Appropriate indications for computed tomography coronary angiography

Despite the various functional tests and biomarkers available for evaluation of patients with coronary artery disease (CAD), we sometimes look for the reassurance of anatomical information by way of a coronary angiogram. However, as an invasive modality, it does carry some risks, and the proportion of patients with normal coronary angiograms has remained relatively stable at 15%.1

In recent years, the rapid development of computed tomography (CT) technology, reduction in radiation levels and introduction of a Medicare Benefits Schedule item number in 2011 have fuelled the growth of multidetector computed tomography (MDCT) around Australia as a non-invasive test for CAD. Therefore, it is important to have a good understanding of the appropriate use and limitations of this new technology in order to prevent its misuse.

The most common CT scanners used for cardiac imaging today have 64 detectors, arranged in the cranial–caudal direction, covering a distance of about 4 cm with each heartbeat and having a spatial resolution of 0.3–0.6 mm. Respiratory motion is eliminated if the patient is able to hold his or her breath for about 10 seconds during the scan. Electrocardiogram (ECG) gating allows the scanner to obtain images during diastole when there is least motion of the coronaries. The computer then aligns the data from the different parts of the heart obtained during those five to seven heartbeats to present a three-dimensional volumetric dataset. Multiplanar reconstructions of the images allow the reporter to cut through this 3-D dataset in any plane to demonstrate the coronaries in different axes. Abnormal movement, breathing, ectopy or arrhythmia during the scan will cause misalignment of the images, resulting in step artefacts, which may hamper interpretation. The general prerequisites for patients undergoing computed tomography coronary angiography (CTCA) in order to achieve optimum image quality are set out in Box 1.

Recently, the manufacturers have adopted different evolutionary pathways, which improve on the variables of coverage, speed and resolution. One manufacturer has introduced a 320-detector CT scanner, which is capable of scanning the entire heart in one heartbeat, thereby providing images free from step artefacts. Another manufacturer has introduced its second-generation dual-source CT scanner, which has two sets of 128-detectors placed 90 degrees apart in the gantry so that it obtains images with only a quarter rotation. This has enabled good quality images at higher heart rates, as it can obtain images in half the time of other manufacturers’ scanners.2 A third manufacturer has introduced a 64-detector CT scanner with improved resolution of 0.23 mm, which enables better discrimination of fine objects like stents.3

The amount of radiation delivered to the patient depends on a number of factors, such as patient size, sex, distance covered and scanning protocols. The traditional method of scanning is called retrospective scanning, where radiation is delivered throughout the cardiac cycle. ECG-gated dose modulation is a setting that decreases radiation during systole, resulting in 25%–40% lower radiation for both men (8 mSv) and women (12 mSv).4,5

In patients who weigh < 85 kg or have a body mass index < 30 kg/m2, lowering the power setting of the scanner (from 120 kV to 100 kV) reduces radiation by up to 60% while maintaining diagnostic quality.6 However, it has not been employed by some diagnostic facilities as they were unaware of this fact.4,6

A recent breakthrough is the prospective scanning technique, which delivers radiation only during a very short period in diastole. The radiation reduction is up to 80%, with doses of 2–5 mSv, which is lower than typical invasive coronary angiograms and nuclear stress scans (Box 2).4,7 However, patients must have stable, low heart rates (< 60 beats per minute) without ectopy or heart rate variability, as there is little margin for error. Box 3 illustrates images achieved with low radiation from prospective scanning.

Meta-analyses of over 45 single-centre studies have consistently shown CTCA to have excellent sensitivity (98%) and very good specificity (88%) compared with invasive coronary angiography for significant disease (stenosis > 50%).5,8,9 The negative predictive values (96%–100%) were better than positive predictive values (93%), demonstrating CTCA to be an excellent tool for ruling out significant disease in patients with low-to-intermediate pretest probability of CAD. Similar results were found in prospective multicentre and multivendor validation trials of CTCA.10-12

The prognostic value of non-obstructive disease on CTCA has been investigated. One study involving 1256 patients with up to 2 years of follow-up found that, of 802 patients with mild disease on CTCA, only one patient (0.12%) had a severe cardiac event in the form of unstable angina.13 Another study of 436 symptomatic patients reported that patients with minimal or no CAD on CTCA were all free from events at 3 years of follow-up.14

Although CTCA can reliably exclude obstructive disease based on excellent negative predictive values, its ability to quantify stenosis severity is not as robust. Studies comparing CTCA to quantitative coronary angiography and intravascular ultrasound found good correlations but large standard deviations (up to ± 25%).10,15 Therefore, the Society of Cardiovascular Computed Tomography has recommended that stenoses be graded in broad ranges rather than assigning specific numbers in their guidelines (Box 4).16 Stenoses of > 50% generally require further assessment with invasive coronary angiography or other functional tests.

In 2011, the CSANZ published comprehensive guidelines on non-invasive coronary artery imaging.17 These are similar to those of the American multisociety18 and European Society of Cardiology19 guidelines. The appropriate indications for performing CTCA are outlined in Box 5. The discussion below includes some of the more common scenarios.

It is now possible to exclude severe coronary artery stenosis non-invasively by CTCA. Current evidence supports its use in symptomatic individuals with select indications. There exists the potential for misuse with this emerging modality, and consideration should be given to other options in light of local resources and expertise (Box 6). The rapid development in technology and further research will clarify and expand the role of cardiac CT in the future.

3 Prospective computed tomography scan images

A: Volume-rendered tomogram of Patient 1, showing location of previous bypass grafts (radiation = 4.2 mSv; 120 kV). B: Computed tomography coronary angiogram of Patient 2 with chest pain, showing moderate LAD stenosis caused by mixed plaque consisting of non-calcified (white arrow) and calcified (black arrow) plaque (radiation = 1.1 mSv; 100 kV, no padding). C: Patient 2, demonstrating anomalous LCx originating from RCA that has mild mixed plaque. LCx takes a “non-malignant” route between the aorta and left atrium, which is unlikely to experience compression. D: Corresponding invasive angiogram of Patient 2, demonstrating anomalous LCx. Ao = aorta. LA = left atrium. LAD = left anterior descending artery. LCx = left circumflex artery. LIMA = left internal mammary artery. RCA = right coronary artery. RV = right ventricle. SVG = saphenous vein graft. Images from Cardiac CT Service, Royal Adelaide Hospital, using a 128-detector computed tomography scanner (SOMATOM Definition AS+, Siemens, Erlangen, Germany).