Comparison of diagnostic algorithms in clinically suspected viral myocarditis: Agreement between cardiovascular magnetic resonance, endomyocardial biopsy, and troponin T.

Abstract

Aims: Myocardial inflammation is increasingly detected noninvasively by tissue mapping with cardiovascular magnetic resonance (CMR). Intraindividual agreement with endomyocardial biopsy (EMB) or markers of myocardial injury, high-sensitive cardiac troponin (hs-cTnT) in patients with clinically suspected viral myocarditis is incompletely understood.

Methods: Prospective multicenter study of consecutive patients with clinically suspected myocarditis who underwent blood testing for hs-cTnT, CMR, and EMB as a part of diagnostic workup. EMB was considered positive based on immunohistological criteria in line with the European Society of Cardiology (ESC) definitions. CMR diagnoses employed tissue mapping using sequence-specific cut-off for native T1 and T2 mapping; active inflammation was defined as T1 ≥2 standard deviation (SD) and T2 ≥2 SD above the mean of normal range. Hs-cTnT of greater than 13.9 ng/L was considered significant.

Results: A total of 114 patients (age (mean ± SD) 54 ± 16, 65% males) were included, of which 79 (69%) had positive EMB criteria, 64 (56%) CMR criteria, and a total of 58 (51%) positive troponin. Agreement between EMB and CMR diagnostic criteria was poor (CMR vs ESC: area under the curve (AUC): 0.51 (0.39-0.62)). The agreement between a significant hs-cTnT rise and CMR-based diagnosis of myocarditis was good (AUC: 0.84 (0.68-0.92); p < 0.001), but poor for EMB (0.50 (0.40-0.61). Hs-cTnT was significantly associated with native T1 and T2, high-sensitive C-reactive protein, and N-terminal pro-hormone brain natriuretic peptide (r = 0.37, r = 0.35, r = 0.30, r = 0.25; p < 0.001), but not immunohistochemical criteria or viral presence.

Conclusion: In clinically suspected viral myocarditis, all diagnostic approaches reflect the pathophysiological elements of myocardial inflammation; however, the differing underlying drivers only partially overlap. The EMB and CMR diagnostic algorithms are neither interchangeable in terms of interpretation of myocardial inflammation nor in their relationship with myocardial injury.