Morphotype stratification of radix entomolaris in mandibular molars

Area of Science:

• Endodontics
• Dental Anatomy
• Radiology
• Machine Learning in Dentistry

Background:

• Radix entomolaris (RE), a supernumerary root in mandibular molars, poses significant endodontic treatment challenges due to complex anatomy.
• Conventional radiography often fails to detect RE, increasing the risk of missed canals and treatment failure.

Purpose of the Study:

• To characterize the morphometric complexity of RE using cone-beam computed tomography (CBCT).
• To develop machine learning (ML) models for classifying root canal morphotypes and predicting anatomical bifurcation in RE.

Main Methods:

• One hundred mandibular first molars with RE underwent high-resolution CBCT scanning.
• Morphometric parameters (canal curvature, area, roundness, volume, root fusion) were extracted and used to train ML models.
• A decision tree classifier predicted bifurcation, and K-means clustering stratified morphotypes.

Main Results:

• The ML model achieved high accuracy (F1-score 0.87) in predicting bifurcation.
• Volumetric canal size was the strongest predictor of bifurcation (AUC=0.81).
• Two morphotypes were identified: simple (round) and complex (irregular/bifurcated/C-shaped); 27% showed mid-root bifurcation.

Conclusions:

• CBCT-derived features, especially canal volume and curvature, effectively predict RE complexity.
• ML models integrating these features offer a framework for personalized endodontic planning.
• AI-assisted diagnostics can improve management of complex root canal anatomy.