Superparamagnetic iron oxides as positive MR contrast agents: In vitro and in vivo evidence

Magnetic Resonance Imaging, Vol. I I , pp. 509-519, 1993 Printed in the USA. All rights reserved.

0730-725X/93 $6.00 + .00 Copyright O 1993 Pergamon Press Ltd.

Original Contribution SUPERPARAMAGNETIC IRON OXIDES AS POSITIVE MR CONTRAST AGENTS: IN VITRO AND IN VIVO EVIDENCE CATHERINE CHAMBON, * OLIVIERCLEMENT, t ALAINLE BLANCHE, t ELISABETH SCHOUMAN-CLAEYS,~ AND GUYFRIJA? *Laboratoire Guerbet, 16/24 rue Jean Chaptal, 93601 Aulnay-Sous-Bois,France ?Service de Radiologie, HBpital Laennec, 42 rue de Skvres, 75007 Paris, France The ability of superparamagnetic iron oxides (SPIO) and ultrasmall superparamagnetic iron oxides (USPIO) to act as positive contrast enhancers due to a marked TI relaxivity was investigated. At low concentrations, an important signal enhancement was observed in vitro, reaching 120% for SPIO and 140% for USPIO in a spin echo 500/22 sequence. The more heavily the sequence was TI-weighted the greater the enhancement. As the concentration increased, the signal dropped. The in vivo study of USPIO in the rat showed that at low doses (14 prmol Fe/kg), the myocardial signal was enhanced by 30070, whereas at high doses (77 pmol Fe/kg), it fell by -50%. These results indicate that in TI-weighted spin echo sequences, the MR signal can be enhanced by low concentrations of superparamagnetic compounds. This effect could be useful in perfusion imaging, and is also important for a better understanding of any possible paradoxical positive enhancement which could occur in perfused organs.

Keywords: Iron; Magnetic resonance, contrast enhancement; Heart, experimental study; Heart, MR.

INTRODUCTION

myocardial perfusion imaging where the circulating magnetic mass may alter the image of the myocardial wall. The aim of this study was therefore to investigate under what conditions positive signal enhancement could be achieved after administration of SPIO or USPIO. This paper reports on:

Due to the marked T2relaxivity of superparamagnetic iron oxides (SPIO) and ultrasmall superparamagnetic iron oxides (USPIO) and to their high magnetic moment which generates microscopic field inhomogeneities, these agents have been used in magnetic resonance (MR) imaging as negative contrast agents. The clinical importance of SPIO has been demonstrated in MRI the in vitro effect of increasing concentrations of of liver and spleen tumors'.2 and perfusion i m a g i ~ ~ g . ~ , ~ SPIO and USPIO on the MR signal with various seHowever, their efficacy on proton relaxation is not quences and at 0.5 and 1.5 T field strengths. confined to T2and magnetic susceptibility-related efthe in vivo effect of increasing doses of USPIO in fects: These contrast agents have a TIrelaxivity which myocardial and liver MR imaging at 0.5 T. is five to six times higher than any of the gadolinium complexes currently available on the market.5,6 AlMATERIAL AND METHODS though this characteristic has been clearly described in Contrast Agents vitro7 it has never been applied, to our knowledge, to SPIO (AMI-25) and USPIO (AMI-227) particles imaging. Positive organ enhancement after administrawere synthesized by advanced Magnetics Inc. (Camtion of superparamagnetic compounds could be inapbridge, MA) and supplied by Laboratoire Guerbet propriate when a negative enhancement is expected. (Aulnay-sous-Bois, France). These nanoparticles are However, this phenomenon could be quite useful in percomposed of an iron oxides crystalline core, measurfusion imaging, since enhancing rather than lowering ing 4-6 nm with an X-ray diffraction technique, sursignal intensity would avoid magnetic susceptibilityrounded by a dextran surfactant. The entire mean related artefacts. This is particularly important for RECENED 8/28/92; ACCEPTED 11/26/92. Address correspondence to Catherine Chambon, Labo-

ratoire Guerbet, 16/24 rue Jean Chaptal, 93601 Aulnay-SousBois CCdex, France.

Magnetic Resonance Imaging

510

particle size is therefore 150 nm for SPIO and 20 nm for USPIO, as measured by laser light scattering. The T2 relaxivity of SPIO and USPIO was 0.7.10~ (mol/l)-I .scl at 37°C and at 20 MHz in plasma. Under the same conditions TI relaxivities were 0.18.10~ (mol/l)-' -s-' and 0.24.10~(mol/l)-' es-', respectively.

Volume 1 1 , Number 4, 1993

formed for regions of interest (ROI) of 100 pixels per tube. Enhancement (ENH) for each tube in relation to the control tube (pure plasma) was calculated using the following formula (Eq. 3): ENH (%) = SI(tube) - SI(pureplasma) x 100

.

(3)

>l(pure plasma)

In Vitro Imaging Solutions. Successive dilutions of SPIO and USPIO were prepared in human plasma obtained from healthy volunteers and placed in eight polyethylene tubes for each contrast agent. The final concentrations were 0, 25,50,70, 140,230,470, and 940 pmol Fe/l for SPIO and 0,20,40,60,120,200,400, and 800 pmol Fe/l for USPIO. Theory. The signal intensity (SI) of the MR image can be plotted as a function of the sequence parameters and the relaxation timess: Spin echo sequences (Eq. 1):

(K = constant, p = proton density)

Animals. Male Sprague-Dawley rats weighing between 250 and 390 grams (Iffa Credo, L'Arbresle, France) were used. Two groups of six rats were used for the pharmacokinetic study and four groups of five rats were used for the in vivo imaging experiment. Effective dose of the contrast agent. Since the acquisition of the MR image was performed approximately 2 hr after the intravenous injection of USPIO, the plasma disappearance curve of USPIO was determined and the Effective Dose (ED, dose remaining in the plasma at the time of imaging) was calculated as a function of the injected dose (ID) (Eq. 4):

. (TI/, = plasma half life)

Gradient echo sequences (Eq. 2):

(p = flip angle)

In Vivo Imaging

.

To calculate the theoretical signal, the relaxation times of the samples were measured on a Minispec PC 20 MR relaxometer (Bruker, Wissembourg, France) at 0.47 Tesla and 37°C. The TI relaxation times were determined with an inversion recovery sequence on eight points and the T2 relaxation times were determined with a Carr-Purcell-Meiboom-Gill (CPMG) sequence and a tau time of 200 psec. Then, the theoretical data points were calculated for the different TRs, TEs and 0 of the imaging experiment using Eqs. (1) and (2).

MRI. The series of tubes bathing in 30 mmol/l copper sulfate solution were imaged the same day on two, 0.5 and 1.5 T, MR scanners (MRMax and Signa, General Electric, Buc, France). Four spin-echo (SE) sequences were performed, with a single echo and the following parameters: 160 msec/20 msec (repetition time "TR"/echo time "TE"), 500/22,1800/40,1800/80. A gradient echo (GE) sequence with a TR of 150 msec, a TE of 15 msec, and a flip angle of 20" was then obtained. For all sequences, the voxel size was 1 x 1 x 10 mm. Signal intensity (SI) measurements were per-

As a comparison, the plasma disappearance curve of SPIO was also determined. The animals were anesthetized with an intramuscular injection of 150 mg/kg of ketamine hydrochloride (Imalgkne, RhBne-Merieux, Lyon, France), and catheters were inserted into the jugular vein and the tail artery. After a bolus intravenous injection of 20 pmol Fe/kg of one of the two contrast agents, blood samples were collected in heparinized vials at regular time intervals for 1 hr for SPIO and for 7 hr for USPIO. After centrifugation, the plasma T2 was measured on a 0.47 T relaxometer with a CPMG sequence using a tau time of 200 psec. The plasma concentrations of the contrast agents were determined at time t with reference to a calibration curve of SPIO or USPIO in plasma. Plasma disappearance curves for SPIO and USPIO were analyzed using a non-compartmental mode1 allowing calculation of the mean residence time (MRT). a one-compartment model was also applied to the kinetics of USPIO, and its elimination half-life (TIl2) and distribution volume (DV) were calculated.

MR Imaging The dose-effect relationship of USPIO in myocardial and liver enhancement was studied in the rat. The four groups received the contrast agent intravenously at doses of 20, 40, 132, and 220 pmol Fe/kg, respec-

Superparamagnetic iron oxides

tively. Doses were determined on the basis of pharmacokinetic and in vitro imaging data. After anesthesia with 70 mg/kg of intraperitoneal sodium pentobarbital (Pentobarbital5, Sanofi, Montpellier, France), a catheter was inserted into the jugular vein and flushed with saline. Oblique sagittal images including the heart and the liver were obtained on the 0.5 T imager with an 8 cm diameter surface coil. A tube filled with 1 mmol/l GdDOTA solution was placed in the field of view to act as a reference phantom. TI-weighted SE images (TR = 500 msec/TE = 22 msec/2 acquisitions) were obtained prior to injection of the contrast agent, without cardiac gating. The animal was then removed from the imager and USPIO was injected as a bolus via the catheter. A second series of images using the same sequence and the same voxel size (0.6 x 0.6 x 3 mm) was obtained about 2 hr after the injection. At this time, the images provided information about both myocardial perfusion and the hepatic uptake of the compound. The rat was then sacrificed in the magnet by an intravenous overdose of pentobarbital. A third series of images was then obtained (immediate postmortem image) using the same parameters. Signal intensity measurements were made with operator defined regions of interest on the liver, the myocardial wall, the ventricular cavity (blood mass) and the phantom. After standardization of the signal intensities to that of the phantom, the enhancement percentage was calculated (Eq. 5): ENH (Yo) =

R I ( p o s t ~-~RI(pre~~p10) ~~~) x 100 (5) RI(preus~10)

(RI = standardized signal intensity) . RESULTS In Vitro Imaging The theoretical concentration-effect curves calculated from the T, and T2 provided a model in which signal enhancement reached a peak regardless of the sequence, then dropped as the concentration increased. In spin echo sequences, the more strongly the sequence was TI-weighted (short TR and TE), the greater and more prolonged the enhancement. In this case the peak was also found at higher concentrations (Fig. 1). Conversely, for T2-weighted SE sequences, weak signal enhancement was observed, followed by an earlier and more marked signal drop. For gradient echo sequences the signal enhancement was weak, similar to that of the T2-weighted SE (1800/40) sequence. The experimental concentration-effect curves obtained at 0.5 T were consistent with the theoretical model. For the most heavily TI-weighted sequence

C. CHAMBON ET AL.

51 1

(160/20), peak enhancement with SPIO was not higher than for the 500/22 sequence (Fig. 2). With USPIO however, peak enhancement was higher the more heavily the sequence was TI-weighted, as predicted by the theory. The greatest enhancement observed at 0.5 T when using a SE 500/22 (TR/TE) sequence reached 120% for SPIO and 140% for USPIO for the 120 pmol Fe/l concentration. The appearance of the curves were not significantly modified when MR images were acquired in a 1.5 T Bo field. The maximum enhancement observed at 1.5 T in the SE 500/22 sequence reached 100% for SPIO and 150% for USPIO (Fig. 3). Pharmacokinetics Compared to SPIO, USPIO cleared slowly from the plasma according to strictly mono-exponential kinetics, characterized by a half-life of 120min (Table 1 and Fig. 4). The calculated effective doses at the time of imaging were therefore 7, 14, 47, and 77 pmol Fe/kg. In Vivo Imaging With a SE 500/22 sequence, the myocardium showed progressive enhancement, reaching a peak of 30% at the effective dose of 14 pmol Fe/kg (Figs. 5 and 7). The enhancement then declined progressively, and a signal drop of -50% was observed at 77 pmol Fe/kg. The signal intensity of the cardiac chambers and the lungs was not affected in vivo.by the injection of USPIO (Figs. 7 and 8). A decrease in the liver signal was observed on images acquired two hours after injection, in all injected doses (Figs. 6-8). Postmortem Imaging In postmortem images, a qualitative study of the blood mass signal was possible due to the cessation of blood flow. Since it was impossible to obtain precontrast standard images, however, the percentage of en-

Table 1 . SPIO and USPIO pharmacokinetic parameters Noncompartmental analysis Mono-exponential model MRT (min) USPIO SPIO

168.36 k 4.67 16.26 k 1.25

T1/2

(min)

DV (ml .kg-')

117.51 + 3.17 28.73 k 1.52

Results expressed as mean + standard error of the mean (n = 6 in each group). MRT = mean residence time, T1/2 = half life, DV = distribution volume.

Volume 11, Number 4, 1993

Magnetic Resonance Imaging

SPIO THEORY

1

CONCENTRATION (pmol FelL)

-

USPlO THEORY

SE 160120 SE 500122 SE 1800140 SE 1800180 GE 150115120"

CONCENTRATION (pmol FelL)

(B) Fig. 1. Theoretical signal enhancement at 0.5 T with SPIO (A) and USPIO (B) as a function of the sequence used and the contrast agent concentration.

hancement could not be calculated. At an effective dose of 14 pmol Fe/kg, the blood mass (cardiac chambers) showed high signal intensities, since SI was similar to that of the standard phantom. Similarly, in the intrahepatic vessels and the lungs, a considerable rise in sig-

nal intensity was observed in comparison to the in vivo images. Conversely, at an effective dose of 77 pmol Fe/kg, the blood mass and the lungs remained dark, since SI was similar to that of the liver (Fig. 8).

Superparamagnetic iron oxides

C. CHAMBON ET AL.

SPIO 0.5 Tesla

0

200

400

600

800

1000

CONCENTRATION (pmol FelL)

(A)

-

USPlO 0.5 Tesla

-

200

SE 160120 SE 500122 SE 1800140 SE 1800180 GE 150115/20"

h

5 100 z W I-

z

0

z 4

I. W

0

-1 00 CONCENTRATION (pmol FelL)

(B)

Fig. 2. In vitro enhancement at 0.5 T with SPIO (A) and USPIO (B) as a function of the sequence and the concentration. Note the good fit with the theoretical curves (Fig. l), except for the 160/20 sequence in (A).

DISCUSSION This study demonstrates theoretically and experimentally that a positive signal enhancement can be achieved with a T,-weighted spin-echo sequence using

low doses of SPIO and USPIO. In vivo, this effect was seen in the myocardium, but less so than in vitro. In the liver where particles accumulate, this effect was not observed. USPIO particles have a prolonged blood half life

Volume 11, Number 4, 1993

Magnetic Resonance Imaging

SPIO 1.5 Tesla

CONCENTRATION (pnol Fell)

(A) USPlO 1.5 Tesla

0

200

400

'

600

800

1000

CONCENTRATION (pmol FeIL)

(B) Fig. 3. In vitro enhancement at 1.5 T with SPIO (A) and USPIO (B) as a function of the sequence and the concentration.

and may act as contrast agents in organs during the initial post-injection period through a "perfusion effect," while they circulate in the blood stream. In organs of the RES like the liver, spleen, and lymph nodes, a "capture effect" is added due to the gradual uptake of the particles by the macrophages. Therefore, in this study, a negative enhancement of the liver was always observed. Accumulation of the compound in the Kupffer

cells resulted in a clustering of the particles, that therefore acted as larger particles and induced more susceptibility effects. It is possible that with rapid image acquisition after the injection a positive liver-signal enhancement can be observed at a time when hepatic uptake has not occurred. A biphasic curve was observed in vitro with signal enhancement at low concentrations followed by a signal

Superparamagnetic iron oxides

C. CHAMBON ET

AL.

TIME (min)

Fig. 4. Plasma disappearance curves for SPIO and USPIO at an injected dose of 20 pmol Fe/kg (logarithmic scale for the y axis). Note that a monoexponential fitting is not adequate for SPIO data. (Mean k SD).

drop at higher concentrations. This result was consistent with the theoretical signal intensity curves calculated from the Tl and T2relaxation times of the same samples, thus suggesting that Tl and T, relaxivities govern this phenomenon in vitro without any magnetic susceptibility-related phenomenon (the maximum enhancement observed was higher with USPIO than with SPIO due to a higher T,relaxivity). The good sirnilar-

ity between the theoretical and experimental curves shows that positive enhancement can be predicted as a result of the high T, relaxivity of the particles. This hypothesis is consistent with recent publications concerning the mechanisms of NMR proton dephasing with superparamagnetic particle^.^ Muller et al. showed that for particles with a small metallic core, such as SPIO and USPIO, proton dephasing occurs by

+MEAN

EFFECTIVE DOSE (pmol Felkg)

Fig. 5. Myocardial enhancement. Enhancement of the myocardium wall 2 hr after injection of USPIO, in relation to remaining circulating dose (effective dose).

Magnetic Resonance Imaging

Volume 1 1 , Number 4, 1993

INJECTED DOSE (pmol FelL)

Fig. 6 . Liver enhancement. Enhancement of the liver 2 hr after injection of USPIO, in relation to injected dose (Mean + SD).

T2relaxation interactions of the outer-sphere type. In contrast, in particles with larger metallic cores, proton dephasing occurs predominantly by a magnetic susceptibility effect and by diffusion in the microscopic field inhomogeneities. Theoretical enhancement curves were calculated with TI and T2 values measured with a relaxometer, and in vitro curves with a clinical magnet. Slight differences between the theory and the in vitro results may be therefore explained by differences in field homogeneities between the two magnets. Although the Tl effect was demonstrated in vitro for both SPIO and USPIO, we studied its application for perfusion imaging using USPIO because of its long blood half-life and higher T, relaxivity. The maximum enhancement difference observed between in vitro (140%) and in vivo results (only 30%) can probably be explained by the development of an additional phenomenon in vivo. A hypothesis which might explain this relative loss of efficacy is an in vivo magnetic susceptibility-related phenomenon caused by two flow-related mechanisms: A blood pool contrast agent is confined to only 25% of the volume of the voxel,1° thereby creating intravoxel magnetic susceptibility inhomogeneities not detected in the in vitro system, where the contrast agent was distributed in 100% of the volume of the voxel. Various authors" have described the development of a magnetic susceptibility phenomenon when magnetic particles travel in the field in different directions from the phase encoding. This is the case for

USPIO circulating in the blood. Comparison between pre- and postmortem images corroborates this hypothesis. When blood circulation and the movement of USPIO particles stop at the time of death, the magnetic susceptibility effect decreases. In the lungs, the considerable rise in SI after death (Fig. 7C) can be explained by the T, effect of USPIO. This effect is probably seen here due to the increase in lungwater content from postmortem edema. Despite the in vivo magnetic susceptibility-related phenomena in the normal myocardium, an increase in the MR signal intensity was observed post-administration of USPIO. These variations in signal intensity were directly correlated with local blood volume and/or blood flow. The clinical applications of this study in myocardial ischemia must be further investigated since various other intricate phenomena (such as the degree of vascularisation or capillary permeability) may affect the applicability of these experimental findings. However, knowledge of the different enhancement properties of superparamagnetic particles is of critical importance since opposite effects can be observed in different organs through the perfusion and active uptake mechanisms. In conclusion, this study demonstrates that the marked Tl relaxivity of superparamagnetic compounds can induce a positive enhancement of the MR signal. The relatively low T1 effect of superparamagnetic particles observed in vivo requires additional study to know if this property can be useful in the assessment of the vascularisation of hypoperfused or infarcted organs.

Superparamagnetic iron oxides

C. CHAMBON ET AL.

517

Fig. 7. Sagittal oblique MR image of a rat (spin echo -500/22-sequence), demonstrating the enhancement pattern following intravenous infusion of a low dose of USPIO. (A) Pre-injection. (B) 2 hr 40 min after intravenous injection of 40 pmol Fe/kg of USPIO, corresponding to an effective dose of 14 pmol Fe/kg: A 30% enhancement of the myocardial wall is observed. Notice the absence of visible enhancement of the lung and the heart cavities, due to entry phenomena. The liver signal intensity is decreased, revealing the hepatic uptake of USPIO. Since the rat was removed from the magnet during the injection procedure, the slice plane is slightly different. (C) Immediate postmortem image: A dramatic enhancement of the lung and the heart cavities is observed due to the TI effect of the low USPIO concentration. The cessation of blood flow following death also prevents any further circulation of the magnet particles. Consequently, a cause for magnetic susceptibility-related artefact is eliminated, and the "TI-effect" of the particles can be exhibited.

518

Magnetic Resonance Imaging

Volume 1 1 , Number 4, 1993

Fig. 8. Sagittal oblique MR image of a rat with a spin echo (500/22) sequence demonstrating the enhancement pattern following IV infusion of a high dose of USPIO. (A) Preinjection. (B) 2 hr and 7 min post-injection of 220 pmol Fe/kg of USPIO, corresponding to an effective dose of 77 pmol Fe/kg, the myocardium and the liver have darkened visibly. (C) Immediate postmortem image: Compared to the immediate pre-mortem image, a recovery of the myocardial signal intensity is observed. Here again, USPIO particles in the blood are no longer moving and induce less susceptibility effect. Furthermore, a slight enhancement of the myocardium compared to the precontrast image is observed due to medium-low myocardial concentrations of USPIO. Conversely, the heart cavities and the lung remain dark due to the negative enhancement caused by high USPIO concentrations.

Superparamagnetic iron oxides Acknowledgment-This work was supported in part by Inserrn grant 900906 and by Laboratoire Guerbet, France.

REFERENCES 1. Stark, D.D.; Weissleder, R.; Elizondo, G.; Hahn, P.F.; Saini, S.; Todd, L.E.; Wittenberg, J.; Ferrucci, J.T. Superparamagnetic iron oxide: Clinical application as a contrast agent for MR imaging of the liver. Radiology 168:297-301; 1988. 2. Weissleder, R.; Hahn, P.F.; Stark, D.D.; Elizondo, G.; Saini, S.; Todd, L.E.; Wittenberg, J.; Ferrucci, J.T. Superparamagnetic iron oxide: Enhanced detection of focal splenic tumors with MR imaging. Radiology 169: 399-403; 1988. 3. Majumdar, S.; Zoghbi, S.S.; Gore, J.C. Regional differences in rat brain displayed by fast MRI with superparamagnetic contrast agents. Magn. Reson. Imaging 6:611-615; 1988. 4. Rozenman, Y.; Zou, X.M.; Kantor, H.L. Cardiovascular MR imaging with iron oxide particles: utility of a superparamagnetic contrast agent and the role of diffusion in signal loss. Radiology 175:655-659; 1990. 5 . Josephson, L.; Lewis, J.; Jacobs, P.; Hahn, P.F.; Stark, D.D. The effects of iron oxides on proton relaxivity. Magn. Reson. Imaging 6547-653; 1988.

C. CHAMBON ET AL.

519

6. Weissleder, R.; Elizondo, G.; Wittenberg, J.; Rabito, C.A.; Bengele, H.H.; Josephson, L. Ultrasmall superparamagnetic iron oxide: Characterization of a new class of contrast agents for MR imaging. Radiology 175:489493; 1990. 7. Gillis, P.; Koenig, S.H. Transverse relaxation of solvent protons induced by magnetized spheres: Application to ferritin, erythrocytes and magnetite. Magn. Reson. Med. 4:323-345; 1987. 8. An introduction to magnetic resonance in medicine. In: P.A. Rinck, R.N. Muller, S.B. Petersen, eds. The Basic Textbook of the European Workshop on Magnetic Resonance in Medicine. Stuttgart: Georg Thieme Verlag; 1990: pp. 68-73. 9. Muller, R.N.; Gillis, P.; Moiny, E; Roch, A. Transverse relaxivity of particulate MRI contrast media: From theories to experiments. Magn. Reson. Med. 22:178-182; 1991. 10. Springer, C.S.; Xu, Y. Aspects of bulk magnetic susceptibility in in vivo MRI and MRS. In: P.A. Rinck, R.N. Muller, eds. New Developments in Contrast Agent Research. Bordeaux, France: European Magnetic Resonance Forum; 1990; pp. 13-25. 11. Ogawa, S.; Lee, T.S. Magnetic resonance imaging of blood vessels at high fields: In vivo and in vitro measurements and image simulation. Magn. Reson. Med. 16:9-18; 1990.