Effect of Cardiac Orientation on Electrical Axis and QRS Duration in Simulated Left Bundle Branch Block

Mohammad Kayyali1, Ana Mincholé2, Fernando Campos3, Shuang Qian1, Gernot Plank4, Aurel Neic5, Luca Azzolin6, Edward Vigmond7, Pablo Lamata1, John Whitaker1, Martin Bishop1
1King's College London, 2University of Zaragoza, 3School of Biomedical Engineering and Imaging Sciences, King's College London, 4Medical University of Graz, 5NumeriCor GmbH, 6Karlsruhe Institute of Technology, 7LIRYC - University of Bordeaux


Abstract

Left bundle branch block (LBBB) is characterised by prolonged QRS duration (QRSd) and altered ECG morphology and reflects heterogeneous substrates of conduction disruption. The transverse electrical axis (φ_E) captures the spatial component of ventricular depolarisation, but its relationship to QRSd and sensitivity to anatomical factors such as cardiac orientation remain unclear.

A personalised biventricular–torso model was constructed from CT imaging. Ventricular activation was simulated using a reaction-eikonal model with fascicular initiation and a fast-conducting subendocardial layer representing His–Purkinje system (HPS) activation. Fascicular activation sites were optimised against the mean 12-lead ECG of the UK Biobank population. LBBB was simulated by removing left fascicular activation and varying HPS engagement. Cardiac orientation was systematically varied using population-derived orientations from UK Biobank. From each simulated ECG, QRS duration was measured and the transverse electrical axis (φ_E) was derived from the vectorcardiogram. Simulated metrics were compared against healthy and LBBB population distributions.

Discrete block of all left fascicles altered early activation and produced posterior shifts in φ_E with limited QRS prolongation. In contrast, removal of HPS engagement resulted in marked QRS widening, more closely matching clinical LBBB, with comparatively smaller additional changes in φ_E. These findings demonstrate a mechanistic dissociation between spatial and temporal ECG metrics, with φ_E primarily determined by the location of earliest activation and QRSd reflecting global propagation delay. Cardiac orientation produced variation in φ_E across all planes, while QRSd remained relatively stable within the clinical LBBB range. Orientation also influenced waveform morphology, including alternation between notching and slurring in lateral leads, without altering overall diagnostic classification. Simplified fascicle-based models are therefore suitable for studying geometric effects such as cardiac orientation. However, their limited representation of distributed HPS impairment restricts physiological reproduction of pathological conduction.