Early Predicting Tribocorrosion Rate of Dental Implant Titanium Materials Using Random Forest Machine Learning Models

Area of Science:

• Biomaterials Science
• Mechanical Engineering
• Computational Science

Background:

• Bio-tribocorrosion is a significant concern for dental implants, potentially leading to mechanical degradation and the need for revision surgery.
• Early detection and prediction of this degradation are crucial for improving implant longevity and patient outcomes.

Purpose of the Study:

• To develop and validate a machine learning (ML) model for classifying and predicting mechanical degradation in dental implant materials due to bio-tribocorrosion.
• To assess the efficacy of ML in identifying potential implant failure through analysis of key tribocorrosion indicators.

Main Methods:

• Utilized machine learning (ML) models to analyze data from pure titanium and titanium-zirconium alloys (5, 10, 15 wt% Zr).
• Key features included corrosion potential, acoustic emission (AE) absolute energy, hardness, and weight-loss estimates.
• Evaluated the predictive accuracy of the developed ML models.

Main Results:

• The ML prototype models demonstrated high suitability for tribocorrosion prediction, achieving an accuracy exceeding 90%.
• The models effectively classified and predicted the likelihood of mechanical degradation in the tested dental implant alloys.

Conclusions:

• The proposed ML system offers a reliable predictive modeling technique for monitoring dental implant health.
• This approach can be further developed into a continuous structural-health monitoring system for dental implants, enhancing safety and reducing revision surgery rates.