Biomedical Engineering Reference
In-Depth Information
near the boundaries may be affected; therefore, postprocessing is used as a final
step in order to produce meshes with no degenerate or poor quality elements.
Postprocessing operations involve optimization of the mesh node locations and
the use of edge-split and edge-collapse operations where necessary to improve
element quality.
Based on brain tumor simulations conducted initially, severe distortion occurs
in the FE mesh due to the large deformation associated with growing large tumors.
This affects the convergence of the FE solution and may cause early termination
of the simulation before the desired realistic tumor sizes are reached. To deal with
this issue, we have included an ability to conduct the tumor growth simulations in
a stepped fashion, and correct the distorted mesh (remesh) between steps. Details
of the employed FE mesh generation and remeshing methods can be found in [99].
4.3.7. Experiments and results
Here, deformations predicted by the described tumor mass-effect model are
compared to deformations observed in four real brain tumor cases. These experi-
ments serve two purposes. First, this dataset of real tumor cases is used to guide
the selection of the parameters of the adopted constitutive material model for brain
tissues. Second, the accuracy of the model in reproducing deformations in real
tumor cases is evaluated.
The rest of this section is organized as follows. A description of the used
dataset is provided in Section 4.3.7.1. In Section 4.3.7.2 the steps involved in
applying the proposed biomechanical model to this dataset are explained. Ex-
periments performed to select the material constitutive parameter values for brain
tissues are described in Sections 4.3.7.3 and 4.3.7.4. Using the material model
parameters and mass-effect model parameters providing the best agreement with
the observed deformation in the real cases, partial validation experiments were
performed. The results of these experiments are reported in Section 4.3.7.5. Fi-
nally, a tumor mass-effect simulation starting with an MR brain image of a normal
subject is demonstrated in Section 4.3.7.6.
4.3.7.1. Dataset
Four brain tumor cases with serial MRI scans were used in the
experiments described here. Some parameters of this dataset can be found in
Table 1, and example 2D images are shown in Figure 14, and later in Figure 18.
Assuming that canine brain tissue properties and tumor growth process are
reasonably representative of their counterparts in humans, three of the studied
tumor cases were dogs with surgically transplanted glioma cells. These dogmodels
were prepared according to the protocol first described by Wodinsky et al. [101].
(The procedure was carried out by Dr. James Anderson and Carolyn Magee
at the Johns Hopkins University School of Medicine and was approved by the
Animal Care and Use Committee. The animal welfare act and guidelines for the
ethical and humane treatment of animal were observed at all times.) A baseline
