Coronary artery disease (CAD) represents a relevant economic burden to modern medicine. In 2003 more than 1.41 million diagnostic coronary angiograms and an additional 1.24 million percutaneous coronary angioplasties were performed in the US alone.1 While coronary angiography is considered the gold standard for diagnosing CAD, several non-invasive techniques have been evaluated for assessing CAD. Over the last decade, multi-slice spiral computed tomography (MSCT) has emerged as a serious alternative for non-invasive cardiac imaging. The rapid technical development from 4-slice technique to dual-source computed tomography (DSCT) fuelled the quest to establish computed tomography (CT) as a routine method for coronary imaging. While 4-slice CT already allowed acceptable visualisation of the proximal coronary segments,2 16-slice CT permitted reliable exclusion of relevant coronary stenosis (>50%) with a negative predictive value (NPV) of 99%.3 The introduction of 64-slice CT with further improved spatial and temporal resolution was the breakthrough for routine implementation of coronary CT-angiography.4 Currently, DSCT is state-of-the-art in cardiac CT imaging. This new generation of CT scanners with two X-ray sources provides a temporal resolution as low as 42msec at submillimetre spatial resolution.5 Thus, DSCT outperforms the spatial and temporal resolution of electron beam CT. Within the first year after its clinical introduction, contrast-enhanced DSCT became a key technology for non-invasive imaging of CAD.
Technical Basics of DSCT
DSCT utilises two tube-detector systems that are mounted at orthogonal orientations in the gantry. The 180° data in parallel geometry are split into two sinograms, which are simultaneously acquired in the same phase of the cardiac cycle at the same anatomical level with a 90° angle between the detectors. The projections for reconstruction of one image slab can thereby be obtained in half the time needed by a single source CT. This results in a constant temporal resolution of 83msec when using data from a single RR-interval only. Avoiding the problems of multi-segment image reconstruction algorithms, the constant high temporal resolution permits an efficient pitch adaptation, which allows a heart rate-dependent dose reduction. Detector design and spatial resolution are identical to the geometry of a 64-slice CT scanner.5
Non-invasive coronary angiography is the most common indication for cardiac DSCT. Since 1998 multiple studies have demonstrated the value of contrast-enhanced MSCT-angiography for the assessment of coronary artery stenosis. Recent studies on 64-slice CT reported sensitivities and specificities in the range of 90–98% for the detection of coronary artery stenoses.4–7 In all of these studies the positive predictive value (PPV) for the detection of relevant coronary artery stenoses was lower than the NPV, which is typically about 98%. Nevertheless, single-source CT is hampered by the limited, heart rate-dependent temporal resolution. Consequently, rigorous patient selection is recommended and even with 64-slice CT, approximately 10% of coronary segments have to be excluded from analysis.
With the introduction of DSCT, the robustness of coronary CT-angiography improved to a degree that allows even patients with heart rates ≥80bpm to be examined at diagnostic image quality (see Figure 1). Several phantom and patient studies confirmed these improvements.8–10 A first study comparing DSCT coronary angiography with invasive coronary angiograms for the detection of coronary artery stenosis in a high pre-test likelihood population confirmed these initial observations. With a sensitivity of 96.4% and a specificity of 97.5%, a NPV of 99.4% for the exclusion of coronary artery stenosis was achieved. Only 1.4% (6/420) of the coronary artery segments had to be excluded from analysis.11 These results are remarkable as the first-time patients with irregular or elevated heart rates were not excluded from the study and no beta-blockers were administered.
So far there are no data on the DSCT assessment of coronary artery bypass grafts (CABGs) or coronary artery stents. However, for both indications only minor improvements are to be expected. While single-source CT already provides excellent results for the assessment of CABG, coronary stent imaging may benefit from the improved temporal resolution. Spatial resolution, which is the key to assessment of the stent lumen, equals that of 64-slice CT with 0.4x0.4x0.4mm3. Thus, the potential for improvement is limited.
Left ventricular (LV) function predicts patient outcome in a variety of conditions, including CAD. While LV function is normally assessed by echocardiography, functional information is also available with retrospectively electrocardiogram (ECG) gated CT of the coronary arteries. In several studies, single source CT has been successfully validated for the quantification of ventricular function.12 A common limitation of functional analysis from CT-angiography data is the administration of beta-blockers to reduce heart rate. Their negative chronotropic and inotropic effects limit the clinical value of functional analysis. With DSCT this limitation can be overcome. An early phantom study indicated the reliability of functional DSCT at heart rates up to 140bpm.13 The reliability of functional data will become independent from the patient’s heart rate.
The extent and degree of myocardial injury determine the patient’s outcome after myocardial infarction (MI).14 Consequently, various methods have been developed for the CT assessment of myocardial viability and perfusion. Areas of MI present as regions of decreased contrast enhancement on contrast-enhanced DSCT angiography. These areas can be detected with 90% accuracy.15 However, this technique does not allow differentiation of MI from hypoperfused but viable myocardium. Therefore, assessment of myocardial viability by myocardial late enhancement is helpful.16 Finally, the assessment of wall thickness helps to differentiate acute MI from chronic MI with left ventricular wall thinning.
Contrast-enhanced cardiac MSCT is also suitable for assessing aortic valve morphology and function. Further indications include the detection of cardiac tumours, congenital anomalies or pericardial disease.
Contrast Material Administration
To successfully perform the above-mentioned examinations, dedicated contrast injection protocols are needed. From a theoretical point of view, the optimal bolus geometry for CT angiography is an immediate increase in enhancement in the target vessel lumen followed by a steady state in which the attenuation does not alter until CT data acquisition is finished. In coronary CT angiography, intravascular attenuation should be between 250 and 400 Hounsfield units (HU).
Lower attenuation values will impair discrimination of plaque components from the vessel lumen, whereas higher attenuation values may obscure coronary calcifications.17 Intracoronary attenuation also affects quantitative assessment of different plaque components due to beam-hardening artefacts.18 For assessment of the ventricular cavities, a certain amount of attenuation in the right heart chambers needs to be preserved (see Figure 2). This may cause problems, as most contrast injection protocols are solely optimised for the coronary arteries.
Several factors are known to influence contrast enhancement after intravenous bolus injection: patient demographics, haemodynamic status, amount of contrast material, injection rate, use of a saline chaser and the site of injection.19 Some authors also state the impact of the iodine concentration on arterial enhancement.20 In the experience of the authors, this does not matter as long as the iodine delivery rate (IDR) and the total amount of iodine are kept constant. In the patient population undergoing cardiac CT, the haemodynamic status is of particular interest as a relevant number of patients suffer from impaired cardiac function, which is associated with increased intravascular attenuation.21 Furthermore, with the reduced scan times of DSCTs there is a need for more compact bolus timing. These requirements emphasise the importance of an elaborate individually adapted contrast administration protocol.
In the early days of cardiac CT a fixed delay technique was used for synchronising contrast administration and data acquisition. This technique was soon abandoned as there is a wide variability in the haemodynamic states, resulting in irreproducible enhancement patterns. There are currently two competing techniques for contrast timing: test-bolus injection and realtime bolus tracking. At the moment the ‘user-friendly’ bolus tracking technique is most commonly used. Moreover, this approach was shown to be superior to test-bolus injection by providing more homogeneous contrast enhancement in the coronary arteries.22 Advantages of a test-bolus injection include prospective planning of the time–attenuation curve and estimation of the cardiac output.23,24
Multi-phasic contrast injections are best suited to achieving optimal bolus geometry.25 For practical purposes a biphasic injection is an acceptable compromise. This includes an IDR of 1.8–2.0gI/sec for the first five seconds of injection that is continued at approximately 80%. IDR is the method of choice for describing contrast injection as both low-concentration contrast material injected at a high flow rate and high-concentration contrast material injected at a low flow rate achieve comparable intra-coronary attenuation.
Besides the IDR, the duration of contrast material injection is crucial. As a role of thumb the duration of contrast injection should equal scan time plus 10%. Therefore, a typical contrast injection protocol for a 12-second cardiac DSCT scan might be as follows: 33ml Iopromide 300 (Ultravist 300, Bayer Schering Pharma) injected at a flow rate of 6.6ml/sec followed by 43ml Iopromide 300 at 5.3ml/sec. This corresponds to an IDR of approximately 2gI/sec for five seconds followed by 1.6gI/sec for another eight seconds, amounting to an injection duration of 13 seconds (12 seconds + 10%). Comparable vascular attenuation values could be achieved using a higher concentration contrast medium if the injection parameters were adapted as follows: 27ml Iopromide 370 (Ultravist 370, Bayer Schering Pharma) injected at a flow rate of 5.4ml/sec followed by 35ml Iopromide 370 at 4.3ml/sec. Currently there is no consensus on a standardised injection protocol. Only recently, multi-phasic contrast injections with diluted contrast material were assessed.26 Dedicated examinations such as the assessment of myocardial viability or perfusion require individually adapted injection delays and different amounts of contrast material.
Finally, the use of a saline chaser must not be forgotten. Saline chasing minimises streak artifacts in the superior vena cava and helps to reduce the amount of contrast material. For cardiac DSCT a 50ml saline chaser injected at the same flow rate as the contrast injection is recommended.
Contrast-enhanced cardiac DSCT is developing at breathtaking speed. While its advantages in coronary CT-angiography are obvious, further improvements are on their way. Owing to its improved temporal resolution, DSCT will facilitate visualisation of the cardiac venous system. Only a few groups have investigated the cardiac veins using MSCT,27,28 but there is a steadily growing interest in venous anatomy due to the increasing number of interventions involving the cardiac veins.
There are also some limitations that have to be carefully considered: the most important are radiation exposure and limited spatial resolution. These problems will not be solved by more detector rows or an increased number of X-ray sources. Spatial resolution as a function of dose will become a border that cannot easily be crossed in clinical routine. The spatial resolution of catheter angiograms will not be reached within the next few years.
In conclusion, contrast-enhanced DSCT permits a comprehensive imaging assessment of the heart. Optimised contrast injection protocols taking the IDR into account are indispensable to achieve high-quality cardiac DSCT examinations. Dedicated contrast injection protocols allow reliable assessment of ventricular function, cardiac veins and myocardial viability.