Researchers publish public virtual cohort of heart models

Per the PLOS ONE article and the King's College London repository, the research team produced a publicly available virtual cohort of twenty-four four-chamber heart meshes built from end-diastolic CT images of heart failure patients, mean age 67±14 years. The cohort was released with linear tetrahedral meshes at an average edge length of 1.1±0.2 mm, and with refined resampled versions at 0.39±0.10 mm to demonstrate adjustable resolution (PLOS ONE; KCL repository).

Per the authors, ventricular fibres were assigned using a rule-based method with orientations of -60° at the epicardium and 80° at the endocardium, and the team ran benchmark simulations for ventricular electrical activation and free mechanical contraction on the 1.1 mm meshes. Reported physiological outputs include a total activation time of 149±16 ms, left ventricular ejection fraction 35±1%, right ventricular ejection fraction 30±2%, and ventricular stroke volumes (PLOS ONE; KCL repository). The cohort and supporting files are archived on public repositories, including Zenodo, to enable reuse.

Public availability of multi-chamber electro-mechanics cohorts is uncommon, and datasets like this lower the barrier for reproducible cohort-scale simulation, validation, and method development. Observed patterns in similar releases show that accessible, well-documented meshes accelerate comparative studies in electrophysiology, parameter-sensitivity testing, and validation of reduced-order or machine-learning surrogate models.

For computational cardiology and device modelling, a 24-patient cohort with both electrical and mechanical simulations provides a middle ground between single-case studies and large-scale statistical atlases. Industry and academic groups working on patient-specific simulation pipelines, inverse problems, or surrogate modelling can use the cohort to test robustness across anatomical and disease variability without requiring access to clinical imaging data or bespoke meshing pipelines.

Observers should track downstream reuse, example: benchmarking studies that adopt these meshes, publications that extend the cohort to healthy controls, or integration of the dataset into community toolchains. Also note whether future releases add electrophysiological heterogeneity, fiber validation against diffusion imaging, or linked electrophysiology recordings, which would increase the cohort's utility for personalization and regulatory evaluation.