Computational fluid dynamics (CFD) modelling of liquid embolic agents (Onyx) used in brain arteriovenous malformation (AVM) treatment to predict the distal penetration behavior

Introduction

Benchtop and animal models have traditionally been used to study the propagation of Onyx Liquid Embolic Systems (Onyx) used in the treatment of brain arteriovenous malformations (AVM). However, such models are costly, do not provide sufficient detail to elucidate how variations in Onyx viscosity alter flow dynamics, and rely on some trial-and-error, resulting in elongated timelines for product development.

Objectives

The goal of this study was to leverage Computational Fluid Dynamics (CFD) simulations to predict the behavior of different Onyx formulations. The key objectives were to: 1) validate the distal penetration results from CFD simulations with existing data from bench experiments, 2) compare the flow characteristics of Onyx formulations with differing viscosities in a blood vessel, 3) elucidate the impact of viscosity on distal penetration, and 4) understand how injection location affects distal penetration.

Methods

Using two-dimensional (2D) CFD simulations, we evaluated the propagation of two Onyx formulations (Onyx 18 and Onyx 34) inside a virtual neurovasculature filled with flowing water to mimic the presence of blood in blood vessels. Onyx was assumed to be a mixture of DMSO and EVOH. A physics-based model was developed to account for the change in viscosity of Onyx resulting from migration of DMSO in Onyx to the surrounding fluid (water). Navier–Stokes equations were solved using the commercially-available, finite-volume CFD code, Ansys Fluent. The mixture multiphase model in Fluent was used to track the evolution of the two fluids (Onyx and water), and a species transport equation was solved to account for mass transfer of DMSO from Onyx to water.

Results

The multiphase, multispecies flow simulations were validated by comparing the distal penetration after reflux with available experimental results from bench tests. The predictions from the simulation capture the lava-like flow behavior of Onyx and closely match the experimental data of distal penetration. As expected, lower viscosity Onyx 18 penetrated more distally than Onyx 34 when evaluated with the same degree of reflux. Next, from the simulation results, the impact of viscosity change and the impact of injection location were analyzed.

Key conclusions

Computational modeling and simulation can be used to create and analyze in-silico models representing physical systems and rapidly perform large numbers of tests to evaluate the different resulting outcomes without the need to build analogous physical prototypes. To the best of our knowledge, this is the first study to provide validation of multiphase CFD simulations against benchtop experimental data for Onyx embolization.