Download Three-Dimensional Coronary Angiography

COVER STORY
Three-Dimensional
Coronary
Angiography
The justification, technology, applications, and barriers
to three-dimensional coronary imaging.
BY EUGENIA P. CARROLL, MD, AND JOHN D. CARROLL, MD
C
oronary angiography is changing in several fundamental ways. The ability to visualize the
human coronary arteries in a three-dimensional (3D) format, a central new functionality, will
soon be a reality in routine patient care. The traditional
two-dimensional (2D) projection image that is ubiquitous in clinical practice will not disappear; rather image
processing to render coronary arteries in 3D and
advanced analysis tools will complement the images of
conventional angiography. This article will examine the
clinical reasons that advances in coronary angiography
are needed, discuss some of the central technical themes,
and provide an overview of what is now available and
what is coming to the clinical arena soon. Finally, change
is not immediate, and the barriers to profound changes
in this field due to advancements in technology will be
discussed.
During the last few decades, progress has occurred in
catheter-based coronary angiography including the transition to standardized digital format and the replacement
of image intensifiers with flat detectors. These changes
have enabled the next major transition, which primarily
involves the processing of these digital files into more
realistic image representations, unique quantitative
analysis, and decision-assistance tools. These advancements will allow the interventionist to understand anatomy better, quantify key features needed for percutaneous
coronary intervention (PCI), and work more efficiently.
The transition to 3D is a major part of this next evolution
in coronary angiography, as has been the transition of 3D
in other modalities, such as computed tomography (CT)
angiography, magnetic resonance angiography, and ultrasound.
38 I CARDIAC INTERVENTIONS TODAY I APRIL/MAY 2009
Figure 1. This is an example of a coronary 3D model created
from two angiographic images. Not shown here are the
other models created over the entire cardiac cycle. An analysis of changes in coronary shape during underlying heart
motion has been done. The areas noted in blue and green
indicate regions in this right coronary artery where curvature
increases and decreases during the cardiac cycle. These are
also called flexion points.
Cardiologists are becoming used to viewing the coronary tree in a 3D format that is routinely available with
CT angiography. They have a new appreciation of the
curvilinear nature of the coronaries, the complexities of
lesions, and the unusual nature of x-ray projection
images that flatten the 3D shape.
COVER STORY
Figure 2. Two examples of left coronary artery reconstructions are shown. The left shows an older version of the reconstruction
algorithm; the right panel shows a different coronary artery reconstruction with reduced noise and artifacts.
The imaging skills of an experienced interventionist are
central to the performance of accurate diagnostic studies
and in guiding the intervention by efficiently and competently using the major modality of x-ray imaging. These
skills are being further put to the test with the movement of PCI into more complex subsets of coronary
artery disease. PCI is a mature therapeutic technique;
many interventionists have years of experience in using
traditional angiography, and the vast majority of interventions are successful and uncomplicated. Are there
really any significant remaining weaknesses in PCI related
to 2D projection images that justify improving the technology of imaging with 3D?
WHY 3D?
Inherent limitations in traditional angiography that play
out in specific ways in the world of PCI involving coronary
stenting1 are summarized in Tables 1 and 2. If the coronary
arterial tree could be presented to the interventionist in
its true 3D shape and if there were quantitative tools to
extract key 3D features, then coronary imaging, PCI planning, and PCI execution could potentially be improved in
multiple ways. Furthermore, the techniques of image
acquisition with traditional angiography are not standardized, but are subjectively chosen and highly dependent on
the 3D visual skills of individual operators working in the
bizarre, 2D, shadow-image format of projection.
Technological advancements should address clinical
problems that interventional cardiologists agree are still
present today. Reduction of radiation is an overriding
patient safety issue and is also relevant to those who
perform the procedure. Contrast volume directly
impacts the occurrence of contrast nephropathy, volume
overload, and the need to stage complex PCI. Prolonged
and complex PCI procedures also have a different complication profile. Enhancing workflow and task completion have implications on the safety and efficiency of the
procedure room; we acknowledge that 5% to 10% of PCIs
are especially long and difficult. Other potential benefits
of technology advancement involve providing the operator with greater confidence in decision making and task
performance before, during, and after PCI. Because the
transition to 3D is central to technological advancement,
the next issue to be addressed is how a 3D coronary
image can be produced.
H OW TO PRODUCE 3D IM AGE S
Projection images can be transformed into 3D models
and volumetric 3D reconstructions in a matter of minutes. The modeling approach and the reconstruction
technique have many differences and will be described
separately.
Coronary Modeling
The 3D modeling technique has been under development for several decades and, in the last decade, modeling
has been introduced in the form of several products. Our
colleague, James Chen, PhD, has been the pioneer in this
field.2-5 Most approaches use two or more angiographic
projections (Figure 1). The second step is called image
segmentation and involves the extraction of the centerline and diameters of all vessels to be included in the
model. Next, the transformation-defining relative location and orientation of the two views is calculated. From
this transformation, the skeleton of the coronary arterial
tree is created and computer graphics are used to display
the tree as a 3D representation. The model can be
manipulated (ie, rotated), allowing viewing that simulates
APRIL/MAY 2009 I CARDIAC INTERVENTIONS TODAY I 39
COVER STORY
TABLE 1. GENERAL IMAGING-RELATED
LIMITATIONS OF CATHETER-BASED
CORONARY ANGIOGRAPHY
TABLE 2. SPECIFIC LIMITATIONS OF
TRADITIONAL CORONARY ANGIOGRAPHY
RELEVANT TO CORONARY STENTING
• Ionizing radiation: Substantial radiation given with both
diagnostic and PCI, especially with prolonged and difficult
procedures. Techniques and technologies to reduce radiation have a high priority.
• Failure to reach the lesion due to underestimation of factors
determining resistance for delivery system advancement,
such as the degree of proximal vessel tortuosity.
• Contrast media: Renal toxicity, especially with pre-existent
renal impairment, remains a major and frequent complication demanding efforts to limit volume.
• Over-reliance on the trial and error system rather than
advanced analysis predicting delivery success versus failure.
• Technique of acquisition: Operator dependency on quality
of study via selection of a variable number of fixed projection views to avoid vessel overlap and minimize lesion foreshortening.
• Geographic miss is common in stent placement due to use
of views producing suboptimal 2D projection images.
• Undersizing and oversizing of stent lengths and diameters
based on “eyeball” estimates from suboptimal 2D images.
• Image format: 2D projection images with inherent problems
of foreshortening, overlap, and misrepresentation of key
anatomical features, such as lesion length, bifurcation angles,
and tortuosity.
• Limited conformability of stents placed in underappreciated
curved segments leading to vessel straightening and edge
kinks.
• Angiographic visualization: Visualization of lumen leads to
underestimation of disease and poor understanding of vessel wall.
• Limited assessment of stent expansion and wall apposition
due, in part, to poor visibility of stents with thinner struts
and larger patients.
• Interpretation and analysis: The subjective nature and semiquantitative analysis using only those views acquired may
lead to errors, especially during earlier years of experience.
• Inability to visualize all but the most gross examples of stent
fracture.
any gantry position. There are three major commercially
available systems that perform coronary modeling: the
CardiOp-B package (Paieon Medical, Rosh Ha’ayin, Israel);
the 3D-CA package (Philips Medical Systems, Bothell, WA);
and the Cardiovascular Angiography Analysis System for
3D Quantitative Coronary Analysis (Pie Medical Imaging,
Maastricht, The Netherlands).6 Dr. Chen’s system has the
most advanced quantitative tools but is not commercially available; it is being used internally in projects to better
quantify 3D and four-dimensional vascular properties in
many locations (arterial and venous) and the effect of
implantable devices.7
a volumetric representation of the vascular tree.8-10 The
algorithm for performing this reconstruction has been
improved with reduced noise and artifacts as shown in
Figure 2. The resulting CT-like datasets can be rendered
on the screen in various ways. Volume rendering is one
visualization technique to display the entire 3D dataset.
Other techniques can be applied, such as surface renderings, maximum intensity projections, and planar
reformats, which allow inspection of the data much as
in CT angiography. Similar to a model, the reconstructed arterial tree can be free-rotated to simulate any vantage point.
Once the interventionist has either a 3D model or a
reconstruction, how can it be used for clinical tasks? The
initial examination of the 3D model or reconstruction by
turning it on a computer screen allows an enhanced
appreciation of 3D geometry. However, can there be clinically important functions and features extracted and
quantified from the data or is it just a nice image with no
practical value?
Coronary Reconstruction
The reconstruction technique is a newer approach
that has not yet been commercially released for coronary applications. Reconstruction is in many ways a
more advanced technology that has multiple advantages over the current modeling techniques, one being
that it is completely automatic without the need for
user interaction to produce a model. Three-dimensional volumetric datasets are typically generated not from
traditional fixed views but from rotational angiography.
These 180º or greater arcs of angiographic projection
images are automatically processed for the creation of
40 I CARDIAC INTERVENTIONS TODAY I APRIL/MAY 2009
Computer Assistance for the Operator: The Coronary
Optimal View Map
Three-dimensional modeling and 3D reconstructions can
provide the datasets that allow development of an optimal
COVER STORY
view map.11,12 The map has
gantry space as the coordinates in
the right anterior oblique/left
anterior oblique and cranial/caudal axes. This optimal view map
can be generated for each coronary segment and provides foreshortening and overlap quantification for all segments for all
gantry positions. Color coding is
used to make it easy to find
regions where a gantry position
would be optimal to minimize
foreshortening of the segment
and avoid overlap with other vessels (Figure 3).
Figure 3. The 3D model of this right coronary artery has been analyzed to determine the
optimal gantry position for visualization of the posterior descending artery bifurcation
region with no overlap and no significant foreshortening. The model has been placed in
a very steep cranial position that has only 4% foreshortening of the bifurcation. The map
colors correspond to the amount of foreshortening in all potential gantry positions. The
red check marks the gantry location on the map that produces this view of the model.
Maps allow prediction of the appearance of coronary angiography and their use should
decrease the radiation dose and contrast volume from the normal process of searching
for a good view through the trial and error technique.
3D Quantitative Coronary
Angiography
Quantifying 3D vessel properties and characteristics of a given
vessel is increasing in contemporary interventional cardiology, as
shown by the development of the
SYNTAX score for selection of appropriate candidates for
PCI with left main and multivessel disease.13 The American
College of Cardiology and the American Heart Association
outline the specific vessel properties that define the procedure outcome risk of a given lesion.14 It is important to
note that these vessel characteristics are inherently 3D in
nature, therefore, a complete analysis requires a 3D evaluation. Quantifications of 3D vessel geometric features,
motion dynamics, and deformation analysis are now possible with 3D coronary images (Figure 1).15-19
BAR R IER S TO 3D
These changes in x-ray coronary angiography represent
a major culture change for the interventional cardiologist, and thus, implementation will not happen
overnight. The barriers to rapid and widespread adoption
of these technologies are multiple, as outlined in Table 3.
New angiographic systems are needed, and the capital
costs are significant. Validation and proof of clinical
value, with improvement in patient outcomes, are also
needed; a few reports have been published.20,21 Unlike
implanted device regulation, imaging devices are
approved without demonstration of improved clinical
outcomes. Finally, education and training may be minimal in using some new technologies, such as flat detectors, but 3D imaging involves new modes of image acquisition, processing, and, most importantly, a new way of
using images.
TABLE 3. BARRIERS TO ADAPTATION OF
3D CORONARY ANGIOGRAPHY
• Evidence-based medicine: Data are needed on how
outcomes are improved.
• Culture and workflow changes: Resistance to change
in long-standing practices is huge. There is an unwillingness
to examine basic assumptions and acknowledge limitations
of current techniques and technologies.
• Training of physicians and staff: There is major effort and
expense related to changing the method of acquisition (ie,
rotational angiography), incorporating processed images
into decision-making, and using new tools.
• Financial: New systems are purchased by many hospitals
only every 5 to 10 years, which slows the acquisition of new
technologies.
CONCLUSI ON
Three-dimensional coronary imaging is emerging from
the research lab into the clinical arena, driven by the
acknowledgment of the limitations of traditional 2D techniques, enabled by many aspects of the digital processing
revolution, and accelerated by the clinical challenges of
more complex anatomical subsets undergoing PCI. As
already demonstrated with rotational angiographic acquisition, 3D coronary imaging should further improve
patient outcomes, starting with reduction of x-ray dose
APRIL/MAY 2009 I CARDIAC INTERVENTIONS TODAY I 41
COVER STORY
and contrast volume, but also including reducing mistakes and errors caused by suboptimal imaging skills and
technology. ■
Eugenia P. Carroll, MD, is Assistant Professor of Medicine,
Division of Cardiology, University of Colorado Denver in
Aurora, Colorado. She has disclosed that she holds no
financial interest in any product or manufacturer herein.
Dr. Carroll may be reached at (720) 848-6508; [email protected].
John D. Carroll, MD, is Professor of Medicine, Division of
Cardiology, University of Colorado Denver in Aurora,
Colorado. He has disclosed that he is a coinventor of
patented 3D modeling and analysis software that is
assigned to the University of Colorado and University of
Chicago. He has also disclosed that he receives research
support, honoraria, and royalties from Philips. Dr. Carroll
may be reached at [email protected].
1. Green NE, Chen SY, Hansgen AR, et al. Angiographic views used for percutaneous
coronary interventions: a three-dimensional analysis of physician-determined vs. computer-generated views. Cathet Cardiovasc Interv. 2005;64:451-459.
2. Chen SY, Metz CE. Improved determination of biplane imaging geometry from two projection images and its application to three-dimensional reconstruction of coronary arterial
trees. Med Phys. 1997;24:633-654.
3. Chen SY, Carroll JD, Messenger JC. Quantitative analysis of reconstructed 3D coronary arterial tree and intracoronary devices. IEEE Trans Med Imag. 2002;21:724-740.
4. Chen SJ, Hoffmann KR, Carroll JD. Three-dimensional reconstruction of coronary arterial tree based on biplane angiograms. SPIE. Med Imag. 1996;2710:103-114.
5. Chen SJ, Carroll JD. 3D reconstruction of coronary arterial tree to optimize angiographic visualization. IEEE Trans Med Imag. 2000;19:318-336.
6. Garcia J, Chen S, Hansgen A, et al. Rotational angiography (RA) and three-dimensional imaging (3DRA): an available clinical tool. Int J Cardiovasc Imag. 2007;23:9-13.
7. Liao R, Chen SY, Green NE, et al. Three dimensional analysis of in vivo coronary-coronary vessel interactions. Int J Cardiovasc Imag. 2004;20:305-313.
8. Jandt U, Schfer D, Rasche V, et al. Automatic generation of time resolved motion
vector fields of coronary arteries and 4D surface extraction using rotational X-ray angiography. Phys Med Biol. 2009;54:47-66.
9. Hansis E, Schfer D, DЪssel O, et al. Projection based motion compensation for
gated coronary artery reconstruction from rotational X-ray angiograms. Phys Med Biol.
2008;53:3807-3820.
10. Blondel C, Malandain G, Vaillant R, et al. Reconstruction of coronary arteries from a
single rotational x-ray projection sequence. IEEE Trans Med Imag. 2006;25:653-663.
11. Chen SYJ, Carroll JD. On-line 3D reconstruction of coronary arterial tree for optimization of visualization strategy using a single-plane imaging system. IEEE Trans Med Imag.
2000;19:318-336.
12. Garcia JA. Optimal angiographic views based on 3D reconstructed models. J Am Coll
Cardiol. 2007;49(suppl B):296A.
13. Sianos G, Morel M, Kappetein A, et al. The SYNTAX Score: an angiographic tool
grading the complexity of coronary artery disease. Euro Interv. 2005;1:219-227.
14. Scanlon PJ, Faxon DP, Audet AM, et al. ACC/AHA guidelines for coronary angiography. A report of the American College of Cardiology/American Heart Association Task Force
on practice guidelines (Committee on Coronary Angiography). Developed in collaboration
with the Society for Cardiac Angiography and Interventions. J Am Coll Cardiol.
1999;33:1756-1824.
15. Messenger JC, Chen SY, Carroll JD, et al. 3D coronary reconstruction from routine
single-plane coronary angiograms: clinical validation and quantitative analysis of the right
coronary artery in 100 patients. Int J Card Imag. 2000;16:413-427.
16. Liao R, Messenger JC, Chen SYJ, et al. Four-dimensional analysis of cyclic changes
in coronary artery shape. Cathet Cardiovasc Interv. 2002;55:344-354.
17. Ding Z, Friedman MH. Quantification of 3D coronary arterial motion using clinical
biplane cineangiograms. Int J Card Imag. 2000;16:331-346.
18. Gross MF, Friedman MH. Dynamics of coronary artery curvature obtained from
biplane cineangiograms. J Biomech. 1998;31:479-484.
19. Chen SYJ, Carroll JD. Kinematic and deformation analysis of 4-D coronary arterial
trees reconstructed from cine angiograms. IEEE Trans Med Imag. 2003;22:710-721.
20. Maddux J, Wink O, Messenger J, et al. Randomized study of the safety and clinical
utility of rotational angiography versus standard angiography in the diagnosis of coronary
artery disease. Cathet Cardiovasc Interv. 2004;62:176-174.
21. Agostoni P, Biondi-Zoccai G, Van Langenhove G, et al. Comparison of assessment of
native coronary arteries by standard versus three-dimensional coronary angiography. Am
J Cardiol. 2008;102:272-279.
42 I CARDIAC INTERVENTIONS TODAY I APRIL/MAY 2009