Landmark annotation and mandibular lateral deviation analysis of posteroanterior cephalograms using a convolutional neural network

Image preprocessing with contrast-limited adaptive histogram equalization improves the segmentation performance of deep learning for the articular disk of the temporomandibular joint on magnetic resonance images

OBJECTIVES

The objective was to evaluate the robustness of deep learning (DL)-based encoder-decoder convolutional neural networks (ED-CNNs) for segmenting temporomandibular joint (TMJ) articular disks using data sets acquired from 2 different 3.0-T magnetic resonance imaging (MRI) scanners using original images and images subjected to contrast-limited adaptive histogram equalization (CLAHE).

STUDY DESIGN

In total, 536 MR images from 49 individuals were examined. An expert orthodontist identified and manually segmented the disks in all images, which were then reviewed by another expert orthodontist and 2 expert oral and maxillofacial radiologists. These images were used to evaluate a DL-based semantic segmentation approach using an ED-CNN. Original and preprocessed CLAHE images were used to train and validate the models whose performances were compared.

RESULTS

Original and CLAHE images acquired on 1 scanner had pixel values that were significantly darker and with lower contrast. The values of 3 metrics-the Dice similarity coefficient, sensitivity, and positive predictive value-were low when the original MR images were used for model training and validation. However, these metrics significantly improved when images were preprocessed with CLAHE.

CONCLUSIONS

The robustness of the ED-CNN model trained on a dataset obtained from a single device is low but can be improved with CLAHE preprocessing. The proposed system provides promising results for a DL-based, fully automated segmentation method for TMJ articular disks on MRI.

Machine learning-based decision support system for orthognathic diagnosis and treatment planning

BACKGROUND

Dento-maxillofacial deformities are common problems. Orthodontic-orthognathic surgery is the primary treatment but accurate diagnosis and careful surgical planning are essential for optimum outcomes. This study aimed to establish and verify a machine learning-based decision support system for treatment of dento-maxillofacial malformations.

METHODS

Patients (n = 574) with dento-maxillofacial deformities undergoing spiral CT during January 2015 to August 2020 were enrolled to train diagnostic models based on five different machine learning algorithms; the diagnostic performances were compared with expert diagnoses. Accuracy, sensitivity, specificity, and area under the curve (AUC) were calculated. The adaptive artificial bee colony algorithm was employed to formulate the orthognathic surgical plan, and subsequently evaluated by maxillofacial surgeons in a cohort of 50 patients. The objective evaluation included the difference in bone position between the artificial intelligence (AI) generated and actual surgical plans for the patient, along with discrepancies in postoperative cephalometric analysis outcomes.

RESULTS

The binary relevance extreme gradient boosting model performed best, with diagnostic success rates > 90% for six different kinds of dento-maxillofacial deformities; the exception was maxillary overdevelopment (89.27%). AUC was > 0.88 for all diagnostic types. Median score for the surgical plans was 9, and was improved after human-computer interaction. There was no statistically significant difference between the actual and AI- groups.

CONCLUSIONS

Machine learning algorithms are effective for diagnosis and surgical planning of dento-maxillofacial deformities and help improve diagnostic efficiency, especially in lower medical centers.

Deep learning for virtual orthodontic bracket removal: tool establishment and application

OBJECTIVE

We aimed to develop a tool for virtual orthodontic bracket removal based on deep learning algorithms for feature extraction from bonded teeth and to demonstrate its application in a bracket position assessment scenario.

MATERIALS AND METHODS

Our segmentation network for virtual bracket removal was trained using dataset A, containing 978 bonded teeth, 20 original teeth, and 20 brackets generated by scanners. The accuracy and segmentation time of the network were tested by dataset B, which included an additional 118 bonded teeth without knowing the original tooth morphology. This tool was then applied for bracket position assessment. The clinical crown center, bracket center, and orientations of separated teeth and brackets were extracted for analyzing the linear distribution and angular deviation of bonded brackets.

RESULTS

This tool performed virtual bracket removal in 2.9 ms per tooth with accuracies of 98.93% and 97.42% (P < 0.01) in datasets A and B, respectively. The tooth surface and bracket characteristics were extracted and used to evaluate the results of manually bonded brackets by 49 orthodontists. Personal preferences for bracket angulation and bracket distribution were displayed graphically and tabularly.

CONCLUSIONS

The tool's efficiency and precision are satisfactory, and it can be operated without original tooth data. It can be used to display the bonding deviation in the bracket position assessment scenario.

CLINICAL SIGNIFICANCE

With the aid of this tool, unnecessary bracket removal can be avoided when evaluating bracket positions and modifying treatment plans. It has the potential to produce retainers and orthodontic devices prior to tooth debonding.

A family-centered orthodontic screening approach using a machine learning-based mobile application

Background/purpose

Skeletal orthodontic deformities can have functional and aesthetic consequences, making early detection critical. This study aimed to address the issue of parents bringing their children for routine orthodontic checkups after the ideal treatment age has passed. To address this, we developed a mobile application that uses machine-learning to make a preliminary diagnosis of skeletal malocclusion using just one photograph.

Materials and methods

A retrospective study was conducted on 524 pre-pubertal children, aged between 5 and 12 years, to evaluate the accuracy of the machine learning based mobile application. The application detects multiple points in photographs taken from the mobile camera and generates a signal indicating the diagnosis of skeletal malocclusion.

Results

The final accuracy of the Class III vs not Class III model deployed to the mobile application was above 81%, indicating its ability to accurately identify skeletal malocclusion. On a separate validation dataset of 145 patients diagnosed by 5 different clinicians, the accuracy of Class II vs Class I model was 69%; And pg 4, ln 61: as Class II vs Class I with 69% accuracy.

Conclusion

The application provides parents with important information about the orthodontic problem, age of treatment, and various treatment options. This enables parents to seek further advice from an orthodontist at an earlier stage and make informed decisions. However, the diagnosis should still be confirmed by an orthodontist. This approach has the potential to improve access to orthodontic care, especially in underserved communities.

Application of Artificial Intelligence in Orthodontics: Current State and Future Perspectives

In recent years, there has been the notable emergency of artificial intelligence (AI) as a transformative force in multiple domains, including orthodontics. This review aims to provide a comprehensive overview of the present state of AI applications in orthodontics, which can be categorized into the following domains: (1) diagnosis, including cephalometric analysis, dental analysis, facial analysis, skeletal-maturation-stage determination and upper-airway obstruction assessment; (2) treatment planning, including decision making for extractions and orthognathic surgery, and treatment outcome prediction; and (3) clinical practice, including practice guidance, remote care, and clinical documentation. We have witnessed a broadening of the application of AI in orthodontics, accompanied by advancements in its performance. Additionally, this review outlines the existing limitations within the field and offers future perspectives.

Performance evaluation of E-nose and E-tongue combined with machine learning for qualitative and quantitative assessment of bear bile powder

Bear bile powder (BBP) is a valuable animal-derived product with a huge adulteration problem on market. It is a crucially important task to identify BBP and its counterfeit. Electronic sensory technologies are the inheritance and development of traditional empirical identification. Considering that each drug has its own specific odor and taste characteristics, electronic tongue (E-tongue), electronic nose (E-nose) and GC-MS were used to evaluate the aroma and taste of BBP and its common counterfeit. Two active components of BBP, namely tauroursodeoxycholic acid (TUDCA) and taurochenodeoxycholic acid (TCDCA) were measured and linked with the electronic sensory data. The results showed that bitterness was the main flavor of TUDCA in BBP, saltiness and umami were the main flavor of TCDCA. The volatiles detected by E-nose and GC-MS were mainly aldehydes, ketones, alcohols, hydrocarbons, carboxylic acids, heterocyclic, lipids, and amines, mainly earthy, musty, coffee, bitter almond, burnt, pungent odor descriptions. Four different machine learning algorithms (backpropagation neural network, support vector machine, K-nearest neighbor, and random forest) were used to identify BBP and its counterfeit, and the regression performance of these four algorithms was also evaluated. For qualitative identification, the algorithm of random forest has shown the best performance, with 100% accuracy, precision, recall and F1-score. Also, the random forest algorithm has the best R² and the lowest RMSE in terms of quantitative prediction.

Detecting the presence of supernumerary teeth during the early mixed dentition stage using deep learning algorithms: A pilot study

BACKGROUND

Supernumerary teeth are a common anomaly and are frequently observed in paediatric patients. To prevent or minimize complications, early diagnosis and treatment is ideal in children with supernumerary teeth.

AIM

This study aimed to apply convolutional neural network (CNN)-based deep learning to detect the presence of supernumerary teeth in children during the early mixed dentition stage.

DESIGN

Three CNN models, AlexNet, VGG16-TL, and InceptionV3-TL, were employed in this study. A total of 220 panoramic radiographs (from children aged 6 years 0 months to 9 years 6 months) including supernumerary teeth (cases, n = 120) or no anomalies (controls, n = 100) were retrospectively analyzed. The CNN performances were assessed according to accuracy, sensitivity, specificity, receiver operating characteristic (ROC) curves, and area under the ROC curves for a cross-validation test dataset.

RESULTS

The VGG16-TL model had the highest performance according to accuracy, sensitivity, specificity, and area under the ROC curve, but the other models had similar performance.

CONCLUSION

CNN-based deep learning is a promising approach for detecting the presence of supernumerary teeth during the early mixed dentition stage.

Analysis of the feasibility of using deep learning for multiclass classification of dental anomalies on panoramic radiographs

The aim of the feasibility study was to construct deep learning models for the classification of multiple dental anomalies in panoramic radiographs. Panoramic radiographs with single supernumerary teeth and/or odontomas were considered the "case" group; panoramic radiographs with no dental anomalies were considered the "control" group. The dataset comprised 150 panoramic radiographs: 50 each of no dental anomalies, single supernumerary teeth, and odontomas. To classify the panoramic radiographs into case and control categories, we employed AlexNet, which is a convolutional neural network model. AlexNet was able to classify whole panoramic radiographs into two or three classes, according to the presence or absence of supernumerary teeth or odontomas. The performance metrics of the three-class classification were 70%, 70.8%, 70%, and 69.7% for accuracy, precision, sensitivity, and F1 score, respectively, in the macro average. These results support the feasibility of using deep learning to detect multiple dental anomalies in panoramic radiographs.

Automated segmentation of articular disc of the temporomandibular joint on magnetic resonance images using deep learning

Temporomandibular disorders are typically accompanied by a number of clinical manifestations that involve pain and dysfunction of the masticatory muscles and temporomandibular joint. The most important subgroup of articular abnormalities in patients with temporomandibular disorders includes patients with different forms of articular disc displacement and deformation. Here, we propose a fully automated articular disc detection and segmentation system to support the diagnosis of temporomandibular disorder on magnetic resonance imaging. This system uses deep learning-based semantic segmentation approaches. The study included a total of 217 magnetic resonance images from 10 patients with anterior displacement of the articular disc and 10 healthy control subjects with normal articular discs. These images were used to evaluate three deep learning-based semantic segmentation approaches: our proposed convolutional neural network encoder-decoder named 3DiscNet (Detection for Displaced articular DISC using convolutional neural NETwork), U-Net, and SegNet-Basic. Of the three algorithms, 3DiscNet and SegNet-Basic showed comparably good metrics (Dice coefficient, sensitivity, and positive predictive value). This study provides a proof-of-concept for a fully automated deep learning-based segmentation methodology for articular discs on magnetic resonance images, and obtained promising initial results, indicating that the method could potentially be used in clinical practice for the assessment of temporomandibular disorders.