Advantages of Contrast
MR angiography can be performed without contrast, and such
techniques often provide excellent information and may, in many
instances, produce images of adequate quality. It is important,
however, to be aware of the limitations of non-contrast techniques.
In Figure 1A, a two-dimensional time-of-flight (TOF) sequence gives
a false appearance of only moderate disease of the carotid
bifurcation. But on the three-dimensional high-dose
gadolinium-enhanced MRA (Figure 2B), it is clear that bifurcation
anatomy is more complex. A maximum intensity projection that zooms
in on the bifurcation shows that, although the degree of luminal
narrowing is only moderate, there is a large ulceration in the
atherosclerotic vessel wall that may be contributing to the
patient's symptoms (Figure 2C). This abnormality is not visualized
by flow-based, non-contrast MRA techniques.
Another shortcoming of flow-based MR techniques, such as 2D TOF,
is the tendency for the signal to drop out in sites of very severe
stenosis as a result of jet flow--so-called spin dephasing.
As shown in Figure 2, 3D gadolinium-enhanced MRA, which is not
flow-based, more accurately depicts the true lumen and enables more
precise appreciation of anatomic details.
Artifact associated with metal objects in the body represents a
challenge for MR imaging. The 3D gradient echo sequence we use for
high-dose gadolinium-enhanced MRA employs a very short echo time
that helps to eliminate this artifact. Nevertheless, some metal
produces such a severe artifact that it cannot be eliminated. The
stainless steel Palmaz stent is an example. In the case of platinum
stents, the metal stent configuration creates a Faraday cage
phenomenon that attenuates the penetration of radiofrequency (RF)
energy both into and out of the stent. By increasing the flip angle
to 75°, however, it is possible to deliver sufficient RF
penetration into the stent to enable visualization of the vessel
lumen (Figure 3).
Another important advantage of 3D contrast-enhanced MRA is the
elimination of in-plane saturation. This is critically important in
patients with alteration of the normal vascular anatomy.
In Figure 4, the patient has an occlusion of the left subclavian
artery, and flow to the left subclavian is provided by a
carotid-to-left subclavian artery graft. Three-dimensional
contrast-enhanced MRA readily depicts the unpredictable course and
direction of such vessels and, in this case, reveals a complex
plaque at the carotid bifurcation.
Patient safety in the MR scanner is an issue that must be
considered. In general, MR is not suited to imaging high-risk
patients. For example, in the evaluation of dissection, MR would
typically be reserved for evaluating Type 3, stable dissections. It
is, however, capable of imaging high-risk patients, and may
represent the best approach even in an emergency if the risk of
subjecting a particular patient to radiation or iodinated contrast
is sufficiently high.
In summary, gadolinium-enhanced MRA is able to visualize the
origins of the great vessels; it is fast and, therefore, produces
less motion artifact than non-contrast techniques; it is associated
with less spin dephasing; it can demonstrate plaque features and
luminal anatomy in enough detail to reveal ulcerations; it is not
hampered by in-plane saturation; it enables the evaluation of small
vessels, such as the vertebral arteries; and postprocessing is very
simple. Moreover, if gadolinium is already being used for brain MR,
there is no extra cost associated with obtaining 3D
gadolinium-enhanced MRA at the same time.
Several factors must be considered when selecting the dose of
First, it is important to realize that the bolus of contrast
evolves as it travels through the body. Initially, at the point of
injection, it is very compact. It spreads out as it passes through
the pulmonary circulation, and becomes even less concentrated after
it passes through the systemic capillary bed.
The selection of gadolinium dose, therefore, must take into
account the distance between the injection site and the area of the
body that will be imaged, as well as how much gadolinium will be
extracted along the way. As might be expected, the higher the dose
of contrast, the greater the visualization of vascular detail. This
is demonstrated in Figure 5. At a gadolinium dose of 0.1 mmol/kg,
it is not possible to see the venous anatomy in the liver, but as
the dose is increased to 0.2 mmol/kg and 0.3 mmol/kg, the portal
venous anatomy can be appreciated.
The injection method itself is also important to consider.
Manual injection has several advantages in MRA, because so many
factors must be coordinated perfectly. The delivery of the bolus,
the breathholding of the patient, and the initiation of the
acquisition all have to occur simultaneously. In my experience, it
is easier to coordinate each of these factors while standing next
to the patient and performing the injection by hand.
In addition, hand injection necessitates consideration of the
viscosity of the contrast agent, which in turn influences decisions
on the size of angiocatheter, length and caliber of intravenous
tubing, and type of gadolinium contrast to use.
In subclavian venography, for example, two operators are
required in order to simultaneously inject the right and left
antecubital veins with dilute gadolinium contrast media (20 mL
gadolinium in 250 mL normal saline) in order to image the
subclavian veins and superior vena cava (Figure 6).
Optimal contrast administration requires an understanding of the
basic physics of k-space and how it influences the features of an
image. In particular, the center of k-space dominates the contrast
features of the image, whereas the periphery of k-space dominates
It is important, therefore, to time the delivery of the contrast
bolus so that arterial contrast is at its maximum during the
central portion of k-space, to achieve the maximum contrast effect.
A perfectly timed injection of contrast may provide excellent
quality with a relatively low dose of contrast media. But with a
small contrast dose, small bolus timing errors can ruin a study.
Increasing the contrast dose allows for a greater margin of error
in bolus timing. In essence, contrast dose can make up for
variations in the skill of the operator. The most talented
operators may be able to use a lower contrast dose. On the other
hand, if the operator is less skilled or performs MRA infrequently,
a larger contrast dose may be warranted in order to demonstrate
more angiographic features.
MRA offers the exciting opportunity to expand beyond a scanner's
usually limited field of view to image extensive regions of
anatomy, simply by moving the table during contrast injection.
For example, after performing an initial image in the pelvis, we
can pull the table and acquire another image of the thigh, then
pull the table again and image the calf. It is also possible to
deliver several small injections of gadolinium contrast as the scan
moves from the feet to the knees, then finish with a bolus chase.
The ability to scan large regions of the body is especially
exciting because of the systemic nature of atherosclerotic
Bolus-chase MRA has been extrapolated to cover arteries of the
entire body, beginning with the aortic arch and carotid arteries,
and moving sequentially down the body to image arteries of the
chest, abdomen, pelvis, thigh, calf, and feet.
The patient in Figure 7 was unable to undergo conventional
angiography, because access through the iliac and common femoral
arteries was not possible. But with MRA, only a peripheral
intravenous line in the arm was needed before imaging the arteries
of the entire body with a single bolus injection.
Contrast-enhanced MRA offers important advantages over other
forms of angiography. It poses no risk of ionizing radiation or
nephrotoxicity. Source data can be reformatted for any view. It can
depict anatomy and physiology in great detail, and it is highly
accurate. As researchers continue to work on refining and advancing
the techniques of MR angiography, its future looks bright. *
Thank you, Dr. Prince. Let me ask a few specifics about how you
administer the contrast, and what doses you use. You mentioned that
you give this by hand injection, and that viscosity is an issue.
Does that effect your decision to use different agents?
Yes, if you need a small caliber angiocatheter, then the higher
viscosity agents are going to be more difficult to inject,
especially by hand. Or especially if you are going to be imaging a
baby with a very small caliber angiocatheter, then the viscosity
becomes very important. So, we would prefer agents that have the
Another important detail is that viscosity is related to
temperature. So if you find that the resistance to injection is too
great, you can warm up the gadolinium to body temperature. And that
will drop its viscosity by approximately a factor of two. Then that
reduces the force required to inject the contrast by hand.
DAVID ROBERTS, MD:
We routinely give high-dose gadolinium for peripheral bolus-chase
MRA applications. We find that the injection protocol is critically
important for these studies. If we inject too quickly, certainly
anything above 1 mL/sec, we notice significant superficial femoral
venous (SFV) enhancement and tibial venous enhancement, which can
really degrade the studies. For that reason, we have found that we
really can't get by with hand injection, because many of the
fellows and residents, despite training and stressing this point,
simply inject too quickly. Therefore, we found power injectors
fairly mandatory for these studies. What's your take on this?
We've had the same experience. When we are performing a
multistation examination with one bolus and are trying to share
that bolus between multiple stations, if we time it for the first
station, our timing for the second and third station may not be
accurate, depending on whether the patient has fast or slow flow.
So we have begun doing a test bolus at the calf. Then that gives us
a better sense of the timing for all three stations. If a patient
has very fast flow, if it is <20 seconds to the calf, for
example, then we won't even attempt to do a bolus-chase exam,
because we know that we just cannot move the table fast enough to
keep up with really fast flow.
How do you address the problem of asymmetry, left to right, in a
Our pattern has been to try to figure out which leg is the
clinically significant leg, which is the one that's on track for
being operated on first or is providing the most difficulty for the
patient. Then we time for that leg, and hope for the best on the
I have a question about a potential pitfall with great vessel and
carotid imaging. Most of these patients, of course, are going to be
injected in one arm or the other and as that tight bolus goes past
the common carotid artery, especially with a left-sided injection,
we sometimes get a false drop-out in signal over a short segment
related to the venous gadolinium passing there. How do you approach
that? Do you routinely inject the right arm? Do you sometimes
inject a femoral vein?
That's an excellent point. It brings up one of the advantages of
the hand injection. We always try to inject the right arm. As we
are injecting, we want to get the bolus in and flushed out of the
vein before we start collecting the central k-space data. So we use
a fluoroscopic triggering technique for our carotids, If we are
injecting and also tracking the pace of the passage of the contrast
through the lungs and as it reaches the great vessels, if we see
that it is going quickly and we are not going to get all the
contrast in and flushed, I may actually terminate the gadolinium
and immediately switch to flush earlier, just to get it flushed out
of the subclavian vein and innominate vein.
If you have to inject the left arm, you may not be able to
escape some venous artifact, superimposing it over the great vessel
origins. Part of the reason for that is that in a lot of older
patients, the aorta is unfolded and very hard and it pinches the
brachycephalic vein between the aorta and the sternum. It causes
the gadolinium to hang up there, and despite with your best efforts
to flush it through, you may not be able to clear that.
LARRY KRAMER, MD:
Martin, you showed in one graph that as you increase your contrast
dose, that you see more and more detail. Is there a point at which
as you increase dose, susceptibility takes over and you actually
decrease signal and detail?
Yes, that's an excellent question. I think it's more related to
injection rate. There's going to be an injection rate at which the
concentration of gadolinium in the blood becomes so high that you
begin to have T2* effects, which actually cause the signal to drop.
In my experience, with hand injecting, it's hard to inject fast
enough to reach that point. But if you are performing venography
and injecting that vein, you have to dilute the gadolinium to avoid
that T2* effect. So that is a very important point.
For venography, how much do you dilute the gadolinium?
I typically take a bottle of 15 or 20 mL of gadolinium and inject
that into a 250 mL bag of normal saline. Then I'll load up four
60-mL syringes. That gives me a very large volume of diluted
contrast to work with, so I can slam that diluted gadolinium in, in
such a way that it will tend to distend the veins. Yet it will not
be excessively concentrated. I'll inject a full 60 mL before
beginning the 3D acquisition. Then I will continue to inject
another 60, then that would be two syringes on each arm, so that I
can get bilateral information.