Introduction

Semisolid parameter quantification is of importance in MRI since these properties underpin observed relaxation phenomena in tissue¹. MR fingerprinting (MRF) sequences are often designed to be maximally sensitive to multiple tissue properties, enabling their simultaneous measurement in an efficient acquisition². Use of ultra-high field (UHF) strength potentially allows high resolution quantification, however specific absorption rate (SAR) limits are significantly more restrictive compared to clinical field strengths, which imposes constraints on contrasts that rely on RF power deposition, such as inhomogeneous magnetization transfer (ihMT).

We propose a combined parameter estimation and motion correction multiband framework at UHF that uses a steady-state sequence with RF pulses comprising a constant on-resonance flip angle but temporally modulated off-resonance bands to create time-variable magnetization-transfer mediated signal modulation³. The signals are sampled using Cartesian encoding and reconstructed with low-rank inversion (LRI)⁴ and DISORDER motion estimation⁵. Two variants are explored: the first aims to generate MT and ihMT contrast and is termed MT-MRF-dipolar (similar to Reference 3), whilst the second aims to measure semisolid T₂ and is termed MT-MRF-$$$T_2^s$$$.

Methods

This work used a steady-state spoiled gradient echo sequence with either single-band (1B) pulses (on-resonance only) or multiband (2B or 3B) pulses which add one or two off-resonance bands that aim to saturate semisolid magnetization; all pulses are spatially non-selective. MT-MRF-dipolar (Figure 1a) uses short batches of either 3B or 2B pulses interspersed with 1B pulses to allow recovery. 3B pulses have symmetric off-resonance bands, 2B have only either a positive or negative offset - the difference between 3B and 2B gives rise to ihMT contrast⁶. MT-MRF-$$$T_2^s$$$ uses a similar multiband RF pulse cycling framework but features 2-band pulses for saturation with variable offset frequencies (-10 to 10kHz) to sample the semisolid lineshape at different positions, thus encoding semisolid T₂ information (Figure 1b).

Phantom and in vivo experiments were performed on a 7T MRI scanner (MAGNETOM Terra, Siemens Healthcare, Erlangen, Germany) with an 8Tx/32Rx channel head coil (Nova Medical, Wilmington, MA, USA). Phantom tubes contained MnCl₂-doped water (0.05mM), cross-linked bovine serum albumin (BSA) or prolipid-161 (PL161). The MnCl₂ phantom should exhibit no semisolid compartment and BSA should have no dipolar order effects. One healthy volunteer (male aged 32) was scanned according to local ethical approval.

Reconstruction of all data was performed using a combined LRI-DISORDER framework⁷. The low-rank basis was discovered by using SVD on a coarse dictionary including all model parameters - the number of singular components was chosen to capture ~99% of the energy. Parameter matching used more finely sampled dictionaries with most parameters held fixed and only the following parameters varied: semisolid fraction ($$$M_0^s$$$) and dipolar T₁ ($$$T_{1D}^s$$$) for MT-MRF-dipolar, and $$$M_0^s$$$ and semisolid T₂ ($$$T_2^s$$$) for MT-MRF-$$$T_2^s$$$. Phantom analysis for the latter used $$$M_0^s$$$ estimates from MT-MRF-dipolar to inform the fit, with frequency-offset of the semisolid line⁸ also estimated. A B₁ map was acquired in vivo using AFI (FA = 60˚) and incorporated during corresponding dictionary matching.

Results

MT-MRF-dipolar phantom results (Figure 2) reveal dipolar contrast in PL161 only and semisolid contrast in both BSA and PL161, as expected. $$$T_{1D}^s$$$ = 24.6±2.0ms in PL161, which concurs with previous measurements at 1.5T³, whilst $$$M_0^s$$$ = 0.15±0.01 in PL161 and $$$M_0^s$$$ = 0.051±0.002 in BSA. $$$T_2^s$$$ maps from MT-MRF-$$$T_2^s$$$ (Figure 2b) show mean values of 8.18±0.39μs in PL161 and 6.03±0.19μs in BSA, with non-negligible semisolid absorption lineshape shift (e.g. -318±10Hz in PL161), though this is likely to depend on B₀, which has not yet been accounted for in the dictionary. Both scenarios yield good signal fits despite many remaining model parameters being fixed - normalized residuals are of the order of a few percent.

MT-MRF-$$$T_2^s$$$ data can also be processed to yield semiquantitative MT ratio (MTR) maps, calculated as the difference in signal between the ends of the 2-band and 1-band periods. DISORDER reconstruction of an in vivo dataset yields motion traces (Figure 3) and MTR maps (Figure 4). Constrained dictionary fitting provides parameter maps (Figure 5). $$$M_0^s$$$ shows strong gray matter-white matter contrast but $$$T_2^s$$$ contrast is mostly flat.

Discussion

MT-MRF-dipolar and MT-MRF-$$$T_2^s$$$ provide reasonable estimates of some important semisolid tissue properties in phantom experiments. Whilst estimation bias certainly exists given the high dimensionality and thus degeneracy of the assumed tissue model, phantom results agreed with previous findings³ at lower field. In vivo estimation of $$$M_0^s$$$ appears robust though $$$T_2^s$$$ shows little contrast. Quantification of $$$T_2^s$$$ remains an open research question; different studies have reported maps with varying degrees of gray matter-white matter contrast^9,10. Future work will involve disentangling the effects of remaining tissue parameters in these MRF methods. T₁ has been fixed to literature values at 7T¹ to perform the above fitting but can be encoded more effectively using inversion pulses, for instance.

Conclusion

This work has demonstrated the feasibility of two novel MRF sequences at 7T both using temporally modulated off-resonance saturation with constant flip angle acquisition. One variant encodes dipolar information and is sensitive to lipid bilayer-like structures, whilst the other has potential for $$$T_2^s$$$ quantification. Future work aims to separate out the effects of other tissue parameters to remove residual estimation biases.

Acknowledgements

This work was supported by core funding from the Wellcome/EPSRC Centre for Medical Engineering [WT203148/Z/16/Z], Wellcome Trust Collaboration in Science grant [WT201526/Z/16/Z] and by the National Institute for Health Research (NIHR) Biomedical Research Centre based at Guy’s and St Thomas’ NHS Foundation Trust and King’s College London and/or the NIHR Clinical Research Facility. The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health and Social Care.

References

1. Wang Y, van Gelderen P, de Zwart JA, Duyn JH. B0-field dependence of MRI T1 relaxation in human brain. Neuroimage 2020;213:1–11 doi: 10.1016/j.neuroimage.2020.116700.
2. Ma D, Gulani V, Seiberlich N, et al. Magnetic resonance fingerprinting. Nature 2013;495:187–192 doi: 10.1038/nature11971.
3. West DJ, Cruz G, Teixeira RPAG, et al. An MR fingerprinting approach for quantitative inhomogeneous magnetization transfer imaging. Magn. Reson. Med. 2021:1–16 doi: 10.1002/mrm.28984.
4. Assländer J, Cloos MA, Knoll F, Sodickson DK, Hennig J, Lattanzi R. Low rank alternating direction method of multipliers reconstruction for MR fingerprinting. Magn. Reson. Med. 2018;79:83–96 doi: 10.1002/mrm.26639.
5. Cordero-Grande L, Ferrazzi G, Teixeira RPAG, O’Muircheartaigh J, Price AN, Hajnal JV. Motion corrected MRI with DISORDER: Distributed and Incoherent Sample Orders for Reconstruction Deblurring using Encoding Redundancy. Magn. Reson. Med. 2020;84:713–726.
6. Malik SJ, Teixeira RPAG, West DJ, Wood TC, Hajnal JV. Steady‐state imaging with inhomogeneous magnetization transfer contrast using multiband radiofrequency pulses. Magn. Reson. Med. 2019;83:935–949 doi: 10.1002/mrm.27984.
7. West DJ, Cordero-Grande L, Teixeira RPAG, Ferrazzi G, Hajnal JV, Malik SJ. Motion corrected magnetization transfer-mediated fingerprinting (MT-MRF) using DISORDER. Proc. ISMRM 2021
8. Jiang X, van Gelderen P, Duyn JH. Spectral characteristics of semisolid protons in human brain white matter at 7 T. Magn. Reson. Med. 2017;78:1950–1958 doi: 10.1002/mrm.26594.
9. Yarnykh VL. Pulsed Z-spectroscopic imaging of cross-relaxation parameters in tissues for human MRI: theory and clinical applications. Magn. Reson. Med. 2002;47:929–939 doi: 10.1002/mrm.10120.
10. Yarnykh VL. Fast macromolecular proton fraction mapping from a single off-resonance magnetization transfer measurement. Magn. Reson. Med. 2012;68:166–178 doi: 10.1002/mrm.23224.