INTRODUCTION

Assessing atherosclerotic plaque burden has long been the goal of cardiologists, considering that the initial presentation of a significant percentage of previously asymptomatic patients is with myocardial infarction (MI) or sudden cardiac death.¹ Large epidemiologic studies have identified a host of “conventional” risk factors that are able to predict only about two-thirds of those patients who will eventually go on to develop coronary artery disease (CAD).² Nearly a third of patients dying of CAD in the United States are classified as being at low risk with the Framingham Risk Index score.³

It is now well known that most of the acute coronary events result from the rupture and/or erosion of non-stenotic but vulnerable plaques.^4,5 To date, invasive coronary angiography remains the gold standard for imaging the coronary arteries and intravascular ultrasonography (IVUS) is the investigation of choice to accurately assess the total atherosclerotic plaque burden and plaque morphology. With the attending risk (though small) involved with an non-invasive technique to image the coronary arteries and accurately assess the total atherosclerotic plaque burden. The exponential improvements in imaging technology in the last 2 decades, most notably the improvements in computed tomography (CT) and the rapid growth of computer technology, have made this goal a reality. There has been a plethora of articles published on the subject of detection of coronary artery calcification (CAC) using electron beam CT (EBCT) and multislice CT (MSCT) and its value in prognosis and risk stratification, particularly in asymptomatic patients or in those patients with a low to intermediate probability of developing CAD.

TYPES OF ARTERIAL CALCIFICATION

Two distinct types of arterial calcification have been described: medial arterial calcification and calcification associated with atherosclerotic plaque.⁶ These two types of calcification differ in terms of their pathophysiology, morphologic features, and clinical significance. In atherosclerotic plaque, calcium deposits are typically found in the intima. In contrast, medial calcification, also known as Mönckeberg’s sclerosis, occurs independently of atherosclerosis. Medial arterial calcification (MAC) occurs in the small and medium sized muscular arteries. It is typically noted in the arteries of the abdominal viscera, breast and the thyroid gland. Medial calcification of the coronary arteries, though rare, has been noted in patients with advanced diabetes and/or chronic kidney disease. MAC is not associated with luminal stenosis but causes stiffening and decreased compliance of the arteries.

PATHOGENESIS OF CORONARY ARTERY CALCIFICATION

Atherosclerosis is known to be a chronic inflammatory process where various components of the immune system are implicated.^7,8 In the majority of cases, the presence of a high plasma lipid concentration fuels the process of atherosclerosis. Endothelial cells, leukocytes, and intimal smooth muscle cells, as well as other contributory factors such as smoking, hypertension, diabetes, and in some cases, high homocysteine levels, all play an important part in the process of atherosclerosis, but the exact nature of their interactions with each other is yet to be fully unraveled.

There have been different views regarding the mechanism of calcium deposition in atherosclerotic plaques.⁹ Some initial theories suggest that calcification results from passive adsorption of Gla-containing proteins with a high affinity for calcium phosphate and hydroxyapatite, whose only known function is to bind calcium.^10,11 This seems unlikely in light of the fact that calcification occurs in only those vessels with atherosclerosis and is absent in normal arteries.¹² Other evidence suggests that coronary calcification is an actively regulated process rather than passive adsorption and precipitation.¹³ In fact, several intriguing similarities have been noted between coronary calcification and bone formation.¹⁴

Calcium deposition within an atherosclerotic plaque lesion starts as early as the second decade of life, when the plaque lesion is no more than a fatty streak¹⁵ (type III plaque lesion). When calcification becomes widespread, it is then classified as a type IV lesion, which contains granules of calcium within the smooth muscle cells and the extracellular matrix. The process of calcification starts in the damaged intracellular organelles of smooth muscle cells that subsequently become apoptotic and act as niduses of further calcification.¹⁶ The extracellular accumulation of lipids within the plaque deep within the intima appears to be the starting point for this process.¹⁷ The small granules of calcium fuse together to form larger deposits in the form of lumps and plates. These are progressively classified as type V, VI, and VII plaque lesions, with increasing quantities of calcium. The stages in the progression of atherosclerotic plaque have been illustrated in Figure 20-1.

Figure 20-1 Stages in the evolution of an atherosclerotic plaque lesion. Schematic representation of different stages in the natural history of atherosclerotic plaque, progressing from fatty streak to advanced vulnerable plaque that is prone to rupture and precipitate occlusive thrombus.

Clinical Relevance of Coronary Artery Calcification

The amount of coronary calcium correlates well with the segmental atherosclerotic plaque burden¹⁸ and is clearly illustrated in Figure 20-2. Histologic study by Rumberger et al. showed that on average, calcium represents about a fifth of the total atherosclerotic burden.¹⁹ However, the amount of calcium does not correlate with the severity of angiographic luminal stenosis, and this is most likely due to the remodeling process, whereby there is an increase in size of the arteries, compensating for the atherosclerotic plaque (Fig. 20-3).²⁰ Furthermore, presence and extent of calcification do not predict the future risk of plaque rupture. In fact, there are conflicting theories regarding the effect of calcium on plaque stability. It has been suggested that calcified plaques are much stiffer and hence less likely to rupture than cellular plaques. Culprit lesions revealed by IVUS were found to contain less calcium than stable lesions,²¹ lending credibility to the notion that calcification stabilizes a plaque. There have also been studies suggesting that calcification is a result of subclinical plaque rupture, a sort of protective mechanism of the plaque, to stabilize and prevent further plaque rupture.

Figure 20-2 Correlation between plaque area and area of calcium in the plaque. Good correlation between the total atherosclerotic plaque area and the area of calcium within the plaque irrespective of the distance of the lesion from the ostium of the vessel.

(From Rumberger JA, Simons DB, Fitzpatrick LA, et al: Coronary artery calcium area by electron-beam computed tomography and coronary atherosclerotic plaque area. A histopathologic correlative study, Circulation 92:2157-2162, 1995.)

Figure 20-3 Complex plaque lesion and vessel wall remodeling. Cross-section through an atherosclerotic plaque lesion: (A) Overall vessel diameter is greater at the level of the lesion as a result of positive remodeling, as compared to the proximal segment, but with a much narrower lumen diameter. Lesion also indicates hemorrhage into the plaque lesion. Similarly, panel (B) shows changes in overall vessel diameter, luminal diameter, and degree of stenosis in a reference segment (left), lipid rich plaque (middle), and calcified plaque (right).

(From Burke AP, Kolodgie FD, Farb A, et al: Morphological predictors of arterial remodeling in coronary atherosclerosis, Circulation 105:297-303, 2002.)

On the contrary, Abedin et al. suggest that the vessel is rendered less vulnerable to rupture only when extensive calcification has occurred, whereas the early stages of calcification may actually enhance plaque vulnerability.²² It has been shown that calcified atherosclerotic plaque is at least 4 to 5 times stiffer than cellular plaque.²³ In the initial stages of calcification, plaques are most prone to rupture at areas of interface between high- and low-density tissue.²⁴ As the degree of calcification increases, the number of interfaces between rigid and distensible plaque initially would increase until the point at which the rigid plaques coalesce. Calcification beyond this point may be associated with decreasing risk of plaque rupture. However, further research is needed to unravel the association between calcification and plaque vulnerability.

CORONARY ARTERY CALCIUM IMAGING USING ELECTRON BEAM COMPUTED TOMOGRAPHY/MULTISLICE COMPUTED TOMOGRAPHY

Though conventional fluoroscopy was the first technique for imaging coronary artery calcification, the advent of EBCT, with its superior temporal and spatial resolution, soon became the method of choice. The guiding principle of CT scanners is the differential attenuation of x-rays passing through various tissues of the body. Unlike the conventional CT scanners, EBCT scanners do not have the disadvantage of having a mechanically rotating gantry. An electron gun shoots the electrons, which are then guided electromagnetically to tungsten target rings. The x-rays thus produced then pass through the patient and are detected on two parallel rings housed within the gantry of the scanner. Usually it takes 100 milliseconds (ms) for one sweep of the target rings. Considering that it requires 30 to 40 slices for an average 3-mm (range 1.5 to 6 mm) thickness to image the entire coronary tree, the entire heart can be imaged in 30 to 40 seconds, which is a realistic goal for one breath hold. The radiation burden and motion artifacts are reduced by prospective gating, whereby the image acquisition is triggered at 60% to 80% of the R-R interval on the ECG, corresponding to the end of diastole. The total radiation delivered during one study is approximately 0.8 to 1.3 mSv. Stable heart rate and rhythm are essential for prospective triggering during image acquisition.

Recent MSCT scanners have an isotropic resolution of 0.4 mm, although for CAC scoring, a reconstructed slice thickness of 3 mm is still usually chosen. Unlike the EBCT, MSCT scanners use a mechanically rotating gantry. The gantry speed has improved through successive scanner generations. The current state-of-the-art scanners can complete one rotation in 330 ms. Further acceleration is likely to be difficult, owing to problems in handling the huge gravitational force generated by the rapid movement of a relatively heavy gantry. The presence of the mechanical gantry makes the temporal resolution of MSCT lower than that of EBCT, except in the newer dual-source CT scanners, which have two x-ray sources and detector arrays incorporated into the gantry, allowing for a twofold improvement in temporal resolution. Images can be acquired by using either prospective or retrospective ECG gating. Retrospective gating requires the retrospective analysis of the entire data set to select the optimum cardiac phase, and also results in larger data sets. The other disadvantage of retrospectively gated image acquisition is the increase in radiation exposure, which can be reduced by ECG-gated tube current modulation (the tube current is reduced substantially during the less informative phases of cardiac cycle). Prospectively gated image acquisition reduces the radiation burden but can result in longer examination times if the heart rate is not regular. CAC scoring using MSCT confers a higher radiation burden than the EBCT scanners, and this is usually in the range of 1 to 5 mSv; however, prospective gating can reduce the radiation dose to 1 to 2 mSv.

A few studies have compared the data variability between these two imaging modalities (EBCT and MSCT scanners), and they show good correlation between the scanners in terms of accuracy and reproducibility. Knez et al. showed an excellent correlation between MSCT and EBCT for quantification of coronary calcium in 99 patients (r = 0.994; P = 0.01).³⁰ Becker et al. compared the two modalities in 100 patients and found a good correlation between the two types of scanners.³¹

ALGORITHMS FOR QUANTIFICATION OF CORONARY CALCIUM

Mass Score

It has also been suggested that calculating absolute calcium mass is a reliable method of quantification of coronary artery calcium. If appropriate calibration is used, either with an external standard or a calibration phantom, the reliability is improved and is independent of the CT hardware or the differing scanning protocols.³⁹ Calcium mass is calculated as a product of lesion volume, average CT density of the lesion in Hounsfield units, and a calibration factor. These newer methods of calcium quantification are not widely being used, because there is no standardized reference database for them as yet. Since the majority of the large-scale trials have used the Agatston score, this method continues to be the clinical standard used in most centers.

The age/sex percentile rank and diagrammatic representation of the coronary arteries with CAC lesions (depicted in black) in a typical CAC score report is shown in Figure 20-4.

Figure 20-4 Typical coronary artery calcification (CAC) report. A, The CAC scores obtained are plotted on this age- and sex-specific nomogram, with CAC scores divided by quartiles. The red star represents the percentile rank for this particular patient. This example has been generated by the TeraRecon (San Mateo, California) workstation. B, Diagrammatic representation of distribution of calcium in the coronary arteries.

Coronary Artery Calcium Score: What Does It Mean?

Age and sex of a patient play an important role in determining the prevalence and extent of CAC. In both the sexes, prevalence of CAC increases with age.⁴⁰ CAC scores vary greatly, so even in subjects of similar age groups, an age/sex nomogram is required to standardize an individual patient value against that of a matched population—or in essence, to suggest a “coronary age.”^41,42 Hoff et al. analyzed 35,246 self-referred, predominantly white, asymptomatic subjects between the ages of 30 and 90 years and categorized them according to percentiles.⁴³ The total CAC scores in this study group were decidedly non-Gaussian. The age/sex nomogram generated from this study is particularly useful in the risk stratification of individuals, depending on their CAC scores. For example, a calcium score of 15 Au in a male subject younger than 40 years of age, though not alarming in itself, places him above the 90th percentile when ranked with the nomogram. From these data, it is also clear that CAC measured in men is similar to the scores measured in women who are 15 years older.

However, a recent study by Budoff and colleagues⁴⁴ demonstrated that absolute CAC score groups were a better predictor of incident CHD events than age/sex and ethnicity-based percentile rank in the MESA cohort of patients. This is clearly illustrated in Figure 20-5. The Kaplan-Meier curves based on absolute CAC scores appear to have a better separation than those based on age-sex-race/ethnicity percentile ranks, indicative of better risk stratification capability in the short term. What this suggests is that patients in the lower absolute CAC score group are at low risk of CHD events irrespective of their age-sex-race/ethnicity percentile rank. This appears to be true for prediction of obstructive CAD as well in both asymptomatic and symptomatic patients.⁴⁵

Figure 20-5 Incident coronary artery disease (CAD) events with different categories of coronary artery calcification (CAC). The Kaplan-Meier curves for absolute CAC categories (A) offer more separation, indicating better risk-stratifying ability compared to the curves for CAC categories based on age-sex-ethnicity percentile rank (B).

(From Budoff MJ, Nasir K, McClelland RL, et al. Coronary artery calcium predicts coronary events better with absolute calcium scores than age-sex-race/ethnicity percentiles: (MESA) Multi Ethnic Study of Atherosclerosis, J Am Coll Cardiol 53:345-352, 2009.)

Severity of CAC score is arbitrarily classified as minimal, mild, moderate, severe, and extensive based on the absolute score, as shown in Table 20-3. Snapshotsof the varying degrees of severity of CAC score in a typical scan are clearly illustrated in Figure 20-6.

Table 20-3 Categories of Coronary Artery Calcification

CAC Score (Agatston Units)	Coronary Calcification Category
0	Absent
1–10	Minimal
11–100	Mild
101–400	Moderate
401–1000	Severe
>1000	Extensive

Figure 20-6 Severity of coronary artery calcification. Clockwise from top left; calcium score of 0, mild calcification in proximal LAD with a CAC score of 36 Au, severe calcification in proximal LAD and LCX with a CAC score of 820 Au, and extensive calcification in proximal LAD and diagonal with a CAC score of 1720 Au. Au, agatson units; CAC, coronary artery calcification; LAD, left anterior descending; LCX, left circumflex.

ROLE OF CAC IMAGING IN RISK STRATIFICATION

Who Should Undergo CAC Screening?

The fundamental principle of preventive cardiology is that the intensity of risk-reduction measures should match the baseline risk of disease,⁶¹ so that the benefits of intervention outweigh the risks. Accordingly, the National Cholesterol Educational Programme (NCEP) Expert Panel⁶² recommends that individuals with multiple cardiovascular risk factors that confer a CAD risk greater than 20% in 10 years should be targeted for intensive lipid-lowering therapy, with a target LDL cholesterol level of under 100 mg/dL, similar to those with established CAD or CAD risk equivalents such as diabetes, peripheral artery disease, and symptomatic carotid artery disease. Hence, high-risk individuals should be targeted for aggressive risk-reduction measures without need for additional risk assessment tests. At the other end of the spectrum, in low-risk individuals (10-year absolute CAD risk < 10%), the likelihood of adverse cardiovascular events is too low to justify any interventions beyond simple lifestyle changes.

In individuals at intermediate risk, it is difficult to make decisions regarding the administration of risk-reduction treatments. Subclinical disease markers such as CAC imaging are most likely to be effective in this group by reclassifying them into either low- or high-risk groups, thereby influencing management decisions. In a prospective observational study,⁶³ Greenland et al. demonstrated that asymptomatic individuals at intermediate risk according to Framingham criteria but with CAC scores above 300 had an annualized hard cardiac event rate of 2.8% (10-year absolute risk > 20%) and therefore would have to be reclassified as high-risk patients. This information would then mandate a more intensive approach to modification of risk factors, including goals for reducing LDL cholesterol. Furthermore, in low-risk (<10% over 10 years) and high-risk (>20% over 10 years) individuals, CAC measurements did not change the risk prediction substantially enough to alter their prognosis and clinical management.

As a caveat, estimation of lifetime risk of cardiovascular disease (CVD) should be considered as an adjunct to 10-year risk calculation in adults younger than 50 years of age. A significant percentage of adults younger than 50 are placed in the low-risk category for developing CVD over a 10-year period, despite a high lifetime risk of developing CVD. Estimation of lifetime risk for CVD has been elucidated by Lloyd-Jones et al. previously.⁶⁴ Berry et al.⁶⁵ studied 2988 subjects aged younger than 50 years at study entry into the MESA cohort and at year 15 of follow-up of the CARDIA trial, respectively, and divided them into three distinct groups. The first group was composed of subjects with low (<10%) 10-year risk of CVD and low lifetime risk (<39%) of developing CVD, the second group was composed of subjects with low 10-year risk but high lifetime risk (>39%) for CVD, and finally those with high 10-year risk or diagnosed diabetes mellitus (DM). In the cohort of subjects with low 10-year risk of CVD, those with a high lifetime risk had a significantly higher prevalence of CAC in men and women, as well as greater carotid intima media thickness. Subjects with a higher lifetime risk also showed greater CAC progression than those with lower lifetime risk. This was true for subjects in both the CARDIA and the MESA trials.

The ACC/AHA expert committee task force analysis⁶⁶ of intermediate FRS patients from four studies^63,67–69 reported annual CAD death/MI rates of 0.4%, 1.3%, and 2.5% for CAC score categories of less than 100, 100 to 400, and =400, respectively. This is illustrated in Figure 20-7 and reflects the discriminatory ability of CAC in this important intermediate-risk group, which represents a third of the U.S. population aged 60 years or older, according to data from the third National Health and Nutrition Examination Survey (NHANES III).⁷⁰

Figure 20-7 Independent and incremental prognostic value of CAC score in intermediate Framingham risk patients. Annual CAD death/MI rates shown by tertile of the Agatston score in patients at intermediate CAD event risk using definitions of an intermediate Framingham Risk Score (FRS) of greater than 1 cardiac risk factor. CAC, coronary artery calcification; CAD, coronary artery disease; MI, myocardial infarction.

(Reproduced from Greenland P, Bonow RO, Brundage BH, et al: ACCF/AHA 2007 Clinical Expert Consensus document on coronary artery calcium scoring by computed tomography in global cardiovascular risk assessment and in evaluation of patients with chest pain: A report of the American College of Cardiology Foundation Clinical Expert Consensus Task Force, J Am Coll Cardiol 49:378-402, 2007.)

THE PROGNOSTIC VALUE OF CAC: AVAILABLE EVIDENCE

A screening test should add to the prognostic information provided by the clinical and demographic data, and the test should be readily available and affordable. This additional prognostic value can be quantified statistically as the increase in the area under the curve of the receiver operating characteristic (ROC), also expressed as C-statistic, after the addition of a new test result. In the past decade, several studies reported on the independent and incremental prognostic value of CAC measurement for the prediction of CAD events in univariable and multivariable models containing data from measured or reported risk factors and CRP measurements.

A meta-analysis⁷¹ of 4 studies^72–75 showed that CAC scores remained predictive of coronary events, even after adjustment for established cardiovascular risk factors. The event rates in individuals with even mild coronary calcification (CAC scores of 1 to 100) were twice as high compared to those with no detectable coronary calcium (RR = 2.1). CAC scores over 400 were associated with very high relative risks (RR = 4.3 to 17) after adjustment for age, sex, and other cardiovascular risk. However, the meta-analysis showed a significant heterogeneity in the quality of these studies, most of which used self-reported or historical as opposed to measured risk factor data and did not use blinded outcome adjudication, which could have overestimated the predictive accuracy of CAC measurements. Selection bias was also an issue, since most of these studies recruited self-referred or physician-referred individuals, who are not necessarily representative of the general population. Furthermore, earlier studies included “soft” endpoints, including coronary revascularization, which could have overestimated the predictive ability of CAC, since CAC results per se could influence the decision to undertake revascularization.

More recent studies have focused on the prediction of “hard” cardiac events (myocardial infarction, CAD death) and are less likely to be subjective. One such study conducted by Ostrom and colleagues⁷⁶ included 2538 consecutive patients referred for CT coronary angiography. Correlation of mortality and absolute CAC score groups (1 to 9, 10 to 99, 100 to 399 and > 400) was assessed in the 1060 patients diagnosed with nonobstructive CAD. The survival rates for different CAC score groups over a follow-up period of 78 ± 12 months were 99.2%, 99%, 96.9%, and 96.1%, respectively. The risk factor adjusted RR of 5.1 (P = 0.003) and 6.2 (P = 0.001) for CAC scores of 100 to 399 and over 400, respectively, elegantly demonstrated the independent and incremental value of the CAC score.

A recent meta-analysis reported as part of the ACC/AHA 2007 Expert Consensus statement⁶⁶ on CAC imaging included six studies^63,67–^69,77,78 published between 2003 and 2005, involving a total of 27,622 patients with 395 hard cardiac events. CAC score of zero was associated with a very low risk (0.4%; 49 events in 11,815 individuals) of CAD death or MI over 3 to 5 years of observation, and presence of any detectable coronary calcium increased the risk fourfold (P < 0.0001). The summary RR increased significantly in those with more severe coronary calcification (7.2 for CAC score of 400 to 1000, and 10.8 for CAC scores >1000, compared to zero CAC score), establishing a continuous graded relationship between CAC scores and risk of coronary events.

A subsequent large observational study⁷⁹ published in 2007 evaluated the effect of CAC on all-cause mortality in a cohort of 25,253 patients followed for 6.8 ± 3 years. The study conclusively established the independent and incremental prognostic ability of CAC to predict all-cause mortality. The 10-year cumulative survival was 99.4% for a CAC score of zero and reduced to 87.8% for a score of over 1000 (P < 0.0001) (Fig. 20-8). ROC analysis showed a significant increase in the area under the curve when CAC score was added to risk factors and age. The improved strength and quality of available evidence were enough for the 2007 Expert Consensus Committee of the AHA/ACC to recommend CAC imaging as a reasonable choice for refining the risk stratification in asymptomatic individuals with intermediate Framingham Risk Score,⁶⁶ which is a significant change from the earlier recommendation made in 2000 by the same body.⁸⁰

Figure 20-8 Survival rates in subjects in progressive quintiles of coronary artery calcification (CAC) scores. The graph shows progressively significant differences in survival rates in quintiles of CAC scores (<10, 11-100, 101-400, 401-1000, and >1000) after 12 years of follow-up. The increased separation of the curves elucidates the significant survival predictive ability of CAC.

(From Budoff MJ, Shaw LJ, Liu ST, et al: Long-term prognosis associated with coronary calcification: Observations from a registry of 25,253 patients, J Am Coll Cardiol 49:1860-1870, 2007.)

COMPARISON OF CAC WITH OTHER DIAGNOSTIC TESTS

To secure its place in the diagnostic armamentarium of a cardiologist, CAC imaging should prove its cost and clinical effectiveness in comparison to other diagnostic tests that are well established and readily available. Earlier studies comparing EBCT with stress electrocardiography (ECG) and scintigraphy indicated that CAC scores were at least as good if not better in predicting angiographically confirmed obstructive coronary disease.^113,114 The anatomic information obtained by CAC imaging can complement the functional information from stress testing and improve the overall diagnostic accuracy for obstructive CAD.

CAC and Myocardial Perfusion Scintigraphy: Role of Synergistic Imaging

For over 3 decades, myocardial perfusion scintigraphy (MPS) by SPECT has been extensively validated for the diagnosis of patients with suspected CAD, and has played a key role in formulating management decisions regarding revascularization or medical therapy in symptomatic patients with established obstructive CAD.^117–124 Attempts have been made to integrate a newer, evolving technology such as CAC imaging with SPECT in investigating patients with suspected or known CAD. Berman et al.¹²⁵ observed the relationship between CAC scores and stress MPS results in 1195 individuals who underwent both these tests within 7.2 + 44.8 days of each other. While only 2% of patients with CAC scores under 100 Au had evidence of ischemia on MPS, 20% of those with CAC scores over 1000 Au had ischemic MPS, indicating significant relationship between these two clinical markers of atherosclerosis and ischemia, respectively. This relationship is elucidated in (Figure 20-9). After multivariable logistic regression analysis, CAC score was found to be the most potent predictor of ischemia. However, 56% of all normal MPS results were noted in those with significant subclinical atherosclerosis, as indicated by CAC score over 100 Au. Furthermore, even among patients with CAC score over 1000 Au, 85% of asymptomatic and 68% of symptomatic individuals had normal MPS.

Figure 20-9 Relation between coronary artery calcium and myocardial perfusion. Composite figure illustrating the advantage of a sequential screening approach, with initial coronary artery calcification (CAC) imaging followed by myocardial perfusion scan if the CAC score is high. These are the scans of a 44-year-old asymptomatic man. On the right is a diagrammatic representation of the CAC score as well as an actual image of his CAC scan showing calcification in the right coronary and left circumflex arteries. On the left is an image of this patient’s myocardial perfusion scan showing extensive reversible ischemia.

In another study to assess the predictive ability of CAC score for the presence of ischemia, Schenker et al.¹²⁶ looked at 621 patients who had no known CAD but were clinically referred to have a PET scan. The frequency of myocardial ischemia showed significant correlation with the severity of coronary calcification; patients with CAC over 400 Au were more likely to have an abnormal scan than those with CAC 1 to 399 Au (48.5% compared to 21.7%; P < 0.001) (Fig. 20-10). But what was interesting was that even in patients with a zero CAC score, 16% had abnormal scans. At the other end of the spectrum, in patients with a CAC score over 1000 Au, 50.6% had a normal perfusion scan. In the same study, CAC score was found to be strongly predictive of survival in patients with abnormal PET scans and also in those with normal scans. In the presence of a normal PET scan, annualized event rates in patients with a CAC score over 1000 Au and those with no CAC were 12.3% and 2.6%, respectively. Similarly, in those patients with demonstrable ischemia on PET scan, CAC scores of over 1000 Au and under 1000 Au were associated with annualized event rates of 22.1% and 8.1%, respectively.

Figure 20-10 Increasing CAC score predicts myocardial ischemia with PET. Bar graph demonstrates increasing frequency of myocardial ischemia with PET, with progressively increasing absolute CAC score categories. CAC, coronary artery calcification; PET, positron emission tomography.

(From Schenker MP, Dorbala S, Hong ECT, et al. Interrelation of coronary calcification, myocardial ischemia, and outcomes in patients with intermediate likelihood of coronary artery disease: A combined positron emission tomography/computed tomography study, Circulation 117:1693-1700, 2008.)

This disparity between CAC and MPS results reflects the fundamental difference in the nature of information provided by these two tests. Coronary calcium allows detection of atherosclerotic lesions, both obstructive and nonobstructive. Secondary to expansive remodeling of the vessel wall, a considerable amount of plaque can accumulate before luminal encroachment starts. Stress scintigraphy (as with all stress imaging methods), in contrast, requires the presence of a hemodynamically significant obstructive lesion, either fixed or dynamic, before an abnormality becomes evident. The study by Ramakrishna et al.¹²⁷ also showed a weak correlation between CAC scores and summed stress scores by SPECT; however, the two tests were complementary in predicting mortality. CAC and MPS appear to have different but complementary roles in risk prediction and diagnosis. While MPS is an excellent tool for predicting short-term risk, thereby guiding decisions regarding revascularization, CAC is a better predictor of long-term risk and hence more useful in the determination of the need for aggressive medical prevention measures. Absence of ischemia on MPS is associated with a low short-term risk of coronary events, even when the CAC scores are high (>1000 Au).¹²⁸

The high sensitivity and negative predictive value of CAC scoring can be combined with relatively high specificity and positive predictive value of MPS to achieve high overall diagnostic accuracy. CAC scores less than 100 Au are typically associated with a low probability (> 2%) of abnormal perfusion on nuclear stress tests¹²⁵ and less than 3% probability of significant obstruction on cardiac catheterization.^107–108 He et al.¹²⁹ evaluated the relationship between the severity of coronary calcification and stress-induced myocardial ischemia in a large cohort of asymptomatic subjects with risk factors for CAD. CAC score was the best predictor of an abnormal SPECT. Nearly half of those with scores over 400 Au had evidence of ischemia, compared to only 6.6% of those with scores under 400 Au.

In the study by Anand et al.,¹³⁰ CAC imaging and MPS were synergistic for the prediction of short-term cardiovascular events in 510 asymptomatic diabetic subjects without prior cardiovascular disease. Those with CAC scores over 100 Au (n = 127) and a random sample of the remaining participants with a CAC score of 100 Au (n = 53) underwent MPS and were followed up. In the multivariable model, CAC score was the only predictor of myocardial perfusion abnormality (P < 0.001); CAC score and extent of myocardial ischemia were the only independent predictors of outcome. Twenty cardiovascular events were noted over a median follow-up of 2.2 years, of which 15 were in those with CAC score over 400 Au, and no events were noted in those with scores less than 10 Au. The CAC score predicted events more accurately than the UKPDS (UK Prospective Diabetes Study) and Framingham Risk Scores (area under the curves were 0.92, 0.74, and 0.60 for CAC, UKPDS, and FRS, respectively; P < 0.0001). In a similar study by Wong et al.,¹³¹ CAC scores below 100 Au were associated with a very low risk (<3%) of ischemic MPS, even in patients with metabolic syndrome and diabetes. The strong and independent relationship between CAC scores and ischemic MPS persisted, even in patients receiving comprehensive medical therapy with adequate control of risk factors.¹³²

In view of this association between the presence of CAC, especially above a threshold value of 400 Au, and abnormal myocardial perfusion study in a wide variety of populations, a sequential screening approach with initial CAC scoring followed by MPS in those with high CAC scores was suggested for maximizing the yield of MPS and improving cost-effectiveness.¹³³ Using CAC imaging prior to MPS seems logical from a pathophysiologic point of view as well, since the development of coronary atherosclerosis almost always precedes the onset of myocardial ischemia in the ischemic cascade. The incremental value of an integrated approach is clearly shown in Figure 20-11.

Figure 20-11 Preselection of asymptomatic diabetic subjects by electron beam computed tomography improves the yield of perfusion abnormalities detected by myocardial perfusion imaging. Left, Myocardial perfusion imaging was performed in asymptomatic diabetic subjects (n = 522) in the DIAD study. In the study of Anand et al,¹³⁰ myocardial perfusion imaging was performed in asymptomatic diabetic subjects with coronary artery calcium scores of greater than 100: unselected patients (middle) and patients selected based on electron beam computed tomography (coronary calcium scores >100 Agatston units) (right).

(Reproduced from Wackers FJ, Young LH, Inzucchi SE, et al: Detection of silent myocardial ischemia in asymptomatic diabetic subjects: The DIAD Study, Diabetes Care 27:1954-1961, 2004; and Lim E, Lahiri A: The importance of the link between coronary artery calcification and myocardial ischemia: A developing argument, J Nucl Cardiol 14:272-274, 2007.)

PROPOSED ALGORITHM FOR SEQUENTIAL SCREENING

Based on the available evidence, a diagnostic algorithm was proposed for assessing patients at risk of CAD and those with atypical symptoms (Fig. 20-12).¹³³ CAC scoring can be used in those subjects assessed to be at intermediate or high risk after an initial clinical evaluation using established risk-assessment systems such as FRS. In individuals with CAC scores above 400 Au, the frequency of ischemia is usually substantial, and MPS is thus likely to be cost-effective. This threshold could be lowered perhaps to 200 Au in diabetic subjects, but more data are required before an exact cutoff value can be agreed upon. In individuals with CAC scores in the intermediate range of 100 to 400 Au, the need for MPS should be determined on the basis of overall risk profile and the likelihood of CAD. The severity of perfusion abnormality on MPS determines the need for angiography. When less than 10% of the myocardium is involved, optimum medical therapy was shown to be better than revascularization; hence, coronary angiography can be avoided in these patients. Patients with more extensive ischemia (>10% myocardial involvement) could be referred for angiography and possible revascularization. Finally, those individuals with severe coronary calcification and normal MPS were shown to have a low frequency of hard coronary events (<1% per year) over 3 years of follow-up¹²⁸; therefore, these patients can be targeted for aggressive medical treatments and spared from invasive interventions.

Figure 20-12 Proposed diagnostic algorithm. Sequential screening approach starting with conventional cardiovascular risk profiling, followed by coronary artery calcification (CAC) imaging in those at intermediate or high risk, and myocardial perfusion imaging in those with high CAC scores or high clinical risk with moderate CAC scores.

(Modified from Yerramasu A, Maggae SV, Lahiri A, Anand DV: Cardiac computed tomography and myocardial perfusion imaging for risk stratification in asymptomatic diabetic patients: A critical review, J Nucl Cardiol 15:13-22, 2008.)

CAC IMAGING IN PATIENTS WITH DIABETES MELLITUS AND CARDIOMETABOLIC SYNDROME

CAD accounts for 70% of the deaths among patients with DM.¹³⁷ Diabetic patients face a two- to fourfold higher risk of cardiac events than their nondiabetic counterparts, and in fact, diabetes is now treated as CAD equivalent¹³⁸ in risk stratifying patients. CAD is often asymptomatic and is associated with worse prognosis in diabetic patients, thus posing a major dilemma in identifying those at increased risk.¹³⁹ CAC imaging in diabetic patients promises to be an effective test for early detection of silent CAD.

Mielke et al.¹⁴⁰ showed that the median CAC score was higher in diabetic subjects compared to their nondiabetic counterparts. In a large study comparing the CAC scores in self-referred diabetic (n = 1075) and nondiabetic subjects (n = 29,829) with no known previous CAD, Hoff et al.¹⁴¹ reported that diabetes was the strongest predictor for having a CAC score in the highest quartile, irrespective of sex.

The PREDICT trial¹⁴² researchers prospectively studied 589 asymptomatic diabetic patients who were followed up for a median of 4 years with first-incident stroke or CAD events as primary endpoints. Similar to previous large observational studies, CAC score was an independent predictor of primary endpoint events, inclusion of which increased the ROC area under the curve from 0.63 to 0.73, using Framingham CAD risk, and from 0.67 to 0.75 for UKPDS CAD risk prediction. What was interesting was that the doubling of CAC score was associated with a 32% increased risk for primary endpoint events. The authors were able to demonstrate this by using a base 2 logarithm (CAC score + 1), as seen in Figure 20-13.

Figure 20-13 Increasing CAD and stroke primary endpoint events in diabetics. Y-axis represents primary endpoint events per 1000 person years, and x-axis represents CAC score variables represented in log 2 (CACS + 1) values. Each unit increase in log 2 (CACS + 1) represents a doubling of CAC score. The traditional CAC score categories of 0-10, 11-100, 101-400, 401-1000, and 1001-10,000, as well as equivalent log transformed value categories of 1-4, 4-7, 7- 9, 9-10, and 11, are shown on the x-axis. The dotted lines on either side represent the 95% confidence intervals. CAC, coronary artery calcification; CAD, coronary artery disease.

(From Elkeles RS, Godsland IF, Feher MD, et al: Coronary artery calcium measurement improves prediction of cardiovascular events in asymptomatic patients with type 2 diabetes: The PREDICT Study, Eur Heart J 29:2244-2251, 2008.)

Diabetic patients with coronary calcification have worse prognosis than nondiabetic subjects with the same extent of coronary calcification (Fig. 20-14).¹⁴³ Furthermore, in the presence of low CAC scores, the risk of coronary events is low even in diabetic patients. The study by Raggi et al.¹⁴³ showed that diabetic patients with no detectable coronary calcium had excellent 5-year survival that was not significantly different from nondiabetic subjects (survival 98.8% versus 99.4%, respectively, P = 0.49). The South Bay Heart Watch Study,¹⁴⁴ in which 1312 diabetic and nondiabetic patients were screened with CAC scoring using an EBCT scanner and followed up for 6.3 ± 1.4 years, with the endpoints of coronary death, nonfatal MI, coronary revascularization, or stroke, demonstrated a nearly fourfold increase in event rates in diabetics with a higher CAC than in the nondiabetics with a lower CAC.

Figure 20-14 Comparison of cumulative survival in diabetics and nondiabetics. Cumulative survival is significantly greater across all quintiles of CAC scores in nondiabetics as compared to diabetics over a follow-up period of 5 years.

(From Raggi P, Shaw LJ, Berman DS, Callister TQ: Prognostic value of coronary artery calcium screening in subjects with and without diabetes, J Am Coll Cardiol 43:1663-1669, 2004.)

One of the theories put forth to explain the increased mortality in diabetic patients as compared to nondiabetics with similar CAC scores is the increased prevalence of noncalcified plaque lesions and the presence of a greater volume of necrotic core with a thin-cap fibroatheroma. Plaque lesions with the aforementioned characteristics are deemed vulnerable for rupture and consequently thrombosis and acute coronary syndromes and/or death.

In patients with diabetes, the presence of metabolic abnormalities increases the likelihood of an abnormal myocardial perfusion scan.¹⁴⁵ Wong et al.¹⁴⁶ examined 4468 asymptomatic individuals who underwent CAC imaging. The prevalence of CAC ranged from 51% in men and 35% in women with neither DM nor cardiometabolic syndrome (CMS) to 75% in men and 58% in women with both DM and CMS. These data concur with the results of the NHANES study¹⁴⁷ that reported the highest prevalence of CAD in patients with both DM and CMS in the United States. The same study also reported higher prevalence of CAD in women with CMS without DM compared to those with DM without CMS (13.9% and 7.5%, respectively).

So far, there are no data to suggest that patients with CMS but no diabetes benefit from aggressive risk-factor modification, as do patients with overt diabetes.¹⁴⁸ This makes a good case for CAC imaging screening of asymptomatic diabetic patients, as well as those with CMS but without overt diabetes, to better guide their clinical management. Metabolic syndrome alone is not a robust marker for 10-year cardiovascular risk, but the presence of metabolic syndrome even in the absence of DM should prompt a reassessment of cardiovascular risk, ideally with imaging for the presence of subclinical atherosclerosis.

EFFECT OF ETHNICITY ON PREVALENCE AND PROGNOSTIC SIGNIFICANCE OF CORONARY CALCIUM

CVD-associated morbidity and mortality varies widely across different ethnic groups, and it is likely that these differences are related to genetic as well as environmental factors. Most of the studies that investigated the prevalence and prognostic significance of coronary calcium were based on a predominantly white population. Earlier studies based on necropsy reports^149,150 and cine-fluoroscopy,¹⁵¹ and more recent studies using EBCT-derived CAC measurements,^152–154 found higher prevalence of coronary calcium in Caucasians compared to African Americans. Using EBCT and coronary angiography, Budoff et al. reported a higher prevalence of coronary atherosclerosis in Caucasians compared to African Americans or Hispanics.¹⁵⁵ On the contrary, the CARDIA study¹⁵⁶ showed no significant difference in the prevalence of coronary calcification in 443 young black and white adults aged between 28 and 40 years. However, the prevalence of CAC was low in this young cohort. Also, the threshold area for CAC detection was 2.06 mm², twice the threshold area used in most of the other studies. In another smaller study of 128 black and 733 white asymptomatic postmenopausal women (mean age, 63 ± 8 years), a similar distribution of CAC scores was found in both the ethnic groups.¹⁵⁷

The Dallas Heart Study evaluated the prevalence and severity of CAC in a multiethnic population-based sample in which blacks were systematically oversampled to make the final sample 50% black.¹⁵⁸ The prevalence of subclinical atherosclerosis was similar in the black and white participants, as were the mean FRSs.¹⁵⁹ However, this study used a score of greater than 10 Au (as opposed to > 0 Au) as the minimum score to define CAC prevalence. Also, greater obesity in black women could have resulted in increased noise, which could be misinterpreted as calcium. In a more recent larger study¹⁶⁰ involving over 16,000 asymptomatic individuals, Budoff and colleagues reported that African American men were least likely to have any coronary calcium, after adjusting for age and risk factors. African American women, on the other hand, had significantly higher odds of having non-zero CAC scores. Caucasians and Hispanics have a significantly higher prevalence of CAC when compared to African and Asian men. Similarly, in the MESA cohort, white men had the highest prevalence of coronary calcification (70.4%), followed by Chinese (59.2%), Hispanic (55.6%), and black men (52.1%). The weight of evidence points to higher prevalence and severity of CAC in Caucasians compared to ethnic minorities, but it is not yet clear if this is due to differences in the plaque burden or due to differing degrees of plaque calcification among different ethnicities.

Two recent studies evaluated the influence of ethnicity on the predictive value of CAC. Nasir et al. reported the prevalence and the prognostic value of CAC in predicting all-cause mortality in a large, ethnically diverse population of over 14,000 asymptomatic individuals with no prior history of CHD.¹⁶¹ After a mean follow-up of 6.8 years, CAC score was the best predictor of time to death among all ethnic groups. Compared to non-Hispanic whites, ethnic minorities have lower odds of having any, as well as increasing, CAC burden (P < 0.0001). Higher CAC scores were associated with greater mortality among ethnic minorities. However, in the MESA study, no interaction was noted between ethnicity and the risk associated with increasing calcium score.¹⁶² Following a median follow-up of 3.9 years, CAC scores were independently predictive of the risk of major coronary events in all ethnic groups; no major differences in the predictive value of calcium scores were detected among different ethnic groups.

The recently concluded London Life Sciences Population (LOLIPOP) study¹⁶³ evaluated the ethnic differences in the prevalence of subclinical atherosclerosis among 2398 asymptomatic participants with no known previous CAD, recruited from the primary care practices in west London, UK. In comparison to the Europeans (n = 1031), Indian Asians (n = 1367) had approximately twofold higher prevalence of hypertension and type 2 diabetes, higher waist-hip ratio and triglycerides, and lower HDL cholesterol. However, there was no difference in CAC prevalence or mean CAC scores between Indian Asians and Europeans either before or after adjustment for the measured cardiovascular risk factors. In the same study, 262 patients with CAC over 100 Au underwent myocardial perfusion scan. There was no difference in prevalence of silent myocardial ischemia between the two ethnic groups, following adjustment for conventional risk factors. These results contrast with an almost twofold higher risk of MI and CAD mortality in Asians compared to Europeans, and indicate the need for more studies to conclusively establish the relationship between ethnicity and CAD.

PROGRESSION OF CORONARY CALCIUM—CAUSES AND CONSEQUENCES

There is a wealth of published data highlighting the prognostic significance of CAC scoring, but there have been very few trials studying the factors responsible for the progression of CAC scores. Yoon et al.¹⁶⁴ showed that the baseline CAC score was the most important determinant of rate of calcium progression. This study showed that neither age nor sex significantly correlated with progression of CAC. But from more robust data from the MESA cohort,¹⁶⁵ annual CAC progression had a linear correlation with age irrespective of race and showed a similar trend in both sexes. Among the 2948 participants (51.2%) without detectable CAC at baseline, 475 (16.1%) went on to develop incident CAC over an average of 2.4 years of follow-up. Thus, the estimated incidence rate was 6.6% per year. The incidence was less than 5% per year at younger ages and more than 12% per year at the highest ages. It is important to note that this study did not control for baseline CAC score, since CAC progression is assumed to be a continuous spectrum of change from the initial focus of calcification in the atherosclerotic plaque lesion; therefore, to control for baseline CAC score would be to ignore the effect of causative factors from initiation untill the baseline scan. The study showed that the age differences were similar across gender and race/ethnicity subgroups. Black women had the highest yearly incident rates at 6.3%. Hispanic, white, and Chinese women had yearly rates of 6%, 5.6%, and 4.7%, respectively. These ethnic differences in yearly incident rates were non significant when adjusted for age. With the exception of Chinese participants, men had higher incidence rates than women (P < 0.001). The yearly incidence rates in decreasing order are 10.3% for white men, 8.2% for black men, 7.2% for Hispanic men, and 4.4% for Chinese men. In the same cohort, after adjusting for age, gender, race, and smoking, systolic blood pressure (SBP) and serum glucose were found to be strong predictors of progression of CAC. Serum glucose level was strongly associated with progression even in patients with no diabetes.

Anand et al.¹⁶⁶ reported a similar finding in a cohort of 398 asymptomatic diabetic subjects. The study revealed that whereas traditional risk factors except smoking and hyperlipidemia were all univariate predictors of CAC progression, suboptimal glycemic control was one of the strongest predictors of CAC progression. Anand et al. also demonstrated that patients with minimal CAC scores did not progress significantly, but those who already had evidence of increased CAC earlier showed considerable CAC progression despite treatment. Though presence of cardiometabolic syndrome with or without diabetes was associated with a higher prevalence of CAC scores, in the recent Rancho Bernardo study,¹⁶⁷ only hypertension and fasting plasma glucose levels significantly predicted the progression of CAC, whereas metabolic syndrome did not.

Yoon et al. also showed that hypertension and diabetes were independent predictors of progression of CAC.¹⁶⁴ In a few studies it has been shown that family history of CAD is a strong predictor for the progression of CAC scores, pointing to hitherto undiscovered genetic factors at play in the progression of CAC scores.^168,169 Surprisingly, most of the studies looking at progression did not find any significant correlation between levels of LDL and progression of CAC.

It has been shown that increased rate of progression of CAC is associated with increased cardiac event rates, as demonstrated by Shemesh et al.¹⁷⁰ in a cohort of high-risk hypertensive patients. The percentage change in CAC score in the three groups of their study population of asymptomatic patients, patients with stable CAD, and patients who suffered a cardiovascular event during the 3 years of follow-up were 118%, 124%, and 180% (P < 0.05). Similarly, Raggi et al.¹⁷¹ also showed that in asymptomatic patients, higher progression rate of calcium volume score was significantly associated with increased risk of coronary events. In a study population of 1310 subjects undergoing sequential EBCT scanning with at least a 1-year interval between scans and average follow up of 2.2 to 2.7 years, 49 subjects suffered from an MI; 90% of the diabetics (n = 9) and 71% of nondiabetics (n = 24) who suffered from an MI had greater than 15% annualized change in their calcium volume score.

It is important to know how frequently a patient should be imaged in order to monitor the progression of CAC, especially when the score is zero. In an interesting study by Gopal et al.,¹⁷² 710 physician-referred patients with an initial CAC of zero were followed up differentially: 35% for 1 to 3 years, 36% for 3 to 5 years, and 29% for over 5 years. A CAC change/year of zero was found in 61% of the study population; in fact 97% of the subjects had a change in CAC/year of less than 10 Au. Only 1% had a CAC change/year of more than 50 Au, and 2% were in the intermediate group of 10 to 50 Au. They concluded that with a baseline score of zero, there was no evidence to suggest repeating the scan before 5 years. These data are supported by an earlier study by Budoff et al., who reported that only 2% of their study population with an initial score of zero had a CAC score of greater than 10 Au at repeat scanning after 1 to 6 years.¹⁷³

EFFECT OF TREATMENT ON PROGRESSION OF CORONARY CALCIUM

The evidence for the importance of CAC imaging as a prognostic indicator and as a method of risk stratification has already been elucidated. Logically, identification of high-risk patients based on CAC imaging would lead to the institution of more intense risk-reduction measures. Some trials have suggested targets for LDL as low as 70 mg/dL for additional benefit in selected populations.^174–176 However, the evidence for the effect of aggressive HMG-CoA reductase inhibitor therapy on the rate of progression of CAC is conflicting. Earlier studies^177,178 that showed slowing of progression of CAC score with HMG-CoA reductase inhibitor therapy are not supported by the data from more recent, larger prospective trials.^179–181 Houslay et al.¹⁸¹ showed that despite aggressive therapy with atorvastatin 80 mg compared with placebo, leading to significant reduction in both LDL (more than half of baseline) and CRP, there was no significant correlation with the progression of CAC. The annualized change in CAC in the atorvastatin group was 26%/year, and in the placebo group it was 18%/year (P = NS). Owing to progression in CAC despite aggressive LDL lowering, they also concluded that there was no significant correlation between LDL levels and progression of CAC. In the special case of familial hypercholesterolemia, it has been shown in a very small study group of 8 patients (7 male and 1 female), that LDL apheresis with an extracorporeal device on a weekly or twice-weekly basis in addition to 80 mg of atorvastatin resulted in a significant reduction in coronary calcium scores (26 ± 14%, P < 0.01) during a 29-month follow-up period.¹⁸² It is postulated that the reduction in cholesterol levels reduced the exposure of macrophages to the toxic effects of oxidized LDL, thus reducing cell death and denying a focus for calcium deposition.

In a larger observational study, Hecht et al.¹⁸³ studied a group of 182 asymptomatic patients who were treated with an HMG-CoA reductase inhibitor, either alone or in combination with niacin. The study population was divided into those who had achieved an LDL of lower than 80 mg/dL and those who had an LDL level of higher than 80 mg/dL. Repeat scan after 1.2 years did not show any significant difference in the annualized progression rates of the calcium score between the two groups (9.3% in the group with LDL < 80 mg/dL and 9.1% in the group with LDL > 80 mg/dL). Similar annual progression was noted in the group of patients given an HMG-CoA reductase inhibitor alone and the group given a combination of HMG-CoA reductase inhibitor and niacin (12.2% and 12.1%, respectively). Similarly, in the St. Francis Heart Study,¹⁷⁹ treatment with 20 mg of atorvastatin, 1g of vitamin C, and 1000 units of vitamin E (alpha tocopherol) induced a significant reduction in LDL but no significant reduction in CAC progression, compared to treatment with placebo. Interestingly, in subjects with a CAC score of over 400, treatment reduced the incidence of atherosclerotic cardiovascular disease events by 42% (P = 0.046). This discrepancy between the effect of HMG-CoA reductase inhibitors on cardiovascular event rates and CAC progression could be due to the fact that they stabilize the plaque by reducing the free lipid core and inducing a fibrovascular transformation of the thin fibrous cap.¹⁸⁴ HMG-CoA reductase inhibitor therapy also reduces the number of macrophages in the plaque,¹⁸⁵ therefore attenuating the inflammatory impetus rendering the plaque vulnerable,¹⁸⁶ consequently reducing the plaque vulnerability and cardiovascular mortality in general. This effect of HMG-CoA reductase inhibitor therapy on stabilization of the plaque has very little effect on the calcium component of the atherosclerotic plaque lesion or in some cases, an increase in the proportion of calcification in the plaque even as the total volume of the lesion regresses.¹⁸⁷

COST-EFFECTIVENESS OF CAC IMAGING

In the current era of inflation beating health care costs and increasing budgetary constraints, imaging services are under growing scrutiny to prove their cost and clinical effectiveness before they can be accepted as part of routine clinical care. Cardiovascular disease already imposes a huge economic burden on the society, costing in excess of $329 billion per year in the United States,¹⁸⁸ and can increase substantially in the future as a result of an aging population and growing prevalence of risk factors. Approximately 300,000 EBCT CAC are scans are performed annually in 79 centers in the United States. In the last 2 decades, the number of cardiac imaging procedures has grown markedly while prevalence of CAD has remained steady.¹⁸⁹ The American College of Cardiology and the American Heart Association now recommend CAC imaging as a reasonable choice in intermediate Framingham risk patients.⁶⁶ The 34th Bethesda Conference¹⁹⁰ on Atherosclerosis Imaging estimated that applying CAC screening to the 34 million intermediate-risk individuals in the United States could result in direct imaging costs of approximately $1 billion, with additional downstream costs from the ensuing tests and treatments. It is estimated that using CAC screening with a cutoff value of above 75% of age- and gender-adjusted percentile, 25% of intermediate-risk individuals could be reclassified into the high-risk group,¹⁹¹ which will have a significant impact on the extra cost of additional aggressive risk-reduction treatments.

To make sense economically, a screening or diagnostic test should provide “incremental” information to clinical history, physical examination, and other low-cost tests (such as lipid analysis and hs-CRP). The incremental cost-effectiveness ratio (ICER, incremental cost/incremental benefit) is usually expressed as cost per quality adjusted life year (QALY) saved, and the U.S. Public Health Service set a threshold of less than $50,000 per QALY for any intervention or investigation to be recommended for routine clinical use. A number of European countries have used lower thresholds of approximately $20,000 per life year saved. However, in the case of CAC imaging, cost per QALY is difficult to assess, since the survival benefit of the test is indirect, dictated largely by the effectiveness of ensuing treatment. Hence, cost to identify CAD death or nonfatal MI was proposed as an intermediate outcome measure.

Most of the evidence for cost effectiveness of CAC imaging comes from analysis based on decision modeling rather than actual trial data. The cost-effectiveness of CAC scoring depends largely on the population being screened. Cost-effectiveness analysis based on a decision model¹⁹² estimated that the cost to identify a CAD event was $73,070 for low-risk and $37,260 for intermediate-risk individuals (P < 0.00001). When the costs and benefits of resulting treatment were taken into account, the cost was $500,000 per life year saved in the low-risk group, whereas in intermediate-risk individuals, cost-effectiveness ratios were much more favorable, ranging between $30,000 and $42,000 per life year saved. Similarly, when CAC imaging is used as a diagnostic test in symptomatic patients, cost-effectiveness depends largely on the disease prevalence and the cutoff CAC score. In a decision-tree model,⁸³ CAC scan with a cutoff score of 168 Au represented the most cost-effective initial test in population groups with low/intermediate disease prevalence (<70%), compared to other investigations such as stress ECG, stress echo, stress scintigraphy, or direct angiography. CAC scan with a cutoff score of greater than 0 proved to be the least cost-effective noninvasive strategy in the same group. In high-prevalence groups, initial angiography with no prior noninvasive testing proved to be the most cost-effective diagnostic strategy. However, an inherent problem with these decision models is that they either are too simplistic in their assumptions or make too many assumptions. Thus, there is a need for prospective studies evaluating the cost performance and economic implications of CAC scoring in the screening and diagnosis of CAD.

FUTURE DIRECTIONS: CAN CT CORONARY ANGIOGRAPHY REPLACE CAC IMAGING?

CAC imaging is already being used extensively in the United States and some European countries, mainly for reclassifying patients who have a 10-year intermediate cardiovascular risk based on conventional risk factors. Despite the extensive research and the wide availability of CT scanners, there is still a reluctance to integrate CAC imaging into the burgeoning investigative arsenal at the disposal of cardiologists today. With the introduction of successive generations of MSCT scanners and gradual improvements in temporal and spatial resolution, it is now possible to make a more comprehensive assessment of coronary atherosclerotic plaques, both calcified and noncalcified. It is well known that although CAC correlates well with the overall atherosclerotic burden, it does not represent the atherosclerotic plaque in its entirety. More important, CAC imaging provides no information regarding the degree of luminal stenosis caused by an individual plaque. CT coronary angiography can provide important information regarding the location,¹⁹³ extent, and morphology of the plaque,¹⁹⁴ vessel wall remodelling,^195,196 and the degree of luminal compromise caused by an individual lesion (Fig. 20-15).¹⁹⁷ Although there is not enough evidence currently to advocate the use of CTA as a screening tool in asymptomatic patients, aggressive marketing strategies and mass media have, on occasions, prompted self-referral from the public, posing additional dilemmas to the physician. A recent study by Choi et al.¹⁹⁸ evaluated the use of CTA as a screening tool in 1000 middle-aged asymptomatic subjects (age 50 ± 9 years; 63% men). Interestingly, 40 participants (4%) had exclusively noncalcified plaque; 10 of these subjects had significant stenosis on CTA (defined as > 50%), and 5 of them had severe stenosis (defined as > 70%). Midterm follow-up of the study population revealed 15 events; however, only one of these events was ACS, and the rest were revascularizations, mostly prompted by the results of the scan. The study showed that the prevalence of significant noncalcified atherosclerotic lesions in asymptomatic individuals is low (1%) but non-negligible. With the additional costs and risks involved with CTA, there is (1) a need for more robust data to prove that CTA can provide prognostic information that is independent of and incremental to CAC imaging and (2) a need to lower the radiation burden significantly before it can replace the latter as a tool for CAD screening and risk stratification.

Figure 20-15 Computed tomography coronary angiography (CTCA) for comprehensive evaluation of coronary atherosclerosis. Left, CT coronary angiogram showing the outline of coronary arteries in a 45-year-old man with atypical chest pain and equivocal exercise ECG test. Severe stenosis was noted in proximal right coronary artery (RCA). Right, Curved multiplanar reformation showing a large, predominantly noncalcified plaque at the site of stenosis. Coronary artery calcification (CAC) imaging showed minimal coronary calcification (<10 Au) in this patient.

There is active ongoing research in the field of atherosclerosis imaging, with an intense pursuit for imaging modalities that can assess not only the extent and severity of the plaque but also its composition and vulnerability.¹⁹⁹ Newer imaging modalities are in various stages of development and mostly rely on the use of specific antibodies and ligands with high affinity for various components of atherosclerotic plaque. Since inflammation plays an important role in all stages of atherosclerosis, macrophages have been an attractive target for most of these modalities. MRI studies using targeted immuno-micelles showed strong correlation with macrophage content in atherosclerotic plaques²⁰⁰; however, these studies were mostly limited to extracardiac vessels. Also, positron emission tomography (PET) studies using fluorodeoxyglucose (FDG) demonstrated increased FDG uptake in atherosclerotic plaques, and there was a significant correlation between the PET signal from carotid plaques and the macrophage staining from corresponding histologic sections (r = 0.70; P < 0.0001).²⁰¹ Similarly, in rabbits, specific uptake of an iodine-containing contrast agent by macrophages allows atherosclerotic lesions to be detected using CT.²⁰² Pending further clinical trials, this contrast agent may become an important adjunct to the clinical evaluation of coronary arteries with CT. These evolving techniques, when successfully validated for use in humans, can potentially transform cardiovascular imaging from a method of diagnosis in symptomatic patients to a tool for noninvasive detection of early subclinical abnormalities. Until such time, CAC imaging, by virtue of its established efficacy and ease of use, remains the method of choice for noninvasive detection of coronary atherosclerosis.