Introduction

Magnetization transfer (MT) effects in mapping T1 or T2/T1 varying the flip angle (VFA) in fast low-angle shot (FLASH) or balanced steady-state free precession (bSSFP) sequences can be stabilized via the average absorption of RF by the “invisible” motionally restricted macromolecular protons¹. Such “controlled saturation magnetization transfer” (CSMT) can be achieved by complementing the root-mean-square of B1 by adding off-resonance sidebands to the RF pulse¹. As such RF pulses are not readily available on clinical scanners, we suggest controlling rmsB1 by pulse duration (PL). This is compared to schemes where RF duration or B_1peak is hold constant in VFA T₁ mapping of human brain at 3T.

Theory

For a given RF pulse shape and flip angle (FA), the rmsB1 of an RF pulse is proportional to FA²/PL, which can be kept constant by increasing PL with FA². If PL is constant, rmsB1 increases with FA². For constant peak B₁, rmsB1 increases with B_1peak*FA. Absorption of RF will impose varying partial saturation to the bound pool, which -- compared to 1–cos(FA) created by the FA on free pool (Figure 1) -- will lead to MT to or from the free pool that influence the steady state signal².

Methods

Measurements were performed on a 3T MAGNETOM Prisma using a 20-channel head-neck coil (Siemens Healthcare, Erlangen, Germany) using a prototype sequence that allowed setting the pulse duration. A healthy male subject (29y) and an agarose phantom (inversion recovery T₁ = 1050ms) were investigated. Sagittal non-selective volumes of 1.25 mm³ isotropic resolution (240x232x180 mm³) were acquired at TE/TR=3.78/10ms in 1:30 minutes using partial Fourier and GRAPPA acceleration. The flip angle of a rectangular pulse was varied through 2°, 3°, 4°, 5°, 6°, 8°, 10°, 12° and 15°, while recording the peak voltage. Three cases were investigated:
A) constant rmsB1: A chosen minimum PL of 24us for 2° was increased to 1800us for 15°. The peak voltage was less than 25% of the hardware limit allowing for 4 times higher FA or 2 times shorter PL.
B) constant B1peak: PL was varied from 80us to 600us (kept within the range of A).
C) A constant PL of 400us was chosen, to keep peak voltages within the range of A.
The VFA signals S(FA) were plotted as Y=S(FA)/FA over X=S(FA)*FA (FA in radian) to enhance signal bias by deviation from a straight line³. Signal amplitude and T₁ were obtained by linear regression. The MR system’s B₁⁺ mapping based on multi-slice saturation recovery⁴ was used to account for local flip angles deviating from the nominal value.

Results and Discussion

At constant rmsB1, the VFA signals followed the Ernst equation, that is, a straight line in Figure 2, better than at constant PL or constant B_1peak. This is in line with results using CSMT pulses¹. Specifically, the latter showed lower signal at higher FA, where the saturation of the bound pool increases.
After each RF pulse, more magnetization will be transferred from the free to the invisible pool, resulting in a lower steady state. In turn, the T₁’s obtained with constant rmsB1 tended to be lower (Figure 3). Even at very long pulse lengths of 1800ms, effects of large B₀ offset were not observed orbito-frontally at 3T.
This proof-of-principle suggests that it can be extended to higher flip angle and use of sinc pulses. Further investigations of incidental MT-effects on T₁ mapping by VFA will be compared to MP-RAGE based isotropic maps of T₁ obtained from fitting the inversion recovery, which has been specifically developed for the purpose of revealing spatial bias (abstract submitted),

Conclusion

By simply increasing the RF duration with the square of the FA, MT effects can be controlled in VFA, reducing systematic signal deviation, but leading to lower T₁ estimates.

Acknowledgements

The sequence was kindly provided by Dr A. Lutti, University of Lausanne, Switzerland. Funding by the Swedish Research Council (NT-2014-6193) and local support by Dr. F. Testud, Siemens Healthcare Sweden, is gratefully acknowledged.