HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Schoepf, U. J. Articles by Yucel, E. K. Search for Related Content PubMed PubMed Citation Articles by Schoepf, U. J. Articles by Yucel, E. K.

CT of Coronary Artery Disease¹

¹ From the Department of Radiology, Brigham and Women’s Hospital, Harvard Medical School, 75 Francis St, Boston, MA 02115 (U.J.S., E.K.Y.); Institute of Clinical Radiology, University of Munich, Germany (C.R.B.); and Siemens Medical Solutions, Division CT, Forchheim, Germany (B.M.O.). Received April 29, 2003; revision requested July 10; revision received July 29; accepted August 25; updated September 17. U.J.S. and E.K.Y. supported in part by research grants from Berlex Laboratories, Wayne, NJ. Address correspondence to U.J.S. (e-mail: schoepf@bwh.harvard.edu).

	ABSTRACT

TOP ABSTRACT INTRODUCTION IMAGE ACQUISITION AND DISPLAY... CLINICAL APPLICATIONS: IMAGING... CLINICAL APPLICATIONS: CONTRAST... CT OF CAD: FUTURE... ESSENTIALS REFERENCES

 
The socioeconomic importance of heart disease provides considerable motivation for development of radiologic tools for noninvasive imaging of the coronary arteries. Current computed tomographic (CT) techniques combine high speed and spatial resolution with sophisticated electrocardiographic synchronization and robustness of use. Application of these modalities for evaluation of coronary artery disease is a topic of active current research. Coronary artery calcium measurements with different CT techniques have been used for determining the risk of coronary events, but the exact role of this marker for cardiac risk stratification remains unclear pending results of population-based studies. Contrast material–enhanced CT coronary angiography has become an established clinical indication for some scenarios (eg, coronary artery anomalies, bypass patency, surgical planning). With current technology, the accuracy of CT coronary angiography for detection of coronary artery stenoses appears promising enough to warrant pursuit of this application, but sensitivity is still not high enough for routine diagnostic needs. The high negative predictive value of a normal CT coronary angiogram, however, may be useful for reliable exclusion of coronary artery stenosis. The cross-sectional nature of CT may allow noninvasive assessment of the coronary artery wall. Use of contrast-enhanced CT coronary angiography for detection, characterization, and quantification of atherosclerotic changes and total disease burden in coronary arteries as a potential tool for cardiac risk stratification is currently being investigated.

© RSNA, 2004

Index terms: Computed tomography (CT), angiography, 54.12111, 54.12112, 54.12115, 54.12116 • Computed tomography (CT), multi–detector row, 54.12111, 54.12115, 54.12117, 54.12118 • Coronary vessels, bypass graft, 54.4544 • Coronary vessels, calcification, 54.812 • Coronary vessels, CT, 54.12111, 54.12112, 54.12115, 54.12116, 54.12117, 54.12118 • Coronary vessels, stenosis or obstruction, 54.4558, 54.76 • Review

	INTRODUCTION

TOP ABSTRACT INTRODUCTION IMAGE ACQUISITION AND DISPLAY... CLINICAL APPLICATIONS: IMAGING... CLINICAL APPLICATIONS: CONTRAST... CT OF CAD: FUTURE... ESSENTIALS REFERENCES

 

EDITOR’S NOTE: In view of the importance of this topic, we are publishing in this issue of Radiology two Special Reviews on CT evaluation of the coronary arteries. Each article presents important perspectives for our readers. Please see also the article by Schoenhagen et al.

Anthony V. Proto, MD, Editor

Coronary artery disease (CAD) remains the leading cause of death in Western nations. The standard of reference for diagnosis of CAD still is conventional coronary angiography. In 1999, more than 1.83 million conventional angiographic examinations were performed in the United States (1). The greatest advantage of conventional angiography is high spatial resolution and the option of direct performance of interventions such as balloon dilatation or coronary stent placement. However, only one-third of all conventional coronary angiographic examinations in the United States are performed in conjunction with an interventional procedure, while the rest are performed only for diagnostic purposes—that is, only for verification of the presence and degree of CAD (1). Thus, both in the face of limited health care resources and in the interest of patients who undergo unnecessary invasive tests, a reliable noninvasive radiologic tool for imaging of the coronary arteries and for early diagnosis of CAD is highly desirable.

Computed tomography (CT) has been embraced as the premier noninvasive modality for vascular imaging of the thorax (2–4). Imaging of the heart, however, has always been technically challenging because of the heart’s continuous motion. CT imaging of the heart moved into the diagnostic realm with the introduction of electron-beam CT (5) and, more recently, of multi–detector row CT (6–8) and with the development of electrocardiography (ECG)-synchronized scanning and reconstruction techniques (9). These modalities permit fast volume coverage and high spatial and temporal resolution, which constitute the sine qua non for successful cardiac imaging (9–14).

For imaging of the heart, CT has been evaluated for assessment of the myocardium (15–18) and myocardial perfusion (19,20) and viability (21), cardiac function and wallmotion (22–24), heart valves (25), and cardiac tumors (26,27). For these applications, however, CT is in considerable competition with less invasive or more comprehensive imaging modalities such as echocardiography or cardiac magnetic resonance (MR) imaging, so that performance of a dedicated CT examination for these purposes is ordinarily reserved for special indications only. Until recently, CT applications for the assessment of CAD per se were almost exclusively directed at the detection and quantification of coronary arterial calcium (28). To date, however, the diagnostic value of CT coronary calcium measurements and the exact role of this marker for cardiac risk stratification remain unclear and controversial (29–32). However, the introduction and ongoing technical improvement of fast ECG-synchronized CT image acquisition in the heart (14) have enabled imaging of the coronary arterial tree with a combination of speed and spatial resolution that has hitherto been unparalleled by other noninvasive imaging modalities. In recent years, considerable interest has accordingly been directed at the beneficial use of high-spatial-resolution contrast medium–enhanced CT angiography for noninvasive interrogation of the coronary arterial tree. To date, the central rationale of this application has been the noninvasive detection and grading of coronary artery stenoses, with the ultimate goal of replacing diagnostic invasive conventional coronary angiography (33–38). Unlike conventional angiography, however, the cross-sectional nature of CT additionally may enable assessment of the vessel wall (39–41). The potential of this technique for noninvasive identification, characterization, and quantification of atherosclerotic lesions and total disease burden within the coronary arteries is currently being evaluated.

For the reasons stated above, dedicated imaging of the gross anatomic and functional sequelae of ischemic heart disease (perfusion, motion, viability, etc) is not the current mainstay of CT, although there may be a greater role for such applications in the future. The focus of this discussion, therefore, will be on the true strength of ever-evolving CT techniques: the combination of unprecedented speed and spatial resolution paired with sophisticated acquisition strategies for imaging the elusive but cardinal anatomy and pathology of the coronary arteries and of CAD.

	IMAGE ACQUISITION AND DISPLAY TECHNIQUES

TOP ABSTRACT INTRODUCTION IMAGE ACQUISITION AND DISPLAY... CLINICAL APPLICATIONS: IMAGING... CLINICAL APPLICATIONS: CONTRAST... CT OF CAD: FUTURE... ESSENTIALS REFERENCES

 
Modalities for CT Imaging of the Heart
In 1984, electron-beam CT was introduced as the first system to enable ECG-synchronized CT imaging of the cardiac anatomy (42). With presently available electron-beam CT scanners, the routine protocol for evaluation of the cardiac anatomy and coronary arteries ordinarily comprises a collimation of 3 mm, a temporal resolution of 100 msec, and prospective ECG triggering (see below) for sequential acquisition of transverse images consistently at the same phase of the cardiac cycle, typically during diastole (43).

In 1998, mechanical spiral CT systems with simultaneous acquisition by four detector rows and a minimum rotation time of 500 msec were introduced (6–8). This development provided a substantial performance increase over the single– and dual–detector row spiral CT systems that had been available until then. For cardiac imaging with four–detector row CT, a collimation of 4 x 2.5 mm or 4 x 1 mm is ordinarily employed. A temporal resolution higher than that of older mechanical spiral CT scanners is enabled by means of faster gantry rotation speed (ie, 500–800 msec) combined with dedicated image reconstruction algorithms (8,44). The strategy that has been pursued since the introduction of the first multi–detector row CT scanner to further improve fast high-resolution volume coverage is to increase the number of sections that are simultaneously acquired. So far, this has resulted in the introduction of eight–, 10–, 16–, 32–, 40–, and 64–detector row CT scanners with further reduced gantry rotation times and minimum beam collimation widths of less than 1 mm (14). However, it seems questionable whether this particular strategy will be pursued indefinitely for increasing scanner performance.

Whereas virtually all CT applications have benefited from improved scanner capabilities, the most striking improvements brought on by the combination of fast rotation time and multi–detector row acquisition with submillimeter spatial resolution were observed in cardiovascular imaging (14,37,45). Coronary arteries are small and complex three-dimensional (3D) structures. The diameter of coronary vessels tapers down from a typical 5 mm in the left main coronary artery to 1 mm luminal diameter in the distal left anterior descending coronary artery (46). For adequate visualization of the small, tortuous, and complex anatomy of the coronary arterial tree and of subtle pathologic conditions along the vessel course, therefore, isotropic (ie, equal voxel dimensions in x, y, and z axes) or near isotropic in-plane and through-plane spatial resolutions of less than 1 mm are necessary. With the protocols that are ordinarily used for high-spatial-resolution imaging of the coronary arteries, current multi–detector row CT scanners provide an in-plane resolution of 0.5 mm and an effective through-plane (z-axis) resolution of 0.6–0.8 mm (14), which approaches the requirements for successful noninvasive imaging of the coronary arteries (47). In order to differentiate a 10%–20% coronary artery stenosis, however, CT systems will need to provide an isotropic spatial resolution of at least 0.3 mm (14).

Motion artifacts that are caused by cardiac pulsation can be minimized in high-spatial-resolution CT studies of the heart by means of scanning or reconstructing projection data at a time point with the least cardiac motion, ordinarily in the diastolic phase of the heart cycle. For most practical purposes, a heart-rate–independent temporal resolution of 100 msec or less allows elimination of most cardiac motion, if images are acquired (prospective triggering) or reconstructed (retrospective gating) during diastole. The heart phases can be determined from a simultaneously recorded ECG signal. Two ECG synchronization techniques are most commonly employed for cardiac CT scanning: prospective ECG triggering and retrospective ECG gating.

ECG-synchronized CT Scan Acquisition: Prospective Triggering
Prospective ECG triggering has long been used in conjunction with electron-beam CT and single–detector row spiral CT (43,48–50). A trigger signal is derived from the patient’s ECG on the basis of a prospective estimation of the present R-R interval, and the scan is started at a defined time point after a detected R wave, usually during diastole. With multi–detector row CT, several sections are obtained simultaneously during one scan acquisition with a cycle time that ordinarily allows image acquisition at every other heartbeat (51). In general, this strategy results in shorter breath-hold times compared with those for single–detector row CT techniques, and respiratory artifacts are less likely to occur.

To improve temporal resolution, scan data are only acquired during a partial scanner rotation (approximately two-thirds of a rotation with 240°–260° projection data), which covers the minimum amount of data required for image reconstruction. Conventional partial reconstruction based on fan-beam projection data results in a temporal resolution that equals the acquisition time of the partial scan. Optimized temporal resolution can be achieved with parallel-beam–based "half-scan" reconstruction algorithms that provide a temporal resolution of half the rotation time in a center area of the scanning field of view (eg, 250 msec for 500-msec rotation time, 210 msec for 420-msec rotation time). In this way, prospective ECG triggering is the most dose-efficient method for ECG-synchronized scanning because only the minimum amount of scan data needed for image reconstruction is acquired (11). However, only rather thick section collimation (3 mm with electron-beam CT, 2.5–3 mm with four– and eight–detector row CT, 1.5 mm with 16—detector row CT) is usually being used for a prospectively ECG-triggered acquisition. Thus, the resulting data sets are less suitable for 3D reconstruction of small cardiac anatomy. Also, the prospectively ECG-triggered technique greatly depends on a regular heart rate of the patient and is bound to result in misregistration in the presence of arrhythmia.

ECG-synchronized CT Scan Acquisition: Retrospective Gating
An alternative approach for ECG-synchronized CT is retrospective ECG gating (Fig 1). This strategy generally enables greater flexibility for phase-consistent image reconstruction when examining a patient with a changing heart rate during acquisition. Retrospective ECG gating requires multi–detector row spiral scanning with a slow table motion and simultaneous recording of the ECG trace, which is used for retrospective linkage of scan data with particular phases of the cardiac cycle (9,12). Retrospectively ECG-gated CT of the heart requires a highly overlapping spiral scan with a spiral table speed adapted to the heart rate to ensure complete phase-consistent coverage of the heart with overlapping image sections.

View larger version (67K): [in this window] [in a new window] [Download PPT slide]   Figure 1. Schematic shows adaptive segmented image reconstruction approach for ECG-gated multi-detector row CT. Dashed lines are z-axis positions of detector rows, which continuously and linearly change position relative to the patient during constant spiral feed. ECG signal is simultaneously recorded during image acquisition and is displayed at bottom of the diagram. At heart rates less than a predefined threshold, one segment of consecutive multisection spiral data is used for image reconstruction. At higher heart rates, two or more subsegments from adjacent heart cycles contribute to the partial scan data segment. In each cardiac cycle, a stack of images is reconstructed at different z-axis positions covering a small subvolume of the heart (dark gray box). The combination of subvolumes from all heart cycles during scanning provides a continuous 3D data set of the entire heart.

Ordinarily, a stack of images is reconstructed at every heartbeat, which enables faster coverage than that for prospectively ECG-triggered multi–detector row CT scanning. Moreover, the continuous spiral acquisition enables reconstruction of overlapping image sections and, thus, a longitudinal spatial resolution of about 20% less than the collimated section width can be achieved (eg, 2.5 mm for 3.0-mm sections, 1.0 mm for 1.25-mm sections, 0.8 mm for 1.0-mm sections, 0.6 mm for 0.75-mm sections) (14,52). For these reasons, a retrospectively ECG-gated acquisition is the preferred method for contrast-enhanced high-spatial-resolution imaging of small cardiac structures, especially the coronary arteries. For image reconstruction at every heart beat, fan-beam data of a partial rotation (usually 240°–260°) are used, which results in a temporal resolution equivalent to half of the rotation time in a centered region of interest (250 msec for 500-msec rotation time, 210 msec for 420-msec rotation time) (9,14,52). A multi–detector row spiral interpolation between the projections of adjacent detector rows is used to compensate for table movement and to provide a well-defined section-sensitivity profile for images without spiral movement artifacts.

Temporal resolution can be improved by using scan data from more than one heart cycle for reconstruction of a single transverse image ("segmented reconstruction") (52–55) (Fig 1). The partial scan data set for reconstruction of one image then consists of projection sectors from multiple consecutive heart cycles. Depending on the relation between rotation time and patient heart rate, a temporal resolution is present between RT/2 and RT/2M, where RT is rotation time and M is the number of projection sectors and the number of used heart cycles. Despite theoretically better temporal resolution, segmented-reconstruction algorithms do not regularly provide superior image quality (55), because the algorithms are very sensitive to changing heart rates.

Diastole is usually chosen for image reconstruction of cardiac and coronary morphology because it is the phase of the cardiac cycle with the least motion; however, owing to the highly overlapping acquisition, image data can be reconstructed for each x-, y- and z-axis position within the scanned volume over the entire course of the cardiac cycle. This allows retrospective selection of reconstruction points that provide the best relative image quality in an individual patient and for anatomy with special motion patterns (56,57). To improve phase consistency in the presence of arrhythmia, individual image stacks can be discarded, or their reconstruction interval can be arbitrarily shifted within the cardiac cycle, so that reconstruction ideally always coincides with the same interval during diastole at each level of the cardiac volume (9). In addition to the structural information that is derived from image reconstruction during diastole, additional reconstructions of the same scan data set at different phases of the cardiac cycle can be used as a by-product of retrospectively ECG-gated acquisitions for analysis of basic cardiac function parameters such as end-diastolic volume, end-systolic volume, and ejection fraction (22–24).

Radiation Dose
Relatively high radiation exposure is involved with retrospectively ECG-gated imaging of the heart because of continuous x-ray exposure and overlapping data acquisition at a slow spiral table feed (58). All acquired data can be used for image reconstruction in different cardiac phases, but, if only a very limited interval (eg, diastolic phase) of the cardiac cycle is targeted during reconstruction, a substantial portion of the acquired data and radiation exposure are redundant and do not contribute to image generation.

Radiation exposure at cardiac CT has received heightened attention recently, especially in light of the current and potentially expanding future use of this modality as a screening tool for CAD in a priori healthy asymptomatic individuals (58–62). There is considerable disagreement in the literature as to the actual radiation dose applied during cardiac CT with different CT scanner types. This disagreement seems mostly related to a lack of standardization of the protocols in use for interrogation of the heart, the absence of a strict distinction between protocols aimed at coronary calcium scoring (low-dose application) and contrast-enhanced CT coronary angiography (high-dose application), failure to adjust measured radiation doses to imaging parameters (eg, section thickness, tube current, tube voltage, scan volume), and hallmarks of image quality (eg, in-plane resolution, through-plane resolution, image noise).

Studies employing rigorous scientific evaluation of the radiation dose applied during cardiac CT (58,62,63) seem to converge at an effective radiation dose of approximately 1 mSv for routine scanner settings and prospectively ECG-triggered acquisitions (primarily used for coronary calcium scoring) with electron-beam CT (3-mm beam collimation) and four–detector row CT (2.5 mm beam collimation). This corresponds to less than one-third of the exposure from natural sources that each individual in the United States receives every year (58). For high spatial resolution (1.00–1.25-mm beam collimation), a retrospectively ECG-gated acquisition (primarily used for contrast-enhanced CT of the heart), and routine scanner settings with four–detector row CT, an exposure limit of approximately 10 mSv is applied, which is two to three times the average annual background radiation in the United States and is comparable to the radiation exposure received during a typical routine diagnostic conventional coronary angiographic examination. As progressively thinner beam collimations are used for scanner types with added detector rows, radiation dose generally increases. However, the addition of detector elements should also improve tube output utilization, compared with that for current four–detector row CT scanners, and reduce the ratio of excess radiation dose that does not contribute to actual image generation (12). Radiation exposure will increase, however, with reduced spiral pitch and extension of the scan volume.

The radiation dose at retrospectively ECG-gated CT of the heart can be substantially reduced by means of online reduction of tube output in each cardiac cycle during phases that are of less importance for ECG-gated reconstruction ("ECG-gated dose modulation") (64). With this approach, the nominal tube output is applied only during the diastolic phase of the cardiac cycle, at which time images are most likely to be reconstructed. For the rest of the cardiac cycle, the tube output is reduced. Depending on the heart rate, an overall exposure savings of 30%–50% can be achieved without compromising image quality. In this way, it should be possible to reduce the radiation exposure for CT coronary angiography to a level of 5–7 mSv, depending on patient heart rate. Where available, therefore, this technology should be used for all patients in the absence of arrhythmia (51,63,64). ECG-gated dose modulation is limited to use in patients with steady heart rates and in whom the time point of the diastolic image reconstruction interval can be predicted with some certainty, since the reduced radiation level during the rest of the cardiac cycle may compromise diagnostic quality.

Contrast Medium Injection
Techniques for contrast medium bolus optimization have been developed in the past (65,66) but have not been widely used because reasonable results could be obtained by adapting strategies for contrast medium administration at single-section CT to dual– and four–detector row CT. However, the introduction of ever-faster CT acquisition techniques now requires careful custom tailoring of the bolus for achieving adequate, consistent, and homogeneous contrast attenuation over the entire course of the coronary arteries in order to facilitate attenuation-threshold–dependent two-dimensional and 3D depiction. Optimal contrast attenuation within the vessel is high enough to allow lesion detection but not so high that it obscures calcified coronary artery wall lesions with higher attenuation (ie, >350 HU).

For CT coronary angiography, an intravenous contrast medium injection can be adjusted by using either a test bolus or an automatic bolus-triggering technique (67). Since scanning times for imaging of the heart with eight– or 16–detector row CT scanners range from 20 to 40 seconds, 80–120 mL contrast medium injected at a rate of 3–5 mL/sec is needed to maintain homogeneous vascular contrast throughout the scan. Forty grams of iodine with a flow rate of 1 g/sec was shown to be a suitable protocol for consistent contrast enhancement of the coronary arteries, with an attenuation that allows assessment of both the contrast-enhanced vessel lumen and potential vessel wall lesions (68). Saline chasing (eg, bolus of 50 mL of saline injected immediately after the iodinated contrast medium bolus) has proved to be helpful for better contrast medium bolus utilization, for high and consistent vascular enhancement, and for prevention of streak artifacts (Fig 2), which frequently arise from dense contrast material in the superior vena cava and right atrium and sometimes interfere with the evaluation especially of the right coronary artery (69).

View larger version (135K): [in this window] [in a new window] [Download PPT slide]   Figure 2. Transverse contrast-enhanced 16-detector row CT scan obtained with retrospective ECG gating at the level of the mitral valve. Complete and homogeneous enhancement of left ventricle and coronary arteries can be achieved with a dedicated contrast medium protocol. Contrast medium has passed the right ventricle, and saline flush results in washout of contrast medium from the right ventricle. This reduces or prevents streak artifacts potentially arising from high-attenuating contrast material in the superior vena cava and right atrium, which may interfere with evaluation of the right coronary artery (arrow).

Maximum intensity projection.—For visualization of the coronary arterial tree at contrast-enhanced CT coronary angiography, maximum intensity projection (71) is widely used and recommended as a robust and easy to perform secondary tool for data visualization (10,34,36,37). Maximum intensity projections or other two-dimensional or 3D visualization methods (see below) for diagnosis not only display coronary artery CT data in a more intuitive format but also condense diagnostic information into a few relevant sections or views, if appropriate strategies are chosen. For routine visualization of large-volume CT coronary angiography data sets, many centers perform three dedicated maximum intensity projection reconstructions to create views of the left (Fig 3a) and right (Fig 3b) coronary arteries and of the entire coronary arterial tree from a cranio-oblique perspective ("spider view") (Fig 3c) (10,36).

View larger version (125K): [in this window] [in a new window] [Download PPT slide]   Figure 3a. Maximum intensity projections for routine visualization of large volume multi-detector row CT coronary angiography data sets. (a) Right anterior oblique view along the interventricular groove shows left anterior descending coronary artery (LAD), with mixed atherosclerotic lesion (arrowhead) with calcified components in the proximal course of the vessel. (b) Left anterior oblique view in plane connecting right coronary artery (RCA) and circumflex coronary artery along the atrioventricular groove shows right coronary artery, with calcified nodules (arrowheads) along the course of the vessel. (c) Left anterior oblique "spider" view along long axis of the heart shows course of the left anterior descending coronary artery (LAD) and its diagonal branches, with soft-tissue-attenuation plaque (arrowhead) in the anterior aspect of the left main coronary artery (LM) wall.

View larger version (140K): [in this window] [in a new window] [Download PPT slide]   Figure 3b. Maximum intensity projections for routine visualization of large volume multi-detector row CT coronary angiography data sets. (a) Right anterior oblique view along the interventricular groove shows left anterior descending coronary artery (LAD), with mixed atherosclerotic lesion (arrowhead) with calcified components in the proximal course of the vessel. (b) Left anterior oblique view in plane connecting right coronary artery (RCA) and circumflex coronary artery along the atrioventricular groove shows right coronary artery, with calcified nodules (arrowheads) along the course of the vessel. (c) Left anterior oblique "spider" view along long axis of the heart shows course of the left anterior descending coronary artery (LAD) and its diagonal branches, with soft-tissue-attenuation plaque (arrowhead) in the anterior aspect of the left main coronary artery (LM) wall.

View larger version (120K): [in this window] [in a new window] [Download PPT slide]   Figure 3c. Maximum intensity projections for routine visualization of large volume multi-detector row CT coronary angiography data sets. (a) Right anterior oblique view along the interventricular groove shows left anterior descending coronary artery (LAD), with mixed atherosclerotic lesion (arrowhead) with calcified components in the proximal course of the vessel. (b) Left anterior oblique view in plane connecting right coronary artery (RCA) and circumflex coronary artery along the atrioventricular groove shows right coronary artery, with calcified nodules (arrowheads) along the course of the vessel. (c) Left anterior oblique "spider" view along long axis of the heart shows course of the left anterior descending coronary artery (LAD) and its diagonal branches, with soft-tissue-attenuation plaque (arrowhead) in the anterior aspect of the left main coronary artery (LM) wall.

View larger version (138K): [in this window] [in a new window] [Download PPT slide]   Figure 4. Curved multiplanar reformation in oblique anterior coronal orientation from contrast-enhanced multi-detector row CT coronary angiography study allows visualization of course of the left anterior descending coronary artery in a patient with CAD. Note significant stenosis (arrow) caused by noncalcified coronary artery lesion proximal to a calcified nodule.

View larger version (86K): [in this window] [in a new window] [Download PPT slide]   Figure 5. Comparison of conventional angiography and contrast-enhanced four-detector row CT coronary angiography in one patient. Left: Anteroposterior cranial projection from conventional selective coronary angiography shows left anterior descending (LAD) and circumflex (Cx) coronary arteries. Right: Volume rendering in anteroposterior cranial projection shows left main coronary artery with its branches, the left anterior descending (LAD) and left circumflex (Cx) coronary arteries.

View larger version (94K): [in this window] [in a new window] [Download PPT slide]   Figure 6. Comparison of conventional angiography and contrast-enhanced four-detector row CT coronary angiography in one patient. Left: Right anterior oblique projection from conventional selective coronary angiography shows right coronary artery (RCA). Right: CT volume rendering shows right coronary artery (RCA) in 30° right anterior oblique projection.

View larger version (106K): [in this window] [in a new window] [Download PPT slide]   Figure 7. Contrast-enhanced 16-detector row CT coronary angiography. Colored volume rendering of right coronary artery (RCA) displayed in slightly cranial right anterior oblique perspective. This method of 3D postprocessing provides an intuitive display and conveys information on the often complicated anatomy of tortuous coronary arteries.

View larger version (133K): [in this window] [in a new window] [Download PPT slide]   Figure 8. Image in left craniolateral perspective was generated by using a dedicated software platform for automated segmentation of the coronary arterial tree from contrast-enhanced CT studies of the heart. On the basis of attenuation thresholds, the course of the left anterior descending coronary artery (LAD, highlighted) and its branches is automatically segmented from volume data.

View larger version (42K): [in this window] [in a new window] [Download PPT slide]   Figure 9. Image generated from same data set as in Figure 8. Intuitive visualization of left anterior descending coronary artery (LAD) is achieved by displaying curved multiplanar reformation of the segmented vessel along an automatically generated centerline. Note noncalcified soft-tissue-attenuation lesion (arrow) in the wall of the left main coronary artery (LM), causing significant stenosis.

	CLINICAL APPLICATIONS: IMAGING OF CORONARY CALCIUM

TOP ABSTRACT INTRODUCTION IMAGE ACQUISITION AND DISPLAY... CLINICAL APPLICATIONS: IMAGING... CLINICAL APPLICATIONS: CONTRAST... CT OF CAD: FUTURE... ESSENTIALS REFERENCES

For detection and quantification of coronary artery calcium, electron-beam CT (28), single-section spiral CT (49,75–77), dual–detector row CT (78,79), and multi–detector row CT (80–83) have been used. Scan data are acquired in the craniocaudal direction from the caudal part of the pulmonary arterial trunk to the apex of the heart (12-cm range). Both ECG-triggered sequential and ECG-gated spiral scanning (discussed earlier) have been used for ECG synchronization of the data acquisition.

For coronary calcium quantification with the use of electron-beam CT scanning, 3.0-mm-thick sections are typically acquired contiguously with prospective ECG triggering in middiastole and an exposure time of 100 msec per section. With four–detector row CT, an ECG-triggered acquisition covers a scan range of 12 cm in 20–25 seconds. ECG-gated spiral scanning provides shorter breath-hold times than does ECG-triggered scanning, and more consistent volume coverage because of the acquisition of overlapping sections (84). With ECG gating, a 12-cm range can be acquired in 15–20 seconds with four–detector row CT and in 6–10 seconds with eight– and 16–detector row CT. The spiral interpolation algorithms used for image reconstruction ordinarily generate 3-mm-thick sections (full width at half maximum), based on a section collimation of 2.5 mm for four– and eight–detector row CT scanners and 1.25–1.5 mm for 16–detector row CT scanners. A tube voltage of 120 kV is routinely applied. For most multi–detector row CT scanners, a 100-mA tube current is used to achieve sufficient signal-to-noise levels for detection of small calcified lesions. The tube current may be increased for obese patients (eg, to 150 mA) to maintain a diagnostic signal-to-noise level (85) at the expense of increased radiation exposure. With multi–detector row CT, every rotation covers a subvolume that consists of several adjacent sections, and the likelihood of interscan misregistration caused by heart movement in the z direction is thus reduced.

The modulation transfer function of the convolution kernel that is used for image reconstruction affects the in-plane spatial resolution and signal-to-noise ratio and thus strongly influences quantitative measurements. For coronary calcium scanning, a medium-sharp convolution kernel ordinarily is used that contains image noise with low-dose scanning techniques and provides an in-plane resolution of about 0.6 x 0.6 mm (12,86). Use of an edge-enhancing convolution kernel is generally avoided, because this may cause overestimation of scores and artifactual "lesions" in the pericardium close to the coronary arteries.

The equivalence of calcium measurements obtained with mechanical CT and with electron-beam CT has been investigated in the past. Reports on the agreement between single-section CT and electron-beam CT for calcium scoring have been conflicting (49,75–77,87), depending on the technology used. Since the introduction of multi–detector row CT, however, there seems to be agreement on the high correlation between and equivalence of results with this latter technology and those with electron-beam CT for coronary artery calcium measurements (80–83,88). It has been argued that databases developed with electron-beam CT studies on age- and sex-based calcium-score percentiles and risk profiles should not be applied to calcium scores obtained with mechanical CT scanners (89). Such notions may be hard to justify, however, given the close correlation of multi–detector row CT and electron-beam CT for measuring the same physical entity, calcified tissue in the wall of the coronary arteries.

The discussion on the equivalence of electron-beam CT and mechanical CT, however, has shed light on more fundamental problems of CT coronary artery calcium measurements—high variability, lack of standardization, and insufficient quality assurance (90–92).

The high interscan, interobserver, and intraobserver variability associated with this test has been recognized (93–98). Most important, this high variability has limited the use of coronary artery calcium measurements for tracking the progression of atherosclerosis (discussed below) (99–104). For electron-beam CT, the main strategy that has been employed to improve accuracy and reproducibility of coronary artery calcium measurements is optimization of the ECG interval chosen for prospectively triggered data acquisition (98,105–107). Reconstruction of overlapping images (with reconstruction increment less than section width) results in better accuracy and reproducibility for both electron-beam CT and mechanical spiral CT (84,108). The lowest interexamination variability and the highest reproducibility for coronary artery calcium measurements have been reported for retrospectively ECG-gated spiral scanning. Recent studies found interscan variability of about 10% or less for repeated retrospectively ECG-gated four–detector row CT scanning (82,84,109,110), which may be accurate enough to enable sensitive detection of changes in the total atherosclerotic disease burden in patients with and in those without specific therapy. In comparison to the prospectively ECG-triggered technique, however, CT acquisition with retrospective ECG gating is associated with a higher effective radiation dose (2.6–4.1 mSv) (58). With sophisticated technical devices, such as ECG-based tube current modulation, however, radiation exposure can be reduced to levels that are comparable to those of the prospectively ECG-triggered acquisition technique (63,64).

For the actual quantification of the amount of calcified tissue in the coronary arterial tree, the semiquantitative score based on section-by-section analysis of CT images, as initially described by Agatston et al (28), has traditionally been used. Authors of more recent studies (82,84,111,112) have described better results for interscan and inter- and intraobserver variability with use of quantitative measures (Ca⁺⁺ volume, Ca⁺⁺ mass) than those obtained with the traditional Agatston scoring method. Advanced software platforms (Fig 10) allow assessment of equivalent volume and total calcified plaque burden in terms of absolute calcium mass, based on actual scanner-specific calibration (82,84,112–114). This latter technique probably has the greatest potential to increase the accuracy, consistency, and reproducibility of coronary calcium assessment (112) and thus may replace traditional scoring methods in the future (92).

View larger version (80K): [in this window] [in a new window] [Download PPT slide]   Figure 10. Screenshot shows software platform for detection and quantification of coronary calcification. Lesions exceeding calcium threshold of 130 HU are identified with 3D-based picking and viewing tools and are assigned to left main (LM), left anterior descending (LAD), left circumflex (CX), or right (RCA) coronary arteries. Coronary calcification is quantified by means of Agatston score, calcium volume, and calcium mass. For calculation of calcium mass, calibration factors are established with phantom measurements and are used to adjust for different scan protocols. Quantitative measurements are displayed and reported in table format. Images shown are in a patient with calcifications in left anterior descending (yellow) and circumflex (blue) coronary arteries.

Another potentially important aspect of coronary artery calcium measurement may be behavior modification in an individual, with adoption of a healthier lifestyle if an elevated amount of coronary artery calcium is detected and made visible with CT imaging (119).

A possibly even more important future application of CT coronary artery calcium measurement may be monitoring of the response of total atherosclerotic plaque burden to statin (lipid-decreasing) therapy (99–104). A noninvasive tool to assess the therapeutic response to statin therapy appears all the more important in light of the potential risks associated with lipid-decreasing drugs that have recently become evident (120).

There have been attempts to use the presence and degree of coronary calcification as markers for the location and extent of stenotic disease (28,30,121) and for identification of patients at risk of hard cardiac events (eg, unstable angina, myocardial infarction, need for revascularization, coronary death) (29,30,122). However, the specificity of coronary calcium measurements for the assessment of the presence and extent of coronary artery stenosis appears to be only moderate, compared with cardiac catheterization (32). Also, early excitement sparked by the potential usefulness of CT coronary calcium measurement as a noninvasive tool for cardiac risk stratification has been tempered by the results of meta-analyses in which prognostic data on the positive predictive value of an elevated calcium score were pooled. According to these meta-analyses, which were based on available published results, there is only a moderately increased risk for hard cardiac events associated with coronary calcifications detected at CT in high-risk asymptomatic populations (32,123).

None of the available studies on the predictive value of coronary artery calcium measurements, however, are based on rigorously performed prospective cohort investigations. Therefore, the true value and usefulness of coronary artery calcium measurements for cardiac risk stratification, or the lack thereof, need to be demonstrated in large, prospective, population-based cohort trials. Three such trials have been launched, with the goal of prospective definition of the potential added prognostic value of coronary artery calcium measurements over traditional risk-factor assessment for cardiac risk stratification (31,124–126). It is hoped that, when these trials are finalized, subgroups can be better defined that might benefit from quantitative assessment of coronary calcium. There are indications that the greatest incremental value of CT coronary artery calcium imaging for risk stratification will most likely be found in the group of asymptomatic patients with an intermediate level of coronary risk (127,128).

Until more definitive results on the usefulness of coronary calcium scoring from cohort trials are available, the current role of CT coronary artery calcium measurements can be summarized as follows, in keeping with American College of Cardiology/American Heart Association expert consensus (32):

(a) A negative CT test makes the presence of atherosclerotic plaque, including unstable plaque, very unlikely. (b) A negative test is highly unlikely in the presence of significant luminal obstructive disease. (c) Negative tests occur in the majority of patients who have angiographically normal coronary arteries. (d) A negative test may be consistent with a low risk of a cardiovascular event in the next 2–5 years. (e) A positive CT confirms the presence of a coronary atherosclerotic plaque. (f) The greater the amount of calcium, the greater the likelihood of occlusive CAD, but there is not a 1-to-1 relationship, and findings may not be site specific. (g) The total amount of calcium correlates best with the total amount of atherosclerotic plaque, although the true "plaque burden" is underestimated. (h) A high calcium score may be consistent with moderate to high risk of a cardiovascular event within the next 2–5 years.

Thus, the degree of coronary artery calcifications may be considered an additional risk factor and, depending on the outcomes of population-based studies currently underway, as such may become part of the traditional Framingham risk-stratification scheme in the future.

More multifaceted current considerations with regard to coronary artery calcium mainly focus on the role of calcium in the pathogenesis of atherosclerotic disease. Coronary atherosclerosis is a systemic disease process. The presence and extent of coronary artery calcifications may be considered indicative of the total disease burden of atherosclerotic lesions ("plaques") in a given individual and, thus, of the likelihood of the presence of potentially unstable "vulnerable" plaques that may trigger an acute coronary event. Histopathologic studies have demonstrated that calcium is a frequent feature of ruptured plaques (ie, culprit lesions associated with acute coronary syndromes), but the presence or absence of calcium does not allow for a reliable distinction between unstable and stable plaques (ie, lesions that do or do not cause acute coronary events) (129–132).

Although coronary calcium is intimately associated with coronary atherosclerotic plaque development, it generally represents a more advanced stage of vascular remodeling in response to atherosclerotic lesions (129). Earlier and more active stages of coronary atherosclerosis, however, appear more frequently associated with noncalcified or mixed-composition plaques consisting of accumulations of extracellular lipid and fibrous tissue (133–137). This may serve to explain the results of clinical studies whose authors argued that acute coronary syndromes are more frequently associated with lesser amounts of coronary calcium and that the presence of more extensive calcification is more characteristic of stable CAD (138–140).

	CLINICAL APPLICATIONS: CONTRAST-ENHANCED CT CORONARY ANGIOGRAPHY

TOP ABSTRACT INTRODUCTION IMAGE ACQUISITION AND DISPLAY... CLINICAL APPLICATIONS: IMAGING... CLINICAL APPLICATIONS: CONTRAST... CT OF CAD: FUTURE... ESSENTIALS REFERENCES

 
CT Coronary Angiography: Technique
Except for contrast medium injection, the strategy for performing contrast-enhanced CT coronary angiography with electron-beam CT does not substantially differ from that for nonenhanced techniques for imaging of coronary artery calcium (100-msec temporal resolution, 3-mm collimation) (33).

CT coronary angiography with multi–detector row CT is ordinarily performed with retrospectively ECG-gated, thin-section spiral scan protocols. Optimization of scan protocols with respect to radiation exposure (discussed previously) is particularly important for contrast-enhanced CT coronary angiography. Sufficiently high spatial resolution and low contrast resolution have to be achieved in normal-sized, as well as larger, patients, with the minimal possible radiation exposure. For CT coronary angiography with four and eight–detector row CT with a section width of 1.3 mm and a rotation time of 500 msec, a tube voltage of 120 kV with a tube current of around 300 mA should be used. The tube current needs to be increased to around 350–400 mA for 16–detector row CT with submillimeter section collimation and faster rotation time.

The overall diagnostic quality of noninvasive CT coronary angiography is largely dependent on the spatial resolution, patient heart rate during the examination, choice of the appropriate reconstruction time point within the cardiac cycle, and contrast enhancement (see above). The coronary arteries and disease manifestation within these vessels are minute and elusive targets for imaging. Thus, the spatial resolution at image acquisition is a crucial factor in the determination of the likelihood with which a lesion (eg, stenosis) can be detected and evaluated with noninvasive imaging techniques. Recent 16–detector row CT scanners provide high-spatial-resolution cardiac scan protocols with submillimeter (0.5–0.75-mm) section collimation and an in-plane spatial resolution of up to 0.5 x 0.5 mm (12–14). For image reconstruction, sections not thinner than 1.0 mm are usually reconstructed with 0.6 x 0.6-mm in-plane resolution to achieve optimal contrast resolution (13).

The duration of the diastolic phase with little cardiac motion is directly related to the heart rate. Since retrospectively ECG-gated CT requires a defined period of time per heart cycle for data acquisition, the heart rate plays an important role in image quality with respect to motion artifacts at high-spatial-resolution CT coronary angiography. Therefore authors of most studies agree that patient heart rate at CT coronary angiography is inversely related to image quality (36,56,141–143). Studies based on four–detector row CT with a 500-msec rotation time found that the upper limit of the heart rate at which motion artifacts can be consistently minimized is 65–75 beats per minute (10,36,56,141–144). At higher heart rates, adequate image quality can also be achieved, but overall results are less consistent and reproducible. For scanners with a faster (<500-msec) rotation time, the robustness of image acquisition with regard to cardiac motion is increased (37); however, slow heart rates are still required to achieve consistently high image quality. Thus, it is recommended that the heart rate of patients undergoing CT coronary angiography be slowed pharmacologically (ie, with oral administration of beta-blockers) if contraindications to such a regimen have been ruled out (10,36,38).

The motion pattern of the left anterior descending and circumflex coronary arteries (Fig 5) tends to follow the left ventricular contraction, whereas the right coronary artery (Figs 6, 7) moves synchronously with the right side of the heart (ie, the right atrium). Because of these different motion patterns, different reconstruction time points over the cardiac cycle can result in optimal display of different coronary arteries (56,57,145). The reliability of multi–detector row CT coronary angiography in patients with cardiac arrhythmia is limited. However, misinterpretation of the ECG signal can be partially compensated for by means of retrospective editing of the ECG trace (discussed earlier). A persistent irregular heart rate, though, such as that in patients with atrial fibrillation, result in interscan discontinuities that prohibit evaluation of CT coronary angiographic images for coronary artery stenosis.

Presence of severe calcification is a limitation of contrast-enhanced CT coronary angiography because beam-hardening and partial-volume effects can completely obscure the cross section of the vessel and prevent assessment of the patency of the coronary artery lumen. Owing to similar effects, metal objects such as stents (discussed in next section), surgical clips, and sternal wires can also interfere with the evaluation of underlying structures. Use of the thinnest possible section width reduces partial-volume artifacts to some extent and improves assessment of calcified coronary segments. In addition, dedicated filtering could be beneficial for more reliable imaging of calcified vessels in the future.

Contrast-enhanced CT of Coronary Artery Anomalies, Bypass Grafts, and Stents
The true prevalence of coronary artery anomalies (Figs 11–13) is difficult to establish (146); however, specific variations (eg, coursing of coronary artery between pulmonary outflow tract and aorta [Fig 11]) of the origin and course of the coronary artery have been recognized as a not infrequent cause of unexpected sudden cardiac death, especially in young athletes (147,148). Cross-sectional imaging has been recognized as the preferred diagnostic strategy for the evaluation of coronary artery anomalies, since the information that is to be obtained from volumetric imaging on the origin and anatomic course of aberrant vessels far surpasses that from conventional angiography (149–153) (Fig 12). MR imaging is limited with regard to determination of the distal coronary arterial course (154). Therefore, CT is the preferred modality for evaluation of small collateral vessels (Fig 12), fistulas (Fig 13), and vessels originating outside the normal sinuses (eg, from a pulmonary artery) (155).

View larger version (120K): [in this window] [in a new window] [Download PPT slide]   Figure 11. Contrast-enhanced four-detector row CT coronary angiography. Transverse thin-section maximum intensity projection reconstructed at level of the aortic valve (AV) shows coronary artery anomaly, with three coronary arteries arising from the origin of the right coronary artery (arrow) and supplying both left and right vascular territories of the myocardium. Both coronary arteries supplying left side of the heart (arrowheads) are compressed between the left ventricular outflow tract and adjacent cardiac cavities.

View larger version (183K): [in this window] [in a new window] [Download PPT slide]   Figure 12a. Patient with superdominant anomalous right coronary artery (AnRCA) supplying the majority of the myocardium. (a) Selective conventional angiographic image and (b) volume-rendered 3D reconstruction (cranial right anterior oblique perspective) from contrast-enhanced 16-detector row CT coronary angiography. Anomalous right coronary artery gives rise to two side branches (arrowheads), which cross over to left anterior surface of the heart, connecting anomalous right coronary artery with left anterior descending (LAD) coronary arterial territory. Native right coronary artery (RCA) is also visualized in its normal anatomic course but is of similar small caliber as left anterior descending artery.

View larger version (128K): [in this window] [in a new window] [Download PPT slide]   Figure 12b. Patient with superdominant anomalous right coronary artery (AnRCA) supplying the majority of the myocardium. (a) Selective conventional angiographic image and (b) volume-rendered 3D reconstruction (cranial right anterior oblique perspective) from contrast-enhanced 16-detector row CT coronary angiography. Anomalous right coronary artery gives rise to two side branches (arrowheads), which cross over to left anterior surface of the heart, connecting anomalous right coronary artery with left anterior descending (LAD) coronary arterial territory. Native right coronary artery (RCA) is also visualized in its normal anatomic course but is of similar small caliber as left anterior descending artery.

View larger version (145K): [in this window] [in a new window] [Download PPT slide]   Figure 13a. (a) Selective conventional coronary angiographic image in right anterior oblique projection and (b) volume-rendered 3D depiction (left posterior perspective) from contrast-enhanced four-detector row CT coronary angiography show coronary artery fistula (arrows) arising from left circumflex coronary artery and connecting to right atrium.

View larger version (125K): [in this window] [in a new window] [Download PPT slide]   Figure 13b. (a) Selective conventional coronary angiographic image in right anterior oblique projection and (b) volume-rendered 3D depiction (left posterior perspective) from contrast-enhanced four-detector row CT coronary angiography show coronary artery fistula (arrows) arising from left circumflex coronary artery and connecting to right atrium.

View larger version (109K): [in this window] [in a new window] [Download PPT slide]   Figure 14. Colored volume-rendered view from anterior perspective, derived from 16-detector row CT angiography, in a patient with three venous bypass grafts to left anterior descending (VCABG-LAD), circumflex (VCABG-Cx), and right (VCABG-RCA) coronary arterial territories and an additional left internal mammary artery bypass graft (LIMA-BG), also to the left anterior descending coronary artery (LAD).

View larger version (176K): [in this window] [in a new window] [Download PPT slide]   Figure 15. Colored 3D volume-rendered view from left anterior oblique perspective, derived from 16-detector row CT coronary angiography, enables visualization of native and graft vessels in their relationship to surrounding thoracic anatomy in a patient with left internal mammary artery (LIMA) bypass graft. Anastomosis has been created between left internal mammary artery and left anterior descending coronary artery (LAD) territory. Note extensive atherosclerotic changes in the native vessels.

Noninvasive imaging of coronary artery stents, stent patency, and in-stent restenosis is another attractive potential application of fast CT techniques (170–173), since the number of patients who undergo coronary angioplasty with stent implantation is rapidly increasing. However, coronary stents have been notoriously difficult to assess with CT. Contrast-enhanced CT can be used to assess stent patency on the basis of contrast enhancement in the course of the artery with the stent, because an unenhanced distal coronary artery lumen usually reflects critical in-stent restenosis (171). However, assessment of the stent lumen for nonocclusive in-stent restenosis due to neointimal hyperplasia remains challenging. Recent improvements in spatial resolution have improved the assessment of the stent lumen (Fig 16). However, improved spatial resolution can only partially compensate for metal artifacts arising from stent struts, which exaggerate the actual size of the stent and obscure subtle in-stent abnormalities within the lumen (172). The clinical value of CT after stent placement is, therefore, currently largely limited to the detection of stent occlusion.

View larger version (124K): [in this window] [in a new window] [Download PPT slide]   Figure 16a. Contrast-enhanced 16-detector row CT coronary angiography in a patient who has undergone percutaneous transluminal coronary angioplasty with stent placement in right coronary artery (RCA). (a) Colored 3D volume-rendered view from right posterior oblique perspective reveals luminal narrowing (arrowhead) of artery proximal to the stent. (b) Maximum intensity projection and (c) multiplanar reformation in oblique coronal planes show patent stent lumen and mixed atherosclerotic lesion (arrow) with calcified and noncalcified components as the cause of high-grade (70%) stenosis proximal to the stent. (d) Conventional angiographic image in left anterior oblique projection confirms stent patency and presence of stenosis (arrow) but fails to elucidate nature of the lesion causing luminal narrowing. (Case courtesy of C. S. Soo, MD, HSC Medical Center, Kuala Lumpur, Malaysia.)

View larger version (161K): [in this window] [in a new window] [Download PPT slide]   Figure 16b. Contrast-enhanced 16-detector row CT coronary angiography in a patient who has undergone percutaneous transluminal coronary angioplasty with stent placement in right coronary artery (RCA). (a) Colored 3D volume-rendered view from right posterior oblique perspective reveals luminal narrowing (arrowhead) of artery proximal to the stent. (b) Maximum intensity projection and (c) multiplanar reformation in oblique coronal planes show patent stent lumen and mixed atherosclerotic lesion (arrow) with calcified and noncalcified components as the cause of high-grade (70%) stenosis proximal to the stent. (d) Conventional angiographic image in left anterior oblique projection confirms stent patency and presence of stenosis (arrow) but fails to elucidate nature of the lesion causing luminal narrowing. (Case courtesy of C. S. Soo, MD, HSC Medical Center, Kuala Lumpur, Malaysia.)

View larger version (158K): [in this window]