Development of Novel Intravascular
MRI Coils
Background
The majority of all deaths
in the United States result from cardiovascular disease. The rupture of an
atherosclerotic plaque is often suspected to be the event precipitating a heart
attack or stroke. Magnetic resonance imaging (MRI) holds great promise for in
vivo plaque
characterization due to its potential for obtaining high-resolution images and
its sensitivity to the compositional characteristics of plaques. In particular,
it has been demonstrated that the canonical constituents of plaques -- lipid,
fibrous tissue, thrombus, and calcium phosphate mineral -- may be detected in ex
vivo specimens by
comparison among several magnetic resonance contrast weightings such as proton
density, T1- and T2-weighted imaging, and chemical shift
imaging. The success of MRI to date for in vivo plaque studies has not been as
profound as in ex vivo studies due to the poor image signal-to-noise ratio (SNR). The low in vivo SNR is a consequence of the small
size of plaques, their relative inaccessibility to external RF coils, the
proximity of flowing blood (which swamps the receiver dynamic range and creates
image artifacts due to fluid flow and coil motion), tissue movement (another
source of motion artifacts), and the short time available to obtain multiple
image acquisitions (the time required to obtain good SNR can easily exceed the
tolerance of the patient). These limitations have driven the development of
intravascular (iv) RF coils to significantly improve the SNR. The SNR is
proportional to the filling factor, which can be loosely defined as the ratio
of the volume of interest to the volume of the coil RF receptivity. To obtain
the very best possible filling factor for plaque MR scanning the coil should be
placed in immediate proximity to the plaque with the coil receptive volume
confined to the volume of interest only, thereby maximizing the plaque signal
strength while minimizing noise pickup from volumes which are not of interest
A wide variety of intravascular (iv) coil designs have been
reported. However, the coil designs developed to date have had limited success
and suffer from a number of common problems that are encountered in in vivo applications: (1) The most sensitive
region of the coil is typically contained in the lumen of the artery and not
the artery wall. This leads to a squandering of the dynamic range of the
receiver by the very intense blood signal, thereby hindering the detection of
the much smaller signals arising from the artery wall and atherosclerotic
plaques; (2) Severe image artifacts are frequently observed due to motion of
the coil during pulsatile blood flow. To reduce motion artifacts the blood
vessel may be occluded with a balloon, however, this can only be done for periods
of time that are too short for many imaging and spectroscopy applications; (3) because of their small size and
inaccessibility, intravascular coils, once positioned, cannot be tuned with
variable mechanical capacitors; (4) the long, small diameter coaxial cable
required leads to significant signal loss. We are adopting a variety of
approaches to overcome these problems, including the design of novel
intravascular coils with sensitive volumes tailored to the cylindrical arterial
wall geometry, as well as incorporating electrically variable tuning and
matching capacitors directly at the iv coil which may be remotely adjusted with
DC voltages, and finally incorporating miniature preamplifiers inside the
catheter in close proximity to the coil.
Meanderline
Coil Design
Intravascular RF coils are
being investigated by several research groups to improve the filling factor and
signal to noise ratio for NMR imaging and spectroscopy of atherosclerotic
plaque. These coils are introduced percutaneously into blood vessels via
catheters in much the same way as are surgical devices used for opening clogged
arteries. In order to obtain the best filling factor in such applications we
are investigating the use of novel RF coil structures with B1 field
profiles tailored to the geometry of the arterial wall. Cylindrical meanderline
coils provide a sensitive detection volume that is restricted to a cylindrical
shell, thereby maximizing the filling factor for the vessel wall, and plaques
in particular, while reducing the blood signal.
|
|
|
Figure 1:
Planar (left) and cylindrical (right) meanderline coil designs. The thickness
of the sensitive volume is determined by the spacing between the conductors. |
Wrapping a meanderline
into a cylinder (Figure 1) takes advantage of the structure’s tight B1
shaping properties, which should minimize blood flow image artifacts and
increase the dynamic range available from plaque signals. The antiparallel
alignment of adjacent conductors of the meanderline coil insures that B1
falls off rapidly with distance from the surface of the coil. This behavior is
demonstrated in Finite Difference Time Domain (FDTD) simulations of the
electromagnetic field (Figure 2) of a 3 mm o.d. cylindrical meanderline coil,
where the region of the most intense RF field is restricted to a thin
cylindrical shell. However, a more distorted, crescent shaped, field symmetry
is observed with increased loading of the coil with saline (s = 1.4, e = 70).
|
|
|
|
|
Figure 2:
FDTD simulations of a 3 mm meanderline coil in (a) free space, (b) immersed
in saline with the coil interior empty, and (c) immersed in saline with the
coil interior full. |
||
Shown in Figure 3 are prototype 3 mm and 2 mm o.d. meanderline
coils (6 conductors, 20 mm long, 6 mm wide, 1.6 mm spacing ) used to acquire
experimental 4.7 Tesla images. The coils are connected to a tuning and matching
pi-circuit via small diameter 50 ohm coaxial cable (1.8 mm o.d.) of length ~l/2. Experimental images acquired with the 3 mm
meanderline coil, shown in Figure 4, are in good qualitative agreement with the
FDTD simulations.
|
|
|
|
Figure 3:
Close-up views of 3 mm (a) and a 2 mm (b) o.d. cylindrical meanderline coils.
The coils are produced by etching copper clad Kapton polyimide flexible
printed circuit substrate. The pictured coils contain six conductors and have
inductances on the order of 50 nanohenries. |
|
|
|
|
|
Figure 4: Image acquired with a 3 mm diameter meanderline coil
immersed in 1.0 mM MnSO4 doped water with the coil interior either empty
(left) or filled (right). The field of view is 20 mm and the in -plane
resolution is 0.16 mm. |
|
The distortion of the field symmetry, observed both in the
simulations and experimentally, can be minimized by incorporating tuning
capacitance into the meanderline coil as demonstrated in both the FDTD
simulations of the magnitude of the magnetic field and the experimental data
shown in Figure 5.
|
|
|
|
|
|
|
Figure 5: FDTD simulations of a 3
mm od cylindrical meanderline coil without (top, left) and with six 60 pF
tuning capacitors distributed in the coil. The simulations are in very good agreement
with the experimental images obtained without any local tuning capacitor
(bottom, left) and with a 10 pF tuning capacitor (bottom, right). |
|
Finally, images of an endarterectomy specimen (see Figure 6) were acquired
at 4.7 Tesla with a 127 mm diameter bird-cage volume coil used in
transmit-receive mode and a 3 mm meanderline surface used in receive mode with
the volume coil used in transmit mode only. The SNR is clearly superior for the
image acquired with the receive-only cyclindrical meanderline coil over that
obtained with the transmit-receive volume coil.
|
|
|
|
Figure 6: Endarterectomy images
acquired at 4.7 T with (left) a 127 mm bird-cage coil and (right) a 3 mm meanderline
coil. The FOV was 25 mm with a 1 mm slice thickness and a 0.1 mm in plane
resolution. |
|
These preliminary studies demonstrate that a cylindrical
meanderline coil provides a cylindrical shell sensitive volume that should be well
suited for imaging blood vessel walls and atherosclerotic plaque. The
cylindrical meanderline coil design provides several potential advantages over
other more commonly employed catheter coil designs: (1) the meanderline coil sensitive volume is limited to a cylindrical
shell, allowing for the selective imaging of arterial walls with minimal or no
signal from blood, thereby reducing dynamic range problems due to the intense
blood signal; (2) the more restricted volume of sensitivity of the meanderline coil
reduces the noise picked up by the coil due to the highly conductive blood,
thereby improving the SNR; (3) because of its readily conformable and
expandable geometry, the cylindrical meanderline coil could be made from
shape-persistent materials or expanded with an annular balloon allowing it to
be deployed similarly to a stent, thereby matching its diameter with that of
the blood vessel wall and minimizing imaging artifacts due to coil motion
during pulsatile blood flow by stabilizing the coil against the vessel wall;
(4) the center of the coil is open and allows for unimpeded blood flow during
image acquisition; (5) since the coil B1 field is always
perpendicular to the long axis of the coil cylinder there will always be some
component of the coil B1 field perpendicular to the applied field B0
for any orientation of the coil and hence a signal will always be observed. In
contrast, signal dropout is observed with other coil designs where, due to
blood vessel tortuosity, some coil orientations result in the coil’s B1
field (or axis of sensitivity) being oriented parallel to the applied static
magnetic field.