Computational methods development for design and optimization of spinal fusion implants

As computers get more powerful, optimizing structures with high-resolution features becomes more and more feasible. This creates fascinating opportunities in computational design e.g. using topology optimization for designing biomechanical structures. In my research, I develop large-scale topology optimization methods with the overall goal of improving patient’s well-being and reducing health-care cost by designing truly optimal patient specific spinal fusion implants.

Reengineering a vertebra using topology optimization:
A domain is defined with idealized dimensions of a human vertebra. A compressive load is applied, representative of the compressive loads in the lumbar spine. The goal of the optimization is to maximize the stiffness of the structure subjected to constraints in porosity and material volume. Figure 1 shows the reengineered vertebra resulting from the topology optimization process.

Multiple loads cases
The human spine is exposed to complex loading in vivo depending on activities. Therefore, it is essential that the topology optimization code can handle multiple load cases. This functionality has been implemented. Figure 2 shows the results of a test case demonstrating the difference between optimizing a structure for all loads being applied in the same load step vs. optimizing a structure with loads being applied in different load steps. The resulting designs show a clear difference between the two approaches. This is an important observation because porous structures are more resistant to fatigue and peak loads (Torres et al., 2019). Therefore, the multi-load functionality contributes to the robustness of the designs that the code can generate.

Back to NU-SPINE Network Members

Last modified: 2021-08-13