Current Status and Future Aspects of Gadolinium Oxide Nanoparticles as Positive Magnetic Resonance Imaging Contrast Agents

Abstract

Although numerous studies have investigated gadolinium oxide (Gd₂O₃) nanoparticles (NPs) as positive (T₁) magnetic resonance imaging (MRI) contrast agents (CAs), comprehensive reviews on this topic remain scarce. Therefore, it is essential to evaluate their current status and outline prospects. Despite promising physicochemical properties such as considerably higher relaxivities compared to 3–5 s⁻¹mM⁻¹ of clinically approved Gd(III)-chelate contrast agents and encouraging results from in vivo animal studies such as highly improved contrast enhancements, drug loading, and tumor targeting, extensive in vivo toxicity assessments including long-term toxicity and formulation advancements suitable for renal excretion (d < ~3 nm) are still required for clinical translation. This review summarizes the synthesis, characterization, in vitro and in vivo toxicity, and in vivo MRI applications of surface-modified Gd₂O₃ NPs as T₁ MRI CAs. Finally, future perspectives on the development of surface-modified Gd₂O₃ NPs as potential next-generation T₁ MRI CAs are discussed.

1. Introduction

Molecular imaging plays a critical role in medical treatments by providing detailed anatomical information through high-resolution imaging [1,2]. Various imaging modalities have been developed, including fluorescence imaging (FI), magnetic resonance imaging (MRI), X-ray computed tomography (CT), positron emission tomography (PET), ultrasound imaging (USI), and single-photon emission computed tomography (SPECT) [3,4]. These techniques enable the detection, diagnosis, and monitoring of diseases. Among them, MRI is particularly valued for its non-invasive nature as it employs non-ionizing radiofrequency radiation and allows for whole-body imaging with unlimited penetration depth in soft tissues, making it a widely adopted modality [5]. However, MRI suffers from relatively low sensitivity due to the small population difference between the two proton spin states [6,7], which limits its ability to detect small or early-stage lesions. This limitation can be mitigated by the use of contrast agents (CAs), which enhance signal intensity by accelerating proton spin relaxation processes, thereby improving detection sensitivity [8,9].

Various types of MRI CAs have been investigated, including Gd(III)- and Mn(II)-chelates [10,11], Gd- [12] and Mn-containing nanoparticles (NPs) [13] for positive (T₁) MRI, and superparamagnetic iron oxide nanoparticles (SPIONs) [14,15] for negative (T₂) MRI. Currently, only Gd(III)-chelates are approved for clinical use as T₁ MRI CAs [16,17]. Although SPIONs were previously commercialized, their use is now minimal [18]. Gd³⁺, with seven unpaired 4f-electrons and a high spin magnetic moment (s = 7/2), is ideal for use in T₁ MRI CAs [19]. Consequently, Gd₂O₃ NPs can induce stronger T₁ proton spin relaxations than Mn-and Fe-based NPs owing to lower electron spin magnetic moments of Mn²⁺ (s = 5/2) and Fe³⁺ (s = 5/2) than Gd³⁺ [16,20,21,22,23,24,25]. For example, this can be noticed from the r₁ value (33.4 s⁻¹mM⁻¹) of free Gd³⁺ ions-grafted NDs [26], which is higher than that (13.5 s⁻¹mM⁻¹) of free Mn²⁺ ions-grafted NDs [27]. However, after PVP coating, r₁ value of free Gd³⁺ ions-grafted NDs dropped to 15.9 s⁻¹mM⁻¹ [28]. T₁ MRI CAs are generally preferred over T₂ MRI CAs because signal enhancement is easier to detect than signal loss [29]. Moreover, SPIONs can cause signal darkening and susceptibility-induced artifacts, which may complicate image interpretation [30,31,32,33]. In contrast, Gd₂O₃ NPs produce the high-quality positive contrast images with less artificial signal voids, thereby offering clearer diagnostic outcomes [34].

However, Gd³⁺ ions are inherently toxic and non-biodegradable through metabolic processes [35]. To mitigate this toxicity, they are commonly administered as Gd(III)-chelates [35,36,37,38,39]. Although Gd(III)-chelates effectively enhance positive contrasts in MRI, they exhibit low sensitivity due to their low longitudinal proton spin relaxivity (r₁) values and lack of specificity. Moreover, they have been associated with serious adverse effects, including nephrogenic systemic fibrosis (NSF) in patients with renal impairment, a condition characterized by skin and organ thickening and darkening, ultimately impairing organ function [40,41,42]. Recent studies have also indicated that repeated administration of Gd(III)-chelates can lead to gadolinium deposition in the brain, potentially causing neurotoxicity [43,44].

With advancements in nanotechnology, Gd₂O₃ NPs have emerged as promising alternatives to Gd(III)-chelates, offering several advantages. These NPs exhibit a high density of Gd³⁺ ions per NP, resulting in higher r₁ values [45,46,47,48], compared to 3–5 s⁻¹mM⁻¹ [17,38] of Gd(III)-chelates. They also demonstrate prolonged blood circulation times due to the enhanced permeability and retention (EPR) effect [46], enabling extended imaging and diagnosing windows [47]. Furthermore, their high surface-to-volume ratios facilitate drug loading and conjugation of targeting ligands, making them suitable for theranostic applications [48,49,50]. However, their drawback is potential toxicity, which needs to be clearly and thoroughly investigated for clinical translations. For biomedical applications, Gd₂O₃ NPs must be surface-modified with hydrophilic and biocompatible ligands to ensure colloidal stability, biocompatibility, and functionalization potential. Various synthesis methods have been reported, including polyol [51,52], dimethyl sulfoxide (DMSO) [53,54], thermal decomposition [55], and hydrothermal methods [56]. Surface modifications have employed a range of biocompatible and hydrophilic ligands such as polyacrylic acid (PAA) [51,57], poly(methyl vinyl ether-alt-maleic acid) (PMVEMA) [58,59,60], poly(acrylic acid-co-maleic acid) (PAAMA) [61,62], polyaspartic acid (PASA) [63], D-glucuronic acid [64], citric acid (CA) [65], dextran [66], polyethylene glycol (PEG) [67], or polyvinylpyrrolidone (PVP) [68]. These ligand coatings are critical not only for stability and functionality but also for reducing toxicity by preventing the release of free Gd³⁺ ions from the NPs.

To date, few comprehensive reviews have focused specifically on surface-coated Gd₂O₃ NPs as T₁ MRI CAs. However, as illustrated in Figure 1, a substantial number of research articles have been reported on surface-coated Gd₂O₃ NPs as T₁ MRI CAs and their biomedical applications. Therefore, it is both timely and necessary to evaluate the current status of this research and to discuss future directions. As shown in Figure 1, the annual growth rate of publications increased up to 2019 and then decreased afterwards. This trend reflects the increase in the development of various types of gadolinium NP formulations, in vivo MRI, and biotoxicity studies up to 2019, and then the decrease afterwards. This review covers key aspects, including synthesis methods, physicochemical characterization, in vitro and in vivo toxicity assessments, and in vivo MRI applications. Finally, future perspectives for clinical translation of surface-coated Gd₂O₃ NPs are outlined.

2. Synthesis of Surface-Modified Gd₂O₃ NPs

To date, various synthetic approaches have been employed to prepare surface-modified Gd₂O₃ NPs for biomedical applications. These include polyol [51,52], DMSO [53], thermal decomposition [55], and hydrothermal methods [56]. As summarized in

Table 1. Summary of synthesis methods with associated advantages and disadvantages.

Method	Solvent	Advantage	Disadvantage	Ref.
Polyol	Polyol such as ethylene glycol, diethylene glycol, triethylene glycol, polyethylene glycol	– surface modification in one-pot synthesis – ultrasmall NPs	– small-scale synthesis – poor crystallinity	[51,52]
DMSO	DMSO	– low temperature synthesis (<60 °C) – surface modification in one-pot synthesis – ultrasmall NPs	– small-scale synthesis – poor crystallinity	[53]
Thermal decomposition	High boiling point organic solvent such as oleic acid, oleylamine	– high crystallinity. – monodisperse NPs with particle size control	– expensive (organic solvent, inert gas) – post synthesis surface modification – solvent waste – small-scale synthesis	[55]
Hydrothermal	Water	– environmentally friendly – large-scale synthesis	– requires autoclave	[56]

, each method presents distinct advantages and limitations in terms of synthesis complexity, scalability, and environmental impact.

2.1. Polyol Method

The polyol method is effective for synthesizing ultrasmall Gd₂O₃ NPs, with average particle diameters of approximately 2.0 nm, via a one-pot process that simultaneously enables surface modification with hydrophilic and biocompatible ligands [34,51,58,61,63,64,69,70,71,72,73,74,75]. High boiling point polyols, such as ethylene glycol (EG), diethylene glycol (DEG), triethylene glycol (TEG), and polyethylene glycol (PEG), are commonly used as solvents. These polyols also function as stabilizing agents, preventing particle aggregation. A general scheme for the one-pot polyol synthesis of surface-coated Gd₂O₃ NPs is displayed in Figure 2a. The process involves a Gd³⁺ precursor, surface-coating ligands, NaOH, and a polyol solvent, all reacted under atmospheric conditions with magnetic stirring. Initially, the Gd³⁺ precursor solution is prepared by dissolving Gd³⁺ salt, such as GdCl₃·xH₂O, and ligands such as PAA in a polyol such as TEG within a three-necked round-bottom flask. This mixture is stirred magnetically at room temperature under atmospheric conditions until complete dissolution. Separately, a NaOH solution is prepared by dissolving NaOH in TEG in a beaker and is then slowly added to the precursor solution until the pH reaches 8–10. The resulting mixture is stirred at ~110 °C for ~12 h to yield ligand-coated Gd₂O₃ NPs.

For example, Lee’s group synthesized ultrasmall Gd₂O₃ NPs using various hydrophilic ligands, such as PAA, PMVEMA, PAAMA, PASA, CA, and D-glucuronic acid, via the one-pot polyol method [51,58,61,63,64]. In another study, Guleria et al. synthesized Gd₂O₃ NPs using polyols of varying chain lengths, including DEG, TEG, tetraethylene glycol (TeEG), and PEG-200, and observed a linear correlation between the glycol chain length and the resulting NP size [70].

2.2. DMSO Method

The DMSO method is suitable for low-temperature synthesis (<60 °C), enabling the production of ultrasmall Gd₂O₃ NPs with uniform particle size and simultaneous one-pot surface modification using hydrophilic and biocompatible ligands [53,76,77]. As shown in Figure 2b, Gd³⁺ precursor such as Gd(acetate)₃·xH₂O is dissolved in DMSO under atmospheric conditions with magnetic stirring until a transparent solution is obtained. Tetramethylammonium hydroxide (TMAH) solution is then added dropwise to adjust the pH to ~8, resulting in a cloudy solution. This mixture is stirred magnetically for 24 h to allow the formation of ultrasmall Gd₂O₃ NPs. For surface coating, a hydrophilic ligand such as PAA is added to the reaction mixture, followed by the addition of TMAH to maintain pH at ~8. The solution is stirred for another 24 h to ensure uniform ligand surface coating. The resulting ligand-coated Gd₂O₃ NPs are purified by repeated ethanol washing and centrifugation to remove unreacted reagents and solvents. Finally, the surface-coated NPs are dispersed in triple-distilled water and subjected to dialysis against water to eliminate residual impurities.

For example, Uvdal and co-workers synthesized Gd₂O₃ NPs (d = 4–5 nm) by dissolving Gd(acetate)₃·xH₂O in DMSO and adding TMAH in ethanol under ambient conditions with magnetic stirring [53]. In another example, Cui et al. synthesized Gd₂O₃ NPs using Gd(acetate)₃·xH₂O and TMAH in DMSO at room temperature and subsequently conjugated the NPs with carboxyfluorescein [FI]-polyethylene glycol [PEG]-bombesin (BBN), resulting in NPs with a particle diameter of 52.3 nm [77].

2.3. Thermal Decomposition

The thermal decomposition method yields monodispersed Gd₂O₃ NPs with controlled particle size; however, it requires high temperatures, organic solvents such as OA and OM, and an inert gas atmosphere, rendering the synthesis costly and environmentally unfriendly [55,78,79,80,81,82]. Additionally, this method does not permit one-pot synthesis for surface modification. As shown in Figure 2(cI), Gd³⁺ precursor is added to a mixture of OA and OM in a three-neck round-bottom flask and heated to ~110 °C under a flow of argon or nitrogen to remove moisture and dissolved gases. The reaction mixture is then refluxed at 290–340 °C for 1–6 h. After cooling to room temperature, the Gd₂O₃ NPs are precipitated by adding ethanol, collected by centrifugation, and redispersed in toluene: this washing procedure is repeated three times. The purified NPs are subsequently dispersed and stored in hexane or chloroform. For surface coating, the hydrophobic ligands on the NP surfaces, primarily OA, with a minor contribution from OM due to the stronger binding affinity of the -COOH group than the -NH₂ group, are replaced with hydrophilic ligands such as polyvinyl pyrrolidone (PVP) through ligand exchange (Figure 2(cII)). The resulting hydrophilic ligand-coated Gd₂O₃ NPs are dispersed in water and further purified through dialysis against water.

For example, Fang et al. synthesized monodispersed ultrasmall Gd₂O₃ NPs (d = 2.9 nm) capped with hydrophobic OA, via thermal decomposition of a Gd(OA)₃ precursor in a mixture of OA and OM, which acted as both solvent and surfactant [78]. The OA-capped Gd₂O₃ NPs were subsequently converted to PVP-Gd₂O₃ NPs through ligand exchange with PVP for use as T₁ MRI CAs. Similarly, Cai et al. prepared OA/OM-capped Gd₂O₃ nanoplates (length × thickness = 10 × 1 nm) via thermal decomposition of Gd(OA)₃ at 340 °C using OA and OM as solvent and capping agents. These nanoplates were encapsulated with N-dodecyl-polyethylene imine (PEI)-PEG polymers to form hydrophilic nanoplate clusters (d = 95 nm) for application as T₁ MRI CAs [55].

2.4. Hydrothermal Method

The hydrothermal method is considered environmentally friendly as it utilizes water as the solvent. The synthesis is controlled by adjusting temperature and pressure [46,56,83,84,85,86]. As illustrated in Figure 2d, the Gd³⁺ precursor is dissolved in deionized water. Separately, a hydrophilic ligand is dissolved in deionized water at room temperature and added to the Gd³⁺ solution under magnetic stirring. An aqueous solution of NaOH or urea is then added dropwise under magnetic stirring to adjust the pH to 9–10. The resulting solution is transferred into an autoclave and subjected to hydrothermal treatment at elevated temperatures (typically above 150 °C) for 6–24 h. The final precipitates are washed three times with ethanol, dispersed in triple-distilled water, and further purified via dialysis against water.

For example, Wu et al. synthesized hyaluronic acid (HA)-coated Gd₂O₃ NPs (d = 105 nm) using the hydrothermal method [46]. In this procedure, HA was dissolved in water under vigorous magnetic stirring at ambient temperature overnight. GdCl₃·6H₂O and NaOH were then added, followed by magnetic stirring for 5 min to form a homogeneous clear solution. The mixture was transferred to an autoclave, sealed, and heated at 120 °C for 6 h. The resulting HA-coated Gd₂O₃ NPs were purified through dialysis against water to remove impurities.

3. Various Characterizations

Following synthesis, various characterization techniques can be employed to determine the physicochemical properties of surface-modified Gd₂O₃ NPs, as summarized in

Table 2. Various characterization techniques and information.

Characterization Technique	Information
TEM	Particle size, morphology
ICP-AES	Gd concentration in aqueous solution sample
XRD	Crystal structure
DLS	Hydrodynamic diameter
FT-IR	Surface coating
TGA	Surface-coating amount
Zeta potential	Surface charge
VSM, MPMS	Magnetic properties
MRI	r1, r2 values
MTT, CCK-8, WST-1	in vitro cellular toxicity

[51,53,61,87]. These include high-resolution transmission electron microscopy (HRTEM) to determine particle size and morphology; dynamic light scattering (DLS) to measure hydrodynamic diameter; X-ray diffraction (XRD) to determine the crystal structure; Fourier transform-infrared (FT-IR) absorption spectroscopy to analyze surface coating; and thermogravimetric analysis (TGA) to estimate the surface-coating content. Additional methods include zeta potential measurements to assess the surface charge; inductively coupled plasma-atomic emission spectroscopy (ICP-AES) to quantify Gd concentration in solution; and vibrating sample magnetometry (VSM) or magnetic property measurement system (MPMS) to measure magnetic properties. MRI is used to determine the r₁ and r₂ relaxivities. In vitro cellular toxicity is assessed using standard assays such as methyl thiazolyl tetrazolium (MTT), cell counting kit (CCK)-8, and water-soluble tetrazolium (WST)-1 to evaluate cell viability. Collectively, these methods constitute a standard characterization protocol for surface-coated Gd₂O₃ NPs.

3.1. Particle Size and Morphology

Particle size and morphology are critical parameters for biomedical applications. These features are typically characterized using HRTEM operated at acceleration voltages above 200 kV. For TEM imaging, NPs are dispersed onto a copper grid coated with a carbon film by drop-casting a solution sample prepared in water or a water/ethanol mixture. High-resolution imaging also enables measurement of lattice fringes, which can aid in NP identification. In addition, elemental mapping provides the spatial distribution of constituent elements, allowing identification of both the NPs and their surface-coating ligands.

For example, Miao et al. synthesized PAA-coated ultrasmall Gd₂O₃ NPs (d_avg = 2.0 ± 0.1 nm) using a one-pot polyol method; the NPs remained colloidally stable in aqueous media without precipitation (Figure 3a) [51]. Ahrén et al. prepared Gd₂O₃ NPs with particle diameters of 4–5 nm via the DMSO method at room temperature (Figure 3b) [53]. These NPs exhibited good aqueous stability, with a zeta potential of +34.7 mV due to the presence of ammonium acetate on their surfaces. Cai et al. prepared OA/OM-capped Gd₂O₃ square nanoplates (thickness × length = 1 × 10 nm) via thermal decomposition of Gd(OA)₃ at 340 °C, using OA and OM as both solvent and capping agent (Figure 3c) [55]. Yang et al. synthesized Gd(OH)₃ nanorods by hydrothermal treatment of GdCl₃ and triethylamine in water at 160 °C for 8 h in an autoclave [56]. Figure 3d displays the TEM image of the as-synthesized Gd(OH)₃ nanorods (length × width = 200 × 10 nm).

3.2. Crystal Structure

XRD analysis provides critical information on the crystal structure of NPs and allows identification of the synthesized products. Powder samples are typically used, and the 2θ scan range is set between 10 and 100°. Diffraction peaks are indexed using Miller indices (hkl), and the unit cell parameters (a, b, c, α, β, γ) are calculated using Bragg’s law, in combination with interplanar spacing equations derived from the assigned crystal system. In addition, the average particle diameter can be roughly estimated using the Scherrer equation, which employs the full width at half maximum (FWHM) of selected XRD peaks [88]. Ultrasmall NPs (d < 3 nm) often exhibit poor crystallinity and may appear amorphous in XRD due to their limited long-range order. The pair distribution function analysis is recommended for ultrasmall NPs with amorphous XRD patterns because it offers the short-range local atomic structure of amorphous NPs [89]. However, heat treatment can induce particle growth and crystallization, allowing crystal structure determination [90].

For example, the XRD patterns corresponding to the gadolinium-based nanomaterials displayed in the TEM images in Figure 3a–d are provided in Figure 3e–h, respectively. The ultrasmall Gd₂O₃ NPs synthesized via the polyol and DMSO methods (TEM images in Figure 3a and Figure 3b, respectively) displayed broad, amorphous XRD patterns, indicating poor crystallinity (bottom spectra in Figure 3e,f) [51,53]. After thermal treatment, the NPs exhibited diffraction peaks corresponding to the cubic phase of Gd₂O₃ (top spectrum in Figure 3e) [51]. The top spectrum in Figure 3f represents the reference pattern of commercial bulk cubic Gd₂O₃ [53]. The Gd₂O₃ nanoplates synthesized via the thermal decomposition method (TEM image in Figure 3c), displayed broad peaks in the XRD pattern (top spectrum in Figure 3g) [55], which were indexed to a poorly developed monoclinic phase of Gd₂O₃ (JCPDS No 43-1015) [80,82,91], as shown in the reference pattern (bottom spectrum in Figure 3g). The XRD pattern of the Gd(OH)₃ nanorods synthesized via the hydrothermal method (TEM image in Figure 3d) is presented in the top spectrum of Figure 3h and matches well with the reference pattern for the hexagonal phase of Gd(OH)₃ (JCPDS No. 83-2037) (bottom spectrum in Figure 3h) [56].

3.3. Hydrodynamic Diameter

The hydrodynamic diameters (a_avg) of NPs are measured using samples dispersed in water. The a_avg values are typically larger than the corresponding d_avg values due to the presence of surface-coating ligands and the hydration layer formed by adsorbed water molecules. The a_avg is influenced by particle size, morphology, and hydrophilicity of the surface ligands. Larger and more hydrophilic ligands attract more water molecules, thereby increasing the a_avg and enhancing colloidal stability. In particular, polymer ligands tend to attract more water molecules than small molecules, resulting in larger a_avg values and improved colloidal stability [92].

For example, Miao et al. synthesized ultrasmall Gd₂O₃ NPs (d = 2.0 nm) coated with PAA of varying molecular weight by the polyol method. They observed an increase in a_avg values (10.3 → 11.1 → 11.3 nm) with increasing PAA size (1200 → 5100 → 15,000 amu) (Figure 4a) [75]. Cui et al. synthesized Gd₂O₃ NPs (d = 52.3 nm) grafted with carboxyfluorescein (FI)-polyethylene glycol (PEG)-bombesin (BBN) via the DMSO method [77], and reported a large a_avg of 90.6 nm (Figure 4b), attributed to the combination of large particle size and bulky polymeric and biomolecular coating. Similarly, Cai et al. synthesized Gd₂O₃@N-dodecyl-PEI-PEG nanoplates (thickness × length = 1 × 10 nm), which exhibited a large a_avg value of 95 nm (Figure 4c) owing to the thick polymer coating [55]. Siribbal et al. synthesized CA-coated hollow Gd₂O₃ NPs (d = 100 nm) using a hydrothermal method [93], reporting a very large a_avg value of 233.7 nm, due to the large particle size and surface modification (Figure 4d).

3.4. Surface-Coating Analysis

The surface coating of surface-modified Gd₂O₃ NPs can be analyzed using FT-IR absorption spectroscopy and TGA [94,95]. Dried powder samples are used for both measurements to minimize the influence of water. Characteristic absorption peaks of surface-coating ligands in the FT-IR absorption spectra provide direct evidence of their presence on the Gd₂O₃ NP surfaces. Additionally, their coating amount can be estimated from TGA data, as most organic ligands undergo thermal decomposition below 450 °C via oxidation in a flow of hot air during TGA [96,97,98]. The coating amount is typically expressed in weight percent (wt.%), calculated from the mass loss observed in the TGA curve, excluding the initial mass loss between room temperature and ~100 °C due to desorption of water and air.

For example, Ho et al. synthesized arginylglycylaspartic acid (RGD)-PAA-coated ultrasmall Gd₂O₃ NPs (denoted as RGD-PAA-UGNPs) for cancer imaging using in vivo T₁ MRI in tumor-model mice [95]. They confirmed the successful coating of RGD-PAA onto UGNPs by recording FT-IR absorption spectra of PAA-UGNPs, RGD-PAA-UGNPs, free PAA, and free RGD (Figure 5a). The C=O stretching peak at 1553 cm⁻¹ in both the PAA-UGNP and RGD–PAA–UGNP spectra confirmed the successful coordination of PAA ligands to the UGNP surfaces. This peak exhibited a red shift from that of free PAA at ~1700 cm⁻¹ due to electrostatic interactions between the COO⁻ groups of PAA and surface Gd³⁺ ions of UGNP. Furthermore, the FT–IR absorption spectrum of RGD–PAA–UGNPs showed characteristic RGD peaks, including N–H bending at 1544 cm⁻¹ and C–N stretching at 1390 cm⁻¹, confirming successful conjugation of RGD to PAA through amide bond formation.

For surface-coating amount estimation, Park et al. measured TGA curves of ultrasmall Gd₂O₃ NPs coated with five different kinds of ligands: PEI, pimelic acid, azelaic acid, suberic acid, and sebacic acid (Figure 5b) [99]. The water and air desorption from the NPs occurred between room temperature and ~105 °C. Mass drops between ~105 and ~450 °C were attributed to the thermal decomposition or oxidative combustion of the coating ligands. The residual mass in TGA curves corresponded to Gd₂O₃ NPs. The measured surface-coating amounts were 52.3% for pimelic acid, 60.9% for suberic acid, 55.7% for azelaic acid, 63.1% for sebacic acid, and 30.5% for PEI-70000.

3.5. Zeta Potentials

The zeta potential (ζ) of surface-coated Gd₂O₃ NPs dispersed in water provides insight into their surface charge in aqueous media. Negative zeta potentials indicate the presence of negatively charged functional groups such as -OH and -COOH on NP surfaces, whereas positive zeta potentials reflect the presence of positively charged functional groups such as -NH₂. Zwitterionic ligands such as amino acids typically result in low zeta potential values [100]. High absolute zeta potential values are needed to ensure good colloidal stability [101,102].

For example, Marasini et al. reported polyaspartic acid (PSA)-coated Gd₂O₃ NPs as a dual-modal T₁ and T₂ MRI CA [63]. Figure 5c shows a highly negative zeta potential of −28.0 mV in aqueous media, consistent with their observed good colloidal stability, evidenced by the absence of NP precipitation following synthesis.

3.6. Magnetic Properties

The magnetic properties of surface-coated Gd₂O₃ NPs are typically assessed by recording magnetization (M) versus applied magnetic field (H) (M–H curves) and M versus temperature (T) (M–T curves) using either VSM or MPMS. The M–H curves provide saturation magnetization, coercivity, and remanence, while the M–T curves yield the blocking temperature. These curves are measured using 10–20 mg of powdered samples. Notably, M values of surface-coated Gd₂O₃ NPs arises predominantly from the spin magnetic moment of Gd³⁺ (s = 7/2), and is largely independent of both ligand type and particle size [21].

For example, Guleria et al. measured M–H and M–T curves of various polyol-coated Gd₂O₃ NPs [70]. Figure 6a and Figure 6b show the M–H curves of DEG-, TEG-, TeEG-, and PEG-coated Gd₂O₃ NPs at 5 and 300 K, respectively. The higher M values at 5 K are attributed to enhanced spin alignment of 4f-electron spins (s = 7/2) at lower temperatures, due to reduced thermal fluctuation. Consistent with the absence of hysteresis in the M–H curves, no magnetic transitions were observed in the M–T curves as depicted in Figure 6c,d, confirming the paramagnetic nature of the Gd₂O₃ NPs down to T = 2 K. Moreover, the negligible variation in magnetization among the DEG-, TEG-, TeEG-, and PEG-coated Gd₂O₃ NPs confirms that the magnetic moment originates primarily from the paramagnetic Gd₂O₃ core, as the surface-coating ligands are nonmagnetic and have minimal influence on the magnetic behavior.

3.7. Water Proton Spin Relaxivities (r₁ and r₂ Values)

In MRI, T₁ (longitudinal or spin-lattice) and T₂ (transverse or spin-spin) relaxation times indicate how quickly the excited proton spins return to their equilibrium state. T₁ describes the recovery time of the longitudinal (i.e., z-component) magnetization along the main magnetic field. T₂ describes the decay time of the transverse (i.e., x-y component) magnetization perpendicular to the main magnetic field [103,104].

The positive (T₁) contrast efficacy increases with increasing r₁ value while keeping r₂/r₁ ratio close to one, whereas negative (T₂) contrast efficacy increases with increasing r₂ value with r₂/r₁ ratio as large as possible. The r₁ and r₂ values are determined by measuring T₁ and T₂ water proton spin relaxation times at various Gd concentrations in aqueous solution samples using an MRI scanner. The inverse 1/T₁ and 1/T₂ relaxation times are plotted as a function of Gd concentration, and r₁ and r₂ values are obtained from the corresponding slopes. The r₂ value is always greater than r₁ because T₁ relaxation always accompanies T₂ relaxation whereas the opposite does not happen, making r₂/r₁ ratio always greater one. Therefore, positive (T₁) contrast efficacy increases with increasing r₁ value with r₂/r₁ ratio close to one, which are preferred to T₁ MRI contrast agents. As shown in Figure 7, T₁ relaxation is primarily induced by the interaction between the Gd³⁺ ion on Gd₂O₃ NP surface and water molecule (called inner sphere interaction model), whereas T₂ relaxation is induced by the interaction between NP magnetic moment and water molecule (called outer sphere interaction model) [104].

For example, Guleria et al. measured the r₁ and r₂ values of Gd₂O₃ NPs coated with various ligands [70]. Figure 8a and Figure 8b depict the plots of 1/T₁ and 1/T₂ relaxation times, respectively, for DEG-, TEG-, TeEG-, and PEG-coated Gd₂O₃ NPs dispersed in water as function of Gd concentration. Both r₁ and r₂ values increased with increasing glycol chain length, as displayed in Figure 8c and Figure 8d, respectively, suggesting that proton relaxation rates are proportional to the thickness of the hydrophilic coating layer, probably due to enhanced access of outer-sphere water molecules facilitated by the thicker coating.

3.8. Toxicity of Surface-Modified Gd₂O₃ NPs

Gd₂O₃ NPs have demonstrated significant potential as T₁ MRI CAs due to their considerably higher r₁ and r₂ values compared to Gd(III)-chelates [47,97,105,106,107,108,109,110,111,112]. However, their clinical translation is hindered by concerns over toxicity, primarily because free Gd³⁺ ions are toxic and cannot be metabolized by the body [113,114]. Therefore, complete excretion of injected Gd₂O₃ NPs, preferably via the renal system, without the release of free Gd³⁺ ions is essential. For clinically approved Gd(III)-chelates, clinical acute and chronic toxicities are well reported [113,114]: acute toxicity includes immediate hypersensitivity allergic-like reactions and physiologic reactions such as coldness, warmth, or pain at the injection site, nausea with or without vomiting, headache, paresthesia, and dizziness and chronic toxicity includes NSF, kidney disease, neurotoxicity, and gadolinium deposition disease such as bone pain and skin and subcutaneous tissue burning pain. However, there exist no clinical studies on acute and chronic toxicity for surface-modified Gd₂O₃ NPs. Therefore, we discussed in vitro cellular and in vivo animal toxicity studies reported so far in this section. Notably, free Gd³⁺ ion has an ionic radius comparable to that of Ca²⁺, enabling them to interfere with Ca²⁺-dependent metabolic process and potentially cause adverse effects [115,116]. To mitigate this risk, Gd₂O₃ NPs must be coated with hydrophilic and biocompatible ligands such as PAA, PMVEMA, PAAMA, PASA, D-glucuronic acid, CA, dextran, PEG, and PVP [51,58,61,63,64,66,68,73,75,93]. This review addresses both in vitro cellular toxicity and in vivo toxicity (i.e., histological analysis, body weight variation, biodistribution, clearance, and immune response).

Although there is a dose conversion formula for drugs between animals and human [117], the FDA-approved dose for gadolinium-based MRI contrast agents is 0.1 mmol/kg for human as well as animals, except for the liver-specific gadoxetic acid dose (0.025 mmol/kg) [118].

3.8.1. In Vitro Cellular Toxicity

In vitro cellular toxicity studies should be conducted prior to in vivo animal experiments to ensure that the surface-coated Gd₂O₃ NPs are nontoxic at the cellular level. Experimental evidence indicates that bare Gd₂O₃ NPs are inherently toxic; thus, surface coating with hydrophilic and biocompatible ligands is necessary to reduce cytotoxicity [69,93,119].

Common techniques for assessing cell viability include MTT, WST-1, and CCK-8 assays [120,121]. The MTT assay is a colorimetric method based on the metabolic activity of the cells, whereby mitochondrial dehydrogenases enzymes reduce the tetrazolium salt MTT into water-insoluble formazan crystals. After incubating cells with surface-coated Gd₂O₃ NPs (typically for 48 h at 37 °C), the resulting formazan crystals are solubilized using an organic solvent or detergent, and absorbance is measured at 570 nm. The WST-1 assay operates on a similar principle, but uses the tetrazolium salt WST-1, which is converted into a highly water-soluble formazan product by mitochondrial dehydrogenase enzymes; absorbance is measured at 450 nm following incubation with NP samples. The CCK-8 assay employs the water-soluble tetrazolium salt WST-8, which is reduced by viable cells into a water-soluble formazan product that also absorbs at 450 nm. Among these assays, WST-8 has been reported as the most sensitive.

For example, Ahmad et al. demonstrated the nontoxicity of PAA-coated Gd₂O₃ NPs in normal human embryonic kidney (HEK293) and human liver cancer (HepG2) cell lines [120]. As shown in Figure 9a, PAA-coated Gd₂O₃ NPs exhibited high cell viability in both cell lines up to a Gd concentration of 500 μM. Figure 9b presents the in vitro cytotoxicity results of PAAMA-coated ultrasmall Gd₂O₃ NPs in human prostate cancer (DU145) and normal mouse hepatocyte (NCTC1469) cell lines, showing no toxicity up to 500 µM Gd [61]. Similarly, Cui et al. evaluated the in vitro cytotoxicity of Gd₂O₃-FI-PEG and Gd₂O₃-FI-PEG-BBN NPs in the prostate cancer (PC)-3 cell line [77]. As shown in Figure 9c, cells treated with various Gd concentrations for both 24 and 48 h, demonstrated good viability for both types of surface-coated NPs. These findings consistently confirm the nontoxicity of surface-coated Gd₂O₃ NPs and underscore the importance of hydrophilic and biocompatible ligand surface coatings in mitigating cytotoxicity.

3.8.2. In Vivo Toxicity

Histology

Histological analysis enables the investigation of the in vivo toxicity of ligand-coated Gd₂O₃ NPs in organs and muscles at the cellular levels. To perform histological studies, mice are intravenously (IV) injected with ligand-coated Gd₂O₃ NPs and sacrificed at designated time points after injection to extract relevant organs and muscles. The harvested tissues are sectioned and imaged using an optical microscope, and the results are compared with those of control animals that did not receive NP injections.

For example, Sun et al. conducted a histological assessment of major organs such as the liver, kidney, spleen, and heart 30 days after IV administration of Gd₂O₃@BSA NPs [102]. As shown in Figure 10, no significant alterations or signs of toxicity were observed compared to control tissues, indicating the nontoxicity of Gd₂O₃@BSA NPs.

Body Weight Change

Body weight is another important indicator of in vivo toxicity and is measured using a balance before and at regular intervals after IV injection of surface-coated Gd₂O₃ NPs into mice tails. In the absence of NP-induced toxicity, body weight should remain stable over time. As shown in Figure 11a, no significant weight loss or gain was observed in mice at any of the tested injection doses (25, 50, and 100 mg/kg) compared to the control group without NP injection, indicating that the Gd₂O₃@BSA NPs did not induce toxicity [102].

Biodistribution

Biodistribution studies enable the tracking of injected NPs within the body. A solution of Gd-containing NPs is IV injected into mice tails, which are then sacrificed at specific time points after injection to extract organs and muscle for quantification of accumulated Gd concentrations. The biodistribution profile of injected NPs depends on particle size, morphology, and the nature of surface-coating ligands. Ideally, the NPs should be completely excreted from the body via the renal system, similar to Gd(III)-chelates, as free Gd³⁺ ions are toxic and cannot be eliminated through normal metabolic processes [123].

For example, Dai et al. investigated the biodistribution of PEG-Gd₂O₃ NPs (hydrodynamic diameter = 30–40 nm) in comparison with commercial Magnetist (Gd-DTPA) as control. They measured Gd accumulation in the heart, lung, liver, spleen, and kidneys at 1 and 12 h after IV injection into mice tails (injection dose = 15 μmol Gd/kg) [124]. As shown in Figure 11b, Magnevist exhibited considerably higher accumulation in the kidneys compared to other organs at 1 h post-injection, indicating rapid renal clearance. Additionally, the Gd concentration in all organs was higher for Magnevist than for PEG-Gd₂O₃ NPs, suggesting faster systemic circulation. At 12 h post-injection, Gd levels in all organs for Magnevist were very low, confirming efficient renal excretion. In contrast, PEG-Gd₂O₃ NPs showed considerable hepatic accumulation at 12 h, with Gd levels in the liver approximately 7.53 times higher than at 1 h. While the Gd concentration in the heart decreased over time, it increased in the lung, spleen and kidneys, indicating prolonged circulation and predominant hepatic clearance of PEG-Gd₂O₃ NPs. As another example, Ashouri et al. investigated the biodistribution of β-cyclodextrin-coated Gd₂O₃ (Gd₂O₃@PCD) NPs (d < 100 nm, a = 967.6 nm), using Omniscan (Gd-DTPA-BMA) and Dotarem (Gd-DOTA) as commercial controls in renally impaired rats [125]. As shown in Figure 11c, Gd₂O₃@PCD NPs exhibited only slightly higher Gd accumulation in all organs compared to Omniscan but considerably higher Gd accumulation than Dotarem, 7 days after tail vein injection (total Gd dose: 1.2 mmol Gd/kg for Gd₂O₃@PCD NPs; 30 mmol Gd/kg for Omniscan and Dotarem, administered as 0.2 or 5.0 mmol/kg per week for 6 weeks). These findings indicate that Dotarem was efficiently excreted through the renal system even in renally impaired rats, whereas both Gd₂O₃@PCD NPs and Omniscan showed notable organ accumulation, particularly in the skin. Yang et al. investigated the long-term biodistribution of the Gd(OH)₃ nanorods in mice up to 150 days post-injection [56]. The Gd(OH)₃ nanorods predominantly accumulated in the liver, spleen, and lungs. They showed that most Gd(OH)₃ nanorods were rapidly cleared from the circulating blood, with accumulation in these organs increasing initially and then decreasing over time.

Clearance

Surface-modified Gd₂O₃ NPs should be completely cleared from the body within appropriate time frames after injection owing to their inherent toxicity. Clearance occurs via two primary pathways: renal excretion and hepatobiliary excretion, with the renal excretion being the preferred route. The feasibility of renal excretion primarily depends on particle size and hydrodynamic diameter. Gd₂O₃ NPs with diameters smaller than 5 nm and hydrodynamic diameter below 10 nm can be efficiently excreted through the renal system [123,126,127], whereas larger NPs are typically sequestered by macrophages in the liver and spleen and cleared via the hepatobiliary pathway [128,129,130].

For instance, for ultrasmall NPs (d < 5 nm), Wu et al. synthesized polymaleic acid (PMA)-coated extremely small gadolinium oxide NPs (ES-GON-PMA) (d = 2.1 nm) and evaluated their clearance following IV tail injection at a dosage of 5.0 mg/kg [128]. Gd levels were measured in urine and feces over a 24 h period. As shown in Figure 11d, urinary Gd content decreased rapidly over time, while fecal Gd levels remained negligible, indicating predominant renal excretion due to the ultrasmall particle size of ES-GdON-PMA. In contrast, for large NPs (d > 5 nm), Tian et al. investigated the excretion profile of larger Gd₂O₃@SiO₂ NPs (d = 37.29 ± 1.03 nm) after IV administration into mice tails (15 µmol/kg) [129]. As shown in Figure 11e, excretion was monitored weekly in both feces and urine over a period of 12 weeks. The results exhibited considerably higher Gd content in feces than in urine, indicating that the large NPs were primarily cleared via the hepatobiliary pathway due to their insufficient renal clearance.

Figure 11. (a) Body weight variation over time in mice (N = 4) following injection with different doses of Gd₂O₃@BSA NPs (25, 50, and 100 mg/kg) compared to a control group (no injection), monitored for 30 days [102]. (b) Biodistribution of Magnevist and PEG-Gd₂O₃ NPs in the heart, lung, liver, spleen, and kidney at 1 and 12 h after IV injection (15 μmol Gd/kg) (N = 3; statistical significance p ** < 0.001 compared with the other group in the same organ; ID = percentage of injection dose per gram of organ) [124]. (c) Biodistribution of Gd₂O₃@PCD NPs, Omniscan, and Dotarem 7 days after tail vein injection (0.2 mmol/kg per week × 6 weeks = total 1.2 mmol Gd/kg for Gd₂O₃@PCD NPs; 5.0 mmol/kg per week × 6 weeks = total 30 mmol Gd/kg for Omniscan and Dotarem) (N = 5) [125]. (d) Time-dependent excretion of ES-GdON-PMA-9 (d = 2.1 nm) via urine and feces after IV injection (5.0 mg Gd/kg) (N = 3; I.D.%/g = percentage of injected dose per gram of urine or feces) [128]. (e) Excretion profile of Gd₂O₃@SiO₂ NPs (d = 37.29 ± 1.03 nm) through feces and urine up to 12 weeks after mice tail vein injection (15 µmol/kg) (N = 3) [129].

Figure 11. (a) Body weight variation over time in mice (N = 4) following injection with different doses of Gd₂O₃@BSA NPs (25, 50, and 100 mg/kg) compared to a control group (no injection), monitored for 30 days [102]. (b) Biodistribution of Magnevist and PEG-Gd₂O₃ NPs in the heart, lung, liver, spleen, and kidney at 1 and 12 h after IV injection (15 μmol Gd/kg) (N = 3; statistical significance p ** < 0.001 compared with the other group in the same organ; ID = percentage of injection dose per gram of organ) [124]. (c) Biodistribution of Gd₂O₃@PCD NPs, Omniscan, and Dotarem 7 days after tail vein injection (0.2 mmol/kg per week × 6 weeks = total 1.2 mmol Gd/kg for Gd₂O₃@PCD NPs; 5.0 mmol/kg per week × 6 weeks = total 30 mmol Gd/kg for Omniscan and Dotarem) (N = 5) [125]. (d) Time-dependent excretion of ES-GdON-PMA-9 (d = 2.1 nm) via urine and feces after IV injection (5.0 mg Gd/kg) (N = 3; I.D.%/g = percentage of injected dose per gram of urine or feces) [128]. (e) Excretion profile of Gd₂O₃@SiO₂ NPs (d = 37.29 ± 1.03 nm) through feces and urine up to 12 weeks after mice tail vein injection (15 µmol/kg) (N = 3) [129].

Immunotoxicity

The immunotoxicity assay involves the measurements of (1) ROS in neutrophils from peripheral blood, (2) innate immunity cluster of differentiation (CD) markers such as CD40, CD206, CD11b, and CD80/CD86 in monocytes/macrophages in peripheral blood, (3) adaptive immunity CD markers such as CD69 in lymphocytes from peripheral blood and lymph node, and (4) cytokines including IL-1β, IL-2, and IL-4 in serum [129,131].

For example, Tian et al. evaluated the immunotoxicity of Gd₂O₃@SiO₂ NPs (Gd-NPs; d = 37.29 ± 1.03 nm) by quantifying ROS production, CD maker expression, and cytokine levels in BALB/c mice divided into three groups: PBS (100 µL, negative control), Gd-DTPA (15 µmol/kg), and Gd-NP (15 µmol Gd/kg), with N = 5 for each group [129]. As shown in Figure 12a–c, a statistically significant difference was observed between the Gd-NP and PBS groups in the expression levels of CD206, CD40, IL-1β, IL-2, and IL-4. However, no significant difference were found between the Gd-NP and Gd-DTPA groups, except ROS generation. As shown in Figure 12c, the Gd-DTPA and Gd-NP groups exhibited increased ROS production of 29% and 75%, respectively, relative to the PBS group. Although the Gd-NP group showed the highest ROS generation, this result is not critical, as ROS is a nonspecific mediator of immune response, capable of promoting either pathogen clearance or immunosuppression during tissue repair. Therefore, the data suggest that Gd-NPs induce minimal immunotoxicity. Similarly, Zheng et al. investigated the immunotoxicity of Gd₂O₃:Eu³⁺ NPs (d = 7.4 nm) in BALB/c mice using PBS and Gd-DTPA as controls [131] and obtained comparable results, further supporting that Gd₂O₃:Eu³⁺ NPs exhibit minimal immunotoxicity.

4. Factors Affecting r₁ and r₂ Values

The r₁ and r₂ values of surface-coated Gd₂O₃ NPs are key factors determining their performance as MRI CAs [5]. These values are influenced by several factors, such as particle diameter, surface-coating ligands, solution pH, temperature, and applied MR field strength [64,97,132,133,134,135].

Table 3. r₁ and r₂ values of surface-coated Gd₂O₃ NPs under various conditions, including coating ligands, particle size, hydrodynamic diameter, applied magnetic field, and temperature.

NP	Synthesis Method	Ligand	Size (nm)		r1 (s−1mM−1)	r2 (s−1mM−1)	H (tesla)	T (°C)	Ref.
			TEM	DLS
Gd2O3	polyol	Succinic acid	1.3	4.11	12.5	15.4	1.5	22	[71]
		Glutaric acid	1.3	4.15	13	13.2
		Terephthalic acid	1.3	4.19	11.6	14.4
		PEI-1300	1.3	12.71	8	9.1
		PEI-10,000	1.3	13.87	5.1	7.6
		D-glucuronic acid	2.4		4.25	27.11
		PAA-1200	2	10.3	5.6	36.6	1.5	22	[75]
		PAA-5100	2	11.1	4.7	31
		PAA-15,000	2	11.3	4	29.2
		PEGD-600	0.9		14.18	19.15	1.5	22	[136]
		D-glucuronic acid	1		12.56	12.95
		Lactobionic acid	0.9		11.57	13.38
		D-glucuronic acid	1	4	9.9	10.5	1.5	22	[64]
		DEG		17 ± 2	2.1	2.8	1.5	25	[137]
		PEG-phosphate (reaction in water)		152 ± 20	10	14.2
		PEG-phosphate (reaction in ethanol)		68 ± 11	11.4	14.5
		PEG-polysiloxane	2.2	3.3	8.8	11.4	7	25	[109]
			3.8	5.2	8.8	28.8
			4.6	8.9	4.4	28.9
		D-glucuronic acid	2.4		4.25	27.11	1.5	22	[138]
		Fluorescein-polyethyleneimine	3.92	7.5	6.76	20.27	1.5	22	[139]
		PEG	1.3	9.1	16.2	17.7	0.47		[73]
					14.2	17.2	1.41
					10.9	15.9	7
					10.4	17.2	11.7
		MSN	2.3	62.25	45.08	48.81	0.5		[140]
					16.95		7
		D-Glucuronic acid	1.3	6.2	24 ± 2	60 ± 5	1.5	22	[97]
		PEGD-250		9.7	5 ± 1	55 ± 2
		PEGD-600		12.1	0.1 ± 0.1	10 ± 1
		3,5-Diiodo-L-tyrosine	2		9.24	38.27	1.5		[141]
		Carbon	3.1	18.9	16.26	24.12	1.5	22	[69]
		PAA5100	2	6.3	31	37.4	1.5	22	[51]
		Polyaspartic acid	2	12.7	19.1	53.7	3	22	[63]
		Apoferritin-D-Glucuronic acid	1.9	29.1	8.7	15.7	1.5	22	[142]
		PVP	2.5		10.28	14.47	3		[68]
		DEG	13		1.14	13.5	3		[70]
		TEG	16		2.6	16.4
		TeEG	19		3.99	22.6
		PEG	21		5.75	28.7
		Dextran	1.5	12.4	12.2	29.3	1.5	22	[66]
		Poly(acrylic acid-co-maleic acid) (PAAMA)	1.8	9	40.6	63.4	3	22	[61]
		Poly(methyl vinyl ether-alt-maleic acid) (PMVEMA)	1.9	19.8	36.2	74	3	22	[58]
		PEG	<3	2.8	9.4	13.4	1.5	19.5	[72]
		DEG	<3		6.4	15.2
		PEG	<40		0.1	7.6
Gd2O3	DMSO		4.7	5–10	6.9	7.9	1.41		[53]
		5(6)-carboxyfluorescein-polyethylene glycol-bombesin	52.3	90.6	4.23		3		[77]
		TEG based carboxylate ligand	4	13	6.4	9.1	1.41		[76]
Gd2O3	Thermal decomposition	PVP	2.9	15.7	12.123	33.184	7	36	[78]
		N-Dodecyl-PEI-PEG	9.1	95	14.13		9.4		[55]
		PAA-OA	2	27.8	47.2	82.4	1.41		[80]
			5	28.1	45.6	75.1
			8	31.2	33.7	70.0
			11	33.0	32.4	67.3
			22	47.1	21.1	32.4
		poly (maleic anhydride-alt-1-octadecene) polymer	10 × 1.1	32	16.95		1.5		[81]
		CTX-PEG-TETT	9.09		8.41		7		[79]
Gd2O3	Hydrothermal	Hyaluronic acid	105		6		3		[46]
		CA	100 ± 20	233.7 ± 10	1.8	5.3	1.41	40	[93]

lists r₁ and r₂ values of various surface-coated Gd₂O₃ NPs prepared by different methods and measured under diverse conditions, illustrating the dependence of r₁ and r₂ values on these factors. This dependence highlights the potential to optimize the efficacy of surface-coated Gd₂O₃ NPs as T₁ MRI CAs through controlled tuning of these variables. Notably, as provided in Table 3, r₁ values of surface-modified Gd₂O₃ NPs are considerably higher than 3–5 s⁻¹mM⁻¹ [17,38] of clinically approved T₁ MRI contrast agents, Gd(III)-chelates.

4.1. Particle Diameter

Monodispersed ultrasmall Gd₂O₃ NPs are preferred for T₁ MRI applications because a higher proportion of surface Gd³⁺ ions is available to interact with surrounding water molecules, thereby enhancing the r₁ value. Previous studies have consistently shown that the r₁ value increases with decreasing particle diameter [80,109,143] and that ultrasmall Gd₂O₃ NPs have higher r₁ values than Gd(III)-chelates, suggesting the existence of an optimal particle size for maximizing r₁ value. This was demonstrated by Park et al. [64] and Rahman et al. [143]. According to Solomon-Bloembergen-Morgan theory [38,134], the T₁ relaxation is due to interaction between Gd³⁺ ion and proton spins. As the number of surface Gd³⁺ ions of NPs increases, T₁ relaxation (or r₁ value) increases for ultrasmall d < ~2.5 nm [64,143] because majority of Gd³⁺ ions in NPs exist on NP surface due to high surface-to-volume (S/V) ratios, whereas T₁ relaxation (or r₁ value) decreases as d increases for d > ~2.5 nm because of low S/V ratios at which majority of Gd³⁺ ions in NPs are internal inactive ones for T₁ relaxations.

Rahman et al. reported that the r₁ value of CA-coated Gd₂O₃ NPs increased as particle size decreased below 3.0 nm at 7 T, reaching a maximum value at approximately 2.3 nm; further size reduction led to a decline in r₁ value (Figure 13a) [143]. Park et al. roughly estimated the optimal particle size to lie between 1.1 and 2.5 nm [64]. In another study, the r₁ values of PAA-octylamine-coated Gd₂O₃ NPs (d = 2, 5, 8, 11, and 22 nm) [80] and PEG-polysiloxane-coated Gd₂O₃ NPs (d = 2.2, 3.8, and 4.6 nm) [109] decreased with increasing particle diameter at 1.41 and 7 T, respectively (Figure 13b). This trend likely results from particle diameters exceeding the optimal range. Conversely, the r₂ value, which is influenced by the NP magnetic moment, tends to increase with particle size due to the corresponding increase in magnetic moment per NP [64].

4.2. Surface-Coating Ligands

Surface coating is another critical factor influencing the r₁ and r₂ values of Gd₂O₃ NPs [71,75,97]. Appropriate selection of coating ligands can enhance the accessibility of water molecules near the Gd₂O₃ NP surface, thereby increasing the r₁ and r₂ values.

For example, Tegafaw et al. investigated the relaxometric properties of Gd₂O₃ NPs coated with various ligands, including small diacids with hydrophobic chains (succinic acid, glutaric acid, and terephthalic acid) and large PEIs (PEI-1300 and PEI-10,000) with hydrophilic chains [71]. As shown in Figure 13c, both r₁ and r₂ values decreased with increasing ligand size. This indicates that small, hydrophilic ligands promote closer interaction between water molecules and the NP surface, thereby enhancing r₁ and r₂ values. Similarly, Miao et al. reported that the r₁ and r₂ values of ultrasmall Gd₂O₃ NPs decreased as the molecular weight of PAA (M_w  =  1200, 5100, 15,000 Da) increased (Figure 13c) [75]. Kim et al. also observed a reduction in r₁ and r₂ values as the molecular weight of PEGD (M_w  =  250 and 600 Da) increased (Figure 13c) [97]. These results indicate that smaller, hydrophilic ligands are preferable for achieving higher r₁ and r₂ values.

4.3. Solution pH

The solution pH influences the surface charge and colloidal stability of surface-coated Gd₂O₃ NPs, thereby affecting their r₁ and r₂ values. pH can be adjusted by adding acid or base. For example, in the case of PAA-coated Gd₂O₃ NPs, the carboxyl groups on PAA ligands become protonated under acidic conditions, potentially leading to NP aggregation. Aggregation increases particle size and reduces the surface-to-volume ratios, which may limit water exchange between bulk water and the Gd₂O₃ NP surface, thereby decreasing the r₁ value. Conversely, under basic conditions, the carboxyl groups are deprotonated, resulting in negatively charged surfaces that promote electrostatic repulsion between NPs. This enhanced repulsion improves colloidal stability and increases water accessibility to the NP surface, potentially enhancing r₁ values. In contrast, the r₂ value, which increases with the NP magnetic moment [64], may rise with NP aggregation due to the increase in the magnetic moment per NP. However, excessive aggregation can still compromise colloidal stability, which may in turn limit this effect.

4.4. Temperature

The r₁ and r₂ values are also influenced by temperature, as key parameters such as the diffusion coefficient (D) and rotational correlation time (τ_R) of water molecules and CAs are temperature-dependent [133,134,144,145]. As temperature increases, D increases and τ_R decreases, thereby shortening the interaction time between water proton spins and the NPs. This reduction in interaction time leads to decreased r₁ and r₂ values. For example, Goussuin et al. demonstrated that the r₁ and r₂ values of Gd(OH)₃ NPs decreased with increasing temperature at 1.41 T, as shown in Figure 13d [144].

4.5. Applied Magnetic Field Strength (H)

Roch et al. theoretically investigated the r₁ and r₂ values of superparamagnetic NPs (d = 5 nm) as a function of the applied magnetic field strength (H) or proton Larmor frequency [133]. As shown in Figure 13e, three distinct regions were identified: Region 1, where r₁ and r₂ values remain nearly constant at H < 1 MHz; Region 2, where r₁ and r₂ values increase with H between 1 and 10 MHz; and Region 3, where r₁ and r₂ values decrease at H > 10 MHz (the clinical region).

For example, Carniato et al. synthesized CA-coated GdF₃ NPs (d = 2.2–2.3 nm) and observed three regions in the r₁ nuclear magnetic resonance dispersion (NMRD) profile, as shown in Figure 13f [135], in agreement with theoretical prediction [133]. Additional experimental r₁ value plots as a function of H in Region 3 are displayed in Figure 13g, demonstrating that the r₁ values of various ligand-coated Gd₂O₃ NPs and Gd(III)-chelates decrease with increasing H in this region [73,107,140,146]. Therefore, the proton signal enhancement by T₁ MRI CAs diminishes with increasing H in the clinical region. However, due to enhanced proton spin alignment along H at higher field strengths, resulting in a larger proton spin population difference, the net proton signal enhancement at high H remains considerably greater than that predicted solely from the T₁ MRI CA contribution.

Figure 13. Plot of r₁ values of (a) CA-coated Gd₂O₃ NPs as a function of particle diameter (1.5–194.0 nm) at 7 T [143] and (b) PAA-octylamine- [80] and PEG-polysiloxane- [109] coated Gd₂O₃ NPs as a function of particle diameter at 1.41 and 7 T, respectively. (c) Plot of r₁ values of various ligand-coated Gd₂O₃ NPs as a function of ligand size at 1.5 T [71,75,97]. (d) Plot of r₁ values of Gd(OH)₃ NPs as a function of temperature at 1.41 T [144]. (e) Theoretical NMRD profile of superparamagnetic NPs (d = 5 nm) [133]. (f) NMRD profile of CA-coated GdF₃ NPs (d = 2.2–2.3 nm) [135]. (g) Plots of r₁ values of various ligand-coated Gd₂O₃ NPs and Gd(III)-DTPA as a function of H [73,107,140,146].

Figure 13. Plot of r₁ values of (a) CA-coated Gd₂O₃ NPs as a function of particle diameter (1.5–194.0 nm) at 7 T [143] and (b) PAA-octylamine- [80] and PEG-polysiloxane- [109] coated Gd₂O₃ NPs as a function of particle diameter at 1.41 and 7 T, respectively. (c) Plot of r₁ values of various ligand-coated Gd₂O₃ NPs as a function of ligand size at 1.5 T [71,75,97]. (d) Plot of r₁ values of Gd(OH)₃ NPs as a function of temperature at 1.41 T [144]. (e) Theoretical NMRD profile of superparamagnetic NPs (d = 5 nm) [133]. (f) NMRD profile of CA-coated GdF₃ NPs (d = 2.2–2.3 nm) [135]. (g) Plots of r₁ values of various ligand-coated Gd₂O₃ NPs and Gd(III)-DTPA as a function of H [73,107,140,146].

5. In Vivo MRI Applications

5.1. Imaging in Normal Mice

The FDA-approved dose for gadolinium-based MRI contrast agents is 0.1 mmol/kg for human as well as animals, except for the liver-specific gadoxetic acid dose (0.025 mmol/kg) [118]. Numerous in vivo studies have reported the use of surface-coated Gd₂O₃ NPs as T₁ MRI CAs, demonstrating their strong potential as T₁ MRI CAs. For example, Ahmad et al. investigated PMVEMA-coated ultrasmall Gd₂O₃ NPs (d = 1.9 nm) as a T₁ MRI CA [58]. Following IV administration of the aqueous NP solution into mice tails, positive (i.e., bright) contrast enhancement was observed in the liver and kidneys, as shown in Figure 14a. Signal-to-noise ratios (SNRs) of regions of interest (ROIs) in these organs initially increased, reached maxima, and then decreased over time, indicating accumulation followed by excretion of the NPs, predominantly through the renal pathway, thus confirming their role as T₁ MRI CAs (Figure 14b). In another example, Yue et al. reported carbon-coated Gd₂O₃ (Gd₂O₃@C) NPs (d = 3.1 nm) as a T₁ MRI CA [69]. Upon IV administration into mice tails, the NPs produced positive contrast enhancement in the liver, kidneys, and bladder (Figure 14c). The SNR-ROIs in these organs, plotted as a function of time (Figure 14d), showed that the contrast enhancements peaked between the point of administration and 30 min post-injection, and subsequently declined. These results further confirm the function of the NPs as a T₁ MRI CA.

5.2. Imaging in Cancer Model Mice

Owing to the EPR effects, NPs can accumulate more readily in cancer cells than in normal cells (called passive targeting) [147]. Furthermore, the accumulation of NPs in cancer cells can be considerably enhanced by conjugating cancer-targeting ligands onto the NP surface (called active targeting).

Dai et al. showed that PEG-Gd₂O₃ NPs (r₁ = 29.0 s⁻¹ mM⁻¹) displayed stronger contrast enhancement of the tumor in in vivo MR images than Magnevist (r₁ = 4.2 s⁻¹ mM⁻¹) at the same amount of Gd injection, demonstrating the superiority of PEG-Gd₂O₃ NPs to commercial molecular contrast agent Magnevist [124]. Shen et al. evaluated T₁ MR images of U87MG tumor model nude mice using exceedingly small (ES), RGD2-conjugated PAA-coated Gd₂O₃ (simply referred to as ES-GON5-PAA@RGD2) NPs [107]. As shown in Figure 15a, the tumor region exhibited a very weak MRI signal prior to IV injection. Following injection, signal intensity increased markedly, peaking at 2 h post-administration. Quantitative analysis in Figure 15b revealed a maximum ΔSNR [=100% × (SNRpost − SNRpre)/SNRpre] of 372 ± 56%, which is considerably higher than the ΔSNR of <80% of typically observed value with conventional Gd(III)-chelates. This remarkably enhanced tumor ΔSNR can be attributed to the combined effects of the high relaxivity value of the NPs (r₁ = 68.7 ± 2.3 mM^–1s^–1, r₂/r₁ = 1.03 ± 0.03 at 1.5 T), the active targeting capability of RGD2 toward integrin αvβ3-overexpressed on tumor cells, and the EPR effect.

6. Conclusions and Future Outlook

Considerable research progress has been made on surface-coated Gd₂O₃ NPs as promising T₁ MRI CAs. However, no comprehensive review on this subject has been reported to data. Therefore, a thorough overview addressing current status and future perspectives of surface-coated Gd₂O₃ NPs as T₁ MRI CAs is highly warranted. Such a review would substantially advance research in this field and assist researchers in understanding the current state of the art and identifying future research directions.

This review has covered synthesis methods (i.e., polyol, DMSO, thermal decomposition, and hydrothermal techniques); characterization techniques (i.e., TEM, DLS, XRD, FTIR, TGA, zeta potential, VSM, and MRI); in vitro cellular and in vivo toxicity assessments (i.e., histology, biodistribution, body weight change, clearance, and immunotoxicity); and in vivo MRI applications of surface-modified Gd₂O₃ NPs. In addition, the influence of various factors (i.e., particle diameter, surface-coating ligands, solution pH, temperature, and applied MR field) on performance parameters of MRI CAs (i.e., r₁ and r₂ values) was discussed, as this information is critical for the rational design of optimized Gd₂O₃ NPs for specific applications. By comparing the strengths and limitations of each synthesis method, researchers can select the most appropriate approach for their intended application. The characterization protocols described herein may also serve as standard procedures for studies. Although previous studies have demonstrated the strong potential of surface-modified Gd₂O₃ NPs as T₁ MRI CAs, further development of advanced formulations that fully meet in vivo toxicity criteria is essential for clinical translation. For instance, nontoxic, renally excretable formulations that prevent the release of free Gd³⁺ ions should be pursued.

While remarkable progress has been achieved in surface-modified Gd₂O₃ NPs, successful translation to clinical trial phases demands standardized methodologies. Future research should focus on conducting comprehensive toxicity assessments, including acute, sub-chronic, and chronic studies aligned with Food and Drug Administration (FDA) and European Medicines Agency (EMA) guidelines [148,149]. Standardized protocols for evaluating renal clearance and nephrotoxicity are also needed for clinical trial phases. Additionally, scalable manufacturing processes compliant with Good Manufacturing Practices (GMPs) are necessary for commercialization. GMP-compliant scale-up strategies involve a systematic approach to increasing production volume while maintaining product quality and regulatory compliance. Key aspects include facility design, process validation, quality control, supply chain management, and workforce training; all tailored to meet GMP requirements.

The surfaced-modified Gd₂O₃ NPs exhibited higher r₁ values compared to surfaced-modified Mn and Fe-based NPs owing to the higher electron spin magnetic moment of Gd³⁺ (s = 7/2) compared to Mn²⁺ (s = 5/2) and Fe³⁺ (s = 5/2). The only disadvantage of surfaced-modified Gd₂O₃ NPs compared to surfaced-modified Mn- and Fe-based NPs is their higher potential toxicity. Therefore, advancement in biocompatible ligand coatings is critical to ensure the safety profile of surfaced-modified Gd₂O₃ NPs. Several studies assessed colloidal stability of surface-modified lanthanide oxide NPs in water, phosphate-buffered saline (PBS), or serum-containing media [51,69,150], demonstrating good colloidal stability overtime and thus, good coating stability.

Over the past decade, extensive efforts have been devoted to developing new T₁ MRI CAs as alternatives to Gd(III)-chelates, which suffer from low sensitivity due to their low r₁ values, as well as non-specificity and short imaging windows resulting from rapid blood clearance. These limitations often necessitate the administration of large doses. Surface-modified Gd₂O₃ NPs have emerged as promising alternatives, offering improved performance by overcoming the shortcomings of Gd(III)-chelates. Moreover, surface-modified Gd₂O₃ NPs can be easily conjugated with cancer-targeting ligands and therapeutic drugs, enabling targeted theranostic applications. They can be further functionalized via integration with other imaging agents such as FI, CT, PET, and SPECT agents to promote their application as multimodal imaging agents. While many of these advantages have been demonstrated in in vivo animal studies, considerable challenges remain in translating these systems into clinical practice. As discussed above, the development of advanced formulations and comprehensive in vivo toxicity evaluations are critical next steps. Achieving clinical translation will require sustained interdisciplinary research efforts.