Aims: To determine the diagnostic yield of tissue characterization by cardiovascular magnetic resonance (CMR) in a large clinical population of patients with suspected acute myocarditis (AM) and to establish its diagnostic value within the 2013 European Society of Cardiology position statement criteria (ESC-PSC) for clinically suspected myocarditis.
Methods and results: In this retrospective study, CMR examinations of 303 hospitalized patients referred for work-up of suspected AM in two tertiary referral centres were analysed. CMR was performed at median 7 days (interquartile range 4-20 days) after clinical presentation and included cine imaging, T2-weighted imaging, and late gadolinium enhancement. CMR images were evaluated to assign each patient to a diagnosis. By using non-CMR criteria only, the 2013 ESC-PSC were positive for suspected myocarditis in 151 patients and negative in 30. In the remaining 122 patients, there was insufficient information available for ESC-PSC assessment, mostly due to lack of coronary angiography (CAG) before the CMR examination (n = 116, 95%). There were no in-hospital deaths. CMR provided a diagnosis in 158 patients (52%), including myocarditis in 104 (34%), myocardial infarction in 44 (15%), and other pathology in 10 patients (3%). Non-urgent CAG (>24 h after presentation) was performed before the CMR examination in 85 patients, of which 20 (24%) were done in patients with subsequently confirmed AM, which could potentially have been avoided if CMR was performed first. ESC-PSC was correct in diagnosing AM before the CMR in 50 of the 151 patients (33%) and was correct in ruling out AM in all the 30 patients (100%). However, ESC-PSC provided an incorrect diagnosis of AM in 27 of the 151 patients (18%), which was corrected by CMR through the identification of new cardiac disease that could explain the clinical syndrome. Patients with insufficient ESC-PSC information had a relatively low pre-test probability of coronary artery disease. In this group, CMR confirmed the diagnosis of AM in a relatively high percentage (44%) but still revealed myocardial infarction in 8% of them.
Conclusion: Tissue characterization by CMR provided a good diagnostic yield in this large clinical population of patients with suspected AM. CMR provided incremental diagnostic value to the ESC-PSC by ruling out the diagnosis of AM on one hand and by potentially sparing AM patients from CAG on the other.