Developing an Artificial Intelligence Solution to Autosegment the Edentulous Mandibular Bone for Implant Planning

Area of Science:

• Artificial Intelligence in Dentistry
• Medical Image Analysis
• Oral and Maxillofacial Surgery

Background:

• Dental implants are crucial for restoring function and aesthetics after tooth loss.
• Accurate surgical planning is vital to avoid damaging anatomical structures during implant placement.
• Manual segmentation of edentulous bone on CBCT images is time-consuming and prone to human error.

Purpose of the Study:

• To develop an artificial intelligence (AI) solution for automated identification and delineation of edentulous alveolar bone on CBCT images.
• To reduce errors and improve efficiency in pre-implant surgical planning.

Main Methods:

• A supervised machine learning approach using a U-Net convolutional neural network (CNN) was employed.
• The model was trained and tested on CBCT images from the University Dental Hospital Sharjah.
• Dice Similarity Coefficient (DSC) was used to measure segmentation accuracy against manual delineations.

Main Results:

• The AI model achieved an average DSC of 0.89 for training and 0.78 for testing.
• Higher accuracy (DSC 0.91) was observed for unilateral edentulous areas compared to bilateral cases (DSC 0.73).
• The model demonstrated the ability to identify missing bone structures, a novel approach in AI object detection.

Conclusions:

• Machine learning-based segmentation of edentulous spans on CBCT images is accurate and efficient.
• This AI solution offers a significant improvement over traditional manual segmentation methods for dental implant planning.
• Future work will focus on addressing data challenges and developing a comprehensive AI solution for automated implant planning.