Related Experiment Video
Updated: Sep 13, 2025
05:49
Author Spotlight: Advancing CBCT and Digital Dental Image Integration with AI-Assisted Digitization
Published on: February 23, 2024
995
Deep learning for tooth detection and segmentation in panoramic radiographs: a systematic review and meta-analysis
M Bonfanti-Gris1, A Herrera1, M P Salido Rodríguez-Manzaneque2
1Department of Conservative and Prosthetic Dentistry, Faculty of Dentistry, Universidad Complutense de Madrid, Plaza Ramón y Cajal S/N, Madrid, 28040, Spain.
BMC Oral Health
|July 30, 2025
Summary
Deep learning methods show high accuracy for detecting and segmenting teeth in panoramic radiographs. A moderate recommendation suggests AI tools can aid dental professionals in these tasks.
Area of Science:
- Dentistry
- Radiology
- Artificial Intelligence
Background:
- Deep learning (DL) shows promise in dental imaging analysis.
- Orthopantomographies are common dental X-rays requiring accurate tooth assessment.
- Evaluating DL performance for tooth detection and segmentation is crucial.
Purpose of the Study:
- To systematically review and meta-analyze the performance of DL methods for tooth detection and segmentation in orthopantomographies.
- To summarize existing evidence on the diagnostic accuracy of DL algorithms in this context.
Main Methods:
- A systematic search of Medline, Embase, and Cochrane databases was conducted up to September 2023.
- Twenty studies were included after screening 2,207 records, with data extracted and quality assessed by two independent reviewers.
- Meta-analysis, including Hierarchical Summary Receiver Operating Characteristic (HSROC) curves, was performed on mesiodens detection and segmentation data.
Main Results:
- The meta-analysis demonstrated high pooled diagnostic performance for DL methods.
- Pooled sensitivity was 0.92 (95% CI, 0.84-0.96) and pooled specificity was 0.94 (95% CI, 0.89-0.97).
- HSROC curves indicated a strong positive correlation between sensitivity and specificity (r=0.886).
Conclusions:
- Deep learning methods exhibit significant accuracy for tooth detection and segmentation in panoramic radiographs.
- A moderate recommendation is provided for dental operators to consider AI-based tools for these applications.
- Further validation and integration of AI tools in dental practice are supported by this evidence.
