Multiparametric cardiac magnetic resonance imaging (CMR) for the diagnosis of Loeffler’s endocarditis: a case report

Endocarditis parietalis fibroplastica (Loeffler’s disease) (EPF) is a disease with poor prognosis, first described by Loeffler in 1936. A diagnosis requires blood eosinophilia and affection of the endocardium as key steps [1]. The EPF belongs to the hypereosinophilic syndromes (HES) with rare epidemiologic data. Spry et al. showed a prevalence of 1:200000 in general population [2]. In more recent studies, the prevalence of EPF is shown from 40 to 50% among patients with HES [3]. Eosinophils must be elevated over 1.5x10⁹ eosinophils per litre for about 6 months to lead to the diagnosis of hypereosinophilia, although blood levels may lack hypereosinophilia in later stages. HES syndrome can be idiopathic or correlated to clonal interleukin 5 (IL-5) over-production of T-helper cells. Once the heart is affected, the disease passes three steps: 1. Acute necrosis followed by 2. thrombosis and 3. end stage of fibrosis [4]. The underlying pathophysiology is a degradation of cytotoxic eosinophilic proteins, which may reveal endocardial thickening, ventricular obliteration by echogenic material, suggestive of fibrosis or thrombosis, atrial dilation, restrictive pattern in echo Doppler or coronary vessel alterations [5]. Still, diagnostic process is heterogeneous and is yet to be standardized. Electrocardiogram (ECG) and transthoracic echocardiography (TTE) are unspecific and these technologies have certain restrictions (i.e. insufficient acoustic window). Endomyocardial biopsy is supposed to be the gold standard of diagnostic tools, but exerts a considerable inherent sampling error and may thus suffer from insufficient diagnostic sensitivity. Cardiovascular magnetic resonance (CMR) might be able to characterize the specific stages of the disease due to its superior native soft tissue contrast and further tissue characterisation by contrast-enhanced sequences.

Cine steady state free precession (SSFP) images revealed normal left and right ventricular function (LV ejection fraction (EF) 61%, RV-EF 61%). As can be seen in Fig. 1a (doubled arrows), LV and RV walls were thickened and displayed subendocardial susceptibility artefacts (arrow). Despite normal LV-EF, myocardial feature tracking analysis (TomTec Imaging Systems, Unterschleißheim, Germany) revealed impaired systolic endocardial and epicardial global longitudinal strain (GLS) compared to literature standards (-17.7% vs. -22.2% for endocardial GLS; -14.5% vs. -20.4% for epicardial GLS) [6]. Peak early diastolic strain rate (SRe) was impaired in epi- and endocardium as well, indicating diastolic impairment (1.2 s^-1 vs. 1.7 s^-1 for endocardial diastolic SRe; 1.0 s^-1 vs. 1.5 s^-1 for epicardial diastolic SRe) (Fig. 1b).

Perfusion images showed an imbalance of transmyocardial perfusion already under resting conditions (Fig. 2a, arrows). This phenomenon was aggravated during regadenoson induced hyperaemia, that showed highly impaired perfusion in endocardial layers, indicating altered coronary microcirculation (Fig. 2b).

To analyse myocardial tissue texture, a parametric CMR was conducted to separate oedema or acute necrosis to other tissue texture components with T2 mapping [7]. As shown in Fig. 3a, T2 mapping revealed a slightly reduced T2 time (55.0 ± 8.4 ms, arrows) compared to an age- and gender-matched control (56.4 ± 3.5), which was still in the normal value range according to current literature [8]. Epimyocardial T2 time was elevated compared to the control (69.6 ± 11.9 ms vs. 60.3 ± 4.6 ms, arrowheads) suspecting acute tissue oedema [8]. Diffuse subendocardial late gadolinium enhancement (LGE) was detected in both chambers in those areas of slightly reduced T2 time completing the previous findings (Fig. 3b, arrows).

Taking the localisation of fibrosis, the T2 map and the perfusion abnormalities into account, the hypotheses of vascular coronary disease or myocarditis were discarded since there was neither an endocardial lesion corresponding to a coronary territory nor an epicardial lesion. Additional details of the patient’s history revealed a hypereosinophilia (max. 11.900/μl, normal value < 440/μl) for more than 6 months. On the basis of the CMR findings and the patient’s history of T-NOS with hypereosinophilia, he was diagnosed to have Endocarditis parietalis fibroplastica Löfflein (EPF). Interestingly, the multiparametric CMR workup revealed different stages of myocardial involvement: reduced GLS, diffuse LGE, perfusion defect and slightly lower T2-time displayed fibrosis and reduced coronary flow of the endocardium already under resting conditions. The epicardium, however, was characterized by abnormal GLS with abnormal diastolic function, lack of LGE, preserved perfusion, but elevated T2 times suggesting an inflammatory stadium. A high-dose corticoid drug (Dexamethasone 40 mg) was added to the existing everolimus-therapy and the patient was transferred to the department of haematology for further diagnostics. His bone marrow showed elevated numbers of eosinophilia up to 10% during completion of diagnostic steps.

Cardiac involvement at different stages of EPF as shown by multiparametric CMR characterized T-NOS as a palliative situation, as recent therapeutic regimes such as immunosuppressive drugs and stem cell transplantation did not lead to a lasting remission.

Since multiparametric CMR workup could not only identify EPF but also characterize the patchy disease state (T2 map, LGE and myocardial strain) and stratify the patient’s individual prognosis, invasive coronary angiography and endomyocardial biopsy were not conducted.

The present case of EPF illustrates that multiparametric CMR can not only identify EPF, but also characterize EPF according to the disease state. CMR is a highly efficient tool for the process of diagnosis if invasive procedures are discarded due to risk/benefit assessment. In this particular case, myocarditis and coronary artery disease could be excluded without EMB or angiography due to multiparametric CMR. In addition, CMR detected advanced endocardial fibrosis with wall thickening as morphological hallmarks of the disease, while those features would have technically impeded adequate EMB.