The evaluation of patients who have had a myocardial infarction
(MI) often starts with functional imaging, which usually consists
of the acquisition of cine images and their analysis for wall
motion abnormalities. Figure 1 demonstrates the use of myocardial
tagging, a technique that aids in the evaluation of regional wall
motion throughout the myocardium. In this case, it reveals severely
disordered contraction of the myocardium. Wall motion analysis is
unable to answer a crucial question, however: whether the areas of
abnormal contraction are still viable or represent infarction. An
assessment of myocardial viability using contrast techniques is,
therefore, the next step.
In assessing myocardial viability, we acquire first-pass
myocardial perfusion images following a single dose of gadolinium
contrast media delivered at a 5 mL/sec, a fairly high injection
rate (Table 1). We use power injectors for all viability and
perfusion studies. Because these patients often are quite sick,
they typically have a large-bore catheter already in place, which
enables us to inject using a large-bore Angiocath.
We image the entire heart every 2 heart beats for about 1 minute
after contrast injection to assess microvascular integrity. Next,
we inject a second dose of gadolinium contrast, wait for 10 to 20
minutes, and obtain myocardial viability images using gradient-echo
techniques with myocardial suppression.
Consider the application of MR viability imaging in a patient
with an extensive MI. Even though many such patients have the
infarct-related artery opened in the cardiac catheterization
laboratory and undergo implantation of a coronary stent, some still
have small-vessel disease or capillary occlusion. MR is the only
noninvasive method that can assess disease at the capillary level
reliably, and this is accomplished during first-pass perfusion
If on the first-pass images the capillaries remain
occluded--evident as a dark area that is not perfused--we inject
the second dose of contrast and wait for approximately 10 minutes.
During this time, as contrast washes out from the rest of the
myocardium, there is a redistribution of contrast between areas of
nonviable and normal myocardium, such that the contrast agent is
retained in the area of infarction. Figure 2 demonstrates this
concept in a patient with large transmural infarction of the
inferior and basal walls.
Even though infarcted myocardium can be mapped on delayed
viability imaging, it is the core of the infarct, which appears on
first-pass images and corresponds to microvascular obstruction,
that predicts patient prognosis. Wu et al
found that patients with evidence of microvascular obstruction
following MI were more likely to experience a cardiovascular event
during follow-up when compared to those without microvascular
obstruction (45% versus 9%). The size and presence of microvascular
obstruction predicted an increase in left ventricular mass and
volume, and a decrease in ejection fraction, at 6 months.
MR can be used to evaluate both acute and chronic MI. The
purpose of the study is often quite different in these two
settings, however. In acute MI, the primary issue is the size of
the infarction, an assessment that can sometimes be difficult to
make clinically. For example, cardiac enzyme levels may be
challenging to interpret in a patient who has received a
thrombolytic agent. In another example, a woman may delay coming to
the hospital for chest pain because her symptoms are atypical, thus
skewing interpretation of the troponin levels.
In chronic MI, patients typically are being evaluated before
revascularization procedures, and the question is one of viability,
rather than infarct size. If the target myocardium is
dysfunctional, but still viable, bypass grafting could markedly
improve cardiac function.
In both acute and chronic MI, cine images demonstrate an absence
of motion in the area of infarction. In the acute infarct, there
may also be areas of dysfunctional myocardium that is stunned but
will eventually recover.
Wall thickness can be analyzed by anatomic imaging. In acute MI,
the myocardium typically does not show thinning, whereas after 3 to
6 months, the chronic infarct is characterized by thinning and
remodeling of the myocardium with scar formation.
Both acute and chronic infarction demonstrate marked enhancement
with gadolinium contrast, although the pathophysiologic mechanisms
are quite different. It is hypothesized that in acute MI, there is
an increase in the distribution volume of the contrast agent in
areas of nonviable myocardium, where myocytes have died. In the
case of chronic infarction, in which the myocardium is replaced by
fibrosis and scar, it is hypothesized that there is a difference in
contrast distribution between the nonviable myocardium and the
scar, when compared with the adjacent myocardium.
Viability imaging can also be useful in the assessment of
patients with idiopathic hypertrophic subaortic stenosis, following
alcohol ablation of the enlarged ventricular septum.
In such patients, the anterior motion of the mitral valve leaflet
is obstructed by the enlarged septum during systole, which
obstructs aortic outflow. The standard treatment at our institution
is coronary alcohol injection to ablate a portion of the
myocardium. These procedures are done empirically and guided by
evidence of decreased wall motion. Gauging how much of the
myocardium has been treated can be quite difficult
Using MR, functional images are obtained before the procedure to
assess outflow obstruction. Following injection of alcohol into the
coronary artery, first-pass perfusion images demonstrate areas
without blood flow (Figure 3). On delayed imaging, approximately 10
minutes after injection, the areas of nonviable myocardium are
clearly mapped out. Patients can also be followed over time to
assess overall cardiac function as well.
Myocardial perfusion imaging is typically used to evaluate
patients with ischemic heart disease, rather than MI. An infusion
of the vasodilator Adenoscan (Fujisawa Healthcare, Inc., Deerfield,
IL), 140 µg/kg/min, augments coronary blood flow. After
approximately 2 minutes, 0.1 mmol/kg gadolinium contrast media is
delivered as a single, rapid-infusion bolus (5 mL/sec), as in the
myocardial viability protocol described previously. Gradient echo,
echo-planar notched interleaved pulse sequences are used to image
the entire myocardium for about 1 minute after a contrast
Because the protocol involves the use of the vasodilator
adenosine to augment coronary blood flow, there are several
contraindications to MR perfusion imaging, including wheezing,
hypertension, the use of caffeine or aminophylline, or greater than
first-degree atrioventricular heart block. Depending on the patient
population, approximately 5% of patients are likely to experience
discomfort during the exam. Approximately 0.5% experience a
potentially serious adverse reaction to adenosine, such as
arrhythmias or heart block. During stress testing, a staff member
with Advanced Cardiac Life Support certification must be
Figure 4 shows a patient with reduced perfusion in the septal
and anterior portion of the myocardium, and another patient with
left ven-tricular hypertrophy and small subendocardial areas of
A major area of investigation involves the quantification of
perfusion defects, as a better alternative to visual estimation of
the extent of stenosis.
Nonetheless, even with visual estimation, the overall correlation
between perfusion abnormalities that are physiologically
significant and angiographic stenosis has been quite good (Table
The imaging of atherosclerosis represents a major new and
developing application of contrast administration. Often, MR
angiography of patients with suspected heart disease shows an
essentially normal coronary lumen. However, there can be tremendous
amounts of disease in the vascular walls. Assessment of the
atheromatous components of plaque represents an important and
active area of investigation today. Key questions in determining
whether a plaque is vulnerable to rupture include whether it has a
large lipid core or is covered by a thin fibrous cap. Both are
thought to increase the risk of plaque rupture, with the ultimate
result being obstruction of the coronary artery and MI.
Figure 5 demonstrates the use of contrast in imaging
atherosclerosis of the carotid artery, where most of the initial
studies have been done.
The area of fibrous cap is well enhanced, enabling assessment of
cap thickness. The lumen is quite small, and adjacent to the
fibrous cap is the lipid core. Within the lipid core is an area of
calcification. All of the plaque components are well delineated on
the MR image, which correlates closely with the histology
The imaging of atherosclerosis can be applied to other vascular
beds as well.
Probably the most interesting and important are the coronary
arteries. Patients with atherosclerosis may have only mild degrees
of coronary arterial narrowing on angiographic sequences, but still
have significant amounts of disease inside the vessel wall (Figure
6). The true extent of disease and risk for subsequent plaque
rupture is probably best assessed using contrast-enhanced MR.
The use of contrast for myocardial viability imaging is a major
new application of MR. Myocardial perfusion, stress imaging, and
the evaluation of myocardial ischemia are active subjects of
investigation as well. Contrast-enhanced imaging for
atherosclerotic disease, particularly in combination with MR
coronary angiography, holds great promise for evaluating the risk
of cardiac events. *
Thank you very much, Dr. Bluemke. Let me ask a question
specifically about the perfusion. You mentioned that you give the
bolus at a high rate of injection. Do you use a power injector for
that, or do you do that by hand?
DAVID BLUEMKE, MD:
We are using power injectors for all of our studies. For first-pass
perfusion studies, we typically inject at about 5 mL/sec. These
patients are quite often very sick patients and they typically have
large-bore catheters already in place. So it is quite feasible to
inject using a large-bore angiocatheter with a high injection
I'd like to ask you about the tendency in nuclear medicine to
obtain rest and stress images. Where do you think that fits in?
Concerning injection, when you have to give dynamic gadolinium
injection and pharmacologic injection simultaneously, then you
would need multiple IVs. Or can they be piggybacked?
As far as the the use of stress and rest images, the initial
studies that were done with perfusion MRI definitely considered
that model of application. The reason for doing stress and rest
images is to distinguish viable from nonviable myocardium, ischemic
myocardium versus infarction. That model was used initially. Now it
is a development of these viability MR sequences, where the normal
myo-cardium is suppressed with inversion recovery pulse
The standard model now is to do stress images during adenosine
infusion, and then wait for 10 to 20 minutes, then look at the
areas that are nonviable on the delayed images. There's been a
movement then to not require multiple injections for those images,
during first-pass conditions.
The reason for that is because you don't need the resting image,
because the viability image tells you where the infarct is, if it
That's right. The nonviable myocardium will light up very nicely,
and can be delineated at about 10 to 20 minutes after contrast
So are you then suggesting that the need to image under stress may
be completely eliminated by these new MR perfusion and viability
No, we think that stress conditions are going to be necessary to
distinguish narrowing prior to infarction. If a patient presents
with chest pain and typical anginal type symptoms, the question is,
is that coronary heart disease or is it some other cause of chest
pain. In those conditions, in which an infarction has not occurred,
it is very likely that only under stress conditions will those
areas of coronary narrowing be revealed on the perfusion images,
the first-pass images.
During the stress imaging, who is monitoring the patients? Do you
do it with the cardiologist?
Yes, during the stress cases, someone who is ACLS certified has to
be there. We typically have a very competent nursing assistant
addressing the medication issues. They can be quite complicated
procedures, although they are of relatively short duration. After 4
to 6 minutes, the stress imaging is completely done and the
adenosine has a very short half life of about 1 to 2 minutes. So
when we are done with that stress component, we can rapidly assess
the patient to determine if there is any evidence of heart block,
or any chest pain associated with the stress condition.
RUSSELL LOW, MD:
What's the incidence of complications or symptoms?
Well, the incidence of symptoms can be relatively high, depending
on the patient population you are studying. Some discomfort can
occur in probably <5% of the patients. As far as overall severe
life-threatening conditions, the incidence for adenosine is on the
order of about 0.5%, at least in the literature. If you talk to
labs that do a lot of these studies, I think they believe that that
number is a little bit high. But if you document a large number of
cases in which arrhythmias and heart block are delineated
accurately, it may very well be that those numbers are quite
accurate. So you have to be prepared to treat patients and get them
out of the room rapidly if complications should occur.
In the first-pass images that you showed, I noticed some
misregistration, which I presumed was motion artifact from
breathing. Do you find that interferes with the creation of color
maps of the mean transit time and other parameters?
Yes, we try to do the first-pass images during an extended
breath-hold and ask the patient to hold his or her breath for as
long possible. For some patients, it is about one image. Other
patients do quite well, and hold their breath for 30 or 40 seconds.
The key time is the first 20 to 30 seconds. We have a pretty good
success rate in having patients coordinated with the technologist
in terms of the injection and the scanning, to be able to hold
their breath during that period of time.
Right now most of these images are evaluated qualitatively. With
quantitative evaluation, it's a huge issue. But there are motion
correction software packages that will re-register the images, so
we can place cursors over the areas of interest to map those signal
How soon after a patient arrives in the emergency room, do you feel
safe to do a stress perfusion study? At some point, you have to
decide whether the patient's had an infarction or not.
Well, there are a couple of issues there. Let's say the patient
comes into the emergency room with chest pain. First of all, there
are not a lot of MR centers that are set up to handle that patient
in the acute setting. There have been studies in which MR scanners
have been sited next to the emergency rooms. They have been used
not so much for stress studies, but for nonstress studies for
viability, since the areas of dead myocardium will light up very
rapidly after the event occurs.
In animal studies with 90-minute occlusions of a LAD,
immediately after imaging, those areas will show on the delayed
enhancement images. The more common scenario is with someone with
chest pain who was triaged and is in the CCU when we image them, if
they need to be imaged at all. Usually it's patients who have more
complicated courses, or who have atypical conditions that are
imaged. Women may present differently; the troponin levels may have
been missed because of delayed presentation in the emergency room.
So usually it's a combination of factors.
Not all patients need viability imaging necessarily. We think
that some routine myocardium infarction cases are treated very
nicely, and they get very little imaging whatsoever in the
hospital. It's probably in the sicker patients, more atypical
presentations, that this will be a useful technique.
Do you think MR has the potential in these emergency room patients
to determine whether there has been a myocardium infarction? Would
a normal MR allow safe discharge of the patient; thereby
eliminating the more common pattern of watching the patient
overnight and checking enzymes and so forth?
Yes, and it is a very significant issue for hospital
administrators, because of the substantial costs of keeping people
in the hospital for 24 hours. So, there is a major reason to triage
these patients quickly. Whether that will have an impact, I think
will still unfold.
There have been studies at Washington Hospital, for example, in
which they have triaged patients using MR to determine if they will
be admitted, or if they will be observed for just a short period of
time. Overall, they've shown reasonable success for mapping areas
of myocardium infarction. One of the issues in the nonstress state
is whether you can assess nonviable myocardium versus ischemic
changes. You probably cannot. So in the acute setting, mapping the
myocardium infarction to determine if an infarct has occurred may
be quite useful. But these patients will still need some triaging
and ischemic testing or stress testing, if they go home.
Howard talked about perfusion imaging the brain, and measuring MTT
blood volume and blood flow. Is there any use for those
measurements in the myocardium?
As I mentioned, most studies have looked at qualitative changes.
There is a lot of investigation in the use of quantitative
evaluation. Personally, I think it would be quite valuable. But the
techniques are not widely applied at this point and the software is
not widely available. There are many people who have subtle
perfusion changes, which may be in the realm of normal, that we'd
like to quantify. We think we can triage between the abnormal
patient with ischemia or infarction, versus the normal