3D maxilla auto segmentation with CBCT using AI

3D morphometric quantification of maxillae and defects for patients with unilateral cleft palate via deep learning-based CBCT image auto-segmentation

Authors: Xiaoyu Wang, Matthew Pastewait, Tai-Hsien Wu, Chunfeng Lian, Beatriz Tejera, Yan-Ting Lee, Feng-Chang Lin, Li Wang, Dinggang Shen, Song Li, Ching-Chang Ko

---

Summary
This study analyzed 3D cone-beam CT images from 60 patients with unilateral cleft lip and palate (UCP) to measure asymmetry of the maxilla. Using a deep learning-based segmentation protocol with manual refinement, the researchers compared the cleft and non-cleft sides. The results showed that the cleft side had significant reductions in maxillary and alveolar volume, length, and height, while the anterior width was increased. Statistical analysis confirmed that defect dimensions were closely related to the variability of the cleft-side maxilla.
In conclusion, the study identified hypoplasia in the pyriform aperture and alveolar crest areas, highlighting that defect structures contribute to maxillary asymmetry in UCP patients.

DOI: https://doi.org/10.1111/ocr.12482

---

Key Words

AI in Dentistry, CBCT, Cleft Lip and Palate, Deep Learning, 3D U-Net, Maxillary Asymmetry, Semantic Segmentation

---

Extracted Data

• Year: 2021
• Modality: CBCT
• Dataset: 60 patients
• Dataset Split: Group 1 = 30 (24 training, 3 validation, 3 test); Group 2 = 30 predicted + refined
• Network Architecture: 3D U-Net
• Metrics: Dice similarity coefficients – maxilla = 0.92 ± 0.01; defect = 0.77 ± 0.06
• AP – Strategy: None reported
• AP – Professional Qty: 2
• AP – Supervisor Presence: No information
• AP – Experience Level: No information
• AP – Expertise Area: Orthodontist
• AP – Tool or System: ITK-SNAP v3.6.0 (www.itksnap.org)
• Task: Semantic Segmentation
• Project Objective: To quantify the 3D asymmetry of the maxilla and investigate defect factors via deep learning-based CBCT auto-segmentation

---

Clinical Relevance

• Clinical relevance: It highlights the specific regions of the maxilla most affected by unilateral cleft lip and palate (pyriform aperture and alveolar crest), which are critical for surgical and orthodontic planning.
• Methodological innovation: By combining deep learning-based segmentation with manual refinement, the authors demonstrated a workflow that balances automation and clinical accuracy — a realistic model for clinical adoption.
• Quantitative insight: The correlation between defect size and maxillary asymmetry provides measurable parameters that can guide treatment decisions and future research.

For dentistry and AI, this work illustrates how artificial intelligence can support craniofacial analysis, improving our ability to objectively evaluate anomalies and their impact, and potentially guide personalized treatment strategies.

Related Video

▶️ Watch on YouTube