Current Applications of Deep Learning and Radiomics on CT and CBCT for Maxillofacial Diseases

CT-based radiomics analysis of different machine learning models for differentiating gnathic fibrous dysplasia and ossifying fibroma

OBJECTIVE

In this study, our aim was to develop and validate the effectiveness of diverse radiomic models for distinguishing between gnathic fibrous dysplasia (FD) and ossifying fibroma (OF) before surgery.

MATERIALS AND METHODS

We enrolled 220 patients with confirmed FD or OF. We extracted radiomic features from nonenhanced CT images. Following dimensionality reduction and feature selection, we constructed radiomic models using logistic regression, support vector machine, random forest, light gradient boosting machine, and eXtreme gradient boosting. We then identified the best radiomic model using receiver operating characteristic (ROC) curve analysis. After combining radiomics features with clinical features, we developed a comprehensive model. ROC curve and decision curve analysis (DCA) demonstrated the models' robustness and clinical value.

RESULTS

We extracted 1834 radiomic features from CT images, reduced them to eight valuable features, and achieved high predictive efficiency, with area under curves (AUC) exceeding 0.95 for all the models. Ultimately, our combined model, which integrates radiomic and clinical data, displayed superior discriminatory ability (AUC: training cohort 0.970; test cohort 0.967). DCA highlighted its optimal clinical efficacy.

CONCLUSION

Our combined model effectively differentiates between FD and OF, offering a noninvasive and efficient approach to clinical decision-making.

Surveying the landscape of diagnostic imaging in dentistry's future: Four emerging technologies with promise

BACKGROUND

Advances in digital radiography for both intraoral and panoramic imaging and cone-beam computed tomography have led the way to an increase in diagnostic capabilities for the dental care profession. In this article, the authors provide information on 4 emerging technologies with promise.

TYPES OF STUDIES REVIEWED

The authors feature the following: artificial intelligence in the form of deep learning using convolutional neural networks, dental magnetic resonance imaging, stationary intraoral tomosynthesis, and second-generation cone-beam computed tomography sources based on carbon nanotube technology and multispectral imaging. The authors review and summarize articles featuring these technologies.

RESULTS

The history and background of these emerging technologies are previewed along with their development and potential impact on the practice of dental diagnostic imaging. The authors conclude that these emerging technologies have the potential to have a substantial influence on the practice of dentistry as these systems mature. The degree of influence most likely will vary, with artificial intelligence being the most influential of the 4.

CONCLUSIONS AND PRACTICAL IMPLICATIONS

The readers are informed about these emerging technologies and the potential effects on their practice going forward, giving them information on which to base decisions on adopting 1 or more of these technologies. The 4 technologies reviewed in this article have the potential to improve imaging diagnostics in dentistry thereby leading to better patient care and heightened professional satisfaction.

Dental Caries Detection and Classification in CBCT Images Using Deep Learning

OBJECTIVES

This study aimed to investigate the accuracy of deep learning algorithms to diagnose tooth caries and classify the extension and location of dental caries in cone beam computed tomography (CBCT) images. To the best of our knowledge, this is the first study to evaluate the application of deep learning for dental caries in CBCT images.

METHODS

The CBCT image dataset comprised 382 molar teeth with caries and 403 noncarious molar cases. The dataset was divided into a development set for training and validation and test set. Three images were obtained for each case, including axial, sagittal, and coronal. The test dataset was provided to a multiple-input convolutional neural network (CNN). The network made predictions regarding the presence or absence of dental decay and classified the lesions according to their depths and types for the provided samples. Accuracy, sensitivity, specificity, and F1 score values were measured for dental caries detection and classification.

RESULTS

The diagnostic accuracy, sensitivity, specificity, and F1 score for caries detection in carious molar teeth were 95.3%, 92.1%, 96.3%, and 93.2%, respectively, and for noncarious molar teeth were 94.8%, 94.3%, 95.8%, and 94.6%. The CNN network showed high sensitivity, specificity, and accuracy in classifying caries extensions and locations.

CONCLUSIONS

This research demonstrates that deep learning models can accurately identify dental caries and classify their depths and types with high accuracy, sensitivity, and specificity. The successful application of deep learning in this field will undoubtedly assist dental practitioners and patients in improving diagnostic and treatment planning in dentistry.

CLINICAL SIGNIFICANCE

This study showed that deep learning can accurately detect and classify dental caries. Deep learning can provide dental caries detection accurately. Considering the shortage of dentists in certain areas, using CNNs can lead to broader geographic coverage in detecting dental caries.

Second mesiobuccal canal segmentation with YOLOv5 architecture using cone beam computed tomography images

The objective of this study is to use a deep-learning model based on CNN architecture to detect the second mesiobuccal (MB2) canals, which are seen as a variation in maxillary molars root canals. In the current study, 922 axial sections from 153 patients' cone beam computed tomography (CBCT) images were used. The segmentation method was employed to identify the MB2 canals in maxillary molars that had not previously had endodontic treatment. Labeled images were divided into training (80%), validation (10%) and testing (10%) groups. The artificial intelligence (AI) model was trained using the You Only Look Once v5 (YOLOv5x) architecture with 500 epochs and a learning rate of 0.01. Confusion matrix and receiver-operating characteristic (ROC) analysis were used in the statistical evaluation of the results. The sensitivity of the MB2 canal segmentation model was 0.92, the precision was 0.83, and the F1 score value was 0.87. The area under the curve (AUC) in the ROC graph of the model was 0.84. The mAP value at 0.5 inter-over union (IoU) was found as 0.88. The deep-learning algorithm used showed a high success in the detection of the MB2 canal. The success of the endodontic treatment can be increased and clinicians' time can be preserved using the newly created artificial intelligence-based models to identify variations in root canal anatomy before the treatment.

Radiographic Imaging for the Diagnosis and Treatment of Patients with Skeletal Class III Malocclusion

Skeletal Class III malocclusion is one type of dentofacial deformity that significantly affects patients' facial aesthetics and oral health. The orthodontic treatment of skeletal Class III malocclusion presents challenges due to uncertainties surrounding mandibular growth patterns and treatment outcomes. In recent years, disease-specific radiographic features have garnered interest from researchers in various fields including orthodontics, for their exceptional performance in enhancing diagnostic precision and treatment effect predictability. The aim of this narrative review is to provide an overview of the valuable radiographic features in the diagnosis and management of skeletal Class III malocclusion. Based on the existing literature, a series of analyses on lateral cephalograms have been concluded to identify the significant variables related to facial type classification, growth prediction, and decision-making for tooth extractions and orthognathic surgery in patients with skeletal Class III malocclusion. Furthermore, we summarize the parameters regarding the inter-maxillary relationship, as well as different anatomical structures including the maxilla, mandible, craniofacial base, and soft tissues from conventional and machine learning statistical models. Several distinct radiographic features for Class III malocclusion have also been preliminarily observed using cone beam computed tomography (CBCT) and magnetic resonance imaging (MRI).

Autologous Transplantation Tooth Guide Design Based on Deep Learning

BACKGROUND

Autologous tooth transplantation requires precise surgical guide design, involving manual tracing of donor tooth contours based on patient cone-beam computed tomography (CBCT) scans. While manual corrections are time-consuming and prone to human errors, deep learning-based approaches show promise in reducing labor and time costs while minimizing errors. However, the application of deep learning techniques in this particular field is yet to be investigated.

PURPOSE

We aimed to assess the feasibility of replacing the traditional design pipeline with a deep learning-enabled autologous tooth transplantation guide design pipeline.

STUDY DESIGN, SETTING, SAMPLE

This retrospective cross-sectional study used 79 CBCT images collected at the Guangzhou Medical University Hospital between October 2022 and March 2023. Following preprocessing, a total of 5,070 region of interest images were extracted from 79 CBCT images.

PREDICTOR VARIABLE

Autologous tooth transplantation guide design pipelines, either based on traditional manual design or deep learning-based design.

MAIN OUTCOME VARIABLE

The main outcome variable was the error between the reconstructed model and the gold standard benchmark. We used the third molar extracted clinically as the gold standard and leveraged it as the benchmark for evaluating our reconstructed models from different design pipelines. Both trueness and accuracy were used to evaluate this error. Trueness was assessed using the root mean square (RMS), and accuracy was measured using the standard deviation. The secondary outcome variable was the pipeline efficiency, assessed based on the time cost. Time cost refers to the amount of time required to acquire the third molar model using the pipeline.

ANALYSES

Data were analyzed using the Kruskal-Wallis test. Statistical significance was set at P < .05.

RESULTS

In the surface matching comparison for different reconstructed models, the deep learning group achieved the lowest RMS value (0.335 ± 0.066 mm). There were no significant differences in RMS values between manual design by a senior doctor and deep learning-based design (P = .688), and the standard deviation values did not differ among the 3 groups (P = .103). The deep learning-based design pipeline (0.017 ± 0.001 minutes) provided a faster assessment compared to the manual design pipeline by both senior (19.676 ± 2.386 minutes) and junior doctors (30.613 ± 6.571 minutes) (P < .001).

CONCLUSIONS AND RELEVANCE

The deep learning-based automatic pipeline exhibited similar performance in surgical guide design for autogenous tooth transplantation compared to manual design by senior doctors, and it minimized time costs.

T2WI-based texture analysis predicts preoperative lymph node metastasis of rectal cancer

BACKGROUND

To prospectively develop and validate the T2WI texture analysis model based on a node-by-node comparison for improving the diagnostic accuracy of lymph node metastasis (LNM) in rectal cancer.

METHODS

A total of 381 histopathologically confirmed lymph nodes (LNs) were collected. LNs texture features were extracted from MRI-T2WI. Spearman's rank correlation coefficient and the least absolute shrinkage and selection operator were used for feature selection to construct the LN rad-score. Then the clinical risk factors and LN texture features were combined to establish combined predictive model. Model performance was assessed by the area under the receiver operating characteristic (ROC) curve (AUC). Decision curve analysis (DCA) and nomogram were used to evaluate the clinical application of the model.

RESULTS

A total of 107 texture features were extracted from LN-MRI images. After selection and dimensionality reduction, the radiomics prediction model consisting of 8 texture features showed well-predictive performance in the training and validation cohorts (AUC, 0.676; 95% CI 0.582-0.771) (AUC, 0.774; 95% CI 0.648-0.899). A clinical-radiomics prediction model with the best performance was created by combining clinical and radiomics features, 0.818 (95% CI 0.742-0.893) for the training and 0.922 (95% CI 0.863-0.980) for the validation cohort. The LN Rad-score in clinical-radiomics nomogram obtained the highest classification contribution and was well calibrated. DCA demonstrated the superiority of the clinical-radiomics model.

CONCLUSION

The lymph node T2WI-based texture features can help to improve the preoperative prediction of LNM.

Application of deep learning and feature selection technique on external root resorption identification on CBCT images

BACKGROUND

Artificial intelligence has been proven to improve the identification of various maxillofacial lesions. The aim of the current study is two-fold: to assess the performance of four deep learning models (DLM) in external root resorption (ERR) identification and to assess the effect of combining feature selection technique (FST) with DLM on their ability in ERR identification.

METHODS

External root resorption was simulated on 88 extracted premolar teeth using tungsten bur in different depths (0.5 mm, 1 mm, and 2 mm). All teeth were scanned using a Cone beam CT (Carestream Dental, Atlanta, GA). Afterward, a training (70%), validation (10%), and test (20%) dataset were established. The performance of four DLMs including Random Forest (RF) + Visual Geometry Group 16 (VGG), RF + EfficienNetB4 (EFNET), Support Vector Machine (SVM) + VGG, and SVM + EFNET) and four hybrid models (DLM + FST: (i) FS + RF + VGG, (ii) FS + RF + EFNET, (iii) FS + SVM + VGG and (iv) FS + SVM + EFNET) was compared. Five performance parameters were assessed: classification accuracy, F1-score, precision, specificity, and error rate. FST algorithms (Boruta and Recursive Feature Selection) were combined with the DLMs to assess their performance.

RESULTS

RF + VGG exhibited the highest performance in identifying ERR, followed by the other tested models. Similarly, FST combined with RF + VGG outperformed other models with classification accuracy, F1-score, precision, and specificity of 81.9%, weighted accuracy of 83%, and area under the curve (AUC) of 96%. Kruskal Wallis test revealed a significant difference (p = 0.008) in the prediction accuracy among the eight DLMs.

CONCLUSION

In general, all DLMs have similar performance on ERR identification. However, the performance can be improved by combining FST with DLMs.

"Air Sign" in Misdiagnosed Mandibular Fractures Based on CT and CBCT Evaluation

BACKGROUND

Diagnostic errors constitute one of the reasons for the improper and often delayed treatment of mandibular fractures. The aim of this study was to present a series of cases involving undiagnosed concomitant secondary fractures in the mandibular body during preoperative diagnostics. Additionally, this study aimed to describe the "air sign" as an indirect indicator of a mandibular body fracture.

METHODS

A retrospective analysis of CT/CBCT scans conducted before surgery was performed on patients misdiagnosed with a mandibular body fracture within a one-year period.

RESULTS

Among the 75 patients who underwent surgical treatment for mandibular fractures, mandibular body fractures were missed in 3 cases (4%) before surgery. The analysis of CT/CBCT before surgery revealed the presence of an air collection, termed the "air sign", in the soft tissue adjacent to each misdiagnosed fracture of the mandibular body.

CONCLUSIONS

The "air sign" in a CT/CBCT scan may serve as an additional indirect indication of a fracture in the mandibular body. Its presence should prompt the surgeon to conduct a more thorough clinical examination of the patient under general anesthesia after completing the ORIF procedure in order to rule-out additional fractures.

Evaluating a Periapical Lesion Detection CNN on a Clinically Representative CBCT Dataset-A Validation Study

The aim of this validation study was to comprehensively evaluate the performance and generalization capability of a deep learning-based periapical lesion detection algorithm on a clinically representative cone-beam computed tomography (CBCT) dataset and test for non-inferiority. The evaluation involved 195 CBCT images of adult upper and lower jaws, where sensitivity and specificity metrics were calculated for all teeth, stratified by jaw, and stratified by tooth type. Furthermore, each lesion was assigned a periapical index score based on its size to enable a score-based evaluation. Non-inferiority tests were conducted with proportions of 90% for sensitivity and 82% for specificity. The algorithm achieved an overall sensitivity of 86.7% and a specificity of 84.3%. The non-inferiority test indicated the rejection of the null hypothesis for specificity but not for sensitivity. However, when excluding lesions with a periapical index score of one (i.e., very small lesions), the sensitivity improved to 90.4%. Despite the challenges posed by the dataset, the algorithm demonstrated promising results. Nevertheless, further improvements are needed to enhance the algorithm's robustness, particularly in detecting very small lesions and the handling of artifacts and outliers commonly encountered in real-world clinical scenarios.

The Application of Deep Learning on CBCT in Dentistry

Cone beam computed tomography (CBCT) has become an essential tool in modern dentistry, allowing dentists to analyze the relationship between teeth and the surrounding tissues. However, traditional manual analysis can be time-consuming and its accuracy depends on the user's proficiency. To address these limitations, deep learning (DL) systems have been integrated into CBCT analysis to improve accuracy and efficiency. Numerous DL models have been developed for tasks such as automatic diagnosis, segmentation, classification of teeth, inferior alveolar nerve, bone, airway, and preoperative planning. All research articles summarized were from Pubmed, IEEE, Google Scholar, and Web of Science up to December 2022. Many studies have demonstrated that the application of deep learning technology in CBCT examination in dentistry has achieved significant progress, and its accuracy in radiology image analysis has reached the level of clinicians. However, in some fields, its accuracy still needs to be improved. Furthermore, ethical issues and CBCT device differences may prohibit its extensive use. DL models have the potential to be used clinically as medical decision-making aids. The combination of DL and CBCT can highly reduce the workload of image reading. This review provides an up-to-date overview of the current applications of DL on CBCT images in dentistry, highlighting its potential and suggesting directions for future research.