Synopsis

Keywords: Low-Field MRI, RF Pulse Design & Fields, Saturation RF pulses

Contrast mechanisms such as T_1ρ and MT are interesting for point-of-care (POC) low field systems, to supplement/augment standard T₁- and T₂- weighting. However they typically require high B₀ homogeneity. Here, we use a high homogeneity Halbach-magnet array POC to perform experiments involving either direct on- or off-resonance pulses, and to calculate T_1ρ and MT ratio in tissue-mimicking phantoms. T_1ρ showed different white matter (WM)/gray matter (GM) contrast compared to T₂. The measured magnetization transfer ratio (MTR) showed higher values in the WM than in the GM.

Introduction

Point-of-care (POC) MRI has a number of advantages including portability, accessibility and reduced financial costs¹, but the magnet intrinsically has a much larger inhomogeneity than clinical systems due to the reduced dimensions. Since obtaining T₁ and T₂ contrast in the brain is challenging at ultralow field (<0.1T)², it is interesting to study other mechanisms such as T_1ρ and magnetization transfer (MT). T_1ρ reflects low-frequency motional processes and could be particularly suited for ultralow field strengths without the SNR penalty of direct measurements at this field. Long on-resonance RF pulses applied after excitation lock the spins, slowing the relaxation process which is determined by T_1ρ³. In MT, off-resonance pulses applied before a standard imaging sequence selectively saturate the pool of protons bound to macromolecules, affecting tissue contrast via transfer of magnetization⁴. Sensitivity to macromolecules such as myelin could improve pathologic specificity in white matter over conventional MRI sequences⁵.
Both spin-locking and MT require relatively high B₀ homogeneity for accurate on-resonance and off-resonance (without direct saturation) measurements. Their potential to manipulate tissue contrast for POC imaging has not yet been fully characterized. In this study, we employ a Halbach-magnet array POC system to explore T_1ρ and MT in tissue-mimicking phantoms.

Methods

Images were obtained using a 46 mT Halbach-magnet based MRI system using a Magritek Kea2 spectrometer⁶. For both spin-lock and MT experiments, after magnetization preparation, a 3D turbo spin-echo (TSE) readout was used, with the following parameters: TR/TE: 1250/20 ms, echo train length: 8, 2x2x10 mm³ resolution, and acquisition bandwidth: 20 kHz.
For the spin-lock sequence, a preparatory module (Figure 1a.) was used to minimize both artefacts from B₀ and B₁ inhomogeneities⁷. Data were acquired with spin-lock durations ranging from 10-80 ms in steps of 10 ms, and were fitted to a monoexponential model to compute T_1ρ maps. T_1ρ maps at different spin-lock frequency (f_SL) were compared with T₂ maps acquired with a conventional variable echo time TSE sequence. The phantoms for this study were a 20-mm thick 3D-printed brain-shaped phantom (termed “brain-T₁/T₂”) and a tube phantom, filled with agarose, copper sulphate and deuterium oxide water to mimic the relaxation times of white matter (WM), grey matter (GM), cerebrospinal fluid (CSF), fat and muscle at 50 mT². For the MT study, as low field imaging has fewer issues with specific absorption rate, a hard RF pulse with B₁=17 μT and Δν=3 kHz was applied continuously for 80 ms (Figure 1b). The amplitude and frequency offset of the MT pulse were optimized to keep the degree of direct saturation of the free protons pool⁸ below 10%. To calculate the MT ratio (MTR), two scans were acquired, with and without the MT pulse. In this experiment, the brain-shaped phantom was filled with cross-linked bovine serum albumin (BSA)⁹ and gelatine (termed “brain-MT”) to mimic the ratio T₁/T₂ and MTR expected in brain at low MR field. For comparison, the same phantom was scanned with a 3T scanner (Philips Achieva), using a clinical MT sequence (sinc-gauss pulse with B1=12 μT, Δν=3 kHz, MT pulse length=16 ms). To calculate the expected MTR, which depends on the ratio of the square root of T₁ and T₂¹⁰, the relaxation times of the phantom were measured at both field strengths.

Results

On the 46 mT system, the measured linewidth over the phantom of 34 Hz enabled the use of on- and off-resonance pulses. T_1ρ-weighted images of the brain-T₁/T₂ are fitted to a T_1ρ map (Figure 2) and compared to the T₂ map. For both WM and GM compartments, the T_1ρ values were higher than T₂ (WM: T₂=84± 3 ms, T_1ρ=94± 4 ms; GM: T₂=103± 5 ms, T_1ρ=107± 7 ms). The same behaviour was also observed in a tube phantom experiment (Figure 3), with the T_1ρ values at higher fSL differing more from the T₂ values, as expected.
The MTR maps of the brain-MT show a higher MTR in the white matter than grey matter, both at 46mT and 3T (Figure 4). The square root of the ratio of T₁ to T₂ changes by 40% in GM and 10% in WM (Table 1) from 46 mT to 3T: therefore the MTR at 46mT is expected to be 40% lower in GM and 10% lower in WM than at 3T. The numbers agree well for the GM but the values measured at 46mT in WM are ~25% below expected, probably because of incomplete saturation of the macromolecular pool, which would have a greater effect due to the higher concentration of BSA in the WM.

Discussions

A Halbach-magnet array POC with optimized ring diameters and magnet positions/orientations was sufficiently homogenous to perform T_1ρ and MT experiments involving either direct on- or off- resonance pulses. T_1ρ images showed a different contrast than T₂, as expected. We note that, in the presence of B₁ inhomogeneity, T_2ρ effects can contaminate the T_1ρ contrast⁷, and so the images may contain some contribution from which will be explored further. In the MT study, the performance of our current RF amplifier limited the duration of the MT pulse to 80 ms. With different hardware, longer RF durations may be used to reach full saturation of the macromolecular pool and higher MTR.

Acknowledgements

No acknowledgement found.

References

1. Sarracanie M, Salameh N. Low-Field MRI: How low can we go? A fresh view on an old debate. Front. Phys. 2020; 8. doi:10.3389/fphy.2020.00172

2. O’Reilly T, Webb AG. In vivo T1 and T2 relaxation time maps of brain tissue, skeletal muscle, and lipid measured in healthy volunteers at 50 mT. Magn Reson Med. 2022;87(2). doi:10.1002/mrm.29009

3. Wáng YXJ, Zhang Q, Li X, Chen W, Ahuja A, Yuan J. T1ρ magnetic resonance: basic physics principles and applications in knee and intervertebral disc imaging. Quant Imaging Med Surg. 2015;5(6). doi:10.3978/j.issn.2223-4292.2015.12.06

4. de Boer RW. Magnetization Transfer Contrast Part 1: MR Physics. Medica Mundi (Philips Healthcare). 1995;(2).

5. Mancini M, Karakuzu A, Cohen-Adad J, Cercignani M, Nichols TE, Stikov N. An interactive meta-analysis of MRI biomarkers of Myelin. Elife. 2020;9. doi:10.7554/eLife.61523

6. O’Reilly T, Webb A. The design of a low-weight homogenous Halbach helmet for imaging the adult brain. In: Proceedings of 31st Annual Meeting of ISMRM & ISMRT. ; 2021.

7. Witschey WRT, Borthakur A, Elliott MA, et al. Compensation for spin-lock artifacts using an off-resonance rotary echo in T1ρoff-weighted imaging. Magn Reson Med. 2007;57(1). doi:10.1002/mrm.21134

8. Hajnal J v., Baudouin CJ, Oatridge A, Young IR, Bydder GM. Design and implementation of magnetization transfer pulse sequences for clinical use. J Comput Assist Tomogr. 1992;16(1). doi:10.1097/00004728-199201000-00003

9. Koenig SH, Brown RD, Ugolini R. Magnetization transfer in cross‐linked bovine serum albumin solutions at 200 MHz: A model for tissue. Magn Reson Med. 1993;29(3). doi:10.1002/mrm.1910290306

10. McGowan JC, Leigh JS. Selective saturation in magnetization transfer experiments. Magn Reson Med. 1994;32(4). doi:10.1002/mrm.1910320415