European Heart Journal (2006) 27, 613–620
doi:10.1093/eurheartj/ehi732
Preclinical research
Feasibility of real-time magnetic resonance-guided
stent-graft placement in a swine model of descending
aortic dissection
Holger Eggebrecht1,2*, Hilmar Kühl2, Gernot M. Kaiser3, Stephanie Aker4, Michael O. Zenge2,
Frank Stock2, Frank Breuckmann1, Florian Grabellus5, Mark E. Ladd2, Rajendra H. Mehta6,
Raimund Erbel1, and Harald H. Quick2
1
Department of Cardiology, University of Duisburg-Essen, Hufelandstrabe 55, 45122 Essen, Germany; 2 Department of
Diagnostic and Interventional Radiology and Neuroradiology, University of Duisburg-Essen, Essen, Germany; 3 Department of
General and Transplantation Surgery, University of Duisburg-Essen, Essen, Germany; 4 Institute of Pathophysiology,
University of Duisburg-Essen, Essen, Germany; 5 Institute of Pathology, University of Duisburg-Essen, Essen, Germany; and
6
Duke Clinical Research Institute, Durham, NC, USA
Received 17 October 2005; revised 20 December 2005; accepted 22 December 2005; online publish-ahead-of-print 23 January 2006
KEYWORDS
Magnetic resonance imaging;
Endovascular aortic
dissection;
Stent-graft intervention
Aims To evaluate the pre-clinical feasibility of real-time magnetic resonance imaging (rtMRI) to guide
stent-graft placement for experimental aortic dissection (AD) and to alleviate disadvantages of ionising
radiation and nephrotoxic contrast media. Endovascular stent-graft placement for thoracic aortic
disease is usually performed under X-ray guidance. The feasibility of rtMRI-guided stent-graft placement
is currently not known.
Methods and results By using a catheter-based technique, dissections of the descending thoracic aorta
were successfully created in eight domestic pigs. Subsequent implantation of commercially available,
nitinol-based stent-grafts was performed entirely under rtMRI guidance. By pre-interventional
MRI, the mean minimal true-lumen diameter was 0.9 (0.825–0.975) cm. rtMRI permitted not only the
successful and safe device navigation within the true lumen from the iliac arteries to the thoracic
aorta, but also the precise positioning and deployment of the stent-graft and safe withdrawal of the
delivery catheter in seven of eight pigs. This was achieved without any other complications. After
the stent-graft placement, MRI demonstrated complete obliteration of the false lumen, which was
confirmed at autopsy. All stent-grafts were well expanded resulting in an increase in the size of the
true-lumen diameter to 2.05 (1.925–2.1) cm (P ¼ 0.066 vs. baseline).
Conclusion In experimental AD, rtMRI-guided endovascular stent-graft placement is feasible and safe
and has the potential for mitigating radiation and contrast-related side effects. Additionally, it allows
not only pre-interventional diagnosis and detailed anatomic diagnosis, but also permits immediate
post-interventional, anatomical, and functional delineation of procedure success that may serve as a
baseline for future comparison during follow-up.
Introduction
Endovascular stent-graft placement is emerging as a promising alternative to medical and surgical treatments of
patients with descending aortic dissection (AD).1 The aim
of endovascular repair in AD is to obliterate the false
lumen by the implantation of a membrane-covered stentgraft into the true lumen across the proximal entry tear.2
Precise placement of the stent-graft, which is currently
performed under X-ray, remains challenging, as there are
several shortcomings to fluoroscopic guidance, beyond that
related to the harmful effect of radiation exposure and
*Corresponding author. Tel: þ49 201 723 4888; fax: þ649 201 723 5480.
E-mail address: [email protected]
nephrotoxic contrast media. With fluoroscopy, it is often
difficult to verify the guide-wire position in the true
lumen, which is mandatory before the stent-graft placement. Contrast angiography, for positioning of the stentgraft, achieves only brief opacification of the large-caliber
aorta and frequent repetition as needed, increases contrast
use resulting in greater potential for contrast-mediated
renal dysfunction. X-ray angiography will not always show
correct location of the entry tears and re-entry tears, and
does not allow consistent differentiation of false from true
lumen as well as the adequate visualisation of the false
lumen. Immediate evaluation of procedure success after
stent-graft placement (i.e. thrombosis of the false lumen)
is thus not possible. Although transoesophageal echocardiography (TEE) and intravascular ultrasound have been
& The European Society of Cardiology 2006. All rights reserved. For Permissions, please e-mail: [email protected]
614
used as adjunct imaging modalities during endovascular
stent-graft procedures to overcome the limitations of angiography, these techniques have not mitigated the need for
fluoroscopy.3–5
Magnetic resonance imaging guidance of vascular interventional procedures offers several potential theoretical
advantages over fluoroscopy-guided techniques, including
image acquisition in any desired orientation, superior threedimensional (3D) soft-tissue contrast with simultaneous visualisation of the interventional device, absence of ionising
radiation, and avoidance of nephrotoxic contrast media.6,7
Magnetic resonance imaging is often used for pre-operative
diagnosis of AD and can provide all relevant information for
the planning of endovascular aortic stent-graft procedures8
and for accurate and immediate post-interventional evaluation.9 Recent studies have also suggested that MRI may be
equivalent or superior to standard techniques [e.g. computed
tomography (CT), TEE] for follow-up examinations.10
The aim of the present study was to evaluate the preclinical feasibility of real-time magnetic resonance imaging
(rtMRI) to guide stent-graft placement in an animal model
of descending thoracic AD, using commercially available
stent-graft devices. We envisioned that if this modality
can be successfully demonstrated to guide stent-graft
placement for AD without increasing risks, a single imaging
modality can potentially be used for diagnosis of AD,
guiding stent-graft placement for its treatment, ascertaining success and procedure-related complications and for
future follow-up of the disease process.
Methods
Animal preparation
In vivo experiments were performed on fully anesthetised domestic
pigs (n ¼ 8) weighing 63–98 kg. The experiments were conducted
in accordance with all regulations set forth by institutional and
governmental agencies. The animals were anesthetised initially by
H. Eggebrecht et al.
intramuscular injection of a combination of ketamin hydrochloride
(30 mg/kg), azaperon (2 mg/kg), and atropine (0.02 mg/kg),
followed by an intravenous continuous infusion of propofol
(160 mg/h), fentanyl (0.4 mg/h), and midazolam (40 mg/h). The
pigs were intubated with an endotracheal tube (6.5–7.5 mm internal
diameter), and ventilation was maintained during the experiment
using a Dräger UV-2 ventilator (Dräger Medical, Lübeck, Germany)
with 100% oxygen. Following the experiments, the animals were
euthanised by bolus injection of 80 mg/kg pentobarbital.
Creation of descending AD
We modified a previously described catheter-based technique for
non-surgical creation of descending ADs.11 A 10-F introducer
sheath (Avanti introducer, Cordis, Miami, FL, USA) was surgically
inserted into the left common carotid artery (CCA). A second surgical access to the right common iliac artery (CIA) was obtained for
later introduction and advancement of the stent-graft delivery
system. Creation of dissection was performed under X-ray in a
cardiac catheterisation laboratory. First, a standard 6-F pigtail catheter (Cordis, Miami, FL, USA) was advanced through the left CCA
sheath and a digital aortogram of the thoracic aorta was obtained
(Figure 1A). Then, the pigtail catheter was exchanged for the 9-F
needle of a Colapinto transjugular liver access set (Cook,
Bloomington, IN, USA) over a standard 0.03500 guide wire with a
length of 260 cm (Cordis, Miami, FL, USA). The guide wire was
pulled back inside the Colapinto needle and the tip of the needle
was positioned against the dorsal wall of the descending thoracic
aorta, 3–4 cm distal to the origin of the left subclavian artery. A
0.01400 coronary guide wire (Floppy II, ACS Guidant, Santa Clara,
CA, USA) was inserted with its stiff end first until 2–3 cm of the
wire end extended beyond the needle tip into the aorta. The
needle/wire combination was used in this manner to protect
against the inadvertent transmural penetration of the needle tip
during the creation of the initial subintimal tear. The needle/wire
combination was then advanced inferolaterally against the aorta
until the needle tip engaged the aortic wall. Approximately 10 mL
of iodinated contrast media (Ultravist 300, Schering, Berlin,
Germany) was forcefully hand-injected through the needle to
create the initial dissection (Figure 1B). The 0.03500 guide wire
was then advanced into the thus created subintimal pocket to
Figure 1 Catheter-based creation of experimental descending AD under X-ray fluoroscopy. (A) Pre-interventional digital angiogram of the descending thoracic
aorta (anterior–posterior projection). (B) The tip of a 9-F Colapinto needle has engaged the posterolateral wall of the thoracic aorta. The initial dissection pocket
(arrow) is created by forceful hand-injection of contrast media. (C and D) The subintimal pocket is enlarged (C) and extended inferiorly to the abdominal aorta
(D) by repeated contrast injections and guide-wire manipulations. (E) Digital angiogram of both true (TL) and false lumen (FL) after creation of the dissection
showing the undulating dissection flap (arrowheads).
MR and aortic stent-grafting
secure the position of the needle tip, and the 0.01400 guide wire was
removed. Further forceful hand injections of contrast media were
performed to enlarge the subintimal pocket (Figure 1C). The
needle was exchanged for the 6-F pigtail catheter over the 0.03500
guide wire, which was coiled in the dissection sac. A total of
10 000 IU of heparin was given through the pigtail catheter, and
the dissection was extended inferiorly by wire manipulation and/
or forceful hand-injection of contrast media. Dissections were
extended into the distal thoracic or abdominal aorta, resulting in
spontaneous creation of a distal re-entry tear (Figure 1D).
Angiograms of both true and false lumens were obtained
(Figure 1E). After creation of the dissection, the stent-graft delivery
system was inserted into the femoral artery and the pigs were transferred to the MR scanner for subsequent stent-graft placement.
Stent-graft device
On the basis of a comprehensive in vitro analysis of MRI characteristics of six commercially available thoracic aortic stent-graft
devices,12 a self-expandable, nitinol-based tubular stent-graft
device (GoreTAG, W.L. Gore Inc., Flagstaff, AZ, USA), which—in
combination with its delivery system—produced the fewest MR artifacts was chosen for the in vivo experiments (Figure 2). Stent-grafts
with lengths of 10 cm and diameters of 2.6 and 2.8 cm, respectively,
were available. The loaded stent-graft was constrained by an
implantable ePTFE sleeve on the leading end of an 18-F polyurethane delivery catheter. Middle-to-end deployment was initiated
by retraction of a GoreTex filament, which rapidly released the
stent-graft from the ePTFE sleeve. No modifications of the commercially available stent-graft device were made.
MRI scanner and sequences
Scanning was performed on a 1.5 T whole-body MRI scanner
(Magnetom Avanto, Siemens Medical Solutions, Erlangen, Germany)
equipped with gradients capable of a maximum amplitude of
45 mT/m and a slew rate of 200 T/m s21. The animals were
placed head first in a supine position inside the scanner on a
spine phased-array radio frequency (RF) coil with three clusters
of coil elements activated for signal reception (Tim Technology,
Siemens Medical Solutions, Erlangen, Germany). Two body flex
phased-array RF coils, each consisting of two clusters of coil
elements, were placed anteriorly on the pig.
615
Pre-interventional MRI evaluation
After creation of descending AD, pre-interventional MRI evaluation
was commenced with steady-state free precession imaging with
ECG gating and retrospective image reconstruction (TrueFISP retro)
within one breath-hold in parasagittal and in axial orientation.
Imaging parameters were repetition time (TR) ¼ 40 ms, echo
time (TE) ¼ 1.1 ms, flip angle ¼ 778, field-of-view (FOV) ¼ 380 330 mm2, matrix ¼ 192 168, slice thickness ¼ 6 mm, bandwidth
(BW) ¼ 930 Hz/pixel. Image acquisition time was 15 s for a single
slice. After data acquisition with an RR-interval of, for example,
680 ms, 20 phases of the cardiac cycle (i.e. one image every 34 ms)
were reconstructed retrospectively. The minimal diameter of the
true lumen was measured on the TrueFISP retro sequences.
Pre-interventional MRI evaluation further encompassed 3D
contrast-enhanced (CE) MR angiography (MRA) with a fast T1weighted spoiled gradient echo sequence (FLASH) in coronal orientation. An automatic contrast injector (Spectris Medrad Inc.,
Indiana, PA, USA) was programmed to inject 32 mL of paramagnetic
contrast agent (Dotarem, Guerbet, Cedex, France) into an ear vein
of the pig, with an injection rate of 2 mL/s followed by a 30 mL
saline flush at the same flow rate. Contrast bolus arrival time in
the aorta was determined with a test bolus technique that provided
one 2D fast gradient echo image per second in a parasagittal orientation. Imaging parameters for the 3D CE FLASH MRA sequence were
TR ¼ 2.5 ms, TE ¼ 1.0 ms, flip angle ¼ 158, FOV ¼ 400 400 mm2,
matrix ¼ 384 246, BW ¼ 685 Hz/pixel. A coronal slab with an
overall slab thickness of 154 mm was acquired and 128 slices with
an interpolated slice thickness of 1.2 mm each were reconstructed.
Image acquisition time was 18 seconds per slab covering the ascending and the descending aorta as well as the aortic arch. Three
successive acquisitions were performed, the first acquiring a native
data set for subsequent image background subtraction, the second
phase acquiring the arterial contrast phase, followed by the third
phase covering the venous phase of the aortic angiogram. Images of
the second (arterial) and the third acquisition phase (venous) were
post-processed to produce maximum-intensity-projections (MIP)
enabling comprehensive 3D angiographic visualisation of the aortic
pathology.
rtMRI interventional guidance
For implantation, the delivery system with the mounted stent-graft
was advanced from the CIA to the thoracic aorta under rtMRI fluoroscopy without use of guide wires. The MRI fluoroscopy was based on
an interactive real-time steady-state free precession sequence
(TrueFISP) with radial k-space filling (TR ¼ 3.0 ms, TE ¼ 1.5 ms,
flip angle ¼ 808, FOV ¼ 360 360 mm2, matrix ¼ 192 192,
BW ¼ 1530 Hz/pixel, slice thickness ¼ 6 mm) during free breathing
and without cardiac triggering. Forty-nine echoes were acquired
for each reconstructed image, resulting in real-time radial projection images with a frame rate of seven reconstructed frames per
second. Images were displayed without detectable image reconstruction delay (,200 ms). The images were projected with an
RF-shielded video projector onto an in-room 6000 high-contrast projection screen (MRI screen, MR-Innovation GmbH, Essen, Germany).
Image slice position, orientation, and contrast parameters could be
changed and adapted interactively from inside the scanner room
while the sequence was running. Imaging position was adjusted to
a parasagittal orientation for anatomic display of the descending
thoracic aorta as derived from the MRI sequence to visualise the
longest course of the thoracic aorta including ascending aorta,
aortic arch, and descending aorta.
Post-interventional MRI evaluation
Figure 2 Photograph of the GoreTAG stent-graft. (A) The stent-graft loaded
on the 18-F polyurethane delivery catheter. (B) The expanded nitinol-based,
membrane-covered stent-graft.
After stent-graft placement, the pre-interventional MRI protocol,
including parasagittal and axial TrueFISP retro sequences as well
as contrast enhanced 3D FLASH MRA in three phases, was repeated
to evaluate procedure success. Additionally, post-contrast
616
T1-weighted 3D volume interpolated breath-hold examination
(VIBE) sequence was acquired in axial orientation to evaluate and
confirm post-interventional thrombus formation in the false lumen
of the treated dissection. Imaging parameters for the 3D VIBE
sequence were TR ¼ 3.1 ms, TE ¼ 1.3 ms, flip angle ¼ 128,
FOV ¼ 400 400 mm2, matrix ¼ 256 134, BW ¼ 560 Hz/pixel, fat
saturation, slice thickness ¼ 3 mm, 128 slices acquired within 21 s.
Similar to pre-interventional evaluation, minimal true-lumen diameter was measured on TrueFISP images. Images were assessed
qualitatively for grade of false-lumen thrombosis and compared
with macroscopic examination of the ex vivo aorta.
Statistical analysis
Continuous variables are presented as medians and inter-quartile
ranges. Diameters of true lumen before and after stent-graft
placement are compared using the non-parametric Wilcoxon
signed-rank test.
Results
Pre-interventional MRI evaluation
Descending ADs were successfully created in all eight pigs.
In the first three pigs, pre-interventional MRI evaluation
revealed that non-communicating, subadventitial dissections of the descending thoracic aorta were created, probably because of deep needle penetration, highlighting the
limitation of creating an experimental AD under flouroscopy.
In one of these pigs, transmural perforation with haemothorax occurred, but the pig survived for the duration of
the experiment. In the remaining five pigs, classic doublebarrel dissections were created. Pre-interventional MRI
showed the typical aspect of communicating true and false
H. Eggebrecht et al.
lumen with an undulating dissection flap. Dissections
extended to the distal thoracic aorta in three and to the
abdominal aorta in two pigs, respectively. The minimal
diameter of the true lumen was 0.9 (0.825–0.975) cm. The
maximum diameter of the aorta (true and false lumens)
was 2.45 (2.175–2.65) cm.
Stent-graft placement under rtMRI
rtMRI allowed for visualisation of both vessel lumen and
delivery system with the mounted stent-graft, providing
image quality sufficient for successful intervention.
Susceptibility artifacts caused by the stent-graft mounted
on the delivery system enabled adequate determination of
the position of the loaded stent-graft in relation to the surrounding anatomy, but caused no undue image distortion. In
one pig, the delivery system was inadvertently inserted into
the false lumen. This was immediately detected by MRI
during advancement of the delivery system (Figure 3, see
Supplementary material online, Movie S1) and corrected.
After repositioning the device into the true lumen, the pig
developed severe arrhythmias with haemodynamic instability, and was euthanised before stent-graft placement could
be attempted. In the remaining seven of the eight pigs, MRI
verified the correct delivery system position within the true
lumen. Subsequently, successful and safe navigation of the
stent-graft delivery system to the thoracic aorta was well
monitored by rtMRI (Figure 4). The susceptibility artifacts
of the loaded stent-graft in combination with the simultaneously visible dissection flap allowed for adequate
positioning over the proximal dissection tear (Figure 4, see
Supplementary material online, Movie S1). Subsequent
Figure 3 (A) rtMRI in parasagittal orientation showing that the stent-graft delivery system (arrows) is inadvertently positioned within the false lumen (TL, true
lumen; FL, false lumen; arrowheads pointing at the dissection flap). (B) The delivery system position in the FL (arrow) is confirmed by the corresponding axial MRI
plane (TL, true lumen). For this purpose, image orientation was changed to an axial slice position from inside the scanner room while the real-time sequence was
continuously running.
MR and aortic stent-grafting
617
Figure 4 (A–E) Safe advancement of the stent-graft delivery system (arrows) up to the level of the dissection (arrowheads) under rtMRI guidance. (F) The
correct stent-graft position (arrows) is confirmed immediately after stent-graft deployment showing complete coverage of the dissection.
stent-graft deployment was successfully completed in seven
of the eight cases and was well visualised by MRI (Figure 4,
see Supplementary material online, Movie S2). After stentgraft deployment, withdrawal of the delivery system was
also adequately visualised under rtMRI. No other complications occurred. MRI time for navigation, positioning, and
stent-graft deployment was 2 (2–4) min.
Post-interventional MRI evaluation
Post-interventional MRI evaluation demonstrated correct
position of the stent-graft in all seven cases. All stentgraft were well expanded. The minimal diameter of the
true lumen had increased to 2.05 (1.925–2.1) cm after
stent-graft placement (P ¼ 0.066 vs. baseline). MRI showed
complete coverage of the proximal dissection tear and
thrombosis of the false lumen (Figures 5–7) as was confirmed by macroscopic examination of the ex vivo aortae
(Figures 7 and 8) in all seven pigs.
Discussion
The present study shows the pre-clinical safety and feasibility of endovascular stent-graft placement for descending
AD performed entirely under rtMRI in a swine model. This
was achieved using commercially available catheter
devices without any modifications made. Beyond detailed
pre-interventional anatomic diagnosis of AD that facilitated
treatment planning, rtMRI permitted (1) verification of
delivery system position within the true lumen, (2) successful and safe device navigation to the thoracic aorta, (3)
precise stent-graft positioning, (4) stent deployment, and
(5) safe catheter withdrawal. After stent-graft placement,
functional imaging with dynamic contrast MRA complemented the anatomic imaging for immediate confirmation of
treatment success. In fact, procedural success as identified
by complete obliteration of the false lumen was observed
in all cases by MR, and was further confirmed by macroscopic examination of the ex vivo aorta at necropsy.
Recently, the feasibility of rtMRI-guided interventions has
been demonstrated for a wide range of vascular interventional procedures in animals, including percutaneous
Figure 5 (A) Pre-interventional high-resolution MRI (TrueFISP retro) in parasagittal orientation showing the dissection flap (arrowheads) in the proximal
descending thoracic aorta. One of 20 acquired frames per RR-interval is
shown. (B) Corresponding post-interventional MRI demonstrating correct position of the stent-graft (arrows) with complete coverage of the dissection.
transluminal angioplasties,13 selective embolisation procedures,14 placement of atrial septal closure devices,15
peripheral as well as coronary stent placements,16 and
most recently, for stenting of aortic coarctation.17 For
peripheral stent placement, the feasibility of MRI guidance
has also been shown in patients.18 In terms of endovascular
aortic stent-graft placement, which is an emerging treatment option for patients with descending thoracic aortic
disease, MRI appears to be particularly useful as it can
provide (1) comprehensive pre-interventional diagnosis and
evaluation of the vascular aortic morphology and pathology
for planning of the intervention,8 (2) interventional image
guidance and monitoring of the therapy itself,6 and (3)
immediate evaluation of treatment success and ascertainment of procedure-related complications and follow-up
examinations.10 Mahnken et al.6 demonstrated the general
feasibility of MRI-guided aortic stent-graft placement.
They implanted commercially available stent-grafts into
618
the infrarenal aorta of healthy pigs under MRI guidance.6
Very recently, Raman et al.19 showed that endovascular
repair of experimental abdominal aortic aneurysm can,
indeed, be successfully performed under rtMRI. Our experiments extend this paradigm to the successful navigation of
the stent-graft delivery system up to the descending
Figure 6 (A) Pre-interventional 3D CE MRA (anterior–posterior projection)
showing true and false lumen (arrowheads). (B) Corresponding contrastenhanced 3D MRA after stent-graft placement showing that the false lumen
(arrowheads) is completely obliterated.
H. Eggebrecht et al.
thoracic aorta and confirm its feasibility under rtMRI guidance even in dissected aortas. Real-time TrueFISP imaging
sequences with radial k-space filling provided excellent
intrinsic blood contrast within the aorta to allow reliable
identification of true and false lumens and further for
successful treatment of complex aortic pathology. Such
steady-state sequences, when used with high-excitation
flip angles, provide high instrument to background contrast,
even without administration of a contrast agent which make
the stent-graft conspicuous in the MR image.20 The temporal
resolution of seven reconstructed images per second,
displayed without detectable image reconstruction delay
on an in-room high-resolution projection screen, proved sufficient for direct control of the catheter instruments and
allowed for precise positioning of the stent-grafts. Stentgraft deployment was also adequately depicted. Neither
respiratory or cardiac triggering nor breath-hold was
required to obtain images without visual degradation by
motion artifacts.
Collection of precise 3D information about soft tissue
anatomy and reliable visualisation of catheter instruments
in relation to the surrounding anatomy, represent the
pre-requisites for safe and successful performance of
vascular interventions under MRI guidance.21 In contrast
to ultrasound, X-ray fluoroscopy, or CT, visualisation of
interventional instruments by MRI has, however, proved
to be difficult.20 The optimal MRI technique to visualise
vascular instruments should provide high spatial and temporal resolution, and should also provide a high-contrast
instrument signature, making it easy to pick out the instrument in the MR image.22 A number of approaches have
been developed for depicting vascular instruments by
MRI. Active device tracking is accomplished by incorporating small RF receiver coils or antennas into the tip or distal
end of a catheter device, which are connected to the
receiver ports of the MR scanner with miniaturised
Figure 7 (A and B) Pre-interventional high-resolution MRI (TrueFISP retro) in axial orientation showing dissection of the thoracic (A) and abdominal aorta (B),
with true and false lumens (*) and the undulating dissection flap. (C and D) Corresponding post-interventional MRI (post-contrast T1-weighted 3D VIBE) showing
the stent-graft within the true lumen (C) while the false lumen is completely thrombosed (*). False lumen thrombosis (*) extends down to the abdominal aorta.
(E and F ) Corresponding macroscopic slices at the level of the in vivo stent-graft position (E, stent-graft removed) and at the level of the abdominal aorta (F)
confirming complete thrombosis (*) of the false lumen (FL).
MR and aortic stent-grafting
619
would allow a more reliable and consistent creation of AD,
as it allows definitive delineation of the various parts of
aortic wall, needs to be determined in future research.
Limitations
Figure 8 Ex vivo specimen of the descending thoracic aorta (view onto the
true lumen) showing the proximal dissection tear (arrow). The intimal
imprints of the stent-graft demonstrate the correct in vivo position of the
stent-graft. Successful obliteration of flow is evidenced by complete thrombosis (*) of the false lumen (FL).
coaxial cables running through the catheter body. The use
of active catheter tracking techniques maybe potentially
hazardous for the patient as long as no safety-tested and
commercially available active devices are available.22 In
contrast, passive device tracking by MRI uses metalinduced susceptibility artifacts of the interventional
instrument for visualisation. Passive tracking requires no
hardware or instrument modifications and thus appears
to be promising in terms of potential clinical applications.
However, passive visualisation requires MR-compatible
instruments with satisfactory susceptibility artifacts. The
artifact has to allow adequate visualisation of the dimensions of the interventional device in relation to the
surrounding tissue (i.e. aortic lumen), but should not
obscure relevant parts of the FOV.23 The artifact should
be as small as possible while maintaining adequate conspicuity.24 The passive commercial stent-grafts used in the
experiments of Raman et al.19 were difficult to visualise
with MRI, limiting operator confidence in positioning and
deployment. Home-made active stent-grafts were found
to facilitate device visualisation and thus positioning.19
On the basis of a previous comprehensive in vitro evaluation of six current thoracic aortic stent-graft devices,12
a nitinol-based endoprosthesis, which in combination with
its delivery system produced the least artifacts, was
chosen for the present experiments. As shown by the
present study, the artifact of the device did not obscure
visualisation of the thin and undulating dissection
lamella. Therefore, rtMRI allowed adequate visualisation
of the delivery device position in relation to the true or
false lumens. Verification of true lumen position of the
device is mandatory before stent-graft deployment, but
is often difficult by X-ray fluoroscopy alone.
Finally, the occurrence of non-communicating, subadventitial experimental dissections of the descending thoracic
aorta, probably due to deep needle penetration, highlights
the limitation of catheter-based creation of experimental
AD under fluoroscopy. This is because X-ray does not
permit differentiation of various aortic layers, thereby providing no guidance with respect to the depth of the needle
insertion for creating a tear in the wall. Whether MRI
In the present experiments, 18-F stent-graft delivery
systems were easily navigated to the proximal descending
thoracic aorta without the use of guide wires. Continuous
monitoring of the delivery system with simultaneous
depiction of the aortic anatomy by MRI allowed for safe
stent-graft navigation and prevented complications, such
as aortic perforation, despite the presence of ADs.
However, the chief obstacle to clinical translation of our
findings is the current unavailability of guide wires suited
for interventional MRI. Localised increases in the RF specific
absorption rate near metallic implants are important potential safety hazards associated with MRI examinations. The
local electric field can be amplified; especially if the
implants are composed of long conducting structures that
potentially can couple significantly with the RF transmit
energy of the body coil. This might lead to excessive local
tissue heating if standing resonant waves are generated,25–27 thus rendering conventional metallic guide wires
as well as guide wires with a nitinol core incompatible for
MR-guided interventions. These resonance effects occur
only when the length of wire-like structures or longitudinal
implants exceeds one-half of the RF wavelength, that is,
around 26 cm for 1.5 T in human tissue.25 For implants
that are small compared with the RF wavelength (e.g.
stents, coils, clips, endografts, etc.), RF resonance is considered unlikely. All endografts used in this study were
10 cm in length; consequently, no significant heating of
these devices is expected.
We used a rather simple model of AD in healthy, nonatherosclerotic swine. This and more complicated surgical
models cannot completely mimic the complex 3D anatomy
and tortuosity encountered clinically in patients with AD.
However, the expected benefits of detailed soft tissue
contrast and 3D representation by MRI may be even more
dramatic in these patients. In particular, safe traversal of
tortuous vascular access routes, one of the procedural
challenges of stent-grafting, might be simplified by using
rtMRI to visualise device-related anatomic distortion and
to guide operator adjustments.
Conclusions
The present experiments demonstrate that endovascular
stent-graft placement for experimental descending AD performed solely under rtMRI guidance is feasible. We used
entirely commercially available catheter devices for passive
visualisation with no hardware modifications to facilitate
MR imaging. MRI provided detailed pre-interventional
evaluation of the dissection, satisfactory monitoring of
stent-graft navigation, positioning, and successful deployment, as well as immediate anatomical and functional
confirmation of treatment success after stent-graft placement that are not available with conventional X-ray. These
advantages of MRI are complimentary to its potential advantage against the harmful effects of fluoroscopic guided
procedure including radiation exposure and nephrotoxicity
of contrast media.
620
Supplementary material
Supplementary material is available at European Heart
Journal online.
Acknowledgements
The authors thank Dr Peter Speier, PhD, of Siemens Medical
Solutions, Erlangen, Germany, for providing support with the interactive real-time ‘Works-in-Progress’ imaging protocols. We acknowledge W.L. Gore Inc., for providing stent-graft devices. H.E. was
supported by a research grant from the University Duisburg-Essen
(IFORES 10 þ 2).
Conflict of interest: none declared.
References
1. Eggebrecht H, Herold U, Kuhnt O, Schmermund A, Bartel T, Martini S,
Lind A, Naber CK, Kienbaum P, Kuhl H, Peters J, Jakob H, Erbel R,
Baumgart D. Endovascular stent-graft treatment of aortic dissection:
determinants of post-interventional outcome. Eur Heart J 2005;26:
489–497.
2. Nienaber CA, Fattori R, Lund G, Dieckmann C, Wolf W, von Kodolitsch Y,
Nicolas V, Pierangeli A. Nonsurgical reconstruction of thoracic aortic dissection by stent-graft placement. N Engl J Med 1999;340:1539–1545.
3. Rapezzi C, Rocchi G, Fattori R, Caldarera I, Ferlito M, Napoli G,
Pierangeli A, Branzi A. Usefulness of transesophageal echocardiographic
monitoring to improve the outcome of stent-graft treatment of thoracic
aortic aneurysms. Am J Cardiol 2001;87:315–319.
4. Moskowitz DM, Kahn RA, Konstadt SN, Mitty H, Hollier LH, Marin ML.
Intraoperative transoesophageal echocardiography as an adjuvant to
fluoroscopy during endovascular thoracic aortic repair. Eur J Vasc
Endovasc Surg 1999;17:22–27.
5. Koschyk DH, Meinertz T, Hofmann T, Kodolitsch YV, Dieckmann C, Wolf W,
Knap M, Kiraly Z, Steffen W, Beythien C, Lund G, Rehders T, Nienaber CA.
Value of intravascular ultrasound for endovascular stent-graft placement
in aortic dissection and aneurysm. J Card Surg 2003;18:471–477.
6. Mahnken AH, Chalabi K, Jalali F, Gunther RW, Buecker A. Magnetic
resonance-guided placement of aortic stents grafts: feasibility with
real-time magnetic resonance fluoroscopy. J Vasc Interv Radiol
2004;15:189–195.
7. Manke C, Nitz WR, Lenhart M, Volk M, Geissler A, Feuerbach S, Link J.
Magnetic resonance monitoring of stent deployment: in vitro evaluation
of different stent designs and stent delivery systems. Invest Radiol
2000;35:343–351.
8. Engellau L, Albrechtsson U, Dahlstrom N, Norgren L, Persson A, Larsson EM.
Measurements before endovascular repair of abdominal aortic aneurysms. MR imaging with MRA vs. angiography and CT. Acta Radiol 2003;
44:177–184.
9. Ayuso JR, de Caralt TM, Pages M, Riambau V, Ayuso C, Sanchez M, Real MI,
Montana X. MRA is useful as a follow-up technique after endovascular
repair of aortic aneurysms with nitinol endoprostheses. J Magn Reson
Imaging 2004;20:803–810.
10. Ersoy H, Jacobs P, Kent CK, Prince MR. Blood pool MR angiography
of aortic stent-graft endoleak. AJR Am J Roentgenol 2004;182:
1181–1186.
H. Eggebrecht et al.
11. Razavi MK, Nishimura E, Slonim S, Zeigler W, Kee S, Witherall HL, Semba
CP, Dake MD. Percutaneous creation of acute type-B aortic dissection: an
experimental model for endoluminal therapy. J Vasc Interv Radiol
1998;9:626–632.
12. Eggebrecht H, Zenge MO, Ladd ME, Erbel R, Quick HH. In-vitro evaluation
of current thoracic aortic stent-grafts for MR-guided placement.
J Endovasc Ther 2006; in press.
13. Buecker A, Adam GB, Neuerburg JM, Kinzel S, Glowinski A, Schaeffter T,
Rasche V, van Vaals JJ, Guenther RW. Simultaneous real-time visualization of the catheter tip and vascular anatomy for MR-guided PTA of
iliac arteries in an animal model. J Magn Reson Imaging 2002;
16:201–208.
14. Fink C, Bock M, Umathum R, Volz S, Zuehlsdorff S, Grobholz R,
Kauczor HU, Hallscheidt P. Renal embolization: feasibility of magnetic
resonance-guidance using active catheter tracking and intraarterial
magnetic resonance angiography. Invest Radiol 2004;39:111–119.
15. Buecker A, Spuentrup E, Grabitz R, Freudenthal F, Schaeffter T, van Vaals
JJ, Gunther W. Real-time-MR guidance for placement of a self-made fully
MR-compatible atrial septal occluder: in vitro test. Rofo 2002;174:
283–285.
16. Spuentrup E, Ruebben A, Schaeffter T, Manning WJ, Gunther RW,
Buecker A. Magnetic resonance–guided coronary artery stent placement
in a swine model. Circulation 2002;105:874–879.
17. Raval AN, Telep JD, Guttman MA, Ozturk C, Jones M, Thompson RB,
Wright VJ, Schenke WH, DeSilva R, Aviles RJ, Raman VK, Slack MC,
Lederman RJ. Real-time magnetic resonance imaging-guided stenting of
aortic coarctation with commercially available catheter devices in
swine. Circulation 2005;112:699–706.
18. Manke C, Nitz WR, Djavidani B, Strotzer M, Lenhart M, Volk M, Feuerbach S,
Link J. MR imaging-guided stent placement in iliac arterial stenoses: a
feasibility study. Radiology 2001;219:527–534.
19. Raman VK, Karmarkar PV, Guttman MA, Dick AJ, Peters DC, Ozturk C,
Pessanha BS, Thompson RB, Raval AN, DeSilva R, Aviles RJ, Atalar E,
McVeigh ER, Lederman RJ. Real-time magnetic resonance-guided endovascular repair of experimental abdominal aortic aneurysm in swine.
J Am Coll Cardiol 2005;45:2069–2077.
20. Quick HH, Kuehl H, Kaiser G, Hornscheidt D, Mikolajczyk KP, Aker S,
Debatin JF, Ladd ME. Interventional MRA using actively visualized catheters, TrueFISP, and real-time image fusion. Magn Reson Med 2003;49:
129–137.
21. Ladd ME, Quick HH, Debatin JF. Interventional MRA and intravascular
imaging. J Magn Reson Imaging 2000;12:534–546.
22. Quick HH, Zenge MO, Kuehl H, Kaiser G, Aker S, Massing S, Bosk S, Ladd
ME. Interventional magnetic resonance angiography with no strings
attached: wireless active catheter visualization. Magn Reson Med
2005;53:446–455.
23. Hilfiker PR, Quick HH, Debatin JF. Plain and covered stent-grafts: in vitro
evaluation of characteristics at three-dimensional MR angiography.
Radiology 1999;211:693–697.
24. Hilfiker PR, Quick HH, Schmidt M, Debatin JF. In vitro image characteristics of an abdominal aortic stent graft: CTA vs 3D MRA. MAGMA
1999;8:27–32.
25. Konings MK, Bartels LW, Smits HF, Bakker CJ. Heating around intravascular guidewires by resonating RF waves. J Magn Reson Imaging
2000;12:79–85.
26. Ladd ME, Quick HH. Reduction of resonant RF heating in intravascular
catheters using coaxial chokes. Magn Reson Med 2000;43:615–619.
27. Nitz WR, Oppelt A, Renz W, Manke C, Lenhart M, Link J. On the heating of
linear conductive structures as guide wires and catheters in interventional MRI. J Magn Reson Imaging 2001;13:105–114.