Convolutional neural network-based automated maxillary alveolar bone segmentation on cone-beam computed tomography...

Area of Science:

• Medical Imaging
• Artificial Intelligence in Dentistry
• Oral and Maxillofacial Surgery

Background:

• Accurate segmentation of maxillary alveolar bone is crucial for dental diagnostics and treatment planning.
• Manual segmentation of 3D cone-beam computed tomography (CBCT) images is time-consuming and prone to variability.
• Developing automated tools can improve efficiency and consistency in analyzing dental CBCT data.

Purpose of the Study:

• To develop and evaluate a novel artificial intelligence (AI)-driven convolutional neural network (CNN) tool for automated 3D segmentation of maxillary alveolar bone.
• To assess the performance and speed of the AI tool compared to manual segmentation methods.
• To analyze the accuracy of AI-generated segmentation for both the alveolar bone and its crestal contour.

Main Methods:

• A CNN model was trained and validated using 141 CBCT scans.
• Automated segmentation was performed, followed by expert refinement for a refined-AI (R-AI) dataset.
• AI segmentation accuracy was compared against manual segmentation on a subset of test data, with time efficiency also recorded.

Main Results:

• The AI tool achieved excellent accuracy metrics for maxillary alveolar bone segmentation.
• Manual segmentation showed slightly superior accuracy (e.g., higher Dice Similarity Coefficient) compared to AI.
• AI-driven segmentation was significantly faster, taking 51.5 seconds compared to 5973.3 seconds for manual segmentation (116x speed increase).

Conclusions:

• The novel AI-driven CNN tool provides highly accurate segmentation of maxillary alveolar bone and its crestal contour.
• Despite slightly lower accuracy than manual methods, the AI tool offers a substantial time-saving benefit.
• This automated approach represents a significant advancement in efficient and accurate analysis of dental CBCT images.