In this chapter, we highlight complementary roles of image analysis and computational biomechanics. We show how medical images are applied to create patient-specific computational biomechanics models for surgical simulation and how such models inform image analysis by computing deformation fields within the human body tissues for nonrigid image registration. Rapid generation of patient-specific models is crucial for the adoption of computational biomechanics in the clinic. We discuss the algorithms that facilitate creation of computational grids for such models without the need for image segmentation and finite element meshing. We also consider methods ensuring accurate solutions for tissues with nonlinear mechanical properties undergoing large deformations. The methods we suggest are robust enough to be used by analysts not being experts in methods of nonlinear computational biomechanics. We illustrate our discussion with two examples of intra-patient neuroimage registration and an example of whole body computed tomography (CT) image registration. The examples show that accurate (within the voxel size of intraoperative image) biomechanics-based registration can be achieved without accurate data about the patient-specific material properties of soft tissues and with only very sparse information about the intraoperative deformations.
|Title of host publication||Handbook of Medical Image Computing and Computer Assisted Intervention|
|Editors||S. Kevin Zhou, Daniel Rueckert, Gabor Fichtinger|
|Publication status||Published - 25 Oct 2019|
Wittek, A., & Miller, K. (2019). Computational biomechanics for medical image analysis. In S. K. Zhou, D. Rueckert, & G. Fichtinger (Eds.), Handbook of Medical Image Computing and Computer Assisted Intervention (pp. 953-977). Elsevier.