Yi-Hang Tung1, Hendrik Mattern1,2,3, and Oliver Speck1,2,3,4
1Department of Biomedical Magnetic Resonance, Institute for Physics, Otto von Guericke University Magdeburg, Magdeburg, Germany, 2German Center for Neurodegenerative Diseases (DZNE), Magdeburg, Germany, 3Center for Behavioral Brain Sciences (CBBS), Magdeburg, Germany, 4Leibniz Institute for Neurobiology (LIN), Magdeburg, Germany
Synopsis
Keywords: Diffusion Acquisition, Challenges
Motivation: Probing brain microstructure with high-b-value brain diffusion imaging was mainly performed at 3T. At higher field strength, the long echo times prohibited a significant SNR boost over 3T. At 7T, a high-performance gradient allows much shorter echo times and harvesting of the field-related signal gain.
Goal(s): Determine SNR and explore feasibility of high b-value DWI.
Approach: The SNR was measured in high b-value DWI at whole-body and high-performance 7T scanners.
Results: The SNR improvement on the high-performance 7T is 2.76 ±0.12 and 1.73 ±0.05 compared to the whole-body 7T and high-performance Connectome 3T respectively, allowing even higher spatial resolution and higher b-value imaging.
Impact: The shortcoming
of 7T for high b-value diffusion MRI can be overcome when leveraging a high-performance
gradient system, effectively reducing scan time 7.6-fold compared to a whole-body
7T scanner. This enables the next generation of diffusion imaging.
Background
7T MRI
enabled higher SNR in many applications, but diffusion MRI remains challenging1,2.
The reason is a reduced apparent transversal relaxation time (T2) and increased
geometric distortion compared to 3T. Due to the required diffusion encoding time
in whole-body 7T systems, the reasonably applicable b-value is lower than at 3T.
b-values larger than 3000 s/mm2 with a spatial resolution higher
than 2 mm isotropic have rarely been applied. A high-performance gradient system
can significantly reduce encoding and readout duration to compensate for
the accelerated signal decay. We compare high b-value DWI at 7T with whole-body and high-performance
gradient scanners and on a 3T Connectome scanner.Methods
Experiments
are performed on two 7T human scanners (Siemens Healthineers, Erlangen, Germany).
The first scanner (Magnetom 7T Plus) is equipped with a conventional whole-body
gradient system (70 mT/m, 200 T/m/s) and an 8 TX/32 RX head coil, referred to as
7T Plus, while the Magnetom Terra.X Impulse Edition has a high-performance
gradient (200 mT/m, 900 T/m/s) and an 8 TX/64 RX head coil, referred to as
Terra.X. Diffusion MRI data of one healthy subject (scanned after giving written,
informed consent, study approved by local institutional review board) were
acquired with the vendor-supplied spin-echo EPI sequence in the latest software
versions (VE12U for 7T Plus and XA70 for Terra.X).
The first imaging
protocol was adapted from the MGH Adult Diffusion Data protocol3
(1.5 mm3, iPAT 3×1, whole-brain coverage, maximum b-value 10 ms/μm2)
with the changes of iPAT 3×2, b-value (1, 2.5, 4, 10 ms/μm2) and
three orthogonal diffusion directions for each b-value. The TR was 8000 ms for
both scanners, and the minimum TE is 110 ms for 7T Plus and 59 ms for Terra.X.
The data was corrected by FSL topup and eddy4. An SNR comparison
with 3T Connectome data (300 mT/m, 200 T/m/s) was also performed. The second
experiment was performed on the Terra.X with a b-value of 16 ms/μm2,
1.4 mm3, iPAT 3×2, 128 diffusion directions, TR/TE 9000/67 ms; the
total scan time was 21 minutes.Results
Figure 1 shows
the minimum TE for different b-values on both 7T systems. On the 7T Plus, TE
increases roughly linearly with b-value. On the Terra.X, TE remains unchanged from
b = 0 to b = 4 ms/μm2, and only then increases linearly to the b-value
of 16 ms/μm2. The TE-slope for Terra.X is only 20% of the 7T Plus.
In Figure
2, the b = 0 images and DW images from both 7T scanners are shown. The white
matter structure is barely detectable on the 7T Plus for b > 2.5 ms/μm2,
while the white matter remains clear up to b = 10 ms/μm2 on the
Terra.X.
In Figure
3, the DW images at b = 10 ms/μm2 are compared between Terra.X and
the 3T Connectome. The 7T images show a visually higher SNR and lower eddy-current
related Nyquist ghost.
Figure 4
displays the SNR evaluation of all data. The noise floor of the 7T Plus data is
1.95 ± 0.06 times higher than Terra.X (Fig. 4a). The noise floor-subtracted SNR
(Fig. 4b) is 2.76 ± 0.12 times higher for the Terra.X compared to the 7T Plus. The
SNR of the 7T Plus is 0.63 ± 0.05 times the SNR of the Connectome 3T and
Terra.X shows 1.73 ± 0.11 time the SNR of the 3T Connectome (Fig. 4c).
The
achievable high quality of the Terra.X for high resolution, high b-value
diffusion imaging is shown in Figure 5 (whole brain DWI with 1.4 mm3
resolution and b = 16 ms/μm2).Conclusions
7T in combination
with a high-performance gradient system allows to realize the envisioned field
strength related SNR gain also for high b-value DWI. The SNR is 2.76 times that
of a whole-body 7T scanner and 1.73 times that of the 3T Connectome scanner. Imaging
parameters were kept as similar as possible in this study. The
high-performance gradient system enables further increase of the readout
bandwidth allowing further reduction of TE, SNR gain, and higher resolution
single-shot EPI.Acknowledgements
Data collection and sharing for the Connectome 3T data was provided
by the Human Connectome Project (HCP). HCP funding was provided by the National Institute of Dental
and Craniofacial Research (NIDCR), the National Institute of Mental Health
(NIMH), and the National Institute of Neurological Disorders and Stroke
(NINDS).References
- Pohmann
R, Speck O, Scheffler K. Signal-to-noise ratio and MR tissue parameters in
human brain imaging at 3, 7, and 9.4 tesla using current receive coil arrays. Magn.
Reson. Med. 2016; 75(2), 801-809.
- Ladd
M, Bachert P, Meyerspeer M, et al. Prog Nucl Magn Reson Spectrosc. 2018;
Dec:109:1-50.
- MGH Adult Diffusion Data Acquisition Details. https://www.humanconnectome.org/study/hcp-young-adult/document/mgh-adult-diffusion-data-acquisition-details.
- Andersson J,
Sotiropoulos S. An integrated
approach to correction for off-resonance effects and subject movement in
diffusion MR imaging. NeuroImage. 2015; 125, 1063-1078.