Prognostic bone fracture healing simulations in an ovine tibia model validated with in vivo sensors.

Abstract

Bone fracture healing is a complex physiological process influenced by biomechanical and biomolecular factors. Mechanical stability is crucial for successful healing, and disruptions can lead to delayed healing or nonunion. Bone commonly heals itself through secondary fracture healing, which is governed by the mechanical strain at the fracture site. To investigate these phenomena, a validated methodology for capturing the mechanoregulatory process in specimen-specific models of fracture healing could provide insight into the healing process. This study implemented a prognostic healing simulation framework to predict healing trajectories based on mechanical stimuli. Sixteen sheep were subjected to a 3 mm transverse tibial mid-shaft osteotomy, stabilized with a custom plate, and equipped with displacement transducer sensors to measure interfragmentary motion over 8 weeks. Computed tomography scans were used to create specimen-specific bone geometries for finite element analysis. Virtual mechanical testing was performed iteratively to calculate strains in the callus region, which guided tissue differentiation and consequently, healing. The predicted healing outcomes were compared to continuous in vivo sensor data, providing a unique validation data set. Healing times derived from the in vivo sensor and in silico sensor showed no significant differences, suggesting the potential for these predictive models to inform clinical assessments and improve nonunion risk evaluations. This study represents a crucial step towards establishing trustworthy computational models of bone healing and translating these to the preclinical and clinical setting, enhancing our understanding of fracture healing mechanisms. Clinical significance: Prognostic bone fracture healing simulation could assist in non-union diagnosis and prediction.