Heart in a knot: unraveling the impact of the nested tori myofiber architecture on ventricular mechanics

Abstract

The intricate three-dimensional organization of cardiac myofibers and sheetlets plays a critical role in the mechanical behavior of the human heart. Despite extensive research and the development of various rule-based myofiber architecture surrogate models, the precise arrangement of these structures and their impact on cardiac function remain subjects of debate. In this study, we present a novel myofiber architecture surrogate inspired by Streeter’s nested tori conjecture, modeling the left ventricle as a series of smoothly twisting toroidal surfaces populated by continuous myofiber and sheetlet fields. Leveraging high-fidelity cardiac computational modeling approaches, we systematically evaluated the biomechanical performance of this nested tori architecture against conventional rule-based nested ellipsoidal models. Our results demonstrate that the nested tori architecture aligns more closely with experimental data on physiological myofiber and sheetlet angles. Notably, it enhances sheetlet mobility—a key mechanism for effective cardiac pumping—resulting in higher ejection fraction, greater global deformation, and a more physiological wall rotation pattern. Additionally, it produces a more homogeneous myofiber stress distribution and increased myofiber shortening during ejection. These findings suggest that the nested tori architecture provides a compelling alternative to conventional nested ellipsoidal models, offering a more physiologically consistent representation of myocardial structure and its functional implications. By enabling improved biomechanical performance in silico, this approach supports further investigation into how detailed myoarchitectural continuity shapes cardiac function. Ultimately, it may open promising avenues for advancing cardiac diagnosis, guiding the design of bioinspired implants and devices, and deepening our understanding of both healthy and diseased cardiac mechanics.