Related Experiment Video
Updated: Jul 6, 2025
The Use of Mixed Reality in Custom-Made Revision Hip Arthroplasty: A First Case Report
Published on: August 4, 2022
Preoperative 3-dimensional computed tomography bone density measures provide objective bone quality classifications
Daniel Ritter1, Patrick J Denard2, Patric Raiss3
1Department of Orthopedic Research, Arthrex GmbH, Munich, Germany; Department of Orthopaedics and Trauma Surgery, Musculoskeletal University Center Munich (MUM), University Hospital, LMU Munich, Munich, Germany.
Preoperative computed tomography (CT) scans can accurately assess proximal humerus bone density for stemless anatomic total shoulder arthroplasty (aTSA). A machine learning model improves bone quality classification, aiding in selecting patients for stemless component implantation.
Area of Science:
- Orthopedic surgery
- Biomedical imaging
- Materials science
Background:
- Lack of reproducible methods for determining bone density for stemless anatomic total shoulder arthroplasty (aTSA).
- Need for objective assessment of proximal humerus bone quality for surgical planning.
- Importance of accurate bone density evaluation for stemless humeral component selection.
Purpose of the Study:
- Evaluate the utility of preoperative computed tomography (CT) imaging for assessing proximal humerus bone density.
- Determine if 3-dimensional (3-D) CT bone density measures can objectively classify bone quality for stemless aTSA.
- Develop an objective tool for preoperative selection criteria for stemless humeral component implantation.
Main Methods:
- Analysis of cadaveric humerus CT scans and micro-CT (μCT) imaging.
- Retrospective application to a clinical cohort (n=150) with preoperative CT and intraoperative assessment.
- Utilized a machine learning model (Support Vector Machine - SVM) with patient-specific calibration for bone density classification.
Main Results:
- Good to excellent accuracy for cancellous bone densities (metaphysis ICC=0.986, epiphysis ICC=0.883).
- Patient-specific calibration significantly reduced biases and variance compared to standard CT scans (P<0.0001).
- SVM model achieved 87.3% accuracy and 0.93 AUC for bone quality classification, outperforming conventional statistics.
Conclusions:
- Preoperative CT imaging provides accurate evaluation of proximal humerus bone densities.
- 3-D regions of interest, patient-specific calibration, and machine learning enable objective bone quality classification.
- This approach offers a valuable tool for extending preoperative selection criteria for stemless humeral components.
