Scan procedure

Multislice CT-scanners consist of a gantry and a patient couch 1. The gantry, which contains an x-ray tube opposite a row of multiple thin detectors, rotates around the patient couch which is moved through the gantry (Figure 1). The tube produces a fan-shaped x-ray beam, which is collimated and hits the detector row after it has passed through the patient. The degree of x-ray absorption is determined by the atomical density of the tissue, with a higher atomic tissue density resulting in a higher attenuation. The attenuation is described as an absolute value measured in CT-numbers, with the attenuation of water as a reference. A high attenuation, such as for iodine and calcium, results into a white appearance on the CT-image. Conversely, lung tissue has a low attenuation and therefore appears black. An intermediate attenuation, such as for myocardium, will result in a grey appearance. A cross-sectional image, or tomogram, can be reconstructed from a 180° rotation. Therefore, the temporal resolution, which is defined as the duration of the reconstruction window during which time a tomogram is reconstructed, equals half the x-ray tube rotation time.

Computed tomography of the coronary arteries: An alternative?

All authors

Matthijs F. L. Meijs, Willem B. Meijboom, Maarten-Jan M. Cramer, Francesca Pugliese, Mathias Prokop, Pieter A. Doevendans & Pim J. De Feyter

https://doi.org/10.1080/14017430701509862

Published online:

12 July 2009

Figure 1.  The x-ray tube and the detectors rotate in an opposing position of the gantry around the patient. A collimated x-ray beam is passed through the patient and the attenuated x-ray beam is collected on the detectors while the patient on the coach is continuously advanced through the gantry. 64-parallel detector rows acquire the data in a very short scan time.

Display full size

Multislice CT-imaging can be performed either in the sequential mode or in the spiral mode. In the sequential mode, one axial image is produced while the table remains motionless. For each subsequent slice, the table is moved forward. In contrast, in the spiral mode the table is moving continuously through the gantry, while the gantry rotates continuously around the patient. This method produces a dataset from which cross-sectional images can be reconstructed.

CTCA can be preceded by a non-contrast enhanced sequential scan for coronary artery calcification measurement (see below). The actual CTCA scan is usually done in the spiral mode, but can be done in the sequential mode. For the CTCA iodine contrast is injected, usually via a cubital vein. To synchronize image acquisition to the movement of the heart, an ECG is recorded simultaneously which is used for either prospective triggering or retrospective gating (Figure 2). In the sequential mode, scanning takes place only during a predefined phase in the cardiac cycle, as determined by prospective triggering. In the spiral mode, the ECG is used to retrospectively select the optimal reconstruction window with the least cardiac motion artefacts. This is usually during end-diastole, when the heart is in the iso-volumetric filling phase and motion is minimal. However, in some patients, especially in those with higher heart rates and with regard to the more mobile right coronary artery, the end-systolic phase can also provide valuable information.

Computed tomography of the coronary arteries: An alternative?

All authors

Matthijs F. L. Meijs, Willem B. Meijboom, Maarten-Jan M. Cramer, Francesca Pugliese, Mathias Prokop, Pieter A. Doevendans & Pim J. De Feyter

https://doi.org/10.1080/14017430701509862

Published online:

12 July 2009

Figure 2.  ECG-synchronized image reconstruction with sequential scan protocols using prospective ECG triggering to synchronize the data acquisition and reconstruction window to the motion of the heart. Spiral CT scanning protocols acquire data continuously and record the ECG during the scan. Isocardiophasic images are reconstructed using retrospective ECG-gating. The reconstruction window can be positioned any where within the R-R-interval.

Display full size

X-ray exposure is expressed as the effective dose, which reflects the potential biological effect of the radiation and is expressed in mSv (Sievert). 64-slice CTCA is associated with 15/21 (male/female) mSv, compared to 4–6 mSv for CCA 2. To reduce the radiation dose associated with CTCA, prospective ECG-triggered x-ray tube dose modulation, also called pulsing, can be applied. In this technique, the full tube current is only applied during a predefined phase, e.g. end-diastole. While prospective triggering and retrospective gating combined with dose modulation can reduce radiation dose by approximately 30 to 50%, they inherently offer less flexibility to cope with artefacts related to coronary motion and arrhythmias.

Image quality parameters

Both a high spatial and temporal resolution is required to visualize the small and rapidly moving coronary arteries. In addition, the scan must be performed during one breath hold to prevent breathing artefacts.

The spatial resolution in the x/y plane of the 64-slice CT scanners is 0.4×0.4 mm. The spatial resolution in the z-axis is dependent on the minimum slice thickness, which is 0.4 to 0.625 mm, depending on the manufacturer. For a better signal-to-noise ratio reconstructions which are slightly thicker than the minimum slice thickness are recommended. Therefore, nearly isotropic (same size in every dimension) voxel size can be reconstructed. This is an enormous improvement compared to previous generations of CT-scanners, yet compared to the 0.2 mm spatial resolution of CCA this is still limited.

As explained above, the temporal resolution is directly dependent on the gantry rotation time. Dependent on the manufacturer, for modern 64-slice CT scanners this means a temporal resolution of 165 to 210 ms. Thus, reliable visualisation of coronary arteries in patients with a stable heart rhythm below 70 beats per minute is possible. The temporal resolution can be improved 2- to 4-fold by multisegmental reconstruction algorithms, that combine data from two to four cardiac cycles. However, these algorithms are vulnerable to arrhythmias, since they assume identical cardiac contraction patterns.

With the 64-slice CT scanners, image acquisition is completed within 6 to 11 s. This is a manageable breath hold for almost all patients.

Clinical considerations

Patients should be screened for contra-indications to x-radiation, e.g. pregnancy, and contra-indications to contrast material, e.g. known contrast-allergy and renal failure. Patients with an irregular heart rhythm, e.g. atrial fibrillation or frequent ventricular extrasystoles, should not be referred for CTCA because of the disturbance of the ECG-guided image acquisition. Good patient instruction is essential to prevent breathing artefacts. Patients with a cardiac frequency above 70 beats per minutes should receive a beta-blocker i.v. or p.o, provided no contra-indications to beta-blockers are present. In patients with a low and stable heart rhythm, dose modulation can be applied to reduce radiation exposure. To a certain degree, minor heart rhythm irregularities can be compensated for by ECG-editing 3. This technique encompasses the adjustment of the position of the reconstruction windows within the cardiac cycle and the exclusion of premature ventricular beats.

CTCA provides clinically relevant information beyond the coronary arteries 4. Therefore, evaluation of CTCA scans should be done in cooperation between a cardiologist and a radiologist. The axial images should remain the basis for the evaluation of coronary artery stenosis, and are also valuable to screen for extracardiac pathology. Using dedicated software, additional reconstructions are available (Figure 3). Volume rendering (VR) images are useful for an overview of general coronary anatomy, e.g. congenital anomalies. In addition, (curved) multiplanar reconstructed ((c)MPR) and Maximum intensity projections (MIP) can be used for additional assessment of a coronary artery stenosis.

Computed tomography of the coronary arteries: An alternative?

All authors

Matthijs F. L. Meijs, Willem B. Meijboom, Maarten-Jan M. Cramer, Francesca Pugliese, Mathias Prokop, Pieter A. Doevendans & Pim J. De Feyter

https://doi.org/10.1080/14017430701509862

Published online:

12 July 2009

Figure 3.  A volume-rendered CTCA image (A) reveals the anatomy of the right coronary artery. A maximum intensity projected CTCA (B) image and a curved multiplanar reconstructed image (C) disclose a short chronic total occlusion in the mid segment of the right coronary artery with distal collateral filling. Furthermore, diffuse coronary artery disease if seen in the distal segment. The conventional coronary angiogram confirms the chronic total occlusion (D).

Display full size

Coronary calcification

An additional non-contrast enhanced low-x-ray dose scan can be used to determine the coronary calcification score. Also in asymptomatic individuals the calcium score predicts coronary ischemic events, independently of traditional risk factors 5. However, the precise additive value of coronary calcium scoring by CT needs further evaluation.

CTCA in stable angina and the acute coronary syndrome

The clinical performance of the successive generations of MSCT-scanners with regard to CTCA has been evaluated extensively in various clinical settings. With ongoing technical progress diagnostic accuracy has increased rapidly.

The most extensively evaluated potential clinical application for CTCA is the possibility to replace CCA in patients with stable angina.

The 16-slice scanner featured a lower temporal (200 ms) and spatial resolution (0.4×04×0.75–0.9 mm) than the 64-slice scanner. Furthermore, the signal-to-noise ratio was lower due to a lower radiation dose which could be maximally generated. In addition, image acquisition took approximately 20 s, leading to a higher incidence of breathing artefacts. The major studies comparing 16-slice CTCA with CAC with regard to the detection of significant coronary stenosis distal coronary segments with a diameter below 1.5 or 2.0 mm were frequently excluded. In addition, approximately 8% of segments were considered not analysable due to motion-artefacts or severe calcification. As displayed in Table I, these studies reported sensitivities ranging from 67 to 98% and specificities from 79 to 98% 6–19. For the remaining 92% of coronary segments, a pooled analysis using a segmental analysis reveals an average sensitivity of 87% and a average specificity of 96% 20. In this pooled analysis, the average positive predictive value is 81% and the average negative predictive value is 98%. Average positive and negative likelihood ratios, which indicate how much a positive test will change the odds of having coronary artery disease, were 21.75 and 0.14, respectively. In a recent multi-center trial the limited diagnostic performance of the 16-slice was confirmed 21. These results argue against routine implementation in clinical practice of the 16-slice scanner.

The improved temporal and spatial resolution and image acquisition time, as described above, and the possibility to use a higher radiation dose have enhanced the clinical performance of the 64-slice scanner 22–30. Small distal segments no longer had to be excluded from analysis. However, a weighted analysis reveals that still approximately 6% of segments were considered not analysable due to motion artefacts or severe calcification. As indicated in Table II, on a segment-based analysis the weighted sensitivity and specificity were 82% and 95%, respectively, and positive and negative predictive values were 72% and 97% respectively 22–30. As displayed in Table III, on a patient-based analysis the weighted sensitivity and specifity were 96% and 90%, respectively, and positive and negative predictive values were 94% and 95%, respectively 22–26, 28–30. At present, complete replacement of CCA by 64-slice CTCA seems not feasible because the positive predictive value of 64-slice CTCA is suboptimal, especially in patients with heavily calcified coronary arteries. However, 64-slice CTCA is highly reliable to rule out coronary artery stenosis. Therefore, CTCA would be valuable as a gatekeeper to CCA in patients with a low to intermediate pre-test risk of significant coronary artery disease.

Such a role for CTCA seems also feasible in the specific setting of pre-operative screening for non-coronary cardiac surgery. In patients undergoing valvular surgery, Meijboom et al. found negative and positive predictive values of 100% and 82% for the presence of coronary artery disease on a patient based analysis 31.

CTCA is also useful in the evaluation of patients with the acute coronary syndrome. Dirksen et al. reported positive and negative predictive values of 85% and 97%, respectively, on a patient-based analysis for the detection of significant coronary artery disease in patients with unstable angina 32. Similarly, Meijboom et al. demonstrated positive and negative predictive values of 96% and 100% in patients with non-ST elevation acute coronary syndrome 2.

In the emergency department, CTCA can be used for the triage of patients with acute chest pain. Hoffman et al. reported a sensitivity of 100% and a specificity of 74% for the detection of significant coronary artery disease in patients with acute chest pain without diagnostic ECG changes and with normal cardiac enzymes on presentation 33. In addition, in the emergency department CTCA may be used to simultaneously rule out coronary artery disease, thoracic aortic aneurysm dissection and pulmonary embolism 34. However, this ‘triple-rule out’ strategy still needs further evaluation.

Patients with acute chest pain who present with diagnostic ECG-changes and/or elevated cardiac enzymes should be referred to the catheterisation-lab immediately.

Performance of CTCA in patients with coronary stents and bypass grafts

The assessment of coronary stents is very difficult. The metal struts appear much larger due to blooming artefacts related to the partial volume effect, which may cause an underestimation of the lumen. In addition, beam hardening artefacts may cause a signal void adjacent to the stent which may resemble a non-calcified plaque. Therefore, the evaluation of neo-intimal hyperplasia or in-stent restenosis is impaired. However, with the advent of the 64-slice CT-scanner the evaluation of in-stent restenosis has become feasible for stents with a diameter over 4 mm (Figure 4). As demonstrated recently by Van Mieghem et al. in patients who underwent CTCA after left main stent implantation, a negative scan could 100% reliably rule out in-stent restenosis 35. However, the limited positive predictive value of 67% still necessitates confirmation of a positive CTCA by CCA.

Computed tomography of the coronary arteries: An alternative?

All authors

Matthijs F. L. Meijs, Willem B. Meijboom, Maarten-Jan M. Cramer, Francesca Pugliese, Mathias Prokop, Pieter A. Doevendans & Pim J. De Feyter

https://doi.org/10.1080/14017430701509862

Published online:

12 July 2009

Figure 4.  A curved multiplanar reconstructed image (A) shows a stent with in-stent restenosis in the circumflex coronary artery, which was corroborated the quantitative conventional coronary angiography (QCA diameter stenosis of 64%) (B).

Display full size

The relatively large diameter and relative immobility of venous bypass grafts make assessment by CTCA less challenging (Figure 5). However, metal artefacts caused by surgical metal clips pose diagnostic difficulty. In addition, the small and often heavily calcified native coronary system of these patients is difficult to assess. In a study by Malagutti et al., sensitivity and specificity for the detection of graft disease were 99% and 96%, respectively, on a segment-based analysis 36. Coronary artery stenoses in native segments distal to the graft were detected with a sensitivity and specificity of 89% and 93%, respectively, on a segment-based analysis.

Computed tomography of the coronary arteries: An alternative?

All authors

Matthijs F. L. Meijs, Willem B. Meijboom, Maarten-Jan M. Cramer, Francesca Pugliese, Mathias Prokop, Pieter A. Doevendans & Pim J. De Feyter

https://doi.org/10.1080/14017430701509862

Published online:

12 July 2009

Figure 5.  (A) Volume-rendered CTCA images show the potential causes of artifacts in bypass imaging. The orifice indicator at the ostium of the venous graft and the vascular clips can cause blooming artifacts which can obscure the underlying lumen or bypass pathology. (B) The two grafts are clearly visible, the LIMA to the distal left artery descending artery bypassing an occluded LAD and the single venous graft to the marginal obtuse. (C) The circumflex artery shows a full metal jacket: the circumflex coronary artery is stented from the proximal left main coronary artery to the distal circumflex and marginal obtuse.

Display full size

Analysis of ventricular function

Provided a sufficient number of phases is obtained, CTCA can additionally measure ventricular ejection fraction, with good correlation to echocardiography and MRI 37.

Technical improvements

Future developments are expected to make CTCA more reproducible and to make it applicable in a broader population of patients. This would require improvements in spatial resolution and acquisition time, but above all in temporal resolution.

An improvement in temporal resolution can be achieved by a decreased rotation time or by an increased number of x-ray tubes. As mentioned above, temporal resolution is directly dependent on the rotation time. However, decreasing the rotation time is associated with an exponential increase in G-forces on the gantry and is therefore technically difficult. To circumvent this, the dual-source scanner was recently introduced 39. This scanner features a gantry composed of two x-ray tubes at an orthogonal angle, each opposite a detector row. To obtain a 360° tomogram, a 90° rotation is now sufficient. Thus, temporal resolution is improved two fold as compared to a conventional scanner. More than two x-ray tubes could improve temporal resolution even further. The performance of the dual-source scanner with a temporal resolution of 83 ms is currently being evaluated. It is expected that the increased temporal resolution will make the use of beta-blockers redundant.

With improving temporal resolution, scan results are expected to become more reproducible. Thus, the number of segments that can not be judged is expected to decrease. Also, a higher reproducibility would enable a wider application of dose modulation or even prospective triggering in order to reduce radiation exposure.

Increasing the spatial resolution by simply decreasing the width of the individual detectors would require an exponential increase in radiation exposure to prevent a decreased signal-to-noise ratio 40. Relatively new techniques such as flat panel detectors may prove valuable in this regard 41. However, these applications are still in the developmental stage.

Scanning the entire heart in one gantry rotation can be achieved by increasing the total width of the detector row. This would eliminate artefacts which occur due to irregular cardiac motion on the transition line between consecutive gantry rotations. As demonstrated by Kondo et al., this seems feasible using a 256-slice CT-scanner 42. In addition, the resulting decrease in acquisition time would reduce the incidence of respiratory artefacts, and reduce the required amount of iodine contrast.

Anatomy versus ischemia

CTCA can provide the location and the degree of coronary artery stenosis. However, as demonstrated by Hacker et al. this anatomical information fails to predict the functional relevance of a stenosis 43. However, CTCA was shown to be highly accurate in ruling out functionally relevant stenosis. In contrast, Single Photon Emission Computed Tomography (SPECT) and Positron Emissiong Tomography (PET) can provide reliable information on myocardial blood flow, but tend to underestimate the extent of atherosclerosis. Therefore, according to guidelines a decision to perform revascularisation therapy should be based on both functional and anatomical information. In patients with stable angina, the combination of 64-slice CTCA and SPECT was compared to the combination of CCA and SPECT with regard to the detection of functionally significant stenosis 44. In a patient-based analysis negative and positive predictive values were 93% and 88%, respectively. Sampson et al. studied the diagnostic performance of a hybrid PET-CT scanner in the detection of obstructive coronary artery disease (defined in this case as >/ = 70% lumen narrowing on CCA) 45. They reported a sensitivity of 93% and a specificity of 83%. Importantly, the combination of CTCA and SPECT or PET is associated with a substantial increase in radiation dose. Therefore, the combination of CTCA and strategies that do not use x-radiation, e.g. stress-echocardiography and stress-MRI, should be investigated further. As an interesting development in a more distant future, scanning the heart in one gantry rotation would create the possibility to perform myocardial perfusion imaging by CT. Likewise, with decreased acquisition time it would also become better feasible to perform delayed enhancement by CT. This technique enables the assessment of myocardial infarction size, and can be used for the evaluation of myocardial viability 46.

Imaging subclinical atherosclerosis

Almost 50% of non-fatal myocardial infarctions arises from plaques with less than 50% lumen narrowing 47. In addition, lipid-rich plaques give rise to a coronary thrombotic event more frequently than fibrous ones, which in turn cause less thrombotic events than calcified plaques 48. Plaque phenotyping may prove to be important in risk-stratification, and the guidance of medical therapy. CCA merely visualizes the vessel lumen. Intravascular ultrasound (IVUS) can be used for plaques phenotyping, yet its invasive nature precludes routine use. Therefore, a non-invasive modality for plaque characterisation is wanted. Coronary angiography by 1.5 T MRI is still of insufficient reproducibility due to the long acquisition time, and has a too low spatial resolution due to the high slice thickness. CTCA has been demonstrated to be able to discern between calcified and non-calcified plaques 49. However, to divide non-calcified plaques further into lipid and fibrous plaques is still difficult due to the limitations in spatial resolution. Further technological developments may lead to improved plaque phenotyping by CTCA.