Cardiac Magnetic Resonance Imaging

Cardiac Magnetic Resonance Imaging

Louis-Philippe David

Panagiotis Antiochos

Raymond Y. Kwong


Cardiac magnetic resonance imaging (CMR) is a noninvasive, nonionizing imaging modality, with high spatial resolution, that is considered to be the gold standard for morphologic assessment of the heart, as well as the assessment of right and left ventricular systolic and diastolic volumes, function, and mass. Using contrast agents and newer mapping techniques, CMR can, furthermore, provide unique information on tissue characterization that can improve diagnostic accuracy in the properly selected patient. The goal of this chapter is to create a comprehensive and practical reference for any general cardiologist, internist, fellow, or trainee who encounters cardiovascular magnetic resonance imaging (MRI) in their practice.


Physics for Clinicians and Instrumentation

Because of its abundance in water and lipid molecules, most clinical applications of MRI target hydrogen to generate a magnetic resonance signal. When a patient is placed inside a strong external magnetic field (B0) such as an MRI scanner, hydrogen atoms align in the direction of B0 resulting in a small net magnetization called Mz (Figure 37.1). In order to generate an MRI signal, energy has to be transferred to the protons. Hence, a radiofrequency pulse B1, a weaker oscillating magnetic field, is applied perpendicularly to Bo at a specific frequency called Larmor frequency. The resulting net magnetization can therefore be divided into two planes: the longitudinal plane Mz (parallel to B0) and transverse plane Mxy (perpendicular to B0). Once the regurgitant fraction (RF) pulse is turned off, hydrogen nuclei return to their equilibrium state and release energy that induces a voltage that the MRI coils can detect.1 This relaxation process has two distinct components that happen at the same time:

  • Longitudinal relaxation corresponds to the recovery of the magnetization along z direction and is defined by an exponential time constant, T1. Water molecules have long T1 values and appear dark on T1-weighted images whereas fat has a short T1 value and will appear bright.

  • Transverse relaxation refers to how quickly the spins exchange energy in the xy direction. This loss of transverse magnetization follows an oscillating pattern referred to as the free-induction decay and follows a time constant: T2. Local inhomogeneities in the magnetic field further accelerate this process and lead to a faster decay called T2*. Typically, water molecules have long T2 values and appear bright on T2-weighted images that can be useful in detecting active inflammation, for example.

System and Signal Encoding

  • The main magnet coils generate a strong constant, yet inhomogeneous, magnetic field (B0). Typically, the field strength of cardiac MRI scanners ranges from 1.5 to 3 Tesla.

  • When turned on, the transmitter coils that are built in the main structure of the magnet can emit an RF (weaker magnetic field) that will excite protons.

  • Three gradient coils built in the main structure of the magnet along the x, y, and z directions can be turned on and off. Gradients create small variations in magnetic field along their axis and are useful for slice selection and localizing the magnetic resonance signal in space.

  • Receiving coils are typically placed on the surface of the patient’s body to maximize signal coming from the heart and are turned on during signal readout.

Cardiac Synchronization and Respiratory Gating

By synchronizing image acquisition with the cardiac cycle using ECG signal from leads attached to the patient, artifacts due to cardiac contractile motion and high velocity blood flow can be minimized. Cardiac gating depends on the quality of the ECG signal and the absence of arrhythmias. Cardiac synchronization can be prospective (R wave is detected and triggers the sequence and acquisition) or retrospective (all the data is continuously acquired over multiple R-R intervals).

Respiratory motion is another potential source of artifacts, so most sequences are performed using fast acquisition imaging techniques during breath-holding. However, some patients cannot breath-hold for 10 to 15 seconds, whereas other sequences may require longer acquisition. Respiratory gating methods such as the use of navigator sequences allow the acquisition of data in a predefined time window based on diaphragm position, at the expense of increased scan time.

Magnetic Resonance Sequences

A detailed explanation of CMR sequences is beyond the scope of this chapter. The interested reader is referred to more detailed publications on the subject.2,3 In CMR, spin echo sequences are the cornerstone of black blood imaging and are mainly used for anatomic imaging. On the other hand, gradient echo sequences produce images where blood appears bright. Gradient recalled echo (GRE) and balanced steady-state free precession (bSSFP) are two types of gradient echo sequences frequently utilized in CMR and are typically used for cine imaging (anatomy, ventricular size, mass, and function) and phase-contrast images. Phase-contrast sequences allow quantification of blood flow and velocity and are particularly useful to assess valvular stenosis, valvular regurgitation, and intracardiac shunts.

Gadolinium-Based Contrast Agents

The use of gadolinium-based contrast agents (GBCA) in CMR provides important diagnostic and prognostic information; they are routinely used in perfusion imaging, assessment of myocardial scar, and cardiomyopathies (CMP). Based on their molecular configuration, GBCA can be classified as linear or macrocyclic agents. Macrocyclic agents are the newer generation of GBCA with a molecular structure substantially more stable in chelating the gadolinium element and thus safer than the linear agents. Allergic reactions may be seen in 0.004% to 0.7%,4 whereas true anaphylactic reactions are rare. The most serious potential complication of exposure to GBCA was nephrogenic systemic fibrosis (NSF), observed in patients with severe kidney disease (estimated glomerular filtration rate [eGFR] ≤ 30 mL/min). The risk of NSF is higher during periods of rapid deterioration of renal function, acute systemic illness, or with repeated administration of GBCA within a short period of time. This irreversible process is characterized by interstitial inflammatory reaction eventually leading to skin induration, contractures, multiorgan fibrosis, and potentially death. After implementation of guidelines restricting the use of GBCA in patients with severe kidney disease and with the use of safer macrocyclic-structured GBCA, there have practically been no new reports of NSF relating to GBCA. The safety profile of GBCA has not been established in pregnant women and in general should also be avoided unless the benefits of performing the study outweighs the potential risk.

GBCA is usually injected through peripheral venous access and reaches the myocardium through coronary artery circulation within 15 to 30 seconds after injection before it diffuses into the extracellular space. Myocardial first-pass perfusion CMR and the majority of magnetic resonance angiography (MRA) are performed during this phase. At 10 to 15 minutes after injection, a wash-in/wash-out equilibrium state is reached between contrast in the blood pool and the extracellular space leading to an optimal imaging window for late gadolinium enhancement (LGE) imaging.

Late Gadolinium Enhancement

Gadolinium shortens T1 time of tissues leading to faster longitudinal recovery and a more intense magnetic resonance signal. First, an inversion time mapping sequence (TI scout) is performed to identify the optimal timing of image acquisition after an inversion recovery pulse where the normal myocardium is dark (nulled). Multiple short-axis images of the left ventricle (LV) are acquired while the inversion time (TI) is progressively increased. In patients with normal myocardium, blood pool, which contains more gadolinium, is nulled before the myocardium. In pathologies such as cardiac amyloidosis (AL) where increased extracellular volume is a hallmark of the disease, there is myocardial retention of gadolinium that may lead to nulling of the myocardium before the blood pool. Once the proper TI has been selected, images should show optimal contrast differentiation between bright LGE and the normal, nulled myocardium. Although people often associate LGE with the presence of fibrosis, it should be clear that LGE does not necessarily mean scarring. LGE reflects abnormally enlarged extracellular myocardial compartment, which can be
seen in infarction, infiltration, replacement fibrosis, and also inflammation/edema.


Over two million patients in the United States have implanted devices, including pacemakers and implantable cardioverter-defibrillators (ICDs), and it is estimated that more than 50% of those will require MRI after device implantation. CMR in patients with magnetic resonance-conditional devices can be performed safely as long as device restrictions are adhered to and safety precautions are taken. Although scanning of patients with non-magnetic resonance-conditional devices—especially legacy devices—is considered high risk and discouraged, it is not an absolute contraindication as long as the risk-to-benefit ratio justifies CMR.6,7 Scanning such patients should only be considered under highly compelling circumstances and where an exit strategy in the event of device failure is firmly in place.

Apart from implanted devices, CMR is potentially problematic in patients with ferromagnetic metallic implants (Tables 37.2 and 37.3). Thorough screening of all patients prior to CMR is mandatory. Devices such as prosthetic heart valves, prosthetic joints, sternal wires, and intravascular stents do not preclude study with CMR at field strengths of 1.5 and 3.0 Tesla.


Chamber Morphology and Function

Accurate quantification of LV and RV volumes can have significant diagnostic and prognostic impacts. TTE (transthoracic echocardiogram) is often the first step for clinicians to assess chamber morphology, although difficult image acquisition because of obesity, unfavorable acoustic windows, and foreshortening of the LV in apical views are potential limitations. LV evaluation by Simpson’s biplane method is based on geometric assumptions and RV evaluation is often just qualitative or based on linear dimensions because of its crescentic shape. CMR can overcome these potential pitfalls, is highly reproducible, and is now considered the gold standard to assess chamber morphology.

First, a stack of short-axis SSFP cine images covering basal to apical segments of both ventricles is acquired using retrospective gating. Each slice is 6 to 8 mm thick and is acquired during breath-holding, combining information over multiple RR intervals to produce the final image. Careful planning with double-oblique orientation is crucial to avoid foreshortening of the LV. For each end-diastolic and end-systolic slices, endocardial and epicardial borders are traced. Because slice thickness is known, Simpson’s method of disks can be used to accurately quantify volumes and mass. Inclusion or exclusion of papillary muscles in the LV mass is considered appropriate, as long as the given normal reference values reflect the methodology that has been used. In patients with high arrhythmia burden or respiratory motion artifacts, the use of faster free-breathing real-time cine imaging can allow evaluation of cardiac function at the expense of lower spatial resolution. Reported volumes are generally indexed to body surface area with normal reference values adapted according to the sex and age of the patient.

Estimation of atrial volumes can also be performed using four- and two-chamber long-axis views (area-length or biplane summation of disks methods). Alternatively, atrial volumes can be obtained from axial or short-axis atrial stacks.

Ischemic Heart Disease, Myocardial Infarction, and Viability

Myocardial Infarction

LGE imaging by CMR is currently the gold standard imaging modality for quantifying MI size and is able to detect small subendocardial MI (as little as 1g) with good accuracy.8 In the acute phase of MI, CMR may detect the presence of microvascular obstruction, which refers to the inability to reperfuse the coronary microcirculation in a previously ischemic region, despite opening of the epicardial vessel. On LGE sequences, microvascular obstruction appears as a dark core within the areas of hyperenhancement (Figure 37.2 with MVO). The extent of microvascular obstruction and infarct size increase substantially over the first 48 hours after an MI and are of prognostic value.

In patients with an acute reperfused MI, the area at risk refers to the territory supplied by the infarct-related artery that would have infarcted after MI if reperfusion had not taken place to salvage viable myocardium. The area at risk includes both the infarcted myocardium and the salvaged myocardium that surrounds it. Myocardial salvage can be calculated by subtracting the MI size from the area at risk and myocardial salvage index refers to the ratio of the myocardial salvage to the area at risk. The myocardial salvage index is considered a sensitive measure for assessing the efficacy of novel cardioprotective therapies compared with MI size alone.9

CMR also offers assessment of myocardial viability before coronary revascularization. The transmural extent of myocardial scar detected by LGE imaging accurately depicts a progressive stepwise decrease in functional recovery despite successful coronary revascularization. Compared to dobutamine exercise treadmill test (ETT), LGE is easy to perform and interpret and a 50% transmurality cutoff is sensitive in detecting segmental contractile recovery.10

Compared to other imaging modalities, CMR is the most accurate to reliably detect unrecognized (silent) MI. In community-based studies that utilized CMR, MI detected by LGE imaging but unrecognized by clinical examination including ECG (thus untreated) were reported to occur in 6% to 17%, with marked increased in patient mortality consistently reported in these patients with unrecognized MI.11

Lastly, CMR is a valuable diagnostic tool in patients who present with acute elevation of serum biomarkers consistent
with myocardial injury, but with nonobstructive coronary arteries (MI with no obstructive arteries [MINOCA]). By providing multicomponent assessment of myocardial structure and physiology, CMR can effectively capture various noncoronary abnormalities and guide differential diagnosis.

Stress Cardiac Magnetic Resonance Imaging for Detecting and Quantifying Myocardial Ischemia

Stress CMR has evolved to an everyday clinical tool, and is considered an effective first-line test for the evaluation, diagnosis, and risk stratification of patients with suspected IHD. Stress CMR is performed using vasodilating (eg, dipyridamole, adenosine, regadenoson) or positive inotropic pharmacologic stress agents. Following this, an intravenous bolus of GBCA is administered and three short-axis slices, each of 10 mm thickness, are acquired per cardiac cycle, at the basal, mid papillary, and apical levels of the LV, with a typical resolution of 2.5 × 2.5 mm.

In a normal scan, the wash-in (first pass) of GBCA into the myocardium can be seen as the myocardium turning from black to mid-gray uniformly throughout the whole of the LV in both the stress and rest scans. In an abnormal scan, an area of the myocardium will turn gray slower than the surrounding tissue because the blood (and hence gadolinium) enters more slowly due to a narrowing of the coronary artery supplying it. This is called a perfusion defect and usually represents myocardial ischemia. It may be seen on both the rest and stress scans in which case it is called a matched perfusion defect and is probably due to an area or scar from a previous MI. If it is only seen on the stress scan, it is called an area of inducible perfusion defect (ischemia). The positions in the LV of the perfusion defects are described using the AHA 17 segment model (Figure 37.3).

Multicenter studies have shown that a negative stress CMR portends to an annualized cardiac event rate of <1% in patients with an intermediate pretest likelihood of IHD.12
There is excellent correlation of stress CMR assessment of ischemia against invasive measurement of fractional flow reserve, showcasing its high accuracy in determining the physiologic significance of coronary stenosis, and safely guiding management in IHD.13 Compared to cardiac single photon emission computed tomography (SPECT) imaging, stress CMR has several technical advantages: (1) it is not limited by attenuation artifacts; (2) it is free from ionizing radiation; (3) it has three- to fourfold higher spatial resolution; (4) it takes 35 to 45 minutes to complete (compared to >2 hours for dual-isotope SPECT); and (5) it performs better than SPECT in detecting single or multivessel coronary disease.14,15,16

Only gold members can continue reading. Log In or Register to continue

May 8, 2022 | Posted by in CARDIOLOGY | Comments Off on Cardiac Magnetic Resonance Imaging
Premium Wordpress Themes by UFO Themes