Coronary artery disease (CAD) continues to be a leading cause of mortality and morbidity, stymieing screening approaches to predict risk for future debilitating events, despite intensive investigation and signiﬁcant interventional advances. More than 1 million Americans and nearly 20 million people worldwide suffer from acute coronary syndrome (ACS).
currently a Resident in the Department of Radiology at the
University of California, San Francisco, CA (UCSF). He received
his medical and graduate training at Boston University School of
Medicine, graduating in 2003, and served as a medical intern at
UCSF. His basic science research interests include investigating
what signals cause a normal cell to proliferate in the setting of
inflammation as well as tumorigenesis, with a mind toward
developing small molecule pharmaceuticals and imaging probes to
monitor and possibly modulate dysregulated growth. His current
clinical focus is on exploring the potential role of imaging in
assessing the permeability of vessels in response to injury.
The atheroma has a central role in coronary artery disease
pathophysiology, focusing interest on noninvasive imaging for
detecting at-risk populations that are harboring vulnerable
plaque. Coronary artery calcium scoring with computed tomography
(CT) is already an established technique, awaiting formal
validation by prospective trials. Cardiac CT angiography (CTA)
can assess for significant luminal stenosis and identify
nonstenotic atherosclerotic plaques. Therefore, alone or in
combination with other modalities, cardiac CTA will potentially
be useful for risk stratification of asymptomatic individuals,
but clinical applications are not yet supported by rigorous
outcome data. Noninvasive characterization and quantification of
plaque burden is likely to have important implications in
prevention of coronary artery disease progression.
Coronary artery disease (CAD) continues to be a leading cause of
mortality and morbidity, stymieing screening approaches to predict
risk for future debilitating events, despite intensive
investigation and significant interventional advances. More than 1
million Americans and nearly 20 million people worldwide suffer
from acute coronary syndrome (ACS).
The estimated healthcare cost to manage end-stage cardiovascular
disease is in excess of $350 billion, which does not account for
the associated cost of loss of productivity.
Therefore, it is very important to develop effective approaches to
identify early CAD.
Rapid progress in cardiac computed tomography (CT) now enables
not only the detection and quantification of coronary artery
calcification (CAC) but also the grading of coronary stenosis.
The contrast resolution afforded by CT potentially may permit
assessment of the blood vessel wall, which would allow physicians
to shift the emphasis in CAD from the "end game"-the
hemodynamically significant lesion-to a more remedial point earlier
in the natural history of the disease. Moreover, since ACS is
believed to be caused by atherosclerotic plaque rupture, CT
evaluation of coronary artery plaque may yield a more effective
approach to risk stratification than with current strategies, which
rely more on assessment of clinical risk factors such as age,
gender, family history, hypertension, dyslipidemia, diabetes, and
The coronary artery plaque that lies at the core of the disease
process has been the subject of intensive research, which has led
to major inroads into understanding the molecular and cellular
underpinnings of atherogenesis and plaque progression.
Inflammation has a central role, with the macrophage as the key
cellular inflammatory mediator of the atheroma, implicated in each
phase of atherogenesis, including plaque initiation, progression,
Initially, circulating monocytes adhere to a region of
dysfunctional endothelium along the arterial wall, maturing into
macrophages that phagocytose lipoproteins, and then morphing into
the well-known foam cells that form the early atheroma. Macrophages
that die in the atheroma contribute to a paucicellular lipid core.
The secretion of proteases by remaining macrophages and other
immune cells can cause weakening of connective tissue in the
atheroma that could predispose to hemorrhage and even frank
disruption of the fibrous cap, eventually leading to plaque erosion
and rupture. Recruitment of other cellular mediators can induce
neovascularization. The aggregate of these changes ultimately can
manifest as dystrophic calcification that can be detected as
calcified plaque and luminal narrowing.
For decades, the paradigm for pathology in CAD focused on the
degree of luminal stenosis, in part because the primary imaging
examination for CAD was angiography and also because clinical
success was noted following revascularization procedures that
targeted critical coronary arterial stenoses. In short, CAD
severity was viewed as dependent on the degree of luminal
narrowing, and atherosclerosis was primarily thought of as a focal
disorder. However, more recent work has triggered a shift in this
thinking and has elucidated the importance of arterial remodeling,
suggesting a new way to improve patient outcomes. It is now
believed that, for much of its life history, the atheroma expands
extrinsically from the vessel lumen, rather than inward, an
adaptive process called "positive remodeling."
The implication is that a nontrivial atherosclerotic plaque burden
can manifest even in the absence of stenosis. Ergo, substantial
disease could exist but remain occult to angiography. Newer
techniques, such as intravascular ultrasound (IVUS), have
substantiated this hypothesis.
Indeed, by the time coronary angiography reveals flow-limiting
stenosis, intimal atherosclerosis is typically present in a diffuse
distribution. Intravascular ultrasound studies have shown
atherosclerotic lesions even in adolescent and young adult
Much research has targeted the so-called "vulnerable plaque," or
the atherosclerotic lesion that is thrombosis-prone, which has a
high probability of undergoing rapid progression and rupturing. The
vulnerable plaque is variably defined but is often described as
evincing a large lipid core, thin fibrous cap, positive remodeling,
fissuring, and/or active inflamma-
The current gold standard for detecting this vulnerable plaque
remains IVUS, but its invasiveness often makes it an inappropriate
tool for asymptomatic patient evaluation.
Therefore, given the current understanding of atherosclerosis,
the foremost goal in noninvasive imaging is to prospectively
identify patients who harbor vulnerable plaque. However, despite
the pace of progress, the ultimate role in predicting cardiac risk
still remains uncertain. The anticipation is high that cardiac CT
potentially will be useful for risk stratification of asymptomatic
individuals, but clinical applications are not yet supported by any
convincing outcomes data.
Current clinical prediction tools have limited accuracy, with up
to 20% of cardiovascular events remaining unexplained by
traditional risk factors.
This prognostic gap may be narrowed by including more recently
recognized coronary risk factors, such as homocysteine, and
inflammatory markers, such as C-reactive protein. Additional
biomarkers, such as coronary calcification and noncalcified plaque
on noninvasive imaging, are also likely to impact cardiovascular
CT and calcified plaque
Earliest imaging of CAD focused on the calcified plaque, which
could be easily detected by plain-film radiography and fluoroscopy.
With the advent in the early 1980s of electron-beam CT (EBCT),
which has excellent temporal resolution, coronary calcification
imaging came into clinical use at various centers, and several
calcium scoring protocols evolved. An increasing number of
prospective studies have shown that the absence of coronary calcium
clearly indicates a low cardiovascular event risk, and the amount
of calcium is directly correlated to the risk of a cardiovascular
event such as myocardial infarction or stroke.
The Agatston scoring system was developed for EBCT
and has been widely used as the reference standard in quantifying
calcified coronary atherosclerotic plaque.
It is defined as the calcified lesion area multiplied by a scaling
coefficient that is determined from the peak attenuation in the
plaque. A constant attenuation threshold minimum of 130 HU is used
to distinguish noncalcified from calcified coronary artery lesions.
While the Agatston score provided good inter- and intraobserver
agreement, the interscan variability of this system is notable,
probably in part because of the variation in the scaling cofactor.
Additional approaches are used to derive a calcium score,
including mass- and volume-based approaches, which provide
estimates of the absolute mass or volume, respectively, of the
calcium in the lesion.
Regardless of the method, a calcium score is generated, and the
patient is provided with an age-related risk. Risk increases
exponentially with increasing calcium in the coronary arteries. The
absence of calcification is normal, small amounts raise the risk
mildly, and an incrementally higher score shifts the risk upward
significantly. Because much of the existing literature has been
based on the Agatston score, most multidetector CT (MDCT) scanners
can generate a score with any number of existing approaches,
including the Agatston system.
A high coronary calcium score is a sensitive, but not specific,
marker for obstructive CAD.
Indeed, coronary calcium can be found even in the absence of any
clinical coronary symptoms.
Whether coronary calcium scoring will be shown to be more accurate
than conventional clinical prediction tools remains to be
determined. Randomized large prospective studies are under way,
with midstream results from one study, the Prospective Army
Coronary Calcium trial (PACC) suggesting that, at least in
asymptomatic, healthy young individuals, CAC screening with EBCT
will prove to be cost-effective.
The 7-year Multi-Ethnic Study of Atherosclerosis (MESA) study is
another such trial, initiated in 2000 to investigate the parameters
of subclinical cardiovascular disease in 6500 men and women. It
prospectively evaluates EBCT or MDCT to determine the viability of
imaging CAC as a predictor of cardiovascular events.
When the outcomes of these trials are known, appropriately designed
cost-benefit analyses can be initiated, thereby addressing the
dearth of extant studies focusing on the cost-effectiveness of
coronary calcium scoring and coronary artery plaque assessment.
Another large study population is examined in the CARDIA study,
formally known as the Coronary Artery Risk Development in Young
Adults study, which is a population-based cohort study begun 2
decades ago with young black and white men and women that included
the CT measurement of calcified coronary artery plaque.
At present, consensus guidelines do not yet endorse routine CT
screening for CAC in the asymptomatic population but suggest that
screening may be warranted in select intermediate-risk subgroups.
With the introduction of electrocardiographically (ECG)-gated
MDCT, a temporal resolution of ≤200 msec made coronary imaging
feasible. It was important to demonstrate that MDCT and EBCT
results were comparable, thereby allowing investigators to leverage
the clinical data that had accumulated for calcium scoring based on
EBCT results (Figure 1). Although these measurements are dependent
on multiple technical factors, one large randomized control trial
(MESA) has yielded preliminary results that suggest that EBCT and
MDCT calcium scores are comparable, and several studies have shown
Additionally, because the signal-to-noise ratio is higher for any
given radiation dose with MDCT than with EBCT, MDCT is more useful
for imaging findings other than coronary artery calcifications,
including the coronary artery lumen (using contrast material), and
structures outside the heart, such as the aorta and the pulmonary
arteries (Figure 2). Thus, there is interest in performing
contrast-enhanced coronary CT angiography (CTA) in addition to
calcium scoring to evaluate patients for atherosclerosis, because
of the additional information on coronary artery caliber and
Like its now-established role in other vascular circulatory
beds, CTA now appears to have a significant role in the assessment
of coronary vasculature (Figure 3). Of the 3 to 4 million
conventional X-ray coronary angiograms that are undertaken annually
in the United States, two thirds are diagnostic studies, performed
mainly with the aim of verifying and quantifying the presence and
extent of CAD.
Unlike X-ray coronary angiography, which some observers have
characterized as "lumenology" or "lumenog- raphy," CT provides
concomitant assessment of the vessel wall. Assessing luminal
caliber with conventional angiography fails to provide a complete
portrait of plaque accumulation, given the ability of the vessel to
adapt by positive remodeling (Figure 4).
At our institution, a noncontrast ECG-gated coronary
calcium-scoring scan is performed as the first part of the coronary
CT examination. After appropriate timing-bolus optimization is
carried out, the ECG-gated CTA component is performed using a
dual-head iodinated contrast/saline injector to ensure a compact
bolus. Retrospective gating of the data is preferred over
prospective gating because it has fewer problems with motion
artifact. All studies are interpreted via the source axial data
set, with 3-dimensional and curved planar reconstructions also
generated for problem solving and for use in consultation with
referring physicians. Other centers will routinely perform image
postprocessing and employ multiplanar reconstructions,
maximum-intensity projections, and 4-dimensional capabilities in
the initial assessment, not just for problem solving or
consultation. With respect to patient preparation for the study,
the general approach taken at UCSF for coronary CTA on a 64-slice
MDCT is to provide pharmacologic heart rate control if there are no
clinical contraindications to beta-blockade, by administering
atenolol orally and/or metoprolol intravenously, as well as
sublingual nitroglycerin for dilation of the vasculature. Important
contraindications include severe asthma, heart block,
hypersensitivity, decompensated cardiac failure, and nitrate usage
(eg, sildenafil or tadalafil) in the preceding 24 to 48 hours. In
our experience, cardiac motion can still decrease the diagnostic
yield of the study if the heart rate is >65 beats per minute,
despite the temporal resolution of the 64-slice scanner.
While initial results suggest that the negative predictive value
of coronary CTA is outstanding,
further work may be necessary to improve its ability to detect
atherosclerosis. Early feasibility studies that used 16-slice CT to
evaluate coronary artery segments in patients scheduled for
conventional angiography found that CT compared favorably in
detecting greater than 50% luminal stenosis, with 95% sensitivity,
86% specificity, negative predictive value of 97%, and positive
predictive value of 80%.
Additional studies corroborated these findings, with positive
predictive values of near 80% and an accuracy >90% in detecting
The investigators reported that approximately 12% of the arterial
segments were unevaluable because of vessel motion or extensive
calcification, likely secondary to the technical constraints of
Studies with 64-detector CT technology, which provides less
motion artifact and improved resolution, have shown greater
accuracy in assessing stenosis and in the characterization of
noncalcified plaque. Leber et al
employed 64-detector CT to evaluate plaque in the proximal coronary
arteries divided into 3-mm sections, with intravascular ultrasound
as the gold standard.
This study reported accurate detection in 83% of the sections that
contained noncalcified plaque, 94% of the sections that contained
mixed plaque, and 95% of the sections that contained calcified
plaque. In 94% of the sections, the absence of atherosclerotic
lesions was also correctly scored. Importantly, the 64-slice MDCT
also detected 70% of the sections with a lipid-rich pool and 90% of
the sections harboring spotty calcifications, characteristic
features that are suggestive of vulnerable plaques.
Yet there is no question that the inherently limited contrast
resolution of CT for differentiating types of tissue affects its
ability to characterize noncalcified plaque. Other imaging
modalities with greater soft tissue contrast, such as IVUS and
magnetic resonance imaging, may provide better evaluation of
noncalcified plaque, but ECG-gated 64-detector CT possesses both
the speed and spatial resolution to investigate the entire coronary
artery tree (Figure 5), giving it significant advantages as a
diagnostic modality. Currently, coronary CTA does not reliably
differentiate between lipid-rich lesions and fibrous tissue within
a coronary artery lesion,
although several initial papers have shown optimistic results when
comparing the ability of CT to assess the internal composition of
atheromas with the ability of IVUS,
finding correlation between plaques with low CT densities and
low-echogenicity lesions. Nonetheless, sensitivity in detecting
noncalcified plaques on a lesion-by-lesion basis for CTA is only
slightly >50% when compared with IVUS,
reinforcing the point that, at this juncture, routine use of
coronary CTA for evaluation of asymptomatic patients cannot be
recommended, especially given the risks of radiation and contrast
However, determining the burden of noncalcified plaque in the
symptomatic population may be useful, particularly since
noncalcified lesions in coronary arteries may represent a thrombus.
Indeed, in symptomatic patients with low pretest probability for
disease, the impressive negative predictive value of CTA can be
useful in deciding whether invasive cardiac catheterization is
needed. Furthermore, it may be that the different types of plaque
may present with dissimilar clinical pictures. In one recent study,
noncalcified plaques were seen more predominantly in patients
presenting with acute myocardial infarction, while those pa-tients
with stable angina were noted to have calcified plaques.
With continued progress of technology, reduced volume-averaging
artifacts, and improved temporal resolution, it may be possible to
improve measurement of different tissue attenuations and to
characterize noncalcified plaque more reliably. One of the most
promising approaches in development appears to be dual-source CT,
which uses 2 X-ray tubes and detectors mounted onto the rotating
gantry at an angular offset of 90˚. This class of scanner will
reduce the temporal resolution to the 50 to 100 msec range from the
current 64-slice standard of approximately 120 to 180 msec; indeed,
initial results in a small series of patients have been reported to
have a temporal resolution of 83 msec in coronary CTA using
ECG-gated single reconstruction, with visualization of 98% of
coronary artery segments without motion artifact.
An additional advantage of the dual-source CT is that the
differential energy sources may permit novel approaches to
subtracting out signal from calcium or stent and, thus, enable
finer evaluation of noncalcified plaque.
Cardiac CT is also being investigated for its ability to assess
response to medical therapy, such as with statins, particularly in
light of the potential risks from ongoing medical therapy, such as
liver abnormalities and myositis, as well as the high cost of
A study from Callister et al,
which used coronary calcium scores that were acquired with EBCT to
evaluate response to statin therapy, offered promising results.
However, although these researchers were able to track stability
and, in some cases, regression of plaque burden,
a main criticism was interscan variability in calcium scores. Newer
MDCT approaches will likely provide a more reliable interscan
measurement, as suggested by early results from the MESA and CARDIA
studies. Therefore, if CTA is eventually able to effectively assess
noncalcified plaque, this is the lesion that should show the most
impact from statin therapy. Finally, detection of vulnerable plaque
in high-risk patients or those refractory to medical management may
trigger invasive therapies such as plaque sealing or stenting.
The potential role of CTA in assessing response to therapy hints
at the promise of personalized medicine in the 21st century. But
the most interesting potential developments lie in its use as a
noninvasive screening tool in high-risk, asymptomatic patients and
in the initial evaluation of a patient presenting acutely with
atypical chest pain.
The conclusions of ongoing prospective cohort trials, such as MESA
and PACC, are eagerly anticipated, and, if favorable, will mean
that the assessment of calcified plaque would be incorporated into
existing risk-stratification and prevention guidelines.
Furthermore, because atherosclerotic disease is not confined to the
coronary bed, the detection of noncalcified coronary plaques might
also suggest associated carotid and/or aortic disease. Therefore,
finding vulnerable patients with ACS who have lesions prone to
rupture would also identify patients at potential risk for
dissection, mesenteric ischemia, and stroke. Lastly, as genetic
profiling of the risk of acute myocardial infarction becomes a
reality, there may be a place for evaluation of the early stages of
atherosclerosis in high-risk patient populations so therapy may be
initiated at a young age. This will be a challenge for MDCT
technology to optimize visualization of the arterial wall and
plaque characterization. The future prospects of non- invasive CT
imaging of coronary artery plaques are exciting, as the pace of
technologic advancements will move investigators beyond feasibility
studies to trials that address how modern high spatial and temporal
resolution MDCT techniques can be cost-effectively integrated into
the evaluation and management of CAD.
The author thanks Dr. Charles Higgins and Dr. Gautham Reddy for
their insightful comments and guidance. He also thanks Dr. Karen
Ordovas for her able assistance in preparing the accompanying