AJR 2003; 181:485-489
© American Roentgen Ray Society
MR Imaging–Guided Vascular Procedures Using CO2 as a Contrast Agent
Frank K. Wacker1,
Robbert M. Maes2,
Jack A. Jesberger1,
Sherif G. Nour1,
Jeffrey L. Duerk1 and
Jonathan S. Lewin1
1 Department of Radiology, University Hospitals of Cleveland, Case Western
Reserve University, 11100 Euclid Ave., Cleveland, OH 44106.
2 Department of Radiology, Gemini-Ziekenhuis, Huisduinerweg 3, 1782 GZ Den
Helder, The Netherlands.
Received December 18, 2002;
accepted after revision February 5, 2003.
Address correspondence to F. K. Wacker
(wackerfrank{at}web.de).
Supported in part by Siemens Medical Solutions and National Cancer
Institute grants R33CA88144 and R01CA81431.
Abstract
OBJECTIVE. The purpose of this study was to test the use of
CO2 as a black blood contrast agent for MR imaging–guided
vascular procedures in an animal model.
MATERIALS AND METHODS. Repeated intraarterial CO2
injections were performed through a catheter located in the aorta and the
renal arteries of three fully anesthetized pigs. Real-time images were
acquired using a steady-state free precession sequence.
RESULTS. During the CO2 injections, the bright blood in
the aorta and the main renal artery was totally replaced, and this procedure
resulted in an immediate, statistically significant signal loss in the vessel
lumen. In more peripheral vessels, CO2 improved the vessel
conspicuity substantially. Confirmation of vessel patency distal to the
catheter tip position was possible.
CONCLUSION. The use of carbon dioxide in combination with a bright
blood MR imaging sequence improves vessel conspicuity and provides immediate
information about blood flow distal to the catheter. This technique may be
used to facilitate MR imaging–guided intravascular procedures.
Introduction
Accurate visualization of target vessels and continuous confirmation of
vessel patency are essential during MR imaging–guided vascular
interventions. Most unenhanced MR angiography sequences, such as
time-of-flight and phase-contrast angiography, have limitations when used for
guidance of vascular interventions. Their acquisition times are longer than 1
sec, which precludes obtaining information about local hemodynamic conditions,
and additional time is thus required for postprocessing.
Contrast-enhanced angiography techniques are much faster. However, the use
of extracellular gadolinium compounds requires repeated injections because of
their short intravascular retention time. Moreover, the large quantities of
gadolinium chelates increase background signal intensity over the course of
the procedure and have been shown to have higher toxicity than modern
iodinated contrast media [1].
Therefore alternatives are needed.
One solution is the use of real-time bright blood MR steady-state free
precession (SSFP) sequences such as true fast imaging with steady-state free
precession (true FISP) [2,
3]. The bright signal of
arterial blood results from both the T2-like contrast and the inflow of
unsaturated protons [2,
4]. The use of this technique
appears to be promising for the visualization of larger vessels
[5]. In addition to its
excellent vessel conspicuity, true FISP imaging offers exceptionally high
signal at a short repetition time, thus making it an ideal imaging method to
guide interventional procedures
[2]. However, it is difficult
to assess the blood flow distal to an interventional device such as a catheter
or a stent when bright blood sequences are used. Confirmation of blood flow
distal to the catheter tip via injection of a signal-enhancing agent does not
generate a sufficient signal difference between flow-related enhanced blood
and that enhanced from the contrast agent to permit easy visualization.
Therefore, in this situation, a dark blood agent seems to be advantageous. In
a recent study, the use of CO2 has been proposed as a black blood
agent for MR angiography [6].
In conventional angiography, CO2 is known to be useful for a
variety of vascular interventional procedures
[7]. The aim of our study was
to evaluate the transcatheter use of CO2 in combination with a
real-time bright blood SSFP imaging sequence for endovascular MR imaging.
Materials and Methods
Animal experiments were conducted on three pigs weighing 20–28 kg.
The experimental protocol was approved by our institutional animal care and
use committee. The animals were anesthetized with an intramuscular injection
of 6–10 mg/kg tiletamine hydrochloride and zolazepam hydrochloride
(Telazol, Fort Dodge Animal Health, Fort Dodge, IA). For maintenance, ketamine
and xylazine were infused, and 1 mg/kg zolazepam hydrochloride was added
intramuscularly every 45–60 min. The animals were positioned in the MR
scanner (1.5-T Magnetom Sonata, Siemens Medical Solutions, Erlangen, Germany)
in a supine position. After acquisition of a localizer, coronal
two-dimensional true FISP images (TR/TE, 3.03/1.52; slice thickness, 5–8
mm; flip angle, 40°; 2 images/sec) were used to localize the aorta and the
renal arteries. The images were displayed on an in-room monitor adjacent to
the magnet; the same near-real-time imaging sequence was subsequently used for
the insertion of guidewires (Radifocus, Terumo, Tokyo, Japan; Ferrotip,
Somatex, Berlin, Germany) and MR imaging–visible catheters (Somatex;
Cordis, Roden, The Netherlands) through the introducer sheath located in the
distal iliac artery into the suprarenal aorta. The intravascular devices were
visualized by means of their susceptibility artifact. For catheter
manipulations, we used a fully MR imaging–visible prototype catheter
(C-1 configuration, 5-French diameter, Somatex) constructed with a ferrite
admixture and a straight catheter with multiple side holes (Cordis) equipped
with dysprosium markers, both of which were designed for susceptibility
artifact–based visualization under MR imaging guidance. For wire
manipulations, we used a steerable hydrophilic-coated guidewire (0.035-inch
Radifocus, Terumo), which generates no MR imaging artifact, and a prototype
Ferrotip guidewire (Somatex), which consisted of a 0.035-inch tapered MR
artifact-free nitinol guidewire shaft and a 2-mm ferromagnetic tip that
induces a round signal void. Neither the mechanical nor the biologic
properties of the catheter and guidewire prototypes used in this study differ
from conventional angiographic devices.
With the catheter in position, seven separate 40 mL CO2 boluses
per pig were manually injected; each bolus injection required 1 sec. The
CO2 was drawn from a closed-bag system. The time between each
injection was at least 1 min. Subsequently, the catheter was introduced into
the renal artery using MR imaging guidance, and three selective renal
CO2 injections (10–15 mL) were administered in each animal,
with MR images acquired throughout the CO2 bolus injection.
In the aorta, the effect of CO2 was quantified by measuring the
time-varying signal-to-noise ratios during the CO2 bolus injection
on every image (frame rate, 2 images/sec). For numeric analysis, the aortic
and background signal intensities were measured retrospectively using the SSFP
images. A circular region of interest was placed in the center of the aortic
lumen and in the background. Mean signal intensities of the regions of
interest were measured at the corresponding locations on the images acquired
during the injections. The area of the aortic region of interest was variable
depending on the target artery and ranged from 1326.00 to 19.00 mm2
(mean, 191.85 mm2). The mean area of the background region of
interest was 1103.15 mm2 (maximum, 1675.00 mm2; minimum,
313.00 mm2). On the basis of these measurements, time profiles were
calculated of mean signal-to-noise–ratio values for arterial regions of
interest. For each injection, three epochs (baseline, CO2 passage,
and signal recovery) were clearly identifiable in the signal-to-noise ratio
versus time plot. The durations of the epochs were variable both within and
between subjects. For consistency, the number of points included from each
epoch in the analysis was held constant and corresponded to the length of
shortest instance of that epoch across all trials. From each trial, the two
points just before the beginning of the CO2 passage were used for
preinjection baseline observations, the first two points of the CO2
passage were included as postinjection observations, and the three points just
after signal recovery (after bolus injection) were used for the last epoch
(Fig. 1). Four instances (from
all 21 boluses) of intermediate signal-to-noise–ratio values occurred at
the transition between the blood signal and the CO2 signal. These
values were interpreted as transitions between epochs rather than as belonging
to a particular epoch and thus were excluded from the analysis. The selected
signal-to-noise–ratio values within each epoch were pooled, averaged,
and plotted (Fig. 2).
View larger version (8K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2. —Graph shows pooled aortic signal-to-noise–ratio values for
each epoch. During CO2 passage, signal-to-noise ratio was
significantly less (p < 0.0005) than that of either baseline or
recovery.
|
|
A repeated-measures analysis of variance with the mean lumen
signal-to-noise ratio as the dependent variable and epoch type as a single
three-level (baseline, CO2 passage, and signal recovery)
within-subjects factor was performed. The statistical significance of the
effect of the independent factor was assessed using the following orthogonal
planned contrasts: baseline versus CO2 passage and CO2
passage versus signal recovery. Pearson's product-moment correlation
coefficients were also calculated between the baseline signal-to-noise ratio
and the signal drop during CO2 passage and between background noise
and the mean signal-to-noise ratio of each of the three epochs. For all tests,
a p value of 0.05 or less indicated a statistically significant
difference.
In the renal vessels, the effect of CO2 was subjectively graded
by two reviewers in consensus. The vascular tree was divided into three
segments: renal artery, first- and second-order branches, and intrarenal
arteries. At each segment interpretation, the reviewers graded the depiction
of the segment before and after CO2 injection on a three-point
scale as follows: A, inadequate (impossible to delineate a vessel); B,
intermediate (image quality is intermediate but vessel can be delineated); and
C, excellent (vessel segment can be clearly identified). In addition, the
effect of the CO2 injection on flow assessment was evaluated using
a cine-loop display. For each segment, the reviewers graded the depiction of
flow during both the CO2 injection and the recovery phase on a
three-point scale as follows: A, inadequate (impossible to delineate flow); B,
intermediate (no continuous vessel delineation but dynamic changes during
CO2 injection and washout can be identified); C, excellent
(CO2 injection and washout can be clearly identified along the
vessel segment).
Results
During bolus injection of CO2 into the suprarenal aorta, the
bright aortic blood was totally replaced for 1–5 sec. The CO2
formed a moving gaseous column resulting in an immediate decrease of the
signal intensity in the aortic lumen promptly after starting the injection
with an instant signal increase as the gas left the vessel segment (Fig.
3A,
3B). Thus, one bolus injection
allowed assessment of the flow not only during the actual contrast injection
but also when the bright blood returned and replaced the low-signal gas. Using
the near-real-time sequence, the CO2 bolus could be tracked into
the renal and splenic arteries (Fig.
3B). Repeated bolus injections showed the excellent
reproducibility of this sharply delineated signal decrease and signal recovery
(Fig. 1). The group
signal-to-noise–ratio means (± SD) for each epoch were baseline
(6.35 ± 1.43), CO2 passage (2.33 ± 0.74), and signal
recovery (5.95 ± 1.54) (Fig.
2). The signal-to-noise ratio during CO2 bolus passage
was statistically significantly less than either the baseline (Fisher's
F-ratio = 274.1, p < 0.0005) or the recovery signal-to-noise ratio
(Fisher's F-ratio = 167.0, p < 0.0005). The magnitude drop in
signal-to-noise ratio was highly correlated with the baseline signal-to-noise
ratio (r = 0.86, p < 0.0005). This correlation was
probably not a result of the signal-scaling differences between trials because
there was no correlation between the background (noise) signal intensities and
the raw signal levels during any of the three epochs (r < 0.12,
p = not significant in all cases). Also, the noise values were stable
across the trials (mean signal intensity of background noise, 30.72; 95%
confidence interval, 29.13–32.3). The most likely explanation is that
signal from a gaseous bolus might be considered equivalent to noise signal in
air outside the body. This would result in a calculated signal-to-noise ratio
of unity. However, the signal-to-noise ratio during CO2 passage was
statistically significantly higher than 1 (2.44 ± 0.76; one sample
t test, t = 8.47, p < 0.0005).
View larger version (135K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3A. —Continuously acquired MR images (TR/TE, 3.03/1.52; flip angle,
40°; 2 images/sec; matrix, 256 x 256; slice thickness, 6 mm)
obtained in pig. Coronal oblique two-dimensional (2D) true fast imaging with
steady-state free precession (true FISP) image shows abdominal aorta before
CO2 injection through catheter located in suprarenal aorta.
|
|
View larger version (141K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3B. —Continuously acquired MR images (TR/TE, 3.03/1.52; flip angle,
40°; 2 images/sec; matrix, 256 x 256; slice thickness, 6 mm)
obtained in pig. Coronal oblique 2D true FISP image shows abdominal aorta
during CO2 injection. Image position is identical to that in
A; both renal (solid arrow) and splenic (open arrow)
arteries can be better appreciated during CO2 injection.
|
|
After selective catheterization of the renal artery, a drop in the signal
intensity could be observed in all pigs. This was most likely caused by
reduced blood flow after catheter insertion
(Fig. 4B). Subsequent selective
CO2 injections into the renal artery facilitated its delineation
and that of the first- and second-order branches. This allowed assessment of
the vessel patency at all levels and improved conspicuity of the intrarenal
arteries, which were hardly visible with the true FISP sequence alone
(Table
1).
,
View larger version (185K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4B. —MR images (TR/TE, 3.03/1.52; flip angle, 40°; 2 images/sec;
matrix, 256 x 256; slice thickness, 5 mm) obtained in pig. Coronal
oblique 2D true FISP image obtained after catheter (arrow) insertion
into right renal artery shows that position of renal artery changed and image
plane had to be adjusted. Compared with A, intraarterial signal
intensity is reduced after catheter insertion.
|
|
View larger version (192K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4A. —MR images (TR/TE, 3.03/1.52; flip angle, 40°; 2 images/sec;
matrix, 256 x 256; slice thickness, 5 mm) obtained in pig. Coronal
oblique two-dimensional (2D) true fast imaging with steady-state free
precession (true FISP) image obtained before catheterization and contrast
injection shows aorta, right renal, splenic, and mesenteric arteries.
|
|
View larger version (186K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4C. —MR images (TR/TE, 3.03/1.52; flip angle, 40°; 2 images/sec;
matrix, 256 x 256; slice thickness, 5 mm) obtained in pig. Coronal
oblique 2D true FISP image obtained during CO2 injection through
catheter located in renal artery shows improved conspicuity of intrarenal
arteries and facilitates assessment of vessel patency distal to catheter.
|
|
Discussion
Both the performance and outcome of vascular interventions are based on
reliable visualization of target vessels and device guidance. Although it is
possible to confirm the location of the catheter tip by creating
susceptibility artifacts using dysprosium markers
[8,
9], ferrite admixture
[10], or electric currents
[11], these artifacts do not
provide subtle visualization during the steering process of the catheter tip
that is necessary to reliably exclude subintimal dissections with subsequent
device displacement. Moreover, selective injection is repeatedly needed during
a vascular intervention to evaluate treatment progress and to provide
information on flow conditions, both of which allow one to rule out
complications such as vessel occlusion or distal embolization. In an MR
imaging environment, the baseline vessel visualization can be achieved using
an imaging sequence with high inherent vessel-to-background contrast, such as
true FISP [2]. Direct
intraarterial injection of diluted paramagnetic compounds such as
gadopentetate dimeglumine have been used for confirmation of catheter position
and assessment of blood flow distal to the catheter
[9,
12–16].
However, in combination with an SSFP sequence, this technique aims to brighten
the already bright blood, which makes it difficult to assess the effect of the
contrast agent injection. Other major limitations of such agents are that they
rapidly diffuse into the extracellular space and result in marked background
enhancement over the course of the procedure and that they impose a negative
effect on kidney function, especially if administered in large quantities
[1,
17]. These limitations also
obviate the use of undiluted gadolinium compounds, which can theoretically be
used as dark blood agents.
Alternatively, the use of CO2 provides significantly improved
contrast to bright blood signal with SSFP imaging techniques. During passage
of CO2, a sharply delineated signal loss and a rapid signal
recovery after CO2 washout could be observed; both were highly
statistically significant. The sharp signal drop resulted from the temporary
replacement of the protons by CO2
[6]. Specifically, the
extremely low signal on the MR images results from the lack of MR signal of
CO2 at 64 MHz at 1.5 T and the displacement of protons by the gas
column. However, the signal-to-noise ratio during the CO2 passage
is still significantly higher than that of the background noise. Because a
signal-scaling difference could be ruled out and considering the slice
thickness and the aortic diameter in pigs
[10,
18], partial volume effects
within the coronal–coronal oblique images used for the
signal-to-noise–ratio measurements or incomplete displacement of blood
protons are the most likely explanations for this difference between noise and
signal-to-noise ratio during CO2 passage.
Because CO2 is a normal blood component, the human organism is
capable of clearing large quantities of CO2 completely during the
first pass through the lungs
[6,
19,
20]. In addition,
CO2 is inexpensive and has a favorable safety profile without known
allergic or nephrotoxic side effects
[19,
20]. In conventional aortic
angiography, CO2 has been used in a total dose up to 3000 mL. In
one case [7], a slight colitis
was found and attributed to entrapment and consequent ischemia in small
arteries. Widespread CO2 use in conventional angiography has been
avoided for two reasons. For diagnostic angiograms, images are often inferior
to those obtained using iodinated contrast agents. For interventional
procedures, the main concern is that valve leakage might result in undetected
air injection, leading to an air embolus with fatal consequences. Safe
CO2 injection is best achieved using a dedicated injector. However,
such delivery systems are not yet approved by the United States Food and Drug
Administration. Although safety was not evaluated in our study, over the past
25 years, many radiology departments have shown that CO2-associated
complications can be prevented by strictly adhering to a well-designed
protocol when performing these procedures in patients, thereby making
CO2 injection a safe procedure
[7,
19,
20].
The clinical utility of CO2 for MR imaging–guided
interventions may be limited if metallic implants such as stents or
embolization coils are used, because their signal void might interfere with
the signal loss induced by CO2. However, as artifact-free implants
become increasingly available
[21], this will no longer be a
concern. Another limitation of our study is that neither a thick slice nor
projection angiography technique was used. Such techniques might be easier to
apply if the use of CO2 can be combined with administration of
blood pool contrast agents that induce an increased vessel-to-background
contrast and thus enable MR angiogram-like imaging techniques.
In conclusion, the intraarterial transcatheter use of CO2 in
combination with an SSFP sequence for real-time MR imaging guidance improves
vessel conspicuity and allows selective assessment of flow distal to a
catheter during an intravascular MR imaging–guided procedure.
Acknowledgments
We thank Bonnie Hami for her invaluable editorial assistance and Elena
DuPont for help with manuscript preparation.
References
- Nyman U, Elmstahl B, Leander P, Nilsson M, Golman K, Almen T. Are
gadolinium-based contrast media really safer than iodinated media for digital
subtraction angiography in patients with azotemia?
Radiology2002; 223:311
–318[Abstract/Free Full Text]
- Duerk JL, Lewin JS, Wendt M, Petersilge C. Remember true FISP? a
high SNR, near 1-second imaging method for T2-like contrast in interventional
MRI at 0.2 T. J Magn Reson Imaging1998; 8:203
–208[Medline]
- Wendt M, Wacker F, Wolf KJ, Lewin JS, Duerk JL. Keyhole-true FISP:
fast T2-weighted imaging for interventional MRT at 0.2 T [in German].
Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr1999; 170:391
–393[Medline]
- Haacke EM, Wielopolski PA, Tkach JA, Modic MT. Steady-state free
precession imaging in the presence of motion: application for improved
visualization of the cerebrospinal fluid. Radiology1990; 175:545
–552[Abstract/Free Full Text]
- Spuentrup E, Ruebben A, Schaeffter T, Manning WJ, Gunther RW,
Buecker A. Magnetic resonance: guided coronary artery stent placement in a
swine model. Circulation2002; 105:874
–879[Abstract/Free Full Text]
- Maes RM, Matheijssen NA, Pattynama PM, Krestin GP. The use of
carbon dioxide in magnetic resonance angiography: a new type of black blood
imaging. J Magn Reson Imaging2000; 12:595
–598[Medline]
- Eschelman DJ, Sullivan KL, Bonn J, Gardiner GA Jr. Carbon dioxide
as a contrast agent to guide vascular interventional procedures.
AJR 1998;171:1265
–1270[Abstract/Free Full Text]
- Bakker CJ, Hoogeveen RM, Hurtak WF, Van Vaals JJ, Viergever MA,
Mali WP. MR-guided endovascular interventions: susceptibility-based catheter
and near-real-time imaging technique. Radiology1997; 202:273
–276[Abstract/Free Full Text]
- Bakker CJ, Bos C, Weinmann HJ. Passive tracking of catheters and
guidewires by contrast-enhanced MR fluoroscopy. Magn Reson
Med 2001;45:17
–23[Medline]
- Wacker FK, Reither K, Branding G, Wendt M, Wolf KJ. Magnetic
resonance-guided vascular catheterization: feasibility using a passive
tracking technique at 0.2 Tesla in a pig model. J Magn Reson
Imaging 1999;10:841
–844[Medline]
- Glowinski A, Kursch J, Adam G, Bucker A, Noll TG, Gunther RW.
Device visualization for interventional MRI using local magnetic fields: basic
theory and its application to catheter visualization. IEEE Trans
Med Imaging 1998;17:786
–793[Medline]
- Frayne R, Omary RA, Unal O, Strother CM. Determination of optimal
injection parameters for intraarterial gadolinium-enhanced MR angiography.
J Vasc Interv Radiol2000; 11:1277
–1284[Medline]
- Omary RA, Frayne R, Unal O, Grist TM, Strother CM. Intraarterial
gadolinium-enhanced 2D and 3D MR angiography: a preliminary study.
J Vasc Interv Radiol1999; 10:1315
–1321[Medline]
- Serfaty JM, Atalar E, Declerck J, et al. Real-time projection MR
angiography: feasibility study. Radiology2000; 217:290
–295[Abstract/Free Full Text]
- Omary RA, Henseler KP, Unal O, et al. Validation of injection
parameters for catheter-directed intraarterial gadolinium-enhanced MR
angiography. Acad Radiol2002; 9:172
–185[Medline]
- Bos C, Smits HF, Bakker CJ, Viergever MA. Selective
contrast-enhanced MR angiography. Magn Reson Med2000; 44:575
–582[Medline]
- Gemery J, Idelson B, Reid S, et al. Acute renal failure after
arteriography with a gadolinium-based contrast agent.
AJR 1998;171:1277
–1278[Free Full Text]
- Dondelinger RF, Ghysels MP, Brisbois D, et al. Relevant
radiological anatomy of the pig as a training model in interventional
radiology. Eur Radiol1998; 8:1254
–1273[Medline]
- Hawkins IF, Caridi JG. Carbon dioxide (CO2) digital
subtraction angiography: 26-year experience at the University of Florida.
Eur Radiol 1998;8:391
–402[Medline]
- Kerns SR, Hawkins IF Jr. Carbon dioxide digital subtraction
angiography: expanding applications and technical evolution.
AJR 1995;164:735
–741[Abstract/Free Full Text]
- Buecker A, Spuentrup E, Ruebben A, Gunther RW. Artifact-free
in-stent lumen visualization by standard magnetic resonance angiography using
a new metallic magnetic resonance imaging stent.
Circulation2002; 105:1772
–1775[Abstract/Free Full Text]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
