Introduction

Myelin water quantification provides specific information about the integrity of the myelin sheath in the brain^1,2. QSM is a contrast mechanism based on intrinsic differences in the magnetic susceptibility of different tissues^3,4. QSM has also provided promising applications in the detection of demyelination⁵. The mGRE sequence can be used to quantify the MWF in the brain⁶ and can also be applied to obtain the magnetic susceptibility information of different tissues⁷. Whereas, the acquisitions of the MWF and QSM maps are usually carried out separately. Moreover, the collection of whole brain mGRE data requires several minutes^8,9. In this study, the feasibility of simultaneous MWF/QSM quantification is investigated using the mGRE acquisition. The potential of obtaining whole brain QSM/MWF quantifications in 1 minute is further tested retrospectively and prospectively.

Methods

Nine healthy subjects (7 males, 2 females, age 25.7 ± 5.3) were scanned using a multi-slice mGRE sequence on a 3T MRI scanner (uMR790, United Imaging Healthcare, Ltd., Shanghai, China). Written consent was obtained before each scan. The following scanning parameters were applied: FOV = 240 mm × 240 mm, matrix = 176 × 176, flip angle = 90°, time repetition = 2 s, first echo time = 1.95 ms, echo spacing = 1.16 ms, echo train length = 30, slice thickness = 3 mm, 25 slices were scanned with a 1.5 mm slice gap. The whole brain scan time was 5.9 min. A variable-density random phase-encoding mask with R=6 was used for retrospective undersampling. The undersampled datasets were reconstructed by the low-rank and sparse regularized tensor dictionary learning algorithm (TDLLS) reconstruction algorithm developed in our lab. The non-negative jointly sparse (NNJS) algorithm was used to estimate the MWF from the T2*WIs¹⁰. A joint 2D and 3D phase processing method was used for quantitative susceptibility mapping calculation⁷. The last ten echoes were used for the QSM estimation. In the prospective undersampling, blip gradients were added to the mGRE sequence at the ramps of a train of readout gradients to cover large Ky-t space in one shot. A pseudo-random sampling pattern with variable sampling density in the Ky direction and constrained blip gradients in the temporal direction was designed. For the prospective undersampling with R = 6, a total of 30 shots were acquired to collect all T2* signals. The 8 central k-space lines were collected at all echoes. The rest 22 Ky lines were designed to pseudo-randomly distributed in periphery k-space, which could be clustered into several groups. Each Ky line in the first group was set to be randomly varied within 6 phase steps along the time direction, and each Ky line in the second group was set to be randomly varied within 9 phase steps at different echoes. The first group included 10 Ky lines, which were closer to the central k-space than the 12 Ky lines in the second group. The phase encodings within each shot covered the Ky-t space complementarily. To reduce the eddy current effects, the randomness of the Ky variation along the time direction was restrained within 2 Ky lines by constraining the max blip gradients. One subject was scanned and the whole brain scan time was 1 min.

Results

Figure 1 and Figure 2 show the MWF and QSM maps obtained from the fully-sampled mGRE data and the retrospectively undersampled data with R=6 from 9 subjects. High-quality MWF and QSM maps can be obtained from the fully-sampled mGRE data. The brain tissue structures can be clearly depicted in both the MWF maps and the QSM maps. The MWF/QSM maps obtained from the retrospectively undesampled data show high similarity with the MWF/QSM results of the fully-sampled data. The profile plots have demonstrated the similar variations in the QSM maps of the fully-sampled data and the TDLLS reconstructed data, as shown in Figure 3. Figure 4 shows the MWF/QSM maps of 8 adjacent slices from one prospectively undersampled datasets with R=6. The MWF/QSM results of the prospective undesampled data also show high similarity with that of the fully-sampled data.

Discussion and Conclusions

The high-quality QSM maps and MWF maps can be obtained simultaneously using the mGRE acquisition. The 1 minute whole brain QSM/MWF quantifications can be achieved by the joint application of the prospective undersampling scheme with pseudo-random and complementary Ky-t space encodings and the TDLLS reconstruction algorithm.

Acknowledgements

No acknowledgement found.