Deep learning-based assessment of the relation between the third molar and mandibular canal on panoramic radiographs using local, centralized, and federated learning in a simulated multi-center setting

Rubak, Johan Andreas Balle and Haghighat, Sara and Jain, Sanyam and Aldesoki, Mostafa and Chaurasia, Akhilanand and Ehsani, Sarah Sadat and Ghanatkaman, Faezeh Dehghan and Ghazali, Ahmad Badruddin and Issa, Julien and Khalil, Basel and Ramani, Rishi and Pauwels, Ruben (2026) Deep learning-based assessment of the relation between the third molar and mandibular canal on panoramic radiographs using local, centralized, and federated learning in a simulated multi-center setting. Applied Sciences, 16 (12). pp. 1-21. E-ISSN 2076-3417

PDF - Published Version Restricted to Registered users only Download (2MB) | Request a copy

Abstract

Impaction of the mandibular third molar in proximity to the mandibular canal increases the risk of inferior alveolar nerve injury. Panoramic radiography is routinely used to assess this relationship. Automated classification of molar–canal overlap could support clinical triage and reduce unnecessary CBCT referrals, while Federated Learning (FL) enables multi-center collaboration without sharing patient data. We compared Local Learning (LL), FL, and Centralized Learning (CL) for binary overlap/no-overlap classification on cropped panoramic radiographs partitioned across eight independent labelers in a simulated heterogeneous multi-center setting. A pretrained ResNet-34 was trained under each paradigm and evaluated using per-client metrics with locally optimized thresholds and pooled test performance with a global threshold. Performance was assessed using area under the receiver operating characteristic curve (AUC) and threshold-based metrics, alongside training dynamics, Grad-CAM visualizations, and server-side aggregate monitoring signals. On the test set, CL achieved the highest performance (AUC 0.831; accuracy ≈ 0.782), FL showed intermediate performance (AUC 0.757; accuracy ≈ 0.703), and LL generalized poorly across clients (AUC range ≈ 0.619–0.734; mean ≈ 0.672). Training curves suggested overfitting, particularly in LL models, and Grad-CAM indicated more anatomically focused attention in CL and FL. Overall, centralized training provided the strongest performance, while FL offers a privacy-preserving alternative that outperforms LL.

Item Type:	Article (Journal)
Uncontrolled Keywords:	dental radiography; third molar; mandibular canal; deep learning; federated learning; data heterogeneity; non-IID
Subjects:	R Medicine > RK Dentistry R Medicine > RK Dentistry > RK318 Oral and Dental Medicine. Pathology. Diseases-Therapeutics-General Works
Kulliyyahs/Centres/Divisions/Institutes (Can select more than one option. Press CONTROL button):	Kulliyyah of Dentistry Kulliyyah of Dentistry > Department of Oral Maxillofacial Surgery and Oral Diagnosis
Depositing User:	Dr Ahmad Badruddin Ghazali
Date Deposited:	29 Jun 2026 12:22
Last Modified:	29 Jun 2026 12:22
Queue Number:	2026-06-Q3789
URI:	http://irep.iium.edu.my/id/eprint/129521

Actions (login required)

View Item