Abstract
Introduction
Short-segment posterior instrumentation for spine fractures is threatened by unacceptable failure rates. Two important design objectives of pedicle screws, bending and pullout strength, may conflict with each other.
Hypothesis
Multiobjective optimization study with artificial neural network (ANN) algorithm and genetic algorithm (GA)
Materials & Methods
Three-dimensional finite element (FE) methods were applied to investigate the optimal designs of pedicle screws with an outer diameter of 7 mm using a multiobjective approach for these two objectives. Based on the FE results on an L25 orthogonal array, two objective functions were developed by an ANN algorithm. Then, the trade-off solutions known as Pareto optima were explored by a GA. The optimal design was validated by mechanical tests.
Results
The knee solutions of the Pareto fronts had simultaneous high bending and pullout strength ranging from 92 to 94 percent of their maxima. The corresponding range of the design parameters was 3.8 to 4.06 mm for inner diameter and 3.21 to 3.3 mm for pitch; 0 mm for beginning position of conical angle, 0.4 mm for proximal root radius, 5 degrees for proximal half angle, and 0.1 mm for thread width. The optimal design was well validated by mechanical tests, comparing with commercially available pedicle screws.
Discussion & Conclusions
The optimal design of pedicle screws obtained could achieve an ideal with high mechanical performance in both bending and pullout tests.
