Out-of-hospital cardiac arrest (OHCA) affects 275,000 people in Europe every year. In Italy alone, 50,000 people experience OHCA annually, with only 9% surviving. Half of the survivors suffer severe brain damage. Immediate CPR and defibrillation by bystanders before the ambulance arrives can save lives, but often, CPR starts only when the ambulance gets there. Additionally, half of all OHCAs occur when the person is alone, causing delays in recognizing the emergency, calling for help, and starting lifesaving actions. Effective chest compressions and defibrillation are crucial but are often not done correctly or are not customized for each patient. Current guidelines recommend the same approach for everyone, which doesn't consider individual needs.

To tackle these issues, we plan to develop artificial intelligence (AI) algorithms, smartphone apps, and new devices. Our main goal is to create tools and technologies to improve the recognition of OHCA and provide timely and effective interventions, ultimately reducing the impact of OHCA and improving survival rates.

First, we aim to create an AI algorithm that can predict major cardiovascular events like heart attacks or cardiac arrests minutes, hours, or days before they happen. We will collect data from wearable devices to train and validate this algorithm, helping us identify individuals at risk. By alerting these individuals, they can seek emergency care and receive treatment before a cardiac arrest occurs. We will also work on recognizing OHCA cases from surveillance camera footage when they happen to people who are alone.

Second, to increase the rate of CPR and defibrillation before ambulances arrive, we will develop a smartphone app that geolocates and alerts nearby citizens to act as first responders. The app will guide them on how to quickly find a defibrillator and use it.

Third, to find the best spots on the chest for compressions and defibrillation, we will study chest scans from CTs and echocardiograms in both elective patients and cardiac arrest victims. This will help us understand the effects of compressing different heart structures and develop a sensor to determine the optimal positions for compressions and defibrillator pads.

Our multidisciplinary team of clinicians, researchers, and engineers will conduct experimental, simulation, and observational studies to develop these technologies, evaluate their potential for patents, design a plan for their use, and test their effectiveness in preventing and recognizing OHCA. We believe that by improving each step in the chain of survival-preventing cardiac events, early recognition, timely CPR and defibrillation, and high-quality advanced resuscitation-we can significantly improve treatment times and reduce the global death and disability rates caused by OHCA.

Out-of-hospital cardiac arrest (OHCA) annually affects 275000 individuals in Europe. In Italy alone, 50000 persons suffer from OHCA each year and only 9% survives1. Half of the survivors are left with severe brain damage. Prompt cardiopulmonary resuscitation (CPR) and defibrillation before ambulance arrival by bystanders can improve outcomes. However, in many cases, CPR only starts when the ambulance arrives. Additionally, half of all OHCAs occur in isolation, meaning that recognition, emergency calls, and lifesaving maneuvers are delayed. Chest compressions and defibrillation are critical for survival, but they are frequently inadequate or not patient-tailored. Current CPR guidelines recommend a uniform approach to chest compressions and defibrillation for all patients, which fails to account for individual differences. To address these unmet medical needs, we will develop artificial intelligence algorithms, smartphone apps, and novel devices. Starting with proof-of-concept approaches that we have already conceived, we will work to improve recognition of OHCA and provide timely and effective interventions. Our goal is to create tools and technologies that can help reduce the burden of OHCA and improve outcomes.

First, we aim to develop an artificial intelligence algorithm that can predict (minutes, hours, or days in advance) major cardiovascular events, such as myocardial infarction or cardiac arrest. To achieve this, we will collect biosignals recorded by wearables to train and validate the algorithm to identify individuals who are at risk of a major cardiovascular event. Alerted individuals will seek emergency medical care and receive treatments before a cardiac arrest occurs. We also aim to recognize OHCAs that occur in isolation from videos of surveillance cameras.

Second, to increase the rate of CPR and defibrillation provided before ambulance arrival, we will develop a smartphone app that will geolocate and alert nearby citizens to act as first responders. The app will also provide guidance on quickly retrieving a defibrillator.

Third, to determine the optimal compressions and defibrillation position on the chest, we will acquire scans of chest computer tomography and transesophageal echocardiography in elective patients and in victims of cardiac arrest. This will allow to determine optimal compression and defibrillator pads position, understanding the effects on outcomes of different cardiac structures compressed, and developing a modern sensor to estimate the optimal compression and defibrillator pads position on the chest.

Through experimental, simulation and observational studies and a multidisciplinary team of clinicians, researchers and engineers, we will develop the proof-of-concept of such technologies, evaluate their patentability, design an exploitation plan, and test efficacy in preventing and anticipating recognition of OHCA, reducing time to CPR and defibrillation, and offering patient-tailored CPR and defibrillation. Our underlying hypothesis is that developing novel methods and technologies to enhance each link in the chain of survival (preventative measures, early recognition, timely initiation of CPR and defibrillation, and high-quality advanced resuscitation) will significantly anticipate lifesaving treatments and reduce the global mortality and disability caused by OHCA.

AIM 1: PREDICTION AND RECOGNITION OF CARDIAC ARREST AIM 1.1: PREDICT A MAJOR CARDIOVASCULAR EVENT

Inclusion criteria:

• Age 18-70 years;
• Being a healthy volunteer (i.e., an individual with no history of cardiovascular events willing to contribute to the project) or a patient (survivors and non-survivors) who experienced major cardiovascular events (i.e., myocardial infarction or cardiac arrest);
• Users of a smartwatch or smartphone that continuously and automatically collect health data;
• Informed consent.

Exclusion criteria:

• Impossibility to access/export data;
• User did not wear the wearable device for periods longer than 24 hours;
• User did not wear the wearable device in the 4 weeks preceding the event.

AIM 1.2 CARDIAC ARREST DETECTION FROM VIDEOS No patient involved.

AIM 2: TECHNOLOGIES TO INCREASE CPR AND DEFIBRILLATION USE BEFORE AMBULANCE ARRIVAL No patient involved.

AIM 3: PATIENT-TAILORED RESUSCITATION AIM 3.1: CLINICAL STUDY IN PATIENTS WHO RECEIVED CPR

Inclusion criteria:

• Adults (≥ 18 years);
• Patients suffering a non-traumatic cardiac arrest treated with chest compressions (both survivors and non-survivors);
• Received a TEE, chest x-ray, or chest CT scan as the standard clinical assessment following cardiac arrest;
• Informed consent.

Exclusion criteria:

- Patients with severe thorax/mediastinal deformity.

AIM 3.2 CLINICAL STUDY IN PATIENTS WHO RECEIVED A CHEST CT SCAN

Inclusion criteria:

• Adults (≥ 18 years);
• Received a chest CT scan for any reasons;
• Informed consent.

Exclusion criteria:

- Patients with severe thorax/mediastinal deformity.

AIM 3.3 MACHINE LEARNING (ML) ALGORITHM No patient involved.

• Healthy volunteers (every adult individual with no history of cardiovascular events willing to contribute to the project) and patients who experienced major cardiovascular events (i.e., myocardial infarction or cardiac arrest). Both groups must have worn a wearable device or used a smartphone able to collect healthcare data and biosignals.
• Patients ≥18 years resuscitated after cardiac arrest or during ongoing cardiopulmonary resuscitation (CPR).
• Patients ≥ 18 years who received a chest CT scan for any reasons.