Document Type : Review
School of Dentistry, Isfahan University of Medical Sciences, Isfahan, Iran
Department of Pharmaceutical Biotechnology, School of Pharmacy and Pharmaceutical Sciences, Isfahan University of Medical Sciences, Isfahan, Iran
Department of Oral and Maxillofacial Radiology, Dental Implants Research Center, Dental Research Institute, School of Dentistry, Isfahan University of Medical Sciences, Isfahan, Iran Department of Neurosciences, Reproductive and Odontostomatological Sciences, University of Naples “Federico II”, Naples, Italy
Dental Research Center, Dental Research Institute, School of Dentistry, Isfahan University of Medical Sciences, Isfahan, Iran
Background: Mandibular fractures are a common trauma in oral and maxillofacial surgery. The accurate diagnosis of these fractures is crucial for successful treatment. However, the interpretation of radiographic scans can be time-consuming and prone to human error. The advent of artificial intelligence (AI), specifically Convolutional Neural Networks (CNNs), has opened up new possibilities for improving the accuracy and efficiency of fracture detection.
Objectives: This review aims to explore the role of AI in detecting mandibular fractures.
Methods: A comprehensive literature search was performed using PubMed, Embase, Web of Science, and Google Scholar databases. Studies were included if they used AI techniques, specifically CNNs or transformers, for the detection of mandibular fractures.
Results: The systematic search yielded 53 studies, with eight studies meeting the inclusion criteria. The AI models across these studies demonstrated a generally high degree of effectiveness in detecting mandibular fractures, with F1 scores ranging from 45% to 100%. Some studies also compared the diagnostic prowess of human clinicians and AI models, with AI models often matching or surpassing human performance.
Conclusion: The application of AI in detecting mandibular fractures represents a promising avenue of research. AI models have the potential to reduce the workload of radiologists, improve the efficiency of fracture detection, and lead to faster diagnosis and treatment. However, further research is needed to validate these findings in larger and more diverse datasets and to address challenges such as the interpretability of AI algorithms and the availability of large, well-annotated datasets.