Diagnostic accuracy of an artificial intelligence-driven platform in assessing periapical healing and endodontic treatment outcomes on panoramic radiographs: a retrospective study

INTRODUCTION. Evaluation of endodontic treatment outcomes through radiographic assessment is subject to interobserver variability and depends heavily on clinician experience. Artificial intelligence (AI) platforms offer potential for standardized, objective assessment of periapical healing.

MATERIALS AND METHODS. This retrospective study analyzed 400 panoramic radiographs from patients who underwent root canal treatment between January 2023 and December 2024. An AI platform developed using TensorFlow and Keras, with model training in PyTorch and validation in MATLAB Deep Learning Toolbox, was employed. Three blinded expert endodontists independently assessed all radiographs, with consensus serving as the gold standard. Outcomes were classified as healed, healing, or diseased based on periapical index criteria. Diagnostic performance metrics including sensitivity, specificity, accuracy, positive predictive value (PPV), and negative predictive value (NPV) were calculated.

RESULTS. The AI platform demonstrated overall accuracy of 89.8% in classifying treatment outcomes. For detecting healed cases, sensitivity was 92.3%, specificity 87.6%, PPV 88.9%, and NPV 91.5%. For diseased / persistent pathology detection, sensitivity was 88.7%, specificity 93.2%, PPV 84.3%, and NPV 95.1%. Agreement between AI and expert consensus was substantial (Cohen’s κ = 0.834, p < 0.001). AI performance was superior in anterior teeth (93.2% accuracy) compared to molars (86.4% accuracy, p = 0.008). Processing time per radiograph averaged 2.3 ± 0.4 seconds.

CONCLUSIONS. The AI-driven platform demonstrated high diagnostic accuracy comparable to expert assessment, with potential for standardized, rapid evaluation of endodontic treatment outcomes. Further prospective validation and clinical integration studies are warranted.

1. Ng Y.L., Mann V., Gulabivala K. A prospective study of the factors affecting outcomes of nonsurgical root canal treatment: part 1: periapical health. Int Endod J. 2011;44(7):583–609. https://doi.org/10.1111/j.1365-2591.2011.01872.x

2. Patel S., Wilson R., Dawood A., Mannocci F. The detection of periapical pathosis using periapical radiography and cone beam computed tomography – part 1: pre-operative status. Int Endod J. 2012;45(8):702–710. https://doi.org/10.1111/j.1365-2591.2011.01989.x

3. Rushton V.E., Horner K., Worthington H.V. The quality of panoramic radiographs in a sample of general dental practices. Br Dent J. 1999;186(12):630–633. https://doi.org/10.1038/sj.bdj.4800182

4. Molander B., Ahlqwist M., Gröndahl H.G., Hollender L. Comparison of panoramic and intraoral radiography for the diagnosis of caries and periapical pathology. Dentomaxillofac Radiol. 1993;22(1):28–32. https://doi.org/10.1259/dmfr.22.1.8508938

5. Orstavik D., Kerekes K., Eriksen H.M. The periapical index: a scoring system for radiographic assessment of apical periodontitis. Endod Dent Traumatol. 1986;2(1):20–34. https://doi.org/10.1111/j.1600-9657.1986.tb00119.x

6. Tarcin B., Gumru B., Iriboz E., Turkaydin D.E., Ovecoglu H.S. Radiologic assessment of periapical health: Comparison of 3 different index systems. J Endod. 2015;41(11):1834–1838. https://doi.org/10.1016/j.joen.2015.08.010

7. Topol E.J. High-performance medicine: the convergence of human and artificial intelligence. Nat Med. 2019;25(1):44–56. https://doi.org/10.1038/s41591-018-0300-7

8. Esteva A., Kuprel B., Novoa R.A., Ko J., Swetter S.M., Blau H.M., Thrun S. Dermatologist-level classification of skin cancer with deep neural networks. Nature. 2017;542(7639):115–118. https://doi.org/10.1038/nature21056 (Erratum in: Nature. 2017;546(7660):686. https://doi.org/10.1038/nature22985).

9. Schwendicke F., Golla T., Dreher M., Krois J. Convolutional neural networks for dental image diagnostics: A scoping review. J Dent. 2019;91:103226. https://doi.org/10.1016/j.jdent.2019.103226

10. Setzer F.C., Shi K.J., Zhang Z., Yan H., Yoon H., Mupparapu M., Li J. Artificial intelligence for the computer-aided detection of periapical lesions in cone-beam computed tomographic images. J Endod. 2020;46(7):987–993. https://doi.org/10.1016/j.joen.2020.03.025 (Erratum in: J Endod. 2026;52(1):156. https://doi.org/10.1016/j.joen.2025.09.022).

11. Fukuda M., Inamoto K., Shibata N., Ariji Y., Yanashita Y., Kutsuna S. et al. Evaluation of an artificial intelligence system for detecting vertical root fracture on panoramic radiography. Oral Radiol. 2020;36(4):337–343. https://doi.org/10.1007/s11282-019-00409-x

12. Ekert T., Krois J., Meinhold L., Elhennawy K., Emara R., Golla T., Schwendicke F. Deep learning for the radiographic detection of apical lesions. J Endod. 2019;45(7):917–922. e5. https://doi.org/10.1016/j.joen.2019.03.016

13. Yamashita R., Nishio M., Do R.K.G., Togashi K. Convo lutional neural networks: an overview and application in radiology. Insights Imaging. 2018;9(4):611–629. https://doi.org/10.1007/s13244-018-0639-9

14. Chen H., Zhang K., Lyu P., Li H., Zhang L., Wu J., Lee C.H. A deep learning approach to automatic teeth detection and numbering based on object detection in dental periapical films. Sci Rep. 2019;9(1):3840. https://doi.org/10.1038/s41598-019-40414-y

15. Jonasson P., Kvist T. Diagnosis of apical periodontitis in root-filled teeth. Clin Dent Rev. 2018;2:15. https://doi.org/10.1007/s41894-018-0029-1