Related Experiment Video
Updated: Jun 26, 2026
09:34
Semi-Automated Planimetric Quantification of Dental Plaque Using an Intraoral Fluorescence Camera
Published on: January 27, 2023
Deep Learning Models for Detection of Periapical Radiolucent Lesions on Panoramic Radiographs: A Systematic Review
Ibrahim Ali Ahmad1,2, Raidan Ba-Hattab3, Sadeq Ali Al-Maweri3
1QU Health, Qatar University, Doha, Qatar.
International Endodontic Journal
|June 25, 2026
Summary
Deep learning (DL) models show high accuracy in detecting periapical radiolucent lesions (PRLs) on dental panoramic radiographs. However, results vary due to study heterogeneity, necessitating further research for reliable generalization.
Area of Science:
- Dentistry
- Radiology
- Artificial Intelligence
Background:
- Panoramic radiographs are standard for dental screening.
- Deep learning (DL) models show promise for analyzing these radiographs.
Purpose of the Study:
- To systematically review and evaluate the diagnostic accuracy of DL models for detecting periapical radiolucent lesions (PRLs) on panoramic radiographs.
Main Methods:
- A comprehensive search of six databases and grey literature was performed.
- Included studies were assessed for bias and applicability using QUADAS-2.
- Meta-analysis synthesized data from six studies using split component synthesis.
Main Results:
- Thirty studies were reviewed; six were meta-analyzed.
- DL models achieved a pooled sensitivity of 0.80 and specificity of 0.98.
- The area under the receiver operating characteristic curve (AUROC) was 0.93, with moderate certainty of evidence.
Conclusions:
- DL models exhibit high accuracy for PRL detection on panoramic radiographs.
- Heterogeneity across studies necessitates cautious interpretation.
- Future research should focus on diverse datasets and standardized reporting for DL model validation.