We will present our research on developing personalised physiology models to simulate and optimise treatment strategies for cardiac diseases. We integrate signal processing, machine learning, and computational modelling to investigate disease mechanisms using clinical imaging and electrical recordings. Our work spans population-level virtual trials and patient-specific models, aiming to translate these tools into the clinical environment. An illustrative application involves utilising machine learning to complement biophysical simulations in an in-silico trial to predict long-term response to treatment strategies for patients with atrial fibrillation. We will also discuss techniques for calibrating cardiac digital twins to patient electrograms and quantifying model uncertainty. Finally, we will introduce our open-source cardiac modelling pipeline, part of the Ecosystem for Digital Twins in Healthcare (EDITH) project, which enables scalable in silico trials and is available for research use.
