DAMOCLES - DigitAl MOdeling of uncertainty for Cardiac abLation trEatmentS

AI Technology: Decision management: Classified and statistical learning methods; Deep Learning.

Atrial fibrillation (AF) represents a critical cardiovascular challenge, often complicated by significant interpatient variability and the limitations of current therapeutic interventions like Pulsed Field Ablation (PFA) and Radiofrequency Ablation (RFA). 

To address these issues, the DAMOCLES project proposes AblaFEMx, a scalable in-silico simulation framework designed to leverage EuroHPC exascale infrastructure for high-resolution multiphysics simulations. Built upon the high-performance finite element library dolfinX, this framework captures coupled biophysical and biological responses across diverse anatomical geometries. 

The project employs a physics-based AI approach, utilizing simulation outputs to train a foundation AI model capable of predicting atrial remodeling and identifying novel imaging-based biomarkers. 

This AI system will be integrated into a multi-fidelity digital twin framework that incorporates anonymized multimodal clinical data, such as MRI, CT, and 4D ultrasound, to enable personalized treatment planning. 

By establishing a probabilistic model for uncertainty quantification and a reusable infrastructure for regulatory-grade simulation studies, DAMOCLES aims to accelerate the transition toward in-silico clinical trials, thereby reducing reliance on animal experimentation and improving clinical decision-making.