Research on multi-class orthodontic image recognition system based on deep learning network model / 中华口腔医学杂志

Objective: To develop a multi-classification orthodontic image recognition system using the SqueezeNet deep learning model for automatic classification of orthodontic image data. Methods: A total of 35 000 clinical orthodontic images were collected in the Department of Orthodontics, Capital Medical University School of Stomatology, from October to November 2020 and June to July 2021. The images were from 490 orthodontic patients with a male-to-female ratio of 49∶51 and the age range of 4 to 45 years. After data cleaning based on inclusion and exclusion criteria, the final image dataset included 17 453 face images (frontal, smiling, 90° right, 90° left, 45° right, and 45° left), 8 026 intraoral images [frontal occlusion, right occlusion, left occlusion, upper occlusal view (original and flipped), lower occlusal view (original and flipped) and coverage of occlusal relationship], 4 115 X-ray images [lateral skull X-ray from the left side, lateral skull X-ray from the right side, frontal skull X-ray, cone-beam CT (CBCT), and wrist bone X-ray] and 684 other non-orthodontic images. A labeling team composed of orthodontic doctoral students, associate professors, and professors used image labeling tools to classify the orthodontic images into 20 categories, including 6 face image categories, 8 intraoral image categories, 5 X-ray image categories, and other images. The data for each label were randomly divided into training, validation, and testing sets in an 8∶1∶1 ratio using the random function in the Python programming language. The improved SqueezeNet deep learning model was used for training, and 13 000 natural images from the ImageNet open-source dataset were used as additional non-orthodontic images for algorithm optimization of anomaly data processing. A multi-classification orthodontic image recognition system based on deep learning models was constructed. The accuracy of the orthodontic image classification was evaluated using precision, recall, F1 score, and confusion matrix based on the prediction results of the test set. The reliability of the model's image classification judgment logic was verified using the gradient-weighted class activation mapping (Grad-CAM) method to generate heat maps. Results: After data cleaning and labeling, a total of 30 278 orthodontic images were included in the dataset. The test set classification results showed that the precision, recall, and F1 scores of most classification labels were 100%, with only 5 misclassified images out of 3 047, resulting in a system accuracy of 99.84%(3 042/3 047). The precision of anomaly data processing was 100% (10 500/10 500). The heat map showed that the judgment basis of the SqueezeNet deep learning model in the image classification process was basically consistent with that of humans. Conclusions: This study developed a multi-classification orthodontic image recognition system for automatic classification of 20 types of orthodontic images based on the improved SqueezeNet deep learning model. The system exhibitted good accuracy in orthodontic image classification.

Automated diagnostic classification with lateral cephalograms based on deep learning network model / 中华口腔医学杂志

Objective: To establish a comprehensive diagnostic classification model of lateral cephalograms based on artificial intelligence (AI) to provide reference for orthodontic diagnosis. Methods: A total of 2 894 lateral cephalograms were collected in Department of Orthodontics, Capital Medical University School of Stomatology from January 2015 to December 2021 to construct a data set, including 1 351 males and 1 543 females with a mean age of (26.4± 7.4) years. Firstly, 2 orthodontists (with 5 and 8 years of orthodontic experience, respectively) performed manual annotation and calculated measurement for primary classification, and then 2 senior orthodontists (with more than 20 years of orthodontic experience) verified the 8 diagnostic classifications including skeletal and dental indices. The data were randomly divided into training, validation, and test sets in the ratio of 7∶2∶1. The open source DenseNet121 was used to construct the model. The performance of the model was evaluated by classification accuracy, precision rate, sensitivity, specificity and area under the curve (AUC). Visualization of model regions of interest through class activation heatmaps. Results: The automatic classification model of lateral cephalograms was successfully established. It took 0.012 s on average to make 8 diagnoses on a lateral cephalogram. The accuracy of 5 classifications was 80%-90%, including sagittal and vertical skeletal facial pattern, mandibular growth, inclination of upper incisors, and protrusion of lower incisors. The acuracy rate of 3 classifications was 70%-80%, including maxillary growth, inclination of lower incisors and protrusion of upper incisors. The average AUC of each classification was ≥0.90. The class activation heat map of successfully classified lateral cephalograms showed that the AI model activation regions were distributed in the relevant structural regions. Conclusions: In this study, an automatic classification model for lateral cephalograms was established based on the DenseNet121 to achieve rapid classification of eight commonly used clinical diagnostic items.

The Experimental Study of Cyasterone on the Treatment of Osteoporosis through the Bidirectional Effect of Inhibiting Osteoclast Differetiation and Promoting Osteoblast Differentiation / 昆明医科大学学报

Objective To explore the influence of cyasterone on the osteoclast and osteoblast differentiation and then to investigate its effect on the bone quality in the osteoporosis mice. Methods CCK8 assay was firstly used to detect the toxic effect of cyasterone on the mouse bone marrow derived mononuclear macrophages (BMMs) and anterior osteoblast lines MC3T3E1. Cell apoptosis was measured by flow cytometry. Then TRAP staining and ALP staining were employed to detect osteoclast differentiation and osteoblast differentiation, respectively. Realtime PCR was carried out to test the expression of osteoclast special gene TRAP and osteogenesis crucial gene ALP. In vivo, 15 mice were divided into three groups: sham-operated group, OVX group and OVX+cyasterone treatment group. In treatment group, cyasterone was used as 5mg/kg every day. Sham-operated group and OVX group were treat with saline solution. After 4 weeks, the tibia was collected for Micro-CT detection to observe the bone quality and microstructure changes. Results Cyasterone with the concentration of less than 10 mg/L had no significant cytotoxicity nor influence on the apoptosis (P＞0.05) . Cyasterone could significantly inhibit the osteoclast differentiation of BMMs (P＜0.05), simultaneously, it also had the effect to promote the osteoblast differetiation of MC3T3E1. Real-time PCR indicated that cyasterone could block the expression of TRAP and increase the expression of ALP (P＜0.05) . In vivo, cyasterone was able to obviously improve the osteoporosis status caused by estrogen deficiency without general toxicity. Conclusion cyasterone could provide a good treatment for osteoporosis through the bidirectional effect of inhibiting osteoclast differetiation and promoting osteoblast differentiation.

Antitumor effect of sphingosine kinase 1 inhibitor in combination with chemotherapy on SGC7901 gastric cancer cells in vitro / 中华肿瘤杂志

<p><b>OBJECTIVE</b>To study the effect of the sphingosine kinase 1 (SphK1) inhibitor N,N-dimethylsphingosine (DMS) in combination with chemotherapeutic drugs (DDP, 5-Fu, MMC) on the proliferation of gastric cancer cells (SGC7901) in vitro, and to evaluate whether SphK1 inhibitors could be used as synergetic agents in chemotherapy.</p><p><b>METHODS</b>SGC7901 cells were incubated in vitro with DMS (1 micromol/L) and 5-Fu, DDP, MMC at different concentrations in combination or separately for 24 h. The effects on the growth and survival of SGC7901 cells were determined by MTT assay. The inhibition rates were assessed by response surface analysis and the interactive relationships between the combined drugs were evaluated on the basis of positive/negative values of the cross product coefficients in the response surface equation.</p><p><b>RESULTS</b>The growth inhibition rate of the gastric cancer cells by treatment with DMS (1 micromol/L) was (10.23 +/- 0.74)%. The growth inhibition rates of the gastric cancer cells treated with 5-Fu (1, 5 and 25 microg/ml) for 24 h were (9.95 +/- 3.24)%, (21.04 +/- 2.19)%, and (45.49 +/- 3.60)%, respectively. The growth inhibition rates of the gastric cancer cells treated with DDP (0.5, 2.5 and 12.5 microg/ml) for 24 h were (9.38 +/- 0.79)%, (19.61 +/- 0.90)%, and (29.83 +/- 0.54)%, respectively. The growth inhibition rates of the gastric cancer cells treated with MMC (0.1, 0.5 and 2.5 microg/ml) for 24 h were (15.35 +/- 0.77)%, (24.72 +/- 0.83)%, and (30.68 +/- 0.28)%, respectively. There were significant differences among the inhibition rates caused by different concentrations of the drugs (P < 0.05). When 1 micromol/L DMS was used in combination with 5-Fu (1, 5, and 25 microg/ml) for 24 h, the growth inhibition rates of the cancer cells were (16.76 +/- 0.41)%, (27.28 +/- 0.29)% and (52.56 +/- 3.60)%, respectively. When 1 micromol/L DMS was used in combination with DDP (0.5, 2.5, and 12.5 microg/ml) for 24 h, the growth inhibition rates of the cancer cells were (15.35 +/- 0.86)%, (25.57 +/- 0.27)%, (36.37 +/- 0.51)%, respectively. When 1 micromol/L DMS was used in combination with MMC (0.1, 0.5, and 2.5 microg/ml) for 24 h, the growth inhibition rates of the cancer cells were (21.02 +/- 0.28)%, (32.10 +/- 0.27)%, (36.36 +/- 0.28)%, respectively. There were also significant differences among the growth inhibition rates caused by different concentrations of the drugs alone and in combination groups (P < 0.05).</p><p><b>CONCLUSIONS</b>DMS can suppress the proliferation of SGC7901 cells in vitro, and there are evident synergetic effects when it is used in combination with chemotherapeutic drugs. The results of this study indicate that SphK1 inhibitors may become novel and promising chemotherapeutic sensitizers.</p>

Influence of different NaCl concentration on passive behavior of Cr alloy in vitro / 中华口腔医学杂志

<p><b>OBJECTIVE</b>To investigate the properties of passive film of three Cr alloys and to analyze their corrosion resistance in the artificial saliva with different NaCl mass fraction in vitro.</p><p><b>METHODS</b>Artificial saliva with 1%, 2% and 3% NaCl mass fraction was prepared. Cobalt-chromium alloy, nickel-chromium alloy and Ti-Ni-Cr alloy were employed as the working electrode in the artificial saliva. Semi-conductor properties of passive film on Cr alloy was analyzed by Mott-Schottky plots. In addition, the factors which affect the semi-conductive compact characteristic of the passive film was also discussed.</p><p><b>RESULTS</b>The passive film exhibits n-type semi-conductive characters. With the increasing of NaCl mass fraction, the carrier density of the Cr alloys also increased. The carrier density of the cobalt-chromium alloy, nickel-chromium alloy and Ti-Ni-Cr alloy are 3.71 x 10⁸, 2.34 x 10⁹, and 7.66 x 10⁹/cm³ respectively. This decreases its donor density and its film stability.</p><p><b>CONCLUSIONS</b>When exposed to saliva environment with high concentration of chlorine ion, corrosion resistance of the three types of Cr alloys decrease. This will reduce the service life of Cr alloy prosthesis.</p>