LVEF Prediction During ACS Using AI Algorithm Applied on Coronary Angiogram Videos

Prospective Validation of a Deep Learning Algorithm for Prediction of the Left Ventricular Ejection Fraction From Coronary Angiogram Videos in Patients With Acute Coronary Syndrome

Left ventricular ejection fraction (LVEF) is one of the strongest predictors of mortality and morbidity in patients with acute coronary syndrome (ACS). Transthoracic echocardiography (TTE) remains the gold standard for LVEF measurement. Currently, LVEF can be estimated at the time of the coronary angiogram but requires a ventriculography. This latter is performed at the price of an increased amount of contrast media injected and puts the patients at risk for mechanical complications, ventricular arrhythmia or atrio-ventricular blocks. Artificial intelligence (AI) has previously been shown to be an accurate method for determining LVEF using different data sources. Fur the purpose of this study, we aim at validating prospectively an AI algorithm, called CathEF, for the prediction of real-time LVEF (AI-LVEF) compared to TTE-LVEF and ventriculography in patients undergoing coronary angiogram for ACS.

Inclusion Criteria:

• Informed consent signed by the participant
• Acute coronary syndrome
• Creatinine clearance ≥30 ml/min/m2 according to MDRD

Exclusion Criteria:

• Creatinine clearance <30 ml/min/m2 according to MDRD
• No indication to perform TTE in the 7 days following coronary angiogram
• Right bundle branch block
• Suspected or confirmed left ventricular thrombus
• Suspected or confirmed aortic dissection
• Ventriculography not feasible
• No left coronary system angiogram