Mechanical optimization of screw constructs for spinal applications

The pedicle screw is one kind of important implant to fix the bone fracture and deformity. However, screw loosening is a crucial problem, especially for osteoporotic patients, which could lead to secondary surgery, increasing the pain of patients and influence the speed of recovery.

In the last decades, different screw designs were tested to output the pullout strength, which is considered to be the most important parameter to indicate the screw stability in cancellous bones. The relation between design parameters of a screw, for example, pitch height, conical angle etc., and the performance were researched based on the experimental data. However, the pedicle screw is still lack of optimization due to the stochasticity of trabecular structure, screw insert position, and numerical complexity of the FE model.

During this project, we want to find the relation between screw stability and pullout data, for instance, pullout strength, stiffness, dissipation energy. Based on the mapping assumption, a numerical optimization routine for the screw stability would be established. FEM and surrogate model would be utilized in the optimization framework, which could combine the advantages of both numerical methods and experiments. In the numerical part, the trabecular and screw geometry would be generated, the contact between trabecular bone and screw would be evaluated, and the fracture mechanisms of trabecular structure would be verified. The FE model should be validated and combined with the experimental data to provide objective values for the optimizer. In the Experimental part, a DOE would be implemented to create the database for several key parameters of screw and bone. After that, a surrogate model would be built by machine learning and numerical optimizer could be used to search the best results, i.e. the most stable screw geometry for specific bone structure. Optimized screws should be manufactured with 3D printing for different working conditions based on the previous techniques, ideally according to BMD of different patients, and be tested in in vitro specimens.

Back to NU-SPINE Network Members

Last modified: 2023-06-09