# Evaluating and improving the efficacy of Extracorporeal Cardiopulmonary Resuscitation (ECPR) in pediatric patients using interactive Machine Learning > **NIH NIH R21** · VILLANOVA UNIVERSITY · 2024 · $110,809 ## Abstract Project Summary/Abstract Pediatric cardiac arrest is a serious life-threatening problem affecting more than 15,000 hospitalized children each year in the US alone. Fewer than 50% of these children survive to hospital discharge, and neurological morbidity is common among those who survive. Importantly, pediatric cardiac arrest survival outcomes plateaued more than a decade ago, and there is hence a critical need for evidence-based and innovative therapeutic approaches. In particular, a significant number of patients fail to achieve return of spontaneous circulation (ROSC) even af- ter 30 minutes of conventional CPR and may be candidates for what is termed Extracorporeal cardiopulmonary resuscitation (ECPR). ECPR is a treatment that involves the use of veno-arterial extracorporeal membrane oxy- genation (VA-ECMO) and has been used successfully for resuscitation from shock or cardiac arrest in adult and pediatric patients. It is often utilized as an alternative resuscitation intervention for in-hospital Cardiac Arrest (IHCA) patients. Currently it is not clear if and which subpopulation of cardiac arrest victims may benefit from this intervention. Hence this proposal aims to develop advanced machine learning and signal processing algo- rithms using a sizeable, high-quality dataset which will identify specific underlying characteristics of the patient who would benefit from ECPR. In particular, in Aim 1, we will develop a model using pre-arrest demographic, physiologic, and biochemical data to predict failure to achieve ROSC within 30 minutes of CPR. We will also de- velop a model using pre-arrest and intra-arrest physiologic data, including continuous invasive and non-invasive waveform data over the first <5, <10, <15, <20, <30 minutes of CPR to predict failure to achieve ROSC. In Aim 2, we will identify pre- and intra-arrest characteristics from discontinuous data and continuous invasive and non-invasive waveform data of ECPR and develop a model to predict survivability to hospital discharge. Such a model would enable initiation of ECPR in critically ill patients who are unlikely to survive otherwise and hence lead to overall improvement of survival for in-hospital CPR patients. ## Key facts - **NIH application ID:** 10797817 - **Project number:** 1R21HL167181-01A1 - **Recipient organization:** VILLANOVA UNIVERSITY - **Principal Investigator:** Chandrasekhar Nataraj - **Activity code:** R21 (R01, R21, SBIR, etc.) - **Funding institute:** NIH - **Fiscal year:** 2024 - **Award amount:** $110,809 - **Award type:** 1 - **Project period:** 2024-01-01 → 2025-12-31 ## Primary source NIH RePORTER: https://reporter.nih.gov/project-details/10797817 ## Citation > US National Institutes of Health, RePORTER application 10797817, Evaluating and improving the efficacy of Extracorporeal Cardiopulmonary Resuscitation (ECPR) in pediatric patients using interactive Machine Learning (1R21HL167181-01A1). Retrieved via AI Analytics 2026-05-25 from https://api.ai-analytics.org/grant/nih/10797817. Licensed CC0. --- *[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*