Original Article
Rapid determination of internal strains in soft tissues using an experimentally calibrated finite element model derived from magnetic resonance imaging
Abstract
Background: Finite element models (FEMs) of medical images can provide information about the underlying tissue that cannot be obtained from the original images. Preforming an accurate simulation requires the careful experimental calibration of boundary conditions. Here we describe a method for deriving a geometric mesh for soft biological materials using a magnetic resonance imaging (MRI) system, and an experimental workflow for calibrating the boundary conditions and optimizing the mesh density in these simulations.
Methods: A three-dimensional image stack of a ballistic sphere gel, a bovine caudal intervertebral disc (IVD), and a human lumbar IVD were generated using a positional MRI system. These images were then segmented using a semi-automated process, converted to a tetrahedral mesh, and then modeled as a linear elastic solid. The mesh density was optimized based on simulation time and convergence with the experimental results. The modulus of the ballistic gel was determined experimentally, while the material properties for the nucleus pulposus (NP) and the annulus fibrosus (AF) within the bovine and human IVDs were assigned from literature. The simulation for the spherical gel and the bovine IVD matched the reaction forces determined experimentally in compression. We then simulated a 0.3 MPa compressive load on the human lumbar IVD at the optimal mesh density and material properties determined from the bovine model and then examined the resultant internal strains.
Results: The scaled mesh density demonstrated excellent correspondence with the experimental results, confirming that accuracy was not compromised. Both the ballistic gel and the IVD samples exhibited a wide range of internal strains. The NP of the IVD underwent greater deformation than the AF under loading.
Conclusions: This study validated a strategy for mesh optimization and FEM of soft biological materials from data generated from MRI scans. This calibrated approach allows for the rapid examination of internal strain distributions medical images that can be performed on the order of minutes.
Methods: A three-dimensional image stack of a ballistic sphere gel, a bovine caudal intervertebral disc (IVD), and a human lumbar IVD were generated using a positional MRI system. These images were then segmented using a semi-automated process, converted to a tetrahedral mesh, and then modeled as a linear elastic solid. The mesh density was optimized based on simulation time and convergence with the experimental results. The modulus of the ballistic gel was determined experimentally, while the material properties for the nucleus pulposus (NP) and the annulus fibrosus (AF) within the bovine and human IVDs were assigned from literature. The simulation for the spherical gel and the bovine IVD matched the reaction forces determined experimentally in compression. We then simulated a 0.3 MPa compressive load on the human lumbar IVD at the optimal mesh density and material properties determined from the bovine model and then examined the resultant internal strains.
Results: The scaled mesh density demonstrated excellent correspondence with the experimental results, confirming that accuracy was not compromised. Both the ballistic gel and the IVD samples exhibited a wide range of internal strains. The NP of the IVD underwent greater deformation than the AF under loading.
Conclusions: This study validated a strategy for mesh optimization and FEM of soft biological materials from data generated from MRI scans. This calibrated approach allows for the rapid examination of internal strain distributions medical images that can be performed on the order of minutes.
