Pulmonary embolism (PE) is a common cause of death in the hospitalized patient, yet the diagnosis of PE can easily be overlooked. The authors discuss the current role of spiral CT scanning in the diagnosis of PE, its role compared with conventional diagnostic methods, and pitfalls in the diagnosis with spiral CT angiography.
Dr. Farag is a Clinical Fellow in the Thoracic Radiology
Division of Brigham and Women's Hospital and Harvard Medical
School, and Dr. Costello is Associate Professor of Radiology at
Harvard Medical School and Director of the Thoracic Radiology
Division of Brigham and Women's Hospital, Boston, MA.
Pulmonary embolism (PE) is a common cause of death in
hospitalized patients with clinical manifestations that are protean
and nonspecific. Patients can present with symptoms ranging from
minimal dyspnea to sudden death. Most often, symptoms and signs
include dyspnea, pleuritic chest pain, cough, hemoptysis,
tachypnea, hypotension, fever, and tachycardia. Given the
non-specific gamut of signs and symptoms, the diagnosis of PE can
easily be overlooked. Standard laboratory tests are also
non-specific with diminished arterial PO
and a widened alveolar-atrial gradient. More recently, D-dimer, a
blood test that measures fibrin degradation products, has been
found to be helpful but remains in development.
The classical clinical approach to a patient with suspected PE
is to combine the degree of clinical suspicion with
ventilation-perfusion (V/Q) scintigraphy, and to perform pulmonary
angiography in cases that remain unclear. The problem arises in
that approximately 75% of patients fall within the "indeterminate"
V/Q category of "diagnosis," requiring pulmonary angiography for a
Pulmonary arteriography is invasive, has an associated small but
measurable morbidity, and is costly. Over the past few years, the
introduction of spiral CT angiography (spiral CTA), a more direct
and relatively non-invasive technique for the diagnosis of PE, has
become generally available.
The addition of spiral CT to the diagnostic armamentarium has
the potential to dramatically alter the approach taken when
assessing patients for PE. Because of the wide availability of
spiral CT systems, CT is generally an easier study to obtain than a
V/Q scan or pulmonary angiogram. With careful attention to CT
technique, direct visualization of the clot in the pulmonary
arteries down to the level of segmental, and even subsegmental,
vessels is possible.
Additionally, spiral CTA is capable of demonstrating a variety of
thoracic pathologies that can "mimic" PE, and may establish an
alternative diagnosis for the patient's symptoms.
This review will explore the current role of spiral (helical) CT
scanning in the diagnosis of PE. We will also attempt to evaluate
where this technique stands with regard to the conventional methods
of diagnosing PE. A brief review of the pitfalls in the diagnosis
with spiral CTA will be given. Finally, diagnostic future
techniques will be examined briefly.
Spiral CT and PE--Over the past 8 years, it has been suggested
that spiral CT can replace V/Q scanning and pulmonary angiography
for the diagnosis of PE. CT criteria for diagnosing acute PE are
similar to criteria used with pulmonary angiography. Clots can be
directly visualized as partial or complete filling defects within
Several studies have reported high sensitivities and specificities
in the identification of PE down to the segmental arterial level
Coche et al
have reported the value of associated pleural and parenchymal
findings that aid in the diagnosis of pulmonary embolism. Ancillary
signs such as wedge-shaped pleural based consolidation, linear
bands, and dilated central or segmental pulmonary arteries can be
associated with the diagnosis of PE and are statistically
significant findings. As such, the detection of these findings may
direct further investigations particularly if a study is
"suboptimal" for assessment of central or segmental vessels.
CT technique--With spiral CT, the x-ray tube and detector
revolve around the patient generating a helix of data while the
patient moves through the gantry. Most newer CT systems have cycle
times of <1 second and have tube capacities that allow the
entire thorax to be examined in <25 seconds.
Recently introduced multi-detector array CT scanners will have a
major influence on PE detection. Through the simultaneous
acquisition of four slices with a rotation speed of 0.5 seconds,
even higher resolution studies of the chest can be acquired with
short breathhold intervals. The shorter scan time reduces cardiac
and respiratory motion artifacts, making vessels adjacent to the
heart easier to visualize.
Intravenous contrast can be administered at
low-concentration/high-flow rate protocols (i.e., 30% concentration
at 5 to 6 mL/sec) or standard concentration and a flow rate of 3
mL/sec can be used for PE detection. Artifacts produced with
high-concentration low-flow rate protocols due to contrast in the
superior vena cava can slightly degrade evaluation of the right
main and right upper lobe pulmonary arteries.
Low-concentration/high-flow rate contrast administration may limit
the evaluation of more distal subsegmental vessels because of
suboptimal vessel opacification. Our protocol uses 100 mL of 60%
iodine concentration infused at 3 mL/sec through an antecubital
vein with a 15-second delay to scan start.
Pulmonary angiography--Pulmonary angiography has long been
considered the "gold standard" for true diagnosis of PE, with a
high spatial resolution that allows for detection of emboli to the
However it is invasive, with morbidity and mortality rates quoted
as 1% to 3% in patients having pulmonary angiograms.
Given its invasive and costly properties, some clinicians are
reluctant to refer all patients with intermediate-probability V/Q
scans for pulmonary angiography, despite the fact that one-third of
these patients will have PE.
Most importantly, the interobserver variability for pulmonary
angiography is significant, and Stein et al
has documented an overall interobserver agreement of only 81%, the
agreement is even lower at the level of the subsegmental
V/Q scintigraphy--The Prospective Investigation of Pulmonary
Embolism Diagnosis (PIOPED) study has evaluated the usefulness of
V/Q scintigraphy in the diagnosis of PE.
In this study, V/Q scans were split into four categories with
regard to the findings and the likelihood of acute PE (normal, low,
intermediate, or high probability). Patients with normal V/Q scans
were shown to have a 2% to 4% chance of having clinically
significant acute PE.
Patients with high-probability V/Q scans were shown to have a 96%
probability of acute PE.
In these settings, clinicians are comfortable with the results and
will withhold or administer anticoagulant therapy
Problems arise with studies that fall within the low- and
intermediate-probability categories. These V/Q diagnoses occur in
up to 75% of patients, and 16% of patients with low probability and
33% of patients with intermediate probability scans will have
angiographically proven PE.
Because of a reluctance to perform pulmonary angiography, this can
translate into patients not receiving needed anticoagulant therapy,
or patients receiving unnecessary anticoagulant therapy.
Interobserver disagreement in the PIOPED study for low- and
intermediate-probability scans was quoted as 25% and 30%,
Compression Doppler ultrasound--Patients with clinically
suspected PE have also been shown to benefit from compression and
color Doppler venous ultrasound of the legs to evaluate for deep
venous thrombosis (DVT).
If DVT is discovered, the patient is treated regardless of the V/Q
scan result. The limits of Doppler ultrasound lie within the deep
pelvic veins, which cannot be assessed adequately. Recently Loud et
described CT evaluation for DVT as an extension of the spiral CT
exam of the chest. Venous phase imaging of the pelvis and upper
legs at the time of spiral CTA was shown to be comparable to venous
sonography for the evaluation of femoropopliteal DVT.
In addition, this technique proved to be useful for evaluation of
clots in the vena cava and iliac veins (vessels poorly assessed
Spiral CT compared with pulmonary angiography and V/Q
scans--Several studies have compared spiral CT with pulmonary
angiography for the diagnosis of acute PE with reported high
sensitivities and specificities in the identification of PE down to
the segmental level.
Data from these studies demonstrates a sensitivity range of 53% to
100% and a specificity of 78% to 100% for spiral CT. Range
variations can be explained by differences in equipment, scanning
protocols, contrast administration, study design, and radiologist
experience. However, when only central or segmental emboli are
considered, sensitivity and specificity of spiral CT approaches 90%
The European Multicenter Trial (ESTIPEP) is a prospective study
of 401 patients that compared spiral CT with pulmonary angiography.
Clinical probability for PE was estimated, and all patients
underwent spiral CT and V/Q scanning. Pulmonary angiography was
performed in cases of low- and intermediate-probability V/Q scans,
and if spiral CT and V/Q diagnoses were contradictory. The study
demonstrated sensitivity of spiral CT to vary between 75% to
Sensitivity of V/Q scans in this study ranged from 36% to 65%.
Spiral CT had the advantage of performing fewer non-diagnostic
exams than V/Q, since the lung parenchyma, mediastinum and chest
wall structures were also assessed resulting in alternative
diagnoses for patient's symptoms. Equally important is the fact
that the interobserver agreement between readers was greater for CT
than for V/Q scanning (*72% versus *39%)
and even better than pulmonary angiography (46%).
The European Multicenter trial also showed interobserver agreement
for CT to be dependent on the quality of the study. For V/Q scans,
interobserver agreement was neither due to study quality nor the
use of PIOPED criteria.
One recently published study highlights the importance of
technical factors when performing spiral CT for PE. Drucker et al
described the sensitivity and specificity of two groups of readers
for detecting PE by spiral CT to be 53% to 60% and 87% to 97%,
respectively. These results, which were significantly worse than
other studies, may be explained by several factors. First,
relatively thick slices were used (5 mm, while most current
techniques utilize 2-mm slice thickness). Second, hard-copy
interpretation makes it difficult to recognize small clots, as many
can only be recognized by varying window levels and tracing vessels
on sequential images. Soft-copy workstation interpretation improves
image viewing since it is interactive and allows for easy window
leveling and vessel tracking.
The recognized limitation of spiral CT for the detection of
subsegmental emboli continues to be debated.
Isolated subsegmental PE are likely infrequent, as most PE split
into an average of 6 to 8 fragments,
of which in most patients one or more should be large enough to be
detected by spiral CT.
In the PIOPED study, 5.6% of PE were limited to subsegmental
In four studies comparing spiral CT to pulmonary angiography, 0% to
36% of pulmonary emboli were isolated and subsegmental.
Spiral CT demonstrated between 0% to 50% of these isolated
Although early results showed limitations of spiral CT in the
detection of subsegmental emboli,
earlier scan protocols (e.g., 5-mm collimation with pitch of 1)
have been surpassed by technological improvements in CT systems.
Narrower collimation (2-mm, and even 1-mm with multi-detector CT
systems) and reconstruction intervals have improved visualization
of peripheral pulmonary arteries (figure 2). Remy-Jardin et al
described a group of 370 patients that demonstrated the percentage
of analyzable subsegmental arteries to be statistically
significantly higher when patients are scanned with 2-mm
collimation and pitch of 2 rather than with 3-mm collimation and a
pitch of 1.7 (65% versus 43%, respectively).
The clinical significance of subsegmental emboli remains
but probably depends on a patient's cardiopulmonary reserve.
In otherwise healthy patients, subsegmental emboli may be
insignificant clinically. However, occlusion of several
subsegmental vessels perfusing the normal lung tissue in a patient
with pre-existing lung disease can lead to respiratory failure.
Pitfalls of spiral CT angiography--Spiral CT and conventional
pulmonary angiography have similar rates of diagnostically
suboptimal examinations, between 2% and 4% of spiral CT studies and
3% of pulmonary angiograms.
Suboptimal exams are most often due to poor image quality arising
from motion related to patient dyspnea.
Breathing artifacts are seen as pseudo-filling defects, mostly at
the level of obliquely oriented arteries due to variable position
of the vessel in the section width in two successive scans. In this
instance, a decrease in attenuation can be seen at the level of the
Routes of intravenous contrast delivery, flow rates, catheter
size and position, and short or long scanning delays may all affect
the quality of pulmonary artery opacification. With prolonged scan
delays, there may not be adequate contrast available to opacify the
vessels at the end of the scanned volume. With short scan delays,
insufficient time has elapsed to allow appropriate opacification of
the pulmonary arteries (especially on the earlier images of the
scanned volume). As such, appropriate consideration must be given
to patient cardiovascular physiology (particularly cardiac output),
anatomy, and the cardiac output. In general, a 15-second delay to
scan initiation is used, but in patients with poor cardiac function
a delay of 20 seconds may be used.
Similarly, increased pulmonary vascular resistance due to
parenchymal abnormalities can lead to false-positive studies.
Additional limitations and possible sources of error in
interpreting spiral CT exams involve elements related to the
interpreter's experience. Confusion of pulmonary veins, hilar lymph
nodes, and mucous-filled bronchi (figure 3) with poorly opacified
pulmonary arteries is often less problematic when the reader is
MRI diagnosis of pulmonary embolism
Gefter et al
advocates the utility of pulmonary magnetic resonance an-giography
(MRA) to evaluate the pulmonary arterial system for PE. The use of
pulmonary MRA is attractive because of its non-invasiveness, and
its ability to provide a three-dimensional evaluation of pulmonary
vasculature. The combination of pulmonary MRA with MR venography of
the pelvis and lower extremities makes the technique more appealing
because of its comprehensive nature.
The role of pulmonary MRA in the clinical algorithm to diagnose PE
has yet to be established, and its cost-effectiveness as a routine
examination is debatable.
Spiral CT angiography is a widely available and cost-effective
technique for the evaluation of the pulmonary arterial system, and
assessment for pulmonary embolism. With improvements in CT
technology (faster gantry speeds and multi-row detector CT),
techniques and protocols are evolving that will lead to a higher
sensitivity and specificity of spiral CTA for PE. Combining spiral
CTA and compression Doppler ultrasound of the legs seems the most
cost-effective method. This technique can also suggest alternate
diagnoses that cannot be established by conventional V/Q
scintigraphy or by pulmonary arteriography. AR
1. Kollef MH, Zahid M, Eisenberg PR: Predilective value of a
rapid semiquantitative D-dimer assay in critically ill patients
with suspected venous thromboembolic disease. Crit Care Med
2. Indik JH, Alpert JS: Detection of pulmonary embolism by
D-dimer assay, spiral computed tomography and magnetic resonance
imaging. Prog Cardiovasc Dis 42:261-272, 2000.
3. The PIOPED Investigators: Value of the ventilation/perfusion
scan in acute pulmonary embolism: Results of the prospective
investigation of pulmonary embolism diagnosis (PIOPED). JAMA
4. Goodman LR, Lipchik RJ, Kuzo RS: Acute pulmonary embolism:
The role of computed tomographic imaging. J Thorac Imag 12:83-86,
5. Goodman LR, Lipchik RJ, Kuzo RS: Reply to opinions. J Thorac
Imag 12:100-102, 1997.
6. Goodman LR: CT of acute pulmonary embolism: Where does it
fit? RadioGraphics 17:1037-1042, 1997.
7. Gurney JW: No fooling around: Direct visualization of
pulmonary embolism. Radiology 188:618-619, 1993.
8. Godwin JD, Webb WR, Gamsu G, et al: Computed tomography of
pulmonary embolism. AJR Am J Roentgenol 135:691-695, 1980.
9. Verschakelen JA, Vanwijck E, Bogaert J, Baert AL: Detection
of unsuspected central pulmonary embolism with conventional
contrast CT. Radiology 188:847-850, 1993.
10. Winston CB, Wechsler RJ, Salazar AM, et al: Incidental
pulmonary emboli detected at helical CT: Effect on patient care.
Radiology 201:23-27, 1996.
11. Coche E, Müller NL, Kim K, et al: Acute pulmonary embolism:
Ancillary findings at spiral CT. Radiology 207:753-758, 1998.
12. Goodman LR, Curtin JJ, Mewissen MW, et al: Detection of
pulmonary embolism in patients with unresolved clinical and
scintigraphic diagnosis: Helical CT versus angiography. AJR Am J
Roentgenol 164:1369-1374, 1995.
13. Remy-Jardin M, Remy J, Wattinne L, Giraud F: Central
pulmonary thromboembolism: Diagnosis with spiral volumetric CT with
single-breath-hold technique: Comparison with pulmonary
angiography. Radiology 185:381-387, 1992.
14. Van Rossum AB, Pattynama PM, Ton ER, et al: Pulmonary
embolism. Validation of spiral CT angiography in 149 patients.
Radiology 201:467-470, 1996.
15. Remy-Jardin M, Remy J, Deschildre F, et al: Diagnosis of
acute pulmonary embolism with spiral CT: Comparison with pulmonary
angiography and scintigraphy. Radiology 200:699-706, 1996.
16. Mayo JR, Remy-Jardin M, Midller NL, et al: Pulmonary
embolism: Prospective comparison of spiral CT and ventilation
perfusion scintigraphy. Radiology 205:447-452, 1997.
17. van Rossum AB, Pattynama PM, Mallens WM, et al: Can helical
CT replace scintigraphy in the diagnostic process in suspected
pulmonary embolism? A retrolective-prolective cohort study focusing
on total diagnostic yield. Eur Radiol 8:90-96, 1998.
18. Kalender WA, Seissler W, Klotz E, Vock P: Spiral volumetric
CT with single breath-hold technique, continuous transport, and
continuous scanner rotation. Radiology 176:181-183, 1990.
19. Stein PD: Wedge arteriography for the identification of
pulmonary emboli in small vessels. Am Heart J 82:618-623, 1971.
20. Stein PD, Athanasoulis C, Alavi A, et al: Complications and
validity of pulmonary angiography in acute pulmonary embolism
21. Mills SR, Jackson DC, Older RA, et al: The incidence,
etiologies, and avoidance of complications of pulmonary angiography
in a large series. Radiology 136:295-299, 1980.
22. Cronan JJ: Venous thromboembolic disease: The role of US.
Radiology 186:619-630, 1993.
23. Polak JF: Doppler ultrasound of the deep leg veins: A
revolution in the diagnosis of deep vein thrombosis and monitoring
of thrombolysis. Chest 99:1655-1725, 1991.
24. Loud PA, Grossman ZD, Klippenstein DL, Ray CE: Combined CT
venography and pulmonary angiography: A new diagnostic technique
for suspected thromboembolic disease. AJR Am J Roentgenol
25. Loud PA, Katz DS, Klippenstein DL, et al: Combined CT
venography and pulmonary angiography in suspected thromboembolic
disease: Diagnostic accuracy for deep venous evaluation. AJR Am J
Roentgenol 174:61-65, 2000.
26. Drucker EA, Rivitz SM, Shepard, et al: Acute pulmonary
embolism: Assessment of helical CT for diagnosis. Radiology
27. Herald C, Remy-Jardin M, Grenier PH, et al: Prospective
evaluation of pulmonary embolism: Initial results of the European
Multicenter Trial (ESTIPEP). Radiology 209:299, 1998. Abstract.
28. The Urokinase Pulmonary Embolism Trial Study Group:
Urokinase pulmonary embolism trial: Phase 1 results. JAMA
29. Teigen CL, Maus TP, Sheedy PF II, et al: Pulmonary embolism:
Diagnosis with electron beam CT. Radiology 188:839-845, 1993.
30. Baghaie F, Remy-Jardin M, Remy J, et al: Diagnosis of
peripheral acute pulmonary emboli: Optimization of the spiral CT
acquisition protocol. Radiology 209(P):299, 1998. Abstract.
31. van Erkel AR, van Rossum AB, Bloem J, et al: Spiral CT
angiography for suspected pulmonary embolism: A cost effective
analysis. Radiology 201:29-36, 1996.
32. Remy-Jardin M, Remy J: Spiral CT angiography of the
pulmonary circulation. Radiology 212:615-636, 1999.
33. Dalen JE: When can treatment be withheld in patients with
suspected pulmonary embolism? Arch Intern Med 153:1415-1418,
34. Gefter WB, Hatabu H, Holland GA, Ostason AW: Pulmonary MR
angiography. Semin Ultrasound CT MR 17:316-323, 1996.