Multimodal Image processing software to Guide Cardiac Ablation Therapy – MIGAT

The development of the software will require frequent interactions with users (radiologists, cardiologists). The optimization process will be iteratively and incrementally performed, following the agile software development method. Requirements and solutions will evolve through collaboration between self-organizing and cross-functional teams. The method will promote adaptive planning, evolutionary development and delivery, a time-boxed iterative approach, and will encourage rapid and flexible response to change. Therefore, foreseen interactions between software design, quality/certification and clinical evaluation will be enhanced. Because the solutions available in current software prototypes are already suitable for clinical transfer, the optimization process will use clinical feedback from the beginning of the project. MIGAT development method is illustrated below.
During MIGAT development, the 5 following tasks will be concomitantly performed, each of which will gather a specific cross-functional team :
- Building MIGAT : this task will consist in integrating currently available research prototypes in MedInria software framework
- Clinical evaluation:this task will consist in testing MIGAT for real-time guidance during cardiac ablation procedures
- Optimize user experience: this task will consist in improving the compatibility of MIGAT with daily clinical use
- Optimize ablation guidance: this task will consist in improving the correlations between MIGAT predictions from non-invasive data, and ablation targets observed during invasive procedures
- Optimize software quality: this task will consist in ensuring the robustness of MIGAT and path the way for medical device certification

Image integration will likely become a clinical standard in the next few years to guide the ablation of complex cardiac arrhythmias such as persistent atrial fibrillation and ventricular tachycardia. This statement is substantiated by preliminary research conducted by our group and others, demonstrating that current research software prototypes have evident impact on daily practice. MIGAT will fill in a technological gap in current clinical electrophysiology. Indeed, the high degree of technology embedded in catheters and localization systems currently used in daily practice contrasts with the poor maturation of image processing tools to assist intra-cardiac navigation. This is likely due to a lack of interaction between 2 distinct fields of healthcare industry (cardiac electrophysiology and medical imaging). Therefore, the methodological developments required to implement this solution in clinical practice will likely emerge from an academic action.

Impact on patient care: By combining cardiac structures, myocardial substrate and non-invasively mapped arrhythmia circuits, MIGAT will allow a comprehensive assessment of patient-specific diseases to be displayed during the procedure to assist ablation. This should lead to increase safety and efficiency of cardiac ablation procedures.

Impact on research: MIGAT will allow multi-modal non-invasive data to be registered to invasive contact mapping data, and will therefore be an ideal tool for research on structure function-relationships and computer-modeling in cardiac diseases.

In addition, MIGAT will have a structuring impact on the cardiac electrophysiology community, harmonizing the methods for image-assisted ablation. This harmonization will be critical to ultimately validate the approach in multi-centric randomized-controlled studies.

Preliminary studies conducted by our group and few others strongly suggest that non-invasive data integration during cardiac ablation will become part of the standard practice in the next few years. The ability to directly display these data during cardiac mapping and ablation procedures will certainly improve patient management. In terms of safety it will likely help to focalize mapping and ablation on previously defined areas, resulting in shorter procedure time, lower radiation exposure, and shorter radiofrequency duration. In addition, the risk of cardiac rupture or extra-cardiac damage (coronary arteries, phrenic nerve, esophagus) will likely be lowered by direct visualization of myocardial thickness and surrounding anatomical structures. In terms of efficiency, the impact of multimodal data integration might differ between ventricular and atrial arrhythmias.
- In scar-related ventricular tachycardia, the integration of imaging data will ensure that all myocardial substrate is covered by mapping prior to ablation. The precise localization of myocardial scar and the thickness of the wall in the area of interest will likely help predicting the most appropriate approach for ablation. In patients with idiopathic ventricular tachycardia or frequent ectopics the impact of body surface mapping might be even higher in the absence of substrate to target the ablation.
- In atrial fibrillation, the method will enable for the first time a precise targeting of fibrillatory sources and rotors, as assessed by body-surface mapping. The ability to display the location of these targets with respect to atrial anatomy and fibrosis as assessed at magnetic resonance imaging will likely lead to a better understanding of the patient-specific substrate, and a better definition of the sources to ablate.

The dissemination of results will be ensured through peer-reviewed publications, communications and posters in French, European and worldwide meetings (AAC, AHA, ESC, HRS, RSNA, ECR, ISMRM, ESCR, MICCAI and FIMH). It will cover the range of scientific communities involved in MIGAT, namely cardiology, electrophysiology, medical imaging, computer sciences and engineering.

There is an important industrial value in MIGAT, as currently no company is able to provide such integrated information during cardiac ablation procedures. MIGAT will introduce a new technological solution in a rapidly growing market (global cardiac electrophysiology market is expected to rise by 500% in the next 10 years). Two innovations currently developed at Inria in collaboration with Liryc on ablation target prediction and interventional guidance have been identified as technological breakthroughs and will be tested in the context of MIGAT. Once mature, we will fill two international patent applications for these technologies (and therefore these cannot be disclosed in this document). MIGAT will be integrated in the MedInria software framework. An FDA and CE cleared version of the MedInria platform is currently developed as an Inria spin-off. Olivier Clatz (a research scientist from the Inria Asclepios team) has recently been awarded the National Start-up Prize for Innovation from the French Government (OSEO emergence). He is now in the creation process -with INRIA support- with an expected creation date in May 2013. This company would be ideally suited to propose MIGAT as a commercial product since all software interfaces will be in common with the MedInria academic platform.

Cardiac electrical disorders are a major cause of human mortality and morbidity worldwide. Catheter ablation therapy has become part of international recommendations for the management of both atrial and ventricular arrhythmias. Advanced catheter localization systems now enable 3-dimensional mapping of cardiac electrical activity. The integration of 3-dimensional imaging data acquired prior to the procedure and its merging with the mapping geometry was shown feasible. Available non-invasive imaging modalities can provide critical complementary information to assist cardiac mapping and ablation. Multidetector computed tomography can provide valuable information on patient’s anatomy (cardiac chambers, pulmonary veins, coronary arteries and veins, phrenic nerve or oesophagus location, epicardial fat thickness). Magnetic resonance imaging and positron emission tomography can provide valuable information on myocardial substrate (scar location), because most arrhythmias occur on structurally diseased hearts. In addition, a body surface mapping technology was recently developed, enabling non-invasive real-time whole-heart 3-dimensional electrophysiological mapping. The system computes unipolar epicardial electrograms from an array of body surface potentials acquired with a multi-electrode vest. Because of its real-time and whole-heart nature, body surface mapping gives access to the assessment of previously non-mappable arrhythmias such as cardiac fibrillation and non-sustained or poorly tolerated arrhythmias. This technology might be of value to preemptively identify ablation targets form epicardial activation (earliest epicardial exit of the circuit in cardiac reentrant tachycardias, location of rotor cores in cardiac fibrillation).

MIGAT objectives are to develop a software able to process and fuse multimodal non-invasive 3-dimensional data and to clinically validate multimodal data integration for the guidance of cardiac ablation. It is part of the scientific project of the Liryc Institute, a multidisciplinary research structure dedicated to cardiac electrical disorders recently created as part of the “Investissement d’Avenir” program. It will largely benefit from the MUSIC equipment combining state-of-the-art electrophysiological mapping technology with magnetic resonance imaging. MIGAT partnership will involve the Liryc Institute (INSERM U1045-University of Bordeaux), the Inria Asclepios Research Team dedicated to computer science (Inria Sophia Antipolis), and the University Hospital of Bordeaux (CHU Bordeaux). Within MUSIC consortium, ongoing collaboration between these partners has already led to the development of software prototypes for image segmentation, body surface mapping data processing, 3-dimensional fusion and modeling. This preliminary research has demonstrated the feasibility of the multimodal approach, and strongly suggested high clinical impact to guide the ablation of complex arrhythmias. MIGAT will aim at integrating currently available prototypes in one stand-alone software, and validating the software during clinical cardiac ablation procedures. We anticipate that a total number of 100 procedures (50 ventricular arrhythmias, 50 atrial arrhythmias) will be necessary to provide sufficient clinical feedback for software validation. Both the pre-clinical and the clinical tasks will be conducted jointly. During each procedure, clinical data will be used to improve both clinical performance (accuracy of non-invasive predictions) and user experience (user-friendliness and time-consumption of data processing, clarity of procedural displays). The optimization of MIGAT software will likely benefit from this continuous interaction between clinicians and engineers.

MIGAT research program is set over a 3 years period, with the objective of a ready-to-use and stand-alone software compatible with subsequent CE/FDA marking and multi-centric randomized-controlled clinical evaluation.