Periodic Reporting for period 3 - SMARTHEART (Directed networks as a novel approach for improving the management of cardiac arrhythmias)
Période du rapport: 2024-02-01 au 2025-07-31
The management of cardiac arrhythmias remains one of the greatest challenges in cardiac electrophysiology. The prevalence of the most common arrhythmia, atrial fibrillation (AF), is expected to rise sharply with the ageing population. Despite decades of intensive research, the underlying mechanisms of AF remain incompletely understood, resulting in suboptimal treatment outcomes. AF ablation often leads to complex atrial tachycardias (AT), which are notoriously difficult to treat. Similarly, ventricular tachycardia (VT) and ventricular fibrillation (VF) remain major causes of sudden cardiac death, and catheter ablation for complex VT cases has achieved only modest long-term success.
There is therefore an urgent need to better understand and precisely localize the sources of arrhythmia in order to improve treatment strategies. In this project, I proposed a radical new approach: applying network theory to study the mechanisms of AT, VT, and AF. Network theory, best known as the mathematical foundation of the Google search algorithm and online social networks, as revolutionized fields across biology, physics, and the social sciences. Yet, before the start of this project, is was never applied to the heart.
Building on our preliminary work, this proposal applied network theory to clinical cardiac arrhythmia data, supported by in-silico simulations. We have developed a novel set of tools to automatically identify the sources of complex AT, AF, and VT enabling accurate prediction of potential ablation targets based on network theory. We demonstrated that our network-based approach could automatically predict ablation sites in AT with accuracy surpassing the most advanced clinical technologies currently available. This translational project is therefore designed not only to generate novel mechanistic insights into cardiac arrhythmias, but also to directly translate these findings into improved patient outcomes.
More information on this project is available at www.dgmapping.com.
There is therefore an urgent need to better understand and precisely localize the sources of arrhythmia in order to improve treatment strategies. In this project, I proposed a radical new approach: applying network theory to study the mechanisms of AT, VT, and AF. Network theory, best known as the mathematical foundation of the Google search algorithm and online social networks, as revolutionized fields across biology, physics, and the social sciences. Yet, before the start of this project, is was never applied to the heart.
Building on our preliminary work, this proposal applied network theory to clinical cardiac arrhythmia data, supported by in-silico simulations. We have developed a novel set of tools to automatically identify the sources of complex AT, AF, and VT enabling accurate prediction of potential ablation targets based on network theory. We demonstrated that our network-based approach could automatically predict ablation sites in AT with accuracy surpassing the most advanced clinical technologies currently available. This translational project is therefore designed not only to generate novel mechanistic insights into cardiac arrhythmias, but also to directly translate these findings into improved patient outcomes.
More information on this project is available at www.dgmapping.com.
Primary Objective
The overarching aim of this project was to use network theory to investigate the underlying mechanisms of three major cardiac arrhythmias—atrial tachycardia (AT), ventricular tachycardia (VT), and atrial fibrillation (AF)—and to advance ablation therapy. To this end, we developed Directed Graph Mapping (DGM), a software package specifically designed for the analysis of diverse cardiac arrhythmias.
1. Development of DGM
The development of DGM was significantly accelerated through the recruitment of a skilled programming team. Our goal was to transform DGM into a widely accessible tool for the scientific community, capable of analyzing electro-anatomical datasets from computational simulations, experimental studies, and clinical recordings. DGM was modularized into Python packages, allowing users to integrate it into their own workflows or to use a complete graphical user interface (GUI). The final version exceeded the original project objectives, offering an extensive set of features for the in-depth analysis and interpretation of arrhythmia dynamics.
Two external partners already applied DGM to analyze computational data, with one joint publication and one expected in the future. We will do a major public software release in the end of this project in 2025, aimed at establishing DGM as a standard tool for reentry analysis in electro-anatomical data.
In addition, the package proved instrumental in achieving substantial progress across all work packages (WPs).
2. Ablation of AT
A key objective was to develop an automated strategy for identifying the sources of complex ATs. This goal was surpassed with the discovery of a novel, topology-based theory that redefined AT ablation.
By modeling the atrium as a sphere with holes, representing anatomical gaps or non-conductive scar tissue, we developed a unified classification system for ATs. Contrary to the prevailing view that ATs were driven by a single reentry loop, our research demonstrated that single loops do not occur. Instead, they always appeared in paired, counter-rotating loops around atrial holes, in line with the index theorem described in the literature over 30 years ago but never applied clinically.
We classified these loops as either:
True loops: directly driving the arrhythmia.
Suppressed loops: not initially active but capable of sustaining slower ATs after ablation if left untreated.
The clinical neglect of suppressed loops explained the recurrence of slower ATs after otherwise successful ablation, an unresolved problem until now. This discovery represented a major advance in arrhythmia mechanisms and directly informed more effective ablation strategies. A white paper outlining this approach was published in the European Heart Journal, one of the leading journals in the field. A subsequent, detailed study was published in Circulation: Arrhythmia and Electrophysiology and has already been downloaded over 5,000 times, an exceptional achievement in our discipline.
3. Ablation of VT
In collaboration with Prof. Lee, we published results in JACC: Clinical Electrophysiology demonstrating the application of DGM for identifying VT-driving circuits. Building on insights gained from AT research, we began exploring the presence and role of suppressed loops in VT, with promising implications for future ablation strategies.
4. Mechanisms of AF
The objective here was to better understand AF dynamics. We demonstrated DGM’s effectiveness in in silico AF data (published in Computers in Biology and Medicine), successfully tracking meandering reentries in basket catheter datasets. A comparative study between DGM and phase mapping was also published in Computer in Biology and Medicine, 2024, with results indicating superior stability for DGM. We also made progress toward directly analyzing raw electrogram signals, bypassing LAT reconstruction. This capability proved essential for the accurate study of highly complex AF datasets, which was published in the Biomedical Signal Processing and Control, 2025.
The overarching aim of this project was to use network theory to investigate the underlying mechanisms of three major cardiac arrhythmias—atrial tachycardia (AT), ventricular tachycardia (VT), and atrial fibrillation (AF)—and to advance ablation therapy. To this end, we developed Directed Graph Mapping (DGM), a software package specifically designed for the analysis of diverse cardiac arrhythmias.
1. Development of DGM
The development of DGM was significantly accelerated through the recruitment of a skilled programming team. Our goal was to transform DGM into a widely accessible tool for the scientific community, capable of analyzing electro-anatomical datasets from computational simulations, experimental studies, and clinical recordings. DGM was modularized into Python packages, allowing users to integrate it into their own workflows or to use a complete graphical user interface (GUI). The final version exceeded the original project objectives, offering an extensive set of features for the in-depth analysis and interpretation of arrhythmia dynamics.
Two external partners already applied DGM to analyze computational data, with one joint publication and one expected in the future. We will do a major public software release in the end of this project in 2025, aimed at establishing DGM as a standard tool for reentry analysis in electro-anatomical data.
In addition, the package proved instrumental in achieving substantial progress across all work packages (WPs).
2. Ablation of AT
A key objective was to develop an automated strategy for identifying the sources of complex ATs. This goal was surpassed with the discovery of a novel, topology-based theory that redefined AT ablation.
By modeling the atrium as a sphere with holes, representing anatomical gaps or non-conductive scar tissue, we developed a unified classification system for ATs. Contrary to the prevailing view that ATs were driven by a single reentry loop, our research demonstrated that single loops do not occur. Instead, they always appeared in paired, counter-rotating loops around atrial holes, in line with the index theorem described in the literature over 30 years ago but never applied clinically.
We classified these loops as either:
True loops: directly driving the arrhythmia.
Suppressed loops: not initially active but capable of sustaining slower ATs after ablation if left untreated.
The clinical neglect of suppressed loops explained the recurrence of slower ATs after otherwise successful ablation, an unresolved problem until now. This discovery represented a major advance in arrhythmia mechanisms and directly informed more effective ablation strategies. A white paper outlining this approach was published in the European Heart Journal, one of the leading journals in the field. A subsequent, detailed study was published in Circulation: Arrhythmia and Electrophysiology and has already been downloaded over 5,000 times, an exceptional achievement in our discipline.
3. Ablation of VT
In collaboration with Prof. Lee, we published results in JACC: Clinical Electrophysiology demonstrating the application of DGM for identifying VT-driving circuits. Building on insights gained from AT research, we began exploring the presence and role of suppressed loops in VT, with promising implications for future ablation strategies.
4. Mechanisms of AF
The objective here was to better understand AF dynamics. We demonstrated DGM’s effectiveness in in silico AF data (published in Computers in Biology and Medicine), successfully tracking meandering reentries in basket catheter datasets. A comparative study between DGM and phase mapping was also published in Computer in Biology and Medicine, 2024, with results indicating superior stability for DGM. We also made progress toward directly analyzing raw electrogram signals, bypassing LAT reconstruction. This capability proved essential for the accurate study of highly complex AF datasets, which was published in the Biomedical Signal Processing and Control, 2025.
Merging Topology with Cardiac Arrhythmia was a significant achievement in this project. By integrating topology, a mathematical field that explores properties of objects and spaces under deformation, with the study of cardiac arrhythmia, we found a new concept how AT can be unique classified and analyzed. We have a fundamental theory that single loops do not exist, and that one always has to look for the second loop. This second loop needs to be taken into account for the ablation therapy. If you do not ablate the second loop, a new (slower) AT will arise. We expected these findings to also influence our two other objectives, namely improving ablation on VT and uncovering the mechanism of AF. We have indeed found that also for VT and AF, reentries must be paired, and have submitted our work.
For this discovery, I was honored the prestigious AstraZeneca Award for Innovation in Personalized Medicine in Cardiovascular Diseases (December 2024, €25,000).
For this discovery, I was honored the prestigious AstraZeneca Award for Innovation in Personalized Medicine in Cardiovascular Diseases (December 2024, €25,000).