In vivo validation of a computational bone adaptation model using open-loop control and time-lapsed micro-computed tomography

Abstract Cyclic mechanical loading augments trabecular bone mass, mainly by increasing trabecular thickness. For this reason, we hypothesized that an in silico thickening algorithm using open-loop control would be sufficient to reliably predict the response of trabecular bone to cyclic mechanical loading. This would also mean that trabecular bone adaptation could be modeled as a system responding to an input signal at the onset of the process in a predefined manner, without feedback from the outputs. Here, time-lapsed in vivo micro-computed tomography scans of mice cyclically loaded at the sixth caudal vertebra were used to validate the in silico model. When comparing in silico and in vivo data sets after a period of four weeks, a maximum prediction error of 2.4% in bone volume fraction and 5.4% in other bone morphometric indices was calculated. Superimposition of sequentially acquired experimental images and simulated structures revealed that in silico simulations deposited thin and homogeneous layers of bone whilst the experiment was characterized by local areas of strong thickening, as well as considerable volumes of bone resorption. From the results, we concluded that the proposed computational algorithm predicted changes in bone volume fraction and global parameters of bone structure very well over a period of four weeks while it was unable to reproduce accurate spatial patterns of local bone formation and resorption. This study demonstrates the importance of validation of computational models through the use of experimental in vivo data, including the local comparison of simulated and experimental remodeling sites. It is assumed that the ability to accurately predict changes in bone micro-architecture will facilitate a deeper understanding of the cellular mechanisms underlying bone remodeling and adaptation due to mechanical loading.