Deep Learning for Caries Detection and Classification

Abstract

Objectives: Deep learning methods have achieved impressive diagnostic performance in the field of radiology. The current study aimed to use deep learning methods to detect caries lesions, classify different radiographic extensions on panoramic films, and compare the classification results with those of expert dentists. Methods: A total of 1160 dental panoramic films were evaluated by three expert dentists. All caries lesions in the films were marked with circles, whose combination was defined as the reference dataset. A training and validation dataset (1071) and a test dataset (89) were then established from the reference dataset. A convolutional neural network, called nnU-Net, was applied to detect caries lesions, and DenseNet121 was applied to classify the lesions according to their depths (dentin lesions in the outer, middle, or inner third D1/2/3 of dentin). The performance of the test dataset in the trained nnU-Net and DenseNet121 models was compared with the results of six expert dentists in terms of the intersection over union (IoU), Dice coefficient, accuracy, precision, recall, negative predictive value (NPV), and F1-score metrics. Results: nnU-Net yielded caries lesion segmentation IoU and Dice coefficient values of 0.785 and 0.663, respectively, and the accuracy and recall rate of nnU-Net were 0.986 and 0.821, respectively. The results of the expert dentists and the neural network were shown to be no different in terms of accuracy, precision, recall, NPV, and F1-score. For caries depth classification, DenseNet121 showed an overall accuracy of 0.957 for D1 lesions, 0.832 for D2 lesions, and 0.863 for D3 lesions. The recall results of the D1/D2/D3 lesions were 0.765, 0.652, and 0.918, respectively. All metric values, including accuracy, precision, recall, NPV, and F1-score values, were proven to be no different from those of the experienced dentists. Conclusion: In detecting and classifying caries lesions on dental panoramic radiographs, the performance of deep learning methods was similar to that of expert dentists. The impact of applying these well-trained neural networks for disease diagnosis and treatment decision making should be explored.

Caries detection with tooth surface segmentation on intraoral photographic images using deep learning

“…Training the DL model for caries detection using these pre-processed images showed improved model performance compared to training using the original photographic images without pre-processing. Similar results have been observed in previous studies [ 21 , 37 – 40 ]. In a study by Lian et al a DL model showed good performance in detection and classification by localising the carious lesion after segmenting only the tooth outline within the entire image of the dental panorama [ 21 ].…”

“…However, relatively large area including several teeth can be captured in one intraoral photograph frequently, which may affect accuracy of CNN model. In case of dental panoramic images, there were some trials to detect carious lesion better by focusing tooth surfaces through image preprocessing with segmentation [ 21 ]. Segmentation task by CNN is a kind of classification at every pixel of the input image and can discriminate different anatomical structures in medical images [ 22 ].…”

Context Aware Convolutional Neural Network for Children Caries Diagnosis on Dental Panoramic Radiographs

“…Various CNN methods have been tested for panoramic radiography caries classification, and relatively good performances have been obtained [12,14,[27][28][29]. For example, Bui et al [14] tested several typical CNN methods including the most famous Alexnet, Googlenet, VGG16, VGG19, Resnet18, Resnet50, Resnet101, and Xception networks.…”

“…First, the proposed context aware CNN model is based on the extraction of each tooth from the dental panoramic radiographs, so extra human annotation is necessary, which limits its usage in clinical applications when performing computer aided diagnosis. Second, caries has different stages [ 28 ], so instead of making a binary classification of whether a tooth is a caries, it is more helpful to assist dentist to make an accurate diagnosis of what degree of the caries is. A follow-up study by considering different stages of caries and an end-to-end extraction and classification framework is necessary and will be studied in the future work.…”

“…Virupaiah and Sathyanarayana [22] utilized Gaussian low pass filter to preprocess dental images, based on which, statistical features are extracted and used by support vector machine classifier for caries classification. Recently, with the breakthrough of deep learning [23,24], especially its outstanding performances of image analysis based on CNN technology [25,26], several works have been proposed for panoramic radiography caries diagnosis based on CNN [12,14,[27][28][29].…”