Using Artificial Intelligence to Guide Chest Compressions During CPR for Adults with Out-of-Hospital Cardiac Arrest

What this Trial Is About

Researchers are exploring a new way to improve survival for adults who experience out-of-hospital cardiac arrest OHCA. This study focuses on using artificial intelligence AI to guide chest compressions during cardiopulmonary resuscitation CPR. Currently, chest compressions follow a standard location, but this may compress the aortic valve in nearly half of patients, reducing the chance of successful resuscitation. The study aims to develop an AI tool that analyzes arterial pressure waveforms in real time to identify the best compression site, shifting CPR towards a more personalized approach. The study involves developing and validating an AI-Enhanced Arterial Waveform Monitor, based on advanced deep learning techniques, including YOLO v8. The project has five phases collecting data from 150 patients, creating algorithms to detect compression waveforms, training the AI to recognize aortic valve compression using patient data, clinically testing the AI on 75 additional patients, and finally assessing the AIs use as a real-time support app in 30 clinical cases. When the AI detects valve compression, rescuers adjust the chest compression position, usually downward and to the left, to improve blood flow. Participants are adults aged 20 or older with non-traumatic OHCA receiving advanced life support. The study collects detailed physiological data, including arterial pressure and transesophageal echocardiography TEE images, to train and test the AI. Outcomes measured include how accurately the AI identifies valve compression, success at repositioning compressions, time taken for adjustments, return of spontaneous circulation ROSC, neurologic outcome at discharge, and chest compression metrics. The study spans about three years, with data collection, AI development, clinical testing, and real-world feasibility assessment phases.

The AIR-CPR Study: AI-Guided Chest Compressions