Recently, the evaluation of myocarditis by CMR has become an important tool in nonischemic cardiomyopathies. In patients presenting with signs and symptoms of heart failure, but without demonstrable evidence of coronary artery disease (CAD), the search for a nonischemic mechanism is possible using CMR. The pattern of late hyperenhancement in these patients is distinctly different from the pattern characteristic of MI (see Fig. 15-1). Importantly, in cases of myocarditis it has been stated that late hyperenhancement is sensitive, but not specific. We note that the phenotype of the delay hyperenhancement distribution with relation to the myocardium gives clues concerning the underlying etiology. Whereas a MI always originates in the subendocardium, extending toward the midwall, and finally, if large enough, become completely transmural, myocarditis patterns are generally much more heterogeneous (see Fig. 15-2). Specifically, myocarditis patterns are generally less focal, much more patchy with skip lesions and, by definition, do not involve a coronary artery distribution (see Fig. 15-3). In some cases, the presentation of myocarditis can provide information about the long-term prognosis. This work is still maturing, but in general, larger, more focal patterns correspond with regional wall motion abnormalities, which will only marginally improve at 6 weeks, with residual contrast still present. Smaller lesions that are patchy or more diffuse, typically in concert with the clinical presentation, resolve in 6 weeks with a parallel improvement in left ventricular wall motion.

Most recently, several authors, including us, have made the interesting observation that the delayed hyperenhancement (DHE) signal in viral cardiomyopathy may demonstrate a linear stripe through the midwall of the myocardium with moderate signal intensity, which is generally a less intense signal than a DHE defined infarct (see Figs. 15-4 and 15-5). This pattern, sparing the epicardial and endocardial layers, may or may not involve more than one segment of the myocardium. Some investigators have performed electron microscopy and viral cultures, and have demonstrated a pattern that may, by purely imaging techniques, identify the offending pathogen. If the linear stripe occurs in the mid anterior and anteroseptal segment, there is a high likelihood that the agent is human herpesvirus 6 (HHV-6), whereas if it occurs in the posterior lateral wall (see Fig. 15-6) the agent is likely to be a parvovirus. CMR appears to be able to aid further in the clinical translation of this diagnostic strategy, because there have been limited observations that the septal pattern has a more ominous prognosis in that there is more left ventricular dysfunction, both globally and segmentally, whereas the posterior pattern has a more benign presentation with less deleterious impact on segmental function. Finally, depending on the location of the pattern, a prediction of interim left ventricular function can be made in that the septal pattern demonstrates limited recovery at several months after infection, whereas
the posterior pattern generally correlates with considerable improvement in left ventricular function (see Fig. 15-7).

Finally, in the absence of a CAD etiology, a finding of any one of the preceding patterns appears to be very helpful in explaining the etiology of the nonischemic cardiomyopathy. Upward of 40% of viral cardiomyopathies appear to have either a midwall stripe and/or a heterogeneous pattern, whereas 15% present with a bright pattern in the right ventricular septal insertion points. Therefore, when a nonischemic patient presents with a viral cardiomyopathy, there is a 50% chance of formally defining, or strongly indicating, the etiology. However, it should be noted that in fully 100% of this population, providing volumetric and left ventricular functional metrics is beneficial.
The finding of the midwall stripe has been proposed to be pathognomonic for a viral cardiomyopathy. However, we have recently seen several cases in which biopsy showed that the midwall stripe to be a forme fruste of idiopathic cardiomyopathy (see Fig. 15-4) and in another case a fatty pattern was seen. Early data supports that a finding of this pattern by itself may denote an adverse cardiovascular prognostic finding.