Periodic Reporting for period 1 - AT-TOP (Using Topology To Revolutionize Atrial Tachycardia Treatment)
Reporting period: 2024-01-01 to 2025-12-31
Atrial tachycardia (AT) is a common cardiac arrhythmia associated with serious health risks, including stroke, heart failure, and long-term atrial dysfunction. Although catheter ablation is an established treatment, complex ATs remain particularly difficult to manage. Current mapping systems rely heavily on manual interpretation by electrophysiologists (EPs), which can lead to variability in diagnosis, suboptimal ablation strategies, and arrhythmia recurrence. In addition, the field lacks a universally accepted classification system for AT subtypes, contributing to diagnostic uncertainty and inconsistent treatment approaches.
The AT-TOP project addresses these limitations through a fundamentally new, topology-based framework for understanding and treating AT. Building on discoveries from the ERC-funded SMARTHEART project, AT-TOP introduces a mathematically grounded classification system based on the concept of complete and incomplete reentry loops. Our research has demonstrated that each AT contains two interconnected loops: a dominant complete loop that sustains the ongoing arrhythmia, and a suppressed incomplete loop that remains latent but can become active if the primary loop is ablated. This dual-loop structure, rooted in topological principles, has been consistently confirmed in computational simulations and across a large set of clinical cases. We define the anatomical boundaries around which the dominant complete loop and the suppressed incomplete loop rotate as critical boundaries.
To translate this insight into clinical practice, we have developed Directed Graph Mapping (DGM), a software tool that automatically detects the critical boundaries, classifies AT types using patient-specific anatomical topology, and predicts how the arrhythmia may evolve after ablation. Importantly, DGM shifts mapping from a purely diagnostic approach to a predictive one: it not only helps terminate the current arrhythmia but also anticipates and prevents recurrence by identifying loops that would otherwise remain undetected.
The primary objectives of the project are:
* To validate the clinical value of DGM in a large-scale prospective study targeting participation from 5 hospitals. Data collection has been successfully initiated across multiple centres, and a substantial multicentre dataset comprising approximately 500 clinical AT cases is currently being analysed.
* To further optimize the DGM algorithms for robust, automated detection and classification of AT based on patient-specific anatomical maps.
* To establish a future roadmap for DGM in terms of valorization
The expected impact of AT-TOP is multifaceted:
* Clinical impact:** Shorter and more accurate ablation procedures, reduced recurrence rates, and improved long-term patient outcomes.
* Operational impact:** Fewer repeat procedures and more efficient use of hospital resources.
* Scientific impact:** Introduction of a mathematically rigorous and clinically validated classification framework that may establish a new standard in arrhythmia management.
* Economic and societal impact:** Broader access to advanced diagnostics, healthcare cost reductions, and improved quality of life for patients.
Beyond engineering and clinical electrophysiology, the project highlights the power of interdisciplinary collaboration. By integrating mathematical topology and network theory into cardiology, AT-TOP provides a common conceptual language for electrophysiologists, supports standardization, and strengthens collaboration across centres.
In summary, AT-TOP translates fundamental scientific insight into a clinically actionable and predictive tool. By redefining how atrial tachycardia is classified and treated, the project has the potential to establish a new paradigm in arrhythmia management and open the door to similar advances in other complex cardiac disorders.
The AT-TOP project addresses these limitations through a fundamentally new, topology-based framework for understanding and treating AT. Building on discoveries from the ERC-funded SMARTHEART project, AT-TOP introduces a mathematically grounded classification system based on the concept of complete and incomplete reentry loops. Our research has demonstrated that each AT contains two interconnected loops: a dominant complete loop that sustains the ongoing arrhythmia, and a suppressed incomplete loop that remains latent but can become active if the primary loop is ablated. This dual-loop structure, rooted in topological principles, has been consistently confirmed in computational simulations and across a large set of clinical cases. We define the anatomical boundaries around which the dominant complete loop and the suppressed incomplete loop rotate as critical boundaries.
To translate this insight into clinical practice, we have developed Directed Graph Mapping (DGM), a software tool that automatically detects the critical boundaries, classifies AT types using patient-specific anatomical topology, and predicts how the arrhythmia may evolve after ablation. Importantly, DGM shifts mapping from a purely diagnostic approach to a predictive one: it not only helps terminate the current arrhythmia but also anticipates and prevents recurrence by identifying loops that would otherwise remain undetected.
The primary objectives of the project are:
* To validate the clinical value of DGM in a large-scale prospective study targeting participation from 5 hospitals. Data collection has been successfully initiated across multiple centres, and a substantial multicentre dataset comprising approximately 500 clinical AT cases is currently being analysed.
* To further optimize the DGM algorithms for robust, automated detection and classification of AT based on patient-specific anatomical maps.
* To establish a future roadmap for DGM in terms of valorization
The expected impact of AT-TOP is multifaceted:
* Clinical impact:** Shorter and more accurate ablation procedures, reduced recurrence rates, and improved long-term patient outcomes.
* Operational impact:** Fewer repeat procedures and more efficient use of hospital resources.
* Scientific impact:** Introduction of a mathematically rigorous and clinically validated classification framework that may establish a new standard in arrhythmia management.
* Economic and societal impact:** Broader access to advanced diagnostics, healthcare cost reductions, and improved quality of life for patients.
Beyond engineering and clinical electrophysiology, the project highlights the power of interdisciplinary collaboration. By integrating mathematical topology and network theory into cardiology, AT-TOP provides a common conceptual language for electrophysiologists, supports standardization, and strengthens collaboration across centres.
In summary, AT-TOP translates fundamental scientific insight into a clinically actionable and predictive tool. By redefining how atrial tachycardia is classified and treated, the project has the potential to establish a new paradigm in arrhythmia management and open the door to similar advances in other complex cardiac disorders.
Objective 1 – Establish a multicentre clinical validation framework
A comprehensive Data Management Plan (DMP) was implemented and the study was preregistered on the Open Science Framework to ensure transparency and reproducibility.
The clinical study was launched in collaboration with five Belgian hospitals. Ethics approvals were obtained at all sites, and Data Transfer Agreements were formalised to enable secure multicentre data exchange.
Data collection has been successfully operationalised, resulting in a structured multicentre dataset of approximately 500 clinical AT cases. Data acquisition is ongoing, and analysis of the collected datasets is currently underway. This marks the transition from study setup to active clinical validation.
Objective 2 – Develop computational infrastructure for topology-based analysis
We improve DGM and made it open source: openDGM. OpenDGM is now a modular Python toolkit for parsing, analysing, and visualising electroanatomical mapping data across multiple platforms (CARTO, Rhythmia, Finitewave, openCARP).
We also created a backend architecture for structured data management and scalable multicentre analysis workflows (currently only for internal use).
Objective 3 – Advance algorithmic automation
Methodological development progressed from semi-automated localisation of critical topological boundaries (the complete and incomplete loop) to the development of fully automated detection algorithms. This significantly improves robustness, scalability, and readiness for clinical integration.
A comprehensive Data Management Plan (DMP) was implemented and the study was preregistered on the Open Science Framework to ensure transparency and reproducibility.
The clinical study was launched in collaboration with five Belgian hospitals. Ethics approvals were obtained at all sites, and Data Transfer Agreements were formalised to enable secure multicentre data exchange.
Data collection has been successfully operationalised, resulting in a structured multicentre dataset of approximately 500 clinical AT cases. Data acquisition is ongoing, and analysis of the collected datasets is currently underway. This marks the transition from study setup to active clinical validation.
Objective 2 – Develop computational infrastructure for topology-based analysis
We improve DGM and made it open source: openDGM. OpenDGM is now a modular Python toolkit for parsing, analysing, and visualising electroanatomical mapping data across multiple platforms (CARTO, Rhythmia, Finitewave, openCARP).
We also created a backend architecture for structured data management and scalable multicentre analysis workflows (currently only for internal use).
Objective 3 – Advance algorithmic automation
Methodological development progressed from semi-automated localisation of critical topological boundaries (the complete and incomplete loop) to the development of fully automated detection algorithms. This significantly improves robustness, scalability, and readiness for clinical integration.
The central scientific and technical breakthrough of this project is the achievement of full automation of the topology-based analysis pipeline. Previously, identification of critical atrial boundaries required manual annotation of anatomical structures to localise conduction barriers. This manual step introduced user dependency, reduced reproducibility, and limited scalability.
A key conceptual advance underlying this automation stems from our refinement of conventional phase mapping techniques. Standard phase mapping methods assume spatial continuity of the cardiac phase field and identify rotational activity through phase singularities. However, in the presence of functional conduction block, fibrosis, or anatomical boundaries, the phase field becomes discontinuous. These discontinuities, which we term phase defects, lead to ill-defined phase indices and consequently to both false positive and false negative detection of rotational drivers.
To address this limitation, we developed an extended phase mapping framework that explicitly detects and accounts for phase defects, enabling robust calculation of the phase index around discontinuous regions. This approach eliminates erroneous detections and resolves previously missed rotations, yielding a physiologically consistent characterisation of both complete and near-complete reentries, including anatomically anchored circuits.
By formally recognising critical phase defects, rather than classical phase singularities, as the fundamental entities governing rotational dynamics, we established the theoretical foundation that enables automated detection of critical atrial boundaries and computation of optimal ablation lines without manual intervention.
This integration of topological theory with automated boundary detection represents a major step toward objective, reproducible, and clinically scalable arrhythmia mapping.
The preprint of the paper we have submitted can be found here: https://www.biorxiv.org/content/10.64898/2026.02.02.703232v1.abstract(opens in new window)
A key conceptual advance underlying this automation stems from our refinement of conventional phase mapping techniques. Standard phase mapping methods assume spatial continuity of the cardiac phase field and identify rotational activity through phase singularities. However, in the presence of functional conduction block, fibrosis, or anatomical boundaries, the phase field becomes discontinuous. These discontinuities, which we term phase defects, lead to ill-defined phase indices and consequently to both false positive and false negative detection of rotational drivers.
To address this limitation, we developed an extended phase mapping framework that explicitly detects and accounts for phase defects, enabling robust calculation of the phase index around discontinuous regions. This approach eliminates erroneous detections and resolves previously missed rotations, yielding a physiologically consistent characterisation of both complete and near-complete reentries, including anatomically anchored circuits.
By formally recognising critical phase defects, rather than classical phase singularities, as the fundamental entities governing rotational dynamics, we established the theoretical foundation that enables automated detection of critical atrial boundaries and computation of optimal ablation lines without manual intervention.
This integration of topological theory with automated boundary detection represents a major step toward objective, reproducible, and clinically scalable arrhythmia mapping.
The preprint of the paper we have submitted can be found here: https://www.biorxiv.org/content/10.64898/2026.02.02.703232v1.abstract(opens in new window)