Deep-learning classification using convolutional neural network for evaluation of maxillary sinusitis on panoramic radiography.

OBJECTIVES: To apply a deep-learning system for diagnosis of maxillary sinusitis on panoramic radiography, and to clarify its diagnostic performance. METHODS: Training data for 400 healthy and 400 inflamed maxillary sinuses were enhanced to 6000 samples in each category by data augmentation. Image patches were input into a deep-learning system, the learning process was repeated for 200 epochs, and a learning model was created. Newly-prepared testing image patches from 60 healthy and 60 inflamed sinuses were input into the learning model, and the diagnostic performance was calculated. Receiver-operating characteristic (ROC) curves were drawn, and the area under the curve (AUC) values were obtained. The results were compared with those of two experienced radiologists and two dental residents. RESULTS: The diagnostic performance of the deep-learning system for maxillary sinusitis on panoramic radiographs was high, with accuracy of 87.5%, sensitivity of 86.7%, specificity of 88.3%, and AUC of 0.875. These values showed no significant differences compared with those of the radiologists and were higher than those of the dental residents. CONCLUSIONS: The diagnostic performance of the deep-learning system for maxillary sinusitis on panoramic radiographs was sufficiently high. Results from the deep-learning system are expected to provide diagnostic support for inexperienced dentists.

Utilization of computer-aided detection system in diagnosing unilateral maxillary sinusitis on panoramic radiographs.

OBJECTIVES: It is unclear whether computer-aided detection (CAD) systems for panoramic radiography can help inexperienced dentists to diagnose maxillary sinusitis. The aim of this study was to clarify whether a CAD system for panoramic radiography can contribute to improved diagnostic performance for maxillary sinusitis by inexperienced dentists. METHODS: The panoramic radiographs of 49 patients with maxillary sinusitis and 49 patients with healthy sinuses were evaluated in this study. The diagnostic performance of the CAD system was determined. 12 inexperienced dentists and 4 expert oral and maxillofacial radiologists observed the total of 98 panoramic radiographs and judged the presence or absence of maxillary sinusitis, under conditions with and without the support of the CAD system. The receiver operating characteristic curves of the two groups were compared. RESULTS: The CAD system provided sensitivity of 77.6%, specificity of 69.4% and accuracy of 73.5%. The diagnostic performance of the inexperienced dentists increased with the support of the CAD system. When the inexperienced dentists diagnosed maxillary sinusitis with CAD support, the area under the curve (AUC) was significantly higher than that without CAD support. When the focus was only on panoramic radiographs in which CAD support led to a correct diagnosis, the AUC of the inexperienced dentists increased to an equivalent level to that of the experienced radiologists. CONCLUSIONS: The CAD system supported the inexperienced dentists in diagnosing maxillary sinusitis on the panoramic radiographs. If the accuracy of the CAD system can be increased, the benefits of CAD support will be further enhanced.