Standardized cardiovascular magnetic resonance imaging (CMR) protocols, society for cardiovascular magnetic resonance: board of trustees task force on standardized protocols

Index 1. General techniques 1.1. Stress and safety equipment 1.2. Left ventricular (LV) structure and function module 1.3. Right ventricular (RV) structure and function module 1.4. Gadolinium dosing module. 1.5. First pass perfusion 1.6. Late gadolinium enhancement (LGE) 2. Disease specific protocols 2.1. Ischemic heart disease 2.1.1. Acute myocardial infarction (MI) 2.1.2. Chronic ischemic heart disease and viability 2.1.3. Dobutamine stress 2.1.4. Adenosine stress perfusion 2.2. Angiography: 2.2.1. Peripheral magnetic resonance angiography (MRA) 2.2.2. Thoracic MRA 2.2.3. Anomalous coronary arteries 2.2.4. Pulmonary vein evaluation 2.3. Other 2.3.1. Non-ischemic cardiomyopathy 2.3.2. Arrhythmogenic right ventricular cardiomyopathy (ARVC) 2.3.3. Congenital heart disease 2.3.4. Valvular heart disease 2.3.5. Pericardial disease 2.3.6. Masses


Other
2. Preparation and regular practice for rapid removal of the patient from the magnet 3. Emergency resuscitation policy in place 4. Defibrillator 5. Drugs for emergency treatment a. Immediately at hand: β-blocker (eg esmolol or metoprolol), nitroglycerin, aminophylline b. In the emergency cart: full set of emergency drugs (including drugs such as: epinephrine, beta blockers, atropine, bronchodilators, antiarrhythmic drugs) 6. For dobutamine -on-line assessment of wall motion during image reconstruction performed immediately after image acquisition 3. Scout to line up short axis images -cine acquisitions are preferable to single shot as long axis motion and inflow should be visualized a. Vertical long axis prescribed orthogonal to transaxial scouts aligned through the apex and center of the mitral valve b. Horizontal long axis aligned orthogonal to the vertical long axis, passing through the apex and center of the mitral valve 4. Steady state free precession short axis cine images, from the mitral valve plane through the apex. The basal most short axis slice should be located immediately on the myocardial side of the atrioventricular junction at enddiastole prescribed from the previously acquired long axis cines.

Stress agents
a. Slice thickness 6-8 mm, with 2-4 mm interslice gaps to equal 10 mm. b. Temporal resolution ≤ 45 ms between phases c. Parallel imaging used as available 5. Steady state free precession long axis cine images a. The 4 chamber long axis is prescribed from the vertical long axis through the apex and center of the mitral and tricuspid valves. This can be cross-checked on basal short axis cines, using the costophrenic angle (margin) of the RV free wall.
b. Vertical long axis, prescribed from the scout already acquired c. LV outflow tract (LVOT) long axis, passing through the apex, the center of the mitral valve and aligned with the center of LVOT to aortic valve, as seen on a basal short axis cine.
d. Optional -a set of more than 3 rotational long axis views can be obtained. 6. Analysis a. All short axis images are evaluated with computer-aided analysis packages for planimetry of endocardial and epicardial borders at end-diastole and end-systole. More advanced software automatically adjusts for systolic atrioventricular ring descent. b. The inclusion or exclusion of papillary muscles in the LV mass should be the same as that used in normal reference ranges used for comparison. c. Care must be used at the 1 or 2 most basal slices. Due to systolic movement of the base towards the apex in normally contractile ventricles, the end-systolic phase will include only left atrium. This may not be the case in a severely dysfunctional LV. Either way, this slice at enddiastole will include LV mass and volume.

Right ventricular structure and function
1. Right ventricular (RV) short axis views can be obtained in a similar fashion to the LV structure and function module. If the short axis is used for quantification, it is important to place the basal short axis slice immediately on the myocardial side of the right ventricle and to take extra care to exclude appropriate amounts of atrial volume from at least one basal slice at end systole.
2. Transaxial stack of cines covering the RV enable best identification of the tricuspid valve plane.
3. Long axis images should include an RV vertical long axis view aligned with tricuspid inflow and a RV outflow tract view (sagittal plane through the pulmonary valve).
4. Analysis a. A similar computer-aided analytic approach is required as for the left ventricle.
b. Care must be taken with RV trabeculations and with the RV outflow tract after repair of tetralogy of Fallot with a consistent approach used for longitudinal comparison.

Gadolinium dosing module
Notes: 1. Volumes and injection rates depend on scan duration: the given values are recommendations for typical scan times.
2. Injection rates are different for 1 molar contrast agents. As a general rule, divide the given injection rates by a factor of 2.
4. Throughout the protocols, the term "gadolinium" refers to gadolinium chelates 5. Injection rate for peripheral angiography with ellipticcentric readouts may be different than those specified below.

First pass perfusion module
1. Scout imaging as per LV structure and function module 2. Saturation-recovery imaging with gradient echo-echo planar (GRE-EPI) hybrid, GRE, or SSFP readout 3. Short-axis view imaging (at least 3 slices per heart beat) a. For ischemia evaluation, must obtain data every heart beat b. Slice thickness 8 mm c. Parallel imaging, 2-fold acceleration, if available d. In-plane resolution, ~< 3 mm e. Readout temporal resolution ~100 -125 ms or shorter as available f. Contrast is given (0.05 -0.1 mmol/kg, 3-7 ml/s) followed by at least 30 ml saline flush (3-7 ml/sec) g. Breathhold starts during early phases of contrast infusion before contrast reaches the LV cavity.
h. Image for 40-50 heart beats by which time contrast has passed through the LV myocardium 1.6. Late gadolinium enhancement module 1. Need at least 10 minute wait after gadolinium injection (0.1-0.2 mmol/kg). Note -The delay may be shorter than 10 minutes if lower doses are used as blood pool signal falls below that of late enhanced myocardium.
2. 2D segmented inversion recovery GRE imaging during diastolic stand-still 3. Same views as for cine imaging (short-and long-axis views) 4. Slice thickness, same as for cine imaging 8. Read-out is usually every other heart beat but should be modified to every heart beat in the setting of bradycardia, and every third heart beat in the setting of tachycardia or arrhythmia.
9. Optional a. Single-shot imaging (SSFP readout) performed as backup for patients with irregular heart beat, difficulty breath holding.
b. 3D sequences with parallel imaging in appropriate patients if signal-to-noise is sufficient.

Disease specific protocols -
f. Adapt the SSFP cine sequence to optimize temporal resolution as needed as the heart rate increases.
g. Stop test for new wall motion abnormality, serious side effect, or achievement of peak heart rate.
3. Analysis a. View cines in multiscreen format, reviewing rest, intermediate stress levels and peak stress at the same time in a synchronized fashion.
c. Report inducible wall motion abnormalities and viability.

Adenosine stress perfusion CMR
1. LV structure and function module (alternatively this can be performed between stress and rest perfusion, although performance immediately after gadolinium infusion may reduce the contrast of the blood-endocardium interface) 2. Two intravenous lines should be available, one for gadolinium and one for adenosine, one in each arm. Preferential site of contrast infusion is antecubital. Blood pressure cuff should be used with care taken not to interfere with gadolinium or adenosine infusion.
3. Adenosine stress perfusion imaging (at least 3 minute infusion of 140 ug/kg body weight/min). Option -initial adenosine infusion may be performed with the patient outside the bore of the magnet. a. Two volumetric acquisitions -one pre-contrast (for subtraction) and one during contrast administration.
b. Gadolinium injected in 2 phases to minimize venous contamination followed by saline bolus.
d. Slices -typically 60-80, as needed to accommodate vessels of interest.
e. Volumes obtained of abdomen/pelvis and thighs may be coarser spatial resolution (larger vessels), while those of the legs preferably are sub-millimeter spatial resolution. The former acquisitions typically require 15-20 seconds, while the leg acquisition may take 60-90 seconds for increased spatial resolution. Elliptical centric k-space acquisition is advantageous for the legs. If available, timeresolved acquisitions are preferred for the legs.
f. Parallel acquisition recommended (multichannel surface coil needed) 5. Analysis a. 3D reconstructions may be helpful for an initial overview and visualizing the vasculature tree, but generally should not be used for primary decision making.
b. Primary diagnoses are made by scrolling through source images (typically coronal and/or sagittal), and using selected thin slab MIP and MPR reconstructions in optimized orthogonal and oblique views for each station. The presence, number, and degree of stenoses are evaluated qualitatively.
Alternative: dual injection protocol e. Consider prone position in overweight patients in order to minimize distance between the surface coil and RV.
Major criteria for diagnosing ARVC include severe dilation and reduction of right ventricular ejection fraction with no or mild left ventricular involvement; localized right ventricular aneurysms (akinetic or dyskinetic areas with diastolic bulgings); severe segmental dilation of right ventricle. Minor criteria include mild global right ventricular dilation or ejection fraction reduction with normal left ventricle; mild segmental dilation of right ventricle; regional right ventricular hypokinesia. All of these are demonstrable by CMR as performed above. CMR also allows assessment of fatty infiltration and myocardial fibrosis. However, the latter two findings are not part of current guidelines.
Note -there is variability of the structure and shape of the RV amongst normals, so there is a tendency for inexperienced observers to overdiagnose RV wall motion abnormalities. For example, relative end-systolic bulging of a thin but contractile region of the RV free wall adjacent to the moderator band can be a normal finding and basal short axis cines may give the impression of inferior wall dyskinesis due to normal through-plane motion. Careful