Artificial Intelligence Models for Diagnosing Pulpitis in Adults Using a Modified Wolters Classification: A...

Area of Science:

• Dentistry
• Artificial Intelligence
• Medical Diagnostics

Background:

• Accurate diagnosis of pulp conditions is crucial for effective dental treatment.
• Traditional diagnostic methods can be subjective, leading to variability.
• Artificial intelligence offers potential for more objective and reproducible diagnostic tools.

Purpose of the Study:

• To evaluate an artificial intelligence (AI) system's performance in diagnosing pulp conditions.
• To utilize a modified Wolters diagnostic classification for enhanced objectivity.
• To compare the performance of various AI models in pulp diagnosis.

Main Methods:

• A cross-sectional study involving 200 teeth from 200 adult patients.
• Clinical evaluation and compilation of a dataset with 21 diagnostic attributes.
• Training and validation of AI models (Decision Tree, SVM, MLP, XGBoost) using K-fold cross-validation and bootstrap resampling.
• Performance assessment using precision, recall, F1-score, and AUC, with attribute importance analysis.

Main Results:

• XGBoost and Support Vector Machine (SVM) models achieved the highest performance with a mean F1-score of 0.85.
• Both XGBoost and SVM demonstrated Area Under the Curve (AUC) values exceeding 0.93 for all diagnostic classes.
• XGBoost exhibited the lowest variability and best average precision (0.81), recall (0.81), and F1-score (0.80).
• Pain in response to cold stimulus was identified as the most relevant diagnostic attribute.

Conclusions:

• The AI-based system demonstrates significant potential for accurate pulp diagnoses.
• AI contributes to more objective diagnostic decisions in dentistry.
• Pain intensity from thermal stimuli is a critical feature for AI-driven pulp condition classification.