Off-Resonance Imaging

 

Introduction:

The development of superparmagnetic nanoparticles, such as monocrystalline iron-oxide (MION), has been central to the development of magnetic resonance molecular imaging. Iron-oxide based contrast agents have been used to both label stem cells for cell tracking, as well as to detect specific and sparsely expressed molecular targets. The strong magnetic dipole field generated by these susceptibility agents both shifts the resonance frequency and decreases the relaxation times of water molecules in the vicinity of the contrast agent. These effects lead to a decrease in the MRI signal intensity in the vicinity of the contrast agent on T2-weighted, T2*-weighted, and steady-state free precession (SSFP) images.

 

The detection of iron-oxides with sequences producing negative or dark signal contrast has proven highly sensitive, capable of detecting contrast agent concentrations in the nanomolar range. However, these techniques do suffer from several potential problems: 1) extremely high local concentrations of MION may produce a complete signal void precluding the accurate quantification of nanoparticle concentration; 2) the detection of MION in regions with low intrinsic signal-to-noise ratio (SNR) is difficult; 3) it is difficult to distinguish MION induced susceptibility changes from those caused by susceptibility artifacts such as air-tissue interfaces; 4) the long echo times required to detect low concentrations of MION complicate the detection of moving objects, such as the heart, where short echo times are optimal.

 

Several off-resonance imaging (ORI) techniques have recently been proposed to create positive contrast from iron-oxide agents and overcome some of the above challenges. These techniques selectively image waters shifted off-resonance by susceptibility agents using a variety of different methods including a gradient dephasing technique [1,2], a spectrally selective RF technique [3,4], an on-resonance water suppression method [5], and an off-resonance saturation method [6].

 

These positive contrast techniques have been used to successfully image high concentrations of iron-oxide, such as those found in exogenously labeled stem cells, at standard clinical field strengths [3,4]. However, there has been little experience with these techniques in the imaging of more dilute concentrations of iron-oxides or imaging at higher field strengths. Furthermore, the quantitative relationships between the contrast agent concentration and the induced chemical shifts, relaxation rates, and signal intensity have not been fully elucidated.

 

The aim of this study was thus to characterize in detail the performance of an off-resonance positive contrast imaging technique for the detection of dilute concentrations of iron-oxides, such as those produced by targeted molecular probes, at high magnetic field strengths.

 

Materials and Methods:

1. MION Phantoms: Three different phantoms were made: (1) A MION phantom consisting of 3 mm NMR tubes loaded with different concentrations of MION in aqueous solution; (2) An air-water phantom consisting of eppendorf tubes filled approximately half way with different concentrations of aqueous MION; (3) An air-fat phantom consisting of an eppendorf tube filled with olive oil and water.

 

2. In Vivo Mouse Model: Myocardial macrophage infiltration was imaged in infarcted C57Bl6 mice (n = 4) 4 days after permanent ligation of the left coronary artery. This model has been shown to produce a macrophage infiltrate in the healing infarct, beginning 48-72 hours after the initial injury. The mice were thus injected 48 hours after coronary ligation with 15 mg/kg MION via the tail vein. Imaging was performed a further 48 hours later to allow the probe sufficient time to be taken up by the macrophages infiltrating the healing infarct. Uninfarcted mice (n = 3) were used as control animals.

 

3. Magnetic Resonance Imaging: The ORI pulse sequence used for both the in vitro and in vivo studies consisted of a variable bandwidth on-resonance water excitation pre-pulse followed by gradient crushers, a subsequent fat-selective excitation pre-pulse also followed by gradient crushers, and a standard slice selective spin-echo imaging sequence. MRI of the MION phantoms was performed on Bruker Avance 14 and 4.7 T scanners using transmit/receive volume coils. T1 and T2 relaxation times were determined for different aqueous MION concentrations from inversion recovery spin-echo data and from multi-echo spin-echo data, respectively. In vivo mouse ORI studies were performed on a 4.7 T Bruker Avance scanner with a custom built solenoid coil (5 turn coil, 30 mm diameter, 70 mm length). Default in vivo imaging parameters included: FOV = 35 x 30 mm, slice thickness = 1 mm, matrix = 128 x 128, on-resonance WSBW = 200 Hz, TR = 2000 ms, TE = 2.9 ms, and 4 signal averages.

 

4. Quantification of Induced Chemical Shifts: The mean chemical shift induced by MION was quantified by varying the on-resonance water suppression bandwidth (WSBW) and measuring the mean signal intensity. The plot of the signal intensity versus WSBW was then fit to a sigmoidal Boltzman function given by:

 


The fit parameter x0 corresponds to the WSBW at which the signal intensity is half way between the maximum (A1) and minimum (A2) values. This corresponds to twice the mean induced chemical shift (d).

 


Results:

1. Aqueous MION Phantom:

Positive contrast images of an aqueous MION phantom are shown in Figure 1. We note that the brightest MION solution does not correspond to the highest MION concentration. Furthermore, a lower MION concentration is detectable at the lower field strength.

 

Figure 1: Spin-echo and off-resonance images of a MION phantom acquired at 4.7 T (top row, a-d) and 14 T (bottom row, e-h). The phantom consists of five 3 mm NMR tubes, with varying MION concentrations, immersed in an undoped water bath. The MION concentrations range from 8-256 ug Fe/ml and are indicated in the respective spin-echo images (a, e). The off-resonance images were acquired with water suppression bandwidths (WSBW) of 250 (b, f), 500 (c, g), and 750 Hz (d, h). The 4.7 T images were acquired with TE/TR = 5.6/1000 ms while the 14 T images were acquired with TE = 3.5/1000 ms.

 

2. Quantification of MION Concentration:

Figure 2 clearly demonstrates that the ORI signal intensity is non-linear. This is attributed to the very short T2 relaxation times of the higher MION concentration samples. In contrast, the induced chemical shift and the off-resonance transverse spin-relaxation rate (R2) vary linearly with MION concentration.

 

 

ORI Signal Intensity

Induced Chemical Shift

Off-Resonance R2

 

 

 

4.7 T

 

 

 

14 T

 

 

 

Figure 2: Dependence of the off-resonance signal intensity, induced chemical shift, and off-resonance R2 on the MION concentration at 4.7 and 14 T. The mean ORI signal intensity was determined for a water suppression bandwidth (WSBW) of 250 Hz. While the ORI signal intensity is clearly non-linear, both the induced chemical shift and the off-resonance R2 vary linearly with MION concentration.

 

3. Susceptibility Shifts at Air-Water and Air-Fat Interfaces:

Significant susceptibility artifacts are observed at air interfaces as shown in Figure 3. A broadband water suppression pulse is needed to reduce or eliminate the off-resonance positive contrast at air interfaces.

 

 

Figure 3: Positive contrast from susceptibility shifts at air-water (a-c) and air-fat (d-f) interfaces. The images shown are: spin-echo image (a), off-resonance image with a 200 Hz WSBW (b), off-resonance image with a 400 Hz WSBW (c), spin-echo image (d), fat-suppressed spin-echo image (e), fat-suppressed off-resonance image with a 250 Hz WSBW (f). All images were acquired at 4.7 T with TE/TR = 5.6/4000 ms.

 

4. In Vivo ORI of MION: Dependence on the Water Suppression Bandwidth

Shown in Figure 4 are off-resonance images of normal and infracted mice. The off-resonance images have been fused with traditional spin-echo images. No off-resonance signal was observed in the myocardium of any of the normal control mice. In contrast, clear evidence of positive contrast was observed in the myocardium of all infarcted mice. Significant susceptibility artifacts are evident in the thoracic wall. Increasing the WSBW reduces much of the background off-resonance signal, however, the sensitivity to MION accumulation in the myocardium is lost.

 

 

Figure 4: Off-resonance imaging of normal (a, b) and infarcted (c, d) mice acquired at 4.7 T with water suppression bandwidths (WSBWs) of 200 (a, c) and 400 Hz (b, d). The off-resonance images have been mapped to a color scale and overlaid onto conventional spin-echo images. Iron-oxide accumulation within macrophages infiltrating the injured myocardium produces positive off-resonance contrast (arrow) for a WSBW of 200 Hz, but not for 400 Hz.

 

5. In Vivo ORI of MION: Dependence on the Echo Time

Positive contrast (arrow) is seen in the anterolateral myocardium at the shorter echo times but not at an echo time of 8.7 ms. The use of short echo times is thus critical to preserve the sensitivity of the off-resonance imaging technique.

 

 

Figure 5: ORI and spin-echo (SE) images of MION accumulation within infarcted myocardium at 4.7 T. Fat suppressed spin-echo images were acquired with echo times of 2.9 (a) and 11.6 ms (b). The negative contrast enhancement produced by MION can be seen in the thinned anterolateral wall (arrow). Off-resonance spin-echo images were acquired using a water suppression bandwidth of 200 Hz and echo times of 2.9 (c) and 8.7 ms (d).

 

6. ORI Contrast-to-Noise Ratio:

As shown in Figure 6, significantly greater CNR was seen in the off-resonance images of the infarcted mice (n=4) than the control mice (n=3), reflecting the known uptake of MION by macrophages infiltrating the healing infarct.

 

 

Figure 6: Quantification of MION uptake by off-resonance imaging. The contrast-to-noise ratio (CNR) between the infarcted anterolateral myocardium and the uninjured septal wall is shown.

 

 

Discussion:

The MION induced chemical shifts are independent of the field strength. However, the on-resonance water linewidth increases with increasing field strength (~15 Hz at 4.7 T and ~45 Hz at 14 T) and provides a lower bound limit on the smallest MION induced chemical shift that can be resolved. The optimal field for ORI will thus be a compromise between maximal SNR and minimal linewidth. For our experiments at 4.7 and 14 T, however, we have found that the spectral linewidths were the limiting factor in detecting low MION concentrations. The lowest MION concentration detectable in our aqueous MION phantom was 8 ug Fe/ml at 4.7 T and 32 ug Fe/ml at 14 T. Increased sensitivity will likely be obtained at lower field strengths, where linewidths are narrower and there is greater spectral resolution between on- and off-resonance water molecules.

 

Due to the short T2 relaxation times observed at high fields, the MION concentration dependence of the ORI signal intensity is highly non-linear (Figure 2). However, the off-resonance R2 and induced chemical shift both vary linearly with MION concentration (Figure 2) and can be used for MION quantification. Due to the short T2 relaxation times, the detection of even relatively dilute MION concentrations depends critically on the spin-echo time (Figure 4). Since ORI works optimally at very short echo times it may be well suited for imaging MION in regions that experience motion, such as the myocardium, where the longer TE’s required for T2*-weighted imaging could result in significant motion artifacts.

 

Large chemical shifts were consistently seen at air interfaces both in vitro (Figure 3) and in vivo (Figure 4). For example, the mean induced chemical shift at the thoracic wall interface was 303 Hz, which is approximately equivalent to the shift induced by an aqueous MION concentration of 151 ug Fe/ml. Suppression of these non-specific background areas of positive contrast could be achieved, but only through the use of broad water suppression bandwidths. However, as shown in Figure 4, the use of broad water suppression bandwidths suppresses signals from dilute concentrations of MION.

 

Conclusion:

Off-resonance imaging of dilute concentrations of iron oxides can be successfully performed at high field strengths but is complicated by the decreased spectral resolution of on and off-resonance spins and the high R2 relaxation rates at high fields. The optimal field strength for off-resonance imaging in small animals, such as mice, will thus be a compromise between maximal SNR and minimal linewidth.

 

 

References:

  1. Seppenwoolde JH, Viergever MA, Bakker CJG. Passive tracking exploiting local signal conservation: the white marker phenomenon. Magn Reson Med 2003;50:784-790.
  2. Mani V, Briley-Saebo KC, Itskovich VV, Samber DD, Fayad ZA. Gradient echo acquisition for superparamagnetic particles with positive contrast (GRASP): sequence characterization in membrane and glass superparamagnetic iron oxide phantoms at 1.5T and 3T. Magn Reson Med 2006;55:126-135.
  3. Cunningham CH, Aria T, Yang PC, McConnell MV, Pauly JM, Conolly SM. Positive contrast magnetic resonance imaging of cells labeled with magnetic nanoparticles. Magn Reson Med 2005;53:999-1005.
  4. Foltz WD, Cunningham CH, Mutsaers AJ, Conolly SM, Stewart DJ, Dick AJ. Positive-contrast imaging in the rabbit hind-limb of transplanted cells bearing endocytosed superparamagnetic beads. J Cardiovasc Magn Reson 2006;8(6):817-823.
  5. Stuber M, Gilson WD, Schaer M, Bulte JW, Kraitchman DL. Shedding light on the dark spot with IRON - a method that generates positive contrast in the presence of superparamagnetic nanoparticles. Proc Intl Soc Magn Reson Med 2005;13:2608.
  6. Zurkiya O, Hu X. Off-resonance saturation as a means of generating contrast with superparamagnetic nanoparticles. Magn Reson Med 2006;56:726-732.