Emerg Radiol

DOI 10.1007/s10140-016-1456-4

ORIGINAL ARTICLE

Detection of unsuspected pelvic DVTs on abdominopelvic CT

scans: a potentially life-saving diagnosis
Mougnyan Cox 1 & Manisha Patel 1 & Zhenteng Li 1 & Sarah Kamel 1 &
Sandeep Deshmukh 1 & Christopher Roth 1 & Laurence Needleman 2

Received: 7 September 2016 / Accepted: 14 October 2016

# American Society of Emergency Radiology 2016

Abstract Venous thromboembolism (VTE) is a serious
common disorder with substantial cost and morbidity to
society and can be life threatening in some cases. The
majority of VTE is diagnosed on lower extremity ultra-
sound or CT pulmonary angiography, but some cases of
deep venous thrombosis (DVT) may be occasionally di-
agnosed on CT of the abdomen and pelvis by the alert
radiologist. The purpose of our study was to determine
the fraction of new/unsuspected DVTs diagnosed on
CTAP and the subsequent management and clinical
course of these patients. After Institutional Review
Board approval, a retrospective search of an institutional
imaging database was performed for all cases of DVTs
diagnosed on CTs of the abdomen and pelvis. Patients
with positive studies were further investigated via clinical
chart review for their subsequent management and clinical
course. The 90-day mortality of the patients diagnosed
with DVT on CTAP was also recorded. Sixty-two patients
met the criteria for positive DVT on CTAP. Of these 62
cases, 26 (42 %) were new. Management was substantial-
ly changed in 24 out of 26 cases (92 %), most commonly
initiation of anticoagulation. The 90-day mortality rate of
patients diagnosed with pelvic DVTs on CTAP in our
cohort was 21 %. Timely detection of pelvic DVTs can
* Mougnyan Cox
mougnyan.cox@gmail.com
1
Department of Radiology, 1087 Main Building, 132 South 10th
Street, Philadelphia, PA 19107, USA
2
Department of Radiology, Jefferson Vascular Center, Noninvasive
Vascular Laboratory, 1087 Main Building, 132 South 10th Street,
Philadelphia, PA 19107, USA
have serious implications for patient management, mor-
bidity, and mortality. The pelvic veins should be included
in the search pattern of all radiologists who review CTs of
the abdomen and pelvis.
Keywords Common . Iliac . Pelvic . Vein . Thrombosis .
Thrombus . DVT . CT . Pelvis
Introduction
Venous thromboembolism (VTE) is a common disorder,
with an estimated annual incidence of 104 per 100,000
patients [1]. Estimated mortality rates for patients diag-
nosed with acute pulmonary embolism (PE) and/or deep
venous thrombosis (DVT) vary, with some estimates as
high as 20 % [2, 3]. Most cases of venous thromboembo-
lism are diagnosed on CT pulmonary angiography
(CTPA) or lower extremity ultrasound (LEUS). Common
iliac vein thrombosis has been described in patients with
extrinsic iliac vein compression syndrome [4, 5], and
there is an associated risk of central embolization and
subsequent PE in this condition (Figs. 1 and 2). The an-
ecdotal experience at our institution is that pelvic DVTs
(thrombi in common iliac, external iliac, or femoral veins)
can occur in patients without evidence of extrinsic com-
mon iliac vein compression, and some of these cases may
only be detected on CT (apart from venograms performed
in the interventional radiology suite). The purpose of our
study was to characterize the clinical characteristics and
subsequent management of patients with unsuspected pel-
vic DVTs discovered on abdominopelvic CTs (CTAP).

Fig. 1 A 41-year-old woman
with post-operative fever and
abdominal pain following
laparoscopic resection of an
endometrioma. The first image
shows post-operative changes in
the uterus with overlying skin
thickening. The second figure
shows a thrombus in the right
common iliac vein (arrow)
Methods
Study design
Institutional Review Board approval was obtained prior to
initiating this study. A retrospective review of an imaging
database at a tertiary-care referral center was performed from
June 2008 to December 2015. Primordial (Primordial Inc.,
San Mateo, CA), a database search engine, was used to inter-
rogate reports of CTAPs performed at our institution. The
following search terms were entered into Primordial: com-
mon, iliac, pelvic, DVT, venous, thrombus, thrombosis, exter-
nal, and internal. Multiple permutations of these words were
entered into the search engine until no new cases of common
iliac DVT were found.
Imaging protocol
Abdominopelvic CTs of patients reported to have pelvic
DVTs were reviewed. A total of 169, 800 CTAPs were per-
formed during the study period, of which 78,500 were
contrast-enhanced CTAP performed in the portal venous
Fig. 2 A 71-year-old woman with no known history of malignancy. This
image nicely illustrates asymmetric enlargement of the left common
femoral vein (arrow), with rim-enhancement of the walls and
penivenular infiltration of the subcutaneous fat when compared with the
normal contralateral right common femoral vein
Emerg Radiol
phase. All CTAPs were performed on a 64-slice multidetector
CT scanner (Lightspeed, General Electric Medical Systems,
Milwaukee, WI, USA and Brilliance, Phillips Healthcare,
Amsterdam, Netherlands). Iodinated contrast was injected in-
travenously at a rate of 4 mL/s for a total volume of 120 mL.
All patients were scanned in the portal venous phase (90 s
delay following intravenous injection of contrast). Positive
oral contrast was administered in approximately half of the
cases for optimal bowel opacification. Axial slices were ob-
tained from the middle of the heart to the lesser trochanter at 5-
mm intervals. Sagittal and coronal reconstructions were per-
formed from 1.25-mm axial slices. All the CTAPs were per-
formed at 120 kVp. The initial CTAPs were read by members
of the body imaging division, comprising of 10 board-certified
attendings with 830 years post-fellowship experience. The
studies identified as positive were reviewed and confirmed
by two board-certified fellowship-trained abdominal radiolo-
gists, with 8 and 13 years of post-fellowship experience.
Positive studies were identified by a filling defect in the com-
mon iliac or pelvic vein that persisted on more than two con-
secutive slices, with or without a thin enhancing venous wall
and perivenular inflammatory change.
Clinical correlation
Patients with positive studies were further investigated by
reviewing their electronic medical records (EMR). While risk
factors for hypercoagulability such as malignancy, history of
prior VTE, oral contraceptives, and recent surgery were noted
when present, the main focus of the medical chart review was
to determine whether the pelvic DVT was a new finding or
already known at the time of detection on CT. The EMR was
also reviewed to determine the management of this patient
following the diagnosis of pelvic DVTs. Specifically, the fol-
lowing parameters were recorded: any confirmatory studies
and their results (for example, lower extremity venograms),
whether anticoagulation was performed, and whether an infe-
rior vena cava filter was placed for embolic protection. The
90-day mortality rate for patients with pelvic DVTs was also
recorded in all cases where follow-up was performed in the
outpatient clinic.
Emerg Radiol

Results Table 2 Summary of study results

Total patients with DVT 62 cases

Sixty-three patients met the criteria for pelvic DVT identified No. of new/unsuspected DVT 26 cases (42 %)
on CTAP by the initial interpreting radiologist. Upon further Cases resulting in new/different anticoagulation 15 cases
review, one case was deemed to be secondary to venous in- No. of cases resulting in new IVC filter 4 cases
flow of unopacified blood and was not included in the analy- No. of unilateral DVT 44 cases (71 %)
sis. A total of 62 patients were reviewed in this study (see
No. of isolated common femoral DVT 17 cases (27 %)
Table 1 for summary of patient demographics). Thirty-one
IVC involvement 12 cases (19 %)
patients were female. The average age was 55.8 years (range
Co-existent PE on CTAP 2 cases (3 %)
21 to 90 years), with a median age of 58. Of the 62 patients
90-day mortality in 56 patients with DVT on CTAP 12 cases (21 %)
with pelvic DVT, 26 cases (42 %) were new and unknown to
the referring clinical team prior to the CTAP interpretation
(see Table 2 for a concise summary of the results). Half of
these cases (13 cases) had a confirmatory ultrasound per- out metastasis/malignancy. Sixteen patients diagnosed with
formed within a day following the CTAP, and all were positive DVT on CTAP had a normal US and no reported history of
for DVT. Twelve of the 26 patients had no confirmatory study DVT within the 12 months preceding their initial CT diagno-
performed, and management was based purely on the findings sis, and an additional 21 patients had no US performed and no
on CT (similar to management of thrombi detected in pulmo- reported history of DVT within the 12 months preceding their
nary arteries). One patient had a confirmatory US performed CT study (total of 33 cases or 53 %).
1 day after the CTAP, which showed dampening of the wave- Forty-four cases of DVT were unilateral, and 18 were
forms and pelvic adenopathy but no definitive thrombus with- bilateral. When only a single vein was involved, the com-
in the limits of the study. mon femoral veins were most commonly implicated (17
Management of the 26 patients was also affected by the cases). The remainder of the cases involved some combi-
CTAP interpretation in the majority of cases. Fifteen of the nation of the common iliac, external iliac, and femoral
26 patients were started on anticoagulation, four had a new veins. Inferior vena cava (IVC) involvement was seen in
IVC filter placed, and three were switched to hospice 12 out of 62 cases, of which 10 had a pre-existing IVC
care/comfort measures only. Four patients were already on filter. Pulmonary artery thrombus was seen on the basilar
anticoagulation at the time of the study due to a remote history slices of the chest on CTAP in 2 patients, consistent with
of DVT, and 2 of these patients were switched to a different acute pulmonary embolism.
anticoagulant due to compliance issues. Overall, management Follow-up at 90 days was documented when available
was substantially changed in 24 out of 26 cases (92 %) in in our group of 62 patients. Six patients had no further
keeping with the well-recognized morbidity/mortality risk as- follow-up. Of the 56 patients in whom follow-up was
sociated with DVT. available, 12 died within 90 days (21 %). All of the 12
Thirty-three patients had a history of cancer, 31 active and patients had active malignancy at the time of diagnosis (4
on therapy, and 2 remote (Table 3). Four patients had a history patients with pancreatic cancer, 3 with lung cancer, 2 with
of recent surgery (hip fracture, tibial plateau fracture, varicose
vein resection, and endometrioma resection). The average in-
ternational normalized ratio (INR) at the time of CT diagnosis Table 3 Prevalence of
different types of active Cancer (CA) Number
was 1.3, and only 7 patients had an INR above 2.0. In terms of cancers in study group of patients
the given indication on the request for the CTAP, only 13 cases
specifically mentioned DVT. The most common indications Pancreatic CA 4
listed on CT were abdominal pain, rule out abscess, and rule Uterine CA 4
Bladder 3
Lung CA 3
Table 1 Summary of patient demographics Lymphoma 3
Total no. of patients with DVT 62 Prostate CA 3
Average age 55.8 (median 58, range Glioma 2
21 to 90 years) Hepatocellular CA 2
a
Active malignancy 31 cases (50 %) Other CA 7
Recent surgery 4 cases (6 %) a
Other CA means one case each of colon
Average INR at diagnosis 1.3 CA, breast CA, ovarian CA, cervical CA,
No. of patients with INR >2 at diagnosis 7 (11 %) gastric CA, melanoma, and peritoneal me-
tastases of unknown primary

lymphoma, and one patient each with glioblastoma, uter-
ine, and gastric cancers).
Discussion
Our study shows that pelvic DVTs are readily identified
on CT of the abdomen and pelvis; a substantial fraction of
which are new or unknown to the referring team and the
patient. Timely diagnosis of these DVTs has the potential
to avert some of the morbidity and mortality associated
with venous thromboembolism, while missing these le-
sions may have grave implications for the patient. CT
diagnosis of new DVT resulted in a change in manage-
ment for a number of our patients (over 90 %), many of
whom had definitive therapy with anticoagulation. The
90-day mortality of our cohort was high at 21 %, but this
number is likely due to the presence of underlying malig-
nancy in these patients. All 44 patients with DVT on
CTAP and without active malignancy at the time of diag-
nosis survived to 90 days.
Lower extremity ultrasound is considered the gold
standard for lower extremity DVT, and a head-to-head
comparison of lower extremity indirect CT venograms
(CTV) to lower extremity ultrasounds showed that CTV
was less sensitive and specific than LEUS [6]. Lower
extremity ultrasound is usually performed from the ingui-
nal ligament down to the level of the calf. As a result,
lower extremity ultrasound is limited for the direct evalu-
ation of the pelvic veins [7], although a central venous
obstruction can be inferred from asymmetric dampening
of the waveforms or loss of respiratory variation or cardi-
ac pulsatility [8]. CTPA performed in the ER has an esti-
mated positivity rate of 10 %, but obviously misses all
cases of lower extremity DVT [9]. Theoretically, interro-
gating the pulmonary arteries and lower extremity deep
veins with a single imaging study will increase the yield
for VTE and decrease the morbidity and mortality from
this disease. While CTPA and CTV can be combined in a
single exam to search for PE and DVT, this protocol is not
commonly performed in many emergency rooms because
of the high gonadal radiation dose and the time/expense
of the procedure [3]. Johnson et al. studied 427 consecu-
tive ER patients with suspected PE and found only one
patient with isolated DVT (DVT without PE), at an esti-
mated cost of 206,400 dollars to detect that single case
[10]. Kalva et al. reviewed CT images of 2074 patients
and found isolated pelvic DVT in only two patients
(0.1 %), leading them to conclude that CT venography
of the pelvis is of questionable benefit and may be limited
to the legs [11]. Cham and his associates had a slightly
higher positivity rate for more proximal lower extremity
venous clots on CTV compared with ultrasound: that
Emerg Radiol
study examined 541 patients and discovered isolated
DVTs only seen on CTV [12]. Similar findings were re-
ported by Loud and associates in a study of 650 patients,
with 4 DVTs found on CTV that were initially missed on
ultrasound, but confirmed on a repeat lower extremity
ultrasound [13]. While we do not advocate screening for
pelvic vein DVT in patients suspected of having PE, di-
agnosing a common iliac thrombus on CT can be of sig-
nificant clinical benefit to the patient, possibly even life
saving if timely and appropriate therapy is undertaken.
Our study had several limitations, including limitations
inherent to all retrospective studies. Probably the most
important limitation of our study was that of patient se-
lection: only the patients with common iliac vein DVTs
that were detected by the alert radiologist at the time of
interpretation were included in the study. The common
iliac vein is a potential blind spot on CT, either because
of poor venous opacification or simply due to the lack of
close visual inspection by the interpreting radiologist.
Also, none of the patients with noncontrast CTAPs were
included in this study; a small number of whom could
have had common iliac DVTs.
Conclusions
The pelvic veins can be a potential blind spot due to
anatomic and technical reasons, and pelvic DVTs are oc-
casionally encountered on CTAP performed for other rea-
sons. Timely detection of pelvic DVTs can have serious
implications for patient management and associated mor-
bidity/mortality, and the pelvic veins should be included
in the search pattern of all radiologists who review CTs of
the abdomen and pelvis.
Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of
interest.
References
1. Spencer F, Emery C, Lessard D et al (2006) The Worcester venous
thromboembolism study: a population-based study of the clinical
epidemiology of venous thromboembolism. J Gen Intern Med 21:
722727
2. Naess I, Christiansen S, Romundstad P et al (2007) Incidence and
mortality of venous thrombosis: a population-based study. J
Thromb Haemost 5:692699
3. Komissarova M, Chong C, Frey K et al (2013) Imaging of acute
pulmonary embolism. Emerg Radiol 20:89101
Emerg Radiol
4. Kwak H, Han Y, Lee Y et al (2005) Stents in common iliac vein
obstruction with acute ipsilateral deep venous thrombosis: early and
late results. J Vasc Interv Radiol 16:815822
5. Oguzkurt L, Ozkan U, Ulusan S et al (2008) Compression of the left
common iliac vein in asymptomatic subjects and patients with left
iliofemoral deep vein thrombosis. J Vasc Interv Radiol 19:366370
6. Garcia-Bolado A, Del Cura J (2007) CT venography vs ultrasound
in the diagnosis of thromboembolic disease in patients with clinical
suspicion of pulmonary embolism. Emerg Radiol 14:403409
7. Laissy J, Cinqualbre A, Loshkajian A et al (1995) Assessment of
deep venous thrombosis in the lower limbs and pelvis: MR venog-
raphy versus duplex doppler sonography. Am J Roentgenol 167:
971975
8. Selis J, Kadakia S (2009) Venus doppler sonography of the extrem-
ities: a window to pathology of the thorax, abdomen, and pelvis.
Am J Roentgenol 193:14461451
9. Parikh N, Morris E, Babb J et al (2015) MDCT diagnosis of acute
pulmonary embolism in the emergent setting. Emerg Radiol 22:
379384
10. Johnson J, Brown M, McCullough N et al (2006) CT lower extrem-
ity venography in suspected pulmonary embolism in the ED. Emerg
Radiol 12:160163
11. Kalva S, Jagannathan J, Hahn P et al (2008) Venous thrombo-
embolism: indirect CT venography during CT pulmonary
angiography-should the pelvis be imaged? Radiology 246:
605611
12. Cham M, Yankelevitz D, Shaham D et al (2000) Deep venous
thrombosis: detection by using indirect CT venography.
Radiology 216:744751
13. Loud P, Katz D, Bruce D et al (2001) Accuracy of combined CT
venography and pulmonary angiography for lower extremity DVT.
Radiology 219:498502