Advertisement
Advertisement
January 2010 Supplement
Three-Dimensional Imaging and Navigation
Three-dimensional CT reconstruction should be included throughout endovascular procedures, enabling better procedure planning and increasing patient satisfaction.
By Alan B. Lumsden, MD; Christof Karmonik, PHD; Miguel Valderabanno, MD; Dipan Shah, MD; and Jean Bismuth, MD
We live and work in a three-dimensional (3D) world but have learned to plan and perform endovascular procedures with two-dimensional (2D) systems. Further, we have learned to perform those procedures very well, to the point that some may question whether developing a 3D operating environment is worth the effort. However, 3D imaging can assist case planning and conduct in two modes: preoperative case planning and intraoperative performance of the case. For preoperative planning, 3D computed tomography (CT) reconstruction has been extensively validated:1 centerline length measurements, evaluating complex aortic necks and iliac tortuosity (Figure 1),2 measuring inner and outer diameters of the aortic arch, understanding complex anatomy in congenital aortic anomalies,3 and planning carotid stenting.4 This has now reached the realm of necessity: we expect and need to be able to view a 3D reconstructed image. It is in response to these demands that 3D Recon, a service from Medtronic Vascular (Santa Rosa, CA) in partnership with Vital Images (Minnetonka, MN), has evolved and provides 3D reconstruction viewing even if this is not available locally to all its customers. To enable 3D Recon from a remote location, the availability of DICOM (Digital Imaging and Communications in Medicine) image transfer becomes important. This too is available through Medtronic's CTeXpress (also in partnership with Vital Images), a service that allows for not only transfer of images but also the ability to utilize these 2D axial images to reconstruct in 3D remotely.
3D Recon provides additional important capabilities: remote viewing and consultation with referring physicians, diameter and centerline measurements, and electronic reports. Physicians can upload CT images to a Web site, where the axial data are converted to 3D using the 3D Recon service. The 3D reconstruction allows for aneurysm measurement that can be used for accurate sizing during endograft placement. Reliable image viewing is consistently accessible. Images can be viewed online before patient arrival in the clinic, resolving the constant irritation of lost discs or incompatible viewers. Frequently, CT scans must be repeated because the patient and images are not present at the same time. This increases radiation exposure for the patient and greatly increases cost. The presence of a 3D image enables better communication with the patient about the nature of the aneurysm and the planned procedure, increasing patient understanding and satisfaction. Similarly, although experienced interventionists have learned to use 2D axial images, 3D images make it easier for referring physicians to understand and further discuss a completed procedure with a patient, as well as explain the necessity for longitudinal follow-up.
Although 3D Recon does provide immediate added value, an additional and perhaps even greater role for 3D reconstruction is its use as a tool to facilitate endoluminal navigation and to understand complex flow patterns within areas of pathology. The following sections will explore these potential benefits. We believe that the future of endovascular intervention will involve use of 3D images to facilitate navigation.
CENTERLINE NAVIGATION
Currently, we talk about navigating a catheter through
a vascular bed. However, navigating a catheter consists of
a series of interactions between the vascular wall and the
catheter. We have remarkably little catheter control:
advancement and rotation. Literally, we bump a specially
shaped catheter off the vessel wall to move the tip
through tortuous anatomy or to enter a vessel orifice. It
is the catheter-wall interaction that generates complications:
embolization, occlusion, and dissection. With current
2D imaging and manual catheter manipulation, it is
impossible to avoid wall contact, and this premise has
not substantially changed for decades. However, centerline
navigation—the same centerline currently created in
3D Recon—fundamentally alters the way we work inside
the vascular system and could prevent damaging wall
interaction. The concept of centerline navigation may
also be thought of as “off-the-wall” navigation—deliberately
trying to minimize catheter-wall interaction until a
therapeutic intervention is delivered. Centerlines have
traditionally been used to improve accuracy in CT measurements
for endografting; let's now think of this as a
navigation strategy. The tools to permit centerline navigation
all exist; the challenge is in making them work in
harmony. Three-dimensional imaging is the imaging core
for such a strategy. Fusion technology and robotic navigation
in a 3D environment are not pipe dreams but currently
exist and are used daily in electrophysiology in an
environment that arguably is much more complex than
an aorta (Figure 2).
FLUORO CT SCANNING
As we replace older angiography suites, the new systems
add powerful imaging modalities, namely fluoro CT or DynaCT capabilities (Figure 3). These systems add a CT
scanner to our hybrid suite, which we never had before.
We believe that this will not only have an important role
in cardiovascular disease but will also revolutionize general
and thoracic surgery by providing en suite CT: CT-guided
biopsies, drainage of collections, or optimizing access port
locations. Imaging companies have developed new combined
angiography/CT suites, which use flat-panel detector
(FD) technology for improved resolution angiography
that is also able to produce improved cone-beam volume
CT images. The system permits 3D rotational digital subtraction
angiography (DSA) or cone-beam volume CT
interchangeably with the same FD C-arm so that patients
do not have to be transferred to a separate unit in order to
obtain both imaging modalities. Real-time feedback of
endovascular procedures is possible for both DSA and CT.
One of the most striking features of this technology is its
simplicity, which allows for efficient and fluid endovascular
procedures.
When comparing DynaCT to a 16-slice multidetector CT scanner (Somatom Sensation 16, Siemens Healthcare), Irie et al found that DynaCT was able to scan a wider area in a shorter period of time while delivering superior quality coronal and sagittal reconstruction images.5 DynaCT allows a contrast resolution of 10 HU (Hounsfield units) as well as a slice thickness and in-plane resolution of < 1 mm.6 It is also ideal if the system is able to boast better coverage, which can be a clear advantage when treating an obese patient but can also serve to decrease exposure times.
One of the concerns with this cone-beam technology is the amount of radiation exposure to the surgeon/interventionist and patient. It was found that the total radiation dose is 236 mGy for FD-based DynaCT, while the dose for 3D DSA using the same system is about 50 mGy.5 Other investigators revealed that the dose of radiation for a conventional head CT was similar to that of DynaCT, namely 60 mGy.7,8
One of the areas where DynaCT has the potential to garner the most advantages is as a navigational tool. As devices become more refined and are able to challenge more complex anatomy, DynaCT will be able to assist obtaining the 3D imaging necessary to situate and guide the instrument to its target. This can be a particularly attractive feature when one starts discussing potential applications for flexible robotics.
REGISTRATION OF AXIAL IMAGING
WITH ANGIOGRAPHY
The fusion of axial imaging and angiography in the interventional
suite will soon enable endovascular navigation in
3D-like perspective. Coregistration of multimodality imaging
in the angiographic suite overcomes some of the weakness
of each separate modality and accentuates the
strengths of both. Angiography provides 2D luminal contour
detail but does not provide extraluminal tissue information.
Combination technology that aligns 2D angiography angiography
with 3D images allows for better visualization of vessel
tortuosity and the relationship of the lumen to surrounding
structures. Because the vessels are 3D structures, visualizing
them in 3D during procedures is more intuitive.
Specialized software now allows for coregistration of the angiographic images with the reconstructed axial images. Initially, most applications for fused multimodality imaging were used for cardiac or intracranial interventions.9-11 In the future, multimodality imaging and processing will likely evolve to become standard tools of vascular specialists.
Multimodality image fusion can be achieved in different ways. One option for coregistration of fluoroscopy and axial imaging is real-time magnetic resonance imaging (MRI) or CT in the interventional suite. However, this currently requires specialized endovascular equipment and poses safety concerns for the treating clinicians. New angiography systems enable CT-like reconstruction of images using rotational angiography (Philips Allura Xper FD20/10 [Philips Healthcare, Andover, MA] and the Siemens zeego). Another option is aligning the diagnostic axial images with fluoroscopic images via externally placed fiducial markers or internal anatomic landmarks. This emerging technique combines reconstructed 3D axial CT or MR images with the 2D imaging in the angiography suite. In one example of bringing fusion imaging into the angiography suite, Gutierrez and colleagues performed both cardiac and peripheral vascular interventions using MRI coregistered to fluoroscopy.10 They placed external fiducial markers on patients before a baseline electrocardiography- gated MRI. The patients were subsequently moved to the angiographic table with the external markers still in place. Several baseline x-rays were acquired, and special software fused the MR and the fluoroscopic data using the markers to align the images. Although the authors found problems coregistering external markers on loose skin and abdominal fat, MRI road mapping proved feasible. Future systems under investigation may include such internal markers as bones.
Once the CT or MRI images are coregistered with the angiographic images, a real-time working overlay or road mapping can be projected for the treating clinician. New software to perform the coregistration with CT and MR is now becoming commercially available, but the quality of the fusion images remains unclear. One unsolved problem is the vessel deformity caused by stiff intraluminal wires, catheters, and devices. The deformity of the vessel causes a mismatch between the preoperative imaging and the live angiogram. However, as this exciting technique evolves, more precise and integrated images will become available.
DYNAMIC MRA
Although the resolution of 3D CT scans is optimal, MRI
has the added advantage of being capable of providing
additional physiologic data. We use dynamic 3D MR
reconstruction in all patients with aortic dissections, often
with computational fluid dynamics overlay (Figure 4).
Although computational simulations in general and of
blood flow in human arteries in particular have been the
topic of research in the last decades,11,12 only recently with
the introduction of advanced clinical imaging techniques
and progressed computing power has it been possible to
tailor these simulations toward the conditions found in a
particular individual.13-16 Initially, computational fluid
dynamics (CFD) simulations were restricted to 2D models
and idealized geometries. Solutions even for these simplified
geometries could only be obtained after many hours
or even days. Continuous technical advances, however,
have now made it possible to convert information from
images acquired during a routine clinical exam into 3D
complex mathematical meshes consisting of hundreds of
thousands of small-volume elements for transient simulation
of the hemodynamics in artery segments of the
human vasculature either in health or in disease.17 The
results of these simulations provide access to hemodynamic
parameters that are currently not reliably measurable
with clinical imaging methods. Arguably one of the
most important of these parameters is the wall shear
stress (WSS) that the flowing blood is exerting onto the
arterial wall. Wall shear is an important determinant of
dissection. Other parameters include dynamic pressures
(dynP) and recirculation patterns; the latter may facilitate
the adhesion of material onto the artery wall and promote
the creation of atherosclerotic lesions, for example,
in the bulb of the carotid bifurcation.
Hemodynamics may play an important role in type B aortic dissection (TB-AD). A recent flow study of a chronic TB-AD demonstrated a direct dependence of systolic and diastolic pressures in the true and false lumen on TB-AD morphology, emphasizing the need for a better understanding of hemodynamic forces in TBAD. 18,19 Toward this goal, we employed CFD simulations to investigate the feasibility of quantifying changes in hemodynamic parameters before and after thoracic endovascular aortic repair (TEVAR) of a type B aortic dissection. 20,21
3D IN THE MODERN ANGIOGRAPHIC OPERATING ROOM
Fusion of preoperative images with live angiography or
simultaneous 3D imaging allows for the delineation of the
vessel lumen centerline during interventions. Especially in
vessels with atherosclerosis or thrombus, manipulation of
endoluminal devices invariably leads to dislodgement of
plaque debris and clot. The catheter or device dragging
along the vessel wall could also cause injury such as dissection
or perforation. Although the clinical significance is
unknown, catheter manipulation in the atherosclerotic
aortic arch is associated with embolic signals on brain
ultrasound.22
Using computer-generated projections of the luminal centerline, could wires, catheters, and devices one day be steered away from the vessel wall? One current system uses a special ferromagnetic-tipped wire to navigate through blood vessels using magnets on either side of the patient. By altering the direction and strength of the magnets, the wire tip can navigate through tortuous vessels and tight stenoses while minimizing vessel wall contact (Axiom Artis dBC [Siemens] and Niobe Magnetic Navigation system [Stereotaxis, Inc., St. Louis, MO]). In addition, new catheter control systems remove the clinician from standing next to the radiation source or the magnet (Sensei Robotic Catheter Control System). Instead, the clinician is seated at a workstation with a controller, and a robotic arm takes the place of the clinician's hand next to the fluoroscopy unit (Figure 5). In addition to reducing radiation doses from decreased fluoroscopy time, this could result in fewer complications due to catheter and wire vessel wall injury (see Advancing Simulation sidebar).
CONCLUSION
CT scanning plays a central role in the diagnosis, treatment,
and follow-up of patients with abdominal aortic
aneurysms. CTeXpress facilitates the electronic exchange
of the 2D images and provides an online image repository.
Furthermore, with 3D Recon, this 2D image database
permits postprocessing of the images to provide 3D rendering,
diameter and length measurements, the key tools
necessary for planning endograft placement. We believe
3D reconstruction will increasingly be used to guide
endovascular procedures.
Alan B. Lumsden, MD, is Professor and Chairman of Cardiovascular Surgery, and Medical Director at Methodist DeBakey Heart & Vascular Center in Houston, Texas. Dr. Lumsden may be reached at (713) 441-6201; ablumsden@tmhs.org.
Christof Karmonik, PhD, is a research scientist at The Methodist Hospital Neurological Institute and DeBakey Heart & Vascular Center in Houston, Texas.
Miguel Valderabanno, MD, is Associate Professor of Medicine at Weill Medical College of Cornell University, and Director, Division of Cardiac Electrophysiology, Department of Cardiology, at Methodist DeBakey Heart & Vascular Center in Houston, Texas.
Dipan Shah, MD, is Assistant Professor of Medicine at Weill Medical College of Cornell University, and Director, Cardiac Magnetic Resonance Imaging at Methodist DeBakey Heart & Vascular Center in Houston, Texas.
Jean Bismuth, MD, is Assistant Professor of Medicine at Methodist DeBakey Heart & Vascular Center in Houston, Texas.
Advertisement
Advertisement
