Maxim Terekhov1, David Lohr1, Michael Hock1, Maya Bille1, Julia Aures1, Ibrahim A. Elabyad1, Florian Schnitter2, Wolfgang Bauer2, Ulrich Hofmann2, and Laura M. Schreiber1
1Chair of Cellular and Molecular Imaging, Comprehensive Heart Failure Center, University Hospital Würzburg, Comprehensive Heart Failure Center, Wuerzburg, Germany, 2Department of Internal Medicine I, Cardiology, University Hospital Würzburg, Wuerzburg, Germany
Synopsis
Cardiac MRI with T2*-contrast at 7T is of
particular interest because it is highly sensitive to cardiac tissue alterations
after myocardial injury. A reliable T2*-contrast
quantification and usage for tissue characterization require minimization of
the inhomogeneity of the B0-field in the heart. In this work, we analyzed the B0-conditions achievable in longitudinal
T2*
measurements in the hearts of domestic pigs which during the experiment
grew from 30 to 90 kg. B0
distribution statistics were summarized for the maps of individual slices. Demands on the 7T-scanners B0-shimming
hardware (up to 3rd-order spherical harmonics) were analyzed.
Introduction
Cardiac MRI with T2*-contrast at 7T is of particular interest
because of high sensitivity to cardiac tissue alterations after myocardial
injury. This includes deposition of iron,
hemorrhage, microvascular obstruction, and other factors influencing the local
magnetic susceptibility [1,2]. The
reliable quantification of a T2* for myocardial tissue
characterization crucially requires minimizing the inhomogeneity of the B0-field
in the heart, especially at the interfaces to the lung and liver tissue. The
main goals of this study were: (1) to analyze
the B0-conditions achievable in the longitudinal T2* measurements in the heart of the domestic pig
of 30 to 90 kg weight range and (2) to
find out the demands put by this B0-conditions for the 7T scanner B0-shimming
hardware and limitations that are met. Methods
In total, the
data of n=8 pigs (German Landrace) were used measured with the approval of the local
Animal Welfare Committee (55.2 DMS 2532-1134-16, Government of Lower Franconia). All measurements were done using a Magnetom™
“Terra” 7T MR-scanner (Siemens, Erlangen, Germany) using custom-designed 8Tx/16Rx
cardiac array coils adapted for a specific animal weight range. The
scanner is equipped with a dedicated shimming system providing shim terms up to 3rd
order spherical harmonics. The total shimming gradients configuration includes
three linear (first-order) terms (X,
Z, Y), five second-order terms (Z2, XY, ZX, ZY, X2-Y2),
and four third-order terms (Z3, Z2X, Z2Y, Z(X2-Y2)).
Each animal was
scanned 4 times within 2 months while growing from 30-35 to 85-90kg. The acquisition of standard cardiac views: long-axis (LA), short-axis (SA), and
two-chambers was done using the cardiac triggered (EasyACT, MRI.Tools system )
GRE-CINE sequence with the retrospective reconstruction of 30 cardiac phases. T2*-measurements and B0-mapping
were performed using a triggered mGRE-sequence with 9 TE-times distributed in
the range [1.1..14.6]ms. A contiguous stack of 10-13 short-axis view (SA)
slices with 6mm thickness and in-plane pixel size of 2.2x2.5mm was acquired. By planning the CINE views, the B0-shimming
volume was set to cover the whole visible heart volume. At the same time, for the T2* measurements, the
shimming volume was adjusted on a slice-by-slice basis to force optimal B0
conditions.
The reconstruction
of B0 maps was done using ROMEO
[3] software. Image segmentation and statistical analysis (including shimming
protocol data) were done by in-house developed Matlab scripts (Mathworks, Natick,
USA). The shim currents data were extracted retrospectively from the header of the DICOM files. Results
Figure 1
shows the examples of B0-maps in the whole SA-view of the heart and in the
segmented left ventricle for the slices from basal to apical and covering
different animal weights. Figure 2 shows the correlation between T2*-data
(a) and B0-maps (b) for the slices at basal, mid-myocardium, and apical
location. One can observe that increased localized B0-gradients
influence significantly the T2* at corresponding segments of the myocardium.
Figure 2c shows the statistics of residual inhomogeneity of B0
remaining after slice-by-slice shimming
(computed over whole measured slices). Figure 3 compares the usage of shim currents for shimming volume placed
over the whole heart and slice-by-slice basis.
Figures 4 and 5 demonstrate the trends of the usage of shim coils
currents for the slice-by-slice B0-shimming in pigs with different heart
dimensions and averaged for the slices at different positions in the myocardium
respectively. Discussion
The results
presented in Figures 1 and 2 confirm the
importance of the B0-shimming for the T2* quantification. Even after dedicated shimming performed for
each slice the local gradients of high orders persist at the interfaces to the
lung (mostly at the posterior and lateral heart walls). The strength of the
local gradients increases in the apical direction with the local curvature of
the myocardial wall. The localization of the high-order spherical harmonics
gradients is in good agreement with earlier results of the study done in human
subjects [4]. Figure 3 demonstrates
that when the targeted shim volume is narrowed to the specific slice the demand
on the compensation of the high order local gradients drastically increases. In
about 50% of measured slices, the
demanded current for 3rd order shimming meets the hardware limitations. The analysis of the shim currents demand for a
pig of different weights shows that the most problematic challenge in terms of 3rd order shim
currents is small hearts with a strong curvature. This demand decreases with the
increasing size of the heart. Finally, the same trend is observed when
considering demand on the 3rd order shimming in the slices at different
positions ( Figure 5). The shim currents reach the hardware limit in up to 30%
of the apical slices with the largest relative interface area to the lung tissue
and strongest curvature. Conclusion
The T2* measurements at 7T covering the whole
heart volume and especially apical regions represent a major challenge for the
scanner’s B0-shimming system. The availability of the 3rd order shimming
hardware is highly recommended for such studies. Slice-by-slice shimming in the
majority of the cases meets the hardware limitations of the available 2nd and 3rd
order shim amplifiers. Development of B0-shimming
technologies based on the usage of the localized shimming coils [5] can open
possibilities for further improvements of the B0-shimming quality in cardiac
MRI at 7T. Acknowledgements
Financial support: German Ministry of Education and
Research (BMBF, grants: 01EO1004, 01EO1504).References
[1] Ghugre, N.R., Ramanan, V., Pop, M., Yang, Y., Barry, J., Qiang, B.,
Connelly, K.A., Dick, A.J., Wright, G.A., 2011. Quantitative tracking of edema,
hemorrhage, and microvascular obstruction in subacute myocardial infarction in
a porcine model by MRI. Magnetic Resonance in Medicine 66, 1129–1141..
doi:10.1002/mrm.22855
[2] Elabyad, I.A., Terekhov, M.,
Lohr, D., Stefanescu, M.R., Baltes, S., Schreiber, L.M., 2020. A Novel
Mono-surface Antisymmetric 8Tx/16Rx Coil Array for Parallel Transmit Cardiac
MRI in Pigs at 7T. Scientific Reports 10.. doi:10.1038/s41598-020-59949-6
[3] Dymerska, B., Eckstein, K., Bachrata, B., Siow, B.,
Trattnig, S., Shmueli, K., Robinson, S.D., 2021. Phase unwrapping with a rapid
opensource minimum spanning tree algorithm (ROMEO). Magnetic Resonance in Medicine 85, 2294–2308..
doi:10.1002/mrm.28563
[4] Hock, M., Terekhov, M.,
Stefanescu, M.R., Lohr, D., Herz, S., Reiter, T., Ankenbrand, M., Kosmala, A.,
Gassenmaier, T., Juchem, C., Schreiber, L.M., 2021. B 0 shimming
of the human heart at 7T. Magnetic Resonance in Medicine 85, 182–196..
doi:10.1002/mrm.28423
[5] Juchem, C., Theilenberg, S., Kumaragamage, C., Mullen, M., Delabarre,
L., Adriany, G., Brown, P.B., Mcintyre, S., Nixon, T.W., Garwood, M., Graaf,
R.A., 2020. Dynamic multicoil technique (DYNAMITE) MRI on human brain. Magnetic
Resonance in Medicine 84, 2953–2963.. doi:10.1002/mrm.28323