A Novel Approach to Identifying C-Shaped Canals in Mandibular Second Molars by Occlusal Photographs: An Artificial Intelligence Analysis

Objectives: Accurate identification of root canal morphology, particularly C-shaped canals in mandibular second molars, is crucial for successful endodontic treatment. Current diagnostic methods rely on radiographic imaging, which has limitations including radiation exposure and cost. To date, no studies have investigated the use of intraoral photographs for identifying C-shaped canals. This study aimed to evaluate the potential of artificial intelligence (AI) in identifying C-shaped canals using occlusal photographs, a novel and non-radiographic approach. Methods: Extracted mandibular second molars (n=231) were collected and mounted. Occlusal surface photographs were taken using a wireless oral camera. Cone-beam computed tomography (CBCT) images served as the gold standard for canal morphology classification. A deep learning model was developed using various convolutional neural network architectures. The model was trained on 80% of the dataset and tested on the remaining 20%. Results: The best-performing networks achieved an accuracy of 85.1% in distinguishing between C-shaped and non-C-shaped root canal morphologies. Conclusion: This study demonstrated the potential of a novel, radiation-free approach to identify complex root canal anatomies via AI analysis of occlusal photographs, useful for initial screening when radiographic equipment is unavailable or contraindicated.