REVIEW ARTICLE

Low-Cost and Portable MRI

Lawrence L. Wald, PhD,1,2,3* Patrick C. McDaniel, MS,1,4 Thomas Witzel, PhD,1,2
Jason P. Stockmann, PhD,1,2 and Clarissa Zimmerman Cooley, PhD1,2

Research in MRI technology has traditionally expanded diagnostic benefit by developing acquisition techniques and instru-
mentation to enable MRI scanners to "see more." This typically focuses on improving MRI’s sensitivity and spatiotemporal
resolution, or expanding its range of biological contrasts and targets. In complement to the clear benefits achieved in this
direction, extending the reach of MRI by reducing its cost, siting, and operational burdens also directly benefits healthcare
by increasing the number of patients with access to MRI examinations and tilting its cost–benefit equation to allow more fre-
quent and varied use. The introduction of low-cost, and/or truly portable scanners, could also enable new point-of-care and
monitoring applications not feasible for today’s scanners in centralized settings. While cost and accessibility have always
been considered, we have seen tremendous advances in the speed and spatial-temporal capabilities of general-purpose MRI
scanners and quantum leaps in patient comfort (such as magnet length and bore diameter), but only modest success in the
reduction of cost and siting constraints. The introduction of specialty scanners (eg, extremity, brain-only, or breast-only scan-
ners) have not been commercially successful enough to tilt the balance away from the prevailing model: a general-purpose
scanner in a centralized healthcare location. Portable MRI scanners equivalent to their counterparts in ultrasound or even
computed tomography have not emerged and MR monitoring devices exist only in research laboratories. Nonetheless,
recent advances in hardware and computational technology as well as burgeoning markets for MRI in the developing world
has created a resurgence of interest in the topic of low-cost and accessible MRI. This review examines the technical forces
and trade-offs that might facilitate a large step forward in the push to "jail-break" MRI from its centralized location in
healthcare and allow it to reach larger patient populations and achieve new uses.
Level of Evidence: 5
Technical Efficacy Stage: 6
J. MAGN. RESON. IMAGING 2019.

W HILE ANY EXPANSION of a healthcare technology

must be viewed under the lens of cost-effectiveness
analysis,1,2 the potential benefits of expanding the accessibility
While accessibility issues are intertwined with other eco-
nomic, social, and technical problems, they are likely to per-
sist as until the relative cost of MRI is reduced relative to
of magnetic resonance imaging (MRI) technology to the other healthcare technologies.
world’s patient population is clear. The range and depth of Because the data compiled by Geethanath and Vaughan
geographical MRI accessibility gaps has been recently clearly spell out the depth of the worldwide accessibility
reviewed by Geethanath and Vaughan, who also review tech- variance,3 our review focuses on assessing progress and pros-
nology development efforts designed to mitigate this prob- pects for cost reduction. Although many factors play a role in
lem.3 There are regions of the world that clearly lag MRI examination costs, or even the marginal cost of adding a
worldwide averages in number of scanners or even any scan- scanner, we focus mainly on the equipment costs and the
ners at all, although there is some evidence that low-income major infrastructure and maintenance cost repercussions of
regions overspend on expensive technology relative to other the equipment. Ongoing steady but gradual cost reductions
infrastructure.4 The demand problem is not limited to devel- of today’s designs will result from the commoditization of
oping countries. Insufficient scanner density in "developed" components, but we focus instead on the potential for radical
countries still results in wait-periods prior to diagnostic care.5 changes in cost and portability. We partition the MRI

View this article online at wileyonlinelibrary.com. DOI: 10.1002/jmri.26942

Received Jun 6, 2019, Accepted for publication Sep 4, 2019.

*Address reprint requests to: L.W., 149 13th St. Rm. 2301, Charlestown, MA 02129. E-mail: wald@nmr.mgh.harvard.edu

From the 1Athinoula A Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital, Charlestown, Massachusetts, USA;
2
Harvard Medical School, Boston, Massachusetts, USA; 3Division of Health Sciences and Technology, Harvard – Massachusetts Institute of Technology,
Cambridge, Massachusetts, USA; and 4Department of Electrical Engineering and Computer Science, Massachusetts Institute of Technology, Cambridge,
Massachusetts, USA

© 2019 International Society for Magnetic Resonance in Medicine 1

Journal of Magnetic Resonance Imaging

scanner into its component groups: 1) the magnet together

with siting considerations, cryogenic systems, and power sup-
plies (if required); 2) the image encoding unit (ie, the gradi-
ent coil and drivers in a traditional scanner); 3) the RF
subsystem including transmit amplifier and coil and receive
coils; and 4) the console (small signal components) and com-
putational system.
We do not provide a detailed accounting of component
costs, as this requires detailed specifications, but attempt a
general overview with a focus on areas of technological inno-
vations/barriers and trends that might significantly reduce
cost. We focus on the cost per scanner, not the cost per scan,
thus omitting some important strategies to reduce diagnostic FIGURE 1: Marie Curie circa 1915 shown with one of her mobile
costs such as increasing patient throughput by increased x-ray units used during WWI. Source: http://www.nobelprize.
imaging speed and workflow advances. Such innovations org/nobel_prizes/physics/articles/curie/images/c_truck.jpg.
might allow a high-cost scanner to economically outperform a
low-cost unit. Finally, we try to separate evolutionary trends portable/POC solutions where some special infrastructure is
by which the canonical scanner architecture is simply refined provided in the room.
from more radical design changes. The latter will likely In this article we make the case that the challenges of
impose limitations on how the images are acquired, for exam- achieving portable and low-cost MRI are rooted in the insen-
ple, limiting the type of images, spatial resolution, and signal- sitivity of MRI, which is typically addressed by using high-
to-noise ratio (SNR). These are important trade-offs that can field, highly homogeneous magnets. The associated bulk and
only be superficially evaluated in this review. Commonly dis- complexity of such magnets, together with their associated
cussed strategies such as reduction of the B0 field below that cryogenic subsystems, creates barriers to lowering cost and
of the market dominating high-field systems (1.5T and 3T) operation as a portable or POC device. Additionally, issues
fall into this category, as do systems employing inhomoge- arise from the safety aspects of the fringe fields and the large
neous or prepolarized B0 field systems. The trade-offs gradient current and heat dissipation and the need to fit
imposed by these steps constitute a major change in the way within conventional "office" space, power, and cooling infra-
MRI is acquired, but are likely needed to significantly expand structure. To achieve all this, the contemporary use-model
MRI’s use into portable or point-of-care (POC) applications, where one scanner is expected to address all possible MRI
or for moving from diagnostics to monitoring. Some examinations might need to be reexamined. Thus, the goal
approaches, such as low-field MRI have been recently becomes low-cost portable scanners that can achieve a useful
reviewed.6 range of clinical services, but perhaps not all. Before exploring
the detailed prospects for dramatically lowering cost and/or
creating a portable or POC MRI, it is valuable to examine
Lessons From Other Modalities
other modalities that have already traversed this path.
Transition from hospital-centric to portable diagnostic devices Although the case-study of portable and POC ultrasound
is nearly as old as radiology itself. The x-ray was discovered in (US) and computed tomography (CT) is informative, it is an
late 1895 and was used clinically within a year. It became a open research question how to achieve the same goal
mobile imaging technology in 1914 when Marie Curie with MRI.
pushed to develop a fleet of x-ray-equipped ambulances for
use in World War I to assist surgeons in field hospitals (see
Fig. 1). By comparison, 45 years after MRI’s introduction we Portable US
still lack a truly portable MRI. Here we define "portable" as a US is the obvious example of a fully portable and POC diag-
device that can be moved room-to-room by a single nurse for nostic imaging modality. It is now readily commercially avail-
use on an individual patient within minutes of arriving. In able in a range of sizes and capabilities, ranging from trolley/
the truly portable scanner case, the power and cooling infra- cart-sized devices found in centralized radiology facilities, to
structure of a typical office or exam room must be sufficient. laptop sized "compact" units, to hand-held POC scanners.7
This contrasts with currently existing "mobile" MRI scanners Compact US units comprised 29% of the units shipped in
inside truck trailers that require a road, dedicated power, and the $6.6 billion USD ultrasound market of 2017, while
cooling infrastructure waiting at the new site, and a minimum handheld devices comprised 3%. Nonetheless, handheld
of tens of hours of setup prior to imaging at the new location. devices were the segment with the largest growth (IHSMarkit
Note, there are also solutions between conventional and true analysis).8 The MRI market is smaller and, no doubt, not a

2
 Wald et al.: Low-Cost and Portable MRI

direct analogy, but the success of the multiple US scanner a second recent technology with potential to extend the reach
sizes and portability ranges delivers several messages. The MR and accessibility of MR. Single-sided depth profilers such as
community can learn several lessons from US. First, all scan- the NMR-Mouse use a built-in field gradient outside the
ners are not required to be equal in their capabilities; the mar- magnet to provide a 1D depth profile from the Fourier trans-
ket is capable of deciding when to use which scanner; form of the recorded MR signal.29 MR depth profilers have
effectively trading-off imaging capability for portability and been used to analyze artwork,30 burn depth,31,32 skin
cost. Second, it is likely that the range of devices significantly layering,33–35 and extended to 2D36 and 3D imaging.37 Simi-
extends the reach of diagnostic US.7 Finally, a significant lar devices operate in the oil-well logging industry, where
equipment price drop (>10×) and extreme portability seems outward-looking NMR spectrometers lowered into the bore-
to be driving the successes of portable US. It remains to be hole record T2 and diffusion information that informs opera-
seen if MR can match these achievements tors of the properties of the porous rock in the well.38 Recent
work has demonstrated 1D, 2D, and 3D imaging over a lim-
Portable CT ited region in a brain-sized instrument,39 and a commercial
Portable CT scanners, although not as small or ubiquitous as low-field single-sided prostate imager is under development
their US counterparts, are readily commercially available.9 for use in urology clinics.40
They bring CT scanning to interventional and interoperative
suites,10,11 intensive care units,12,13 emergency departments,14
and ambulances.15,16 Although they have shielding chal- Cost of an MRI Scanner
lenges, they can operate with standard power and cooling Before reviewing the prospects for lowering the costs of the
infrastructure. However, the use of CT, especially in pediat- subsystems found in MRI scanners, or assessing alternative
rics, may carry risks. In a 15-year retrospective study of pedi- MRI acquisition strategies, it is worth reviewing the goal set
atric CT scans, Miglioretti et al estimate that the ~4 million by conventional 1.5T scanners. A simple-sounding concept
pediatric CT scans will cause 4870 additional cancers in the such as "cost" is surprisingly multifaceted.41 We consider pri-
U.S., although simple dose reduction might significantly marily the potential purchase price of the scanner itself,
decrease this.17,18 The success of portable CT, together with knowing full-well that considerable additional cost factors are
the complementary diagnostic value possible from MRI, sug- at play. For this purpose, we consider the low-end, but highly
gests an MRI scanner with similar portability could find a role functional end of the 1.5T superconductor based MRI seg-
in these settings now uniquely occupied by CT scanners. ment as well as low field (0.2T to 0.35T) clinical vertical-field
"open" systems. It is important to note that as the initial pur-
Desktop and Single-Sided NMR Spectrometers chase price of these systems is reduced, siting, infrastructure,
Although not a medical diagnostic tool, the recent advent of and operating costs are also relevant.
"desktop" nuclear magnetic resonance (NMR) spectrometers19–22 Unfortunately, most market information is proprietary.
merits discussion as a technological close-cousin of MRI. For A 2011 report by the UK National Health Service reported
many years, the NMR spectrometer market was nearly exclusively that they paid an average of $1.4M USD (in the exchange
the domain of superconducting solenoid-based magnet systems, rate at the time) to purchase and site each of the 267 new
usually operated in a central NMR core service facility. Thus, the MRI scanners it installed in the previous 10 years (predomi-
NMR spectrometer world of 10 years ago was analogous to the nantly 1.5T scanners).42 A 2010 study of 28 Belgian hospitals
clinical MRI world today. However, recent advances in perma- found that the average initial cost for five new 1.5T scanners
nent magnet design,23–25 and the sophistication of a field pro- installed in 2006–2008 was $1.5M USD (in the exchange
grammable gate array (FPGA) and similar electronics has allowed rate at the time).43 New "low-end" (but fully functional and
the deployment of low-cost, easy to site, bench-top systems with general purpose) 1.5T scanners can reportedly be currently
footprints of less than 1 m2.26 These systems do not provide the purchased for between $600,000 and $800,000 USD.44
spectral resolution or sensitivity offered by ultrahigh-field sup- Therefore, a likely cost target for a quantum leap in afford-
erconducting magnet systems, but allow the chemist to analyze a ability is likely a factor of 2× below this; more if the scanner
range of samples at the bench, without the trouble and expense of has reduced applicability or performance. Many MR
using the considerably costlier high-end systems. They also allow researchers think of equipment costs in terms of a "parts-
NMR to be more easily incorporated as a monitoring device in cost." For medical equipment, this is typically 4× to 5× lower
the chemical reaction process. Bench-top spectrometers have been than the purchase price. Therefore, we are looking for a solu-
applied to food analysis,27 interoperative flow,22 and exist as "sin- tion with a hardware parts cost of below $75,000 USD. Sit-
gle-sided" devices, allowing NMR spectra to be recorded for sam- ing costs become a significant factor, with RF-shielded rooms
ples outside the magnet.21,28 costing as much as $100,000 USD. Additional siting costs
The creation of successful MR depth-profilers that mea- stem from installing helium quench vents, large electrical
sure a 1D image of samples external to the magnet constitutes feeds, and cooling water.

3
Journal of Magnetic Resonance Imaging

The "parts costs" of the individual components of the Magnets

MR system (such as magnet, RF amplifier, etc) are considered Superconducting Magnets
proprietary information by manufacturers and are not readily Persistent mode superconducting solenoid designs and vari-
available. While it is possible for individual researchers to pur- ants comprise the vast majority of MRI magnet systems. Ben-
chase these items, what is available is generally more general- efits include stability, homogeneity, and potential for high
purpose laboratory-grade versions from companies catering to magnetic field generation.45 Wire costs dominate the magnet
that market and not mass-produced components used by the construction costs and liquid helium (LHe) and refrigeration
clinical companies. For example, we can have some idea what system (cold-head) maintenance dominate magnet operating
a magnet company would charge for a single laboratory use costs. Siting and portability is compromised by the size and
1.5T magnet, but the incremental cost of a magnet in a batch weight of the magnet (typically on the order of 3000–6000
of 1000 is presumably much cheaper. Despite the difficulty kg for a 1.5T actively shielded whole-body magnet),46 a rela-
of obtaining data on these proprietary component costs, tively large magnetic footprint even for actively shielded sole-
which also depends, of course, on exact specifications, Fig. 2 noid designs, the fragile nature of the cryogenic system, and
estimates the relative component group costs for a low-end the infrastructure for the LHe quench vent.
1.5T superconducting magnet scanner. The magnet cost is largely a function of wire type and
The total cost of ownership for the MR system includes length. For a given field strength, wire length is dictated by a
many additional factors beyond the equipment costs. Many design trade-off between the diameter of the homogeneous
of these likely have strong regional variation (such as building spherical volume (DSV), the magnet bore diameter (D), and
renovation and service/maintenance costs) and are not consid- the magnet length (L). The conductor placement is typically
ered here. We include discussion of such factors only to the optimized to achieve a minimum cost design with an accept-
extent that they can be impacted by system design choices, able trade-off between these three parameters (DSV, D, and
such as the potential to eliminate RF shielding, cryogen L). Xu et al looked at 100 constrained designs for a DSV
quench vents, cooling infrastructure, or specialized electrical with 1 ppm homogeneity.47 They show that for a given
feeds. homogeneity ppm specification (eg, 1 ppm), the problem can
be cast using only two independent variables (D/DSV) and
(L/DSV). Optimizing magnet cost (wire length) formed clas-
sic "l-curve" trade-offs between cost and the two variables
(D/DSV) and (L/DSV). Selecting the "knee" of the l-curve as
the design point for Lopt, they found a simple relationship for
cylindrical magnets: Lopt = 1.18 DSV + 0.77 D, and show
that modern designs adhere to this law. While magnets with
the desired homogeneity specifications can be designed with
shorter lengths (<Lopt) the wire cost rises exponentially (see
Fig. 3). They conclude that the best way to decrease cost is to
decrease bore size. This, in turn, puts pressure on the gradient
and RF engineers to cede territory.
Of course, lowering field strength will lower magnet
cost, as does relaxing homogeneity constraints. Both of these
will be examined below. Other ways to reduce cost with con-
ventional solenoid designs include finding an affordable wire
alternative with higher current density or reducing cryogenic
costs. New types of superconducting wires do not appear
often and copper-sheathed NbTi remains the mainstay, but
its 9.3K critical temperature requires 4K operation. In con-
FIGURE 2: Estimate of the relative component costs within a low-
end conventional 1.5T scanner. Costs are estimated for high-
trast, MgB2, ReBCO, YBCO, and BSCCO are potential
volume commodity production (runs of ~1000 per year) and entrants with higher critical temperatures,48 although the
would be considerably higher for one-off, or laboratory price per kAm increases dramatically. The wire cost of these
instrumentation. Note that a low-end system with a market price
high-temperature superconductors is such that it is currently
of about $800,000 USD and a 4× markup over components cost
suggests a total components cost of $200,000 USD leaving a more economical to operate them at low temperature where
cost of about $75,000 USD for the magnet and cryostat. Costs, their critical current is higher, thus reducing wire length.
of course, depend on the detailed specifications, and are In cryogenics, the recent decade has brought revolution-
treated as proprietary figures in the industry. Thus, these
estimates are based on the authors’ intuition and not ary advances in cryogenic cooling systems. Modern cold-
specific data. heads can now operate at temperatures of 4K with sufficient

4
 Wald et al.: Low-Cost and Portable MRI

will quickly warm above the critical temperature of the sup-

erconducting wire if the cryocooler stops operating. This
increases system vulnerability to power or cooling water loss.
The ride-thru time can be extended by increasing the cold-
mass; however, this would also result in longer startup/
quench recovery times.
Modern GM cryocoolers are cheaper pulse-tube
cryocoolers, but produce mechanical vibrations associated
with the well-known "washing-machine" sound heard in
every clinical scanner. In a traditional cryostat with LHe bath,
the cryocooler is mechanically uncoupled from the magnet
cold-mass; the cryocooler is connected only to the cryostat’s
radiation shields and a second stage to a recondenser plate
which reliquefies the rising gas in a zero boil-off design. GM
cryocoolers also require maintenance every ~2 years, a signifi-
cant operating cost. Pulse-tube cryocooler designs can be
FIGURE 3: Magnet wire cost as a function of magnet length L made without moving parts in the low-temperature subsec-
and bore diameter D (both expressed as a fraction of the
diameter of the homogeneous region, DSV) for superconducting
tion, offering mechanical stability and an improved mainte-
solenoid designs. This design analysis shows the rapid rise in nance life-cycle. There is a nearby moving piston, however,
cost for short magnets and constructing a homogeneous and unfortunately it usually contains a relatively strong para-
magnet shorter than Lopt quickly becomes expensive (a 1 ppm
magnet, potentially disrupting the B0 field. These problems
homogeneity in the DSV was used in this analysis). For all the
designs studied, the optimal length followed the relationship notwithstanding, it seems likely that conduction cooled clini-
Lopt = 1.18 DSV + 0.77 D. From Xu et al. In: Proc 7th Annual cal MRI magnets will become available.
Meeting ISMRM, Philadelphia; 1999. p 475.47 Reducing the diameter of superconducting magnets and
focusing on a specific body part reduces magnet cost size and
heat removal to allow modern clinical magnets to operate weight and has been applied to head, extremity, and neonatal
with "zero boil-off." Thus, although they contain 500 or scanners at conventional field strengths (1.5T and 3T). For
more liters of LHe, the cryocooler effectively recondenses the example, the GE neonatal 1.5T scanner has a 52-cm bore
evaporating liquid, eliminating the need to replenish LHe length and weighs 408 kg.51 A recent commercial effort has
under normal operation. This is fortuitous, since the world’s coupled these strategies in a small solenoid head-only,
helium supply, while apparently sufficient in principal, is conduction-cooled 1100 kg magnet at 0.5T.52 Another effort,
dominated by a small number of suppliers and has proven funded by the NIH Brain Initiative, utilizes a 1.5T asymmet-
highly volatile, with escalating costs49; a situation that is ric "hair-dryer" style design with relaxed homogeneity specifi-
unlikely to improve in the future. cations and high-temperature superconductors.53
The next step is a "nearly-dry" magnet using a suffi- Adjusting the standard superconducting solenoid design to
ciently small LHe volume to safely eliminate the quench vent permit true portability would pose several challenges if it was
system. The final step is the so-called "dry" or "conduction intended to be wheeled around while at field. Solenoid designs
cooled" magnet; a magnet containing no LHe. Multiple com- have large fringe fields, which would complicate movement
mercial small-bore dry magnets are available, and at least one logistics, and the design would have to be highly shielded or
nearly-dry clinical system. Modern MRI cryocoolers (possibly readily ramp-able. Extensive shielding with superconducting
with two per magnet) can achieve the heat removal needed windings leads to reduced efficiency, increased wire costs, and
for a dry or nearly-dry operation, utilizing either the Gifford- larger size (the shielding windings are typically large diameter).
McMahon (GM) cycle or a pulse-tube system50 to remove a Shielding with an iron yoke adds considerable weight. Movable
watt of heat at 4K. The truly dry mode is appealing since superconducting high-field magnets have been achieved over a
cryogens are not even needed at installation, and the LHe ves- limited movement range (such as the IMRIS interoperative sys-
sel can be omitted from the cryostat. But dry operation tem), but quench ventilation and cryogenic stability are challeng-
requires a direct heat path (mechanical connection) between ing. Because of these issues, true portability will likely require a
the cryocooler cold stage and the magnet thimble (thus the low-field or very low-field device where resistive magnets or per-
term "conduction cooling"). In this case, special care must be manent magnets lacking the cryogenic concerns become feasible.
taken to avoid mechanical vibrations from the cryocooler
from being directly transmitted to the magnet, which could Low-Field Permanent Magnets
induce field instabilities. Additionally, dry magnets are depen- Although low-field superconducting magnets are an option for
dent on continuous operation of the cryocooler. The magnet reducing cost and footprint, once the field is lowered into the

5
Journal of Magnetic Resonance Imaging

<0.3T range, the cryogenic system will likely dominate cost which have recently been reviewed.57 The Halbach cylinder is
and other cheaper, simpler options are typically turned to an attractive choice for MRI. RF engineers will recognize it as
including resistive magnets and permanent magnets. Resistive being the permanent magnet analog of a birdcage RF coil.
magnets incur heat dissipation challenges and add the cost of a Phasing the magnetization direction angle from 0 to 4π creates
stabilized power supply. Permanent magnet designs benefit a uniform field transverse to the cylinder axis.
from the stored energy in the magnetized material and power- For an idealized Halbach cylinder, the B0 strength is sim-
free and cryogen-free operation, although they require some ply related to the remnant magnetization (Br) of the material
form of temperature control to stabilize the B0 field. and the inner (ri) and outer (ro) radius of the cylinder: B0 = Br
Strong permanent "rare-earth" magnets have been available ln(ro/ri). Thus, B0 = 0.51 T for a head sized magnet with ro
since the early 1980s when the first sintered NdFeB rare-earth = 36 cm and ri = 25 cm using N50-grade NdFeB (Br = 1.4 T).
magnets were introduced.54 Early clinical low-field systems used a In practice, the fields are lower due to gaps between the dis-
permanent magnet dipole design with iron yoke flux return such cretized magnet segments and the finite cylinder length. An ide-
as the GE Signa Profile (0.2T), the Siemens Magnetom C! alized Halbach cylinder also has no flux density outside the
(0.35T), and the Hitatchi AIRIS (0.3T).55 Although known for cylinder, suggesting that these are very self-shielded designs.
being "open," the iron yokes and the rare-earth material itself made Figure 4 shows flavors of the Halbach cylinder. Useful variants
these systems heavier than 1.5T superconducting solenoid mag- include oppositely oriented NdFeB and SmCo Halbach units
nets. Nonetheless, with specialized applications and homogeneity in a configuration that cancels the temperature coefficient.23
relaxation, the magnet weight can be reduced to the portable level. Some form of permanent magnet array has formed the
For example, a 200 kg, 0.2T dipole permanent magnet with iron basis of several current small-bore (preclinical) systems
yoke has been recently operated in a minivan for imaging elbow designed for cost and easy siting, including the Aspect Imag-
injuries in baseball.56 ing M3 and M7 and the Bruker Biospin Icon 1T scanner.
Since these initial dipole-based low-field permanent MRI Aspect Imaging also makes a 510k cleared neonatal scanner
systems, a series of array-based designs have appeared following (Embrace) and wrist scanner (Wristview). Although the cost
the work of Halbach.24 Assembling the magnetized blocks into of rare-earth materials is volatile, the materials cost of a head-
arrays leads to a wealth of interesting configurations, many of sized magnet is relatively cheap (under $10,000 USD) and

FIGURE 4: Halbach cylinder designs of potential interest for low-field MRI. In the Halbach cylinder, a nearly uniform transverse field
is produced inside the cylinder if the magnetic moment of the magnetized material is phased from 0 to 4π azimuthally. Note that
this is similar to the phase relationship for a birdcage coil where a 0 to 2π azimuthal phase relationship is used. Top row shows an
ideal cylinder with continuous magnetization and a more practical approximation comprising keystone-shaped sections. Far top right
shows a simpler to construct configuration using only identical rectangular blocks and with all the magnetization vectors normal to a
face. The phase relationship comes only from rotations of the blocks. The bottom row shows further optimizations allowing degrees
of freedom to be adjusted to achieve a target field pattern (typically either a uniform field or a monotonic gradient) despite the
imperfections of the array (eg, finite cylinder and sparse population). One option is to maintain linear rungs of material but vary the
material. A second approach is to maintain rings of material but allow varying diameters.

6
 Wald et al.: Low-Cost and Portable MRI

the ability to choose where to place their resources. The low-

power electronics in the receive system may better commodi-
tize, allowing cost reduction from reduced gradient coil
performance with accompanying reductions in electrical
power and cooling infrastructure.
If the clinical use cases can be restricted, tailoring the
gradient design to specific clinical applications (such as
extremity or brain) is an attractive direction for reducing the
size, cost, and power needs. In applications where the size of
the coil can be reduced, the stored energy in the magnetic
field is greatly lowered, since this scales as the volume of the
gradient coil. The smaller coil delivers considerably more effi-
ciency (field gradient per ampere of current). Thus, a knee or
even head-only gradient is multifold more efficient than a
whole-body system, lowering gradient driver costs, conductor
cost, and cooling subsystem complexity.
FIGURE 5: A prototype portable brain MRI scanner based on the A second level of cost reduction could potentially come
Halbach permanent magnet described in Cooley et al (IEEE
Trans Magn 2018;54)58 and configured for rotational encoding
from building in a "permanent" gradient into the magnet and
as in Cooley et al (Magn Reson Med 2015;73:872–883).59 The utilizing that field for either readout or slice select. The
magnet weighs ~125 kg and achieves an 80 mT B0 field. MagneVu (Carlsbad, CA) wrist MRI used a 0.2T U-shaped
magnet weighing about 90 kg with a static gradient in the B0
can weigh as little as 120 kg.58 Several Halbach cylinders have operating as a slice select gradient.78–80 Recent work has used
been constructed with the goal of a low-cost brain MRI58–60 the built-in gradient of a head-sized Halbach magnet for read-
including with the use of magnet motion for image out encoding of a RARE spin-echo train sequence, including
encoding.58–64 Figure 5 shows a prototype 80 mT brain scan- rotating the magnet to acquire the additional radial projec-
ner based on an optimized Halbach array58 with built in gra- tions needed for a 2D image.59 The use of the built-in gradi-
dient for rotational projection encoding.59 ent for a readout requires a spin-echo to refocus the
magnetization but it eliminates the need for a switched gradi-
ent coil in this direction. Note that the readout gradient has a
Ultra-Low Field (ULF) Magnets
special role in MRI in that it must dominate the other gradi-
Further reduction of the B0 field below the ~20 mT lower
ents present, for example, from an inhomogeneous magnet.
limit of "low field MRI" results in further savings in magnet
Phase encode gradient blips surrounding the 180 refocusing
costs. Here, resistive electromagnets and very large field of
pulses in a spin-echo sequence do not have this requirement,
view or very light-weight permanent magnet configurations
since the spin-echo mechanism refocuses the magnet inhomo-
are possible.65,66 But challenges arise from limited SNR. The
geneity. In this way, phase encoding works even in very inho-
work on ULF systems has been elegantly reviewed by Kraus
mogeneous fields.36
et al.67 The SNR issues have been addressed by
Thus, for a system whose magnet cost has been reduced
prepolarization methods,66,68,69 SQUID, or atomic magne-
by relaxing homogeneity constraints, the use of a built-in gra-
tometer detection,70–72 the use of hyperpolarized media,73,74
dient and spin-echo-based sequence is attractive because it
or very efficient pulse sequences such as balanced steady-state
eliminates one of the three gradient coils and amplifier chan-
free precession (bSSFP).75 In some cases the "fix" to the SNR
nels. A second gradient channel can be eliminated by rotating
problem might be more costly than a higher strength magnet.
the magnet a small amount (1–3 ) per repetition time
Beyond the SNR problem, challenges arise from concomitant
(TR) period to achieve radial in-plane encoding.59 Silent or
gradient fields and the bandwidth of RF coils (even modest
near-silent operation is a side benefit of eliminating the large
Q tuned circuits have a narrower bandwidth than desirable).
switching readout gradient. Most of the desired MR contrasts
are readily available with spin-echo or RARE-type techniques,
Image Encoding (Gradient) System including proton density, T2, FLAIR prepped T2, and T1
Image encoding typically employs a 3-axis gradient system. (with an inversion prep). Even diffusion weighting can be
However, modern parallel imaging acquisitions omit a sub- seen. This is common in the well-logging industry where dif-
stantial fraction of gradient encoding steps in k-space and fusion weighting from the effects of diffusion in the perma-
make up for this with geometric information from the RF nent field gradient is uncovered by modulating the RARE
receive coil array.76,77 This trade-off between the gradient echo-train timing.38 However, gradient echo sequences such
and receive subsystems potentially affords system designers as the T2*-weighted images are not possible with a static

7
Journal of Magnetic Resonance Imaging

readout, but little T2* contrast is possible at low field not available for many parts of the MRI console. Although sev-
anyway. eral inexpensive ($2,000 to $10,000) MRI consoles have been
Finally, although shielded gradients were an enabling developed with other approaches,89–91 many of the MRI special-
technology to today’s high-field MRI scanners, significant ized RF signal chain components and signal processing modules
efficiency improvements and accompanying cost reduction can be implemented with a modern FPGA. Implementing fully
could be achieved by eliminating the shielding windings. The digital transmitters and receivers in FPGA hardware means that
shielding windings greatly reduced the eddy currents gener- less specialized analog electronics are needed. The console then
ated on the magnet bore and cold conducting structures essentially becomes a software project with substantially lower
within the cryostat and magnet. As such, cryogenic-free hardware development costs. For example, an open-source
designs are perhaps more amenable to this, since they lack FPGA MRI console based on the $349 USD STEMLab/Red
the cold, conducting structures. Other approaches might also Pitaya device has recently been described for controlling and
impact this problem, including improved preemphasis, acquiring spin-echo, gradient echo, single shot echo-planar imag-
dynamic field monitoring systems81 coupled with incorpora- ing, and single-shot spiral imaging.92,93 This system uses Xilinx
tion into image reconstruction.82 Zynq 7010 SoC, two 125 Msps 14-bit resolution ADCs, and
two 125 Msps 14-bit resolution DACs and 16 DIO pins con-
nected to the programmable logic. The entire pulse sequence
RF Subsystem, Console, and Computational execution and real-time control are performed by the integrated
System Dual-Core ARM9 CPUs running at a clock speed of 866 MHz
To a first approximation, the RF transmit and receive needs of and running software on an embedded Linux OS. The result is
portable and low-cost MRI are identical to conventional sys- two fully programmable transmit and two receive channels that
tems at that field strength, since high sensitivity and some directly synthesize and digitize between signals between DC and
degree of parallel imaging are likely still needed. We note that 40 MHz (suitable for low-field systems). The $349 price tag
POC use by less expert users would suggest an emphasis on underscores how the console of a low-cost system could be dom-
simplicity, and a specialized scanner (for a particular set of clini- inated entirely by software development costs.
cal applications) might not require the breadth of coils found in Another aspect of console cost for conventional clinical
a standard clinical MRI suite. The RF subsystem adds cost systems is the rather large software applications environment
mainly through the high power RF amplifier system needed for and the cost of development and maintenance of this soft-
spin excitation. Typically, >20 kW of peak power is needed, ware. This perhaps exceeds the cost of the hardware. Low-
but this scales with the excitation coil volume and can thus be field portable MRI systems are often designed for specific
greatly reduced for dedicated scanners. For example, only about purposes, and therefore should be able to operate with a
5 kW is needed for a head scanner and less for an extremity much smaller focused software environment, rather than pro-
scanner. Reducing the B0 field also increases the power effi- viding all the applications a general-purpose clinical MRI
ciency of excitation, since tissue losses dominate the power dis- device demands. Additionally, open-source sequence and image
sipation (above ~0.5T) and scale as the square of the frequency. reconstruction projects are underway94 and can potentially
Low-cost RF power amplifiers have been introduced to try to reduce software development costs. Finally, we note that some
support accessible MRI efforts, including an ~3000 € 1 kW of the strategies proposed for low-cost systems will require
open-source effort.83 Other approaches include efficient model-based iterative image reconstruction methods.95 This
switched mode modular RF amplifiers with the potential to be will likely entail retaining a full-power image-processing com-
placed directly on the RF transmit coil.84–86 puter or moving the image reconstruction to a cloud-based
The RF receive system costs can escalate with a large system.
number of parallel channels. To this end, commoditization of
these small-signal units could further reduce cost. Full-scale Other Components
integration efforts are underway to reduce an MRI preampli-
Other components, such as the patient table, depend more
fier and receiver (including ADC) into a single CMOS chip
on driving the cost out of manufacturing. RF coils tend to
for mounting on each receive coil element.87,88 These single-
have a high cost relative to component costs, which likely
chip MR receiver systems might aid the development of
reflects development costs.
hand-held (single-sided) MR devices, as well as their original
goal of creating wireless receive coils.
Conventional MRI consoles retain their high cost despite Potential Clinical Applications of Low-Cost,
continued advances and cost reductions in the electronics behind Portable, and POC MRI
them. The high cost results from nonrecurring development Broad and impactful applications are needed to drive the con-
costs and low-volume production of specialized hardware com- siderable technical effort required to significantly reduce cost
ponents in the console, as suitable off-the-shelf components are or create truly portable or POC MRI. The potential

8
 Wald et al.: Low-Cost and Portable MRI

directions are likely different for each of these three cases. 2. Singer ME, Applegate KE. Cost-effectiveness analysis in radiology.
Radiology 2001;219:611–620.
Low-cost MRI could be immediately useful in providing
3. Geethanath S, Vaughan JT Jr. Accessible magnetic resonance imaging:
access to all the many existing clinical applications of MRI A review. J Magn Reson Imaging 2019. Epub 2019/01/15. doi: https://
for those patients for whom insufficient scanner resources doi.org/10.1002/jmri.26638.
exist.3 A lower cost and the improved footprint associated 4. Lazaro P, Beerra A, Luengo S, Moteagudo JL, editors. Health care
with portability also allows MRIs to be sited in facilities nor- expenditures and expensive medical technology; the paradox of low-
income countries. In: Annual Meeting Int Soc Technol Assess Health
mally lacking MRI scanners, such as Emergency Depart- Care; 1994. p 31.
ments, ICUs, and recovery rooms, as well as smaller health
5. Emery DJ, Forster AJ, Shojania KG, Magnan S, Tubman M, Feasby TE.
clinics. A sufficiently inexpensive scanner does not need to be Management of MRI wait lists in Canada. Healthc Policy 2009;4:76–86.
as heavily utilized to recoup investment. Note that a low-cost 6. Marques JP, Simonis FFJ, Webb AG. Low-field MRI: An MR physics
device can add value even if it cannot perform all possible perspective. J Magn Reson Imaging 2019;49:1528–1542.
MRI examinations. 7. Moore CL, Copel JA. Point-of-care ultrasonography. N Engl J Med
Clinical applications for portable and POC devices are 2011;364:749–757.

harder to foresee. Since truly portable or POC MRI scanners 8. The global ultrasound equipment market in 2018: IHS Markit; 2018.
Available from: https://cdn.ihs.com.
are not yet commercially available, we can only speculate that
they could take over some of the applications currently served 9. Rumboldt Z, Huda W, All JW. Review of portable CT with assessment
of a dedicated head CT scanner. AJNR Am J Neuroradiol 2009;30:
by portable CT, where MRI can offer better soft-tissue con- 1630–1636.
trast or repeated x-ray doses are undesirable. The safety of a 10. Wallace MJ, Kuo MD, Glaiberman C, Binkert CA, Orth RC, Soulez G,
repeated application of MRI suggests that POC MRI devices Technology Assessment Committee of the Society of Interventional R.
might find applications more resembling patient monitoring Three-dimensional C-arm cone-beam CT: Applications in the interven-
tional suite. J Vasc Interv Radiol 2008;19:799–813.
than a traditional diagnostic role. The difficulty of intracranial
11. Gupta R, Cheung AC, Bartling SH, et al. Flat-panel volume CT: Funda-
monitoring with other modalities suggests a possible role for mental principles, technology, and applications. Radiographics 2008;
POC MRI. 28:2009–2022.

12. Peace K, Wilensky EM, Frangos S, et al. The use of a portable head CT
scanner in the intensive care unit. J Neurosci Nurs 2010;42:109–116.
Conclusion
13. Butler WE, Piaggio CM, Constantinou C, et al. A mobile computed
A series of technical design advances and system topology tomographic scanner with intraoperative and intensive care unit appli-
innovations are significantly shifting the landscape of what cations. Neurosurgery 1998;42:1304–1310; discussion 1310–1311.
the MRI of the future can look like. Based on the successes 14. Novelline RA, Rhea JT, Rao PM, Stuk JL. Helical CT in emergency radi-
other medical imaging modalities have gained by increasing ology. Radiology 1999;213:321–339.

accessibility and addressing POC markets, any morphology 15. Ebinger M, Winter B, Wendt M, et al. Effect of the use of ambulance-
based thrombolysis on time to thrombolysis in acute ischemic stroke: A
changes that MRI scanners can undertake are likely to pay off randomized clinical trial. JAMA 2014;311:1622–1631.
by extending benefits to existing patients and reaching new
16. Ebinger M, Grittner U, Audebert HJ. Ambulance-based thrombolysis in
populations of patients. Although no truly portable commer- acute ischemic stroke—Reply. JAMA 2014;312:961–962.
cial MRI scanner has been introduced to date, considerable 17. Smith-Bindman R, Miglioretti DL, Johnson E, et al. Use of diagnostic
work has been put into the feasibility of such a system, imaging studies and associated radiation exposure for patients enrolled
in large integrated health care systems, 1996-2010. JAMA 2012;307:
including light-weight, low-field extremity and brain magnets
2400–2409.
with the needed mobility. Further work in this direction is
18. Miglioretti DL, Johnson E, Williams A, et al. The use of computed
poised to provide insight into the clinical utility of such tomography in pediatrics and the associated radiation exposure and
systems. estimated cancer risk. JAMA Pediatr 2013;167:700–707.

19. Blumich B, Singh K. Desktop NMR and its applications from materials
science to organic chemistry. Angew Chem Int Ed Engl 2018;57:
6996–7010.
Acknowledgment
20. Danieli E, Mauler J, Perlo J, Blumich B, Casanova F. Mobile sensor for
Contract grant sponsor: National Institute of Biomedical high resolution NMR spectroscopy and imaging. J Magn Reson 2009;
Imaging and Bioengineering of the National Institutes of 198:80–87.
Health; Contract grant number: R01EB018976. 21. Perlo J, Demas V, Casanova F, et al. High-resolution NMR spectros-
The authors thank Matthew Rosen for many useful copy with a portable single-sided sensor. Science 2005;308:1279.

conversations about low field, ultra-low field, and low- 22. Perlo J, Silletta EV, Danieli E, et al. Desktop MRI as a promising tool for
mapping intra-aneurismal flow. Magn Reson Imaging 2015;33:
cost MRI. 328–335.

23. Danieli E, Perlo J, Blumich B, Casanova F. Highly stable and finely

References tuned magnetic fields generated by permanent magnet assemblies.
Phys Rev Lett 2013;110:180801.
1. Hutubessy R, Chisholm D, Edejer TT. Generalized cost-effectiveness
analysis for national-level priority-setting in the health sector. Cost Eff 24. Halbach K. Design of permanent multipole magnets with oriented rare
Resour Alloc 2003;1:8. earth cobalt material. J Nucl Instrum Methods 1980;69:1–10.

9
Journal of Magnetic Resonance Imaging

25. Raich H, Blumler P. Design and construction of a dipolar Halbach array 48. Parizh M, Lvovsky Y, Sumption M. Conductors for commercial MRI
with a homogeneous field from identical bar magnets: NMR man- magnets beyond NbTi: Requirements and challenges. Supercond Sci
dhalas. Conc Magn Reson 2004;23B:16–25. Technol 2017;30:014007.

26. Johns M, Fridjonsson EO, Voigh S, Haber A (eds.). Mobile NMR and 49. Anderson ST. Economics, helium, and the U.S. Federal Helium Reserve:
MRI, developments and applications. Cambridge, UK: Royal Society of Summary and outlook. Nat Resour Res 2018;27:455–477.
Chemistry; 2016.
50. Hirayama T, Morie T, Xu M. Experimental investigation of cooling
27. Voda MA, van Duynhoven J. Benchtop-NMR-food: Solid fat content capacity of 4K pulse tube cryocooler in magnetic fields. IOP Conf Ser:
determination and emulsion droplet sizing. In: Johns M, Fridjonsson EO, Mater Sci Eng 2018;502:012035.
Voigh S, Haber A, eds. Mobile NMR and MRI, developments and applica-
tions. Cambridge, UK: Royal Society of Chemistry; 2016. p 86–106. 51. Merhar SL, Tkach JA, Woods JC, et al. Neonatal imaging using an on-
site small footprint MR scanner. Pediatr Radiol 2017;47:1001–1011.
28. Perlo J, Casanova F, Blumich B. Ex situ NMR in highly homogeneous
fields: 1H spectroscopy. Science 2007;315:1110–1112. 52. Panther A, Thevathasen G, Connell IRO, et al. A dedicated head-only
MRI scanner for point-of-care imaging. In: Proc 27th Annual Meeting
29. Blumich B, Blumler P, Eidmann G, et al. The NMR-mouse: Construc- ISMRM, Montreal; 2019. p 3679.
tion, excitation, and applications. Magn Reson Imaging 1998;16:
479–484. 53. Servick K. Cheap, portable scanners could transform brain imaging. But
how will scientists deliver the data? Science 2019 [Epub ahead of
30. Blumich B, Casanova F, Perlo J, Presciutti F, Anselmi C, Doherty B. print].
Noninvasive testing of art and cultural heritage by mobile NMR. Acc
Chem Res 2010;43:761–770. 54. Tenaud P, Lemaire H, Vial F. Recent improvements in NdFeB sintered
magnets. J Magn Magn Mater 1991;101:328–332.
31. He Z, He W, Wu J, Xu Z. The novel design of a single-sided MRI probe
for assessing burn depth. Sensors 2017;17(3). 55. Miyamoto T, Sakurai H, Takabayashi H, Aoki M. A development of a
permanent magnet assembly for MRI devices using Nd-Fe-B material.
32. Koruda MJ, Zimbler A, Settle RG, et al. Assessing burn wound depth IEEE Trans Magn 1989;25:3907–3909.
using in vitro nuclear magnetic resonance (NMR). J Surg Res 1986;40:
475–481. 56. Nakagomi M, Kajiwara M, Matsuzaki J, et al. Development of a small
car-mounted magnetic resonance imaging system for human elbows
33. Van Landeghem M, Danieli E, Perlo J, Blumich B, Casanova F. Low- using a 0.2T permanent magnet. J Magn Reson 2019;304:1–6.
gradient single-sided NMR sensor for one-shot profiling of human skin.
J Magn Reson 2012;215:74–84. 57. Blumler P, Casanova F. Hardware developments: Halbach magnet
arrays. In: Johns M, Findjonson EO, Vogt S, Haber A, eds. Mobile NMR
34. Backhouse L, Dias M, Gorce JP, Hadgraft J, McDonald PJ, and MRI; developments and applications. Cambridge UK: Royal Soci-
Wiechers JW. GARField magnetic resonance profiling of the ingress of ety of Chemistry; 2016. p 133–155.
model skin-care product ingredients into human skin in vitro. J Pharm
Sci 2004;93:2274–2283. 58. Cooley CZ, Haskell MW, Cauley SF, et al. Design of sparse Halbach
magnet arrays for portable MRI using a genetic algorithm. IEEE Trans
35. McDonald PJ, Akhmerov A, Backhouse LJ, Pitts S. Magnetic resonance Magn 2018;54(1).
profiling of human skin in vivo using GARField magnets. J Pharm Sci
2005;94:1850–1860. 59. Cooley CZ, Stockmann JP, Armstrong BD, et al. Two-dimensional
imaging in a lightweight portable MRI scanner without gradient coils.
36. Casanova F, Blumich B. Two-dimensional imaging with a single-sided Magn Reson Med 2015;73:872–883.
NMR probe. J Magn Reson 2003;163:38–45.
60. O’Reilly T, Wouter T, Winter L, Webb A (eds.). Design of a homoge-
37. Perlo J, Casanova F, Blumich B. 3D imaging with a single-sided sensor: neous large-bore Halbach array for low field MRI. In: Proc 27th Annual
An open tomograph. J Magn Reson 2004;166:228–235. Meeting ISMRM, Montreal; 2019.
38. Hurlimann MD, Heaton NJ. NMR well logging. In: Johns M,
61. Blumler P. Proposal for a permanent magnet system with a constant
Findjonson EO, Vogt S, Haber A, eds. Mobile NMR and MRI; develop-
gradient mechanically adjustable in direction and strength. Conc Magn
ments and applications. Cambridge, UK: Royal Society of Chemistry;
Reson 2016;46:41–48.
2016. p 11–79.
62. Vogel MW, Guridi RP, Su J, Vegh V, Reutens DC. 3D-spatial encoding
39. McDaniel PC, Cooley CZ, Stockmann JP, Wald LL. The MR Cap: A
with permanent magnets for ultra-low field magnetic resonance imag-
single-sided MRI system designed for potential point-of-care limited
ing. Sci Rep 2019;9:1522.
field-of-view brain imaging. Magn Reson Med 2019 (in press).
63. Sarty GE. Cyclic generalized projection MRI. Magn Reson Imaging
40. Promaxo. MRI for live diagnosis and treatment in a physicians office
2015;33:304–311.
2019. Available from: https://promaxo.com/.
64. Ren ZH, Mu WC, Huang SY. Design and optimization of a ring-pair per-
41. Rubin GD. Costing in radiology and health care: Rationale, relativity,
manent magnet array for head imaging in a low-field portable MRI sys-
rudiments, and realities. Radiology 2017;282:333–347.
tem. IEEE Trans Magn 2019;55:1–8.
42. Managing high value capital equipment in the NHS in England. In:
65. Lother S, Schiff SJ, Neuberger T, Jakob PM, Fidler F. Design of a
Health Do, ed. London: National Audit Office, Stationery Office; 2011.
mobile, homogeneous, and efficient electromagnet with a large field
43. Obyn C, Cleemput I. The capital cost and productivity of MRI in a Bel- of view for neonatal low-field MRI. MAGMA 2016;29:691–698.
gian setting. JBR-BTR 2010;93:92–96.
66. Obungoloch J, Harper JR, Consevage S, et al. Design of a sustainable
44. Panel discussion. ISMRM Workshop on Accessible MRI; 2019. New prepolarizing magnetic resonance imaging system for infant hydro-
Dehli, India. cephalus. MAGMA 2018;31:665–676.

45. Lvovsky Y, Stautner EW, Zhang T. Novel technologies and configura- 67. Kraus RH, Espy MA, Magnelind PE, Volegov PL. Ultra-low field nuclear
tions of superconducting magnets for MRI. Supercon Sci Technol 2013; magnetic resonance, a new MRI regime. Oxford, UK: Oxford University
26(093001). Press; 2014.

46. Cosmus TC, Parizh M. Advances in whole-body MRI magnets. IEEE 68. Macovski A, Conolly S. Novel approaches to low-cost MRI. Magn
Trans Appl Supercond 2011;21:2104–2109. Reson Med 1993;30:221–230.

47. Xu H, Conolly SM, Scott GC, Macovski A (eds.). Fundamental scaling 69. Matter NI, Scott GC, Venook RD, et al. Three-dimensional prepolarized
relations for homogeneous magnets. In: Proc 7th Annual Meeting magnetic resonance imaging using rapid acquisition with relaxation
ISMRM, Philadelphia; 1999. p 475. enhancement. Magn Reson Med 2006;56:1085–1095.

10
 Wald et al.: Low-Cost and Portable MRI

70. Lee SK, Mossle M, Myers W, et al. SQUID-detected MRI at 132 microT 83. Blucher C, Han H, Hoffmann W, et al. (eds.). COSI transmit: Open
with T1-weighted contrast established at 10 microT–300 mT. Magn source soft-and hardware transmission system for traditional and
Reson Med 2005;53:9–14. rotating MR. In: Proc 25th Annual Meeting ISMRM, Honolulu; 2017.
p 1055.
71. Savukov IM, Zotev VS, Volegov PL, et al. MRI with an atomic magne-
tometer suitable for practical imaging applications. J Magn Reson 84. Gudino N, de Zwart JA, Duan Q, et al. Optically controlled on-coil
2009;199:188–191. amplifier with RF monitoring feedback. Magn Reson Med 2018;79:
72. Zotev VS, Matlashov AN, Volegov PL, et al. Microtesla MRI of the 2833–2841.
human brain combined with MEG. J Magn Reson 2008;194:115–120.
85. Gudino N, Duan Q, de Zwart JA, et al. Optically controlled switch-
73. Zotev VS, Owens T, Matlashov AN, Savukov IM, Gomez JJ, Espy MA. mode current-source amplifiers for on-coil implementation in high-field
Microtesla MRI with dynamic nuclear polarization. J Magn Reson 2010; parallel transmission. Magn Reson Med 2016;76:340–349.
207:78–88.
86. Twieg M, Griswold MA. High efficiency radiofrequency power amplifier
74. Tsai LL, Mair RW, Rosen MS, Patz S, Walsworth RL. An open-access, module for parallel transmit arrays at 3 Tesla. Magn Reson Med 2017;
very-low-field MRI system for posture-dependent 3He human lung 78:1589–1598.
imaging. J Magn Reson 2008;193:274–285.
87. Sporrer B. Integrated broadband receivers for magnetic resonance
75. Sarracanie M, LaPierre CD, Salameh N, Waddington DE, Witzel T,
imaging. Zurich, Switzerland: ETH Zurich; 2017.
Rosen MS. Low-cost high-performance MRI. Sci Rep 2015;5:15177.

76. Griswold MA, Jakob PM, Heidemann RM, et al. Generalized 88. Sporrer B, Wu L, Bettini L, et al. A fully integrated dual-channel on-coil
autocalibrating partially parallel acquisitions (GRAPPA). Magn Reson CMOS receiver for array coils in 1.5-10.5 T MRI. IEEE Trans Biomed Cir-
Med 2002;47:1202–1210. cuits Syst 2017;11:1245–1255.

77. Pruessmann KP, Weiger M, Scheidegger MB, Boesiger P. SENSE: Sen- 89. Tang W, Wang W, Liu W, et al. A home-built digital optical MRI con-
sitivity encoding for fast MRI. Magn Reson Med 1999;42:952–962. sole using high-speed serial links. Magn Reson Med 2015;74:578–588.

78. Duer-Jensen A, Vestergaard A, Dohn UM, et al. Detection of rheumatoid 90. Stang PP, Conolly SM, Santos JM, Pauly JM, Scott GC. Medusa: A scal-
arthritis bone erosions by two different dedicated extremity MRI units able MR console using USB. IEEE Trans Med Imaging 2012;31:
and conventional radiography. Ann Rheum Dis 2008;67:998–1003. 370–379.
79. Gold G, Theodorou D, Blair T, et al. (eds.). MR imaging of the wrist
91. Hasselwander CJ, Cao Z, Grissom WA. gr-MRI: A software package for
with a portable extremity scanner. In: Proc 7th Annual Meeting ISMRM,
magnetic resonance imaging using software defined radios. J Magn
Philadelphia; 1999. p 2039.
Reson 2016;270:47–55.
80. Crues JV, Shellock FG, Dardashti S, James TW, Troum OM. Identifica-
tion of wrist and metacarpophalangeal joint erosions using a portable 92. Anand S, Stockmann JP, Wald LL, Witzel T (eds.). A low-cost (<$500
magnetic resonance imaging system compared to conventional radio- USD) FPGA-based console capable of real-time control. In: Proc 26th
graphs. J Rheumatol 2004;31:676–685. Annual Meeting ISMRM, Paris; 2018. p 948.

81. Dietrich BE, Brunner DO, Wilm BJ, Barmet C, Pruessmann KP. Continu- 93. Ocra open MRI. Available from: https://openmri.github.io/ocra/.
ous magnetic field monitoring using rapid re-excitation of NMR probe
sets. IEEE Trans Med Imaging 2016;35:1452–1462. 94. Winter LB. Open source magnetic resonance imaging: From the commu-
nity to the community 2019. Available from: opensourceimaging.org.
82. Vannesjo SJ, Graedel NN, Kasper L, et al. Image reconstruction using a
gradient impulse response model for trajectory prediction. Magn Reson 95. Fessler J. Model-based image reconstruction for MRI. IEE Signal Proc
Med 2016;76:45–58. Mag 2010;27:81–89.

11