Femoral Bone Strength Prediction Using Isotopological B-Spline-Transformed Meshes

Lukas Steiner, Alexander Synek, Dieter H. Pahr

Research output: Journal article (peer-reviewed)Journal article


Finite element (FE) analysis can predict proximal human femoral strength. Automated meshing and identifying subregions with high relevance for strength prediction could reduce the laborious modeling process. Mesh morphing based on free-form registration provides a high level of automation and inherently creates isotopological meshes. The goals of this study were to investigate if FE models based on free-form transformed meshes predict experimental femoral strength as well as manually created FE models and to identify regions and parameters with highest correlation to femoral strength. Subject-specific meshes and FE models were created from a set of quantitative CT images (QCT) using a B-Spline registration-based algorithm. Correlation of FE-predicted bone strength and local parameters with experimental bone strength were investigated. FE models based on transformed meshes closely resembled manually created counterparts, with equally strong correlations with experimental bone strength ((Formula presented.) vs. (Formula presented.)). The regional analysis showed strong correlations ((Formula presented.)) of experimental strength with local parameters. No subregion or parameter lead to stronger correlation than FE predicted bone strength. B-spline-transformed meshes can be used to create FE models, able to predict femoral bone strength and simplify FE model generation. They can be used to reveal relations of local parameters with failure load.

Original languageEnglish
Pages (from-to)125-137
Number of pages13
Issue number1
Publication statusPublished - Mar 2022


  • B-Spline transformation
  • finite element
  • human femora
  • isotopology
  • mean model
  • QCT

ASJC Scopus subject areas

  • Rehabilitation
  • Biomedical Engineering
  • Orthopedics and Sports Medicine


Dive into the research topics of 'Femoral Bone Strength Prediction Using Isotopological B-Spline-Transformed Meshes'. Together they form a unique fingerprint.

Cite this