Artificial Intellegence Rivals Digital Bitewing in Detect Secondary Caries

Official title

AI Rivals Traditional Bite Wing Radiography in Detecting Proximal Secondary Caries in A Group of Egyptian Patients at Cairo University, Faculty OF Dentistry Hospital (Diagnostic Accuracy Study)

Overview

This study uses digital bitewing radiography as a standard for diagnosing proximal secondary caries. Patients will undergo imaging with a parallel technique and fixed settings to ensure high-quality, consistent images. Radiographs are interpreted by experienced dental professionals to maintain diagnostic accuracy. Machine learning models YOLO and Mask-RCNN will analyze these images in three phases: pre-analytical, analytical, and post-analytical. A dataset of 322 labeled images, annotated by experts, is used to train these models. Data augmentation methods enhance model performance, and accuracy is assessed against radiographic results to confirm reliability.

Detailed description

Dental caries are chronic diseases that results in the destruction of the hard tooth tissues. It is a multifactorial condition that often goes undiagnosed, especially when it is hidden or in its initial stages. Detecting non-cavitated lesions is crucial for their early management. The standard visual-tactile inspection often fails to identify early lesions on hard-to-reach surfaces, such as proximal areas and beneath restorations. Detecting proximal caries early is crucial for implementing effective treatments and achieving optimal outcomes. A common supplementary method for detecting early lesions on proximal surfaces and assessing their extent is bitewing radiography. The routine diagnostic approach combines clinical examination with radiographic evaluation. To increase the detection rate of proximal secondary caries, experts recommend integrating visual and clinical examinations with bitewing radiography. Intraoral bitewing radiographs can be captured using either film or digital sensors, with preference for digital systems due to their benefits of reduced patient exposure, time savings, image enhancement, and ease of image storage, retrieval, and transmission. Although more sensitive for detecting early lesions than visual-tactile assessments, bitewing evaluations comes with significant variance between examiners and a considerable proportion of false-positive or false-negative detections. Recent literature has explored the use of artificial intelligence (AI), a field of computer science focused on developing machines capable of mimicking human cognitive abilities, as a diagnostic tool for detecting caries lesions using dental (digital radiographic) images. As AI technology advances, an increasing number of studies have examined the diagnostic performance of AI-based models, emphasizing the importance of creating reliable tools like AI to enhance the diagnostic process. Numerous studies have assessed the performance of AI models on diverse types of dental radiographs, with a significant focus on bitewing radiographs (BWR). AI has been used for various applications in oral and dental health, including the detection of dental caries, endodontic treatment and diagnosis, periodontal issues, and the detection of oral lesion pathology. A reference dataset of caries diagnoses from bitewing radiographs by different examiners created this benchmark which serves as a crucial tool for comparing the diagnostic performance of AI models against human examiners, emphasizing the potential improvements in accuracy and reliability that AI can bring to dental diagnostics.

Interventions

• Other artificial intelligence models (YOLO and Mask-RCNN)
  machine learning model will used to detect secondary caries around restorations by comparing the results with digital bitewing radiography

Primary outcome measures

• two deep learning models, YOLO and Mask-RCNN, will be trained on this dataset to accurately detect and classify images showing signs of secondary caries [Time frame: baseline]

Eligibility criteria

Inclusion criteria

• Adult Patients Aged 22-60 Patient
• Males or females.
• Patients have proximal restorations.
• Co-operative patients who show interest in participating in the study.

Exclusion criteria

• Patients with orthodontic appliances, or bridge work that might interfere with evaluation
• Patients with no caries.
• Systematic disease that may affect participation.
• Patients not willing to be part of the study or ones who refuse to sign the informed consent.

Criteria are shown verbatim from the registry (in English). Final eligibility is always assessed by the study center.

Healthy volunteers: Yes

Study design

Observational model

Other

Study locations

Center list to be confirmed — check the primary protocol.

Publications

• Chaves ET, Vinayahalingam S, van Nistelrooij N, Xi T, Romero VHD, Flugge T, Saker H, Kim A, Lima GDS, Loomans B, Huysmans MC, Mendes FM, Cenci MS. Detection of caries around restorations on bitewings using deep learning. J Dent. 2024 Apr;143:104886. doi: 10.1016/j.jdent.2024.104886. Epub 2024 Feb 9. PMID 38342368
• Mohammad-Rahimi H, Motamedian SR, Rohban MH, Krois J, Uribe SE, Mahmoudinia E, Rokhshad R, Nadimi M, Schwendicke F. Deep learning for caries detection: A systematic review. J Dent. 2022 Jul;122:104115. doi: 10.1016/j.jdent.2022.104115. Epub 2022 Mar 30. PMID 35367318
• Mertens S, Krois J, Cantu AG, Arsiwala LT, Schwendicke F. Artificial intelligence for caries detection: Randomized trial. J Dent. 2021 Dec;115:103849. doi: 10.1016/j.jdent.2021.103849. Epub 2021 Oct 14. PMID 34656656
• Chen X, Guo J, Ye J, Zhang M, Liang Y. Detection of Proximal Caries Lesions on Bitewing Radiographs Using Deep Learning Method. Caries Res. 2022;56(5-6):455-463. doi: 10.1159/000527418. Epub 2022 Oct 10. PMID 36215971

Identifiers

NCT: NCT06667986 · AI in detect dental caries

Primary sources (government registries)

• ClinicalTrials.gov (US NIH)

View this study on ClinicalTrials.gov ↗