Related Experiment Video
Updated: Jan 8, 2026
09:10
Digital Hybrid Model Preparation for Virtual Planning of Reconstructive Dentoalveolar Surgical Procedures
Published on: August 5, 2021
2.2K
Deep learning for automated alveolar cleft segmentation and bone graft volume estimation in cone-beam computed
António Vicente1, Kuo Feng Hung2, Zineng Xu3
1Department of Oral and Maxillofacial Radiology, Faculty of Odontology, Malmö University, Malmö, Sweden.
Oral Surgery, Oral Medicine, Oral Pathology and Oral Radiology
|December 13, 2025
Summary
A new deep learning tool accurately segments unilateral alveolar clefts from cone-beam computed tomography (CBCT) scans, automating bone graft volume estimation and significantly reducing processing time.
Area of Science:
- Medical Imaging
- Artificial Intelligence in Medicine
- Craniofacial Surgery
Background:
- Alveolar clefts require bone grafting for reconstruction.
- Accurate volume estimation is crucial for successful grafting.
- Manual segmentation of clefts from CBCT is time-consuming.
Purpose of the Study:
- To develop and validate a deep learning tool for automated alveolar cleft segmentation.
- To enable automatic estimation of bone graft volume using CBCT.
- To improve efficiency in planning alveolar cleft reconstruction.
Main Methods:
- Trained 3D U-Net models on 88 CBCT scans of unilateral clefts.
- Used manual segmentations as ground truth.
- Validated automated segmentation using Dice Similarity Coefficient (DSC) and observer assessments.
Main Results:
- Achieved a DSC of 0.78 between automated and manual segmentations.
- Automated segmentation was deemed acceptable in 82%-94% of cases by observers.
- Automated segmentation took seconds, compared to minutes for manual methods.
Conclusions:
- The developed deep learning tool accurately segments unilateral alveolar clefts.
- The tool effectively estimates required bone graft volume.
- This automated approach offers a significant time-saving advantage in clinical practice.
