Cardiac investigation—nuclear and other imaging techniques.
- Nuclear imaging
- Cardiac CT
- Further reading
Myocardial perfusion scintigraphy
Three radioisotopic tracers are routinely used in single photon emission computed tomography (SPECT) imaging: thallium-201 and technetium-99 m (bound to either sestamibi or tetrofosmin). Imaging can be performed at rest or with stress (exercise or pharmacological), comparison allowing determination of whether regional perfusion is normal, or there is ischaemia, or there is infarction/scar.
Myocardial perfusion imaging is minimally invasive, and—in contrast to other methods of investigation—is not limited by exercise capacity, airways disease, abnormalities of the resting ECG, pacemakers, or acoustic windows.
In the investigation of the patient with possible coronary artery disease, a normal SPECT study is very reassuring, predicting a very low chance of a major cardiac endpoint event in the following few years (<1% per year). High-risk markers on SPECT provide additional prognostic value to clinical and electrocardiographic variables, and decisions about revascularization can be usefully informed by SPECT imaging.
ECG-gated SPECT allows images to be taken throughout the cardiac cycle, when comparison of end-systolic and end-diastolic images then allows volumetric analysis and calculation of left ventricular ejection fraction.
Positron emission tomography (PET)
Perfusion can be assessed with nitrogen-13 ammonia or rubidium-82, and metabolism with fluorine-18 fluorodeoxyglucose (FDG). Cardiac PET studies tend to be confined to research institutions, with the metabolic tracer FDG considered to be the ‘gold standard’ for assessment of myocardial viability.
Cardiac MRI can reveal images of spectacular similarity to anatomical cross-sections and is the best method available for quantifying ventricular volumes, ejection fraction, myocardial mass, and differentiating viable (preserved myocytes) from nonviable (fibrotic) myocardium (although echocardiography—which is cheaper and more readily available—remains the first choice in routine clinical practice for many of these indications).
Cardiac MRI is also very useful in assessing patients with congenital heart disease and is particularly indicated for those with complex conditions or in whom it is difficult to obtain good echocardiographic pictures.
Multislice spiral computed tomography (MSCT) is indicated to assess pericardial thickening/calcification and is a fast and noninvasive method for the visualization of the coronary arteries. It can also be used to quantify the amount of coronary and aortic valve calcium.
Cardiac CT does not yet match invasive coronary angiography, but many studies have shown a very high negative predictive value, hence cardiac CT appears to be a reasonable test to rule out coronary stenoses in patients with low-to-intermediate likelihood of disease. However, with further developments it is likely that coronary CT will replace invasive coronary angiography for diagnostic purposes.
Within cardiovascular medicine nuclear imaging is an important technique with the following capacity:
- ◆ Identification of ischaemic heart disease, with ability to make a prognostic assessment
- ◆ Identification of the quantity of viable myocardium in a patient with heart failure
- ◆ Assessment of regional and global myocardial function
- ◆ Provision of insights into molecular processes via targeted imaging
The procedure is versatile and minimally invasive, and is not limited by exercise capacity, airways disease, abnormalities of the resting ECG, pacemakers, or acoustic windows. Indeed, it is very difficult to identify any patient who is not suitable for nuclear perfusion imaging, and as a result the technique has matured into an almost comprehensive procedure for assessment of coronary artery disease. Over 5 million nuclear cardiology procedures were undertaken in the USA in 2001.
Myocardial perfusion scintigraphy
An intravenous injection of a radiopharmaceutical tracer is administered, which is taken up by intact myocardial cells, the cellular uptake being dependent on the myocardial blood flow at the time of the injection. Rest and peak stress images are required to assess for reversible ischaemia, and comparison of these images determines whether regional perfusion is normal (no hypoperfusion), or due to infarction/scar (hypoperfusion at rest and stress) or ischaemia (hypoperfusion at stress only).
There are currently three radioisotopic tracers used in single photon emission computed tomography (SPECT) imaging. The oldest is thallium-201, which is an analogue of the potassium ion; it has a half-life of 73 h and emits photons of varying energies (predominantly 68–80 keV). Uptake is dependent upon an intact Na+,K+-ATPase membrane pump. A dose of 80 Mbq is injected at peak stress (exercise or vasodilator), with imaging after 5 to 10 min. Myocardial perfusion defects fill with tracer (by redistribution) 3 to 4 h after initial injection, although areas subtended by severely stenotic coronary arteries may not redistribute for many hours and delayed imaging up to 24 h later may be required. Reinjection of thallium-201 at rest is sometimes added to protocols after redistribution imaging to improve the overall accuracy.
The other two tracers contain technetium-99m, which emits γ-rays at 140 keV and has a half-life of 6 h, bound to either sestamibi or tetrofosmin before intravenous injection. Both tracers enter viable myocardial cells and are fixed—there is no/minimal redistribution—hence separate injections (250–1000 MBq) are required for rest and peak stress imaging. Imaging can then occur at a more convenient time after stress, e.g. 45 to 90 min, reducing motion artefact from overbreathing.
Photons emitted from the patient are identified by a gamma camera and reconstructed to form an image on a computer workstation. The gamma camera rotates around the patient in a 180° arc from right anterior oblique (RAO) to left posterior oblique (LPO). Acquisition usually takes less than 15 min, newer solid state cameras promise two minute acquisition times. The images are ECG gated to allow functional as well as perfusion analysis and are reconstructed and displayed in the following three planes to allow analysis of each region in more than one view—short axis (SA), horizontal long axis (HLA), and vertical long axis (VLA).
Principles of stress testing
The varied stress modalities available to nuclear cardiology are one of its major advantages. Exercise (or physiological) stress can be achieved with a treadmill or bicycle following a specified protocol, e.g., Bruce protocol. The preferred method, which mimics ‘real world’ stress and also provides valuable ECG data, is for the isotope to be injected at peak stress and the patient asked to maintain exercise for a further minute to allow circulation of the radiotracer.
Patients unable to exercise can undergo pharmacological stress instead. Vasodilators such as adenosine or dipyridamole can be injected intravenously to induce hyperaemia and provide flow heterogeneity between coronary vascular beds. These vasodilators are contraindicated in patients with significant airways disease, significant atrioventricular node disease, hypotension, methylxanthine exposure in last 48 h, and acute myocardial infarction (<2 days). Selective adenosine A2A receptor agonists have fewer serious side effects. If a patient cannot undergo vasodilator stress, then inotropic stress with escalating doses of dobutamine (± atropine) can be used. GTN-enhanced rest technetium-99m studies have also been shown to demonstrate viable myocardium. It is very unlikely that any patient cannot undergo some form of stress imaging with nuclear cardiology techniques.
Some practical considerations
The overall radiation exposure of a patient undergoing a stress-rest technetium study is approximately 10 mSv, which is greater than a diagnostic coronary angiogram but without the invasive and vascular complications.
Cost-effectiveness analyses have been performed with SPECT. Studies in Europe and the United States of America have confirmed that utilizing a strategy involving these techniques is associated with cost savings as well as fewer invasive tests. This has helped to drive a significant increase in the number of SPECT procedures performed worldwide.
Clinical uses of nuclear imaging
Investigation of coronary artery disease
In a large meta-analysis of 33 studies the sensitivity and specificity of myocardial perfusion imaging were 87% and 73% respectively. The normalcy rate, which removes the referral bias of false-positive patients being referred on for coronary angiography, was 91%. Similar results are available for vasodilator and dobutamine stress. More importantly, there is a wealth of prognostic data available. The value of a normal SPECT study is beyond doubt, with a meta-analysis including just under 21 000 patients followed up for 2.3 years demonstrating a major cardiac endpoint event rate of 0.7% per year. Follow-up studies extending up to 7 years have demonstrated similar low event rates.
High-risk markers on SPECT have incremental prognostic value over electrocardiographic and clinical variables. These include multivessel disease patterns, large burdens of ischaemia (>10% of myocardium), transient ischaemic left ventricular dilatation, left ventricular ejection fraction (LVEF) <0.4 (see below) and lung uptake (only with thallium-201). SPECT is also able to further risk stratify when risk scores such as the Duke treadmill score are applied to exercise ECG variables, and can provides additional prognostic value in specific populations such as patients after myocardial infarction or with diabetes mellitus, women and patients with an abnormal ECG. e.g. left bundle branch block.
Nuclear techniques are well suited to the identification of myocardial viability and predict functional recovery (identified by echocardiography) in approximately 80% of segments after revascularization. This means that decisions about revascularization can be usefully informed by SPECT: studies have clearly demonstrated that patients with low ischaemic burdens on SPECT have the same cardiac event rate if treated with either medical therapy or revascularization, and that those with significantly abnormal scans have lower event rates with revascularization compared to medical therapy. Comparative studies with low dose dobutamine echocardiography, positron emission tomography (PET), and cardiovascular magnetic resonance (CMR) have been performed. Each test is broadly similar in its ability to predict functional recovery. SPECT has also been used to assess success of revascularization procedures.
In the acute setting, resting SPECT injections have been performed in patients attending Emergency Departments with chest pain and a nondiagnostic initial ECG. A normal perfusion scan was associated with a lower risk of future events, lower likelihood of requiring cardiac catheterization, and lower costs owing to the shorter hospital stay and fewer subsequent investigations.
Assessment of left ventricular volume and function
ECG-gated SPECT allows images to be taken throughout the cardiac cycle, when comparison of end-systolic and end-diastolic images then allows volumetric analysis and calculation of LVEF. There are three methods for calculation of volumes, regional function, and ejection fraction with nuclear techniques—first pass radionuclide ventriculography, equilibrium radionuclide ventriculography, and gated SPECT.
First pass radionuclide ventriculography
This relies on an intravenous injection of technetium-99 m DTPA or pertechnetate while the patient is already lying under the gamma camera. The circulation of isotope is studied in the right- and then left-sided cardiac chambers. The radioactivity is proportional to the volume of blood in the chamber. The technetium is not bound to sestamibi or tetrofosmin and therefore no myocardial uptake occurs. Right-to-left shunts can be detected by this method.
Equilibrium radionuclide ventriculography
This technique, also known as multigated acquisition (MUGA), relies on technetium-99 m pertechnetate radiolabelled erythrocytes, which are prepared in vivo to accept the technetium with a preceding injection of stannous pyrophosphate. The radiolabelled erythrocytes are allowed to circulate within the circulating volume with acquisition after 10 min. All four cardiac chambers are identified at the same time and so accurate camera positioning is required to reduce overlap. Images are gated to the R wave and many hundreds of cycles are acquired to produce an average. Regional motion can be identified on end systolic and diastolic frames. Left ventricular and right ventricular ejection fractions are also calculated. Acquisition time for both techniques is short (typically 10–15 min).
This relies on endomyocardial border definition from techneiutm-99 m sestamibi or tetrofosmin to produce end-systolic and diastolic frames. Regional analysis, volumes, and LVEF are calculated.
The first two techniques are well established and validated and until recently were considered the ‘gold standard’ for volumetric analysis. Gated SPECT provides accurate assessment without requiring a blood pool injection. Alternatives such as echocardiography and X-ray ventriculography are not as reproducible, although cardiac MRI is now considered the gold standard for volumetric analysis.
The addition of volumes and LVEF to SPECT increases the prognostic value of myocardial perfusion imaging. Changes in post-stress LVEF are likely to represent sub-endocardial ischaemia, which may help to assess patients with ‘balanced’ multivessel ischaemia who may not have obvious regional perfusion defects.
Positron emission tomography (PET)
PET relies on coincidence detection of 512-keV photons. Perfusion can be assessed with nitrogen-13 ammonia or rubidium-82. Metabolism is assessed with fluorine-18 fluorodeoxyglucose (FDG). Nitrogen-13 ammonia and FDG have a short half-life and need to be produced in a cyclotron, which restricts their use. Nevertheless these tracers allow quantitation of myocardial flow, and the metabolic tracer (FDG) is considered to be the gold standard for assessment of myocardial viability. Cardiac PET studies tend to be confined to research institutions, but the increase in oncological studies requiring combined PET/CT machines may increase the availability of this technique for cardiac studies.
Comparison with other techniques
For myocardial perfusion the alternatives to PET include exercise electrocardiography, stress (exercise or dobutamine) echocardiography, and CMR (pharmacological stress). The exercise electrocardiogram is inferior mainly due to its dependence on exercise ability and a normal resting ECG. Stress echocardiography is a rapidly improving technique with a slightly lower sensitivity but a superior specificity in comparative studies: it is physician intensive and operator dependent, but harmonic imaging and microbubble contrast agents have made an enormous difference to the technique so that it is comparable to PET in expert hands. Cardiac MRI can assess regional and global LV wall motion during dobutamine infusion, similar to stress echocardiography. Another MRI technique looks at myocardial perfusion using gadolinium as a contrast agent (although recent experience suggests that gadolinium is not entirely benign, see below). So far the different methods for assessment of myocardial ischemia are regarded equivalent in clinical practice, but local expertise may vary. Multislice CT (MSCT) may provide ischaemia/infarct imaging, but this is still at an experimental phase. As mentioned earlier, CMR is the gold standard for volumetric and functional analysis, although echocardiography is more versatile and accessible. Functional imaging using nuclear techniques, stress echocardiography or MRI is recommended in the latest NICE guidelines for assessment of patients with chest pain.
For cardiac investigations standard MRI scanners have to be upgraded with a cardiac program (ECG gating, etc.) to allow for the movement of the heart, but can then reveal images of spectacular similarity to anatomical cross-sections. MRI also has the advantage over other techniques described in this article of not exposing the patient to radiation, although recently nephrogenic systemic fibrosis has been reported after application of the MRI contrast agent gadolinium. Those with impaired renal function are most susceptible to this complication, hence a patient’s serum creatinine, allowing derivation of estimated glomerular filtration rate (eGFR), should be obtained before an MRI scan. In patients with an eGFR below 30ml/min exposure to gadolinium MRI contrast agents should be avoided if possible, but imaging with gadolinium may be performed if no other imaging modalities are available to make the diagnosis and the diagnostic information is vital.
Traditional concerns about MRI of patients with implantable cardiac devices include possible movement of the device, programming changes, and induced lead currents that might cause heating and cardiac stimulation. The presence of a permanent pacemaker or implantable cardioverter defibrillator (ICD) is therefore currently considered a relative contraindication to MRI. However, if the information provided by cardiac MRI cannot be obtained using other imaging modalities, it may be necessary to take the small risk of MRI imaging. Clinical studies have shown that—as long as the patient is not dependent on antibradycardic pacing—noncardiac and cardiac MRI can be performed safely in patients with selected implantable pacemaker and defibrillator systems.
Clinical uses of cardiac MRI Assessment of ventricular function and mass
For this indication no contrast agent is needed and the technique has become the gold standard for quantifying ventricular volumes, ejection fraction, and myocardial mass. The image quality of native MRI recordings is usually excellent, with substantial differences in contrast between blood and myocardial tissue, which allows for accurate contour finding that is the prerequisite for reliable measurements of volumes and mass, particularly when the ventricular shape deviates from the assumed geometric model, as in patients with aneurysms or other major wall motion abnormalities.
With the availability of cardiac MRI, MUGA scanning—although also very accurate—has become less important for assessment of left ventricular volumes and function because of the radiation exposure that it requires. Cardiovascular MRI yields more accurate values for left ventricular parameters than planar imaging methods such as two-dimensional echocardiography or angiography. Newer echocardiographic techniques such as three-dimensional echocardiography and contrast echocardiography come close to cardiac MRI for assessment of left ventricular volumes and left ventricular function, but for accurate measurement of left ventricular mass MRI still appears to be superior to echocardiography. Therefore, in routine clinical practice the cheaper echocardiography remains first choice, but cardiac MRI is indicated when echocardiographic image quality is suboptimal.
By using the MRI-specific contrast agent gadolinium it is possible to display small areas of damaged myocytes or loss of myocytes and replacement by scar tissue. Gadolinium chelates are extracellular tracers that cannot cross cell membranes. In normal myocardium the myocytes are densely packed and the extracellular space and vascular volume represents less than 15% of the myocardial volume, hence after injection of gadolinium there are only few gadolinium molecules in a myocardial sample volume. By contrast, when the membranes of myocytes rupture, gadolinium molecules can penetrate into the myocytes and stay there, even late after gadolinium injection, such that in scar tissue the interstitial space is expanded and increased gadolinium concentration is found.
Late gadolinium enhancement (LGE)
Late after injection of gadolinium the intravascular contrast molecules are washed out and only the few gadolinium molecules in the interstitial space persist in normal myocardium. More gadolinium molecules remain in acutely damaged myocardium and scar tissue. The display of acute myocardial damage and scar tissue is independent from the underlying disease, hence LGE cannot distinguish between acute and chronic infarcts, but the pattern may differ in different ischaemic and inflammatory heart diseases.
The need for cardiac MRI in acute myocardial infarction is limited, but LGE reflects irreversibly injured myocardium and MRI is an excellent method for displaying the extent of myocardial infarction. For larger infarcts this is not clinically relevant because SPECT and echocardiography can do the same, but due to its higher spatial resolution LGE appears to be the best method for detection of small subendocardial infarctions—even the small infarcts that may occur during percutaneous coronary interventions.
LGE imaging is of more clinical relevance in chronic ischaemic disease and in many centres it has become the method of choice for assessing myocardial viability in chronic coronary artery disease. Alternative methods are SPECT and PET, which expose the patient to radiation, or echocardiography, which needs to be performed with dobutamine stress.
In patients with coronary artery disease impaired myocardial contraction can be due to necrosis of myocytes and subsequent fibrosis, or hibernating or stunned myocardium. Cardiac MRI appears to be the best imaging modality to differentiate viable (preserved myocytes) from nonviable (fibrotic) myocardium. Such assessment of viability is an important step in the consideration of patients with ischaemic cardiomyopathy, because percutaneous coronary interventions or bypass grafting are only indicated in vessels that supply viable myocardium. Myocardial tissue showing LGE is not likely to recover after revascularization of the supplying coronary artery and the transmural extent of LGE correlates negatively with the outcome after revascularization, but the threshold is not clear. When more than 50% of the wall thickness shows late enhancement further tests may be considered. To improve the sensitivity and specificity for prediction of functional recovery additional assessment of contractile reserve using low dose dobutamine may be considered, with imaging by echocardiography or MRI.
Subendocardial or transmural LGE related to the territories supplied by coronary arteries is typical for ischemic cardiomyopathy, but can also found in patients with diffuse reduction in contractility and only minor plaques seen by coronary angiography. This is probably caused by transient obstruction or spasm of the corresponding coronary arteries. By contrast, in dilated cardiomyopathy there may be no LGE, or a patchy or streaky pattern in the mid wall of the left ventricle.
LGE—often multifocal and patchy—is a frequent finding in patients with hypertrophic cardiomyopathy, but the clinical implication is not yet understood.
The diagnosis of acute myocarditis is often difficult using laboratory findings, ECG, and echocardiography. Cardiac MRI can confirm a clinical suspicion, typically revealing LGE in the myocardium in a patchy epicardial distribution, or band-like in the inferolateral wall. Cardiac involvement in sarcoidosis, which is otherwise difficult to diagnose, can also show a patchy LGE pattern.
Amyloidosis is another indication for cardiac MRI, with the gadolinium contrast agent diffusing into the amyloid that is deposited in the interstitium: hyperenhancement is diffuse, but more pronounced in the endocardial layers. In Fabry’s disease enhancement is diffuse or focal in the inferolateral midwall.
In many patients echocardiography is sufficient for clinical purposes, but all the information obtained in an echocardiographic examination is provided by MRI, and MRI images are more comprehensive because the window to the heart is not limited, and the images are often of better quality than the corresponding echocardiographic recordings.
Sedation is required in small children and monitoring is demanding in critically ill infants, hence CMR is usually performed following, and as an adjunct to, transthoracic echocardiography in neonates and infants. When readily available, it becomes a first-line technique in older children, adolescents, or adults, in patients with more complex anatomy, and at any age after surgery because body habitus and interposition of scar tissue and lungs become an increasing problem for transthoracic echocardiography. As discussed previously, cardiac MRI provides precise and reproducible quantification of ventricular volumes, mass, and function, and this is especially the case for the right ventricle, which is usually the chamber implicated in and stressed by repair of congenital heart disease. Cardiac MRI is also very effective for the evaluation of anomalies of the thoracic aorta and conduits.
MRI is indicated to define congenital or inflammatory changes of the coronary arteries. The spatial resolution and image quality of MRI are inferior to multislice CT, but for disease and anomalies affecting the proximal coronary arteries cardiac MRI appears to be the first choice as it does not entail exposure to radiation.
Cardiac MRI offers two methods for diagnosis of inducible myocardial ischaemia: firstly, assessment of left ventricular wall motion and thickening using dobutamine stress to provoke wall motion abnormalities in a territory supplied by a stenosed coronary artery; secondly, direct display of myocardial perfusion using gadolinium combined with adenosine stress, similar to nuclear imaging. Both methods compete with stress echocardiography and nuclear imaging, and CMR has similar sensitivity and specificity for detection of coronary disease. The most recent NICE guidelines on assessment of myocardial ischemia include both MRI methods, together with stress echocardiography and SPECT. The choice of perfusion imaging technique depends in practice on local availabilities, with stress echocardiography and nuclear imaging usually less expensive and more widely available than MRI techniques.
Multislice spiral computed tomography (MSCT) has become a valuable diagnostic procedure for a variety of diseases. It is indicated to assess pericardial thickening/calcification and is a fast and noninvasive method for the visualization of the coronary arteries, with the entire coronary circulation revealed in a single breath-hold with modern techniques. MSCT provides a unique opportunity to exclude significant coronary stenosis and to quantify the amount of coronary and aortic valve calcium.
For cardiac indications a volume data set is acquired, covering the distance from the carina to the diaphragmatic side of the heart. Native scanning is sufficient for assessment of coronary calcium. For display of the coronary lumen the investigation is longer and exposes the patient to more radiation, equivalent to the amount required for conventional coronary angiography, and intravenous infusion of a large amount of contrast agent (currently 100 ml) is needed. Other limitations of the technique are shown in Bullet list 1
Bullet list 1 Limitations of coronary CT
- ◆ Radiation exposure
- ◆ Calcium scoring 1.5–2 mSv
- ◆ Coronary angiography 4–12 mSv
- ◆ Needs iodinated contrast agents
- ◆ Needs low heart rate to avoid motion artefactsa
- ◆ Calcium/stents can impair the judgement of the lumen
a For visualization of coronary artery lumen only.
The high spatial resolution and contrast between the myocardium/valves and the blood pool means that an excellent display of various cardiac structures can be achieved. For many indications cardiac MRI and echocardiography are preferable because they do not require radiation or contrast agents, but there are two areas where cardiac CT can play a major role—the display of coronary calcification and arteriosclerotic lesions, and luminal obstructions of the coronary arteries.
The accuracy of CT techniques to detect and quantify calcified structures is unbeaten by other imaging techniques. The presence of calcifications is assumed if contiguous pixels with a density of more than 130 Hounsfield units (HU) are found within the coronary artery system. The Agatston score is used to quantify the amount of coronary calcium, which takes into account the area and the CT density of calcified coronary lesions.
Coronary calcium is a surrogate marker for coronary arteriosclerosis, only patients with renal failure have coronary calcifications without arteriosclerotic plaques. However, not every coronary plaque is calcified, and coronary calcium does not indicate or exclude instability of the plaque. The amount of coronary calcium correlates moderately closely with the amount of coronary arteriosclerosis, but there is only a weak correlation with the angiographic severity of obstructive coronary lesions. Significant obstruction of the coronary arteries is very unlikely, but not impossible, if no calcium is found in the coronary artery tree. Measurement of the coronary calcium score can be used for risk stratification of asymptomatic patients or patients with atypical angina. In the recent NICE guidelines measurement of coronary calcium is recommended in patients with chest pain and low likelihood of coronary disease (see Table 1).
|Table 1 Assessment of patients with stable chest pain according to NICE. CAD coronary artery disease, CT computer tomography, MPS myocardial perfusion scintigraphy|
|Estimated likelihood of CAD*|
|<10 %||consider non cardiac causes of chest pain|
|10-29%||offer CT calcium score:
1-400 offer CT angiography
< 400 follow pathway > 60% CAD
|30-60%||offer non-invasive functional imaging (MPS with SPECT or stress echocardiography or first-pass contrast-enhanced magnetic resonance perfusion or magnetic resonance imaging for stress-induced wall motion abnormalities)
if positive offer invasive coronary angiography
|<60%||invasive coronary angiography|
*according type of chest pain, age and sex (see NICE clinical guideline 95)