Deep learning for automated alveolar cleft segmentation and bone graft volume estimation in cone-beam computed tomography imaging : a multicenter study
(2026) In Oral Surgery, Oral Medicine, Oral Pathology and Oral Radiology 141(3). p.400-408- Abstract
Objective To train and validate a deep learning-based diagnostic tool capable of accurately segmenting the alveolar cleft region and automatically estimating the required bone graft volume using cone-beam computed tomography (CBCT) imaging. Study Design Eighty-eight CBCT scans from patients with nonsyndromic unilateral clefts were divided into training (n = 45), validation (n = 10), and test (n = 33) sets. Two annotators performed manual segmentations, and the intersection of these served as the ground truth for training three-dimensional (3D) U-Net models. The dice similarity coefficient (DSC) was calculated to validate the tool by comparing manual and automated segmentations. Three observers evaluated the resulting deep learning model... (More)
Objective To train and validate a deep learning-based diagnostic tool capable of accurately segmenting the alveolar cleft region and automatically estimating the required bone graft volume using cone-beam computed tomography (CBCT) imaging. Study Design Eighty-eight CBCT scans from patients with nonsyndromic unilateral clefts were divided into training (n = 45), validation (n = 10), and test (n = 33) sets. Two annotators performed manual segmentations, and the intersection of these served as the ground truth for training three-dimensional (3D) U-Net models. The dice similarity coefficient (DSC) was calculated to validate the tool by comparing manual and automated segmentations. Three observers evaluated the resulting deep learning model using 33 CBCT scans and performing subjective assessments in terms of shape and size. Results The dice similarity coefficient (DSC) between the two annotators was 0.66, and between the automated and manual segmentations, 0.78. The observers considered the automated segmentations acceptable in 82%-94% of the cases. The deep learning-based tool took approximately second seconds to perform an automated segmentation, while manual segmentation by the annotators required 14 and 6.5 minutes. Conclusion The deep learning-based tool that was developed in the present study can accurately perform automated segmentations of unilateral alveolar clefts and estimate the required bone graft volume.
(Less)
- author
- organization
- publishing date
- 2026-03
- type
- Contribution to journal
- publication status
- published
- subject
- in
- Oral Surgery, Oral Medicine, Oral Pathology and Oral Radiology
- volume
- 141
- issue
- 3
- pages
- 9 pages
- publisher
- Elsevier
- external identifiers
-
- scopus:105025147915
- pmid:41390267
- ISSN
- 2212-4403
- DOI
- 10.1016/j.oooo.2025.10.020
- language
- English
- LU publication?
- yes
- additional info
- Publisher Copyright: © 2025 The Authors.
- id
- 4c8aa870-c8b0-4c5a-985a-88d8d9189d33
- date added to LUP
- 2026-02-26 16:22:36
- date last changed
- 2026-02-26 16:22:58
@article{4c8aa870-c8b0-4c5a-985a-88d8d9189d33,
abstract = {{<p>Objective To train and validate a deep learning-based diagnostic tool capable of accurately segmenting the alveolar cleft region and automatically estimating the required bone graft volume using cone-beam computed tomography (CBCT) imaging. Study Design Eighty-eight CBCT scans from patients with nonsyndromic unilateral clefts were divided into training (n = 45), validation (n = 10), and test (n = 33) sets. Two annotators performed manual segmentations, and the intersection of these served as the ground truth for training three-dimensional (3D) U-Net models. The dice similarity coefficient (DSC) was calculated to validate the tool by comparing manual and automated segmentations. Three observers evaluated the resulting deep learning model using 33 CBCT scans and performing subjective assessments in terms of shape and size. Results The dice similarity coefficient (DSC) between the two annotators was 0.66, and between the automated and manual segmentations, 0.78. The observers considered the automated segmentations acceptable in 82%-94% of the cases. The deep learning-based tool took approximately second seconds to perform an automated segmentation, while manual segmentation by the annotators required 14 and 6.5 minutes. Conclusion The deep learning-based tool that was developed in the present study can accurately perform automated segmentations of unilateral alveolar clefts and estimate the required bone graft volume.</p>}},
author = {{Vicente, António and Hung, Kuo Feng and Xu, Zineng and Yang, Jiegang and Li, Jian and Wiedel, Anna Paulina and Becker, Magnus and Brogårdh-Roth, Susanne and Ding, Peng and Hellén-Halme, Kristina and Shi, Xie Qi}},
issn = {{2212-4403}},
language = {{eng}},
number = {{3}},
pages = {{400--408}},
publisher = {{Elsevier}},
series = {{Oral Surgery, Oral Medicine, Oral Pathology and Oral Radiology}},
title = {{Deep learning for automated alveolar cleft segmentation and bone graft volume estimation in cone-beam computed tomography imaging : a multicenter study}},
url = {{http://dx.doi.org/10.1016/j.oooo.2025.10.020}},
doi = {{10.1016/j.oooo.2025.10.020}},
volume = {{141}},
year = {{2026}},
}
