Three-dimensional files featuring patients’ geometry can be obtained through common tools in dental practice, such as an intraoral scanner (IOS) or Cone Beam Computed Tomography (CBCT). The use of 3D files in medical education is promoted, but only few methodologies were reported due to the lack of ease to use and accessible protocols for educators. The aim of this work was to present innovative and accessible methodologies to create 3D files in dental education. The first step requires the definition of the educational outcomes and the situations of interest. The second step relies on the use of IOS and CBCT to digitize the content. The last “post-treatment” steps involve free software for analysis of quality, re-meshing and simplifying the file in accordance with the desired educational activity. Several examples of educational activities using 3D files are illustrated in dental education and discussed. Three-dimensional files open up many accessible applications for a dental educator, but further investigations are required to develop collaborative tools and prevent educational inequalities between establishments.

No abstract

Expert determination of the third molar maturity index (I3M) constitutes one of the most common approaches for dental age estimation. This work aimed to investigate the technical feasibility of creating a decision-making tool based on I3M to support expert decision-making. Methods: The dataset consisted of 456 images from France and Uganda. Two deep learning approaches (Mask R-CNN, U-Net) were compared on mandibular radiographs, leading to a two-part instance segmentation (apical and coronal). Then, two topological data analysis approaches were compared on the inferred mask: one with a deep learning component (TDA-DL), one without (TDA). Regarding mask inference, U-Net had a better accuracy (mean intersection over union metric (mIoU)), 91.2% compared to 83.8% for Mask R-CNN. The combination of U-Net with TDA or TDA-DL to compute the I3M score revealed satisfying results in comparison with a dental forensic expert. The mean ± SD absolute error was 0.04 ± 0.03 for TDA, and 0.06 ± 0.04 for TDA-DL. The Pearson correlation coefficient of the I3M scores between the expert and a U-Net model was 0.93 when combined with TDA and 0.89 with TDA-DL. This pilot study illustrates the potential feasibility to automate an I3M solution combining a deep learning and a topological approach, with 95% accuracy in comparison with an expert.