Mark J. van Uden1, Bart Philips1, Miriam Lagemaat1, and T.W.J. Scheenen1,2
1Radiology and Nuclear Medicine, Radboud university medical center, Nijmegen, Netherlands, 2Erwin L. Hahn Institute for Magnetic Resonance Imaging, Essen, Germany
Synopsis
Multiparametric
MRI (mpMRI) is a valuable tool to assess
aggressiveness and stage of localized prostate cancer. When combined with
metabolic information from 1H and 31P spectra valuable
information on the metabolism of the disease can be derived. Here we
demonstrate a dual element endorectal coil with 1H Rx and 31P
TxRx capabilities combined with an external 8-channel transmit-receive body
array. One patient (Gleason score 4+4) was measured. High quality proton
spectra show elevated total choline in tumor tissue, which corresponded with
increased glycerophosphocholine, glycerophosphoethanolamine and phosphocholine in
the 31P spectra of the tumor.Introduction
In prostate cancer (PCa) management multiparametric(mpMRI) proton MRI
forms an important tool to assess the aggressiveness and stage [1] of the
disease. Advancing this technique to 7 Tesla using a 1H endorectal
coil (ERC) in combination with an 8-channel body coil already showed promising
results for imaging and spectroscopy [2]. Combing the mpMRI results with
metabolic information derived from 31P spectra can add valuable
information on the metabolism of the disease.
Surface coils have the disadvantage of an extremely inhomogeneous B1
field which is especially unfavorable for techniques like fast or turbo spin
echo imaging (TSE). An earlier described double tuned ERC [3] with TxRx
capabilities on 31P and 1H has this drawback. Here we
demonstrate a dual element ERC with a 1H Rx element and a 31P
TxRx loop combined with an external 8-channel body array for homogeneous B1-shimmed
1H transmit fields at 7 Tesla.
Methods and Materials
The ERC is a modified commercially available 3T ERC (Medrad) and consist
of a 31P TxRx loop [4] and an asymmetric microstrip 1H receive
element. The 1H receive element is positioned on the rod (fig. 1) of
the ERC. The ground plane of the microstrip is rotated around the center of the
rod to focus the Rx field. Detuning was done with a series PIN-diode, that adds
some extra noise but ensures safe operation if not actively controlled. Via a 4:1
half-wave coax balun and a 75cm cable the element was connected to a
preamplifier. A separate 75cm cable was used to connect the 31P
element to a home-built transmit-receive switch with an integrated
preamplifier. A single floating proton cable trap enclosing both cables ensured
the absence of common mode currents. For transmitting on proton a multi-transmit
8-channel body coil was used [4,5].
The safety of the setup was ensured by EM-field simulations and
measurements and extensive safety measurements on the MR-system.
After patient-specific B0 and B1+ shimming of the prostate with the
multi-Tx external coil array a mpMRI protocol was performed with 1H
imaging, and both 1H and 31P spectroscopy. Transversal T2-weighted
and diffusion weighted imaging were performed as anatomical and functional
references. 1H-MRSI was done with a PRESS-like (TR=1000ms, TE=135ms,
FOV=84x70x70, matrix=12x10x10, 50% Hamming filter, NA=1, true voxel size 0.94
cc, acquisition time 7:01 minutes) sequence with spectral-spatial refocusing pulses
[6]. Only the spectral region in the VOI ranging from 2.3-3.3 ppm was refocused.
Lipids from outside the prostate were
saturated with two saturation bands in the ventral side.
The 31P MRSI sequence was played out with a non-localized
BIR-4 45 degree excitation pulse (TR=1500ms, FOV=120x100x100, matrix=12x10x10, NA=4,
true voxel size 4.2cc, acquisition time 13:09 minutes).
Initial measurements were performed on one patient with histologically
proven Gleason 4+4 prostate cancer.
Results
Safety tests showed no temperature increase or disturbances in the proton
B1+ field of the external body array when the ERC was present. Next to that the
EM-field simulations and measurements showed no change in EM-field
distributions of the dual-element ERC compared to the 31P TxRx ERC
(fig 2).
The T2W and ADC images acquired with the ERC clearly show the tumor in
the peripheral zone of the patients’ prostate (fig 3a&b). 1H and
31P MRSI voxels representing tumor and healthy tissue were picked
based on the 7T images. Because of Hamming
filtering of the MRSI k-space matrix, the true voxel shapes after 3D Fourier
transform were best represented by spheres, indicated by the colored circles (fig 3c-f).
Proton spectra (fig 3c&d) show elevated total choline (Cho) in the
cancer lesion which is almost absent in healthy tissue, in which large
polyamine and citrate signals are present. In the right peripheral zone of the
prostate (with tumor tissue) the B1-shim was not optimal, resulting in some
signal dropout in the ADC image as well as a distorted citrate shape in the
shown tumor voxel. The corresponding voxels in the phosphorous spectrum reveal
big differences in relative phospho-monoester-and -diester signals between the
tumor and the healthy tissue. Glycerophosphocholine (GPC), glycerophosphoethanolamine
(GPC) and phosphocholine (PC) were elevated in tumor compared to healthy tissue at equal
distance from the ERC (fig 3e&f).
Conclusion/discussion
In this work we present a
31P/
1H dual-element
endorectal coil combined with an 8-channel multi-TxRx body coil capable of
acquiring proton imaging, proton spectra and phosphorous spectra of the human
prostate in a single scan session at 7 Tesla. Care should be taken for
homogeneous B1-shimming of the whole prostate and an optimal balance between
voxel size, measurement time and SNR of the
31P MRSI experiment.
Acknowledgements
No acknowledgement found.References
[1] Hoeks,C.
ea. Radiology(2011)
[2] Maas MC. ea
ISMRM 2012 (#0964)
[3] Luttje
M.P. ea MRM(2014)
[4] Kobus,T.
ea. MRM(2012)
[5] Orzada,S.
ea ISMRM 2009 (#2999)
[6] Lagemaat MW et al, MRM 2015