3089

Real-time motion and retrospective receiver sensitivity correction for gluCEST using volumetric navigators (vNavs) at 7T
Esau Poblador Rodriguez1, Philipp Moser1, Sami Auno2, Korvinian Eckstein1, Stephan Gruber1, Benjamin Schmitt3, Andre van de Kouwe4, Siegfried Trattnig  1, and Wolfgang Bogner1
1High Field MR Centre, Biomedical Imaging and Image-guided Therapy, Medical University of Vienna, Vienna, Austria, 2Department of Physics, University of Helsinky, Helsinki, Finland, 3Siemens Healthineers, Sydney, Australia, 4Harvard Medical School, Boston, MA, United States

Synopsis

CEST is susceptible to subject motion not only due to misregistration of MR images acquired over time and subsequent alterations of B0 field, but also due to the related receiver sensitivity (B1-) inhomogeneities. We propose to correct for ∆B1- from the images generated by an interleaved EPI navigator used for prospective motion correction. We show the benefit of the proposed method, in combination with B1+ and dynamic B0 corrections, to restore a glutamate weighted CEST map in an active healthy volunteer.

Introduction

Chemical Exchange Saturation Transfer (CEST) MRI is a contrast mechanism that allows the detection of small amounts of metabolites with enhanced sensitivity including glutamate, the main excitatory neurotransmitter in the human brain (1). However small movements, field artifacts and can lead to Z spectra alterations that complicate the CEST quantification (2). We propose to incorporate a retrospective receiver sensitivity (B1-) correction step to avoid distortions arising from subject movement closer or further away from the coil elements.

Methods

A 2D Turbo-Flash CEST acquisition optimized to be corrected intrinsically for B0 inhomogeneities was interleaved with a volumetric EPI navigator (vNav) (3). The navigator images were used for both, real-time motion correction and to map B1- fluctuations over time (Figure 1). The B1- correction was done retrospectively by normalizing each navigator image acquired prior each CEST offset acquisition with the first navigator of the sequence. This relative maps was subsequently smoothed, masked, extrapolated and resized to match the slice location of the CEST images (4). B0 inhomogeneities were corrected based on the CEST images by the dynamic method CEST‐GRE‐2TE (3). B1+ correction was based on Z-B1-correction (5) where a pre acquired a flip angle map is used to correct each CEST image.
The imaging workflow (Figure 1) consisted on a Flip angle 2D Turbo-FLASH-based flip angle pre-scan necessary for B1+ correction, followed by four CEST measurements. In the first two a healthy volunteer was asked to remain still with and without real-time motion correction respectively, and the last two he was instructed to rotate the head 6 times again with/without enabling motion correction. The rotations were performed during the 8s long pauses placed between the acquisition of each CEST offset to allow for T1 relaxation. MTR asymmetry maps were calculated in the range between 2.75-3.25 ppm and normalized by M0 acquired at -100 ppm.
The study was performed on a whole-body 7T MR system (Magnetom, Siemens) with a 1H 32-channel head coil (Nova Medical) after informed written consent was obtained. Image processing, coil combination, B0/B1-/B1+ maps calculation were conducted in MATLAB (MathWorks). B0/B1-/B1+ corrections and asymmetric magnetic transfer ratio (MTRasym) analysis were performed voxel wise. The 3D vNavs had a resolution of 7x7x7 mm3 and the CEST sequences 1.75x1.75x5.0mm. GRE readout TR/TE1/TE2=9.5/1.74/5.16 ms and measurement TR=11s. The CEST saturation consisted on 3 Gaussian pulses with tp=100ms, td=12ms and amplitude 5.6uT applied at 34 frequencies alternating from ±4 down to 0 ppm.

Results:

The navigator based motion correction was able to correct accumulated head rotations up to 7º with minimal residual pseduo CEST effects due to remaining misregistrations (Fig. 2). The efficaccy of the retrospective dynamic B1- correction is demonstrated for the vNav images with values ~100% with respect to the reference prior to movement (Fig 3) and applied to the CEST images (Fig 4). In the most severly affected frequencies, for CEST weighted images of metabolites resonating ~2 ppm (creatine), the current correction of a pseudo CEST effect of in the selected ROI of 1% would have reached to 2%.

Discussion and Conclusion:

The real-time motion correction based on vNavs is able to substantially reduce rotations up to 7º, however remaining misregistration may be reduced by acquiring 3D CEST data. Receiver sensitivity correction based on vNavs data is able to correct hyper/hypo intensities on the MTR maps within the brain, there is still space for improvement in areas close tho the borther of the brain. This is explained because the navigator images show some misregistration among the CEST acquisition.
Retrospective B1- correction is a viable method, sinergic with the use of vNav for prospective motion correction that significantly reduce motion related artifacts in CEST MRI. In combination with B0 and B1+ corrections provide the opportunity to aply this technique with motion prone patients.

Acknowledgements

No acknowledgement found.

References

1. Cai K, Singh A, Roalf DR, Nanga RPR, Haris M, Hariharan H, Gur R, Reddy R. 2013. Mapping glutamate in subcortical brain structures using high-resolution glucest mri. NMR in biomedicine. 26(10):1278-1284.

2. Zaiss M, Herz K, Deshmane A, Kim M, Golay X, Lindig T, Bender B, Ernemann U, Scheffler K. 2019. Possible artifacts in dynamic cest mri due to motion and field alterations. J Magn Reson. 298:16-22.

3. Poblador Rodriguez E, Moser P, Dymerska B, Robinson S, Schmitt B, van der Kouwe A, Gruber S, Trattnig S, Bogner W. 2019. A comparison of static and dynamic b0 mapping methods for correction of cest mri in the presence of temporal b0 field variations. Magn Reson Med. 82(2):633-646.

4. Auno S, Poblador Rodriguez E, Moser P, v.d.Kouwe A, Gruber S,Trattnig S, Bogner W. Real-time motion and dynamic transmit/receive B1 correction of CEST in the human brain at 7T [Abstract]. In ISMRM 26th Annual Meeting & Exhibition: 2018 abstract nr 0661

5. Windschuh J, Zaiss M, Meissner JE, Paech D, Radbruch A, Ladd ME, Bachert P. 2015. Correction of b1-inhomogeneities for relaxation-compensated cest imaging at 7 t. NMR in biomedicine. 28(5):529-537.

Figures

Fig 1: Protocol with four CEST scans. In the first two the volunteer is asked to remain still and motion correction is turned ON and OFF respectively. The last two are identical but the volunteer is instructed to rotate the head during the CEST acquisition. AutoAlign scans are interleaved to maintain the slice position. The workflow start with the B1- correction of the magnitude CEST images, subsequently the Z spectral points are shifted according to the ∆B0/∆B1+ maps before the MT asymmetry map is derived.

Fig 2: Effect of the real-time motion correction on the MTR asymmetry map. When no motion was instructed the tracked rotation remained between ±0.2º (D) and they grey matter was highlighted (A). For the case of no motion correction the MTR map presented severe artifacts (B) for a rotation of 3º (E). When motion correction is applied, MTR is improved substantially even for a rotation of 7º (F), but B1- inhomogeneities are apparent (C).

Fig 3: Effect of B1- correction on the magnitude vNav images (not affected by saturation) and the magnitude CEST images. On the left side the effect of the B1- fluctuation is clear between the 10th (after motion) and 1st measurement (reference). On the right side the relative B1- map corresponding to the 10th measurement is applied to both navigator and CEST images. The navigator image show a ratio around 100% within the brain borders confirming the proper correction. The CEST image display values <100% due to partial saturation.

Fig 4: Effect of B1- correction on the MTR asymmetry. MTR maps without (A) and with (B) B1- correction. The effect of the rotation of the subject was an apparent hyperintensity on the right side of the MTR map and hypointensities in the left side, induced by the motion closer/further from the coil elements. The final corrected glutamate weighted MTR corresponds to the right map (B) and the correction effect in the ROI, delineated covering a region of GM, is displayed for a wither range of frequencies (C). This distortion would have been stronger at ~2 ppm.

Proc. Intl. Soc. Mag. Reson. Med. 28 (2020)
3089