Non invasive coronary imaging with computed tomography

The technological advances made by computed tomography (CT) since its initial clinical application in the 1970s have been staggering.

The development of slip ring technology in the late 1980s allowed the continuous acquisition of data.
However it was not until the mid 1990s that sub-second gantry rotation became available. This allowed high quality examinations of the thorax or abdomen to be obtained in a single breath hold. Major advances were required however to usefully image small fast moving structures.
Initially electron beam CT showed promise. While this offered excellent temporal resolution (approximately 50 – 100 milliseconds (ms)), the technology was limited by spatial resolution (1 – 2 mm), high cost and poor coverage of large imaging volumes. Overall it had limited attractiveness to radiology departments and was never likely to become widely available. 
Multislice CT (MSCT) has begun to provide the answers. Current technology offers a gantry rotation speed of 0.33 seconds (giving a true temporal resolution of 165 ms – or even 83 ms with dual source CT), sub-millimeter collimation allowing 0.33mm isotropic voxels (essential for optimal 3 dimensional reconstructions), all acquired in less than a ten second breath hold. Under optimal circumstances MSCT can now rival invasive coronary angiography.

Since MSCT can image the whole body in a breath hold at a reasonable cost it has found wide acceptance and therefore wide availability. This has revolutionised the availability of MSCT for cardiac patients.

Calcium Scoring

Recently published guidelines suggest a role for calcium scoring in patients with intermediate risk profiles for sudden cardiac death. In this patient cohort, an increased calcium score adds to the risk profile and can trigger preventive treatment [1]. The routine assessment of the coronary calcium score is still under debate [2]. While additional information, independent of the conventional risk scores can be obtained, population screening would prove expensive, and comes with the additional risk of radiation exposure. It remains to be seen, whether larger population data will further support a more liberal use for calcium scoring.

Measurement of coronary calcification may have a role as a gate keeper to CT coronary angiography (CTCA). This is potentially important as significant calcification produces beam hardening artefact and impairs the ability to accurately diagnose and assess stenoses. Heuschmid et al [3] showed (in 37 patients using 16 slice CTCA) that if analysis was limited to patients with an Agatson score equivalent of = 1000, sensitivity for significant stenoses was improved from 59 to 93%, specificity 87 to 94%, positive predictive value 61 to 68% and negative predictive value from 87 to 99%.

The data would suggest that an initial calcium score should be used as a quality control measure to limit CTCA to patients who are likely to have a high quality and reliable diagnostic study.

Coronary Angiography

Numerous studies have evaluated the use of 4, 16 and 64 slice CTCA for the detection of coronary artery stenoses [e.g. 4-8].  CTCA overestimates coronary stenoses compared to invasive coronary angiography (ICA) and has a positive predictive value of about 80% and negative predictive value of about 95%. As many studies included patients with a high pretest probability of significant coronary artery disease, the results in low risk groups may be even better.
 
Lim et al [9] showed that with a 40 slice machine, CTCA had a sensitivity, specificity and negative predictive value of 98 – 99 % for showing the 94 significant stenoses in 480 segments detected on ICA. Ferencik et al [10] produced similar results with a 64 slice MDCT while Achenbach et al [11] showed that, in selected patients, 98% of coronary artery segments could be visualised free of motion artefact using a dual source CT.

For bypass grafts, CTCA offers high diagnostic value. Several studies have shown a 100% sensitivity for the detection of graft occlusion and high sensitivities and specificities for significant graft stenoses [12-15]. Pache et al [15], including patients with atrial fibrillation and uncontrolled heart rates showed that with CTCA all grafts and 93/96 of distal anastomoses could be visualised. The three non-visualised anastomoses were due to metal-clip artefact. Three grafts were missed by ICA, but not by CTCA. Overall sensitivity for detecting significant stenoses for CTCA was 97.8% with a specificity of 89.3%.

For the evaluation of coronary stents data is generally limited. Cademartiri et al [16] showed that in 74 stents, 67/68 were correctly identified as having no significant re-stenosis (specificity 98.5%). The two errors were in 2 mm vessels. 5/6 significant stenoses were detected by 16 slice MDCT (sensitivity 83% but the confidence interval was wide).

It is recommended that before commencing a CTCA service that the referrer and provider consider in some detail exactly what they wish to achieve and the quality of the available equipment. All machines are not equal. In general a 64 slice CT is preferred though good results can be obtained in certain patient groups with 16 slice CT. A high quality software platform is essential to avoid excessively time consuming analysis.

Who Should Have CT Coronary Angiography?

Before contemplating CTCA the cardiologist should consider the pretest probability of ischaemic heart disease and the availability and efficacy of alternative strategies that could be used to evaluate any given patient.
We should keep in mind certain principles:

1. CTCA should reduce the number of diagnostic ICAs that do not result in intervention.
2. The public should not be exposed to unnecessary ionisation radiation. (In Europe this principle is reinforced by law). CTCA can therefore rarely be justified in the asymptomatic and should not be offered as a screening examination.
3. CTCA should not delay access to revascularisation.
4. CTCA should not add to costs.

In general

• There is no point in performing CTCA in patients with a high probability of new or recurrent myocardial ischaemia (for example men greater than 40 years of age, and women greater than 60 years of age with typical angina) as most will have significant coronary artery disease and require intervention.
• Similarly patients with a very low probability of myocardial ischaemia may in many instances be safely reassured that they have a low cardiac risk without CTCA.
• Patients unlikely to have a diagnostic study should usually not have CTCA. This would include patients with large amounts of coronary calcium (frequently elderly males and determined by a coronary calcium score prior to CTCA) and those with atrial fibrillation and an uncontrolled ventricular response, or multiple ventricular ectopics. The patient must be able lie still, hold their breath for 10 – 20 seconds and to raise both hands (actively or passively) above their head. Very obese patients (body mass index > 40 kg/m2) are often unsuitable as the machine may not be able to generate  sufficient electric current to produce a diagnostic study.
• Patients who are diabetic with impaired renal function may be more safely managed with an alternative strategy (e.g. perfusion imaging).
• Younger patients (for example < 40-50 years of age) who have a significant lifetime risk of fatal cancer induction by the ionising radiation. These also may be equally well managed with an alternative strategy (e.g. magnetic resonance perfusion imaging). This is controversial – the cardiologist and radiologist should understand the radiation dose [17] (and its risk) their equipment generates. Recent work has suggested that dual source CT may offer CTCA at half the radiation dose of single source CT [18]
• 16 slice CT is inadequate for detailed assessment of coronary stents =3mm in diameter [19]. 64 CTCA currently does not appear reliable in detecting instent restenosis < 50% [20].

While this may seem disappointing it is imperative that CTCA is “rolled out” in a sensible fashion. There do remain numerous potential applications. Indeed the ACC/AHA guidelines (21) identify 11 appropriate indications for cardiac CT. These guidelines are reasonable and provide useful guidance. Relating to coronary heart disease a simplified appropriate referral strategy for CTCA would include:

1. Intermediate pretest probability of coronary artery disease (CAD) in chronic chest pain syndrome when the unable to interpret the ECG for ischaemia or the patient cannot exercise.
2. Intermediate pretest probability of CAD in acute chest pain with no ECG or biochemical abnormalities.
3. Suspected anomalous coronary arteries.
4. Uncertain stress test result.
5. Dilated cardiomyopathy if thought unlikely to be ischaemic in origin.

Into the Future

Current best CTCA has a temporal resolution of 83 ms and 0.35 mm near isotropic voxel scanning giving an in plane spatial resolution of up to 20 line pairs per cm, and a z axis resolution of about 12 line pairs per cm. Contrast resolution is excellent. As all relevant examinations can now, with a 64 slice CT, be achieved in a comfortable breath hold of less than 10 seconds a simple increase in the number of acquired slices will not provide much clinical benefit in terms of the breath hold time.

The trend in the future will rather be to concentrate on improved resolution. Increasing gantry rotation speed further (to improve temporal resolution) is difficult (unless there is a dramatic reduction in scanner weight). Multisource scanners will have a role to play in this. Increasing the number of detectors does not on its own improve spatial resolution – to do this narrower collimation will be needed. As the radiation dose of CTCA is already around double that of ICA, detector design needs further improvement before thinner collimation (and hence better spatial resolution) can be offered. Otherwise the CTCA radiation dose (or noise level) will rise uncomfortably high. 256 detector row CT is being developed, but as yet does not offer thinner collimation. Also because of its increased mass, gantry rotation speeds are likely to be inferior to 64 slice CT [22] until there is another engineering leap forward. Flat panel detectors are under development which produce a true isotropic 0.25 mm voxel and in plane and out of plane spatial resolution of about 22 line pairs per cm. However currently temporal resolution is inadequate for cardiac imaging. Further great engineering improvements of the CT era will hopefully be realised.

The content of this article reflects the personal opinion of the author/s and is not necessarily the official position of the European Society of Cardiology.

Conclusion:

CTCA is a valuable clinical tool that has the potential to improve the investigation of cardiac patients. However if used injudiciously it is unlikely to be helpful and could be harmful. CTCA will undoubtedly become both more prevalent and more useful as access and technology improves.