Modelling tools for data extraction

Early Stage Researcher: Luca Barbarotta

Host institution: TU/e


The ESR will develop tools to extract and translate row imaging data into data sets that directly can be used for modelling and simulation purposes. Once available these tools will be used to postprocess the outcome of the image analysis of Task 2.1. Based on geometrical and functional features optimal computational grids including boundary conditions will be generated and stored in standard data formats optimized for Finite Element Analysis. Material properties need to be derived from image data as well as functional measurements such as strain and perfusion. The Finite Element simulation tools that will be deployed in Task 3.2 will be used as a target application. Also the graphical visualization methods of Task 2.3 will be projected into a finite element context such that not only different features but also uncertainty quantification will be accessible for the modelling tools of Task 3.2. To this end the ESR will develop modelling tools to complement quantified image information from Task 2.1 with estimations of clinically relevant information that cannot be retrieved from the images directly and can be used as input for patient specific predictive modelling for clinical decision making. An inverse approach will be followed using mathematical models of cardiac function describing the causal relation between morphology, tissue properties and function on the basis of physical and physiological principles. The ESR will design tools to render generic models, developed at TU/e, and patient-specific models. The tools will involve iterative tuning of tissue properties until model predicted functional characteristics match experimental characteristics quantified in Task 2.1. In this way important patient-specific modelling input such as passive electrical conductivity, myocardial stiffness and myocardial contractility but also myocardial fibre orientation and volume can be deduced from medical images just showing myocardial deformation. Estimated spatial maps of myocardial tissue properties represent direct and quantitative diagnosis of the pathology at tissue level.

Expected Results

i) A comprehensive software tool that enables the actual translation from cardiac image analysis to patient specific computational cardiac modelling ii) A module that enables to visualize computational results and imaging data of cardiac function in a similar way.