You are here

SAFE-PLAI

SAFE-PLAI
Signal and image-based fast simulations of heart electrophysiology for computer-assisted planning of clinical interventions (SAFE-PLAI)

Funding agency: Spanish Ministry of Science and Innovation through the CICyT program (TIN2011-28067). 2012-2014

Partners: Universitat Pompeu Fabra (Oscar Camara, Daniel Romero, David Soto, Ali Pashaei, Rosa Maria Figueras i Ventura, Xènia Albà), Universitat de València (Rafael Sebastian), Hospital Clínic i Provincial de Barcelona (Juan Fernández-Armenta)

Project description: During recent years, the incorporation of computer science methodologies applied to imaging data in clinical routine has substantially improved the diagnosis, treatment and therapy planning of complex cardiomyopathies. On the other hand, computational modeling has not been clinically adapted yet in the same way, neither for in silico testing nor for planning of different treatment delivery options.
There are a large variety of computational models with different levels of complexity. Two of the main issues preventing the use of complex models in a clinical environment are the difficulties to to make them patient-specific and the associated high computational costs due to the large amount of parameters guiding the simulations. However, the use of reduced models with few number of parameters that can be personalized by data assimilation techniques and patientspecific data can make simulations more realistic with a reasonable computational burden.
In this project, we aim to develop accurate anatomo-functional models of the heart that are fast enough to be used in a clinical environment for in silico planning of cardiac treatments related to arrhythmias.
In order to improve the realism of these fast simulations, finite-element anatomical meshes will be generated, from patient-specific information automatically extracted from computed tomography and magnetic resonance images. In addition, the heterogeneity of the cardiac tissue will be taken into account by coupling multi-scale physiological models: 1D fast electrical conduction system; ischemic tissue models; tissue anisotropy models; myocardial tissue propagation models; and a torso-based ECG model that takes into account several organs. Finally, the most relevant parameters of the simplified models will be personalized using data assimilation techniques and patient-specific electrophysiological signals such as invasive electro-anatomical maps and noninvasive surface ECG data. Further improvements to the models will be based on ex-vivo confocal microscopy and histology data.
Personalized models will then be used for reproducing different scenarios of cardiac interventions that present electrical disturbances such as the choice of ablation sites in ventricular tachycardia or characterizing different bundle block patterns in dyssynchrony patients. In addition, in silico testing of different pacing sites with different scar configurations will be performed on large-scale simulation data generated from atlas-based virtual populations.