Multimodal Machine Learning for Risk Stratification of Low-Voltage Intracardiac Areas in Patients with Atrial Fibrillation

• ETH Zurich
• Medical Data Science

This Master’s thesis proposal, beginning in February 2025, aims to develop a non-invasive AI algorithm to estimate low-voltage areas (LVAs) in atrial fibrillation (AF) patients. Identifying LVAs, typically done through invasive intracardiac mapping, is crucial for tailoring ablation strategies. By combining 12-lead ECGs, clinical data, and imaging features, the proposed model seeks to localize LVAs, supporting personalized treatment and minimizing the need for invasive mapping. Advanced techniques such as variational inference, diffusion models, and transformers will be explored to enhance AF management and patient outcomes.

• Arithmetic and Logic Structures, Biomedical Engineering, Computer Vision, Electrical and Electronic Engineering, Knowledge Representation and Machine Learning
• ETH Zurich (ETHZ), Master Thesis