The purpose of this study is to evaluate and compare the performance of six state-of-the-art convolutional neural network (CNN)-based deep learning models for cervical vertebral maturation (CVM) on lateral cephalometric radiographs, and implement visualization of CVM classification for each model using gradient-weighted class activation map (Grad-CAM) technology. A total of 600 lateral cephalometric radiographs obtained from patients aged 6–19 years between 2013 and 2020 in Pusan National University Dental Hospital were used in this study. ResNet-18, MobileNet-v2, ResNet-50, ResNet-101, Inception-v3, and Inception-ResNet-v2 were tested to determine the optimal pre-trained network architecture. Multi-class classification metrics, accuracy, recall, precision, F1-score, and area under the curve (AUC) values from the receiver operating characteristic (ROC) curve were used to evaluate the performance of the models. All deep learning models demonstrated more than 90% accuracy, with Inception-ResNet-v2 performing the best, relatively. In addition, visualizing each deep learning model using Grad-CAM led to a primary focus on the cervical vertebrae and surrounding structures. The use of these deep learning models in clinical practice will facilitate dental practitioners in making accurate diagnoses and treatment plans.

In this study, we aimed to develop and evaluate the performance of deep-learning models that automatically classify mesiodens in primary or mixed dentition panoramic radiographs. Panoramic radiographs of 550 patients with mesiodens and 550 patients without mesiodens were used. Primary or mixed dentition patients were included. SqueezeNet, ResNet-18, ResNet-101, and Inception-ResNet-V2 were each used to create deep-learning models. The accuracy, precision, recall, and F1 score of ResNet-101 and Inception-ResNet-V2 were higher than 90%. SqueezeNet exhibited relatively inferior results. In addition, we attempted to visualize the models using a class activation map. In images with mesiodens, the deep-learning models focused on the actual locations of the mesiodens in many cases. Deep-learning technologies may help clinicians with insufficient clinical experience in more accurate and faster diagnosis.

Dental caries are one of the chronic diseases caused by organic acids made from oral microbes. However, there was a lack of knowledge about the oral microbiome of Korean children. The aim of this study was to analyze the metagenome data of the oral microbiome obtained from Korean children and to discover bacteria highly related to dental caries with machine learning models. Saliva and plaque samples from 120 Korean children aged below 12 years were collected. Bacterial composition was identified using Illumina HiSeq sequencing based on the V3–V4 hypervariable region of the 16S rRNA gene. Ten major genera accounted for approximately 70% of the samples on average, including Streptococcus, Neisseria, Corynebacterium, and Fusobacterium. Differential abundant analyses revealed that Scardovia wiggsiae and Leptotrichia wadei were enriched in the caries samples, while Neisseria oralis was abundant in the non-caries samples of children aged below 6 years. The caries and non-caries samples of children aged 6–12 years were enriched in Streptococcus mutans and Corynebacterium durum, respectively. The machine learning models based on these differentially enriched taxa showed accuracies of up to 83%. These results confirmed significant alterations in the oral microbiome according to dental caries and age, and these differences can be used as diagnostic biomarkers.

Objectives: The purpose of this study is to develop and evaluate the performance of a model that automatically sets a region of interest (ROI) and diagnoses mesiodens in panoramic radiographs of growing children using deep learning technology. Methods: Out of 988 panoramic radiographs, 489 patients with mesiodens were classified as an experimental group, and 499 patients without mesiodens were classified as a control group. This study consists of two networks. The first network (DeeplabV3plus) is a segmentation model that uses the posterior molar space to set the ROI in the maxillary anterior region with the mesiodens in the panoramic radiograph. The second network (Inception-resnet-v2) is a classification model that uses cropped maxillary anterior teeth to determine the presence of mesiodens. The data were divided into five groups and cross-validated. Deep learning model were created and trained using Inception-ResNet-v2. The performance of the segmentation network was evaluated using accuracy, Intersection over Union (IoU), and MeanBFscore. The overall network performance was evaluated using accuracy, precision, recall, and F1-score. Results: Segmentation performance using posterior molar space in panoramic radiographs was 0.839, IoU 0.762, and MeanBFscore 0.907. The mean values of accuracy, precision, recall, F1-score, and area under the curve for the diagnosis of mesiodens using automatic segmentation were 0.971, 0.971, 0.971, 0.971, and 0.971, respectively. Conclusions: The diagnostic performance of the deep learning system using posterior molar space on the panoramic radiograph was sufficiently useful. The results of the deep learning system confirmed the possibility of complete automation of the classification of mesiodens.

The aim of the study was to evaluate the effects of polymeric computer-aided design/computer-aided manufacturing CAD/CAM materials on antagonistic primary tooth wear. Five CAD/CAM polymeric materials were examined: Vipi Block Monocolor (VBM), Yamahachi polymethylmethacrylate (PMMA) (YAP), Mazic Duro (MZD), Vita Enamic (ENA), and Pekkton (PEK). All of the specimens were tested in a thermomechanical loading machine with the primary canine as the antagonist (50 N, 1.2 × 10⁵ cycles, 1.7 Hz, 5/55 °C). The wear losses of the antagonist tooth and the restorative materials were calculated using reverse modelling software and an electronic scale. VBM and ENA showed significantly higher antagonist tooth wear than PEK (p < 0.05), but there was no significant difference observed among VBM, YAP, MZD, and ENA (p > 0.05). PEK showed the largest value in both material volumetric and weight losses. In terms of material volumetric losses, there was no significant difference between all of the groups (p > 0.05). In terms of material weight losses, PEK was significantly larger than ENA (p < 0.05), but there was no significant difference between VBM, YAP, MZD, and ENA (p > 0.05). Volumetric and weight losses of materials showed similar wear behaviour. However, the wear patterns of antagonists and materials were different, especially in PEK.

Some resin polymers available for three-dimensional (3D) printing are slightly elastic, which may be advantageous when used for full crown coverage of the primary teeth. This study was performed to evaluate the mechanical properties of two types of 3D-printed resin crowns in terms of strength and surface characteristics. Polymer resins used for temporary crowns (TCs) and temporary flexible dentures (TFDs) were tested. Digitally designed crowns with different thicknesses (0.4 and 0.6 mm) were 3D-printed. Milled zirconia crowns were used as the control. The static and dynamic fracture loads of the crowns were measured. The crown surface was evaluated using scanning electron microscopy. The average strength did not differ between the types of crowns. The differences between the dynamic and static fracture loads were insignificant. In the TC group, thicker crowns showed lower strength both under static and dynamic loads. After thermomechanical loading, microcracks and dropouts of macrofillers were detected on the surface of all types of resin crowns. The deposition of abraded debris occurred more in the TFD group. The 3D-printed resin crowns were thought to endure biting forces in children. However, some limitations of the material itself should be improved for consideration as a new treatment option in pediatric dentistry.