Synopsis
Mechanisms to create a
homogenous T2 contrast at 7T should be robust to both B0
and B1 inhomogeneities. This issue was addressed by applying a zero-degree
BIR-4 adiabatic pulse for T2 preparation. Simulation and in vivo experiments in mouse brain verified the robustness and contrast of the preparation
scheme at 7T. Application of this method was further demonstrated in a study of a patient with a brain
tumor. The adiabatic T2 weighting showed high intensity in the region
of the tumor, suggesting that the proposed method is likely to be useful in screening
and characterizing tumors.Purpose
Mechanisms to create homogenous
T
2 contrast at 7T should be robust to both B
0 and B
1
inhomogeneities. Previous studies showed the T
2 contrast can be
generated by a long and segmented 0°- B
1-independent rotation (0°-BIR4)
adiabatic pulse [1-4]. In this study, we applied
this strategy to provide adiabatic T
2 preparation in combination with the widely used fast 3D gradient echo
imaging (GRE). The resulting adiabatic T
2-prepared 3D GRE was further
applied to a patient with a brain tumor in order to evaluate its potential for improving diagnostic performance.
Methods
Adiabatic T2 preparation design: A BIR4 RF pulse was
designed according to the parameters: bandwidth = 1 kHz, duration = 56 ms,
maximum B1 = 9 μT, sechn modulations, β = 184 rad/s, μ = 8, order =
8 and flip angle = 0°. Bloch simulations for using this as a T2
preparation pulse were performed with varied B0 and B1.
3D fast gradient echo (GRE) sequence: The sequence was
modified from a standard segmented fast GRE sequence (GE Healthcare) with the
inversion pulse being replaced by the T2 preparation pulse,
acquiring multiple ky and kz lines per each T2
preparation.
In vivo mouse brain experiment: This
was performed on a 7T whole-body MRI scanner (GE
Healthcare, Waukesha, WI) with an RF surface coil, which induced a varied B1
field. The 3D GRE images with and without T2-preparation
were acquired with TE/TR = 3.7/20.3 ms, bandwidth = 62.5 Hz, matrix size =
256×128×28, FOV = 60×60×10 mm3, NEX = 1 and FA = 1°.
Patient study: Brain tumor patient scans were performed
on the same 7T MRI. The study was performed with informed consent under a
protocol approved by the UCSF Committee on Human Research. A 32-channel volume RF coil (NOVA Medical) was used. 3D GRE images with
and without T2 preparation were acquired
with parameters: TE/TR = 1.3/3.7 ms, bandwidth = 62.5 Hz, matrix size = 256×256×32,
FOV = 240×240×30 mm3, NEX = 1 and FA = 4°. High resolution T2-prepared 3D GRE images were also
acquired, with TE/TR = 2.3/7.1
ms, matrix size = 512×512×32, FOV = 240×240×60 mm3 and NEX = 2, and other
parameters were identical with the previous scans. A 2D fast spin echo (FSE) scan was also performed
with effective TE/TR = 85.5/3000 ms, matrix size = 512×512, number of slices = 12,
slice thickness = 3 mm, slice gap = 4 mm, FOV = 240×240 mm3, NEX =
2, echo train = 16 and FA = 90°.
Results
In Figure 1, Bloch
simulations show that homogenous T
2 contrast can be achieved across
a large region with varied B
0 (i.e., -500 Hz < Δ f < 500 Hz)
and B
1 (70% < RF amplitude < 110%). Figure 2 shows results of the in
vivo mouse brain experiment at 7T with inhomogeneous B
1 field
that was caused by the surface coil used. The adiabatic T
2
preparation pulse provided homogeneous T
2 contrast, demonstrating
robustness to B
1 variations. Brain tumor patient scans
shown in Figures 3 and 4 demonstrate a similar high intensity in the tumor
region with either T
2-weighted spin echo or adiabatic T
2
weighting. It should be noted that homogenous T
2 contrast was
achieved with the adiabatic preparation even in the presence of the B
0
inhomogeneity in the frontal lobe (≈72 Hz) and B1 inhomogeneity of 14%
(std/mean) across the brain (known as “central brightening”). In Figure 4, the
T
2 contrast generated by the adiabatic T
2
preparation was observed to be comparable with
that of T
2-weighted FSE.
Discussion
B
0
and B
1 inhomogeneities are central concerns for clinical MRI studies
at high field. Although adiabatic pulses are known to be effective tools for
addressing this problem, intensively applying adiabatic pulses
can drastically
increase the SAR, and limitations in the RF hardware at high field (e.g., B
1
threshold = 9 μT in this study) impose a further challenge in applying adiabatic
pulses. To overcome these issues, we integrated a long (and strongly time
modulated) adiabatic T
2 preparation with a fast 3D GRE sequence,
which is an extension of the conventional hard-pulse-based T
2
preparation scheme.
This
approach can be applied for 7T patient studies and is immune to both
B
0 and B
1 inhomogeneities.
Conclusion
Animal
and human brain experiments verified the robustness of adiabatic T
2
preparation at 7T. The value of the adiabatic T
2 preparation was
demonstrated on a clinical brain tumor study at 7T. Adiabatic T
2
weighting showed high intensity in the tumor region, suggesting that the
proposed method may be useful for tumor screening and characterization.
Acknowledgements
No acknowledgement found.References
[1] Jenista
E. R., Magnetic resonance in medicine 2013;70:1360-8.
[2] Guo J.
and E. C. Wong Magnetic resonance in medicine 2012;68:1458-71.
[3] Nezafat
R., Magnetic resonance in medicine 2009;61:1326-35.
[4] Nguyen
T. D., Magnetic resonance in medicine 2015.