2891
Anthropomorphic Spinal Cord Phantom with Respiratory B0 Field Fluctuations1Translational and Molecular Imaging Institute, Icahn School of Medicine at Mount Sinai, New York, NY, United States, 2Department of Radiology, Icahn School of Medicine at Mount Sinai, New York, NY, United States, 3Graduate School of Biomedical Sciences, Icahn School of Medicine at Mount Sinai, New York, NY, United States, 4Department of Biomedical Engineering, The City College of New York, New York, NY, United States, 5Department of Neuroscience, Icahn School of Medicine at Mount Sinai, New York, NY, United States
Synopsis
The spinal cord exists in an unfavorable magnetic field environment; the lungs produce strong B0 field inhomogeneities that vary over time. We have designed and built a phantom that simulates these temporal field distortions to aid in the development of spinal cord imaging methods. The phantom consists of an acrylic tank, two lung simulants, a spinal cord and canal phantom, and a microcontroller-governed air pump. The respiratory waveform is customizable. This phantom accurately reproduced the ~20Hz respiratory-induced field shifts observed in vivo at the C3 vertebral level at 7T, and, being fully synthetic, is stable and replicable.
Introduction
The human spinal cord exists in a uniquely unfavorable magnetic field environment, rendering advanced MRI of the spinal cord technically challenging [1]. Air, which is paramagnetic, creates strong magnetic field inhomogeneities; as the volume of air in the lungs changes during respiration, the B0 field in the adjacent spinal cord can vary by up to 74Hz at 3T [2] and over 100Hz at 7T [3]. EPI-based sequences, commonly used in functional and diffusion MRI, are particularly sensitive to B0 field inhomogeneity. In order to simulate these temporal B0 field fluctuations and thereby aid in the development of spinal cord imaging methods that are robust to these effects, de Tillieux et al. designed a spinal cord respiratory phantom composed of a preserved human vertebral column and balloons for lung simulants [4]. Here, we present an alternative design using fully synthetic material.Methods
An acrylic tank approximating the size and shape of the human head, neck, and torso was built to interface with a custom brainstem/cervical spinal cord RF coil [5]. The tank was filled with water, and a previously-presented spinal cord/canal phantom [6] was affixed inside the tank between two lung simulants (Fig. 1a-d). Each lung simulant comprises an acrylic cylinder 11cm in both diameter and height, open at the bottom and capped with a funnel.
The lung simulants are driven by an air pump composed of a linear actuator (Fig. 1e) and a neoprene bellows. The actuator is controlled by a microcontroller (Uno, Arduino LLC), allowing control of tidal volume, rate, and timing. The top port of each lung was connected to the pump by 10m of 0.5-in inner diameter vinyl tubing, joined at a T-junction 70cm from the lungs. At rest, the air/water level inside the lung simulants is at the same level as the air/water level in the tank. When air is pumped into the lung simulants, water is displaced and exits through the open bottom of the lung simulants into the tank.
The actuation paradigm used for the imaging experiments was 2s inspiration, 2s expiration, and 4s rest at end expiration, looping continuously. The tidal volume (~1L) was empirically tuned to achieve a target field perturbation of 20Hz between inspired and expired states at the C3 level of the cord [3]. Static shimming was performed with respiration disabled at full expiration.
All images were acquired on a 7T AS whole-body scanner (Magnetom, Siemens) using a custom 22-channel brainstem/cervical spinal cord RF array coil. To rapidly capture magnetic field and signal magnitude changes induced by the simulated respiration, single-shot gradient-echo (GRE)-EPI images were acquired with the following protocol: 1x1x3mm3, 18 axial slices, TR/TE=625/15.2ms, flip angle=50°, phase encoding AP, GRAPPA R=2, partial Fourier=6/8, BW=1430Hz/px, echo spacing=0.82ms, one coronal saturation slab anterior to the cord. Respiration was disabled until 20s into the scan. Respiration-induced frequency fluctuations were calculated relative to the mean of the initial 20s baseline.
To demonstrate the effect of respiratory induced field fluctuation on structural image quality, multi-echo GRE images were acquired with the following protocol: 0.3x0.3x3.0mm3, 9 axial slices, TR/TE1/TE2/TE3/TE4=293/3.3/9.2/15.2/21.1ms, flip angle=20°, phase encoding RL, GRAPPA R=3, BW=260Hz/px, acquisition time=3min. Echoes were combined by root-sum-of-squares. One acquisition was performed with respiration off, and another with respiration on (including GRAPPA reference scans).
Results
Magnitude (Fig. 2a,b) and phase (Fig. 2c,d) GRE-EPI images show more artifacts and greater phase variation in fully inspired (Fig. 2b,d) than expired (Fig. 2a,c) states. Traces of frequency and signal magnitude offsets at the center of the spinal cord at the C3 level show a 20Hz frequency offset and 30% change in signal magnitude over the simulated respiratory cycle (Fig. 3). The detrimental effects of the simulated respiration on high-resolution multi-echo GRE images are demonstrated in Figure 4.Discussion
This synthetic phantom reproduced the respiratory-induced field shifts measured by Vannesjo et al. in vivo at 7T [3] and is capable of simulating comparable respiratory tidal volumes to de Tillieux’s phantom. Compared to the spinal cord respiratory phantom presented by de Tillieux et al. [4], our design prioritized stability and replicability of the phantom without using biological specimens. Similarly, Bolwin et al. [7] used only synthetic materials in their phantom, which is designed to simulate cardiac and respiratory motion of organs.
We will combine this respiratory system with the previously-presented static field inhomogeneity phantom [6] to simulate both spatially- and temporally-periodic field variations produced in the spinal cord by the vertebral column and lungs, respectively.
Conclusion
This phantom will accelerate the development of imaging methods in the spinal cord.Acknowledgements
This study was supported by National Institutes of Health (NINDS) award number K01NS105160 (ACS).References
[1] Stroman PW, Wheeler-Kingshott C, Bacon M, et al. The current state-of-the-art of spinal cord imaging: methods. NeuroImage 2014;84:1070-1081.
[2] Verma T and Cohen-Adad j. Effect of respiration on the B0 field in the human spinal cord at 3T. Magn. Reson. Med. 2014;72:1629-1636.
[3] Vannesjo SJ, Miller KL, Clare S, Tracey I. Spatiotemporal characterization of breathing-induced B0 field fluctuations in the cervical spinal cord at 7T. NeuroImage 2018;167:191-202.
[4] De Tillieux P, Topfer R, Foias A, et al. A pneumatic phantom for mimicking respiration-induced artifacts in spinal MRI. Magnetic Resonance in Medicine 2018;79(1):600-605.
[5] Zhang B, Seifert AC, Kim JW, Borrello J, Xu J. 7 Tesla 22-Channel Wrap-Around Coil Array for Cervical Spinal Cord and Brainstem Imaging. Magnetic Resonance in Medicine 2017;78(4):1623-1634.
[6] Seifert AC, Patel V, Grace M, et al. Anthropomorphic Spinal Cord Phantom with Induced Field Inhomogeneity. In: Proceedings of the 25th Annual Meeting of the International Society for Magnetic Resonance in Medicine; 2017 April 22-27; Honolulu, HI. Abstract 920.
[7] Bolwin K, Czekalla B, Frohwein LJ, et al. Anthropomorphic thorax phantom for cardio-respiratory motion simulation in tomographic imaging. Physics in Medicine & Biology 2018;63:035009.
Figures
