Related Experiment Video
Updated: May 2, 2026
10:42
A Postoperative Evaluation Guideline for Computer-Assisted Reconstruction of the Mandible
Published on: January 28, 2020
6.6K
Mandiblemath: External validation of a predictive tool for mandibular reconstruction
Tania Camila Niño-Sandoval1, Belmiro Ce Vasconcelos2
1Department of Oral and Maxillofacial Surgery, Universidade de Pernambuco - School of Dentistry (UPE/FOP), Recife, Brazil.
Summary
Mandiblemath, an AI tool, predicts 3D mandibular shape from 2D data for reconstructive surgery. It shows clinical potential for accurate, low-cost, and reproducible surgical planning.
Area of Science:
- Biomedical Engineering
- Artificial Intelligence in Medicine
- Surgical Planning
Background:
- Mandibular reconstruction presents challenges in accurately predicting 3D morphology.
- Existing methods may lack efficiency and reproducibility in surgical planning.
Purpose of the Study:
- To develop and validate Mandiblemath, an AI-driven software for predictive mandibular reconstruction.
- To infer 3D mandibular morphology from 2D craniofacial data for surgical planning.
Main Methods:
- Developed Mandiblemath using Python, integrating scientific computing, 3D visualization, and supervised learning.
- Trained sex-specific classifiers (random forest, SVM, gradient boosting) on 91 CT scans.
- Validated predictions using 18 independent CT scans and surface-based metrics.
Main Results:
- Achieved high accuracy in external validation (88.8% for group identification, 66.6-70.4% for ranking).
- Predicted models showed clinically acceptable discrepancies (RMSD 1.2-3.3 mm) and strong overlap (F1@2.5 mm ≥ 0.85).
- Demonstrated technical feasibility and clinical potential as a decision-support tool.
Conclusions:
- Mandiblemath offers a low-cost, reproducible method for predicting mandibular morphology.
- The software has potential for integration into digital surgical workflows for reconstructive planning.
- Further multicenter validation is supported by these findings.
