University of Twente Student Theses


Optimizing the reference radius for preoperative surgical planning of distal radius osteotomy.

Heide, Stein van der (2022) Optimizing the reference radius for preoperative surgical planning of distal radius osteotomy.

[img] PDF
Abstract:Introduction: Distal radius osteotomies for distal radius malunion can preoperatively be planned. Currently, the affected radius is compared with the contralateral unaffected radius to find what corrections are needed. However, problems arise if no healthy contralateral radius or asymmetry is present. Therefore, there is a need for an accurate reference radius. This thesis aims to describe the magnitude of asymmetry and offers an alternative reference radius in the form of a predicted radius. Methods: Length asymmetry in the forearm was measured on 50 pairs of unaffected ulnae. Further, full shapes of radii were predicted given the shaft and proximal radius. This was done using a shape completion model based on Gaussian process regression. The model was trained on 57 healthy radii, and its performance was validated using a leave one out method. Differences between the predicted and original radius were expressed in mean and Hausdorff distances, rotational and translational differences, heatmaps of local differences, and as principal components of the differences. To assess how the model performs when a larger part of the radius is predicted, validation steps were repeated for 5%, 10% and a 15% predicted radius. Results: An absolute left/right length difference in the ulna of 2.4 ± 2.3 mm was found. Further, the mean and Hausdorff distances between the 10% predicted radius and the original radius were 0.83 ± -0.24 mm and 3.11 ± 0.69 mm. The algorithm performed statistical significant better when 5% of the radius was predicted and worse if 15% was predicted. The rotational differences between the predicted and original distal radius around the x, y, and z-axis were: 0.19 ± 2.64 deg, 0.15 ± 2.49 deg, 0.07 ± 3.19 deg. Translational differences over the x, y, and z-axis were: 0.33 ± 4.74 mm, -0.23 ± 4.72 mm, and 0.31 ± 0.88 mm. Locally, differences were mainly present in the radial styloid. Discussion: Significant left/right length asymmetry is present in the forearm, therefore, a length correction is needed when used for correction osteotomy. Further, when no contralateral radius is present, the shape of the distal radius can accurately be predicted. Because the shape of the radial styloid was predicted less accurately, it is advised to use other parts of the radius as a template for osteotomy. Lastly, an as large as possible part of the radius should be used to predict the radius.
Item Type:Essay (Master)
Faculty:TNW: Science and Technology
Programme:Technical Medicine MSc (60033)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page