Myrte Barthels - ZOL / Hasselt University & Pieter Vandervoort - Ziekenhuis Oost-Limburg

Prof. Dr. Pieter Vandervoort, cardiologist and Myrte Barthels, data scientist both work at Ziekenhuis Oost-Limburg, Genk. Together with Imec they launched Heartfelt, an innovative project that utilizes artificial intelligence (AI) and federated learning to tackle the critical issue of early detection of atrial fibrillation (AF).  
AF is a heart condition affecting up to 20% of the elderly population and significantly increases the risk for stroke and other complications. Unfortunately, AF is often asymptomatic and under-diagnosed  making early detection crucial for proper and timely treatment. 
Currently, AF is diagnosed  using a 12-lead electrocardiogram (ECG). However, AI has shown the possibility  to predict from a normal ECG showing regular sinus rhythm the presence of dormant paroxysmal AF, which even the most highly trained cardiologists is not able to infer.  
In this presentation we will  show how Heartfelt tries to further exploit the power of AI to build a robust prediction model able to identify early signs of AF risk within the normal ECG. To address the challenge of training data-hungry AI models in a highly regulated healthcare environment, Heartfelt will use federated learning techniques,  
This presentation will showcase how collaboration between healthcare professionals and engineers can unlock the full potential of data-driven healthcare solutions, ultimately resulting in innovative solutions for preventative care.