This retrospective study was approved by the Institutional Review Boards of our hospital(No.2022-L-128). Forty-seven patients (35 males, 12 females) were retrospectively enrolled from 2015 to May 2019 in our hospital. All the patients had been clinically suspected acute myocarditis (mean symptom duration before referral: 5.2 ± 4.1 days, < 14 days for all the patients) and underwent CMR examination. They were diagnosed with myocarditis for having clinical symptoms and according to the corresponding clinical criteria based on 2013 ESC guidelines (15). Clinically suspected myocarditis refers to symptomatic patients (chest pain, dyspnea, palpitation, or other relevant cardiac symptoms) based on one or more diagnostic criteria (abnormal electrocardiography, elevated troponin, and functional or structural abnormalities confirmed by echocardiography or MR) and asymptomatic patients based on two or more of the above criteria. Coronary artery disease was ruled out prior to CMR in all patients (Table 1). The control group consisted of 39 age- and gender-matched healthy controls (HCs) (26 males, 13 females), who were selected for their uneventful medical histories, absence of symptoms indicative of cardiovascular dysfunction, lack of abnormalities in electrocardiograms, echocardiography and CMR, and no history of inflammatory diseases.
CMR was performed on a 3.0-T MR system (Achieva, Philips Healthcare, Best, Netherland) using a standard 16-element cardiac phased array coil and a four-lead vectorcardiogram. For functional analysis, electrocardiography-gated balance turbo field echo (b-TFE) cine images were obtained in the short-axis, four-chamber, and two-chamber views (TR, 3.2 ms; TE, 1.53 ms; FOV, 320 mm × 320mm; reconstruction matrix, 320 × 320; flip angle, 45°; slice thickness, 8 mm; space, 0 mm). Edema-sensitive black-blood T2-weighted images with fat saturation (triple inversion recovery turbo spin echo sequences with inversion pulses for fat and blood suppression) were acquired in the short-axis orientation covering the entire left ventricle (TR, 1200 ms; TE, 60 ms; FOV, 320 mm x 320 mm; reconstruction matrix, 352 × 352; slice thickness, 8 mm; space, 2 mm). Late gadolinium enhancement (LGE) imaging was performed 10 minutes after the injection of 0.1 mmol/kg gadobutrol (Gadovist, Bayer Healthcare, Leverkusen, Germany) with inversion time (300–340 mm) adjusted according to a Look-Locker inversion recovery-prepared T1-weighted phase-sensitive inversion recovery sequence (TR, 6.1 ms; TE, 3.0 ms; TI, 400 ms; FOV, 320 mm × 320 mm; reconstruction matrix, 320 × 320; flip angle, 25°; slice thickness, 8 mm; space, 2 mm) in the short-axis, four-chamber, and two-chamber views, which corresponded with cine images. CMR scanning was performed according to the standardized protocols recommended by the Society for Cardiovascular Magnetic Resonance (SCMR).(16)
Three radiologists experienced in CMR diagnosis blindly analyzed the data, performed the measurements, and reached agreement regarding the consequences.
Cardiac function analysis
Cardiac function analysis was performed offline based on the acquired b-TFE cine images by using dedicated software (CVI 42 v. 5.6, Circle Cardiovascular Imaging Inc., Calgary, AB, Canada). Endocardial and epicardial contours of the left ventricle were manually delineated at the end-systolic and end-diastolic phases using a three-dimensional short tool to calculate volume changes and left ventricular ejection fraction (LVEF). Based upon the body surface area, LV end-diastolic volume index, LV end-systolic volume index, and myocardial mass index were quantified. Patients with acute myocarditis were divided into two subgroups according to LVEF, including the impaired-LVEF group (LVEF < 55%; n = 12) and the preserved-LVEF group (LVEF ≥ 55%; n = 35) (9).
The updated LLC
Image analysis of the updated LLC was performed using the CVI 42 software (CVI 42 v. 5.6, Circle Cardiovascular Imaging Inc., Calgary, AB, Canada). The myocardium was divided into 16 segments according to the American Heart Association segmentation (17). Every segment was evaluated for the following tissue characterizations: 1) T2-based marker for myocardial edema with either T2-weighted imaging or T2 mapping, and 2) T1-based marker for associated myocardial injury: one of the three methods, namely, LGE, T1-mapping or extracellular volume (ECV).
CMR diagnosis of myocarditis was based on the edema-sensitive CMR (T2-weighted images or T2 mapping) and at least one additional T1-based tissue characterization technique (8). As T1 and T2 mapping and ECV are not routine sequences in our institution, the approach chosen by our study included only T2-weighted imaging and LGE. T2-weighted imaging is identified visually on T2-weighted black-blood imaging and by calculating the T2 ratio of ≥ 1.9 (signal intensity normalized to skeletal muscle in the same slice) (8, 18). The patterns of LGE are commonly and typically located in the subepicardial and midmyocardial regions (8). All the myocardial segments of the enrolled patients were divided into three subgroups based on the severity of their myocardial injury (19), viz. segments with non-involvement (SNi; n = 509), segments with edema (SE; n = 89), and segments with both edema and LGE (SE+LGE; n = 154). The SE subgroup was localized using T2-weighted images; the SE+LGE subgroup was localized using both T2-weighted images and LGE images; and the SNi subgroup was considered as normal in comparison with the two preceding subgroups. A cohort of segments of HCs (SHCs; n = 272) served as the control group.
Myocardial strain analysis using CMR-FT
CMR-FT was performed offline based on the acquired b-TFE cine images using CVI 42 software (CVI 42 v. 5.6, Circle Cardiovascular Imaging Inc., Calgary, AB, Canada). Endocardial and epicardial contours were drawn manually in the end-diastolic phases, and myocardial strain was automatically tracked by CVI 42 throughout the cardiac cycle. Global peak longitudinal strain (GLS) was averaged from the measurements of the two-, three-, and four-chamber views. Circumferential (GCS) and radial (GRS) peak strain parameters were determined in the short-axis view covering the entire left ventricle (Fig. 1). Every segment was evaluated for myocardial strain parameters to obtain segmental peak radial strain (PRS), peak circumferential strain (PCS), and peak longitudinal strain (PLS).
Normality was tested with the Kolmogorov-Smirnov test. Continuous variables were presented as means ± standard deviations and compared using the Student’s t-test for normally distributed data or the Mann-Whitney U-test for non-normally distributed data. Parametric data of more than two groups were compared using one-way analysis of variance testing. Post-hoc testing was performed with the least significant difference test (homogeneity of variance) or Dunnett’s T3 test (heterogeneity of variance). Categorical group data presented as percentages were compared using the chi-squared test or Fisher’s exact test, as appropriate. Diagnostic performance of the strain parameters was analyzed by plotting receiver operating characteristic curves and comparing the areas under those curves. Cutoff values were chosen by maximizing reclassification accuracy for the predictive variables, and reclassification sensitivity, specificity, and accuracy were calculated. To combine the single predictive variables, scores were derived from logistic regression analysis. The level of statistical significance was set to p < 0.050. Statistical analysis was performed using SPSS software(v.25.0, IBM SPSS Statistics, Armonk, NY, USA).