Artificial Intelligence Scalable Solution for ST Myocardial Infarction

Delays in triage and diagnosis of patients presented with chest pain or other symptoms suggestive of Acute coronary syndrome (ACS) can be fatal. This study aims to improve those two aspects in acute ischemic disease care: reducing the delays to intervention and improving the accuracy of initial diagnosis, which are of paramount importance in cases of ACS, especially in ST-elevation myocardial infarction (STEMI). The former plays a critical role in minimizing in-hospital mortality rates, which have been shown to decrease proportionally with reduction of times to intervention. The latter relies on a correct interpretation of the ECG, first-line diagnostic tool in the assessment of patients with suspected ACS. The current standard of care for ACS includes a 12-lead ECG that should be performed within the first 10 minutes from the first medical contact. The ECG must be interpreted by a qualified physician, who will alert the on-call cardiologist to confirm or not the activation of the "infarction code", based on the ECG and clinical presentation. Such activation will mainly entail immediate transfer of the patient to the nearest hospital with the possibility of emergency coronary angiography (if not present in the initial institution), and eventual percutaneous coronary intervention (PCI). Regarding the diagnosis of Acute Myocardial Infarction (AMI), an accurate and rapid interpretation of the first ECG is critical for the differential diagnosis between STEMI, NSTEMI or unstable angina; and follow the proper standard of care guidelines. The largest delays occurs between the first ECG and the transportation for the cardiac catheterization laboratory, which has prognostic implications. In recent years, automatic digital tools based on artificial intelligence (AI) have been proposed as a solution to support physicians in the ECG interpretation, reducing their workload and time-to-diagnosis, suggesting the beneficial impact of AI-platforms for accurate diagnosis of AMI. In this setting, the AI-platforms should be able to automatically detect ECG patterns linked to unfavorable coronary anatomy and poor outcomes. It is also essential to have the capacity to identify more subtle ECG patterns, not obvious during physicians' interpretation, but indicating high-risk coronary anatomy. Additionally, the platform should assist the prediction of most severe coronary lesions, especially obstructive stenosis. This ability to detect coronary lesions could be useful in preventing unnecessary or premature activation of the catheterization laboratory, mainly in NSTEMI setting. The ASSIST clinical study is a cross-sectional, multicenter study aiming to collect data to develop the Willem™ platform, an AI-based tool for ECG analysis. This plataform could improve the accuracy for AMI diagnosis, particularly the differentiation between STEMI and NSTEMI, and early identification of patients with Occlusion Myocardial Infarction (OMI).