DOI:10.2214/AJR.07.2702
AJR 2008; 190:294-299
© American Roentgen Ray Society
Gas Bubble Motion Artifact in MDCT
Franklin Liu1,
Carlos Cuevas,
Albert A. Moss,
Orpheus Kolokythas,
Theodore J. Dubinsky and
Paul E. Kinahan
1 All authors: Department of Radiology, University of Washington Medical Center,
Box 357115, Seattle, WA 98195.
Received June 8, 2007;
accepted after revision September 10, 2007.
Address correspondence to F. Liu.
Abstract
OBJECTIVE. The purpose of our study was to characterize the imaging
features of an MDCT artifact caused by gas bubble motion.
MATERIALS AND METHODS. For the period 2002–2006, the CT images
of 10 patients that revealed a curvilinear artifact thought to be due to a gas
bubble moving during CT acquisition were retrospectively identified and
reviewed by an attending body radiologist. The clinical images were acquired
on MDCT scanners. A phantom containing water and moving air bubbles (gas
injection rates of 0.1 and 0.5 mL/s) was designed, built, and scanned using a
16-MDCT scanner. Computer simulation was used to test our hypothesis that
these observed artifacts originated from moving gas bubbles.
RESULTS. Semicircular tubular CT artifacts with attenuation close to
that of air were observed in all 10 clinical cases. The acquired MDCT images
from the phantom and the computer simulations showed tubular air-attenuation
semicircular artifacts that closely matched the imaging findings on the 10
clinical cases. Increased rates of bubble injection multiplied the artifact.
Based on the computer simulation, the precise appearance of the artifact
depends on the bubble location and velocity relative to the rotation of the CT
scanner.
CONCLUSION. Gas bubble motion during CT generates a semicircular
air-attenuation artifact that has not been previously described to our
knowledge. The artifact is typically found in bowel and other
liquid-containing structures in which a bubble of gas floats up through a
liquid medium during CT.
Keywords: abdominal imaging • artifact • CT • gas bubble • motion
Introduction
Artifacts of CT have been well studied
[1]. Documented artifacts range
from those due to limitations inherent in the physical processes of CT
acquisition, such as beam hardening and partial volume effect, to those caused
by the patient, such as motion. Identification of these artifacts is important
for evaluation of image quality and accurate interpretation of CT images.
Motion artifacts are often easily recognizable because of the organs or
body parts they arise from (e.g., heart, lungs, extremities) and because of
their characteristic blurring and distortion of large parts of the
reconstructed image. Much work has been done on motion artifacts that result
in misdiagnoses on CT [2,
3] and on techniques to reduce
these artifacts
[4–6].
However, the artifact produced by the isolated motion of a small structure,
such as a moving gas bubble, has not been previously described in the
literature to our knowledge. Such an artifact would be small and discrete and
could potentially be mistaken for a clinically relevant finding.
In this article, we describe a tubular semicircular artifact with
attenuation close to that of air, seen clinically in CT scans of the
gastrointestinal tract. To further understand the cause and nature of this
artifact, we built a CT phantom in which gas bubbles were injected into a
container with liquid content at different rates. We were also able to
reproduce the artifact and study its behavior using computer simulations.
Materials and Methods
Clinical Images
For the time period of 2002–2006, 10 adult patients were identified
whose abdominal or pelvic CT scans showed a tubular air-attenuation
semicircular artifact. These patients were referred to our institution for CT
of the chest, abdomen, or pelvis. All of the patients were scanned with 4-,
8-, or 64-MDCT scanners(LightSpeed Plus and LightSpeed VCT, GE Healthcare)
using our standard scanning protocols: helical scanning with pitch of 1.375,
reconstructed slice thickness of 2.5 mm, 120 kVp, 0.5–0.8 second per
rotation, and automatic exposure-controlled mAs. These diagnostic CT scans
were obtained in patients both with and without the ingestion of oral contrast
material. All images were reviewed by one of three attending radiologists
whose experience ranged from 5 to 20 years, at the body imaging department at
our institution.
Permission from our institutional review board and waiver of consent were
obtained to do a retrospective review of the clinical MDCT examinations for
the patient component of this HIPAA-compliant study.
Phantom Construction
To perform a phantom study, a model was designed to generate gas bubbles
moving through a liquid medium. For a container, we used a noncircular NEMA
IEC Body Phantom (Data Spectrum) with an approximate diameter of 30 cm and
axial length of 18 cm (Fig.
1A). Tubing was affixed to the bottom of the container for air
injection to produce air bubbles. The phantom was filled with tap water. The
phantom was then placed on the CT table. A CT injector (Medrad Stellant dual
power injector) was filled with 60 mL of room air and used to inject air into
the tubing at a controlled rate.
The phantom was scanned on the 16-MDCT component of a PET/CT scanner (DSTE,
GE Healthcare) using the same acquisition and reconstruction parameters as for
the patient images, but with a current of 100 mAs. Images were obtained with
air injection rates of 0.1 and 0.5 mL/s from the contrast injector.
Computer Simulation
A computer simulation of the bubble artifact was developed by one of the
authors using IDL (Interactive Data Language, RSI Inc.). The simulation method
used a basic mathematic model that was analogous to a simplified version of
the approach used by Comtat et al.
[7]. The simulation replicated
the relative velocities of the bubble and scanner rotation. Choosing an upward
velocity for the bubble is problematic. After formation, a small (e.g., 3-mm
diameter) bubble rapidly reaches its terminal velocity in less than 0.1 second
[8]. In pure water at 20°C,
the maximum terminal velocity is approximately 25–35 cm/s
[8,
9]. This velocity is strongly
affected by several parameters, including diameter, temperature, and presence
of contaminants (surfactants). The presence of acidic or basic surfactants in
the water reduces the terminal velocity to as low as 15 cm/s, depending on
concentration [9]. For bubbles
larger than a few millimeters, the terminal velocity can also depend on the
degree of bubble oscillation during motion. For the purposes of our study, we
assumed a bubble terminal velocity of 20 cm/s and a diameter of 5 mm.
The scanner rotation period was varied between 0.35 and 2.0 seconds per
rotation. The patient cross-section was modeled as a 33 x 25 cm ellipse
filled with water, and a parallel beam monoenergetic beam X-ray transform
model was used. For each scanner rotation period, the simulation was repeated
with the relative motion of the bubble used during the forward projection of
the sinogram (a pictorial depiction of the collected data as the CT gantry
rotates around an object). After forward projection, the sinogram data were
reconstructed using filtered back-projection, with only a small amount of
smoothing because photon and other noise effects were not included in the
simulation.
Results
Clinical Images
We observed the presence of a curvilinear tubular air-attenuation artifact
in all 10 clinical cases performed between 2002 and 2006 at our institution.
The most common anatomic location of the artifact was the stomach in eight of
10 patients (Table 1) at the
level of the gastroesophageal junction (Fig.
2A,
2B,
2C,
2D). It was also observed in
other liquid-filled structures such as the cecum at the level of the ileocecal
valve (Fig. 3A,
3B). In seven of the 10
patients, the semicircular tubular shape of the artifact appeared to pass
through the gastric or bowel wall (Table
1, Fig. 2A,
2B,
2C,
2D). The presence of the
artifact did not appear to depend on the slice thickness, number of scanner
rows, scanning phase, or the presence or absence of IV or oral contrast
material (Table 1). The degree
of arc, length, and thickness of the artifact varied widely among the clinical
cases.
View larger version (128K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2A —Gas bubble artifacts in four patients. Axial CT images of
abdomen at level of stomach show multiple examples of gas bubble artifacts
(arrows) in patient 1 (A), patient 9 (B), patient 3
(C), and patient 4 (D). Gas bubble originates from air passing
through gastroesophageal junction. Note how artifact extends beyond gastric
wall.
|
|
View larger version (141K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2B —Gas bubble artifacts in four patients. Axial CT images of
abdomen at level of stomach show multiple examples of gas bubble artifacts
(arrows) in patient 1 (A), patient 9 (B), patient 3
(C), and patient 4 (D). Gas bubble originates from air passing
through gastroesophageal junction. Note how artifact extends beyond gastric
wall.
|
|
View larger version (104K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2C —Gas bubble artifacts in four patients. Axial CT images of
abdomen at level of stomach show multiple examples of gas bubble artifacts
(arrows) in patient 1 (A), patient 9 (B), patient 3
(C), and patient 4 (D). Gas bubble originates from air passing
through gastroesophageal junction. Note how artifact extends beyond gastric
wall.
|
|
View larger version (123K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2D —Gas bubble artifacts in four patients. Axial CT images of
abdomen at level of stomach show multiple examples of gas bubble artifacts
(arrows) in patient 1 (A), patient 9 (B), patient 3
(C), and patient 4 (D). Gas bubble originates from air passing
through gastroesophageal junction. Note how artifact extends beyond gastric
wall.
|
|
Phantom Images
CT images of the phantom showed a tubular air-attenuation semicircular
structure that corresponded to the moving bubble of gas
(Fig. 4A). The appearance was
similar to the artifacts observed in the clinical images. Increasing the rate
of gas bubbling from 0.1 to 0.5 mL/s resulted in the appearance of multiple
semicircles on each image due to image capture of multiple moving gas bubbles
(Fig. 4B). The number of
artifacts was proportional to the number of bubbles seen at any given moment
in the phantom. Any bubble generated a single semicircular artifact, and the
more bubbles injected, the more artifacts were produced. Parameters such as
slice thickness did not appear to change the artifact.
View larger version (77K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4A —Images of phantom. CT images of phantom obtained at base of
phantom with air injection rates of 0.1 mL/s (A) and 0.5 mL/s
(B) from contrast injector. Shape of artifact was the same as observed
in patients.
|
|
View larger version (72K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4B —Images of phantom. CT images of phantom obtained at base of
phantom with air injection rates of 0.1 mL/s (A) and 0.5 mL/s
(B) from contrast injector. Shape of artifact was the same as observed
in patients.
|
|
Computer Simulations
We simulated a series of sinograms to better evaluate how this artifact was
produced. The sinogram was later reconstructed into the conventional 2D CT
image using filtered back-projection. A subset of the simulation results is
shown in Figure 5A,
5B,
5C,
5D,
5E. A point in image space
corresponds to a sinusoidal locus of data values in the sinogram; thus, in the
absence of bubble movement, the sinogram of the gas bubble contains a
consistent sinusoidal path (Fig.
5A, center), which is reconstructed correctly. However, with
upward movement of the gas bubble during acquisition, this sinusoidal path is
deflected, resulting in inconsistent sinogram data and a reconstructed image
showing an arc pattern or artifact (Fig.
5B). Using a slower gantry rotation speed (during which the bubble
would travel farther), there is an even more pronounced reconstructed arc
pattern (Figs. 5C and
5D). At a 90° gantry
offset, a different, but still inconsistent, sinogram is acquired, leading to
a difference in the shape of the artifact
(Fig. 5E).
View larger version (41K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5A —Simulations of artifacts generated by bubble motion during CT
acquisition. First column shows acquisition process; center column, resulting
sinogram after filtering; and last column, reconstructed image. Three viewing
directions (V1 = 0°, V2 = 90°, V3 = 180°) are shown for 180°
acquisition and for corresponding rows in sinogram. Arrows indicate distance
traveled by bubble during acquisition. Stationary bubble (A) and bubble
velocity of 20 cm/s with gantry rotation period of 0.35 second (B).
|
|
View larger version (35K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5B —Simulations of artifacts generated by bubble motion during CT
acquisition. First column shows acquisition process; center column, resulting
sinogram after filtering; and last column, reconstructed image. Three viewing
directions (V1 = 0°, V2 = 90°, V3 = 180°) are shown for 180°
acquisition and for corresponding rows in sinogram. Arrows indicate distance
traveled by bubble during acquisition. Stationary bubble (A) and bubble
velocity of 20 cm/s with gantry rotation period of 0.35 second (B).
|
|
View larger version (33K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5C —Simulations of artifacts generated by bubble motion during CT
acquisition. First column shows acquisition process; center column, resulting
sinogram after filtering; and last column, reconstructed image. Three viewing
directions (V1 = 0°, V2 = 90°, V3 = 180°) are shown for 180°
acquisition and for corresponding rows in sinogram. Arrows indicate distance
traveled by bubble during acquisition. Gantry rotation periods of 0.5
(C) and 1.0 (D) second.
|
|
View larger version (32K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5D —Simulations of artifacts generated by bubble motion during CT
acquisition. First column shows acquisition process; center column, resulting
sinogram after filtering; and last column, reconstructed image. Three viewing
directions (V1 = 0°, V2 = 90°, V3 = 180°) are shown for 180°
acquisition and for corresponding rows in sinogram. Arrows indicate distance
traveled by bubble during acquisition. Gantry rotation periods of 0.5
(C) and 1.0 (D) second.
|
|
View larger version (38K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5E —Simulations of artifacts generated by bubble motion during CT
acquisition. First column shows acquisition process; center column, resulting
sinogram after filtering; and last column, reconstructed image. Three viewing
directions (V1 = 0°, V2 = 90°, V3 = 180°) are shown for 180°
acquisition and for corresponding rows in sinogram. Arrows indicate distance
traveled by bubble during acquisition. Acquisition process is the same as in
C, with gantry rotation period of 0.5 second, but with 90° gantry
offset for start of acquisition as shown.
|
|
Discussion
Artifacts have always been important in medical imaging. The correct
identification of certain artifacts may help to characterize and diagnose an
abnormality, whereas others may sometimes obscure the findings or induce a
wrong diagnosis if not properly recognized. An article by Barrett and Keat
[10] divides CT artifacts into
four general groups: physics-based (which result from the physical processes
involved in the acquisition of CT data), patientbased (caused by patient
movement or metallic materials within the patient), scanner-based (which
result from imperfections in scanner function), and helical or multisection
(produced by the image reconstruction process) artifacts. To our knowledge,
the artifact described in our study has not been previously characterized and
could be classified as a combination of a motion and reconstruction artifact.
The cause of the artifact was determined to be a bubble of gas moving in a
liquid-filled structure during the image acquisition process.
Intuitively, one might expect CT of a gas bubble to result in a single
circular spot, like a snapshot, or possibly a straight cylindric tube, as
could be seen with a photograph with long exposure time. However, we showed
that CT of a moving gas bubble produces a curvilinear artifact. We have termed
this the "gas bubble motion artifact."The production of the gas
bubble artifact is a result of the image acquisition and reconstruction
processes: The initial beam attenuation data are collected from a rotating
gantry that continuously measures the gas bubble during its travel in liquid;
the data processing uses a reconstruction technique that assumes the
attenuation measurements were obtained from a static object. The vertical, or
nondependent, motion of the bubble during acquisition results in inconsistent
data and generates a systematic error, producing the final curved shape of the
artifact. For the purposes of our study we evaluated the rising nondependent
motion of the gas bubble. Although the motion of bubbles in other
directions—for example, transversely or dependently—through forced
action through a tube could result in other artifacts, this seems extremely
unlikely to occur in clinical practice.
The simulations performed were a simplified mathematic model of the CT
acquisition process and the gas bubble motion. The computer simulations
allowed us to replicate our observed phenomena, thus increasing our confidence
that the artifacts come from gas bubble motion, and to view the effect of
varying scanning parameters, in particular gantry rotation speed and relative
start position.
In an article by McCollough et al.
[11], CT motion artifacts were
generated through the use of a moving acrylic cylinder containing
high-contrast test objects, including air bubbles, that generated similar
artifacts to the ones described in our study. The objective of that article
was to study the relationship between time resolution and motion artifacts. It
is the only publication, to our knowledge, that includes images of moving gas
bubbles. However, the gas bubble motion artifact was not characterized in
their article.
One of the limitations of our study was that we did not reproduce our
findings experimentally in an actual patient or animal. However, the
correlation between the clinical images, the phantom, and the computer
simulation is strong evidence as to the cause of the bubble artifact. We also
did not vary the medium through which the gas bubbles traveled. The viscosity
of the liquid affects gas bubble velocity and thus may affect the size or
shape of the artifact. For the purposes of this study, we assumed that the
liquid in the bowel of the patients is comparable to water. Our results were
concordant with this assumption because the size and shape of the artifacts
obtained from the phantom and simulation studies were similar to the ones seen
in patient images, which also suggests that gas bubble velocities were
similar.
The clinical importance of the artifact can be related to two kinds of
possible interpretation errors: The artifact could obscure a lesion such as a
small polyp or an ulcer in the gastric or bowel wall; or the artifact might
simulate an abnormality such as bowel wall perforation by a tube. Although
this kind of artifact should be easily recognized, it is important to promote
education to prevent misinterpretations. Eliminating the presence of gas
bubbles in the patient is not possible, so any attempt to prevent the
generation of the artifact should be focused on the scanning method. The major
parameters that appear to determine the appearance of the gas bubble artifact
are the bubble size and velocity, scanner rotation speed, and relative
position of the X-ray tube. Theoretically, a significant decrease in scanning
times to 20 milliseconds or less should eliminate a motion artifact caused by
an object moving at the bubble speed (15–35 mm/s)
[11,
12]. Although the fastest
available CT scanners currently have an acquisition time of 83 milliseconds,
scanning times of 20 milliseconds will likely be reached in the near
future.
The gas bubble motion artifact is an air-attenuation smooth circular
structure that results from reconstruction of a moving gas bubble through a
liquid medium. Recognition of its distinctive characteristics can avoid a
potentially delayed or wrong diagnosis, and unnecessary further workup.
References
- Hsieh J. Computed tomography: principles, design,
artifacts, and recent advances. Bellingham, WA: SPIE Press,2003
- Katoh M, Wildberger JE, Gunther RW, Buecker A. Malignant right
coronary artery anomaly simulated by motion artifacts on MDCT.
AJR 2005; 185:1007
–1010[Abstract/Free Full Text]
- Tarver RD, Conces DJ, Godwin JD. Motion artifacts on CT simulate
bronchiectasis. AJR 1988;151
:1117
–1119[Free Full Text]
- Ritchie CJ, Hsieh J, Gard MF, Godwin JD, Kim Y, Crawford CR.
Predictive respiratory gating: a new method to reduce motion artifacts on CT
scans. Radiology 1994;190
: 847–852[Abstract/Free Full Text]
- Rubin GD, Leung AN, Robertson VJ, Stark P. Thoracic spiral CT:
influence of subsecond gantry rotation on image quality.
Radiology 1998;208
: 771–776[Abstract/Free Full Text]
- Arac M, Oner AY, Celik H, Akpek S, Isik S. Lung at thin-section CT:
influence of multiple-segment reconstruction on image quality.
Radiology 2003;229
: 195–199[Abstract/Free Full Text]
- Comtat C, Kinahan PE, Defrise M, Michel C, Townsend DW. Simulating
whole-body PET scanning with rapid analytical methods. Proceedings
of 1999 IEEE Nuclear Science Symposium and Medical Imaging
Conference 1999; 3:1260
–1264[CrossRef]
- Leifer I, Patro RK, Bowyer P. A study on the temperature variation
of rise velocity for large clean bubbles. J Atm Ocean
Tech 2000; 17:1393
–1402[CrossRef]
- Malysa K, Krasowska M, Krzan M. Influence of surface active
substances on bubble motion and collision with various interfaces.
Advan Colloid Interface Sci 2005;114–115
:205
–225[CrossRef]
- Barrett JF, Keat N. Artifacts in CT: recognition and avoidance.
RadioGraphics 2004;24
:1679
–1691[Abstract/Free Full Text]
- McCollough CH, Bruesewitz MR, Daly TR, Zink FE. Motion artifacts in
subsecond conventional CT and electron-beam CT: pictorial visualization of
temporal resolution. RadioGraphics 2000;20
:1675
–1681[Abstract/Free Full Text]
- Ritchie CJ, Godwin JD, Crawford CR, Stanford W, Anno H, Kim Y.
Minimum scan speeds for suppression of motion artifacts in CT.
Radiology 1992;185
: 37–42[Abstract/Free Full Text]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
|
|
 
D. J. Emby and K. C. Ho
Gas Bubble Motion Artifact
Am. J. Roentgenol.,
December 1, 2008;
191(6):
W312 - W312.
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
