Postmenopausal osteoporosis is a disease manifesting in degradation of bone mass and microarchitecture, leading to weakening and increased risk of fracture. Clinical trials are an essential tool for evaluating new treatments and may provide further mechanistic understanding of their effects in vivo. However, the histomorphometry from clinical trials is limited to 2D images and reflects single time points. Biochemical markers of bone turnover give global insight into a drug's action, but not the local dynamics of the bone remodeling process and the cells involved. Additionally, comparative trials necessitate separate treatment groups, meaning only aggregated measures can be compared. In this study, in silico modeling based on histomorphometry and pharmacokinetic data was used to assess the effects of treatment versus control on μCT scans of the same biopsy samples over time, matching the changes in bone volume fraction observed in biopsies from denosumab and placebo groups through year 10 of the FREEDOM Extension trial. In the simulation, treatment decreased osteoclast number, which led to a modest increase in trabecular thickness and osteocyte stress shielding. Long‐term bone turnover suppression led to increased RANKL production, followed by a small increase in osteoclast number at the end of the 6‐month–dosing interval, especially at the end of the Extension study. Lack of treatment led to a significant loss of bone mass and structure. The study's results show how in silico models can generate predictions of denosumab cellular action over a 10‐year period, matching static and dynamic morphometric measures assessed in clinical biopsies. The use of in silico models with clinical trial data can be a method to gain further insight into fundamental bone biology and how treatments can perturb this. With rigorous validation, such models could be used for informing the design of clinical trials, such that the number of participants could be reduced to a minimum to show efficacy. © 2021 The Authors. JBMR Plus published by Wiley Periodicals LLC on behalf of American Society for Bone and Mineral Research.

Mechanical loading allows both investigation into the mechano-regulation of fracture healing as well as interventions to improve fracture-healing outcomes such as delayed healing or non-unions. However, loading is seldom individualised or even targeted to an effective mechanical stimulus level within the bone tissue. In this study, we use micro-finite element analysis to demonstrate the result of using a constant loading assumption for all mouse femurs in a given group. We then contrast this with the application of an adaptive loading approach, denoted real time Finite Element adaptation, in which micro-computed tomography images provide the basis for micro-FE based simulations and the resulting strains are manipulated and targeted to a reference distribution. Using this approach, we demonstrate that individualised femoral loading leads to a better-specified strain distribution and lower variance in tissue mechanical stimulus across all mice, both longitudinally and cross-sectionally, while making sure that no overloading is occurring leading to refracture of the femur bones.

No abstract