Radiology

Matthew D. Cham, MD
David F. Yankelevitz, MD
Claudia I. Henschke, PhD,
MD
Published online
10.1148/radiol.2342021656
Radiology 2005; 234:591594

Thromboembolic Disease
Detection at Indirect CT
Venography versus CT
Pulmonary Angiography1

Abbreviation:
DVT deep venous thrombosis
1

From the Department of Radiology,

Strong Memorial HospitalUniversity
of Rochester School of Medicine and
Dentistry, 620 Park Ave, PMB 244,
Rochester, NY 14607 (M.D.C.); and
Department of Radiology, New York
Presbyterian HospitalWeill Medical
College at Cornell University, New
York, NY (D.F.Y., C.I.H.). From the
2002 RSNA Annual Meeting. Received
December 9, 2002; revision requested
February 6, 2003; nal revision received
April 12, 2004; accepted April 28. Address correspondence to M.D.C.
(e-mail: matthew_cham@urmc.rochester
.edu).
Authors stated no nancial relationship to disclose.

PURPOSE: To assess the incremental increase in thromboembolic disease detection

at indirect computed tomographic (CT) venography versus CT pulmonary angiography and to determine the importance of scan interval for indirect CT venography
on the basis of thrombus length.
MATERIALS AND METHODS: Institutional review board approval was obtained,
and informed consent was not required. The study included 1590 consecutive
patients undergoing CT pulmonary angiography for the suspicion of pulmonary
embolism. Two minutes after completion of pulmonary angiography, a contiguous
indirect CT venography was performed from the iliac crest to the popliteal fossa. The
presence of pulmonary embolism or deep venous thrombosis (DVT) was recorded
for all patients. The lengths of all deep venous thrombi found in the rst 378
consecutive patients were recorded.
RESULTS: Pulmonary embolism was detected in 243 (15%) of 1590 patients at CT
pulmonary angiography, and DVT was detected in 148 (9%) patients at indirect CT
venography. Among 148 patients with DVT, pulmonary embolism was detected in
100 patients at CT pulmonary angiography. Thus, the addition of indirect CT
venography to CT pulmonary angiography resulted in a 20% incremental increase
in thromboembolic disease detection compared with that at CT pulmonary angiography alone (99% condence interval: 17%, 23%). Among the 378 patients, DVT
was present in 33 patients at indirect CT venography. Two (6%) of 33 patients had
clots measuring 2 cm or less, six (18%) had clots measuring 3 4 cm, and 25 (76%)
had clots measuring more than 4 cm in length.
CONCLUSION: The addition of indirect CT venography to CT pulmonary angiography incrementally increases the detection rate of thromboembolic disease by
20%. Performance of indirect CT venography by using contiguous section intervals,
with a section width of 1 cm, is recommended to accurately detect DVT.

Author contributions:
Guarantors of integrity of entire study,
D.F.Y., M.D.C., C.I.H.; study concepts
and design, D.F.Y., C.I.H.; literature
research, M.D.C., D.F.Y., C.I.H.; clinical studies, D.F.Y., M.D.C.; data acquisition, M.D.C., D.F.Y.; data analysis/
interpretation, M.D.C., D.F.Y., C.I.H.;
statistical analysis, C.I.H., M.D.C.,
D.F.Y.; manuscript preparation, denition of intellectual content, editing, revision/review, and nal version approval, M.D.C., D.F.Y., C.I.H.

RSNA, 2005

RSNA, 2005

Pulmonary embolism is a life-threatening disease that affects between 40 000 and 200 000
patients annually in the United States alone (13). Pulmonary embolism is a well-recognized sequela of deep venous thrombosis (DVT) of the lower extremities (4,5). Findings of
studies have shown that inadequately treated DVT is associated with recurrent pulmonary
embolism (6 8). To accurately diagnose thromboembolic disease in patients suspected of
having pulmonary embolism, several investigators have developed helical computed tomographic (CT) protocols that depict both pulmonary embolism and lower extremity DVT
(9,10). This technique has been referred to as combined CT pulmonary angiography and
indirect CT venography.
Indirect CT venography allows examination of the pelvis and lower extremities by using
only the contrast material already in circulation from the preceding CT pulmonary
angiography, thus obviating additional contrast material, which is associated with both
traditional and direct CT venography (9,10).
The addition of indirect CT venography to the standard CT pulmonary angiographic
591

protocol requires only an additional

3

minutes to perform, potentially obviating a separate lower extremity examination that can further delay the tur
naround time of results (9,10). In our prior
study, we compared indirect CT venography with lower extremity sonography in
116 patients and found a disagreemen
t
rate of only 3% (11). Authors of a doubleblinded prospective study involving 70
consecutive patients undergoing both
lower extremity sonography and combined CT pulmonary angiography and
indirect CT venography reported a sensitivity and specicity of 100% and 97
%,
respectively, for indirect CT venography
(12). Authors of other nonrandomized
retrospective studies have reported sensi
tivities of 71%94% and specicities o
f
93%94% for indirect CT
venography
(13,14).
Several investigators have also note
d
that, like CT pulmonary angiography, indirect CT venography also has the potential to provide alternative nonvascular
diagnoses for the patients clinical pre
sentation (1517). The widespread use of
this examination has been further supported by studies in which a moderately
good interobserver agreement, a consi
stently high level of venous enhancement, and a low radiation risk were r
eported (18 21).
Authors of several large prospecti
ve
multicenter studies have found that the
addition of indirect CT venography to
the basic CT pulmonary angiographi
c
protocol increased the diagnosis of
thromboembolic disease by 15%38%
(11,17,22).
Another consideration in the increasingly common use of combined CT pu
lmonary angiography and indirect CT
venography is the optimization of the indirect CT venographic portion of the examination. From a technical standpoint, it
is necessary to consider two important pa
rameters: the time delay prior to initiatin
g
indirect CT venography and the sectio
n
intervals. We previously evaluated t
he
former and found that the 2-minute delay
following CT pulmonary angiography produced near maximum opacication in the
majority of patients (11). Consideration o
f
the latter necessitates obtaining infor
ma-

tion about thrombi length, b

ecause long
(ie, 5-cm) section intervals
would potenbolic disease detection at indirect CT
tially miss small thrombi. On
venography versus CT pulmonary anthe other
hand, increasing the section giography and to determine the importance of scan interval for indirect CT
interval reduces radiation exposure, wh venography on the basis of thrombus
length.
ich is currently slightly lower than tha
MATERIALS AND METHODS
t of a standard pelvic CT (21).
Patients
The purpose of this study wa
Our study included 1590 consecutive
s to assess
the incremental increase in thr patients who had undergone combined CT
pulmonary angiography and indirect CT
omboemvenography between June 2, 1998, and October 31, 2001. Institutional review board
approval was obtained for the study (IRB#
0699746). Informed consent was not required by our institutional review board;
patient identiers were deleted.
Of 1590 patients (age range, 1899
years), 707 (44%) were men (median age,
64 years; mean age, 61 years) and 883
(56%) were women (median age, 64 years;
mean age, 62 years). A total of 1324 (83%)
patients were from the inpatient setting,
while 266 (17%) were from the emergency
department and outpatient setting.
From June 2, 1998, to June 27, 1999,
we documented the lengths of deep venous thrombi in 378 consecutive patients. We also recorded whether these
378 patients and an additional 1212 consecutive patients had positive ndings
for pulmonary embolism or DVT at combined CT pulmonary angiography and
indirect CT venography, for a total of
1590 patients.
All patients were referred to the New
York Presbyterian-HospitalWeill Medical College at Cornell University because
of the clinical suspicion of pulmonary
embolism. All studies were ordered by
physicians as clinically indicated and
were not inuenced by the study protocol. Because no additional contrast agent
was necessary to perform indirect CT
venography, all patients with a serum
creatinine level of less than 1.5 mg/dL
(133 mol/L) were eligible to undergo
combined CT pulmonary angiography
and indirect CT venography. Patients
who could not complete the study
because of known allergic reactions, inadequate intravenous access, or renal insufciency without hemodialysis were
excluded from the study.

Imaging
Our imaging protocol was the standard
hospital protocol used for routine clinical
care, without modications. All patients
underwent combined CT pulmonary angiography and indirect CT venography. A

Findings of the examination were categorized as inconclusive when poor examination quality resulted in a nondiagnosEach study had been read by one of six tic impression.
attending chest radiologists (including
D.F.Y. and C.I.H.). Each attending radiolStatistical Analysis
ogist had at least 10 years of experience.
The parameter of interest was the increFor both CT pulmonary angiography and
mental increase in detection of thromboindirect CT venography, thrombi were
dened as low-attenuating partial or embolic disease when combined CT pulmonary angiography and indirect CT
complete intraluminal lling defects surrounded by a high-attenuating ring of venography was performed versus when
enhanced blood and were seen on at least only CT pulmonary angiography was pertwo consecutive transverse images. If formed. The 99% condence interval for
only one transverse image demonstrated this incremental increase was calculated by
a lling defect, then a thrombus would using the standard approach.
not be diagnosed. Thrombus length was
measured by counting the number of
RESULTS
consecutive images on which the deep
venous thrombus was visible. Because a
A typical example of a deep venous thrombus
section width of 10 mm was used for
at indirect CT venography is shown in Figure
indirect CT venography, each image rep1. Among the rst 378 consecutive patients,
resents 1 cm of the thrombus length.

Image Interpretation

Radiology

140 mL dose of iohexol (Omnipaque 300;

Nycomed-Amersham, Princeton, NJ) was
injected at a rate of 3 mL/sec. Helical CT
scanning was performed (Hi-Speed Advantage CT/i; GE Medical Systems, Milwaukee, Wis), with a scan delay of 28
seconds. Images were obtained from the
diaphragm to the aortic arch, with a section width of 3 mm and a pitch of 1.6:1.
Scanning of the pelvis started from the
iliac crest 120 seconds after completion
of CT pulmonary angiography and continued to the popliteal fossa, with a section width of 10 mm and a pitch of 1:1.
Reconstruction for the CT pulmonary angiographic portion was performed at
1-mm intervals. Indirect CT venography
was evaluated by using standard 10-mmthick nonoverlapping images.
592

Radiology

February 2005

Cham et al

DVT, 43 (3%) had DVT in the absence of

pulmonary embolism, and 139 (9%) had
pulmonary embolism in the absence of
DVT (Table). Thus, the 43 (3%) patients

Results of CT Pulmonary Angiography and Indirect CT Venography

CT Pulmonary Angiography

Indirect CT
Venography
Positive for DVT
Negative for DVT
Inconclusive for DVT
Total

Positive for
Pulmonary
Embolism

Negative for
Pulmonary
Embolism

Inconclusive
for Pulmonary
Embolism

Total

100
139
4

43
1263
24

5
12
0

148 (9)
1414 (89)
28 (2)

243 (15)

1330 (84)

17 (1)

1590 (100)

Note.Data are the numbers of patients. Numbers in parentheses are percentages.

Figure 1. Indirect CT venogram of the left common femoral vein

shows a deep venous thrombus (arrow).

with DVT in the absence of pulmonary embolism and the ve (0.3%) patients with
DVT and nondiagnostic CT pulmonary anFigure 2.

Bar graph illu

strates the relative frequency of

thrombi of various lengths
among 378 consecutive patients
suspected of having pulmo
nary embolism. Among 33 patients
with DVT at indirect CT
venography, 24% had maximum
thrombus lengths of less th
an 5 cm.

phy into the CT pulmonary angiographic examination (11,21,22).

In previous studies, indirect CT venography has been shown to increase the diagnosis
of thromboembolic disease by 15%38%,
compared with the basic CT pulmonary angiographic examination (11,17,22). In this
study, we found that combined CT pulmonary angiography and indirect CT venography increases thromboembolic disease detection by 20% (99% condence interval: 17%,
DISCUSSION
23%) compared with CT pulmonary angiography alone. This result is similar to our previous ndings, where the detection rate of
Over the past years, CT pulmonary angiograthromboembolic disease was increased by
phy has become a commonly ordered exam18% with use of indirect CT venography in
ination. In many specic clinical situations, it
541 patients (11). Our current results lend
has replaced ventilation-perfusion scanning
further support to the consistent diagnostic
as the initial examination in the work-up of
yield that can be expected from indirect CT
pulmonary embolism (23,24). Likewise, comvenography.
bined CT pulmonary angiography and indiThere were several limitations in our study.
rect CT venography has become increasingly
First, only one radiologist read each image
common, with many institutions worldwide
from the combined CT pulmonary angiograroutinely incorporating indirect CT venograDVT was detected at indirect CT venograph enteen (1%) patients had inconclusive nd- phy and indirect CT venography. However,
y
ings at CT pulmonary angiography, and 28 this replicates real-world conditions. Second,
in 33 patients, a thrombus was seen on con- (2%) patients had inconclusive ndings at in- our criterion for thrombus diagnosis requires
tiguous sections in 21 patients, and two dis- direct CT venography. Twenty-two of the 28 that lling defects are seen on at least two
tinct thrombi were seen in 12 patients each. patients had poor boluses of the contrast consecutive sections; otherwise, possible llThe longest thrombus length for each pa- agent, which resulted in poorly opacied ing defects seen on only one section would be
tient is shown in Figure 2. Thrombus lengths veins, while six had lower extremity prosthe- attributed to partial volume averaging. Thus,
ranged from 2 to 81 cm. In eight (24%) pa- ses that resulted in nondiagnostic indirect CT there is potential for small clots of less than 1
tients, the maximum thrombus lengths mea venographic studies. In four of the 28 pa cm to be undiagnosed. Such an underestimation of clot length would however only add
sured 4 cm or less.
tients, pulmonary embolism was detected at credence to the need for contiguous section
Among the 1590 patients enrolled, onl CT pulmonary angiography.
intervals. Third, there were no standard comy
Of 1590 patients, 243 (15%) had pulmo- parisons performed in this study. Findings of
282 (18%) had positive ndings for pulmo- nary embolism, 148 (9%) had DVT, 100 published prospective studies in which indinary embolism, DVT, or both, while 126 (6%) had both pulmonary embolism and
3
(79%) patients had negative ndings for bot
h
pulmonary embolism and DVT (Table). Sev-

Radiology

giographic ndings adds up to 48 patients,

which is 20% of the 243 patients found to
have pulmonary embolism at CT pulmonary angiography. The addition of indirect
CT venography to CT pulmonary angiography resulted in a 20% incremental increase
in the detection rate of thromboembolic
disease (99% condence interval: 17%,
23%).

Volume 234

Number 2

rect CT venography was directly compar

ed
with sonography have conrmed our initial
ndings that indirect CT venography is similar to sonography in sensitivity and specic
ity (11,12,18).
As indirect CT venography becomes increasingly used at more institutions, there is
great need to optimize its effectiveness (11)
.
The radiation dose to the pelvis and gonads
is
an important consideration when perform
ing indirect CT venography. Using a com
bined CT pulmonary angiographic and indirect CT venographic protocol similar to ours,
Rademaker et al (21) measured patient
gonadal doses on the order of 2.110.7
mSv,
with variation between individuals and sex.
They found that the addition of indirect CT
venography increases the gonadal radiation
dose by 500- to 2000-fold compared with CT

Thromboembolic Disease Detection

pulmonary angiography alone.

Fortunately,
this increase in gonadal dose is
well below the
thresholds for deterministic radia
tion effects
provided in the International C
ommission
on Radiological Protection Public
ation 60, or
ICRP-60, guidelines (25). There
are several
stochastic effects that may arise
from irradiation of the pelvis during indirect C
T venography, such as leukemia and heri
table genetic
disease. Given the ICRP-60 stoch
astic risk estimate of 5% per sievert and a ca
lculated effective dose of about 2.5 mSv for
indirect CT
venography, the risk of radi
ation-related
death from leukemia is on the or
der of 1:8000
(21). Given the ICRP-60 stochasti
c risk estimate of 1% per sievert, this corre
sponds to a

593

genetic risk of about 1:15 000 among pa

tients
in the reproductive age group undergoin
g indirect CT venography (21). This genetic r
isk
does not exist for the majority of pa
tients
undergoing indirect CT venography who,
on
average, are older than 60 years. Like all
other
forms of x-raybased imaging, the ris
k of
mortality and morbidity from thrombo
embolic disease should be considered in the
context of these risks, especially for patients
in
the reproductive age group.
Some investigators have suggested th
e use
of discontinuous transverse sections,
with
5-cm gaps between sections (9,22). The
use of
discontinuous sections can reduce the in
tegral dose by as much as 80% but h
as the
potential to decrease specicity owing t
o in-

terpretive pitfalls (16,26,27). Noncontiguous

section intervals can adversely affect the interpretation of images along the iliac an
d
popliteal veins that run oblique to the trans- Even in the absence of interpretive pitfalls,
verse plane and are thus prone to partial volour current data suggest that use of a 5-cm
scan interval would result in a 40% chance o
ume artifacts that mimic a clot. Since contraf
missing the thrombi that measure less than
st
enhancement in the legs can be relatively lo 5
cm. The lengths of deep vein thrombi among
w
and there are subtle differences in the atten patients suspected of having pulmonary embolism strongly favor the use of contiguous
uation distribution, it is useful to see if a sub indirect CT venography for the pelvis and
lower extremities.
tle lling defect persists on more than o Some investigators have found that the fre
ne
image. This approach is used routinely in th quency of isolated pelvic DVT is relativel
y
e
interpretation of CT pulmonary angiographic uncommon, comprising 1%4% of all cases
studies, where it is often necessary to identi positive for DVT (22,28). Thus, another possible imaging option would be to scan only
fy
a clot on two or more consecutive images fo the legs in patients who are in the reproductive age group.
r
Several new technologies have the potena denitive diagnosis.
tial to decrease the radiation dose during
combined CT pulmonary angiography and
indirect CT venography, while hopefully
maintaining diagnostic effectiveness. Most
multidetector row CT scanners are now
equipped with dose-reducing applications
that automatically modulate the tube current
in the z-, x-, or y-axis. Depending on
patients
body habitus, this technology can reduce radiation dose by as much as 50%. In the future, clinical studies will be needed to determine whether a small reduction in radiation
dose, at the expense of a slightly increased
noise, will adversely affect the detection
of
thromboembolic disease.
References
1.

Goldhaber SZ. Pulmonary embolism. N Engl

J Med 1998; 339:93104.
2. Anderson FA Jr, Wheeler B, Goldberg RJ, et al.
A population-based perspective of the hospital
incidence and case fatality rates of deep veno
us
thrombosis and pulmonary embolism: the
Worcester DVT study. Arch Intern Med 1991;
151:933938.
3. Beyth RJ, Cohen AM, Landefeld CS. Long-term
outcomes of deep vein thrombosis. Arch Intern Med 1995; 155:10311037.
4. Gould MK, Dembitzer AD, Sanders GD, Garber
AM. Low-molecular-weight heparin compared
with unfractionated heparin for the treatment
of acute deep venous thrombosis. Ann Intern
Med 1999; 130:789799.
5. Moser KM. Venous thromboembolism. Am
Rev Respir Dis 1990; 141:235249.
6. Hull R, Delmore T, Genton E, et al. Warfarin
sodium versus low-dose heparin in the longterm treatment of venous thrombosis. N Engl
J Med 1979; 301:855858.
7. Hull R, Delmore T, Carter C, et al. Adjusted
subcutaneous heparin versus warfarin sodium
in the long-term treatment of venous thrombosis. N Engl J Med 1982; 306:189194.
8. Hull R, Hirsh J, Jay R, et al. Different intensities
of oral anticoagulant therapy in the treatment
of proximal-vein thrombosis. N Engl J Med
1982; 307:16761681.
9. 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 1998; 170:951954.
10. Yankelevitz DF, Gamsu G, Shah A, et al. Optimization of combined pulmonary CT angiography with lower extremity CT venography.
AJR Am J Roentgenol 2000; 174:6769.
11. Cham MD, Yankelevitz DF, Shaham D, et al.

Indirect CT venography in the detection of

17.

12.

13.

14.

15.

16.

deep vein thrombosis. Radiology 2000; 216:

744751.
Garg K, Kemp JL, Wojcik D, et al. Thromboembolic disease: comparison of combined CT
pulmonary angiography and venography with
bilateral leg sonography in 70 patients. AJR
Am J Roentgenol 2000; 175:9971001.
Duwe KM, Shiau M, Budorick NE, et al. Evaluation of the lower extremity veins in patients
with suspected pulmonary embolism: a retrospective comparison of helical CT venography
and sonography. AJR Am J Roentgenol 2000;
175:15251531.
Peterson DA, Kazerooni EA, Wakeeld TW, et
al. Computed tomographic venography is specic but not sensitive for diagnosis of acute
lower extremity deep venous thrombosis in
patients with suspected pulmonary embolus. J
Vasc Surg 2001; 34:798804.
Katz DS, Loud PA, Klippenstein DL, et al. Extrathoracic ndings on the venous phase of combined CT venography and pulmonary angiography. Clin Radiol 2000; 55:177181.
Ciccotosto C, Goodman LR, Washington L, et

18.

19.

20.
21.

22.

al. Indirect CT venography following CT pulmonary angiography: spectrum of CT ndings. J Thorac Imaging 2002; 17:1827.
Coche EE, Hamoir XL, Hammer FD, et al. Using dual-detector helical CT angiography to
detect deep venous thrombosis in patients
with suspicion of pulmonary embolism: diagnostic value and additional ndings. AJR Am J
Roentgenol 2001; 176:10351039.
Garg K, Kemp JL, Russ PD, Baron AE. Thromboembolic disease: variability of interobserver
agreement in the interpretation of CT venography with CT pulmonary angiography. AJR
Am J Roentgenol 2001; 176:10431047.
Bruce D, Loud PA, Klippenstein DL, et al. Combined CT venography and pulmonary angiography: how much venous enhancement is routinely obtained? AJR Am J Roentgenol 2001;
176:12811285.
Szapiro D, Ghaye B, Willems V, et al. Evaluation of CT time-density curves of lower-limb
veins. Invest Radiol 2001; 36:164169.
Rademaker J, Griesshaber V, Hidajat N, et al.
Combined CT pulmonary angiography and
venography for diagnosis of pulmonary embolism and deep venous thrombosis: radiation
dose. J Thorac Imaging 2001; 16:297299.
Loud PA, Katz DS, Bruce DA, et al. Deep

23.
24.

25.

26.

27.

28.

venous thrombosis with suspected pulmonary embolism: detection with combined

CT venography and pulmonary angiography. Radiology 2001; 219:498502.
Goodman LR, Lipchik RJ. Diagnosis of acute
pulmonary embolism: time for a new approach. Radiology 1996; 199:2527.
Holbert JM, Costello P, Federle MP. Role of
spiral computed tomography in the diagnosis
of pulmonary embolism in the emergency department. Ann Emerg Med 1999; 33:520528.
International Commission on Radiological
Protection. 1990 recommendations of the International Commission on Radiological Protection. ICRP Publication no. 60. Ann ICRP
1991; 21:6777.
Ghaye B, Szapiro D, Willems V, et al. Pitfalls in
CT venography of lower limbs and abdominal
veins. AJR Am J Roentgenol 2002; 178:1465
1471.
Ghaye B, Dondelinger RF. Non-traumatic thoracic emergencies: CT venography in an integrated diagnostic strategy for pulmonary embolism and venous thrombosis. Eur Radiol
2002; 12:19061921.
Goodman LR. Venous thromboembolic disease: CT evaluation. Q J Nucl Med 2001; 45:
302310.

594

Radiology

February 2005

Cham et al