Coronary CT Angiography for the Detection of Obstructive Coronary Artery Disease
- First Online:
- Cite this article as:
- Maurovich-Horvat, P., Ghoshhajra, B. & Ferencik, M. Curr Cardiovasc Imaging Rep (2010) 3: 355. doi:10.1007/s12410-010-9045-5
Contemporary CT scanners offer high temporal and spatial resolution, permitting visualization of the rapidly moving heart and coronary arteries. The imaging of coronary artery lumen and detection of obstructive coronary artery disease is feasible with 64-detector-row and higher generation CT scanners. The diagnostic accuracy of coronary CT angiography as compared to invasive coronary angiography is good (sensitivity of 85%–100%, specificity of 85%–99%). The major strength of coronary CT angiography is the high negative predictive value (96% to 99%) that permits excluding significant coronary artery stenosis with high accuracy, when optimal image quality is achieved. Therefore, coronary CT angiography is an appropriate diagnostic test for a selected patient population with a low to intermediate probability of coronary artery disease.
KeywordsCardiac computed tomographyCoronary computed tomography angiographyCoronary artery diseaseAtherosclerosisDiagnostic imagingCoronary stenosis
Introduction: Current and Emerging Techniques of Coronary CT Angiography
Since the introduction of electron beam CT in the early 1980s and subsequent multidetector-row CT technology in the late 1990s, the field of cardiac CT has rapidly evolved. Contemporary CT scanners offer high temporal and spatial resolution, permitting visualization of the rapidly moving heart. Scanners with 64 detector rows are now considered a minimum requirement to perform noninvasive coronary CT angiography studies. The latest scanners include up to 320 detector rows and up to two tube-detector pairs, each offering image quality beyond 64-detector-row CT.
These scanners provide a slice collimation (z-axis resolution) of 0.50–0.75 mm and rotation times of 270–420 msec. All state-of-the-art scanners offer an isotropic spatial resolution at approximately 0.5 × 0.5 × 0.5 mm, and a temporal resolution of 75–210 msec. Acquisition of a complete high-resolution, volumetric dataset of the heart now takes less than 10 s (and often <0.5 s), therefore offering diagnostic scans even in patients with limited breath-hold capacity. Coronary CT angiography depends on adequate enhancement of the vessel lumen, and this is achieved by intravenous injection of 50 to 120 mL iodinated contrast medium at flow rates up to 8 mL/sec.
The image quality strongly depends on adequate patient preparation in addition to the advanced hardware and image post-processing software. The most important predictor of image quality is the heart rate of the patients and is preferably lowered to <65 beats per minute, although a stable heart rhythm is equally important. In addition to the heart rate control (usually with oral and/or intravenous β-blockers), sublingual administration of short-acting nitroglycerin immediately prior to the scan is used to dilate the coronary arteries, thus improving the visualization of small vessels.
The emerging technologies in coronary CT imaging include dual-source CT, wide-area coverage multidetector-row CT, and high-definition CT scanners. Dual-source CT scanners have two x-ray tubes and detectors 90 degrees apart. A dual-source CT gantry architecture provides improved temporal resolution, resulting in decreased motion artifacts [1•, 2]. With this higher temporal resolution, β-blocker premedication is sometimes not necessary. At higher heart rates, higher pitch values (table feed per gantry rotation) are possible, thus paradoxically reducing radiation dose if all other factors are held constant [1•, 3]. However, at high heart rates, advanced dose reduction techniques such as prospective triggering, or high-pitch coronary CT angiography might not be feasible [4•, 5].
Further technologic advances include the introduction of the wide-area coverage multidetector-row CT, such as the 256 and 320-detector-row scanners [6–8]. The wide-area coverage scanners allow simultaneous imaging of the entire heart, thereby alleviating the stair-step artifacts that result from misalignment of the heart at successive tube rotations. Utilizing the prospectively ECG-triggered scan mode, significant reduction in radiation dose can be achieved.
High accuracy of coronary CT angiography depends on obtaining motion-free images of the heart and theoretically requires a temporal resolution of less than 50 msec, thus noninvasive coronary CT angiography cannot be expected to widely replace the invasive coronary imaging techniques in the near future, especially in patients with high and irregular heart rates. Furthermore, image interpretation can be difficult or even impossible in patients with pronounced coronary calcification. The high-density calcium deposits appear enlarged (or bloomed) in part because of partial volume averaging effect and may obscure the adjacent coronary lumen. The motion and blooming artifacts can be reduced by increased spatial and temporal resolution. Thus, further improvements in scanner technology may lead to a reduction in calcium-related image artifacts in the near future. A high-definition 64-slice CT scanner with improved in-plane spatial resolution was recently introduced, which may allow reduced “blooming” artifacts caused by high-density structures such as coronary calcification and stents . Promising results have been published using novel iterative reconstruction algorithms to reduce blooming artifacts and to improve the contrast-to-noise ratio of the images [9, 10].
Because ECG synchronization is mandatory for coronary CT angiography, patients with atrial fibrillation and other severe arrhythmias will pose a challenge for CT imaging, often requiring prohibitive radiation doses. At high temporal resolutions such as with dual-source CT, imaging of patients with atrial fibrillation may be feasible. Preliminary data suggest that dual-source CT allows scanning patients with rapid and/or irregular heart rate due to the high temporal resolution of the system [11–13]. Coronary CT angiography performed with 256 and 320-detector-row CT scanners may also be beneficial in patients with severe arrhythmia .
Evaluation of Coronary CT Angiogram
Qualitative Assessment of Coronary Stenosis
Using mostly quantitative assessment, coronary CT angiography has been shown to consistently achieve high sensitivity and specificity for the detection of hemodynamically significant coronary artery stenosis.
Early data derived from 4-detecor-row and 16-detector-row CT scanners established the role of CT in the field of coronary imaging. Current-generation 64-, 256-, 320-detector-row, and dual-source CT scanners are now widely available for routine coronary examinations.
Multicenter studies comparing 64-detector-row CT to invasive coronary angiography for detecting coronary artery stenoses
Dual-source CT permits imaging with improved temporal resolution and reduction of motion artifacts. The diagnostic performance of coronary CT angiography using first-generation dual-source CT was reported at 88% to 100% sensitivity and 90% to 98% specificity [1•, 30–32]. Furthermore, the number of non-evaluable coronary segments has decreased while the high negative predictive value is maintained.
Quantitative Assessment of Coronary Stenosis
The majority of studies comparing coronary CT angiography with invasive coronary angiography have uniformly used a binary cut point (stenosis >50% or >70%) to define obstructive coronary artery disease (CAD). However, more exact grading of stenosis severity may be more appropriate for clinical practice and even more important for research studies. Early studies that attempted to assess stenosis severity used visual assessment or semiautomatic approaches.
The studies with 16-detector-row and 64-detector-row CT scanners found contradicting results. The results reflected problems with image quality, image artifacts, and limited reproducibility and accuracy of the quantitative assessment of stenosis. Cury et al.  reported an excellent correlation between 16-detector-row CT and quantitative coronary angiography (QCA). In contrast, an analysis performed by Leber et al.  showed only a moderate correlation between 64-detector-row CT and QCA. The discrepancy between lesion quantification by two techniques was most pronounced in the intermediate range (20%–70%) . Furthermore, several studies have described a tendency of coronary CT angiography to underestimate stenosis severity in the presence of noncalcified plaque and to overestimate stenosis severity in the presence of calcified plaque [34, 35].
It may be more appropriate to use intravascular ultrasound (IVUS)–based quantification of coronary stenosis when assessing the accuracy of coronary CT angiography. Indeed, recently Joshi et al.  demonstrated that the stenosis severity as assessed by coronary CT angiography correlated with IVUS, but not with QCA. Voros et al.  compared the diagnostic accuracy of coronary CT angiography against IVUS with radiofrequency backscatter analysis (IVUS/VH) and QCA. Again, the study demonstrated that IVUS might be more appropriate than QCA for the comparison of the quantification of stenosis severity by coronary CT angiography .
The new CT scanner generations with wide area detectors and improved temporal and spatial resolution have a higher accuracy for quantification of coronary stenosis. Cheng et al.  reported an excellent correlation between dual-source CT and QCA coronary stenosis quantification using a multi-tiered grading system. A strong correlation between coronary CT angiography and QCA stenosis assessment was reported using a 256-slice CT scanner with a semiautomatic stenosis quantification tool .
With the advent of automated dedicated quantitative coronary CT angiography (QCCTA) tools, more accurate and highly reproducible stenosis quantification might be anticipated. In a recently published study, Boogers et al.  reported promising results regarding automated QCCTA that yielded an improved diagnostic accuracy over the visual stenosis assessment.
Role of Coronary CT Angiography in Clinical Practice
The very high negative predictive value (95% to 99%) that has been consistently shown by accuracy studies highlights the clinical role of noninvasive coronary CT angiography. The high negative predictive value suggests that 64-detector-row and higher generation scanners can reliably exclude significant coronary artery stenosis. Thus, if a coronary CT angiogram with good image quality demonstrates no coronary luminal narrowing, the patient therefore has a very low likelihood of having a hemodynamically significant coronary artery stenosis. An important caveat is that most CT studies enrolled a patient population with relatively low prevalence of significant CAD (in per coronary segment analysis), thus results may apply only for similar patient populations (eg, for patients with low pretest likelihood of CAD). Large multicenter, multivendor, randomized clinical trials are currently being conducted to further clarify the clinical role of coronary CT imaging. Given the noninvasive nature and the high negative predictive value of this diagnostic test, the most reasonable application is excluding significant CAD in patients with low to intermediate pre-test likelihood of significant CAD . Currently, the unselected screening of asymptomatic individuals is not indicated. Patients with high pre-test likelihood of significant CAD should undergo invasive coronary angiography, in part because of the possibility of an immediate percutaneous coronary intervention.
Recommendations of Professional Societies
Evaluation of chest pain syndrome in patients with intermediate pre-test probability of CAD in a patient with uninterpretable ECG or unable to exercise (appropriate, score 7);
Evaluation of acute chest pain in patients with intermediate pre-test probability, no ECG changes, and negative serial cardiac biomarkers (appropriate, score 7);
Evaluation of chest pain syndrome in patients with prior test results when stress test (exercise, perfusion, or stress echocardiogram) was uninterpretable or equivocal (appropriate, score 8).
The imaging of patients that are asymptomatic or have high pre-test probability of CAD is inappropriate according to the expert consensus document [42••]. However, several prospective studies are ongoing at this time to evaluate the incremental value of contrast-enhanced coronary CT angiography in addition to coronary calcium scanning in asymptomatic subjects with high Framingham risk (eg, Detection of Subclinical Atherosclerosis in Asymptomatic Individuals [Decide CTA]; NCT00862056).
The benefit of coronary CT angiography is likely to be greatest and is reasonable for symptomatic patients who are at intermediate risk for CAD after initial risk stratification, including patients with equivocal stress test results (Class IIa, level of evidence B).
A similar principle of selection of symptomatic patients with intermediate likelihood of CAD for coronary CT angiography is included in the more recent consensus documents by the American College of Cardiology, the American Heart Association, the Society of Cardiovascular Computed Tomography, and the European Society of Cardiology [41••, 44•, 45].
Coronary Atherosclerotic Plaque Imaging
One of the major strengths of coronary CT angiography is the ability to detect nonobstructive coronary plaques [8, 46, 47]. Beyond detection, the classification of atherosclerotic plaques into calcified, partly calcified, and noncalcified lesions is feasible. However, data supporting the role of CT plaque assessment for the prediction of occurrence of future acute coronary syndromes are sparse. Large prospective imaging studies similar to those available for myocardial stress testing are warranted. There are preliminary data suggesting that presence of nonobstructive atherosclerotic lesions in all three coronaries is associated with increased mortality . Another study investigating a symptomatic patient population showed that detectable atherosclerotic plaques in at least five coronary segments are associated with increased mortality rate [49•]. In a recently published study, CT characteristics of plaques subsequently causing acute coronary syndrome were described. The study demonstrated that plaques with low CT attenuation and positive remodeling are at particularly high risk for causing future cardiovascular events . On the other hand, negative coronary CT angiography scan and/or minimal burden of CAD predict a very low cardiovascular event rate during the follow-up of up to 2 years . Ongoing multicenter, randomized studies will provide further refinement regarding the prognostic values of coronary CT angiography examinations and appropriate populations.
Coronary Artery Bypass Graft Assessment
Reliable coronary CT angiography in patients with prior coronary artery bypass graft (CABG) surgery is possible because of the grafts’ larger diameter (typically 3–4 mm), and reduced mobility. Assessment of arterial grafts can be challenging due to smaller vessel diameter (typically 1–2 mm) and artifacts caused by metallic surgical clips. These features may obscure the lumen. Nevertheless, significant stenosis or occlusion of bypass grafts can be detected with very high accuracy [52–54]. The most recent meta-analysis (15 studies, 723 patients) on the diagnostic accuracy of 16 or 64-detector-row CT scanners to detect significant stenosis or occlusion of bypass grafts showed sensitivity of 98%, specificity of 97%, positive predictive value of 93%, and negative predictive value of 99% [55•]. Notably, 8% of the grafts were excluded from the analysis due to artifacts.
Assessing the native coronary arteries in symptomatic patients with history of CABG surgery is of clinical importance. A limited number of studies has been performed to assess the performance of coronary CT angiography in the evaluation of both bypass grafts and native coronary arteries [52, 56, 57]. The analysis of native coronary arteries was challenging due to the frequently small vessel sizes, extensive calcification, and presence of metallic surgical clips at the anastomoses. Consequently, 9% to 34% of native coronary arteries were excluded from the analysis due to image artifacts. The hampered accuracy for the detection of significant stenosis in native coronary arteries was reflected in the lower sensitivity (79%–86%) and specificity (72%–76%) values compared to the graft assessment. Therefore, routine assessment of both bypass grafts and native coronary arteries is challenging.
Evaluation of bypass grafts and coronary anatomy by coronary CT angiography in subjects with chest pain syndrome (uncertain, score 6);
Evaluation of bypass grafts and coronary anatomy by coronary CT angiography in asymptomatic patients less than 5 years after the surgery (inappropriate, score 2);
Evaluation of bypass grafts and coronary anatomy by coronary CT angiography in asymptomatic patients greater than 5 years after the surgery (inappropriate, score 3).
Evaluation of Coronary Stents
The diagnostic visualization of stents is challenging with current CT scanners. The three most common artifacts that may prevent stent evaluation with coronary CT angiography are beam-hardening artifacts, partial volume averaging, and motion artifacts. Beam-hardening artifacts are caused by the metallic stent struts, which absorb much more of the lower-energy portion of the x-ray photons, thus leaving only high-energy photons which pass through the neighboring areas with little absorption. This results in a low-density area in the reconstructed image, which can mimic in-stent restenosis. Partial volume averaging affects the voxels immediately adjacent to the stent struts. Due to the limited spatial resolution of the CT scanners, the individual voxel might contain density information from both the low-attenuation tissue and the high-attenuation stent material. Thus, the stent appears much larger (ie, “bloomed”) and may obscure the adjacent coronary lumen. Sharper image reconstruction kernel filters and thinner reconstructed slices (0.4–0.6 mm) may reduce these artifacts and improve image quality. Motion artifacts are the most common reason for non-evaluable stented segments. Furthermore, motion artifacts tend to worsen the artifacts described above. True improvement might be expected in the future from scanners with increased spatial and temporal resolution and with the advent of advanced image reconstruction algorithms .
Currently, only a modest diagnostic accuracy of coronary CT angiography for the detection of in-stent restenosis can be achieved [58–61]. Notably, a significant portion of the imaged stents (5%–42%) was excluded from the analysis due to artifacts. In general, larger stents are easier to evaluate. However, even with stents with diameters greater than 3 mm, up to one in five stents proved to be non-evaluable [61, 62]. Stent design (ie, material and type) can also impact the stent visualization by CT . A recently published meta-analysis (14 studies, 895 patients) of 64-detector-row CT for the detection of in-stent restenosis confirmed a modest diagnostic accuracy with a sensitivity of 87%, specificity of 84%, positive predictive value of 53%, and negative predictive value of 97% .
Evaluation of stents by coronary CT angiography in subjects with chest pain syndrome (uncertain, score 5);
Evaluation of stents in asymptomatic patients (inappropriate, score 2).
Reduction of artifact caused by stents is critical prior to the implementation of CT for stent assessment in the clinical routine.
Radiation Dose in Coronary CT Angiography
Coronary CT angiography requires at least a moderate radiation dose with current techniques; like other diagnostic modalities that involve ionizing radiation, the clinical need for the test must be weighed against any risks, and appropriate steps must be taken to use an appropriate radiation dose. The radiation dose required to obtain a diagnostic coronary CT angiogram varies with the acquisition parameters, patient parameters, and the specific equipment available in the CT laboratory. In recent years, an increasing number of dose-saving measures have also become available to the cardiac imager. As a result, radiation doses have generally decreased, sometimes dramatically.
Radiation doses for 64-detector-row and higher cardiac CT are generally similar or less than that of a dual-isotope nuclear myocardial perfusion scan (12–15 mSv) . The largest driver of patient dose is the heart rate. Although the retrospective gating required to perform ECG-synchronized cardiac CT initially required a relatively high dose, ECG-based tube current modulation has allowed significant dose savings by reducing the dose during systole. If heart rate and rhythm permits, prospective triggering can be performed, where acquisition only occurs intermittently at the desired cardiac phase (sacrificing systolic images since they will likely not be needed) . Although no amount of radiation should be considered negligible, some coronary CT angiograms can even be performed with doses less than 1 mSv, an amount considered “low dose” by the American College of Radiology .
Dose savings have also been achieved through other means, such as using a lower x-ray tube voltage (ie, low-kVp settings), and by lowering the tube current (mA) based on patient size (including automatic adjustments using the patient’s scout radiographs) [4•, 32, 68]. By tailoring the examination to the individual patient, diagnostic image quality can be maintained with significant dose reduction beyond empiric settings [32, 69, 70]. The anatomy scanned should be limited as closely as possible, and patient positioning can also reduce dose (such as moving female breasts away from the scan field of view) .
Future dose reduction methods include iterative reconstruction algorithms, increasing detector efficiency, and faster temporal resolution (allowing higher heart rates with maintained image quality) .
Contemporary CT scanners with high temporal and spatial resolution permit the visualization of the rapidly moving heart and coronary arteries. Scanners with 64-detector-rows are now considered the minimum requirement to perform noninvasive coronary CT angiography studies. Coronary CT angiography can rule out significant coronary artery stenosis with high accuracy, when optimal image quality is achieved. However, sufficient experience and skill in both image acquisition and interpretation is mandatory to achieve reliable results. Large prospective trials are currently underway to clearly establish the role of coronary CT angiography imaging in the triage of patients with suspected CAD. Coronary CT angiography is an appropriate diagnostic test for a selected patient population with a low to intermediate probability of CAD. However, the broad use of diagnostic CT angiography in all patients with suspected obstructive CAD is not feasible nor appropriate with today’s scanner technology.
No potential conflicts of interest relevant to this article were reported.