The development of helical CT and multidetector-row CT technology has revolutionized the noninvasive imaging of the thoracic aorta. These advances, however, mandate careful attention to scan acquisition parameters. This article addresses proper scan methadology, diagnostic pitfalls, and the imaging findings of aortic diseases.
Dr. Gotway is an Assistant Professor of Radiology at the
University of California, San Francisco and Chief of Thoracic
Imaging at San Francisco General Hospital, San Francisco,
The development of helical CT and, recently, multidetector-row
CT technology, has revolutionized the ability to image the thoracic
aorta non-invasively. Helical CT aortography (HCTA) possesses
several advantages over other modalities frequently used to
investigate thoracic aortic disease, such as transesophageal
echocardiography (TEE), MRI, and aortography. The speed, ease, and
widespread availability of helical CT make it the examination of
choice for acute aortic pathologies such as dissection, intramural
hematoma, aneurysm, and acute traumatic injury. Helical CT is less
operator-dependent than TEE and is easier and faster to perform
than either MRI or aortography. These considerations are
particularly advantageous in acutely ill patients. Another benefit
of helical CT that these other methods do not provide is the
ability to image the entire thorax. This is particularly important
because cardiovascular abnormalities frequently present with
nonspecific symptoms, and nonvascular etiologies of the patient's
presentation may be obvious with helical CT. Finally, the
volumetric data obtained with helical CT aortography is readily
processed using two-dimensional, three-dimensional, or
volume-rendering techniques to provide angiogram-like images or
images in any desired plane or obliquity.
Helical CT aortography technique
Although no standardized techniques exist for helical CT
evaluation of the thoracic aorta, and practices vary from
institution to institution, certain general principles may be
applied to optimize techniques. Variables that must be considered
include collimation, pitch, field-of-view, reconstruction
increment, amount, rate and timing of contrast administration,
distance to be scanned, kVp, mA, and tube rotation time. These
variables must be optimized to provide the highest possible scan
quality. Occasionally, a compromise among these parameters is
required to achieve this goal.
Collimation and pitch--When a scan protocol is constructed, one
must generally strike a compromise between the collimation and the
pitch. Recall that pitch reflects the ratio of table speed to slice
thickness (collimation) per tube rotation, or:
Pitch = Table transport distance
A useful general rule of thumb is that effective slice thickness
increases by 30% when the pitch is doubled. Thus, a 3-mm scan with
a pitch of 2 has an effective slice thickness of 3.9 mm. Apply this
concept to the following situation:
What technical parameters provide the best scan quality if the
distance to be scanned is 10 cm? If one uses 5-mm collimation with
a pitch of 1, the scan will take about 20 seconds (depending on
tube rotation time) with an effective slice thickness of about 5
mm. If 3-mm collimation with a pitch of 2 is employed, the scan
will require just under 17 seconds with a 3.9 mm effective slice
thickness. Thus, whenever possible, it is preferable to maximize
spatial resolution using thinner collimation while increasing the
pitch to cover the entire scan distance more expediently. A faster
scan will in-crease scanner throughput and minimize respiratory
Field-of-view--The field-of-view (FOV) deserves special
consideration for two reasons. First, because the conditions that
prompt helical CT examination of the aorta are often nonspecific,
evaluation of the entire thorax is important. The diagnosis may be
found in the pleural space or peripheral lung parenchyma, not the
aorta. Additionally, using small FOVs will decrease photon flux in
the imaging volume, contributing increased image noise. This may be
partially compensated for by increasing the mA, but the scanner may
not permit a sufficient increase in mA if the length of the scan is
long. For these reasons, it is advisable to measure the FOV from
outer rib to outer rib at the widest portion of the thorax. This
FOV creates a good compromise between a large enough FOV to
visualize the entire thorax yet provides a larger image with
improved resolution that is easier to read.
Reconstruction increment--A reconstruction increment that
provides nearly a 50% overlap in slice thickness is generally
sufficient to generate excellent quality images. Recent work
suggests that there is limit at which overlapping the
reconstructions no longer provide improved longitudinal resolution.
Beyond this point, further overlap only generates more images and
thus only increases scan reconstruction time and storage
requirements. The optimal overlap depends on the pitch employed. As
a compromise, a 50% overlap generally provides optimal resolution
and enhances the quality of post-processing techniques. One should
avoid using reconstruction increments that are exact harmonics of
the collimation employed (i.e., a 2.5-mm reconstruction increment
with a 5-mm collimation scan). In such circumstances, artifacts may
be introduced if post-processing techniques are used.
Iodinated contrast--Amount and concentration: There is no
consensus regarding what iodine concentration should be routinely
employed for helical CTA of the aorta. The common practice of using
undiluted (300 mg I/mL or 360 mg I/mL) is usually sufficient. In
one study, an iodine concentration of 125 mg I/mL was found to be
unacceptable at any flow rate.
In this same study, an iodine concentration of 150 mg I/mL provided
satisfactory images, when the flow rate was adjusted. This iodine
concentration had the added benefit of decreasing the amount of
streak artifact emanating from thoracic venous structures. However,
questions regarding the safety of diluted contrast preparations
remain because such dilute mixtures are not generally commercially
available. Obviously, there are issues of practicality as well.
Generally, we employ undiluted contrast (300 mg I /mL) for all
thoracic helical CTA applications.
Rate: There is considerable variation among investigators
regarding the rate at which contrast is delivered for helical CTA
of the thorax. Generally speaking, a rate of 3 cc/sec is
sufficient, although some investigators use rates as high as 5
cc/sec. Some consider higher rates as less practical because they
may require larger bore catheters; however, rates of 4 to 5 cc/sec
may be achieved using 20-gauge catheters. Generally, an injection
rate of 3.5 cc/sec is satisfactory for helical CTA applications in
the chest. For such injection rates, 22-gauge catheters are
adequate. Indeed, flow rates of 4 to 4.5 cc/sec may be used through
such catheters when required. It should be noted that when higher
injection rates are used, a larger volume of contrast must be
delivered to maintain the injection throughout the scan
Injection timing: An appropriate scan delay is critical for
excellent vascular imaging. As a general rule of thumb, a standard
delay of 25 seconds for helical CTA of the thoracic aorta will
usually suffice. Patients with cardiomyopathy or an unsuspected
stenosis of the injected vein may require longer delays to achieve
optimal aortic opacification. Patients with hyperdynamic cardiac
function may require a shorter delay. The time to peak aortic
opacification from patient to patient varies considerably. Van Hoe
found that this delay varied from 11 to 30 seconds, with an average
of 20 seconds. One may perform a test injection to determine the
time to optimal aortic opacification in a given patient. With this
method, a 20-cc bolus of contrast is administered, followed by a 10
second delay. After the 10 second delay, one image every 2 seconds
is acquired at the same level (often the cranial aspect of the
descending thoracic aorta) for a total of 30 seconds. The image
with the greatest contrast density is used to select the proper
scan delay. This method is somewhat cumbersome to perform
routinely, and has the added detraction of an increased radiation
dose as well as a slight increase in visceral background
attenuation due to the injected contrast.
As an alternative, bolus timing software packages may be
purchased for most modern CT scanners. Such programs allow the time
to peak enhancement to be monitored without a separate test
injection. These programs monitor the attenuation of the vessel in
question during the contrast injection, and display the attenuation
graphically in real-time. Once the graph demonstrates a sharp rise
in attenuation, the scan sequence is triggered manually. When
possible, we prefer to use bolus timing software to ensure proper
arterial opacification. With even minimal experience, the CT
technologist can learn to achieve excellent arterial opacification
without the direct supervision of the radiologist.
Distance to be scanned--The entire distance to be scanned is
important for several reasons. Perhaps, most importantly, the
contrast injection must be maintained for all but the last few
seconds of the scan acquisition. Failure to do so may allow
unopacified blood to enter the imaging volume, thus creating
artifacts and diminishing the quality of the study. Second, the
longer the distance to be scanned, the greater the heat loading on
the tube, necessitating a lower mA be used. This is particularly
problematic in larger patients, who may require all the photons the
tube can muster. Third, imaging beyond the desired structures will
provide additional information, but at the expense of more tube
exposures. As the x-ray tube has a finite life span, extra tube
exposures may come at significant cost. Finally, imaging beyond the
desired volume also contributes to increased patient radiation
kVp and mA--There is usually no reason to adjust the kVp from
the standard values of 120 to 140 kVp. However, the mA used is
critically important for achieving excellent quality studies. While
it is preferable to keep mA as low as possible to reduce patient
dose, little is accomplished if the patient receives a somewhat
lower radiation dose at the expense of a poor quality study. The
contrast resolution on CT scanners is heavily dependent on mA.
Noisy images resulting from increased quantum mottle often result
from using mA values that are too low for a particular patient.
Although for the average patient, mA is often not a major
consideration, for larger patients, an upward adjustment in the mA
is critical to achieve a high quality study. Often this adjustment
must be performed manually because modern CT scanners are often
programmed to minimize the mA to reduce tube heat loading and
patient dose. When larger patients are being scanned, the mA should
be adjusted manually to near maximum to diminish image noise and
provide the highest quality imaging. Careful attention to FOV and
scanning distance in this circumstance is also required.
Tube rotation time--Modern scanners now allow tube rotation
times as low as 0.5 seconds. GE scanners (GE Medical Systems,
Milwaukee, WI) have a minimum tube rotation time of 0.8 seconds.
There are options to increase the tube rotation time that may have
a role in other thoracic applications (e.g., HRCT). Helical CTA of
the aorta, however, is best accomplished with the fastest allowable
tube rotation time.
Normal aortic anatomy
An extensive discussion of mediastinal anatomy is beyond the
scope of this review, but normal aortic anatomy will be considered
briefly. Of course, the size and shape of the aorta differs
significantly among individuals and even within the same individual
at different ages. The thoracic aorta may be divided into 5
segments: the aortic root, the ascending aorta, the proximal aortic
arch, the posterior aortic arch, and the descending thoracic
The aortic root is the short segment of the aorta arising from
the base of the heart and containing the valve, the annulus, and
the sinuses of Valsalva. The right and left coronary arteries arise
from the right and left sinuses of Valsalva, respectively. The
third, posteriorly located, sinus of Valsalva is termed the
noncoronary sinus. The normal diameter of the adult aorta just
cranial to the root averages 3.6 cm (range 2.4 to 4.7 cm).
The ascending aorta extends from the root to the origin of the
right brachiocephalic artery. The average diameter of the adult
ascending aorta is 3.5 cm (range 2.2 to 4.7 cm).
The aortic arch begins at the origin of the right
brachiocephalic artery and ends at the attachment of the ligamentum
arteriosum. It is divided into two segments: the proximal arch and
the posterior (distal) arch. The proximal arch extends from the
origin of the right brachiocephalic artery to the origin of the
left subclavian artery, also giving rise to the left common carotid
artery. The posterior or distal arch encompasses that segment of
the arch from the origin of the left subclavian artery to the
ligamentum arteriosum and is also known as the aortic isthmus. This
segment is occasionally slightly narrower than the proximal
descending thoracic aorta, particularly in infants.
The descending thoracic aorta begins after the ligamentum
arteriosum and extends to the aortic hiatus in the diaphragm. The
most cranial aspect of the descending thoracic aorta may appear
slightly dilated and is known as the "aortic spindle." This finding
is more commonly encountered in children than adults. The
mid-portion of the descending thoracic aorta has an average
diameter of 2.48 cm (range 1.6 to 3.7 cm). The distal descending
thoracic aorta (just above the diaphragm) has an average diameter
of 2.42 cm (range 1.4 to 3.3 cm).
Obviously, the average aortic diameters are less in children.
Normal aortic diameter measurements for children have been
The ascending aorta is normally always larger than the descending
Generally, the aortic wall is only a few millimeters thick, and
usually is not separable from the unenhanced aortic blood pool.
Occasionally, the wall of the aorta is visible in anemic patients,
appearing as a thin rim of increased attenuation that is smooth and
of uniform thickness around the circumference of the vessel.
Thoracic aorta: Normal variants
Congenital anomalies of the aorta and the great vessels are
beyond the scope of this review. The frequency and imaging
appearance of such anomalies have been well described elsewhere.
Normal variations in the aortic contour that will be discussed
include the aortic spindle, ductus diverticulum, branch vessel
infundibula, and pseudocoarctation.
Aortic spindle--As mentioned previously, the aortic spindle is
the region of the posterior aortic arch located between the origin
of the left subclavian artery and the ligamentum arteriosum. It
appears as a circumferential bulge below the region of the isthmus
Ductus diverticulum--The ductus diverticulum is the term applied
to a focal, convex bulge along anterior undersurface of the isthmic
region of the aortic arch (figure 2). The major significance of
this structure lies in distinguishing it from a posttraumatic
pseudoaneurysm. The ductus usually is a smooth convexity that
creates obtuse angles with the aortic wall, as opposed to the acute
angles created by a pseudoaneurysm. Additionally, a ductus
diverticulum does not cause delayed contrast washout, as might a
Branch vessel infundibula--Infundibula of aortic branch vessels,
including the brachiocephalic artery, left common carotid artery,
left subclavian artery, and intercostal arteries may simulate
traumatic injuries or aneurysms. Infundibula are recognized by
their anatomic configuration and smooth margins and by the presence
of a vessel emanating from the apex of the infunibulum. Infundibula
also tend to occur in characteristic locations, commonly affecting
the left subclavian artery and the third right intercostal artery.
Pseudocoarctation--Pseudocoarctation results from elongation of
the aortic arch and kinking at the site where the aorta is tethered
by the ligamentum arteriosum (figure 3). Although the configuration
of the aorta in pseudocoarctation somewhat resembles true
coarctation, because there is no pressure gradient across this
region, collateral circulation does not develop.
Pitfalls in helical CT aortography: Normal anatomic
Normal pericardial recesses, the left brachiocephalic vein, the
left inferior pulmonary vein, the left superior intercostal vein,
the right atrial appendage, and normal thymus constitute the most
frequently encountered normal anatomic structures that may simulate
aortic pathology such as dissection or acute traumatic aortic
Pericardial recesses--The preaortic and retroaortic portions of
the superior pericardial recesses are intimately related to the
ascending aorta and may occasionally simulate dissection or
The preaortic recess is found along the anterior aspect of the
ascending aorta, and the retroaortic portion of the superior
pericardial recess is located posterior to the ascending aorta near
the level of the left pulmonary artery. The homogeneous water
attenuation and characteristic locations are clues to the proper
Left inferior pulmonary vein--The left inferior pulmonary vein
is in close proximity to the descending thoracic aorta near the
former's entry into the left atrium. Where it is in close contact
with the aorta, this structure may mimic dissection of the
descending thoracic aorta. Knowledge of the normal appearance and
course of this vessel and following it superiorly to its
termination in the left atrium ensure the proper diagnosis.
Left brachiocephalic vein--The unopacified left brachiocephalic
vein may simulate a mediastinal hematoma in the setting of trauma.
When opacified and low-lying, the left brachiocephalic vein may
occasionally mimic a Stanford type A dissection of the ascending
Left superior intercostal vein--Occasionally, the left superior
intercostal vein is visible adjacent to the left lateral aspect of
the aortic arch as it courses anteriorly from the left-sided
accessory hemiazygos vein to the left brachiocephalic vein (figure
4). When the left upper extremity is injected, this vessel may
enhance vigorously and can mimic dissection. The focal nature of
the abnormality as well as knowledge of the normal course of this
vessel will allow the proper diagnosis.
Right atrial appendage--The right atrial appendage is normally
visible on helical CT aortography anterior to the proximal
ascending aorta just above the aortic root, where it may simulate
aortic pathology (figure 5). Following this structure into the
right atrium on contiguous images will demonstrate its true nature.
Thymus--The thymus is normally located anterior to the aortic
arch and ascending aorta. In younger patients, residual thymus may
resemble hematoma or a false lumen of an aortic dissection. The
characteristic location and triangular configuration in a younger
patient are clues to the correct diagnosis.
Although the lack of mass effect and focal nature of the thymus
help to distinguish it from mediastinal hematoma, differentiating
the normal thymus from mediastinal hematoma in pediatric trauma
patients may be difficult.
Pitfalls in helical CT aortography: Periaortic
Periaortic pathology such as medial left lower lobe atelectasis,
left pleural effusion, mediastinal masses or adenopathy, and
pericardial effusion may simulate aortic dissection,
intramural hematoma, or mediastinal hematoma. Additionally, anemia
may render the aortic wall visible and may resemble intramural
Left lower lobe atelectasis--Enhanced, atelectatic lung
immediately adjacent to the descending thoracic aorta may simulate
dissection (figure 6). Additionally, atelectatic lung in this
location on unenhanced scans may resemble intramural hematoma. Lack
of a constant relationship to the aorta as well as the presence of
air bronchograms visible on lung windows
will suggest the proper diagnosis.
Left pleural effusion--Left pleural effusion may resemble
periaortic hematoma, particularly in the acutely injured patient.
Usually effusion is apparent elsewhere, and occasionally
attenuation measurements may be useful for distinguishing effusion
from aortic pathology. When required, decubitus views may allow the
Mediastinal masses and adenopathy--Mediastinal masses and
adenopathy may mimic mediastinal hematoma in the trauma patient.
Usually the focal and lobulated appearance and location of
mediastinal masses and adenopathy will suggest the correct
Pericardial effusion--Pericardial effusion may simulate aortic
dissection or mediastinal hematoma. Pericardial fluid is usually of
homogeneous low attenuation, unless complicated by hemorrhage.
Knowledge of the normal pericardial recesses also aids in
Anemia--The aortic wall may become visible as a thin, uniform
ring of high attenuation in anemic patients (figure 7).
This appearance may occasionally cause diagnostic difficulty on
unenhanced scans because it may resemble intramural hematoma. The
smooth, uniform, homogeneous appearance and absence of periaortic
fat stranding or hematoma as well as the tendency to primarily
affect the ascending aorta allow the proper diagnosis.
Normal aortic anatomy: Technical difficulties
Poor contrast enhancement--Poor contrast enhancement may result
from improper contrast bolus timing due to a wide variety of
factors, including poor cardiovascular function, stenosis of the
injected vein, or slow administration of contrast media. Poor
vascular opacification could obscure the diagnosis of dissection
A timing bolus or bolus timing software are useful techniques for
ensuring proper contrast delivery.
kVp and mA--As discussed previously, kVp, section thickness,
FOV, and, especially, mA have a profound influence on image
quality. Careful attention to the technical parameters of the scan
can improve image quality by reducing noise.
Streak artifacts--Streak artifact results from dense, highly
concentrated contrast causing beam hardening and obscuring photon
transmission. When high-density contrast material markedly reduces
photon transmission, the scanner detectors may record no
transmission. CT scanner reconstruction programs do not recognize
this phenomenon, and streak artifacts will result. Streak artifact
commonly affects the left brachiocephalic vein and superior vena
cava, potentially obscuring the aortic arch and ascending aorta.
Streak artifacts may also result from pacemakers, surgical staples,
external monitoring devices, or positioning the patient's arms at
their sides during the scan.
These artifacts are readily recognized by the radial orientation of
the artifact and by observing that the dense contrast column is the
source of the artifact (figure 9). Streak artifacts may be reduced
by using dilute contrast mixtures, scanning caudal to cranial, or
injecting via the right upper extremity.
Motion artifact--Motion artifacts can be quite problematic,
particularly at the base of the heart. Motion artifacts may either
obscure the diagnosis or simulate the appearance of aortic
dissection (figure 10). Noting motion elsewhere on the scan, as
well as the fact that the artifact is usually absent or markedly
different on immediately adjacent levels, may allow recognition of
Aortic pathology: Atherosclerotic vascular
Aortic atherosclerotic vascular disease (ASVD) is increasingly
recognized as a source of stroke and visceral and lower extremity
ischemia resulting from peripheral embolization.
Ulcerated atherosclerotic plaques in the ascending aorta and aortic
arch have been noted with increased frequency in patients with
stroke or death due to cerebrovascular disease,
and contribute to cerebrovascular events among patients undergoing
Features of aortic ASVD include smooth intimal plaques, ulcerated
plaques, calcified plaques, mobile thrombi, and "protruding"
atheromas (figure 11).
Plaques thicker than 4 mm have been associated with an increased
risk of stroke.
While calcified plaque may be relatively stable, uncalcified or
ulcerated plaque may pose an increased risk for embolization.
Thrombi may be superimposed on atherosclerotic plaques; in
particular, protruding atheromas are prone to embolize.
Such thrombi may resolve with anticoagulation.
HCTA nicely demonstrates the location and morphology of
atherosclerotic plaque. Atherosclerotic plaque typically appears as
low attenuation material internal to the aortic wall (figure 11).
Calcified plaque is readily apparent, and plaque ulceration may
also be seen. Although TEE is used more commonly for the specific
investigation of aortic ASVD, disease in the cranial aspect of the
ascending aorta and arch may be difficult to identify with TEE. In
contrast, HCTA demonstrates plaque in the ascending aorta and
aortic arch well.
Aortic pathology: Aortic dissection
Aortic dissection results from a primary tear in the intima of
the aorta that allows blood to gain access to the aortic media;
blood may then propagate proximally and/or distally within the
media, creating a false lumen. The false lumen may or may not
re-enter the true aortic lumen at a point removed from the primary
intimal tear. The intimal tear itself may be a primary inciting
factor in aortic dissection
or it may be the result of primary weakening of the aortic media
due to a spontaneous intramural hematoma.
Aortic Dissection: Etiologies--Predisposing factors for aortic
dissection include hypertension, aortic valve disease, coarctation
of the aorta, connective tissue disorders (e.g., Marfan's syndrome,
Ehlers-Danlos syndrome), aortic aneurysm, cystic medial necrosis,
cardiovascular surgery, infections (syphilis and bacterial), and
non-infectious causes of arteritis.
Clinical evaluation--Patients with aortic dissection often
present with severe chest pain that may mimic myocardial infarction
or pulmonary embolism. Pulmonary edema resulting from cardiac
failure may occur due to acute aortic insufficiency. Cardiac
tamponade may occur with intrapericardial rupture of the
dissection. Neurologic sequelae resulting from aortic branch vessel
occlusion, such as paraplegia, stroke, and syncope, may be
presenting manifestations of aortic dissection. Although uncommon,
branch vessel occlusion may cause mesenteric or limb ischemia to be
the initial manifestation of aortic dissection. Differential blood
pressures may be noted on physical examination.
Classification schemes--Two well known classification schemes
for the extent of aortic dissection exist: the DeBakey
and Stanford classifications. DeBakey type I dissections (figure
9B) begin in the ascending aorta and extend for a variable distance
beyond the aortic arch. DeBakey type II dissections involve the
ascending aorta only, and DeBakey type III dissections involve the
descending thoracic aorta beginning distal to the left subclavian
artery. The Stanford classification scheme primarily considers
involvement of the ascending aorta and has largely replaced the
DeBakey classification. A Stanford A dissection involves the
ascending aorta, with or without descending thoracic aortic
involvement. Stanford B dissection is confined to the descending
thoracic aorta, beyond the origin of the left subclavian artery.
Classification is important because most Stanford A (DeBakey type I
and II) dissections are treated surgically,
whereas Stanford B (DeBakey type III) dissections are usually
managed medically, provided evidence of end-organ ischemia is not
of HCTA imaging are to:
1) identify the intimal flap (figures 12-19, 21);
2) identify any branch vessel involvement (great vessels,
mesenteric and renal arteries) (figures 13 and 14);
3) identify the presence of pericardial fluid (intrapericardial
rupture) (figure 15) or periaortic hematoma (aortic rupture)
4) identify extent of dissection;
5) evaluate size of the aorta (figure 17); and
6) evaluate the patency of the false lumen and degree of true
lumen compression (figures 18 and 19).
HCTA findings of aortic dissection--The HCTA diagnosis of aortic
dissection is based on findings on both unenhanced and enhanced
scans. Unenhanced scans are valuable for demonstrating aortic
intramural hematomas (figure 20). Intramural hematomas (IM) appear
as focal, eccentric thickening of the aortic wall that is high in
Unenhanced scans may also demonstrate displacement of aortic
intimal calcifications (figure 20). However, this finding may be
difficult to assess in the presence of marked aortic tortuosity.
Furthermore, the displaced intimal calcifications of aortic
dissection are often not easily distinguished from calcified
atherosclerotic plaque and thrombus, nor does this sign allow one
to reliably distinguish acute from chronic dissections.
The single best criterion for the diagnosis of aortic dissection
on enhanced HCTA scans is the demonstration of an intimal flap
separating the true and false lumens (figures 12- 19, 21). Contrast
enhanced studies are also valuable for demonstrating the patency of
branch vessels as well as revealing differential flow rates between
the true and false lumens (figures 17 and 18). Enhanced scans also
are useful for estimating the size of the aorta (figure 17).
Occasionally, it can be difficult to distinguish true from false
lumen. The most specific indicator for the false lumen is the
presence of irregular strands within the lumen, known as "cobwebs."
Cobwebs represent residual fragments of aortic media separated
during the dissection.
Frequently, the true lumen is located anteriorly in the descending
thoracic aorta (figures 12, 17-19). Often the true lumen is smaller
than the false lumen due to compression by the latter (figures 12,
14-19). It has been reported that the true lumen may assume a
concave orientation with respect to the false lumen when the latter
generates relatively higher pressure, a situation that is
associated with an increased likelihood of impaired end-organ
perfusion (figure 19).
Older data suggests that the intimal flap is visualized on
dynamic CT in only 63% to 70% of cases;
this figure may be higher with the rapid volumetric acquisitions
possible with HCTA.
In cases when an intimal flap is not seen, the diagnosis of aortic
dissection must be made on the basis of ancillary criteria, such as
distortion of the aortic contour, intramural high attenuation
(figure 20), periaortic hematoma (figures 16 B and 16C), or
displaced intimal calcifications (figure 20).
HCTA sensitivity and specificity for aortic dissection--Dynamic,
contrast-enhanced CT scanning (a non-volumetric technique) has
demonstrated a sensitivity of 93.8% and a specificity of 87.1% for
the diagnosis of aortic dissection.
Recently, Sommer et al,
comparing HCTA with TEE and MRI, demonstrated sensitivities of 100%
for all three techniques. Specificity was 100%, 94%, and 94% for
HCTA, TEE, and MRI, respectively. Disadvantages of HCTA compared
with these other modalities include the inability to demonstrate
aortic insufficiency or involvement of the coronary arteries.
Multiplanar reformations and rendering techniques (figures 12B, 17,
and 21) are useful for demonstrating dissections in a format more
familiar to clinicians and surgeons, although such techniques
rarely add information that is not apparent on the axial
Aortic pathology: Aortic aneurysms
Aneurysm of the aorta is defined as dilation of the aorta
involving all three wall layers equaling or exceeding twice the
Pseudoaneurysms (false aneurysms), however, represent saccular
dilations that do not contain aortic intima. Typically, such
aneurysms are associated with blunt chest trauma or penetrating
atherosclerotic ulcers or, less commonly, with infection or
Etiologies of true thoracic aortic aneurysms include
atherosclerosis, infection (mycotic aneurysms), and cystic medial
necrosis (annuloaortic ectasia).
Certain etiologies of thoracic aortic aneurysms tend to be
associated with particular locations. For example, causes of
annuloaortic ectasia (figure 22) and syphilis usually affect the
ascending aorta, whereas atherosclerotic aortic aneurysms (figure
23) most often affect the descending thoracic aorta.
Growth rate and risk of rupture--Thoracic aortic aneurysms have
been associated with a growth rate of 0.42 cm/year as opposed to
the 0.25 cm/ year growth rate suggested for abdominal aortic
More than 25% of thoracic aortic aneurysms are accompanied by
infrarenal abdominal aortic aneurysms.
The most dreaded consequence of thoracic aortic aneurysm is rupture
(figure 24). The risk of rupture increases with increasing aneurysm
size, with ruptured ascending aortic aneurysms often measuring
>6 cm in diameter.
The frequency of rupture varies from 40% to 70% depending on the
and 30% to 50% of deaths related to thoracic aortic aneurysms are
due to rupture.
Because elective thoracic aortic aneurysm repair is associated with
a lower mortality than emergent repair, (9% versus 22%,
and the risk of rupture increases with increasing size of the
aneurysm, size criteria for operative intervention have been
suggested. Coady et al
noted that the median size of rupture or dissection of ascending
and descending aortic aneurysms was 5.9 cm and 7.2 cm,
respectively. Therefore, these authors advocate operative
intervention for thoracic aortic aneurysms when aneurysm size
exceeds 5.5 cm for ascending and 6.5 cm for descending thoracic
CT evaluation of thoracic aortic aneurysms--HCTA accurately
identifies the extent and size of thoracic aortic aneurysms.
Atherosclerotic aortic aneurysms typically appear as a fusiform
dilation of the aorta, commonly affecting the descending aorta.
Causes of cystic medial necrosis, (especially Marfan's syndrome and
Ehler's-Danlos syndrome), result in an aorta with a conical
appearance that tapers superiorly; this appearance has been likened
to the shape of a pear. The sinuses of Valsalva are often involved
in such cases. Saccular aortic aneurysms may be due to infections
(mycotic aneurysms) (figure 25). The morphology of a saccular
aneurysm must always raise the possibility of pseudoaneurysm.
HCTA clearly defines the extent of the enhancing lumen as well
as the degree and morphology of associated thrombus. HCTA also
clearly delineates the degree and location (intimal or within the
thrombus itself) of calcification of the aneurysm. The local effect
of thoracic aortic aneurysms (compression or erosion of adjacent
structures) is also clearly defined with HCTA. Branch vessel
involvement is also clearly depicted with HCTA.
The HCTA findings of rupture of thoracic aortic aneurysms
include high attenuation fluid in the pleural or pericardial spaces
or within the wall of the aorta. Mediastinal fat stranding
representing hematoma also suggests the possibility of aneurysm
rupture (figure 24). The "draped aorta" sign has been suggested as
a sign of an early, contained rupture. This sign is considered
present when the posterior wall of the aorta is either not
identifiable as distinct from adjacent structures or when it is
closely applied to and follows the contour of the adjacent
Multiplanar reformations and rendering techniques--Multiplanar
reformations, shaded surface displays, and volume-rendering
techniques (figures 1B to 3, 11C, 12B, 17B, 21, and 25B) are useful
methods for providing alternate formats for viewing large data
sets. Such techniques are particularly valuable for demonstrating
pathology to nonradiologists in a more familiar format (i.e., a
sagittal oblique view that resembles conventional angiography).
However, such techniques rarely change the initial interpretations
of the axial source images.
Aortic pathology: Intramural hematoma
Intramural hematoma (IMH) represents localized hemorrhage within
the aortic media. Several methods for the formation of IMH have
been proposed: 1) rupture of the vasa vasorum resulting in
weakening of the aortic wall; 2) spontaneous thrombosis of the
false lumen of an aortic dissection; and 3) penetrating
atherosclerotic ulcer induced by rupture of an intimal
atherosclerotic plaque, allowing blood to gain access to the aortic
It is unclear which mechanism predominates. Indeed, it is likely
that all of the above mechanisms play some role in the development
CT of IMH--On CT scans, IMH appears as crescenteric, high
attenuation material within the wall of the aorta (figures 20, 26
to 28). Although the hyperattenuation may be perceived on
postcontrast images (figures 26B to 28), the finding is usually far
better appreciated on unenhanced images (figures 20 and 26A). Thus,
protocols for aortic imaging should begin with noncontrast helical
images of the chest. IMH may present as focal thickening of the
aortic wall with internal displacement of intimal calcifications
(figures 20 and 28). Unlike atherosclerotic plaque, IMH generally
creates a smooth margin with the contrast-enhanced aortic lumen
(compare figures 11 and 26, 28).
Natural history of IMH--The natural history of IMH is not
completely clear. Several fates may befall an intramural aortic
hematoma. Pathologically, IMH, no matter the mechanism, results in
weakening of the aortic wall; this weakening may predispose to
dissection or frank rupture. This assertion is evidenced by several
series demonstrating IMH associated with either hematoma within the
mediastinum or pericardial sac,
or conversion of IMH to either imaging or surgically proven
aneurysms. Factors that predict the conversion of IMH to
dissection, rupture, or saccular aneurysm are unclear, and this
uncertainty results in ambiguity regarding proper management.
Several series indicate that IMH in the ascending aorta (figure 26)
portends a poor prognosis and thus requires urgent surgical repair.
IMH involving the descending thoracic aorta is managed as a type B
IMH that is not managed surgically should be periodically followed
for early detection of complications such as saccular aneurysm,
pseudoaneurysm, or classic dissection. Several authors have
advocated imaging every 3 months in the first year after diagnosis
and continued long-term follow-up to detect the development of
Aortic pathology: Penetrating atherosclerotic ulcer of
the thoracic aorta
Penetrating atherosclerotic ulcer (PAU) occurs when
atherosclerotic plaque penetrates through the intima and internal
elastic membrane of the aorta, allowing blood to gain access to the
aortic media. PAU results in IMH (figure 27), and can progress to
pseudoaneurysm (figure 29),
or frank aortic rupture.
Clinical presentation of PAU--PAU, like dissection, may be
associated with chest or back pain,
although this entity is most often detected in asymptomatic
patients imaged for some other reason. Patients with PAU are
usually older than patients with dissection.
PAU is most often located in the descending thoracic aorta,
although they may present in the arch or, less commonly, the
HCTA of PAU--PAU manifests on unenhanced helical CT as an
intramural hematoma (figure 27). Aortic wall thickening may be
and displaced intimal calcifications
may also be present. Enhanced images demonstrate a focal contrast
collection projecting beyond the confines of the lumen of the aorta
Aortic wall enhancement
and, rarely, active contrast extravasation, may be encountered.
PAU is frequently multiple, and evidence of complications, such as
and pseudoaneurysm (figure 29),
may coexist. PAU may also be associated with aortic rupture,
evidenced by pericardial or pleural high attenuation fluid
or mediastinal hematoma.
Natural history of PAU--Early surgical intervention is advocated
for patients with PAU of the ascending aorta or cases of
complicated PAU who are operative candidates. Conservative therapy
is generally employed for patients with PAU of the descending
thoracic aorta that is stable or asymptomatic,
or in patients who are too unstable for operative intervention.
It should be understood that surgery for PAU is generally more
extensive than that for dissection.
For cases that are managed conservatively, imaging surveillance is
recommended for early detection of potential complications such as
saccular or fusiform aortic aneurysm, dissection, or rupture.
Aortic pathology: Aortitis
Aortitis usually refers to inflammatory condition affecting the
aortic wall. Takayasu's arteritis is the most common such
inflammatory disorder, although Reiter's syndrome, psoriasis,
ankylosing spondylitis, relapsing polychondritis, giant cell
arteritis, and infections (particularly syphilis) are other
Takayasu's arteritis is a vasculitis that results in
inflammation involving the media and adventitia, with reactive
change in the intima.
This condition is commonly associated with, though not limited to,
young Asian women and affects women 8 to 10 times as often as men.
This disease is may be classified based on the pattern of anatomic
1) aortic arch and branch vessels; 2) segmental involvement of the
thoracic and abdominal aorta and branch vessels; 3) a combination
of types I and II; and 4) involvement of the pulmonary
Others classification schemes exist.
Takayasu's arteritis typically results in multifocal stenoses and
although aneurysms (fusiform more than saccular) may occur.
Aortic wall thickening and enhancement has been reported as a
manifestation of early disease, whereas calcification of the aortic
wall, vascular occlusions, and collateral vessel formation are
generally late findings.
Aortic pathology: Mycotic aneurysms
Mycotic aneurysms are the result of infection of the aortic wall
and occlusion of the vasa vasorum by septic emboli that results in
weakening of the aortic wall (figure 25), predisposing to aneurysm
formation. Etiologies of mycotic aneurysm include septic emboli
from intravenous drug use and in-dwelling catheters, infection of
prosthetic valves, or infection of atherosclerotic plaque. Common
infectious agents include Salmonella (figure 25), Staphylococcus
aureus, streptococci, tuberculosis, and fungi, such as Candida and
Mycotic aneurysms are commonly saccular (figure 25) and may grow
rapidly. They may be associated with surrounding periaortic fat
infiltration and inflammatory changes or even gas formation. A
mycotic etiology of a saccular aneurysm is often suggested when an
aneurysm is found in an atypical location. However, the most common
location is the undersurface of the aortic arch, near the
where a mycotic aneurysm may be mistaken for a post-traumatic
pseudoaneurysm. Mycotic aneurysms are prone to rupture.
Aortic pathology: Acute traumatic aortic injury
Acute traumatic aortic injury (ATAI) is usually the result of
severe deceleration injury, such as high speed motor vehicle
accidents (MVA) or falls from significant heights. ATAI accounts
for 10% to 20% of fatalities in MVAs and results in immediate death
in 80% to 90% of cases.
Among patients with ATAI who survive transport from the initial
trauma scene, the mortality of untreated ATAI is high: 30% die in
the ensuing 6 hours,
72, 77, 78
nearly 50% by 24 hours,
and 90% within the following 4 months.
Chronic pseudoaneurysm may result in the small fraction of
untreated long-term survivors.
Several mechanisms for ATAI have been postulated. Theories such
as the "osseous pinch," the "water-hammer effect," and horizontal
and/or vertical deceleration have been proposed. The osseous pinch
theory holds that the aorta is "pinched" between the bones of the
anterior thorax and the spine, resulting in injury. The
"water-hammer" theory suggests that blunt trauma causes aortic
compression, resulting in an acute and dramatic rise in
intravascular pressure. Horizontal and vertical deceleration
mechanisms cause injury by creating bending forces and shear stress
at sites where the aorta is relatively fixed (especially the
isthmus). All of these possibilities may play some role in any
The most common site of injury is the aortic isthmus (90%),
followed by the ascending aorta (5% to 10%), and the descending
thoracic aorta near the diaphragmatic hiatus (1% to 3%).
Great vessel injuries may coexist. Pathologically, a nearly
circumferential laceration is usually present, although partial
tears may occur.
HCTA of ATAI--HCTA findings of ATAI are classified as direct or
indirect (figures 30 to 34). One of the most commonly encountered
direct signs of HCTA is a pseudoaneurysm (figure 30). Other direct
findings include abnormal aortic contours or abrupt caliber
changes, pseudocoarctation, occlusion of a segment of aorta (figure
31), and an intimal flap. Indirect findings of ATAI include
mediastinal or retrocrural hematomas. Though indirect findings
suggest the possibility of ATAI, they may be the result of
mediastinal venous bleeding. When a mediastinal hematoma is
encountered, its relationship to the aorta is of paramount
importance (figures 32 and 33). Hematomas that obliterate the fat
plane surrounding the aorta or great vessels remain suspicious for
occult ATAI and are usually evaluated with aortography (figure 32).
Mediastinal hematomas that do not directly contact the aorta or
great vessels usually represent mediastinal venous bleeding, and
aortography is not required (figure 33). In the rare circumstance
that a patient survives undiagnosed ATAI, a chronic pseudoaneurysm
may result (figure 34).
Overall, HCTA is 100% sensitive and 82% specific for the
diagnosis of ATAI.
Pate et al
recently evaluated 6169 patients with blunt chest trauma, of whom
47 were diagnosed with ATAI. None of the cases in which the CT scan
was read as "normal" were associated with proved ATAI, and the use
of HCTA resulted in significantly reduced utilization of
The postoperative aorta
Aortic aneurysms and dissections are often repaired using one of
two general methods: an interposition graft or an inclusion graft.
Use of the interposition graft technique implies resection of the
pathologic aortic segment and reconstruction of the vessel with
graft material, commonly made of Dacron. Composite interposition
Dacron grafts containing a mechanical aortic valve may be used when
When an interposition graft is used, necessary vasculature, such as
the coronary arteries, is reimplanted into the graft. Occasionally,
high attenuation felt rings (figures 35 and 36) are employed to
reinforce the anastamosis of interposition grafts; these rings
indicate the site of anastamosis.
If such rings are not used, the site of the anastomosis may be
surmised by noting an abrupt change in caliber of the aorta or an
abrupt change in an atherosclerotic native aorta versus a
Felt pledgets are also sometimes employed with interposition grafts
to reinforce the sites of cannula placement. These felt pledgets,
like felt rings, are high in attenuation and can resemble a
pseudoaneurysm to the unwary.
The coronary artery anastomoses represent another pitfall of
composite interposition grafts that may simulate pathology. Often
the coronary arteries are reimplanted into the aortic graft with a
portion of native aortic root known as coronary buttons. These
buttons may occasionally appear somewhat prominent and may simulate
a pseudoaneurysm if the true nature of the finding is not
The use of inclusion graft technique entails insertion of graft
material within the native, diseased aorta. This form of
reconstruction creates a potential space between the graft material
This potential space may thrombose or contain flowing blood. Blood
flow within the perigraft space does not mandate surgery unless the
situation is associated with hemodynamic instability.
Serious postoperative complications include dehiscence of the
surgical suture line, more commonly at the proximal site.
Dehiscence may result in pseudoaneurysm formation. Pseu-doaneurysm
at the site of coronary artery reimplantation may result in
myocardial ischemia and infarction. Aneurysm formation and
dissection may also occur, particularly in patients with cystic
medial necrosis. CT or MR surveillance of aortic grafts is
therefore routinely recommended.
Innocuous findings commonly encountered following aortic repair
include left lower lobe atelectasis, pleural and pericardial
effusions, and mediastinal adenopathy. These findings usually
With interposition grafts, low attenuation material within the
mediastinum surrounding the graft is often encountered, even months
or years following surgery. This material may represent hematoma
organizing into fibrous tissue, and it may never completely resolve
in some patients.
An intimal flap is often identified beyond the graft site in
patients following dissection repair because surgical
reconstruction frequently intentionally maintains blood flow to
both the true and false lumens. If the false lumen continues to
enlarge, a pseudoaneurysm of the false lumen may result.
Helical CT aortography is a rapid and accurate method for the
evaluation of a wide variety of diseases affecting the thoracic
aorta. Recent advances in CT technology have mandated careful
attention to scan acquisition parameters. Proper scan methodology
will ensure excellent scan quality, and an awareness of diagnostic
pitfalls and the imaging findings of aortic diseases should allow
an accurate scan interpretation in the vast majority of cases.