Deep Learning for Automated Segmentation of the Edentulous Region on Cone Beam Computed Tomography for Dental Implant Planning: A Narrative Review

Rene Agatha¹, Eha Renwi Astuti^2*, Alhidayati Asymal²

Abstract

Background: Edentulism remains a considerable concern in dental practice, and dental implant is favored among rehabilitation options for its stability and long-term survival. The outcome of implant therapy depends heavily on pre-surgical planning, which requires precise identification of the edentulous area, the residual alveolar bone, and adjacent vital structures on Cone Beam Computed Tomography (CBCT). Conventional manual segmentation of these structures is time-consuming, requires substantial operator expertise, and is prone to inter-operator variability. Deep learning-based automated segmentation has therefore been increasingly explored as a solution to these limitations. Purpose: This narrative review discusses the current application, performance, and limitations of deep learning for automated segmentation of the edentulous area on CBCT in supporting dental implant planning. Review: A literature search was conducted through PubMed, Scopus, ScienceDirect, IEEE Xplore, and Google Scholar for articles published between 2015 and 2026 using combinations of keywords related to artificial intelligence, deep learning, CBCT, segmentation, and dental implant. Convolutional neural network-based architectures, particularly U-Net and YOLO-derived models, have been applied to automatically delineate the edentulous alveolar ridge together with the mandibular canal and maxillary sinus, achieving accuracy that approaches manual segmentation performed by experienced clinicians while considerably shortening analysis time. Dataset limitations, cross-institutional variability, and the continued need for clinician oversight remain the main challenges. Conclusion: Deep learning offers meaningful potential to improve the efficiency and consistency of CBCT-based edentulous area segmentation, although its clinical use should remain supervised by trained clinicians.

Keywords

artificial intelligence; cone beam computed tomography; deep learning; dental implant; image segmentation

Cite This Article

Agatha. R., Astuti, E. R., Asymal, A. (2026). Deep Learning for Automated Segmentation of the Edentulous Region on Cone Beam Computed Tomography for Dental Implant Planning: A Narrative Review. International Journal of Scientific Advances (IJSCIA), Volume 7| Issue 5: Sep – Oct 2026, Pages 483-489 URL: https://www.ijscia.com/wp-content/uploads/2026/09/Volume7-Issue5-Sep-Oct-No.1067-483-489.pdf

Volume 7 | Issue 5: Sep – Oct 2026