A deep-learning model for idiopathic osteosclerosis detection on dental panoramic radiographs

Abstract

<h3>Objectives</h3> The purpose of the study was to propose an automatic idiopathic osteosclerosis (IO) detection model based on the convolutional neural network (CNN) algorithms in panoramic radiographs using GoogleNet Inception V2 architecture and assess the performance of the artificial intelligence (AI) model compared to the human observer. <h3>Study Design</h3> A total of 493 anonymized panoramic radiographs were used to develop the AI system (CranioCatch, Eskisehir-Turkey) for the detection of IOs. The panoramic radiographs were acquired from the Radiology Archive of the Department of Oral and Maxillofacial Radiology of the Faculty of Dentistry of Eskisehir Osmangazi University. GoogleNet Inception V2 model implemented with TensorFlow library was used for the detection of IOs. A confusion matrix was used to predict model achievements. <h3>Results</h3> 50 IOs were detected by the AI model from the 52 test images, which had 57 IOs. The sensitivity, precision, and F-measure values were 0.88, 0.83, and 0.86, respectively. <h3>Conclusion</h3> Deep learning-based AI models have the potential to detect IOs accurately on panoramic radiographs. AI systems can be used to support clinicians for baseline diagnosis of panoramic radiographs and may eventually replace human observers' evaluation and support performance in the future. <b>Statement of Ethical Review</b> Human/Animal subjects were used and this study was approved by an institutional ethics panel