Cardiac Magnetic Resonance Imaging Detects Long-Term Fibrotic Changes in Patients with Primary Eosinophilic Disorders: A Cross-Sectional Study

Myocardial involvement is a major cause of morbidity and mortality in the setting of primary eosinophilia. We sought to compare conventional diagnostic modalities, such as clinical examination, electrocardiography and echocardiography to cardiac magnetic resonance imaging (CMR) for the evaluation of cardiac involvement in patients with primary eosinophilia.

Consecutive patients with primary eosinophilia and end-organ involvement were considered eligible. Secondary causes of eosinophilia were excluded using a standardized diagnostic protocol. Clonality was assessed by a combination of diagnostic methods including bone marrow smear, biopsy examination, immunophenotyping (using an extensive panel of T-cell markers), cytogenetic and molecular analysis, including RT-PCR for the detection of eosinophilia related fusion genes [M-BCR-ABL, t(9;22); m-BCR-ABL, t(9;22); μ-BCR-ABL, t(9;22); BCR-FGFR1, t(8;22); ZNF198-FGFR1, t(8;13); FOP-FGFR1, t(6;8); CEP110-FGFR1, t(6;8); BCR-PDGFR, t(4;22); FIP1L1-PDGFRA, del(4)(q12); ETV6-PDGFRB, t(5;12); H4-PDGFRB, t(5;10)], as well as ARMS PCR for the detection of JAK2 V617F. Echocardiography and electrocardiography were performed using standard methods. CMR scans were performed and interpreted by an assessor blinded to the clonality status of patients, using a 1.5 T Philips Intera CV MR scanner (Philips Medical Systems, Best, The Netherlands). Patients were evaluated for the presence of myocardial inflamation using T2-weighted, T1-weighted before and after contrast media injection and late enhanced images. ECG-triggered, T2-weighted (T2-W), multi-slice spin-echo sequence was performed in axial orientation. ECG-triggered, T1-weighted (T1-W) multi-slice spin-echo images were obtained in axial orientation with identical parameters before and after an intravenous bolus of 0.1mmol/kg Gd-DTPA. Measurements after Gd-DTPA were started within 1 minute of injection. Immediately after the second set of T1-W images, 0.1mmol/kg Gd-DTPA was given again and delayed-enhanced images were taken 15 min later, using an inversion recovery sequence. The patients’ results were compared with those of 20 healthy volunteers.

A total of 15 patients (2 male/13 female) were enrolled in this study. Five patients had been diagnosed with chronic eosinophilic leukemia (CEL; 3 being positive for JAK2 V617F and 2 carrying clonal cytogenetic abnormalities) and 10 with hypereosinophilic syndrome (HES). All patients were receiving treatment to control eosinophilia. 13 were in complete hematologic response (less than 5% eosinophils and less than 10.000 white blood cells/mm³) and 2 displayed mildly elevated eosinophil counts. Five patients had evidence of myocardial involvement by conventional diagnostic modalities. Despite being in hematological and clinical remission, 6 patients, 5 of whom were diagnosed with CEL, had abnormalities detected by CMR, including endomyocardial fibrosis in 5 (all with CEL) and active myocardial inflammation in one HES patient. Overall, cardiac involvement and CMR-detected endomyocardial fibrosis were more common among patients with CEL compared to patients with HES (Fisher’s exact test p-value= 0.0088, and 0.0037, respectively). None of the healthy volunteers had evidence of inflamation of fibrosis.

Our study demonstrates that CMR detects endomyocardial fibrosis in a substantial proportion of patients with primary eosinophilia, despite adequate disease control. Myocardial involvement and endomyocardial fibrosis in particular, are associated with clonal primary eosinophilia.