A deep learning model based on concatenation approach to predict the time to extract a mandibular third molar tooth

Area of Science:

• Oral and Maxillofacial Surgery
• Artificial Intelligence in Medicine
• Predictive Modeling

Background:

• Accurate estimation of surgical time is crucial for dental procedures.
• Mandibular third molar extraction is a common surgical intervention.
• Predicting extraction duration aids in pre-operative planning and resource allocation.

Purpose of the Study:

• To develop a practical deep learning model for predicting mandibular third molar extraction time.
• To assess the accuracy of the predictive model against actual surgical durations.
• To enhance pre-surgical planning for dental extractions.

Main Methods:

• Utilized 724 panoramic X-ray images and clinical data for AI model training.
• Incorporated patient demographics, clinical measurements, and surgeon experience.
• Employed data augmentation and weight balancing to optimize AI learning.
• Compared AI-predicted extraction times with actual recorded times.

Main Results:

• The combined deep learning model (CNN + MLP) demonstrated strong predictive performance.
• Achieved an R-squared value of 0.6839 and an R value of 0.8315.
• Reported a mean absolute error (MAE) of 2.95 minutes on the test dataset.
• Statistical significance was confirmed with a p-value < 0.0001.

Conclusions:

• The developed deep learning model accurately predicts mandibular third molar extraction time.
• The model shows high potential for practical application in clinical settings.
• This AI-driven approach can improve efficiency and predictability in oral surgery.