Cardiac resynchronisation therapy

Lead – Universitat Pompeu Fabra

Clinical background – The increasing prevalence of congestive heart failure (CHF) is mainly caused by the steadily increasing number of heart attack survivors. However CHF has a terrible prognosis, worst than most malignancies, with 50% mortality in the first 3 years after diagnosis. In total, there are about 10 million patients treated for heart failure in the EU, corresponding to about 4% of the adult population, and resulting in 2% of the total health care costs. In a large subgroup, CHF is associated with abnormal electrical activation of the left ventricle. In these patients, Cardiac Resynchronisation Therapy (CRT) has recently been shown to be an effective treatment. However, since about one third of the patients do not respond to this very expensive (>€20.000) therapy and might even worsen their symptoms, significant controversy remains regarding optimal patient selection. Additionally, their might be a number of potentially successful responders who are currently not considered for the therapy. The ability to predict an individual response to CRT, thus reducing the unsuccessful response rate, and identifying the best methods of delivering this treatment (e.g. lead positioning, pacemaker setting), will have a real impact on CHF treatment.

Intended work – This application package will use image processing techniques and computational models jointly with algorithms to personalise multi-scale electro-mechanical models of the heart. The model and the personalisation strategy will be validated using near simultaneous intra cardiac electrical recordings with a non-contact mapping system and MRI derived cardiac geometry, tissue characteristics (Gadolinium late enhancement scar) and motion (tagging) in addition to changes to invasive pressure recordings and 3D ultrasound derived myocardial contraction from different lead positioning and pacemaker settings.

The results will be used to characterise the phenomena leading to heart failure and to derive new indices for patient selection thus improving the specificity of the current selection criteria for CRT. Finally, this simulation environment will be used to optimize the patient specific therapy response with respect to the positioning of the leads and setting of the pacing device including using an image guided system to ensure the leads are placed in the optimal position.

The concepts developed in this work package will be demonstrated in a small multi-centre trial involving three clinical centres (INSERM, KCL, and HSCM) from three different countries. This study will cover 120 patients and will allow collecting evidence of the clinical benefit of patient-specific simulations in CRT.

Progress – The generation of patient-specific models requires a tight, coordinated and collaborative effort between multidisciplinary teams. The main components of the workflow for the generation of the personalized computational models required for the CRT planning platform to be developed in the euHeart project are described below. Some of these components, mainly the ones related to the anatomical building, are already integrated into a common software framework, GIMIAS [1], which is a workflow-oriented framework extensible through plug-ins, which has been designed for the development of biomedical imaging prototypes.

Cardiovascular geometry: the overall cardiac geometry can be extracted from automatic and semi-automatic segmentation techniques based on atlases with population data, and that can be applied to whole-heart CT or MR images [2][3]. In addition, the user has manual correction tools if the segmentation is not accurate enough. Functional information such as tissue viability, which is crucial for interventional planning, can also be extracted from late enhancement MRI [2]. Finally, the coronary vein tree is extracted from vascular segmentation techniques [4] since it will constrain the lead placement.

Anatomical model building: from the previous segmentations, the finite-element volumetric meshes required for the simulations can be built and additional sub-structures relevant for CRT such as fiber orientation or the Purkinje system [5] can be synthetically added. The acquisition of human in vivo DT-MRI is being investigated within euHeart, but no patient-specific data is available yet.

Electromechanical simulations: mechanical and electrophysiological cellular models, most of them already available in the CellML repository, are coupled to monodomain activation and finite deformation mechanics within continuum-based simulation codes optimised for high performance parallel implementation. The key step is the personalization of conductivity and stiffness parameters as well as identifying appropriate boundary conditions, since these variables cannot be measured in vivo and are very critical for having more realistic electromechanical simulations. Most of these codes have been implemented in Open Source platforms such as  and Chaste, making collaboration between different research groups lot easier. Finally, the modelling approach to develop fast computational efficient cardiac models suitable for real-time interaction and for data assimilation procedures is also investigated [8].

Post-processing and validation of models: Once electromechanical simulations are obtained, they are post-processed and compared with information and other indices extracted from imaging and signal data, such as deformation fields from echo images [7], contact mapping data [8] or pressure information.