Wittram C h e s t I m a g i n g • Pe r s p e c t i v e

CT Pulmonary Angiography

How I Do It: CT Pulmonary

Angiography
Conrad Wittram1 OBJECTIVE. The purpose of this article is to describe the techniques to improve motion ar-
tifacts, vascular enhancement, flow artifacts, body habitus image noise, vascular opacification in
Wittram C parenchymal lung disease, streak artifacts, and the indeterminate CT pulmonary angiogram. In
addition, this article will illustrate the diagnostic criteria of acute and chronic pulmonary emboli.
CONCLUSION. Pulmonary embolism is the third most common acute cardiovascular
disease, after myocardial infarction and stroke, and it leads to thousands of deaths each year
because it often goes undetected. For the more than 25 years that the direct signs of pulmonary
American Journal of Roentgenology 2007.188:1255-1261.

embolism have been available to the radiologist on CT, this noninvasive technique has produced
a paradigm shift that has raised the standard of care for patients with this disease.

ulmonary embolism is the third raphy have a sensitivity of 90% and specificity

P most common acute cardiovascu-

lar disease, after myocardial in-
farction and stroke, and results in
of 95% for the detection of venous thromboem-
bolic disease [7]. The PIOPED II study found
that patients with a low or intermediate clinical
an estimated 200,000–300,000 hospitaliza- probability of pulmonary embolism and normal
tions and 37,000–44,000 deaths per year in the results on CTPA had a high negative predictive
United States [1]. In 1980, Godwin et al. [2] value for PE (96% for patients with a low prob-
were among the first to describe pulmonary ability and 89% for patients with an intermedi-
embolism on contrast-enhanced CT. In 1990, ate probability); however, the negative predic-
the Prospective Investigation of Pulmonary tive value was 60% in patients with a high
Embolism Diagnosis (PIOPED) study results probability before CTPA. The positive predic-
were published [3]. This large multicenter trial tive value of abnormal findings on CTPA was
compared ventilation–perfusion (V/Q) scintig- high (92–96%) in patients with an intermediate
raphy with pulmonary angiography and estab- or high clinical probability but much lower
lished the diagnostic characteristics of pulmo- (58%) in patients with a low likelihood of pul-
nary embolism on V/Q scintigraphy. The monary embolism. Therefore, additional test-
sensitivity of V/Q scintigraphy was found to be ing is recommended when the clinical probabil-
Keywords: chest, CT arteriography, CT technique, 98%, with a specificity of 10% [3]. The poten- ity is inconsistent with the imaging results [7].
embolism tial of the noninvasive technique, CT pulmo- A limitation of the PIOPED II study was
nary angiography (CTPA), has now been real- that the composite gold standard was not 100%
DOI:10.2214/AJR.06.1104 ized at most institutions; it has become the test accurate for the diagnosis of venous throm-
Received August 18, 2006; accepted after revision
of choice and thus the de facto standard of care boembolic disease; it therefore follows that the
November 8, 2006. [4]. Recent studies have shown the sensitivity performance of CT was likely better than the
of thin-slice MDCTPA to be 90–100% and the results indicate. In the PIOPED II study,
1Department of Thoracic Radiology, Massachusetts specificity to be 89–94% for the detection of among 824 patients with a reference diagnosis
General Hospital, Founders 202, 55 Fruit St., Boston, pulmonary emboli to the level of the subseg- and a completed CT study, CTPA was incon-
MA 02114. Address correspondence to C. Wittram.
mental arteries, using pulmonary angiography clusive in 51 because of poor image quality [7].
CME as the gold standard [5, 6]. A recent study that evaluated the causes of in-
This article is available for CME credit. See www.arrs.org A much larger multicenter study has been re- determinate CTPA findings found an indeter-
for more information. cently published: The PIOPED II study, which minate rate of 6.6% [8]. The most common
AJR 2007; 188:1255–1261
used a composite gold standard, showed that cause was motion artifacts in 74% of the cases;
CTPA has a sensitivity of 83% and specificity other reasons included poor enhancement
0361–803X/07/1885–1255
of 96% for the detection of pulmonary embo- (40%), patient habitus (7%), parenchymal dis-
© American Roentgen Ray Society lism and that combined CTPA and CT venog- ease (12%), and streak artifacts (7%) [8]. The

AJR:188, May 2007 1255

 Wittram

TABLE 1: 16-MDCT Pulmonary Angiography Protocol can be a cause of misdiagnosis of pulmonary

Parameter Normal-Sized Patient Large Patient (> 250 lb [113 kg]) embolism. They are best seen on lung window
Detector width/reconstruction (mm) 1.25/1.25 2.5/1.2.5 settings that show composite images of vessels
[11]. A rapid change in position of vessels on
Table speed/rotation (mm) 13.75 6.88
contiguous images also confirms motion artifact.
Pitch 1.375:1 0.562:1 A low-density abnormality that simulates pul-
Peak kilovoltage 140 140 monary embolism may result from partial vol-
Milliamperes 380 380 uming of vessel and lung [11]. Motion artifact
Rotation time (s) 0.5 1.0 renders the diagnosis of pulmonary embolism at
the affected anatomic level indeterminate. The
Algorithm Standard Standard
frequency of examinations devoid of motion ar-
Scanning field of view Large Large
tifacts is significantly higher for MDCT, which
Display field of view Rib to rib Rib to rib has a shorter breath-hold than single-detector CT
[14, 15]. At the moment, the breath-hold re-
quired for 16-MDCT is approximately 10 sec-
TABLE 2: 64-MDCT Pulmonary Angiography Protocol onds, and for 64-MDCT, less than 3 seconds. In
Parameter Normal-Sized Patient Large Patient (> 250 lb [113 kg]) dyspneic patients, oxygen supplementation can
Detector width/reconstruction (mm) 0.625/1.25 0.625/1.25/2.5 help the patient provide the desired period of ap-
nea. The implementation of higher order MDCT
Table speed/rotation (mm) 55 55
scanners should lower the indeterminate CTPA
Pitch 1.375:1 1.375:1
rate due to respiratory motion.
American Journal of Roentgenology 2007.188:1255-1261.

Peak kilovoltage 140 140

Milliamperes 380 380 Pulmonary Artery Enhancement
Rotation time (s) 0.5 1.0 Theory
Algorithm Standard Standard
An increase in the attenuation of blood on
CT may be obtained with intravascular con-
Scanning field of view Large Large
trast material containing the atoms of iodine or
Display field of view Rib to rib Rib to rib gadolinium. Previous work has defined the at-
tenuation values of acute and chronic pulmo-
purpose of this article is to describe the tech- (window width, 700 H; window level, 100 H) nary emboli [16]. Combining these values with
niques used to improve the quality of CT pul- settings because pulmonary embolism can be experimental work by Meaney et al. [17], it is
monary angiography and to illustrate the diag- missed when a case with very bright contrast is possible to calculate the minimum amount of
nostic criteria of acute and chronic pulmonary viewed only on mediastinal window settings IV attenuation required to perceive pulmonary
emboli. Indirect CT venography will not be [11]. The pulmonary embolism–specific set- emboli on CT. Meaney et al. showed that the
dealt with in detail in this article. tings also help to differentiate between a sharp detection of a low-contrast abnormality is not
margined embolus and an ill-defined artifact. accurate when the SD of the mean of the abnor-
CT Technique However, modified window settings can also mality exceeds the difference in the means of
At the moment, at our institution, Light- increase the conspicuity of artifacts caused by the lesion and the surrounding region [17]. For
speed (GE Healthcare) 16- and 64-MDCT image noise and flow. acute pulmonary emboli, the mean attenuation
scanners are used to acquire the images of the Multiplanar reformation images through the value is 33 H (SD, 15 H) [16].
thorax in a caudal–cranial direction. The cau- longitudinal axis of a vessel can be used to Because it is important to detect all pulmo-
dal–cranial direction is used because most em- overcome some of the difficulties encountered nary emboli, we should calculate the highest
boli are located in the lower lobes and, if the with axial-orientated images of obliquely or possible attenuation of an acute pulmonary em-
patient breathes during image acquisition, axially orientated arteries [12]. Also, reformat- bolism to be the mean plus 3 SDs; this would in-
there is more excursion of the lower lobes com- ted images can help to differentiate between clude 99.75% of all acute emboli, which equates
pared with the upper lobes. For IV access, the some patient, technical, anatomic, and patho- to 78 H. According to Meaney et al. [17], we
antecubital vein and an 18- or 20-gauge cathe- logic factors that mimic pulmonary embolism need attenuation in the artery of at least one more
ter is preferred. The CT parameters are given in and true pulmonary embolism [11]. SD; the final figure therefore equals 93 H. The
Tables 1 and 2. Images are viewed on a PACS Contrast-enhanced helical CT of the veins mean attenuation and SD values for chronic pul-
monitor using IMPAX version 4.1 (AGFA) be- of the lower extremities is performed using monary embolism are 87 and 31 H, respectively.
cause there is improved accuracy in viewing the same contrast bolus as used for chest CT. Therefore, the highest possible attenuation value
chest CT cases on a workstation compared Images of the iliac, femoral, and popliteal of chronic pulmonary emboli with 3 SDs is cal-
with hard-copy film [9, 10]. The images are veins are obtained 3 minutes after the onset of culated to be 180 H. The minimum attenuation
displayed with three different gray scales for the initial contrast injection [13]. of adjacent opacified blood to identify this outly-
interpretation of lung window (window width, ing chronic thrombus is 211 H. The theoretic
1,500 H; window level, –600 H), mediastinal How to Reduce Motion Artifacts minimum attenuations of blood required to see
window (window width, 350 H; window level, Respiratory motion artifacts are the most all acute and chronic pulmonary venous throm-
40 H), and pulmonary embolism–specific common cause of an indeterminate CTPA and boemboli are 93 and 211 H, respectively.

1256 AJR:188, May 2007

 CT Pulmonary Angiography

TABLE 3: Empirical Timing Delay for CT Pulmonary Angiography After IV study [25], whereas in the study by Gosselin et
Administration of 370 mg I/mL al. [24], it was present in 37% of the study
Timing Delay (s) group. An interesting major difference between
16-MDCT 64-MDCT the studies, and a possible explanation of the
difference in frequency, is that the patients in
Normal-Size Large Patient Normal-Size Large Patient
Injection Patient (> 250 lb [113 kg]) Patient (> 250 lb [113 kg]) the study by Wittram and Yoo were instructed
to “take a breath in and hold it” before image
Amount injected (mL) 110 130 110 130
acquisition. The patients in the study by Gosse-
Rate of injection (mL/s)
lin et al. were instructed to have five respiratory
4 22 30 26 31 cycles of hyperventilation followed by a com-
3.5 26 34 30 35 mand of full inspiration 2 seconds before initial
3 32 40 36 42 images were obtained [24]. The hyperventila-
2.5 39 49 43 51
tion before inspiration and the breath-hold is
likely the exacerbating factor of this artifact.
2 50 61 53 63
Both studies used the same injection rate, but
Gosselin et al. used single-detector CT whereas
Wittram and Yoo used MDCT. However, the
To detect abnormalities with low differ- heart, the enhancement threshold might never number of detectors should not affect the ap-
ences in CT contrast, and to improve pulmo- be reached; this leaves the technologist uncer- pearance of this artifact.
nary embolism conspicuity, it is necessary to tain as to when to start image acquisition. The solution to transient interruption of con-
adjust the display window widths and levels Empiric scanning delay also has the advan- trast flow of the pulmonary arteries is to reduce
American Journal of Roentgenology 2007.188:1255-1261.

[17–19]. Also, the decision of the reviewer to tage of reducing operator error and motion arti- the volume of unopacified blood entering the
interpret a study as adequate or indeterminate facts by removing the added complexity of when right atrium from the inferior vena cava. Pres-
will be affected by the interplay of factors that to start the study based on a threshold value. To canning hyperventilation is likely the cause;
include the size of the suspected embolism, comprehensively evaluate for venous throm- with the implementation of faster scanners,
the anatomic level of the vessel being evalu- boembolic disease, patients need to receive a prescan hyperventilation should be dropped.
ated, and the amount of image noise. large contrast material bolus to evaluate the Because the venous return from the inferior
lower-limb veins [7]. Using an empiric scanning vena cava to the right atrium is exaggerated with
Timing of Bolus delay on 16- and 64-MDCT scanners, one aims
Several techniques are available for con- to be midscan at the peak of pulmonary artery
trast delivery on CT studies. A high injection enhancement; therefore, the start of the scanning
rate with a uniphase injection bolus of 4 mL/s is calculated to equal the injection time minus
of contrast material is preferred [20]; this al- half the scanning time. If the size of the IV access
lows a high intensity of contrast enhancement catheter does not allow 4 mL/s, then the delay
in the pulmonary arterial system. The injec- needs to increase, as illustrated in Table 3.
tion duration has an important influence in If an indeterminate scan occurs with stan-
optimizing contrast delivery in CT. dard delay due to poor enhancement, there is
Injection of contrast material can be consid- no extravasation of contrast material, and the
ered in two components: first pass and recircu- timing is appropriate, then poor venous flow
lation. The first-pass effect is optimized by the due to stenosis or obstruction may be a factor
use of contrast material with 370 mg I/mL. As [8], in which case a different venous access
the injection duration increases, the recircula- site may be necessary. A repeat CTPA after
tion of contrast material causes a cumulative ef- hydration of the patient is recommended.
fect on enhancement over time [21], so that an
increase in time increases the enhancement of Flow Artifacts
the pulmonary arteries during the injection. A transient interruption of contrast material
This enhancement advantage is most optimally consists of a portion of the pulmonary artery
used with the empiric delay technique, whereas that shows relatively poor enhancement be-
bolus tracking starts the CT scan earlier on the tween areas of higher attenuation both proxi-
rise of the enhancement curve and results in mally and distally [24, 25] (Fig. 1). Comparing
worse pulmonary artery enhancement. Al- patients with this artifact with age- and sex-
though no published data as yet can validate matched controls, Wittram and Yoo [25]
this statement, preliminary work appears to showed that the artifact results from an increase Fig. 1—Transient interruption of flow of contrast material
support this observation [22, 23]. One could ar- in flow of unopacified blood from the inferior in 59-year-old woman. Coronal oblique reformatted image
gue that when the triggering threshold for bolus vena cava. What can be done to avoid this flow through right posterior basal segmental artery from CT
tracking is increased, CT would start later on phenomenon? A review of the literature shows pulmonary angiography shows segment of poor
opacification (arrow) between areas of higher
the rise of the enhancement curve. However, in that the transient interruption of contrast artifact attenuation both proximally and distally. Interface
cases with poor function of the right side of the was seen in 3% of the study population in that between low- and high-attenuation areas is ill-defined.

AJR:188, May 2007 1257

 Wittram

Streak Artifacts
Streak artifact that obscures pulmonary ves-
sels because of metallic implants can make a
study indeterminate, a repeat CT will not im-
prove this problem, and additional imaging
with V/Q scintigraphy or pulmonary angiogra-
phy may be necessary. Streak artifact from
high-density contrast material in the superior
vena cava can obscure adjacent pulmonary ar-
teries. The frequency of this artifact can be re-
duced by using a saline bolus immediately af-
ter the contrast material injection [30].
A B
Fig. 2—Localized increase in vascular resistance in 69-year-old woman with breast cancer who has right-sided The Indeterminate CTPA
talc pleurodesis. This article discusses the solutions to the
A, Staging CT was performed with injection of 65 mL of Isovue 370 (iopamidol, Bristol-Myers Squibb) at rate of 1.5 mL/s
using scan delay of 35 seconds. Right lower lobe shows volume loss and consolidation. Note good opacification of common causes of an indeterminate CTPA. In
left lower lobe pulmonary arteries (arrowheads). However, also note poor opacification of right lower lobe pulmonary practice, if a diagnosis of pulmonary embolism
arteries (arrows), indicating localized increase in vascular resistance in right lower lobe arteries. cannot be confidently confirmed or refuted and
B, CT pulmonary angiogram 3 days after A using 110 mL of Isovue 370 at 4 mL/s and 22-second scanning delay. Note
good opacification of right lower lobe pulmonary arteries (arrows). This image illustrates that peripheral vascular
the study is indeterminate, it is recommended
resistance can be overcome with large volume of contrast material injected rapidly and by acquiring images at that the radiologist decide at which anatomic
very end of injection. level the study is indeterminate; for example, if
American Journal of Roentgenology 2007.188:1255-1261.

the radiologist can clear the vessels to the level

of the segmental arteries, and the subsegmental
heightened respiratory movements [26], we ver- image noise and improve scan quality by in- arteries are indeterminate, the clinician might
bally instruct our patients not to perform an ex- creasing reconstruction width to 2.5 mm. How- not require further imaging in cases with a low
aggerated inspiration and the CT technologist ever, the reconstruction width will decrease the clinical pretest probability for pulmonary embo-
prompts the patient to “hold your breath” before sensitivity of pulmonary embolism detection lism. However, some patients with indetermi-
image acquisition. Further study is required to [28]. In larger patients, for optimal pulmonary nate CTPA findings will need further imaging,
assess the possible benefits of these maneuvers. artery enhancement, the quantity of contrast ma- with ultrasound scan of the legs after hydration,
Localized increase in vascular resistance can terial needs to be adapted to the patient’s size a repeat CTPA, V/Q scintigraphy (if the lungs
result from lung consolidation or atelectasis [29]; to simplify the protocol, 110 mL of 370 mg are clear on CT), or pulmonary angiography.
[27]. The focal slow pulmonary artery flow can I/mL contrast material is used for patients weigh-
be a cause of an indeterminate CTPA (Fig. 2) ing 250 lb (113 kg) or less and 130 mL of 370 mg Direct Signs of Acute and
and can be a cause of misdiagnosis of pulmo- I/mL contrast material is used for those weight- Chronic Embolism
nary embolism [11]. Recognition of this phe- ing more than 250 lb (113 kg) (Tables 1 and 2). Both acute and chronic pulmonary emboli
nomenon is important because the poorly opac- For pregnant patients, the volume of con- are identified as intraluminal filling defects that
ified vessel may be normal or the poor contrast trast material should be reduced to 70 mL and show a sharp interface with IV contrast mate-
enhancement may obscure thrombus. A region- the timing adjusted accordingly (Table 4). rial. The diagnostic criteria for acute pulmo-
of-interest measurement may be helpful in this The reason for this rationale is that the legs nary embolism include, first, complete arterial
decision if the attenuation is greater than 78 H, and pelvis are not imaged and that the quan- occlusion with failure to opacify the entire lu-
which is the upper value of acute pulmonary tity of iodine to the fetus is also reduced. men; the artery may be enlarged in comparison
emboli [16]. Further imaging may be neces- with pulmonary arteries of the same order of
sary, either repeating CTPA with an increased Parenchymal Disease branching [31–33] (Fig. 3); second, a central
delay or pulmonary angiography. Although our Consolidation can cause a focal increase in
experience is anecdotal, this is an uncommon vascular resistance and focal poor vascular
artifact with empiric timing delay; it is likely opacification [27]. However, the frequency of TABLE 4: Empirical Timing Delay for
due to the wider temporal window of contrast this artifact will be reduced with the use of CT Pulmonary Angiography
injection that occurs with empiric timing delay empiric timing delay (Fig. 2) because image for Pregnant Patients After
IV Administration of 70 mL
compared with other techniques (Fig. 2). acquisition is performed at the end of the in-
of 370 mg I/mL
jection. As for reviewing vessels surrounded
Patient Habitus by consolidation, as with all radiology inter- Rate of Injection Timing Delay (s)
Two major issues are related to imaging pul- pretation, it is important to be systematic and (mL/s) 16-MDCT 64-MDCT
monary arteries of large patients: image noise review one vessel at a time and ignore the 4 12 16
and the volume of IV contrast material. For pa- consolidation or any other pathology that 3.5 15 18
tients weighing more than 250 lb (113 kg), it is might distract the attention of the reviewer. In
3 18 21
necessary to increase the radiation dose to de- this manner, any case with adequate enhance-
crease the amount of image noise. In addition, ment and no or minimal motion can be confi- 2.5 23 26
the protocol is modified to help decrease display dently interpreted. 2 30 33

1258 AJR:188, May 2007

 CT Pulmonary Angiography

Fig. 3—Acute pulmonary embolism in 27-year-old Indirect Signs of Acute and

woman. CT pulmonary angiogram shows thrombus ChronicPulmonary Embolism
(arrow) that expands diameter of right posterior basal
subsegmental artery compared with pulmonary These signs include nonuniform arterial
arteries of same order of branching (arrowheads). perfusion for both acute and chronic pulmo-
nary embolism; this radiologic sign is diffi-
cult to identify in cases of acute pulmonary
embolism but manifests as mosaic attenuation
in cases of chronic pulmonary embolism. A
mosaic pattern of lung attenuation is identi-
fied on the lung window settings.
The three major causes of mosaic lung atten-
uation are airways disease, chronic pulmonary
embolism (in which the abnormal region is more
radiolucent), and interstitial lung disease (in
which the abnormal lung is more opaque).
Oligemia, or a decrease in the flow rate due to
acute pulmonary embolism, is often identified
on angiography [34, 35]. In my experience, this
finding is more often seen on angiography than
on CT; this discrepancy is thought to be related
to the larger temporal window of IV contrast ma-
American Journal of Roentgenology 2007.188:1255-1261.

terial for CT as compared with angiography. Oc-

casionally, a large acute central pulmonary embo-
lism can cause oligemia and a reversible decrease
in vessel diameter; this CT equivalent of the Wes-
A termark sign has been previously illustrated [36].
Nonuniform arterial perfusion due to acute
pulmonary embolism can uncommonly mani-
Fig. 4—Acute pulmonary embolism in 27-year-old woman. fest as a mosaic pattern of attenuation on CT.
A, Centrally located thrombus, in right posterior basal
segmental artery, has well-defined margins and is
Additional indirect signs seen in chronic pul-
completely surrounded by contrast material (arrow). monary embolism include poststenotic dilata-
Acute emboli are also noted in right lateral basal tion, tortuous vessels, enlargement of the main
segmental and left posterior basal subsegmental arteries. pulmonary artery, and enlargement of the bron-
B, Curved reformatted image of posterior basal
segmental artery of right lower lobe shows that central chial arteries [36]. For a long time we have been
arterial filling defect (seen in A) cannot occur in at a stage at which the direct radiologic signs, as
isolation without embolism draping over vessel branch shown on CT angiography, are required to
point or touching vessel wall at some point. Axial image
of thrombus (A) was acquired at level of arrow. make a diagnosis of acute or chronic pulmo-
B nary thromboembolic disease. Because the in-
direct signs have a differential diagnosis, they
are helpful only as indicators of the sites of the
arterial filling defect surrounded by IV contrast direct radiologic signs of pulmonary embolism.
material [31] (Fig. 4); and third, a peripheral in-
traluminal filling defect that makes an acute an- Severity of Acute
gle with the arterial wall [32, 33] (Fig. 5). Pulmonary Embolism
The diagnostic criteria for chronic pulmo- After the initial embolic event, the patient
nary embolism include complete occlusion of may be at risk for circulatory collapse second-
a vessel that is permanently smaller than pul- ary to right heart failure, and a subsequent em-
monary arteries of the same order of branch- bolism may be fatal. It has been suggested that
ing [32, 33] (Fig. 6), a peripheral eccentric the early detection of acute right ventricular
filling defect that makes an obtuse angle with failure allows the implementation of the most
the vessel wall [32, 33] (Fig. 7), contrast ma- appropriate therapeutic strategy [37]. Right
terial flowing through apparently thick- ventricular strain or failure is optimally moni-
walled arteries that are smaller due to recanal- tored on echocardiography. However, some
ization [32, 33] (Fig. 8), a band or web in a morphologic abnormalities that indicate right
Fig. 5—Acute pulmonary embolism in 28-year-old contrast-filled artery [32, 33] (Fig. 9), and an ventricular failure can be quantified by CTPA.
woman. Eccentrically located embolism (arrow) forms
intraluminal filling defect with an acute pul- The most robust CT sign is right ventricular di-
acute angle with vessel wall. Emboli are also noted in
right lower lobar and left anteromedial basal monary embolism morphology that has been lation (in which the greatest right ventricle
segmental arteries. present for more than 3 months [16]. short-axis measurement is wider than the max-

AJR:188, May 2007 1259

 Wittram
American Journal of Roentgenology 2007.188:1255-1261.

Fig. 6—Chronic pulmonary embolism in 37-year-old Fig. 7—Chronic pulmonary embolism in 60-year-old Fig. 8—Chronic pulmonary embolism in 65-year-old man.
woman. Curved coronal reformatted CT image viewed man. Axial CT image obtained at level of right lower Curved coronal reformatted CT image viewed on maximum
on lung window setting shows pouch defect (arrow) of lobe pulmonary artery shows broad-based, smoothly intensity projection shows abrupt vessel narrowing that
anterior basal segmental artery of right lower lobe. margined, eccentric filling defect (arrow) that forms affects posterior basal segmental artery of right lower lobe.
Contracted or obliterated artery (arrowheads) is obtuse angle with vessel wall. Note abrupt convergence of contrast material, leading to
shown peripheral to site of chronic obstruction. thin column of more distal IV contrast material (arrow). In
addition, organized thrombus is identified surrounding
column of contrast material (arrowheads).

A B
Fig. 9—Chronic pulmonary embolism in 54-year-old Fig. 10—Acute right ventricle dilatation in 33-year-old woman with large acute pulmonary embolism clot burden.
man. Axial CT image of right lower lobe pulmonary A, Maximum short-axis diameter (black rule) of right ventricle measures 5.2 cm.
artery shows band or web (arrow) surrounded by B, At more cephalad level, maximum short-axis diameter (black rule) of left ventricle measures 3.2 cm. CT
contrast material. Subcarinal and right hilar pulmonary angiography right ventricle–to–left ventricle ratio equals 1.6.
lymphadenopathy is also noted, which is associated
with chronic pulmonary embolism.

imum left ventricle short-axis measurement) death [39]. A ratio of 1.0 is associated with a Conclusion
[38] (Fig. 10). The greater the right ventri- 5% chance of death; 1.3, 10%; 1.7, 20%; 1.9, For more than 25 years, the direct signs of pul-
cle–to–left ventricle short-axis ratio in acute 30%; 2.1, 40%; and a ratio of 2.3 is associated monary embolism have been available to the ra-
pulmonary embolism, the greater the risk of with a 50% chance of death [39]. diologist on CT, and this noninvasive technique

1260 AJR:188, May 2007

 CT Pulmonary Angiography

has produced a paradigm shift that has raised the zation of combined CT pulmonary angiography the essentials. Baltimore, MD: Williams & Wilkins,
standard of care for patients with this disease. with lower extremity CT venography. AJR 2000; 1980:53–75
This article outlines the approaches necessary to 174:67–69 27. Goodman LR, Curtin JJ, Mewissen MW, et al.
improve the quality of CT pulmonary angiogra- 14. Remy-Jardin M, Tillie-Leblond I, Szapiro D, et al. Detection of pulmonary embolism in patients
phy and summarizes the diagnostic criteria for CT angiography of pulmonary embolism in patients with unresolved clinical and scintigraphic diag-
acute and chronic pulmonary emboli. with underlying respiratory disease: impact of mul- nosis: helical CT versus angiography. AJR 1995;
tislice CT on image quality and negative predictive 164:1369–1374
value. J Eur Radiol 2002; 12:1971–1978 28. Schoepf UJ, Holzknecht N, Helmberger T, et al.
References 15. Boiselle PM, Hasegawa I, Nishino M, Raptopoulos Subsegmental pulmonary emboli: improved detec-
1. Horlander KT, Mannino DM, Leeper KV. Pulmo- V, Hatabu H. Comparison of artifacts on coronal ref- tion with thin collimation multi-row spiral CT. Ra-
nary embolism mortality in the United States, ormation and axial CT pulmonary angiography im- diology 2002; 222:483–490
1979–1998: an analysis using multiple-cause mor- ages using single-detector and 4- and 8-detector 29. Bae KT, Tao C, Hong C, Zhu F, Milite M, Hildebolt
tality data. Arch Intern Med 2003; 163:1711–1717 multidetector-row helical CT. Acad Radiol 2005; CF. Degree of contrast enhancement in pulmonary
2. Godwin JD, Webb WR, Gamsu G, Ovenfors CO. 12:602–607 CTA with MDCT: association with body weight
Computed tomography of pulmonary embolism. 16. Wittram C, Maher MM, Halpern E, Shepard JO. and volume of contrast medium. (abstr) RSNA sci-
AJR 1980; 135:691–695 Hounsfield unit values of acute and chronic pulmo- entific assembly and annual meeting program. Chi-
3. The PIOPED investigators. Value of the ventila- nary emboli. Radiology 2005; 235:1050–1054 cago, IL: RSNA, 2005:546
tion/perfusion scan in acute pulmonary embolism: 17. Meaney TF, Raudikivi U, McIntyre WJ, et al. De- 30. Vogel N, Kauczor HU, Heussel CP, Reis BG,
results of the Prospective Investigation of Pulmo- tection of low-contrast in computed body tomogra- Thelen M. Artifact reducing in diagnosis of lung
nary Embolism Diagnosis (PIOPED). JAMA 1990; phy: an experimental study of simulated lesions. embolism using CT with saline bolus. Rofo 2001;
263:2753–2759 Radiology 1980; 134:149–154 173:460–465
American Journal of Roentgenology 2007.188:1255-1261.

4. Wittram C, Meehan MJ, Halpern EF, et al. Trends in 18. Brink JA, Woodard PK, Horesh L, et al. Depiction 31. Ghaye B, Remy J, Remy-Jardin M. Non-traumatic
thoracic radiology over a decade at a large academic of pulmonary emboli with spiral CT: optimization thoracic emergencies: CT diagnosis of acute pul-
medical center. J Thorac Imaging 2004; 19:164–170 of display window settings in a porcine model. Ra- monary embolism—the first 10 years. Eur Radiol
5. Qanadli SD, Hajjam ME, Mesurolle B, et al. Pul- diology 1997; 204:703–708 2002; 12:1866–1905
monary embolism detection: prospective evalua- 19. Bae KT, Mody GN, Balfe DM, et al. CT depiction 32. Gottschalk A, Stein PD, Goodman LR, Sostman
tion of dual-section helical CT versus selective pul- of pulmonary emboli: display window settings. Ra- HD. Overview of Prospective Investigation of Pul-
monary arteriography in 157 patients. Radiology diology 2005; 236:677–684 monary Embolism Diagnosis II. Semin Nucl Med
2000; 217:447–455 20. Bae KT, Tran HQ, Heiken JP. Multiphasic injec- 2002; 32:173–182
6. Winer-Muram HT, Rydberg J, Johnson MS, et al. tion method for uniform prolonged vascular en- 33. Washington L, Goodman LR, Gonyo MB. CT for
Suspected acute pulmonary embolism: evaluation hancement at CTA: pharmacokinetic analysis and thromboembolic disease. Radiol Clin North Am
with multi-detector row CT versus digital subtrac- experimental porcine model. Radiology 2000; 2002; 40:751–771
tion pulmonary arteriography. Radiology 2004; 216:872–880 34. Sagel SS, Greenspan RH. Nonuniform pulmonary
233:806–815 21. Bae KT. Peak contrast enhancement in CTA and arterial perfusion: pulmonary embolism? Radiol-
7. Stein PD, Fowler SE, Goodman LR, et al. Multide- MRA: when does it occur and why? Radiology ogy 1970; 99:541–548
tector computed tomography for acute pulmonary 2003; 227:809–816 35. Dalen JE, Brooks HL, Johnson LW, et al. Pulmo-
embolism. N Engl J Med 2006; 354:2317–2327 22. Sahagun O, Wittram C. A population-controlled nary angiography in acute pulmonary embolism: in-
8. Jones SE, Wittram C. The indeterminate CT pulmo- comparison of fixed-delay and Smart-Prep en- dications, techniques, and results in 367 patients.
nary angiogram: imaging characteristics and patient hancement for CT pulmonary angiography. (abstr) Am Heart J 1971; 81:175–185
clinical outcome. Radiology 2005; 237:329–337 RSNA scientific assembly and annual meeting pro- 36. Wittram C, Kalra MK, Maher MM, Greenfield A,
9. Seltzer SE, Judy PF, Adams DF, et al. Spiral CT of gram. Chicago, IL: RSNA, 2005:546 McLoud TC, Shepard JO. Acute and chronic pul-
the chest: comparison of cine and film-based view- 23. Chang SA, Matthews CC, Kay D. CT pulmonary monary emboli: angiographic–CT correlation. AJR
ing. Radiology 1995; 197:73–78 angiography for the evaluation of pulmonary em- 2006; 186[6 suppl 2]:S421–S429
10. Reiner BI, Eliot L. Siegel EL, Hooper FJ. Accuracy bolus: fixed delay versus bolus tracking techniques 37. Goldhaber SZ. Echocardiography in the manage-
of interpretation of CT scans: comparing PACS of contrast administration. (abstr) RSNA scientific ment of pulmonary embolism. Ann Intern Med
monitor displays and hard-copy images. AJR 2002; assembly and annual meeting program. Chicago, 2002; 136:691–700
179:1407–1410 IL: RSNA, 2005:639 38. van der Meer RW, Pattynama PM, van Strijen MJ,
11. Wittram C, Maher MM, Yoo AJ, Kalra MK, Shep- 24. Gosselin M, Rassner UA, Thieszen SL, et al. Con- et al. Right ventricular dysfunction and pulmonary
ard JO, McLoud TC. CT angiography of pulmonary trast dynamics during CT pulmonary angiogram: obstruction index at helical CT: prediction of clini-
embolism: diagnostic criteria and causes of misdi- analysis of an inspiration associated artifact. J Tho- cal outcome during 3-month follow-up in patients
agnosis. RadioGraphics 2004; 24:1219–1239 rac Imaging 2004; 19:1–7 with acute pulmonary embolism. Radiology 2005;
12. Remy-Jardin M, Remy J, Cauvain O, et al. Diagno- 25. Wittram C, Yoo A. Transient interruption of con- 235:798–803
sis of central pulmonary embolism with helical CT: trast on CT pulmonary angiography: proof of mech- 39. Ghaye B, Ghuysen A, Willems V, et al. Severe pul-
role of two-dimensional multiplanar reformations. anism. J Thorac Imaging (in press) monary embolism: pulmonary artery clot load
AJR 1995; 165:1131–1138 26. Smith JS. Kampine JP. Pressure and flow in the ar- scores and cardiovascular parameters as predictors
13. Yankelevitz DF, Gamsu G, Shah A, et al. Optimi- terial and venous systems. Circulatory physiology: of mortality. Radiology 2006; 239:884–891

F O R YO U R I N F O R M AT I O N

This article is available for CME credit. See www.arrs.org for more information.

AJR:188, May 2007 1261