Abstract 13293: Predicting Successful ECMO Decannulation - A Novel Machine Learning Approach

• DOI: 10.1161/circ.148.suppl_1.13293
• Publication Date: 2023-12-19T07:50:21Z
• Authors: Elizabeth Hutchins; Al Rahrooh; Jeffrey Feng; Neha Chandra; Jeffrey J Hsu; Alex Bui
• Keywords: 03 medical and health sciences; 0302 clinical medicine

Introduction: Extra Corporeal Membrane Oxygenation (ECMO) has emerged as a crucial life support intervention, yet the decision-making process for safe decannulation remains challenging due to paucity of data. Methods: 199 patients who underwent venoarterial ECMO cardiogenic shock between 2015 and 2021 were identified. Demographic, hemodynamic, laboratory, echocardiographic data obtained within 24 hours collected. Successful was defined survival without relapse mechanical circulatory or heart transplant 30 days. The dataset randomly split into 80/20 train-validation splits, four machine learning models employed. 5-fold cross-validation error analysis performed. Feature importance derived from best performing model. Results: 103 successfully weaned off ECMO. Table 1 provides patient characteristics, which used features in models. Among tested, Random Forest (RF) exhibited highest performance (accuracy 85.0%, precision 90.0%, recall 85.7%, AUROC 0.94). Figure depicts each model well ranked feature Receiver Operating Curve RF most important mixed venous oxygen saturation, systolic blood pressure, flow rate pulmonary diastolic pressure. Conclusion: This study demonstrates successful application predicting weaning outcomes. Future research will add enhance with goal developing clinical tool assist physicians caring patients.