Contrast Agents in Cardiac Magnetic Resonance

Currently only gadolinium-based contrast agents (GBCAs) are used in clinical practice. When injected as an intravenous bolus, a GBCA takes 15 to 30 seconds for transit through the cardiac chambers and blood vessels (first-pass phase) before it diffuses into the extracellular space. At approximately 10 to 15 minutes after injection, a transient equilibrium between contrast washing-in into the extracellular space and washing-out to the blood pool is reached. Myocardial perfusion CMR and most magnetic resonance angiography (MRA) examinations are performed during the first-pass phase, whereas late gadolinium enhancement (LGE) images are obtained during the equilibrium phase. Several GBCAs are commercially available in the United States; however, their use in CMR imaging is considered off-label. Mild reactions from GBCAs occur in approximately 1% of patients receiving these agents, but severe or anaphylactic reactions are very rare. All GBCAs are chelated to make the compounds nontoxic and to allow renal excretion. Exposure to the nonchelated component of GBCAs (Gd³⁺) has been associated with a rare condition known as nephrogenic systemic fibrosis (NSF), which is an interstitial inflammatory reaction that leads to severe skin induration, contracture of the extremities, fibrosis of internal organs, and even death. Risk factors for development of NSF include high-dose (>0.1 mmol/kg) GBCA regimens with estimated glomerular filtration rate (eGFR) less than 30 mL/min/1.73 m², need for hemodialysis, an eGFR less than 15 mL/min/1.73 m², use of gadodiamine (Omniscan, General Electric Healthcare, Chalfont St. Giles, Buckinghamshire, United Kingdom), acute renal failure, and presence of concurrent proinflammatory events. With the use of weight-based dosing and pretest screening, recent data suggest that NSF is now extremely rare. Previously, an incidence of 0.02% in 83,121 patients exposed to GBCAs over 10 years was noted; however, with current eGFR screening guidelines that have been widely practiced since 2006, a near-zero incidence has been reported.²

Technical Aspects of Cardiac Magnetic Resonance Pulse Sequences

CMR uses a range of strategies to overcome technical difficulties caused by cardiac, respiratory, and blood flow motion. Synchronized gating to the electrocardiogram (ECG) is routinely performed. Cardiac gating can be either prospective (triggering by an ECG waveform followed by a fixed period of acquisition during all cardiac cycles) or retrospective (continuous data acquisition with subsequent reconstruction based on ECG timing). For cine imaging, retrospective gating is preferred because it covers the entire cardiac cycle and is less prone to artifacts. To reduce blurring from cardiac motion, many CMR techniques fractionate the data for an image to acquire data only within a narrow window of the cardiac cycle (segmented approach). Currently, patient breath-holding remains the most common method to contain respiratory motion during CMR data acquisition, although navigator-based techniques (tracking of diaphragmatic motion to control respiratory motions) and respiratory motion averaging are options in some pulse sequences. Finally, by rapidly acquiring data of an entire image within a cardiac cycle, single-shot imaging and real-time cine imaging (continuous acquisition of single-shot images) can overcome both respiratory and cardiac motions, but at the expense of reduced temporal and spatial resolution. Table e17-1 shows a summary of the most common clinical CMR pulse sequence techniques at our center. Minor variations exist in these parameters between centers and vendors. CMR uses bright-blood cine imaging or dark-blood fast spin-echo (FSE) imaging to assess cardiac morphology and structure. Cine CMR is the modality that serves as a reference standard for quantifying ventricular volumes. Among the cine techniques, cine steady-state free precession (SSFP) is the technique of choice. It can acquire a cine movie at a high temporal resolution of 30 to 45 milliseconds during a breath-hold of less than 10 seconds, thereby capturing the whole heart in motion volumetrically in 3 to 5 minutes (Fig. 17-1; Video 17-1). For dark-blood techniques, T1-weighted FSE is used for assessing morphology of cardiac chambers, vascular structures, and pericardium and for imaging of fat (Fig. 17-2). T2-weighted FSE with fat suppression is used for imaging of myocardial edema occurring as a result of ischemia, infection, or infiltration. Three main techniques have been developed to quantify intramyocardial motion: myocardial grid or line tagging, phase contrast velocity mapping of myocardial motion, and displacement encoding with stimulated echoes (DENSE). Tagging assesses myocardial strain by marking the myocardium with parallel dark lines or a grid so that myocardial deformation can be visualized or quantified. Circumferential and radial strain also can be calculated and displayed with a color-coded scale. Although myocardial tagging is the most widely available, phase contrast velocity mapping and DENSE techniques can be completed at higher spatial resolution.

TABLE e17-1

Summary of Common Clinical Cardiac Magnetic Resonance Pulse Sequence Techniques at Brigham and Women’s Hospital

TECHNIQUE	PULSE SEQUENCE OPTIONS	DARK-/ BRIGHT-BLOOD	CONTRAST WEIGHTING	TYPICAL IN-PLANE SPATIAL/TEMPORAL RESOLUTIONS AND OTHER IMAGING PARAMETERS	BREATH-HOLD REQUIRED	GADOLINIUM CONTRAST REQUIRED	RELATIVE MERITS OF THE PULSE SEQUENCE OPTIONS	IMAGE EXAMPLE
Cine cardiac structure and ventricular function	Cine SSFP* Cine FGRE Real-time cine SSFP	Bright	T2W/T1W for cine SSFP and real-time cine SSFP; T1W for FGRE	1.5-2.5 mm/30-45 msec per phase • Adjust number of lines of k-space per cardiac cycle (segments) to balance temporal resolution with duration of patient breath-holds	Yes for ECG-gated cine SSFP and FGRE No for real-time cine	No	Cine SSFP has higher SNR and CNR (between endomyocardium and blood) than FGRE but is sensitive to field inhomogeneity (especially at 3T), giving rise to banding artifact FGRE has weaker endocardial definition than cine SSFP but is an alternative with severe artifact in cine SSFP Real-time cine SSFP: use in patients with significant arrhythmia or difficulty breath-holding; it has the lowest spatial and temporal resolutions	Cine SSFP
Quantitative regional myocardial strain	Myocardial tagging (newer but less widely available techniques for regional strain exist, (see text)	Bright	T1W	• Tag spacing 5-10 mm • Temporal resolution ~50 msec	Yes	No	Tissue tracking quantitation of intramyocardial motion Disadvantages: fading of tag lines near end of cardiac cycle and time-consuming strain analysis (postprocessing)
Structure, morphology, and fat imaging	Standard FSE* Single-shot (SS) FSE (or HASTE)	Dark	T1W ± fat suppression	0.8-1.5 mm every cardiac cycle	Yes for standard FSE No for SS FSE	No	Standard FSE has better image quality but relatively long scan time Fat suppression can be achieved by fat saturation pulse (more specific) or by suppressing tissues with short T1 (a technique known as STIR, which is less specific for fat). SS FSE covers the whole heart quickly and is useful in patients with arrhythmia or limited breath-holding
Myocardial scar by LGE imaging	Standard 2D segmented FGRE* 2D SS SSFP technique 3D whole-heart techniques (breath-hold or navigator-guided) PSIR (phase sensitive image reconstruction)	Bright	T1W (10-30 min after 0.1-0.2 mmol/kg GBCA injection)	1.5-2.0 mm/ 150-200 msec (for standard 2D) Adjust inversion time and time delay after ECG detection to null “normal” myocardium signal and to image in diastole, respectively	Yes for standard 2D technique No for SS technique	Yes	Standard 2D technique has higher spatial and temporal resolutions than the SS technique 2D SS technique covers the whole heart quickly and is useful in patients with arrhythmia or difficulty breath-holding PSIR is inversion time–insensitive and is more robust in nulling normal myocardium signal New 3D application using navigator-guidance yields higher SNR than 2D and can achieve spatial resolution of <1 mm without the need for breath-holding Refer to Table 17-e3 for patterns of LGE in various cardiomyopathies	2D segmented LGE
Myocardial perfusion imaging	Saturation prepared gradient-echo–based 2D techniques: – FGRE* – hybrid GE-echoplanar (EPI) – SSFP	Bright	T1W	2.0-3.0 mm 130-180 msec per slice 3-4 locations every cardiac cycle or 6-8 locations every 2 cardiac cycles during vasodilator stress and rest. 0.05-0.1 mmol/kg IV GBCA injected at 4 or 5 mL/sec (qualitative assessment only)	No, but breath-hold is preferable	Yes	Breath-holding useful to track contrast enhancement in specific segments. Parallel imaging acceleration and sparse sampling to reduce acquisition time per slice and extend slice coverage of the heart
Myocardial edema imaging	T2W FSE* STIR FSE T1W EGE ratio T2 prep SSFP T2 map (SSFP readout)	Dark (FSE-based), Bright (SSFP-based)	T2W + fat suppression (for T2W techniques) T1W (for EGE technique)	In-plane spatial and temporal resolutions similar to standard FSE Slice thickness 7-10 mm to improve SNR Algorithm to correct for distance of the heart from the receiver surface coils is required	Yes	No Yes for EGE ratio	Myocardial edema appears as a transmural area of high SI on T2W images In FSE techniques, beware of artifacts from slow flow, especially adjacent to regional wall motion abnormality or the LV apex, which may mimic edema Regional myocardial signal variation from phase array coils may mimic edema In absence of LGE, T2W edema reflects reversible myocardial injury Using T2W FSE techniques, an SI ratio of myocardium to skeletal muscle >1.9 has been reported to be abnormal in myocarditis An EGE ratio between myocardium and skeletal muscle of ≥4 or an absolute myocardial SI increase of 45% after contrast is considered abnormal in myocarditis The bright-blood SSFP-based technique has improved CNR and is less susceptible to slow-flow artifact. T2 map is insensitive to surface coil–related signal inhomogeneity and slow-flowing blood related artifact
Myocardial iron content imaging	T2*W multiple echo times FGRE	Bright	T2*W	2.0-3.0 mm/~100-150 msec One short-axis midventricular location A series of images with 8 echo times that goes from ~4 msec to 35 msec Axial ungated acquisition of the liver for comparison	Yes	No	Measurement is most accurate and reproducible in the midseptum T2* value describes the exponential decay of myocardial SI as the echo time increases T2* value of <20 msec with LV dysfunction (without other obvious cause) indicates iron overload cardiomyopathy
Cardiac thrombus	LGE with long inversion time EGE imaging	Bright	T1W	In-plane spatial and temporal resolutions similar to those with LGE imaging EGE is acquired within the first 5 min after gadolinium injection	Yes	Yes	LGE imaging with inversion time set at 600 msec or longer or EGE imaging can detect thrombus indicated by an intense “black” regions Look for thrombus in locations of stagnant flows
Cardiac blood flow	Phase contrast imaging cine GE	Bright	Velocity-related phase shift	1.5-2.5 mm/50 msec per phase Keep number of lines of k space per cardiac cycle (segments) low to improve temporal resolution during free-breathing studies	No (multiple signal averages used)	No	Multiple averages can reduce ghosting artifacts from respiratory motion during free breathing. Should keep velocity encoding strength slightly greater than the highest expected flow velocity, to avoid aliasing while maximizing accuracy
Coronary MRA	3D whole-heart volume using SSFP or FGRE* Target vessel approach	Bright	T2-prepared 3D SSFP or FGRE technique	~ 0.6-1.0 mm in plane Free-breathing navigator-guided 3D technique is currently most widely used	No, but Yes for target vessel approach	Yes or No (with SSFP-based technique in 1.5 T)	Compared with the target-vessel approach, 3D coronary MRA has higher SNR and provides volumetric whole-heart coverage T2-prepared SSFP sequence with suppression of the adjacent epicardial fat provides the strong blood vessel contrast Contrast-enhanced FGRE based technique is used in 3T
Anatomy for electrophysiologic mapping of the pulmonary vein	3D FGRE MRA of the left atrial volume and pulmonary veins	Bright	T1W FGRE	1.5- to 2.5-mm isotropic volume Timing bolus is required to achieve proper timing of imaging during first-pass transit of the contrast bolus Gating is optional but may improve border definition at the expense of prolonging breath-hold	Yes	Yes	Subtraction mask scan is necessary to enhance the MRA images Coronal (more common) or axial 3D MRA of the entire left atrium and the pulmonary vein is generated for electrophysiologic mapping; use same parameters as in the subtraction mask scan

^* More commonly used option.

CNR = contrast-to-noise ratio; EGE = early gadolinium enhancement; EPI = echoplanar imaging; FGRE = fast gradient-recalled echo; GE = gradient-echo imaging; HASTE = half-Fourier acquisition single-shot turbo spin echo; SE = spin echo imaging; SI = signal intensity; SNR = signal-to-noise ratio; SS = single-shot; T = Tesla; 3D = three-dimensional; 2D = two-dimensional; T1W = T1-weighted; T2*W = T2*-weighted.

(Note: Dark-blood techniques and myocardial iron content determination by T2* imaging should be performed before administration of gadolinium contrast.)

T1-weighted imaging techniques such as LGE imaging can detect accumulation of GBCA into the extracellular compartment of the myocardium secondary to infarction, infiltration, or fibrosis. LGE is detected 5 to 15 minutes after an intravenous injection of GBCA (0.1 to 0.2 mmol/kg) (hence the designation “late”). LGE data can be captured in two- or three-dimensional representation. Several technical improvements in LGE imaging have emerged. Phase-sensitive inversion recovery (PSIR) reference imaging incorporates the phase polarity information that enhances myocardial tissue contrast. Single-shot LGE imaging offers an option to overcome motion when cardiac gating or patient breath-holding is not possible. Navigator-guided LGE eliminates the need for breath-holding and allows three-dimensional acquisition with in-plane resolution below 1 mm (Fig. 17-3, Video 17-2). CMR perfusion imaging examines the first-pass transit of an intravenous bolus of GBCA as it travels through the coronary circulation. Several perfusion techniques are available, which are fast bright-blood gradient-echo imaging sequences in which three to five short-axis slices of the heart are acquired every cardiac cycle, during the injection of a GBCA bolus. Gadolinium provides strong signal enhancement in well-perfused regions, compared with hypoenhancement (dark regions) in poorly perfused myocardium. At a spatial resolution of approximately 2 mm in plane, CMR perfusion imaging can provide information of myocardial blood flow at the endocardial/epicardial or at a segmental level (Fig. 17-4; Video 17-3). Dynamic three-dimensional perfusion imaging can provide greater myocardial coverage and improved image quality and has shown promising preliminary clinical results.³ T2-weighted imaging detects myocardial edema, from ischemic injury or inflammation, and it has been shown to have high correlation to the area at risk after acute myocardial infarction (MI). It also complements LGE imaging in determining the chronicity of an MI and allowing for accurate measurement of salvageable myocardium. The pulse sequence options for T2-weighted imaging include black-blood short TI inversion recovery (STIR) FSE and the newer SSFP-type methods,⁴ and their merits are listed in Table e17-1. T2* is a transverse relaxation parameter sensitive to tissue iron content. T2* imaging is a well-validated method for measuring tissue iron content. A T2* less than 20 milliseconds (value for normal myocardium, approximately 40 to 50 milliseconds) is diagnostic of myocardial iron overload, and a T2* less than 10 milliseconds is evidence of severe iron overload⁵ (Fig. 17-5; Video 17-4). Despite challenges posed by small luminal sizes and cardiac and respiratory motions, technical advances in coronary MRA imaging have favored the use of whole-heart three-dimensional acquisition (with or without navigator guidance), with promising preliminary clinical results⁶ (Fig. 17-6). Similar to Doppler echocardiography (see Chapter 14), phase contrast imaging allows quantitation of velocities of blood flow and myocardial motion and intravascular flow rates. Parallel imaging is a family of techniques that speeds up CMR (k-space) data acquisition by combining information obtained separately from each element of the surface receiver coils. Incorporating parallel imaging can reduce acquisition time, improve temporal resolution, or even eliminate certain artifacts. The main disadvantage of parallel imaging is a reduction in signal-to-noise ratio resulting from undersampling of the k-space data.