handheld MRIs

[Rohit Khare]

http://www.its.caltech.edu/~sciwrite/2000-01/ferguson.htm
[Core I science writing program at Caltech; reprinted in Engineering &=20=

Science, the alumni magazine, in 2001 issue 2]
[this article is missing from Caltech's website now, so I'm appending=20
the google-cache version... RK]

------------------------------------------------------------------------
The Promise of Portable MRI

John Ferguson

Mentored by Professor Michael Tyszka


Over the last 20 years, Magnetic Resonance Imaging (MRI) scans have=20
become a standard medical diagnostic tool. Every major hospital owns at=20=

least one scanner and normally has no problem scheduling it to capacity.=20=

MRI=92s popularity arises from the non-invasive nature of NMR technology=20=

and the versatility of measurements. However, the scanner requires a=20
large dedicated space, magnetic shielding, liquid nitrogen/helium=20
cooling systems, and complex electronics.


To find an alternative to the resource demands of an MRI scanner, a=20
group of German scientists have built a portable version, called an=20
NMR-MOUSE. Its capabilities are extremely limited, as the device is=20
still in its infancy. But with sufficient improvements, the NMR-MOUSE or=20=

a similar device could revolutionize the world of medicine. The progress=20=

of portable MRI technology is compared to the development of the=20
portable ultrasound scanner, which now enjoys immense success as a=20
non-invasive medical diagnostic device.



In the summer of 1881, America=92s 20th president, James A. Garfield, =
lay=20
on his bed slowly dying. Somewhere in his body was an assassin=92s =
bullet.=20
Over a period of 80 days, 16 different doctors and surgeons tried to=20
locate the bullet. They would poke fingers and metal probes into the=20
bullet hole without success. Even the famous American inventor,=20
Alexander Graham Bell, tried his hand in locating the bullet by creating=20=

a crude metal detector. After some time, he claimed he found the bullet,=20=

and the doctors rushed to operate on Garfield to excise it. However,=20
Bell did not realize that the president=92s bed contained metal springs.=20=

The doctors, of course, could not find the bullet, but created more=20
complications which ultimately led to the president=92s death.

If this happened fifteen years later, the doctors would have had no=20
difficulty finding the bullet. A new technology had been found which=20
revolutionized the world of medicine. The German physicist, Wilhelm=20
Roentgen, discovered x-rays by observing electric currents in a=20
partially evacuated glass tube. These rays, now known to be high-energy=20=

electromagnetic radiation, would pass through objects that light could=20=

not. Roentgen experimented with the tube observing metal and non-metal=20=

objects. He could even see through doors. But more amazing than that=20
were the results when he looked at his wife=92s hand. In this first=20
medical x-ray image, her bones and metal wedding ring could be seen=20
clearly.

X-rays are still very common today. But with the development of=20
computers, 3-D images can be reconstructed from signals received by=20
x-ray detectors that rotate around the subject=92s body. This technique,=20=

known as CT (Computed Tomography) or, more popularly, =93CAT scanning=94=20=

(for Computed Axial Tomography) was introduced in 1972 by Godfrey=20
Hounsfield, who was honored with won a Nobel prize in 1979 and a unit in=20=

his name, the Houndsfield. A modern X-ray CT can collect and reconstruct=20=

a high-resolution =93slice=94 of the body in half a second 1. Computers =
are=20
also used to create images for Positron Emission Tomography (PET) and=20
Single Photon Emission Computerized Tomography (SPECT). Both imaging=20
techniques involve injecting the body with a radioactive tracer and=20
monitoring the emissions as the radioactive substance travels around the=20=

body. However, overexposure to ionizing radiation and x-rays has been=20
shown to be dangerous and unhealthy.


Two alternative imaging techniques have become immensely popular due to=20=

their mildness and versatility: ultrasound and Magnetic Resonance=20
Imaging (MRI). These work through sound waves and magnetic fields=20
respectively, which are safe for most people (unless they have a=20
pacemaker, cochlear implants, or aneurysm clips when using MRI). Doctors=20=

can prescribe these diagnostic tests as frequently as desired without=20
concern about the patient=92s long-term health. However, the largest=20
drawback of these two modalities is their size and their price. A common=20=

ultrasound scanner can be wheeled bed-to-bed around the hospital on a=20
cart and costs $150,000 while a MRI scanner needs a shielded room=20
dedicated to MRI and a cryogenic cooling system to operate a=20
superconducting magnet and costs about two million dollars.

But on the horizon, two new products show the next direction of medical=20=

imaging: portable, non-ionizing, imaging devices at a fraction of the=20
current prices. A highly portable ultrasound scanner is already being=20
sold; a promising hand-held MRI scanner is under development. If=20
successful, the MRI scanner could revolutionize medicine by creating an=20=

incredible medical tool. We would have a device very similar to a=20
medical tricorder from Star Trek; it could measure anything, anywhere.


MRI history
But to see how this hand-held device works and to realize its=20
possibilities and limitations requires a good understanding of the=20
history of the technology that led to its current state. Thousands of=20
years ago, people in China and Greece independently discovered=20
lodestones and used them for fortune telling and navigation. These=20
useful but seemingly magical motions of the lodestone were not=20
understood until the 19th century. In 1802, the Italian scientist=20
Romagnosi found that a lodestone would move when a nearby wire conducted=20=

current; unfortunately, this result lay unnoticed for nearly twenty=20
years until =D8rsted published the finding. Suddenly the floodgate was=20=

opened for scientific discovery. Countless people contributed findings=20=

to further the field of electromagnetism including Ampere, Weber,=20
Faraday, Hertz, Curie, and even the American patriot Ben Franklin. In=20
1873, Maxwell combined all of these results into four simple equations=20=

that can be summarized as follows: electricity and magnetism are=20
inescapably tied together. A moving charge (i.e., current) creates a=20
magnetic field; a changing magnetic field induces current. The picture=20=

seemed complete and fully understood.


However, as in all of physics, the 20th century uprooted many firmly=20
established ideas giving rise to new possibilities. Quantum Mechanics,=20=

as described by Pauli in 1924, gave a more complete explanation for the=20=

nature of electricity and magnetism. Nuclei of atoms display a property=20=

called spin, which is in a sense small-scale angular momentum. Normally,=20=

the orientation of these spins is quite random in a material. However,=20=

in the presence of a magnetic field, the nuclear spins line up parallel=20=

and antiparallel to the field, with a tiny excess pointed parallel (a=20
few nuclei per million). It is these few spins that can be detected, and=20=

the way they react to changing magnetic fields gives rise to the=20
predecessor of MRI, Nuclear Magnetic Resonance (NMR).


The official birthday of NMR was in 1946 when two American teams led by=20=

Bloch and Purcell independently discovered that by adding another,=20
smaller field to the original, larger magnetic field interesting results=20=

would follow. By adding the second oscillating field, known as a=20
radiofrequency (RF) pulse, at the proper frequency for a short period of=20=

time, some of the nuclei would absorb the energy (resonate). After the=20=

RF pulse turned off, the nuclei would try to return to their original=20
energies (relaxation) and give off a signal that provides a great deal=20=

of information. These would include details of the chemical composition=20=

and density, movement within the sample, and with the use of a third,=20
spatially varying magnetic field (produced from gradient coils),=20
location. Answers to what, how much, when, and where would be given=20
without touching or damaging the sample. NMR remains to this day a=20
premier instrument to investigate chemical structures. But it wasn=92t =
for=20
twenty years after the discovery of NMR that the medical diagnostic=20
possibilities were realized.

A short time after Bloch discovered NMR, he put his finger inside the=20
magnet and noticed a strong output signal. Unfortunately, he did not=20
pursue this result any further. It wasn=92t until 1972 when a paper by=20=

Damadian, showing that tumors could be distinguished from normal tissue=20=

by relaxation times, woke the world up to the possibilities of medical=20=

NMR. Another scientific floodgate opened. Numerous people started=20
noticing all sorts of medically relevant phenomena that could be=20
detected with NMR. In 1978, an imaging device was created based on the=20=

differences in relaxation times of different tissues. Instead of naming=20=

the new device =93Nuclear Magnetic Resonance Imaging=94, the word =
=93nuclear=94=20
was dropped due to its negative connotations with nuclear warfare and=20
nuclear radiation and MRI was born. Now, MRI scanners can be found in=20
most hospitals and are frequently used to get detailed images of the=20
body that also contain functional information. Since the first device in=20=

1978, technological improvements in MRI have been true triumphs of both=20=

science and engineering.


The main magnet, the most vital component of any MRI system, creates=20
enormous design demands. The most common MRI scanner found in hospitals=20=

has a field strength of 1.5 Tesla, 30000 times the earth=92s magnetic=20
field. To get such high field strength, a superconducting magnet is used=20=

that requires a sophisticated cooling system made from layers of Liquid=20=

Helium and vacuums. Precautions must be taken against a quench, a=20
violent expansion of the helium due to insufficient cooling, and a safe=20=

release system in case the quench occurs. Also, shielding must be=20
constructed to protect neighboring rooms (and floors) from the strong=20
magnetic field which can erase credit cards, affect computer memory and=20=

displays, and also kill people with pacemakers. This can involve about=20=

20 tons of material to block the fields and adding additional magnets=20
that reduce the outside fields sacrificing some of the desired field=20
strength inside the magnet. Other systems that involve complicated=20
design are the shim system, to create magnetic fields that are=20
homogenous; gradient systems to give better spatial resolution and=20
faster imaging; RF systems that would resonate the nuclei within given=20=

specifications. All of these issues contribute to the overwhelming cost=20=

of an MRI system, averaging $2 million for a new 1.5 Tesla installation=20=

with about $300,000 in yearly maintenance. In addition, many people=20
believe that the best way to improve the MRI scanner is by increasing=20
the field strength to 3 or 4 Tesla. This will improve resolution and=20
scan time but with a significant increase in size, complexity,=20
protective equipment, and, of course, price.

The NMR-MOUSE
Enter the NMR-MOUSE (MObile Universal Surface Explorer). The hand-held,=20=

one-kilogram unit holds promise as a portable MRI device for less than=20=

$1000. Its design is very simple: two antiparallel magnets held apart=20
with a block of iron and a RF coil and two gradient coils in the gap=20
between them. The two magnets are made of a rare-earth metal which=20
generate a high magnetic field strength for their size; the yoke serves=20=

to increase the field strength. The RF and gradient coils serve the same=20=

purpose as in a normal MRI scanner, to stimulate nuclei within a=20
specified region. However, the most important feature in the NMR-MOUSE=20=

is the absence of extra equipment found in a typical scanner. There are=20=

no shim coils, no shielding, and no cryogenic cooling system since two=20=

permanent magnets are used instead of a superconducting magnet. This=20
resulted in an entirely different philosophy from mainstream MRI, namely=20=

using smaller field strength, exploiting the stray field as the main=20
field, and converting the =93disadvantage=94 of inhomogeneity (where the=20=

magnetic field is not perfectly uniform) into an advantage.

The inventors of the NMR-MOUSE, Bl=FCmer and Bl=FCmich, began the design =
and=20
construction of the NMR-MOUSE in 1993 with two simple realizations:=20
=93typical MRI contrast (relaxation, diffusion) doesn=92t rely on=20
homogeneous fields=94 and =93localization procedures imply inhomogeneous=20=

fields=94. Instead of using the traditional approach of creating a=20
completely homogenous field and adding systemic variations, they found=20=

that that a naturally inhomogenous field with consistent responses would=20=

work as well. Unknowingly, they had just entered the world of=20
fringe-field NMR. The first few magnets that were used in NMR had=20
significant inhomogenities; scientists designed their experiments around=20=

this fact. However, as technology improved, the vast majority of=20
researchers looked to bigger and more homogenous magnets. Even so, a few=20=

people, most notably Jasper Jackson, did research with smaller magnetic=20=

fields, down to just the earth=92s natural magnetic field, and strange=20=

combinations of magnets that would create regular fields outside of two=20=

magnets. These devices found applications in detecting signals in=20
dangerous places, like using the magnet through a wall to peer into a=20
dangerous room and well logging (dropping an NMR device down a very deep=20=

bore hole to make moisture measurements and detect oil in rocks miles=20
below the surface). These instruments were providing solutions that the=20=

larger, superconducting NMR spectrometers could never attempt. So, what=20=

seemed like an original idea by Bl=FCmer and Bl=FCmich to use a small =
system=20
to make measurements outside of the main field was actually closely=20
related to work done 40 years before. Nevertheless, with the use of more=20=

powerful rare-earth magnets and efficient computer-aided designs, new=20
possibilities have opened up, most notably in imaging.


The NMR-MOUSE brings to the table a small yet potent design. With=20
dimensions of 9cm x 2cm x 2cm and weight of 1.25 kg, it is truly=20
handheld. To operate it only needs to be connected to a computer with a=20=

cable, so it is highly portable. The NMR-MOUSE can image a sensitive=20
region of 16cm x 1cm x 2cm in the shape of an ellipsoid. At the surface,=20=

a field strength of .5 Tesla can be found. Incredibly, as an imaging=20
tool, it can scan with 100-micron (0.1 mm) resolution. However, the=20
NMR-MOUSE=92s biggest limitation is penetration depth; it can only image=20=

objects very close by. Currently, the maximum penetration depth is 5mm=20=

with closer objects generating stronger signals. This creates a large=20
restriction on the types of objects the NMR-MOUSE can image and its=20
general applications.

Even so, the NMR-MOUSE has found both traditional and novel uses. The=20
inventors, Bl=FCmer and Bl=FCmich, see applications in non-destructive=20=

materials testing, process control, agriculture, food processing, and=20
medicine. Theoretically, any material that contains protons can be=20
detected by the NMR-MOUSE, but polymers (i.e., plastic), elastomers=20
(i.e., tires), and biological materials (i.e., humans) give the best=20
results. Two new applications are possible now due to the portable and=20=

inhomogenous-field nature of the detector. First, imaging can be done on=20=

substances that contain ferromagnetic materials, such as steel-belted=20
tires. An ordinary NMR device would not be able to image such a thing=20
without serious damage and/or signal interference occurring. In fact,=20
strangely enough, the steel cords in tires have been shown to be=20
beneficial in improving the signal for the NMR-MOUSE. Second, because=20
the NMR-MOUSE can be placed in any desired position, directional=20
patterns (anisotropies) can be measured. Tendons have dense collagen=20
structures which are normally very difficult to measure using=20
traditional NMR. Even so, the NMR-MOUSE has successfully performed=20
accurate measurements on the Achilles=92 tendon in human subjects. This =
is=20
because the scanner can be manually adjusted to the =93magic-angle=94, =
54.7=20
degrees, where a robust measurement of anisotropies (directional=20
dependencies) can be made. More experiments are required to develop=20
these measurements into a useful diagnostic tool. As far as traditional=20=

imaging, the NMR-MOUSE has produced an image of the cross-section of a=20=

pork leg obtained from a butcher where muscle, bone, and marrow can be=20=

easily distinguished. Of course, this is not on the same par as modern=20=

MRI images which take incredibly detailed pictures of human structures.=20=

However, compared to older MRI images produced by multi-ton machines,=20
the 1.25 kg NMR-MOUSE performs quite impressively.

There is another imaging modality that is enjoying success with its=20
portability. Ultrasound has been used as a medical device since the=20
1950=92s. Since ultrasound uses sound waves, it is like MRI in being=20
non-invasive and non-ionizing. Ultrasound is safe enough to view the=20
fetus inside a pregnant woman=92s body without any concern for the baby.=20=

It can detect solid structures in the human body and analyze the=20
movements of fluids like blood. However, its biggest disadvantage is=20
that it cannot see behind bone or gas (i.e., air in the lungs). A modern=20=

ultrasound scanner has two components: a transducer that emits and=20
receives the sound waves and a computer-based data processing unit. The=20=

transducer is handheld while the computer is normally very large and=20
needs a skilled technician to operate. However, Sonosite has just=20
developed an ultrasound scanner which weighs 2.5 kg (for the transducer,=20=

computer, and display) and is portable and easier to operate than=20
traditional systems 2. Doctors can take the scanner anywhere to make=20
measurements without carrying a large computer-based system. In=20
addition, this new scanner costs only about $25,000 compared to the=20
$150,000 to $300,000 price tag for larger scanners. How can the=20
NMR-MOUSE compete with this new device?

At the moment, the NMR-MOUSE cannot compare with the new Sonosite=20
scanner in terms of usage and portability. First of all, it is important=20=

to realize that ultrasound and MRI are different imaging modalitites;=20
they each have significant advantages and disadvantages. Also, even=20
though Sonosite=92s ultrasound scanner is much better developed than the=20=

NMR-MOUSE, there is a large company behind Sonosite=92s new system which=20=

has provided all of the necessary capital to miniaturize the device.=20
Recall that traditional ultrasounds have handheld transducers and big=20
computers. Sonosite=92s main contribution has been to make the large=20
computer smaller, they haven=92t changed the fundamental ultrasound=20
technique. On the other hand, the NMR-MOUSE was developed by the=20
intellectual curiosity of two scientists with minimal funding who had to=20=

develop new techniques and designs. They were trying to find=20
alternatives to large and expensive MRI scanners. Fortunately, Bruker, a=20=

giant in the NMR and MRI industries, has showed interest in the=20
NMR-MOUSE and has begun collaborating with the scientists. With this=20
large corporate sponsorship, the NMR-MOUSE has an opportunity to really=20=

blossom. New designs will be considered to minimize the penetration=20
limitations that have given the NMR-MOUSE so many restrictions. Maybe in=20=

the near future, we can be diagnosed and treated by doctors in their=20
offices by portable MRI and ultrasound devices without having to go to=20=

the hospital and be charged a few thousand dollars. Even though=20
physicians will not carry Star Trek medical tricorders anytime soon, we=20=

are definitely a step closer. Now we just have to work on teleportation.
=09


President Garfield is assassinated at a train depot
 =46rom Frank Leslie's Illustrated Newspaper. July 16, 1881.




The first x-ray image


1. General Electric=92s medical devices page: http://www.
gemedicalsystems
.com/index.html



Felix Bloch (left) and Edward Purcell (right) are credited with=20
discovering NMR.



A normal MRI scanner setup
 =46rom Hornack, J. Basics of MRI



Schematic of the NMR-MOUSE
B0 is the main field from the magnets.
B1 is the varying field from the RF coil.
The gray rectangle below the magnets
represents the iron yolk.
 =46rom the University of Kent at Canterbury=92s NMR-MOUSE website



 =46rom the University of Kent at Canterbury=92s
NMR-MOUSE website



MEAT BONE MARROW
A pork leg, as seen by the NMR-MOUSE
 =46rom Bl=FCmler, P et al



The portable Sonosite ultrasound scanner


References
Bl=FCmich, B et al. The NMR-MOUSE: construction, excitation, and=20
applications. Mag. Res. Imaging. 16: 479-484; 1998.
Bl=FCmler, P et al. Spatially Resolved Magnetic Resonance. Wiley-VCH,=20
Weinheim; 1998.
Prado, P; Bl=FCmich, B; and Schmitz, U. One-Dimensional Imaging with a=20=

Palm-Size Probe. J. Magn. Res. 144: 200-206; 2000
Vlaardingerbroek, M. and den Boer, J. Magnetic Resonance Imaging.=20
Springer, Berlin; 1996.

Further Information
Hornack, J. Basics of MRI http://www
.cis.rit.edu/htbooks/mri/
Hornack, J. Basics of NMR http://www
.cis.rit.edu/htbooks/nmr/
Brown, G. Introduction to MRI Hardware
http://www.users.on.net/
vision/papers/hardware/hardware.htm
University of Kent at Canterbury=92s NMR-MOUSE website=20
http://wwwnmr.ukc.ac.uk/nmr/mouse/index.htm
University of Aachen=92s NMR-MOUSE website
http://www.nmr-mouse.de/