Abstract text (incl. figure legends and references)
Introduction: The analysis of differences in the shape of 3D structures plays an important role in a large variety of studies. For instance in genetic modified mouse models the assessment if the modified gene causes a wanted or unwanted skeletal phenotype is important to predict the involvement of the gene in other pathways. To study other biological process acting on the skeleton such as the Dehnel effect, the skeleton of different specimens need to compared, which requires an analysis that removes the influence of the size of the specific animal from the analysis of the shape of the skeleton. Another typical task is the generation of facsimiles by 3D printing. Dependent on the used technology and the structure of the object to print, deformations and/or changes in the size can occur, which need to quantified by comparison with the original.
Methods: We applied a variety of different computed tomography device ranging from synchrotron microCT to clinical CT scanners depending on the size of the specimens of interest. CT allows a rapid, easy and dimensionally stable assessment of the 3D structure of e.g. bone, teeth and other dense materials. Here we show results for the analysis of the Dehnel effect in shrews, the phenotyping of genetic modified mice and the comparison of 3D printed teeth facsimiles with CT scans of the original teeth. We combine methods of calculation of surface meshes, calculation of higher order shape descriptors, 3D registration, calculation of local distances between two meshes as well as different types of supervised and unsupervised classification schemes.
Results: Using this general pipeline we were not only able to show clear differences between genetic modified mice (knock-out of the Discoidin Domain Receptor 1 and 2) but also to discover mild effects caused by the gene-dose effect in heterozygous mice, which were previously unknown. In addition, the analysis pipeline validated the known Dehnel effect in shrews (the shrinkage of the brain and brain case volume by up to 40% in winter) by clearly demonstrating a change in the skull shape predominately in the brain case region (Figure 1). Moreover, the analysis pipeline was successfully applied to quantify the differences in the root channel geometry of 3D printed teeth in order to establish working with tooth facsimiles for the education of dentist.
Conclusion: We developed an analysis scheme based on CT data sets of hard-tissue specimens to analysis differences in complex 3D shapes which could be easily adapted to other use cases.