·To investigate the predictive value of myocardial strain in left ventricular remodeling (LVR) after acute ST-segment elevation myocardial infarction (STEMI).

·Forty-seven patients with STEMI who were admitted to Department of Cardiology of Shanghai Sixth People's Hospital, Shanghai Jiao Tong University from December 2018 to December 2019 and successfully received emergency primary percutaneous coronary intervention were included. Cardiac magnetic resonance (CMR) was performed at the acute stage of STEMI and 1 year later. Left ventricular stroke volume(LVSV), left ventricular mass index (LVMI), global longitudinal strain (GLS), global circumferential strain (GCS) and global radial strain (GRS) were analyzed. Biochemical indexes such as peak hypersensitive serum cardiac troponin I (hs-cTnI) and peak brain natriuretic peptide precursor (proBNP) and clinical data such as age, sex, blood pressure, heart rate (HR) were collected. LVR was defined as the following 2 models. The first one was that the left ventricular end diastolic volume (LVEDV) increased by ≥20% after 1 year of STEMI compared with the baseline. The second one was that the left ventricular end systolic volume (LVESV) increased by ≥15% after 1 year of STEMI compared with the baseline. Patients with STEMI were divided into LVR group and non-LVR group according to the above two models. The differences of CMR index, biochemical index and clinical data between the two groups were compared. Logistic regression was performed and receiver operating characteristic curve was drawn.

·In the model where LVEDV increased by ≥20% 1 year after the acute stage of STEMI, the LVR group and the non-LVR group had statistically significant differences in baseline HR (P=0.003), LVSV (P=0.034), LVMI (P=0.029), GLS (P=0.003), GCS (P=0.016) and GRS (P=0.019). Logistic regression analysis showed that baseline HR and GLS were significantly correlated with LVR (β=-0.070, OR=0.932, P=0.043; β=-0.334, OR=0.716, P=0.031). When the cut-off value of GLS was -9.835, the sensitivity and specificity of LVR prediction were 75.0% and 71.0%, respectively. In the model where LVESV increased by≥15% 1 year after the acute stage of STEMI, there were statistically significant differences in peak hs-cTnI (P=0.035), peak proBNP (P=0.044), LVMI (P=0.034) and GLS (P=0.010) between the LVR group and the non-LVR group. Logistic regression analysis showed that peak hs-cTnI and baseline GLS were significantly correlated with LVR (β=-0.017, OR=0.982, P=0.024; β=-0.304, OR=0.738, P=0.041). When the cut-off value of GLS was -11.145, the sensitivity and specificity for LVR prediction were 94.1% and 50.0%, respectively.

·In patients with STEMI, baseline GLS has a good predictive value for LVR, while GCS and GRS have no significant predictive value.