Artificial intelligence guided tools in pediatric extracorporeal membrane oxygenation: Implications for clinical...

Area of Science:

• Biomedical Engineering
• Clinical Informatics
• Pediatric Critical Care

Background:

• Pediatric extracorporeal membrane oxygenation (ECMO) is a complex life support therapy with high complication rates.
• Effective ECMO management necessitates continuous data integration for critical decision-making.
• Neurologic injury, bleeding, thrombosis, and weaning challenges remain significant drivers of morbidity and mortality.

Purpose of the Study:

• To review current artificial intelligence (AI) and machine learning (ML) applications in pediatric ECMO.
• To evaluate the potential of AI/ML as decision-support tools for clinicians.
• To assess AI/ML's role in improving ECMO practice and patient outcomes.

Main Methods:

• A narrative review of English-language literature from 2020-2026 was performed.
• Databases searched included PubMed, MEDLINE, and Scopus.
• Studies focused on AI/ML in pediatric ECMO outcomes, complication prediction, circuit surveillance, and weaning support were identified.

Main Results:

• AI/ML applications predominantly use supervised or deep learning with integrated patient and device data.
• Emerging AI/ML tools show potential in predicting neurologic risk, bleeding, and transfusion needs.
• AI/ML demonstrates promise in detecting circuit anomalies and assessing weaning readiness.
• Progress includes validated pediatric models and time-series frameworks, though data heterogeneity persists.

Conclusions:

• AI-guided tools for pediatric ECMO are advancing from feasibility to clinical decision support.
• Rigorous validation, transparent reporting, and strong governance are crucial for AI implementation.
• AI has the potential to enhance situational awareness and care consistency in ECMO, supporting expert clinical judgment.