Deep Learning Risk Assessment of Myocardial Scars in Dilated Cardiomyopathy Patients using Electrophysiology Simulations


Dilated cardiomyopathy is a dangerous heart disease which can involve scarring of the heart tissue (myocardium). Such myocardial scars often have a fine scale microstructure which can trap electrical currents, leading to re-entrant electrical waves, heart rhythm disorders, and in the worse case sudden cardiac arrest. Knowing which scars are dangerous is important for patient risk assessment and treatment planning.


In this thesis you will train a neural network classifier to determine whether a particular scar is able to generate a dangerous re-entrant electrical activation. The overall shape of the scars can be seen in magnetic resonance images, but due to resolution issues we will need to guess what the scar microstructure looks like. The results of electrophysiology simulations will serve as training data, so that you will be able to simulate as much data as you need.

Learning outcome

• Hands on experience with deep learning and cardiac electrophysiology simulation.
• Experience with medical image analysis.
• Good potential to coauthor and publish a scientific paper.


• Data science and deep learning
• Python Programming
• Knowledge of biophysics desirable but not necessary


  • (Not listed)
  • Gabriel Balaban
  • Molly Maleckar


"Fibrosis Microstructure Modulates Reentry in Non-ischemic Dilated Cardiomyopathy: Insights From Imaged Guided 2D Computational Modeling"

"Late-Gadolinium Enhancement Interface Area and Electrophysiological Simulations Predict Arrhythmic Events in Patients With Nonischemic Dilated Cardiomyopathy"

Contact person