REVIEW

Review

Role of MRI in clinical cardiology

Godwin Constantine, Kesavan Shan, Scott D Flamm, Mohan U Sivananthan

Rapid progress has been made in cardiac MRI (CMRI) over the past decade, which has firmly established it as a
reliable and clinically important technique for assessment of cardiac structure, function, perfusion, and myocardial
viability. Its versatility and accuracy is unmatched by any other individual imaging modality. CMRI is non-invasive and
has high spatial resolution and avoids use of potentially nephrotoxic contrast agent or radiation. It has been
extensively studied against other established non-invasive imaging modalities and has been shown to be superior in
many scenarios, particularly with respect to assessment of cardiac and great vessel morphology and left ventricular
function. Furthermore, its clinical use continues to expand with increasing experience and proliferation of CMRI
centres. As worldwide prevalence of cardiovascular disease continues to rise, CMRI provides opportunity for improved
and cost-effective non-invasive assessment. Continued progress in CMRI technology promises to further widen its
clinical application in coronary imaging, myocardial perfusion, comprehensive assessment of valves, and plaque
characterisation.

During the past decade, cardiac MRI (CMRI) has discontinued, the magnetisation vector starts to recover
emerged as an important procedure in investigation of back to its former position, releasing signal in the form of
cardiovascular disease. Many clinical and experimental radiowaves. This relaxation of net vector is attributable to
studies have established its clinical use over other two distinct but simultaneous processes, referred to as the
modalities in various situations. CMRI is non-invasive, longitudinal (T1) and the transverse (T2) relaxations.1,2
has high spatial resolution, and does not use ionising A pulse sequence consists of a series of radiofrequency
radiation. Furthermore, it can characterise tissue to a pulses of varying duration or strength and application of
greater degree than other modalities. With technological magnetic-field gradients that are adjusted to highlight
advancements, we can now rapidly acquire imaging desired tissue characteristics.1,2 Basic pulse sequences used
datasets that are essential for imaging the heart. Here, we in CMRI are spin-echo and gradient-echo sequences, or
outline present clinical indications of CMRI and its their faster hybrids (table).
potential for the future. Spin-echo sequences are usually used for assessment of
morphology, and flowing blood typically appears black.
The technique Gradient-echo sequences have fairly low soft-tissue
MRI is based on the principle of nuclear magnetic contrast compared with spin-echo sequences, and flowing
resonance and has intrinsic attributes well suited for blood is represented by high signal intensity and
cardiac imaging. It is a tomographic technique that can turbulence as areas of signal void. Gradient-echo
acquire images in virtually any orientation. sequences are used in assessment of valvular lesions,
Images are derived from signals produced by protons shunts, and great vessels, and of ventricular function and
(hydrogen nuclei), which are present in abundance wall-motion characteristics.
because the human body consists mainly of water. The Flowing blood across a magnetic gradient can be
proton behaves like a small magnet when placed in a encoded to quantify flow. These phase-encoded velocity
magnetic field: it aligns with the field and precesses with a maps are based on the principle that magnetic vectors of
given frequency that depends on field strength. Protons flowing protons acquire a phase shift that is proportional
will align parallel and antiparallel to the direction of the to flow velocity. This technique is used to assess flow
primary field, with a small excess of parallel protons that volumes and velocity profiles across valves and shunts.
gives rise to a net magnetisation vector. This net vector CMRI can be used to visualise intrinsic contractile
can be altered—ie, rotated—by application of a secondary patterns of the myocardium by tagging sequences. Tag
temporary radiofrequency pulse. Once this pulse is lines and grids are produced by a combination of
restricted localised radiofrequency pulses and gradients to
Lancet 2004; 363: 2162–71 modulate magnetisation.

British Heart Foundation Cardiac MRI Unit, General Infirmary at

Leeds, Leeds, UK (G Constantine MD, M Sivananthan FRCR); Search strategy and selection criteria
Department of Cardiology, Southwest Memorial Hospital and We did an extensive search for original articles and reviews
University of Texas Medical School, Houston, TX, USA from 1987 to March 15, 2003. We used EMBASE and
(K Shan MD); and Department of Cardiology and Radiology, PubMed in addition to manual journal search. The keyword
St Luke’s Episcopal Hospital and Texas Heart Institute, Houston, “cardiac magnetic resonance imaging”, with a combination
TX, USA (S D Flamm MD) of subtitles such as “myocardial viability” and other related
Correspondence to: Dr Mohan Sivananthan, Room D170, terms, were used for the search. We included major
Jubilee Building, General Infirmary at Leeds, Gt George Street, Leeds studies, studies published from institutions specialising in
LS1 3EX, UK specific conditions, and important review articles.
(e-mail: [email protected])

2162 THE LANCET • Vol 363 • June 26, 2004 • www.thelancet.com

For personal use. Only reproduce with permission from The Lancet Publishing Group.
 REVIEW

Description Vendor-specific names

Technique
Spin-echo pulse sequence: Standard Spin echo, SE (G, P, S)
black blood Fast (breath-hold technique) Fast spin echo (G); turbo spin echo, TSE (P,S)
Ultrafast (single heartbeat technique) SSFSE (G); single-shot TSE (P); HASTE (S)
Gradient-echo pulse sequence: Standard (conventional, non-breath-hold technique) Spoiled GRASS, SPGR (G); fast field echo, FFE (P); FISP (S)
bright blood Fast (breath-hold segmented K-space technique) Fast SPGR (G); turbo field echo, TFE (P); TurboFLASH (S)
Steady-state free-precession (breath-hold segmented FIESTA (G); balanced-FFE (P); TrueFISP (S)
K-space technique, high SNR/CNR, real time)
Segmented EPI (real time) Segmented EPI (G); TFE-EPI (P)
Parallel imaging Accelerated imaging (by sensitivity encoding) ASSET (G); SENSE (P); IPAT (P)
G=General Electric (MI, USA). P=Philips (Best, Netherlands). S=Siemens (Erlangen, Germany). SSFSE=single-shot fast spin echo. HASTE=half-Fourier acquisition single-
shot turbo spin echo. GRASS=gradient recalled acquisition in a steady state. FISP=fast imaging with steady-state precession. SNR=signal to noise ratio. CNR=contrast
to noise ratio. FLASH=fast low angle shot. FIESTA=fast imaging employing steady aquisition. EPI=echo planar imaging. ASSET=array spatial sensitivity encoding
technique. SENSE=sensitive encoding. IPAT=integrated parallel acquisition technique.
Commonly used CMRI sequences1,2

Cardiac and respiratory motion artifacts are not unique towards the fixed base. This complex intramyocardial
to CMRI, but can be especially problematic. Fortunately, contractile pattern can be assessed non-invasively by
electrocardiograph-gating, breath-hold navigation, and magnetic resonance myocardial tagging,10,12 and is
real-time techniques have been developed and refined to especially useful when studying abnormalities in contrac-
favourably synchronise data acquisition and keep to a tion in ventricular hypertrophy and cardiomyopathies. It
minimum or eliminate such artifacts and allow clear also enables qualitative and quantitative assessment of
precise images of morphology and function. However, regional wall-motion abnormalities and overall cardiac
cardiac dysrhythmias can still interfere with synchronised function.
data acquisition. Cardiac haemodynamic assessment can be best studied
Metallic implants can lead to artifacts, which might by velocity mapping. Three-dimensional velocity meas-
degrade images. Most cardiac prosthetic valves and urements allow accurate depiction of peak velocities
coronary stents presently in use are safe for CMRI at unhindered by direction of flow. This technique is
available clinical field strengths. However, safety of new typically used in preoperative and postoperative assess-
stents and high field strengths needs to be ascertained ment of congenital cardiac diseases, coarctation of aorta,
individually. Cardiac pacemakers and cardioverter- and valvular lesions.13
defibrillators remain contraindications. Children and
patients with claustrophobia and anxiety may need Aortic disease
sedation or anaesthesia during study.3,4 With modern Findings of several studies have established the place of
ultrafast sequences, even patients with acute cardiac CMRI in the screening, diagnosis, and follow-up of aortic
conditions can undergo CMRI with careful precautions pathology. The large field of view, flexible imaging planes,
and appropriate monitoring.5 and ability to image surrounding structures make CMRI
Gadolinium-based contrast agents used in CMRI are an ideal method for diagnosis of diseases of the aorta.14,15
safer than iodinated ones used in radiography. In a study comparing transoesophageal echocardiography,
Gadolinium agents are not nephrotoxic and have much CT, transthoracic echocardiography, and CMRI, CMRI
lower incidence of allergic reactions than radiography was shown to have equal if not greater sensitivity and
contrast agents.6 superior specificity over other imaging modalities in
diagnosis of acute aortic dissection.16
Assessment of morphology and function Although choice of diagnostic modality depends on
The clinical role of CMRI is well established for availability and local expertise, CMRI is the imaging
morphological assessment of complex cardiac anomalies modality of choice in stable patients with suspected aortic
and great vessels. This role is partly attributed to the large dissection.17,18 CMRI can delineate extent, site of entry,
field of view of CMRI, its unlimited scanning planes, and and involvement of arch vessels, and detect renal and
good tissue contrast. Furthermore, CMRI is a precise and other visceral arterial involvement in patients with aortic
highly reproducible method for measurement of right dissection (figure 1). It has enhanced sensitivity and
ventricular and left ventricular systolic and diastolic specificity in diagnosis of dissection in patients with
function, and mass.7–10 previous aortic disease.19
In echocardiography, mass and volumes are derived Intramural haematoma of the aorta is a variant of
from a limited two-dimensional dataset with assumptions dissection, for which there is no detectable entry site and
of cardiac outline based on a geometrical model. This no flow in the false lumen. This disorder is potentially
model can potentially lead to discrepancies in hearts with lethal, with frequent progression to aortic rupture, which
abnormal anatomy or contraction patterns. However, in is increasingly recognised after introduction of CMRI.20
CMRI, actual myocardial mass and volumes are obtained Need for early detection of postoperative complications
with Simpson’s algorithm from three-dimensional data in and long-term surveillance of patients undergoing aortic
the form of series of thin slices with no geometric surgery has been emphasised in various studies. Serious
assumptions. Thus, CMRI is judged the non-invasive late complications after surgery to the ascending aorta—in
reference standard for cardiac function and mass and the sometimes symptom-free patients—are readily identified
preferred imaging modality in situations in which accurate by CMRI. These include periprosthetic haematomas,
variables are vital. CMRI is especially suited for research false aneurysms, and persistent residual dissections.
studies that need longitudinal assessment of ventricular Therefore, CMRI is the technique of choice for
variables since it considerably reduces the sample size monitoring of the aorta after surgery.21–23 A combination of
needed because of its high reproducibility.11 spin-echo and gradient-echo techniques can be used to
Cardiac contraction happens in a three-dimensional diagnose nature and extent of thoracic aortic aneurysms
rotational fashion starting at the apex and moving irrespective of their cause.24

THE LANCET • Vol 363 • June 26, 2004 • www.thelancet.com 2163

For personal use. Only reproduce with permission from The Lancet Publishing Group.
REVIEW

gradients determined by doppler echocardiography and

cardiac catheterisation (r=0·96 and r=0·97,
respectively).37
Since CMRI can reproducibly determine ventricular
function and myocardial mass, it enables accurate
assessment of ventricular remodelling. CMRI can be used
to follow up patients with valvular disease who might need
surgical intervention in the future. Its potential in clinical
use in valvular disease needs further study.

Congenital heart disease

Management of congenital heart diseases depends on
precise definition of cardiac morphology and altered
haemodynamics. Echocardiography and cardiac
catheterisation are the standard procedures used to assess
these variables. However, because CMRI can completely
assess cardiac and vascular morphology, venoatrial
connections, and visceral situs and extracardiac
abnormalities, it is effective in anatomic assessment of
complex congenital heart defects.38,39 Although cardiac
catheterisation cannot be entirely replaced by CMRI in
these patients, need for diagnostic angiography and its
extent and duration can be kept to a minimum.40,41
Introduction of post-processing of contrast-enhanced
magnetic resonance angiography and fast cine sequences
Figure 1: Magnetic resonance angiography with intravenous are having a great effect on investigation of congenital
gadolinium contrast, arterial phase (frontal view), in a patient heart disease. Contrast enhancement helps to differentiate
with type A dissection between the vessel lumen and vessel wall independent of
Dissection is confined to the ascending aorta, and true (T) and false (F) flow effects, and enables small tortuous vessels to be
lumen are clearly seen. Inset: axial spin-echo image at mid-thoracic level, clearly visualised. Contrast-enhanced magnetic resonance
showing dilated ascending aorta and normal-sized descending aorta (DA)
with evidence of intramural haemorrhage (arrows). Arrowheads indicate angiography data can be used to create three-dimensional
dissection flap in ascending aorta. and four-dimensional reconstructions, and, along with
techniques such as maximum intensity projections,
Aortic elastic property has been recorded by CMRI, further image interpretation is possible.42–44
either by calculation of aortic distensibility or by In coarctation of aorta, high-resolution anatomic
measurement of pulse-wave velocity through the aorta. definition of the obstruction and collateral vessels can be
Assessment of elastic properties can aid early diagnosis of shown with various CMRI techniques. Contrast-
and screening for Marfan’s syndrome.25,26 In patients with enhanced magnetic resonance angiography has replaced
this disorder who have undergone repair of the ascending conventional angiography as the preferred mode of
aorta, a more than 3 mm rise in diameter of the
thoracoabdominal aorta during follow-up was reported to
be relevant and associated with need for further surgery.27
CMRI can image sequelae of aortic atheroma,
including penetrating ulcers, and can differentiate
intramural haematoma from atherosclerotic plaque.15,20,28

Valvular heart disease

Echocardiography and catheterisation remain the
mainstay of assessment for valvular heart disease, despite
their limitations. However, CMRI is rapidly gaining
acceptance as an accurate and highly reproducible
method for best assessment of structural and functional
valvular variables. Although direct visualisation of valve
leaflets is difficult because of limited temporal resolution,
new techniques are showing promise (figure 2).29
Specificity, sensitivity, and diagnostic accuracy of cine
gradient-echo CMRI for detection of mitral and aortic
regurgitation is greater than 93%, 89%, and 92%,
respectively.30 Phase-encoded magnetic resonance velocity
mapping is highly accurate and reproducible for
measurement of regurgitant volume and fraction (r>0·97)
in patients with chronic aortic regurgitation.31,32 The
regurgitant jet can also be semiquantitatively graded, with Figure 2: CMRI of aortic valve disease
CMRI comparing well with transoesophageal echo- Systolic (A) and diastolic (B) short axis views of bicuspid dysplastic valve
cardiography (for jet length, r=0·85; for area r=0·91).33–35 with stenosis, showing strikingly thickened valve leaflets (arrowheads)
However, subjective assessment of jet characteristics in and fused commissure (arrow). (C) Systolic frame of a patient with
systemic lupus erythematosus, showing mild thickening of leaflets. (D) In
CMRI varies greatly depending on scanning variables and diastole there is absence of coaption of cusps leading to a central
sequences used.36 Alternatively, the stenotic pressure regurgitant jet (black arrow). LA=left atrium. RA=right atrium. RVOT=right
gradients measured by CMRI correlate closely with ventricular outflow tract.

2164 THE LANCET • Vol 363 • June 26, 2004 • www.thelancet.com

For personal use. Only reproduce with permission from The Lancet Publishing Group.
 REVIEW

investigation in cases of isolated

coarctation (figure 3).45,46 Magnetic
resonance velocity mapping can detect
peak velocity across the stenosis, with
good correlation with continuous wave
doppler (r=0·95).47 An alternative
method is to measure volume flow in
the descending and ascending aorta,
thereby determining collateral
contribution. An important relation
exists between amount of flow increase
in the distal aorta because of
collaterals—as measured by CMRI—
and reduction in diameter at the
coarctation site (r=0·94) and clinical
gradient (r=0·84).48 Coarctation of
aorta needs lifelong postoperative
surveillance, for which CMRI is the
most cost-effective method.49–51
Clear visualisation and assessment of
right ventricle and pulmonary artery
morphology and function are
important in assessment of many
congenital heart diseases. Assessment
of right ventricular structure and Figure 3: Contrast magnetic resonance angiography (maximum intensity projection) in
function by echocardiography is a patient with interrupted aortic arch (block arrow) in the frontal (A) and lateral (B)
difficult because of its proximity to the projections
chest wall and complex shape. Arrowheads show some of the intercostal collaterals. LV=left ventricle.
Furthermore, the right ventricle is
usually distorted in these conditions. CMRI can assess and asymptomatic patients with hypertrophic cardio-
structure and function accurately and reproducibly in myopathy by contrast-enhanced techniques.68–70 The
such situations.52,53 This method has been extensively precise importance of the distribution of contractile
evaluated in the assessment of pulmonary valve, abnormalities, perfusion defects, and fibrosis remains to
pulmonary arterial tree, and related haemodynamics in be determined.
congenital heart disease and is the preferred non-invasive Arrhythmogenic right-ventricular cardiomyopathy
modality.54,55 presents a special case for use of CMRI in diagnosis and
CMRI is an invaluable technique in assessment of follow-up.71 Most features of this disorder seen with CMRI
postoperative patients, especially in disorders in which the are dependable, without much interobserver variability.72
right ventricle has been affected. CMRI is often superior Magnetic resonance detects the characteristic high signal
to echocardiography and catheterisation, because these intensity of fat on spin-echo images within the right
techniques can be hindered by post-surgical fibrosis and ventricular myocardium. It identifies diffuse and focal
anatomical distortion—complications that do not impede thinning and local aneurysms of the right ventricle along
CMRI.56–60 with both systolic and diastolic dysfunction. Fat-
suppression techniques can be used to correctly confirm
Cardiomyopathies fat within the wall. Diastolic dysfunction is an early marker
CMRI is well established in diagnosis and follow-up of of disease and progression is typically characterised by
cardiomyopathies. It can characterise tissue and provide development of systolic dysfunction.73 Although diagnosis
good morphological definition. Also, accurate assessment is traditionally based on presence of major and minor
of function aids diagnosis and prognosis of various types criteria encompassing genetic, electrocardiographic,
of cardiomyopathies.
Although echocardiography gives
adequate morphological and functional
data, CMRI is especially valuable in
diagnosis of rare forms and atypical
patterns. Moreover, since it can
accurately assess left ventricular mass
and distribution, CMRI is valuable in
distinguishing physiological hyper-
trophy and dilatation from cardio-
myopathies.61–64
CMRI is an important method to
study pathogenesis of hypertrophic
cardiomyopathy (figure 4). Reduced
three-dimensional myocardial shorten-
ing, enhanced left ventricular torsion,
and abnormal strain rates have been Figure 4: Short axis views in diastole and systole in a patient with hypertrophic
shown in this disorder.65–67 Myocardial cardiomyopathy
perfusion abnormalities and fibrosis There is hypertrophy of anterior wall and upper interventricular septum (small arrows) and a normal
have been shown in both symptomatic area in posterolateral wall (large arrow).

THE LANCET • Vol 363 • June 26, 2004 • www.thelancet.com 2165

For personal use. Only reproduce with permission from The Lancet Publishing Group.
REVIEW

pathophysiological, and histopathological features, positive Myocardial iron deposition is typically preferentially seen
findings in CMRI should be included as important subepicardially and results in low signal intensity on T2-
additional criteria in clinical diagnosis of arrhythmogenic weighted images (because of magnetic susceptibility
right-ventricular cardiomyopathy.71,74 effects).84,85
CMRI is also useful to study pathogenesis of
dilated cardiomyopathy and to differentiate it from Cardiac tumours
myocarditis. Post-gadolinium T1-weighted sequences can Magnetic resonance has an increasing part to play in
distinguish early nodular and late diffuse enhancement management of cardiac neoplasms. Apart from atrial
patterns in myocarditis.75,76 myxomas and primary valvular tumours, for which
Restrictive cardiomyopathy and constrictive pericarditis echocardiography is adequate for diagnosis and planning
have many similar clinical and pathophysiological of treatment, CMRI provides additional anatomical
characteristics. Thickened pericardium has been described information within the heart and evidence of extension
as the most useful distinguishing feature of constrictive into extracardiac structures.86,87
pericarditis. Sensitivity, specificity, and accuracy for Selective sequences—such as fat-suppression tech-
diagnosis of constriction with CMRI are high (88%, niques—can be used to further characterise tumours such
100%, and 93%, respectively), with pericardial thickening as lipomas or to accurately define extent of lesions when
seen in 88%.77–79 However, calcification of the pericardium, there is surrounding adipose tissue.88 Use of contrast
which is an important feature, cannot be directly shown better delineates tumour mass and invasion and
with CMRI. characterises tissue, as shown by differential enhancement
In sarcoidosis, detection of infiltrates in myocardium attributable to variation in vascularity and capillary
carries an adverse prognosis. Infiltrates are seen as areas of permeability. Contrast enhancement also can help to
amplified T2 signal intensity (attributable to inflammation, differentiate thrombi from tumours because thrombi do
oedema, and granulomas) and they enhance after not usually enhance.89,90
administration of gadolinium. CMRI is also a useful non- Primary malignant cardiac tumours are typically
invasive method in assessment of response to steroid sarcomas. On CMRI, sarcomas appear as heterogeneous,
treatment.80–82 CMRI also has a potential role in assessment broad-based large masses that usually occupy most of the
of cardiac amyloidosis by detection of the presence of affected cardiac chamber and extend into adjacent
thickening of the right atrial and right ventricular wall and chambers. Contrast enhancement generally shows
differentiation of it from hypertrophic cardiomyopathy.83 heterogeneity, with non-enhancing necrotic areas.
Preliminary work has shown the use of CMRI in Pericardial irregularity and effusion due to tumour
diagnosis and assessment of response to treatment in invasion and extracardiac extension are typical, with
patients with cardiac involvement in haemochromatosis. CMRI providing good definition for surgical planning.91,92

Ischaemic heart disease

Evaluation of ischaemic heart disease
includes not only definition of coronary
anatomy but also assessment of
ventricular morphology and function,
myocardial perfusion and viability, and
coronary flow reserve.

Coronary artery imaging

Coronary arteries are imaged routinely
by conventional angiography. Non-
invasive assessment of these arteries
would be a great advance in investi-
gation and follow-up of ischaemic
heart disease. Coronary magnetic
resonance angiography (CMRA) has
been the focus of technical develop-
ment in the past few years and could
become a promising non-invasive
diagnostic method. However, the fairly
small size of coronary arteries, their
complex course, and constant motion
with cardiac contraction and res-
piration makes CMRA technically
difficult.
Several two-dimensional and three-
dimensional magnetic resonance
techniques have been proposed for
CMRA. Sensitivity and specificity
varies widely depending on the
technique used and vessels and
Figure 5: Examples of anomalous coronary arteries segments analysed.93–96 In a multicentre
(A, B) Single coronary artery arising from right sinus of the aorta (Ao). Right coronary artery (black study,97 use of three-dimensional
arrow) has a normal course whereas left main stem (arrowheads) is running in between aorta and
right ventricular outflow tract (RVOT). (C, D, E) Coronary arteries arising separately from left sinus
CMRA was compared with elective
with normal course of left main stem (white arrow), while proximal portion of right (black arrow) is X-ray coronary angiography in
running between outflow tracts. LA=left atrium. DA=descending aorta. 109 patients. Results showed that 636

2166 THE LANCET • Vol 363 • June 26, 2004 • www.thelancet.com

For personal use. Only reproduce with permission from The Lancet Publishing Group.
 REVIEW

of 759 proximal and middle segments of coronary arteries CMRI perfusion reserve improves after revascular-
could be interpreted on magnetic resonance angiography. isation.112 Development of multislice techniques with fast
In these segments, 83% of clinically important lesions near-real-time sequences has resulted in near-complete
(⭓50% luminal stenosis) were detected by CMRA. myocardial coverage, making first-pass CMRI perfusion a
Accuracy of this technique in diagnosis of coronary artery viable alternative to scintigraphy.
disease was 72% (95% CI 63–81). Sensitivity, specificity, Phase relaxation measurements (T2*) in myocardium
and accuracy of detection of patients with disease of the attributable to blood oxygenation level-dependant effect,
left main stem or three-vessel disease were 100%, 85%, based on the paramagnetic properties of deoxyhaemo-
and 87%, respectively.97 globin, has been reported.113 This experimental technique,
Studies comparing CMRA with conventional done without use of contrast, has shown lower T2* values
angiography need to account for rapid technological in segments supplied by the stenotic artery.113
advances in this area, and many patients must be
recruited to have any effect on day-to-day clinical Myocardial viability
practice. Available data suggest that CMRA might be a Viable myocytes might be present in severely
non-invasive method for assessment of the proximal few dysfunctional myocardium because of an acute ischaemic
centimetres of coronary arteries in ischaemic heart insult (stunning) or in areas of long-term downregulated
disease. metabolism due to chronic ischaemia (hibernation).
Coronary-artery bypass grafts are fairly easy to image Detection of myocardial viability is important to predict
because of their fixed position and large lumen size potential for functional recovery and, therefore, overall
(especially vein grafts). However, presence of sternal risk and prognosis.
metal clips and markers used at surgery can cause Various variables, derived from resting and functional
artifacts. Findings of early MRI studies with low spatial CMRI, can be used as markers of myocardial viability,
resolution have shown specificities of 85–100% and including end-diastolic wall thickness, systolic wall
sensitivities of 88–93%.98–100 In a study of high resolution thickening, and signal intensity with and without contrast-
three-dimensional CMRA, sensitivity and specificity for enhancement (figure 6). These characteristics alone and
detection of graft occlusion and stenosis of 70% or more in combination have been investigated to predict presence
were 83% and 99% and 73% and 83%, respectively.101 of viable tissue.114–116
CMRA is a useful investigation in non-invasive Early detection of the transmural extent of infarction is
detection of anomalous coronary arteries (figure 5). The the cornerstone of myocardial salvage in patients with
clinical importance of an anomalous coronary artery acute myocardial infarction. CMRI enables detection of
depends on its origin and initial course, which is infarct tissue by delayed contrast enhancement
sometimes difficult to delineate by conventional techniques.114,117 An association between transmural extent
angiography.102–104 detected by magnetic resonance and functional recovery

Coronary flow reserve and myocardial

perfusion
The physiological relevance of a
coronary stenosis can be established by
the ratio of hyperaemic to baseline
flow, known as the coronary flow
reserve. Many studies have validated
use of CMRI techniques to assess
coronary flow and coronary flow
reserve in native and graft vessels.
CMRI can evaluate coronary flow
reserve by assessment of myocardial
perfusion with first-pass contrast
enhanced imaging or by determining
coronary artery blood flow via fast
velocity encoded cine MRI before and
after pharmacologically induced
hyperaemia.105–108 Findings of compar-
ative coronary flow reserve studies
have shown close agreement between
velocity encoded cine MRI and
radioisotope studies.109,110
Myocardial perfusion and perfusion
reserve can be used as indirect
measures of coronary flow. CMRI
first-pass perfusion has been reported
to have high diagnostic accuracy for
detection of coronary artery disease.
Myocardial perfusion reserve can be
established by assessment of signal Figure 6: MRI in myocardial infarction
intensity–time curves, with sensitivity, Gradient-echo four-chamber images in diastole (A) and systole (B) from a patient with recent
specificity, and diagnostic accuracy of anterolateral myocardial infarction. Systolic wall thickening is only evident in posterolateral wall, while
most of the septum and apex are akinetic with wall thinning (arrowheads). Left ventricular (LV) short
90%, 83%, and 87%, respectively, for axis (C) and four-chamber (D) from contrast delayed-enhancement sequences show transmural high
detection of coronary artery stenosis signal intensity in parts of the septum (thick arrow) and the apex, and subendocardial high signal
(⭓75%), in a one-slice method.111 intensity in anterior and lateral wall (thin arrows). LA=left atrium. RV=right ventricle.

THE LANCET • Vol 363 • June 26, 2004 • www.thelancet.com 2167

For personal use. Only reproduce with permission from The Lancet Publishing Group.
REVIEW

after revascularisation has been shown. Most (77%) its fibrous content, bordered by high signal intensities of
dysfunctional non-infarcted myocardium and segments epicardial and pericardial fat. Pericardial thickness of up
with small transmural infarction (1–25% of the thickness) to 4 mm is judged normal.123
improve function on revascularisation (77% and 67%, Congenital pericardial defects, tumours, effusions, and
respectively). In a study,114 a recovery rate of 35% was constrictive pericarditis are diagnosed precisely with
noted in areas of transmural infarction involving 51–75% CMRI. Localised pericardial thickening and encysted
of the thickness after revascularisation, which fell to 5% pericardial effusions are visualised well.124 Appearance of
with transmural extent involving 76–100%. pericardial fluid on CMRI depends on the nature of the
In chronic coronary artery disease, a combination of effusion.125 Transudates appear as low signal on
fibrosis and viable non-functional myocardium exists. T1-weighted images and high signal on T2-weighted and
Recovery of impaired left ventricular function depends on gradient-echo images. Exudates have an intermediate
extent of fibrosis. In a study,118 ventricular contractility signal on both sequences. Haemorrhagic effusions show
improved in 78% of dysfunctional segments with no various signal intensities on spin-echo sequences, from
hyper-enhancement compared with less than 2% of those low to high depending on the age of the effusion.
with hyper-enhancement of more than 75% of tissue.
Mean transmural extent of hyper-enhancement was 10% Interventional MRI
(SD 7) for the group with improved contractility and 41% Ultrafast imaging techniques are being investigated to
(14) when there was no improvement.118 extend the potential role of CMRI from diagnostic use to
Functional CMRI, which applies the same principles as an alternative real-time technique to conventional
stress echocardiography, can also be used in assessment of fluoroscopy. Magnetic resonance fluoroscopy is being
viability. Diastolic wall thickness, endocardial motion, investigated in interventional cardiology and related areas
and systolic wall thickening are used to evaluate response with transcatheter techniques, because CMRI has the
to dobutamine stress. Diastolic wall thickness of 5·5 mm advantage of three-dimensional imaging with high spatial
or more and systolic wall thickening of 2 mm or more resolution and good soft-tissue contrast. Furthermore, it
have been shown to predict reversibility of contractile might provide pathological information such as plaque
function after revascularisation. Stress-induced systolic characteristics, which could modify therapeutic
wall thickening and diastolic wall thickness had sensitivity procedures.126–129
and specificity of 89% and 94% and 92% and 56%, Further developments with MRI-compatible devices
respectively.119,120 such as intravascular coils, guidewires, and catheters are
CMRI offers potential advantages over radioisotope needed before interventional MRI is accepted for a wider
techniques in assessment of myocardial viability because it clinical role.130
provides superior spatial resolution, does not use ionising
radiation, and has no constraints in visualising the entire Future
heart. Introduction of parallel acquisition methods that allow
simultaneous generation of multiple image sets will make
One-stop assessment of coronary artery disease near real-time high-resolution MRI possible.131,132 Several
Comprehensive assessment of sequelae of coronary artery other pulse sequences, such as spiral imaging, multiecho
disease includes outlining of remodelled anatomy and imaging, wavelet-encoded data acquisition, keyhole
function of the ventricles, assessment of myocardial imaging, etc, achieve faster image acquisition times with
perfusion at rest and after stress, detection of myocardial some compromise in spatial resolution, but show promise
viability, and definition of the coronary anatomy and in magnetic resonance fluoroscopy.133–136
valvular pathology. This plethora of information allows CMRI is a potential non-invasive imaging method for
informed decision making in patients’ management. atherosclerotic plaque characterisation. Fayad and
CMRI has been proposed as a non-invasive modality that colleagues,126,137 with high-resolution black-blood magnetic
could provide a one-stop evaluation of coronary artery resonance technique, showed the possibility of assessment
disease. of morphological characteristics of coronary arteries.
Few integrated protocols have been proposed for Other magnetic resonance techniques for assessment of
comprehensive assessment of coronary artery disease.121,122 the coronary arterial wall have also been explored.138
Plein and co-workers122 have proposed a CMRI protocol Cardiac magnetic resonance spectroscopy is an evolving
integrating multiphase gradient-echo cine magnetic technique. It can detect abnormalities in myocardial
resonance for function, first-pass myocardial perfusion metabolic variables and can risk-stratify patients with
imaging at rest and during adenosine stress, delayed cardiomyopathies with potential for use in other
contrast enhancement for assessment of myocardial disorders.139,140
viability, and three-dimensional respiratory navigator- The role of CMRI in day-to-day clinical practice has
gated CMRA.122 The entire protocol has been undertaken been limited more by absence of trained personnel than
in about 1 h. A similar integrated CMRI protocol has by availability of suitable magnetic resonance scanners. As
been tried in risk assessment of patients presenting with proficiency in CMRI encompasses knowledge of cross-
chest pain at the emergency department with acute sectional imaging, physics, and understanding of unique
coronary syndrome, with sensitivity and specificity of and complex morphology and pathophysiology of the
85%.5,116 cardiovascular system, well directed, policed, and
If comprehensive assessment of coronary artery disease accredited training programmes are necessary. These
by CMRI becomes a clinical reality, this modality will be courses will ensure CMRI matures into a robust clinical
one of the most important investigations in ischaemic imaging technique with widespread use, superior to
heart disease. traditional modalities used in many disorders.

Pericardial disease Conflict of interest statement

GC and KS declare no conflicts of interest. SDF declares research
CMRI is the investigation of choice in diagnosis of collaboration with or grant support from the following commercial
pericardial disease. On T1-weighted spin-echo imaging, institutions: Philips Medical Systems, Siemens Medical Solutions, Tyco
pericardium appears as a thin low-signal band because of Healthcare/Mallinckrodt, Amersham Health, and Bracco Diagnostics.

2168 THE LANCET • Vol 363 • June 26, 2004 • www.thelancet.com

For personal use. Only reproduce with permission from The Lancet Publishing Group.
 REVIEW

MUS declares research collaboration with or grant support from the 22 Mathieu D, Keita K, Loisance D, Cachera JP, Rousseau M,
following commercial institutions: Philips Medical Systems, Guerbet Vasile N. Postoperative CT follow-up of aortic dissection.
Medical, Pfizer Pharmaceuticals, Medtronic, and Centocor. Neither SDF J Comput Assist Tomogr 1986; 10: 216–18.
nor MUS received substantial personal or financial benefit from any of 23 Pucillo AL, Schechter AG, Moggio RA, Kay RH, Tenner MS,
these relations. Herman MV. Postoperative evaluation of ascending aortic
prosthetic conduits by magnetic resonance imaging. Chest 1990;
Acknowledgments 97: 106–10.
We thank all medical and technical staff in our units for help in 24 Dinsmore RE, Liberthson RR, Wismer GL, et al. Magnetic
preparation of this manuscript; John Ridgway for review of technical resonance imaging of thoracic aortic aneurysms: comparison with
content; and Scott Reid and Barbara Corney for help in formatting the other imaging methods. AJR Am J Roentgenol 1986; 146: 309–14.
text. 25 Fattori R, Bacchi Reggiani L, Pepe G, et al. Magnetic resonance
imaging evaluation of aortic elastic properties as early expression of
Marfan syndrome. J Cardiovasc Magn Reson 2000; 2: 251–56.
26 Metafratzi ZM, Efremidis SC, Skopelitou AS, De Roos A. The
References clinical significance of aortic compliance and its assessment with
1 McRobbie DW, Moore EA, Graves MJ, Prince MR. MRI from magnetic resonance imaging. J Cardiovasc Magn Reson 2002; 4:
picture to proton. Cambridge: Cambridge University Press, 2003. 481–91.
2 Elster AD, Burdette JH. Questions and answers in MRI. London: 27 Kawamoto S, Bluemke DA, Traill TA, Zerhouni EA.
Mosby, 2001. Thoracoabdominal aorta in Marfan syndrome: MR imaging findings
3 Kanal E, Borgstede JP, Barkovich AJ, et al. American College of of progression of vasculopathy after surgical repair. Radiology 1997;
Radiology white paper on MR safety. AJR Am J Roentgenol 2002; 203: 727–32.
178: 1335–47. 28 Yucel EK, Steinberg FL, Egglin TK, Geller SC, Waltman AC,
4 Draud KS. Magnetic resonance imaging bioeffects and safety: a Athanasoulis CA. Penetrating aortic ulcers: diagnosis with MR
review. Appl Radiol 2002; 31 (12 suppl A): 14–21. imaging. Radiology 1990; 177: 779–81.
5 Kwong RY, Schussheim AE, Rekhraj S, et al. Detecting acute 29 Arai A, Epstein FH, Bove KE, Wolff SD. Visualization of aortic valve
coronary syndrome in the emergency department with cardiac leaflets using black blood MRI. J Magn Reson Imaging 1999; 10:
magnetic resonance imaging. Circulation 2003; 107: 531–37. 771–77.
6 Holland GA, Dougherty L, Carpenter JP, et al. Breath-hold 30 Higgins CB, Wagner S, Kondo C, Suzuki J, Caputo GR. Evaluation
ultrafast three-dimensional gadolinium-enhanced MR angiography of valvular heart disease with cine gradient echo magnetic resonance
of the aorta and the renal and other visceral abdominal arteries. imaging. Circulation 1991; 84 (3 suppl): I198–207.
AJR Am J Roentgenol 1996; 166: 971–81. 31 Dulce MC, Mostbeck GH, O’Sullivan M, Cheitlin M, Caputo GR,
7 Bottini PB, Carr AA, Prisant LM, Flickinger FW, Allison JD, Higgins CB. Severity of aortic regurgitation: interstudy reproducibility
Gottdiener JS. Magnetic resonance imaging compared to of measurements with velocity-encoded cine MR imaging. Radiology
echocardiography to assess left ventricular mass in the hypertensive 1992; 185: 235–40.
patient. Am J Hypertens 1995; 8: 221–28. 32 Honda N, Machida K, Hashimoto M, et al. Aortic regurgitation:
8 Myerson SG, Montgomery HE, World MJ, Pennell DJ. Left quantitation with MR imaging velocity mapping. Radiology 1993; 186:
ventricular mass: reliability of M-mode and 2-dimensional 189–94.
echocardiographic formulas. Hypertension 2002; 40: 673–78. 33 Deutsch HJ, Bachmann R, Sechtem U, et al. Regurgitant flow in
9 Bellenger NG, Burgess MI, Ray SG, et al. Comparison of left cardiac valve prostheses: diagnostic value of gradient echo nuclear
ventricular ejection fraction and volumes in heart failure by magnetic resonance imaging in reference to transesophageal two-
echocardiography, radionuclide ventriculography and cardiovascular dimensional color Doppler echocardiography. J Am Coll Cardiol 1992;
magnetic resonance: are they interchangeable. Eur Heart J 2000; 21: 19: 1500–07.
1387–96. 34 Kizilbash AM, Hundley WG, Willett DL, Franco F, Peshock RM,
10 Petrank YF, Dong SJ, Tyberg J, Sideman S, Beyar R. Regional Grayburn PA. Comparison of quantitative Doppler with magnetic
differences in shape and load in normal and diseased hearts studied resonance imaging for assessment of the severity of mitral
by three dimensional tagged magnetic resonance imaging. regurgitation. Am J Cardiol 1998; 81: 792–95.
Int J Card Imaging 1999; 15: 309–21. 35 Hundley WG, Li HF, Willard JE, et al. Magnetic resonance imaging
11 Bellenger NG, Davies LC, Francis JM, Coats AJ, Pennell DJ. assessment of the severity of mitral regurgitation: comparison with
Reduction in sample size for studies of remodeling in heart failure invasive techniques. Circulation 1995; 92: 1151–58.
by the use of cardiovascular magnetic resonance. 36 Suzuki J, Caputo GR, Kondo C, Higgins CB. Cine MR imaging of
J Cardiovasc Magn Reson 2000; 2: 271–78. valvular heart disease: display and imaging parameters affect the size
12 Lorenz CH, Pastorek JS, Bundy JM. Delineation of normal of the signal void caused by valvular regurgitation.
human left ventricular twist throughout systole by tagged cine AJR Am J Roentgenol 1990; 155: 723–27.
magnetic resonance imaging. J Cardiovasc Magn Reson 2000; 37 Eichenberger AC, Jenni R, von Schulthess GK. Aortic valve pressure
2: 97–108. gradients in patients with aortic valve stenosis: quantification with
13 Underwood SR, Firmin DN, Klipstein RH, Rees RS, Longmore DB. velocity-encoded cine MR imaging. AJR Am J Roentgenol 1993; 160:
Magnetic resonance velocity mapping: clinical applications of a new 971–77.
technique. Br Heart J 1987; 57: 404–12. 38 de Roos A, Roest AA. Evaluation of congenital heart disease by
14 Meijboom LJ, Groenink M, van der Wall EE, Romkes H, Stoker J, magnetic resonance imaging. Eur Radiol 2000; 10: 2–6.
Mulder BJ. Aortic root asymmetry in Marfan patients: evaluation by 39 Higgins CB, Sakuma H. Heart disease: functional evaluation with
magnetic resonance imaging and comparison with standard MR imaging. Radiology 1996; 199: 307–15.
echocardiography. Int J Card Imaging 2000; 16: 161–68. 40 Kersting-Sommerhoff BA, Diethelm L, Stanger P, et al. Evaluation of
15 Sawhney, NS, DeMaria, AN, Blanchard, DG. Aortic intramural complex congenital ventricular anomalies with magnetic resonance
hematoma: an increasingly recognized and potentially fatal entity. imaging. Am Heart J 1990; 120: 133–42.
Chest 2001; 120: 1340–46. 41 Hoppe UC, Dederichs B, Deutsch HJ, Theissen P, Schicha H,
16 Nienaber CA, von Kodolitsch Y, Nicolas V, et al. The diagnosis of Sechtem U. Congenital heart disease in adults and adolescents:
thoracic aortic dissection by noninvasive imaging procedures. comparative value of transthoracic and transesophageal
N Engl J Med 1993; 328: 1–9. echocardiography and MR imaging. Radiology 1996; 199: 669–77.
17 Bansal RC, Chandrasekaran K, Ayala K, Smith DC. Frequency and 42 Keller PJ, Drayer BP, Fram EK, Williams KD, Dumoulin CL,
explanation of false negative diagnosis of aortic dissection by Souza SP. MR angiography with two-dimensional acquisition and
aortography and transesophageal echocardiography. J Am Coll Cardiol three-dimensional display: work in progress. Radiology 1989; 173:
1995; 25: 1393–401. 527–32.
18 Nienaber CA, Spielmann RP, von Kodolitsch Y, et al. Diagnosis of 43 Ladd ME, Gohde SC, Steiner P, Pfammatter T, McKinnon GC,
thoracic aortic dissection: magnetic resonance imaging versus Debatin JF. Virtual MR angioscopy of the pulmonary artery tree.
transesophageal echocardiography. Circulation 1992; 85: 434–47. J Comput Assist Tomogr 1996; 20: 782–85.
19 Khan IA, Nair CK. Clinical, diagnostic, and management 44 Vick GW III. Three- and four-dimensional visualization of magnetic
perspectives of aortic dissection. Chest 2002; 122: 311–28. resonance imaging data sets in pediatric cardiology. Pediatr Cardiol
20 von Kodolitsch Y, Csosz SK, Koschyk DH, et al. Intramural 2000; 21: 27–36.
hematoma of the aorta: predictors of progression to dissection and 45 Simpson IA, Chung KJ, Glass RF, Sahn DJ, Sherman FS,
rupture. Circulation 2003; 107: 1158–63. Hesselink J. Cine magnetic resonance imaging for evaluation of
21 Schepens MA, Dossche KM, Morshuis WJ. Reoperations on the anatomy and flow relations in infants and children with coarctation of
ascending aorta and aortic root: pitfalls and results in 134 patients. the aorta. Circulation 1988; 78: 142–48.
Ann Thorac Surg 1999; 68: 1676–80. 46 Oshinski JN, Parks WJ, Markou CP, et al. Improved measurement of

THE LANCET • Vol 363 • June 26, 2004 • www.thelancet.com 2169

For personal use. Only reproduce with permission from The Lancet Publishing Group.
REVIEW

pressure gradients in aortic coarctation by magnetic resonance hypertrophic cardiomyopathy and the Asp175Asn mutation of the
imaging. J Am Coll Cardiol 1996; 28: 1818–26. alpha-tropomyosin gene. Radiology 2003; 226: 129–37.
47 Mohiaddin RH, Kilner PJ, Rees S, Longmore DB. Magnetic 70 Tanaka M, Fujiwara H, Onodera T, Wu DJ, Hamashima Y,
resonance volume flow and jet velocity mapping in aortic coarctation. Kawai C. Quantitative analysis of myocardial fibrosis in normals,
J Am Coll Cardiol 1993; 22: 1515–21. hypertensive hearts, and hypertrophic cardiomyopathy.
48 Steffens JC, Bourne MW, Sakuma H, O’Sullivan M, Higgins CB. Br Heart J 1986; 55: 575–81.
Quantification of collateral blood flow in coarctation of the aorta by 71 Kayser HW, van der Wall EE, Sivananthan MU, Plein S,
velocity encoded cine magnetic resonance imaging. Circulation 1994; Bloomer TN, de Roos A. Diagnosis of arrhythmogenic right
90: 937–43. ventricular dysplasia: a review. Radiographics 2002; 22: 639–48.
49 Manganas C, Iliopoulos J, Chard RB, Nunn GR. Reoperation and 72 Bluemke DA, Krupinski EA, Ovitt T, et al. MR imaging of
coarctation of the aorta: the need for lifelong surveillance. arrhythmogenic right ventricular cardiomyopathy: morphologic
Ann Thorac Surg 2001; 72: 1222–24. findings and interobserver reliability. Cardiology 2003; 99: 153–62.
50 Riquelme C, Laissy JP, Menegazzo D, et al. MR imaging of 73 Ricci C, Longo R, Pagnan L, et al. Magnetic resonance imaging in
coarctation of the aorta and its postoperative complications in adults: right ventricular dysplasia. Am J Cardiol 1992; 70: 1589–95.
assessment with spin-echo and cine-MR imaging. Magn Reson 74 Hamid MS, Norman M, Quraishi A, et al. Prospective evaluation of
Imaging 1999; 17: 37–46. relatives for familial arrhythmogenic right ventricular
51 Therrien J, Thorne SA, Wright A, Kilner PJ, Somerville J. Repaired cardiomyopathy/dysplasia reveals a need to broaden diagnostic
coarctation: a “cost-effective” approach to identify complications in criteria. J Am Coll Cardiol 2002; 40: 1445–50.
adults. J Am Coll Cardiol 2000; 35: 997–1002. 75 Young AA, Dokos S, Powell KA, et al. Regional heterogeneity of
52 Helbing WA, Bosch HG, Maliepaard C, et al. Comparison of function in nonischemic dilated cardiomyopathy. Cardiovasc Res
echocardiographic methods with magnetic resonance imaging for 2001; 49: 308–18.
assessment of right ventricular function in children. Am J Cardiol 76 Laissy JP, Messin B, Varenne O, et al. MRI of acute myocarditis: a
1995; 76: 589–94. comprehensive approach based on various imaging sequences.
53 Helbing WA, Rebergen SA, Maliepaard C, et al. Quantification of Chest 2002; 122: 1638–48.
right ventricular function with magnetic resonance imaging in 77 Masui T, Finck S, Higgins CB. Constrictive pericarditis and
children with normal hearts and with congenital heart disease. restrictive cardiomyopathy: evaluation with MR imaging. Radiology
Am Heart J 1995; 130: 828–37. 1992; 182: 369–73.
54 Vick GW 3rd, Rokey R, Huhta JC, Mulvagh SL, Johnston DL. 78 Sechtem U, Higgins CB, Sommerhoff BA, Lipton MJ, Huycke EC.
Nuclear magnetic resonance imaging of the pulmonary arteries, Magnetic resonance imaging of restrictive cardiomyopathy.
subpulmonary region, and aorticopulmonary shunts: a comparative Am J Cardiol 1987; 59: 480–82.
study with two-dimensional echocardiography and angiography. 79 Hancock EW. Differential diagnosis of restrictive cardiomyopathy and
Am Heart J 1990; 119: 1103–10.
constrictive pericarditis. Heart 2001; 86: 343–49.
55 Beekman RP, Beek FJ, Meijboom EJ. Usefulness of MRI for the pre-
80 Matsuki M, Matsuo M. MR findings of myocardial sarcoidosis.
operative evaluation of the pulmonary arteries in Tetralogy of Fallot.
Clin Radiol 2000; 55: 323–25.
Magn Reson Imaging 1997; 15: 1005–15.
81 Vignaux O, Dhote R, Duboc D, et al. Clinical significance of
56 Kersting-Sommerhoff BA, Seelos KC, Hardy C, Kondo C,
myocardial magnetic resonance abnormalities in patients with
Higgins SS, Higgins CB. Evaluation of surgical procedures for
sarcoidosis: a 1-year follow-up study. Chest 2002; 122: 1895–901.
cyanotic congenital heart disease by using MR imaging.
AJR Am J Roentgenol 1990; 155: 259–66. 82 Shimada J, Shimada K, Sakane T, et al. Diagnosis of sarcoidosis and
evaluation of the effects of steroid therapy by gadolinium–DTPA-
57 Hardy CE, Helton GJ, Kondo C, Higgins SS, Young NJ,
enhanced magnetic resonance imaging. Am J Med 2001; 110: 520–27.
Higgins CB. Usefulness of magnetic resonance imaging for
evaluating great-vessel anatomy after arterial switch operation 83 Fattori R, Rocchi G, Celletti F, Bertaccini P, Rapezzi C, Gavelli G.
for D-transposition of the great arteries. Am Heart J 1994; 128: Contribution of magnetic resonance imaging in the differential
326–32. diagnosis of cardiac amyloidosis and symmetric hypertrophic
cardiomyopathy. Am Heart J 1998; 136: 824–30.
58 Duerinckx A, Atkinson D, Klitzner TS, Perloff J, Drinkwater D,
Laks H. MR imaging of surgical complications of systemic-to- 84 Anderson LJ, Holden S, Davis B, et al. Cardiovascular T2-star (T2*)
pulmonary artery shunts. Magn Reson Imaging 1996; 14: 1099–105. magnetic resonance for the early diagnosis of myocardial iron
overload. Eur Heart J 2001; 22: 2171–79.
59 Fogel MA, Hubbard A, Weinberg PM. A simplified approach for
assessment of intracardiac baffles and extracardiac conduits in 85 Waxman S, Eustace S, Hartnell GG. Myocardial involvement in
congenital heart surgery with two- and three-dimensional magnetic primary hemochromatosis demonstrated by magnetic resonance
resonance imaging. Am Heart J 2001; 142: 1028–36. imaging. Am Heart J 1994; 128: 1047–49.
60 Helbing WA, de Roos A. Clinical applications of cardiac magnetic 86 Freedberg RS, Kronzon I, Rumancik WM, Liebeskind D.
resonance imaging after repair of tetralogy of Fallot. Pediatr Cardiol The contribution of magnetic resonance imaging to the evaluation of
2000; 21: 70–79. intracardiac tumors diagnosed by echocardiography. Circulation 1988;
61 Park JH, Kim YM, Chung JW, Park YB, Han JK, Han MC. 77: 96–103.
MR imaging of hypertrophic cardiomyopathy. Radiology 1992; 185: 87 Araoz PA , Mulvagh SL, Tazelaar HD, Julsrud PR, Breen JF. CT and
441–46. MR imaging of benign primary cardiac neoplasms with
62 Sardanelli F, Molinari G, Petillo A, et al. MRI in hypertrophic echocardiographic correlation. Radiographics 2000; 20: 1303–19.
cardiomyopathy: a morphofunctional study. J Comput Assist Tomogr 88 Gatehouse PD, Bydder GM. Magnetic resonance imaging of short T2
1993; 17: 862–72. components in tissue. Clin Radiol 2003; 58: 1–19.
63 Sharma S, Maron BJ, Whyte G, Firoozi S, Elliott PM, McKenna WJ. 89 Gomes AS, Lois JF, Child JS, Brown K, Batra P. Cardiac tumors and
Physiologic limits of left ventricular hypertrophy in elite junior thrombus: evaluation with MR imaging. AJR Am J Roentgenol 1987;
athletes: relevance to differential diagnosis of athlete’s heart and 149: 895–99.
hypertrophic cardiomyopathy. J Am Coll Cardiol 2002; 40: 1431–36. 90 Funari M, Fujita N, Peck WW, Higgins CB. Cardiac tumors:
64 Maron BJ, Pelliccia A, Spirito P. Cardiac disease in young trained assessment with Gd-DTPA enhanced MR imaging.
athletes: insights into methods for distinguishing athlete’s heart from J Comput Assist Tomogr 1991; 15: 953–58.
structural heart disease, with particular emphasis on hypertrophic 91 Siripornpitak S, Higgins CB. MRI of primary malignant
cardiomyopathy. Circulation 1995; 91: 1596–601. cardiovascular tumors. J Comput Assist Tomogr 1997; 21: 462–66.
65 Young AA, Kramer CM, Ferrari VA, Axel L, Reichek N. Three- 92 Araoz PA, Eklund HE, Welch TJ, Breen JF. CT and MR imaging of
dimensional left ventricular deformation in hypertrophic primary cardiac malignancies. Radiographics 1999; 19: 1421–34.
cardiomyopathy. Circulation 1994; 90: 854–67. 93 Manning WJ, Li W, Edelman RR. A preliminary report comparing
66 Maier SE, Fischer SE, McKinnon GC, Hess OM, Krayenbuehl HP, magnetic resonance coronary angiography with conventional
Boesiger P. Evaluation of left ventricular segmental wall motion in angiography. N Engl J Med 1993; 328: 828–32.
hypertrophic cardiomyopathy with myocardial tagging. Circulation 94 Post JC, van Rossum AC, Hofman MB, Valk J, Visser CA. Three-
1992; 86: 1919–28. dimensional respiratory-gated MR angiography of coronary arteries:
67 Beache GM, Wedeen VJ, Weisskoff RM, et al. Intramural mechanics comparison with conventional coronary angiography.
in hypertrophic cardiomyopathy: functional mapping with strain-rate AJR Am J Roentgenol 1996; 166: 1399–404.
MR imaging. Radiology 1995; 197: 117–24. 95 van Geuns RJ, Wielopolski PA, de Bruin HG, et al. MR coronary
68 Choudhury L, Mahrholdt H, Wagner A, et al. Myocardial scarring in angiography with breath-hold targeted volumes: preliminary clinical
asymptomatic or mildly symptomatic patients with hypertrophic results. Radiology 2000; 217: 270–77.
cardiomyopathy. J Am Coll Cardiol 2002; 40: 2156–64. 96 Huber A, Nikolaou K, Gonschior P, Knez A, Stehling M, Reiser M.
69 Sipola P, Lauerma K, Husso-Saastamoinen M, et al. First-pass MR Navigator echo-based respiratory gating for three-dimensional MR
imaging in the assessment of perfusion impairment in patients with coronary angiography: results from healthy volunteers and patients

2170 THE LANCET • Vol 363 • June 26, 2004 • www.thelancet.com

For personal use. Only reproduce with permission from The Lancet Publishing Group.
 REVIEW

with proximal coronary artery stenoses. AJR Am J Roentgenol 1999; with non-ST segment-elevation acute coronary syndromes: feasibility
173: 95–101. study. Radiology 2003; 226: 717–22.
97 Kim WY, Danias PG, Stuber M, et al. Coronary magnetic resonance 117 Kim RJ, Fieno DS, Parrish TB, et al. Relationship of MRI delayed
angiography for the detection of coronary stenoses. N Engl J Med contrast enhancement to irreversible injury, infarct age, and
2001; 345: 1863–69. contractile function. Circulation 1999; 100: 1992–2002.
98 Rubinstein RI, Askenase AD, Thickman D, Feldman MS, 118 Kim RJ, Wu E, Rafael A, et al. The use of contrast-enhanced
Agarwal JB, Helfant RH. Magnetic resonance imaging to evaluate magnetic resonance imaging to identify reversible myocardial
patency of aortocoronary bypass grafts. Circulation 1987; 76: 786–91. dysfunction. N Engl J Med 2000; 343: 1445–53.
99 White RD, Pflugfelder PW, Lipton MJ, Higgins CB. Coronary artery 119 Baer FM, Theissen P, Schneider CA, et al. Dobutamine magnetic
bypass grafts: evaluation of patency with cine MR imaging. resonance imaging predicts contractile recovery of chronically
AJR Am J Roentgenol 1988; 150: 1271–74. dysfunctional myocardium after successful revascularization.
100 Aurigemma GP, Reichek N, Axel L, Schiebler M, Harris C, J Am Coll Cardiol 1998; 31: 1040–48.
Kressel HY. Noninvasive determination of coronary artery bypass 120 Baer FM, Voth E, LaRosee K, et al. Comparison of dobutamine
graft patency by cine magnetic resonance imaging. Circulation 1989; transesophageal echocardiography and dobutamine magnetic
80: 1595–602. resonance imaging for detection of residual myocardial viability.
101 Langerak SE, Vliegen HW, de Roos A, et al. Detection of vein graft Am J Cardiol 1996; 78: 415–19.
disease using high-resolution magnetic resonance angiography. 121 Sensky PR, Jivan A, Hudson NM, et al. Coronary artery disease:
Circulation 2002; 105: 328–33. combined stress MR imaging protocol-one-stop evaluation of
102 Page HL, Engel HJ, Campbell WB, Thomas CS Jr. Anomalous myocardial perfusion and function. Radiology 2000; 215: 608–14.
origin of the left circumflex coronary artery: Recognition, 122 Plein S, Ridgway JP, Jones TR, Bloomer TN, Sivananthan UM.
angiographic demonstration and clinical significance. Circulation Coronary artery disease: assessment with a comprehensive MR
1974; 50: 768–73. imaging protocol—initial results. Radiology 2002; 225: 300–07.
103 McConnell MV, Ganz P, Selwyn AP, Li W, Edelman RR, 123 Smith WH, Beacock DJ, Goddard AJ, Bloomer TN, Ridgway JP,
Manning WJ. Identification of anomalous coronary arteries and their Sivananthan UM. Magnetic resonance evaluation of the pericardium.
anatomic course by magnetic resonance coronary angiography. Br J Radiol 2001; 74: 384–92.
Circulation 1995; 92: 3158–62. 124 Breen JF. Imaging of the pericardium. J Thorac Imaging 2001; 16:
104 Post JC, van Rossum AC, Bronzwaer JG, et al. Magnetic resonance 47–54.
angiography of anomalous coronary arteries: a new gold standard for 125 Kovanlikaya A, Burke LP, Nelson MD, Wood J. Characterizing
delineating the proximal course? Circulation 1995; 92: 3163–71. chronic pericarditis using steady-state free-precession cine MR
105 Saeed M, Wendland MF, Lauerma K, et al. First pass contrast imaging. AJR Am J Roentgenol 2002; 179: 475–76.
enhanced inversion recovery and driven equilibrium fast GRE 126 Fayad ZA, Fuster V, Nikolaou K, Becker C. Computed tomography
imaging studies: detection of acute myocardial ischemia. and magnetic resonance imaging for noninvasive coronary
J Magn Reson Imaging 1995; 5: 515–23. angiography and plaque imaging: current and potential future
106 Sakuma H, Koskenvuo JW, Niemi P, et al. Assessment of coronary concepts. Circulation 2002; 106: 2026–34.
flow reserve using fast velocity-encoded cine MR imaging: validation 127 Spuentrup E, Ruebben A, Schaeffter T, Manning WJ, Gunther RW,
study using positron emission tomography. AJR Am J Roentgenol Buecker A. Magnetic resonance-guided coronary artery stent
2000; 175: 1029–33. placement in a swine model. Circulation 2002; 105: 874–79.
107 Clarke GD, Eckels R, Chaney C, et al. Measurement of absolute 128 Rickers C, Jerosch-Herold M, Hu X, et al. Magnetic resonance
epicardial coronary artery flow and flow reserve with breath-hold cine image-guided transcatheter closure of atrial septal defects. Circulation
phase-contrast magnetic resonance imaging. Circulation 1995; 91: 2003; 107: 132–38.
2627–34. 129 Lardo AC, McVeigh ER, Jumrussirikul P, et al. Visualization and
108 Sakuma H, Saeed M, Takeda K, et al. Quantification of coronary temporal/spatial characterization of cardiac radiofrequency ablation
artery volume flow rate using fast velocity-encoded cine MR imaging. lesions using magnetic resonance imaging. Circulation 2000; 102:
AJR Am J Roentgenol 1997; 168: 1363–67. 698–705.
109 Koskenvuo JW, Sakuma H, Niemi P, et al. Global myocardial blood 130 Lenfant C. Report of the task force on research in pediatric
flow and global flow reserve measurements by MRI and PET are cardiovascular disease. Circulation 2002; 106: 1037–42.
comparable. J Magn Reson Imaging 2001; 13: 361–66. 131 Pruessmann KP, Weiger M, Scheidegger MB, Boesiger P. SENSE:
110 Keijer JT, Van Rossum AC, Van Eenige MJ, et al. Magnetic sensitivity encoding for fast MRI. Magn Reson Med 1999; 42: 952–62.
resonance imaging of regional myocardial perfusion in patients with 132 Sodickson DK. Tailored SMASH image reconstructions for robust in
single-vessel coronary artery disease: quantitative comparison with vivo parallel MR imaging. Magn Reson Med 2000; 44: 243–51.
(201) Thallium-SPECT and coronary angiography. 133 Meyer CH, Hu BS, Nishimura DG, Macovski A. Fast spiral coronary
J Magn Reson Imaging 2000; 11: 607–15. artery imaging. Magn Reson Med 1992; 28: 202–13.
111 Al-Saadi N, Nagel E, Gross M, et al. Noninvasive detection of 134 Reeder SB, Atalar E, Faranesh AZ, McVeigh ER. Multi-echo
myocardial ischemia from perfusion reserve based on cardiovascular segmented k-space imaging: an optimized hybrid sequence for
magnetic resonance. Circulation 2000; 101: 1379–83. ultrafast cardiac imaging. Magn Reson Med 1999; 41: 375–85.
112 Al-Saadi N, Nagel E, Gross M, et al. Improvement of myocardial 135 Weaver JB, Xu Y, Healy DM, Discoll JR. Wavelet-encoded MR
perfusion reserve early after coronary intervention: assessment with imaging. Magn Reson Med 1992; 24: 275–87.
cardiac magnetic resonance imaging. J Am Coll Cardiol 2000; 36: 136 van Vaals JJ, Brummer ME, Dixon WT, et al. “Key-hole” method for
1557–64. accelerating imaging of contrast agent uptake. J Magn Reson Imaging
113 Wacker CM, Hartlep AW, Pfleger S, Schad LR, Ertl G, Bauer WR. 1993; 3: 671–75.
Susceptibility-sensitive magnetic resonance imaging detects human 137 Fayad ZA, Fuster V, Fallon JT, et al. Noninvasive In vivo human
myocardium supplied by a stenotic coronary artery without a coronary artery lumen and wall imaging using black-blood magnetic
contrast agent. J Am Coll Cardiol 2003; 41: 834–40. resonance imaging. Circulation 2000; 102: 506–10.
114 Choi KM, Kim RJ, Gubernikoff G, Vargas JD, Parker M, Judd RM. 138 Botnar RM, Kim WY, Bornert P, Stuber M, Spuentrup E,
Transmural extent of acute myocardial infarction predicts long-term Manning WJ. 3D coronary vessel wall imaging utilizing a local
improvement in contractile function. Circulation 2001; 104: 1101–07. inversion technique with spiral image acquisition. Magn Reson Med
115 Lauerma K, Niemi P, Hanninen H, et al. Multimodality MR 2001; 46: 848–54.
imaging assessment of myocardial viability: combination of first-pass 139 Schaefer S. Magnetic resonance spectroscopy in human
and late contrast enhancement to wall motion dynamics and cardiomyopathies. J Cardiovasc Magn Reson 2000; 2: 151–57.
comparison with FDG PET-initial experience. Radiology 2000; 217: 140 Kim RJ, Judd RM, Chen EL, Fieno DS, Parrish TB, Lima JA.
729–36. Relationship of elevated 23Na magnetic resonance image intensity to
116 Chiu CW, So NM, Lam WW, Chan KY, Sanderson JE. Combined infarct size after acute reperfused myocardial infarction. Circulation
first-pass perfusion and viability study at MR imaging in patients 1999; 100: 185–92.

THE LANCET • Vol 363 • June 26, 2004 • www.thelancet.com 2171

For personal use. Only reproduce with permission from The Lancet Publishing Group.