Comparing the findings and diagnostic accuracy of cardiac magnetic resonance imaging in patients with biopsy confirmed acute myocarditis presenting with infarct-like and heart failure symptoms

Abstract

Abstract Introduction Cardiac magnetic resonance imaging (CMR) is considered the reference imaging modality in providing a non-invasive diagnosis of acute myocarditis (AM), as it allows for the detection of myocardial injury associated with AM. However, the diagnostic accuracy and pattern of CMR findings appear to differ according to clinical presentation. Methods This is a retrospective cross-sectional study. Consecutive adult patients presenting to a single tertiary centre in South Africa between August 2017 and November 2021 with endomyocardial biopsy (EMB) confirmed AM presenting with infarct-like (IL) or heart failure (HF) symptoms who had undergone CMR were included. Results Thirty-eight patients were enrolled. Twenty-two presented with IL symptoms and sixteen with HF symptoms. On CMR, the IL group had significantly higher ejection fractions of both ventricles (LVEF 54.3±14.9% vs 35.4±14.0%, p<0.001; RVEF 56.8±10.9% vs 43.6±18.6%, p=0.03), despite similar end diastolic dimensions (LVEDVi 83.2±30.9 ml/m2 vs 104.4±38.0 ml/m2, p=0.07; RVEDVi 74.4±22.1 ml/m2 vs 85.8±29.9 ml/m2, 0=0.19). Myocardial oedema was detected more frequently on T2-weighted imaging (90.9% vs 62.5%, p=0.05) and in more myocardial segments (3.0 (IQR 2.0–4.0) vs 1.0 (IQR 0–1.0), p 0.003) in the IL group. Despite no significant difference in the prevalence of late gadolinium enhancement (LGE) between the two groups (95.5% vs 68.8%, p=0.06), the IL group had LGE detectable in significantly more ventricular segments (4.5 (IQR 2.3–6.0) vs 2.0 (IQR 0–3.3), p=0.01) and in a different distribution. The sensitivity of the original Lake Louise Criteria (LLC) was 90.9% in IL patients and 56.3% in HF patients, which improved to 95.5% and 68.8% with the updated LLC respectively. Conclusion The pattern of CMR findings and its diagnostic accuracy appears to differ in AM patients presenting with IL and HF symptoms. Although the sensitivity of the LLC improved with the addition of parametric mapping in the HF group, it remained lower than that of the IL group, and suggests that EMB should be strongly considered in patients with clinically suspected AM presenting with HF. Funding Acknowledgement Type of funding sources: None.