Session: 16-01-01: Government Agency Student Poster Competition

Paper Number: 149764

149764 - Coupled Multiphysics Finite Element Models Using Fluid Flow, Mass Transfer, and Structural Deformation of In-Vitro Tumor Models to Estimate the Sampling Volume of a Fine Needle Aspiration Biopsy 

Introduction:

Fine Needle Aspiration (FNA) is a common medical procedure where a thin needle is inserted into a tissue of interest to collect a liquid sample containing biological for cancer diagnosis.

Despite FNA’s critical role in the diagnosis of solid tumors, the volume which the sample is collected from during this procedure is undefined and has not been previously modelled.

We investigated the use of coupled multiphysics finite element models (FEMs) to numerically simulate and estimate the sampling volume of FNA within a solid tumor mimic.

Materials and Methods:

A 2D finite element model (FEM) (19,500 elements) was generated using Comsol Multiphysics to represent a tumor on a chip device containing a 1 x 1 cm tumor mimic. Mesh refinement was calibrated to the dominating physics phenomena for each domain.

Spatial tumor heterogeneity and porosity information was incorporated into another 2D FEM (11,929 elements) by rasterizing an image of a mouse tumor stained with hematoxylin and eosin and assigning each mesh element an individualized porosity value via linear bilaterial interpolation (MatLab).

Time-dependent Eulerian-Lagrangian methods were used to model interstitial fluid, material transfer, and particle dynamics with Navier-Stoke’s, Darcy’s, and Newtonian First Order Formulation equations respectively, and the mechanics of structural pore deformation due to needle suction was implemented. Particles were assumed to be average cell size [GS1] and particle inertial forces were assumed to be negligible.

Hydrogel material properties (storage and loss modulus) were calibrated using data collected from Rheometer instrumentation (Discovery HR-3 Hybrid Rheometer) and modeled as a hyperelastic, incompressible, porous material using 1^st order Ogden with parameters available in literature defining hydrogel FEMs (α=4.18, μ₀=2.25 kPa, κ = 1 GPa).

Clinically relevant FNA suction parameters (19-gauge needle, 5 mL syringe, -25 kPa pressure, 20s) were used as outlet boundary conditions for all models, and the edges of the device and tumor were set as wall conditions at atmospheric pressure.

Multifrontal Massively Parallel Sparse Direct Solver (MUMPS) and Parallel Sparse Direct Solver (PARDISO) were the solution schemes for the tumor on a chip device and tumor slice FEMs respectively, and both simulations utilized a nonlinear Newton iteration for the termination method.

Simulations were validated by performing FNA on in-vitro tumor hydrogels and microfluidic hydrogel devices.  

Results and Discussion:

In the tumor device FEM, constant suction at -25 kPa was observed to produce a mass flow rate of 16 uL/s of liquid at the entrance of the needle and 20 seconds of constant negative pressure was shown to access material from a 1.3mm radius at the tip of the needle. Due to the low induced pressure the fluid velocity remained well within the laminar flow region, resulting in negligible pore deformation (maximum of 1.3nm) during the FNA procedure.

The 2D tumor FEMs showed that under constant -25 kPa suction, various pores in the heterogeneous tumor model experience a positive pressure (20 Pa), whereas the pores within homogeneous tumor model exclusively experience negative pressures. The heterogeneous tumor model has larger average and median pore pressures (5% and 5%, respectively) and counterintuitively shows a more uniform flow pattern than the homogeneous tumor model.

Conclusion:

Computational models of tumor and tumor mimic devices can be used to estimate the FNA sample volumetric region of a tumor. This can help cancer diagnosis by better defining the regions of interest and fine-tuning needle placement prior to an FNA procedure.

Presenting Author: Mary Chase Sheehan University of Massachusetts Amherst

Presenting Author Biography: Mary Chase is a doctoral student in mechanical and industrial engineering. She is interested in developing multidisciplinary mathematical models to simulate biological fluid flow, mass transport, and structural deformation.

Authors:

Mary Chase Sheehan University of Massachusetts Amherst
David Schmidt University of Massachusetts Amherst
Govind Srimathveeravalli University of Massachusetts Amherst