Yuan Zheng1, Michael Marx1, Rachelle R. Bitton1, and Kim Butts Pauly1
1Radiology, Stanford University, Stanford, CA, United States
Synopsis
We have demonstrated a method for shear wave
elasticity imaging (SWEI). A shear wave was generated by a short focused
ultrasound (FUS) pulse, and was tracked by collecting images with different
delays (tdelay) between the FUS pulse and bipolar motion-encoding
gradients (MEG). The time-of-flight (TOF) at each pixel was determined by the
zero-crossing of the image phase as a function of tdelay. Based on
the TOF map, a shear wave velocity map was generated in polar coordinates. Introduction
Tissue elasticity is often affected by
pathological changes, and could provide diagnostic value. Tissue elasticity can
be probed using shear waves, which can be generated by applying a short pulse
of focused ultrasound (FUS)
1. The speed at which a shear wave propagates
in the plane perpendicular to the FUS beam direction is directly related to the shear
modulus, and can be measured by tracking the propagation of the wave. Shear
wave tracking has been done mostly using ultrasound imaging techniques, but can
also be achieved with MRI. The advantage of using MRI instead for shear wave
tracking is that the tissue elasticity map can be directly registered to other
MR images, such as temperature maps acquired during MR-guided focused
ultrasound (MRgFUS) treatment. Previous work by Souchon et al. used an spGRE
pulse sequence with bipolar motion-encoding gradients (MEG) to visualize shear
waves
2. Bitton et al. tracked shear wave propagation using a
spin-echo pulse sequence with unipolar MEG
3, and calculated an
average shear wave speed for each radial spoke covering a few degrees, based on
a time-of-flight (TOF) map constructed by peak detection. In this work, shear
wave tracking was achieved using an spGRE pulse sequence with bipolar MEG. A
TOF map was generated by finding the zero-crossing of the motion-encoded phase
at each pixel as a function of the delay time (t
delay) between the
FUS pulse and MEG, and shear wave velocity was calculated as a function of both
polar angle and radial distance.
Methods
An spGRE pulse sequence with bipolar MEG and
spiral readout was used to track shear wave propagation in an isotropic acrylamide
phantom, by changing t
delay between the FUS pulse and the second
lobe of MEG from 0 ms to 27 ms in steps of 1 ms (Fig. 1). MR images were
acquired on a 3T GE MR750 scanner (GE Healthcare, Milwaukee, WI) with a single-channel
local coil. Pulse sequences were developed and implemented on the scanner using
RTHawk (HeartVista, Inc., Menlo Park). 2D images were collected with FA/TR/TE=30°/500
ms/13.5 ms, slice thickness=3 mm. The spiral readout had 24 interleaves with
BW=62.5 kHz, FOV=28 cm and resolution=1.5 mm. Each lobe of the MEG was 6 ms
long, with plateau magnitude of 4 G/cm. FUS was generated by a multi-element
phased-array transducer (Exablate 2000, Insightec, Haifa, Israel). The pulse
duration was 6 ms and the acoustic power was 21.6 W. The HIFU beam was parallel
to the MEG and perpendicular to the imaging slice. A full image was collected
at each t
delay. After subtracting the phase of a baseline image that
was acquired without FUS, phase images characterizing the shear wave
propagation were generated.
The
phase of each pixel is the integral of tissue displacement with the amplitude
of MEG
4: $$$\phi=\int x(t)G(t)dt$$$.
Therefore
ϕ is sensitive to t
delay since it determines how x(t) and G(t)
overlap. ϕ=0 when the
displacement partly overlaps with both gradient lobes and phases accrued during
the two lobes exactly cancel each other. To calculate TOF, we estimated the
time when ϕ is 0 by performing 1d gridding between the maximum and minimum measured
phases (Fig 2). A shear wave velocity map was generated by dividing the 2d TOF map
into 8° sectors
3, and calculating the shear wave speed at radial
increments of 2 mm by linear fitting (using a sliding Kaiser-Bessel kernel with
w=10 mm, b=6, FWHM≈5 mm).
Results
Fig. 3a shows the TOF map, and Fig. 3b shows
the points that fall within an 8° sector. The TOF increases roughly linearly
with radial distance, confirming that the phantom is largely isotropic in the
radial direction. The shear wave speed map is shown in Fig 4 in polar coordinates
from r=4 mm to r=26 mm. The arithmetic mean of wave speed was 1.51 mm/ms, with
standard deviation of 0.04 mm/ms. Note the standard deviation is affected by
the width and shape of the linear fit kernel.
Discussion
We have demonstrated the feasibility of a
method for SWEI with MRI using bipolar MEG. A high quality TOF map was
generated using the zero-crossing time of the motion-encoded phase, which is
relatively simple and robust. Shear wave speed as a function of the radial
distance and polar angle was calculated, which might be used to detect
pathological changes in tissue. Although we implicitly assumed the wavefront is always
perpendicular to radial spokes in calculating the speed map, this assumption is
not fundamentally necessary and more advanced algorithms can be used to
determine shear wave velocity from the TOF map.
Acknowledgements
The authors acknowledge our funding sources: P01 CA159992, InSightec and General Electric.References
1, Palmeri ML, Nightingale KR. Acoustic
radiation force-based elasticity imaging methods. Interface Focus. 2011 Aug
6;1(4):553-64.
2, Souchon R, Salomir R, Beuf O, Milot L,
Grenier D, Lyonnet D, Chapelon JY, Rouvière O. Transient MR elastography
(t-MRE) using ultrasound radiation force: theory, safety, and initial
experiments in vitro. Magn Reson Med. 2008 Oct;60(4):871-81.
3, Bitton R, Kaya E and Butts Pauly K.
Shear Wave Tracking in Cadaveric Breast Using MR-ARFI. Proceedings of ISMRM,
2012 (557).
4, McDannold
N, Maier SE. Magnetic resonance acoustic radiation force imaging. Med Phys.
2008 Aug;35(8):3748-58.