Chapter 29: Cardiovascular Magnetic Resonance

Cardiovascular magnetic resonance is an advanced noninvasive cardiovascular imaging technique that has become well established but continues to evolve. It has some advantages over SPECT myocardial perfusion imaging (MPI) in the evaluation of ischemic heart disease but also has limitations that restrict the potential patient population. Technical advances in hardware and software protocols are unlocking new diagnostic and prognostic possibilities. In many areas of cardiovascular disease, it is the reference standard. The versatility of CMR allows one scan to provide information on cardiovascular structure, function, fibrosis, perfusion, and tissue characterization. CMR is a robust technique for evaluating ischemic heart disease and cardiomyopathies and has a role to play in evaluating other conditions, such as valvular and pericardial disease. We will provide a brief overview of the methods of CMR imaging, discuss its advantages and limitations, and outline its use in specific disease states.

Magnetic resonance imaging is performed by having a patient lie in the bore of a large hollow magnet in which a magnetic field is generated, typically at 1.5 Tesla. The hydrogen atoms in the body, predominately in water and fat, behave like magnets and possess "spin." When the hydrogen protons are exposed to the magnetic field they align their spins.^1,2 Radiofrequency pulses are generated by the magnet and excite the protons in specific planes of predetermined size and location so that their spins are aligned in a higher-energy state. Relaxation of the hydrogen nuclei to the lower-energy state gives off an electromagnetic signal that is detected by the scanner and processed into an image through a technique known as Fourier transformation.

There are many substantive advantages of CMR compared with other imaging modalities, including its high spatial resolution and high signal-to-noise ratio. CMR can also provide a 3D assessment of the heart, allowing the selection of any imaging plane, and is not susceptible to attenuation artifacts. Using different CMR sequences can also aid in tissue characterization, allowing multiple aspects of the myocardium to be assessed in one study (Table 29-1).

There are several limitations to the routine use of CMR. Despite recent improvements, there continues to be limited hardware availability, and cardiac-specific software is necessary. CMR remains an expensive imaging modality; this is challenging in the era of cost containment, although stress CMR and stress radionuclide imaging (RNI) are fairly similar in cost. Functional images are acquired over multiple beats gated to the electrocardiogram, which prolongs study times. Study times can be lengthy and are typically on the order of 40 to 60 minutes. Electrocardiographic gating can also result in artifacts in the setting of arrhythmias, particularly irregular rhythms such as atrial fibrillation and frequent premature ventricular contractions. Gating can be more difficult at higher field strengths. Patients must lie still, as all images are dependent on a fixed position relative to the magnetic field and RF coils. This can create compliance issues. Patients must hold their breath for most sequences, which can be challenging for symptomatic patients with limited respiratory reserve and also results in reduced compliance.¹ The confined nature of the magnet bore may be an issue for morbidly obese patients and can trigger claustrophobia, which can accentuate all of these issues. The magnetic field can interact with implanted devices. Some newer devices are potentially MRI compatible, but careful patient selection, well-defined safety protocols, and additional large-scale testing are needed prior to the widespread use of CMR in this population and in those without MRI-compatible devices. Performing protocols on the MRI machine and software requires highly trained staff. Adequate training is also required to perform the image processing that is required to quantify volumes and function. However, newer software is allowing these processes to be increasingly automated.³ The use of gadolinium-based contrast agents (GBCAs) are contraindicated in patients with stage 4 and 5 kidney disease due to the association with nephrogenic systemic fibrosis.⁴ Finally, in all but a few specialized laboratories, stress imaging can only be performed with pharmacologic stress, eliminating the most powerful prognostic marker in stress imaging, exercise capacity.⁵ Rapid advances in CMR hardware, software, and MR sequences have the potential to improve many of these limitations in the near future.

A CMR study is made up of a series of specific sequences that reveal different aspects of the heart and vascular system. The ordering physician supplies a clinical question and the performing laboratory determines the optimal CMR protocol to answer that question. Unique sequences can assess heart structure, myocardial function, tissue characterization, perfusion, valvular function, and angiography.

Structure and Function

Steady-State Free Precession Cine Imaging

CMR is the gold standard for measuring left ventricular (LV) mass and LV and right ventricular (RV) chamber volumes and ejection fraction. Steady-state free precession (SSFP) imaging sequences are most commonly used to assess volumes and function, as they have excellent spatial and temporal resolution, a high signal-to-noise ratio, and short breath hold times.⁶ Stacking multiple short-axis slices together allows precise volume and function assessment without geometric assumptions.¹ The endocardial and epicardial borders are traced in end-diastole and end-systole, and the myocardial mass, stroke volume and ejection fraction can be calculated (Fig. 29-1).^6,7 Regional function can be quantified using a 17-segment model based on coronary distribution.⁸

In a comparison of LVEF between CMR, radionuclide ventriculography, and 2D echocardiography, the closest limits of agreement were with radionuclide ventriculography, which has also been considered a gold standard.⁹ There are wide variances with 2D and 3D echo, with systematic underestimation by 3D echo due to insufficient definition of endocardial trabeculae.¹⁰ CMR can orient the heart in any direction to assess function and precisely quantify RV and atrial function. CMR is particularly well suited for repeated studies for disease progression monitoring compared with echocardiography due to its accuracy and reproducibility, and RNI for the lack of radiation exposure.³

Several techniques are available to assess regional myocardial function and strain include myocardial tagging, harmonic phase imaging (HARP), and cine displacement encoded with stimulated echoes (DENSE).^3,6 Dobutamine CMR with tagging improves the identification of significant coronary lesions.¹¹ HARP is a low spatial resolution approach to tag analysis that has been used widely in population studies.¹² Cine DENSE has higher temporal and spatial resolution and will likely become dominant as the technology matures (Table 29-2).¹³

Tissue Characterization

T1-, T2-, and T2*-Weighted (W) Imaging

Adjustment of the electrical gradients, RF pulses, and timing of signal acquisition can emphasize different properties of the myocardium and aid in tissue characterization. The time it takes tissue to recover from excitation, the T1 relaxation time, is prolonged in fibrotic tissue and reduced in lipid-rich myocardium.¹⁴ T2 describes the time for the proton spins to lose their alignment in adjacent tissue (spin–spin relaxation time). T2-W accentuates tissue with a high water content, such as edema in early states of myocardial injury.¹⁵ T2* sequences can identify iron overload, such as in hemochromatosis or in conditions such as thalassemia after multiple transfusions.¹⁶ T1, T2, and T2* mapping are rapidly replacing static images due to their quantitative nature and robustness to artifacts.^17,18

Late Gadolinium Enhancement

Late gadolinium enhancement (LGE) is a method of scar assessment useful in assessing ischemic heart disease and cardiomyopathies. Patterns of gadolinium uptake during specific time periods following injection can help identify pathology. An absence of gadolinium uptake on delayed imaging indicates severe perfusion defects such as from microvascular obstruction after acute myocardial infarction (MI).¹⁹ Increased gadolinium uptake and delayed washout occur in infarcted myocardium due to intracellular accumulation in injured myocytes due to ruptured cell walls and in fibrotic scar due to its small intracellular fraction.¹⁹

Blood Flow

Blood flow quantification is performed in CMR through velocity-encoded imaging. Protons at motion develop a specific phase shift directly proportional to their velocity. Velocity-encoded imaging (also called phase-contrast) produces images with signal proportional to the velocity and direction of each individual pixel.⁷ Cumulative antero- or retrograde flows through specific vessels can be calculated by assessing blood flow through the region of interest over the entire cardiac cycle. Forward and regurgitant flow can be quantified and the regurgitant fraction calculated (Fig. 29-2).

Myocardial Perfusion

CMR perfusion imaging involves administering GBCAs and performing first-pass imaging, typically in several short-axis slices. Imaging is performed at rest and after stress (typically using a vasodilator agent) to assess for ischemia. Images are typically assessed qualitatively for visible defects. However, quantitative perfusion analysis can be performed and improves differentiation of moderate and severe stenoses and improves the identification of three-vessel CAD.²⁰ In combination with LGE imaging, ischemia and infarct can be differentiated with high accuracy (Fig. 29-3).²¹ The direct imaging of scar with LGE is an advantage over nuclear imaging.

CMR Coronary Angiography

SSFP imaging is used for coronary evaluation without contrast given the high blood T2/T1 ratio.²² SSFP is not susceptible to calcium, allowing imaging with heavy calcification. Images are limited by reduced spatial resolution (1–1.5 mm) and long imaging times requiring free-breathing imaging with respiratory gating.

Ischemic Heart Disease

Although relatively new in the evaluation of ischemic heart disease, stress CMR has many unique advantages and a rapidly expanding literature base supporting its use. Its high spatial resolution allows CMR to detect small subendocardial perfusion defects often missed by other techniques.⁶ The standard CMR-guided evaluation of ischemic heart disease includes assessments of function, perfusion, extent and location of scar, and myocardial viability. Advanced techniques such as absolute flow assessment and angiography are also available in centers with expertise in these areas.

Function

Resting global systolic left ventricular function and segmental wall-motion can be assessed readily by CMR and can assist in differentiating acute coronary syndromes from noncardiac conditions. The improved spatial resolution and lack of attenuation from ribs and other structures can be particularly useful in assessing certain regions like the posterior wall.²³

Quantitative strain analysis can also be useful to assess function in ischemic heart disease.²⁴

Stress-induced wall-motion abnormalities have high diagnostic accuracy for CAD. Exercise stress CMR is challenging logistically and is only performed in a handful of centers. Dobutamine is a sympathomimetic drug that increases myocardial blood flow and contractility through chronotropic and inotropic effects and can be used in stress CMR.²⁵ A meta-analysis of 1183 patients showed a sensitivity of 83% and a specificity of 86% of dobutamine stress CMR for CAD at the patient level, which is higher than studies with stress echocardiography.^26,27 The prognostic impact of inducible wall-motion abnormalities on stress CMR was studied in patients in a bicenter study by Kelle et al.²⁸ Patients without abnormalities had a 0.9% rate of cardiac death/nonfatal MI. Inducible WMAs were an independent predictor with hazard ratio of 6.5, p < 0.001. Dobutamine functional stress CMR has largely been replaced at most centers by vasodilator stress perfusion CMR due to the short half-life of the vasodilators, patient comfort, as well as the use of GBCAs to assess LGE.

Perfusion

Stress CMR perfusion imaging has replaced dobutamine stress at most centers due to its high diagnostic accuracy, short vasodilator half-life, patient comfort, and excellent scar analysis. A meta-analysis of 166 articles (including 37 analyzing CMR) found an excellent patient-level sensitivity of 89% and specificity of 76% for the identification of 50% stenosis of an epicardial coronary artery.²⁹ The sensitivity of CMR was comparable to SPECT (88%) and PET (84%). However, the specificity was substantially improved over SPECT (61%). The CE-MARC trial compared stress SPECT MPI and CMR in 752 patients with suspected angina and at least one risk factor. In this study, while specificity was similar, sensitivity was increased with CMR (86.5% vs. 66.5%, p < 0.001).³⁰

The degree of inducible ischemia is prognostically significant. Shah et al.³¹ examined 815 consecutive patients undergoing stress CMR and showed that inducible ischemia was strongly associated with MACE (HR 14.66, p < 0.0001) and reclassified 91.5% of patients at moderate pretest risk to low risk (65.7%, MACE 0.3%/year) or high risk (25.8%, MACE 4.9%/year). A negative stress CMR is associated with a very low risk of cardiovascular death and MI (0.8%/year).³² A recent analysis of follow-up data in the CE-MARC trial population showed no difference in major adverse cardiovascular event rates between the CMR and SPECT groups and similar rates of unnecessary coronary angiography.³³

A substantial potential benefit of stress CMR over SPECT MPI and stress echocardiography is the ability to quantify absolute blood flow and calculate myocardial flow reserve with values similar to PET MPI.³⁴ Reduced absolute stress flow and flow reserve are associated with multivessel disease and increased cardiac events.^35,36 The high resolution of CMR imaging allows quantification of absolute flow across the layers of the myocardium.⁶ The limitations of operator expertise and long postprocessing times will be reduced through automated software, allowing more mainstream adoption.

Assessment of Scar

The presence, location, and extent of scar are critical for diagnosing prior infarction and for risk stratification. LGE has been shown to correlate closely with the distribution of myocardial necrosis by TTC staining and microsphere analysis in animal models.³⁷ In contrast, ischemic myocardium that is not irreversibly damaged does not exhibit LGE.³⁸ Wagner et al.³⁹ performed both LGE CMR and SPECT in 91 patients. Although SPECT identified all patients with nearly transmural infarction, it did not identify subendocardial infarction in 47% of segments and 13% of patients. Although PET and CMR infarct size correlated well (r = 0.81, p < 0.0001), a small number of segments read as normal on PET MPI had evidence of LGE on CMR.⁴⁰ Scar location can also be used to identify optimal lead location in cardiac resynchronization therapy and likelihood of functional recovery.¹⁴ Kwong et al.⁴¹ showed that the presence of LGE in patients without known prior MI has incremental prognostic value beyond clinical factors, coronary stenosis, and left ventricular function. This same group found that diabetic patients with silent MI have increased mortality.⁴²

Myocardial Viability

Noninvasive imaging can help identify which areas of myocardium are viable postinfarction. Kim et al.⁴³ performed CMR with LGE in 50 patients with left ventricular dysfunction. In an analysis of all 804 dysfunctional segments, there was a stepwise decrease in likelihood of functional improvement with increasing transmurality of LGE (Fig. 29-4). There have not been studies of long-term outcomes stratified by extent of viability identified on CMR imaging, though similar studies in patients undergoing SPECT and PET have been favorable.^44,45 The complementary information provided in the identification of metabolically active myocardium by FDG-PET and scar by CMR suggests there may be benefit to a comprehensive hybrid PET-MR viability assessment.¹⁹

Another index of viability available by CMR is the end-diastolic wall thickness. The high resolution of CMR allows assessment of end-diastolic wall thickness, with values ≤5.5 mm associated with a low likelihood of myocardial viability.⁴⁶ Low-dose dobutamine induces inotropy that recruits stunned and hibernating myocardium but not necrotic tissue.²⁵ Multiple studies comparing LGE and contractile reserve assessment by CMR disagree on the best method to predict wall-motion improvement after revascularization. A meta-analysis by Romero et al.⁴⁷ found LGE to have the highest sensitivity and negative predictive value and dobutamine contractile reserve to have the highest specificity and positive predictive value. In particular, dobutamine contractile reserve appears to be beneficial in those with intermediate LGE thickness 1% to 75%, in which LGE is less definitive.⁴⁸

Tissue Characterization

Tissue characterization by CMR can be particularly helpful in the setting of acute or subacute MI.¹⁹ T2-W imaging can identify myocardial edema in areas with acute injury. The combination of this tool with LGE can help differentiate acute from chronic MI.⁴⁹ The extent of T2 enhancement identifies the area at risk in acute injury, and the area without subsequent LGE represents salvaged myocardium.^50,51 These findings are comparable to SPECT estimates of area at risk.⁵² T2-W imaging acutely postinfarction can also identify dark zones of intramyocardial hemorrhage from particularly severe ischemic damage.⁵³ Intramyocardial hemorrhage is associated with adverse long-term left ventricular remodeling.^25,54

MR Coronary Angiography

MR coronary angiography can be combined with stress perfusion imaging to provide a comprehensive anatomic and functional assessment of ischemic heart disease. The absence of susceptibility to calcium artifacts is a substantial advantage, and CMR coronary angiography may have improved diagnostic accuracy in patients with high calcium scores.⁵⁵ Poor spatial resolution and long imaging times limit full coronary evaluation currently. These limitations have resulted in a rating of "inappropriate" for the exclusion of significant CAD in patients at intermediate risk with chest pain.⁵⁶ However, the sites of origin and proximal courses of the coronaries can be evaluated quickly and with sufficient accuracy.²² Therefore, clinical use currently centers on the identification of anomalous coronary arteries and evaluation of coronary artery aneurysms in Kawasaki disease.²² However, future advantages such as use of high field-strength 3T MRI, 32-channel coils, and high parallel imaging factors may allow full coronary evaluation with diagnostic accuracy similar to that obtained with a 64-slice CT scan (Fig. 29-5).⁵⁷

Heart Failure/Cardiomyopathies

CMR is considered "appropriate" in the 2013 ACC/AHA Appropriate Utilization of Cardiovascular Imaging in Heart Failure guidelines for evaluation in newly suspected or diagnosed heart failure, as well as in those meeting criteria for ICD or CRT implantation.⁵⁸ It plays an important role in multiple etiologies of heart failure, including dilated (DCM), hypertrophic (HCM), and restrictive cardiomyopathies (RCM), arrhythmogenic right ventricular cardiomyopathy (ARVC), iron overload, and left ventricular noncompaction.

LV and RV volumes and function may be ascertained with a lack of radiation, which facilitates repeat testing to monitor disease course and effects of therapy. The high spatial resolution facilitates the detection of LV thrombus and identification of LV noncompaction.^14,59 Excellent RV wall visualization aids in identifying global dysfunction and focal segmental abnormalities in ARVC.²³

In DCM, tissue characterization and patterns of LGE can aid in differentiation of ischemic from nonischemic cardiomyopathy, suggest etiology, and inform prognosis (Fig. 29-6). In 90 patients with DCM, all with CAD had subendocardial LGE, while those with a nonischemic etiology had no enhancement (59%) or midwall or epicardial LGE.⁶⁰ LGE in DCM is associated with increased mortality and risk of sudden cardiac death.⁶¹ CMR can identify myocarditis with moderate sensitivity (67%) and excellent specificity (91%) using edema presence and LGE patterns.⁶² Nuclear MPI can assess LV, and to a lesser extent RV function and assess the microvasculature, but there is currently minimal role for the assessment of nonischemic DCM using this modality.⁶³

CMR structural imaging can identify the multiple patterns of HCM, including classic asymmetric septal involvement, but also atypical midwall and apical variants that can be missed by echocardiography. T1 mapping can identify the extent of fibrosis.^64,65 LGE is often present with patchy midwall involvement.⁶⁶ In a recent meta-analysis of 1063 patients, LGE was present in 60% with HCM and was closely associated with increased cardiovascular mortality.⁶⁷ LGE involvement may predict sudden cardiac death and its prognostic role is being further evaluated in a large, multicontinent NIH-funded study.⁶⁸

CMR structural assessment readily demonstrates the diffuse hypertrophy and biatrial enlargement in RCM. Amyloid can present with multiple LGE patterns, but the characteristic diffuse subendocardial involvement has 80% sensitivity and 94% specificity for amyloid and indicates a worse prognosis with increased 1-year mortality.^69,70 In similar fashion to Tc-pyrophosphate imaging, T1 mapping by CMR can differentiate between ATTR and AL amyloid subtypes.⁷¹ LGE uptake in a noncoronary distribution is also present in some patients with cardiac sarcoidosis (13% and 26% in studies by Nagai et al. and Greulich et al., respectively).^72,73 LGE has been shown to be one of the best predictors of prognosis in this population, including death, defibrillator shocks, and need for a pacemaker.^73,74 In contrast to the scar identified by CMR LGE evaluation, ¹⁸F-FDG PET imaging assesses active inflammation, which is more useful for assessment of a patient's current clinical status or response to therapy. There is some evidence that T2-W CMR imaging could play a similar role.⁷⁵

Valvular Heart Disease

Echocardiography remains the primary means of evaluating valvular heart disease, but CMR has a role in certain circumstances. CMR can provide a highly detailed anatomic assessment in any plane, assisting in mechanism evaluation.⁷ It is particularly helpful in evaluating right-sided valves, which are not evaluated well by echocardiography. CMR is beneficial in assessing valvular disease severity, particularly regurgitant volume and fraction by velocity-encoded imaging, in which it is superior to echocardiography.⁷⁶ The precise quantification of chamber volumes by SSFP imaging provides ideal assessment of the valvular disease consequences.

Cardiac Masses and Pericardial Disease

The combination of high spatial resolution for SSFP images and tissue characterization make CMR the ideal study to assess cardiac masses and pericardial disease. Combinations of T1- and T2-W imaging, perfusion, and LGE help identify the type of tumor and correctly classify 95% as benign or malignant.⁷⁷ Invasion of tumors from surrounding tissue is readily apparent. Spin echo sequences with T1-weighting visualize the extent and thickness of the pericardium, while T2-W images assess pericardial effusions and edema of the pericardium in cases such as inflammatory pericarditis.⁷⁸ Real-time SSFP images assess ventricular interdependence and myocardial tagging evaluates reduced myocardial–pericardial slippage in constriction.^78,79 LGE can be used both to assess inflammation in the pericardium, but also to assess myocardial involvement in myopericarditis.⁸⁰

The multiple sequences available in CMR imaging harness advantages including high spatial resolution and signal-to-noise ratio and limitless imaging planes to provide a comprehensive assessment of anatomy, function, perfusion, and tissue characterization of the myocardium and associated structures. CMR assesses ischemic heart disease with similar accuracy to nuclear MPI and has a robust and growing prognostic literature. In other conditions such as noninvasive evaluation of heart failure and cardiac masses, CMR is the gold standard. While some limitations with CMR imaging remain, including claustrophobia and restrictions in renal dysfunction, many are decreasing with hardware and software advances. Given these advances, CMR is playing an increasing role in the routine evaluation of cardiovascular disease.