Author + information
- Received November 16, 2015
- Revision received February 3, 2016
- Accepted February 8, 2016
- Published online April 19, 2016.
- Philipp Lurz, MD, PhDa,∗ (, )
- Christian Luecke, MDb,
- Ingo Eitel, MDc,d,
- Felix Föhrenbach, MDa,
- Clara Frank, MDb,
- Matthias Grothoff, MDb,
- Suzanne de Waha, MDc,d,
- Karl-Philipp Rommel, MDa,
- Julia Anna Lurz, MDe,
- Karin Klingel, MDf,
- Reinhard Kandolf, MDf,
- Gerhard Schuler, MDa,
- Holger Thiele, MDc,d and
- Matthias Gutberlet, MDb
- aDepartment of Internal Medicine/Cardiology, University of Leipzig—Heart Center, Leipzig, Germany
- bDepartment of Diagnostic and Interventional Radiology, University of Leipzig—Heart Center, Leipzig, Germany
- cUniversity Heart Center Luebeck, University of Schleswig-Holstein, Medical Clinic II (Cardiology, Angiology, Intensive Care Medicine), Luebeck, Germany
- dGerman Centre for Cardiovascular Research (DZHK), Partner Site Hamburg/Kiel/Lübeck, Luebeck, Germany
- eDepartment of Electrophysiology, University of Leipzig—Heart Center, Leipzig, Germany
- fDepartment of Molecular Pathology, University Hospital Tuebingen, Tuebingen, Germany
- ↵∗Reprint requests and correspondence:
Dr. Philipp Lurz, Department of Internal Medicine/Cardiology, Heart Center of the University of Leipzig, Struempellstrasse 39, 04289 Leipzig, Germany.
Background Data suggest that T1 and T2 mapping have excellent diagnostic accuracy in patients with suspected myocarditis. However, the true diagnostic performance of comprehensive cardiac magnetic resonance (CMR) mapping versus endomyocardial biopsy (EMB) has not been determined.
Objectives This study assessed the performance of CMR imaging, including T1 and T2 mapping, compared with EMB in an unselected, consecutive patient cohort with suspected myocarditis. It also examined the potential role of CMR field strength by comparing 1.5-T versus 3.0-T imaging.
Methods Patients underwent biventricular EMB, cardiac catheterization (for exclusion of coronary artery disease), and CMR imaging on 1.5- and 3-T scanners. The CMR protocol included current standard Lake Louise criteria (LLC) for myocarditis as well as native T1, calculation of extracellular volume fraction (ECV), and T2 mapping (only on 1.5-T). Patients were divided into 2 groups according to symptom duration (acute: ≤14 days vs. chronic: >14 days).
Results A total of 129 patients underwent 1.5-T imaging. In patients with acute symptoms, native T1 yielded the best diagnostic performance as defined by the area under the curve (AUC) of receiver-operating curves (0.82) followed by T2 (0.81), ECV (0.75), and LLC (0.56). In patients with chronic symptoms, only T2 mapping yielded an acceptable AUC (0.77). On 3.0-T, AUCs of native T1, ECV, and LLC were comparable to 1.5-T with no significant differences.
Conclusions In patients with acute symptoms, mapping techniques provide a useful tool for confirming or rejecting the diagnosis of myocarditis and are superior to the LLC. However, only T2 mapping has acceptable diagnostic performance in patients with chronic symptoms. (Magnetic Resonance Imaging in Myocarditis [MyoRacer]; NCT02177630)
Diagnosing myocarditis still represents a conundrum in modern cardiology (1); due to the variety of clinical manifestations, clear-cut diagnostic criteria are lacking (2,3). The reference diagnostic standard is considered to be endomyocardial biopsy (EMB) (1), with diagnostic performance improved when taking myocardial samples from both the left ventricle (LV) and right ventricle (RV) (4,5). However, EMB is an invasive procedure with a risk of relevant complications and is not widely available.
Tremendous efforts have been undertaken to use cardiac magnetic resonance (CMR) imaging to diagnose myocarditis (6–11). One approach is the proposed Lake Louise criteria (LLC) (11), consisting of late-enhancement sequences, T2-weighted edema images, and T1-weighted sequences before and after contrast injection (early enhancement). Although helpful, especially in patients with angina-like symptoms and recent symptom onset, the LLC are insufficient when applied in patients with heart failure (HF) symptoms and chronic manifestation (12–14). However, most CMR studies were performed on 1.5-T scanners. The potential benefit of increased signal when using these sequences on a 3-T scanner remains unknown.
Recently, novel quantitative T1 and T2 mapping techniques, including the quantification of extracellular volume (ECV), have been suggested to overcome some limitations of the LLC, creating high expectations about their diagnostic utility (13,15–18). However, the majority of T1 and T2 mapping studies compared CMR results of patients with suspected myocarditis with healthy control subjects (13,15–17). The true diagnostic performance of the latest CMR techniques in patients with various pathologies, with EMB serving as the reference standard, has been determined in a single recently published trial in a small population of patients (18). Therefore, we sought to assess the diagnostic performance of a comprehensive CMR protocol, including the LLC, native T1 mapping, quantification of ECV, and T2 mapping on a 1.5-T scanner in a large, unselected, consecutive patient cohort in comparison with biventricular EMB. We also evaluated the potential role of CMR field strength in diagnostic accuracy by comparing CMR imaging results using 1.5- and 3-T.
In this prospective study, patients with clinical suspicion of myocarditis were included. Myocarditis was suspected in patients who fulfilled all of the following criteria: 1) new onset or persistent symptoms suggestive of myocarditis (shortness of breath, effort intolerance, fatigue, palpitations, or chest pain); 2) evidence of recent or ongoing myocardial damage (LV dysfunction, electrocardiographic abnormalities, or elevated troponin); 3) history of systemic viral disease; and 4) exclusion of relevant coronary artery disease (CAD) on angiography.
Patients with a history of CAD or contraindication to cardiac catheterization, EMB, or CMR were excluded. Additionally, patients with nondiagnostic EMB or CMR were excluded from the final analysis.
Patients were divided into 2 groups according to duration of symptoms from onset to hospital admission (acute symptoms ≤14 days; chronic symptoms >14 days).
Patients underwent biventricular EMB, CMR imaging on 1.5-T, as well as CMR imaging on 3-T within 36 h. CMR imaging on 1.5-T preceded 3-T imaging, whereas EMB was performed at any time point within the 36-h diagnostic window. Additionally, 10 healthy volunteers underwent CMR imaging and served as a control group (Online Appendix and Online Figure 1).
The study was approved by the local ethics committee, and all patients gave written informed consent.
Imaging, biopsy, and immunohistological acquisition and analyses
CMR was performed using a 1.5-T scanner (Intera CV, Philips, Best, the Netherlands) and 3-T scanner (Verio, Siemens Healthcare, Erlangen, Germany). The CMR protocol consisted of cine sequences, T2-weighted short-tau inversion-recovery (T2w), T1-weighted spin-echo (T1w), and phase-sensitive inversion recovery (PSIR) imaging for late enhancement assessment after intravenous administration of 0.15 mmol gadobutrol per kg of body weight. T2 mapping was performed before contrast media application with free-breathing, navigator-gated multi-echo sequence. T1 mapping was performed with a modified Look-Locker inversion recovery sequence using a 3(3)5 scheme before and 15 min after contrast application. Cine and mapping were acquired in vertical long-axis, horizontal long-axis, and short-axis orientations; T1w in transversal; and T2w and PSIR in short-axis orientation. Cine, T2w, T1w, and PSIR were realized as multislice stacks, covering the whole LV, whereas mapping was performed as single slices. Detailed descriptions of the CMR protocol and data analysis are provided in the Online Appendix.
Selective angiograms of the left and right coronary artery were acquired to exclude CAD. Biventricular EMB sampling was achieved via femoral venous and arterial access using myocardial biopsy forceps (Teleflex Medical Tuttlingen GmbH, Tuttlingen, Germany). Using fluoroscopic guidance, 6 to 7 EMBs were taken from different locations within the LV; RV biopsies were taken exclusively from the septal or apical regions. In case of thrombi formation within the LV or inadequate biopsy forceps position in the RV septal region, only RV or LV EMB was performed. Histological, immunohistological, and molecular pathological analyses were performed as previously described (4,6,10,12,19,20) and according to consensus recommendations (1). The Online Appendix provides a detailed description of EMB analysis.
The degree of myocardial fibrosis was quantified in a subset of 77 patients. Analyses of CMR imaging and EMB were performed in a blinded fashion by experienced operators.
Data were tested for normal distribution by the Kolmogorov-Smirnov test. Normally distributed data are expressed as mean ± SD; not normally distributed data are expressed by median and interquartile range (IQR). Proportions are expressed as number of patients and percentages. Unpaired samples between patient groups were analyzed with the 2-tailed unpaired Student t test or Mann-Whitney U test. Categorical variables were compared using the Fisher exact test. All statistical testing was on the basis of a 2-sided α = 0.05 significance level. A receiver-operating characteristic analysis was performed to calculate optimal thresholds and areas under the curves (AUCs). The Youden index was used to depict optimal cutoff values from the receiver-operating curves, and AUCs were compared by the DeLong method.
Statistical testing and data analysis was performed with SPSS version 16 (SPSS Inc., Chicago, Illinois), GraphPad Prism version 5.0b (GraphPad Software, San Diego, California), and MedCalc Statistical Software version 15.4 (MedCalc Software bvba, Ostend, Belgium).
From August 2012 until May 2015, 138 patients were enrolled into this study. Due to various reasons (Figure 1), EMB or results of 1.5-T imaging were not available in 9 patients, leaving 129 patients for the final analysis of 1.5-T imaging.
On the basis of symptom duration, 61 patients presented with acute and 68 patients with chronic symptoms. Cardiovascular risk factor profile was similar between the 2 groups (Table 1). Patients with chronic symptoms reported more dyspnea (71% vs. 44% in group 1; p < 0.001), whereas chest pain was more frequent in the acute patients (64% vs. 41%; p = 0.001). The LV ejection fraction was significantly lower in patients with chronic symptoms (27%; IQR: 17% to 44%) compared with acute patients (48%; IQR: 28% to 54%; p < 0.001).
Patients presenting with acute symptoms had a higher frequency of elevated creatine kinase-MB and C-reactive protein levels, whereas troponin was similar between groups.
Biventricular EMB was performed in 93% of patients (120 of 129). In 6 patients, only RV specimens were obtained and in 3, only LV specimens were available for analysis. Per EMB results (Table 2), myocarditis was the most common diagnosis on EMB analysis, with no differences between the acute and chronic groups (70% vs. 71%). Signs of significant myocardial inflammation with evidence of myocardial injury/necrosis (acute myocarditis) were observed in 4% and were found exclusively in patients with acute symptoms. Viral genomes were detected in 37% (48 of 129) of the total patient population with no significant difference on the basis of symptom duration (Online Figure 2).
The extent of LV fibrosis (collagen volume fraction) in 77 patients was significantly higher in patients with chronic symptoms as in patients with acute symptoms (14.2 ± 8.7% vs. 9.8 ± 6.0%; p = 0.03).
Results of 1.5-T CMR imaging
Edema ratio and early relative enhancement, as well as presence of late enhancement, did not differ between patients with biopsy-proven myocarditis and absence of myocarditis (Figure 2). According to the LLC, the CMR diagnosis of myocarditis was established in 66% of patients with EMB-proven myocarditis and in 53% of patients without evidence of myocarditis on EMB (p = 0.39) (Table 3). In contrast, values for native T1, ECV, and T2 were significantly higher in patients with EMB-proven myocarditis versus patients with no evidence of myocarditis on EMB.
Diagnostic performance, as determined by the AUCs of receiver-operating curves, was significantly higher for native T1 mapping (0.82; p = 0.002), ECV (0.75; p = 0.04), and T2 (0.81; p = 0.001) as compared with LLC (0.56), with no significant differences between mapping techniques (Central Illustration). Native T1 mapping yielded the highest diagnostic accuracy (81%) followed by T2 mapping (80%), ECV (75%), and LLC (59%) (Table 4).
In patients with chronic symptoms, conventional imaging techniques included within the LLC did not allow for differentiation between EMB-proven myocarditis and absence of inflammation (Figure 2). Importantly, values of native T1 mapping and ECV calculation did not differ between patients with and without myocarditis. Only values of T2 mapping were significantly different between groups (Table 3). The AUC for T2 mapping was superior to LLC (p = 0.002) and native T1 (p = 0.04) (Central Illustration).
The highest diagnostic accuracy was achieved by T2 (73%) followed by ECV (67%), LCC (59%), and native T1 mapping (45%) (Table 4). Comparison of patients with myocarditis and acute versus chronic symptoms demonstrated no differences in native T1 (p = 0.51), ECV (p = 0.23), or T2 (p = 0.18) (Table 3).
Results of 3-T CMR imaging
Of 129 patients within the 1.5-T population, 111 underwent subsequent 3-T CMR imaging (Figure 1 provides the reasons why 18 did not).
Results of 3-T imaging in patients with acute and chronic symptoms mirrored those of 1.5-T imaging apart from results of T2 mapping, which was inadequate on 3-T due to technical reasons and therefore excluded from the study. Values for native T1 and ECV, but not LLC, were significantly higher in patients with acute symptoms and EMB-proven myocarditis than in patients with no myocarditis in EMB (Table 3). The AUCs were significantly higher for native T1 (p = 0.004) and ECV (p = 0.002) as compared with LLC (Figure 3). There were no significant differences between T1 and ECV (p = 0.78) (Figure 3). In patients with chronic symptoms, there were also no significant differences in values of any imaging techniques (Figure 3).
Comparison of patients with myocarditis and acute versus chronic symptoms demonstrated no differences in native T1 (p = 0.15) or ECV (p = 0.38) (Table 3). AUCs of T1 mapping, ECV, and LLC in patients with acute and chronic symptoms were comparable on 1.5- and 3-T imaging with no significant differences (Central Illustration).
To the best of our knowledge, this is the first adequately sized study assessing the diagnostic performance of novel T1 and T2 mapping CMR techniques in consecutive patients with suspected myocarditis using biventricular EMB as the reference standard. Furthermore, the potential effect of CMR field strength was assessed by comparing 1.5- and 3-T CMR.
There are several core findings of this study. In patients with acute symptoms, native T1, ECV, and T2 exhibit diagnostic performance superior to the LCC and add substantially to CMR’s ability in confirming or rejecting the diagnosis of myocarditis. In contrast, only T2 mapping provided an adequate diagnostic performance in patients with chronic symptoms (on 1.5-T only, as T2 mapping on 3-T was inadequate due to technical reasons). Results of CMR imaging on 1.5- and 3-T were comparable, with no evidence of added value with 3-T imaging.
When interpreting this study’s results, it is important to appreciate the characteristics of the population studied. The vast majority of patients had reduced LV ejection fraction, with few patients presenting with infarct-like myocarditis. The challenge of CMR in such a patient population (especially in the chronic group) is differentiating between dilated cardiomyopathy (DCM) with inflammation versus DCM without inflammation, which is probably the most difficult scenario to face.
Lake louise criteria
The difficulties arise from the observation that the 3 tissue-based CMR markers of LLC (edema, hyperemia, and late enhancement) are not specific to myocarditis but also occur in other noninflammatory cardiomyopathies; this is particularly true for late enhancement. Although useful to differentiate between diseased and healthy populations (6), late enhancement’s low specificity produces much poorer accuracy when applied to HF patients (10,12–14). At the same time, the observation of sometimes only subtle and diffuse myocyte necrosis or fibrosis limits the sensitivity of late enhancement. Likewise, edema and hyperemia on T2w and T1w are not specific to myocarditis and can occur in other pathologies (21–23). Additionally, several methodological and technical problems with T1w and T2w imaging exist, such as susceptibility to artifacts (11), low signal-to-noise ratio (11), and the necessity of reference signal intensities in the skeletal muscle. The latter raises the risk of false negative results in the presence of accompanied skeletal muscle myositis (24). Consequently, LLC’s low diagnostic performance in our population was not just secondary to inherent limitations of these imaging techniques, but also a reflection of the studied patient population with very few patients with infarct-like myocarditis but predominantly HF-like myocarditis, in which LLC seems to perform inadequately (12,14).
Native T1 mapping and ECV
With the advent of native and contrast-enhanced T1 mapping techniques, some limitations of edema and fibrosis imaging in myocarditis seemed to have been overcome. When compared with healthy control subjects, native T1 mapping and ECV demonstrated good diagnostic performance for diagnosing myocarditis (13,15,17). Currently, T1 values are understood to be increased in the presence of intracellular or extracellular edema, hyperemia, and myocardial necrosis/fibrosis (25,26), which are all part of the pathological alterations seen in myocarditis. Therefore, it is plausible that T1 mapping might strongly discriminate between diseased and healthy subjects. Ideally, however, what we wish to differentiate between is inflammatory and noninflammatory pathologies.
In the early stage of myocarditis, native T1 might be predominantly increased due to intracellular and extracellular edema as well as hyperemia. Likewise, besides fibrosis, ECV reflects increased intracapillary plasma volume and extracellular edema as signs of an early inflammatory response (25). Although not solely specific, these alterations are more common in acute myocarditis than in DCM, explaining the significantly higher values of native T1 and ECV in myocarditis patients versus patients with other pathologies in the acute symptoms cohort. Consequently, in patients with acute symptoms, native T1 and ECV had superior and clinically more useful diagnostic performance compared with LLC.
In line with our results, Bohnen et al. (18) demonstrated in patients presenting with chronic HF symptoms that T1 values and the calculated ECV did not differ significantly between patients with and without inflammation. Our data confirmed that in patients with chronic symptoms, native T1 and ECV do not add significantly to the LLC; the finding that they were useful in patients with acute but not chronic symptoms might be explained by a shift in histological pathology. Acute inflammatory responses, such as edema and hyperemia, are believed to regress as expansion of the extracellular space due to cellular debris and diffuse fibrosis progresses (3). This results in decreasing native T1 values over time, as suggested in animal (27) and human studies (17).
However, diffuse fibrosis is seen in almost all cardiac pathologies at later stages and is, therefore, not helpful in differentiating between disease entities. Consequently, when the discriminating component of native T1 and ECV (the “inflammatory response”) diminishes and the component shared by many pathologies (“diffuse fibrosis”) increases, these imaging techniques can no longer differentiate between myocarditis and noninflammatory DCMs in patients with chronic symptoms, as reflected in our study by the higher percentage of interstitial fibrosis in patients with chronic (versus acute) symptoms.
In our study, T2 mapping was the only CMR imaging technique that produced acceptable diagnostic accuracy in both patients with acute and chronic symptoms, echoing 2 previous reports (18,28).
The overall superiority of T2 mapping could be explained by its sensitivity for detecting myocardial edema. T2 mapping is currently considered the most suitable CMR technique for detecting free tissue water content (29). As outlined earlier, native T1 and ECV can lose their discriminating power in myocarditis versus DCM in the presence of confounding extracellular fibrosis. In contrast, T2 should be increased by myocardial edema, irrespective of myocardial fibrosis. Assuming that myocarditis differs from noninflammatory cardiomyopathies in extent of myocardial edema but not fibrosis, T2 times will be increased in myocarditis patients to a greater extent than in patients with cardiomyopathies without edema, resulting in a higher specificity for myocarditis of T2 mapping than T1 mapping or ECV.
The diagnostic performance of T2 mapping in our study was better in the acute setting, with higher T2 values of nonmyocarditis patients in the chronic group, which led to some attenuation of T2’s discriminating power. Bohnen et al. (18) demonstrated that increased T2 can be also present in DCM patients without inflammation, mirroring our results. The reason for increased T2 in noninflammatory DCM remains speculative. Previously, it has been explained by myocardial edema secondary to inflammation (30), which seems unlikely in our nonmyocarditis patients, because inflammation was excluded on EMB. Alternatively, increased T2 might be secondary to subtle edema caused by repetitive ischemia. Stenosis of epicardial coronaries as a cause for ischemia was excluded on coronary angiography. In contrast, microvascular dysfunction as a cause for repetitive ischemia has been described in various nonmyocarditis cardiomyopathies and could have also been present in some of our patients with nonmyocarditis cardiomyopathy (31). Although the reason for T2 increase in the absence of myocarditis remains unknown, edema due to microvascular dysfunction can also be present in noninflammatory cardiomyopathies.
Nevertheless, T2 emerges as the only mapping technique allowing for discrimination between inflammatory and noninflammatory cardiomyopathy, even in the presence of increased fibrosis.
Bohnen et al. (18) reported excellent sensitivity with T2 mapping, suggesting that a T2 value of <60 ms indicates a very low probability of myocarditis. The ability to exclude myocarditis in patients with T2 below a certain cutoff was less prominent in our trial, but our cutoff value of 59 ms compared satisfactorily to previously described cutoff values of 60 ms (18) and 59 ms (29). Although definite conclusions and interinstitutional cutoff values remain elusive, there is increasing evidence that T2 mapping adds significantly to confirming or rejecting the diagnosis of myocarditis on CMR, even in patients with chronic inflammatory versus chronic noninflammatory DCM.
1.5- versus 3-T CMR imaging
Our study’s secondary aim was to assess the effect of CMR field strength on diagnostic performance. The potential benefits of higher field strength and improved signal-to-noise ratio on 3-T would suggest superior diagnostic performances for numerous CMR applications (32,33). In contrast, our 3-T results were comparable to those of 1.5-T imaging, with no further benefit or improvement in diagnostic performance. This might be more a reflection of nonspecific diagnostic criteria than technical limitations: late enhancement was more often detected on 3-T than on 1.5-T imaging in myocarditis patients, but also in nonmyocarditis patients. At least for the LLC, its potential superiority in detecting pathologies, such as late enhancement, are negated by the nonspecific nature of late enhancement in such a patient population. This should apply, too, for T1 mapping and ECV, which yielded equal but not superior AUC on 3-T. Encouragingly, the higher values of T1 on 3-T correspond well to what has been described previously (17).
Although EMB is widely accepted as the reference standard for diagnosing myocarditis, a potential sampling error can lead to underdiagnosing myocardial inflammation. We aimed for a reference standard as robust as possible, and therefore performed biventricular EMB as suggested previously (4,5), but cannot exclude some missed cases of myocarditis on EMB. Also, the imaging protocol applied in this study represented the knowledge and methodology available in mid-2012. Since then, consensus recommendations were published (25) and numerous technical innovations of CMR mapping have been described (34). Therefore, our results might differ when applying newer mapping sequences and developments. Moreover, across equal setups and parameters of CMR sequences, algorithms used on different scanners can vary, which could affect results, especially cutoff values. Our study results also must be viewed in the light of the studied patient population of predominately HF patients, who are likely to differ from a population with predominantly infarct-like myocarditis.
Also, in light of recent findings (35), assumptions about different CMR phenomena explaining histological pathologies should be made with caution. Although there are some data relating increased T1 or T2 mapping with fibrosis and edema, these mapping techniques might resemble a variety of cardiac pathologies, more or less specific and sensitive to alterations seen in myocarditis. Because values for native T1, ECV, or T2 did not differ significantly between myocarditis patients with acute versus chronic symptoms, currently applied mapping techniques might not help in distinguishing between acute and chronic myocarditis. This study focused on global and diffuse myocardial alterations and pathologies only. Further studies and analyses need to clarify whether a segmental approach to T1 and T2 mapping will provide an incremental diagnostic value. Finally, when comparing results of 1.5- and 3-T imaging, it is important to emphasize that this was not a pure comparison of CMR field strength. Differences in sequences, vendors, and timing of imaging on 1.5- and 3-T imaging are all potential confounders. Additionally, the lack of T2 mapping results on 3-T markedly limits comparison of 1.5- and 3-T results in our study. Further studies will be necessary before drawing any definite conclusions about the effect of CMR field strength in patients with suspected myocarditis.
Our results suggested that duration of symptoms is crucial when using CMR techniques to confirm or reject myocarditis. This seemed to be a consequence of the shift in inflammatory pathology over the disease course. Native T1 mapping and ECV were helpful in patients with acute symptoms and suspected myocarditis, but only T2 mapping added significantly to LCC when assessing patients with chronic symptoms. There was no effect of CMR field strength on diagnostic accuracy in our patient population. However, the exact effect of CMR field strength requires further evaluation before definite conclusions can be drawn.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: The diagnostic value of CMR in patients with myocarditis depends on the duration of symptoms and histological pathology. Although a variety of CMR techniques are useful in patients with recent onset of symptoms, for those with heart failure in the chronic phase of myocarditis, T2-weighted edema imaging has greater diagnostic sensitivity.
TRANSLATIONAL OUTLOOK: Additional research is needed to better define specific CMR acquisition protocols to distinguish the inflammatory, edematous, and fibrotic stages of myocarditis and guide clinical management.
This work was supported by the Deutsche Forschungsgemeinschaft (SFB TR19) and the Federal Ministry of Education and Research (01EZ0817) to Drs. Klingel and Kandolf. Drs. Philipp Lurz and Luecke contributed equally to this work. Drs. Thiele and Gutberlet contributed equally to this work, and are joint senior authors.
- Abbreviations and Acronyms
- area under the curve
- cardiac magnetic resonance
- extracellular volume
- endomyocardial biopsy
- Lake Louise criteria
- phase-sensitive inversion recovery
- T1-weighted spin echo
- T2-weighted short-tau inversion recovery
- Received November 16, 2015.
- Revision received February 3, 2016.
- Accepted February 8, 2016.
- 2016 American College of Cardiology Foundation
- Caforio A.L.,
- Pankuweit S.,
- Arbustini E.,
- et al.
- Kindermann I.,
- Barth C.,
- Mahfoud F.,
- et al.
- Yilmaz A.,
- Kindermann I.,
- Kindermann M.,
- et al.
- Chimenti C.,
- Frustaci A.
- Mahrholdt H.,
- Wagner A.,
- Deluigi C.C.,
- et al.
- Abdel-Aty H.,
- Boye P.,
- Zagrosek A.,
- et al.
- Friedrich M.G.,
- Strohm O.,
- Schulz-Menger J.,
- Marciniak H.,
- Luft F.C.,
- Dietz R.
- Friedrich M.G.,
- Sechtem U.,
- Schulz-Menger J.,
- et al.
- Lurz P.,
- Eitel I.,
- Adam J.,
- et al.
- Radunski U.K.,
- Lund G.K.,
- Stehning C.,
- et al.
- Francone M.,
- Chimenti C.,
- Galea N.,
- et al.
- Ferreira V.M.,
- Piechnik S.K.,
- Dall'Armellina E.,
- et al.
- Hinojar R.,
- Foote L.,
- Arroyo Ucar E.,
- et al.
- Bohnen S.,
- Radunski U.K.,
- Lund G.K.,
- et al.
- Mahrholdt H.,
- Goedecke C.,
- Wagner A.,
- et al.
- Greaves K.,
- Oxford J.S.,
- Price C.P.,
- Clarke G.H.,
- Crake T.
- Moon J.C.,
- Messroghli D.R.,
- Kellman P.,
- et al.
- Moon J.C.,
- Treibel T.A.,
- Schelbert E.B.
- Jeuthe S.,
- Wassilew K.,
- h-Ici D.O.,
- et al.
- Thavendiranathan P.,
- Walls M.,
- Giri S.,
- et al.
- Mordi I.,
- Carrick D.,
- Bezerra H.,
- Tzemos N.
- Spoladore R.,
- Fisicaro A.,
- Faccini A.,
- Camici P.G.
- Kim H.W.,
- Van Assche L.,
- Jennings R.B.,
- et al.