Deep Learning (DL) and Artificial Intelligence (AI) tools have shown great success in different areas of medical diagnostics. In this paper, we show another success in orthodontics. In orthodontics, the right treatment timing of many actions and operations is crucial because many environmental and genetic conditions may modify jaw growth. The stage of growth is related to the Cervical Vertebra Maturation (CVM) degree. Thus, determining the CVM to determine the suitable timing of the treatment is important. In orthodontics, lateral X-ray radiography is used to determine it. Many classical methods need knowledge and time to look and identify some features. Nowadays, ML and AI tools are used for many medical and biological diagnostic imaging. This paper reports on the development of a Deep Learning (DL) Convolutional Neural Network (CNN) method to determine (directly from images) the degree of maturation of CVM classified in six degrees. The results show the performances of the proposed method in different contexts with different number of images for training, evaluation and testing and different pre-processing of these images. The proposed model and method are validated by cross validation. The implemented software is almost ready for use by orthodontists.
Many environmental and genetic conditions may modify jaws growth. In orthodontics, the right treatment timing is crucial. This timing is a function of the Cervical Vertebra Maturation (CVM) degree. Thus, determining the CVM is important. In orthodontics, the lateral X-ray radiography is used to determine it. Many classical methods need knowledge and time to look and identify some features to do it. Nowadays, Machine Learning (ML) and Artificial Intelligent (AI) tools are used for many medical and biological image processing, clustering and classification. This paper reports on the development of a Deep Learning (DL) method to determine directly from the images the degree of maturation of CVM classified in six degrees. Using 300 such images for training and 200 for evaluating and 100 for testing, we could obtain a 90% accuracy. The proposed model and method are validated by cross validation. The implemented software is ready for use by orthodontists.
In the field of orthodontics, digital dental arch models are increasingly replacing plaster models in orthodontic diagnostics. This change in interface from physical plaster models to digital image-based models raises the question of how orthodontists interpret intra- and inter-arch relationships from an image on a screen. In particular, the issue of the accuracy of the distances estimated on numerical models is crucial since the estimation of distances is the basis of the therapeutic decision-making process. Studies have shown that distances were well estimated on plaster models, but no study has verified this point on numerical models. This is the question that our study addresses. The experimental protocol consisted of collecting estimates of measurements made by orthodontists using digital models. The reliability of these measurements was then assessed by comparing them to the actual physical distances. We asked 31 orthodontists (19 women and 12 men; an average age of 37 years) to generate 3D model-based measurements of seven different elements: mandibular congestion, the maxillary intermolar distance, Spee’s curve, 16/26 symmetry, the right canine class, overbite, and overjet. These values were then compared to the actual measurements calculated using Insignia® software (ORMCO Corporation: Brea, CA, USA), using single sample t-tests. This test makes it possible to compare a distance estimated by the participants with a reference value, which corresponds here to the real distance. The results indicate that, overall, the distance estimates made on the 3D models differ significantly from the actual distances measured using the Insignia® software. This was particularly so for mandibular crowding (test value = 0; t (30) = 10.74; p ≤ 0.01), test value = 1; t (30) = 6.23; p ≤ 0.01). Although no study has focused on distance estimation on numerical models in the field of orthodontics, our results agree with the conclusions of studies showing that distances are not estimated in the same way in real environments and virtual environments. Additional studies will make it possible to identify more clearly the parameters (individual factors, equipment, etc.), which make it possible to improve the estimation of distances in the practice of orthodontics. In any case, these studies are necessary to improve the training of future practitioners in the use of virtual models for decision-making and to support them in the digital transition.
Mandibular retrognathia (C2Rm) is one of the most common oral pathologies. Acquiring a better understanding of the points of impact of C2Rm on the entire skull is of major interest in the diagnosis, treatment, and management of this dysmorphism, but also permits us to contribute to the debate on the changes undergone by the shape of the skull during human evolution. However, conventional methods have some limits in meeting these challenges, insofar as they require defining in advance the structures to be studied, and identifying them using landmarks. In this context, our work aims to answer these questions using AI tools and, in particular, machine learning, with the objective of relaying these treatments automatically. We propose an innovative methodology coupling convolutional neural networks (CNNs) and interpretability algorithms. Applied to a set of radiographs classified into physiological versus pathological categories, our methodology made it possible to: discuss the structures impacted by retrognathia and already identified in literature; identify new structures of potential interest in medical terms; highlight the dynamic evolution of impacted structures according to the level of gravity of C2Rm; provide for insights into the evolution of human anatomy. Results were discussed in terms of the major interest of this approach in the field of orthodontics and, more generally, in the field of automated processing of medical images.
Orthodontists have seen their practices evolve from estimating distances on plaster models to estimating distances on non-immersive virtual models. However, if the estimation of distance using real models can generate errors (compared to the real distance measured using tools), which remains acceptable from a clinical point of view, is this also the case for distance estimation performed on digital models? To answer this question, 50 orthodontists (31 women and 19 men) with an average age of 36 years (σ = 12.84; min = 23; max = 63) participated in an experiment consisting of estimating 3 types of distances (mandibular crowding, inter-canine distance, and inter-molar distance) on 6 dental models, including 3 real and 3 virtual models. Moreover, these models were of three different levels of complexity (easy, medium, and difficult). The results showed that, overall, the distances were overestimated (compared to the distance measured using an instrument) regardless of the situation (estimates on real or virtual models), but this overestimation was greater for the virtual models than for the real models. In addition, the mental load associated with the estimation tasks was considered by practitioners to be greater for the estimation tasks performed virtually compared to the same tasks performed on plaster models. Finally, when the estimation task was more complex, the number of estimation errors decreased in both the real and virtual situations, which could be related to the greater number of therapeutic issues associated with more complex models.
Accurate determination of skeletal maturation indicators is crucial in the orthodontic process. Chronologic age is not a reliable skeletal maturation indicator, thus physicians use bone age. In orthodontics, the treatment timing depends on Cervical Vertebral Maturation (CVM) assessment. Determination of CVM degree remains challenging due to the limited annotated dataset, the existence of significant irrelevant areas in the image, the huge intra-class variances, and the high degree of inter-class similarities. To address this problem, researchers have started looking for external information beyond current available medical datasets. This work utilizes the domain knowledge from radiologists to train neural network models that can be utilized as a decision support system. We proposed a novel supervised learning method with a multi-scale attention mechanism, and we incorporated the general diagnostic patterns of medical doctors to classify lateral X-ray images as six CVM classes. The proposed network highlights the important regions, surpasses the irrelevant part of the image, and efficiently models long-range dependencies. Employing the attention mechanism improves both the performance and interpretability. In this work, we used additive spatial and channel attention modules. Our proposed network consists of three branches. The first branch extracts local features, and creates attention maps and related masks, the second branch uses the masks to extract discriminative features for classification, and the third branch fuses local and global features. The result shows that the proposed method can represent more discriminative features, therefore, the accuracy of image classification is greater in comparison to in backbone and some attention-based state-of-the-art networks.