Due to the diversity of its clinical presentation, the diagnosis of myocarditis remains one of the most challenging in cardiology.1,2 The current gold standard, endomyocardial biopsy, is limited by its periprocedural risks and low diagnostic sensitivity due to sampling error.1,2 Therefore, cardiac magnetic resonance (CMR) has now become the gold standard non-invasive diagnostic modality in suspected myocarditis. Due to its unique ability to combine morphological and functional imaging with myocardial tissue characterization, CMR enables the detection of the typical features of acute inflammation, such as edema, hyperemia, and necrosis, that are usually associated with myocardial dysfunction.1–3 However, the current Lake Louise criteria for CMR-based diagnosis of myocarditis4 still lack diagnostic accuracy.3 Recently, novel quantitative CMR techniques such as tissue tracking (TT)-based strain analysis have emerged as potential novel diagnostic tools,5,6 aiming at improved diagnostic accuracy in myocarditis.7

The aims of this study were to characterize myocardial deformation alterations in patients with myocarditis by applying CMR-TT, and to correlate systolic and diastolic deformation mechanics with left ventricular ejection fraction (LVEF), extent of regional wall motion abnormalities (WMA), and extent of late gadolinium enhancement (LGE).

In this study, we assessed the value of CMR-TT derived myocardial deformation parameters in patients with myocarditis. The main findings of our study were: (1) TT analysis was feasible in the majority of myocarditis patients, with good inter- and intraobserver reproducibility; (2) all myocardial deformation parameters were significantly correlated with LVEF and LV WMA; (3) myocardial deformation parameters were considerably reduced, compared to previously published normal values; (4) global peak systolic radial strain, strain rate and displacement, and global peak systolic circumferential strain were the only parameters to be closely correlated with extent of LV LGE.

Tissue-tracking technologies such as speckle-tracking echocardiography and CMR-TT have enhanced the noninvasive assessment of myocardial deformation in clinical practice. The TT technique raises the possibility of retrospective analysis of previously acquired SSFP datasets. With this post-processing approach, no additional dedicated CMR scans are required. Although good agreement has been established between speckle tracking echocardiography and CMR-TT,10,11 difficulty in obtaining adequate echocardiographic views, observer dependency, signal noise and angle dependency can hinder strain assessment by echocardiography.12 Hence, tomographic imaging modalities like CMR are a clinically valuable alternative to overcome such shortcomings. In our study, TT software delivered outputs of myocardial strain, strain rate, velocity and displacement parameters and was able to track existing cine sequences in most patients (90.7% of all myocarditis patients).

While comparative analysis shows that measures of deformation parameters are similar between CMR-TT and speckle-tracking echocardiography, it also shows that TT analysis has superior reproducibility.11 Regarding variability, we found good inter- and intraobserver agreement for all deformation parameters. Global systolic deformation parameters showed an inter- and intraobserver CoV ranging from 1.7% to 10.6%, which is comparable to previously reported values.11,13 Although slightly lower agreement was achieved in diastolic strain rates (CoV ranging from 4.6% to 13.0%), there are currently no data on the reproducibility of diastolic function by CMR-TT analysis.

In this study, it is worth noting that there was a linear relation between all strain parameters and LVEF as well as the extent of regional WMA, as shown in Figures 3 and 5. Although these relations are mainly weak, it is clear that they may be a contributing reason why TT analysis is additive to segmental changes in contractility and to global myocardial dysfunction. The absence of follow-up data precludes determination of the clinical prognostic value of CMR-TT. In addition, a clinically meaningful cut-off value remain to be assessed. This might be an objective of future trials.

Normal reference values for TT-derived systolic and diastolic deformation parameters have been published. In a contemporary systematic review and meta-analysis of 18 studies, the pooled mean of global longitudinal strain derived from three apical views was -19.4% and the global radial and circumferential strains derived from three SAX were 39.1% and -23.6%, respectively.9 In our myocarditis population, mean global longitudinal, radial and circumferential strain values were significantly lower than in the published reference values. Moreover, strain parameters were reduced even when only patients with preserved LVEF were considered. This observation is in agreement with a previous CMR-TT study7 that detected significant group differences between myocarditis patients with preserved ejection fraction and control subjects. Overall these results thus suggest that myocardial strain parameters represent new quantitative indices of cardiac deformation that are thought to be capable of detecting subtle alterations in myocardial function, and that the addition of global peak systolic strains to LVEF may improve the diagnostic accuracy of CMR for myocarditis.

Further systolic deformation parameters such as strain rate, velocity and displacement deformation parameters are rarely assessed. A study assessing a cohort of 145 healthy volunteers reported slightly lower ranges than those found in our population, however different software was used,13 which precludes direct comparisons.

Diastolic dysfunction, as assessed by echocardiography, has been reported to be prevalent in acute myocarditis.14 In a seminal paper, Eurlings et al.14 reported that inflammation induces fibrosis, cardiomyocyte stiffening and endothelial dysfunction, which could explain the impairment of myocardial diastolic properties in acute myocarditis. Assessment of diastolic function by CMR-TT has already taken its first steps in other cardiomyopathies, notably in hypertrophic cardiomyopathy (HCM). In a study of patients with HCM, Nucifora et al.15 showed that LV diastolic mechanics was influenced by the extent of replacement and interstitial fibrosis. The present study is, to the best of our knowledge, the first to investigate diastolic function parameters by CMR-TT in the setting of myocarditis. In our study, diastolic deformation parameters, as determined by strain rate, were significantly related to the extent of WMA and LVEF. In view of the otherwise limited available data, understanding of the role of diastolic function parameters, as assessed by CMR-TT, still remains elusive and our results should encourage future studies to reappraise the possible importance of diastolic function parameters in the setting of myocarditis.

We also revealed that worse global peak systolic radial strain, strain rate and displacement, and global peak systolic circumferential strain showed a significant correlation with the extent of LV LGE, regardless of the type of LGE pattern. Similarly, a case-control study found that global peak systolic circumferential strain displayed good accuracy in detecting LGE, a value worse than -20.5% detecting LGE with 73% sensitivity and 77% specificity.16 TT analysis may therefore serve as a surrogate for detection of fibrotic alterations of the myocardium. This might also have significant prognostic value in patients with suspected myocarditis, since the presence of LGE has been found to be a predictor of major adverse cardiac events.17,18 Although LGE has been shown to be related to poor prognosis and TT analysis identifies the presence of LGE, the direct prognostic implications of CMR-TT assessment in myocarditis remain to be established.

Other applications of myocardial deformation analysis could be useful in patients in whom gadolinium administration is inadvisable (for example in renal insufficiency), enabling indirect quantification of myocardial fibrosis, as administration of contrast media is unnecessary for this analysis. Furthermore, tissue characterization by CMR-TT may also overcome the technical limitations of the Lake Louise criteria, as T1 and T2 mapping parametric mapping techniques appear to do.19 In view of all these advantages, it can be expected that CMR-TT will be increasingly used in published studies.

The present study has several limitations. Firstly, there are the inherent drawbacks of the retrospective study design and potential selection bias, as only patients who underwent CMR were included. Future larger multicenter studies with prospective selection of patients with myocarditis should be performed in order to adequately validate the present report. Secondly, endomyocardial biopsy, the reference standard, was not systematically performed to diagnose myocarditis. Instead, the presence of myocarditis was defined by CMR-based tissue characterization, combining typical clinical features, elevated biomarkers and exclusion of coronary artery disease, which have already proved to have good diagnostic accuracy.20 Thirdly, WMA and LGE were assessed visually and quantified manually as the percentage of segments involved. Better reproducibility is usually obtained for LGE quantification using available software that delineates areas of interest, automatically traced in all slices and quantified as proportion of total LV myocardial mass,21 however there is no current consensus on the best method of LGE quantification.22 Finally, the long waiting times from symptom onset to CMR acquisition (median 11 days; IQR: 2-138 days), mainly due to low resources and high demand, may have influenced the results obtained. Supplementary material presents an analysis of the 19 patients in whom CMR was performed within 15 days of the acute event (Supplementary Tables 1 and 2).

CMR-TT myocardial strain analysis is an objective method for the quantification of myocardial function. Patients with myocarditis showed slightly reduced myocardial strain, and the deformation parameters correlate with LVEF, along with WMA and extent of myocardial fibrosis. Future studies are needed to determine the clinical relevance of these findings for the management of myocarditis patients.

The authors have no conflicts of interest to declare.