Author + information
- Received March 4, 2011
- Revision received May 20, 2011
- Accepted June 3, 2011
- Published online August 1, 2011.
- Vasim Farooq, MBChB,
- Patrick W. Serruys, MD, PhD⁎ (, )
- Jung Ho Heo, MD,
- Bill D. Gogas, MD,
- Takayuki Okamura, MD, PhD,
- Josep Gomez-Lara, MD,
- Salvatore Brugaletta, MD,
- Hector M. Garcìa-Garcìa, MD, MSc, PhD and
- Robert Jan van Geuns, MD, PhD
- ↵⁎Reprint requests and correspondence
: Dr. Patrick W. Serruys, Interventional Cardiology Department, ThoraxCenter, Erasmus University Medical Centre,'s-Gravendijkwal 230, 3015 CE Rotterdam, the Netherlands
Coronary artery bifurcations are a common challenging lesion subset accounting for approximately 10% to 20% of all percutaneous coronary interventions. The provisional T-stenting approach is generally recommended as the first-line management of most lesions. Carina shift is suggested to be the predominant mechanism of side-branch pinching during provisional T-stenting and has been indirectly inferred from bench work and other intravascular imaging modalities. Offline 3-dimensional (3D) reconstructions of patients studied in the first-in-man trial of the high-frequency (160 frames/s) Terumo optical frequency domain imaging system were undertaken using volume-rendering software. Through a series of 3D reconstructions, several novel hypothesis-generating concepts are presented.
Coronary bifurcations are a challenging lesion subset accounting for approximately 10% to 20% of all percutaneous coronary interventions (PCI). Historically, they have been associated with lower rates of procedural success, higher restenosis rates, in particular at the ostium, and adverse events compared with the treatment of simpler, nonbifurcation lesions (1–3).
The current prevailing opinion in their management is one of a “simpler is better” approach, with provisional T-stenting recommended as the first-line strategy in most lesions (1,2). The traditional concept of plaque shift as the predominant mechanism in the pinching of the side branch (SideB) during this technique has recently been challenged and replaced by carina shift, with the suggestion that up to 30% of bifurcations have involvement of plaque shift, since atheroma is rarely seen at the carina alone because of it being a high wall shear stress area (2,4). The concept of carina shift has been indirectly inferred from bench work, in vivo longitudinal and cross-sectional intravascular ultrasound intravascular (IVUS) imaging, and computed tomography (CT) imaging modalities (1,4–8).
Both IVUS (image resolution: 100 to 150 μm) and CT (resolution: 300 to 500 μm) lack the imaging resolution to fully appreciate the complex architecture of the bifurcation compared with optical coherence tomography (OCT; image resolution: 10 to 20 μm). Fusion of CT and IVUS in obtaining 3D reconstructions of human coronary bifurcation have previously successfully been undertaken to allow for wall shear stress analyses; this system, however, lacks the resolution obtainable with OCT (9).
Current-generation Fourier-domain OCT (FD-OCT) allows rapid pullback speeds and has allowed visualization of the coronary bifurcation with 2-dimensional (2D) images in great detail (10–12). One of the major limitations of this technology, however, has been the lack of 3-dimensional (3D) images. Trying to mentally reconstruct a complex 3D structure from 2D images is difficult; 2D imaging may not allow a full appreciation of the anatomic features of the bifurcation and the effects of PCI.
Three-dimensional, offline FD-OCT reconstructions were first described by Tearney et al. (13) and more recently by Okamura et al. (14) in the assessment of jailed SideB by bioresorbable vascular scaffolds.
Prototypes of current-generation “real-time” (i.e., periprocedural) 3D FD-OCT are experimental, have not yet entered conventional clinical practice, and appear to have a limited resolution compared with the offline 3D optical frequency domain imaging (OFDI) reconstructions; this may be related to the lower frame rate of commercially available systems: the intracoronary Terumo OFDI system (Terumo Corporation, Tokyo, Japan) is currently the only system with a frame rate as high as 160 frames/s (11,14,15).
Through a series of images demonstrating 3D OFDI reconstructions of coronary artery bifurcations, we aim to demonstrate several novel, hypothesis-generating concepts with regard to the anatomic characteristics of this complex structure.
Three-dimensional reconstructions of patients who underwent conventional PCI from the original first-in-man study of the intracoronary Terumo OFDI system were undertaken (15). The high-speed Terumo OFDI system is capable of acquiring 160 frames/s during the catheter pullback up to a maximum speed of 40 mm/s; all images were acquired with a pullback speed of 20 mm/s. The 3D coronary angiography images were constructed from their respective biplane 2D images (CAAS 5.9, Pie Medical Imaging, Maastricht, the Netherlands) (16). All patients studied had stable angina; 3D reconstructions of coronary bifurcations and the consequent effects of provisional T-stenting were performed.
The methodology for the 3D reconstructions has previously been described (14). In brief, offline bitmap sequences (704 × 704 pixels) were generated from prior 2D OFDI imaging. Manual detection of every strut in each 2D cross section were undertaken, and 3D reconstructions of the coronary vessel pre- and post-intervention were performed using volume-rendering software (INTAGE Realia, KGT, Tokyo, Japan).
Nomenclature for 3D FD-OCT reconstructions
“Fly-through” views indicate a selected still image of an internal view of a vessel looking either downstream (proximal-to-distal vessel) or upstream (distal-to-proximal vessel). An “orientation” figure is located as an inset figure within the 3D reconstruction to best illustrate where the endoluminal point of view is electronically located, and in which direction it is pointing. In 3D reconstructions, the 3D rendering software provides x-, y-, and z-axes within the coronary vessel to allow precise assessment of the location of the endoluminal point of view. Longitudinal and, in some cases, cross-sectional 2D OFDI frames of the vessel and bifurcation, with a blue arrow superimposed on it, are used to orient the reader within the vessel: the base and direction of the blue arrow indicates from and in which direction, respectively, the 3D image is visualized from within the 2D plane(s).
Anatomy of the left main stem, circumflex, left anterior descending, diagonal, and septal branches
Figure 1A demonstrates a fly-through view, looking downstream, from the distal left main stem showing the ostia of the left anterior descending coronary artery (LAD), circumflex, first diagonal (Fig. 1A, D1), and septal branches. Note the appearance of the opening of the diagonal vessel and its relationship to the LAD vessel opening. The main branch (MainB) and SideB appear to diverge parallel to each other at their respective origins, with the carina (labeled) appearing to be “interposed” between both vessel openings at this point of divergence.
Figure 1B demonstrates a close-up fly-through view of the same vessel looking downstream from the proximal LAD and aimed towards the diagonal ostium (upper right image); the left circumflex coronary artery (LCx) orifice appears between 12 and 3 o'clock in relation to the diagonal ostium. For comparison, note the corresponding 2D OFDI frames (lower image). The slit-like, elliptical appearance of the diagonal opening is clearly visible, and when bifurcation is viewed perpendicular to the vessel wall (inset left image—use the orientation figure to allow assessment of the endoluminal point of view), the carina is predominantly visualized, with the proximal course of the diagonal vessel appearing to be hidden behind the rim of the carina so that the diagonal orifice appears, in-depth, as a dead end. This is further suggestive of the proximal parallel course of the diagonal with the LAD at the point of divergence. The yellow arrows in the fly-through and perpendicular views are pointing in identical directions, namely, the direction of the opening of the diagonal vessel.
Figure 1C demonstrates a fly-through view looking further downstream into the same LAD (upper left image). Observe the large septal branch orifice opening; when viewed perpendicular to the vessel wall (upper right image), and contrary to the observations made with the diagonal branch, the endoluminal opening of the septal branch is entirely visible and does not appear to be concealed behind the carina. This corresponds to the characteristic, almost perpendicular takeoff of the septal branch from the LAD. Corresponding 2D OFDI frames are displayed below the panel.
Anatomy of the proximal, mid, and distal LAD–diagonal branches
Figure 2A demonstrates downstream fly-through views of the LAD, looking distally (Image 1) and aimed towards the proximal diagonal ostium (Image 2). When the bifurcation is visualized perpendicular to the vessel wall (Image 3), the diagonal orifice appears to have a circular appearance with the rim of the carina appearing to be concealing the proximal course of diagonal, so that the diagonal orifice once again appears, in-depth, as a dead end. Yellow arrows (Images 2 and 3) point in identical directions at the opening of the same diagonal vessel. Corresponding 2D OFDI frames are shown below the panel for comparison.
Figure 2B demonstrates fly-through views (looking downstream) of the mid (D2, left image) and distal (D3, right image) LAD–diagonal bifurcations; yellow arrows point at the diagonal vessel openings. The corresponding 2D OFDI frames of each bifurcation are displayed below their respective 3D reconstructions.
All of these images are further suggestive of the proximal parallel course of the diagonal vessels relative to the LAD at their point of divergence.
Anatomy of diagonal branch originating perpendicular to the LAD
Figure 3 demonstrates a coronary angiogram (left image), suggesting the diagonal vessel (asterisk) originates almost perpendicular to the LAD at the point of divergence of both vessel origins. The 3D coronary angiography images (left inset figures) confirmed these findings with a bifurcation angle of 85°. A fly-through view of the LAD looking downstream (upper right image) demonstrates the diagonal vessel opening (asterisk), note the elliptical shape of the vessel opening and the observation that the diagonal vessel opening is fully visible, and not concealed by the carina, when visualized perpendicular to the vessel wall (lower right image). Observe how the stent is able to divide the opening of the diagonal vessel into at least 3 segments. Effectively, the diagonal branch with an almost perpendicular takeoff appears to have similar characteristics to the septal branch as described in Figure 1C.
The right ventricular branch of the right coronary artery
Figure 4 demonstrates how the right ventricular (RV) branch of the right coronary artery (RCA) appears to exhibit the phenomenon of a “parallel bifurcation” as described for the diagonal vessel. The downstream fly-through view of the proximal RCA demonstrates the opening of the RV branch (upper left image); note how the opening of the RV branch (yellow arrows) appears to be concealed behind the proximal rim of the carina in the perpendicular (upper middle image) and retrograde (upper right image) endoluminal point of views. In support of this concept is that the bifurcation angle is 57° on the corresponding 2D and 3D coronary angiograms (lower left images). Corresponding 2D OFDI frames are displayed (lower right images).
Bifurcation treated with provisional-T approach: comparing the diagonal and septal branches on 2D and 3D FD-OCT
Figure 5A demonstrates a long segment of disease in the proximal-mid LAD (upper left image) extending across the bifurcation with a diagonal branch. This was treated with 2 overlapping Xience V stents (Abbott Vascular, Abbott Park, Illinois) that lead to “pinching” of the SideB ostium with Thrombolysis In Myocardial Infarction flow grade 3 (middle left image with corresponding 3D coronary angiogram). Kissing balloon post-dilation (KBPD) was subsequently successfully performed with improvements in the pinched angiographic appearances (lower left image with corresponding 3D coronary angiograms) (Online Video 1). In the fly-through view (looking downstream) of the LAD post-intervention (right image), note the characteristic perpendicular and parallel takeoffs of the septal (asterisk) and diagonal (white arrow) branches at the point of divergence from the LAD lumen, respectively.
Figure 5B demonstrates fly-through views looking downstream at the diagonal and septal branches; note the differences between both branches as a more perpendicular view of the vessel wall is seen. Parallel yellow arrows represent the parallel courses of the diagonal and LAD vessels at their point of divergence. Medina et al. (6) have previously hypothesized the concept of carina shift pinching the SideB by the displacement of the carina so that it appears as an “eye-brow” sign on longitudinal 2D IVUS imaging (17). As shown on the 3D reconstruction, this effect can be more easily appreciated and may be hypothesized to have led to the near closure of the diagonal ostium following post-dilation of the MainB stent; subsequent KBPD may have reopened the SideB ostium by displacing the carina towards the lumen of the MainB. This principle may not be applicable to the septal branch because of the differing appearances of the carina.
Figure 5C demonstrates the corresponding 2D OFDI frames of the septal (asterisks) and diagonal branches (arrows). The almost parallel course of the diagonal vessel (yellow arrow indicates the diagonal vessel in cross-sectional view) in relation to the LAD lumen at the point of divergence appears to determine the 2D FD-OCT characteristics; without careful observation, the diagonal vessel may have been misinterpreted as an area of stent malapposition on the 2D FD-OCT imaging. With the septal branch, this appears to originate perpendicular to the vessel wall giving its characteristic 2D FD-OCT appearances as shown (yellow asterisk).
Bifurcation treated with provisional-T approach: the concept of a parallel bifurcation and carina shift
Figure 6A demonstrates the concept of carina shift with pre- and post-intervention images: the solid line indicates the position of the carina pre-intervention (left image); broken lines demonstrate the position of the carina pre- (left image) and post- (right image) intervention. Two-dimensional (Online Video 2) and 3D coronary angiograms (inset images) demonstrate disease of the LAD with involvement of the diagonal ostium not evident on subsequent 3D reconstruction.
Implantation of a Xience V stent (Abbott Vascular) in the mid LAD led to angiographic “pinching” of the SideB ostium with Thrombolysis In Myocardial Infarction flow grade 3 maintained in the SideB. Post-dilation of the MainB alone was performed with the deploying stent balloon; no KBPD was performed. The angiographic appearance of the pinching of SideB ostium improved (Online Video 2).
Figure 6B (upper images) demonstrates the appearance of the SideB opening when viewed perpendicular to the vessel wall (Image 3), illustrating the principle of a “parallel bifurcation” with the concealment of most of the opening of the diagonal vessel by the rim of the carina. It may be speculated that continued post-dilation of the MainB stent alone, especially with larger balloons, may have risked further carina shift and SideB closure.
Note how the MainB struts appear to overhang the carina over the SideB opening, especially evident in the perpendicular view (Image 3), because of the structure of the parallel bifurcation and carina. For comparison, the 2D OFDI frames (lower images) with the stent struts at the SideB ostium are illustrated. Asterisk indicates thrombus and the shadow it casts on the vessel wall, in both the 2D and 3D FD-OCT imaging.
On the basis of the offline 3D OFDI reconstructions, we have proposed several hypothesis-generating concepts relating to bifurcation anatomy and the effects of PCI: 1) the concept of the parallel or perpendicular bifurcations and the corresponding 2D and 3D appearances—a reassessment of current interpretations of the 2D FD-OCT imaging of the bifurcation may be warranted; 2) the hypothesis that the angle of divergence between the MainB and SideB at their respective origins will ultimately determine the anatomic features of the carina and how it potentially interacts with the SideB vessel orifice during MainB stenting; and 3) the potential clinical application of instantaneous 3D FD-OCT in coronary bifurcation treatment.
Based on the appearances of the carina on the 3D reconstructions and the effects of PCI, a more perpendicular takeoff of the SideB from the MainB may be less prone to the effects of carina shift whereas a shallower angle of divergence appears to be more susceptible to the effects of carina shift. This concept is supported by a study suggesting a specific measure of SideB angulation on the coronary angiogram is associated with a higher incidence of SideB compromise (18). Asakaura et al. (19) also demonstrated, in a small case series, that a shallower angle, with a cutoff of 80°, between the LCx and LAD was predictive of LCx ostial impairment after stenting within the LAD ostial region—the authors had presumed at the time that the mechanism of SideB (LCx) closure was secondary to plaque shift or coronary dissection. Both of the aforementioned studies are, however, limited by a lack of 3D quantitative coronary angiography to calculate the bifurcation angulation (16). An awareness of the potential increased risk of SideB closure during MainB stenting of coronary bifurcations with a shallower bifurcation angle may be justified—further study into this phenomenon is required.
Angiographic appearances of the SideB ostium after MainB stenting has previously been demonstrated to be unreliable, with only a quarter (27%) of cases with a residual angiographic narrowing of ≥75% in the SideB being found to have a functionally significant narrowing in pressure wire studies (1,8). The main limitation of this technique is the potential risk of dissecting the SideB ostium with a less flexible, less torquable, and less hydrophilic pressure wire and the increase in procedural time in rewiring the SideB through the MainB stent. As the 3D FD-OCT reconstructions can visualize the carina shift and the SideB vessel opening, quantification of the SideB opening as an area may allow the operator to assess whether the SideB ostium is hemodynamically compromised, without the need to perform a pressure wire study. The addition of quantitative measurements to the 3D software as well as the requirement for instantaneous online 3D FD-OCT availability of a high-enough resolution would, however, be necessary requirements.
Within the parallel bifurcation, the positioning and geometric relationship of the MainB stent over the SideB opening suggested “overhanging” of the struts over the edge of the carina (Fig. 7); this was especially evident in the views perpendicular to the vessel wall (Figs. 5B and 6B) and contrary to the appearances on the corresponding 2D images where a “jailed” appearance of the stent struts covering the SideB opening was suggested; the jailed appearance of the struts covering the SideB ostium was, however, evident with perpendicular bifurcations (Figs. 3 and 5).
Current recommendations during provisional T-stenting suggest the recrossing of the coronary wire (after MainB stenting) into the SideB through the most distal cell of the MainB stent covering the SideB opening (11,20,21). If the coronary wire is passed through a proximal cell into the SideB, this may potentially provide no scaffolding to the SideB ostium and leave many struts unopposed adjacent to the carina. The adoption of this principle to bifurcations with 3D FD-OCT has recently been shown to be potentially feasible in humans (11,22). To help further appreciate this potential application, a 3D OFDI reconstruction was performed in a patient separate from this study, utilizing the OFDI system from an ongoing trial (23): the case is of a proximal LCx–obtuse marginal bifurcation in the context of an acute coronary syndrome requiring manual aspiration thrombectomy; malapposition was evident immediately after MainB stent implantation requiring further post-dilation (not illustrated) (Fig. 7).
Based on the 3D reconstructions, the malapposed struts located in close vicinity to the proximal ostial rim (i.e., in the “takeoff” position) are overprojected on the true orifice of the SideB, giving the illusion that the SideB orifice is jailed when the orifice is viewed obliquely (Fig. 7A); this, however, appears to be the ideal endoluminal point of view to best select the distal cells of the stent over the parallel bifurcation to pass the coronary wire and perform KBPD (white arrow, Fig. 7A). When moving further downstream along the axis of the MainB, the optical visual perspective of overprojection on the orifice of the SideB is gradually reduced (Fig. 7B) and eliminated when a downstream endoluminal point of view at the level of the carina is selected (Fig. 7C). The struts located in front of the edge of the carina are now seen prominently as a metallic extension of the carina and not covering the SideB opening. The potential to advance a coronary wire beneath the malapposed struts in the proximal vicinity of the ostial rim appears to be very real, and if SideB dilation or stenting were performed, would result in multiple unappposed struts taking off from the carina.
Other properties of 3D FD-OCT reconstructions that may have a clinical application include the addition of tissue characterization; this has previously been performed offline (13,24,25). Tissue characterization within 3D FD-OCT imaging (25) or fusion imaging of 2D FD-OCT and IVUS virtual histology (26) would have the potential to aid in the identification of clinically useful areas of interest such as fibrocalcific plaque, lipid pools, and vulnerable plaque. A recent study utilizing longitudinal high-resolution intravascular ultrasound has suggested that that the effects of carina shift may be limited by the presence of fibro-calcific plaque at the carina, presumably because of resistance to expansion—using 3D FD-OCT, this concept appears easier to appreciate; tissue characterization of intravascular imaging may ultimately aid in identifying these areas (2,17).
The potential for the clinical application of 3D FD-OCT as a complementary tool to 2D imaging is demonstrated. A reassessment of the understanding of 2D FD-OCT imaging may be warranted in light of the 3D findings. Real-time, instantaneous, high-resolution 3D FD-OCT, with the addition of quantitative and possible tissue characterization properties, are required from industry to validate and apply this technology in conventional PCI practice. This may aid in the further understanding of the complexities of the coronary bifurcation.
The first author wishes to thank the Dickinson Trust Travelling Scholarship, Manchester Royal Infirmary, Manchester, England, UK. The authors wish to express their thanks to Evelyn Regar, MD, PhD, Carl Schultz, MD, PhD, Willem J. van der Giessen, MD, PhD, and Jurgen Ligthart, BSc, of the ThoraxCenter, Erasmus University Medical Centre, Rotterdam, the Netherlands, part of the team who undertook the original first-in-man study of the intracoronary Terumo optical frequency domain imaging (OFDI) system. The authors also thank Dr. Dragica Paunovic and Dr. Vladimir Borovicanin of Terumo Corporation and Terumo Europe N.V., and Glenda van Bochove, Ravindra Pawar, and Monique Schuijer of Cardialysis BV, Rotterdam, the Netherlands, for their invaluable technical support.
For supplementary videos, please see the online version of this article.
All authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- computed tomography
- Fourier-domain optical coherence tomography
- intravascular ultrasound intravascular
- kissing balloon post-dilation
- left anterior descending coronary artery
- left circumflex coronary artery
- main branch
- optical coherence tomography
- optical frequency domain imaging
- percutaneous coronary intervention
- right ventricular
- side branch
- Received March 4, 2011.
- Revision received May 20, 2011.
- Accepted June 3, 2011.
- American College of Cardiology Foundation
- Colombo A.,
- Al-Lamee R.
- Hildick-Smith D.,
- de Belder A.J.,
- Cooter N.,
- et al.
- Koo B.K.,
- Kang H.J.,
- Youn T.J.,
- et al.
- Bezerra H.G.,
- Costa M.A.,
- Guagliumi G.,
- Rollins A.M.,
- Simon D.I.
- Di Mario C.I.I.,
- van der Giessen W.J.,
- Foin N.,
- et al.
- Gonzalo N.,
- Garcia-Garcia H.M.,
- Regar E.,
- et al.
- Tearney G.J.,
- Waxman S.,
- Shishkov M.,
- et al.
- Okamura T.,
- Onuma Y.,
- García-García H.M.,
- et al.
- Onuma Y.,
- Girasis C.,
- Aben J.P.,
- et al.
- Medina A.,
- Pan M.,
- Martin P.,
- Suarez de Lezo J.
- Gil R.J.,
- Vassilev D.,
- Formuszewicz R.,
- Rusicka-Piekarz T.,
- Doganov A.
- van Geuns R.J.,
- Gogas B.D.,
- Farooq V.,
- Regar E.,
- Serruys P.W.
- Serruys P.W.,
- Marco J.
- Gonzalo N.,
- Tearney G.J.,
- van Soest G.,
- et al.
- Waxman S.,
- Freilich M.I.,
- Suter M.J.,
- et al.
- Diletti R.,
- Garcia-Garcia H.M.,
- Gomez-Lara J.,
- et al.