Yijun Zhou: Mechanical optimization of orthopaedic bone screw constructs

  • Date: 7 December 2023, 09:15
  • Location: Polhemsalen, Lägerhyddsvägen 1, Uppsala
  • Type: Thesis defence
  • Thesis author: Yijun Zhou
  • External reviewer: Richie Gill
  • Supervisors: Cecilia Persson, Benedikt Helgason, Philip Procter
  • Research subject: Engineering Science with specialization in Biomedical Engineering
  • DiVA

Abstract

Orthopaedic screw implants are crucial in surgical procedures for bone structures, with a longstanding issue being screw loosening. Improving screw stability requires an enhanced understanding of the bone-implant interactions and associated failure mechanisms. This knowledge can be used to optimize screw designs.

In this thesis, a first step towards this end was taken through the development of enhanced trabecular bone models. Due to the trabecular bone's complex structure and the difficulties in obtaining micro-structural information, we proposed a numerical model based on Voronoi tessellation to mimic the trabecular bone morphology across varying porosities. This model's mechanical properties aligned well with analytical formulas and finite element modelling of real bone specimens, showing strong agreement with experimental results.

Further investigation into screw-trabecular bone interaction was carried out using two numerical models. The explicit finite element models were able to replicate experimental screw push-in results. While only one of the models accounted for the screw insertion step, both models showed strong congruence with key experimental results. The two-step simulation however led to a more physically plausible Young’s modulus for trabecular bone material.

Utilizing a validated numerical model, a numerical optimization process was initiated, where novel screw designs were proposed. Several neural network surrogate models were introduced, reducing evaluation costs while maintaining prediction accuracy. We found that screw insertion position, trabecular bone porosity, and orientation were significant factors, explaining about 96% of the variance in predicted response. 

Furthermore, a neural network workflow was developed to generate super-resolution trabecular bone models from clinical CT data, improving accuracy by up to 700% both morphologically and mechanically, using micro-CT models as a benchmark.

Lastly, the potential enhancement of screw's primary stability was explored, by injection of hyaluronic acid mixed with hydroxyapatite particles. The augmentation effect was influenced by the hydroxyapatite particles' size and shape. Considering that these particles can also promote bone growth, a particle-laden hydrogel injection could potentially enhance screw stability throughout its lifespan.

In conclusion, this study proposed methods to elucidate bone-implant interaction and enhance screw stability in trabecular bone, encompassing trabecular bone modelling, bone-screw interaction investigations, screw geometric optimization, super-resolution trabecular bone model generation from clinical CT data, and hydrogel optimization for stability. The results provide an enhanced understanding as well as optimization of complex geometric interactions, and could lead to future enhancements of clinical practice in terms of screw stability.

FOLLOW UPPSALA UNIVERSITY ON

facebook
instagram
twitter
youtube
linkedin