1154

Dixon-Based B0-Navigation to Correct B0 Drift and B0 Fluctuations in Radial Stack-Of-Stars Multi-Echo Gradient Echo Imaging

Jonathan K. Stelter¹, Mingming Wu¹, Johannes Raspe¹, Philipp Braun¹, Christof Boehm¹, Kilian Weiss², and Dimitrios C. Karampinos¹
¹Department of Diagnostic and Interventional Radiology, Technical University of Munich, Munich, Germany, ²Philips Healthcare, Hamburg, Germany

Synopsis

Keywords: Artifacts, Motion Correction

Multi-echo gradient-echo imaging is known to be affected by temporally varying B0 effects, including B0 drifts and respiratory motion-induced B0 fluctuations. The radial stack-of-stars trajectory enables the oversampling of the k-space center and has been previously employed for B0-navigation without the consideration of fat. The present work develops a methodology for Dixon-based B0-navigation to correct for B0 drift and B0 fluctuations in stack-of-stars multi-echo gradient-echo imaging for body imaging. Simulations and in vivo measurements show the advantage of the Dixon-based B0-navigation in correcting quantification errors in proton density fat fraction, T2* and field-map, when using stack-of-stars multi-echo gradient-echo acquisitions.

Introduction

Multi-echo gradient-echo imaging is established in quantitative body imaging settings, including fat quantification¹, fatty acid composition determination², T2* mapping³ and quantitative susceptibility mapping⁴. With increasing number of echoes and maximum echo time, temporally varying B0 effects induced by hardware-originating B0 drifts and respiratory motion-originating B0 fluctuations have been shown to affect the parameter quantification even in body regions not primarily affected by gross motion⁵^,6,⁷. The radial stack-of-stars (SoS) trajectory allows oversampling of the k-space center over time and has been extensively used for self-navigation in moving organs⁸. Furthermore, SoS acquisitions have been proposed to correct temporal B0 effects based on the phase of the k-space center^9,10, but without accounting for fat. An 1D-Dixon analysis would be ideal to process the multi-echo self-navigator of a SoS acquisition to estimate the temporal B0-navigation independent of the underlying water-fat composition. Therefore, the present work develops a methodology for Dixon-based B0-navigation to correct B0 drift and B0 fluctuations in SoS multi-echo gradient-echo imaging.

Methods

B0 self-navigation and correction method
The proposed method determines B0 changes based on the oversampled k-space center of the SoS trajectory (Fig.1). The center of each spoke was selected and projection magnitude and phase images were computed based on the IFFT along the partition direction. System imperfections were corrected (e.g. gradient delays, eddy currents) to compensate for a low-frequency signal if the spokes are ordered based on the angle increment¹¹. All magnitude and phase projections for all uneven echoes and individually for all coils were selected and a graph-cut-based water-fat separation¹² was performed yielding water and fat signals as well as the field-map. Given the spatial variation of B0 fluctuations, the $$$\Delta$$$B0 map was estimated by computing the coil-wise B0 difference between each spoke and the last acquired spoke and weighting this information based on the relative signal of each coil to use the coils' different spatial information.
Two different strategies were applied to correct for B0 drifts and fluctuations found in the navigator signal: A global correction using the median along the partition direction to increase the robustness of the navigator signal ($$$\Delta\text{B0} = \text{median}_{z}(\Delta\text{B0}')$$$ with $$$z$$$ being the slices in feet-head (F/H) direction) or a slice-wise correction using the $$$\Delta$$$B0 map. Each k-space spoke $$$s'_{jkz}$$$ in shot $$$j$$$ at echo time $$$\text{TE}_k$$$ was corrected:
$$s_{jkz}=s'_{jkz}e^{-i2\pi\Delta\text{B0}{jz}\text{TE}_k}$$
After performing the correction, coil sensitivities maps were computed using ESPIRiT¹³and images were reconstructed using a nonuniform fast Fourier transform in Julia¹⁴.
Phantom and in vivo measurements
Measurements were performed at 3T (Ingenia Elition, Philips Healthcare) on a water-fat phantom with varying T1/T2 (Calimetrix, Madison, USA) and two volunteers. Respiratory motion-induced B0 fluctuations were simulated by modulating the phase of each spoke of the phantom scan with a sinusoidal phase (f=15min^-1, spatially constant B0, amplitude=3Hz). The neck (FOV=400x400x120mm³, voxel size=2x2x5mm³, 601/301 spokes per partition with scan time=5:15/2:45min) and pelvis (FOV=450x450x210mm³, voxel size=1.95x1.95x10mm³, 361 spokes per partition, scan time=2:51min) regions were scanned with a 12-echoes multi-echo SoS acquisition (FA=3°, TR/TE1/$$$\Delta$$$TE=13/1.13/0.9ms). The pelvis scan parameters were used for the phantom (FOV=450x450x120mm³, voxel size=1.95x1.95x5mm³, 541 spokes per partition, scan time=5min). For the pelvis, a Cartesian reference scan with similar parameters was acquired (scan time=1:47min).
Evaluation
The phantom and neck scans were processed with the global correction for $$$\Delta$$$B0 and the pelvis scan was processed with the slice-wise correction for $$$\Delta$$$B0. The simulation, phantom and in vivo results were evaluated similarly. Uncorrected and corrected magnitude and phase images were compared and difference images were computed. Furthermore, PDFF, T2* and the field-map were estimated using a graph-cut algorithm¹⁵.

Results

Simulations in Fig.2 show a decrease in T2* due to motion-induced B0 fluctuations. Fig.3 shows the effect of the proposed $$$\Delta$$$B0 correction on the B0 drift in a phantom experiment, reporting an increase in T2* and a constant offset in the field-map for the corrected parameters. In vivo results in the neck (Fig.4) show results in agreement with the phantom results with additionally observed differences in PDFF in the order of 2.5%. Fig.5 presents the pelvis results, where a slice-wise $$$\Delta$$$B0 correction was performed due to the observed differences of the B0 fluctuations in F/H direction.

Discussion

Simulation, phantom and in vivo results in two anatomies showed the feasibility of the proposed Dixon-based B0 self-navigation to detect and correct the B0 drift and B0 fluctuations. The SoS acquisition is generally motion robust and the magnitude differences in the echo images due to B0 drift and fluctuations are generally low, increasing for later echo times. However, B0 drift and fluctuations effect the quantification of PDFF, T2* and the field-map with highest implications on the T2* accuracy, similar as previously shown in Cartesian acquisitions of the neck⁷. The quantification accuracy of the in vivo measurements in the study may be additionally confounded by system imperfections, partial volume effects due to the low resolution in F/H-direction or other types of motion (e.g. bowel motion). Further investigations are warranted also with respect to a reference standard.

Conclusion

The proposed Dixon-based B0-navigation enables the correction of B0 drifts and fluctuations in different anatomies in the body which has been shown to be an important correction for an accurate quantification of PDFF, T2* and the field-map.

Acknowledgements

The present work was supported by the TUM International Graduate School of Science and Engineering (TUMICL Joint Academy of Doctoral Studies) and the German Research Foundation (project number 455422993, FOR5298-iMAGO-P1). The authors also acknowledge research support from Philips Healthcare.

References

[1] Reeder SB, Cruite I, Hamilton G, Sirlin CB. Quantitative assessment of liver fat with magnetic resonance imaging and spectroscopy. J Magn Reson 2011; 34:729–749. 10.1002/jmri.22775.

[2] Berglund J, Ahlström H, Kullberg J. Model-based mapping of fat unsaturation and chain length by chemical shift imaging-phantom validation and in vivo feasibility. Magnetic Resonance in Medicine 2012; 68:1815–1827. 10.1002/mrm.24196.

[3] Hernando D, Levin YS, Sirlin CB, Reeder SB. Quantification of liver iron with MRI: State of the art and remaining challenges. Journal of Magnetic Resonance Imaging 2014; 40:1003–1021. 10.1002/jmri.24584.

[4] Wang Y, Spincemaille P, Liu Z, Dimov A, Deh K, Li J, Zhang Y, Yao Y, Gillen KM, Wilman AH, Gupta A, Tsiouris AJ, Kovanlikaya I, Chiang GCY, Weinsaft JW, Tanenbaum L, Chen W, Zhu W, Chang S, Lou M, Kopell BH, Kaplitt MG, Devos D, Hirai T, Huang X, Korogi Y, Shtilbans A, Jahng GH, Pelletier D, Gauthier SA, Pitt D, Bush AI, Brittenham GM, Prince MR. Clinical quantitative susceptibility mapping (QSM): Biometal imaging and its emerging roles in patient care. Journal of Magnetic Resonance Imaging 2017; 46:951–971. 10.1002/jmri.25693.

[5] Baboli M, Storey P, Sood TP, Fogarty J, Moccaldi M, Lewin A, Moy L, Kim SG. Bilateral gradient-echo spectroscopic imaging with correction of frequency variations for measurement of fatty acid composition in mammary adipose tissue. Magnetic Resonance in Medicine 2021; 86:33–45. 10.1002/mrm.28692.

[6] Vannesjo SJ, Miller KL, Clare S, Tracey I. Spatiotemporal characterization of breathinginduced B0 field fluctuations in the cervical spinal cord at 7T. NeuroImage 2018; 167:191–202. 10.1016/j.neuroimage.2017.11.031.

[7] Wu M, Held C, Diefenbach MN, Meineke J, Mishre ASS, Weiss K, Kan HE, Junker D, Hauner H, Karampinos DC. Breathing-induced B0 fluctuations bias proton density fat fraction and T2* mapping of brown adipose tissue in the supraclavicular fossa. Proc Intl Soc Mag Reson Med 29 2021; p. 3854.

[8] Feng L, Axel L, Chandarana H, Block KT, Sodickson DK, Otazo R. XD-GRASP: Golden-angle radial MRI with reconstruction of extra motion-state dimensions using compressed sensing. Magnetic Resonance in Medicine 2015; 75:775–788. 10.1002/mrm.25665.

[9] Jonathan SV, Grissom WA. Volumetric MRI thermometry using a three-dimensional stack-of-stars echoplanar imaging pulse sequence. Magnetic Resonance in Medicine 2017; 79:2003–2013. 10.1002/mrm.26862

[10] Svedin BT, Payne A, Bolster BD, Parker DL. Multiecho pseudo-golden angle stack of stars thermometry with high spatial and temporal resolution using k-space weighted image contrast. Magnetic Resonance in Medicine 2017; 79:1407–1419. 10.1002/mrm.26797.

[11] Rosenzweig S, Scholand N, Holme HCM, Uecker M. Cardiac and Respiratory Self-Gating in Radial MRI Using an Adapted Singular Spectrum Analysis (SSA-FARY). IEEE Transactions on Medical Imaging 2020; 39:3029–3041. 10.1109/tmi.2020.2985994.

[12] Boehm C, Diefenbach MN, Makowski MR, Karampinos DC. Improved Body Quantitative Susceptibility Mapping By Using a Variable-layer Single-min-cut Graph-cut for Field-mapping. Magn Reson Med 2020; 85:1697–1712. 10.1002/mrm.28515.

[13] Uecker M, Lai P, Murphy MJ, Virtue P, Elad M, Pauly JM, Vasanawala SS, Lustig M. ESPIRiT-an eigenvalue approach to autocalibrating parallel MRI: Where SENSE meets GRAPPA. Magnetic Resonance in Medicine 2013; 71:990–1001. 10.1002/mrm.24751.

[14] Knopp T, Grosser M. MRIReco.jl: An MRI reconstruction framework written in Julia. Magnetic Resonance in Medicine 2021; 86:1633–1646. 10.1002/mrm.28792.

[15] Stelter JK, Boehm C, Ruschke S, Weiss K, Diefenbach MN, Wu M, Borde T, Schmidt GP, Makowski MR, Fallenberg EM, Karampinos DC. Hierarchical Multi-Resolution Graph-Cuts for Water-Fat-Silicone Separation in Breast MRI. IEEE Transactions on Medical Imaging 2022; 41:3253–3265. 10.1109/tmi.2022.3180302.

Figures

Fig.1: Schematic overview of the proposed Dixon navigator and ∆B0 correction scheme for a radial stack-of-stars acquisition. The self-navigation signal was extracted from the central k-space and the projections in feet-head direction were decomposed in the water and fat signal and the B0. The ∆B0-map was used to correct for B0 drift and fluctuations.

Fig.2: Phantom results for the proposed ∆B0 correction in the presence of both B0 drift and respiratory motion-induced B0 fluctuations. A sinusoidal B0 fluctuation was modulated on the observed B0 drift. Results show a smaller magnitude and an increasing phase offset for the simulated image with increasing echo time. T2* is the most heavily affected extracted parameter.

Fig.3: Phantom results for the proposed ∆B0 correction in the presence of only a B0 drift. The corrected images show an increase in the magnitude signal and a phase offset increasing with the echo time. The uncorrected T2* map shows lower T2* values whereas differences in PDFF are small. There is a constant offset between the corrected and uncorrected field-map.

Fig.4: In vivo results in the neck for two volunteers. In (A), the scan time was increased to over 5 minutes to evaluate the effect of the B0 drift. B0 differences are mainly caused by the B0 drift showing large phase differences and also a field-map difference of 10Hz comparing the uncorrected and corrected images. In (B), respiratory breathing-induced B0 fluctuations are the major contribution to ∆B0. Differences in T2* and field-map are considerably smaller than for (A) but differences up to 2.5% can be observed for the PDFF map.

Fig.5: In vivo results in the pelvis for one volunteer comparing the free-breathing radial stack-of-stars acquisition with a Cartesian free-breathing acquisition. The stack-of-stars acquisition shows generally a higher motion robustness whereas the Cartesian acquisition has motion artifacts. The ∆B0 signal might not only cover the respiratory motion but there might be also other contributions from bowel motion. Differences between the uncorrected and corrected images are especially large near to the air-tissue interface.

Proc. Intl. Soc. Mag. Reson. Med. 31 (2023)

1154

DOI: https://doi.org/10.58530/2023/1154