Zhiyong Zhang1,2, Amir Seginer1, and Lucio Frydman1
1Chemical Physics, Weizmann Institute of Science, Rehovot, Israel, 2Electronic Science, Xiamen University, Xiamen, China, People's Republic of
Synopsis
Magnetic resonance imaging (MRI) near metallic implants remains an
unmet need because of severe artifacts, which mainly stem from large
metal-induced field inhomogeneities. The single-scan cross spatiotemporal
encoding (xSPEN) technique delivers in-plane distortion-free 2D images under
such large field inhomogeneity condition, while the slice-plane displacement,
“signal voids” and “pile-up” effects are proposed to be solved by applying t1-evolution-time encoding on
the multi-slicing 2D xSPEN technique. Compared to the popular “SEMAC” and
“MAVIC” techniques, the remarkable time efficiency of this t1-encoding xSPEN thus enable many advanced MRI
applications near metal implants with another additional dimension, such as diffusing
MRI, function MRI.Purpose
The single-scan cross spatiotemporal encoding
(xSPEN) technique delivers in-plane distortion-free 2D images under highly
inhomogeneous fields, such as those found near metal implants. This
study introduces an xSPEN variant incorporating a t
1 evolution encoding time, which together allows one to
solve the in-slice residual distortions of the third dimension. By removing
“signal voids” and “pile-up” effects the ensuing slice-encoded xSPEN (SexSPEN) method delivers fast,
quality 3D images.
Introduction
SEMAC [1] and MAVRIC [2] are ingenious imaging techniques for handling distortions
arising near metal implants. Both of these are multi-scan-based sequences with
relative long acquisition times, leading to complications in their use for
advanced MRI applications such as diffusion-weighted imaging (DWI). We have
recently introduced a single-shot method, xSPEN, with exceptional resilience
to in-plane field heterogeneities. The
present work proposes an xSPEN-based t
1
encoding scheme the deals with residual in-plane distortions that still arise in
xSPEN’s slice-selection, based on SEMAC/MAVRIC-related ideas. The SexSPEN method can thus reinstate all spins in a 3D region-of-interest
to their actual spatial locations, in a highly time-efficient fashion.
Methods
SexSPEN’s sequence is shown in Fig.1. In-plane
field inhomogeneities are dealt with by xSPEN using a G
z-gradient that cross spatiotemporally encodes and reads
out the desired imaging information along
y
axis, coupled with a strong oscillating readout G
x gradient sampling this dimension’s
kx-space. xSPEN then obtains a
2D MR image by applying a Fourier transform only on
kx, in a single-shot fashion that is remarkably
insensitive to in-plane heterogeneities. Field inhomogeneities will, however,
still distort xSPEN’s images due to inhomogeneity-imposed distortions in the
slice excitation process –as they do in all slice-selective MR experiments.
SexSPEN solves this by adding a t
1-encoding
that maps the frequency information within each slice. By allowing a repositioning
of all spins in a targeted region-of-interest back to their actual spatial
locations, through-plane distortions are corrected within a small number of
scans.
Results & Discussion
Figure
2 demonstrates the xSPEN’s exceptional resilience to in-plane field
heterogeneities, with a 7T scan targeting a phantom incorporating two
features that are notoriously challenging to single-shot acquisitions: a
titanium screw of the kind used in orthopedic implants, and a sample composed
of several chemically-shifted sites.
Figures 2c-2f illustrate images arising from these six slices, using a
variety of pulse sequences. Notice that –apart
from sensitivity considerations dictated by their single-shot scan– xSPEN-derived
images are virtually indistinguishable from those arising from their multiscan
counterparts. This, despite the fact that the entire xSPEN set was acquired
within 360 ms (60 ms per slice, no relaxation delays within slices, no need for
navigator scans or field maps). Notice, however, the residual influence of “pile-up”
effects and slice-plane displacements in both
xSPEN and multi-scan images, reflecting through-plane distortions
affecting the initial slice selection.
These artifacts can be solved by the SexSPEN procedure.
Figure 3
illustrates this with 3T scans on a gel phantom containing a hip implant (panel
a). A 20-step t1-encoded SexSPEN experiment with TR/TE=10000/61 ms,
40 slices, 3 mm isotropic resolution, and a 3 min 20 sec total scan time, leads
to 3D images that are free from all through-plane and in-plane distortions. Images
arising after Fourier transforming the xSPEN data along kx (Fig. 3b) still show in-plane distortions; after FT
processing the t1
dimension, frequency maps can be obtained that resolve the through-plane distortions
(Fig. 3c). Slice-corrected, repositioned images inside a region-of-interest (Fig.
3d) clearly depict the metal position and the details of the phantom. Parallel
imaging and partial Fourier reconstruction along the readout dimension may be
used to further improve the in-plane resolution and image quality.
Conclusion
A t
1-encoding
SexSPEN sequence that can deliver both in-plane and through-plane
distortion-free images in a fast manner is introduced and demonstrated. This
could enable advanced MRI applications near metal
implants, including diffusion, flow, real-time and functional studies.
Acknowledgements
This work was
funded by the Israel Science Foundation grant 795/13, by ERC-2014-PoC grant #
633888, by the Kimmel Institute of Magnetic Resonance, and by the generosity of
the Perlman Family Foundation. References
[1] Lu W, Pauly KB, Gold GE, Pauly JM, Hargreaves
BA. SEMAC: Slice encoding for metal artifact correction in MRI. Magn Reson Med
2009;62:66-76.
[2] Koch KM, Lorbiecki JE, Hinks RS, King KF. A
multispectral three-dimensional acquisition technique for imaging near metal
implants. Magn Reson Med 2009;61:381-390.