AJR 2000; 175:447-449
© American Roentgen Ray Society
Modified Four-Coil Phased Array Assembly for High-Resolution MR Imaging of the Cerebellopontine Angle
J. H. M. Chan1,
W. C. G. Peh2,3,
K. P. C. Wong1,
S. H. Luk1,
E. Y. K. Tsui1 and
M. K. Yuen1
1
Department of Diagnostic Radiology, Tuen Mun Hospital, Tuen Mun, Hong Kong,
China.
2
Department of Diagnostic Radiology, The University of Hong Kong, Pokfulam,
Hong Kong, China.
3
Present address: Department of Diagnostic Radiology, Singapore General
Hospital, Outram Rd., Singapore 169608.
Received June 1, 1999;
accepted after revision January 6, 2000.
Address correspondence to W. C. G. Peh.
Introduction
High-resolution three-dimensional (3D) MR imaging of the internal auditory
canals is acquired using two 3-inch surface coils configured as a dual phased
array coil assembly
[1,2,3].
The depth sensitivity of the two-coil phased array assembly is limited by the
effective field of view of the 3-inch coil, making visualization of deep
structures such as the midbrain, cerebellar vermis, and deep temporal lobes
difficult. The design and construction of various types of four-coil phased
arrays for high-resolution MR imaging of the human brain and orbits have been
reported
[4,5,6,7].
To our knowledge, all existing four-coil phased arrays require installation of
four independent radiofrequency (RF) receivers. Because four independent RF
receivers are not standard equipment in most MR imaging centers, the aim of
the study was to design and test a four-coil phased array assembly comprising
two conventional 3-inch surface coils and two conventional 5-inch surface
coils, and also to quantitatively and qualitatively compare the new coil
assembly with conventional 3-inch surface coils.
Materials and Methods
MR imaging was performed using a clinical 1.5-T MR scanner (Signa Horizon
Echospeed, version 5.6 software; General Electric Medical Systems, Milwaukee,
WI). Ten consecutive adult patients (mean age, 46 years) with symptoms of
hearing loss and tinnitus were examined using both conventional two-coil
phased array and the new four-coil phased array assemblies. Informed consent
was obtained from all patients. Two 3-inch surface coils were connected to a
dual phased array adapter, which was then plugged into the phased array coil
port on the head carriage. The two 3-inch coils were centered bilaterally over
the external auditory canals. After the coronal fast spin-echo T2-weighted
localizer images (TR/TE, 2000/85; echo train length, 16) were obtained,
high-resolution axial T2-weighted images were acquired using 3D fast spin-echo
pulse sequence (TR/TE, 4000/131; echo train length, 64; receiver bandwidth,
31.2 kHz; field of view, 13 x 13 cm; number of partitions, 30; partition
thickness, 0.8 mm; matrix, 256 x 256; and excitation, one).
Two 5-inch surface coils were then connected to a dual combiner box plugged
into the surface coil port on the head carriage. These 5-inch coils were
positioned above and below the region of interest (ROI) in addition to the two
3-inch surface coils. The image acquisition was then repeated using the same
pulse sequences, imaging parameters, and coil configuration file as the dual
3-inch phased array coil. All MR images were qualitatively assessed by two
radiologists. After removal of annotations, MR images were randomized before
being independently evaluated by each radiologist. The seventh and eighth
cranial nerves, the midbrain, and the temporal lobes were graded using the
following criteria: poor visualization, grade 1; adequate visualization
sufficient for diagnosis, grade 2; good visualization, grade 3. The signal
intensity was measured by placing an ROI cursor over the deep portion of each
temporal lobe and the center of the midbrain
(Fig. 1). Care was taken to
ensure that the ROIs were of the same size and shape, and that the same
anatomic site in each patient was assessed. The signal-to-noise ratio (SNR)
obtained at each site using the two- and four-coil phased arrays was
calculated and compared using the Student's t test. Image uniformity
was estimated by comparing the variation in SNR obtained by the two different
coil assemblies.
The performance of the modified four-coil phased array assembly was also
assessed qualitatively and quantitatively by imaging a spherical phantom to
simulate the shape and tissue of the brain. MR imaging was performed with the
four-coil array (Fig. 2) and
then repeated with the dual 3-inch phased array using identical imaging
parameters. The 20-cm-diameter spherical phantom, approximately the size of a
human head, was filled with 0.1 mol of nickel chloride solution. Only
conventional spin-echo T1-weighted axial images (TR/TE, 500/8) were acquired.
Signal intensities were measured by placing an ROI cursor over the central 80%
of the phantom image. The maximum signal intensity (Smax) and the
minimum signal intensity (Smin) inside the ROI cursor were noted
and used to compute the image uniformity by the formula (Smax -
Smin) / (Smax + Smin). The signal intensities
at the center, top, bottom, right, and left regions of the phantom were
measured by placing an ROI cursor at these sites. The SNR values obtained with
the two- and four-coil phased arrays at each site were compared.
View larger version (96K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2. —Photograph shows modified four-coil phased array assembly placed
around phantom. Two 3-inch surface coils (either side of phantom) are
connected to phased array coil adapter and two 5-inch surface coils (top and
bottom of phantom) are connected to surface coil combiner box.
|
|
Results
The radiologists' scores in assessment of the midbrain and the temporal
lobes on images obtained by the modified four-coil phased array assembly (mean
scores of 2.85 for the midbrain and 2.75 for the temporal lobes) were
significantly better (p < 0.05) than those obtained by the
two-coil array (mean scores of 2.05 for the midbrain and 1.85 for the temporal
lobes). Scores in assessment of the seventh and eighth cranial nerves were not
statistically significant (p > 0.05) for both coils (mean scores
of 2.60 for four-coil and 2.35 for two-coil). There was no interobserver
difference in quantitative grading.
The SNR measurements at the midbrain and the right and left temporal lobes
showed that the two-coil phased array assembly was inferior to the modified
four-coil phased array assembly. The SNR of the midbrain and the right and
left temporal lobes revealed an improvement of approximately 38%, 25%, and
22%, respectively, using the modified four-coil phased array assembly. The
seventh and eighth cranial nerves were equally well seen on the images
acquired with the two-coil assemblies in all but two patients, who had large
acoustic schwannomas. Visualization of the cranial nerves was not possible
because of the presence of these large tumors. However, using the modified
four-coil phased array assembly, the overall image quality was superior in
terms of image uniformity, with better display of the cerebellar vermis and
deep temporal lobes (Fig.
3A,3B).
Using the modified four-coil phased array assembly, the phantom study yielded
an improvement of about 30% in the SNR at the center of the phantom and an
improvement in image uniformity of about 70%. Measurements of the SNR at
different regions of the phantom showed an approximately twofold increase at
the top and bottom regions of the phantom caused by the close proximity of the
two 5-inch coils.
View larger version (150K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3A. —40-year-old woman with pulsatile tinnitus. Axial fast spin-echo MR
images (TR/TE, 4000/131; echo train length, 64) with conventional two-coil
phased array assembly (A) and modified four-coil phased array assembly
(B) show cerebellar vermis and deep temporal lobes. Note better
resolution and gray-white differentiation on B.
|
|
View larger version (177K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3B. —40-year-old woman with pulsatile tinnitus. Axial fast spin-echo MR
images (TR/TE, 4000/131; echo train length, 64) with conventional two-coil
phased array assembly (A) and modified four-coil phased array assembly
(B) show cerebellar vermis and deep temporal lobes. Note better
resolution and gray-white differentiation on B.
|
|
Discussion
In principle, a four-coil phased array requires the MR imaging system to be
equipped with four independent RF receivers. This allows MR signals from each
independent surface coil to be combined using a sum-of-squares method on a
pixel-by-pixel basis. Because only the coil configuration file for dual phased
array is selected during prescription, only the two 3-inch coils receive RF
signals, which are then transferred to the two corresponding RF receivers.
When our modified four-coil phased array assembly is used, the two 3-inch
coils work only as signal-receiving coils and the two 5-inch coils act only as
signal-reflecting coils. During spin relaxation, MR signals are generated and
induced in the two 5-inch coils. The signal current flows to the terminals of
the circuit and, because it is not passed to the RF receiver, it is reflected
back to the 5-inch coils. Hence, the 5-inch coils act as RF antennae, emitting
electromagnetic waves when the signal current passes through them.
Subsequently, the two 3-inch coils pick up the electromagnetic waves, which
are then transmitted to the RF receivers. The end result is that the two
5-inch coils reflect the MR signals emitted from the top and bottom parts of
the anatomy onto the two 3-inch coils and are picked up by them.
After using our modified four-coil phased array assembly, we see no need to
purchase two extra RF receivers and no requirement for engineering
modifications. This coil assembly can be used on any MR system with two RF
coil ports on its gantry, one for the surface coil and one for the phased
array coil. We have shown, using both a phantom and human subjects, that the
modified four-coil phased array assembly is superior to the conventional
two-coil phased arrays in terms of SNR and image uniformity. We recommend
validating the use of this modified phased array assembly in a larger group of
patients with various types of lesions affecting the cerebellopontine angle
and surrounding areas. We believe that, in addition to providing
high-resolution MR images of the internal auditory canals, the four-coil array
assembly can potentially be used for high-resolution MR imaging of other
anatomic regions such as the hippocampus, orbits, and optic nerves.
References
-
Lee JN, King BD, Parker DL, et al. High resolution 3D imaging of
the inner ear with a modified fast spin echo pulse sequence. J Magn
Reson Imaging 1996;6:223
-225[Medline]
-
Naganawa S, Itoh T, Fukatsu H, et al. Three-dimensional fast
spin-echo MR of the inner ear: ultra-long echo train length and half-Fourier
technique. AJNR
1998;19:739
-741[Abstract]
-
Stone JA, Chakeres DW, Schmalbrock P. High-resolution MR imaging of
the auditory pathway. Magn Reson Imaging Clin N Am
1998;6:195
-217[Medline]
-
Wald LL, Carvajal L, Moyher SE, et al. Phased array detectors and
an automated intensity-correction algorithm for high-resolution MR imaging of
the human brain. Magn Reson Med
1995;34:433
-439[Medline]
-
Hayes CE, Mathis CM, Yuan C. Surface coil phased arrays for
high-resolution imaging of the carotid arteries. J Magn Reson
Imaging 1996;6:109
-112[Medline]
-
Breslau J, Dalley RW, Tsuruda JS, Hayes CE, Maravilla KR.
Phased-array surface coil MR of the orbits and optic nerves.
AJNR
1995;16:1247
-1251[Abstract]
-
Schmalbrock P, Pruski J, Sun L, Rao A, Monroe JW. Phased array RF
coils for high-resolution MRI of the inner ear and brain stem. J
Comput Assist Tomogr
1995;19:8
-14[Medline]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Introduction
Materials and Methods
