Boichenko O. M., Bublii T. D., Ivanytskyi I. O.

ARTIFICIAL INTELLIGENCE IN THE DIAGNOSIS OF HIDDEN PROXIMAL CARIES: COMPARISON WITH THE TRADITIONAL METHOD

---

Show/Download

About the author:

Boichenko O. M., Bublii T. D., Ivanytskyi I. O.

Heading:

METHODS AND METHODOLOGIES

Type of article:

Scientific article

Annotation:

Proximal caries is difficult to detect early because contact surfaces are poorly accessible to direct clinical inspection. Digital bitewing radiography expands diagnostic capability, while artificial intelligence may support image interpretation. Objective. To compare the diagnostic characteristics of a traditional visual-instrumental method and an artificial intelligence system for detecting proximal caries of different depths and to assess the potential value of a clinicianplus- AI second-opinion model. A clinical-radiological study was conducted involving 150 patients aged 18–45 years, with 600 proximal tooth surfaces assessed. Following clinical examination, digital bitewing radiographs were obtained. The results of traditional diagnosis were compared with those obtained by automated radiographic analysis using an artificial intelligence system. Diagnostic performance was assessed using sensitivity, specificity, overall accuracy, and the numbers of true-positive, false-positive, true-negative, and false-negative results. McNemar’s test was used to compare paired categorical data. Results. For the detection of carious lesions confined to enamel, the sensitivity of the traditional method was 49.5% (95% CI: 42.6–56.4), compared with 95.0% (95% CI: 90.8–97.4) for the AI system. Overall accuracy was 81.3% and 94.7%, respectively, while the number of false-negative results was 100 and 10. The difference between the methods was statistically significant (χ²=72.4; p<0.001). For lesions involving the enamel-dentin junction and dentin, sensitivity was 82.8% (95% CI: 75.7–88.1) for the traditional method and 97.9% (95% CI: 93.8–99.3) for the AI system, while overall accuracy was 95.0% and 96.9%, respectively. The number of false-negative results was 25 for the traditional method and 3 for AI-assisted diagnosis; the difference was statistically significant (χ²=15.2; p<0.01). At the same time, the specificity of the traditional method remained slightly higher than that of the AI system. The artificial intelligence system demonstrated higher sensitivity and overall diagnostic accuracy for proximal caries detection than the traditional method, particularly for early enamel lesions. AI-assisted diagnosis substantially reduced the number of false-negative results. These findings support the potential use of AI as a “second opinion” tool in dental practice. However, the final diagnostic decision should remain with the dentist and should be based on the integration of clinical and radiographic findings.

Tags:

Bibliography:

1. World Health Organization. Sugars and dental caries. Geneva: WHO; 2025.
2. Ismail AI, Sohn W, Tellez M, Amaya A, Sen A, Hasson H, et al. The International Caries Detection and Assessment System (ICDAS): an integrated system for measuring dental caries. Community Dent Oral Epidemiol. 2007;35(3):170-178. DOI: https://doi.org/10.1111/j.1600-0528.2007.00347.x
3. Ammar N, Kühnisch J. Diagnostic performance of artificial intelligence-aided caries detection on bitewing radiographs: a systematic review and meta-analysis. Jpn Dent Sci Rev. 2024;60:128-136. DOI: https://doi.org/10.1016/j.jdsr.2024.02.001
4. Salehizeinabadi M, Math SY, Ameli N, Demirturk H, Pacheco-Pereira C. Effectiveness of human-artificial intelligence collaboration for interproximal caries detection using bitewing radiographs: a systematic review. Oral Surg Oral Med Oral Pathol Oral Radiol. 2026. DOI: https://doi.org/10.1016/j.oooo.2026.07.005
5. Wang Y, Fu D, Zhou G, Zou L. AI-based detection of proximal caries across radiographic modalities: a systematic review and metaanalysis. Dentomaxillofac Radiol. 2026. DOI: https://doi.org/10.1093/dmfr/twag046
6. Whiting PF, Rutjes AW, Westwood ME, Mallett S, Deeks JJ, Reitsma JB, et al. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med. 2011;155(8):529-536. DOI: https://doi.org/10.7326/0003-4819-155-8-201110180-00009
7. Shivakumar KM, Prasad S, Chandu GN. International caries detection and assessment system: a new paradigm in detection of dental caries. J Conserv Dent. 2009;12(1):10-16. DOI: https://doi.org/10.4103/0972-0707.53335
8. World Medical Association. World Medical Association Declaration of Helsinki: Ethical Principles for Medical Research Involving Human Participants. JAMA. 2025;333(1):71-74. DOI: https://doi.org/10.1001/jama.2024.21972
9. Meléndez Rojas P, Rodríguez Luengo M, Durán Anrique M, Niklander S, Villalobos Dellafiori MF, Jamett Rojas J, et al. Artificial Intelligence Tools for Dental Caries Detection: A Scoping Review. Oral. 2025;5(4):102. DOI: https://doi.org/10.3390/oral5040102
10. Cohen J. A coefficient of agreement for nominal scales. Educ Psychol Meas. 1960;20(1):37-46.

Publication of the article:

«Bulletin of problems biology and medicine», 2026 Issue 3, 182, 385-389 pages, index UDC 616.314-002-07-08:378.147

DOI:

10.29254/2077-4214-2026-3-182-385-389