Quantitative coronary angiography
|
Coronary Angiography | |
|
General Principles | |
|---|---|
|
Anatomy & Projection Angles | |
|
Normal Anatomy | |
|
Anatomic Variants | |
|
Projection Angles | |
|
Epicardial Flow & Myocardial Perfusion | |
|
Epicardial Flow | |
|
Myocardial Perfusion | |
|
Lesion Complexity | |
|
ACC/AHA Lesion-Specific Classification of the Primary Target Stenosis | |
|
Lesion Morphology | |
Editor-In-Chief: C. Michael Gibson, M.S., M.D. [1]
Overview
While “on-line” quantitative angiographic is somewhat cumbersome to use in the catheterization laboratory, “off-line” quantitative angiography has proven invaluable for research investigation in determining the effect of new drugs and devices on lumen dimensions early and late after percutaneous coronary interventions.
Notably, for clinical decision making in intermediate lesions, neither trained visual estimates or on-line quanitative angiography are substitutes for precise physiologic measurements of stenosis severity, such as fractional flow reserve or coronary Doppler measurements.
Computer-Assisted Quantitative Angiography
Quantitative coronary angiography was initiated nearly 30 years ago by Brown and colleagues who magnified 35-mm cineangiograms obtained from orthogonal projections and hand traced the arterial edges on a large screen. After computer-assisted correction for pincushion distortion, the tracings were digitized and the orthogonal projections were combined to form a three-dimensional representation of the arterial segment, assuming an elliptical geometry. While the accuracy and precision were enhanced compared with visual methods, the time needed for image processing limited its clinical use.
Several automated edge-detection algorithms were then developed and applied to directly acquired digital images or to 35-mm cinefilm digitized using a cine-video converter. Subsequent interations of these first degeneration devices have utilized enhanced microprocessing speed and digital image acquisition to render the end-user interface more flexible and substantially shortened the time required for image analysis.
Quantitative coronary angiography is divided into several distinct processes, including film digitization (when applicable), image calibration, and arterial contour detection. For processing 35-mm cinefilm, a cine-video converter is used to digitize images into a 512 × 512 (or larger) × 8-bit pixel matrix. Optical, or less preferred digital, magnification results in an effective pixel matrix up to 2458 × 2458.
Image Calibration
For estimation of absolute coronary dimensions, the diagnostic or guiding catheter is generally used as the scaling device. In general, a nontapered segment of the catheter is selected, and a centerline through the catheter is drawn. Linear density profiles are then constructed perpendicular to the catheter centerline, and a weighted average of the first and second derivative function is used to define the catheter edge points. Individual edge points are then connected using an automated algorithm, outliers are discarded, and the edges are smoothed. The diameter of the catheter is then used to obtain a calibration factor, expressed in millimeters per pixel. The injection catheter dimensions may be influenced by whether contrast or saline is imaged within the catheter tip, and by the type of material used in catheter construction. As the high-flow injection catheters have been developed, more quantitative angiographic systems using contrast-filled injection catheters for image calibration.
Quantitative Coronary Analysis
The automated algorithm is then applied to a selected arterial segment, and absolute coronary dimensions are obtained from the minimal lumen diameter (MLD) reference diameter, and from these, the percent diameter stenoses are derived. For most angiographic systems, interobserver variabilities are 3.1% for diameter stenosis and 0.10 to 0.18 mm for MLD for cineangiographic readings; variabilities are slightly higher (< 0.25 mm) for repeated analyses of the digital angiograms due the the slightly lower resolution compared with cineangiography. The two most commonly used quantitative angiographic systems are described below:
Cardiovascular Angiography Analysis System (CAAS)
Cardiovascular Angiography Analysis System is a quantitative angiographic system developed for off-line cineangiographic analysis (Pie Data Medical B.V., Maastricht, The Netherlands). The edge-detection algorithm incorporates an optional correction for pincushion distortion; its edge detection uses a weighted (50%) sum of the first and second derivatives of the mean pixel density; and it applies minimal cost criteria for smoothing of the arterial edge contours. In addition to reporting a interpolated reference diameter and a minimal lumen diameter (MLD), a subsegment analysis provides mean, minimum, and maximum subsegment diameters. Specific reporting algorithms have been developed for drug-eluting stents, patients undergoing radiation brachytherapy, and in those undergoing peripheral intervention.
Coronary Measurement System (CMS)
Specific features of the CMS include two-point user-defined centerline identification, arterial edge detection using a weighted (50%) sum of the first and second derivatives of the mean pixel density, arterial contour detection using a minimal cost matrix algorithm, and an “interpolated” reference vessel diameter (MEDIS, Leiden, The Netherlands). One limitation of the minimal cost algorithm used with the first-generation CMS (and CAAS-II) system has been its inability to precisely quantify arterial lumen contours characterized by abrupt changes. The CMS-GFT is an algorithm that is not restricted in its search directions, incorporating multidirectional information about the arterial boundaries for construction of the arterial edge that is suitable for the analysis of complex coronary artery lesions. Specific reporting algorithms have been developed for bifurcation lesions, drug-eluting stents, patients undergoing radiation brachytherapy, and in those undergoing peripheral intervention.
Factors Contributing to Variability Using QCA
Variability associated with measurements of the MLD and reference diameter is affected by a number of factors:
- The biologic differences among lumen diameters (e.g., reference vessel size, vasomotor tone, thrombus).
- Inconsistencies in radiographic image acquisition parameters (e.g., quantum mottling, out-of-plane magnification, foreshortening).
- Angiographic measurement variability (e.g., frame selection, factors affecting the edge-detection algorithm). These factors should be controlled in order to improve on the overall diagnostic accuracy of quantitative angiography
Quantitative Angiographic Indices
- Angiographic Success
- Binary Angiographic Restenosis
- Late Lumen Loss
Risk Assessment Using Specific Lesion Morphologic Criteria
Despite the value of risk scores in estimating aggregate procedural risk, there are several limitations of these criteria as applied to individual patients. Identification of lesion characteristics, such as eccentricity, irregularity, angulation, and tortousity, is limited by substantial inter-observer variability. Agreement with ACC/AHA classification was noted in only 58% of lesions in one series, with disagreement by two classification grades noted in nearly 10% of lesions.
Accordingly, rather than a composite score, description of individual morphologic features may be more predictive of early and late outcome following PCI. Some ACC/AHA morphologic features are associated with a complicated procedure (e.g., thrombus, saphenous vein graft [SVG] degeneration, and angulated segments), whereas others are associated with an unsuccessful but uncomplicated procedure (e.g., chronic total occlusions or diffuse disease). As an alternative to providing a composite lesion complexity score, estimation of procedural risk based on the presence of one or more specific adverse morphologic features may be more useful.
Definitions of Preprocedural Lesion Morphology
Feature: Definition
- Eccentricity: Stenosis that is noted to have one of its luminal edges in the outer one quarter of the apparently normal lumen
- Irregularity: Characterized by lesion ulceration, intimal flap, aneurysm, or “sawtooth” pattern
- Ulceration: Lesion with a small crater consisting of a discrete luminal widening in the area of the stenosis is noted, provided it does not extend beyond the normal arterial lumen
- Intimal flap: A mobile, radiolucent extension of the vessel wall into the arterial lumen
- Aneurysmal dilation: Segment of arterial dilation larger than the dimensions of the normal arterial segment
- 'Sawtooth pattern: Multiple, sequential stenosis irregularities
- Lesion length: Measured “shoulder-to-shoulder” in an unforeshortened view
- Discrete: Lesion length < 10 mm
- Tubular: Lesion length 10–20 mm
- Diffuse: Lesion length ≥ 20 mm
- Ostial location: Origin of the lesion within 3 mm of the vessel origin
- Lesion angulation: Vessel angle formed by the centerline through the lumen proximal to the stenosis and extending beyond it and a second centerline in the straight portion of the artery distal to the stenosis
- Moderate: Lesion angulation ≥ 45 degrees
- Severe: Lesion angulation ≥ 90 degrees
- Bifurcation stenosis: Present if a medium or large branch (>1.5 mm) originates within the stenosis and if the side branch is completely surrounded by stenotic portions of the lesion to be dilated
- Lesion accessibility (proximal tortuosity)
- Moderate tortuosity: Lesion is distal to two bends ≥ 75 degrees
- Severe tortuosity: Lesion is distal to three bends ≥ 75 degrees
- Degenerated saphenous vein graft: Graft characterized by luminal irregularities or ectasia comprising > 50% of the graft length
- Calcification: Readily apparent densities noted within the apparent vascular wall at the site of the stenosis
- Moderate: Densities noted only with cardiac motion prior to contrast injection
- Severe: Radiopacities noted without cardiac motion prior to contrast injection
- Total occlusion: TIMI 0 or 1 flow
- Thrombus: Discrete, intraluminal filling defect is noted with defined borders and is largely separated from the adjacent wall; contrast staining may or may not be present.