3255
Partially spoiled non-balanced oscillating steady state imaging - A new T2-mapping approach1Department of Diagnostic and Interventional Radiology, University Hospital Würzburg, Würzburg, Germany
Synopsis
Keywords: Pulse Sequence Design, New Signal Preparation Schemes
Motivation: Determination of T2 typically depends on spin-echo based relaxometry. Partially spoiled non-balanced gradient echo imaging also allows T2-mapping but requires two data sets from two different steady states, making it prone to motion artefacts.
Goal(s): To determine T2 from a single T2-weighted steady state.
Approach: A combination of partial spoiling and oscillating steady state imaging was optimized using Bloch simulations, implemented on a human whole-body MR-scanner and tested on a test object and in vivo.
Results: Partially spoiled oscillating steady state imaging allowed generating phase images similar to partially spoiled non-balanced SSFP imaging without the need to sequentially establish two different steady states.
Impact: Partially spoiled gradient echo imaging allows T2-mapping without using large flip angles but requires data from two different steady states. By combining partial spoiling with oscillating steady state imaging T2 can be determined from one single oscillating steady state.
Introduction
Typically, T2 is determined from spin echo based acquisitions resulting in prolonged examinations. Partial spoiling of non-balanced gradient echoes [1] allows fast mapping of T2 [2, 3], but requires establishing two different steady states consecutively. Gradient echoes, where the phase [4] or the amplitude [5] of the excitation pulse or the repetition time [6] are varied periodically (with n steps) lead to oscillating steady states. That means, the signals of n consecutively acquired echoes differ, but every n-th echo is equal. Or, in other words, n different “projections” of one established steady state develop. In this work, we combined partial spoiling and non-balanced oscillating steady state imaging to determine T2 from different projections of one oscillating steady state.Methods
Bloch simulations were performed to calculate the signal intensities of bSSFP, partially spoiled bSSFP, oscillating steady states and partially spoiled oscillating steady states. For non-balanced SSFP sequences the signal was calculated by integrating over a dephasing of 2π per TR. Simulation parameters (unless otherwise stated in the figures) were: T1=700 ms, T2=60 ms, TR=6 ms, TE=0.1 ms, flip angle α=12°, flip angle variation Δα=2°, quadratic phase increment Θ=0.7°. On a 1.5 T scanner (Avanto Fit, Siemens Healthcare, Erlangen, Germany) experiments were performed using non-balanced oscillating steady states with partial and full spoiling, i.e. quadratic phase increments of 0.7/1° and 117°, respectively. The sequence diagram is shown in fig 1. All sequences were equipped with a spiral read out optimized to avoid peripheral nerve stimulations [7]. As a test object the Essential System Phantom (Caliber MRI, Boulder Co, USA) was used [8]. Here the acquisition parameters were TR=6 ms, TE=1 ms, flip angle α=7.5°, flip angle variation Δα=2.5°, quadratic phase increment Θ=0.7°. Additionally, in vivo images were acquired from the head of a healthy volunteer with the following parameters: TR=5.4 ms, TE=0.6 ms, flip angle α=12°, flip angle variation Δα=2°, quadratic phase increment Θ=1°. Images were reconstructed using convolution gridding with a Voronoi density compensation [9].Results
Figure 2 shows the complex signal intensities of bSSFP, partially spoiled bSSFP, oscillating steady states and partially spoiled oscillating steady states. The signal intensities of the non-balanced sequence are displayed as crosses in figure 2 and 3 (top left). The imaginary part of the signal is similar for both echoes of the non-balanced oscillating steady state sequence and the corresponding non-balanced SSFP sequence and the real part is similarly increased and decrease from SSFP to the even and odd echoes of the oscillation steady states. Figure 3 also displays simulations showing the dependency of the complex signal amplitudes on T1, T2 and flip angle attenuations. It can be recognised that the imaginary part of the signal strongly depends on T2, but is largely independent of T1 or the flip angle attenuation. Additionally, for an appropriate choice of parameters the imaginary parts of both echoes appear to be similar. That means, the phase of difference between the complex signals of the two echoes of the oscillating steady state can be used as an estimate for the background phase. Figure 4 presents a reconstruction of the “T2”-slice of the Essential System Phantom using a partially spoiled non-balanced oscillating steady state sequence. The phase images generated using either the difference of the two echoes or an additional reference measurement (using a quadratic spoiling of 117°) appear similar. A quantitative evaluation presents agreement within the errors of the two methods. The results of the quantitative evaluation in the phantom could be used as a basis for a look-up table to convert phase images to quantify T2-maps in the future [3]. In figure 5 the magnitude and phase images (using only the partially spoiled non-balanced oscillating steady state or the reference measurement with quadratic spoiling of 117°) are shown. Again, the differently generated phase images appear similar, however with more noise in the image with the oscillating steady state phase reference. Both images show the expected T2-dependence.Discussion and Conclusion
Partially spoiled oscillating steady states were simulated and tested on a clinical scanner. The implementation of a single oscillating steady state, where two different projections can be acquired at odd and even echoes, respectively, allow to continuously sample T2-weighted information and retrieve the background phase. The use of a spiral read-out allowed a short echo time and thus reduced T2*-weighting but is not mandatory for partially spoiled non-balanced oscillating steady state imaging. In summary, partially spoiled oscillating steady states allow generating T2-maps without the need to sequentially settle two different steady states.Acknowledgements
No acknowledgement found.References
1. Zur, Y., M.L. Wood, and L.J. Neuringer, Spoiling of Transverse Magnetization in Steady-State Sequences. Magnetic Resonance in Medicine, 1991. 21(2): p. 251-263.
2. Bieri, O., et al., Quantitative Mapping of T-2 Using Partial Spoiling. Magnetic Resonance in Medicine, 2011. 66(2): p. 410-418.
3. Wang, X.K., D. Hernando, and S.B. Reeder, Phase-based T-2 mapping with gradient echo imaging. Magnetic Resonance in Medicine, 2020. 84(2): p. 609-619.
4. Vasanawala, S.S., J.M. Pauly, and D.G. Nishimura, Fluctuating equilibrium MRI. Magnetic Resonance in Medicine, 1999. 42(5): p. 876-883.
5. Absil, J., V. Denolin, and T. Metens, Fat attenuation using a dual steady-state balanced-SSFP sequence with periodically variable flip angles. Magnetic Resonance in Medicine, 2006. 55(2): p. 343-351.
6. Leupold, J., J. Hennig, and K. Scheffler, Alternating repetition time balanced steady state free precession. Magnetic Resonance in Medicine, 2006. 55(3): p. 557-565.
7. Schad, O., T. Wech, and H. Köstler, Consideration of peripheral nerve stimulation in the optimization of spiral k-spacetrajectories. Proceedings of the ISMRM, 2023. 31: p. 2872.
8. Caliber MRI, Model 106 Essential System Phantom Spec Sheet. Last retrieved: 02.11.2023. Available at: https://qmri.com/wp-content/uploads/2023/05/Model-106-Essential-System-Phantom-specs-Rev-B.pdf
9. Jeffrey A. Fessler, Michigan Image Reconstruction Toolbox. Last retrieved: 06.11.2023. Available at: https://web.eecs.umich.edu/~fessler/code/.
Figures
