Iron Oxide Nanoparticles: An Alternative for Positive Contrast in Magnetic Resonance Imaging

Iron oxide nanoparticles have been extensively utilised as negative (T2) contrast agents in magnetic resonance imaging. In the past few years, researchers have also exploited their application as positive (T1) contrast agents to overcome the limitation of traditional Gd3+ contrast agents. To provide T1 contrast, these particles must present certain physicochemical properties with control over the size, morphology and surface of the particles. In this review, we summarise the reported T1 iron oxide nanoparticles and critically revise their properties, synthetic protocols and application, not only in MRI but also in multimodal imaging. In addition, we briefly summarise the most important nanoparticulate Gd and Mn agents to evaluate whether T1 iron oxide nanoparticles can reach Gd/Mn contrast capabilities.


Introduction
Iron oxide nanoparticles (IONPs) are one of the most used nanomaterials in biomedicine. Among the reasons justifying this interest, their biocompatibility and magnetic properties are probably the most important. These properties have boosted their use in hyperthermia cancer treatment and, as imaging probes, in magnetic resonance imaging (MRI). When IONPs are prepared using "traditional" synthetic methods they show superparamagnetic properties. In other words, these nanoparticles show a very strong magnetic response when placed under the influence of a magnetic field, turning to zero when the magnetic field is off. Because of this, when placed inside MRI equipment, IONPs act as "small magnets", suppressing the signal and, therefore, appearing as a dark spot, the so-called negative contrast. Due to the strong magnetic response, the concentration needed for an in vivo application is often low. However, based on this, IONPs have been, for a long time, the never fulfilled eternal promise to change the current clinical scenario in MRI. Currently, Gd-based compounds are the standard probes when an MRI scan is performed. It is well-known that, under certain circumstances, Gd compounds show important toxicity. This is particularly important for patients suffering from kidney problems. Besides the toxicity problems, Gd probes normally have a small molecular weight and are, after injection, rapidly extravasated, excluding them from many applications that require long circulating times. If Gd-based probes present these problems, why have IONPs not displaced them from clinical practice? Basically, because the signal provided by Gd compounds is much more useful for in vivo diagnosis than that provided by IONPs for many diseases. This is due to the dark, negative signal that traditional IONPs generate. Frequently, in many diseases, hypointense (dark) areas appear naturally in an MR image. If the image probe generates a dark signal over a dark background, diagnosis gets complicated. For this reason, in recent years, researchers have searched for an alternative that can join the good physicochemical properties of IONPs with the outstanding imaging properties of Gd compounds. This has led to numerous synthetic developments producing extremely small iron oxide nanoparticles that, being more paramagnetic than superpramagmetic, are capable of generating bright, positive contrast in MRI. Here, we will critically review these developments, highlighting achievements and considering what is left to accomplish to reach a point at which the use of IONPs in clinics is as frequent as the use of gadolinium compounds.

Spin Density and Relaxation Times
Magnetic resonance imaging (MRI) uses differences in spin density or relaxation properties (T 1 , longitudinal or spin lattice relaxation time, T 2 , transversal or spin-spin relaxation time, T 2 *, apparent transversal relaxation time) to generate signal-to-noise ratio and contrast between different soft tissues in an image. Apart from these relaxation times that are interrogating different features of molecular dynamics and physical mechanisms, each soft tissue (liver, brain, heart, lung, etc.) will have a different chemical composition and different spin densities to encode an image using MRI. Water or fat are the usual predominant content of tissues; therefore, spin densities of tissues are often given by the concentration of water weighted by the relaxation properties and the acquisition parameters and type of sequence used. T 1 is normally much longer than T 2 and does not exhibit proportionality among them. Table 1 shows typical relaxation time values for different tissues at normal magnetic fields for clinical applications.

T 1 -Weighted or Positive Contrast Using Gradient and Spin Echo Sequences
For T 1 -contrast, imaging is performed by emphasising the differences in longitudinal components of magnetisation. This information can be coded with typical spin-echo or gradient-echo sequences. In the spin-echo sequences, two radiofrequency (RF) pulses (one at 90 • followed by another at 180 • or refocusing pulse) are used to acquire the signal. T 1 -weighted imaging in a spin-echo sequence is determined by the repetition time (i.e., the temporal distance between two consecutive 90-degree pulses, also called T R ) whilst echo times (named T E hereinafter) are kept at the lowest possible value. The spin-echo sequences have the value of being immune to off-resonance artefacts caused by B 0 inhomogeneities and to magnetic susceptibility shifts due to heterogenous tissues (such as multiple air-tissue interfaces in lung tissue or in some brain or abdominal regions) or to the presence of magnetic impurities (such as the presence of iron oxide nanoparticles).
In contrast to this alternative, a gradient-echo [1] image is formed with a single pulse (flip angle (θ) generally inferior to 90 • ), combined with the application of two lobes, also called dephasing and rephasing gradients, and the absence of a 180 • refocusing pulse, per each RF excitation. Differently to the spin-echo methods, local susceptibilities are not refocused in this approach, so image quality is normally inferior and signal reduction is anticipated at gradient-echo sequences.
For our purpose here, to help to understand the contents of this review, for T 1 -contrast or positive contrast with nanoparticles, keeping a short T E in the pulse sequence should not be a problem for most MRI systems. Longer T E enables a T 2 * weighting contrast, which for certain nanoparticles can still be recommendable. However, for positive contrast or T 1 -contrast, T E should be kept as short as possible. The selection of the best θ to maximise the T 1 contrast will then be critical for final optimisation. This flip angle finally enables T 1 weighting and must be investigated to allow fast imaging (short T R ) with this possibility. To do so, we need to take into consideration the expected T 1 and T 2 values of the different tissues for the magnetic field used in the study (see Table 1). As we can see from this table, T 1 values generally will increase with B 0 , whereas T 2 remain constant, so the properties as contrast agent are normally very different at low (<1.5 T) and high fields (>3 T).

Nanoparticles for Positive Contrast MRI
Despite the high resolution of anatomical features presented by MRI, sensitivity is one of the weak points. MRI relies on the differences in tissue proton density and therefore differences in tissues' relaxation times to generate contrast. These differences allow discrimination between bone, air, and soft tissues in vivo. Nevertheless, discrimination of certain tissues and diseased areas gets complicated when these differences do not generate enough contrast. For this reason, contrast agents are regularly used to facilitate the diagnosis and characterisation of pathologies at cellular and molecular levels [2]. T 1 or positive contrast agents shorten the longitudinal relaxation time in the areas or tissues where accumulation occurs, making brighter images. The majority of clinically available T 1 contrast agents are paramagnetic metallo-chelates, composed of Gd, whose potential toxicity and short circulating times have driven the quest for improved positive contrast agents that overcome these limitations [3].
Nanostructured materials exhibit unique properties by virtue of their size. At this scale, quantum effects dominate material behaviour, conferring the size-dependent magnetic, electrical and optical behavior of nanomaterials. At the nanoscale, materials present an enhanced surface-to-volume ratio, extremely useful for bioconjugation purposes and targeted imaging. There is a broad variety of nanoparticulate T 1 contrast agents mainly based on the incorporation of paramagnetic Gd or Mn ( Table 2). Some of the most remarkable are summarised below.

Paramagnetic Gd 2 O 3 Nanoparticles
Gadolinium-based contrast agents are the gold standard T 1 contrast agents for MRI; however, their toxicity and reduced number of applications boosted the quest for alternatives. Most of Gd-based T 1 contrast agents are based on organic molecules chelating Gd 3+ ions to prevent toxicity. However, inorganic nanoparticles made up of Gd are increasingly common. Gadolinium oxide [4,5], gadolinium fluoride [6,7] and gadolinium phosphate [8] NPs have been synthesised and tried as T 1 contrast agents.
Most recent examples are focused on targeted T 1 Gd 2 O 3 NPs [9], dual T 1 -T 2 MRI probes [10], hybrid MRI/fluorescent probes [11,12] and theranostic (theraphy + diagnostic) probes [13] for tumour imaging. A Gd 2 O 3 -NP hybrid CT/MRI probe functionalised with bisphosphonate was used by Mastrogiacomo et al. to visualise calcium phosphate bone cement [14]. Dai et al. carried out a comparison between their PEGylated-Gd 2 O 3 nanoparticles and the commercially available Magnevist, observing that their NPs presented a long half-life in blood and efficient MRI contrast, lower hepatic and renal toxicity and greater accumulation at the tumour site [15].

Paramagnetic MnO Nanoparticles
Manganese ions have emerged as a potential alternative to gadolinium as T 1 enhancer; however, their toxicity, affecting the central nervous and the cardiovascular systems, has determined its clinical application [16,17]. Manganese oxide nanoparticles have emerged as the most suitable alternative to overcome these toxicity issues [18]. Size and shape control allow fine tuning of the relaxivity values of these nanoparticles. PEGylated MnO NPs have been demonstrated by several groups to be useful and non-toxic T 1 contrast agents [19,20]. Their use for tumour detection is fairly popular, both in a non-specific and specific manner. Wang et al. achieved MnO NP accumulation in gliomas by elongating the circulation time of their MnO NPs, functionalising their surface with cysteine [21]. Chen et al. also managed to visualise mouse gliomas using MnO NPs, however using a glioma-specific moiety: folic acid [22]. Gallo et al. achieved M21 tumour visualisation in mice using RGD (arginine-glycine-aspartate peptide)-functionalised MnO nanoparticles [23]. Renal carcinoma T 1 MR imaging was accomplished by Li et al., targeting MnO NPs using AS1411 aptamer [24]. Manganese oxide NPs have also been used as theranostic platforms for drug delivery and photothermal therapy at the tumour site [25][26][27]. MnO NPs have served as a platform to synthesise hybrid molecular imaging probes for both PET (Positron Emission Tomography)/MRI for tumour vasculature imaging using 64 Cu [28] and fluorescence/MRI with Cy7.5 for lymph node mapping [29].

Organic Nanostructured Materials
Organic nanostructured materials have a long history as contrast agents for MRI. However, their use has been limited, as the relaxivity values of most of them, such as albumin-and dextran-based MR probes, are often insufficiently high.
Dendrimers are polymeric molecules with monomers branching out radially from a central core, forming a tree-like architecture [30]. As their synthesis is stepwise, resultant structures present narrow polydispersity, and terminal groups on the surface of the dendrimer can be included in a controlled manner [31]. The most commonly used monomers are polyamidoamine (PAMAM), polypropylimine (PPI), poly(ether imine) (PETIM) and poly-l-lysine (PLL) [32]. Due to plentiful anchoring sites for paramagnetic ions in their structure, they are appropriate nanoplatforms to integrate paramagnetic ions in their structure. Gadolinium (Gd 3+ ) is one of the most repeatedly used ions in dendrimers for T 1 MRI. The macromolecular size of these ions increases the rotational correlation times of integrated Gd 3+ , resulting in relaxivities larger than most of the clinically approved Gd-based contrast agents with low molecular weight. The use of these macromolecular probes has been demonstrated in lymphatic imaging [33], tumour detection [34], liver fibrosis staging [35], and colon cancer and brain theranostics [36]. Mn 2+ -based compounds have been also proposed as non-toxic alternatives to Gd chelates to image hepatocellular carcinoma and atherosclerosis [37,38]. Fan et al. have recently shown that Cu 2+ can be integrated in dendrimeric structures to form a platform for tumour/metastasis imaging and chemotherapy [39].

Silica Based Nanoparticles
One of the main applications for SiO 2 nanoparticles is drug delivery that, combined with the possibility of carrying paramagnetic ions in their pores or surface, yields excellent theranostic agents. Kim et al. used Mn-doped silica NPs to detect hepatocellular carcinoma [40]. Li et al. made use of a theranostic agent based on mesoporous manganese silica NPs loaded with doxorubicin to image and treat a breast cancer xenograft murine model [41]. Gd 3+ -containing silica NPs have also been used in vivo and/or in vitro as a T 1 contrast agent [42][43][44]. Recently, Carniato et al. summarised the most remarkable examples of Gd-based mesoporous silica nanoparticles for MRI. In this work, the influence over the relaxometric properties of important factors such as the porosity, the localisation of the paramagnetic chelate and the surface properties of the mesoporous silica nanoparticles are discussed in detail [45]. For instance, Davis et al. reported mesoporous silica nanoparticles doped with Gd in the inner or outer part of the pore and in the surface of the particle showing great differences in the relaxivity values depending on the Gd localisation [46].

Liposomes
Liposomes are spherical structures formed by one or several concentric lipid bilayers with an aqueous phase inside [64]. Due to their amphiphilic composition, they can integrate hydrophobic and/or hydrophilic molecules. This, added to their outstanding biocompatibility, has boosted their use as nanocarriers in drug delivery and molecular imaging. There are numerous examples of liposome-based MRI probes carrying paramagnetic agents, mostly composed of Gd 3+ and Mn 2+ in the aqueous lumen [47,[65][66][67][68]. Alternatively, some other liposomal contrast agents used in MRI carry the paramagnetic molecule in their lipid bilayer [48,[69][70][71][72]. Recent applications of these probes include imaging of tumours [49][50][51]73], atherosclerotic plaque [52] and blood brain barrier permeability [74].

Iron Oxide Nanoparticles for MRI
Iron oxide nanoparticles have been mostly used as negative (T 2 ) contrast agent. Their superparamagnetic behaviour, driven by the magnetic mono-domains at nanometric scale in the appropriate iron oxide phase, along with the great saturation magnetisation values, provide excellent T 2 shortening in MRI images [75]. The limitations of T 2 -driven diagnosis has stimulated the development of iron oxide-based T 1 agents [76]. IONPs act as T 1 agent when certain physicochemical properties are fulfilled. These features, along with the synthetic method to achieve them, are summarised below.

Physicochemical Properties
IONP components, the core and coating, play an important role in the contrast behaviour. The composition, size and shape of both components must be controlled since they determine the iron oxide phase, crystallinity, magnetic properties and the hydrodynamic size. All of these properties are key to develop an IONP-based T 1 contrast agent.
In terms of size, we have to consider both the size of the core and the hydrodynamic size (core + coating). As a rule of thumb, in IONPs, the smaller the core the better the positive contrast. When the core size is decreased (<5 nm) the magnetic single domains decrease, leading to a spin canting effect, which provides five oriented and unpaired Fe 3+ /Fe 2+ electrons, making the particle more paramagnetic than superparamagnetic ( Figure 1). Smaller cores usually imply less crystallinity, with a subsequent decrease in the saturation magnetisation values (similar to paramagnetic materials). Therefore, maghemite (γ-Fe 2 O 3 ) is often preferred for T 1 contrast rather than magnetite (Fe 3 O 4 ) where the crystallinity is usually higher. Iron oxide nanoparticles have been mostly used as negative (T2) contrast agent. Their superparamagnetic behaviour, driven by the magnetic mono-domains at nanometric scale in the appropriate iron oxide phase, along with the great saturation magnetisation values, provide excellent T2 shortening in MRI images [75]. The limitations of T2-driven diagnosis has stimulated the development of iron oxide-based T1 agents [76]. IONPs act as T1 agent when certain physicochemical properties are fulfilled. These features, along with the synthetic method to achieve them, are summarised below.

Physicochemical Properties
IONP components, the core and coating, play an important role in the contrast behaviour. The composition, size and shape of both components must be controlled since they determine the iron oxide phase, crystallinity, magnetic properties and the hydrodynamic size. All of these properties are key to develop an IONP-based T1 contrast agent.
In terms of size, we have to consider both the size of the core and the hydrodynamic size (core + coating). As a rule of thumb, in IONPs, the smaller the core the better the positive contrast. When the core size is decreased (<5 nm) the magnetic single domains decrease, leading to a spin canting effect, which provides five oriented and unpaired Fe 3+ /Fe 2+ electrons, making the particle more paramagnetic than superparamagnetic ( Figure 1). Smaller cores usually imply less crystallinity, with a subsequent decrease in the saturation magnetisation values (similar to paramagnetic materials). Therefore, maghemite (γ-Fe2O3) is often preferred for T1 contrast rather than magnetite (Fe3O4) where the crystallinity is usually higher. Core size and composition are essential but so is an appropriate coating. The stability of the nanoparticles is crucial to avoid aggregation, which triggers multiple particles to act as a single magnetic domain and, hence, increases the T 2 effect. In this regard, different coatings such as small molecules, macromolecules or proteins have served as stabilisers [77][78][79][80]. In addition, the size of the coating has been demonstrated to be a key factor in the T 1 effect [81]. Large size coatings (i.e., large hydrodynamic size, >30-40 nm) restrict water access to the nanoparticle core increasing the outer-sphere contribution to the relaxation mechanism. Coatings providing ultrasmall hydrodynamic sizes (<7 nm) usually confer poor stability, with aggregation increasing the T 2 effect. This is, however, an interesting challenge due to the renal clearance of ultrasmall IONPs, which increase their translational potential. Coatings rendering medium hydrodynamic sizes (10-20 nm) provide good colloidal stability with a thin coating that increases the water exchange rate, boosting the T 1 effect.
Although these are general considerations, there is no general rule established to describe the best-case scenario for each T 1 -IONP. An optimal T 1 -weighted sequence, taking into account the physicochemical and relaxometric properties of the nanoparticles and selecting the experimental parameters to highlight them appropriately for each magnetic field, is mandatory. A great number of works have shown a wide variety of formulations, sizes and coatings with the magnetic field, pulse sequences, and acquisition parameters not always optimised for T 1 -weighted imaging. Table 3 summarises some of the reported T 1 -IONPs, depicting these differences and the influence over the relaxometric properties. Table 3. Hydrodynamic diameter (D H , nm), core size (nm) and r 1 ,

Synthesis
Co-precipitation and thermal decomposition are the most frequently used methods to produce iron oxide nanoparticles. In addition, the extensive work on iron oxide nanoparticles over the last decade is bringing new procedures every year for the synthesis of nanoparticles of different sizes, shapes and composition and, hence, magnetic behaviour. In this section, the most remarkable methods are briefly described pointing out those which render nanoparticles with positive contrast capabilities.

Co-Precipitation
The co-precipitation protocol is the most used throughout the literature. It is based on the precipitation of iron oxides throughout a mixture of aqueous solutions of ferrous and ferric salts at a 2:1 ratio under basic conditions. The success of the co-precipitation method lies in its simplicity, flexibility and the hydrophilicity of the nanoparticles. However, it also shows some associated drawbacks such as the lack of control over the uniformity of the nanoparticles [103]. Modifications in the co-precipitation protocol can render ultrasmall iron oxide nanoparticles (USPIO) for T 1 -MRI. A high increase in the temperature of the reaction, combined with different polymers, has been used to develop nanoparticles between 3 and 8 nm with associated longitudinal relaxivities of up to 9 mM −1 s −1 at 4.7 T and 31 mM −1 s −1 at 1 T [90,104].

Thermal Decomposition
This well-known method, based on the decomposition of organic precursors at very high temperatures, has been widely utilised because of the high uniformity, crystallinity and control over the size of the nanoparticles. These nanoparticles are hydrophobic and therefore only stable in hydrophobic solvents, which implies an extra reaction step to stabilise the nanoparticles in physiological media. Despite this inconvenience, the control over the hydrodynamic size allows the obtention of USPIO with potential in positive contrast [87]. Wei et al. described the thermal decomposition of Fe(oleate) 3 in the presence of 1-tetradecene, 1-hexadecene and 1-octadecene, rendering nanoparticles from 2.5 to 7.0 nm of a maghemite core oxidised with trimethylamine N-oxide. The nanoparticles, stabilised in water using a ligand exchange reaction with a zwitterionic dopamine sulfonate (ZDS), showed a modest r 1 = 5.2 mM −1 s −1 at 1.5 T and 1.5 mM −1 s −1 at 7 T [78]. Another remarkable example describes the synthesis of USPIO by decomposition of Fe(acac) 3 at 300 • C in the presence of oleic acid, oleylamine, hexadecanediol and oleyl alcohol as solvent. Ligand exchange with different bisphosphonate-based ligands provided nanoparticles with a magnetic core of 3.6 nm and a maximum r 1 = 11 mM −1 s −1 at 1.5 T [98].

Polyol Synthesis
Polyol synthesis has been gaining attention in the recent years for the synthesis of IONPs. This method allows for an easy scaling-up of the reaction in a single-step reaction producing hydrophilic nanoparticles, although aggregation often happens. The method essentially consists of the reduction of the organometallic precursor in the presence of different polyols such as trimethylene glycol, propylene glycol or ethylene glycol. The role of the polyol has been reported not only as stabiliser but also as reducing agent [105].
Concerning IONPs with T 1 capabilities, the reduction of Fe(acac) 3 in diethylene glycol at 200 • C, under an inert atmosphere provided nanoparticles with 8 nm of hydrodynamic size and 3 nm of core size. These small particles exhibited r 1 = 12 mM −1 s −1 at 1.41 T with a low r 2 /r 1 = 2.4 [106]. Another reported example depicts a one-pot reaction of 5.4 nm IONPs by reduction of Fe(acac) 3 in the presence of triethylene glycol and HOOC-PEG-COOH at 260 • C. Under these conditions, the synthesised particles showed a r 1 = 19.7 mM −1 s −1 at 1.5 T with a r 2 /r 1 ratio of 2.0 [84].

Microwave Assisted Synthesis
Due to the simplicity, fast kinetics and reproducibility of the reactions, microwave-assisted synthesis of IONPs has grown lately. The use of microwaves ensures a very fast and homogeneous heating in the sample, which translates in a narrow size distribution of the nanoparticles [107]. Regarding the synthesis of IONPs for T 1 contrast, our group has been deeply involved in the use of microwaves. Essentially, extremally small iron oxide cores (~2.5 nm) can be achieved with FeCl 3 as iron source. The reaction is conducted at 100 • C for 10 min in the presence of hydrazine hydrate (reducing agent) and a surfactant, usually dextran or sodium citrate. In a first approach, a r 1 = 5.97 mM −1 s −1 at 1.5 T was obtained using (FITC)-dextran or dextran 6KDa as surfactant [108,109]. After these first approaches, better longitudinal relaxivities were observed using sodium citrate as surfactant. In this case, an increase in the temperature of the reaction turns over the contrast capabilities from T 2 to T 1 with an r 1 = 11.9 mM −1 s −1 at 1.5 T when the reaction is performed at 120 • C [81]. Very recently, Fernandez-Barahona et al. increased this value up to 15.7 mM −1 s −1 , doping the iron oxide core with 4% mol of Cu [99].

In Vivo Applications
The use of iron oxide nanoparticles as contrast agents for MRI has usually been associated with T 2 -MRI. This is one of the main reasons why the use of iron oxide-based contrast agents for MRI is far from standard in clinical practice. The dark (T 2 ) signal makes them difficult to distinguish from naturally occurring hypointense areas in many diseases-caused by calcium deposits, bleeding, other metals or any signal void present in the area of interest. Given this, the need for nontoxic positive contrast agents has led to intense research in the development of iron oxide nanoparticles that produce high T 1 -weighted MRI signals, becoming therefore, a very active topic.

Iron-Based and Iron Oxide Nanoparticles
In recent years, several research groups have been developing new iron-based probes, mostly for T 1 -MRI but also for dual contrast MRI. Regarding T 1 -MRI, some researchers have been able to show, in vivo, a positive contrast effect. In 2013 Ju et al. [110] elaborated non-toxic synthetic melanin-like nanoparticles complexed with paramagnetic Fe 3+ ions and stabilised with PEG (PEGylated Fe 3+ -MelNPs), which were inspired by the MRI signal-enhancing capability of natural melanin. These nanoparticles showed relaxivity values of r 1 = 17 mM −1 s −1 and r 2 =18 mM −1 s −1 at 3 T, which are higher than those of existing T 1 -MRI contrast agents based on gadolinium (Gd) or manganese (Mn). In vivo, PEGylated Fe 3+ -MelNPs showed a positive signal enhancement in the spleen and liver of healthy mice within 0.5 and 1.5 h, respectively, after intravenous injection. During the same year, Peng et al. [111] synthesised antiferromagnetic α-iron oxide-hydroxide (α-FeOOH) nanocolloids, with diameters of 2-3 nm, which were placed inside the mesopores of worm-like silica nanoparticles. These nanocomposites exhibited a low r 2 /r 1 ratio of 1.9, making them suitable as T 1 -weighted contrast agents. The in vivo experiments carried out showed a positive enhancement in the brain, bladder and kidneys of healthy mice. Moreover, in 2015, Iqbal et al. [85] produced biocompatible silica-coated superparamagnetic IONPs with diameters of~4 nm, which showed in vivo T 1 contrast enhancement in the heart, liver, kidney and bladder of healthy mice. In addition, Macher et al. [93] developed an innovative T 1 MRI contrast agent, ultrathin iron oxide nanowhiskers. These nanostructures, with dimensions of 2 × 20 nm, possessed a high surface-to-volume ratio, leading to a strong paramagnetic signal, a property suitable for T 1 contrast. In vivo experiments showed a positive contrast enhancement in rat models after intraperitoneal (IP) and subcutaneous injection of the nanowhiskers. In 2019, Tao et al. [77] synthesised (poly(acrylic acid)-poly(methacrylic acid) iron oxide nanoparticles (PMAA-PTTM-IONPs) 4.34 nm in diameter with a low r 2 /r 1 ratio of 2.78, adequate for T 1 -weighted contrast agents. In vivo experiments in mouse models showed biocompatibility and T 1 contrast enhancement in liver and kidney. Even though the previously described probes were shown to provide contrast enhancement during in vivo T 1 -MRI, no specific in vivo application was described for them; thus, only their biodistribution was tested. Nevertheless, different research groups have developed probes for particular applications such as magnetic resonance angiography (MRA) and tumour imaging.
MRA has been shown to be a very helpful technique in clinical imaging. With its use, several diseases could be detected including myocardial infarction, renal failure, atherosclerotic plaque, thrombosis and tumour angiogenesis. Regarding research carried out on MRA, in 2011 Kim et al. [87] synthesised ultrasmall iron oxide nanoparticles (ESION, size < 4 nm) capped using poly(ethylene glycol)-derivatised phosphine oxide (PO-PEG) ligands, which enabled clear observation of various blood vessels, with sizes down to 0.2 mm, during in vivo T 1 -MRI in rat models. They were able to maintain the bright signal of blood vessels for 1 h on dynamic time-resolved MR angiography, showing that this type of probe can be used for T 1 enhanced blood pool MRI (Figure 2).  Furthermore, in 2014 Chan et al. [83] produced multidentate block-copolymer-stabilised ultrasmall superparamagnetic iron oxide nanoparticles (MDBC-USPIOs), with diameters <5 nm and a r2/r1 ratio of 4.74 to test them as a promising T1-positive contrast agent for in vivo MRI. Results from in vivo MRI showed a strong blood signal enhancement after their intravenous injection in mouse models. During the same year, Liu et al. [91] developed glutathione-coated iron oxide nanoparticles (GSH-IONPs) with sizes of around 3.75 nm and a r2/r1 ratio of 2.28 as a novel T1-MRI contrast agent. The in vivo results showed a strong vascular enhancement at the carotid artery and superior sagittal sinus of healthy mice models, making it a promising contrast agent for thrombus detection.
In 2015, Bhavesh et al. [108] proposed an extremely fast microwave synthesis of fluoresceinlabelled dextran-coated extremely small IONPs for their use as a contrast agent for T1-weighted MRI. This method yielded very small NPs with hydrodynamic diameters of 21.5 nm and with a r2/r1 ratio of 4.7 suitable for T1-MRI. In vivo MRA of healthy mice showed a clear depiction of the main vascular architecture; furthermore, high quality visualisation of small vessels was maintained even 90 min post NP injection, highlighting the advantage of these NPs to be used as a contrast agent for blood pool imaging applications. In 2017, Pellico et al. [81] described the changes produced on the relaxometric properties of IONPs when the thickness of their organic coating is modified. For that purpose, this group synthesised all IONPs using a microwave-driven synthesis method; however, in Furthermore, in 2014 Chan et al. [83] produced multidentate block-copolymer-stabilised ultrasmall superparamagnetic iron oxide nanoparticles (MDBC-USPIOs), with diameters <5 nm and a r 2 /r 1 ratio of 4.74 to test them as a promising T 1 -positive contrast agent for in vivo MRI. Results from in vivo MRI showed a strong blood signal enhancement after their intravenous injection in mouse models. During the same year, Liu et al. [91] developed glutathione-coated iron oxide nanoparticles (GSH-IONPs) with sizes of around 3.75 nm and a r 2 /r 1 ratio of 2.28 as a novel T 1 -MRI contrast agent. The in vivo results showed a strong vascular enhancement at the carotid artery and superior sagittal sinus of healthy mice models, making it a promising contrast agent for thrombus detection.
In 2015, Bhavesh et al. [108] proposed an extremely fast microwave synthesis of fluorescein-labelled dextran-coated extremely small IONPs for their use as a contrast agent for T 1 -weighted MRI. This method yielded very small NPs with hydrodynamic diameters of 21.5 nm and with a r 2 /r 1 ratio of 4.7 suitable for T 1 -MRI. In vivo MRA of healthy mice showed a clear depiction of the main vascular architecture; furthermore, high quality visualisation of small vessels was maintained even 90 min post NP injection, highlighting the advantage of these NPs to be used as a contrast agent for blood pool imaging applications. In 2017, Pellico et al. [81] described the changes produced on the relaxometric properties of IONPs when the thickness of their organic coating is modified. For that purpose, this group synthesised all IONPs using a microwave-driven synthesis method; however, in order to change the coating layer thickness, different syntheses were carried out heating at different temperatures. Results showed that IONPs with a thinner coating yielded an excellent positive T 1 -MRI contrast, more specifically NPs synthesised at 120 • C. These NPs, with a hydrodynamic size of 7.9 nm and a r 2 /r 1 ratio of 1.9 showed great contrast enhancement in both body and brain MRA of healthy mice, providing the possibility of better visualising their vasculature. In 2018, Vangijzegem et al. [98] synthesised very small PEGylated iron oxide nanoparticles (VSIONPs) with core sizes of 3.5 nm that showed positive contrast in in vivo MRA, enabling the heart chamber and the vena cava of healthy mice to be observed.
Regarding research on T 1 -weighted MR tumour imaging, in 2014 Wu et al. [112] produced mesoporous silica NPs with drug-labelled USPIONs confined within the mesoporous matrix (Fe-MSNs) as a pH-responsive theranostic platform. The tumour accumulation, driven by the enhanced permeability and retention effect (EPR), was confirmed by the T 1 enhancement in the affected site. Moreover, the unique metal-ligand coordination bonding between Fe species and the anticancer drug molecules provided the carrier with a pH-responsive drug release feature, triggering a controlled drug release under the acidic microenvironment in the tumour area. Moreover, Shen et al. [94] in 2017 synthesised extremely small 3.6 nm magnetic iron oxide nanoparticles (ES-MIONs) functionalised with dimeric RGD peptide (RGD 2 ) and PEG methyl ether (mPEG), which were afterwards loaded with the anticancer drug doxorubicin hydrochloride (DOX), as a tumour targeting theranostic platform. In this probe, the ES-MIONs serve as the contrast agent for T 1 -MRI, while RGD 2 is used for tumour targeting and DOX for chemotherapy. In vivo experiments on tumour-bearing mice showed an enhanced T 1 -MRI signal when using this platform together with partial regression of tumours due to active targeting and chemotherapy. Furthermore, Li et al. [88] in 2019 developed dynamically reversible iron oxide nanoparticle assemblies (IONAs) consisting of ES-IONs cross-linked by small molecular aldehyde derivative ligands. The linkage of these ligands is cleaved in the presence of acidic environments, such as tumours, releasing ES-IONs that are capable of producing contrast in T 1 -weighted MRI (Figure 3a). To demonstrate the specificity of the tandem linkage-cleavage, pH-insensitive cross-linked iron oxide nanoparticles (Ins-IONAs) and micelle-like pH sensitive polymer-assisted iron oxide nanoparticle assemblies (PIONAs) of similar sizes were used as controls (Figure 3a). In vivo experiments on tumour-bearing mice clearly indicated that the IONAs could amplify the T 1 -MRI signal, whilst Ins-IONAs and PIONAs showed significantly lower intensities (Figure 3b).
In addition to the undertaken research to find new probes that serve as contrast agents for T 1 -weighted MRI, there have been some research groups that have focused on developing probes for dual contrast MRI, that is, probes that are capable of producing both negative (T 2 ) and positive (T 1 ) contrast. In MRI, T 1 -weighted images typically provide better spatial resolution, while T 2 -weighted images can provide enhanced detection of lesions. For this reason, dual contrast probes could potentially provide better imaging information leading to higher diagnostic accuracy. Regarding research in this field, Jung et al. [86] in 2014 synthesised 5 nm SPIONs that, as shown using in vivo MRA experiments in rat models, were capable of producing both positive and negative contrast in MRA by using different acquisition pulses. On the one hand, ultrashort echo (UTE) sequence, which differently to the sequences explained in Section 2 use a radial sampling of the k-space to minimise T E , positively enhanced vascular signals in MR angiography, providing highly resolved vessel structures. In addition, typical gradient sequences such as fast low angle shot (FLASH) acquisition yielded strong negative vessel contrast, resulting in a higher number of discernible vessel branches. Moreover, in 2018 Alipour et al. [82] developed 11 nm silica coated cubic SPIONs as a dual-mode contrast agent for T 1 and T 2 MRI. In vivo investigations on a 3 T MRI scanner demonstrated both positive and negative contrast enhancement 70 min post intravenous injection in healthy rat models.

Doped Iron Oxide Nanoparticles
Throughout the past years, some metals have been integrated into IONPs in order to get better T1 enhancement. Among different metals, focus remains on gadolinium (Gd), which are not included since our goal is focusing on alternatives, and manganese (Mn); however, there are some studies that use other metals such as copper (Cu) and europium (Eu).
Manganese-based contrast agents, similarly to Gd contrast agents, can significantly enhance the positive contrast in T1-weighted MRI. Furthermore, they can serve as contrast agents for visualising sub-anatomic structures and activities of brain and heart by taking advantage of Mn 2+ as a biological calcium analogue. Nevertheless, their high toxicity limits their clinical applications, and for that reason, it is highly desirable to develop alternative non-toxic manganese contrast agents for clinical MRI. Several research groups have been recently developing manganese-containing iron oxide nanoparticles as alternative positive MRI contrast agents. In 2013, Li et al. [113] synthesised nontoxic ultrasmall manganese ferrite (MnFe2O4) nanoparticles of 2-3 nm that were shown to enhance the T1 contrast in the liver, kidneys and brain of healthy mice during in vivo MRI. Similarly, in 2014 Huang et al. [114] produced Mn-doped IONPs of around 5 nm in diameter capable of functioning as a contrast agent for T1-weighted MRI having a low r2/r1 ratio of 2.6. Post injection, in vivo T1-MRI showed a brighter signal in the liver regions of healthy mice. Furthermore, in 2015 Zhang et al. [115] synthesised bovine serum albumin-coated manganese-doped IONPs (MnIO-BSA), 5 nm in size and with a low r2/r1 ratio of 2.18, as a theranostic platform for tumour targeting. The probe, once accumulated in the tumour, was able to provide contrast in T1-weighted MRI and be heated using an

Doped Iron Oxide Nanoparticles
Throughout the past years, some metals have been integrated into IONPs in order to get better T 1 enhancement. Among different metals, focus remains on gadolinium (Gd), which are not included since our goal is focusing on alternatives, and manganese (Mn); however, there are some studies that use other metals such as copper (Cu) and europium (Eu).
Manganese-based contrast agents, similarly to Gd contrast agents, can significantly enhance the positive contrast in T 1 -weighted MRI. Furthermore, they can serve as contrast agents for visualising sub-anatomic structures and activities of brain and heart by taking advantage of Mn 2+ as a biological calcium analogue. Nevertheless, their high toxicity limits their clinical applications, and for that reason, it is highly desirable to develop alternative non-toxic manganese contrast agents for clinical MRI. Several research groups have been recently developing manganese-containing iron oxide nanoparticles as alternative positive MRI contrast agents. In 2013, Li et al. [113] synthesised nontoxic ultrasmall manganese ferrite (MnFe 2 O 4 ) nanoparticles of 2-3 nm that were shown to enhance the T 1 contrast in the liver, kidneys and brain of healthy mice during in vivo MRI. Similarly, in 2014 Huang et al. [114] produced Mn-doped IONPs of around 5 nm in diameter capable of functioning as a contrast agent for T 1 -weighted MRI having a low r 2 /r 1 ratio of 2.6. Post injection, in vivo T 1 -MRI showed a brighter signal in the liver regions of healthy mice. Furthermore, in 2015 Zhang et al. [115] synthesised bovine serum albumin-coated manganese-doped IONPs (MnIO-BSA), 5 nm in size and with a low r 2 /r 1 ratio of 2.18, as a theranostic platform for tumour targeting. The probe, once accumulated in the tumour, was able to provide contrast in T 1 -weighted MRI and be heated using an external NIR light source for photothermal therapy (PTT) purposes. The in vivo MRI experiments using a tumour-bearing mouse model exhibited significant signal enhancement (about two times) at the tumour site. Furthermore, it was demonstrated that hyperthermia caused by the photothermal effect of the MnIO-BSA nanoparticles under NIR laser irradiation resulted in significant death of the cancer cells.
Only a few studies use copper or europium as doping metals. In 2019, our group synthesised extremely small citrate-coated iron oxide nanoparticles doped with Cu as a T 1 -weighted contrast agent [99]. After studying several Cu doping amounts, IONPs doped with 4% mol of Cu (Cu4-NPs) were chosen as the best probe. Cu4-NPs presented a hydrodynamic diameter of~15 nm and a core size of~3.5 nm. Moreover, they showed a high r 1 value of 15.7 mM −1 s −1 at 1.5 T and a low r 2 /r 1 ratio of 2.1. Interestingly, the reason for this increase in the r 1 values was due to the distribution of Cu atoms within the iron oxide structure. In vivo MRA showed that Cu4-NPs provided high-quality images with fine details of the vasculature up to 30 min post injection (Figure 4). In order to explore a different probe application, RGD molecules were conjugated to the surfaces of Cu4-NPs for active tumour targeting. In vivo T 1 -MRI experiments on tumour-bearing mice showed an increased positive signal on the tumour after intravenous injection of the RGD-Cu4-NP, further confirming their use as a positive contrast agent. Europium was used by Yang et al. [100] in 2015; this group produced citrate-coated europium-doped iron oxide nanocubes, 14 nm in size, with high r 1 values of 36.8 mM −1 s −1 at 0.5 T and a low r 2 /r 1 ratio of 2.65 for T 1 -weighted MRI. In vivo results showed an increase in signal brightness in the heart of healthy mice. Only a few studies use copper or europium as doping metals. In 2019, our group synthesised extremely small citrate-coated iron oxide nanoparticles doped with Cu as a T1-weighted contrast agent [99]. After studying several Cu doping amounts, IONPs doped with 4% mol of Cu (Cu4-NPs) were chosen as the best probe. Cu4-NPs presented a hydrodynamic diameter of ~15 nm and a core size of ~3.5 nm. Moreover, they showed a high r1 value of 15.7 mM −1 s −1 at 1.5 T and a low r2/r1 ratio of 2.1. Interestingly, the reason for this increase in the r1 values was due to the distribution of Cu atoms within the iron oxide structure. In vivo MRA showed that Cu4-NPs provided high-quality images with fine details of the vasculature up to 30 min post injection (Figure 4). In order to explore a different probe application, RGD molecules were conjugated to the surfaces of Cu4-NPs for active tumour targeting. In vivo T1-MRI experiments on tumour-bearing mice showed an increased positive signal on the tumour after intravenous injection of the RGD-Cu4-NP, further confirming their use as a positive contrast agent. Europium was used by Yang et al. [100] in 2015; this group produced citratecoated europium-doped iron oxide nanocubes, 14 nm in size, with high r1 values of 36.8 mM −1 s −1 at 0.5 T and a low r2/r1 ratio of 2.65 for T1-weighted MRI. In vivo results showed an increase in signal brightness in the heart of healthy mice.

Multimodal T1-Iron Oxide Nanoparticles
The combination of MRI with complementary imaging techniques such as PET or SPECT (Single Photon Emission Computed Tomography), allows protocols to be developed that exploit their synergy. Although a great number of multimodal IONPs have been described for T2-weighted MRI, only a few examples have been reported for IONPs acting as positive contrast agents [116].
These examples are focused on the combination of biocompatible iron oxide nanoparticles with radionuclides, such as Galium-68 ( 68 Ga) or Technetium-99m ( 99m Tc) for PET and SPECT, respectively. 68 Ga has been successfully used with IONPs in several multimodal experiments. In 2016, Pellico et al. developed a T1-weighted MRI/PET platform consisting of extremely small 68 Ga-doped IONPs synthesised using a microwave-driven protocol, which were diversely functionalised for their use in several applications. For example, 68 Ga core-doped IONPs functionalised with RGD were synthesised for tumour angiogenesis targeting. This probe presented a hydrodynamic size of 20.6 nm with core sizes of 2.5 nm. Furthermore, it showed a r1 value of 5.7 mM −1 s −1 at 1.5 T and a low r2/r1 ratio of 3.9, making it suitable for T1-weighted MRI. In vivo MRI experiments in tumour-bearing mouse models

Multimodal T 1 -Iron Oxide Nanoparticles
The combination of MRI with complementary imaging techniques such as PET or SPECT (Single Photon Emission Computed Tomography), allows protocols to be developed that exploit their synergy. Although a great number of multimodal IONPs have been described for T 2 -weighted MRI, only a few examples have been reported for IONPs acting as positive contrast agents [116].
These examples are focused on the combination of biocompatible iron oxide nanoparticles with radionuclides, such as Galium-68 ( 68 Ga) or Technetium-99m ( 99m Tc) for PET and SPECT, respectively. 68 Ga has been successfully used with IONPs in several multimodal experiments. In 2016, Pellico et al. developed a T 1 -weighted MRI/PET platform consisting of extremely small 68 Ga-doped IONPs synthesised using a microwave-driven protocol, which were diversely functionalised for their use in several applications. For example, 68 Ga core-doped IONPs functionalised with RGD were synthesised for tumour angiogenesis targeting. This probe presented a hydrodynamic size of 20.6 nm with core sizes of 2.5 nm. Furthermore, it showed a r 1 value of 5.7 mM −1 s −1 at 1.5 T and a low r 2 /r 1 ratio of 3.9, making it suitable for T 1 -weighted MRI. In vivo MRI experiments in tumour-bearing mouse models showed a brighter signal in the tumour 24 h post NP injection. 114 In addition, in 2018 the group synthesised a biorthogonal nano-radiotracer for in vivo pretargeted molecular imaging of atherosclerosis. This probe was based on the in vivo tetrazine ligation of an imaging and a targeting moiety, which were functionalised with a tetrazine (Tz) and a trans-cyclooctene (TCO), respectively. The imaging part consisted of 68 Ga-doped IONPs that provided simultaneous PET and T 1 -MRI signals, while the atherosclerosis targeting part consisted of an oxidised LDL (Low-Density Lipoprotein)-targeting IgM antibody (E-06). 68 Ga-IONP-Tz presented a hydrodynamic diameter of 15.5 nm with a core size of 2.8 nm. Moreover, they showed a high r 1 value of 7.1 mM −1 s −1 at 1.5 T and a low r 2 /r 1 ratio of 2.5, making them useful as contrast agents for T 1 -weighted MRI. Ex vivo T 1 -MRI of the aorta of an atherosclerotic mouse model showed a brighter signal on the atherosclerotic plaque [117].
Regarding research carried out on 99m Tc-doped IONPs, in 2013 Sandiford et al. [118] produced PEGylated bisphosphonate-coated USPIOs to use as a multimodal platform in T 1 -MRI and SPECT. This probe showed a high r 1 value of 9.5 mM −1 s −1 at 3 T and a low r 2 /r 1 ratio of 2.9, enabling their use as positive contrast agents. In vivo MRI experiments in healthy BALB/C mice showed a strong T 1 effect post NP injection resulting in a substantial increase in the signal from blood, making vessels, the heart compartments and other highly vascularised organs such as the spleen visible ( Figure 5a). This biodistribution was also observed using SPECT/CT with the nanoparticles circulating 40 min after i.v. injection (Figure 5b). showed a brighter signal in the tumour 24 h post NP injection. 114 In addition, in 2018 the group synthesised a biorthogonal nano-radiotracer for in vivo pretargeted molecular imaging of atherosclerosis. This probe was based on the in vivo tetrazine ligation of an imaging and a targeting moiety, which were functionalised with a tetrazine (Tz) and a trans-cyclooctene (TCO), respectively. The imaging part consisted of 68 Ga-doped IONPs that provided simultaneous PET and T1-MRI signals, while the atherosclerosis targeting part consisted of an oxidised LDL (Low-Density Lipoprotein)-targeting IgM antibody (E-06). 68 Ga-IONP-Tz presented a hydrodynamic diameter of 15.5 nm with a core size of 2.8 nm. Moreover, they showed a high r1 value of 7.1 mM −1 s −1 at 1.5 T and a low r2/r1 ratio of 2.5, making them useful as contrast agents for T1-weighted MRI. Ex vivo T1-MRI of the aorta of an atherosclerotic mouse model showed a brighter signal on the atherosclerotic plaque [117].
Regarding research carried out on 99m Tc-doped IONPs, in 2013 Sandiford et al. [118] produced PEGylated bisphosphonate-coated USPIOs to use as a multimodal platform in T1-MRI and SPECT. This probe showed a high r1 value of 9.5 mM −1 s −1 at 3 T and a low r2/r1 ratio of 2.9, enabling their use as positive contrast agents. In vivo MRI experiments in healthy BALB/C mice showed a strong T1 effect post NP injection resulting in a substantial increase in the signal from blood, making vessels, the heart compartments and other highly vascularised organs such as the spleen visible (Figure 5a). This biodistribution was also observed using SPECT/CT with the nanoparticles circulating 40 min after i.v. injection (Figure 5b).
Besides the use of T1-IONPs for dual MRI(T1)/PET or SPECT imaging, G. Wang et al. reported an example for MRI(T1)/CT imaging. In this case, ultrasmall magnetite nanoparticles were combined with Au nanocages rendering F-AuNC@Fe3O4. The ultrasmall Fe3O4 particles (2.2 nm) showed a r1 = 6.3 mM −1 s −1 , providing a T1-enhancement in MRI, whilst a strong X-ray attenuation was observed due to the AuNC [119].

Conclusions and Future Perspectives
In the last decade, significant effort has been made in the design of iron oxide formulations for T1 contrast due to the associated toxicity in the application of the traditional Gd 3+ contrast agents. We have summarised the most important features to optimise IONPs as a T1 contrast agent, i.e., ultrasmall core size with moderate crystallinity (usually maghemite (γ-Fe2O3)) and high colloidal stability with hydrodynamic sizes ranging from 5 to 20 nm. With the focus on the physicochemical Besides the use of T 1 -IONPs for dual MRI(T 1 )/PET or SPECT imaging, G. Wang et al. reported an example for MRI(T 1 )/CT imaging. In this case, ultrasmall magnetite nanoparticles were combined with Au nanocages rendering F-AuNC@Fe 3 O 4 . The ultrasmall Fe 3 O 4 particles (2.2 nm) showed a r 1 = 6.3 mM −1 s −1 , providing a T 1 -enhancement in MRI, whilst a strong X-ray attenuation was observed due to the AuNC [119].

Conclusions and Future Perspectives
In the last decade, significant effort has been made in the design of iron oxide formulations for T 1 contrast due to the associated toxicity in the application of the traditional Gd 3+ contrast agents. We have summarised the most important features to optimise IONPs as a T 1 contrast agent, i.e., ultrasmall core size with moderate crystallinity (usually maghemite (γ-Fe 2 O 3 )) and high colloidal stability with hydrodynamic sizes ranging from 5 to 20 nm. With the focus on the physicochemical characteristics of T 1 -IONPs such as iron phase, core size, type of coating or the surface charge, there is still room for improvement using the described synthetic procedures or novel strategies. Although the optimisation of these properties brings associated higher T 1 performance, the longitudinal relaxivities are still far from those obtained in some Gd 3+ -based nanoparticles such as gadofullerens, gadolinium oxide particles, Gd-nanocages or Gd 3+ conjugated to mesoporous silica nanoparticles [120]. It is noteworthy that the application of Gd 3+ -based agents has been thoroughly studied for many years, whilst the potential of IONPs as T 1 agents is an emerging field of research. Therefore, there is still a lack of consensus to determine the best scenario for T 1 -IONPs. To date, the variability of the particles showing different physicochemical properties and MRI responses is high. However, a broad number of applications have been reported including not only MRI but also multimodal imaging for molecular imaging applications or metal-doped particles for therapeutic purposes. In addition, other issues such as biocompatibility, pharmacokinetics or delivery pathways must be studied in advance to guarantee their clinical translation. In conclusion, there is great potential in the development of T 1 -IONPs with several concerns to be evaluated in detail. Further studies must bring a clear conclusion about the best properties and then adapt the formulations to clinical requirements.

Conflicts of Interest:
The authors declare no conflict of interest.