Extensive research has been conducted in the field of interventional magnetic resonance imaging. There are several factors driving this research, including the possibility that it may replace traditional fluoroscopy. There are many advantages and disadvantages of the modalities currently used for image guidance. This paper will address the various guidance methods and discuss the advantages and disadvantages of each, as well as current research efforts and developments.
Extensive research has been conducted in the field of
interventional magnetic resonance imaging. There are several
factors driving this research, including the possibility that it
may replace traditional fluoroscopy. There are many advantages
and disadvantages of the modalities currently used for image
guidance. This paper will address the various guidance methods
and discuss the advantages and disadvantages of each, as well as
current research efforts and developments.
Significant advances in the field of interventional magnetic
resonance (MR) imaging have resulted from an interdisciplinary
approach involving radiologists, physicists, surgeons, and computer
According to Gunther et al,
the most interesting question in conjunction with interventional MR
imaging is whether or not it will replace standard X-ray
fluoroscopy. The current clinical applications are numerous;
however, many of the procedures are performed efficiently and with
relatively low risk using the standard methods. The clear
challenges of vascular interventions include accurate and rapid
visualization and sequence acquisition in a safe and reproducible
manner. In situations in which the targeted lesions are seen only
with MR imaging, such as breast lesions, MR-guided intervention is
the ideal technique. This review addresses the potential advantages
and limitations of interventional MR imaging, discusses catheter
tracking and visualization techniques, reviews present device
development, and assesses current investigational and clinical
Potential advantages of interventional MR imaging
versus conventional fluoroscopy
Interventional radiology originated with fluoroscopically guided
percutaneous biopsies and angiography. These X-rayguided
procedures were soon coupled with, and thereafter often replaced
by, computed tomography (CT)-guided and ultrasound-guided
procedures. Both of these modalities offer the advantage of
improved visualization as well as a reduction in radiation
exposure. CT guidance provides greater anatomic detail and imaging
in the axial plane; ultrasound allows for variation in imaging
planes, lower cost, and no exposure to ionizing radiation. In
addition to these advantages, both of these techniques offer the
radiologist real-time image guidance similar to fluoroscopy.
As the field of interventional radiology progressed, MR imaging
was incorporated as a method of guidance. Several contributing
factors to the use of MR as a modality for imaging guidance
included multiplanar imaging capabilites and a decrease in the
radiation exposure to patients and personnel.
Additionally, MR imaging, exclusively and in combination with
high-performance CT, permitted the integration of three-dimensional
(3D) volume rendering and interactive localization. MR imaging also
provided optimal tissue contrast, thus allowing the visualization
of diseased vessels and their relationship to the surrounding
parenchyma without the use of intravenous contrast material.
MR guidance also offers the radiologist rapid access to
physiologic information, including quantification of flow,
perfusion, diffusion, and temperature.
Real-time information is available to the physician, illustrating
the immediate effect of the procedure or intervention.
Such immediate feedback may allow for prompt decision-making and
consideration of further treatment options that may not have been
assessable with conventionally available fluoroscopic
The combination of fluoroscopy and MR imaging in a tandem system
using MRI-compatible catheters and guidewires has provided an
alternative to single modalities. This tandem system consists of a
fluoroscopy unit, complete with digital subtraction angiography and
road mapping, as well as a closed, short-bore 1.5 T magnet. The
patient is placed on a table that is MR compatible and transparent
to X-rays, therefore, there is no need to reposition the patient.
The fluoroscopic system may be used for guidewire and catheter
placement, and the magnet can be used for multiplanar imaging,
anatomic visualization, tissue characterization, and physiologic
This combined technique incorporates the advantages of both imaging
modalities and avoids some of the difficulties associated with a
solely MR-guided procedure. This system also provides a significant
financial advantage over a dedicated interventional magnet because
each of the imaging suites can be used independently.
Using a somewhat different approach, Wen and colleagues
have developed a method of modifying a real-time fluoroscopic unit
for use in conjunction with an interventional magnet. This system
consists of a standard X-ray source (X-ray tube and housing), a
high-voltage generator, an X-ray detector, a detector power supply,
data acquisition electronics, and a display. The X-ray fluoroscopy
unit is placed completely within the bore of an interventional
system. Both X-ray fluoroscopy and MR images are available to the
interventionalist; however, one system is inactive while the other
is active. This technique provides an adjunct method that is
particularly useful as a bridging point to entirely MR-guided
Current limitations of interventional MR imaging
Despite the numerous advantages of interventional MR imaging,
interventional radiologists are still constrained by the
limitations of current equipment and devices. The current MR
imaging systems for interventional procedures may be classified as
either dedicated open systems, open or closed standard imaging
systems, or hybrid units consisting of MR imaging combined with
standard X-ray fluoroscopy.
Because open-bore magnets are limited by lower field strengths and
weaker gradients, present research efforts are focused on the
development of open magnets with higher field strengths.
An innovative development was the design of a magnet that allowed
relatively unrestricted access to the patient, similar to that
provided in a CT scanner or fluoroscopy suite.
This dedicated open magnet is a GE Signa 0.5 T (MRT) "Double
Doughnut" (GE Medical Systems, Waukesha, WI) that is open
vertically and offers relatively unhindered access to the patient
(Figure 1). A lower field-strength magnet results in decreased
susceptibility artifact from interventional devices; however, a
standard closed high-field strength magnet offers the advantage of
rapid image acquisition with higher temporal and spatial
resolution. Several alternative design options have been
considered, including shortening the magnet to provide access to
the patient from either end of the table.
Although new magnet designs have resulted in improved image
quality and increased accessibility, one of the most challenging
aspects of interventional MR imaging involves the development and
manufacture of interventional catheters, needles, and guidewires.
Standard interventional guidewires are composed of stainless steel
and other ferromagnetic materials that are incompatible with the MR
environment. Some devices may cause significant artifact and
obscure the field of interest. Additionally, metallic devices may
heat up secondary to current induction. To overcome the problem of
heating, several decoupling methods that are frequently used in
surface coils can be implemented.
Similarly, to address the concerns of ferromagnetic properties and
artifacts, several alloys and ceramic materials have been
As researchers design new catheters and interventional devices,
methods of image localization and guidance techniques must be
developed concurrently. Researchers have focused on the development
of rapid imaging sequences for the standard MR image acquisition
and interventional procedures.
Presently, experimental and clinically investigative devices are
visualized using passive and active methods.
There are three methods of passive visualization used to localize
catheters and guidewires. One method involves filling the catheter
lumen with contrast agents and causing the catheter to appear
brighter than the surrounding unenhanced tissues.
This technique was used by Strother and colleagues
in the endovascular treatment of experimental canine aneurysms.
Catheters were placed using fluoroscopic guidance and subsequently
tracked and depicted with MR imaging using commercially available
catheters filled with a gadopentetate dimeglumine solution.
Catheter movement was visualized when either a catheter or the
coaxial space between a catheter and a guidewire was filled with
contrast. Superimposing the reconstructed images obtained during
catheter manipulation onto a previously acquired MR angiogram
localized the catheter.
Unal and colleagues
used gadolinium-coated catheters and guidewires for passive
tracking and visualization in phantoms and animals. This coating
technology, in combination with a passive tracking technique,
offered several advantages over other active and passive tracking
techniques. These advantages included visualization of the entire
device independent of its orientation in the magnetic field. These
preliminary results demonstrated the potential use of MR-visible
Another method of passive visualization involves capitalizing on
the differences in magnetic susceptibility of the catheter and the
surrounding tissue, which results in intravoxel dephasing. This
method does not require contrast; however, the geometric distortion
that accompanies this effect can limit the practicality of this
Active tracking is an additional method of catheter
localization. It is based on imaging with either a local coil or
the "loopless" antenna of the guidewire. In active tracking, a
small receiver coil in the tip of the catheter or guidewire
generates the signal. The signal is then created and actively
detected or emitted by a device to identify its spatial location.
The position of the coil is shown in three dimensions by the
depiction of a small colored dot on a previously acquired image.
During the procedure, the coil may be tracked up to 20 times per
As methods of visualization, both passive and active tracking
have limitations when compared with conventional fluoroscopy and
digital subtraction angiography. These methods of visualization and
navigation require substantial deviation from the accustomed
fluoroscopic methods. As previously mentioned, passive
visualization may be limited by geometric distortion and may
require special coating. The constraints of active tracking include
the ability to identify the coil position only; thus, the entire
length of the device cannot be visualized unless multiple coils are
Significant progress toward clinical applications has been made
as a result of early experiments. The first human interventional MR
imaging experiment, conducted in 1996 by Smits and colleagues,
showed clear delineation and visualization of movement of a solid
catheter with 5 dysprosium markers.
The authors described the early endovascular experiments showing
the feasibility of placement of nitinol stents in an arterial graft
model under MR imaging guidance. Standard catheters with the tip
markers removed were visualized in a phantom. Further iterations
involved attaching the flow phantoms to human volunteers for
further imaging sequences.
As equipment and tracking methods improve, the development of
MR-compatible catheters, guidewires, and stents becomes integral to
the continued advancement of MR imaging guidance. Considerable
research efforts have focused on MR-guided cardiovascular and
stenting procedures. Currently, cardiovascular MR imaging is
performed using external receivers in external, phased-array
radiofrequency (RF) and surface coils that are typically placed in
the chest and back of the patient. Coronary atherosclerotic plaques
have been visualized using this technique; however,
characterization of plaque composition and vascular wall detail is
not optimal. In order to achieve a desirable level of detail, a
high signal-to-noise ratio and high spatial resolution is required.
To this end, Hurst et al
developed intravascular receiver coils. Endovascular placement of
these devices improves the spatial resolution by generating an
extremely high signal-to-noise ratio in the voxel immediately
adjacent to the coil. According to Lardo,
most of these early designs were limited by their rapid radial
signal intensity fall-off, sensitivity to motion, and poor
flexibility. A recent design by Ocali and Atalar consists of a
loopless (dipole) antenna.
This design offers a smaller profile and a higher degree of
flexibility than previous designs. The loopless catheter antenna is
used as a transmit/receive probe for transmitting RF pulses and
receiving MR signal. It can also be used as a receive-only signal
probe. The signal intensity is also high enough to allow remote
placement of the associated electronics.
In parallel with the loopless catheter antenna, alternative
catheter designs incorporating balloon technology have been
developed. Ladd and colleagues
describe a new intravascular design consisting of a loop of wire
encased by a polymeric water-filled balloon. Expansion of the
balloon increases the coil diameter and improves imaging
capabilities. Using an ex vivo technique in a rabbit model, the
authors demonstrated enhanced visualization of vessel wall layers
and plaque structure. Other investigators have used either a
loopless or balloon-expandable transesophageal probe to obtain
high-resolution vascular images (Figure 2).
One promising design, a vascular 0.030-inch internal MR coil
(Intercept Vascular 0.030" Internal MR Coil, SurgiVision, Inc.,
Columbia, MD) that is similar to a guidewire in construction, has
recently received Food and Drug Administration approval for
high-resolution vascular imaging (Figure 3) (Ingmar Viohl, personal
communication, June 26, 2002). The recent developments of small
intravascular coils, guidewires, and active and passive methods of
visualization have led to the investigation of monitoring balloon
angioplasty under MR guidance. Yang et al
recently reported the feasibility of MR-guided balloon angioplasty
in an animal model of aortic stenosis. This study was performed
entirely under MR guidance with a 1.5F intravascular loopless
antenna guidewire as described by Ocali and Atalar in conjunction
with a standard 4F, 4-cm balloon angioplasty catheter.
The balloon catheter was inserted into the vessel over a 0.018-inch
guidewire, which was subsequently removed during MR imaging.
Passive tracking was used to monitor the margin of the balloon with
tantulum markers placed proximal and distal to the balloon. Vessel
dilatation was monitored with an intravascular antenna as a
high-resolution probe using an MR fluoroscopy sequence.
In conventional fluroscopically guided procedures, balloon
angioplasty is often followed by placement of endovascular stents.
The design and development of endovascular stents is an active area
of research in interventional MR imaging. A new MR-guided stent has
been tested in in vitro and sheep models.
The ingenuity of this stent design allows it to serve not only as a
mechanical device to maintain vessel patency, but also as a
receiving antenna capable of generating the 3D coordinates of the
stent coupled with high-resolution images of the vessel wall.
Current investigational and clinical
Newer techniques have shown promise as alternatives to
traditional invasive surgical procedures and current interventional
methods. Jolesz and colleagues
have integrated MR image guidance with computer-assisted surgery.
This system allows the surgeon to use the magnet while in the
operating room to allow accurate active localization and anatomical
definition. The first open brain surgery using intraoperative MR
imaging was performed in 1996. Lesions currently treated with this
method include hemorrhage, neoplasms, cavernous hemangiomas, and
A significant benefit of intraoperative MR imaging is the immediate
evaluation of the effectiveness of resection with specific
sequences and spectroscopic curves, which may help to prevent the
need for a second surgery. This same group has performed >300
craniotomies, >130 cases of MR-guided prostate brachytherapy,
and >20 cryosurgical procedures for tumor ablation with
preoperative diagnostic imaging and intraoperative therapy
Kee and colleagues
showed successful hepatic-to-portal vein puncture in a swine model
using MR and fluoroscopic guidance. A nitinol guidewire was used to
access the right hepatic vein using a right jugular vein approach
under C-arm guidance. The pigs were then moved into an open
configuration Signa SP/i 0.5T MR imaging unit (GE Medical Systems).
A nitinol transjugular intrahepatic portosystemic shunt (TIPS) set
was subsequently advanced into the right hepatic vein over a
guidewire. Multishot echoplanar images were obtained and used to
guide puncture of the portal vein by passive artifact from the TIPS
set. Successful puncture of the portal vein was confirmed by MR
imaging followed by portal venogram and fluoroscopy for additional
confirmation. This study concluded that this technique "may in the
future prove the standard method of performing what is now a
'blind' access procedure."
Clinical MR guidance was first applied to needle aspirations
that were subsequently followed by MR monitoring of thermal
Breast biopsies and lesion localizations are currently performed
clinically. Daniel and colleagues
described a freehand interactive MR imaging technique for
preoperative localization of breast lesions (Figure 4). Nineteen MR
image-guided breast localization procedures were performed in 17
patients using an open platform breast coil in either a closed-bore
1.5 T magnet or an open-bore 0.5 T magnet. Rapid imaging sequences
(fast-spin echo, water selective fast-spin echo, or water-specific
three-point Dixon gradient echo) were used to localize the lesion.
The needle was manipulated between sequences to the desired
location. Accurate localization of the lesion was achieved within 9
mm of the target in all cases and 10 lesions were confirmed with
mammography or ultrasound. The remaining 9 lesions were visible
only with MR imaging. This study showed a simple and accurate
technique for interactive MR imaging-guided localization of breast
lesions without the need for stereotactic equipment. This method
offers the additional advantage of providing access to lesions
close to the chest wall and in the anterior portion of the breast.
Clinical work conducted by Roberts et al
recently showed successful placement of carotid stents in 7
patients using a combined MR imaging/X-ray angiography suite.
Patients underwent a preprocedural MR imaging including MR
angiography, phase-contrast flow assessment, and perfusion and
diffusion imaging. Stents were then placed using standard
fluoroscopic methods and patients were imaged after the procedure.
MR imaging provides immediate evaluation of intimal detail and
flow, which allows for relatively immediate feedback of the success
of the initial procedure and the evaluation for the need of further
When contrasting interventional MR with the traditional methods
of image guidance, relevant issues include: imaging plane
capabilities, exposure to ionizing radiation, and available devices
and equipment. The question as to whether interventional MR will
routinely replace or be combined with standard imaging guidance
methods remains. The few clinical trials discussed here have proven
successful. Further clinical progress can be anticipated as new
devices and methods of imaging localization and guidance
The author wishes to thank Dr. Robert Herkfens and Dr. Bruce
Daniel for their advice and guidance, and Dr. Viohl for kindly
allowing the use of his images. The author also wants to thank Mark
Riesenberger for his images. A special thanks to Kristina Benjamin
for her help.