1090

Orientation Dependence of Magnetization Transfer and “Inhomogeneous Magnetization Transfer” in Spinal Cord
Niklas Wallstein1, André Pampel1, Carsten Jäger2,3, Roland Müller1, Jens Stieler3, Sven Martin3, Markus Morawski2,3, and Harald E. Möller1,4
1NMR Methods & Development Group, Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany, 2Department of Neurophysics, Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany, 3Center of Neuropathology and Brain Research, Medical Faculty, University of Leipzig, Paul Flechsig Institute, Leipzig, Germany, 4Felix Bloch Institute for Solid State Physics, Leipzig University, Leipzig, Germany

Synopsis

Keywords: Magnetization Transfer, Magnetization transfer, Inhomogeneous Magnetization Transfer, Orientation Dependence, Spinal Cord

Motivation: Magnetization transfer (MT) and inhomogeneous MT (ihMT) are assumed to report on myelin content. Recently, anisotropy of MT and ihMT have been demonstrated in model systems and in white matter.

Goal(s): Our goal was to quantify orientation effects in (ih)MT, as they closely relate to the microstructure and serve to confirm assumptions about the relaxation mechanism.

Approach: Comprehensive (ih)MT investigations were performed in fixed spinal cord with variation of the fiber-to-field angle (θFB).

Results: Unambiguous orientation dependence was observed for (ih)MT. The variation depends strongly on the offset frequency, which was quantitatively predicted in simulations of the BSBM with a realistic fiber model.

Impact: This study investigates the orientation dependency of magnetization transfer and related model parameters. Some subtle orientation effects are observed for the first time and are quantitatively explained by using a reasonable model for the RF saturation lineshape.

Introduction

Magnetization transfer (MT) and the so-called inhomogeneous MT (ihMT) are assumed to report on myelin content. Quantitative analysis is usually based on the binary spin-bath model (BSBM) with a water and macromolecular pool,1,2 and a dipolar reservoir if necessary.3,4 Recently, anisotropy of MT and ihMT have been demonstrated in model systems5 and in white matter (WM) in vivo6,7 and post mortem.8,9 Understanding orientation effects in (ih)MT quantitatively is of interest as it provides a direct link to microstructure (i.e., WM fibers) and validation of assumptions about the relaxation mechanism. Due to the large number of (partially correlated) model parameters, this requires a detailed experimental characterization, which is easier achieved in small tissue samples. Therefore, comprehensive (ih)MT investigations were performed in fixed spinal cord with systematic variation of the fiber-to-field angle (θFB) under well-controlled conditions.

Methods

Sample: A cylindrical piece of hamster spinal cord (approx. 22 mm long, 4 mm diameter) was dissected immediately after death, fixed in paraformaldehyde (4%; 3 weeks), washed in phosphate-buffered saline (pH 7.4) and placed in a 5mm NMR tube (filled with Fomblin).

MR experiments: 1D MT data (20µs 90° rectangular readout pulse, TE=7.5 ms, readout along sample axis, 306µm nominal resolution) were acquired at 3T (MAGNETOM Sykrafit; Siemens Healthineers) with a tiltable Helmholtz coil10 inside a thermically insulated box at 37.9±1.1 °C (9 hours; fiber-optic temperature monitoring) and reorientation of the sample (randomized variation of angles 0°–90° relative to B0, adjusted remotely). MT preparation schemes (2ms Gaussian pulses, 250µs separation) comprised:
(i) Dynamic acquisitions with incrementation of the number of MT pulses (NRF=0–280, amplitude γB1,rms+= 2π·750 rad/s, offset Δf=10 kHz).
(ii) ihMT acquisitions including no MT preparation (MToff), single-sided irradiation (MT+, MT–), dual-sided irradiation with alternating offset (MTdual) or cosine-modulated pulse (MTcos) with NRF=250, γB1,rms+=2π·500/1000 rad/s, Δf=8/12/15 kHz).
(iii) Z-spectra acquisitions with 28 offsets (Δf=0.5–60 kHz, NRF=200, γB1,rms+=2π·500 rad/s).
T1 measurements (40µs rectangular inversion pulse; 23 inversion times, TI=770μs–10s; TR=13 s) and diffusion-weighted imaging (DWI; TE=80 ms, b=1500 s/mm2, 60 directions) were also performed.

Analysis: A 10-voxel ROI with high anisotropy (FA≈0.65) was chosen for the analysis. Timing and pulse shapes were imported from the scanner and the data were fitted using in-house developed routines to the BSBM11 with a dipolar reservoir12,13 (Matlab). Initially, parameters of the ‘classical BSBM’ were obtained from MTcos data to ensure perfect decoupling from the dipolar reservoir. Subsequently, all experiments were simultaneously analyzed (Figure 1), whereby T2b (macromolecular pool; super-Lorentzian lineshape) was readjusted and T1,D and ωD describing the dipolar reservoir were determined (Figure 2). Furthermore, the lineshape describing the RF absorption was calculated using two different approaches: (i) using the standard assumption of a super-Lorentzian and (ii) derived from a more realistic fiber model where the distribution of WM fiber orientation was described by a scaled Bingham function (parameters κ1 and κ2, see Figure 3)1.

Results and Discussion

Even with the current characterization by about 200 well-controlled acquisitions with high SNR and simultaneous fitting, correlations of multiple parameters remain, making a complete quantitative characterization of all model parameters difficult.
Unambiguous orientation dependence was consistently observed for single- and dual-sided MT pulse application (Figures 3 and 4). Notably, the variation depended strongly on the offset frequency, with the strongest effect (8–10%) at Δf>10 kHz but only moderate effect (≈3%) for Δf<7 kHz. This underlines the importance of Δf in experiments aimed at orientation-dependent (ih)MT. A maximum was found with θFB≈30–40° (Figure 3) and, consistently, for T2b (Figure 2), which is in line with previous results in-vivo.6
An unexpected finding was the remarkable change in the orientation dependence of the ihMT ratio (ihMTR). Recent statements12,14 about a higher ihMTR in fibers running parallel to B0 was valid only for Δf<15 kHz), whereas the opposite was observed at higher offsets with a maximum ihMTR at θFB≈90° (Figure 4). Of note, this behavior has been quantitatively predicted in simulations of the BSBM, particularly with the realistic fiber model (Figure 5).

Conclusion

Remarkable variations of MT and ihMT parameters with θFB are found in fixed spinal cord as a WM model with high structural order. These orientation effects depend on the offset frequency, being more pronounced at higher frequencies than those utilized in in-vivo experiments. Even more complex are the results for ihMT, with opposite trends at lower (e.g., 10 kHz) or higher offset frequencies (e.g., 20 kHz). These effects may explain some inconsistencies in the literature. Additionally, an appropriate description of fiber dispersion beyond a single main direction is required for a reliable interpretation of the results.

Acknowledgements

No acknowledgements found.

References

  1. Henkelman RM, Huang X, Xiang Q-S, Stanisz GJ, Swanson SD, Bronskill MJ. Quantitative interprepation of magnetization transfer. Magn. Reson. Med. 1993; 29: 759–766.
  2. Morrison C, Henkelman RM. A model for magnetization transfer in tissues. Magn. Reson. Med. 1995; 33: 475–482.
  3. Yeung HN, Adler RS, Swanson SD. Transient decay of longitudinal magnetization in heterogeneous spin systems under selective saturation. IV. Reformulation of the spin-bath-model equations by the Redfield-Provotorov theory. J. Magn. Reson. A 1994; 106: 37–45.
  4. Morrison C, Stanisz G, Henkelman RM. Modeling magnetization transfer for biological-like systems using a semi-solid pool with a super-Lorentzian lineshape and a dipolar reservoir. J Magn. Reson. B 1995; 108: 103–113.
  5. Morris SR, Frederick R, MacKay AL, Laule C, Michal CA. Orientation dependence of inhomogeneous magnetization transfer and dipolar order relaxation rate in phospholipid bilayers. J. Magn. Reson. 2022; 338: 107205.
  6. Pampel A, Müller DK, Anwander A, Marschner H, Möller HE. Orientation dependence of magnetization transfer parameters in human white matter. NeuroImage 2015; 114: 136–146.
  7. Morris SR, Vavasour IM, Smolina A, MacMillan EL, Gilbert G, Lam M, Kozlowski P, Michal CA, Manning A, MacKay AL, Laule C. Myelin biomarkers in the healthy adult brain: Correlation, reproducibility, and the effect of fiber orientation. Magn. Reson. Med. 2023; 89: 1809–1824.
  8. Marschner H, Metere R, Geyer S, Pampel A, Möller HE. Further evidence of an orientation dependence of magnetization transfer parameters from investigations in post-mortem marmoset brain. Proceedings of the 23rd Annual Meeting of ISMRM, Toronto, ON, Canada, 2015; p. 0995.
  9. Olivier MG, Prevost VH, Mchinda S, Varma G, Alsop DC, Duhamel G. Anisotropy of inhomogeneous magnetization transfer (ihMT) in white matter. Proceedings of the 25th Annual Meeting of ISMRM, Honolulu, HI, USA, 2017; p. 0472.
  10. Wallstein N, Pampel A, Müller R, Möller HE. Radiation damping at clinical field strength: Characterization and compensation in quantitative measurements. Magn. Reson. Med. (in press) DOI: 10.1002/MRM.29934.
  11. Müller DK, Pampel A, Möller HE. Matrix-algebra-based calculations of the time evolution of the binary spin-bath model for magnetization transfer. J. Magn. Reson 2013; 230: 88–97.
  12. Varma G, Girard OM, Prevost VH, Grant AK, Duhamel G, Alsop DC. Interpretation of magnetization transfer from inhomogeneously broadened lines (ihMT) in tissues as a dipolar order effect within motion restricted molecules. J. Magn. Reson. 2015; 260: 67–76.
  13. Provotorov BN. Magnetic resonance saturation in crystals. Sov. Phys. JETP 1962; 14: 1126–1131.
  14. Ercan E, Varma G, Mädler B, Dimitrov IE, Pinho MC, Xi Y, Wagner BC, Davenport EM, Maldjian JA, Alsop DC, Lenkinski RE, Vinogradov E. Microstructural correlates of 3D steady-state inhomogeneous magnetization transfer (ihMT) in the human brain white matter assessed by myelin water imaging and diffusion tensor imaging. Magn. Reson. Med. 2018; 80: 2402–2414

Figures

Figure 1: Subset of data recorded at 50° orientation (experimental results are indicated as red triangles). The simulated data (blue circles) were calculated by the BSBM assuming a super-Lorentzian lineshape and show excellent overall agreement (deviations are less than 2% compared to the experiment). Data points used for the first fitting step (MTcos) are shaded gray.


Figure 2: Overview of adjusted parameters. Left, results from initial fit (using only MTcos data, shaded in Fig. 1) are shown. These parameters were used as fixed values (except T2B) in subsequent simultaneous analysis of all acquired data sets. Right, parameters describing the dipolar reservoir (T1,D & ωD) and readjusted T2B are shown. This “complete fitting” was done according to both approaches (i, ii) resulting in similar values for T1,D and ωD, but opposite trends for T2B. Parameter L(5,5) (matrix element) was used as a substitute to describe the effect of the inversion pulse on MzB.


Figure 3: Orientation dependence of MT saturation (= SMT/S0) by MTcos preparation for different offset frequencies. Error bars indicate standard deviations across the ROI. Results calculated with the classical BSBM (super-Lorentzian) using the parameters from Fig. 2 are shown as blue circles. In comparison, the green curve represents the expected orientation dependence as a result of a lineshape variation according to the realistic fiber model (approach ii, κ1≈ 6.92, κ2≈ 1.74) assuming no orientation dependence for all other BSBM parameters (mean values are used as in Fig. 2).


Figure 4: MT saturation obtained with single-sided (blue symbols, MTpos) and dual-sided (yellow symbols, MTdual) off-resonant saturation and the calculated ihMTR (red triangles) - dashed lines to guide the eye. Data are shown for four offset-frequencies, indicating how important these are for evaluating the observed orientation dependence. Note that the trend for the ihMTR is completely reversed depending on ΔωR. Namely, the maximum value is observed at θFB = 0° for small offset-frequencies (ΔωRF < 15 kHz), while the maximum is reached at θFB = 90° for larger offsets.


Figure 5: Calculated ihMTR for three offset frequencies. Red triangles indicate the experimental results (Fig. 4). The simulation was performed according to the classical BSBM (i) assuming a super-Lorentzian lineshape for the semi-solid pool and fitted parameters as shown in Fig. 2 (blue). Alternatively, the calculation of the lineshape according to the more realistic fiber model (ii, green, Bingham parameters as given in Fig. 3) was used, again in combination with mean values of all other parameters (Fig. 2) describing the BSBM, except for those describing the dipolar reservoir.


Proc. Intl. Soc. Mag. Reson. Med. 32 (2024)
1090
DOI: https://doi.org/10.58530/2024/1090