Active learning-assisted deep learning for automated maxillary sinus segmentation in dental cone-beam computed tomography

Authors

• Hsuan Chen, College of Dentistry, National Yang-Ming University, Taipei, Taiwan
• Khoa Dang Nguyen, College of Dentistry, National Yang-Ming University, Taipei, Taiwan
• Chih-Ling Chang, College of Dentistry, National Yang-Ming University, Taipei, Taiwan
• Hong Thi-Phuong Doan, College of Dentistry, National Yang-Ming University, Taipei, Taiwan
• Ya-Chu Chang, Department of Mechanical Engineering, Chinese Culture University, Taipei, Taiwan
• Thantrira Porntaveetus, Faculty of Dentistry, Chulalongkorn University, Bangkok, Thailand
• Guang Hong, School of Dentistry, Tohoku University, Sendai, Japan
• Wen-Liang Lo, College of Dentistry, National Yang-Ming University, Taipei, Taiwan
• Ding-Han Wang, College of Dentistry, National Yang-Ming University, Taipei, Taiwan
• Ming-Lun Hsu, College of Dentistry, National Yang-Ming University, Taipei, TaiwanFollow

Abstract

Background/purpose: Accurate automated segmentation of the maxillary sinus in cone-beam computed tomography (CBCT) scans is crucial for dental implant planning, sinus lift procedures, and pathological assessment, yet there remains a notable lack of publicly available models and datasets for this task. This study developed and evaluated a deep learning model for automated maxillary sinus segmentation using an active learning approach to minimize annotation effort while achieving high accuracy.

Materials and methods: We utilized 249 CBCT scans from the publicly available CTooth+ and STS-Tooth datasets. Ground truth annotations were generated through semi-automated segmentation in 3D Slicer with validation by five expert dentists. An active learning strategy was employed using the nnU-Net 3D full-resolution architecture trained with five-fold cross-validation. Performance was evaluated using Dice similarity coefficient (Dice).

Results: The model achieved a mean Dice of 0.985 (standard deviation (SD): 0.026, median: 0.992) and mean intersection over union (IoU) of 0.972 (SD: 0.044) across 249 cases. Performance remained stable across varying sinus volumes, with only rare outliers (minimum Dice: 0.712) corresponding to unusual anatomical presentations or image artifacts.

Conclusion: This study demonstrates that automated maxillary sinus segmentation using nnU-Net with active learning can achieve highly accurate and consistent results on diverse CBCT scans. The approach shows strong potential for clinical integration to enhance efficiency in dental treatment planning and supports the advancement of dental radiology driven by artificial intelligence (AI) through utilization of publicly available datasets.

Recommended Citation

Chen, Hsuan; Nguyen, Khoa Dang; Chang, Chih-Ling; Doan, Hong Thi-Phuong; Chang, Ya-Chu; Porntaveetus, Thantrira; Hong, Guang; Lo, Wen-Liang; Wang, Ding-Han; and Hsu, Ming-Lun (2026) "Active learning-assisted deep learning for automated maxillary sinus segmentation in dental cone-beam computed tomography," Journal of Dental Sciences: Vol. 21: Iss. 4, Article 32.
Available at: https://jds.ads.org.tw/journal/vol21/iss4/32

Publication Date

2026

Received Date

Apr 1 2026

Accepted Date

May 7 2026

Final Revision Date

May 7 2026