PB0875 - A parallel kinetic Monte Carlo simulation of human stenotic thrombus growth with single platelet resolution from plaque rupture to occlusion

Background: Modeling thrombus growth in pathological flows would allow evaluation of risk under patient-specific pharmacological, hematological, and hemodynamical conditions.

Aims: To develop a 3D, parallelizable multiscale computational model for the prediction of thrombus growth under flow on a spatially resolved surface presenting collagen and tissue factor (TF).

Methods: The multiscale framework is composed of four coupled modules: a Neural Network (NN) that accounts for platelet signaling, a Lattice Kinetic Monte Carlo (LKMC) simulation for tracking platelet positions, a Finite Volume Method (FVM) simulator for solving convection-diffusion-reaction equations describing agonist release and transport, and a Lattice Boltzmann (LB) flow solver for computing the blood flow field over the growing thrombus. Computationally efficient parallel simulations were achieved by using open-source software where applicable: Palabos (LB), OpenFOAM (FVM), and Multiscale Universal Interface for module coupling. Parallel versions of LKMC and NN were achieved by employing a novel parallel platelet-decomposition approach.

Results: The model accurately captured the evolution of the growing thrombus; predictions agreed well with fluorescent micrographs of thrombi observed in experiments. The modularity and generalizability of the 3D multiscale solver allowed the simulations of important clinical situations such as cylindrical blood vessels and stenoses. Moreover, parallel simulations demonstrated reasonable speedup in simulation time with increase in processor cores. We were able to execute simulations in ~100-fold larger geometries (~0.25mm diameter vessels, ~0.5million adhered platelets; see Fig 1) than previously studied under single-platelet resolution (e.g., 60 micron vessels were considered by Shankar et al, PLoS Comp Biol, 2022). Enhanced platelet-platelet bonding at pathologically high shear rates (e.g., von Willebrand factor unfolding) was required for accurately describing thrombus growth in stenotic flows.

Conclusion(s): The present model, when suitably trained, may be useful in predicting the patient-specific effect of antiplatelet drug treatments and impacts of blood disorders such as von Willebrand disease in realistic 3D settings within the vasculature.