Related Experiment Video
Updated: May 29, 2026
Establishment of a Murine Pulp Exposure Model with a Novel Mouth-Gag for Pulpitis Research
Published on: October 27, 2023
Deep learning-driven intraoperative assessment of pulp stumps for precision pulpotomy
Qianli Zhang1, Meiyu Hu2, Junran Peng3
1Department of the Fourth Clinical Division, Peking University School and Hospital of Stomatology & National Center of Stomatology & National Clinical Research Center for Oral Disease & National Engineering Laboratory for Digital and Material Technology of Stomatology, Beijing 100081, China.
An artificial intelligence (AI) framework using DINOv2 accurately assesses pulp stumps for dental capping. This AI tool matches expert judgment, improving endodontic treatment precision and preventing overdiagnosis.
Area of Science:
- Dentistry
- Artificial Intelligence
- Computer Vision
Background:
- Intraoperative assessment of pulp stumps is crucial for successful pulpotomy.
- Current methods rely on subjective visual interpretation, leading to variability.
- Bridging the clinical experience gap is essential for consistent outcomes.
Purpose of the Study:
- To develop and validate an AI framework for objective classification of pulp stump suitability for capping.
- To utilize the self-supervised DINOv2 vision transformer architecture for this task.
- To compare the AI model's performance against human experts.
Main Methods:
- Trained a DINOv2-based AI model on 443 microscopic pulp stump images.
- Evaluated performance on an internal test set (n=93) and an external literature set (n=21).
- Conducted comparative analysis against expert endodontists and novice practitioners.
Main Results:
- The DINOv2 model achieved high cross-validation accuracy (0.9398) and outperformed traditional methods.
- Achieved 0.9570 accuracy on the internal test set and 0.8571 on the external set.
- AI model's accuracy was comparable to experts and superior to novices, with balanced specificity.
Conclusions:
- The DINOv2 framework serves as a proof-of-concept for expert-level pulp stump assessment.
- This AI decision-support system enhances precision and predictability in endodontic treatments.
- The tool helps bridge the clinical experience gap and mitigate overdiagnosis.
