Current patient selection for cardiac resynchronization therapy (CRT) relies on global measures of electrical conduction delay (the electrocardiographic QRS duration) and overall systolic function (the left ventricular ejection fraction) combined with clinical assessment of functional status (New York Heart Association class). Approximately one-third of patients diagnosed with heart failure demonstrate electrical conduction abnormalities ^1,2,3 by surface electrocardiography, usually in the form of a complete left bundle branch block (LBBB) or nonspecific intraventricular conduction delay (IVCD).⁴ Such electrical conduction abnormalities have been shown to be associated with structural changes resulting in impaired left ventricular function⁵ as well as increased morbidity and mortality.⁴ In large multicenter trials, responders to CRT have experienced improvements in exercise tolerance, heart failure symptoms, heart failure hospitalizations,^6,7 and mortality.^8,9 These benefits of CRT have been accompanied by reductions in LV chamber size, improvements in ejection fraction, and decreases in the severity of mitral regurgitation.^10,11

If dyssynchrony is a measurable condition that identifies patients who should benefit from resynchronization, it may have different facets depending on the technique used for measurement. For instance, QRS duration and tissue-Doppler indices (TDI) of dyssynchrony bear only modest overlap. In a study of heart failure patients with LVEF <35%, up to 40%, of patients with QRS duration >120 ms did not exhibit left ventricular dyssynchrony by TDI, and 27% of patients with TDI >60 ms between the septum and lateral wall had normal QRS duration.¹² Some studies have shown a better response to CRT in patients with wider QRS,^{13, 14} while others have shown baseline QRS duration to be a poor predictor of long-term response.^{6, 15,16,17} Achilli et al., for example, found a favorable response to CRT seen in patients with dyssynchrony as assessed by echocardiography and QRS <120 ms.¹⁶

Thus, a potential role for cardiac magnetic resonance (CMR) in patient assessment for CRT emerges from two observations. First, approximately one-third of patients undergoing CRT device implantation do not respond to therapy.^{6, 18} Second, the currently used electrocardiographic and echocardiographic measures of electrical and mechanical dyssynchrony perform poorly in discriminating responders from nonresponders. However, postdevice CMR remains relatively contraindicated pending uniform use of MR-compatible lead systems, limiting the use of CMR in CRT to evaluating patients prior to device placement. Here, too, a further limitation becomes apparent when recognizing that many patients referred for CRT already have implanted pacemakers or defibrillators that currently preclude elective magnetic resonance examination. Despite these limitations, CMR identifies important myocardial characteristics not feasible with other approaches that, when feasible, should be performed to improve patient selection for CRT. In addition to the assessment of dyssynchrony, CMR allows for the determination of the size, shape, and function of the LV as well as the presence and transmurality of scar tissue in the location where the LV lead should be positioned.^19,20 Better patient selection should improve outcomes twofold for patients with heart failure as it 1) reduces the cost and complications associated with invasive device implantation in patients who are unlikely to respond and 2) identifies patients who may benefit from CRT yet are deemed ineligible according to conventional ECG and echo parameters. In this chapter we will discuss CMR-based assessment of myocardial strain and their applications to dyssynchrony measurement for CRT. Specific CMR techniques including myocardial tagging, velocity-encoded cine, and late postgadolinium enhancement (LGE) will be examined. We have also included details of a novel approach developed in our laboratory that uses frequency domain-based segmentation of routinely acquired cine CMR images that may overcome some of the variability seen in time domain-based dyssynchrony evaluation. Finally, the safety of CMR in patients with implanted devices will also briefly be discussed.