Comparison of algorithms for automated femur fracture reduction

Int J Med Robot. 2018 Feb;14(1). doi: 10.1002/rcs.1864. Epub 2017 Sep 24.

Abstract

Purpose: We designed an experiment to determine the comparative effectiveness of computer algorithms for performing automated long bone fracture reduction.

Methods: Automated reduction of 10 3D fracture models was performed using two computer algorithms, random sample matching (RANSAM) and Z-buffering (Z-Buffer), and one of five options of post-processing: none; iterative closest point algorithm (ICP); ICP-X1; ICP-X2; and ICP-X3. We measured the final alignment between the two fragments for each algorithm and post-processing option.

Results: The RANSAM algorithm combined with postprocessing algorithm ICP-X1 or ICP-X3 resulted in the most accurate fracture reduction in the translational plane. No discernible difference was observed in the rotational plane. Automated reduction had more accurate translational displacement than telemanipulated manual reductions.

Conclusion: This study supports the use of the RANSAM algorithm for automated fracture reduction procedures. The use of ICP algorithms provides further optimization of the initial reduction.

Keywords: femur fracture; intramedullary nailing; robotic fracture reduction; telemanipulated fracture reduction.

Publication types

  • Comparative Study

MeSH terms

  • Algorithms
  • Automation
  • Cadaver
  • Electronic Data Processing*
  • Femoral Fractures / surgery*
  • Femur / surgery
  • Fracture Fixation / instrumentation*
  • Fracture Fixation / methods*
  • Fracture Healing
  • Humans
  • Models, Statistical
  • Motion
  • Normal Distribution
  • Reproducibility of Results
  • Rotation
  • Signal Processing, Computer-Assisted*
  • Software
  • Tomography, X-Ray Computed
  • User-Computer Interface