Detaillierte Beitrags-Information

Zurück zur Liste

Beitragstitel Computational model predicts risk of spinal screw loosening in patients
Beitragscode P18
  1. Marie-Rosa Fasser Balgrist University Hospital, University of Zurich Vortragender
  2. Gabriela Gerber Institute for Biomechanics, ETH Zurich, Zurich, Switzerland
  3. Caroline Passaplan Department of Orthopaedics, Balgrist University Hospital, Zurich, Switzerland
  4. Frédéric Cornaz Department of Orthopaedics, Balgrist University Hospital, Zurich, Switzerland
  5. Jess Snedeker Universitätsklinik Balgrist
  6. Mazda Farshad Balgrist University Hospital
  7. Jonas Widmer University Spine Center Z8 Zurich, Switzerland; Institute for Biomechanics, ETH Zuürich, Balgrist University Hospital, University of Zurich, 800rich, Zurich, Switzerland
Präsentationsform Poster
  • A03 - Wirbelsäule
Abstract Introduction: Pedicle screw loosening after spondylodesis surgery is a common complication and it has been reported to occur in more than 50% of cases among patients with osteoporotic bone. Pre-operative knowledge about the risk of insufficient screw fixation might lead to more personalized and therefore successful treatment. In this case-control study, the use of computational biomechanical models for quantifying screw loosening risk was evaluated.

Methods: Based on pre-operative and post-operative imaging data, individual biomechanical models were retrospectively generated for 16 patients who underwent spondylodesis surgery. Screws that successfully integrated into the vertebral bone (control group) were compared to loose screws (case group). For this purpose, finite element models rendering the entire instrumented region of each patient’s lumbar spine were generated. Joint loads obtained from patient-specific musculoskeletal simulations were applied to render physiological loading during standing. The mean Hounsfield unit values, von Mises Stress, and a parameter quantifying the relationship between the predicted stress and bone strength (“loading factor”) around the screw were evaluated. Additionally, models consisting of only a single vertebra and a screw were prepared for 26 vertebrae and the screws were loaded along the longitudinal axis to simulate pull-out.

Results: The mean Hounsfield unit and the stress values in the vicinity of the screws differed significantly between the case and the control group (p = 0.017 and p = 0.029 for the Hounsfield unit and stress read-out, respectively). However, discrimination and predictability were best when considering the loading factor (p = 0.001, R2Pseudo = 0.34, AUC = 0.87). Specifically, a loading factor higher than 35% suggests a high risk of loosening, while a factor lower than 10% indicates an almost 0 risk of loosening to occur. Finite element analysis of screw pull-out models suggests that no association exists between pull-out strength and implant loosening (p = 0.7953).

Conclusion: The results of this study indicate that the risk of implant fixation failure could be reliably quantified with a biomechanics-based score describing the relationship between local bone mechanical properties and the simulated stress. With the outlined approach, a cost-efficient improvement of patient care might be possible. Next, a prospective study is required to further assess the clinical feasibility of the method.