Selection Criteria For an MR System
Selecting an MRI System is often a lengthy and difficult process. In
today's environment, multiple vendors or distributors can bombard the potential
user with a combination of marketing fact and fiction, all of which can
make the process quite confusing. In this brief summary, we will look at
some of the important selection criteria for MR, and provide the reader
with a "guide" to the decision variables.
he
purchase of a MR system can be a complicated, confusing and difficult process.
When the typical Physician or Administrator is asked to participate in an
equipment selection committee, the process frequently becomes a mix of lengthy,
conflicting and emotional arguments. But this need not be the case. By following
some simple guidelines and separating the important issues from marketing
buzzwords, making the right choice should become that much easier.
MRI Basics
One of the first tasks for the potential buyer is a consideration of the
type of work that the new MR system is expected to perform. Historical data
indicates that neurologic exams are one of the primary driving forces behind
MR system purchase, but the applications tend to diversify quickly into
vascular, musculo-sketetal and body imaging once the equipment becomes available.
It is wise to think through the desired types of exams and the Referring
Physician population to be served, as these factors will certainly govern
the number and types of software packages and dedicated surface coils required
to satisfy the expectations for the new machine.
Generally speaking, all MRI systems can be thought
of as having similar components or block elements. They are: Magnet, Radio
Frequency Subsystem (RF), Gradient Subsystem, and the Computer/ Array Processor
Subsystem. While each of these subsystems are important (and we will look
at each in turn), it is prudent to understand that most modern MR systems
also require some environmental modifications. These modifications include,
but are not limited to: RF screen room, air conditioning, power and lighting
modifications, and for heavier systems, structural reinforcement. In large
part, the sitting considerations are related to the magnet type, size and
weight. In a worst case scenario, a very heavy magnet or one with a large
fringe field can add an additional 10-20% in costs, depending on the site
p; over and above the "normal" MR purchase price.
The Magnet Selection
The heart of an MR System is the Magnet; it is
this subsystem that is responsible for "persuading" nuclei to
align with or against (parallel or anti-parallel) the field direction. The
discussion of field strength, type of magnet and direction of field are
all likely to arise during your purchase cycle. The simplest way to look
at field strength is this: the stronger the filed of the magnet (higher
Tesla), the greater the potential signal-to-noise return in clinical studies.
This is indeed the simplest way to look at the equation, because other variable
and subsystems play important roles in signal capture. Overall, the strength
of the MR signal is the "currency" with which we will clinically
"buy" the parameters that govern the overall image quality, resolution
and time. Higher field magnets will therefore exhibit thinner slices, routine
use of higher resolution matrices p; all with much shorter imaging times.
Shorter imaging times are beneficial in that they may be parleyed into increased
throughput p; and increased Return on Investment (ROI) p; or can
be used to perform more exhaustive patient studies.
Superconducting magnets utilise a principal in which the resistance in certain
types of wire can be reduced to near zero as long as the temperature is
at or very close to Absolute Zero (0° Kelvin). One of the key elements
in the construction of a superconducting magnet is the cryostat (thermal
"can" that slows the loss of cryogen's - normally liquid Helium
(@ 4° K). The cryostat is quite expensive, and the reason that few
vendors offer a low field Supercon unit is that this additional manufacturing
cost cannot usually be justified below 0.5 Tesla. The normal clinical operating
range for such magnets is from 0.5 T to 2.0T, with modern technology allowing
high-field installation in areas that were inaccessible with the previous
generation of magnets.
While potentially yielding the highest signal-to-noise,
purchasing the highest field available is not always advisable. Consider,
for example, that even though 2.0 Tesla is roughly 30% higher field than
a 1.5 Tesla magnet, the 1.5 T systems are vastly more popular, largely sue
to fewer imaging artifacts, hence greater sequence flexibility and better
image quality. About 25-30% of all systems sold world-wide are 1.5T. Mid-field
(usually 0.5T) magnets remain very popular. They offer signal-to-noise return
better than lower field alternatives yet have historically offered lower
entry and life cycle costs when compared to high field. This has made the
systems popular enough that approximately 30 - 40% of all units sold fall
into this range. 1.0T is a more recent market that began as a "niche",
but has grown dramatically, apparently driven by a physics perspective,
but more importantly a competitive marketing argument. The 1.0T systems
are considered closer to a "High Field" system, and offer a differential
over mid field without 1.5T associated costs. Just over 20% of all systems
sold are 1.0 Tesla units. (Above left: GE
Medical's Signa Horizon 1.0T).
The standards for the medical industry are the 0.5T, 1.0T and 1.5T, and
as such, products in these categories tend to offer the buyer a heightened
Security of Investment.
Permanent magnets are constructed of magnetised
iron (Ferrite) or rare earth alloys. Permanents systems offer some advantage
in that they generally have zero life-cycle magnet cost versus the need
for liquid Helium to cool the coils of superconducting magnets. The main
disadvantages are the weight of the iron required to create the field, the
limited field attainable, and the environmental requirements in order to
maintain a stable field. Permanent magnets typically weigh in a range from
9 - 12 tons, and due to floor loading, structural reinforcement is generally
required to site such systems. These types of magnets are generally available
in clinical operating range 0.06 T to 0.3 T, and previous remarks made regarding
the signal return and the concomitant imaging trade-off with a lower field
will apply. (Above right: Philips'
popular Gyroscan NT).
Resistive Magnets are, as the name infers, magnets that create field via
the resistance to electrical current in wire that is wrapped around an iron
core. This type magnet was relatively popular in the early days of MR, however
from an environmental standpoint the early systems had disadvantages. Resistive
magnets require a very constant and stable flow of electricity, and since
the resistance produces heat, they also require water cooling. Newer systems
may include power stabilisers and closed loop cooling systems, but the costs
associated are still quite high for the benefit returned to the user. These
systems are typically operated in the 0.15 - 0.25 Tesla range.
About 10 - 15% of the world-wide volume in MR units falls into the low-field
(Permanents and Resistive) range. Perhaps the most significant benefit offered
by these systems is a geometry that differs from the traditional cylindrical
shape of superconducting magnets, and may allow for a more friendly patient
interface.
RF Subsystems
Nearly all modern MR systems utilise - or claim, to utilise - digital RF
technology. Digital RF permits precise control over the RF wave form, complex
RF pulses and higher signal to noise capture. The wise buyer will ask for
this feature, and is able to check functionally via RF "spoiled"
T1 Gradient Echo sequences, availability of EPI (Echo Planar Imaging) and
sequence that is a combination of Gradient and Spin Echo, sometimes called
GRaSE. Given the rate of change in MR applications, the buyer should also
insist on a system, compatible with coil mounted pre-amplifiers, with a
range of flexible and Quadrature surface coils to increase Signal to-Noise
Ratios (SNR) and to cover the desired anatomic areas of interest.
From a hardware perspective, most of the newer systems are offering some
type of phased array surface coils but the growth in MR technology makes
distinctions even here. The buyer should look for Quadrature versus Linear
Phased Array Technology, and methods for extending the number of coils in
a given array so that increased resolution and increased anatomic coverage
is possible. The benefits and flexibility provided by Quadrature Phased
Array will probably change MR image acquisition practices over the next
few years.
Gradient Subsystems
This is one of the most misunderstood areas in MR technology today. Many
buyers automatically assume that "bigger is better", and base
selection on the highest gradient amplitude available from various manufacturers.
This attitude overlooks how the gradients are used, and, indeed, how they
will be applied in clinical cases. Gradient amplitude governs the system
ability to differentiate spatial location, and as such covers Field of View,
Matrix of Acquisition and Slice Thickness. A combination of magnetic field
homogeneity and digital control over the gradient amplitude permits a moving
locus for the field of view, or off-centre FOV. Gradient rise time, or the
actual time taken for the amplifiers to energise the coil(s) to full amplitude
governs the speed of the imaging sequence. Taken together, the amplitude
and rise time can be expressed in a performance figure known as the "slew
rate" of the gradient subsystem. Gradient coils are really resistive
magnets that alter the main magnetic field for purposes of creating images.
Systems that exhibit high slew rates are relatively power hungry, and of
course, produce heat that requires dissipation... generally with liquid
cooling (see resistive magnets). As such, the big gradients are not for
every clinical site, regardless of what the latest marketing story from
commercial vendors might suggest. The buyer should look for systems that
are upgradable to higher slew rate gradients in the future, once truly clinical
applications are designed and proven.
The buyer who is actively involved in research may wish to begin at once
with high slew rate gradients, however, even then there are bio-limitations
that will set a maximum rate of change in keeping with patient safety. The
buyer who must have this feature is recommended to check the flexibility
of the gradient subsystem and ask pertinent questions. Does the benefit
of high slew rate extend to "normal" imaging as well as EPI? Can
the user take advantage of the increase amplitude to perform MR microscopy
that requires high-resolution matrix and small FOV? Is MR fluoroscopy or
real time imaging permitted? Does the system permit physician/ patient interaction
and/or intervention under real-time MR guidance? Are there provisions for
viewing the "real time" images while working with the patient
in the magnet room? If required, is the system fast enough to produce motion
free cardiac images? The buyer is urged to take the time to study their
needs and the offering of the vendors and "beware" offers that
seem too good to be true, as they usually are.
Computer Subsystem
Early generations of MR Systems all used proprietary Operator Consoles,
customised computers and array processors to run the systems and process
images. In later years, the customisation became quite elegant, with touch
screens and specialised displays. Given the speed of change in the computer
and electronics industry, however, these must be considered strategic dead
ends as any future upgrade will be quite expensive. The buyer should look
for a work station-like environment utilising easily attached and easily
upgradable off-the-shelf components to minimise the planned obsolescence
that characterises many of the available systems today. (Left: One of
a range of workstations from Picker International
configured from standard components).
The buyer should investigate the computer and array processor capability
to "keep up" with the data stream coming from the acquisition
module of the equipment. It does not make sense to purchase a high performance
system that can acquire 10 images per second (or more) if the reconstruction
capability is only two images per second, as a "log jam" of data
will certainly ensue. A system which is balanced in design will permit smooth
operation and allow for an upgrade path in the design, which in turn increases
the security of investment in the MR system.
Closing Thoughts
Looking at the first decade or so of MR innovation,
it seems clear that the buyer must require that the manufacturer provide
substantial proof of upgrades of the MR system in order to keep pace with
the ever-changing technology. While it may not seem to matter..a system
which has been on the market for several years with few changes in the basic
building blocks (e.g. Magnet, RF, Gradient and Computer subsystems) may
lead to a technological and competitive "dead end" long before
the systems can be automised by the hospital. Capital equipment upgrades
must be affordable, and involve more than a re-packaging of the same old
technology with a new name. It is always wise to plan for upgrades at the
time of purchase, with provisions in the equipment finance package or in
future capital budgets that can help insure that the hospital remains on
the cutting edge of technology. Ask for the vendor's history in providing
upgrades, and ask what additional benefits are derived from the "new"
system over previous versions. (Above right: The Magnetom Impact from
Siemens, the industry's most recent
success story with over 900 units sold in four years).
As stated at the outset, purchase of an MR system is complex and can cause
anxiety during the process p; and buyer remorse long after the acquisition
has been completed. The buyer should look for a vendor with a consultative
style rather than a "hard sell" agent who is only looking for
the next sales trophy. While this is never an easy task, the buyer should
take the time to evaluate systems on the basis of appropriateness to their
need, a match to the budget, and a company that will stand behind the equipment
for the entire life cycle.
© MCMXCVI Calyx Productions