Patient-specific cardiac digital twins require physiologically realistic representation of the cardiac conduction system, yet most whole-heart electrophysiology models simplify or omit key conduction structures. We present a subject-specific human whole-heart model with an anatomy-guided and mechanistic reconstruction of the complete cardiac conduction system for efficient simulation of sinus activation and body-surface ECG.
A four-chamber heart-torso geometry was generated from magnetic resonance imaging of a human subject. The model explicitly incorporated the sinoatrial node with discrete exit pathways, major interatrial conduction pathways, the atrioventricular node with dual-pathway organisation, the His bundle, bundle branches, and a structured Purkinje system with regionally differentiated subendocardial conduction. Electrophysiology was simulated using a reaction-eikonal formulation, and a standard 12-lead ECG was computed with image-based electrode positions and a lead-field approach.
The integrated framework reproduced physiologically coherent whole-heart activation. Atrial excitation originated in the sinoatrial node, propagated through preferential right atrial and interatrial pathways, and converged at the atrioventricular junction. Ventricular activation proceeded through rapid His-Purkinje recruitment with near-synchronous biventricular endocardial breakthrough followed by physiologic epicardial activation patterns. Simulated 12-lead ECGs showed close qualitative agreement with measured recordings in both P-wave and QRS morphology.
This work establishes a computationally efficient, subject-specific cardiovascular digital twin framework with explicit conduction-system representation. By embedding strong anatomical and biophysical constraints, the model narrows the space of plausible activation sequences and provides a platform for studying conduction disorders, pacing response, and future patient stratification strategies based on personalised electrophysiological mechanisms.