1215

A totally balanced spin lock preparation module for accurate and artifact-free T1ρ-mapping
Maximilian Gram1,2, Daniel Gensler1,3, Anton Xu2,3, Peter Nordbeck1,3, Wolfgang Rudolf Bauer1,3, Peter Michael Jakob2, and Michael Seethaler2,3

1Department of Internal Medicine I, University Hospital Würzburg, Würzburg, Germany, 2Experimental Physics V, University of Würzburg, Würzburg, Germany, 3Comprehensive Heart Failure Center (CHFC), University Hospital Würzburg, Würzburg, Germany

Synopsis

To date, one of the main challenges of accurate and artifact-free T1ρ-mapping is the sensitivity of the required spin lock preparation module against field imperfections. Already established methods are sensitive to B0 and B1 field inhomogeneities. In this work, we present a novel spin lock preparation module that aims to be totally balanced, meaning that every pulse is being compensated by a correspondent pulse with opposite phase. Our new method proves to be highly robust to both types of field inhomogeneities. The superiority over common methods is demonstrated by Bloch simulations and measurements of a glucose phantom at 7.0T.

Introduction

In the last decade, T1ρ-weighted imaging as well as T1ρ-mapping have become increasingly popular imaging methods with additional improved contrast mechanisms, as large numbers of clinical studies have shown [1]. The T1ρ relaxation mechanism exhibits a high sensitivity for slow motional processes like low frequency processes at the molecular and cellular level and is therefore a useful tool for achieving new specific tissue contrasts. The information provided by T1ρ cannot be obtained by conventional spin-lattice or spin-spin relaxation techniques [2]. However, a major problem for an accurate and artifact-free T1ρ-mapping arises from the spin lock preparation module. To date, currently available techniques show sensitivity to B0- and/or B1 field inhomogeneities [3,4,5]. Hence, new improved techniques are needed to enable robust and artifact-free T1ρ imaging, potentially increasing its application in clinical routine.

In the current study, a new preparation module is presented that aims to achieve a high degree of balance by compensating every pulse with a correspondent pulse of opposite phase. To prove the performance of this new preparation module, several simulations as well as measurements have been performed indicating a significantly improved behavior compared to commonly available techniques.


Methods

All measurements were performed on a 7.0T small animal imaging system Bruker BioSpec 70/30 (Bruker BioSpin MRI GmbH, Ettlingen, Germany) and the simulations were done in MATLAB R2017a (The MathWorks, Massachusetts, USA).

Our novel preparation module consists of two 90-degree pulses to tip the magnetization into the transverse plane and back. The spin lock itself is done with four separated pulses with alternating phases. To compensate for B0 imperfections, two 180-degree refocusing pulses with opposite phases were included. This new preparation module was compared to common preparation methods (Fig. 1). The quantification error of T1ρ was calculated for the different modules using a Bloch simulation (Fig. 2). The coefficient of determination (R-squared) was used to investigate whether the modules show a pure mono-exponential relationship with T1ρ. Afterwards, the theoretical predictions were validated by measurements on a homogeneous phantom using certain B0 and B1 field imperfections (Fig. 3). The phantom consisted of a homogeneous cylindrical sample filled with a 50% solution of glucose and water. B0 and B1 inhomogeneities were increased by off-resonance irradiation (up to +1ppm) and incorrect flip angles (up to -50%), respectively.

For the imaging readout module, a standard turbo spin echo sequence (TSE) has been used. Imaging parameters were: TR=5000ms, TE=7ms, turbo factor: 2, slice thickness: 5mm, fov: 30x30mm2, matrix: 96x96


Results

The simulation results (Fig. 2) show that the standard sequence (S-SL) only provides a good T1ρ quantification if the spin lock condition is perfectly met. The rotary-echo module (RE-SL) is suitable to compensate pure B1 field imperfections and the composite module (C-SL) provides good T1ρ quantification for B0 imperfections. However, the C-SL module does not show satisfactory accuracy in the case of B1 imperfections because of the sensitivity of the 180-degree refocusing pulse. The simulation results demonstrate that only the totally-balanced module (TB-SL) effectively combines both, the B0 and the B1 compensation mechanism. The quantification accuracy of T1ρ and the R-squared value remain stable within a large range. These results are confirmed by measurements on the glucose phantom (Fig. 3). The measurements illustrate banding artifacts that typically occur in T1ρ-based imaging. In the case of incorrect flip angles, artifacts occur with the S-SL and the C-SL module. In case of off-resonant spin locking, the S-SL and the RE-SL module show artifacts. Only the TB-SL module is able to produce artifact-free images in both configurations.

Discussion

In the current work, we presented a very robust T1ρ preparation module. Simulations and measurements show that our totally balanced setup (TB-SL) enables accurate T1ρ quantification and artifact-free T1ρ imaging even for distinct B1 and B0 imperfections. In comparison, the previously proposed methods cannot compensate both field inhomogeneities at once, resulting in even more intense imaging artifacts under certain circumstances. However, a drawback of our new method is the increased SAR due to the additional 180-degree pulse. In conclusion, our novel sequence enables artifact-free T1ρ imaging, which potentially could pave the way for robust T1ρ-mapping in clinical routine. Currently, first in-vivo investigations show promising results and work is underway on a large comparative study to statistically verify the advantages of the TB-SL sequence.

Acknowledgements

This work was supported by the Federal Ministry for Education and Research of the Federal Republic of Germany (BMBF 01EO1504, MO6).

References

1. Wáng, et al. Quant Imaging Med Surg. 2015 Dec; 5(6): 858–885.

2. Spear and Gore. NMR in Biomedicine. 2016 Sep;29(9):1258-73

3. Witschey, et al. J Magn Reson. 2007 May;186(1):75-85.

4. Schuenke, et al. Magn Reson Med. 2017 Jul;78(1):215-225

5. Berisha, et al. PLoS One. 2016 Mar 22;11(3):e0151144.

Figures

Fig. 1) The standard spin lock module (S-SL) consists of two 90-degree pulses and one spin lock pulse. The rotary echo module (RE-SL) uses two spin lock pulses with opposite phases in order to reduce artifacts due to B1 inhomogeneities. The composite module (C-SL) is able to compensate B0 inhomogeneities using a 180-degree refocusing pulse. Our novel totally balanced module (TB-SL) uses both strategies. Additionally, the 180-degree pulses compensate each other due to their opposite phase, leading to a highly increased insensitivity against B1 imperfections.

Fig. 2) T1ρ quantification accuracy for various B0- and B1 inhomogeneities. The Bloch simulations were performed with a T1ρ:T2ρ ratio of 3:1 for a B1 inhomogeneity up to ±50% and a B0 inhomogeneity up to ±300Hz. The dashed lines delineate the regions where R² of the mono-exponential behavior is higher than 0.95. The S-SL module gives good results if the spin lock condition (θ=β) is met. The RE-SL sequence shows good compensation for B1 inhomogeneities and the C-SL sequence for B0 inhomogeneities. The TB-SL sequence effectively compensates for both B0 and B1 inhomogeneities.

Fig. 3) Measurements of the homogeneous glucose phantom using different preparation modules. Incorrect flip angles (-50%) were used to increase B1 inhomogeneities (left). To illustrate the effect of B0 imperfections, the spin lock pulse was irradiated off-resonant (+1ppm). As predicted from the theoretical considerations, the RE-SL is able to compensate B1 inhomogeneities but shows distinct artifacts for B0 inhomogeneities and the C-SL can compensate B0 inhomogeneities but shows artifacts for B1 inhomogeneities. In contrast, our TB-SL module provides artifact-free images for both B0- and B1 inhomogeneities combining the strengths of the two other modules. spin lock time: TSL=50ms, spin lock amplitude: FSL=1458Hz



Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)
1215