Dual-energy CT angiography of the lungs: comparison of test bolus and bolus tracking techniques for the determination of scan delay.

Please cite
techniques
ARTICLE IN PRESSGModelEURR-4843; No.of Pages7
European Journal of Radiology xxx (2010) xxx–xxx
Contents lists available at ScienceDirect
European Journal of Radiology
journa l homepage: www.e lsev ier .co
Dual-e pa
trackin can
Thomas Kris
Stefan H
Department of al Facu
Theodor-Kutze
a r t i c l
Article history:
Received 20 January 2010
Received in revised form 14 June 2010
Accepted 16 June 2010
Keywords:
Test bolus
Bolus tracking
Pulmonary em
Dual-energy C
Contrast agent
est bo
pulmonary dual-energy CT angiography in patients with suspected pulmonary embolism.
Materials and methods: 60 consecutive patients referred for CTA for exclusion of PE were randomized
either into a test bolus group or into a bolus tracking group. All exams were performed on a 64-channel
dual source CT scanner. A standard single-acquisition dual-energy CTA was performed after injection of
100ml Iomeprol 400 followed by a saline chaser of 4ml/s. The scan delay was determined using either
1. Introdu
Dual-en
tive visuali
in the lung
ies but also
examinatio
racy of pulm
embolism (
strated that
with a good
“perfusion”
provide add
∗ Correspon
E-mail add
(T. Henzler).
0720-048X/$ –
doi:10.1016/j.this article in press as: Henzler T, et al. Dual-energy CT angiography of the lungs: Comparison of test bolus and bolus tracking
for the determination of scan delay. Eur J Radiol (2010), doi:10.1016/j.ejrad.2010.06.023
bolism
T
administration
test bolus (n=30) or bolus tracking (n=30). Test bolus was performed using an additional 20ml Iomeprol
400 injected with a rate of 4ml/s during acquisition of a series of dynamic low-dose monitoring scans
followed by injection of a saline bolus of 20ml using the same flow rate. For DECT angiography of the
lungs 100ml Iomeprol 400 was injected with an injection rate of 4ml/s followed by a saline chaser of
20ml using the same flow rate. Attenuation profiles of different vascular segments (pulmonary arteries,
pulmonary parenchyma, aorta, all 4 heart chambers) were measured to evaluate the timing techniques.
Overall image quality of dual-energy “perfusion” maps and virtual 120kV CTA images was evaluated by
two radiologists regarding the present of artifacts.
Results: In all patients an adequate and homogeneous contrast enhancement ofmore than 400Hounsfield
units (HU) was achieved in the different vascular districts. No statistically significant difference between
test bolus and bolus tracking was found regarding vessel attenuation or overall image quality.
Conclusion: A homogeneous opacification of the different vascular territories and the pulmonary
parenchyma as well as a sufficient image quality can be achieved with either bolus tracking or test
bolus techniques.
© 2010 Elsevier Ireland Ltd. All rights reserved.
ction
ergy CT angiography (DE CTA) of the lungs allows selec-
zation of the distribution of iodinated contrast media
parenchyma [1]. As not only the pulmonary arter-
the capillary “perfusion” can be assessed in a single
n, the technique may improve the diagnostic accu-
onary CT angiography for the detection of pulmonary
PE). In a recently published study, Thieme et al. demon-
DECTA is able todisplaypulmonary “perfusion” defects
agreement to scintigraphic findings by using DE CTA
maps [2]. Therefore, DE CTA “perfusion” maps might
itional value to detect small peripheral or subsegmen-
ding author. Tel.: +49 621 383 2067; fax: +49 621 383 3817.
resses: Thomas.Henzler@umm.de, thomas.henzler@gmx.de
tal PE. However, DE CTA “perfusion” maps of the lungs may suffer
from misleading artificial “perfusion” defects caused by streak arti-
facts from dense contrast material in the great thoracic vessels and
heart chambers.
To optimize the results of DE CTA of the lungs, the imaging
and injection protocols have to be adapted to each other, as the
technique requires both a high vessel attenuation of the central
pulmonary arteries and sufficient attenuation of the downstream
pulmonary capillary bed. Consequently, the optimum scan start
for DE CTA of the lungs should be slightly delayed compared to
a standard single-energy CTA protocol without missing the main
contrast material bolus within the pulmonary arteries. However,
without knowing the patient’s hemodynamic profile, an optimal
timing of the scan start may be difficult to predict, especially in
patients with PE. It is well known that PE can lead to an increase in
thepulmonary arterial resistancewhichhampers thepassageof the
contrast material from the pulmonary arteries into the pulmonary
capillary system [3]; on the other hand, the contrast bolus may be
see front matter © 2010 Elsevier Ireland Ltd. All rights reserved.
ejrad.2010.06.023nergy CT angiography of the lungs: Com
g techniques for the determination of s
Henzler ∗, Mathias Meyer, Miriam Reichert, Radko
aneder, Stefan O. Schoenberg, Christian Fink
Clinical Radiology and Nuclear Medicine, University Medical Center Mannheim, Medic
r-Ufer 1-3, D-68167 Mannheim, Germany
e i n f o a b s t r a c t
Objective: To prospectively compare tm/locate /e j rad
rison of test bolus and bolus
delay
sak, John W. Nance Jr.,
lty Mannheim, Heidelberg University,
lus and bolus tracking for the determination of scan delay of

Please cite y of
techniques .ejrad
ARTICLE IN PRESSGModelEURR-4843; No.of Pages7
2 T. Henzler et al. / European Journal of Radiology xxx (2010) xxx–xxx
missed if the scan delay is extended too much, causing inadequate
attenuation of the pulmonary arteries.
Excluding a fixed empirical delay, bolus tracking (BT) and test
bolus (TB) are the most frequently used techniques for the deter-
mination of scan delay in clinical routine. Until now it is unclear
whether one technique is more suitable for DE CTA of the lungs in
patients with suspected acute PE. Therefore, the aim of this study
was to compare both intravenous contrast media administration
techniques with regard to arterial enhancement and image quality
of the CTA as well as enhancement and image quality of DE iodine
maps.
2. Materials and methods
2.1. Patients
Prior to enrollment, the study protocol was approved by the
institutional ethics committee of our institution. Written informed
consentwasacquired fromall subjects after thenatureof theproce-
dure had been fully explained. 60 consecutive patients (27 women,
33 men, mean age 62.8±16.6 years) who were referred for CTA
to our department for exclusion of PE were prospectively enrolled
in this study. After enrollment all patients were divided either
into a TB group or into a BT group for the determination of scan
delay.
2.2. Scanning protocol
All exam
ner (Somat
Germany).
correspond
angular offs
provides a
array (corre
For CTA
Tube voltag
datasets we
and 30% of
pensate for
current wa
matic tube current modulation (CARE Dose 4D) was used in all
patients. The detector collimation was set to 14mm×1.2mm. A
separate dataset for each tube kV as well as one linearly weighted
average dataset (“virtual 120 kV”)was calculatedwith a slice thick-
ness of 2mm and a reconstruction increment of 1.4mm using a
soft tissue kernel (D30f). For DECT angiography of the lungs 100ml
Iomeprol 400 (Imeron 400, Bracco Imaging S.p.A., Milan, Italy) was
injected over an antecubital vein using a power injector (Stellant®
D CT Injection System MEDRAD, INC, Warrendale, USA) with an
injection rate of 4ml/s followed by a saline chaser of 20ml using
the same flow rate. All scans were performed in caudo-cranial scan
direction during breathhold at a mid-inspiratory level. Two differ-
ent methods were used to determine the scan delay for DE CTA.
2.2.1. Test bolus technique
TB was performed in 30 patients. 20ml Iomeprol 400 (Imeron®
400) were additionally injected with a rate of 4ml/s during the
acquisition of a series of dynamic low-dose monitoring scans
(120kV, 20mAs). The contrast injection was followed by injec-
tion of a saline bolus of 20ml using the same flow rate. ROIs were
placed within the pulmonary trunk and the left atrium to calculate
enhancement/time curves. Acquisition of the dynamic monitoring
scans started 5 s after the beginning of the injection. The delay
between each monitoring scan was set to 1 s. The intersection of
both attenuation curves was assumed to be the optimal scan delay
with an additional delay of 5 s for breathing instructions (Fig. 1). DE
CTA of the lungs was performed within a single breathhold using
100ml Iomeprol 400 at a flow rate of 4ml/s followed by injection
line b
Bolus
was
tion
ny).
follow
te. R
cont
ectio
he R
hres
an st
Fig. 1. Typical term
to 100HU (A). itiona
the monitorin ) show
the attenuatio al scan
interpretation rsion othis article in press as: Henzler T, et al. Dual-energy CT angiograph
for the determination of scan delay. Eur J Radiol (2010), doi:10.1016/j
swereperformedona64-channel dual-source CT scan-
om Definition, Siemens Health Care Sector, Forchheim,
The system is equipped with two X-ray tubes and two
ing detectors which are oriented in the gantry with an
et of 90◦. One detector array (corresponding to tube A)
field of view (FOV) of 50 cm, while the other detector
sponding to tube B) is restricted to a FOV of 26 cm [4,5].
of the lungs, a single-acquisition, DE protocol was used.
es were set to 140kV (tube A) and 80kV (tube B). DECT
re calculated by merging 70% of the 140kV spectrum
the 80kV spectrum (weighting factor 0.3). To com-
the lower photon output of tube B the reference tube
s set to 235mA for tube B and 50mA for tube A. Auto-
of a sa
2.2.2.
BT
applica
Germa
4ml/s
flow ra
of the
the inj
1.5 s. T
gered t
The sc
time attenuation profiles of bolus tracking (A) and test bolus (B) exams used for de
After the trigger threshold was reached the scan started automatically with an add
g scans was set to 1.5 s. (B) The attenuation profile of a test bolus exam. Curve 1 (red
n within the left atrium. The intersection of both curves was assumed as the optim
of the references to color in this figure legend, the reader is referred to the web vethe lungs: Comparison of test bolus and bolus tracking
.2010.06.023
olus of 20ml using the same flow rate.
tracking technique
performed in 30 patients using a dedicated software
(CARE bolus Siemens Health Care Sector, Forchheim,
100ml Iomeprol 400 were injected with a flow rate of
edby injectionof a salinebolus of 20mlusing the same
eal time low dose monitoring scans (120kV, 20mAs)
rast media were performed 5 s after the beginning of
n. The delay between each monitoring scan was set to
OI was placed within the pulmonary trunk. The trig-
hold within the ROI was set to ≥100HU above baseline.
arted automatically after the triggered threshold was
ination of scan delay. The trigger threshold for bolus tracking was set
l delay of 5 s for breathing instructions. The interscan delay between
s the attenuation within the pulmonary trunk. Curve 2 (blue) shows
delay with an additional delay of 5 s for breathing instructions. (For
f the article)

Please cite
techniques
ARTICLE IN PRESSGModelEURR-4843; No.of Pages7
T. Henzler et al. / European Journal of Radiology xxx (2010) xxx–xxx 3
Fig. 6. Subjectively perceived image quality of the virtual 120kV dataset which was calculated by merging 70% of the 140kV spectrum and 30% of the 80kV spectra. Score:
1 =poor; 2 = fair; 3 =moderate; 4 = good; 5= excellent.
reached with an additional delay of 5 s for breathing instructions
(Fig. 1).
2.3. Image
The ima
by two rad
tion in che
to the used
multi-moda
processing
Forchheim,
“perfusion”
maps the io
decomposit
enhanceme
All imag
enhanceme
placing ROI
pulmonary
heart cham
CTA datase
placed within the lung parenchyma of the right lower lobe exclud-
ing pulmonary vessels. From these measurements the mean HU
as well as the “overlay value” within the ROIs were collected. The
y va
ithi
ted i
cau
0 lun
ent w
strea
ithi
the
jecti
rtual
int s
ce of
; 5:
atisti
istic
soft
Fig. 7. Subjectanalysis
ge analysis was performed in a consensus reading
iologists (board certified radiologist with specializa-
st radiology, 3rd year resident) who were blinded
technique. All scans were processed and read on a
lityworkstation equippedwith commercial DECTpost-
software (Syngo VA11, Siemens Health Care Sector,
Germany). Using the PBV algorithm, 3D color-coded
maps of the lungs were calculated. This algorithm
dine content of lung tissue based on a three material
ion for air, soft tissue and iodine [5]. The iodine-related
nt can be quantified in HU and color-coded.
es were analyzed regarding presence of PE. Contrast
nt was measured on the “virtual 120kV” dataset by
s within ascending aorta, pulmonary trunk, left main
artery, rightmain pulmonary artery andwithin the four
bers. Parenchymal attenuationwasmeasured on theDE
ts using the PBV algorithm. Three different ROIs were
“overla
ation w
calcula
defects
of 119
consist
the up
ation w
within
Sub
the “vi
a 5-po
presen
4: good
2.4. St
Stat
tisticalthis article in press as: Henzler T, et al. Dual-energy CT angiography of
for the determination of scan delay. Eur J Radiol (2010), doi:10.1016/j.ejrad
ivelyperceived imagequalityof the3DDECTcolor-coded“perfusion”mapsusing thePBValue” quantifies the contribution of iodine to the attenu-
n the ROI (in HU) and therefore represents the pure DE
odine distribution within the parenchyma. “Perfusion”
sed by PE were analyzed on a segmental basis (in a total
g segments). “Perfusion” defects were considered to be
ith PE if thrombembolic material was present within
m vessel on the “virtual 120kV” CTA dataset. Attenu-
n “perfusion” defects was measured by placing a ROI
“perfusion” defect.
ve image quality of the DE “perfusion” maps and of
120 kV” CTA datasets were rated separately by using
core concerning homogeneous vessel attenuation and
beam hardening artifacts (1: poor; 2: fair; 3: moderate;
excellent) (Figs. 2–5 and 8).
cal analysis
al calculations were performed by using SAS JMP8® sta-
ware (SAS Institute, Cary, North Carolina, USA). To rulethe lungs: Comparison of test bolus and bolus tracking
.2010.06.023
lgorithm.Score:1 =poor;2 = fair; 3 =moderate;4 = good;5= excellent.

Please cite
techniques
ARTICLE IN PRESSGModelEURR-4843; No.of Pages7
4 T. Henzler et al. / European Journal of Radiology xxx (2010) xxx–xxx
Fig. 2. Dual energy CTA of 62-year-old man with suspected pulmonary embolism in axial and sagittal orientation. Scan delay was determined using bolus tracking. The
“perfusion” map was rated as excellent (5) and shows no relevant artifacts regarding the diagnostic quality.
Fig. 3. Dual en termi
rated as good
out signific
variance w
weight. Con
deviation. C
sels, the hea
Fig. 4. Dual en
Scan delaywa
rialwithin the
leading to false
A. Pulmonary
caused a true
black arrowhe
multiple streaergy CTA of a 58-year-old man in axial and sagittal orientation. Scan delay was dethis article in press as: Henzler T, et al. Dual-energy CT angiography of
for the determination of scan delay. Eur J Radiol (2010), doi:10.1016/j.ejrad
(4) as there were only minor artifacts beside the right atrium and in the right upper lobe
ant differences between the two groups, an analysis of
as applied to the following parameters: age, sex and
tinuous variables are presented as mean± standard
omparison of CT density values (HU) within the ves-
rt chambers, lung parenchyma and within “perfusion”
ergy CTA of a 70-year-old femalewith bilateral pulmonary embolism.
s determinedwith bolus tracking. Highly concentrated contrastmate-
heart chambers andmediastinal vessels caused severe streak artifacts
positive “perfusion” defects indicated by the white arrows on image
embolism indicated by the white arrows on the images B and C
positive “perfusion” defect in the right lower lobe indicated by the
ad on image B. The image quality was rated as moderate (3) because
k artifacts that caused false positive artificial “perfusion” defects.
defects we
distribution
variables (i
quality of t
with 25–75
Kruskal–W
sidered sign
Fig. 5. Dual e
test bolus. Con
the vena cava
homogeneous
and not as pooned using a test bolus. The image quality of the “perfusion” map wasthe lungs: Comparison of test bolus and bolus tracking
.2010.06.023
caused by dense contrast material.
re performed using a two sample t-test after normal
was confirmed using the Shapiro–Wilk test. Ordinal
mage quality of the “virtual 120kV” dataset and image
he DECT “perfusion” maps) are presented as median
% interquartile ranges and are compared using the
allis analysis of variance. A p value of ≤0.05 was con-
ificant.
nergy CTA of a 60-year-old man. Scan delay was determined with a
trast material is concentrated within the right subclavian vein and
superior leading to severe streak artifacts in the upper lung and a
“perfusion” of the lower zone. The image quality was rated as fair
r because of the homogeneous “perfusion” in both lower lobes.

Please cite
techniques
ARTICLE IN PRESSGModelEURR-4843; No.of Pages7
T. Henzler et al. / European Journal of Radiology xxx (2010) xxx–xxx 5
Table 1
Mean attenuation (HU), standard deviation within different vascular districts.
Test bolus Bolus tracking t-Test twest bolus vs. bolus tracking (p-value)
Superior vena cava 640.7 ± 43.3 605.5 ± 27.5 n.s.
Pulmonary trunk 590.3 ± 37.5 595.2 ± 36.9 n.s.
Right main pulmonary artery 577.9 ± 33.7 579.3 ± 32.6 n.s.
Left main pulmonary artery 584.3 ± 33.3 592.9 ± 32.8 n.s.
Ascending aorta 446.7 ± 23.5 592.9 ± 32.8 n.s.
Right ventricle 597.6 ± 34.7 585.9 ± 34.1 n.s.
Left ventricle 433.9 ± 26.9 478.9 ± 26.5 n.s.
Right atrium 746.1 ± 43.4 653.3 ± 42.7 n.s.
Left atrium 461.1 ± 26.2 506.6 ± 25.8 n.s.
Pulmonary parenchyma −785.4 ± 10.8 −794.1 ± 10.6 n.s.
Pulmonary parenchyma DE overlay 33.1 ± 2.1 34.0 ± 2.0 n.s.
3. Results
Patient d
the two gr
exams were
patients. Th
fully diagno
men) patien
the BT grou
3.1. Contra
While th
and overlay
reached sta
(Table 2). B
ment in th
arteries, lef
the TB grou
within the
The DECT ca
higher usin
regions is g
3.2. “Perfus
“Perfusio
of the TB gr
“perfusion”
a total of 17
within “per
and −836.7
valueof the
andBT, resp
value for pu
33.2±2.1 a
3.3. Image
The med
dataset was
(p=0.289) (
3 for
sion”
tively
pecti
cussi
form
s ass
ution
zatio
relev
e th
ents
is he
thic p
ulmo
nary
the s
ta ac
d to
as on
illar
sion”
of D
E, all
sting
oten
mis
aim
on o
ted P
a m
vess
perfu
h th
ins
s cho
el low
terv
st m
ation
Table 2
Summary of p
Age
Weight
Height
BMI
Scan delaya
a Note thatemographics were not significantly different between
oups in terms of age, sex and weight (Table 1). All
acquired successfully and without complications in all
ere were no technical errors and all examinations were
stically evaluable. PE was diagnosed in 6 (3 women, 3
ts of the TB group and 4 (1 women, 3 men) patients of
p.
st enhancement
ere were small differences in contrast enhancement
values between groups, none of these differences
tistical significance in any anatomic region (p>0.5)
T provided a non-significantly higher mean enhance-
e pulmonary trunk, left and right main pulmonary
t atrium, left ventricle, and lung parenchyma, while
p provided slightly higher mean contrast enhancement
superior vena cava, right heart and the right ventricle.
lculated overlay value of lung parenchyma was slightly
g BT (p>0.5). Mean attenuation for different anatomic
iven in Table 2 for both techniques.
ion” defects
n” defects were present in 5 of the 6 patients with PE
oup for a total of 18 “perfusion” defects. In the BT group
defects were present in all of the 4 patients with PE for
PE related “perfusion” defects. The mean attenuation
fusion” defects was −777.9±118.6HU in the TB group
±124.8HU in the BT group. The calculated DE overlay
“perfusion”defectswas4.6±1.5 and6.5±3.4HU for TB
ectively. In comparison, themeancalculatedDEoverlay
lmonary parenchyma outside “perfusion” defects was
nd 34.0±2.0HU for TB and BT, respectively.
quality
ian score for image quality of the “virtual 120kV”
4 (3–4) for the TB group and 4 (3–5) for the BT group
Fig. 6). The maximum/minimum score were 5/2 for TB
and 5/
“perfu
respec
BT, res
4. Dis
Per
taneou
distrib
visuali
namic
increas
In pati
fusion
idiopa
bolic p
pulmo
tion of
and da
require
where
the cap
“perfu
sibility
with P
the exi
facts, p
may be
The
minati
suspec
allows
higher
DECT “
Wit
plotted
value i
gle lev
time in
contra
attenuthis article in press as: Henzler T, et al. Dual-energy CT angiography of
for the determination of scan delay. Eur J Radiol (2010), doi:10.1016/j.ejrad
atient demographics and scan delay for the bolus tracking and the test bolus group.
Bolus tracking T
64.8±17.6 95% CI: 58.3–71.36 6
80.8±17.4, 95% CI: 74.4–87.2 8
168.9±9.5, 95% CI: 165.3–172.4 1
28.25±5.2, 95% CI: 26.3–30.1 2
13.4±2.6, 95% CI: 12.1–14.4 1
the given scan delay is without the additional 5 s delay for breathing instructions.BT. The median score for the image quality of the DECT
mapswas 3 (3–4) and 3 (3–4) for TB andBT (p=0.3913),
. The maximum/minimum score were 5/2 for TB and
vely (Fig. 7).
on
ing chest CTA in DE mode has the advantage of simul-
essment of pulmonary vessels and parenchymal iodine
as a surrogate of lung perfusion. In patients with PE,
n of lung perfusion can help to assess the hemody-
ance of intravascular clots detected with CTA and to
e detection rate of small peripheral embolism (Fig. 8).
with pulmonary hypertension, the visualization of per-
lpful to differentiate different disease entities such as
ulmonary hypertension (iPAH) or chronic thromboem-
nary hypertension (CTEPH). One critical issue of DECT
angiography is that the technique requires optimiza-
ynchronization between the arrival of contrastmaterial
quisition. On the one hand, a high vessel attenuation is
visualize pulmonary emboli in the pulmonary arteries,
the other hand the contrast bolus needs time to perfuse
y bed in order to enable the assessment of pulmonary
[6]. Although initial studies have demonstrated the fea-
ECT for the detection of “perfusion” defects in patients
studies have uniformly reported the susceptibility of
DE postprocessing algorithms to beam hardening arti-
tially causing false positive “perfusion” defects which
interpreted as “perfusion” defects caused by PE [1,2,7].
of this study was to compare TB and BT for the deter-
f scan delay for DE CTA of the lungs in patients with
E. Our initial assumption was that the TB technique
ore accurate timing of the contrast bolus leading to
el attenuation and an improved image quality of the
sion” maps.
e BT technique for CTA of the lungs, a ROI is usually
ide the lumen of the pulmonary trunk. A trigger HU
sen, and before starting the CTA data acquisition, sin-
-dose dynamic scans are performed at predetermined
als during the injection of contrast material. When the
aterial arrives at the level of the ROI, the change in
is detected, and a CT scan is started after reaching thethe lungs: Comparison of test bolus and bolus tracking
.2010.06.023
est bolus p-Value
0.7±15.5, 95% CI: 53.3–70.2 0.443
1.5±17.1, 95% CI: 75.0–87.9 0.838
71.3±9.6, 95% CI: 167.7–174.9 0.197
7.7±5.5, 95% CI: 25.7–29.8 0.699
2.5±3.3, 95% CL: 11.7–13.5 0.644

Please cite
techniques
ARTICLE IN PRESSGModelEURR-4843; No.of Pages7
6 T. Henzler et al. / European Journal of Radiology xxx (2010) xxx–xxx
Fig. 8. Dual-e tual 1
right lower lob
triggering H
effective m
among them
and the aor
nique are tr
the delay b
the start of
tion delay c
are not per
is used to g
is the time
intersection
Using a sho
contrast bo
vessel atten
The major d
that the con
frequently
tion about p
shownan in
enhanceme
the more d
vascular res
For this
scan delay u
angiograph
nique by pl
We injected
rate as the
were perfornergy CTA of a 65-year-old patient with acute dyspnea and elevated D-dimers. “Virthis article in press as: Henzler T, et al. Dual-energy CT angiography of
for the determination of scan delay. Eur J Radiol (2010), doi:10.1016/j.ejrad
e (A, B and C). Color-coded DECT “perfusion” map demonstrates a typical wedge shaped
U value. The BT technique has been shown to be an
eans to optimize contrast timing for many applications,
CTA of the pulmonary arteries, the coronary arteries,
ta [8,9]. Parameters of great importance for the BT tech-
ansition delay and interscan delay. Transition delay is
etween the time at which the threshold is reached and
the actual CT scan. With novel CT scanners the transi-
an be reduced down to 4 s, even if the monitoring scans
formed at the same level of the scan start. This delay
ive the patient breathing instructions. Interscan delay
between the consecutive dynamic scans. The reported
delay in the literature ranges from 0.33 s to 6.0 s [10].
rt transition and interscan delay allows monitoring the
lus very closely, which reduces significant changes in
uationbetween themonitoring scans and themainCTA.
rawback of using BT in CT pulmonary angiography is
trast bolus is monitored within only one vessel, most
the pulmonary trunk. Therefore, BT gives no informa-
atients’ individual pulmonary circulation. Studies have
verse relationship between cardiac output and contrast
nt which could lead to an insufficient opacification of
istal pulmonary arteries in patients with an increased
istance [8,10].
reason, many groups prefer to determine an individual
sing a TB especially in coronary, cervical, and cerebral
y [11–13]. To address this problem,weused theTB tech-
acing ROIs in the pulmonary trunk and the left atrium.
a small amount of contrastmaterial (20ml) at the same
main bolus while single level low-dose dynamic scans
med with an interscan delay of 1 s.
Several g
with mixed
between TB
single-ener
TB and BT
BT yields a
technique [
niques forD
no statistic
contrast en
the subject
dataset and
compared t
parameters
of variance
itively seem
determinat
nation for t
tracking as
of the TB a
moderate c
two other
must consid
capacitance
contrast ma
Indeed, Han
highly sign
shifting tim
the right, w
[18].20kV” dataset images revealed only a small segmental PE within thethe lungs: Comparison of test bolus and bolus tracking
.2010.06.023
“perfusion” defect within the right lower lobe (D, E and F).
roupshave comparedTBandBT invarious applications,
results. Johnson et al. found no significant differences
and BT for ECG-gated CTA of the chest using a standard
gy CT scanner [14], while Cademartiri et al. compared
for CT coronary angiography and demonstrated that
more homogenous enhancement compared to the TB
8]. To our knowledge, no study has compared the tech-
ECTpulmonary angiography. Our results demonstrated
ally significant difference between TB and BT regarding
hancement in any anatomic region. Likewise, although
ively perceived image quality of the “virtual 120kV”
the DE “perfusion” maps were rated superior for BT
o TB, therewere again no significant differences in these
between techniques using the Kruskal–Wallis analysis
(p>0.05). While the test bolus technique may intu-
to represent a more accurate method of scan delay
ion, there is prior work that offers a potential expla-
he lack of improvement using this method. Test bolus
sumes a close correlation between the bolus geometry
nd the main bolus, and while one group did display a
orrelation between TB and main bolus geometry [15],
studies did not find such a correlation [16,17]. One
er that a lower volume lacks driving force in the high-
venous system, which can result in pooling of the
terial and hence an artificially lengthened delay time.
et al. demonstrated that contrast volume does have a
ificant effect on bolus geometry, with higher volumes
e enhancement curves not only upward, but also to
hich would obviously affect the optimal scan delay

Please cite y of
techniques .ejrad
ARTICLE IN PRESSGModelEURR-4843; No.of Pages7
T. Henzler et al. / European Journal of Radiology xxx (2010) xxx–xxx 7
As we used highly concentrated contrast material, one may
expect more artifacts within the subclavian veins, the vena cava
superior and the right ventricle. In order to avoid those artifacts,
we used a saline bolus of 20ml with a flow rate of 4ml/s after
contrast agent administration in both groups. Moreover, all scans
were performed in a caudocranial scan direction to give the con-
trast material additional time to flow out of the subclavian vein.
However, with both techniques the highest attenuation was still
found in the superior vena cava which frequently led to misleading
artificial “perfusion” defects caused by streak artifacts. Future stud-
ies have to evaluate whether a shorter contrast material bolus at a
higher flow rate helps to reduce these artifacts which are mainly
caused by the high contrastmaterial concentration in the great tho-
racic vessels. The total amount of the saline bolus used in this study
was relatively small compared to other groups [2,7,14,19]. A criti-
cal issue of injecting larger amounts of saline with a high flow rate
after the injectionof the contrast bolus is thatpatientswithacutePE
oftensuffer acute rightventriculardysfunction. Therefore thesaline
bolus should not be unnecessarily increased. However, future stud-
ies in patients without PE are necessary to evaluate whether larger
volumes of saline bolus can increase the image quality by pushing
the dense contrast bolus effectively out of the subclavian veins.
A limitation of this study which must be considered is that
only data from a limited number of patients with pulmonary
embolism w
sionswheth
“perfusion”
ber of patie
the diagnos
assessment
Another
empirical m
the TB grou
peak enhan
for DE CTA
5. Conclus
A homog
ries and the
quality can
Conflict of
The auth
References
[1] Fink C, Johnson TR, Michaely HJ, et al. Dual-Energy CT angiography of the lung
in patients with suspected pulmonary embolism: initial results. Rofo 2008.
[2] ThiemeSF, BeckerCR,HackerM,NikolaouK, ReiserMF, JohnsonTR.Dual energy
CT for the assessment of lungperfusion–correlation to scintigraphy. Eur J Radiol
2008;68(3):369–74.
[3] Smulders YM. Pathophysiology and treatment of haemodynamic instability in
acute pulmonary embolism: the pivotal role of pulmonary vasoconstriction.
Cardiovasc Res 2000;48(1):23–33.
[4] Flohr TG, McCollough CH, Bruder H, et al. First performance evaluation of a
dual-source CT (DSCT) system. Eur Radiol 2006;16(2):256–68.
[5] Johnson TR, Krauss B, Sedlmair M, et al. Material differentiation by dual energy
CT: initial experience. Eur Radiol 2007;17(6):1510–7.
[6] Hoey ET, Gopalan D, Screaton NJ. Dual-energy CT pulmonary angiography: a
new horizon in the imaging of acute pulmonary thromboembolism. AJR Am J
Roentgenol 2009;192(6):W341–2 (author reply).
[7] Thieme SF, Johnson TR, Lee C, et al. Dual-energy CT for the assessment of con-
trast material distribution in the pulmonary parenchyma. AJR Am J Roentgenol
2009;193(1):144–9.
[8] Cademartiri F, Nieman K, van der Lugt A, et al. Intravenous contrast material
administration at 16-detector row helical CT coronary angiography: test bolus
versus bolus-tracking technique. Radiology 2004;233(3):817–23.
[9] Kirchner J, Kickuth R, Laufer U, Noack M, Liermann D. Optimized enhance-
ment in helical CT: experiences with a real-time bolus tracking system in 628
patients. Clin Radiol 2000;55(5):368–73.
[10] Cademartiri F, van der Lugt A, Luccichenti G, Pavone P, Krestin GP. Param-
eters affecting bolus geometry in CTA: a review. J Comput Assist Tomogr
2002;26(4):598–607.
[11] Anders K, Baum U, Gauss S, et al. [Initial experience with prospectively trig-
gered, sequential CT coronary angiography on a 128-slice scanner]. Rofo
2009;181(4):332–8.
l M,
gered
9;44(
l M, M
acqui
initia
nson
t mat
phic
–71.
tee R
al arte
n dela
ist To
tt JF, R
: corr
J Roe
Hoe L
ay tim
ist To
JK, K
nt at C
put A
sch M
essel
iograp
nary ethis article in press as: Henzler T, et al. Dual-energy CT angiograph
for the determination of scan delay. Eur J Radiol (2010), doi:10.1016/j
as available. Therefore, the data does not allow conclu-
er TBorBT is superior for thedetectionof PEand related
defects. Because of this, further studies in a larger num-
nts are required to compare both techniques regarding
tic accuracy of DECT pulmonary angiography for the
of PE.
potential limitation of this study is that we used an
odel for the determination of the optimal scan delay in
p. Until now, it is unclear whether the time between the
cement in the pulmonary artery or the aorta is optimal
of the lungs.
ion
eneous opacification of the different vascular territo-
pulmonary parenchyma as well as a sufficient image
be achieved with both, the BT and the TB technique.
interest
ors state no conflict of interest.
[12] Lel
trig
200
[13] Lel
ral
and
[14] Joh
tras
gra
265
[15] Kaa
ren
sca
Ass
[16] Pla
phy
Am
[17] van
del
Ass
[18] Han
me
Com
[19] Jelt
of v
ang
mothe lungs: Comparison of test bolus and bolus tracking
.2010.06.023
Hinkmann F, Anders K, et al. High-pitch electrocardiogram-
computed tomography of the chest: initial results. Invest Radiol
11):728–33.
arwan M, Schepis T, et al. Prospectively ECG-triggered high-pitch spi-
sition for coronary CT angiography using dual source CT: technique
l experience. Eur Radiol 2009;19(11):2576–83.
TR, Nikolaou K, Wintersperger BJ, et al. Optimization of con-
erial administration for electrocardiogram-gated computed tomo-
angiography of the chest. J Comput Assist Tomogr 2007;31(2):
, Van Leeuwen MS, De Lange EE, et al. Spiral CT angiography of the
ries: should a scan delay based on a test bolus injection or a fixed
y be used to obtain maximum enhancement of the vessels? J Comput
mogr 1998;22(4):541–7.
eige KA, Ellis JH. Aortic enhancement during abdominal CT angiogra-
elation with test injections, flow rates, and patient demographics. AJR
ntgenol 1999;172(1):53–6.
, Marchal G, Baert AL, Gryspeerdt S, Mertens L. Determination of scan
e in spiral CT-angiography: utility of a test bolus injection. J Comput
mogr 1995;19(2):216–20.
im AY, Lee KY, et al. Factors influencing vascular and hepatic enhance-
T: experimental study on injection protocol using a canine model. J
ssist Tomogr 2000;24(3):400–6.
, Klein S, Juchems MS, Hoffmann MH, Aschoff AJ. Objective evaluation
attenuation in multidetector-row computed tomographic pulmonary
hy using high-density contrast material for the detection of pul-
mbolism. J Comput Assist Tomogr 2008;32(6):893–8.