The evaluation of ischaemic heart disease requires the measurement of regional and global ventricular function, identification of the presence and extent of myocardial ischaemia, visualisation of luminal narrowing of the coronary arteries and assessment of myocardial infarction (MI) and viability. Cardiovascular magnetic resonance imaging (MRI) has evolved over the last decade into a valuable tool for the diagnosis and management of a wide spectrum of cardiovascular disorders. It provides anatomical and functional information in acquired and congenital heart disease and is the most precise technique for the quantification of ventricular volumes, function and mass.
Considerable technical and practice advances have been made since MRI was first introduced, and clinicians in the field are showing an unprecedented level of interest in it. MRI is non-invasive, has high spatial resolution and requires no potentially nephrotoxic contrast agent or radiation. It has been compared extensively with other established non-invasive imaging modalities such as electrocardiography (ECG) and nuclear cardiology (single photon emission computed tomography [SPECT]), and has been shown to be superior in many scenarios, particularly for the assessment of cardiac morphology and function.
Pathophysiological processes such as MI and ischaemia, stunning and hibernation and scarring and fibrosis can be identified easily using quick and simple protocols. Recently, late gadolinium-enhanced MRI and stress myocardial perfusion MRI have been shown to be useful in detecting infarct tissue and myocardial viability and in patient prognosis.
We present here an overview of the role of MRI in interventional cardiology, distinguishing two groups of patients: those with known and those with unknown coronary artery disease (CAD).
Evaluation of Patients with Coronary Artery Disease
The evaluation of ischaemic heart disease requires the assessment of ventricular morphology and function, myocardial perfusion and viability and coronary flow reserve.
MRI is the ‘gold standard’ for quantifying ventricular volumes, ejection fraction and myocardial mass.1 With ECG-gated and breath-hold sequences, the development of steady-state free precession sequences has provided substantially improved blood/myocardial contrast on cine MR images in a reduced imaging time, allowing more accurate and reproducible delineation of the endocardial borders. MRI uses volumetric quantification based on Simpson’s rule. Endocardial and epicardial contours are drawn during post-processing, generating end-diastolic, end-systolic and stroke volumes, ejection fraction and myocardial mass. In contrast to planar imaging modalities (2D ECG and ventriculography), MRI provides more accurate left ventricular (LV) parameter values, especially when ventricular shape deviates from the assumed geometrical model, as in ischaemic or dilated cardiomyopathy (DCM),2,3 and is the most reliable way to assess regional and global right ventricular (RV) function.4
Stress MRI has become a well-established modality for diagnosing myocardial ischaemia. Stress-induced wall-motion abnormalities can be detected with significantly greater diagnostic accuracy on dobutamine-stress cine MRI than on dobutamine-stress ECG in patients with CAD. This greater accuracy is generally explained by improved delineation of endocardial and epicardial borders and the higher quality of MR images in patients with obesity and pulmonary emphysema. Sensitivity and specificity of high-dose dobutamine-stress MRI in detecting significant CAD were reported to be, respectively, 83–86 and 83–86%.5,6 High-dose dobutamine stress is superior to adenosine stress for detecting significant coronary artery lumen narrowing as a stress-induced wall-motion abnormality.7 The event rate is low when dobutamine-stress MRI is normal,5,8,9 and higher in the presence of ischaemia.9 Cardiovascular MRI has also been used for pre-operative risk assessment.10
The accurate assessment of haemodynamic coronary artery stenosis is important for evaluating patients with chest pain syndromes and in managing patients after therapeutic interventions. Dynamic MRI following a bolus injection of contrast material enables the assessment of first-pass myocardial enhancement during pharmacological stress, which can provide information on the presence and extent of CAD. Rest perfusion MR images are usually analysed semi-quantitatively, and remain unaltered unless epicardial coronary artery luminal diameter stenosis exceeds 90%. Vasodilator stress with adenosine or dipyridamole relaxes arteriolar tonus and induces a perceptible difference between normal and ischaemic myocardial perfusion.11 The improved spatial resolution of MRI enables the detection of slight subendocardial ischaemia in mild to moderate CAD and delineation of diffuse subendocardial hypoenhancement in multivessel CAD. The validity of myocardial perfusion MRI has been shown by comparison with other techniques such as SPECT,12–14 positron emission tomography (PET)15 and coronary angiography.16–18 Stress perfusion MRI showed sensitivity of 88–91%, specificity of 75–94% and accuracy of 85–89% in detecting CAD.13,18–20
In patients with known or suspected CAD, myocardial ischaemia detected by stress MRI can be used to identify those at high risk of subsequent cardiac death or non-fatal MI. For patients with normal stress MRI, three-year event-free survival was 99.2%.21
In patients with acute and chronic ischaemic heart disease, assessing reversible and irreversible myocardial ischaemia, visualising acute MI (AMI) and detecting viable myocardium are important diagnostic tasks.
MI can be detected with high accuracy and sensitivity using late gadolinium-enhanced MRI. Gadolinium is introduced intravenously and, after 10–20 minutes, MRI is performed using an inversion recovery sequence, whereby the inversion time is chosen to obtain a null myocardial signal (dark signal). In areas of MI, the gadolinium shifts towards the expanded extracellular compartment, leading to a higher concentration on the late enhancement scan, which shows as a bright signal (‘hyperenhanced’) compared with the low signal of the normal heart (whence the aphorism ‘bright is dead’). Late gadolinium-enhanced MRI has been shown to accurately detect both Q-wave and non-Q-wave MI.22
After a successful primary angioplasty in AMI, microvascular obstruction (blockage of capillaries by microemboli and endothelial oedema), known as the no-reflow phenomenon, may be encountered. In such cases, a particular pattern of late gadolinium enhancement is seen, e.g. dark subendocardial areas of microvascular obstruction where contrast cannot penetrate, surrounded by a rim of hyperenhancement (‘halo sign’). This microvascular obstruction detected by MRI has been linked to ventricular remodelling and adverse cardiovascular events.23,24
Because the technique is so sensitive, MRI has been shown to identify subendocardial MI when wall motion and perfusion on SPECT are normal.25 When the transmural extent of infarction is <50%, the likelihood of functional recovery in AMI is good.26 MRI accurately measures wall thickness, which is reduced in chronic transmural MI, allowing the presence of viable myocardium to be excluded in chronic infarcts, with excellent correlation to fluorodeoxyglucose PET findings27,28 and greater accuracy than thallium SPECT.29 Radionuclide techniques expose patients to a substantial amount of ionising radiation, and PET is performed by relatively few specialised centres.
Coronary Artery Imaging
Coronary MR angiography (MRA) has been the subject of intense research over the past few years. Small vessel size and motion artefacts during the cardiac cycle make imaging the coronary arteries difficult. Coronary MRA has been the focus of technical developments in the past few years and could become a promising non-invasive diagnostic method.
MRI is currently a useful means of non-invasive detection of anomalous coronary arteries and their origin and initial course, without radiation or contrast injection.30 Coronary MRA also enables the non-invasive detection and measurement of coronary artery aneurysms in patients with Kawasaki disease. Coronary artery bypass grafts (and especially vein grafts) are fairly easy to image because of their fixed position and large lumen size. However, the presence of sternal metal clips and markers used in surgery can cause artefacts.
Recent studies indicate that MRI for the detection of CAD is not as far from clinical feasibility as many physicians assume. With the introduction of new sequences, blood signal intensity on 3D coronary MRI is considerably augmented without the use of an MR contrast medium. This enables the acquisition of a 3D axial volume encompassing the entire heart without losing arterial contrast. A whole-heart approach visualises all three major coronary arteries, thereby reducing total examination time. Coronary MRI acquired with this approach is useful in assessing lumen narrowing in coronary arteries exhibiting heavy calcification in the vessel wall on multislice CT.
In a recent study evaluating 113 patients,31 whole-heart coronary MRI allowed the non-invasive detection of significant narrowing in coronary arterial segments with a diameter of 2mm or more, with moderate sensitivity of 82% and high specificity of 90% in patient-based analysis; in vessel-based analysis, sensitivity was 85% for the right coronary artery, 77% for the anterior descending artery and 70% for the circumflex artery. However, MR acquisition was not successful in approximately 14% of patients, who had unstable breathing patterns.
Detection of Coronary Artery Disease
MRI can be useful in detecting CAD in some indications, as follows.
The morphological and functional abnormalities of DCM are clearly demonstrated and quantified by MRI. These findings, however, may fail to distinguish DCM from other forms of LV dysfunction, such as those resulting from CAD. This differentiation is important because the treatment is not the same. LV dysfunction due to CAD may benefit from revascularisation and secondary preventative pharmacotherapy with statins and aspirin. An advantage of MRI over ECG is the possibility of using late gadolinium enhancement, which shows no uptake in the majority of DCM patients, making MRI a useful non-invasive clinical tool for distinguishing LV dysfunction related to CAD so as to perform invasive coronary angiography in those cases only.32
In some DCM patients, late gadolinium enhancement is seen, but only in the mid-myocardium in a non-coronary pattern, which is clearly distinguishable from CAD, and is recognised by pathologists as mid-wall fibrosis seen at post mortem. Late gadolinium enhancement is associated with a significantly worse outcome in idiopathic DCM.32
Acute Chest Pain in the Hospital Setting
Managing chest pain in the emergency department remains a challenge with current diagnostic strategies. Recent estimates indicate that about 2% of patients with MI are discharged inappropriately from the emergency room.33,34 Patients with suspected acute coronary syndrome (ACS) with a low or intermediate probability of CAD, in whom follow-up 12-lead ECG and cardiac biomarker measurements are normal, undergo functional cardiac testing (resting nuclear scan or ECG) and/or stress testing (treadmill, stress ECG or stress nuclear testing) or non-invasive coronary imaging (multislice CT scan). In these cases, MRI has been used in the emergency room in the assessment of chest pain, and showed a sensitivity and specificity of, respectively, 84 and 85% in identifying patients with CAD. Multivariate analysis, including standard clinical tests (ECG, troponin and Thrombolysis in Myocardial Infarction risk score), showed that MRI was the strongest predictor of CAD and added diagnostic value over clinical parameters, including identification of enzyme-negative unstable angina.35
Acute Coronary Syndrome with ST Elevation and Normal Coronary Angiography
Some patients have various combinations of chest pain, haemodynamic instability, ischaemia-like ECG changes, biochemical marker elevation and segmental wall motion abnormalities at presentation associated, however, with normal coronary angiography. It still may be difficult to distinguish AMI yielding normal coronary angiographic findings (owing to temporary coronary artery occlusion followed by spontaneous reperfusion, or to a prolonged coronary artery spasm) from myocarditis or apical ballooning syndrome (Takotsubo cardiomyopathy). Treatment and evolution therefore strongly depend on the formal differential diagnosis. Cardiac MRI can add valuable information for those patients and is currently the only non-invasive technique able to make the aetiological diagnosis.
Formal differential diagnosis between AMI and acute myocarditis is based on positive findings of endomyocardial biopsy, usually considered as the gold standard. However, this procedure is not suitable for the diagnosis and treatment of myocarditis cardiomyopathy due to its dependence on sample quality (frequent false negative results), considerable interobserver variability, severe adverse effects (questionable risk–benefit ratio) and failure to identify areas suspected of being affected by an ischaemic process.36-38
We therefore need an accurate non-invasive procedure to diagnose acute myocarditis that can be repeated during follow-up to assess complete resolution of inflammation. MRI with early first-pass perfusion and delayed enhancement sequences plays an important role in the differential diagnosis of myocarditis versus AMI, particularly in cases of normal coronary angiographic findings. Myocarditis is characterised by normal first-pass perfusion MRI and nodular delayed enhancement in a diffuse, predominantly inferolateral subepicardial location in non-vascular territories; AMI is associated with early subendocardial perfusion defects and subendocardial or transmural delayed enhancement of a smaller number of segments, all in a vascular distribution. In a delayed enhancement sequence, myocardial abnormalities were mainly in subepicardial or mid-wall locations, whereas the AMI abnormalities were subendocardial or transmural.39-40
Apical Ballooning Syndrome (Takotsubo Cardiomyopathy)
Clinically, apical ballooning syndrome is characterised by transient LV apical wall motion abnormality or, less often, mid-ventricular dysfunction, chest pain with altered ECG and minimal myocardial enzymatic release, mimicking AMI, but without significant CAD. Prognosis is generally favourable.
MRI with early first-pass perfusion and delayed enhancement sequences may be used for distinguishing between different causative aetiologies, including MI and myocarditis. Apical ballooning syndrome is characterised by normal firstpass perfusion MRI, and the absence of any pathological signal activity on late enhancement imaging rules out MI or myocarditis.41,42 MRI disclosed, for the first time, RV involvement in more than one-third of the reported cases with Takotsubo syndrome.43,44 Significant bilateral pleural effusion was exclusively associated with RV involvement, of which it was a reliable indicator. In addition, patients with RV involvement had significantly lower LV ejection fraction, which may have a significant impact on morbidity and outcome.
MRI is very safe, and no long-term adverse effects have been demonstrated. Claustrophobia may be problematic in about 2% of patients. The magnetic fields are always active, posing risks for patients and staff and requiring meticulous safety procedures such as preventing the introduction into the scanner area of ferromagnetic objects, which could otherwise become projectiles.
Neurovascular clips, cochlear implants, metal in the eye and retained shrapnel are thus all contraindications for MRI, as are pacemakers, automatic implantable defibrillators and neurostimulators, although certain models may be safe.45 However, intracoronary stents, coronary artery bypass graft surgery and all mechanical heart valves (except Starr-Edwards valves 100 and 6000 and Carpentier-Edwards rings 4400 and 4500) are considered safe. When in doubt, various resources, such as www.mrisafety.com,46 are available to check a device’s safety before MRI.
The benefits of high-field MRI systems (3T and over) have been thoroughly explored for neuroimaging, functional imaging and spectroscopy of the head. The most important benefits of cardiac imaging at 3T arise from the increased signal-to-noise ratio (SNR) and contrast-to-noise ratio. The latter improves background suppression, contrast and detection of small defects in first-pass perfusion imaging and delineation of infarct extent in viability imaging.47,48 The increased SNR and benefits of 3T for parallel imaging can be used to accelerate imaging in order to provide broader coverage or higher spatial or temporal resolution.49–51 Coronary MRI at 3T did not significantly improve overall image quality score on visual assessment, and no significant difference in diagnostic accuracy was observed between 1.5 and 3T.52,53
Gains with higher field strength are somewhat offset by the presence of more severe artefacts in many cardiac MRI sequences. Susceptibility-related local magnetic field variations, increased sensitivity to motion artefacts and impaired ECG R-wave triggering are cited as factors related to increased artefaction at 3T. Another major limitation of 3T for cardiac MRI is radiofrequency-induced heating.
With further advances in hardware and the optimisation of pulse sequences for high-field cardiac MRI, 3T MRI may become a platform of choice for high-resolution cardiac MRI.
There have been considerable technical and clinical advances in cardiac MRI in the last several years. Although there is overlap with other cardiac imaging modalities, MRI often works in a complementary fashion to these other techniques or resolves residual diagnostic dilemmas in interventional cardiology. The strengths of MRI lie in its ability to comprehensively image cardiac anatomy, function, perfusion, viability and physiology in one-stop testing and to provide high-quality diagnostic information without the need for radiation.