Endovascular stent-grafting is an increasingly common treatment for aortic aneurysm. Early on, digital subtraction angiography and intravascular ultrasound were used intraoperatively to take critical measurements of aortic length and diameter, thus guiding selection and placement of the aortic stent-graft
is a Professor of Radiology, Medical College of Wisconsin,
Aortic aneurysm occurs in 3% of people over the age of 50, and
5% to 8% of men over the age of 65. Men who smoke and people with
hypertension are at highest risk. If left untreated, an aortic
aneurysm with a diameter exceeding 6 cm has a 5-year mortality rate
of 25% to 40%.
Endovascular stent-grafting is an increasingly common treatment
for aortic aneurysm. Early on, digital subtraction angiography and
intravascular ultrasound were used intraoperatively to take
critical measurements of aortic length and diameter, thus guiding
selection and placement of the aortic stent-graft. Today, CT
angiography (CTA) can be used prior to the stent-grafting to create
an anatomic model of the aorta. Quantitative measurements enable
preprocedural selection of the stent-graft and planning for its
placement. CT angiography also evaluates patient characteristics
that infiuence procedural success, including the viability of iliac
access, and the presence and quantity of calcification and
atheromatous debris in the superior neck.
This article will discuss screening for and surveillance of
aortic aneurysms; the role of CTA in patient selection and planning
for endograft therapy; the performance and interpretation of CTA,
including the use of quantitative analysis tools; and the
recognition and management of stent-graft complications.
Aortic aneurysm can be defined in various ways. A practical
definition is an abdominal aorta with an external diameter >4 cm
(in the case of a fusiform aneurysm); or a focal dilatation >3
cm in conjunction with a proximal aortic luminal diameter <1.5
Ultrasound is the preferred technique for screening patients for
aortic aneurysm. It is recommended that patients who have an
abdominal aortic aneurysm ≥4 cm in diameter undergo surveillance
sonography at intervals of 6 to 12 months. Once the diameter
reaches approximately 5.5 cm, an intervention should be considered,
balancing the risks of treatment against those of leaving the
aneurysm untreated. Another indication for intervention is an
abdominal aortic aneurysm that increases by >5 mm in diameter in
a 6-month interval. Due to the inherent risks of therapy and
reduced life expectancy, a larger aortic diameter may be required
in an older patient be-fore recommending intervention.
An aortic aneurysm can be repaired surgically or with
endovascular stenting. Endovascular stenting is becoming more
common because it avoids many of the disadvantages of surgery,
including the need for laparotomy. Laparotomy involves aortic
cross-clamping, which, in the case of juxtarenal or suprarenal
aneurysms, can result in ischemia of the intestine and the kidneys;
surgical opening of an aneurysmal sac causes obligatory blood loss;
and the surgeon may inadvertently lacerate major vessels, such as a
retroaortic left renal vein.
Several critical anatomic issues must be evaluated before
endovascular stent-grafting. These include the availability of
iliac access, which is necessary for introduction of the
stent-graft device; the length and diameter of the proximal
placement zone, between the renal vascular pedicle and the upper
margin of the aneurysm; the degree of atheromatous burden in the
superior neck of the aneurysm; the angle of the neck in relation to
the aneurysm and its implications for fixation of the device; and
the length and diameter of the distal landing zone, which, in most
cases, is the common iliac arteries.
There are three goals of CTA of the aorta. First, there should
be arterial enhancement of 250 to 300 HU, achieved through rapid
injection of contrast material (5 mL/sec). This level of arterial
enhancement will enable discrimination of intravascular contrast
from aortic calcification and thrombus, while providing adequate
contrast for longitudinal resolution of vessels parallel to the
slice plane, particularly the renal arteries. The second goal is
contrast injection intervals and image acquisition intervals of
similar length. The third goal is first-pass cephalocaudad coverage
from the supraceliac aorta to the proximal thigh, achieved within a
sustainable breath-hold interval.
On a GE 16-channel scanner (GE Healthcare, Waukesha, WI), the
imaging protocol would call for 16 detector rows with a collimation
of 0.625 mm, resulting in a beam width of 10 mm. Using a pitch of
1.375:1, and a 0.5-second scan rotation, scan speed would be 27.5
mm/sec. Given an average cephalocaudad distance of at least 30 cm
from the supraceliac aorta to the proximal thigh, the scan would
take approximately 12 seconds. Acquisition interval will vary with
required anatomic coverage (depending on patient body habitus) and
scan rotation speed.
Once the image acquisition interval has been calculated, the
patient's circulation time can be determined through use of either
bolus tracking software or by injection of a preliminary mini-bolus
and measurement of the time to peak aortic enhancement. This
represents the optimal delay between contrast injection and image
acquisition. The injection interval can then be made to equal the
calculated image acquisition interval (usually approximately 12
During image analysis, we use an automated technique to measure
diameters and lengths at predefined points along the aorta. This
technique uses seed points and automated edge detection to enable
centerline tracking of the aorta and the iliac arteries. A slider
is moved up and down while the model of the aorta is rotated,
selecting specific anatomic points at the level of the renal
arteries, the upper and lower margins of the aneurysm, and the
iliac bifurcation. The system then calculates diameters
(perpendicular to the centerline of the aorta), cross-sectional
areas, and the lengths between the selected points.
In Figure 1, measurement of the superior neck of the aneurysm,
from the renal vascular pedicle to the upper margin of the
aneurysm, is defined. The image nicely reveals the aneurysm itself,
thrombus and calcification in the wall and, to some extent,
atheromatous material. Figure 2 illustrates measurements from the
renal vascular pedicle to each iliac bifurcation, enabling an
estimate of the length that must be covered by the endovascular
stent-graft (not including the superior neck of the aneurysm), and
measurement of the distal placement zone.
There are several contraindications to aortic stent-graft
placement. In the case of juxtarenal aneurysm (Figure 3), the
covered portion of the stent-graft would not only exclude the
aneurysm but would also overlay and block blood fiow to the renal
arteries. Development of fenestrated stent-grafts may overcome this
problem in the future.
A short superior aneurysmal neck is a contraindication with some
types of stent-grafts, as it does not provide an adequate proximal
landing zone. Newer types of stent-grafts that enable transrenal
fixation and placement of the upper covered margin immediately
underneath the renal vascular pedicle now make it possible to treat
such an aneurysm endovascularly, however.
Distal stent-graft placement that would occlude both hypogastric
arteries (as would occur if the patient had bilateral common iliac
aneurysms extending to the iliac bifurcations requiring distal
extension of the stent to the external iliac arteries) represents
another contraindication. It would lead to relative ischemia of the
pelvis and rectum. This is of particular concern in patients who
have stenosis of the superior mesenteric artery.
A relative contraindication to stent-graft placement is the
presence of large amounts of atheromatous material in the aorta. In
Figure 4, atheromatous and atherothrombotic material account for
>25% of circumference of the superior neck of the aneurysm. The
accompanying orthogonal longitudinal projection reveals a
significant amount of atheromatous material in the aneurysm
Concerns also arise in patients who have a history of kidney
surgery. For example, in a patient who has undergone nephrec-tomy,
it becomes critical that placement of the endovascular graft not
traumatize the renal artery supplying the single remaining kidney.
Similarly, in a patient with a transplant kidney, passing a large
endovascular device through the iliac artery could injure the renal
The Guidant Ancure system (Guidant Corp., Indianapolis, IN) was
one of the earliest stent-graft devices. Although it is no longer
available commercially, patients implanted with this device still
undergo CT follow-up (Figure 5). It uses barb fixation superiorly
and inferiorly and is often supported by internal Wallstents
(Boston Scientific, Natick, MA) at the level of the iliac arteries.
The aorto-uni-iliac device was used in the early days of
stent-grafting in patients with a relatively wide superior
aneurysmal neck (these were fabricated locally according to
individual patient need). This device had only one iliac limb;
therefore, the opposite iliac limb had to be occluded and
circulation to the contralateral lower extremity restored with a
surgical femorofemoral crossover graft.
The Medtronic AneuRx stent-graft (Medtronic, Inc., Minneapolis,
MN) uses friction seal for proximal and distal fixation and
accommodates the same aortic dimensions as the Ancure device.
The Gore-Excluder (W.L. Gore & Associates, Inc., Flagstaff,
AZ) (Figure 6) has lower-profile introducer (18F) than many other
devices and accommodates a 26-mm superior neck. The Zenith
stent-graft (Cook, Inc., Bloomington, IN) (Figure 7) uses
suprarenal fixation by barb devices, which enables better placement
in certain types of patients. The device can be placed in superior
necks up to 32 mm in diameter. It also has a bell-bottom
configuration that enables it to expand in the iliac arteries for
The Medtronic Talent stent-graft, not yet commercially available
in the United States, is also able to accommodate a wider superior
aneurysmal neck for proximal placement and wider iliac arteries for
Two major complications of endovascular stent-grafting are
atheroembolism during the procedure and Type I endoleak. A Type I
endoleak results from inadequate attachment of the stent-graft at
either its upper or lower margin (Figure 8).
This complication requires early treatment.
A Type II endoleak (Figure 9) or refiow endoleak results from
infiow from either the inferior mesenteric or lumbar arteries. The
infiow maintains an elevated pressure in the aneurysmal sac, which
prevents it from becoming smaller over time, or even causing it to
A Type III endoleak is caused by a defect in the graft or to
modular disruption. A Type IV endoleak is due to fabric porosity,
and is only a potential issue in the initial period after
Generally, patients are imaged within 1 week of stent-graft
placement; again at 3, 6, 12, and 18 months; and annually
thereafter. During postendograft evaluation, patients have a
precontrast imaging study first, so that calcium in the organized
wall of an aneurysm sac can later be differentiated from contrast
material in the aneurysm sac, which would indicate an endoleak.
Patients then undergo a timed first-circulation study. This
yields aortoiliac CTA images and enables measurement of both
stent-graft dimensions and the distance between the stent-graft and
related anatomy. For example, we may measure clearance between the
renal vascular pedicle and the upper margin of the stent-graft, as
well as between the lower margin of the stent-graft and the iliac
bifurcation to make sure that there has been no interval migration
of the stent-graft. The patient also undergoes an immediately
delayed study, which, in combination with the initial first
circulation study, enables better detection of Type II
The CTA displays we use are familiar ones: stacked axial scans,
multiplanar reformations, maximum-intensity projections,
volume-rendered displays, curved planar reformations, and
quantitative vessel measurements.
Endoleaks can be treated by translumbar needle insertion into
the aneurysmal fiow channel. Another option is selective
transfemoral catheterization, either using a perigraft approach to
catheterization of the lumbar artery, or direct catheterization of
the superior mesenteric artery, followed by catheterization of the
inferior mesenteric artery. Acrylate injection is used to occlude
the perigraft aneurysm sac.
CT angiography is a robust and reproducible technique that
provides both three-dimensional imaging and quantitative
information critical to the pre- and postintervention evaluation of
patients who are candidates for aortic endovascular
LEO P. LAWLER, MD, FRCR:
Dennis, thank you; that was very good. There is one area I was
wondering about, though. When you are following up, some of you do
a noncontrast study, then an arterial-phase study, and then a
delayed study. Although I think the multidimensional approach can
help in terms of assessment of either stent integrity or migration,
certainly in the coronal images, what do you think about the idea
that one should really strive to use larger collimation, higher
pitch, much lower dose, and limit your z-axis for the noncontrast
study and the postarterial phase? In truth, do you really need to
have the same kind of coverage? There is at least one paper showing
that the dose differences are huge if you can limit yourself that
way. But, of course, you are going to be left with 5-mm collimation
and so forth.
W. DENNIS FOLEY, MD:
Actually, we use thin section only for the angiographic phase and
thicker sections for the precontrast and immediate delayed
Right. But they all have metal components for the sites of
purchase. Can you standardize?
BRIAN R. HERTS, MD:
Do you have problems trying to distinguish between contrast and
calcification in the wall of the aneurysm on the bolus phase if you
are using thick sections on the precontrast image?
It is still a pretty good registration. So, yes, we used to look at
those images at 5-mm and now at 2.5-mm image thickness for the
precontrast and immediate delayed postcontrast studies. We use thin
0.625-mm sections for the first circulation run.
GEOFFREY D. RUBIN, MD:
Actually it is for that reason specifically that we use thin
sections for the noncontrast phase especially.
At least in our experience, we find that some of this
calcification can be pretty tough to separate from small areas of
endoleak, and the confident identification of that calcification
before the contrast is delivered can sometimes make or break the
diagnosis. I think, within the context of radiation dose and
exposure, you have to look at your demographics. We are dealing
with people who have almost zero probability of developing a
radiation-induced neoplasm in their life, especially people with
aortic aneurysms. So I really do not see dose as a consideration at
all for these people.
I find that it is also easier for the technologist to set up the
same scan range three times.
I think that is probably true. On the other hand, when one is
arguing over "is it calcium or is it contrast," you are generally
dealing with something that is very small. Since the majority of
endoleaks actually resolve themselves in 6 months or so, I wonder
whether we should really be so worried about them in terms of
chasing them, especially the small things, to determine whether
there is just a little bit of contrast or is it a little bit of
calcium. I understand if it is large and the leak is coming back
from the lumbar or something. I find they are fairly easy to
What is interesting, at least in our experience, is that the actual
size of the endoleak, particularly if it is a type II endoleak,
does not necessarily associate with the likelihood of a sac
shrinking or not. So if the sac does not shrink, even the smallest
of endoleaks has to be suspect and of concern, whereas if you have
a large endoleak but the sac is shrinking, then the significance of
that endoleak is not necessarily so great. I think it is important
to keep our eye on the ultimate outcome here, and that is the sac
shrinkage and, ultimately, nonrupture.
But do you also find that with sac stability that lack of growth
rather than shrinkage is the more common finding? Or do you find
most of your sacs shrink?
That is a good point. Actually the likelihood of shrinkage seems to
be device-dependent. Ultimately, you like to avoid growth and a lot
of them will stay stable.
So what I meant to imply is that with some of these type II
endoleaks, even if they continue to grow, that growth is not
necessarily associated with the size of the endoleak. This issue of
sac shrinkage and device-dependence is something that I think we
will need to learn more about. There actually have been some
reports of rupturing sacs in the absence of any endoleak. And
people are beginning to implicate ePTFE over Dacron as being a
semiporous material that allows hygromas to develop in the sac,
which results in the so-called endotension phenomenon. In some
cases, they have gone on to rupture. Although in that case, the
question is what is really rupturing and does it have the same
Dennis, do you believe that endotension exists?
We have had a case similar to what Geoff is describing in which we
could not find a cause, yet the aneurysm sac expanded.
U. JOSEPH SCHOEPF, MD:
So what is the protocol in your institution? How often do you image
those patients and how do you react if you see the sac growing or
We image the patients immediately post stent-graft and then at 3,
6, 12, and 18 months and 1 year intervals after that. But if they
have an endoleak even in the second year, then I would go with
6-month intervals rather than the 1-year intervals. If a patient
really is a critical issue, then I will go with 3 months. That
patient I mentioned who had endotension got converted to an open
Dennis, I have two questions. You have great information there.
First, can you tell us all what images and data you are providing
to the surgeon? My second question is one that I hear a lot of
radiologists ask at meetings. When you are performing a CTA study,
do you look at the other organs? Should we report the full
abdominal and pelvic CT if we are only looking at the aorta?
The answer to the second question is yes. We are obligated to look
at everything, although we make the disclaimer that it is based on
an arterial-phase acquisition. Some things may be missed but, yes,
we certainly do look.
In response to your first question, we provide a lot of
quantitative information. Although there is a question of how much
of that is really utilized.
I try to keep it to what I demonstrated; ie, the superior neck
measurements of length, diameter, and angle (ie, angle in relation
to aneurysm), and the length and diameter of the common iliac
arteries. That is rather critical information. Otherwise, you could
build up a large table of information that just sits there not
ELLIOT K. FISHMAN, MD:
Let me ask one last question about selection of contrast agents.
Many of your images showed patients with aortic aneurysms, or in
the age group in which they have atherosclerotic disease throughout
many vessels, and many of them had decreased renal size. In
choosing agents, is the se-lection of an isosmoloar agent in those
patients something to consider?
One of the issues is that we do get patients with marginal renal
function. On the other hand, the amount of contrast we give is
So you can do a study with 60 to 70 mL of contrast plus a
preliminary mini-bolus. We do use the isosmolar contrast agent when
their creatinine level is >2, which I think other people would
suggest. But, otherwise, <2 we are not really concerned about
the type of contrast agent because we are using a relatively small
Right. I guess the question to follow it up might be: Is there any
concern with the smaller volumes of contrast? I know in terms of
doing measurements, it is not an issue. But following a patient
with suspected endoleak, is there some volume of contrast that
makes you feel comfortable that you are picking up those leaks?
You may have hit me on a weak point there because, yes, we do use
the same technique. It is possible, and I suspect you may be right,
that particularly with that second-pass immediate delayed imaging
sequence, you probably are more sensitive if you inject more
contrast to begin with.