(329d) A 3D Multiscale Model for Prediction of Patient-Specific Platelet Function Under Flow

Atherosclerotic plaque rupture exposes blood to a highly procoagulant surface containing tissue factor, causing platelets to become activated. Activated platelets release soluble agonists such as ADP and thromboxane, causing further activation and platelet recruitment, resulting in thrombus growth. In this work, we have extended our previous 2D multiscale model of thrombosis¹ in a microfluidic channel to develop a fully spatially resolved 3D framework with capabilities to simulate patient-specific phenotypes and blood vessel geometries. The multiscale framework is composed of four modules: Neural Network for platelet signaling, Lattice Kinetic Monte Carlo for platelet motion and binding, Lattice Boltzmann for blood flow field, and Finite Volume Method for tracking soluble agonist concentrations. Thrombus growth dynamics and morphology predicted by the model agree well with experiments. Moreover, the model clearly overcomes artifacts that were present in our previous 2D model¹ by allowing for flow around bound platelets rather than obstructing flow altogether. Furthermore, the model is largely agnostic to the geometry of the channel and can easily be applied to fully resolved simulations of thrombus growth in arbitrary geometries, such as stenoses and bifurcations. To our knowledge, our model is the first of its kind that accounts for patient-specific platelet phenotypes to perform robust 3D simulations of thrombosis given any arbitrary geometry. This framework facilitates the identification of patient-specific thrombotic risks and drug responses.

References:

1. Lu et al., Math. Med. Biol., 2017, Vol. 34(4), pp. 523-546