EUVIMEDEUVIMED entdecken
Crossref

Automatic multi-class classification of 3D relationships between the mandibular third molar and canal on cone-beam computed tomography using a geometry-aware network

Hyomin Kim, Ji Yong Han, Su Yang, Jo-Eun Kim iD, Kyung-Hoe Huh iD, Sam-Sun Lee iD, Min-Suk Heo iD, Jeong Joon Han, Joo-Young Park, Won-Jin Yi iD

Dentomaxillofacial Radiology · 01.07.2026 · 55 (5) · 460-476

DOI: 10.1093/dmfr/twag010 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Dentomaxillofacial Radiology, online erschienen am 24.03.2026. Verfasst von Hyomin Kim und 9 weiteren Autor:innen. Beteiligte Einrichtung: Department of Oral and Maxillofacial Radiology and Dental Research Institute, School of Dentistry, Seoul National University, Seoul 03080, Republic o…. Hauptthemen laut MEDLINE: Molar, Third, Cone-Beam Computed Tomography, Imaging, Three-Dimensional, Mandibular Canal. Die Arbeit nennt 74 Literaturangaben. Gefördert von: National Research Foundation of Korea, Technology Innovation Program, Ministry of Trade Industry & Energy.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Objectives

Accurate 3D evaluation of the spatial relationship between the mandibular third molar (M3) and the mandibular canal (MC) is critical for assessing the risk of nerve injury during M3 extraction. The purpose of this study was to automatically classify the diverse spatial relationships between M3 and MC into 5 distinct categories based on the degree of anatomical contact or involvement as well as spatial proximity on cone-beam computed tomography (CBCT) images using a geometry-aware deep learning framework.

Methods

The 3D spatial relationships between M3 and MC were categorized into 5 distinct types. The proposed framework consisted of 2 stages: first, the modified mAttUNet, an improved 3D U-Net architecture augmented with an attention mechanism was used for segmentation of M3 and MC; second, the proposed DenseAttNet was developed for multi-class classification. By incorporating dense attention mechanisms and signed distance map (SDM) inputs, the proposed network effectively captured both geometric and anatomical features, leading to more accurate and reliable multi-class classification.

Results

The mAttUNet outperformed other conventional models, achieving the highest segmentation performance with average precision scores of 0.96 for MC and 0.84 for M3. The proposed DenseAttNet demonstrated superior and consistent performance, achieving an overall AUC of 0.97 and maintaining high accuracy across all 5 relationship types, effectively and reliably distinguishing the various spatial relationships between MC and M3.

Conclusions

This automated and accurate classification of M3-MC spatial relationships offers valuable clinical utility, supporting enhanced risk evaluation and optimized surgical planning, and ultimately helping to reduce complications such as inferior alveolar nerve injury.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Molar, ThirdCone-Beam Computed TomographyImaging, Three-DimensionalMandibular CanalDeep LearningHumans

Stichwörter der Autor:innen

3D relationshipCBCT imagegeometry-aware networkmandibular canalmandibular third molarmulti-class classification

Autor:innen und Institutionen

  • Department of Oral and Maxillofacial Radiology and Dental Research Institute, School of Dentistry, Seoul National University, Seoul 03080, Republic of Korea.
  • Interdisciplinary Program in Bioengineering, Graduate School of Engineering, Seoul National University, Seoul 08826, Republic of Korea.
  • Department of Medical Informatics, Kyungpook National University, Daegu 41566,Republic of Korea.
  • Department of Oral and Maxillofacial Surgery and Dental Research Institute, School of Dentistry, Seoul National University, Seoul 03080, Republic of Korea.

ORCID-Profile: Jo-Eun Kim, Kyung-Hoe Huh, Sam-Sun Lee, Min-Suk Heo, Won-Jin Yi

Förderung

  • National Research Foundation of Korea
  • Technology Innovation Program
  • Ministry of Trade Industry & Energy

Zitieren

APA

Kim, H., Han, J. Y., Yang, S., Kim, J. E., Huh, K. H., Lee, S. S., Heo, M. S., Han, J. J., Park, J. Y., & Yi, W. J. (2026). Automatic multi-class classification of 3D relationships between the mandibular third molar and canal on cone-beam computed tomography using a geometry-aware network. Dentomaxillofacial Radiology, 55(5), 460-476. https://doi.org/10.1093/dmfr/twag010

Vancouver

Kim H, Han JY, Yang S, Kim JE, Huh KH, Lee SS, et al. Automatic multi-class classification of 3D relationships between the mandibular third molar and canal on cone-beam computed tomography using a geometry-aware network. Dentomaxillofacial Radiology. 2026;55(5):460-476. doi:10.1093/dmfr/twag010

BibTeX
@article{Kim2026,
  author = {Kim, Hyomin and Han, Ji Yong and Yang, Su and Kim, Jo-Eun and Huh, Kyung-Hoe and Lee, Sam-Sun and Heo, Min-Suk and Han, Jeong Joon and Park, Joo-Young and Yi, Won-Jin},
  title = {Automatic multi-class classification of 3D relationships between the mandibular third molar and canal on cone-beam computed tomography using a geometry-aware network},
  journal = {Dentomaxillofacial Radiology},
  year = {2026},
  volume = {55},
  number = {5},
  pages = {460--476},
  doi = {10.1093/dmfr/twag010}
}
Kostenlos starten →