Abstract
Abstract Background/Introduction Electrocardiogram (ECG) changes after primary percutaneous coronary intervention (PCI) in ST-segment elevation myocardial infarction (STEMI) patients, such as the recovery of ST elevation, are associated with prognosis, but quantifying them is challenging. Deep-learning algorithms (DLA) can diagnose STEMI and quantify the probability as a bio-index. However, the prognostic value of this index has yet to be well studied. Purpose This study aims to investigate the feasibility of predicting the prognosis of patients with STEMI by analyzing serial ECGs post-primary PCI using an DLA designed to detect AMI and STEMI. Methods We developed a DLA to detect acute myocardial infarction (AMI) and STEMI using 12-lead ECGs from hospital A. The DLA was based on convolutional neural network and presented the index for probability of AMI and STEMI as a number between 0 and 1. We subsequently analyzed serial ECGs from hospital B, taken more than one month prior to the STEMI event (baseline), at pre-PCI (at emergency department [ED]), immediate post-PCI, 6h, 24h, 1 month after PCI, using the DLA. The primary endpoint was a post-STEMI heart failure defined as left ventricular ejection fraction (LVEF) less than 50%. Results A total of 22,259 ECGs (13,916 ECGs of non-AMI and 8,343 ECGs of AMI) from 15,113 patients were used to develop and validate the DLA. The area under the curve (AUC) of the final DLA was 0.950 for AMI and 0.954 for STEMI. Serial ECGs of 639 STEMI patients were analyzed. The mean index value at different time points were as follows: 0.061 ± 0.158 in baseline ECG, 0.801 ± 0.333 in ECG at ED, 0.525 ± 0.429 in immediate post-PCI, 0.467 ± 0.425 in 6-hour, 0.482 ± 0.423 in 24-hour, and 0.273 ± 0.348 in 1-month after PCI (ANOVA P-value <0.001). To evaluate the predictive value of the DLA index, we stratified the study population into low, intermediate, and high-risk group based on the DLA index, with thresholds of 0.03 and 0.45 of the immediate post-PCI AI indices. The mean LVEF was 53.0 ± 10.1% in the low-risk group, 49.6 ± 10 %, and 47.4 ± 10.2 % in the intermediate and the high-risk group (ANOVA P-value < 0.001). The prevalence of post-STEMI heart failure was 30.6%, 46.4%, and 60.6%, respectively (P < 0.001). Furthermore, the DLA index was significantly associated with the recovery of left ventricular function. The prevalence of heart failure in 3-month follow-up echocardiography was highest in the high-risk group (15.4% vs. 27.1% vs. 31.8%, respectively, P = 0.022). Conclusion Our study shows that the DLA-derived bio-index is an effective tool to quantify ECG changes after primary PCI in STEMI patients and predict post-AMI heart failure. The DLA index is also associated with left ventricular function recovery, indicating its potential as a prognostic marker. Further studies are needed to validate our results and explore the use of the DLA index in assessing myocardial reperfusion after primary PCI.
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call