Applied Radiology

Dr. Becker is an Associate Professor and the Section Chief of Body CT, Department of Clinical Radiology, University Hospital Munich, Grosshadern, Germany.

In the United States, there are 1.5 million heart attacks annually and 1 of 2 adults dies from heart and vascular disease. The leading cause of unheralded heart attacks is rupture of vulnerable atherosclerotic plaques, resulting in thrombosis and sudden coronary occlusion. However, many more of the pre-existing vulnerable atherosclerotic lesions are in nonstenotic rather than in stenotic coronary artery segments and, therefore, do not cause any symptoms prior to an event. ¹

Because of this fact, vulnerable plaques may not be reliably assessed by coronary angiography or by electrocardiogram (ECG) stress testing. Noninvasive ultrasound and Doppler are able to assess atherosclerosis in several different arterial vessel regions but not in the coronary arteries. With computed tomography (CT), coronary calcifications can be detected easily; therefore, it allows for a fast and noninvasive assessment of coronary atherosclerosis. ²

Histopathology of coronary calcium

The presence of arterial calcifications is specific for atherosclerosis. In rare cases (eg, hypervitaminosis D, Mönckeberg's sclerosis, and infantile calcifications), vessel calcifications may mimic classical atherosclerotic calcifications. In classical atherosclerotic lesions, precipitates of calcium phosphate develop between the intimal and medial layers of the vessel wall by mechanisms similar to those known from osteoneogenesis. From the pathologic point of view, vascular calcifications are found in fibroatheromas, an advanced and nonreversible stage of atherosclerosis. ³ From the clinical point of view, coronary artery calcifications correlate neither with the site nor with the degree of coronary artery disease (CAD) with vessel stenoses and myocardial ischemia. ²

Coronary calcium and coronary artery disease

Agatston et al ⁴ first reported the detection of coronary calcifications by electron beam CT (EBCT). They investigated 475 patients with CAD and 109 patients without CAD by EBCT as well as by fluoroscopy, and divided the patients into different age groups (30 to 39, 40 to 49, 50 to 59, 60 to 69 years). In every patient, a scan of the first 60 mm of the heart was acquired, with 3-mm consecutive slices in a breath-hold period of 30 seconds. To reduce cardiac motion artifacts, the images were acquired with prospective ECG triggering and exposure time of 100 milliseconds. To distinguish between coronary calcification and noise in EBCT images, a threshold of 130 Hounsfield units (HU) was used. Fluoroscopy and EBCT detected coronary calcifications in 90% and 52% of patients, respectively. Every plaque in the EBCT images was ranked by its HU density (1 = 130 to 199 HU, 2 = 200 to 299 HU, 3 = 300 to 399 HU, 4 = >= 400 HU) and multiplied by the area of the plaque. A score for every patient was determined by the sum of all plaque numbers. They reported that the score in every age group was significantly different in patients with and without CAD.

When scanning for coronary calcifications became more popular, the American Heart Association released guidelines for coronary calcium screening. ⁵ According to these guidelines, there is no indication for screening symptomatic patients with obvious myocardial ischemia. Coronary screening may be indicated only for symptomatic patients with atypical chest pain, for asymptomatic patients with cardiovascular risk factors, and for a follow-up investigation in patients under lipid-lowering therapy.

Laudon et al ⁶ demonstrated that in patients presenting to an emergency department with atypical complaints, the exclusion of coronary calcifications by EBCT has a high negative predictive value (95% to 98%) in ruling out CAD. If coronary calcifications are present, these patients still require further investigations, such as ECG stress testing. In younger patients suffering from atypical chest pain, coronary calcifications may be a rare finding, and therefore EBCT fails to rule out CAD. These patients may have noncalcified plaques only, resulting in coronary artery stenoses. Therefore, in our institution, we have decided to perform CT angiography in any symptomatic patient to investigate calcified, as well as noncalcified, lesions.

Coronary calcium and future cardiac events

Conventional risk-factor assessment can help to predict the risk for future cardiac events in a cohort population, but not in the individual patient. The most challenging aspect of the detection of coronary calcification by CT may be the ability to predict cardiac events in each asymptomatic patient. Arad et al ⁷ followed 1173 asymptomatic patients over a 19-month period. In this patient population, they observed 26 soft (revascularization and bypass grafting) and hard (myocardial infarction and death) events in 18 patients. On the base of the data collected, an odds ratio of 35.4 was calculated for a cardiac event in patients with a score above 160.

Raggi et al ⁸ published score percentiles on the base of 9728 patients investigated with an EBCT. In a control group study of 172 patients with myocardial infarction and 632 without, they found that 70% of all events took place in patients who were above the 75 ^th score percentile.

Wong et al ⁹ followed 928 asymptomatic patients over a period of 3.3 years and observed 28 soft and hard events. Patients were assigned to the 4 ^th quartile if they were above the 75 ^th score percentile from 21,000 patients (relative ranking) or above the score of 105 (absolute ranking). They found that the odds ratio was not significantly different between the relative and absolute ranking (6.02 and 6.37, respectively). Because of the ease of use, we perform absolute ranking for risk stratification in asymptomatic patients.

All of the currently available literature on the predictive value of coronary calcium is limited by many different factors. Most of the patients enrolled in these studies were not randomized but self-referred; many of the patients were aware of their scan results and may have changed their lifestyle accordingly; the study end points were coronary revascularization and bypass grafting instead of only hard events. ¹⁰

The PACC, ¹¹ RECALL ¹² and MESA ¹³ trials are prospective randomized cohort studies with patients who are unaware of their scan results and use hard events only as the final end point. They will enroll a total of more than 10,000 patients to prove the hypotheses that coronary calcifications have a predictive value for hard events independent of, and superior to, conventional risk-factor assessment. The final results of these trials will be available no earlier than 2010.

Progression of coronary calcium

Callister et al ¹⁴ reported that the traditional scoring algorithm is inappropriate for following the progression of coronary calcifications over time. Therefore, he introduced the volume calcium measurement from EBCT images and followed 149 asymptomatic patients over a period of 12 to 15 months. ¹⁵ A total of 65 patients with a mean low-density lipoprotein (LDL) <120 mg/dL and 40 patients with a mean LDL >120 mg/dL received lipid-lowering medication. The remaining 44 patients presented with a mean LDL >120 mg/dL and did not receive any medication. The effectively treated patient group demonstrated a 7% regression in calcium volume, while the noneffectively treated patient group had a moderate progression of 25% of the calcium volume. A mean progression rate of 52% was found in the nontreated control group. It should be emphasized that the variability of a repeated coronary calcification measurement with EBCT is >10% and a regression of coronary calcifications is very unlikely from the pathologic point of view. ¹⁶

Coronary calcium with multidetector row CT

We were able to demonstrate that conventional CT is as sensitive as EBCT in detecting coronary calcifications. ¹⁷ Adjusting conventional CT scan parameters to acquire image quality similar to EBCT also allows for determination of the traditional calcium score. ¹⁸ However, the score originally defined by Agatston et al ⁴ is a dedicated semiquantitative quantification algorithm designed for the EBCT and it is difficult to reproduce by conventional CT when scan parameters are changed.

After helical scanning with multi-detector row CT (MDCT), retrospective ECG gating with overlapping incremental reconstruction is performed to improve reproducibility of the calcium measurement. Multiplying the area, slice increment, and HU-density allows for the determination of the mass equivalent of calcified plaques in MDCT images. Using a phantom with calcium inserts of known mass allows for determination of a calibration factor for different MDCT scan protocols.

Multiplying the mass equivalent by the calibration factor obtained from the calibration phantom results in the absolute mass of coronary calcifications in milligrams of calcium hydroxy-apatite (mg CaHA). Multidetector row CT is able to detect coronary calcification with a sensitivity of 1 mg CaHA. In patients with extensive coronary calcifications, a calcium mass of more than 1.5 g CaHA may be found.

The traditional score is approximately 5 times higher than the calcium mass. Only with the calibrated mass measurement, however, can MDCT scan protocols be modified with the same results. For example, we found that the absorption for coronary calcium is higher, at 80 kV compared with 120 kV. Using 80 kV instead 120 kV allows reduction of radiation for screening by 30% without loss of information. Changing the tube voltage will influence the density of plaques and therefore the traditional score. The calibrated calcium mass, however, will remain the same with both CT protocols.

Practical consideration

We still follow the recommendations of the American Heart Association with the following modifications.

Patients with unstable angina tend to have fewer calcified plaques, but more noncalcified plaques, than patients with stable angina. In younger symptomatic patients, noncalcified lesions may be present even in the absence of any calcified lesions. For example, in a patient with atypical chest pain, we observed a noncalcified plaque with a low density (20 HU) that seemed to correspond to an intracoronary thrombus. ¹⁹ Therefore, we decided to perform a contrast-enhanced MDCT angiography in any symptomatic patient referred for CT investigation of the coronary arteries.

In asymptomatic patients, we first perform nonenhanced CT studies of the coronary arteries. In cases in which no calcifications are present, we do not recommend any treatment and simply reassure these patients of their low-risk. In asymptomatic patients with a positive coronary calcium scan and a calcium mass <20 mg CaHA (traditional score of 100), we recommend that patients modify their cardiovascular risk factors and follow the progression of the calcifications within the next 5 years. If the amount of coronary calcium exceeds 20 mg CaHA, we recommend performing further testing for CAD. We never recommend performing a cardiac catheter investigation on the basis of an unenhanced coronary CT investigation alone.

However, neither the presence nor the amount of coronary calcium accurately reflects the extent of coronary atherosclerosis. In asymptomatic patients, contrast-enhanced CT angiography can detect calcified as well as noncalcified lesions. We have found that noncalcified lesions with low CT densities (40 HU) may correspond to lipid-rich plaques (atheromas), which are more prone to rupture and to cause an acute event than are plaques with high CT densities (90 HU) that more likely contain fibrotic tissue (fibroatheromas).

Conclusion

Currently, many MDCT scanners are widely available. Since we have shown that MDCT is as feasible as EBCT for the detection and quantification of coronary calcifications, non-invasive assessment of coronary atherosclerosis is no longer limited to special centers with an EBCT scanner. Current clinical applications of coronary MDCT may include the evaluation of symptomatic patients with atypical chest pain to rule out CAD, as well as the investigation of asymptomatic patients to predict the risk of future cardiac events and to follow the effectiveness of risk-factor modification. Assessment of coronary atherosclerosis and CAD may further improve with the use of contrast-enhanced MDCT studies. AR