Expert review document part 2: methodology, terminology and clinical applications of optical coherence tomography for the assessment of interventional procedures.

Introduction

This document is complementary to an Expert Review Document on Optical Coherence Tomography (OCT) for the study of coronary arteries and atherosclerosis.[1] The goal of this companion manuscript is to provide a practical guide framework for the appropriate use and reporting of the novel frequency domain (FD) OCT imaging to guide interventional procedures, with a particular interest on the comparison with intravascular ultrasound (IVUS).[1-4]

Technique for optical coherence tomography imaging

In the OCT Expert Review Document on Atherosclerosis, a comprehensive description of the physical principles for OCT imaging and time domain (TD) catheters (St Jude Medical, Westford, MA, USA) was provided.[1]

The main advantage of FD-OCT is that the technology enables rapid imaging of the coronary artery, using a non-occlusive acquisition modality. The FD-OCT catheter (DragonflyTM; St Jude Medical) employs a single-mode optical fibre, enclosed in a hollow metal torque wire that rotates at a speed of 100 r.p.s. It is compatible with a conventional 0.014″ angioplasty guide wire, inserted into a short monorail lumen at the tip. The frequency domain optical coherence tomography lateral resolution is improved in comparison with TD-OCT, while the axial resolution did not change. These features, together with reduced motion artefacts and an increased maximum field of view up to 11 mm, have significantly improved both the quality and ease of use of OCT in the catheterization laboratory.[3,4] However, the imaging depth of the FD-OCT is still limited to 0.5–2.0 mm.[5]

Methodological issues: image acquisition, safety, and effectiveness

The main obstacle to the adoption of TD-OCT imaging in clinical practice is that OCT cannot image through a blood field, and therefore requires clearing or flushing of blood from the lumen.[1] The 6 Fr compatible DragonflyTM FD-OCT catheter is so far the only one in the market, as two other systems from Volcano and Terumo, which have function similar to the DragonflyTM, are not yet available. The DragonflyTM catheter is first advanced over a regular guide wire, distal to the region of interest. A dedicated marker, located 10 mm distal to the OCT lens, enables the pull-back starting point selection.

The acquisition of an OCT image sequence requires a bolus of crystalloid solution (usually contrast) injected through the guiding catheter. The acquisition speed can be set up in a range between 5 and 40 mm/s, based on the used OCT system. Most expert users advocate the use of automated contrast injection to optimize the image quality.

The previous experience with TD and FD-OCT technology shows OCT acquisition to be safe,[1,4,6] effective,[1,4,6-8] and highly reproducible for the assessment of the luminal areas and length.[9-11] A fair correlation between OCT and IVUS quantitative measurements of the lumen areas was reported,[9-11] despite comparative studies showing that IVUS tends to slightly overestimate lumen areas, while stent and neointimal areas are slightly higher on OCT.[9-11] Frequency domain optical coherence tomography image quality depends on an accurate acquisition technique and proper guiding catheter engagement is needed to optimize directional contrast flushing.

Procedural definitions

Metallic stent struts are highly reflective structures that lead to specular reflection with the excessive signal usually resulting in ‘blooming’ of the strut surface. Optical coherence tomography shows only the reflection of the luminal surface of the stent strut and is unable to provide a direct measurement of the strut thickness.[8,12]

Two features indicate a strut: a highly reflective spot and/or an associated shadow behind the strut. The presence of only one of these two defining features is sufficient for strut identification.

For bioresorbable stents, the assessment of the whole thickness is possible, since light can usually penetrate non-metallic materials. They can be followed by serial images; the degradation process, characterized by the presence of empty spaces replacing stent struts, was documented for the first generation design,[13] but was not visible with the second-generation model, which maintains its mechanical properties.[14] Based on preliminary unpublished data, the bioabsorbable magnesium stent has different properties; it is visible after implantation and tends to disappear at the follow-up due to magnesium degradation.

Stent overlap occurs in the cross sections in which two superimposed layers of stent struts are visible (Figure ). Incomplete stent apposition or malapposition occurs if there is a separation of a stent strut from the vessel wall. Malapposition can be addressed at a cross-section level and expressed as an area, or can be evaluated at the strut level analysis.[15] In the latter case, it is defined as a measured distance greater than the strut thickness for bare metal stents or greater than the sum of the thickness of the strut plus polymer for drug eluting stents (DES)[8,15] (Figure ). Acute malapposition is diagnosed immediately after stent deployment (Figure ),[14] late malapposition at the follow-up, while late acquired malapposition requires the comparison of the post-intervention and follow-up images. Intra-stent tissue protrusion is defined as tissue prolapsed between stent struts extending inside a circular arc connecting adjacent struts (Figure ).[16] A thrombus is identified as an intra-luminal mass, with no direct continuity with the surface of the vessel wall or as a highly backscattered luminal protrusion in continuity with the vessel wall and resulting in signal-free shadowing.[17] However, the discrimination between a tissue prolapse and intra-stent thrombosis is not simple and sufficiently validated. A dissection flap is a linear rim of tissue with a clear separation from the vessel wall, plaque, or the stent struts.[18] Dissections are also frequent at the stent edges (Figure ) and are defined by their longitudinal extension (mm), circumferential extension (degrees or quadrants), and width.

Left panel: Stent overlapping in the left anterior descending artery. Despite an optimal angiographic result, optical coherence tomography shows a malapposition of the proximal edge of the distal stent, with a 430 µ distance from the inner (arrow heads) to the outer struts (arrows). Right panel: the optical coherence tomography image of the same cross section obtained after additional high pressure inflation with a non-compliant oversized balloon shows a correction of the malapposition.

An optical coherence tomography cross section of a stent immediately following implantation in the right coronary artery. The stent struts are clearly visible. There are struts that are apposed to the vessel wall. However, what is clearly evident is a group of stent struts (between 4 and 12 o'clock) which are grossly malapposed to the vessel wall. Optical coherence tomography cannot penetrate through metallic structures and hence only shows the endoluminal aspect of the stent strut. To be classified as malapposed, the strut must not be in contact with the vessel wall with the distance between the endoluminal aspect and the vessel wall greater than the known thickness of the stent strut itself, including the polymer thickness (arrow in the right panel and dotted line in the left panel with a magnified view).

Examples of intra-stent thrombotic formations and a plaque prolapse. (A and B) A tissue protrusion. Arrows show tissue prolapsed between stent struts and extending inside a circular arc, connecting adjacent struts. (C) Intra-luminal globular protrusions at stent strut locations with no direct continuation from the surface of the vessel wall (arrows). (D) A marked intra-stent thrombosis observed in a drug eluting stents at a late follow-up. Arrows show an intra-stent mass with an irregular inner border. Images in (A) and (D) were obtained with frequency domain optical coherence tomography, those in (B) and (C) with time domain optical coherence tomography.

Left panel: Cypher stent deployed in the mid-right coronary artery with an optimal final angiographic result. Optical coherence tomography (right inferior panels) shows a moderate dissection distal to the stent ending, missed by angiography. The distance between the rim of the dissection and the vessel wall is 250 µm (line), slightly above the resolution of intravascular ultrasound, while the circumferential extension of the plaque dissection is 90°. The right upper panels are obtained at the stent minimum lumen area and show a well-apposed stent with an MLA of 6.82 mm2.

Clinical applications of optical coherence tomography and comparison with intravascular ultrasound

Impact of plaque composition on coronary intervention

Calcification

Unlike angiography, IVUS identifies calcium with high accuracy,[19] but is unable to measure its thickness.[1] Infrared light penetrates calcium better, but calcific components with a thickness >1–1.3 mm or a deep intra-plaque location can prove impossible to penetrate.

In the presence of an important circumferential calcification direct stenting should be avoided and the strategy can span from a careful pre-dilatation to test expansion, to cutting balloons for very short calcific rings or to pre-treatment with a rotational atherectomy.[19,20]

As OCT accurately measures the radial and longitudinal extension of a lipid pool, its use may be encouraged to also predict the risk of embolization of plaque components;[1] this is in line with the previous experience with VH-IVUS.[21,22]

Unlike angiography or IVUS, OCT holds promise in identifying thrombi, measuring their dimensions and guiding their removal[23] (Figures  and ).

The figure shows optical coherence tomography potentialities to detect components of ruptured plaques. Left panel shows a suboccluded right coronary artery in a patient with an inferior myocardial infarction. Optical coherence tomography cross sections (B) and (C) were obtained at the site of a plaque rupture and show a plaque with mixed composition (calcium + lipid) (arrow). At the shoulder of the plaque, at the site of minimum thickness of the fibrous cap (arrow head), an intra-luminal thrombotic formation is shown (dotted arrow). (A) A severe reduction in the lumen area due to a marked concentric thrombotic formation. As the penetration of optical coherence tomography through a thrombus is limited, vessel media cannot be studied (arrow).

The application of a score based on the semi-quantitative assessment of a thrombus (number of the involved quadrants on the cross-sectional optical coherence tomography images) and the longitudinal extension of the thrombus itself. By applying this semi-quantitative grading, in each cross section, a thrombus is classified as absent or subtending 1, 2, 3, or 4 quadrants. The score is then calculated as the sum of each cross-section score.

Assessment of ambiguous angiographic lesions and deferral of interventions

In angiographically hazy lesions, OCT often detects ruptured plaques with a thrombus attached to the site of rupture of the fibrous caps over a partially emptied lipid pool.[1,6] Under these circumstances, the decision for the need to proceed with treatment can stem more from these morphological observations than from the absolute measurement of the lumen area.

A technical drawback of both TD-OCT and FD-OCT in this application is that a plaque located at the very ostium of the left or right coronary arteries cannot be accurately imaged.[1] As for IVUS, preliminary data indicate that OCT can change the operator's intention to treat, avoiding unnecessary interventional procedures or modifying the strategy in some cases.[6,24,25]

As FD-OCT probes have a slightly thinner profile than IVUS probes and pull-back imaging can be done at very high speeds (normally 20 mm/s), a significant fraction of severely diseased target lesions can be imaged without causing luminal obstruction, with symptomatic ischaemia being less likely. However, in the presence of subocclusive lesions the OCT probe can cause a luminal obstruction and it can be more convenient to perform OCT after gentle pre-dilatation.

Identification of vulnerable plaques

A number of IVUS studies attempted to characterize the appearance of vulnerable plaques. Recently, the PROSPECT trial, based on a signal radiofrequency analysis of the IVUS backscatter, showed for the first time that IVUS can identify plaques with a risk of events at 3 years.[26] OCT, due to its high accuracy in the detection of superficial plaque components, can directly measure the fibrous cap thickness and can possibly detect plaques with a risk of a rupture with greater accuracy. The utility of a combined approach based on the use of OCT and VH-IVUS has been recently proposed to better characterize deep lesion components.[27,28] Possibly, in the near future, a real-time application of OCT algorithms to characterize plaque components and identify local signs of inflammation will facilitate OCT detection of plaque vulnerability.[29,30]

Advantages and pitfalls of optical coherence tomography for guidance of coronary interventions

Optical coherence tomography easily enables the comparison of the minimal stent area with the reference area, which is the most often used IVUS criterion for optimal stent expansion. The demonstration of the usefulness of an IVUS-guided approach of a bare metal stent expansion to reduce restenosis is more by meta-analyses[31] than by single randomized studies.[32,33] There is some evidence that IVUS guidance can improve the clinical outcome in the presence of left main disease[34] and tackles the occurrence of thrombosis.[35,36] The ability of OCT to address luminal areas and identify underexpansion, malapposition, uneven stent strut distribution, or small intra-stent thrombotic formations makes the technique a very attractive tool for the prevention of thrombosis. However, future studies are required to determine whether OCT is as useful or more useful than IVUS in this regard.

Acute malapposition can contribute to a stent thrombosis by disturbing the normal laminar blood flow along the vessel wall, by promoting the deposition of platelets and fibrin[37] and reducing re-endothelialization and neointima formation.[38] However, to date, IVUS data suggest that acute and late-stent malapposition do not increase the risk of major adverse cardiac events.[39,40] Nevertheless, both studies detected a surprisingly small number of malapposed stents (<7.2%) using the IVUS-derived criterion of at least one malapposed strut. Optical coherence tomography studies following stent implantation have demonstrated a much greater proportion of malapposed struts, even after optimal high-pressure post-dilatation, with this phenomenon being particularly evident in regions of stent overlap,[41] after deployment of DES or long stents, and in type C lesions.[15]

In patients with acute coronary syndromes, the occurrence of an in-stent tissue protrusion, due to the presence of residual thrombus, is a common finding on OCT[1,8,42] that may elevate the risk for an acute and subacute stent thrombosis.

The need of serial OCT acquisitions to guide the selection of balloons and stents and to correct underexpansion means repeated contrast injections which may significantly increase the total amount of procedural contrast. The main drawback of OCT is its inability, unlike IVUS, to outline the vessel architecture, and measure the external elastic membrane, and the longitudinal extent of plaque burden in lesions with thicknesses exceeding 1.0–1.5 mm (Figure ). This drawback may have some clinical implications, as the accepted criterion for the identification of reference segments by IVUS is a plaque burden <40%.[33] Therefore, the pre-intervention use of OCT to select the appropriate stent size and length in diffusely diseased vessels is questionable.

Optical coherence tomography is very well versed at detecting microstructures in close proximity to the vessel wall. Its limited tissue penetration (<1.5 mm), however, means that the extent of plaque is unable to be determined. In this example, diffuse, poorly reflective tissue is seen consistent with a lipid rich plaque (arrow), however, its extent beyond 1 mm from the vessel wall is limited.

Optical coherence tomography has an easier application for the treatment of late in-stent restenosis because the strong reflective power of the stent struts allows their detection through thick layers of hyperplasia, allowing optimal sizing of high-pressure balloons to correct underexpansion and facilitating the use of cutting balloons.[15]

Stent follow-up

Delayed healing and poor endothelialization are common findings in the pathological specimens of vessels treated with DES,[43] and recent post-mortem studies demonstrated that a late-stent thrombosis was strongly associated with the ratio of uncovered/total stent struts.[44]

Several experimental and clinical studies have demonstrated the high correlation between OCT and histological measurements of tissue coverage of stent struts,[45,46] although OCT is unable to identify the endothelium.[1] In a subanalysis of the ODESSA trial, 8% of the stented segments with no detectable neointima by IVUS were found to have neointimal coverage by OCT.[41]

Follow-up OCT data revealed that most of the DES, including those with a biodegradable polymer, were covered with thin neointima, but few revealed complete coverage.[47-51] Incomplete stent apposition without neointimal hyperplasia was significantly associated with the presence of an OCT-detected thrombus at the follow-up, and may constitute a potent substrate for a late-stent thrombosis;[52] however, so far a subclinical thrombus has not been related to the risk of major adverse cardiac events.[53]

Furthermore, incomplete stent apposition and the absence of OCT tissue coverage are more frequently found in the setting of acute coronary syndromes, particularly after DES deployment (Figure ).[54] It is difficult to offer any recommendation at this stage for the use of OCT for the late follow-up of individual patients but anecdotal cases of OCT application to rule out the need for the prolongation of a dual antiplatelet treatment in patients requiring undeferrable surgery have been reported. The main application is, at present, the comparison of different stent platforms, assuming that a more uniform strut coverage could improve the late outcome.

Malapposed struts tend not to heal with tissue coverage unlike well-apposed ones. Optical coherence tomography images obtained 1 month after Taxus stent deployment in the right coronary artery. Optical coherence tomography showed malapposition of the drug eluting stents with many struts not adhering at the vessel wall. The vast majority of malapposed struts (from a to l) were uncovered while only two of them (arrows) had some tissue coverage. On the other hand the two well-apposed struts (m, n) were covered.

Standardized methodology for core laboratory optical coherence tomography image analysis

Optical coherence tomography is being increasingly used as a surrogate method for evaluating stents.[15] The ability to provide serial images and analyse a large number of OCT cross section and stent struts represent an advantage over standard histopathology. Core laboratory OCT image analysis should, however, be performed according to a standardized methodology and after a deep understanding of the method.[15] Many OCT readers suggest to address images not only at a strut level, but including also cross-sectional and in-stent level analyses, to further study an in-stent coverage uniformity and individual propensity towards incomplete stent healing. Furthermore, if available, automated tools should be applied to reduce possible intra- and inter-observer' variability.[55]

Ideally, for each strut information on malapposition, coverage, presence of thrombotic formations, and tissue prolapse should be given. Different classifications have been adopted so far by OCT readers and data on intra- and inter-observer variability for the applied methods are not available. As a rationale solution, the following classification can be adopted to define stent struts throughout the pull-back image sequence: embedded covered struts: covered by tissue with at least 50% of the strut boundary below the level of the luminal surface; protruding covered struts: covered by tissue and with the strut boundary located above the level of the luminal surface; uncovered apposed struts: for those not covered by tissue but abutting the vessel wall; uncovered and malapposed struts: for those not covered by tissue and not abutting the vessel wall.[1,15,41]

Some OCT readers explored different patterns of stent coverage[56,57] describing a homogeneous vs. heterogeneous pattern of intimal thickening.[56] An intra-stent thrombus is typically shown as an in-stent globular appearance and can be reasonably identified. In the presence of stent malapposition, the lumen area should be divided into an in-stent lumen area and an extra-stent lumen area. Stents overlapping segments and bifurcation lesions with major side branches should be noted.

Conclusions

Optical coherence tomography enables an accurate vessel lumen and stent assessment, providing a high image quality with fast pull-back speeds, procedural safety, and patient tolerability. Optical coherence tomography offers unique insights into stenting procedures that otherwise would be missed using conventional angiography or IVUS. Whether such anatomical details will lead to superior or complementary information with respect to IVUS in guiding coronary intervention, particularly in the era of DES, remains to be defined.

Funding

Funding to pay the Open Access publication charges for this article was provided by CLI Foundation.

Conflict of interest: F.P. has received consultant fees from St Jude Medical. G.G. has received consultant fees from Boston Scientific, St Jude Medical, Volcano, and Cordis; and research grants from Boston Scientific, Medtronic Vascular, Abbott, and LightLab Imaging. G.S.M. has received consultant fees from Volcano and St Jude Medical; and grant support from Volcano and Boston Scientific. M.C. has received consultant fees from Abbott, St Jude Medical, LightLab Imaging, and Biosense Webster; and is a scientific advisor for Scitech. G.J.T. has received consultant fees from Samsung and Merck; royalties from Terumo and MIT; and has received sponsored research from Terumo. I.K.J. has received consultant fees and research grant from St Jude Medical. H.G.B. has received consultant fees and honoraria from St Jude Medical. D.D. has received consultant fees and research grants from Abbott, Adamed, AstraZeneca, Biotronik, Balton, Bayer, BBraun, BioMatrix, Boston Scientific, Boehringer Ing, Bristol Myers Squibb, Cordis, Cook, Eli-Lilly, EuroCor, Glaxo, Invatec, Medtronic, Medicines Company, Merck, Nycomed, Orbus-Neich, Pfizer, Possis, Promed, sanofi-aventis, Siemens, Solvay, Stentys, Terumo, and Tyco; and is an advisory board member. P.M. has received consultant fees from St Jude Medical and Terumo. T.A. has received consultant fees and carried out educational activity for St Jude Medical. C.D.M. has received speaker fees from St Jude Medical.