Introduction

Quantitative mapping is a promising tool for understanding the biophysical mechanisms related to diseases. Previously the clinical utility of MAGiC a method based on a single multi-delay multi-echo 2D fast spin echo scan and synthetic MRI data processing have been demonstrated for several clinical applications1,2. 3D QALAS 3,4(3D-quantification using an interleaved Look-Locker acquisition sequence with T2 preparation pulse) is a rapid quantitative mapping technique to provide 1mm3 iso-voxel high resolution volumetric images and 3D T1 and T2 , PD parameter maps from a 6-7 min single scan of the whole brain5 lending to multiplanar reformats and 3D parametric maps. However, to be clinically applicable, reliable quantitative measurement is crucial for reducing the incidence of motion artifact caused by uncooperative populations such as those with dementia or even by cooperative subjects with physiological motion such as respiratory-related movements and swallowing movement6. In this work, a real-time prospective motion correction (PROMO)7was integrated into a 3D acquisition based off QALAS and feasibility study was performed aiming for reliable parameter mapping.

Method

Figure 1 shows schematic diagrams of our proposed method which is based off on 3D QALAS with 2D single-shot spiral navigator pulse sequences inserted into the dead times of the acquisition such that there is no scantime penalty. The three orthogonal navigators (SpNav) were played out just after 3D segmented data acquisition blocks 3, 4, and 5. The reconstructed images from three SpNavs were passed as input to the extended Kalman filter (EKF) algorithm for motion tracking and correction. Six dimensional motion parameters for translations and rotations in x-y-z space were calculated. The navigator and imaging sequences were updated in real-time for frequency, phase and slice encoding of the image volume for prospective rigid motion correction. Rescans were performed when large k-space inconsistency happens because transient motion can’t be captured between SpNavs.
Our proposed method obtains five image contrasts using 3D gradient echo sequence in Fig.1. The first acquires MR data after T2 preparation pulse. The second and later readouts are performed during the following T1 relaxation after Inversion pulse. Five data collections are executed at equal time intervals of 1 second. The 3D data is acquired with segmented Cartesian k-space sampling. Quantitative T1, T2 and proton density maps were reconstructed from the obtained images using a research version of SyMRI (SyntheticMR, Linkoping, Sweden).
SpNav uses spiral acquisition in the following parameters: TR/TE=15.3/0.9 ms, FOV 32cm, 30 mm thickness, rBW 125kHz, reconstructed MAT 128x128. Twenty-four axial spiral images are acquired at the beginning of the scan and registered to a reference to determine the initial position of the navigators and the location of an anatomical mask applied to the navigators to localize motion detection to the brain.
To minimize the impact of signal saturation by SpNav on the image volume, flip angle of SpNav was optimized with cylindrical phantoms including white matter (T1: 588ms, T2: 79ms), gray matter(T1: 808ms, T2: 99ms) and CSF(T1:4061, T2:1639ms). Navigator slice thickness 10mm was used to show the saturation effect clearly. Flip angles of 3,6,10 and 30 degree were used.
To assess the effectiveness of motion correction, healthy volunteer scan was performed with motion correction on and off, where the subject was asked to repeat staged motions during scan. A 3.0 T System (Discovery MR 750w, GE Healthcare, Waukesha, WI, U.S.A.) with 12 channel coils (GEM HNU, GE Healthcare) was used. Other Imaging parameters were: spatial resolution 1x1x1.4mm3, TR/TE 6.5/2.2 ms, bandwidth 62.5kHz, flip angle 4 degree. Parallel imaging factor 2, scan time 4:55.

Results

Fig.2 showed the result of signal saturation effect by SpNav. The use of flip angle 3 deg showed no banding artifact on source image of Acquisition 4 in the QALAS and similar signal profile to flip angle 0 deg which means that no SpNav is used. Other #Acquisitions of the QALAS had the same result which is not shown here. Fig.3 displayed representative source images of the QALAS with motion correction on and off. Motion correction was effectively applied to reduce motion artifact. Dominant motion was translation x as the left-right direction and rotation z as the inferior-superior direction. Maximum translation and rotation were 1.94 mm and 3.3 degree, respectively. Fig.4 and Fig.5 were synthetic multi-contrast images and quantitative mappings generated with syMRI. Motion artifact from source image was propagated to the final synthetic images.

Discussions and conclusion

Motion still poses one the biggest challenges to consistent MR image quality since some amount of head motion is typical in clinical subjects, and frequent in uncooperative cohorts. We successfully demonstrated the effectiveness of prospective motion correction in multi-parametric imaging such that signal intensity in source image was not impacted by SpNav acquisition. Scan time was maintained by inserting SpNav into the dead time of the QALAS except for rescan functionality, which is a reasonable trade-off to ensure more reliable measurement with small time penalty, instead of having to repeat the entire scan. The proposed prospectively motion corrected 3D multicontrast method yielding volumetric images of multiple contrasts such as T1, T2 , T2 FLAIR, PSIR and 3D parametric maps could significantly improve scan workflow, productivity and information/scantime and would be feasible in clinical settings.

Acknowledgements

No acknowledgement found.