Detection of Proximal Caries in Bitewing Radiography Using Artificial Intelligence

NCT07404007 | Completed DMC

• 1 other identifier: CONS_ Rad 1180
• Study Type: observational
• Target: 2,000
• Locations: 1 country
• Sites: 1

Brief Summary

Using a sequence of bitewing radiographs, Artificial intelligence assists in identifying interproximal caries. For the identification of dental caries in bitewing, periapical, and panoramic radiographs, a trained deep learning network will be created This study aimed to investigate the reliability of a novel Artificial Intelligence model based on deep learning in the detection of Proximal Caries using Digital Bitewing Radiographs. (BW).

Conditions

Proximal Caries; Tooth Caries

Keywords

Artificial intelligence; Proximal caries; A Diagnostic accuracy Study

Outcome Measures

Primary Outcomes (1)

• Reliability of the artificial intelligence model in detecting proximal caries on digital bitewing radiographs

  cross-sectional assessment at baseline, with no follow-up period

Study Arms (2)

Group 1 Artificial Intelligence Deep learning that is applied in Diagnosis of the proximal Caries

Diagnostic Test: Artificial Intelligence (AI): Deep learning that is applied in Diagnosis of the proximal Caries

Group 2 : Digital Bitewing manually annotated by human experts

Diagnostic Test: Manual annotation of Digital Bitewing Radiograph by human experts

Interventions

Manual annotation of Digital Bitewing Radiograph by human experts DIAGNOSTIC_TEST

Digital bitewing radiographs were manually annotated by calibrated human experts to identify the presence and location of proximal caries. Annotations were performed using standardized diagnostic criteria and dedicated imaging software to mark suspected lesions. These expert markings served as the reference standard for comparison with the artificial intelligence outputs. Inter-examiner agreement was assessed, and disagreements were resolved by consensus.

Group 2 : Digital Bitewing manually annotated by human experts

Artificial Intelligence (AI): Deep learning that is applied in Diagnosis of the proximal Caries DIAGNOSTIC_TEST

Artificial intelligence was used as a deep-learning diagnostic tool to detect proximal caries on digital bitewing radiographs. The system analyzed images and generated probability scores and visual markers for suspected lesions. Its performance was compared with expert examiner diagnoses as the reference standard. AI results were used for evaluation only and did not influence patient treatment decisions.

Group 1 Artificial Intelligence Deep learning that is applied in Diagnosis of the proximal Caries

Eligibility Criteria

• Age: 18 Years - 70 Years
• Sex: all
• Healthy Volunteers: Yes
• Age Groups: Adult (18-64), Older Adult (65+)
• Sampling Method: Non-Probability Sample

Study Population

Patients attending the Faculty dental out patient clinic who required bitewing radiographic examination for routine diagnosis or treatment planning.

You may qualify if:

• Patients having all Permanent premolars and molars (maximum one tooth missing on each side)

You may not qualify if:

• Dental Anomalies →Amelogenesis Imperfecta, Dentinogenesis Imperfecta, taurodontism 2- Severe crowding which prevent visualization of teeth Contacts 3-Orthodontic wires bonded to Enamel of the tooth

Sponsors & Collaborators

• Cairo University: lead
• Ain Shams University: collaborator

Study Sites (1)

Ain Shams University

Cairo, 11331, Egypt

Location

MeSH Terms

Conditions

Dental Caries

Interventions

Artificial Intelligence

Condition Hierarchy (Ancestors)

Tooth Demineralization; Tooth Diseases; Stomatognathic Diseases

Intervention Hierarchy (Ancestors)

Algorithms; Mathematical Concepts

Study Design

• Study Type: observational
• Observational Model: CASE CONTROL
• Time Perspective: PROSPECTIVE
• Sponsor Type: OTHER
• Responsible Party: PRINCIPAL INVESTIGATOR
• PI Title: Associate Professor, Conservative dentistry department

Study Record Dates

• First Submitted: February 4, 2026
• First Posted: February 11, 2026
• Study Start: January 15, 2023
• Primary Completion: September 15, 2025
• Study Completion: December 1, 2025
• Last Updated: February 11, 2026

Record last verified: 2026-02

Data Sharing

• IPD Sharing: Will not share

Locations

EG (1)