Predicting sequenced dental treatment plans from electronic dental records using deep learning

Area of Science:

• Artificial Intelligence in Dentistry
• Machine Learning for Healthcare
• Deep Learning in Clinical Decision Support

Background:

• Increasing prevalence of partial edentulism in aging populations necessitates improved clinical dental treatment planning.
• The complexity of dental treatments requires advanced tools for accurate plan generation.

Purpose of the Study:

• To develop a model that predicts sequential dental treatment plans using electronic dental records.
• To provide a decision support system for dentists managing partial edentulism.

Main Methods:

• Construction of the MultiTP model, integrating convolutional neural networks and recurrent neural networks.
• Utilizing an attention mechanism to optimize treatment plan prediction.
• Exploring the unique topology of dental data and treatment variations.

Main Results:

• MultiTP achieved high performance with an AUC of 0.9079 and an F-score of 0.8472.
• The model demonstrated interpretability, capable of extracting clinical knowledge from text data.
• Successful prediction across five distinct treatment plans.

Conclusions:

• MultiTP offers a novel framework for applying deep learning in dental healthcare.
• The model can be integrated into existing electronic dental record systems for clinical practice.
• This technology supports dentists in making informed decisions for partial edentulism treatment, ultimately benefiting patients.