Kitware is proud to present the Dental Segmentator extension, based on the AI segmentation model of Dot G, Chaurasia A, Dubois G, et al., and supported by  FFO (Fédération Française d’Orthodontie) and the Fondation des Gueules Cassées.

The Dental Segmentator extension allows the fully automatic segmentation of 5 anatomic structures on DMF CT and CBCT scans: maxilla and upper skull, mandible, upper teeth, lower teeth, and mandibular canal.

The extension allows running the segmentation on a selected CT or CBCT volume, editing, and exporting it directly from the Dental Segmentator module.

DentalSegmentator model

Dental Segmentator is based on the nnU-Net framework. It has been trained on 470 dento-maxillo-facial CT and CBCT scans, and evaluated on a hold-out test dataset of 256 CT and CBCT scans from 7 institutions.

The results obtained on this dataset demonstrate the ability to provide fully automatic and robust multi-class segmentation for dento-maxillo-facial (CB)CT scans, independently of the field of view of the scan.

The performance and generalization of the tool were evaluated by comparing segmentations provided by experts and automatic segmentations on 2 hold-out test datasets: an internal dataset of 133 CT and CBCT scans acquired before orthognathic surgery and an external dataset of 123 CBCT scans randomly sampled from routine examinations in 5 institutions.

The model obtained a mean Dice score of 92.2 on the internal test dataset and a normalized surface distance of 98.2. The mean results on the external test dataset were a Dice score of 94.2 and a normalized surface distance of 98.4.

Using the extension

The extension can be downloaded, installed, and used directly from 3D Slicer from the Slicer 5.7 preview release onward.

The following video shows how to install and use the extension :

Learn more

To learn more about the model, please look at the Dental Segmentator paper here: Dot G, Chaurasia A, Dubois G, et al. DentalSegmentator: robust deep learning-based CBCT image segmentation. Published online March 18, 2024:2024.03.18.24304458. doi:10.1101/2024.03.18.24304458

The extension is publicly available on GitHub.

Credits

Authors: G. Dot (Université Paris Cité, AP-HP, Arts-et-Métiers), L. Gajny (Arts-et-Métiers), R. Fenioux (Kitware SAS), T. Pelletier (Kitware SAS)

Supported by the FFO (Fédération Française d’Orthodontie) and the Fondation des Gueules Cassées.

Please contact the authors for more information about the model and Kitware for your medical AI needs, with or without 3D Slicer. We provide numerous solutions adapted to each.