Neural Orthodontic Staging: Predicting Teeth Movements With a Transformer.

Area of Science:

• Biomedical Engineering
• Artificial Intelligence in Healthcare
• Orthodontics

Background:

• Orthodontic treatment planning, or staging, is complex due to multiple possible solutions.
• Accurate prediction of tooth movement is crucial for effective orthodontic staging.

Purpose of the Study:

• To develop a novel learning-based method for predicting tooth movements in orthodontic treatment path planning.
• To progressively generate orthodontic staging sequences using a Transformer model.

Main Methods:

• A Transformer model predicts teeth movements iteratively over a set number of steps.
• Spatial and temporal attentions with relative positional encoding capture inter-tooth and inter-step dynamics.
• A tooth-wise shape encoder integrates 3D tooth morphology features into the model.

Main Results:

• The proposed method outperforms state-of-the-art approaches on a large dataset of 10K orthodontic cases.
• The AI-generated staging predictions are favored by orthodontists.
• The model effectively refines predictions iteratively to achieve target alignment.

Conclusions:

• The novel learning-based method offers a significant advancement in AI-driven orthodontic staging.
• Integrating tooth shape features enhances the prediction of inter-tooth dynamics.
• This approach provides a promising tool for optimizing orthodontic treatment planning.