0943

A 6.3kg Single-Sided Magnet for 3D, Point-of-Care Brain Imaging

Patrick C McDaniel^1,2, Clarissa Z Cooley², Jason P Stockmann², and Lawrence L Wald^2,3

¹Massachusetts Institute of Technology, Cambridge, MA, United States, ²Athinoula A Martinos Center for Biomedical Imaging, Charlestown, MA, United States, ³Harvard Medical School, Boston, MA, United States

Synopsis

MRI, as currently used, requires transporting the patient to the scanner. A truly point-of-care MRI device, possibly even hand-held, could increase the utility of MRI extending its reach and enabling new applications, such as continuous bedside monitoring. In this work, we design and construct a light-weight (6.3kg), single-sided permanent magnet designed to image the cortical region it is positioned over (~8cm x 8cm x 3cm ROI). We describe the magnet optimization and compare the predicted and measured B0 field pattern and validate its imaging potential by acquiring 1D depth profiles in a phantom.

Introduction

Size, expense and siting issues prohibit conventional MRI systems from point-of-care use in almost all clinical settings. Although effort is building toward inexpensive, easily-sited systems for rural, developing-world, or even bedside settings^1,2 and considerable work has been put into single-sided MR devices for rock and materials characterization^3–7 only a few studies⁸ have focused on highly portable, or hand-held devices for medical applications. In this work, we assess the feasibility of a lightweight single-sided device capable of imaging a few centimeters into the human brain. We optimize a rare-earth permanent magnet array for this purpose in a “cap-like” configuration, and show that a ~8x8x3cm3 imaging region can be achieved with reasonable field gradient in depth and a mean

$\bar{B}_{0}$ of about

$64mT$ , proton freq. = 2.7MHz. We validate the magnet performance with 1D depth profile images from a multiple-disc phantom.

Methods

The starting point for the cap-shaped magnet was an equatorial portion of a “Halbach Sphere”⁹ . This magnetization pattern was discretized into 37 blocks, whose compositions and positions were then optimized using a genetic optimization in Matlab (Mathworks, Natick MA). Blocks were allowed to use one of 7 easily procurable material/size combinations varying from empty/non-magnetic through an N52 NdFeB material, 1”x1”x1.375” block. The genetic algorithm also had the ability to shift all blocks along

$\hat{x}$ by

$\pm 1cm$ , and to shift 6 blocks along

$\hat{y}$ by

$\pm 1cm$ . The cost function employed the percent variation in magnetic field over an ROI; mean

$|\bar{B}_{0}|$ was constrained to be at least

$50mT$ ; and magnet symmetry about the x-y and x-z planes was imposed. A hemi-ellipsoidal ROI with

$4cm$ major radii and a

$3cm$ minor radius was used. This ROI penetrates

$3cm$ into cerebral cortex (Figure 1) and roughly matches the excitation region of a loop Tx coil.

$|\bar{B}_{0}|$ maps were analyzed using COMSOL (COMSOL Inc, Burlington MA). After the optimal design was chosen, a former to hold the magnets was constructed with 3D printing (Formlabs Form2, Somerville MA). The prescribed NdFeB magnets (Applied Magnets, Plano TX), were then epoxied into the former. A

$|\bar{B}_{0}|$ map was acquired using a 3-axis Hall-effect magnetometer (Metrolab, Geneva, Switzerland) moved by a stepper robot. A , 5-turn Tx/Rx coil was tuned to

${f}_{c}=2.685 MHz$ and matched (

$BW (3dB)=150kHz$ ). The coil fit closely about a phantom containing three

$D=10mm$ ,

$h=5mm$ discs of 0.09% Gd-DPTA solution spaced

$5mm$ apart (Figure 5A). The unshielded phantom and coil were placed in the sensitive ROI of the magnet, and data were acquired using a hard pulse TSE sequence (

${f}_{c}=2.69MHz$ ; echo train length = 6;

${N}_{ave}=64$ ; 128 samples;

$BW=1221 \frac{Hz}{Px}$ ;

$TR=923ms$ ; pulse lengths

${t}_{90}$ /

${t}_{180}$ =

$2\mu s$ /

$4\mu s$ ).

Results

The chosen design utilized blocks with 4 different material/size combinations (Figure 2A), and resulted in a

$11.3cm\times 22.5cm\times 21.8cm$ ,

$6.3kg$ magnet (Figure 2B). The necessary magnetic material cost under $450 (USD). The constructed magnet could be held and moved by hand, and fit on an 85^th-percentile adult male head phantom (Figure 3). Figure 4 shows simulated and measured field maps are shown in the x-y and y-z planes. The magnet’s field was

$67.5mT$ at the ROI center with a field range of

$4.77mT$ across the ROI in the simulations. For the constructed magnet, the corresponding center field and range were

$63.6mT$ (

$2.71MHz$ ) and

$4.40mT$ (

$187kHz$ ). The simulated built-in

$|\bar{B}_{0}|$ gradient along the center axis of the magnet varied from

$154\frac{mT}{m}$ to

$198\frac{mT}{m}$ as one moves away from the magnet. For the constructed magnet, the corresponding limits were

$88\frac{mT}{m}$ and

$174\frac{mT}{m}$ Figure 5 shows the acquired spectrum of the 3-disc phantom centered

$6cm$ from the bottom of the magnet “bowl”. Three lobes are visible in the projection, corresponding to three regions of water in the phantom. The local

$|\bar{B}_{0}|$ gradients calculated from these data vary between

$95\frac{mT}{m}$ (near the magnet) and

$143\frac{mT}{m}$ (far from the magnet). For this acquisition, this corresponds to a depth resolution between

$0.30mm$ and

$0.20mm$ . The three 1D projection lobes are of different heights, likely resulting from the phantom signal bandwidth (

$135kHz$ ) being near the Tx/Rx coil bandwidth (

$150kHz$ ).

Discussion

We have demonstrated a novel design of a low-cost, lightweight (<$450 (USD), 6.3kg) single-sided magnet for point-of-care 3D brain imaging. As the initial work towards 3D imaging, we have shown the ability of this magnet to perform high-resolution (0.2 to 0.3mm) depth profiling along the 2.5cm length of phantom. Our next steps will be to add single-sided gradient coils for phase encoding along the other two spatial dimensions and enable full 3D imaging.

Acknowledgements

NIH: 5T32EB1680, R01EB018976; Thomas Witzel for help with the Apollo console

References

1. Vaughan JT, Wang B, Idiyatullin D, Sohn S, Jang A, BelaBarra L, Garwood M. Progress Toward a Portable MRI System for Human Brain Imaging. In: ISMRM. ; 2016. p. 498.

2. Cooley CZ, Stockmann JP, Armstrong BD, Sarracanie M, Lev MH, Rosen MS, Wald LL. Two-dimensional imaging in a lightweight portable MRI scanner without gradient coils. Magn. Reson. Med. [Internet] 2015;73:872–83. doi: 10.1002/mrm.25147.

3. Jackson JA. Remote NMR well logging. Geophysics 1981;46:415.

4. Goga NO, Demco DE, Kolz J, Ferencz R, Haber A, Casanova F, Blu B. Surface UV aging of elastomers investigated with microscopic resolution by single-sided NMR. 2008;192:1–7. doi: 10.1016/j.jmr.2007.10.017.

5. Judeinstein P, Ferdeghini F, Oliveira-silva R, Zanotti J, Sakellariou D. Low-field single-sided NMR for one-shot 1D-mapping : Application to membranes. J. Magn. Reson. [Internet] 2017;277:25–29. doi: 10.1016/j.jmr.2017.02.003.

6. Casanova F, Perlo J, Blümich B. Single-Sided NMR. 1st ed. Berlin Heidelberg: Springer-Verlag; 2011.

7. Perlo J, Casanova F, Blümich B. 3D imaging with a single-sided sensor: An open tomograph. J. Magn. Reson. 2004;166:228–235. doi: 10.1016/j.jmr.2003.10.018.

8. Landeghem M Van, Danieli E, Perlo J, Blümich B, Casanova F. Low-gradient single-sided NMR sensor for one-shot profiling of human skin. J. Magn. Reson. [Internet] 2012;215:74–84. doi: 10.1016/j.jmr.2011.12.010.

9. Leupold H, Potenziani II E. Novel High-Field Permanent-Magnet Flux Sources. IEEE Trans. Magn. 1987;MAG-23:3628–3629.

Figures

Figure 1: Magnet geometry, sensitive ROI, and simulated B₀ map overlain on an adult brain MRI acquired on a high-field scanner. The sensitive region of the magnet extends 3cm deep into cortex from the epidural surface.

Figure 2: (A) Magnet design; colors indicate size/material combinations for magnet blocks; black arrows indicate the direction of magnetization; the shaded red-outlined region is the optimization ROI. (B) Selected parameters characterizing the size, strength, and cost of the magnet.

Figure 3: (A) The hand-held potential of the constructed magnet is shown, showing the concave region intended for the head. (B) Magnet shown on an 85^th-percentile male adult head phantom, demonstrating that it fits as desired by analogy with Figure 1.

Figure 4: (A) Simulated (top) and measured (bottom) B₀ maps in the x-y plane at z=0. Note the different scales for (A) and (B). The measured B₀ map corresponds to the white dashed region on the simulated map. (B) Simulated (top) and measured (bottom) B₀ maps in the y-z plane at x=1cm and x=2.5cm. In (A) and (B) the optimization ROI is indicated by the black dashed line. Different color scales are used to elucidate the similar morphologies of the simulated and measured field maps.

Figure 5: (A) Tx/Rx coil and phantom. The three blue regions contain a dyed water/Gd-DPTA solution. Labeled ticks are in units of cm (B) Acquired 1D projection of the phantom onto frequency.

Proc. Intl. Soc. Mag. Reson. Med. 26 (2018)

0943