Penetration depth	Example nanomaterials
X-ray contrast agents
Computed tomography	25 μm–5 mm	Seconds to minutes	10−3 M	No limit	Metal NPs (iodine, gold, tantalum, bismuth, lanthanides)
Mammography	50 μm–1 mm	Seconds	10−3 M	No limit	Metal NPs (iodine, silver, molybdenum, gold)
Nuclear contrast agents
Positron emission tomography	1–5 mm	Seconds to minutes	10−10 M	No limit	89Zr-labelled HDL, 64Cu-labelled CuS NPs
Single photon emission computed tomography	1–10 mm	minutes	10−10 M	No limit	99mTc-labelled IONPs/liposomes
Magnetic contrast agents
Magnetic resonance imaging	25 μm–5 mm	Minutes to hours	10−5 M	No limit	IONPs, Gd-dendrimers, MnO NPs
Magnetic particle imaging	500 μm	Milliseconds	10−6 M	No limit	IONPs, FeCo NPs
Optical contrast agents
Fluorescence imaging	1–3 mm	Seconds to minutes	10−12 M	<1–5 cm	QDs, AuNPs, carbon nanotubes, dye-loaded micelle
Photoacoustic imaging	50 μm–1 mm	Seconds to minutes	10−11 M	<5 cm	AuNPs, semiconducting polymer
SERS	1 μm–1 mm	Seconds to hours	10−15 M	<0 mm	Gold and SERS NPs
Ultrasound	0.1–1 mm	Seconds	10−4 M	5–30 cm	Lipid, polymer or protein-coated nanobubbles

Values listed span clinical and preclinical systems. Sensitivities are on a per element basis. HDL, high-density lipoprotein; IONPs, iron oxide nanoparticles; NP, nanoparticle; QD, quantum dot; SERS, surface-enhanced Raman spectroscopy.

Magnetic resonance imaging (MRI) is a non-ionizing modality that generates tissue contrast on the basis of the relaxation processes of neighbouring water protons in the presence of an external magnetic field. Most MR contrast agents are either paramagnetic, which yields contrast for T₁ (longitudinal)-weighted imaging, or superparamagnetic, which provides contrast in T₂/T₂* (transverse)-weighted imaging (where T₁ and T₂ are the water relaxation times). Notably, by modulating the core size, the same type of nano-based contrast agent can be used for either T₁-weighted or T₂-weighted imaging²⁶. T₁ agents, such as those based on gadolinium and manganese, result in positive contrast (brightening) on MR images²⁷, whereas T₂-shortening agents, such as iron oxide nanoparticles (IONPs), give rise to negative contrast (darkening) on MR images^28,29. IONPs have shown promise in clinical diagnosis as they are frequently used as an ‘off-label’ MR contrast agent. Another class of MR agent, which does not rely on proton relaxivity, produces a signal through chemical exchange saturation transfer (CEST) by enhancing the difference in the chemical shift of exchangeable protons. CEST agents can be classified as paramagnetic (PARACEST) agents that involve paramagnetic lanthanide ions, such as europium and ytterbium, or diamagnetic (diaCEST) agents that involve liposomes or polymers^30,31. One advantage of CEST is the ability to turn the contrast on or off depending on the pulse sequence used, which allows for hot-spot imaging. However, it is of lower sensitivity than conventional MRI. Another variant of MRI is ¹⁹F MRI, in which specialized coils are used to detect signal from agents loaded with fluorine. As humans have low levels of fluorine, this also results in hot-spot imaging. ¹⁹F MRI agents are usually emulsions or NPs composed of inorganic fluoride salts^32,33. Drawbacks of ¹⁹F MRI include lower sensitivity compared with conventional MRI and the need for additional equipment. Magnetic particle imaging (MPI) is a recently developed technique in which a shaped magnetic field is scanned over the region of interest³⁴. Magnetic NPs can be detected by this technique, owing to signals emitted when they enter the field-free zone as the field moves. MPI is a high spatial resolution, temporal resolution and sensitivity technique³⁵ that provides readouts on the spatial distribution of magnetic NPs, such as IONPs, but as it offers no signal from endogenous tissue, it is often paired with another technique, such as CT, to provide anatomical information. Limitations of MPI include sparse preclinical availability and a lack of widespread clinical use.

Nuclear imaging approaches, including positron emission tomography (PET) and single photon emission computed tomography (SPECT), are based on the detection of ionizing γ-photons emitted from radionuclides during radioactive decay. PET signals arise from electron–positron annihilation that releases two γ-rays travelling in opposite directions³⁶. PET imaging is an appealing approach for measuring NP biodistribution in vivo over time³⁷. SPECT has the potential for multiplexed imaging owing to its ability to discriminate isotopes based on the characteristic energies of emitted γ-photons. Essentially, any nanomaterial can be modified with radioisotopes through chelation, intrinsic labelling and chelator-free post-synthetic labelling³⁸. Commonly used radioisotopes for NP imaging include ⁶⁴Cu and ⁸⁹Zr for PET imaging and ^99mTc and ¹¹¹In for SPECT imaging³⁹. Although nuclear imaging suffers from much lower spatial resolution, it has a better detection limit towards its contrast agents than MRI and CT.

Fluorescence imaging is a rapid, high-throughput modality that detects photon emission in the visible or near-infrared (NIR) range upon energetic excitation. This technique offers real-time imaging and is highly sensitive towards contrast agents but has low depth penetration (<1 cm) owing to visible light absorption and scattering in tissues. To improve tissue penetration issues, NIR emitting semiconducting QDs, such as those based on cadmium selenide (CdSe) and silver sulfide (Ag₂S), can be synthesized as tissue absorbs less light in the NIR region⁴⁰. The NIR range is further categorized into NIR-I (700–950 nm) and NIR-II (1,000–1,700 nm) windows, where the latter is more advantageous for bioimaging owing to lower tissue autofluorescence and photon scattering and absorption, thus achieving higher signal-to-noise ratio (SNR) and deeper tissue penetration. Owing to the quantum confinement effect, the core size of a QD dictates their absorbance and emission wavelengths⁴¹. Other frequently reported nano-based fluorescent or optical agents include gold nanoparticles (AuNPs), fluorophore-conjugated polymer-based or lipid-based NPs, carbon nanomaterials and lanthanide-doped upconverting NPs.

Photoacoustic imaging (PAI), also known as optoacoustic imaging, is based on the generation of broadband ultrasonic waves from transient thermoelastic expansion of tissues upon direct laser light irradiation. PAI is less limited by tissue thickness and thus achieves better penetration depth (<5 cm) compared with conventional optical imaging techniques. Exogenous nano-based PA contrast agents have been engineered to possess efficient photothermal conversion, high photostability and optimal NIR absorption with large cross-sections⁴². These agents include gold nanostructures, Ag₂S NPs, dye-decorated nanomaterials and semiconducting polymeric NPs^43,44.

Ultrasound imaging uses hand-held transducers that emit sound waves that are typically in the 1–20 MHz range. These sound waves are reflected, scattered or absorbed by tissues and tissue interfaces. From the sound waves that return to the transducer, images of the subject can be formed. Ultrasound provides high spatial and temporal resolution, is portable and low cost and is widely available. On the contrary, it is operator-dependent, has low soft tissue contrast and provides limited fields of view. Contrast agents for ultrasound have historically been based on gas microbubbles⁴⁵, which are outside the scope of this Primer. However, in recent years, contrast agents for ultrasound based on nanobubbles have been reported. These structures are composed of a gas core, which is often a perfluorocarbon, enclosed in a shell made of lipids, polymers or proteins⁴⁶.

This Primer aims to inspect the different facets of bioimaging using nanomaterial-based contrast agents and should give readers the tools to design and evaluate nanomaterials as contrast agents. The Primer examines various synthetic approaches used to produce nano-based contrast agent platforms by focusing on their primary design considerations. The array of analytical techniques used for characterizing NP formulations and assessing their contrast properties among other attributes is outlined, and the main biomedical applications of nano-based contrast agents, including structural and functional imaging, are highlighted using relevant examples. Key factors that affect the reproducibility of nanomaterials and reliability of data are described, and potential guidelines and repositories for data reporting are discussed. Finally, the Primer offers perspectives on current challenges and future developments in the field of nano-based contrast agents.

Experimentation

Some of the most frequently studied nano-based contrast agent formulations for medical imaging applications include inorganic solid core NPs (for example, metal and metal alloy), lipid-based structures (for example, liposomes, emulsions, micelles and lipoproteins), polymeric NPs (for example, PEG, dextran and synthetic polymers) or a combination of these different NPs. This section details the various synthetic strategies and purification approaches, as well as surface modification and conjugation methods, that are most commonly implemented to fabricate these nano-based contrast agents (Fig. 2). Recent advances in synthetic techniques are also discussed, and factors dictating the design of an NP contrast agent are described.

Fig. 2 |. Overview of the approaches used for synthesizing, conjugating and purifying nanomaterial-based contrast agents.

a, Nano-based contrast agents have traditionally been synthesized using aqueous and organic solvents. Recent syntheses based on green chemistry, biological precursors, microfluidics and artificial intelligence (AI) have become increasingly utilized for nanomaterial fabrication. b,c, Surface modification (part b) and/or functionalization (part c) is performed to render the agents viable for bioimaging applications. d, The agents are further purified to remove impurities before in vitro and in vivo evaluation. EDC, N-ethyl-N’-(3-dimethylaminopropyl) carbodiimide; MWCO, molecular weight cut-off; NHS, N-hydroxysuccinimide; PEG, polyethylene glycol. Part c reprinted from ref. 86, CC BY 4.0.

Nanoparticle synthesis

At the broadest level, NP synthesis can be classified as top-down (milling) or bottom-up (self-assembly), of which the latter is more commonly used for contrast agents. The top-down approach involves ball milling, laser ablation, photolithography or sputter deposition to derive NPs from bulk materials. These physical processes are not cost-effective as they require expensive and complicated instruments, as well as high power consumption. On the contrary, bottom-up syntheses entail the ionization of precursor chemicals in liquid or vapour phase, which leads to the nucleation and growth of nuclei into nanoclusters, or other approaches such as precipitation or self-assembly. For low-cost, high-throughput fabrication of nano-based contrast agents, aqueous syntheses are ideal as they use fewer toxic reactants and are simple, economical, energy-efficient and eco-friendly. In many cases, hydrophobic or organic syntheses are adopted to control the nanomaterial size, composition and shape with high precision. However, subsequent surface modification is necessary to improve the solubility and stability of the obtained nanostructures. Recently, there has been growing interest in applying natural templates, green chemistry, microfluidics and artificial intelligence for NP synthesis^47–50 (Fig. 2a). Nevertheless, NP formulation methodology and the required associated apparatus are highly specific to each NP type.

Inorganic nanocrystals.

Inorganic nanocrystals are typically synthesized from a mixture of metal salts and capping ligands either in the aqueous or hydrophobic phase⁵¹. For instance, the most commonly used method to synthesize AuNPs is that of Turkevich, which involves adding a reducing agent (for example, sodium citrate) to an aqueous gold salt. The size of AuNPs can be fine-tuned by varying the amount of reductants used or by using reductants with different reducing strengths — stronger reducing agents typically result in smaller NPs (for example, sodium borohydride versus hydroquinone)⁵². Furthermore, the Brust method transfers gold salts from water to toluene or other organic solvents, while capping AuNPs (1–5 nm) with dodecanethiol or other hydrophobic ligands⁵³. With this method, the core size can be controlled by adjusting the amount of ligands used in the synthesis. The morphology can also be tuned by adding cetyl trimethylammonium bromide to gold seeds to produce gold nanorods (AuNRs) or nanostars⁵⁴ or by depositing gold seeds onto a template (for example, silica spheres) to yield gold nanocages or nanoshells⁵⁵.

Many hydrophobic phase-based syntheses have been established to produce various metal nanocrystals with controllable and monodisperse diameters. For example, iron salts are reacted with sodium oleate to form iron oleate complexes, which are then slowly heated in octadecene to produce oleic-acid-coated iron oxide cores⁵⁶. The same strategy has been adapted to form manganese oxide (MnO) NPs by using manganese precursor salts⁵⁷. Furthermore, methods such as hot injection and thermal decomposition can be used to synthesize conventional QDs with tunable fluorescence emission spectra and high quantum yields following a similar procedure. For example, a solution of cadmium and selenium precursor salts is quickly injected into hot trioctylphosphine oxide solvent to fabricate CdSe QDs with low size dispersity⁵⁸. However, these syntheses are very sensitive to the slightest changes in reaction conditions, which may affect the quality and reproducibility of the resulting products. They also require the use of inert gas and complex equipment. Therefore, there has been a shift towards implementing syntheses based on green chemistry to provide a straightforward production process and a sustainable alternative to traditional organic syntheses⁵⁹. Most of the reaction solvents (for example, water and ionic liquids), reductants (such as vitamin C) and capping ligands (for example, biomolecules and plant extract) used are water-soluble, naturally occurring and safe^60,61. For instance, IONPs can be made directly by mixing iron salts in the presence of dextran, a natural polysaccharide, and subsequently raising the pH via the addition of ammonia⁶². Bismuth–iron oxide nanohybrids and cerium oxide NPs can also be synthesized following similar co-precipitation procedures^63,64. Another example of green synthesis involves the use of natural templates, such as DNA aptamers and viral proteins, to facilitate the assembly of inorganic nanocrystals⁶⁵.

Lipid-based nanomaterials.

Lipids are amphiphilic molecules that contain a hydrophilic head group and a hydrophobic tail group. Depending on the relative sizes of these groups, lipids can self-assemble into a liposomal bilayer surrounding an aqueous core or a micellar monolayer (or an emulsion) surrounding a hydrophobic core. A standard protocol for forming liposomes comprises mixing of two or more types of phospholipids that usually possess two hydrocarbon chains, such as distearoylphosphatidylcholine and 1,2-distearoylsn-glycero-3-phosphoethanolamine-PEG (DSPE-PEG), with cholesterol (to promote formulation stability) in chloroform, thin film formation, followed by sonication and solubilization in an aqueous medium⁶⁶. By contrast, micelles are formed using a proportion of phospholipids with a larger head group than its hydrophobic component. There are several preparation methods for micelles, including oil-in-water emulsion, solvent evaporation and lyophilization¹³. Passive encapsulation of and lipidic labelling with contrast-generating substances are the common methodologies for formulating liposomal, emulsion and micellar NP contrast agents^67,68. Examples include the encapsulation of hydrophobic payloads, such as hydrocarbon-coated nanocrystals, fluorophores or perfluorocarbons (liquid or gas), as well as use of gadolinium chelate or fluorophore-modified lipids. A few formulations, such as liposomes, require post-preparation downsizing to decrease size and dispersity. Techniques such as extrusion, ultrasonication, field-flow fractionation and centrifugation are available for reducing NPs to smaller size ranges.

Biologically derived structures, such as lipoproteins and cell membranes, are naturally occurring nanomaterials that can be repurposed as contrast delivery platforms⁶⁹. There are different types of lipoproteins, such as high-density lipoproteins (HDLs) and low-density lipoproteins (LDLs), but they all have a similar basic structure — a hydrophobic core of triglycerides and cholesterol that is covered with an apolipoprotein containing phospholipid layer^70,71. Lipoproteins can therefore be considered a form of micelle or nanoemulsion. In general, lipoproteins are extracted from blood serum via ultracentrifugation with increasing densities of potassium bromide solution. HDLs and LDLs are generally in the size range of 7–13 and 22–27 nm, respectively. Moreover, contrast-generating materials can be loaded in the hydrophobic core of lipoproteins⁷², for instance, by replacing triglycerides and cholesterol ester with nanocrystals. Contrast-generating materials can also be included in the lipid coating by attaching contrasting elements to the protein constituent or by incorporating contrast-labelled amphiphiles in the formulation⁷³.

An alternative approach for the assembly of lipid-like NPs is in vivo self-assembly. In a recent example, monomers containing an FDA-approved drug, olsalazine, were dimerized in situ via a reaction catalysed by the enzyme furin, which is overexpressed in colon tumours⁷⁴. The dimer aggregates into NPs, which can be detected via CEST or Raman imaging, and has anti-cancer effects, all of which arises from the olsalazine component⁷⁵.

Polymeric nanoparticles.

Polymeric nanomaterial-based contrast agents are typically composed of natural polymers (for example, dextran and chitosan) and synthetic polymers (for example, polyethy-lenimine and poly(lactic-co-glycolic acid)), which are further divided into multiple subclasses, such as polymersomes and dendrimers. Conventional polymeric NPs are typically synthesized via covalent conjugation, cross-linking or physical encapsulation⁷⁶. Single or multiple contrast agents can be encapsulated within the carrier matrix to enable multimodality imaging. Furthermore, some polymer matrices may provide intrinsic contrast, such as semiconducting polymers with PA and NIR fluorescence imaging capabilities^77,78. Such polymers can self-assemble into NPs by interacting via the hydrophobic effect. Finally, a microfluidic chip could be used to crosslink water-soluble polymers with hydrophilic contrast cargoes via ionic interactions⁷⁹.

Dendrimers (for example, poly-l-lysine and polyamidoamine (PAMAM)) are highly branched, tree-like nanostructures that result from a sequence of reaction steps using divergent or convergent methods⁸⁰. Each step leads to an additional generation of branching as well as an increase in the number of available surface functional groups. Therefore, generation 5 PAMAM dendrimers display more primary amino groups compared with previous generations, allowing for incorporation of targeting ligands in addition to contrast agents⁸¹. For instance, generation 5 PAMAM dendrimers were used to entrap 2 nm AuNPs inside PEG cavities as well as attach Fluo-4 via covalent conjugation for labelling T cells and thereby imaging and sensing their location and activity⁸².

Nanoparticle modification and conjugation

For use in biological settings, toxic surfactants and hydrophobic ligands need to be displaced to render colloidal nanocrystals hydrophilic with good stability and biocompatibility. Approaches to this include direct ligand exchange, embedding in a carrier matrix (for example, silica and polymer) or coating with amphiphilic molecules (for example, micelles and vesicles) (Fig. 2b). For example, citrate ions act as a capping agent for gold nanocrystals and are typically substituted with thiolated compounds, such as thiol-PEG or glutathione, through direct incubation owing to the strong affinity of sulfur with gold⁸³. However, post-synthesis modification via transfer to the aqueous phase can be quite cumbersome and may affect certain NP characteristics (such as pharmacokinetics) that are highly dependent on the surface structure.

Surface functionalization with targeting moieties, cargoes and chelators can be achieved using several established methods^84–86 (Fig. 2c). One commonly utilized strategy is the N-ethyl-N’-(3-dimethylaminopropyl)carbodiimide/N-hydroxysuccinimide-mediated carboxylate–amine coupling. More recently, click chemistry via copper-catalysed azide–alkyne cycloaddition is a preferred bioconjugation method⁸⁷. Click chemistry is most suited for the conjugation of targeting ligands to NPs as it forms highly oriented linkages and ensures site-specific binding. However, as most biomolecules and NP surfaces do not normally contain azide and alkyne groups, pre-activation is required to carry out the reaction, which can increase the complexity of the work. Bioconjugation can also be achieved by using an intermediate crosslinker, such as bissulfosuccinimidyl suberate and sulfosuccinimidyl 4-(N-maleimidomethyl)cyclohexane-1-carboxylate, to facilitate linkage between two complementary or different functional groups⁸⁸. In addition, the specific and stable interaction between streptavidin and biotin, which is less likely to alter or interfere with the biomolecule activity, can be used. Nevertheless, several factors, such as cost, effectiveness, NP stability and bioactivity change associated with each method, should be taken into consideration when choosing a bioconjugation strategy that is most suitable for the NP–biomolecule pair.

Nanoparticle purification

The high density of many NPs allows unreacted molecules or unconjugated moieties to be easily removed via repeated centrifugation, thus producing a highly concentrated and purified product. Density gradient centrifugation via potassium bromide or sucrose is used when the density difference between the product and by-products is low, such as LDL loaded with 3 nm AuNPs⁸⁹. Another purification method involves the use of molecular weight cut-off (MWCO) concentrator tubes, which separate impurities and NPs into different compartments via centrifugation based on the relative sizes. In a related approach, MWCO dialysis membranes or ultrafiltration columns are used together with a large volume of fresh buffer to allow free ions to diffuse into the reservoir over time. The NP products are retained by the MWCO filters given the appropriate pore size (for example, 10 kDa MWCO for 15 nm AuNPs). Finally, NPs can be precipitated using an incompatible or insoluble solvent and washed using a solvent in which the impurities are soluble (Fig. 2d).

Results

Nanomaterial designs can vary widely in size, shape and composition, among other physicochemical properties. It is generally accepted that these factors strongly influence the contrast and bio-interactions of nano-based agents. In this section, analytical characterization methods used to identify the physical and chemical features of various nanomaterial types are discussed⁹⁰. Although each technique has its own advantages and use cases, a combination of multiple techniques is essential to provide conclusive nanomaterial characterization, as well as information related to cellular specificity and in vivo fate and performance of nano-based contrast agents. Imaging approaches that can be used to assess the behaviour of the agent in vitro and in vivo are also described. Finally, how these properties define the suitability of an NP for different imaging applications and alter its interactions at the nano–bio interface is examined.

Physicochemical characterization

Size and morphology.

In nanomedicine research, the results section typically begins with NP characterization data. Size and morphology are key characteristics that can affect contrast generation, pharmacokinetics, biodistribution and other agent properties. Synthesis and characterization often occur in a feedback loop, in which results from characterization are used to inform changes to synthetic conditions in an effort to generate NPs that meet design criteria or have high performance. Micrographs from transmission electron microscopy (TEM) and scanning electron microscopy (SEM) show the NP core (Fig. 3a) and surface morphology, respectively⁹¹. NP shape or geometry can be easily visualized by both techniques. High-resolution TEM is sometimes performed to characterize the lattice fringes of nanocrystals. Freeze-fracture TEM can be used to study the structure of NPs made of soft materials such as lipids or proteins. NP surface topography can be assessed by atomic force microscopy⁹². Moreover, dynamic light scattering (DLS) is the main method to determine the hydrodynamic diameter (D_h) and size distribution of NPs in suspension⁹¹ (Fig. 3b), although it should be noted that DLS typically assumes spherical morphology; therefore, it is less informative for non-spherical agents^93,94. A homogeneous NP size distribution contributes to consistent in vivo behaviour; therefore, the NP size range, or the polydispersity index (PDI), should be noted. A PDI value of <0.1 typically indicates a properly monodisperse NP population, whereas larger PDI values are acceptable under certain circumstances, for example, NPs produced from complex multistep syntheses. NP tracking analysis (NTA) and fluorescence correlation spectroscopy are also useful tools for measuring the size and number of NPs in solution, respectively^94–96. For certain formulations, the porosity of NPs is examined via focused ion beam tomography in conjunction with electron microscopy⁹⁷.

Fig. 3 |. Examples of physical characterization, imaging assessments and biological interactions of nanomaterial-based contrast agents.

a, Transmission electron microscopy (TEM) assesses the nanoparticle (NP) core diameter. b, Dynamic light scattering (DLS) measures the NP hydrodynamic diameter. c, Energy-dispersive X-ray spectroscopy (EDX) identifies the elemental composition of NP. d, Example of a computed tomography (CT) phantom (inset) and attenuation curve derived from the corresponding CT phantom images. e, NPs based on gadolinium (Gd) and iron (Fe) are commonly used as T₁-weighted and T₂-weighted contrast agents for magnetic resonance imaging (MRI), respectively. Yellow arrow indicates brightening of a brain metastasis due to Gd via T₁-weighted MRI. Orange arrows indicate darkening of mammary gland tumours due to Fe via T₂-weighted MRI. f, In vitro intracellular reactive oxygen species (ROS) generation is commonly investigated by the 2′,7′-dichlorofluorescein diacetate (DCFH-DA) assay. Data are presented as mean ± s.d. ****P < 0.0001. g, Subcellular resolution multiphoton microscopy (MPM) evaluates cellular internalization of the NP. h, Scanning electron microscopy (SEM) visualizes the interactions between the NP and cell surface. i, In vivo biodistribution (percentage of injected dose per gram of tissue, %ID g⁻¹) is easily determined using positron emission tomography (PET) imaging with radiolabelled NPs. j, Serum biomarkers from blood samples indicate potential NP toxicity and organ damages, such as alanine transaminase (ALT) and blood urea nitrogen (BUN) for liver and kidney functions, respectively. AST, aspartate transferase; CeONP, ceriumoxide nanoparticles; HepG2, hepatocellular carcinoma; NS, not significant; Renca, renal cell carcinoma; RFU, relative fluorescence units. Parts a and b reprinted from ref. 91, Springer Nature Limited. Parts c and f reprinted with permission from ref. 5. Copyright 2022 American Chemical Society. Part d reprinted from ref. 110, Springer Nature Limited. Part e reprinted with permission from ref. 112, Wiley. Part g reprinted from ref. 123, Springer Nature Limited. Part h reprinted from ref. 124, Springer Nature Limited. Part i reprinted from ref. 125, CC BY 4.0. Part j reprinted with permission from ref. 64. Copyright 2021 American Chemical Society.

Chemical composition.

The elemental composition of NPs can be assessed using several techniques such as inductively coupled plasma (ICP) spectroscopy, energy-dispersive X-ray spectroscopy, X-ray diffraction and X-ray photoelectron spectroscopy. ICP is frequently used to measure the concentration of metal elements in samples, which is particularly relevant given that metals are commonly present in NP contrast agents. Energy-dispersive X-ray spectroscopy identifies the presence of certain elements in a sample by analysing the emitted characteristic X-rays⁵ (Fig. 3c). X-ray diffraction examines the crystallinity and lattice planes of dried NPs by recording the diffraction patterns of scattered X-rays. X-ray photoelectron spectroscopy operates on the basis of the principle of photoelectric effect to determine the core and surface elemental composition, as well as the binding or oxidation states of the elements⁹⁸. Other methods can be used, depending on the NP composition. For formulations that contain phosphate groups, for example, phospholipid-based NPs, colorimetric assays can be used to assess the amount of inorganic phosphate and thus to determine the final phospholipid inclusion⁹⁹. The protein content of NPs can be measured by the Bradford assay, for example. Such methods are important for understanding the degree of conjugation success or composition-contrast generation relationships.

Surface chemistry.

Characterization of NP surface chemistry is important as the surface has direct exposure to biological components, which can affect the overall NP stability, toxicity and functionality. The net surface charge of NPs is presented as zeta potential, which can vary widely with the solvent type and pH of the solution. High surface charge increases repulsive forces between NPs, thus improving the colloidal stability and promoting interactions with oppositely charged molecules of interest¹⁰⁰. Fourier transform infrared spectroscopy is the main technique for identifying functional groups or surface ligands and for detecting changes in surface properties. Nuclear magnetic resonance (NMR) can also be used for this purpose; however, various effects from the attachment of ligands to NP surfaces (for example, superparamagnetism of IONPs or decreased ligand rotation from surface immobilization) can render NMR data hard to obtain or interpret. Nevertheless, for some types of NPs, NMR can yield key insights into surface chemistry⁵⁰. Thermogravimetric analysis can be used to quantify the density of targeting moieties, or the number of organic ligands present on the NP surface.

Stability.

Stability should also be assessed for weeks to months after synthesis, depending on the storage conditions and suspension fluids, to determine the shelf-life and ideal time frame (absence of aggregation) for performing subsequent studies. Nano-based contrast agents are incubated in simulated biological fluids, such as phosphate-buffered saline with 10% fetal bovine serum, to assess their stability under physiological conditions. NTA, DLS, UV–Vis measurements and asymmetrical field-flow fractionation are used to evaluate changes owing to agglomeration and formation of protein corona¹⁰¹. Formulation stability can also be examined via the release of payloads or metal ions owing to degradation or leaching. This can be assessed using ICP and the aforementioned analytical techniques.

Contrast properties

NPs can be used as contrast agents based either on their inherent contrast properties or on functionalization with contrast-generating moieties. To characterize the contrast properties, a series of increasing concentrations of the contrast agent together with relevant controls, such as water, saline, air and comparator agents (for example, FDA or EMA-approved agents such as Gd-chelates for MRI and iodinated molecules for CT), are inserted into a phantom, which is a device composed of tissue-equivalent materials and can be scanned to evaluate the imaging performance of agents. The qualities desired for these phantoms vary by scanner type and their designs range from simple to complex (phantoms used for CT include a bucket of water to 3D-printed chest mimics), with the goal that more complex phantoms will be a better predictor of in vivo contrast generation, result in better image quality or otherwise provide advantages^102–105. The phantom is then scanned with the desired imaging instrument using a set (or multiple sets) of acquisition parameters. Contrast properties of the NP formulation of interest should be measured within the linear portion of the standard curve. A broad range of concentrations should be tested to determine the optimal dose for in vivo administration as NPs are expected to accumulate in various organs at differing concentrations. The doses needed for in vivo studies may be estimated from previous, similar experiments, although some trial and error (such as dose optimization) may be needed for any new agent. Depending on the imaging application, specialized devices may be used to characterize signal output in vitro. Finally, quantification of in vivo contrast enhancement is generally performed via advanced image analysis software, such as ImageJ and MATLAB. Regions of interest are drawn over certain features (such as tumours), and results are typically presented as contrast-to-noise ratio or tissue-to-background ratio. Artificial intelligence methods for both image reconstruction and image analysis have drawn considerable attention recently and may have a larger role in the future^106,107.

X-ray imaging.

CT phantoms can be constructed simply by securing vials of NP samples in plastic racks and submerging in water to simulate beam hardening, or the rise in average energy of an X-ray beam as it is attenuated by the body¹⁰⁸. Anthropomorphic phantoms can also be utilized, such as chest phantoms that mimic the human organs in the thorax, in terms of thickness, density and mass attenuation¹⁰⁵. Furthermore, phantoms for mammography are fabricated using varying compositions of glandular and adipose tissue equivalent materials¹⁰⁹. These phantoms are scanned using a range of X-ray tube voltages, beam filters, current, slice thickness, exposure time and field of view. CT attenuation, as displayed in Hounsfield units (HU), increases linearly with NP concentration¹¹⁰ (Fig. 3d). The slope of the linear regression line, or attenuation rate (HU·ml·mg⁻¹), is then calculated to characterize CT contrast. Mammo-graphic contrast is shown as contrast-to-noise ratio or signal difference-to-noise ratio by measuring signal intensity from the sample (or tissue of interest) and background (for example, body of the phantom).

MRI.

MRI phantoms can be prepared by inserting NP samples in an agarose gel doped with manganese, nickel or other salts^63,111. Depending on the mode of detection, the phantom is scanned using T₁-weighted or T₂-weighted sequences at multiple echo times, repetition times or inversion times¹¹² (Fig. 3e). Signal intensity is measured at each echo time and inversion time for the determination of T₂ and T₁ relaxation times, respectively; r₁ and r₂ relaxivities can then be calculated through curve fitting of the inverse of T₁ and T₂ relaxation times as a function of NP concentration. Alternatively, relaxivity measurements can be performed with a benchtop device known as a relaxometer. For in vivo MR contrast measurement, pixel-wise curve fitting of the relaxation times can be used to generate colour-coded maps of R₁ and R₂ relaxation rates. The contrast generation of ¹⁹F MRI agents can similarly be assessed by measurements of r₁ and r₂ using NMR spectrometers or other systems¹¹³. Alternatively, samples can be scanned at a range of concentrations and the SNR analysed¹¹⁴. The contrast generated by CEST agents can be assessed using sets of MR pulse sequences to determine the ppm shift from water where the maximum chemical exchange is observed¹¹⁵. Water exchange rates are also determined.

Nuclear imaging.

PET radioisotopes are often produced in a cyclotron using nuclear reaction processes, with a specific activity at the end of bombardment. Radionuclides can be attached to the NP surface or interior using chelator-based or chelator-free labelling methods¹¹⁶. Radiolabelled NPs can usually be purified by size exclusion column chromatography with phosphate-buffered saline as the mobile phase. Radiolabelling yield can be determined by direct radioactivity measurement or by using an autoradiograph of a thin layer chromatography plate with EDTA as the mobile phase¹¹⁷. Radiolabelling stability is usually evaluated after incubation with serum at various time points. Before animal studies, radioactivity of a sample is assessed using a well counter to calculate the injection dose needed for adequate detection in vivo. Regions of interest are drawn on PET images to measure biodistribution in vivo, whereas the major organs and tissues are often collected to quantify biodistribution ex vivo using a γ-counter¹¹⁸. Finally, owing to the high detection sensitivity of nuclear imaging, low labelling densities are required in comparison to the high labelling required for MRI-active NPs, for example.

Fluorescence imaging.

The signal generation of fluorescent NPs should be evaluated in tissue-mimicking phantoms with defined light scattering properties to nullify issues related to fluorescence (self) quenching and light absorption in biological tissues. Assessments can be done with the scanner intended for in vivo use, as well as fluorometers. Combinations of excitation and emission wavelength pairs should be tested to determine optimal imaging parameters. Fluorescence intensity usually exhibits a non-linear relationship with NP concentration, so a broad range of concentrations should be tested to identify the detection saturation threshold. The effects of target depth can be assessed by including differing thicknesses of tissue-mimicking materials in the phantom between the samples and the scanner optics.

Photoacoustic and ultrasound imaging.

PA phantoms can be created by mounting tubes of NP samples into a plastic holder that is immersed in deionized water free of air bubbles¹¹⁹. More complex phantom designs can be created from moulded plastics, for example, in which depth penetration can be explored with phantoms of varying thicknesses¹²⁰. The transducer is placed above the phantom (in contact with the water or using ultrasound gel to provide contact with plastic phantoms), and PA signals are acquired using an NIR excitation wavelength with a defined PA gain. PA spectra (intensity versus wavelength) are collected to characterize NP signal generation within the NIR range and to determine optimal imaging conditions. Integrated pixel density of the diseased region before and after injection can be quantified to show PA contrast enhancement. Ultrasound agent assessment can be done with similar physical set-ups to PAs. Samples are tested over a range of concentrations and frequencies. Dedicated sample testing systems can be used, as well as preclinical or clinical systems¹²¹.

Biological characterization

In vitro studies.

To examine cytotoxic effects, primary cells and/or cell lines of interest are treated with NPs at a range of concentrations, and cell viability is analysed using various assays, such as MTS and LIVE/DEAD staining. Two-dimensional cell cultures are commonly reported for this purpose; however, more sophisticated in vitro models, such as organoids and organ-on-a-chip, more closely mimic disease pathophysiology and local microenvironments¹²². This allows for better prediction of in vivo NP cytotoxicity and imaging and treatment efficacy. More detailed examination of the effects of NP contrast agents on cells in vitro can be performed to better understand their potential toxicities. For example, NP surface reactivity can be investigated by the ferric reduction of serum assay or intracellular oxidative stress (Fig. 3f) assays, such as DCFH-DA and TBARS⁵. Should toxicities arise, such assays can inform on methods that can be used to reformulate to avoid toxicity, such as changing surface ligand or polymer molecular weight.

For targeted imaging applications, it is important to evaluate the affinity and specificity of targeted NP contrast agents for their intended targets. Receptor binding assays should be conducted to demonstrate the ligand specificity. Cells are pre-treated with free variants of the targeting ligand (or receptor blocker), followed by targeted nano-based contrast agents, a method known as competitive inhibition. Additional methods to test cell-specific NP targeting include incubating the cells of interest with NPs that have varying amounts of the targeting ligand at low and high temperatures to inform on uptake mechanisms, as well as under static and flow conditions to provide insights on which tissues will have greater uptake. Assessment of cellular uptake of nanomaterials could be performed via techniques such as fluorescence microscopy (Fig. 3g), SEM and TEM (Fig. 3h), flow cytometry or contrast quantification (for example, γ-counting or CT attenuation of cell pellets)^123,124. NPs can be taken up by the cell via different endocytic pathways. Blocking of these pathways using inhibitors or cells with genetic knockouts can reveal the contribution of each endocytic pathway in the uptake process and determine the ability of NPs to undergo endosomal escape.

Preclinical in vivo studies.

Appropriate animal models, in terms of species, age and gender, should be selected for each disease application⁹⁰. A power analysis is then performed to calculate the number of animals required for each cohort. Many imaging studies use on average five to seven animals per treatment group. For most applications, intravenous injection is the standard route for nanomaterial administration. Depending on the specific imaging purpose, other injection methods can be used, such as interstitial and orotracheal administration for mapping of sentinel lymph node and lung activity, respectively. Furthermore, NPs can be monitored at defined time points post injection (minutes to weeks) to estimate their circulation time and tissue uptake. Nuclear imaging (using radiolabelled NPs) is the best technique for evaluating pharmacokinetic and biodistribution profiles owing to its high detection sensitivity (Fig. 3i) and ability to longitudinally image any depth in the patient¹²⁵. The elimination route of NPs can be easily determined by analysing faecal (via hepatobiliary system) and urine (via renal system) samples after injection. The amount of NPs found at the target site should be presented as percentage of injected dose per gram of tissue (%ID g⁻¹) to normalize data reporting for more direct comparison.

Nano-based contrast agents should be thoroughly assessed for their in vivo biocompatibility. Typically, multiples of the intended dose of NPs are injected in healthy cohorts, and the animals are monitored regularly over a long period of time. Blood samples are drawn and analysed to detect biomarkers that are indicative of potential inflammatory responses as well as damages to liver (ALT and AST) and kidney (BUN) functions⁶⁴ (Fig. 3j). Routine welfare checks are performed to identify changes to the overall physical and mental health (for example, body weight, behaviour and signs of weakness) of the animals. Upon euthanization, major organs are collected and subjected to extensive histological analyses to detect any phenotypic changes.

Applications

Nano-based contrast agents have been demonstrated to visualize various pathologies and gain insight into underlying disease mechanisms through structural and functional imaging. Some multimodal NP diagnostics have been developed to possess additional therapeutic functions, which are also known as theranostics. In this section, representative applications of nano-based contrast agents in several key areas of medical imaging research are presented, and nano-based contrast agents that are approved or currently under clinical evaluation are discussed.

X-ray nanomaterials

Delineation of the vasculature has become a major application of CT¹²⁶. These agents are typically large (D_h > 10 nm) with bulky surface ligands to bypass renal clearance⁸. The most frequently reported CT blood pool agents are PEGylated AuNPs and iodine-loaded liposomes or nanoemulsions^127–129. Moreover, iohexol, an approved iodinated molecule, was crosslinked with a homopolymer to improve blood circulation and to enhance tumour contrast owing to passive accumulation via the EPR effect¹³⁰. Similar findings have been reported for breast cancer imaging with DEM using PEGylated gold–silver alloy NPs, PEGylated silver telluride NPs (PEG-Ag₂Te) and micelle-encapsulated Ag₂S NPs^131–133. Notably, PEG-Ag₂Te has been found to produce stronger X-ray contrast compared with elemental silver NPs owing to the inclusion of tellurium. In addition to imaging the blood pool and tumours, CT in conjunction with metal (such as gold and cerium) NPs has increasingly been utilized for detecting inflammatory diseases, such as colitis and abdominal aortic aneurysms^23,134. Several studies have also shown the feasibility of quantifying inflammatory macrophage content in atherosclerotic plaques via spectral CT (capable of resolving individual photon and classifying them into energy bins to produce quantitative elemental maps) with gold cores in PEG coating or HDL^135,136 (Fig. 4).

Fig. 4 |. Spectral computed tomographic imaging of atherosclerotic plaque composition with gold-loaded high-density lipoprotein.

a, Schematic illustration of macrophage-targeted nanoparticle computed tomographic (CT) contrast agent based on Au-high-density lipoprotein (Au-HDL). b, Au-HDL on a negative-stain micrograph from transmission electron microscopy. c, Conventional CT image of thorax and abdomen in a mouse with atherosclerosis after injection of Au-HDL. d, Spectral CT image showing signal from gold and accumulation of Au-HDL in the arteries of the mouse. e, Overlay of conventional and gold images. Reprinted with permission from ref. 136, Radiological Society of North America.

With rapid development in cell-based therapies, the topic of cell tracking with CT has garnered considerable interest, although high contrast payloads are needed to overcome the low sensitivity of CT, such as >10 pg per cell^1,137. Many nano-based CT contrast agents have been engineered to facilitate uptake by various cell types, such as stem cells and immune cells^138–140. For example, it was demonstrated that CT contrast is linearly proportional to the number of AuNP-labelled mesenchymal stem cells, highlighting the quantitative nature of CT imaging¹⁴¹. Furthermore, AuNPs with various sizes and coatings were examined with CT scans for optimal uptake by monocytes. It was found that 50 and 75 nm AuNPs with carboxylic acid ligands provided the most internalization¹⁴². Note that inherent CT contrast generation is dependent on elemental mass concentration but independent of NP size⁵².

Several high Z elements, specifically those that make potent CT agents, can induce DNA damage and cell death through radio-sensitization. For example, RGD-targeted AuNPs can augment radiotherapy by delivering radiation dose preferentially to endothelial cells of tumour neovessels¹⁴³. PEGylated bismuth NPs can also induce antitumour effects by allowing CT image-guided radiotherapy and photothermal therapy (PTT)¹⁴⁴. Finally, advances are being made to develop novel CT agents that generate the most optimal contrast, such as ytterbium and tantalum NPs, which can be further utilized for multiplexed imaging with spectral CT^{110,145–147}.

Magnetic nanomaterials

Most magnetic nanomaterials reported to date are based on gadolinium or iron^148–150. Gadolinium is paramagnetic and such agents provide excellent T₁ contrast and are frequently based on chelated gadolinium ions. For instance, HDL NPs were labelled with gadolinium chelates and conjugated with a collagen targeting peptide to monitor the compositional changes in atherosclerotic plaque regression via MRI¹⁵¹. Interestingly, poly-l-lysine-coated NaGdF₄ nanodots were synthesized, resulting in a dual-modality contrast agent that is capable of anatomical and functional imaging¹⁵². These nanodots possess high longitudinal relaxivity for T₁-weighted MRI and also exhibit an excellent CEST effect for pH mapping of brain tumours. However, free gadolinium can deposit in the brain with unknown long-term effects and can also induce nephrotoxicity in patients with renal disease. Other T₁ agents, such as MnO NPs, have been developed with the hypothesis that they will offer better patient safety^153,154. For example, hollow MnO NPs with expanded water-accessible surface areas were used for transplant tracking and drug delivery in addition to enhanced T₁ MRI^155,156. Such manganese-based NPs can yield T₂ or dual-mode T₁/T₂ contrast through appropriate size control¹⁵⁷.

IONPs are superparamagnetic and are the most studied T₂-shortening contrast agent^158,159. They have been used for MR detection of cancer and atherosclerosis, imaging of circulating cells and tracking of labelled stem and immune cells^160–162. In the clinical setting, dextran-coated ultrasmall superparamagnetic IONPs have been used to target intraplaque macrophages to evaluate the efficacy of novel cardio-vascular therapeutics¹⁶³. As a preclinical example, the M13 filamentous bacteriophage template with a glycoprotein targeting peptide was used to assemble IONPs for improved MRI of prostate cancer¹⁶⁴. In another study, two complementary IONPs (6 nm) were designed to self-assemble into larger nanoclusters (800 nm) and boost T₂ relaxivity after undergoing a biorthogonal reaction when cleaved by tumour enzymes¹⁶⁵. Notably, IONPs can transition into a T₁ contrast agent by decreasing their size. T₂ effects can also be increased by doping IONPs with transition metals, such as manganese, or by forming alloys with other metals (such as FeCo)¹⁶⁶. In addition to their MRI applications, IONPs can be utilized to track stem cells via MPI, an emerging technique that exploits nonlinear magnetization of NPs to generate positive contrast images¹⁶⁷. Notably, FeCo NPs have recently been shown to improve the detection sensitivity of MPI owing to higher magnetic saturation¹⁶⁸.

Radiolabelled nanomaterials

Nanomaterials have traditionally been labelled with radionuclides through coordination chemistry using exogenous chelators. DOTA, NOTA (or their derivatives) and deferoxamine (DFO) are the commonly used chelating agents for complexation of ⁶⁴Cu and ⁸⁹Zr radioisotopes, respectively¹⁶⁹. These chelators are conjugated to the NP surface to facilitate disease detection and cell tracking via PET^170–173. For instance, polymeric dextran NPs or HDL NPs were labelled with DFO for ⁸⁹Zr chelation and were used to detect resident macrophages in assessing atherosclerotic plaque inflammation via PET imaging^174,175. Moreover, selenium-doped carbon QDs with reactive oxygen species (ROS) scavenging properties were labelled with ⁸⁹Zr via DFO and demonstrated superior efficacy in the treatment of acute kidney injury¹²⁵. Their uptake in the kidneys was quantitatively measured by PET imaging. Similar chelation strategies have been implemented for radiolabelling nanomaterials for SPECT¹⁷⁶. For example, ^99mTc-labelled nanostructured lipid carriers were found to specifically target brown fat tissue via SPECT, which may prove beneficial for imaging metabolic disorders in humans¹⁷⁷. Interestingly, NOTA, which has better stability compared with DOTA¹⁷⁸, can chelate ⁶⁴Cu and be further conjugated onto RGD-functionalized superparamagnetic IONPs for targeted imaging of tumour neovessels via PET and MRI¹⁷⁹. Antibody-conjugated mesoporous silica NPs and unimolecular micelles that carry chemotherapeutics can also be radiolabelled in a similar fashion to enable targeted tumour imaging and drug delivery^180–182.

Chelator-free, intrinsic radiolabelling has recently been extensively investigated as an alternative to traditional chelation techniques. Several synthetic methods, such as hot-plus-cold precursors, specific trapping and cation exchange, can be used to acquire intrinsically labelled nanomaterials¹⁸³. The hot-plus-cold precursors approach, such as integrating ⁶⁴Cu into CuS NP cores, has led to a multifunctional agent with PET imaging and PTT capabilities^184,185. Other radioactive metal ions, such as ⁸⁹Zr and ⁶⁹Ge, have displayed strong binding affinity towards metal oxide NPs with high labelling yield and good serum stability^186,187. For instance, ⁶⁹Ge-labelled superparamagnetic IONPs were synthesized by simple mixing and used for lymph node mapping via PET and MRI¹⁸⁸ (Fig. 5). In another example, ⁶⁴Cu²⁺ ions were adsorbed onto MoS₂ nanosheets without chelating molecules owing to the affinity between copper and sulfur at molybdenum defect sites^189,190. This post-synthesis labelling approach yielded a theranostic nanoplatform that provides PET and PAI-guided tumour ablation.

Fig. 5 |. Superparamagnetic iron oxide nanoparticles labelled with ⁶⁹Ge for lymph node mapping via positron emission tomography and MRI.

a, Schematic illustration of chelator-free synthesis of ⁶⁹Ge-iron oxide nanoparticles (IONPs). b, Micrograph of IONPs from transmission electron microscopy. c, Lymph node (white arrow) imaging with positron emission tomography after injection of ⁶⁹Ge-IONPs into left paw (red arrow) of a mouse. d, Lymph node (green circle) mapping with MRI before and after injection of ⁶⁹Ge-IONPs into left paw of a mouse. Contralateral lymph node is indicated by red circle. %ID g⁻¹, percentage of injected dose per gram of tissue; PEG, polyethylene glycol. Reprinted with permission from ref. 188, Wiley.

Optical nanomaterials

Metal-based semiconducting QDs with narrow band gaps are perhaps the most sought-after nanomaterials for fluorescence imaging applications^191,192. By modulating the reaction time and solvent, the core diameter of Ag₂S QDs was tuned from 1.5 to 9 nm, which corresponded to an emission wavelength from 500 to 1,200 nm, allowing for detection in the NIR-I and NIR-II regions¹⁹³. Furthermore, RGD-conjugated, NIR-emitting, mercury QDs with a membrane impermeable etchant can be quenched in excess via cation exchange, thus achieving tumour-specific fluorescence imaging with high sensitivity¹⁹⁴. However, many QDs contain heavy metals, such as mercury or cadmium, which are strongly toxic at low concentrations and may therefore cause safety issues that hinder their clinical translation.

Carbon QDs, an organic nanomaterial with minimal toxicity, can be synthesized to display semiconductor band-gap-like fluorescence via surface passivation-induced defects¹⁹⁵. For instance, carbon QDs with surface doping of zinc sulfide have similar quantum yield compared with well-established CdSe QDs¹⁹⁵. Other carbon-based nanomaterials, such as graphene, nanodiamonds and nanotubes, possess exceptional optical properties in the NIR-II window, where relatively deep tissues (for example, brain microcapillaries and lung stem cells) can be visualized^196–198. NIR-II fluorescence imaging is also useful in detecting brain vasculature and blood flow with fluorophore-conjugated polymeric NPs and rare-earth downconverting nanocrystals^199,200.

Gold nanomaterials also have use in fluorescence and other optical imaging applications. NIR fluorescent gold nanoclusters (AuNCs) were encapsulated in red blood cell membranes to improve blood circulation for enhanced tumour accumulation and contrast²⁰¹. Targeted kidney imaging has been achieved using AuNCs-loaded ferritin nanocages²⁰². Moreover, surface plasmon resonance (SPR) in metallic nanostructures (gold, silver and copper) can induce SERS effects for Raman imaging^203,204. For example, colon cancer detection was performed using silica-coated SERS AuNPs with a dedicated Raman-based endoscope²⁰⁵. SERS-enhanced gold nanostars were synthesized to support surgical resection with Raman-assisted guidance to ensure complete tumour removal²⁰⁶. In addition, the amount of metallic nanomaterials needed for clinical SERS imaging can be greatly reduced by embedding additional Raman dyes in the surface coating²⁰⁷. It is worth noting that, unlike any other biomedical imaging modalities, nano-based SERS imaging is capable of profiling molecular targets with highly multiplexed capabilities²⁰⁸.

Photoacoustic nanomaterials

Gold nanostructures are a class of contrast agents with appealing PAI properties, owing to their tunable SPR peaks via size and geometric modifications^209,210. For instance, sub-5-nm AuNPs were encapsulated in larger biodegradable polymeric nanospheres, which provided significant PA contrast enhancement in vivo owing to a red shift in their absorption from interparticle plasmon coupling⁷⁹. AuNPs-loaded ROS responsive polymeric nanospheres provided differential PA contrast (signal loss) in ROS-rich microenvironments, such as inflammation and cancer²¹¹. The presence of ultrasmall AuNPs makes them a dual-modality PA and CT contrast agent with the potential for swift renal elimination upon degradation. Polymeric NPs can also carry and deliver therapeutic drugs via PAI guidance²¹². In addition, AuNPs conjugated with epidermal growth factor receptor (EGFR) antibodies and coated with NIR-active molecules allowed for targeted tumour imaging with bimodal PAI and SERS imaging²¹³. Furthermore, a new class of PA contrast agents based on NIR-absorbing semiconducting polymers has been introduced^214,215. A study showed that ROS levels can be measured by quantifying PA signals from ROS-sensitive dye-decorated semiconducting polymeric NPs at different input laser wavelengths⁷⁷.

Other nanomaterials apart from AuNPs also make potent PA contrast agents²¹⁶. AuNRs have been extensively applied in PAI owing to tunable transverse and longitudinal SPR peaks that can shift into the NIR region¹¹⁸. It was reported that miniature AuNRs can significantly enhance PA contrast in tumour compared with larger AuNRs, indicating a size effect on PA signal generation²¹⁷. PAI can also be used for tracking and quantifying mesenchymal stem cell in rodent muscle tissue with silica-coated AuNRs²¹⁸. Similarly, Ag₂S NPs (Fig. 6) and TiS₂ nanosheets exhibit intense NIR light absorption and are excellent for theranostic applications through PAI (and NIR fluorescence imaging) and PTT^219,220. Finally, an unconventional class of PAI agents, which is based on upconversion nanomaterials, has recently emerged²²¹. NaYF₄ upconverting NPs doped with lanthanide ions and complexed with α-cyclodextrin showed remarkable PA signal enhancement in vivo owing to luminescence quenching from non-radiative relaxation.

Fig. 6 |. Ag₂S nanoparticles with theranostic functionality for tumour imaging and treatment.

a, Schematic illustration of synthesis of serum albumin-coated Ag₂S nanoparticles. b,c, In vivo tumour detection via near-infrared-II fluorescence imaging (part b) and photoacoustic imaging (part c). d, Infrared thermography of tumours showing elevated temperature when irradiated with a near-infrared laser. NDs, nanodots; NHS, N-hydroxysuccinimide; PBS, phosphate-buffered saline. Adapted with permission from ref. 220. Copyright 2017 American Chemical Society.

Reproducibility and data deposition

The condition sensitivity of nanomaterial synthesis and nano–bio interactions can affect data reliability and reproducibility, leading to variability in experimental findings. Moreover, inconsistent results at the preclinical stages may present challenges for clinical translation. To address such issues, the nanomedicine community has proposed to adopt the minimum information reporting in bio–nano experimental literature (MIRIBEL) guidelines for standardizing data reporting^222–224. This section describes best practices for producing high-quality, reproducible results and discusses potential repositories for accessing experimental data related to nano-based contrast agents research.

For some nanomaterial types, the synthetic process can be quite extensive and may involve multiple reagents and complicated preparation protocols. To minimize batch-to-batch variations, it is important to provide detailed synthesis and purification procedures and note any sensitive or challenging operations²²⁵. Photographs showing the experimental set-up, each successive reaction step and appearance of the final products could be included to allow for more accurate synthesis replication. Reagent information such as manufacturer, lot number, expiration date, storage requirement, stability and purity could be listed, as well as quality acceptance criteria²²⁶.

Rigorous reporting of NP characterization data is crucial for improving data reproducibility. According to the MIRIBEL guidelines, physicochemical properties including the size, morphology, dispersity and zeta potential should be reported as a minimum. Readouts on three or more individually produced NP batches should be included to inform the robustness of the synthesis and reproducibility of the product. Measurements should also be conducted in biologically relevant media as changes may occur owing to aggregation. More than two independent characterization methods should be utilized to fully understand the material properties of an NP contrast agent with a detailed description of how a parameter is assessed. This includes sample preparation steps, controls, equipment information, instrument calibration procedures, acquisition protocols and analysis software version. It is essential to provide a distribution (or a range) of the measured parameter instead of a single average value. NP contrast properties should be investigated using well-constructed, contrast-embedded phantoms, and details of phantom design and image reconstruction and analysis should be provided. For in vivo imaging experiments, information detailing the animal strain, disease model, administration route, injected dose and volume, injection rate, NP suspension buffer and image processing methods (for example, thresholding) should be documented. Finally, researchers should make use of self-checklists, validated assays and standard operating procedures to ensure accurate and reproducible outcomes.

As recommended by the MIRIBEL guidelines, it would be beneficial if raw data and experimental reports were provided together with manuscript submission or deposited in online open-access repositories. Funding bodies such as the NIH are placing a greater emphasis on scientific data sharing. However, the development of large-scale, trusted data repositories for nanomaterial and medical imaging research is still at its inception. Thus, researchers should take advantage of currently available data sharing platforms such as GitHub and preprints (for example, ChemRxiv) to promote early data exchange and to facilitate timely communication. Further experimental details, results and statistical analyses should be included in the supplementary information to improve transparency. Finally, dedicated journals that outline specific synthetic protocols, characterization techniques and imaging methods may ensure thorough evaluation of published studies and establish successful standardization practices²²⁷.

Limitations and optimizations

Even though nano-based agents have great potential to provide rich imaging contrast, certain design elements should be considered and further optimized before their clinical implementation. This includes ensuring reproducible and cost-effective NP synthesis in large scale, maximizing their imaging sensitivity, resolving biocompatibility and toxicity concerns and enabling effective delivery to target sites.

Fabrication

As a rule of thumb, simpler synthetic methods generate particles with a wider particle size distribution, which can result in batch-to-batch inconsistencies and may require secondary size selection steps. Fortunately, recent advancements in analytical capabilities, such as TEM and NTA, allow for deeper understanding of nanomaterial properties and easier optimization of synthetic protocols. For inorganic NP fabrication, a two-step process is often involved: NP creation followed by stabilization to prevent aggregation and improve biocompatibility. However, regulatory agencies will typically favour easier one-step aqueous processes, such as the fabrication of liposomes, which are biodegradable and have already been FDA-approved. Dendrimers, in comparison, require a costly multistep process. Preclinical studies investigating new nanomaterials are often not systematically designed or optimized for clinical translation, and most academic laboratories are not equipped with certified good manufacturing practice capacities or the ability to scale up to meet clinical needs. In addition, few commercial manufacturers have previous experience with the synthesis and scale up for biological applications of many of the types of nanomaterials that are used as contrast agents. As a result, manufacturing and characterization complexities of nano-based imaging agents can lead to poor product reproducibility and reliability²²⁸. Prioritizing commercialization challenges, assessing manufacturing complexities, considering available scale-up resources and defining rigorous measures for quality control at an early stage of development can increase overall potential for success in clinical translation. To avoid risks of overengineering and fabrication complexity for FDA or EMA approval, careful consideration should be made early on towards pairing the right NP type with the appropriate clinical application.

Imaging performance

Medical imaging modalities, such as PET, CT, MRI and PAI, that provide anatomical information and visualize cellular and biological processes have potential to transform patient care^205,229,230. Each of these imaging techniques has its own unique advantages. However, no single imaging modality currently meets all the criteria of high sensitivity, high specificity, high spatial and temporal resolution, high multiplexing capacity, deep tissue penetration, low cost and high throughput. Therefore, considerable efforts have been made to develop multimodal imaging approaches that aim to circumvent the inherent limitations of each modality. Nanomaterials are attractive vehicles for multimodal imaging as they may inherently generate more than one type of contrast or can easily be loaded with multiple forms of contrast in a single platform^{131,231–238}. However, a challenge for such multimodal agents is generating detectable contrast for each technique with the same NP dose.

Theranostic NP agents may result in better patient outcomes via simultaneous disease detection and treatment planning. For example, owing to high X-ray attenuation, high light absorption and high plasmon resonance effect, certain designs of AuNPs produce strong CT, PA and Raman contrast, as well as enhance the efficacy of radiation therapy and phototherapy^{79,208,239,240}. IONPs induce negative contrast, which may be difficult to identify with certainty. For instance, if the target is expressed in very low concentrations and tissues produce inhomogeneous MR signals, it may not be possible to distinguish the contrast arising from IONPs from the background. Image acquisition sequences that produce positive contrast from IONPs can mitigate this difficulty. As each imaging modality can offer varying ways to quantify image signal intensity, researchers should take care when comparing and interpreting their data across the field, particularly when optimizing nano-based contrast agents for improved detection sensitivity.

Toxicity

To achieve eventual clinical approval, nano-based contrast agents are subjected to thorough assessments of their long-term biodistribution, degradability, clearance and toxicity. Lipid-based and silica-based NPs have shown to be biodegradable in the body, but there are concerns about the biocompatibility and stability of many inorganic NPs, with a few exceptions such as iron oxides and manganese oxides, which are generally regarded as reasonably biodegradable and biocompatible^241,242. QDs often contain heavy metals, such as cadmium and lead, that are strongly toxic at low concentrations²⁴³. Thus, QDs that have biocompatible coatings or are free of heavy metals are more likely be considered for eventual clinical translation^244,245. Moreover, AuNPs are chemically inert and non-toxic, but their long-term effects on health are still uncertain, if a portion of the dose is retained in the body²⁴⁶. Recently, atomic AuNCs, which are ultrasmall clusters of 10–100 gold atoms (having a total diameter of <2 nm), have emerged as promising agents in cancer diagnosis and treatment. Owing to their small size, these AuNCs result in fast renal clearance with minimum retention in the MPS organs, such as liver and spleen²⁴⁷. As a result, renally clearable and biodegradable NP formulations are preferred for in vivo imaging applications to avoid any long-term toxicity concern. Although non-biodegradable NPs are being evaluated for their safety in vivo, they can still offer valuable information in both in vitro and ex vivo imaging studies. Nonetheless, the physicochemical characteristics of non-biodegradable NPs, such as composition and surface modifications, should be further investigated to understand the issue of toxicity effects and concerns related to their excretion profiles before they can be implemented for use in patient care. Current efforts should therefore be focused on developing NP structures that subsequently break down into nontoxic materials and harmless by-products for efficient elimination.

Delivery

One of the biggest hurdles for clinical translation of any nano-based contrast agent is their effective delivery in a quantity that enables sensitive detection and localization²⁴⁸. Large amounts of NP agent are often required at the target site, especially when using less-sensitive imaging modalities, such as CT²⁴⁹. In particular, the EPR effect is a powerful delivery mechanism that allows NPs to passively accumulate in diseased tissues with sustained angiogenesis and a permeable vasculature (such as tumours, inflammation and atherosclerosis), although the magnitude of the EPR effect is variable depending on the individual²⁵⁰. Retention of these NPs at diseased sites has also been demonstrated as a result of their uptake by local pathologically associated macrophages²⁵¹. It is essential that the NPs have long enough blood circulation times and stealthy coatings to effectively reach their target and evade uptake by other organs. Enhanced imaging specificity can be achieved by conjugating the NP surface with targeting moieties that have a high degree of molecular specificity. There are also other factors to take into consideration when optimizing for targeted NP delivery. For example, antibodies are highly specific towards their targets but can be expensive, and their attachment may increase the overall NP size by as much as 10 nm, thereby changing the pharmacological characteristics of the NP. Addition of antibodies and other biological linkages to an NP can result in further regulatory testing and increased developmental cost. Some linkages, such as biotin-streptavidin, may generate immunologic responses in patients. In such cases, ligands can be covalently attached to NPs using various methods, such as novel protein Z-mediated expressed protein ligation (EPL)-click reaction, that do not elicit an immune response in humans, but may still require additional regulatory testing²⁵².

Commercial considerations

Most of the current NPs that are FDA-approved or in clinical trials involve therapeutics, whereas those for imaging agents are fewer^253,254. This disparity may, in part, be due to the higher market potential for developing new therapies as opposed to new imaging agents. Unfortunately, the earning potential of diagnostic imaging contrast agents is much lower than that of therapeutic agents. The recently approved nano-based COVID vaccines achieved sales over US$50 billion. The market clearly favours the therapeutic route²⁵⁴. It is a challenge to justify the costs to develop nano-based imaging agents when the return on the investment is low. Fortunately, there are several new and exciting opportunities for imaging to offer important insights to enable better therapeutic drug discovery and help guide therapeutic decisions. This opens up an entirely new market for the development of nano-based imaging contrast agents to offer rich molecular information about the disease of an individual patient, enabling a more effective treatment response and justifying the cost for clinical translation. Additionally, with a new generation of NP therapeutics likely to be forthcoming, spurred by the resounding success of the NP COVID vaccines, the clinical translation of new NP imaging contrast agents may be driven in part by the need for companion diagnostics (companion diagnostics are commonly required for targeted radiotherapeutics to determine whether the disease of the patient has the target and therefore whether the radiotherapeutic is likely to be effective).

Outlook

Many preclinical studies have supported the use of nanomaterials as imaging agents given their tunable physicochemical properties and desired contrast enhancement. From the mid-1990s to mid-2000s, several IONP MRI contrast agents designed to diagnose and characterize focal liver lesions or liver metastasis were successfully translated into the clinic²⁵⁵. These NPs have been shown to outperform their small-molecule counterparts, such as ¹⁸F-FDG and gadolinium chelates, especially in hepatic cancer imaging; however, some formulations were withdrawn owing to low sales. Several gold nanomaterials, such as AuroShells, have also been investigated in clinical trials as cancer therapeutics²⁵⁶. These advances are considered a recent major milestone for the field, given that only a handful of NP formulations has been approved as diagnostics to date (Table 2). However, translational activity has since been low, partly owing to poor economic outlook for diagnostic agents and current market emphasis on nano-therapeutics. Another hindrance for clinical translation is that there is a lower tolerance for side effects for contrast agents than for therapeutics, which raises the bar for safety. Although contrast agents have lower standards for efficacy than therapeutics, simply binding a target, as opposed to increasing survival compared with current therapies, may render approval more likely.

Table 2 |.

Summary of several nanomaterial-based contrast agents that are approved or trialled in the clinic

Imaging mode	Candidate	Approved trade name or ClinicalTrials.gov identifier	Phase
PET/SPECT	99mTc sulfur colloids	Technecoll (USA)	Approved
PET/SPECT	125I albumin colloids	Jeanatope (USA)	Approved
PET/SPECT	99mTc SnF2 colloids	Hepatate (France)	Approved
PET/SPECT	99mTc Re2O7 colloids	Nanocis (EU)	Approved
PET/SPECT	89Zr cRGDY silica	NCT03465618	1
PET/SPECT	Gd liposome	NCT05453539	1
MRI	Iron oxides	Feridex (USA)	Approved
MRI	Polysiloxane (AGuIX)	NCT04789486	1
Fluorescence	cRGDY silica (C dots)	NCT02106598	2
Fluorescence	pH-sensitive micelle (ONM-100)	NCT05048082	2
X-ray CT	Hafnium oxide (NBTXR3)	NCT04892173	3
X-ray CT	Ethiodized oil	Lipiodol (USA)	Approved

CT, computed tomography; MRI, magnetic resonance imaging; PET, positron emission tomography; SPECT, single photon emission computed tomography.

With the recent surge in targeted immunotherapies, clinicians are eager for newly developed imaging and monitoring tools, including novel nanomaterial-based contrast agents and relevant detection techniques, to help predict patient response and personalize disease therapy on an individual basis. Non-invasive imaging strategies have the added benefit of offering longitudinal information to better determine the effectiveness of a given treatment over time. The ability of NPs to actively target tumour areas can help differentiate inflammation from tumour progression, thus providing a more reliable way to evaluate for treatment response. Furthermore, surgery is usually the first line of treatment for most patients with cancer diagnosed at stages I to III. Incomplete tumour resection often results in negative patient outcomes and decreased survival, as well as increased health-care costs owing to additional surgical interventions. Nano-based contrast agents have been shown to improve tumour margin delineation and assist in surgical resection through intraoperative, real-time image guidance with greater sensitivity and specificity. Notably, a recent phase II clinical study has shown that ONM-100 (or pegsitacianine), a pH-sensitive, micellar fluorescence agent, could enhance the ability of a surgeon to detect residual malignancies following cytoreductive surgery of peritoneal metastases⁸⁵. With these encouraging results, more research interest and funding support for this agent type in the coming years may be expected. Overall, intraoperative image guidance may considerably decrease the frequency of repeat surgeries and markedly reduce patient expenses, while justifying further development of these nano-based contrast agents with the possibility to overcome their market challenges.

In summary, nano-based imaging contrast agents have great potential to vastly improve disease detection and patient outcomes if further developed²⁵⁷. The field would benefit from greater interactions and collaboration among material scientists, academic researchers, radiologists, industries and regulatory agencies to guide the advancement of nano-based contrast agents, meet clinical demands and satisfy regulatory requirements. Nano-based imaging strategies can offer so much more than disease diagnosis. New studies have demonstrated their ability to present crucial insights to direct drug development and provide predictive information to guide therapeutic decisions. With this in mind, it is imperative to continue to promote the development of new nano-based contrast agents that can offer physicians invaluable structural and functional information about the disease progression within an individual.

Glossary

Brust method

A commonly used method to synthesize gold nanoparticles in non-aqueous solutions

Contrast agents

Substances administered in medical imaging procedures to facilitate disease diagnosis via altering image contrast

Enhanced permeation and retention (EPR) effect

The phenomenon of elevated retention of nanomaterials in certain diseased tissues (for example, tumours and sites of inflammation) owing to leaky vessels and poor lymphatic drainage

Hot-spot imaging

Imaging of certain tracers, such as fluorinated nanoparticles with ¹⁹F-magnetic resonance imaging, allows for specific detection in vivo without endogenous background signal

Nano-based contrast agents

Contrast agents that are within the nano-size range, typically 1–400 nm

Phantom

A device composed of tissue-equivalent materials that is used to evaluate imaging performance of contrast agents

Quantum confinement effect

When a critical threshold size (2–10 nm) is reached, nanomaterials present tunable fluorescence properties reflecting small differences in particle size

T₁ agents

Contrast agents for magnetic resonance imaging that are used to shorten T₁ relaxation. They typically involve gadolinium or manganese

T₂-shortening agents

Contrast agents for magnetic resonance imaging that are used to shorten T₂ relaxation. They are most frequently iron oxide nanoparticles