The ELEM V.HEART platform, also known as ELVIS, is an integrated system designed for cardiac modelling. Its core architecture comprises the WebApp (which hosts both frontend and backend systems), the Virtual Population Generator, and the high-fidelity simulation engine, Alya Red. Significant development progress has been reported across all components.
The ELVIS WebApp
The Elvis WebApp has reached a production milestone with the successful deployment of the first version of V.HEART on commercial cloud infrastructure, ensuring access security through a two-step authentication process. This cloud-based product is now utilized by customers to study the cardiac safety of their proprietary compounds, and internal cardiotoxicity study results are also being migrated onto the platform for direct analysis. The backend system managing simulations and data feeding has been substantially improved; a rigid, unstructured architecture was replaced with a flexible, scalable alternative that provides granular control over simulation configurations. V.HEART currently integrates a high-fidelity 3D cardiac simulation engine with a newly developed AI-driven predictive model capable of forecasting QT prolongation by analyzing compound characteristics across a simulated population of virtual hearts. Future efforts will concentrate on stakeholder validation, deeper integration into the wider Virtual Human Platform, and scientifically expanding the predictive model's capabilities.
The simulation engine, Alya Red
The Alya Red simulation engine is fully operational for modeling cardiac in silico clinical trials. It is capable of efficiently solving large 3D models, each representing a virtual patient from a synthetically generated cohort. The current version allows for the simulation of cardiac electrophysiology in healthy patients and includes new implementation and model parameterization for diseased patients, leading to a better representation of real populations. Furthermore, the models have been expanded to include electromechanical coupling—adding new circulatory system models and improving the coupling mechanism—and the atrial model, moving towards a 4-chamber cardiac model.
The Virtual Population Generator
The Virtual Population Generator utilizes advanced technology to create virtual populations for conducting in silico clinical trials. Key methodological developments include a Statistical Shape Model (SSM), applied primarily to biventricular cardiac anatomies, which allows for the generation of random heart shapes similar to those observed in the training dataset. An improved rule-based fiber generation algorithm has been developed to incorporate geometry-independent cardiac coordinates, enabling the standardized representation, comparison, and transfer of patient information. An automatic pipeline prepares the computational mesh for in silico trials, generating a high-quality tetrahedral mesh and adding necessary information such as myocardial fiber orientation, cell heterogeneity, activation locations, and the placement of ECG leads for pseudo-ECG measurements. Finally, the generator allows for defining a population starting from a given heart by altering the myocyte phenotype, material properties, and activation patterns, thereby supporting the generation of a wide range of hearts and diseases.Several in silico trials have been carried out on various virtual populations produced over the year, covering conditions such as LBBB, oncological issues, myocardial infarction, and heart failure.