In recent years, human in silico trials have gained significant attention as a valuable approach proposed by various research and regulatory bodies to evaluate the effects of drugs, clinical interventions, and medical devices. By utilizing cardiac modelling and patient-specific simulations, in silico trials minimize patient risks and reduce reliance on animal testing.
However, the implementation of cardiac in silico trials presents several time-consuming challenges. It requires the generation of a large number of patient-specific cardiac models and conducting numerous simulations for each individual geometry, driven by diverse uncertainties such as clinical protocols, drug dosages or electrophysiological variability. These variations lead to an exponential growth of the number of simulations. Therefore, automatic methods for the generation of personalized models and simulation files are crucial for handling these computational demands.
In this study, we present an open-source pipeline tailored for automated execution of in silico trials in patient-specific biventricular geometries, primarily focusing on electrophysiological and electromechanical simulations. The pipeline generates closed biventricular surface meshes and automatically adds labels to the surfaces, apexes, and valves. From these surface meshes, it creates tetrahedral and hexahedral meshes with the desired resolution and generates the most common fields required for the simulations such as fibre information, cellular heterogeneity, or heart gradients. Additionally, the pipeline integrates with algorithms for Purkinje tree generation and electrophysiological personalisation of patient data. We demonstrate the application of the proposed methodology by performing simulations on two different databases: 1) electromechanical simulations on meshes from 100 UK Biobank patients, using Alya Software, and 2) assessing drug effects through electrophysiological simulations on 20 open-source geometries using MonoAlg3D solver. This highlights its versatility and potential for effectively managing different large databases.