Download Prospective and Retrospective ECG Gating for Thoracic CT

6A S C U L A R A N D ) N T E R VE N T I O N A L 2 A D I O L O G Y s / R I G I N A L 2 E S E A R C H
Wu et al.
ECG Gating for Thoracic CT Angiography
FOCUS ON:
Vascular and Interventional Radiology
Original Research
Wenhui Wu1
Joseph Budovec2
W. Dennis Foley 2
Wu W, Budovec J, Foley WD
Keywords: cardiac gating, CT angiography, emergency
radiology, hemodynamics, thoracic CT angiography,
vascular disease
DOI:10.2214/AJR.08.2158
Received November 24, 2008; accepted after revision
February 2, 2009.
W. Wu was at the Medical College of Wisconsin at the
time of the study for a visiting fellowship that was
sponsored in part by GE Healthcare.
W. D. Foley has an Investigator Agreement with GE
Healthcare CT Division.
1
Department of Radiology, Cardiovascular Institute &
Fuwai Hospital, Chinese Academy of Medical Sciences &
Peking Union Medical College, Beijing, China.
2
Department of Radiology, Medical College of Wisconsin,
9200 W Wisconsin Ave., Milwaukee, WI 53226. Address
correspondence to W. D. Foley ([email protected]).
AJR 2009; 193:955–963
0361–803X/09/1934–955
© American Roentgen Ray Society
AJR:193, October 2009
Prospective and Retrospective
ECG Gating for Thoracic CT
Angiography: A Comparative Study
/"*%#4)6% The objective of our study was to compare radiation dose, contrast load,
thoracic aortic attenuation value, and image quality parameters of MDCT thoracic aortography performed with prospective and retrospective cardiac gating.
-!4%2)!,3 !.$ -%4(/$3 Studies were performed on 80 patients (prospective
ECG gating, n = 40; retrospective ECG gating, n = 40) either being evaluated for thoracic
aortic aneurysm (n = 23) or aortic dissection (n = 36) or undergoing postsurgical or postintervention follow-up (n = 21). Image acquisition parameters and radiation dose (CT dose index
volume [CTDIvol] and dose–length product [DLP]) were obtained from image archival data.
Contrast load and aortic attenuation values were obtained from a data registry. The comparative degrees of motion artifact and banding artifact were assessed on parasagittal maximumintensity-projection (MIP) images and reformatted images in the plane of the aortic valve.
2%35,43 CTDIvol and DLP in the prospective ECG-gating group was 28.8 ± 2.12 mGy
(mean ± SD) and 833.7 ± 115.77 mGy/cm, respectively, which are significantly lower (p <
0.001) than those values in the retrospective ECG-gating group (74.7 ± 13.42 mGy and
2,547.3 ± 553.27 mGy/cm). The average contrast load in the prospective gating group was
109.1 ± 14.74 mL and in the retrospective gating group, 101.3 ± 10.45 mL (p < 0.05). The average aortic attenuation values (in Hounsfield units) for the prospective and retrospective
ECG-gated groups were 447.6 and 350.2 HU, respectively, for the mid ascending aorta, 413.6
and 325.7 HU for the mid aortic arch, 418.2 and 327.6 HU for the mid descending aorta, and
355.0 and 306.2 HU for the supraceliac aorta. Subjective scores of motion artifact and banding artifact were equivalent between the two groups.
#/.#,53)/. Compared with retrospective ECG-gated thoracic CT angiography, prospective ECG-gated thoracic CT angiography was associated with a lower radiation dose, slightly increased contrast load, increased aortic attenuation values, and equivalent image quality.
M
DCT angiography is a technically robust, noninvasive imaging
technique for the evaluation of
vascular disease, including diseases affecting the thoracic aorta. Advances
in image acquisition technology have resulted in improvements in image acquisition
speed and spatial resolution. Consequently,
patient studies can now be performed rapidly
for both emergent and elective indications.
MDCT of the thoracic aorta is used to diagnose various acute and chronic conditions including aortic aneurysm, aortic dissection,
penetrating atherosclerotic ulcer, intramural
hematoma, traumatic aortic injury, aortic
rupture, inflammatory disorders, and congenital malformations [1, 2]. By creating a
3D image, CT angiography may be superior
to conventional angiography [3, 4].
MDCT using a 64-channel system is used
not only for the diagnosis of aortic disease
but also for surgical planning and postintervention evaluation [5]. In addition, CT angiography is less invasive, consumes fewer
resources, and is more cost efficient than conventional angiography, especially after the
introduction of 64-MDCT scanner [6, 7].
In patients undergoing CT evaluation for
suspected aortic dissection or ascending aortic
aneurysm, vessel motion can result in blurred
aortic outlines, particularly of the aortic root
and proximal ascending aorta. In addition, the
aortic valve and coronary arteries frequently
are not clearly displayed even when images
are reoriented in the longitudinal and shortaxis plane of these structures.
ECG-gated data acquisition addresses these
issues. Roos et al. [8] reported that image
955
7U ET AL
quality of the aorta was superior using retrospective ECG-gated acquisition compared with
non-ECG-gated acquisition. The benefit of
ECG-gated synchronization was most pronounced in the proximal ascending aorta and
was less so at the aortic arch and proximal descending aorta [1]. ECG gating improves the
diagnosis of type A aortic dissections, particularly the proximal extent of the mural flap in
relationship to the aortic root, coronary ostia,
and aortic valves. In addition, it improves the
delineation of the proximal margin of ascending aortic aneurysms.
Thoracic aortic CT angiography with cardiac gating may now be considered the preferred technique for thoracic CT angiography when diagnostic concern relates to the
aortic sinus, ascending aorta, or coronary arteries. An alternative technique is gated thoracic aortic MR angiography using steadystate free precession acquisition. Advances
in parallel imaging techniques permit imaging of the thoracic aorta and left ventricular
outflow tract with a 5- to 6-second breathhold [9–11]. A more comprehensive evaluation of the thoracic aorta and aortic valve
requires multiplanar, multisequence imaging with ECG and respiratory gating, including sagittal oblique images through the aorta and aortic root, that prolong acquisition
time. Thoracic aortic MR angiography is not
used in patients presenting with acute chest
pain and is limited in patients with implanted cardiac devices (e.g., automatic implantable cardioverter-defibrillators, pacemakers)
and prosthetic valves. Although MR angiography can define the anatomy of the aortic
valve and aortic root, the technique is not
well suited to coronary arteriography.
Until recently, the initial implementation
of cardiac gating for CT angiography was
with a retrospective ECG-gating technique.
Compared with nongated techniques, retrospective gating has an acquisition duration
and radiation dose that are relatively high
because of the low pitch value. Prospective
ECG gating is a newer technique in which
the tube current is triggered for only a short
segment of the R-R interval at a preset time
after the R wave. Consequently, radiation is
emitted every alternate heartbeat and the table moves without radiation being emitted
during the intervening heartbeat [12].
The purpose of this study was to compare prospective and retrospective ECG-gated MDCT data acquisition techniques for CT
angiography of the thoracic aorta with regard
to radiation dose, contrast load, thoracic aortic attenuation, and image quality parameters.
956
Materials and Methods
Patient Population
The institutional review board approved this
study with a waiver of authorization for patient informed consent. Retrospective review of CT examinations performed between March 2007 and
August 2007 yielded 40 patients who had undergone MDCT thoracic CT angiography using prospective ECG gating. A comparison group of 40
patients was selected from patients who had undergone MDCT thoracic CT angiography using
retrospective ECG gating performed from June
2006 to March 2007.
The 80 patients in the study group had been referred for evaluation of suspected thoracic aortic
disease, either dissection (n = 36) or aneurysm (n =
23), or for postsurgical or postintervention followup (n = 21). Nine patients in the prospective ECGgated group and 11 patients in the retrospective
ECG-gated group were referred from the emergency department.
Imaging Techniques
The CT examinations were performed using a
64-MDCT system (LightSpeed VCT, GE Healthcare). All examinations were performed during
patient inspiration.
Circulation timing was determined via a preliminary mini bolus in which 15 mL of contrast material, either iodixanol (Visipaque 320, 320 mg I/mL,
GE Healthcare) or iopamidol (Isovue 370, 370 mg
I/mL, Bracco Diagnostics), was administered IV at
6 mL/s using an 18-gauge catheter placed in an antecubital vein. Visipaque was used for patients with
marginal renal function (i.e., estimated glomerular filtration rate [GFR] < 60 mL/min/1.73 m2). A
time–attenuation curve was derived from the comparative attenuation values of 10 sequential images obtained from the mid ascending aorta over a
20-second interval beginning 5 seconds after the
beginning of contrast injection.
The contrast arrival time in the ascending aorta
was calculated as the time from the beginning of
the injection until peak attenuation on the time–
enhancement curve after the mini bolus injection.
The delay time between contrast arrival and the
beginning of acquisition was set at 4 seconds for
both protocol groups.
The contrast injection interval was the combination of the time between contrast arrival in the
ascending aorta and the beginning of the acquisition plus the acquisition time (Fig. 1).
The imaging volume extended from the proximal brachiocephalic arteries in a cephalad to caudad direction to the level of the origin of the celiac
artery in the upper abdominal aorta.
Retrospective ECG-gated acquisition used a
detector configuration of 64 × 0.625 mm, scan rotation speed of 350 milliseconds, and pitch adjust-
ed from 0.18 to 0.24 on the basis of the patient’s
heart rate. The average heart rate during acquisition was 63 beats per minute (bpm). Table speed
averaged 2.5 cm/s. ECG modulation of x-ray tube
current was used, decreasing tube current during
systole and early diastole and maintaining full selected tube current in mid to late diastole.
Prospective ECG-gated acquisition was performed in a sequential mode with a beam collimation of 40 mm (detector configuration, 64 × 0.625
mm). Prospective gating used axial plane acquisition in an incremental dynamic mode; 75% of
the R-R interval was chosen as the midpoint for
the data acquisition time of 175 milliseconds, with
an additional padding factor of 125 milliseconds.
This padding factor added 125 milliseconds both
before and after the initial designated acquisition
time. The data acquisition occurs during every alternate heartbeat with incremental table motion of
3.5 cm in the intervening heartbeat. For a heart
rate of 60 bpm, beam collimation of 4 cm, and acquisition of every alternate heartbeat, the effective table speed approximates 2 cm/s. The average heart rate in the prospective ECG-gated group
was 62 bpm.
Beta-blocker medication was administered to
patients with sinus rhythm and heart rates between
70 and 100 bpm. Patients with arrhythmia were
excluded. For elective studies, patients received
either 100 mg of metoprolol orally or, if on current E-blocker medication, 50 mg of metoprolol
orally unless contraindicated. After 45 minutes
of observation, if the heart rate remained above
70 bpm, an additional 50 mg of metoprolol was
administered orally. Additional E-blocker medication (5 mg of IV metoprolol) was administered
if necessary in the CT scanner suite to a maximum dose of 15 mg. For patients presenting to the
emergency department with acute symptoms, IV
E-blocker medication (5 mg of metoprolol repeated × 2 as required) was administered. Contraindications to metoprolol therapy included severe
aortic stenosis, systolic blood pressure less than
100 mm Hg, and a history of asthma requiring
daily nebulizer therapy.
The image acquisition parameters for prospective ECG-gated acquisition and retrospective
ECG-gated acquisition are detailed in Table 1.
Radiation Dose Estimation
CT dose index volume (CTDIvol) and dose–
length product (DLP) were recorded as direct data
output from prospective ECG-gated examinations.
Only the CTDIvol and DLP values obtained from
the contrast-enhanced CT angiographic studies
were used in study analysis. Radiation from the
preliminary localization CT radiographs and from
single-level dynamic scanning to determine circulation time was not included. DLP in milligrays
AJR:193, October 2009
ECG Gating for Thoracic CT Angiography
Contrast
Contrast
Saline
Saline
10 seconds
10
20
30
10
20
30
6 mL/s, 14 seconds
6 mL/s, 16 seconds
First circulation
First circulation
6 mL/s, 8 seconds
6 mL/s, 8 seconds
64 × 0.625 mm, 0.20, 0.35 second, 2.5 cm/s
64 × 0.625 mm, 1.0, 0.35 second, 2 cm/s
A
"
Fig. 1—Diagrams show injection/acquisition technique retrospective and prospective gating for thoracic CT angiography. Cephalocaudad coverage is 25 cm.
A, Schematic diagram illustrates injection acquisition technique for thoracic CT angiography using retrospective gating. Time in seconds is indicated on x-axis. Arrow
indicates calculated arrival time of injected contrast in ascending aorta using aortic peak on time–attenuation curve after mini bolus injection (6 mL/s for 15 mL). There is
4-second delay between calculated arrival time and beginning of acquisition. Acquisition interval is calculated at 10 seconds for 25-cm cephalocaudad acquisition using
image acquisition parameters detailed on bottom line: detector configuration, 64 × 0.625 mm; average pitch value, 0.20; scan rotation speed, 0.35 second; table speed, 2.5
cm/s. Contrast injection rate is set at 6 mL/s with injection duration equal to time delay between aortic arrival and beginning of acquisition (4 seconds) plus acquisition
interval (10 seconds). Contrast injection is followed by saline flush at same flow rate over 8-second interval. Aim of this injection and acquisition protocol is to capture
image data during first circulation of injected bolus.
B, Schematic diagram illustrates injection and acquisition parameters for prospective cardiac gating. Time in seconds is indicated on x-axis. Arrow represents aortic
arrival of contrast material after preliminary mini bolus injection (6 mL/s for 16 seconds) and is calculated as time to peak on time–attenuation curve in ascending aorta.
There is 4-second delay time between aortic arrival and beginning of acquisition. Acquisition uses step-and-shoot, or incremental dynamic, mode in which 4 cm of axial
scan data is acquired in 1 rotation at fixed preset time after R wave of ECG; incremental table motion occurs during next heartbeat, and in third heartbeat, another 4
cm of axial data is again acquired at same preset time after R wave of ECG. There is 0.5-cm overlap of image anatomy between each incremental acquisition. Process
is replicated until required cephalocaudad dimension is imaged. Acquisition parameters are detector configuration of 64 × 0.625 mm and scan rotation speed of 0.35
second, 2 cm/s. At heart rate of 60 beats per minute, table speed approximates 2 cm/s. Contrast medium is injected at 6 mL/s over time interval equal to delay between
aortic arrival and beginning of acquisition plus acquisition time. After contrast injection, saline is injected at same rate to enable effective utilization of injected bolus and
clear superior vena cava. As with retrospective gating, injection and acquisition technique are aimed to provide first circulation acquisition of imaged anatomy.
per centimeter (mGy/cm) was converted to an estimated effective dose in millisieverts (mSv) using
a standard conversion equation that includes the
patient’s age and sex [13].
Patient radiation dose for retrospective ECGgated MDCT thoracic angiography was estimated using the lowest and highest tube current (mA)
values and on times in the R-R interval during acquisition to define the estimated CTDIvol. DLP in
mGy/cm was then determined by the total length
of the acquisition. As with prospective ECG-gated acquisition, the DLP in mGy/cm was converted
into an estimated dose in mSv using a standard
conversion equation.
Contrast Material Load
The contrast material for the examinations was
either Isovue 370 or Visipaque 320. A data registry recorded the type and amount of contrast material for each separate patient study. Visipaque
was used in patients with marginal renal function,
which was defined as an estimated GFR of less
than 60 mL/min/1.73 m 2. Before and after the procedure, oral hydration (two glasses of 16 oz [471
mL] of tap water) was given for renal protection in
all patients who received Visipaque.
AJR:193, October 2009
Thoracic Aortic Attenuation
The mean CT attenuation values (in Hounsfield
units [HU]) of the mid ascending aorta, mid aortic
arch, mid descending aorta, and supraceliac abdominal aorta were measured using a circular region of
interest approximating 50% of the aortic cross-sectional area. The mean and SD of attenuation at all
four levels in both patient groups were recorded.
Image Quality Parameters
Image quality parameters were assessed subjectively; motion artifact and resultant blurring, as
well as misregistration or stairstep artifact, affecting both the thoracic aorta and aortic valve were
evaluated.
Motion artifact—Motion artifact causing image blurring of the thoracic aorta and brachiocephalic arteries was assessed from parasagittal
subvolume maximum-intensity-projection (MIP)
reformatted images using the following scale
adapted from the work of Roos et al. [8]: 1, no motion artifact, optimum for diagnosis; 2, minimum
aortic wall blurring, satisfactory for diagnosis; 3,
moderate motion artifact, adequate for diagnosis;
4!",% )MAGING 0ARAMETERS FOR 0ROSPECTIVE AND 2ETROSPECTIVE %#''ATED
Techniques
CT Mode
Peak Tube
Potential
(kVp)
Tube Current
(mA)
Sequential prospective
ECG-gated
120–140
a
Helical retrospective
ECG-gated
120–140
ECG dose
modulationa,b
Pitch
0.18–0.24
Rotation
Time (s)
Detector
Configuration
(mm)
Beam
Width
(mm)
0.35
0.625 × 64
40
0.35
0.625 × 64
40
aTube current selection was based on patient weight: 350–450 mA for small patients (< 120 lb [54 kg]), 450–550
mA for medium patients (120–220 lb [54–99 kg]), and 600–750 mA for large patients (> 220 lb [99 kg]).
bIf heart rate was less than 65 bpm, peak mA was 70–80% R-R interval; if heart rate was greater than 65 bpm,
peak mA was 40–80% R-R interval.
957
7U ET AL
A
"
C
D
Fig. 2—Prospective (A, B) and retrospective (C, D) ECG-gated acquisitions in patients with suspected aortic
dissection. Optimal image quality without motion blurring or misregistration is shown on both types of
examinations.
A, Prospective ECG-gated thoracic CT aortography in 75-year-old woman with suspected aortic dissection
using parasagittal subvolume maximum intensity projection (MIP). Left anterior oblique projection shows there
is no motion or banding artifact. Study is considered optimum for diagnosis. Differential opacity of aortic root
as compared with mid and distal ascending aorta represents time interval of two heartbeats approximating 2
seconds with differential aortic opacity between two acquisitions. Resultant linear band across aortic lumen
does not represent stairstep or banding artifact. This patient had prior sternotomy and coronary artery bypass
grafting (arrow) to distal right coronary artery.
B, Prospective ECG-gated thoracic aortic CT angiography of same patient shown in A. Reformatted image
through plane of aortic valve shows sharp delineation of trileaflet valve and aortic wall with normal leaflet
coaptation. Left atrium is on posterior aspect of aortic root. Small segment of proximal left main coronary
artery (arrow) is opacified. There is mild uniform opacity of right atrium and right ventricle after bolus passage
through right heart.
C, Retrospective ECG-gated thoracic CT aortogram study in 61-year-old man with suspected aortic dissection
using parasagittal subvolume MIP. Left anterior oblique projection shows neither motion artifact nor banding
artifact. Study is considered optimum for diagnosis.
D, Retrospective ECG-gated thoracic aortic CT angiography in 59-year-old man undergoing surveillance study
after aortic graft repair of type A dissection. Reformatted image of aortic root through plane of valve leaflets
shows sharp definition of normal trileaflet valve and aortic wall and normal coaptation of leaflets. Opacified left
atrium marginates right coronary and noncoronary aortic sinuses. Small segment of proximal right coronary
artery (arrow) is imaged. Mild residual uniform enhancement of right heart chambers and pulmonary outflow
tract is seen after bolus passage through right heart.
4, distinct motion artifact, questionable for diagnosis; and 5, severe motion artifact, diagnostically
uninterpretable.
Aortic valve motion artifact was determined
from in-plane batch reformatted images using the
following 5-point scale: 1, no motion artifact, aortic valve leaflets, orifice, and annulus clearly vis-
958
ible; 2, minimum motion artifact, satisfactory visualization of the annulus, leaflets, and orifice; 3,
moderate motion artifact, adequate visualization of
the annulus, orifice, and leaflets; 4, distinct motion
artifact, inadequate visualization of the annulus,
orifice, and leaflets; 5, severe motion artifact, no
anatomic detail of the aortic wall or aortic valve.
Banding artifact—Banding artifact occurs because of anatomic misregistration of contiguous
cardiac structures obtained during sequential heartbeats. This artifact is identified as one or more
parallel or straight lines in which the cardiac or
aortic contour showed abrupt steplike distortions.
In the thoracic aorta, banding artifact was assessed as follows: 1, no banding artifact, distinct
aortic wall delineation and continuity; 2, minor
banding artifact, slight discontinuity and misregistration on one margin of the aorta at one level;
3, moderate banding artifact, distinct misregistration and discontinuity of both margins of the aorta
at one level; 4, severe banding artifact, stairstep
artifact at multiple levels.
In the aortic valve, banding artifact was assessed as follows: 1, no banding artifact, aortic
annulus and leaflets intact and continuous; 2, minor banding artifact, slight misregistration and
discontinuity of aortic annulus but the three aortic
leaflets are clearly defined; 3, moderate banding
artifact, distinct misregistration and discontinuity
of the aortic annulus and three aortic valve leaflets
are in the same cardiac phase; 4, severe banding
artifact with prominent banding artifact across the
aortic annulus and different segments of the aortic
valve visualized in different cardiac phases.
The differences in average heart rate and heart
rate variability of the patients with images showing no banding artifact and those with images
showing banding artifact were recorded.
Statistical Analysis
Statistical analysis was performed on commercially available software (SPSS, version 14.5,
SPSS) for Microsoft Windows. Continuous variables were expressed as means and SDs. Comparative analysis for the numeric variables of two
groups used the unpaired Student’s t test or the
Mann-Whitney U test. For nonparametric data, the
Kruskal-Wallis H test was used to compare scores
of image quality of the thoracic aorta and aortic
valve between the two groups. A p value of less
than 0.05 was considered statistically significant.
Results
The prospective ECG-gating and retrospective ECG-gating techniques were successful in all patients without any complications (Fig. 2). The 40 patients in the
prospective ECG-gated group (24 men, 16
women) ranged in age from 34 to 88 years
(mean ± SD, 60.3 ± 13.78 years). All the patients had normal sinus rhythm at the time of
examination. The retrospective ECG-gated
group was composed of 40 patients (22 men,
18 women) who ranged in age from 25 to 93
years (mean, 55.3 ± 14.42 years). The mean
AJR:193, October 2009
ECG Gating for Thoracic CT Angiography
4!",% 2ADIATION $OSE OF 4WO 'ROUPS
Value (mean ± SD)
Radiation Dose
CTDIvol (mGy)
Prospective Group
Retrospective Group
p
28.8 ± 2.12
74.7 ± 13.42
< 0.001
DLP
mGy/cm
mSv
< 0.001
833.7 ± 115.77
2,547.3 ± 553.27
14.2 ± 1.96
43.3 ± 9.4
Note—CTDIvol = CT dose index volume, DLP = dose–length product.
weight of the patients was 194 ± 50.64 lb (87 ±
22.79 kg) in the prospective ECG-gated group
and 197 ± 60.34 lb (89 ± 27.15 kg in the retrospective ECG-gated group. No statistically significant difference between the two groups in
either age or body weight was detected.
The mean delay time between the beginning of the contrast injection and scanning
acquisition was 24 ± 3.2 seconds in the prospective ECG-gated group and 23 ± 3.4 seconds in the retrospective ECG-gated group.
The duration of acquisition in the prospective
ECG-gated group (mean ± SD) was 15.4 ± 2.2
seconds and in the retrospective ECG-gated
group, 13.4 ± 2.2 seconds.
In the prospective ECG-gated group, 12
patients had a thoracic aortic aneurysm, eight
of which involved the ascending aorta, and
11 patients had aortic dissection, only one of
which involved the ascending aorta. One patient with a type B dissection had a postoperative pseudoaneurysm of the aortic root.
In the retrospective ECG-gated group, 10
patients had an aortic aneurysm, five of which
involved the ascending aorta and an additional two postoperative pseudoaneurysms of the
aortic root. Of the seven aortic dissections,
one involved the ascending aorta.
Radiation Dose
The average CTDIvol was 28.8 mGy in the
prospective ECG-gated group and 74.7 mGy
in the retrospective ECG-gated group. The
DLP was 833.7 mGy/cm in the prospective
ECG-gated group and 2,547.3 mGy/cm in
the retrospective ECG-gated group. The radiation dose in the prospective ECG-gated
group was significantly reduced compared
with that of the retrospective ECG-gated
group (p < 0.001) (Table 2).
Contrast Volume
In the prospective ECG-gated group, 34 patients received Visipaque 320 and six patients
received Isovue 370. In the retrospective
ECG-gated group, 28 patients received Visipaque and 12 patients received Isovue 370.
AJR:193, October 2009
The average total contrast load in the prospective ECG-gated group was 109.1 ± 14.74 mL,
which was significantly greater (p < 0.05)
than that in the retrospective ECG-gated
group (101.3 ± 10.45 mL).
Contrast Enhancement
Table 3 lists the attenuation values in the
mid ascending aorta, mid aortic arch, mid
descending aorta, and supraceliac abdominal aorta for both the prospective and retrospective ECG-gated groups. The mean attenuation value was significantly greater in
the prospective ECG-gated group in the mid
ascending aorta, mid aortic arch, and mid
descending aorta (p < 0.001). There was no
significant difference in the mean attenuation value at the level of the supraceliac abdominal aorta between the two groups (p =
0.122).
Image Quality
There was no difference in the extent of
motion artifact in the thoracic aorta or at the
aortic valve annulus between the prospective
and retrospective ECG-gated groups (Fig. 3).
Motion artifact scores in the thoracic aorta
did not exceed 2 in either group.
In the prospective ECG-gated group, one
patient had an aortic valve motion artifact
score of 2. All the other patients has scores
of 1.
Regarding banding artifacts, there was no
statistically significant difference in the overall score between the two groups (p = 0.46).
Banding artifacts were observed in studies
of 10 patients in the prospective ECG-gated group and in studies of 13 patients in the
retrospective ECG-gated group (Fig. 4). No
significant difference was recorded between
the prospective and retrospective ECG-gated
groups in the banding artifact scores of the
aortic valve (1.38 ± 0.774 vs 1.43 ± 0.78, respectively; p = 0.869) or of the thoracic aorta
(1.38 ± 0.93 vs 1.45 ± 0.93; p = 0.712).
The average heart rate of patients in the
prospective ECG-gated group was 62.3 bpm.
In patients whose images showed banding
artifact, the average heart rate was 68.9 bpm,
which was significantly higher than the average heart rate in those whose images were
free of banding artifact (60.2 bpm; p < 0.05).
In the retrospective ECG-gated group, the
average heart rate was 63.4 bpm. In those
with banding artifact, the average heart rate
was 69.62 bpm, whereas it was 60.4 bpm in
patients without banding artifact (Table 4).
The heart rate variability of patients with
banding artifact in the prospective ECG-gated group (18.0 bpm) was significantly greater than those without banding artifact (6.3
bpm; p < 0.05). In the retrospective ECGgated group, the average heart rate variability was higher in patients whose images
showed banding artifact (11.18 bpm) than in
those whose images did not show banding
artifact (6.1 bpm). This difference was not
statistically significant.
Discussion
Nongated thoracic aortic CT angiography is typically performed with a pitch value
of 1.0–1.5. In a system with a 40-mm beam
collimation, total acquisition can be performed within 3 seconds. However, motion
of the ascending aorta can result in blurring
of the vessel outlines and apparent thickening of the aortic wall, which could simulate
a thrombosed dissection or obscure a mural
flap (Fig. 5). In addition, the aortic valve and
coronary arteries are not clearly displayed.
Thoracic aortic CT angiography with cardiac
gating can allow diagnostic evaluation of the
4!",% %NHANCED !ORTIC !TTENUATION 6ALUE AT &OUR -EASURED 3ITES
Attenuation Value (HU) (mean ± SD)
Prospective Group
Retrospective Group
p
Mid ascending aorta
Site
447.6 ± 93.02
350.2 ± 71.09
< 0.001
Mid aortic arch
413.6 ± 92.68
325.7 ± 67.13
< 0.001
Mid descending aorta
418.2 ± 94.98
327.6 ± 65.25
< 0.001
Supraceliac abdominal aorta
355.0 ± 129.88
306.2 ± 83.08
0.122
57.3 ± 34.43
31.5 ± 21.12
< 0.001
SD
959
7U ET AL
A
"
C
D
Fig. 3—Prospective (A, B) and retrospective (C, D) ECG-gated acquisitions in patients with suspected aortic
dissection demonstrate minimal motion blurring in both examinations.
A, Prospective ECG-gated thoracic CT angiography with subvolume maximum intensity projection (MIP) in
65-year-old man with suspected aortic dissection in left anterior oblique projection. Minimum aortic wall
blurring is noted, but image is satisfactory for diagnosis.
B, Prospective ECG-gated thoracic CT aortography of same patient as shown in A. Reformatted image in
plane of aortic valve shows minimal motion blurring of valve leaflets and aortic wall with normal coaptation of
leaflets. Note small segment of proximal left coronary artery in left atrioventricular groove (arrow). Opacified
left atrium is posterior to aorta, and there is residual contrast medium in right atrium, right ventricle, and
pulmonary outflow tract.
C, Retrospective ECG-gated thoracic aortic CT angiography in 61-year-old woman with suspected aortic
dissection. Subvolume MIP in left anterior oblique projection shows minimal motion blurring of ascending
aorta. Image is satisfactory for diagnosis.
D, Retrospective ECG-gated CT thoracic aortography image shows minimal motion blurring in same patient
as shown in C. Reformatted image in plane of aortic valve shows minimal motion blurring of valve leaflets and
aortic wall. Opacified left atrium is left and posterior to aortic valve. Short segment of proximal left coronary
artery (arrow) is imaged. Note mild uniform attenuation of right heart chambers after passage of contrast bolus.
entire thoracic aorta, including the aortic sinus and aortic valve, and of the coronary arteries. Consequently, thoracic aortic CT angiography shows promise as a method for the
comprehensive evaluation of the entire chest
within a single breathhold including evaluation of the thoracic aorta, coronary arteries,
and pulmonary arteries [14]. This practice
mode includes both elective patient studies and evaluation of patients for acute chest
pain in the emergency department in whom
960
E-blocker medication is administered IV by
emergency department physicians.
Radiation Dose
The radiation dose from CT examinations has recently come to national attention.
In 1989, CT accounted for 4% of diagnostic radiologic examinations performed in the
United Kingdom, but CT contributed 40%
of the collective population radiation dose
from medical radiation [15]. In 1999, inves-
tigators from one North American institution reported that 11.1% of the department
workload was CT examinations but that CT
contributed 67% of the collective patient radiation dose [16]. More recent reports indicate that CT accounts for 17% of departmental workload but accounts for 70–75% of the
collective dose from medical radiation [17,
18]. These figures mirror the introduction
of the first single-detector CT scanner in the
late 1980s and MDCT in the late 1990s [19].
Radiation dose is becoming an issue of increasing importance for contrast-enhanced
cardiac MDCT (coronary CT angiography)
because 64-MDCT has shown promising results for coronary artery evaluation [20, 21].
Both the CTDIvol and the DLP were significantly reduced in the prospective ECGgated group compared with the retrospective
ECG-gated group.
In this study, ECG-dependent tube current
modulation was used in the retrospective
ECG-gated group. ECG tube current modulation is currently the most effective technical tool for radiation dose reduction during
retrospective ECG-gated acquisitions without loss of image quality. Tube current is reduced during phases that are not usually necessary for interpretation of cardiac structure
(i.e., phases associated with heart motion).
During mid to late diastole, when motionfree images of the coronary artery and aortic valve may be acquired, the tube current
is increased to the full prescribed radiation
dose. ECG-dependent dose modulation may
reduce radiation dose in retrospective ECGgated CT angiography by up to 50% [22].
Even with optimum ECG-dependent tube
current modulation, the radiation dose from
retrospective ECG-gated acquisition is still
higher than that of prospective ECG-gated
acquisition. In prospective ECG-gated acquisition, a single image is acquired sequentially during the same phase of the cardiac cycle
at contiguous anatomic levels. A single exposure time is used for one image. In our study,
the data acquisition time was lengthened beyond the required 175 milliseconds to allow
the possible use of different midpoints of the
R-R interval for data acquisition in the alternate heartbeats. Apart from phases selected
for acquisition, the patient does not receive
unnecessary radiation exposure during other
phases of the cardiac cycle.
Relatively slow heart rates (< 65 bpm) are
optimal for both retrospective and prospective
cardiac gating. A relatively slow heart rate
with retrospective cardiac gating allows the
AJR:193, October 2009
ECG Gating for Thoracic CT Angiography
ble translation so that contiguous anatomy
can be imaged in every alternate heartbeat. At
faster heart rates, imaging may occur every
third rather than every second heartbeat, thus
prolonging the overall image acquisition.
A
"
C
D
Fig. 4—Prospective (A, B) and retrospective (C, D) ECG-gated acquisitions in patients with suspected thoracic
aortic aneurysms. Banding artifacts are shown on both types of examinations.
A, Prospective ECG-gated thoracic CT angiography in 69-year-old woman with suspected thoracic aortic
aneurysm. Subvolume maximum-intensity-projection (MIP) image in left anterior oblique projection shows
misregistration stairstep artifact at junction of ascending aorta and aortic arch involving both anterior
and posterior wall (left arrow). Identical stairstep artifact can be seen at same anatomic level in proximal
descending thoracic aorta (right arrow). Additional minor stairstep artifacts are visible in mid and distal
descending aorta.
B, Prospective ECG-gated CT thoracic aortography with banding artifact in same patient as shown in A.
Reformatted image in plane of aortic valve shows stairstep discontinuity of aortic wall predominantly anterior
(single arrow). Diastolic phase is recorded in left half of aortic sinus with coapted right and left aortic valve
leaflets, and systolic phase is recorded in right side of aortic sinus with indistinct and relatively blurred aortic
valve leaflets. Left anterior descending coronary artery and proximal circumflex coronary artery (double
arrows) are imaged in proximal atrioventricular groove. Left atrium is posterior to aortic root. Mild uniform
attenuation of right heart chamber secondary to passage of contrast bolus is seen.
C, Retrospective ECG-gated thoracic aortic CT angiography in 41-year-old woman with bicuspid aortic valve
and aortic stenosis. Minimal banding artifact of aortic root and proximal ascending aorta is seen. Slight
stairstep misregistration artifact (arrow) of anterior aortic wall through sinotubular junction and proximal
ascending aorta are evident.
D, Retrospective ECG-gated thoracic CT aortography with banding artifact of same patient as shown in C.
Image shows partially calcified bicuspid aortic valve. Stairstep artifact (arrow) of right lateral margin of aortic
root and mild blurring of anterior aspect can be seen. Remainder of aortic root and visualized valve leaflets are
sharply delineated. Left atrium is posterior to aortic root. Mild uniform opacity of right heart chambers is seen
after passage of contrast bolus.
ECG-dependent tube modulation technique
to be applied for the highest mA only in a relatively restricted mid to late diastolic phase of
acquisition. With faster heart rates and a single-source system, a higher mA is usually applied to a longer segment of the cardiac cycle
AJR:193, October 2009
because the optimum phase for image acquisition may be late systolic or mid to late diastolic in these circumstances; as a result, the
patient receives a higher radiation dose.
With prospective cardiac gating, heart
rates of less than 65–70 bpm allow rapid ta-
Contrast Load
Adequate and uniform aortic enhancement
throughout the entire period of image acquisition is highly desirable for the purpose of image processing and display, in which 3D postprocessing is frequently based on a threshold
CT attenuation value. In our study, contrast
material was injected at a fixed rate (6 mL/s)
in both protocol groups, and the overall contrast volume was dependent on the combination of delay time from aortic arrival of contrast material until image acquisition plus the
image acquisition interval. Aortic arrival of
contrast medium can be determined by either
a preliminary mini bolus or bolus-tracking
software. In our practice, we use a preliminary mini bolus because it should be more
precise. The image acquisition interval was
longer in the prospective ECG-gated group
than the retrospective ECG-gated group because the table speed in the prospective ECGgated group approximated 2 cm/s, whereas
the table speed was 2.5 cm/s in the retrospective ECG-gated group. Faster table speed results in a shorter acquisition time. Aortic arrival of contrast material may be determined
by a preliminary mini bolus technique or bolus-tracking software. The preliminary mini
bolus technique was used in this study because it should be more precise. The increase
in total contrast load in the prospective ECGgated group (109.1 ± 14.74 mL) versus the
retrospective ECG-gated group (101.3 ±
10.45 mL) reflects the difference in the acquisition interval.
Thoracic Aortic Attenuation
The mean attenuation values in the mid
ascending aorta, mid aortic arch, and mid
descending aorta were higher in the prospective ECG-gated group than in the retrospective ECG-gated group. This increase in
aortic attenuation, with the longer duration
of acquisition and the longer injection interval, with prospective gating is supported
by the work of Bae [23]. Bae reported that
peak aortic enhancement was increased with
a more prolonged contrast injection duration
and that the time to peak aortic enhancement
corresponded to the weighted sum of the injection duration and the time to peak test bolus enhancement. Of note in our study is that
961
7U ET AL
4!",% (EART 2ATE AND (EART 2ATE 6ARIABILITY "ETWEEN THE 4WO 'ROUPS
Value (mean ± SD)
Heart Rate
Banding Artifact
No Banding Artifact
p
68.9 ± 10.58
60.2 ± 9.86
< 0.05
69.62 ± 13.49
60.4 ± 10.06
< 0.05
Prospective group
18.0 ± 19.4
6.3 ± 5.59
< 0.05
Retrospective group
11.18 ± 11.85
6.1 ± 6.35
NS
Rate (bpm)
Prospective group
Retrospective group
Variability (bpm)
Note—NS = not significant.
A
"
Fig. 5—Non-ECG-gated acquisition in 34-year-old woman with Ehlers-Danlos syndrome and left hemothorax.
Neither thoracic aortic dissection nor aneurysm is present.
A, Axial image of nongated acquisition for CT thoracic aortography shows pseudodissection (arrows) of
ascending thoracic aorta with apparent curvilinear thickening of aorta wall on anterior and right posterior wall
secondary to motion artifact.
B, Nongated thoracic aortic CT angiography with subvolume maximum intensity projection in left anterior
oblique projection shows motion blurring predominantly of anterior wall of ascending aorta and pseudointimal
flap (arrow) at junction of aortic sinus and ascending aorta.
no significant difference in aortic enhancement between the two groups was found at
the level of the supraceliac aorta. This may
be due to the longer scan duration of prospective ECG-gated acquisition, with the
resultant data acquisition in the supraceliac aorta occurring after a relative plateau of
thoracic aortic enhancement.
With prospective ECG-gated acquisition,
the time window for each segmental image
data acquisition is separated by a separate
heartbeat for table feed from the next image acquisition. The 2-second lag time between sequential acquisitions may result in
prominent changes in aortic attenuation at
the two sequential levels, resulting in prominent attenuation changes with linear margins
at contiguous anatomic levels on subvolume MIP images obtained in the parasagittal plane. This appearance does not affect the
diagnostic quality of the study.
962
Image Quality and Motion Artifact
Motion artifacts that cause diagnostic uncertainty on CT images may occur in various parts of the body. Cardiac motion creates a challenge for diagnostic cardiac CT
angiography. Cardiac motion also affects the
perceived image quality of the adjacent anatomic structures—that is, the aortic valve
and ascending aorta [24]. Pseudodissection
of the thoracic aorta is a well-known pitfall.
This artifact occurs at the left anterior and
right posterior aortic circumference. A summation effect of both pendular and circular
aortic motion produces a curvilinear thickening of the aortic wall, which simulates a
dissection or hampers detailed visualization
of an intraluminal flap [25–27].
Roos et al. [8] used a 4-MDCT scanner to
show a significant reduction in motion-related artifacts for the entire thoracic aorta with
ECG-gated MDCT. Maximum reduction
was achieved at the level of the heart that
physiologically moves the most—the aortic valve—followed by the ascending aorta. Thomas et al. [28] have also shown that
ECG gating can be performed without delaying the workflow in an emergency setting,
resulting in improved MDCT images of the
thoracic aorta.
In our study with both prospective and retrospective ECG-gated acquisitions, the aortic valve and ascending thoracic aorta were
generally free of artifact with good to excellent image quality, reflecting fast gantry rotation speed and ECG-gated acquisition.
Banding artifact is caused by misregistration between the software-detected ECG signal and cardiac motion. This artifact is also
called synchronization or stairstep artifact.
This artifact is identified as one or more parallel or straight lines along which the cardiac
and aortic contours show abrupt steplike distortions. The most common reason that this
artifact occurs is nonoptimal phase selection during either prospective or retrospective ECG-gated acquisition because of either
tachycardia or cardiac arrhythmia [29].
Theoretically, prospective ECG gating is
more dependent on a regular heart rate than
retrospective ECG gating, and artifacts and
misregistration occur in the presence of cardiac arrhythmias [29]. In this study, however, no significant difference between the
two protocol groups was seen either in the
number of patients with banding artifacts or
in the banding artifact score. An increased
heart rate was more likely to induce banding artifact in both groups. More prominent
heart rate variation in both groups was observed in patients whose images were compromised by banding artifact than in those
whose images were not.
Because a heart rate of less than 65–70
bpm is recommended for optimal image
quality in both prospective and retrospective
ECG gating, E-blocker medication may be
administered at the time of study.
Study Limitations
This study was not a prospective, randomized evaluation. Comparable patient groups
were studied at different time intervals. Two
different types of contrast material were used
in both patient groups, although the distribution
of isoosmolar and low-osmolar contrast agents
was comparable between the two groups.
The study was also limited to evaluating
the thoracic aorta and did not evaluate the coronary arteries, which may be reconstructed
AJR:193, October 2009
ECG Gating for Thoracic CT Angiography
with equal fidelity using either a prospective
or retrospective ECG-gated technique. Recently published studies have attested to the
value of prospective cardiac gating in decreasing radiation dose for coronary artery CT angiography examinations while maintaining
adequate image quality [12].
Retrospective cardiac gating may be used
for quantitative analysis of left ventricular
function and for studies of aortic valve motion. Thus, more complete patient information may be derived from a retrospective
ECG-gated study. Nevertheless, a relatively
low-dose prospective ECG-gated CT angiographic examination of the coronary arteries and thoracic aorta may be combined with
either echocardiography, for functional evaluation of the left ventricle and aortic valve,
or nuclear scintigraphy, for perfusion evaluation of the left ventricle.
In conclusion, prospective ECG-gated thoracic CT angiography is a robust technique
allowing high-quality diagnostic studies to
be obtained with equivalent contrast material load but with a significantly reduced radiation dose compared with retrospective
ECG-gated imaging. Optimum image quality for either technique is obtained with a relatively slow heart rate, which may require
E-blocker medication.
References
1. Hayter RG, Rhea JT, Small A, Tafazoli FS, Novelline RA. Suspected aortic dissection and other
aortic disorders: multi-detector row CT in 373
cases in the emergency setting. Radiology 2006;
238:841–852
2. Bhalla S, West OC. CT of nontraumatic thoracic
aortic emergencies. Semin Ultrasound CT MR
2005; 26:281–304
3. Yoshida S, Akiba H, Tamakawa M, et al. Thoracic
involvement of type A aortic dissection and intramural hematoma: diagnostic accuracy—comparison of emergency helical CT and surgical findings. Radiology 2003; 228:430–435
4. Yamada I, Nakagawa T, Himeno Y, Numano F,
Shibuya H. Takayasu arteritis: evaluation of the
thoracic aorta with CT angiography. Radiology
1998; 209:103–109
AJR:193, October 2009
5. Matalanis G, Durairaj M, Brooks M. A hybrid
technique of aortic arch branch transposition and
antegrade stent graft deployment for complete
arch repair without cardiopulmonary bypass. Eur
J Cardiothorac Surg 2006; 29:611–612
6. Rubin GD, Armerding MD, Dake MD, Napel S.
Cost identification of abdominal aortic aneurysm
imaging by using time and motion analyses. Radiology 2000; 215:63–70
7. Vrtiska TJ, Fletcher JG, McCollough CH. Stateof-the-art imaging with 64-channel multidetector
CT angiography. Perspect Vasc Surg Endovasc
Ther 2005; 17:3–8
8. Roos JE, Willmann JK, Weishaupt D, Lachat M,
Marincek B, Hilfiker PR. Thoracic aorta: motion
artifact reduction with retrospective and prospective electrocardiography-assisted multi-detector
row CT. Radiology 2002; 222:271–277
9. Bi X, Deshpande V, Simonetti O, et al. Three-dimensional breathhold SSFP coronary MRA: a
comparison between 1.5T and 3.0T. J Magn Reson
Imaging 2005; 22:206–212
10. Weber OM, Pujadas S, Martin AJ, et al. Free-breathing, three-dimensional coronary artery magnetic
resonance angiography: comparison of sequences.
J Magn Reson Imaging 2004; 20:395–402
11. Pereles FS, McCarthy RM, Baskaran V, et al.
Thoracic aortic dissection and aneurysm: evaluation with non-enhanced true FISP MR angiography in less than 4 minutes. Radiology 2002; 223:
270–274
12. Earls JP, Berman EL, Urban BA, et al. Prospectively gated transverse coronary CT angiography
versus retrospectively gated helical technique:
improved image quality and reduced radiation
dose. Radiology 2008; 246:742–753
13. Morin RL, Gerber TC, McCollough CH. Radiation dose and computed tomography of the heart.
Circulation 2003; 107:917–922
14. White CS, Kuo D, Kelemen M, et al. Chest pain
evaluation in the emergency department. AJR
2005; 185:533–540
15. Shrimpton PC, Edyrean S. CT scanner dosimetry.
Br J Radiol 1998; 71:1–3
16. Mettler FA Jr, Wiest PW, Locken JA, Kelsey CA.
CT scanning: patterns of use and dose. J Radiol
Prot 2000; 20:353–359
17. Baker SR. Musings at the beginning of the hyperCT era. Abdom Imaging 2003; 28:110–114
18. Linton OW, Mettler FA Jr; National Council on
Radiation Protection and Measurements. National
conference on dose reduction in CT, with an emphasis on pediatric patients. AJR 2003; 181:321–
329
19. Paterson A, Frush D. Dose reduction in pediatric
MDCT: general principles. Clin Radiol 2007; 62:
507–517
20. Leschka S, Alkadhi H, Plass A, et al. Accuracy of
MSCT coronary angiography with 64-slice technology: first experience. Eur Heart J 2005;
26:1482–1487
21. Mollet NR, Cademartiri F, van Mieghem CA, et
al. High-resolution spiral computed tomography
coronary angiography in patients referred for diagnostic conventional coronary angiography. Circulation 2005; 112:2318–2323
22. Paul JF, Abada HT. Strategies for reduction of radiation dose in cardiac multislice CT. Eur Radiol
2007; 17:2028–2037
23. Bae KT. Peak contrast enhancement in CT and
MR angiography: when does it occur and why?
Pharmacokinetic study in a porcine mode. Radiology 2003; 227:809–816
24. Fleischmann D, Rubin GD, Paik DS, et al. Stairstep artifacts with single versus multiple detectorrow helical CT. Radiology 2000; 216:185–196
25. Qanadli SD, El Hajjam M, Mesurolle B, et al. Motion artifacts of the aorta simulating aortic dissection on spiral CT. J Comput Assist Tomogr 1999;
23:1–6
26. Loubeyre P, Angelie E, Grozel F, Abidi H, Minh
VA. Spiral CT artifact that simulates aortic dissection: imaging reconstruction with use of 180°
and 360° linear-interpolation algorithms. Radiology 1997; 205:153–157
27. Duvernoy O, Coulden R, Ytterberg C. Aortic motion: a potential pitfall in CT imaging of dissection in the ascending aorta. J Comput Assist Tomogr 1995; 19:569–572
28. Thomas S, Thomas E, Simon W, et al. Comparison of retrospectively ECG-gated and non-gated
MDCT of the chest in an emergency setting regarding workflow, image quality, and diagnostic
certainty. Emerg Radiol 2005; 12:19–29
29. Hsieh J. Advanced CT applications. In: Hsieh J.
Computed tomography: principles, design, artifacts, and recent advances. Bellingham, WA:
SPIE Optical Engineering Press, 2003:341–387
963