The ethical board of Heinrich-Heine University Düsseldorf approved the present study (application number 4307). All participants signed informed consent. The study complies with the declaration of Helsinki.
Forty-six patients meeting inclusion criteria of sAMC according to ESC guidelines were prospectively enrolled between 2013 and 2016: clinical symptoms <14 days (dyspnoea, chest pain, fatigue) and one additional diagnostic criterion in terms of ECG abnormalities (ST-changes, conduction defects), hsTNT-elevation or new global/regional wall motion abnormalities and exclusion of coronary artery disease.
Exclusion criteria of the study were 1) coronary artery disease (coronary stenosis >50% proven by angiography 2) pre-existing other cardiac disease that could explain symptoms and 3) contraindications against CMR.
All patients underwent CMR between day 0 and day 5 after initial presentation at hospital. Right ventricular EMB was conducted whenever feasible (n = 40, 72 ± 12 h after presentation).
Patients were scheduled after informed consent for a follow-up assessment 6 to 18 months after the initial hospitalisation. Left ventricular systolic function at follow-up was assessed by CMR or 3D-echocardiography. Sixty age, sex and cardiovascular risk factor -matched volunteers served as controls.
CMR imaging and analysis
CMR was performed using a 1.5 tesla scanner (Achieva, Philips, Best Netherlands) with a 32-channel phased array coil. After scout and reference scans, CINE-loops in continuous short axis slices covering the whole ventricle were analyzed to calculate left ventricular function and volumes.
For edema imaging, a T2-weighted TSE-STIR sequence was used and T2 Mapping was conducted with a GRASE sequence in three short axis slices (basal, midventricular and apical). This sequence combines the TSE and echo-planar imaging methods by using a train of refocusing 180° pulses and an odd number of additional gradient echoes for each spin echo. This sequence was used with cardiac triggering and respiration navigator gating with the following parameters: TR = 1 RR interval, number of echo images = 15, echo spacing 10 ms, leading to an echo train of 150 ms, number of gradient echoes for segmented acquisition = 3 (EPI factor), FA = 90°, spatial resolution: 2 × 2 × 10 mm3, parallel imaging (SENSE) with an accelerating factor of 2, k-space data acquired with Cartesian encoding scheme. For blood saturation a double inversion (black-blood) pulse was used .
Late Gadolinium Enhancement (LGE) imaging was performed eight to ten minutes after gadolinium contrast injection (ProHance®, Bracco Imaging, 0.2 mmol/kg) using a 3-dimensional–gradient spoiled turbo fast-field-echo sequence with a non-selective 180° inversion-recovery pre-pulse triggered to end-diastole acquired in the short axis, 4–3- and 2-chamber view to cover the whole ventricle.
For post-processing, left ventricular function and LGE were assessed off-line using a commercial software (Extended Workspace, Philips Healthcare). LGE was evaluated by visual assessment and assignment of certain myocardial parts according to the 17-segment model of the AHA .
T2 maps were generated according to Bönner et al. . T2 maps were generated off-line using software based on the graphical programming language LabVIEW (National Instruments, Austin, TX). An exponential decay curve was fitted to the intensity decline of each pixel within the images obtained from the multi echo sequence. We used a 2-parameter (amplitude and damping) fit model with a constant off-set. The bias was calculated from the noise of all echo images and assumed to be constant, so that the problem could be linearized and the regression coefficient (R2) could be used as a goodness-of-fit parameter in order to exclude accidental values. If R2 was not within a tolerance interval chosen to be 0.7-1, the corresponding T2 value was not considered for further calculations. The resulting T2 constants were colour-coded using the spectral look-up table. Endo- and epimyocardial region of interests (ROI) were drawn manually in the first echo image of the echo-train.
The myocardial ROI was segmented according to the AHA 17-segment model  and T2 values were calculated for 16 segments (segment 17 was not considered). In order to detect regional edema more reliably we quantified the myocardial fraction with T2 time over several cut-offs. ROC analysis suggested that a selected cut-off over 80 ms was the best parameter to distinguish between biopsy-proven inflamed myocardium and healthy myocardium. This cut-off was determined in a previous study including 26 patients with bpAMC and 60 healthy controls . The myocardial fraction exceeding 80 ms was quantified and expressed as percentage of the whole myocardium and graphically highlighted in white. To avoid spill over of high T2 values due to epicardial fat and endocardial slow flow artefacts a T2 time limit of 110 ms was chosen in the final absolute T2 time and fractional quantification.
Endpoints and clinical follow-up
Patients were followed over time and a combined clinical endpoint was defined as a composite of major adverse cardiac events (MACE) and hospitalisation due to heart failure. MACE was defined as a composite of all-cause death, cardiac death, cardiac transplantation and ventricular assist device implantation. The length of follow-up was determined by occurrence of an endpoint or the last clinical follow-up examination.
Statistical analysis was performed with Sigma Stat2010 (Systat GmbH) and GraphPad Prism 7. Unless otherwise stated, data are presented as mean value ± standard deviation. Data were statistically analysed by the paired or unpaired Student’s t-test. Welch’s correction was used when unequal SD was assumed. Bonferroni correction was applied for multiple comparisons. Fisher’s exact test was used to examine the significance of the association between two kinds of classification. Receiver operator characteristic (ROC) analysis was performed to generate threshold values with respect to optimal sensitivities, specificities and areas under the curve (AUC). Kaplan-Meier curves were calculated for visualizing cumulative event free survival of patients with global and regional T2 time exceeding several cut-offs. P-values below 0.05 were assumed to be significant.