ObjectiveTo explore the effects of the Tai Chi training on bone density， bone turnover markers， and heart rate variability for people with high-risk osteoporosis， and to provide evidence for the prevention of osteoporosis at early stage. MethodsSixty-six cases of people with high risk of osteoporosis were included， and they were divided into 33 cases each in the intervention group and the control group using the random number table method. The control group received osteoporosis health education three times a week， and the intervention group received Tai Chi training under the guidance of a trainer three times a week for 40 mins each time on the basis of the control group， and both groups were intervened for 12 weeks. Dual-energy X-ray absorptiometry was used to measure the bone density of L₁~L₄ vertebrae， bilateral femoral necks and bilateral total hips in the two groups before and after the intervention； enzyme-linked immunosorbent assay was used to determine bone turnover markers before and after the intervention， including pro-collagen type Ⅰ pro-amino-terminal prepropyl peptide （P1NP） and β-collagen type Ⅰ cross-linking carboxy-terminal peptide （β-CTX）. Seven cases with good compliance in the intervention group were selected. After wearing the heart rate sensor， they successively performed Tai Chi training and walking activities recommended by the guideline for 20 mins each， and the heart rate variability （HRV） during exercise was collected， including time-domain indexes such as standard deviation of normal sinus intervals （SDNN）， root-mean-square of the difference between adjacent RR intervals （RMSSD）， frequency-domain metrics such as low-frequency power （LF）， high-frequency power （HF）， and low-frequency/high-frequency power ratio （LF/HF）， as well as nonlinear metrics such as approximate entropy （ApEn）， sample entropy （SampEn）. ResultsFinally， 63 cases were included in the outcome analysis， including 30 cases in the intervention group and 33 cases in the control group. After the intervention， the differences of L₁~L₄ vertebrae， bone density of bilateral femoral neck and bilateral total hip in the intervention group were not statistically significant when compared with those before intervention （P＞0.05）， while the bone density of all parts of the control group decreased significantly compared with that before intervention （P＜0.05）， and the difference in the bone density of the L₁~L₄ vertebrae， bilateral femoral neck， and the right total hip before and after the intervention of the intervention group was smaller than that of the control group （P＜0.05）. The differences in P1NP and β-CTX between groups before and after intervention was not statistically significant （P＞0.05）. Compared with walking exercise， LF decreased， HF increased and LF/HF decreased during Tai Chi exercise （P＜0.05）； the time domain indexes and non-linear indexes between groups had no statistically significant difference （P＞0.05）. ConclusionTai Chi exercise can maintain lumbar， hip， and femoral bone density and improve sympathetic/parasympathetic balance in people at high risk for osteoporosis， but cannot significantly improve bone turnover markers.

ObjectiveTo investigate the differences in clinical efficacy of different medication regimens of Wangbi Tablets （尪痹片） for knee osteoarthritis （KOA） in a real-world setting， providing a basis for rational clinical use of Wangbi Tablets. MethodsA prospective registry study was conducted， involving 2，999 KOA patients registered in 30 hospitals nationwide from January 26th， 2019， to December 17th， 2021. Based on the use of Wangbi Tablets during the observation period， patients were divided into a monotherapy group （1，507 cases） and a combination therapy group （1，492 cases）， and the combination group can be further divided into Wangbi Tablets plus Chinese medicine （CM）， Wangbi Tablets plus western medicine （WM）， and Wangbi Tablets plus Chinese and western medicine （CM+WM） subgroups. The baseline data of patients in the monotherapy group and the combination group were compared， including age， gender， body weight， medication time， clinical stage， K-L grade， and others. Efficacy indicators included the Visual Analog Scale （VAS） score， Western Ontario and McMaster Universities Osteoarthritis Index （WOMAC） score， and EuroQol five-dimensional （EQ-5D） health index， which were evaluated before and after 4-， 8- and 12-week treatment， and the difference before and after treatment was calculated after 4， 8 and 12 weeks of treatment. The difference between the baseline and 12 weeks of treatment of all the above indicators was used as the dependent variables， and gender， age， body mass index （BMI）， course of disease， K-L grade， and clinical stage were used as independent variables， when multiple linear regression was taken to explore the influencing factors of the efficacy. At the same time， the occurrence of major symptoms （including morning stiffness， joint swelling， soreness of waist and knees， fear of wind， and fear of cold） was counted， and the disappearance of symptoms at each time point was counted after 4， 8， and 12 weeks of treatment. ResultsAt baseline， there were no statistically significant differences in gender and age distribution between the monotherapy and combination therapy groups （P＞0.05）； the proportion of patients in the acute stage and recovery stage was higher in the monotherapy group than in the combination therapy group， while the proportion in the remission stage was lower （P＜0.05）； the VAS score was higher in the monotherapy group， and the EQ-5D index was lower （P＜0.01）， with no statistically significant difference in total WOMAC score between the two groups （P＞0.05）. Compared to those measured before treatment and at previous timepoint， the VAS score and WOMAC total score significantly decreased in both groups， while EQ-5D value increased （P＜0.05）. The difference in VAS score between baseline and after 12-week treatment was higher in the monotherapy group than the combination group， while the differences in WOMAC total score and EQ-5D value between baseline and after 4-， 8- and 12-week treatment were higher in the combination group （P＜0.05）. Multiple linear regression showed that VAS score before treatment had greatest impact on pain improvement （P＜0.01）， and compared to Wangbi Tablets monotherapy， the combination of Wangbi tablets with WM or CM had larger associations with pain improvement （P＜0.05）； and Wangbi Tablets had better efficacy when the course of treatment was ＞28 days （P＜0.01）. Wangbi Tablets plus WM had a better effect on improving the overall function of the knee joint than Wangbi Tablets alone （P＜0.01）； and the efficacy of Wangbi Tablets with a course of treatment ＞28 days was better （P＜0.05）. The improvement of quality of life of patients in the attack and remission stages was more obvious than that in the recovery stage （P＜0.01）； Wangbi Tablets plus WM or CM had a better effect on improving quality of life than Wangbi Tablets alone （P＜0.05）. Before treatment， the proportion of patients with morning stiffness， soreness of waist and knees， fear of wind and chills in the monotherapy group was higher than that in the combination group （P＜0.01）. The proportion of main symptoms in both groups decreased after 4， 8 and 12 weeks of treatment （P＜0.05）. After 4 weeks of treatment， the disappearance rate of each main symptom in the combination group was higher than that in the monotherapy group， and after 12 weeks of treatment， the disappearance rate of fear of wind in the monotherapy group was higher than that in the combination group， while the disappearance rate of joint swelling and soreness of waist and knees was lower （P＜0.05）. ConclusionWangbi Tablets， whether used alone or in combination with other medications， is effective throughout the course of KOA， with greater benefits in improving joint function and quality of life during the acute and remission stages compared to the recovery stage. Combination therapy had a faster onset of effect， but began to converge with monotherapy after 8 weeks. The best efficacy was observed with the combination of Wangbi Tablets with WM， followed by combination with CM.

Objective:To explore the effectiveness of a hysteroscopic endometrial lesion diagnosis model de-veloped based on deep learning(DL)algorithm combined with gradient-weighted class activation mapping(Grad-CAM)visualization technology.Methods:303 hysteroscopy videos(4781 images)of 291 patients who un-derwent hysteroscopy examination in the Department of Gynecology,Renmin Hospital of Wuhan University from June 1,2021 to December 31,2022 were selected.The dataset was divided into a training set(3703 images)and a test set(1078 images)by weight sampling method.After the training set was used for model learning and train-ing,two model architectures,residual neural network(ResNet18)and efficient neural network(EfficientNet-B0),were selected to verify the model in the test set by five-class and two-class classification tasks,respectively.Tak-ing histopathology as the gold standard,the diagnostic efficacy was evaluated to select the optimal model,and the Grad-CAM layer was embedded in the optimal model to output hysteroscopy images of Grad-CAM.Results:①In the five-class classification tasks,the accuracy of EfficientNet-B0 model(93.23％)was higher than that of Res-Net18 model(84.23％);the area under the curve(AUC)of EfficientNet-B0 model in the diagnosis of five disea-ses,including atypical endometrial hyperplasia,endometrial polyps,endometrial cancer,endometrial atypical hy-perplasia,and submucous myoma,was slightly higher than that of ResNet18 model,and the AUC of both models was almost above 0.980.②In the binary classification task of accuracy and the evaluation of specificity,the two models were similar,both above 93.00％,and the sensitivity of EfficientNet-B0 model(91.14％)was significantly better than that of ResNet18 model(77.22％).③EfficientNet-B0 model combined with Grad-CAM algorithm could identify the abnormal areas in the image.After biopsy and pathological examination,it was confirmed that about 95％of the marked areas in the model's output heatmap were lesion areas.Conclusions:The hysteroscopy di-agnostic model developed by EfficientNet-B0 model combined with Grad-CAM has high diagnostic accuracy,sen-sitivity,and specificity,and has application value in the diagnosis of endometrial lesions.

Objective:To explore the effectiveness of a hysteroscopic endometrial lesion diagnosis model de-veloped based on deep learning(DL)algorithm combined with gradient-weighted class activation mapping(Grad-CAM)visualization technology.Methods:303 hysteroscopy videos(4781 images)of 291 patients who un-derwent hysteroscopy examination in the Department of Gynecology,Renmin Hospital of Wuhan University from June 1,2021 to December 31,2022 were selected.The dataset was divided into a training set(3703 images)and a test set(1078 images)by weight sampling method.After the training set was used for model learning and train-ing,two model architectures,residual neural network(ResNet18)and efficient neural network(EfficientNet-B0),were selected to verify the model in the test set by five-class and two-class classification tasks,respectively.Tak-ing histopathology as the gold standard,the diagnostic efficacy was evaluated to select the optimal model,and the Grad-CAM layer was embedded in the optimal model to output hysteroscopy images of Grad-CAM.Results:①In the five-class classification tasks,the accuracy of EfficientNet-B0 model(93.23％)was higher than that of Res-Net18 model(84.23％);the area under the curve(AUC)of EfficientNet-B0 model in the diagnosis of five disea-ses,including atypical endometrial hyperplasia,endometrial polyps,endometrial cancer,endometrial atypical hy-perplasia,and submucous myoma,was slightly higher than that of ResNet18 model,and the AUC of both models was almost above 0.980.②In the binary classification task of accuracy and the evaluation of specificity,the two models were similar,both above 93.00％,and the sensitivity of EfficientNet-B0 model(91.14％)was significantly better than that of ResNet18 model(77.22％).③EfficientNet-B0 model combined with Grad-CAM algorithm could identify the abnormal areas in the image.After biopsy and pathological examination,it was confirmed that about 95％of the marked areas in the model's output heatmap were lesion areas.Conclusions:The hysteroscopy di-agnostic model developed by EfficientNet-B0 model combined with Grad-CAM has high diagnostic accuracy,sen-sitivity,and specificity,and has application value in the diagnosis of endometrial lesions.

Objective To analyze the characteristics of pattern elements and the distribution of patterns of femoral head necrosis, hip osteoarthritis, and hip rheumatoid arthritis, in order to provide a theoretical basis for the "different diseases with the same pattern" of chronic bone diseases. Methods A cross-sectional survey was conducted to select patients with femoral head necrosis Association Research Circulation Osseous Ⅰ-Ⅱ stages, hip osteoarthritis Kellgren & Lawrence Ⅰ-Ⅱ stages, and acute or subacute hip rheumatoid arthritis who visited the Minimally Invasive Arthrology Department, Traumatology Department, and Rheumatology and Immunology Department of Beijing University of Chinese Medicine Third Affiliated Hospital from June 2020 to June 2022. The " case report form - traditional Chinese medicine pattern manifestation scale" previously developed by our team was used to collect the pattern manifestations, which were included into an Excel 2020 spreadsheet to establish a database. SPSS 20.0 software was used for factor analysis and cluster analysis to extract pattern element information, such as disease nature and location, in order to summarize the characteristics of pattern elements, the distribution of patterns, and the similarities and differences of the three kinds of early hip joint diseases.Results A total of 410 patients were included, including 150 patients with femoral head necrosis, 160 patients with hip osteoarthritis, and 100 patients with hip rheumatoid arthritis. The pattern elements of the disease nature of femoral head necrosis include phlegm (dampness), blood stasis, yang deficiency, essence deficiency, and qi deficiency. The pattern types were initially divided into four categories: syndrome of meridian obstruction (43.33％), syndrome of phlegm and blood stasis blocking collaterals (38.00％), syndrome of liver and kidney deficiency (12.00％), and syndrome of kidney essence deficiency (6.67％). The pattern elements of the disease nature of hip osteoarthritis include phlegm (dampness), blood stasis, qi deficiency, essence deficiency, yang deficiency, and cold (dampness). The pattern types were preliminarily divided into five categories: syndrome of spleen and kidney deficiency (37.50％), syndrome of meridian obstruction (26.87％), syndrome of cold and dampness obstruction (18.75％), syndrome of phlegm and blood stasis obstruction (9.38％), and syndrome of liver and kidney deficiency (7.50％). The pattern elements of the disease nature of hip rheumatoid arthritis include phlegm (dampness), blood stasis, qi deficiency, yin and yang deficiency, cold (dampness), and essence deficiency. The pattern types were preliminarily divided into four categories: syndrome of phlegm and blood stasis obstruction (34.00％), syndrome of cold and dampness obstruction (28.00％), syndrome of blood stasis blocking collaterals (23.00％), and syndrome of liver and kidney deficiency (15.00％). Overall, the top five pattern manifestations of the three kinds of hip joint diseases were hip joint pain (96.59％), tenderness (93.90％), fixed pain (87.56％), heavy joints (85.37％), and sourness of lower limbs (75.37％). The pattern elements of the disease nature include phlegm (dampness), blood stasis, qi deficiency, etc. The pattern types were preliminarily divided into five categories: syndrome of phlegm stasis blocking collaterals (33.17％), syndrome of meridian obstruction (31.95％), syndrome of cold dampness obstruction (21.46％), syndrome of liver and kidney deficiency (7.32％), and syndrome of spleen and kidney deficiency (6.10％). There were 21 similar pattern manifestations in the three kinds of early hip joint diseases, with blood stasis and spleen deficiency being the main pattern.Conclusion The common pattern characteristics of three kinds of early hip joint diseases are spleen deficiency and blood stasis. In addition, femoral head necrosis is accompanied with phlegm-dampness pattern, hip osteoarthritis is accompanied with kidney deficiency and phlegm-dampness pattern, hip rheumatoid arthritis is accompanied with kidney deficiency and cold-dampness pattern.

Objective To develop an intelligent model for differential diagnosis of atrioventricular nodal re-entrant tachycardia(AVNRT)and atrioventricular re-entrant tachycardia(AVRT)using 12-lead wearable electrocardiogram devices.Methods A total of 356 samples of 12-lead supraventricular tachycardia(SVT)electrocardiograms recorded by wearable devices were randomly divided into training and validation sets using 5-fold cross validation to establish the intelligent classification model,and 101 patients with the diagnosis of SVT undergoing electrophysiological studies and radiofrequency ablation from October,2021 to March,2023 were selected as the testing set.The changes in electrocardiogram parameters before and during induced tachycardia were compared.Based on multiscale deep neural network,an intelligent diagnosis model for classifying SVT mechanisms was constructed and validated.The 3-lead electrocardiogram signals from Ⅱ,Ⅲ,and V1 were extracted to build new classification models,whose diagnostic efficacy was compared with that of the 12-lead model.Results Of the 101 patients with SVT in the testing set,68 were diagnosed with AVNRT and 33 were diagnosed with AVRT by electrophysiological study.The pre-trained model achieved a high area under the precision-recall curve(0.9492)and F1 score(0.8195)for identifying AVNRT in the validation set.The total F1 scores of the lead Ⅱ,Ⅲ,V1,3-lead and 12-lead intelligent diagnostic models in the testing set were 0.5597,0.6061,0.3419,0.6003 and 0.6136,respectively.Compared with the 12-lead classification model,the lead-Ⅲ model had a net reclassification index improvement of-0.029(P=0.878)and an integrated discrimination index improvement of-0.005(P=0.965).Conclusion The intelligent diagnostic model based on multiscale deep neural network using wearable electrocardiogram devices has an acceptable accuracy for classifying SVT mechanisms.

Objective:To explore the effectiveness of a hysteroscopic endometrial lesion diagnosis model de-veloped based on deep learning(DL)algorithm combined with gradient-weighted class activation mapping(Grad-CAM)visualization technology.Methods:303 hysteroscopy videos(4781 images)of 291 patients who un-derwent hysteroscopy examination in the Department of Gynecology,Renmin Hospital of Wuhan University from June 1,2021 to December 31,2022 were selected.The dataset was divided into a training set(3703 images)and a test set(1078 images)by weight sampling method.After the training set was used for model learning and train-ing,two model architectures,residual neural network(ResNet18)and efficient neural network(EfficientNet-B0),were selected to verify the model in the test set by five-class and two-class classification tasks,respectively.Tak-ing histopathology as the gold standard,the diagnostic efficacy was evaluated to select the optimal model,and the Grad-CAM layer was embedded in the optimal model to output hysteroscopy images of Grad-CAM.Results:①In the five-class classification tasks,the accuracy of EfficientNet-B0 model(93.23％)was higher than that of Res-Net18 model(84.23％);the area under the curve(AUC)of EfficientNet-B0 model in the diagnosis of five disea-ses,including atypical endometrial hyperplasia,endometrial polyps,endometrial cancer,endometrial atypical hy-perplasia,and submucous myoma,was slightly higher than that of ResNet18 model,and the AUC of both models was almost above 0.980.②In the binary classification task of accuracy and the evaluation of specificity,the two models were similar,both above 93.00％,and the sensitivity of EfficientNet-B0 model(91.14％)was significantly better than that of ResNet18 model(77.22％).③EfficientNet-B0 model combined with Grad-CAM algorithm could identify the abnormal areas in the image.After biopsy and pathological examination,it was confirmed that about 95％of the marked areas in the model's output heatmap were lesion areas.Conclusions:The hysteroscopy di-agnostic model developed by EfficientNet-B0 model combined with Grad-CAM has high diagnostic accuracy,sen-sitivity,and specificity,and has application value in the diagnosis of endometrial lesions.

Objective:To explore the effectiveness of a hysteroscopic endometrial lesion diagnosis model de-veloped based on deep learning(DL)algorithm combined with gradient-weighted class activation mapping(Grad-CAM)visualization technology.Methods:303 hysteroscopy videos(4781 images)of 291 patients who un-derwent hysteroscopy examination in the Department of Gynecology,Renmin Hospital of Wuhan University from June 1,2021 to December 31,2022 were selected.The dataset was divided into a training set(3703 images)and a test set(1078 images)by weight sampling method.After the training set was used for model learning and train-ing,two model architectures,residual neural network(ResNet18)and efficient neural network(EfficientNet-B0),were selected to verify the model in the test set by five-class and two-class classification tasks,respectively.Tak-ing histopathology as the gold standard,the diagnostic efficacy was evaluated to select the optimal model,and the Grad-CAM layer was embedded in the optimal model to output hysteroscopy images of Grad-CAM.Results:①In the five-class classification tasks,the accuracy of EfficientNet-B0 model(93.23％)was higher than that of Res-Net18 model(84.23％);the area under the curve(AUC)of EfficientNet-B0 model in the diagnosis of five disea-ses,including atypical endometrial hyperplasia,endometrial polyps,endometrial cancer,endometrial atypical hy-perplasia,and submucous myoma,was slightly higher than that of ResNet18 model,and the AUC of both models was almost above 0.980.②In the binary classification task of accuracy and the evaluation of specificity,the two models were similar,both above 93.00％,and the sensitivity of EfficientNet-B0 model(91.14％)was significantly better than that of ResNet18 model(77.22％).③EfficientNet-B0 model combined with Grad-CAM algorithm could identify the abnormal areas in the image.After biopsy and pathological examination,it was confirmed that about 95％of the marked areas in the model's output heatmap were lesion areas.Conclusions:The hysteroscopy di-agnostic model developed by EfficientNet-B0 model combined with Grad-CAM has high diagnostic accuracy,sen-sitivity,and specificity,and has application value in the diagnosis of endometrial lesions.

Objective:To explore the effectiveness of a hysteroscopic endometrial lesion diagnosis model de-veloped based on deep learning(DL)algorithm combined with gradient-weighted class activation mapping(Grad-CAM)visualization technology.Methods:303 hysteroscopy videos(4781 images)of 291 patients who un-derwent hysteroscopy examination in the Department of Gynecology,Renmin Hospital of Wuhan University from June 1,2021 to December 31,2022 were selected.The dataset was divided into a training set(3703 images)and a test set(1078 images)by weight sampling method.After the training set was used for model learning and train-ing,two model architectures,residual neural network(ResNet18)and efficient neural network(EfficientNet-B0),were selected to verify the model in the test set by five-class and two-class classification tasks,respectively.Tak-ing histopathology as the gold standard,the diagnostic efficacy was evaluated to select the optimal model,and the Grad-CAM layer was embedded in the optimal model to output hysteroscopy images of Grad-CAM.Results:①In the five-class classification tasks,the accuracy of EfficientNet-B0 model(93.23％)was higher than that of Res-Net18 model(84.23％);the area under the curve(AUC)of EfficientNet-B0 model in the diagnosis of five disea-ses,including atypical endometrial hyperplasia,endometrial polyps,endometrial cancer,endometrial atypical hy-perplasia,and submucous myoma,was slightly higher than that of ResNet18 model,and the AUC of both models was almost above 0.980.②In the binary classification task of accuracy and the evaluation of specificity,the two models were similar,both above 93.00％,and the sensitivity of EfficientNet-B0 model(91.14％)was significantly better than that of ResNet18 model(77.22％).③EfficientNet-B0 model combined with Grad-CAM algorithm could identify the abnormal areas in the image.After biopsy and pathological examination,it was confirmed that about 95％of the marked areas in the model's output heatmap were lesion areas.Conclusions:The hysteroscopy di-agnostic model developed by EfficientNet-B0 model combined with Grad-CAM has high diagnostic accuracy,sen-sitivity,and specificity,and has application value in the diagnosis of endometrial lesions.

Objective:To explore the effectiveness of a hysteroscopic endometrial lesion diagnosis model de-veloped based on deep learning(DL)algorithm combined with gradient-weighted class activation mapping(Grad-CAM)visualization technology.Methods:303 hysteroscopy videos(4781 images)of 291 patients who un-derwent hysteroscopy examination in the Department of Gynecology,Renmin Hospital of Wuhan University from June 1,2021 to December 31,2022 were selected.The dataset was divided into a training set(3703 images)and a test set(1078 images)by weight sampling method.After the training set was used for model learning and train-ing,two model architectures,residual neural network(ResNet18)and efficient neural network(EfficientNet-B0),were selected to verify the model in the test set by five-class and two-class classification tasks,respectively.Tak-ing histopathology as the gold standard,the diagnostic efficacy was evaluated to select the optimal model,and the Grad-CAM layer was embedded in the optimal model to output hysteroscopy images of Grad-CAM.Results:①In the five-class classification tasks,the accuracy of EfficientNet-B0 model(93.23％)was higher than that of Res-Net18 model(84.23％);the area under the curve(AUC)of EfficientNet-B0 model in the diagnosis of five disea-ses,including atypical endometrial hyperplasia,endometrial polyps,endometrial cancer,endometrial atypical hy-perplasia,and submucous myoma,was slightly higher than that of ResNet18 model,and the AUC of both models was almost above 0.980.②In the binary classification task of accuracy and the evaluation of specificity,the two models were similar,both above 93.00％,and the sensitivity of EfficientNet-B0 model(91.14％)was significantly better than that of ResNet18 model(77.22％).③EfficientNet-B0 model combined with Grad-CAM algorithm could identify the abnormal areas in the image.After biopsy and pathological examination,it was confirmed that about 95％of the marked areas in the model's output heatmap were lesion areas.Conclusions:The hysteroscopy di-agnostic model developed by EfficientNet-B0 model combined with Grad-CAM has high diagnostic accuracy,sen-sitivity,and specificity,and has application value in the diagnosis of endometrial lesions.