YP Liao1, A.-M. Oros-Peusquens1, J. Lindemeyer1, N. Lechea1, C. Weiss2, G. Stoffels1, C. Filss1, K.J. Langen1, and N.J. Shah1,3
1Institute of Neuroscience and Medicine-4, Forschungszentrum Juelich, Juelich, Germany, 2Department of Neurosurgery, University of Cologne, Cologne, Germany, 3Department of Neurology, JARA, Faculty of Medicine, RWTH Aachen University, Aachen, Germany
Synopsis
The availability of combined MR-PET scanners opens
new opportunities for the characterisation of the tumour environments. In this
study, MR-based simultaneous water content, electrical conductivity and
susceptibility mapping in meningioma patients was implemented based on a
multi-echo gradient echo sequence. The information was complemented by
characterisation of the tumour with simultaneous FET-PET. This is a powerful
combination of parameters which reflect important aspects of tissue physiology
and also characterise to a large extent, tumour electromagnetic (EM)
properties. This multi-parametric information helps to understand pathological
tissue and can be applied to planning nonionizing
EM hyperthermia therapy.Purpose
Compared to traditional T
1
and T
2 contrasts, MR-derived quantitative water content, electrical
conductivity and magnetic susceptibility more directly reflect physiological tissue properties (water, ion and ferritin content) and physical tissue properties
describing its interaction with electromagnetic fields. Moreover, electrical
conductivity can provide additional valuable information for SAR estimation in
ultra high field MRI and for patient specific non-ionizing EM hyperthermia
therapy
1. This study was based on a single multi-echo GRE sequence
with long TR and total acquisition time of 7 ½ minutes
2 and thus
completely compatible with clinical protocols. To the best of our knowledge
this is the first report of simultaneous acquisition of water content,
conductivity and susceptibility in tumour patients featuring moreover tumour
characterisation by dynamic PET with O-(2[18F]-fluoroethy1)-L-tyrosine tracer
(FET-PET). Water content information was mainly derived from the magnitude of
GRE images while conductivity and susceptibility maps were retrieved from the
multi-echo phase profiles
3.
Methods & Materials
Two meningioma patients (62-year-old
male and 64-year-old Female) were investigated in a hybrid MR-PET 3T scanner as
part of the pre-surgery planning. The prototype 3T MR-PET scanner comprised a
commercially available 3T Siemens Trio MR system and a custom-built
MR-compatible Brain PET scanner
4. Two dedicated head coils for MR
including an outer birdcage coil for transmit and an inner 8 channels coil for
receive were placed in the PET detector. The MR Brain PET delivers PET images
with an optimal resolution of 3mm. For quantitative
mapping, a 2D multi-echo gradient echo sequence was employed with TR=10s and
nominal flip angle of 90°. FOV=200×162mm
2; slice thickness=1.5mm;
TE1=3.87ms;echo separation ΔTE=4.08ms; 12 echoes; acceleration factor=2. To calculate the water content and its complementary proton density, T
1,
T
2* and transmit-receive field corrections usually need to be
applied. However, in our long TR protocol T
1 saturation can be
neglected for brain tissue, which at 3T has T
1<2s. It is still present for cerebral spinal
fluid (CSF) and can be
calculated using its known T
1 of 4.3s
5. T
2*
mapping is performed by a mono-exponential fitting to the signal decay of GRE.
The combined effect of transmit and receive RF field inhomogeneity can be
corrected using e.g. the bias field correction algorithm implemented in SPM
2.
The water content map is obtained by calibrating the T
2* and bias
field-corrected tissue values to the saturation-corrected CSF signal intensity. Reconstruction of the electrical conductivity is
based on electrical properties tomography: σ=(ΔΦ)/(2μω), where Δ represents the
Laplacian operator, μ the magnetic permeability, and ω the Larmor frequency.
The transceiver phase was calculated from the 12 multi-echo GRE phase profiles
by interpolating them at the TE=0 after unwrapping. A local quadratic fitting
of the transceiver phase was performed before applying the second-order
derivative
6. For
susceptibility reconstruction, the B0 field map was estimated by unwrapping the
phase of all GRE echoes and applying linear regression, the effect of
background field shifts was eliminated using the in-house software MUBAFIRE
7.
Susceptibility estimation was performed using a minimisation strategy with
Tikhonov- and gradient-regularisation
8. In addition, pre- and
post-contrast-enhanced T
1 weighted (MPRAGE) and T
2 weighted
(FLAIR) images were obtained and co-registered to the electrical conductivity
map. For
further clinical analysis and comparison, tumour regions were identified and
segmented based on the PET data.
Results & Discussion
Fig1a)-
c) show the reconstructed water content, electrical conductivity and
susceptibility mapping in one patient. The water content values, defined as
mean and SD of the distribution over all voxels assigned to a given tissue
class, were 70.3±3.5 % in white matter and 83.5±6.3 % in grey matter. The
conductivity values were 0.41±0.15 S/m in the white matter, 0.83±0.32 in Grey
matter and 2.11±0.43 S/m in CSF. The reconstructed PET map (average tracer
uptake over 30 minutes) of one patient is shown in Fig1d), Fig1e)-f) represents
the enhanced T
1weighted MPGRAGE map and T
2 weighted
FLAIR. Water content in tumour region defined by PET is 85.54± 0.91%, while
conductivity is 0.91±0.18 S/m. Compared to PET and contrast
enhanced T
1 weighted images, both the water content and conductivity
maps are able to differentiate the tumour region from surrounding tissue.In
both cases, the values are higher in the tumour than in surrounding tissue. Fig 2 plots the water content versus conductivity in four region of interests including water matter, grey matter, CSF, and tumour.
Susceptibility also provides useful information for defining the edge of the
tumour region and thus provides complementary information to conductivity.
Conclusion
Simultaneous mapping of water
content, conductivity and susceptibility based on multi-echo GRE is feasible.
This method makes full use of the magnitude and phase profiles of a standard
sequence and can be applied to investigate and understand pathological tissues.
Acknowledgements
No acknowledgement found.References
[1] Esrick et al., Phys.Med.Biol. 1994,39: 133-144.
[2] Oros-Peusquens A-M et al., Nucl Instrum Meth A., 2014,734:185-90.
[3] DH Kim et al., MRM, 2014, 71:1144-50.
[4] Weirich C et al., IEEE TMI, 2012,31:1372-80.
[5] Rooney W et al., MRM, 2007,57:308:318.
[6] Katcher U et al., ISMRM, 2012,p.3428.
[7] Lindemeyer J et al., PLOS ONE. 2015, Epub.
[8] de Rochefort.L et al., MRM., 2010,63:194-206.