Biomedical Engineering Reference
In-Depth Information
1
Introduction
Neurosurgical planning for image-guided interventions is typically conducted using
high quality preoperative radiographic images. Craniotomy tends to distort the
preoperative geometry and lead to misalignment between the positions of pathology
determined from preoperative images and their actual positions [
1
]. In the past, the
prediction of such intraoperative deformations relied solely on image-based
methods and did not consider the biomechanics of the brain [
2
]. In recent years,
biomechanical models that ensure plausibility of the predicted intraoperative
deformations have become a viable alternative to purely image-based methods
for image-guided surgery [
3
].
The majority of biomechanical models for predicting such deformations utilize
the finite element method (FEM) to discretize the domain and solve sets of partial
differential equations governing the mechanical behavior of the analyzed soft
organ. Results from these finite element models are promising, demonstrating that
a high level of precision can be achieved, and solution times are well within the
real-time constraints of image-guided surgery [
3
]. However, the efficient generation
of the patient-specific computation grids (finite element meshes) for finite element
models from medical images remains a major bottleneck that prevents widespread
application of computational biomechanics in clinical flow.
Traditionally, the process of generating patient-specific computation grids contains
numerous independent steps as shown in Fig.
1a
. Therein, segmentation and high
Fig. 1 Pipeline of computational grid generation for finite element model (a), and Statistical
meshless model (b)
