0764
Time-Resolved MRI of the Human Brain with 3.5 ms per Frame1Institute for Biomedical Engineering, University of Zurich and ETH Zurich, Zurich, Switzerland
Synopsis
MR imaging with shortest possible acquisition times is targeted by single-shot spiral readouts and extremely rapid gradient encoding. The implementation involves a first demonstration of spiral imaging using a gradient insert to achieve a slew rate of 1200 T/m/s. The achievable imaging performance is evaluated and in-vivo results are presented. The setup permitted imaging with more than 280 frames per second, which also allowed the extraction of rudimentary spectral imaging information from a single MR readout. The newly available temporal resolution in MR may be utilized to gain new insights in brain function and physiology.
Introduction
The achievable minimum time required to encode an MR image defines to which extent dynamic processes can be resolved. Moreover, shortest possible imaging readouts maximize robustness against motion. Single-shot spiral MRI [1-3] is among the fastest ways to encode an MR image and thus a candidate to best achieve these goals. A combination with parallel imaging can further accelerate acquisition speed and has previously enabled in-vivo single-shot acquisition of the brain [4-6]. Moreover, faster gradient hardware may be employed. This became possible with a recently developed head gradient insert [7] that allows producing a gradient strength of 100 mT/m with a slew rate of 1200 T/m/s. In this work, ultra-fast time-resolved MRI using accelerated single-shot MRI with extremely rapid gradient encoding is targeted. The achievable imaging performance is evaluated and first in-vivo data are presented.
Methods
MR scanning was performed on at 3T Achieva (Philips Healthcare, Best, The Netherlands) using an 8-channel transmit-receive coil array [8]. A custom-built head gradient insert was used with 100 mT/m and 1200 T/m/s. A key element of the design lies in the short gradient length, which allows to effectively limit peripheral nerve stimulation [7]. To assess the applied encoding as well as to perform faithful image reconstruction, the k-space evolution of all sequences were recorded using a dynamic field camera (Skope MR Technologies, Zurich, Switzerland). To capture eddy current effects of local spatial extent, a 10th order spherical harmonic k-space model was calculated from the field camera recordings at 11 different positions. Concomitant fields [9] were incorporated using the known relationship to the transverse fields. A gradient-echo sequence was performed, from which SENSE and B0 maps were calculated. Image reconstruction was performed using an iterative higher-order SENSE reconstruction [10]. To evaluate the encoded resolution per acquisition time and the achievable image quality, the MR system’s implementation of a 2D-spiral readout was played out (resolution: 1mm, FOV: 220 mm, SENSE: R=3, acquisition duration: 18 ms). For a second experiment, a multi-echo spiral trajectory with 10 echoes (resolution: 2.8 mm, FOV: 220 mm, SENSE: R=3) was targeted, for which the shortest possible gradient time-course was calculated [11]. This resulted in an acquisition duration of 3.3 ms and 0.2 ms to move back to the k-space center. The sequence was repeated with and without fat-suppression.Results
The gradient evolution (Fig.1a) shows that for the targeted 1 mm resolution, acquisition speed is predominantly limited by the achievable slew rate rather than gradient strength. The targeted resolution of 1 mm was reached (Fig.1b). This is confirmed in Fig.1c, which also depicts that decent image resolution can be achieved in only a few milliseconds. A significant contribution to the image encoding came from concomitant fields Bc (Fig.1c). At a gradient strength of 100 mT/m, Bc amounts to an off-resonance up to the kHz range (Fig.1d). Neglecting concomitant field effects caused strong blurring artifacts (Fig.1d), which were largely removed by incorporating all encoding information (Fig.2e). Decent image quality was also achieved for the low resolution image that was reconstructed from the first 3.9 ms of the acquired data. The effectively reached resolution for the multi-echo spiral data (Fig.2) was 3.3 mm. Ringing artifacts on the first echoes may arise from unsuppressed off-resonant fat signal. The magnitude images (Fig2.a) depict the evolution of T2*-decay on a 3.5 ms time grid. The signal decay was quantified (Fig2.d) for different regions of interest (Fig2.c). In addition, rudimentary spectral information could be extracted from the non-fat-suppressed data (Fig.2e), where the water peak and an alias of the fat peak are visible in the spectrum from the scull region.Discussion and Conclusion
This is the first demonstration of in-vivo spiral imaging at a slew rate of 1200 T/m/s. Apart from providing robustness against off-resonance artifacts for high resolution (spiral) single-shot MRI, an extreme temporal resolution for multi-echo MRI at decent resolution (3.3 mm) was achieved without applying any image regularization. The acquisition with a rate of more than 280 images per second also allowed the extraction of rudimentary spectral information from a single MR readout. Further acceleration by increased SENSE undersampling as well as prolonged readout train may be feasible. Performing spectroscopic image reconstruction might as well be tested. Moreover, such rapid imaging could also be performed as single-echo scans with minimum TR on one slice to study long term dynamics at very high frame rates. The newly available temporal resolution in MR may be utilized to gain new insights in brain function and physiology.Acknowledgements
No acknowledgement found.References
1. Ahn et al., IEEE Trans Med Imaging 1986;5:2–7.
2. Glover et al., Neuroimage 2012;62:706–712.
3. Meyer et al., Magn Reson Med 1992;28:202–213.
4. Weiger et al., Magn Reson Med 2002;48:860–866.
5. Heberlein, and Hu, Magn Reson Med 2006;55:619–625.
6. Heidemann et al., Magn Reson Med 2006;56:317–326.
7. Weiger et al., Magn Reson Med 2018 https://doi.org/10.1002/mrm.26954
8. Roesler et al., 2019, submitted to ISMRM
9. Bernstein et al., Magn Reson Med. 1998 Feb;39(2):300-8.
10. Wilm et al., Magn Reson Med. 2017 Jan;77(1):83-91.
11. Lustig et al., IEEE Trans Med Imaging. 2008 Jun; 27(6): 866–873.
Figures
