Coupled personalization of cardiac electrophysiology models for prediction of ischaemic ventricular tachycardia

Abstract

In order to translate the important progress in cardiac electrophysiology modelling of the last decades into clinical applications, there is a requirement to make macroscopic models that can be used for the planning and performance of the clinical procedures. This requires model personalization, i.e. estimation of patient-specific model parameters and computations compatible with clinical constraints. Simplified macroscopic models can allow a rapid estimation of the tissue conductivity, but are often unreliable to predict arrhythmias. Conversely, complex biophysical models are more complete and have mechanisms of arrhythmogenesis and arrhythmia sustainibility, but are computationally expensive and their predictions at the organ scale still have to be validated. We present a coupled personalization framework that combines the power of the two kinds of models while keeping the computational complexity tractable. A simple eikonal model is used to estimate the conductivity parameters, which are then used to set the parameters of a biophysical model, the Mitchell -Schaeffer (MS) model. Additional parameters related to action potential duration restitution curves for the tissue are further estimated for the MS model. This framework is applied to a clinical dataset derived from a hybrid X-ray/magnetic resonance imaging and non-contact mapping procedure on a patient with heart failure. This personalized MS model is then used to perform an in silico simulation of a ventricular tachycardia (VT) stimulation protocol to predict the induction of VT. This proof of concept opens up possibilities of using VT induction modelling in order to both assess the risk of VT for a given patient and also to plan a potential subsequent radio-frequency ablation strategy to treat VT.

Velocity-based cardiac contractility personalization from images using derivative-free optimization

“…Maps in sinus rhythm and during specific stimulation protocols were acquired, but only those in sinus rhythm were used as imaging can only be done in this rhythm. The extracted depolarization and repolarization isochrones then serve as input information to an electrophysiology personalization method (Relan et al, 2011) which minimizes the discrepancy between measured and simulated isochrones. Its output is a set of global parameters and local parameters (electrical conductivities) of the Mitchell-Schaeffer electrophysiology model (Mitchell and Schaeffer, 2003) which allow to interpolate, extrapolate and regularize the acquired isochrones.…”

“…The heart geometry was segmented from the image frame at mid-diastole, and a FEM mesh with synthetic fiber orientations was obtained (Figure 1(a)). The personalized T d and T r derived from the patient noncontact endocardial electrical maps were used in (3) (Relan et al, 2011). The passive mechanical parameters used are shown in Table 1.…”

“…The corresponding in-plane resolutions are 1.56, 1.45, and 1.52 mm 2 , and all images have 10 mm slice thickness. All data sets have the endocardial activation maps measured with the Ensite balloon (St. Jude Medical, MN), which were extrapolated to the myocardial volume using an electrophysiological model to provide the subject-specific T d and T r in (3) for the experiments (Relan et al, 2011). The left-ventricular blood pressures were also measured invasively.…”

Efficient probabilistic model personalization integrating uncertainty on data and parameters: Application to Eikonal-Diffusion models in cardiac electrophysiology

“…Different methods have been proposed to adjust cardiac electrophysiology models, including for instance genetic algorithms for the fast conduction system, Camara et al (2010), Maximum A Posteriori state estimation, Wang et al (2011), or in similar conditions using a deterministic approach and a trust-region minimizations, Chinchapatnam et al (2008); Relan et al (2011). Up to the best of our knowledge none of these approaches use a full probabilistic treatment of the problem.…”

Traversed Graph Representation for Sparse Encoding of Macro-Reentrant Tachycardia

“…Although the wellestablished CARTO and EnSite technologies are preferred in clinical practice, the electroanatomical models can also provide the cardiologists with activation time maps and potentially voltage information. Among the least computationally expensive frameworks are the eikonal model for conduction parameter estimation at macro-scale [1,2], but also simplified biophysical ionic channel models [3] or mono-domain models such as Lattice-Boltzmann [4]. Fast Marching, an adaptation of the graph traverse Dijkstra algorithm, is typically used to solve the differential equations in these models.…”