Deep convolutional neural network-based automated segmentation of the maxillofacial complex from cone-beam computed...

Area of Science:

• Medical Imaging Analysis
• Artificial Intelligence in Dentistry
• Computational Anatomy

Background:

• Maxillofacial bone segmentation from Cone Beam Computed Tomography (CBCT) is crucial for dental applications.
• Manual segmentation is time-consuming and prone to inter-observer variability.
• Deep learning offers potential for automating complex medical image segmentation tasks.

Purpose of the Study:

• To evaluate a novel deep convolutional neural network (CNN) model for automated maxillofacial bone segmentation.
• To assess the accuracy, consistency, and time-efficiency of the CNN model compared to manual segmentation.
• To determine the clinical utility of automated segmentation for digital workflows.

Main Methods:

• A dataset of 144 CBCT scans was utilized, divided into training, validation, and testing sets.
• A three-dimensional (3D) U-Net convolutional neural network (CNN) model was developed for automated segmentation.
• Automated segmentation results were compared against manual segmentation using Dice Similarity Coefficient (DSC) and time measurements.

Main Results:

• The CNN model achieved an average automated segmentation time of 39.1 seconds, a 204-fold reduction from manual segmentation (132.7 minutes).
• High accuracy was demonstrated with a Dice Similarity Coefficient (DSC) of 92.6% for bony structure identification.
• The model exhibited 100% consistency, and expert-corrected automated segmentation showed an excellent inter-observer DSC of 99.7%.

Conclusions:

• The proposed CNN model provides a highly accurate, consistent, and time-efficient solution for CBCT-based maxillofacial complex segmentation.
• Automated segmentation serves as a viable alternative to conventional methods, enhancing digital workflow efficiency.
• This approach facilitates the creation of accurate 3D models for patient-specific treatment planning in orthodontics, surgery, and implant dentistry.