4911

An inhomogeneous magnetization transfer (ihMT) quantification method robust to B1 and T1 variations in magnetization prepared acquisitions

Gopal Varma¹, Fanny Munsch¹, Olivier M Girard², Guillaume Duhamel², and David C Alsop¹

¹Radiology, Division of MR Research, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA, United States, ²Aix Marseille Univ, CNRS, CRMBM, Marseille, France

Synopsis

Standard MT and ihMT ratio (ihMTR) measures can be sensitive to B₁and T₁, making them less specific to tissue microstructure. Using the inverse signal, i.e. one divided by the signal, and a high flip-angle reference image in calculation of an ihMTR metric has been proposed as a metric with improved insensitivity to T₁ and B₁ in steady-state gradient-echo sequences. We present a modified method for use in prepared sequences such as magnetization prepared rapid gradient echo (MPRAGE). The sensitivity of ihMT MPRAGE metrics to T₁ and B₁ was tested using simulations and acquisitions in brains of healthy volunteers.

Introduction

The inhomogeneous magnetization transfer (ihMT) signal can be used to characterize myelinated tissues and extract information on dipolar order¹. IhMT sensitive images can be acquired either by interleaving RF irradiation into a stead-state acquisition², e.g. spoiled gradient echo (SPGR)³, or by applying irradiation in a preparation module prior to acquisition⁴, such as magnetization prepared rapid gradient echo acquisition⁵(MPRAGE) (Fig. 1a). IhMT prepared MPRAGE (or ihMTRAGE) can utilize the low duty cycle preparation that increases the ihMT signal while maintaining or lowering specific absorption rates⁴.

The approach to quantifying ihMT can affect its specificity and sensitivity. The ratio of the ihMT signal over that with no MT preparation, i.e. ihMTR, provides a semi-quantitative metric (Fig. 2a), but is dependent on a combination of quantitative parameters¹. Use of the inverse ihMT signal along with a high flip-angle (FA) reference in a steady-state ihMT sequence has been proposed as a more B₁and T₁ independent measure^3,6. The B₁ independence depends on a linear relationship of the inverse subtraction with power, as achieved in low duty cycle preparations⁴. In this work, we demonstrate the (in)dependence of the ihMTRAGE sequence to T₁and B₁ using the inverse ihMT signal multiplied by a reference acquisition with an SPGR like preparation. We carry out simulations and in vivo experiments to study the combination of the inverse ihMT signal with an SPGR preparation before the MPRAGE acquisition to mimic a high FA reference.

Methods

The MT preparation for ihMTRAGE consisted of 5ms pulses (rectangular/trapezoid-like for simulations/experiments) repeated ten times (every 100ms for 1s), with peak B₁ of 14μT (for single frequency RF irradiation at off-resonance Δ) (Fig. 1a). The SPGR-like module consisted of RF-spoiled on-resonance pulses with a 25ms repetition time (TR) applied 40 times over 1s (Fig. 1b). Several FAs for the on-resonance pulses were evaluated. An adiabatic spectrally selective inversion (ASPIR) pulse was also included in the last TR of the reference to mitigate fat signal contamination in the reference. IhMT signals were simulated using numerical integration of the differential equations associated with a model for ihMT, and parameters based on white matter (WM)⁴.

All experiments were carried out on a 3T scanner with a 32-channel receive coil. Data were collected from the brains of healthy volunteers (n=5). IhMTR and inverse ihMT, ihMTR_inv,θmetrics were calculated (Figs. 2a-b). Acquisition of high FA SPGR-like preparations with two FAs (25° and 45°) along with the zero power reference allowed construction of a pseudo-T₁map⁷. B₁mapping was conducted based on the Bloch-Siegert shift⁸and calculated as the ratio of the actual FA to that prescribed.

Results

Simulations provide a guide as to the response of the ihMT signal metrics to variations in parameters. Both ihMTR and the inverse ihMT signal show an increase with B₁and T₁(Figs. 3a-b). Multiplication of the inverse ihMT by the signal from SPGR-like preparations (Fig. 2b) compensates for changes in simulated T₁. The sensitivity of ihMTR_inv,θto B₁ was dependent on FA (Figs. 3c-d): IhMTR_inv,45°showed good insensitivity around the nominal B₁. IhMTR_inv,25° showed insensitivity to increases in B₁, but greater sensitivity to decreases in B₁ than ihMTR_inv,45°.

IhMTRAGE data were of good image quality but ihMTR_inv,45° images appeared noisier than ihMTR_inv,25° (Figs. 4b-c). This is a result of the low signal from extreme T₁ weighting of the 45° reference data. Both ihMTR_inv metrics showed greater insensitivity to the variation in B₁ in WM, with a variance divided by the mean value across manually selected WM regions in all subjects of 3x10^-4, compared with 4x10^-3 for ihMTR (Figs. 4d-f). WM in and around the CC only was used to assess the effect of T₁ variation to limit contribution from B₁ inhomogeneity (Figs. 5a-c). IhMTR_inv,25° provided the smallest variance divided by mean value of 7x10^-4. A negative correlation was observed between T₁ and all three ihMT metrics (Figs. 5d-f), which was strongest for ihMTR_inv,45°. The implications of this correlation are uncertain, since higher myelin density should decrease T₁ and increase ihMT measures.

Discussion

Inverse subtraction and a high FA reference can provide a simple, more robust measure for myelin imaging. Though simulations suggest use of a 45° FA is more robust to B₁, the poor SNR of the resultant acquisition makes use of a lower FA reference, such as 25°, preferable. Further validation of the inverse metric by comparison with quantitative ihMT model fitting should be evaluated.

Conclusion

The inverse ihMT signal from the ihMTRAGE acquisition continues to provide a metric that is robust to T₁ and B₁, and thus more specific to microstructural properties of myelin that determine dipolar order relaxation. For magnetization prepared acquisitions, there is a trade-off between simulated robustness and SNR.

Acknowledgements

No acknowledgement found.

References

1. Varma, G. et al. Interpretation of magnetization transfer from inhomogeneously broadened lines (ihMT) in tissues as a dipolar order effect within motion restricted molecules. J. Magn. Reson. 260, 67–76 (2015).

2. Mchinda, S. et al. Whole brain inhomogeneous magnetization transfer (ihMT) imaging: Sensitivity enhancement within a steady-state gradient echo sequence. Magn. Reson. Med. 79, 2607–2619 (2018).

3. Geeraert, B. L. et al. A comparison of inhomogeneous magnetization transfer, myelin volume fraction, and diffusion tensor imaging measures in healthy children. Neuroimage [Epub ahead of print] (2017). doi:10.1016/j.neuroimage.2017.09.019

4. Varma, G. et al. Low duty-cycle pulsed irradiation reduces magnetization transfer and increases the inhomogeneous magnetization transfer effect. J. Magn. Reson. (2018). doi:10.1016/j.jmr.2018.08.004

5. Saranathan, M., Tourdias, T., Bayram, E., Ghanouni, P. & Rutt, B. K. Optimization of white-matter-nulled magnetization prepared rapid gradient echo (MP-RAGE) imaging. Magn. Reson. Med. 73, 1786–1794 (2015).

6. Varma, G., Girard, O. M., Prevost, V., Duhamel, G. & Alsop, D. C. Extracting a robust inhomogeneous magnetization transfer (ihMT) rate parameter, ihMT-Rex. in Proceedings 23rd Scientific Meeting, International Society for Magnetic Resonance in Medicine 3357 (2015).

7. Preibisch, C. & Deichmann, R. Influence of RF Spoiling on the Stability and Accuracy of T1 Mapping Based on Spoiled FLASH With Varying Flip Angles. Magn. Reson. Med. 61, 125–135 (2009).

8. Sacolick, L. I., Wiesinger, F., Hancu, I. & Vogel, M. W. B 1 mapping by Bloch-Siegert shift. Magn. Reson. Med. 63, 1315–1322 (2010).

Figures

Figure 1 Diagram of sequences used to acquire data for: a) ihMT, and b) reference using an SPGR-like preparation. Both are based on an MPRAGE sequence that utilized a radial-fan-beam acquisition (10° FA; 90 readouts; TE/TR=1.8/4.6ms; 1.6mm isotropic resolution; 160x160x100 matrix; 2x2 ARC). The duration of both types of preparation is 1s, and the TR for each MPRAGE module is 2s. Dual frequency RF irradiation was achieved using cosine modulation of the pulses for the same power.

Figure 2 Semi-quantitative metrics for the ihMT signal are normally formed from the signal following MT preparations of: zero power (S₀); single frequency off-resonance (S_single); and, dual frequency off-resonance (S_dual) RF pulses. We utilize the signals from SPGR-like references with FA=θ (S_θ) on-resonance pulses to define an inverse ihMT ratio.

Figure 3 The semi-quantitative metrics: a) ihMTR; b) inverse ihMTR, ihMTR_inv; c) ihMTR_inv,25°; and d) ihMTR_inv,45°, in WM are simulated as functions of T₁, and B₁ relative to the ideal values in the MT and SPGR-like preparations. Parameters used in the simulations by numerical integration mimic those used in experiments, i.e. following 1s of: 14μT B_1,peak, 5ms rectangular off-resonance pulses every 100ms for ihMT, and instantaneous on-resonance pulses every 25ms for the SPGR-like references.

Figure 4 Representative maps of: a) ihMTR; b) ihMTR_inv,25°; and, c) ihMTR_inv45° from a healthy volunteer, as acquired in the sagittal plane, and reformatted in the coronal and axial planes. The set of 4 ihMT acquisitions were repeated three times, and single zero power MT and high FA preparations were acquired once for a total scan time of 22mins. Regions of interest manually drawn in WM to correspond with slices from B₁ mapping were used to form scatter plots of B₁ and: d) ihMTR; e) ihMTR_inv,25°; and, f) ihMTR_inv45°. Pearson’s correlation coefficient, r were: r=0.20 for ihMTR; r=-0.17 for ihMTR_inv,25°; and, r=-0.25 for ihMTR_inv,45°.

Figure 5 Representative maps of: a) T₁ estimated using data from 25° and 45° SPGR-like preparations; and, b) B₁ from multi-slice 2D acquisition using Bloch-Siegert shift (field-of-view-matched; matrix=64x64; 6.4mm slice; FA=15°; TE/TR=13/29ms) from a healthy volunteer. Scatter plots of T₁ estimates and: c) B₁; d) ihMTR; e) ihMTR_inv,25°; and, f) ihMTR_inv,45° using data from healthy volunteer subjects in regions in and around the corpus callosum to reduce effects of B₁ inhomogeneity on T₁ estimates. Measures of variance divided by the mean and Pearson’s correlation coefficient, r were: 9x10^-3 and r=-0.71 for ihMTR; 8x10^-4 and r=-0.76 for ihMTR_inv,25°; and, 7x10^-4 and r=-0.81 for ihMTR_inv,45°.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)

4911