Artificial intelligence for reducing metal artifacts in dental CBCT images: a systematic review.
Authors
Affiliations (9)
Affiliations (9)
- Department of Oral and Maxillofacial Radiology, Dental Implants Research Center, Dental Research Institute, School of Dentistry, Isfahan University of Medical Sciences, Isfahan, Iran.
- Department of Neurosciences, Reproductive and Odontostomatological Sciences, University of Naples Federico II, Naples, Italy.
- Department of Endodontics, Dental Research Center, Dental Research Institute, School of Dentistry, Isfahan University of Medical Sciences, Isfahan, Iran.
- Student Research Committee, School of Dentistry, Isfahan University of Medical Sciences, Isfahan, Iran.
- Department of Engineering, University of Sannio, Benevento, Italy.
- Department of Electrical Engineering and Information Technology, University of Naples Federico II, Naples, Italy.
- Department of Oral and Maxillofacial Surgery, Dental Research Center, Mazandaran University of Medical Sciences, Sari, Iran.
- Global Research Cell, D.Y. Patil Dental College and Hospital, D.Y. Patil Vidyapeeth, Pimpri, Pune, India.
- Oral and Maxillofacial Radiology, Department of Odontology, Umeå University, Umeå, Sweden.
Abstract
The presence of metallic restorations introduces severe artifacts that compromise diagnostic accuracy of cone beam computed tomography (CBCT) images. This systematic review aims to evaluate the effectiveness of artificial intelligence (AI) techniques in reducing metal artifacts in dental CBCT images. A comprehensive literature search was conducted across six databases up until April 2025, using keywords related to AI and CBCT artifact reduction. Studies were selected based on the eligibility criteria, focusing on AI models trained on human CBCT scans targeting dental regions. Data extraction and risk of bias assessment were performed independently by several reviewers accordingly. Risk of bias was independently assessed using the QUADAS-2 tool. Ten studies published between 2019 and 2025 met the eligibility criteria. The majority of included studies had a high risk of bias. The most employed deep learning architectures wereU-Net, GANs, and transformer-based models. Quantitative metrics like SSIM, PSNR, and RMSE demonstrated consistent improvements in image quality, while qualitative assessments confirmed superior artifact suppression and anatomical detail preservation compared to conventional methods. Transformer-based and physics-informed dual-domain networks perform superior to traditional image-only U-Nets in both quantitative fidelity and visual artifact suppression. AI-based approaches show promising potential in enhancing CBCT image fidelity by mitigating metal-induced distortions effectively. Despite encouraging results, the broader clinical adoption of such techniques requires standardized evaluation protocols, training models by open-access datasets, and further trials to validate diagnostic impact and generalizability.