Accuracy and Reproducibility of a Manual 3-Dimensional Superimposition Method to Register Dental Model Images to Cone-Beam Computer Tomography (CBCT) Image-A Preliminary Study

Purpose: The purposes of this study was to evaluate the registration error and its reproducibility for implementation of the laser-scanned dental model images onto the CBCT scan data by using the manual 3D superimposition.Materials and methods: Twelve arches out of 8 consecutive patients were enrolled in this study. Both CBCT and dental impression were taken as orthodontic records and the dental casts were then laser-scanned into dental model images with STL file format. In SimPlant OMS software, we register the dental model images onto CBCT teeth images with manual 3D superimposition technique. Afterwards, validation was performed in 3dMDvultus and the accuracy was presented in terms of root mean square difference (RMSD). Wilcoxon Two-Sample Test and pair T test was used for post-hoc comparisons to evaluated the intra-operator and inter-operator difference.Results: The RMSD of shell/shell deviation was 0.31 mm in operator A and 0.23 mm in operator B. There was significant difference between these two. The RMSD of shell/shell deviation in operator A was both 0.31 mm in 1st and 2nd time operations. There was no significant difference between these two operations.Conclusions: Free-hand 3D superimposition is able to achieve good accuracy of registration. However, outlier still exists and would deteriorate clinical results. So validation of the registration is highly recommended after each registration with the use of this manual 3D superimposition technique.