Stability of MRI contrast agents in high-energy radiation of a 1.5T MR-Linac

Abstract

Background:

Gadolinium-based contrast is often used when acquiring MR images for radiation therapy planning for better target delineation. In some situations, patients may still have residual MRI contrast agents in their tissue while being treated with high-energy radiation. This is especially true when MRI contrast agents are administered during adaptive treatment replanning for patients treated on MR-Linac systems.

Purpose:

The purpose of this study was to analyze the molecular stability of MRI contrast agents when exposed to high energy photons and the associated secondary electrons in a 1.5T MR-Linac system. This was the first step in assessing the safety of administering MRI contrast agents throughout the course of treatment.

Materials and methods:

Two common MRI contrast agents were irradiated with 7 MV photons to clinical dose levels. The irradiated samples were analyzed using liquid chromatography-high resolution mass spectrometry to detect degradation products or conformational alterations created by irradiation with high energy photons and associated secondary electrons.

Results:

No significant change in chemical composition or displacement of gadolinium ions from their chelates was discovered in samples irradiated with 7 MV photons at relevant clinical doses in a 1.5T MR-Linac. Additionally, no significant correlation between concentrations of irradiated MRI contrast agents and radiation dose was observed.

Conclusion:

The chemical composition stability of the irradiated contrast agents is promising for future use throughout the course of patient treatment. However, in vivo studies are needed to confirm that unexpected metabolites are not created in biological milieus.

Magnetic resonance imaging (MRI) is increasingly being used throughout the radiation therapy process, from treatment planning to response assessment [1–4]. Some types of scans require the use of MRI contrast agents for better visualizing and delineating the extent of the disease to be treated [5]. The mean half-lives for distribution and elimination of such agents are typically 30–90 min for patients with normal renal function but can be much longer (up to 30–40 h) for patients with poor renal function [6,7]. Thus, in theory, some residual MRI contrast agent could still be present in patients who receive radiation therapy shortly after MRI scanning. Furthermore, integrated MR-guided radiation therapy systems (MR-Linac) are increasingly being used in clinical practice [8–12], and MRI contrast administration can aid in patient replanning for adaptive treatment [13]. In such conditions, the contrast agent is still in the patient’s system during subsequent irradiation, and thus the stability and potential toxicity of contrast agents in these circumstances have become critical issues in patient care and safety in radiation oncology [14]. The presence of strong magnetic fields in the MR-Linac systems during irradiation with high-energy radiation photons (as well as secondary scattered electrons) may have additional unknown toxicologic effects via chemical alteration of the contrast agents.

Most of the MRI contrast agents approved for clinical use in the United States are based on the lanthanide metal gadolinium (Gd). In their chelated form, these Gd-based contrast agents (GBCA) are considered safe and lead to adverse reactions in less than 1% of cases [15,16]. Affinity of Gd for its chelating ligand at equilibrium is expressed by thermodynamic and conditional stability. The stability of ionic compounds is greater than non-ionic compounds due to the electrostatic interactions between the metal and ligand through their carboxylate groups [17]. Moreover, the stability of macrocyclic compounds is greater compared with linear compounds as described by the macrocyclic effect of wrapping the chelating atoms tightly around the Gd³⁺ ion. This requires the near simultaneous breaking of all Gd-N bonds to release the Gd³⁺ ion whereas linear bonds can be broken sequentially [18]. Lastly, the kinetic inertia, which describes the rate of dechelation, is significantly improved in macrocyclic compounds compared with linear compounds, often preventing macrocyclic compounds from reaching thermodynamic equilibrium in biological settings prior to excretion [19].

Fe³⁺, Zn²⁺, Cu²⁺, and Ca²⁺ are competing cations for the Gd³⁺ chelating ligand, which can lead to the release of free Gd³⁺ through transmetallation [20]. Free Gd³⁺ ions are insoluble and are potent calcium antagonists. Thus, they can block the activation of voltage-gated calcium channels, which are involved in skeletal and cardiac muscle contraction and nerve impulse transmission [21,22]. Gd³⁺ ions can also be precipitated by endogenous anions such as phosphate, hydroxide, and carbonate and deposited into the skin, bone, and other organs, leading to interstitial fibrosis [17]. It has been suggested that Zn²⁺ is the only endogenous cation that can lead to significant transmetallation in vivo. The endogenous concentration of Cu²⁺ and the affinity of calcium for chelating ligands are both too low to pose serious concerns [23]. Thus, chelation stability is the primary safety consideration for the use of GBCA, especially in patients with compromised renal function.

The most serious side effect associated with GBCA dechelation is nephrogenic systemic fibrosis (NSF), which was first described in 2000 as the thickening and hardening of the skin in the extremities of renal dialysis patients and reminiscent of scleromyxoedema [24]. In extreme cases, this can lead to impaired movement and disability. The first reports potentially linking GBCA administration and NSF in patients with renal impairment was published in 2006 by two institutions, independently [25,26]. NSF symptoms were reported in end-stage renal disease patients 2–4 weeks post-administration of the nonionic linear GBCA Magnevist or Omniscan. Reports of Gd deposition in skin biopsies of NSF patients further implicated GBCA as a contributing factor to NSF [27,28]. Additional data suggested that cumulative Omniscan dose correlated with NSF severity [29]. GBCA with linear chelates have lower thermodynamic stability and shorter dissociation lifetimes compared to those with macrocyclic chelates. This makes them more likely to undergo transmetallation, and as a result, they were alleged to be a contributing factor in majority of NSF cases [30]. While rare, a small number of NSF cases included patients who were administered macrocyclic Gd-CA, though the link to NSF is much less convincing than that of linear GBCA [31–33].

So far, very little research has been conducted on the physical effects of high-energy radiation on GBCA in vivo. It has been hypothesized that high-energy radiation can potentially break the chemical bonds between Gd ions and their ligands, releasing Gd³⁺ ions from their chelates and leading to similar toxic side effects as transmetallation. One study used nuclear magnetic resonance spectroscopy to evaluate the effects of radiation and temperature on the imaging characteristics (relaxivities) of GBCA [34]. For a dose of 20 Gy, r₁ and r₂ relaxivity of the GBCA were altered by <1%, which is an order of magnitude less than the effects of varying the temperature by 4 K. The authors postulated that these changes in relaxivities were caused by the production free Gd³⁺ ions at a rate of <1 μM/Gy. However, the concentrations of free Gd³⁺ were not experimentally verified, and changes in GBCA structure or potential toxicity were not evaluated.

Therefore, additional research into the molecular effects of high-energy radiation on chelated GBCA needs to be conducted. The International Society for Magnetic Resonance in Medicine (ISMRM) has called for additional data-driven research investigating the safe use of GBCA, including during treatments such as chemotherapy and radiation therapy [35]. In order for physicians to make informed decisions on the administration of GBCA prior to radiation therapy, the identity of any chemical products and their biological effects should be characterized. As a first step in addressing this gap in knowledge, we used mass spectrometry to detect possible new compounds formed by high-energy radiation (at doses commonly used in radiation therapy) of two commonly used MRI GBCA. One ionic linear and one macrocyclic compound were chosen to compare the effects of radiation between both categories of GBCA.

Materials and methods

In this prospective study, we studied two MRI GBCA commonly used in the authors’ clinic: macrocyclic GBCA Gadavist (gadobutrol [Gd-DO3A-butrol]; Bayer HealthCare Pharmaceuticals, Whippany, NJ, USA) and ionic linear GBCA MultiHance (gadobenate dimeglumine [Gd-BOPTA]; Bracco Diagnostics, Monroe Township, NJ, USA). Both agents are usually administered in doses of 0.1 mmol per kg of body weight for clinical use.

MRI contrast agent sample preparation

Gadavist was supplied in solution of 604.72 mg of gadobutrol per 1 mL (or 1 mmol/mL). The typical patient dose is 0.1 mmol/kg of body weight, so the dose for a representative 80-kg adult patient with a blood volume of 5100 mL would be approximately 8 mmol of Gadavist. Assuming that the injected compound is evenly distributed over the entire circulating blood volume, the final Gadavist concentration would be approximately 1.6 mmol/L, or 1.3 mg/mL. To approximate this concentration, the original solution was diluted with 5.5 × 10⁻⁶ S/m (18 MΩ cm) deionized water at a ratio of 1:620.

MultiHance was also supplied in a solution of 529 mg gadobenate dimeglumine salt per 1 mL (or 0.5 mmol/mL). The recommended patient dose is the same as for Gadavist (0.1 mmol/kg of body weight), so an 80-kg patient would be given 8 mmol of MultiHance for a circulating blood concentration of 1.6 mmol/L, or 1.7 mg/mL. To approximate this concentration, the original solution was diluted with deionized water at a ratio of 1:310.

Irradiation conditions

The diluted contrast agent samples were placed in separate vials (2-mL clear glass liquid chromatography-mass spectrometry [LC-MS] vials [#5183–2067], Agilent Technologies, Santa Clara, CA, USA). Each vial was placed at the isocenter of a 1.5-T Unity MR-Linac (Elekta, Crawly, UK) and irradiated with 7 MV photons to doses of approximately 2 Gy, 8 Gy, 15 Gy, or 30 Gy; two non-irradiated vials were used as baselines for comparison (controls).

Liquid chromatography-high resolution mass spectrometry (LC-HRMS) analysis of radiation-induced changes in contrast agents

Both irradiated and control samples of Gadavist (gadobutrol) and MultiHance (gadobenate dimeglumine) were diluted to a final concentration of 10 μM in 50% (v/v) acetonitrile/water and analyzed by liquid chromatography-high resolution mass spectrometry (LC-HRMS). The LC-HRMS analyses were done on a Waters Acquity I-Class ultraperformance liquid chromatography (UPLC) system (Waters Corp, Milford, MA, USA) coupled with a linear-trap quadrupole (LTQ) Orbitrap Elite mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) operated in electrospray ionization mode (ESI+). Detection conditions of the mass spectrometer were as follows: Heat Temp 375 °C, Sheath Gas Flow Rate 45, Aux Gas Flow Rate 10, Sweep Gas Flow Rate 3, I Spray Voltage 4.10 kV, Capillary Temp 320 °C, and S-Lens RF Level 55%. The contrast solutions were separated by using a Waters Acquity UPLC HSS T3 Column (1.8 μm, 2.1 mm × 100 mm). The injection volume was 5 μL. The LC mobile phase A was 0.1% acetic acid-water and phase B was 0.1% acetic acid-acetonitrile. The gradient was 0% B (0–2 min), 5–100% B (2–17 min), 100% B (17 to 20 min), 100–0% B (20–22 min), and 0% B (22 to 24 min), and the flow rate was 0.1 mL/min (for Gadavist) or 0.2 mL/min (for MultiHance). The column temperature was 40 °C. The photodiode array wavelength was set at 190–450 nm. Under these conditions, the retention times were 2.5 min for Gadavist and 7.1 min for MultiHance.

Additional LC-HRMS quantification was performed on a spectrometer equipped with a more sensitive detector to verify the findings from the primary analysis. The system included an Agilent 1200 HPLC equipped with a Jupiter 5 μm C18 300 Å guard column (Phenomenex, Torrance, CA) and a Kinetex 2.6 μm C18 100 Å 100×2.1 mm (Phenomenex), the pump (G1312B), Autosampler (G1329A) and evaporative light scattering detector (ELSD, G4260B). The injection volume was 2 μL. Column was thermostated at 54 °C. Mobile phase A was 1.5 mM aqueous trifluoroacetic acid (TFA, Acros); phase B - acetonitrile (BDH) with 1.5 mM TFA. Mobile phase was delivered at a flow rate of 0.5 mL/min with gradient program: 99%A in 0–0.25 min, 99% to 100%A in 0.25–1 min, 100% to 0%A in 12 min. ELSD settings were: 1.6 SLM for N₂ nebulizer gas, 65 °C for nebulizer and 80 °C for evaporator, a sampling rate of 80 Hz, signal smoothing parameter of 30, LED power of 100% and a photomultiplier tube gain set to 1. To avoid the carry-over on subsequent injections, injection needle was cleaned after each injection in a wash-vial filled with 50% aqueous methanol. All analyses were repeated at least 5 times with consistent peak parameters. All data were acquired and processed with OpenLab C.01.10 software (Agilent).

Statistical analysis

The peaks associated with Gadavist and MultiHance in the LC-HMRS chromatograms were integrated to quantify the concentration of the contrast agent for each of the irradiated conditions. GraphPad Prism (Version 8, GraphPad Software, San Diego, CA) was used for all statistical analysis. For each contrast agent, the average and standard deviation of the peak areas across the irradiation conditions was calculated along with the coefficient of variation (CV) to measure the dispersion of the values. The correlation of peak area with dose was measured by calculating the Pearson correlation coefficient, and correlations that produced two-tailed p-values <0.05 were deemed significant.

Results

The purity of Gadavist (gadobutrol) and MultiHance (gadobenate dimeglumine) samples were analyzed by LC-HRMS 7 days after irradiation to the dose levels mentioned above (2 Gy, 8 Gy, 15 Gy, or 30 Gy) and compared with that of both reference and unirradiated samples.

A total of five detectable peaks (P1–P5) were observed on each of the LC-HRMS chromatograms of the reference, unirradiated, and irradiated Gadavist samples (Fig. 1A). Qualitatively, the LC-HRMS profiles of the reference and unirradiated samples were identical to that of the irradiated samples. The mass spectra of the total ions of each individual Gadavist sample (Fig. 2) indicated that none of these samples showed any detectable new ions following irradiation. Among the five detectable peaks (P1–P5) observed on the LC-HRMS chromatograms (Fig. 1A), P3 was identified as Gadavist (calculated m/z = 606.1425) (Fig. 3), and P1, P2, P4, and P5 (Fig. S1–4) were found to be unrelated to Gadavist (confirmed by spectral analysis of dry runs, where no GBCA was injected). These unrelated peaks may be explained by the presence of UV-absorbing impurities in HPLC solvents (which were all HPLC grade) which may concentrate when applying the mobile phase gradients. Another group of peaks originates from a gradual column deterioration due to accumulation of impurities on column silica gel. Thus, returning to the LC-HMRS chromatograms, the area under peak P3 was calculated for the unirradiated and irradiated samples (Fig. 1B). Average peak area was 5,751,672 ± 520,986 (CV = 0.09), and there was no significant correlation between dose and peak area (R² = 0.00, p = 0.90). Collectively, these results indicate that there is no molecular change in Gadavist as a result of irradiation to clinical dose levels.

Fig. 1.

Liquid chromatography-high-resolution mass spectrometry (LC-HMRS) chromatograms of Gadavist samples exposed to various doses of radiation. The entire profile is displayed in (A) where 5 peaks are visible. Peak P3 was identified as Gadavist, which is enlarged and quantified in (B).

Fig. 2.

Mass spectra of total ions (m/z range 100–1000) identified in Gadavist samples exposed to various doses of radiation.

Fig. 3.

Mass spectra of peak 3 (P3), which was identified as the Gadavist peak, for Gadavist samples exposed to various doses of radiation.

A total of four detectable peaks (P1–P4) were observed on each of the LC-HRMS chromatograms of the reference, unirradiated, and irradiated MultiHance samples (Fig. 4A). Similar to the results from the Gadavist irradiation, the LC-HRMS profiles of the reference and unirradiated Multihance samples were comparable to that of the irradiated samples. The total ions of each individual MultiHance sample (Fig. 5) indicated that none of these samples showed any detectable new ions following irradiation. Among the four detectable peaks (P1–P4) observed on the LC-HRMS chromatograms (Fig. 4A), P2 was identified as MultiHance (calculated m/z = 669.1038) (Fig. 6), and P1, P3, and P4 (Fig. S5–7) were found to be unrelated to MultiHance. Returning to the LC-HMRS chromatograms, the area under peak P2 was calculated for the unirradiated and irradiated samples (Fig. 4B). Average peak area was 3,257,574 ± 106,097 (CV = 0.03), and there was no correlation between dose and peak area (R² = 0.11, p = 0.52). LC-UV chromatograms were also obtained for the reference, unirradiated, and irradiated MultiHance samples (Fig. S8), which produced one ultraviolet peak that was identified as MultiHance. This peak area was calculated for the unirradiated and irradiated samples, which are listed in Fig. 4B. Average peak area was 0.053 ± 0.003 (CV = 0.07), and there was no correlation between dose and peak area (R² = 0.12, p = 0.50). Thus, these results from both LC-HRMS and LC-UV indicate that there is no molecular change in MultiHance as a result of irradiation to clinical dose levels.

Fig. 4.

Liquid chromatography-high-resolution mass spectrometry (LC-HMRS) chromatograms of MultiHance samples exposed to various doses of radiation. The entire profile is displayed in (A) where 5 peaks are visible. Peak P2 was identified as MultiHance, which is enlarged and quantified in (B).

Fig. 5.

Mass spectra of total ions (m/z range 100–1000) identified in MultiHance samples exposed to various doses of radiation.

Fig. 6.

Mass spectra of peak 2 (P2), which was identified as the MultiHance peak, for MultiHance samples exposed to various doses of radiation.

In order to verify the results from this primary analysis, each of the Gadavist and MultiHance samples were reanalyzed with LC-HRMS using a spectrometer equipped with a more sensitive detector (ESLD). Chromatographic peaks were characterized by the retention time (t_R) and quantified by peak area (A). For the Gadavist samples, only a single peak was observed, with average t_R = 0.919 ± 0.006 min (Fig. S9). Average peak area was 1974 ± 44.21 (CV = 0.02), and there was no correlation between dose and peak area (R² = 0.03, p = 0.79). In the MultiHance samples, two peaks were observed: t_R1 = 0.840 ± 0.002 (Gd gadobenate), t_R2 = 5.28 ± 0.022 (meglumine) (Fig. S10). The last peak was found to be identical to the peak observed with injection of the solution of meglumine sample (Alfa Æsar). Average peak area of Gd gadobenate was 1637 ± 10.26 (CV = 0.01), and there was no correlation between dose and peak area (R² = 0.00, p = 0.87).

Additionally, Gd chloride (Strem) and trifluoroacetate (prepared from gadolinium(III) carbonate hydrate, Gd₂(CO₃)₂, Alfa Æsar, and trifluoroacetic acid) were subjected to HPLC analysis under the aforementioned conditions. Both produced an identical peak with t_R = 0.623 ± 0.014 and w_1/2 = 0.059. No peak with similar retention time and half-width was identified in any of analyzed samples. The results from this secondary analysis confirmed the initial results found in the primary analysis.

Discussion

The use of Gd contrast agents (GBCA) in MRI to help delineate tumors and define the target region has become common practice in many radiation oncology clinics. These GBCA may still circulate in the patient’s bloodstream at the time of treatment for those with poor renal function, or in the cases when they are administered during on-the-fly adaptive replanning in MR-Linac systems. Because the stability of these GBCA under high-energy irradiation is poorly understood, we performed mass spectrometry analysis to determine if molecular products were created as a result of the irradiation.

There were two main takeaways from the results of these experiments. Qualitatively, there was no visible change in the liquid chromatography-high resolution mass spectrometry (LC-HRMS) and liquid chromatography ultraviolet (LC-UV) chromatograms between the irradiated samples and unirradiated/reference samples for either GBCA. The absence of new peaks in the irradiated samples suggests that no significant amounts of degradation compounds or conformational alterations were created as a result of the high-energy irradiation. This was also observed in the mass spectra of total ions and the mass spectra of individual peaks. Second, the quantification of the Gadavist/MultiHance peak areas was uniform among all the samples, and there was no significant correlation of peak area with dose. The consistency between the peak areas and lack of negative correlation with dose suggests that the concentration of the contrast agent was not significantly reduced due to potential conversion to new products. This was verified with secondary LC-HRMS analysis with a more sensitive detector. Collectively, the chemical composition of the GBCA tested in these experiments were generally stable in their original solution environment when exposed to high-energy irradiation in the presence of a strong magnetic field, which had minimal and insignificant effects.

Our experimental results are consistent with the only study that we could find regarding the effects of radiation on GBCA, which found a change in relaxivity of <1% [34]. The authors of that study proposed that any changes caused by radiation would be several orders of magnitude lower than changes caused by temperature. We studied the effect of irradiation from the perspective of chemical composition changes using mass spectrometry, which should theoretically be more direct and more sensitive. Our findings were consistent in that the effect of radiation, if any, was insignificant. There have also been studies that explored the effect of repeat GBCA administration on previously irradiated tissue with mixed results. It is hypothesized that blood-brain barrier (BBB) breakdown, due to a disease or irradiation thereof, leads to increased Gd deposition in the brain, which is detected by signal hyperintensities in non-contrast-enhanced exams. While some studies reported a positive correlation of signal hyperintensities in the brain with radiation history [36,37], others suggested that this increase in potential Gd deposition was more closely correlated with the number of contrast-enhanced exams rather than with radiation exposure [38,39]. As of now, there are no peer-reviewed data that links the signal hyperintensities reported in these cases with any toxicities or adverse effects [40,41]. These studies are inherently different than our study, as the authors researched the effect of tissue irradiation on Gd deposition, rather than irradiation of the GBCA themselves. Furthermore, Gd-based nanoparticles have been studied as a potential theragnostic [42,43]. In addition to their contrast-enhancing properties, the high atomic number, and thus increased photoabsorption cross section, of Gd leads to increased emission of Auger electrons, which enhances the radiosensitivity of surrounding tissues in a similar manner as gold nanoparticles [44–46]. While this effect is desired in the nanoparticle theragnostic agents, it could be an undesired consequence of residual GBCA in the vascular and extracellular space of radiation therapy patients. However, the range of these Auger electrons (and resulting enhanced radiosensitivity) is approximately 5 nm [47]. Thus, intracellular localization of Gd and close proximity to cell nuclei is necessary for radiosensitization [48]. Unlike nanoparticles, GBCA remain extracellular in their chelated form, so enhanced radiosensitivity to off-target tissue is negligible [49]. Nonetheless, these studies are an important when considering the safety of administering GBCA concurrently with radiation therapy.

There has been a vast amount of published literature which report the chemical stability of linear and macrocyclic GBCA with regards to transmetallation. A variety of methods including pH-potentiometry, spectrophotometry, nuclear magnetic resonance, capillary electrophoresis, high performance liquid chromatography, inductively coupled plasma atomic emission spectroscopy, and mass spectrometry have all been utilized to measure thermodynamic stability and kinetic inertia of GBCA and the release of free Gd³⁺ ions in vitro in aqueous solutions with Fe³⁺, Zn²⁺, Cu²⁺, and Ca²⁺ ions as well as plasma models with endogenous metals [50–58]. The shared findings between these studies indicated that transmetallation products were more readily detected in nonionic linear GBCA compared with ionic linear GBCA, and that negligible transmetallation was evident in macrocyclic GBCA. These findings were corroborated in ex vivo animal biodistribution studies [18,59,60] and patient biopsy and urine samples [27,28,61,62]. Several international prospective trials have further reported that NSF presentation was linked to nonionic linear GBCA administration in renally-impaired patients and not ionic linear or macrocyclic GBCA [31,63–65].

Since 2010, the use of linear GBCA Magnevist, Omniscan, and OptiMARK has been restricted in patients with AKI or with chronic, severe kidney disease due to their link to NSF [66,67]. In 2017, the European Medicines Agency confirmed recommendations to suspend the use of Magnevist, Omniscan, and OptiMARK completely and restrict the use of MultiHance and Primovist to liver scans due to their propensity for dechelation and accumulation in tissues including the brain [68]. In 2018, the United States Food and Drug Administration issued a new class warning and patient medication guide for all GBCA which acknowledged Gd retention and accumulation, but because no direct link of adverse health effects in patients with normal renal function had been established, no new restrictions were mandated [69].

Although the experiments in this study were performed on the MR-Linac, they are generally applicable to patients, who were recently administered GBCA, treated on conventional linear accelerators. However, it is important to note that the presence of the 1.5T magnetic field of the MR-Linac does lead to dosimetric differences. High-energy photons will generate secondary electrons upon interacting with tissue. In a magnetic field, these electrons will experience a Lorentzian force and make a curved trajectory, which in some circumstances leads to the electron-return effect (ERE) [70,71]. This can lead to discrepancies in dose of up to 20% around air-tissue interfaces, when compared with conventional linear accelerator treatments [72]. The range of radiation doses measured in this study is large enough to encompass clinical dose levels on conventional linear accelerators, even with dose discrepancies due to the ERE.

One caveat to this study is that the time between irradiation and the mass spectrum measurement in this study was not immediate (approximately 7 days). Thus, the possibility still exists that any chemical composition changes in response to the irradiation in strong magnetic fields are transient or not stable for an extended period of time, and they may already have reversed during the interval between the irradiation and the time of mass spectrometry analysis. Thus, in future experiments, a range of times between irradiation and analysis will need to be evaluated. Furthermore, questions about the effects of radiation in vivo remain unanswered, particularly if there are metabolites or other derivatives from endogenous enzymes/compounds. In vitro models provide useful theoretical calculations of stability and selectivity for approximate estimates of in vivo behavior. However, the numerous published studies incorporate a variety of experimental conditions and analytical techniques which leads to some uncertainty on the absolute values of these constants. Furthermore, no in vitro model can fully capture all of the intermediate and alternative equilibriums that take place with endogenous molecules [17]. For example, the extent of observed transmetallation can vary significantly between aqueous solutions containing metal ions and human plasma containing endogenous metals and metalloenzymes [57]. This preliminary study was a necessary first step to assess molecular changes of GBCA in response to high-energy irradiation. However, in order to gauge the full physiologic effects of irradiated GBCA and any associated influences on transmetallation in vivo, subsequent investigations will include full-scale toxicity assessments in animal models.

In conclusion, our mass spectrometry analysis indicated that MRI contrast agents were minimally affected by high-energy radiation in the presence of strong magnetic fields. Further studies are needed to confirm the stability and any transient effects on these agents in vivo by full toxicity studies.

Abbreviations:

Gd

gadolinium

GBCA

gadolinium-based contrast agents

LC-HRMS

liquid chromatography-high-resolution mass spectrometry

LC-UV

liquid chromatography-ultraviolet

ELSD

evaporative light scattering detector