Multiparametric cardiovascular magnetic resonance characteristics and dynamic changes in myocardial and skeletal muscles in idiopathic inflammatory cardiomyopathy

This was a prospective observational cohort study involving 44 patients with newly diagnosed IIM who were consecutively enrolled from January 2017 to April 2018. The diagnosis of IIM was made according to the criteria of the European Neuromuscular Centre (ENMC) International Workshop (2004) [14]. All patients were assessed by clinical symptoms, laboratory tests, electromyography (EMG) and muscle or skin biopsy. For patients with muscle weakness with or without rash, the elevated creatine kinase (CK) levels and abnormal EMG findings indicated muscle damage, and muscle biopsy was performed to confirm the diagnosis. IIM was classified as polymyositis (without rash), dermatomyositis (with rash), or autoimmune necrotizing myositis (without rash) according to specific muscular pathology. For patients without muscle weakness but rash, IIM was classified as amyopathic dermatomyositis if CK and EMG were normal but skin biopsy demonstrated reduced capillary density and deposition of the C5b-9 complement membrane attack complex on small blood-vessels along the dermal-epidermal junction. The exclusion criteria were as follows: (1) patients with other autoimmune or inflammatory diseases; (2) patients who had been diagnosed as having IIM and were treated with corticosteroids or immunosuppressive drugs; (3) patients with any known cardiovascular diseases, including coronary artery diseases, cardiomyopathy, valvular diseases, long-term poor controlled hypertension or other chronic cardiac conditions; (4) patients who were unable to tolerate CMR examination or had a contraindications to CMR; (5) patients with renal impairment defined by glomerular filtration rate < 30 ml/min/1.73m²; and (6) patients with incomplete or poor-quality images. For all patients with IIM, baseline image acquisition was performed before the administration of regular treatment. Thirty age- and sex-matched healthy controls without any known cardiac disease or hypertension, diabetes mellitus, cerebrovascular disease, nervous system disease, or pulmonary or renal diseases were included. The first CMR scan was performed within 1 week of the IIM diagnosis and the follow-up CMR scans were performed among patients who received at least 1 year of regular treatment.

All participants underwent the same CMR protocol and provided written informed consent. The study complied with the Declaration of Helsinki and ethical approval was obtained from the local institutional ethics committee at West China Hospital (2016355).

All participants underwent CMR examinations on a 3 T scanner (MAGNETOM, Trio a Tim system; Siemens Healthineers, Erlangen, Germany) using an eight-channel phased-array body coil. All the images were acquired by breath-holding and ECG gating. A balanced steady-state free precession (bSSFP) sequence was used for cine images with continuous sampling from the basal to the apical levels on short-axis views and 2-, 3-, 4-chamber long-axis views. The typical scan parameters were as follows: repetition time (TR), 34 ms; echo time (TE), 1.3 ms; flip angle, 50°; field of view (FOV), 280 × 340 mm²; matrix size, 162 × 192; average temporal resolution, 35–45 ms; and slice thickness, 8 mm with no gap. Cardiac T2-weighted (T2w) images were acquired with the short TI inversion recovery (STIR) sequence on three short-axis views (basal, mid-ventricular and apical) and the 4-chamber long-axis view. The typical scan parameters were as follows: TR = 2 RR intervals during breath-hold; TE, 67 ms; flip angle, 180°; FOV, 280 × 340 mm²; matrix size, 166 × 256; and slice thickness, 10 mm. Phase-sensitive inversion recovery (PSIR) sequence with a segmented FLASH readout scheme was performed 10–15 min after gadolinium (Magnevist; Bayer Schering Pharma, Berlin, Germany) injection with 0.15 mmol/kg per bolus on consecutive short-axis views from the basal to apical level of LV and 2-, 3-, and 4-chamber long-axis views to obtain the LGE images. The typical scan parameters were as follows: TR, 700 ms; TE, 2.0 ms; delay time after the inversion pulse, 300–380 ms; flip angle, 20°; FOV, 260 × 340 mm²; matrix size, 116 × 192; and slice thickness, 8 mm. T1 mapping images were acquired on three short-axis and 4-chamber views, as in the previous sequence, by using the Modified Look-Locker Inversion-recovery (MOLLI) sequence with a 5b(3b)3b (b stands for heartbeat) scheme and a 4b(1b)3b(1b)2b scheme before and 10–15 min after the gadolinium injection. The typical scan parameters were as follows: TR, 300 ms; TE, 1.2 ms; flip angle, 35°; FOV, 270 × 320 mm²; matrix size, 144 × 256; and slice thickness, 8 mm. T2 mapping images were acquired by T₂-prepared single-shot bSSFP technique before administration of gadolinium on the short- and long-axis views identical to those associated with T1 mapping images. The typical scan parameters were as follows: TR, 240 ms; TE, 1.0 ms; flip angle, 12°; FOV, 320 × 340 mm²; matrix, 114 × 176; slice thickness, 8 mm; and T2 preparation pulse with 0, 30, 55 ms spin echo times.

We enrolled 46 IIM patients, among which two patients (4.3%) were excluded due to poor image quality. The study cohort consisted of 44 newly diagnosed IIM patients, and the demographic, clinical, and CMR parameters of IIM patients and healthy controls are summarized in Table 1. At baseline, IIM patients showed a significantly lower Hct, higher LV mass index and right ventricular (RV) diastolic volume index, and lower RV ejection fraction (RVEF) in comparison with age- and sex-matched controls. In the analysis of myocardial tissue characteristics, newly diagnosed IIM patients showed significantly longer native T1 (1275 ± 82 vs. 1199 ± 47 ms, p < 0.001) and T2 (44.2 ± 4.0 vs. 38.3 ± 2.9 ms, p < 0.001) relaxation times, and higher ECV (32.3% ± 6.2% vs. 26.6% ± 2.4%, p < 0.001) than normal controls. Eleven patients (25%) were LGE-positive with an average extent of 12.0% ± 3.1% by the threshold of 3SD. For the evaluation of skeletal tissue characteristics, IIM patients showed significantly longer native T1 (1287 ± 1499 vs. 1107 ± 70 ms, p < 0.001) and T2 (34.4 ± 3.1 vs. 20.6 ± 1.4 ms, p < 0.001) relaxation times, and higher ECV (31.1% ± 12.8% vs. 13.6% ± 3.1%, p < 0.001) at baseline compared to normal controls.

The increased myocardial native T1, T2, and ECV values were significantly correlated with a larger LV volume index, higher LV mass index, and lower LV ejection fraction (EF) Table 2. Meanwhile, elongation of the myocardial T2 relaxation time was significantly correlated with a decrease in RVEF. The myocardial native T1 (r = 0.547, p < 0.001), T2 (r = 0.406, p = 0.005), and ECV (r = 0.313, p = 0.037) values showed moderate positive correlation with the log NT-proBNP. Among the multiparametric tissue characteristics of skeletal muscles, we found that skeletal T2 relaxation times showed significant positive correlation with the log ESR (r = 0.430, p = 0.004), log CK-MB (r = 0.471, p = 0.001), and log cTnT (r = 0.563, p < 0.001) levels, while skeletal native T1 and ECV did not show significant correlations with the levels of laboratory biomarkers. The results are presented in Fig.2.

In the study, we also found that elevation of myocardial T2 value was significantly correlated with the skeletal T2 value (r = 0.317, p = 0.036), while no significant correlation was found among T1 or ECV values on myocardium with those on skeletal muscles. The scatterplots are presented in Fig. 3.

During the follow-up, 28 IIM patients were arranged to consecutively undergo repeated scans with a median CMR interval of 14.5 mo (interquartile: 13.2–15.5 mo). Those who have not been treated regularly (n = 2), had a treatment time of less than 1 year (n = 6), or were reluctant to undergo a second CMR for personal reasons were excluded from the subgroup analysis. All 28 patients reached clinical remission during the CMR follow-up, with no patients demonstrating proximal muscle weakness or skin rash. Patients also showed significantly decreased CRP, CK, CK-MB, TnT and NT-proBNP levels and increased Hct during follow-up in comparision with those at baseline. While a significant decrease in LV mass index was observed (baseline: 51.2 ± 12. vs. follow-up: 45.2 ± 10.2 g/m², p = 0.011), no significant changes were observed in the biventricular volume index or ejection function. Seven LGE-positive patients at baseline were still LGE-positive at CMR follow-up with a trend of a reduced extent of LGE with no statistically significant difference (baseline: 11.8 ± 2.9% vs. follow-up: 8.1 ± 2.0%, p = NS). Patients showed significantly decreased myocardial T1 mapping (baseline: 1289 ± 53 vs. follow-up: 1262 ± 59 ms, p = 0.013), T2 mapping (baseline: 44.4 ± 4.0 ms vs. follow-up: 41.8 ± 4.2 ms, p = 0.001), and ECV (baseline: 30.7 ± 4.1% vs. follow-up: 27.6 ± 4.4%, p = 0.001) during the follow-up CMR. However, the myocardial native T1 and T2 values were still significantly higher than those for normal controls. The comparison between baseline and follow-up CMR of the subgroup IIM patients is summarized in Table 3 and Fig. 4.

In the assessment of skeletal tissue characteristics, patients showed significantly decreased skeletal native T1 relaxation time (baseline: 1268 ± 103 vs. baseline: 1154 ± 73 ms, p < 0.001), T2 relaxation time (baseline: 35.5 ± 3.5 follow-up: 29.7 ± 2.8 ms, p < 0.001) time, and ECV (baseline: 29.2% ± 11.1% vs. baseline: 21.9% ± 11.2%, p = 0.011), with the native T1, T2 and ECV values of skeletal muscles still significantly higher than those of normal controls. The results are summarized in Table 4 and presented in Fig. 4.

Table 4 shows the inter- and intra-reproducibility of the measurements obtained with native T1 mapping, T2 mapping, and ECV for myocardium and skeletal muscles. We found the reproducibility of skeletal mapping parameters (CoV ranging from 2.12 to 4.43%) to be slightly worse than those of myocardial mapping parameters (CoV ranging from 0.86 to 2.19%).

It is not uncommon for the myocardium to be affected in systemic inflammatory diseases. In this context, the clinical value of CMR parametric mapping techniques in quantifying myocardial tissue alterations has been gaining recognition in many diseases [17]. Ntusi et al. demonstrated that T1 mapping could provide added value as a biomarker for diffuse myocardial fibrosis in rheumatoid arthritis patients [18]. Biesbroek et al. found that ECV and T1 mapping were significantly associated with the disease activity in ankylosing spondylitis patients and may be a potential marker for disease monitoring [19]. In the present study, newly diagnosed IIM patients showed significantly elevated T1, T2, and ECV values in comparison with normal controls, and these findings are consistent with recent CMR studies of IIM patients [10, 12]. We also found that the myocardial native T1, T2 and ECV values were significantly correlated with the NT-proBNP level, which may suggest that.

quantitative mapping parameters can be used as potential imaging markers for detecting subclinical cardiac involvement in newly diagnosed IIM patients. The significant correlation between myocardial ECV and serum CRP level, as well as the significant correlation between myocardial and skeletal T2 values, may indicate that the degree of myocardial inflammation is positively correlated with the systematic inflammation level. In addition, for the assessment of skeletal tissue characteristics, elevated skeletal T2 relaxation time was shown to be correlated with biomarkers such as the ESR, CK, CK-MB, and cTnT levels. Our findings suggested that the skeletal inflammation reflected by CMR multiparametric mapping techniques is related to the disease activity.

CMR mapping technologies have shown good reproducibility and sensitivity in dynamic monitoring for different diseases. Hinojar et al. found that lupus myocarditis patients had elevated myocardial native T1 and T2 relaxation times and showed a significant decrease after intensive therapy, which suggests an attenuation of myocardial inflammation [20]. Bohnen et al. found that native myocardial T1 and T2 measurements provide excellent performance for monitoring the healing of myocarditis with a significant decline based on the stage of acute myocarditis [21]. Spieker et al. also found T2 mapping to be a good tool for monitoring myocardial inflammation in patients with suspected acute myocarditis [22]. Alkhalil et al. demonstrated that the anatomical area at risk may best be quantified by T1 mapping undertaken in the first week after reperfusion [23]. However, some previous studies have presented discrepant findings regarding the dynamic changes associated with cardiac involvement in IIM patients. Allanore et al. studied four patients and found markedly reduced LGE areas and decreased hypokinesia after 6 months of corticosteroid and immunosuppressive therapy [24]. In contrast, Mavrogeni et al. found that the signs of myocarditis were still positive on CMR reevaluation after 3 months of steroid treatment with clinical remission [25]. Péter et al. reevaluated IIM patients 3 months after starting treatment by using echocardiography and found a significant improvement in LV and RV systolic function, despite also observing deterioration of diastolic dysfunction [25]. The uncertain conclusions in previous studies may be attributable to differences in patient selection, approaches used to evaluate cardiac abnormalities, different follow-up intervals, and the insufficient sample sizes.

In the present study, we found that newly diagnosed IIM patients had significantly elevated T1, T2, and ECV values on both myocardium and skeletal muscles, which are consistent with the findings of previous studies [10, 12]. In addition, we also demonstrated for the first time that multiparametric CMR can detect dynamic changes in both myocardial and skeletal tissue characteristics after effective treatment.

Although the cellular and molecular pathophysiology underlying the myocardial involvement in IIM patients is not clear, the increased T2 relaxation time at baseline suggested that myocardial inflammation might be a major pathophysiological change in IIM patients [26]. Consistent with this speculation, we found a significant decrease in the myocardial T2 relaxation time at follow-up, which paralleled the remission of clinical symptoms and improvements in serum inflammatory biomarkers. In addition, the significant reduction in the myocardial mass index could be partially explained by the remission of myocardial inflammation. Skeletal T2 also showed a significant decrease under clinical remission. This finding suggests that IIM patients have both myocardial and skeletal inflammation at the newly diagnosed period, and that acute inflammation and edema are reversible after effective treatment. Both myocardial and skeletal T2 values served as feasible and sensitive markers for detecting myocardial and skeletal inflammation and monitoring the therapeutic effects in IIM patients. Although we found that the clinical application of these changes may be limited due to their small absolute change values, the study still provides clues for the potential application of individualized dynamic monitoring of the effect of treatment.

We found that IIM patients had elevated T1 values at both newly diagnosed and clinical remission stages in comparison with normal controls. While a previous study found that native T1 shows high diagnostic accuracy in discriminating the acute and convalescent stages of myocarditis [27], the inconsistent results of the current study indicate that the myocardial involvement in IIM patients may not only include acute inflammation but also diffuse fibrosis. The histological studies had demonstrated that the native T1 mapping relaxation time is affected by diffuse fibrosis [28, 29] and myocardial inflammation [30]. Whether the myocardial diffuse fibrosis in IIM patients is permanent or reversible over a longer term is not clear, and observational studies with a longer follow-up period are needed.

One previous study had demonstrated that in cases with coexisting inflammation and diffuse fibrosis, the ECV would reflect the sum of both pathologies [31]. ECV has shown excellent discriminative value in differentiating myocarditis patients from healthy subjects [21, 32], and declined significantly with the recovery of myocarditis [21]. In our study, myocardial ECV provided good discriminative value between newly diagnosed IIM patients and normal controls, with a significant decrease in the ECV of both the myocardium and skeletal muscle at follow-up CMR. Thus, ECV is also a promising tissue characterization marker for detecting subtle abnormalities and monitoring the recovery from inflammation of myocardial and skeletal tissue.

In contrast to the findings for quantitative T1, T2, and ECV techniques, the T2W images showed limited ability to distinguish IIM patients from normal controls. This lack of sensitivity may be explained by the concurrent abnormality on skeletal muscles [33]. The T2 mapping technique offered the advantages of quantitative measurement of T2 relaxation times without the need for skeletal muscles as reference. Our finding suggests that T2W images have limited value for evaluation of IIM patients, consistent with the findings of a previous study [26].

In the current study, IIM patients with positive LGE at baseline still showed LGE during follow-up, without a significant decrease in the LGE extent. This may be explained by the fact that LGE reflects irreversible myocardial injury [34]. Previous studies have indicated that the presence of LGE may reflect chronic and active myocardial inflammation [34, 35].

In the current study, we found that elevated myocardial native T1, T2, and ECV significantly correlated with LV systolic dysfunction and enlargement, while the myocardial T2 value also correlated with RV dysfunction. These results suggest that myocardial inflammation and the possible presence of myocardial fibrosis in IIM patients can adversely impact cardiac function and structure.

This study had several limitations. First, this is a single-center study with a relatively small number of patients. Further studies with larger sample sizes and external validation are needed. Second, not all IIM patients enrolled at baseline underwent follow-up CMR scans. Although this could potentially lead to study bias, we focused on the subsequent observation CMR scans to analyze the dynamic changes in both myocardial and skeletal tissue characterizations. Third, the imaging data for the myocardium were not supported by biopsy findings. Future studies on the prognostic value of multiparametric CMR in IIM patients are required to validate these findings.