Coronary artery disease (CAD) represents the major cause of morbidity and mortality in Western populations.1 The prime diagnostic tool that allowed the development of rational treatment techniques for this disease is invasive coronary angiography ((CA) an X-ray fluoroscopy guided procedure), which is associated with a low rate of life-threatening complications.2 More than 40% of the invasive CA studies are also carried out for the purpose of ruling out CAD.3 Non-invasive cardiac assessment has therefore been a goal of investigators for decades; however, the coronary arteries in particular prove a formidable challenge because of their small size (3-4mm), tortuous course, and susceptibility to cardiac and respiratory motion. In addition, the evaluation of cardiac function requires excellent temporal resolution on the order of less than 100ms. Echocardiography (ECG), nuclear medicine techniques, and MRI have been used non-invasively for a variety of cardiac indications, although no single technique provides a comprehensive assessment.
The prospect of imaging the heart and coronary arteries using computed tomography (CT) has been anticipated since the development of CT more than three decades ago. The lack of speed and poor temporal resolution of previous generations of CT scanners prevented meaningful evaluation of the coronary arteries and cardiac function. Most early assessments of the coronary arteries with CT were performed with electron beam computed tomography (EBCT), developed in the early 1980s.4 EBCT has been mostly used for the non-invasive evaluation of coronary artery calcium (CAC), but other applications, including assessment of coronary artery stenosis (CAS), have been reported in limited cases; however, EBCT is expensive and is not widely available.
Recent advances in CT technologies, especially multiple-row detector computed tomography (MDCT), have dramatically changed the approach to the non-invasive imaging of cardiac disease. With sub-millimeter spatial resolution (less than 0.75mm), improved temporal resolution (50-200ms), and ECG gating, the current generation of CT scanners (16-64-row detectors) makes imaging possible, and has the potential to accurately characterize the coronary tree. The purpose of this article is to provide a perspective on the developments in MDCT technology that have made non-invasive cardiac imaging with CT a reality, and compare it with conventional imaging techniques, such as CA performed under X-ray fluoroscopy guidance.
Technical Differences Between CA Performed Under X-ray Fluoroscopy and MDCT
Even though both techniques utilize X-ray radiation, there are potential risk differences (stochastic and non-stochastic risks) due to the nature of exposure.5,6 There are also fundamental technical differences between CA performed invasively with X-ray fluoroscopy (conventional method) and non-invasively with MDCT.
The conventional angiographic image represents an instantaneous, two-dimensional (2-D) planar projection of the 3-D contrast filled vessel lumen resulting in tissue superposition. During selective angiography, the operator obtains a 3-D understanding of the anatomy by repetitive injection and visualization of the artery of interest in different planes oblique to the body axis. Since the procedure is carried out under the guidance of continual or pulsed X-ray fluoroscopy, the spatial resolution (less than 0.2mm) and temporal resolution (less than 1s) are quite high and are considered the gold standard for comparing imaging capabilities of other modalities, such as CT. On the other hand, CT acquires multiple axial tomographic image slices, which are combined into 3-D volumetric data sets. Subsequent image reformation or reconstruction can provide 3-D or 4-D images for volumetric visualization. In addition, tomographic image acquisition during CT angiography (CTA) provides additional information about the arterial wall and structures surrounding the arteries, which are not part of the conventional angiographic image.
Technological Developments in MDCT
By late 1998, all major CT manufacturers launched MDCT scanners capable of providing at least four slices/sections per rotation with minimum gantry rotation times of 0.5s.7 This enabled volumetric data eight times faster than the earlier single-row detector CT to be obtained with a scan time of 1s.8 Irrespective of the number of detector rows in the longitudinal (Z-axis) direction, the number of slices obtained per CT gantry rotation depends on the number of data acquisition system (DAS) channels. The drive toward an increased number of thinner detector dimensions is mainly due to the demand for obtaining high spatial resolution in the longitudinal direction over a large scan volume, so as to obtain isotropic resolution in all three dimensions.9
Current MDCT scanners are capable of obtaining 16-64 slices per gantry rotation10 with slice thickness in the longitudinal direction as thin as 0.5mm. A number of novel image reconstruction algorithms are developed to handle the large volume data sets.11-13 The improved longitudinal (Z-axis) resolution, along with improved temporal resolution due to ECG gating, provides a scan technique considered well suited for CT imaging of the heart and other moving organs.14,15
The key issues for successful cardiac imaging are that the imaging modality should have the capability to provide high spatial and temporal resolution. It is therefore appropriate to examine the technological advances enabling MDCT to perform cardiac imaging.
Spatial
There are a number of factors that can influence the spatial resolution in a CT image. The transaxial (X-Y plane) resolution in CT has been quite high from the beginning, which is dependent on the image matrix and the field of view and is in the order of 0.5-0.25mm (one to two line-pairs/mm). The challenge concerns resolution in the longitudinal direction, which is influenced by the MDCT detector array design, slice thickness, reconstruction algorithms and increments, pitch, patient motion, and other technique factors.
Starting with four slices per rotation, the detector designs quickly migrated to 16 thin slices and have rapidly advanced to yield up to 64 thin slices (see evolution of detector array designs in Figure 1). The longitudinal resolution in modern MDCT scanners is in the order of 0.7-0.3mm (0.7 to 1.5 line pairs/mm), and is rapidly approaching the resolution achievable in transaxial direction. The technology is fast advancing with the goal of obtaining isotropic resolution. This is accompanied by faster scan times resulting in extended volume coverage making angiographic techniques feasible with MDCT scanners. Even though the spatial resolution of conventional cardiac catheter remains unchallenged, the advances in MDCT technology are impressive (see Figure 2).16,17
Tempora l Resolution
High temporal resolution is needed to minimize motion artifacts caused by cardiac pulsation. Since rapid movement is present during the systole phase, imaging is performed during the diastole phase. Desired temporal resolution for motion-free cardiac imaging in the diastole phase ranges from 150-250ms in order to image heart rates between 70 and 100 beats per minute (BPM), and less than 50ms to image during other phases. The temporal resolution in conventional X-ray fluoroscopy is high since the images are acquired with rapid exposure rates ranging from 7.5 to 30 frames per second.
The temporal resolution in MDCT depends on the gantry rotation times, type of ECG triggering, reconstruction methods, pitch, and other factors. Current MDCT scanners are capable of obtaining up to 64 slices per gantry rotation, and have a gantry speed as low as 330ms. One way to achieve high temporal resolution is by ECG triggering. Most cardiac CT procedures are performed with either prospective ECG triggering or retrospective ECG gating. Prospective ECG triggering has long been used in conjunction with EBCT and more recently with single-slice spiral CT.18 A prospective trigger signal is derived from the patientÔÇÖs ECG and the scan is started at a defined point in time, usually during diastole. MDCT allows the simultaneous acquisition of several slices within one heartbeat. The data is acquired from only part of the cardiac cycle and is the most dose-efficient way of ECG synchronization; however, the ECG-triggering technique greatly depends on a regular heart rate and is bound to result in misregistration and motion artifacts. On the other hand, retrospective ECG gating effectively overrides the limitations of prospective ECG triggering by acquiring data throughout the cardiac cycle and allowing image reconstruction on selected part of the cardiac cycle.19 Retrospective ECG gating creates image stacks reconstructed at exactly the same phase of the heart cycle. The downside to this method is the radiation exposure; only partial data is used in the image reconstruction, and the rest is discarded.
Temporal resolution in MDCT is further improved by the type of image reconstruction; namely partial or segmented reconstruction. During partial image reconstruction, the data is either acquired or used in only part of the gantry rotation (half plus fan angle) resulting in a temporal resolution of up to half the gantry rotation speed, i.e. as low as 200ms. On the other hand, even higher temporal resolution is achieved with multisegmented image reconstruction, where partial data from multiple heart cycles yields a temporal resolution of less than 100ms (see Figure 3).19,20 However, multi-segment reconstructed images are prone to reduced spatial resolution due to the variation in heart cycles. Although EBCT shows a favorable temporal resolution (up to 50ms), it is outperformed by MDCT due to their limitation in spatial resolution and poor contrast-to-noise ratio.
Pitch
The concept of pitch was introduced with the advent of spiral CT and is defined as the ratio of table increment per gantry rotation to X-ray beam width. Pitch values of less than one implies tissue overlapping and higher patient dose and pitch values greater than one imply extended imaging and reduced patient dose.8 However, in cardiac imaging the need for high spatial and temporal resolution demands that the pitch values need to be as low as 0.2-0.4, implying a tissue overlap of 50% to 75%, resulting in significant radiation exposure to patients.
Radiation Dose
One of the disadvantages of cardiac imaging with MDCT is its use of ionizing radiation. The radiation dose is highly dependent on the protocol used in cardiac CT.21 Among the most widely known protocols such as calcium scoring studies, the dose is relatively small (1-3milliSieverts (mSv)). However, for retrospective gating used for coronary stenosis assessment and CTA,much higher doses of 8mSv22 and greater than 20mSv, are reported. By comparison, the radiation dose of an uncomplicated CA is 4-6mSv. Across the board, the radiation doses are higher with MDCT compared with the doses delivered with EBCT and catheter CA (CCA) for similar procedures.22 Efforts to reduce the high doses accumulated in retrospective gating (dose modulation) are directed at reducing the tube current during parts of the cardiac cycle, particularly systole, where poorer image quality can be tolerated because the assessment of coronary arteries is sub-optimal. With the dose modulation technique, a 30% to 50% dose reduction can be achieved; however, it has to be evaluated per protocol. Any step to reduce radiation exposure should not jeopardize the image quality, as poor image quality may require repeat scans resulting in additional radiation doses to patients.
Future
It is certain that future technological advances will further enhance the role of MDCT in cardiac imaging. Manufacturers already have prototype scanners with 256-row detectors and flat panel technologies that can scan the entire heart in single CT gantry rotation. Improvements in data processing and reconstructing are forthcoming, and will further enhance MDCT capability, not just in imaging but also in quantifying cardiac functions. Research in the areas of multiple tubes to acquire cardiac data with superior temporal resolution is being worked on. In the near future, because of shorter examination times, improved spatial and temporal resolution, and cross-sectional imaging capabilities, CA procedures performed for diagnosis purposes in catheterization laboratories will be better suited for MDCT scanners.
Conclusion
Cardiac imaging with MDCT is evolving rapidly with technological advances. Widespread availability, shorter examination times of a non-invasive nature, and increasing numbers of studies demonstrating high sensitivity and specificity in diagnosing early onset of cardiovascular diseases (CVDs) will enable MDCT systems to play even greater role in diagnosis and follow-up treatment for CAD in the near future.
At the same time, since the number of cardiac CT examinations is increasing and the examinations involve substantial radiation doses, strong indications for the procedure and improved and standardized scanning protocols are essential for further advancement of MDCT for cardiac imaging.