Date of Award
Paul J. Rullkoetter, Ph.D.
Calibration and validation, FEA, Finite element analysis, Probabilistic analysis, Spine, Total disc replacement
Validated finite element (FE) models of the functional spinal unit (FSU) and lumbar spine are essential in design-phase device development and in assessing the mechanics associated with normal spine function and degenerative disc disease (DDD), as well as the impact of fusion and total disc replacement (TDR). Although experimental data from fully intact specimens can be used for model calibration and validation, the contributions from the individual structures (disc, facets, and ligaments) may be inappropriately distributed. Hence, creation of decompression conditions or device implantations that require structure removal may not have the proper resulting mechanics. An explicit FE formulation may be advantageous compared to standard analysis due to efficiency in handling complex, changing contact conditions and the ability to evaluate either rigid or deformable body contact. Also, probabilistic studies based on these deterministic FE formulations are of great interest currently as model input parameters (such as properties of nucleus, annulus, ligament stiffness and facet material and geometric orientation) have been characterized experimentally, but contain substantial variability. The use of these FE formulations is not only valuable from an intact spine point of view, but also relevant in understanding and improving the design outcome of procedures like the total disc replacement (TDR). It has been shown that clinical outcome and the incidence of adjacent level disease is linked to the range of motion achieved by the introduction of these disc replacement devices like the ProDisc-L (Synthes Spine, West Chester, PA). Placement of the spherical center of the device as close as possible to the anatomical axis of rotation of the segment is essential in achieving optimal performance.
Accordingly, an explicit FE model of the lumbosacral spine and FSU's L2-L3 and L4-L5 using subject-specific in vitro data was developed using sequential transection of each structure. In addition, the objective of this dissertation was to develop a computationally efficient, probabilistic explicit FE model of the lumbar spine, to evaluate spine mechanics for the FSU's L2-L3 and L4-L5. This probabilistic modeling approach was used to assess the capability of efficient probabilistic analyses to predict performance incorporating disc and ligament material variability as well as geometric variability of the facet joint. A well calibrated deterministic and probabilistic model can be used as an excellent computational tool to predict the behavior of the spine with implants like the ProDisc-L. This dissertation also investigates the effect of altering the position of the Prodisc-L implanted in a FE model on ROM during flexion-extension, lateral bending, and axial rotation. Specifically, ROM, bone impingement, implant impingement, and facet forces were evaluated with varying anterior-posterior and medial-lateral placement of the TDR implant.
The uniqueness of this work is the method developed to tune the individual structures in calibrating the FE model using sequential sectioning. This strong calibration against subject-specific in vitro data developed confidence in the predictive power of this FE model. For an applied torque, the rotational root mean squared error between the model predictions and the experimental results were within 0.15deg averaged during flexion and extension. The probabilistic analysis compared some of the advanced reliability and probabilistic techniques with the Monte Carlo simulation which is considered the gold standard. The efficient methods accurately estimated the results from Monte Carlo simulation in approximately 5% of computational time. This study on the implanted spine performed on four different spine models showed the importance of using FE techniques as a pre-op templating tool in decision making process for spinal procedures.
Rao, Milind, "Explicit Finite Element Modeling of the Human Lumbar Spine" (2012). Electronic Theses and Dissertations. 906.
Received from ProQuest
Mechanical Engineering, Biomechanics