Prognostic value of cardiac magnetic resonance global T2 mapping in patients with takotsubo syndrome

European Heart Journal - Cardiovascular Imaging

7 September 2026
Organised by: Logo
ESC Journals IMAGING Cardiac Magnetic Resonance (CMR) VALVULAR, MYOCARDIAL, PERICARDIAL, PULMONARY, CONGENITAL HEART DISEASE Myocardial Disease

Abstract

AbstractAims

Myocardial oedema is a hallmark of takotsubo syndrome (TTS). Cardiac magnetic resonance (CMR) T2 mapping quantifies myocardial oedema, but its prognostic value in TTS is unclear. We evaluated the association between CMR-derived global myocardial T2 and major adverse cardiovascular and cerebrovascular events (MACE).

Methods and results

We retrospectively identified consecutive TTS patients undergoing CMR with tissue mapping within 30 days of admission. Imaging was re-analysed, blinded to clinical data. The primary endpoint was MACE (all-cause death, heart failure hospitalization, recurrent TTS, stroke, or transient ischaemic attack). Time-dependent receiver operating characteristic (ROC) analysis derived a T2 threshold. Prespecified covariates were summarized by ridge-penalized Cox into a single linear predictor for adjustment. A Firth-penalized Cox model tested robustness. Among 110 patients (median age 68 years; 91% women), median global T2 was 56 ms (IQR 53–60), and 17 (15.5%) experienced MACE over a median follow-up of 763 days (IQR 525–1357). T2 ≥ 63 ms predicted 1-year MACE (AUC 0.671, sensitivity 56%, specificity 89%) and separated Kaplan–Meier curves (log-rank P = 0.003). In ridge-penalized Cox models, each 5 ms increase in global T2 was associated with MACE (HR 1.40, 95% CI 1.06–1.87; P = 0.020). This association was also observed after adjustment in the Firth-penalized model (HR 1.70, 95% CI 1.20–2.40; P = 0.004). An internally derived threshold of T2 ≥ 63 ms was also associated with MACE but showed modest 1-year discrimination (AUC 0.671).

Conclusion

Global myocardial T2 may provide incremental prognostic information beyond prespecified clinical and imaging indices in TTS and refine risk stratification.

Contributors

Yoshito Kadoya
Yoshito Kadoya

Author

Faculty of Medicine - University of British Columbia Vancouver , Canada

Andrew M Crean
Andrew M Crean

Author

University of Manchester Manchester , United Kingdom of Great Britain & Northern Ireland

D Ian Paterson
D Ian Paterson

Author

University of Ottawa Heart Institute Ottawa , Canada