Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

Objective: Machine learning (ML) has been reported to have better predictive capability than traditional statistical techniques. The aim of this study was to assess the efficacy of ML algorithms and logistic regression (LR) for predicting depressive symptoms during the COVID-19 pandemic. Methods: Analyses were carried out in a national cross-sectional study involving 21,916 participants. The ML algorithms in this study included random forest (RF), support vector machine (SVM), neural network (NN), and gradient boosting machine (GBM) methods. The performance indices were sensitivity, specificity, accuracy, precision, F1-score, and area under the receiver operating characteristic curve (AUC). Results: LR and NN had the best performance in terms of AUCs. The risk of overfitting was found to be negligible for most ML models except for RF, and GBM obtained the highest sensitivity, specificity, accuracy, precision, and F1-score. Therefore, LR, NN, and GBM models ranked among the best models. Conclusion Compared with ML models, LR model performed comparably to ML models in predicting depressive symptoms and identifying potential risk factors while also exhibiting a lower risk of overfitting.

Aim To explore the effects of Lycium berry seed oil on Nrf2/ARE pathway and oxidative damage in testis of subacute aging rats. Methods Fifty out of 60 male SD rats, aged 8 weeks, were subcutaneously injected with 125 mg • kg"D-galactosidase in the neck for 8 weeks to establish a subacute senescent rat model. The presence of senescent cells was observed using P-galactosidase ((3-gal), while testicular morphology was examined using HE staining. Serum levels of testosterone (testosterone, T), follicle-stimulating hormone ( follicle stimulating hormone, FSH ) , luteinizing hormone ( luteinizing hormone, LH ) , superoxide dis-mutase ( superoxide dismutase, SOD ) , glutathione ( glutathione, GSH) and malondialdehyde ( malondial-dehyde, MDA) were measured through ELISA, and the expressions of factors related to aging, oxidative damage, and the Nrf2/ARE pathway were assessed via immunohistochemical analysis and Western blotting. Results After successfully identifying the model, the morphology of the testis was improved and the intervention of Lycium seed oil led to a down-regulation in the expression of [3-gal and -yH2AX. The serum levels of SOD, GSH, T, and FSH increased while MDA and LH decreased (P 0. 05) . Additionally, there was an up-regulated expression of Nrf2, GCLC, NQOl, and SOD2 proteins in testicular tissue ( P 0. 05 ) and nuclear expression of Nrf2 in sertoli cells. Conclusion Lycium barbarum seed oil may reduce oxidative damage in testes of subacute senescent rats by activating the Nrf2/ARE signaling pathway.

Aim To analyze the anti-A549 and HI299 lung ade-nocarcinoma activities via using examples of baicalin, astragalo-side, hesperidin and cisplatin based on real time cellular analysis (RTCA) technology, and to build a new strategy for EC50 e-valuation reflecting the time-dimensional characteristic. Methods Using RTCA Software Pro for data analysis and GraphPad Prism and Origin Pro plotting, the in vitro anti-A549 and H1299 lung adenocarcinoma activities of baicalin, astragaloside, hesperidin, and cisplatin were characterized using the endpoint method and time dimension, respectively. Results (X) There were significant differences in EC50 values of A549 and H1299 cells at 24 h and 48 h endpoint methods. (2) The correlation coefficient of the curve fitted with the four-parameter equation was > 0. 9, and the dynamic change of EC50 remained relatively stable (the linear fitting of EC50 at adjacent 4 points I slope 1^1) used to calculate the EC50 value within this time dimension. The EC50 of baicalin, astragaloside, hesperidin and cisplatin on A549 cells was 52. 97 ±1.75 плпо! • L~1(16~48 h) , 62.88 ± 2.91 ijunol • L"1 (32.25 -48 h) , 78.84 ±0.33 плпо1 • L"1 (21.5 -29.75 h), 13.57 ±1.54 плпо1 • L_1(27.5 -48 h), respectively; the EC50 of baicalin, astragaloside, hesperidin and cisplatin on H1299 cells was 43. 71 ± 1. 26 |лто1 • L_1 ( 19. 5 -48 h), 47.23 ±1. 19 |лто1 • L_1(14 -48 h) , 39.45 ±0.24 плпо1 • L"1 (12.75 -46.25 h), 25.97 ±4.76 плпо1 • L"1 (10. 25 -48 h) , respectively. The results showed that the time window for the anti-tumor effect of the test solution/drug was different. Conclusions Based on RTCA technology, it is more accurate and reasonable to select EC50 data that exhibit better fitting, stable changes, and time-dimensional characteristics for the evaluation of anti-tumor activity. In addition, this method of distinguishing different effective time of antitumor drugs can provide a reference for the timing of clinical combination drugs, and this approach will also provide a reference for further related studies.

OBJECTIVE To evaluate the cost-effectiveness of trastuzumab biosimilars （Hanquyou） versus original drug （Hesaiting） in the treatment of recurrent/metastatic human epidermal growth factor receptor-2 （HER-2） positive breast cancer. METHODS A partitional survival model was constructed based on the NCT03084237 trial data. The simulation period was 3 weeks， and the simulation time was 10 years. Using costs and quality-adjusted life year （QALY） as the output indicator， the cost- utility analysis method was used to evaluate the cost-effectiveness of the two schemes mentioned above. Univariate and probabilistic sensitivity analyses were performed to verify the robustness of the basic analysis. RESULTS The costs of the trastuzumab biosimilars group and original drug group were 111 516.72 yuan and 111 122.30 yuan respectively， with health utility values of 1.52 QALYs and 1.36 QALYs， and ICER of 2 465.12 yuan/QALY， which were less than 3 times China’s per capita gross domestic product （GDP） in 2023 as the threshold for willingness-to-pay （WTP） （268 200 yuan/QALY）. Univariate sensitivity analysis showed that the cost of the trastuzumab biosimilars and original drug had a great impact on the ICER. The probabilistic sensitivity analysis showed that the probability of trastuzumab biosimilars being cost-effective was 100% at WTP threshold of 14 902 yuan/QALY. CONCLUSIONS When WTP threshold is 3 times China’s GDP in 2023 （268 200 yuan/QALY）， compared with original drug， trastuzumab biosimilars have good cost-effectiveness in the treatment of recurrent/metastatic HER-2 positive breast cancer.

BACKGROUND:As a leading technique in the treatment of primary stenosis by posterior spinal endoscopy through unilateral approach and bilateral decompression using single channel endoscopy,the long-term efficacy needs to be further observed.There are few reports on the scope of intraoperative resection and few relevant studies on biomechanics and finite element analysis. OBJECTIVE:A three-dimensional finite element model was established to evaluate the effects of bilateral lumbar canal decompression under a one-hole split endoscope on lumbar range of motion and intradiscal pressure,to provide suggestions for clinical operation and theoretical basis for further clinical research. METHODS:A complete L3-L5 vertebral body model was reconstructed by CT images of nine healthy volunteers,which was used as the preoperative model M1.The simulated surgical resection range of L4-L5 was performed,and 1/4,1/3 and 1/2 of bilateral facet joints were removed respectively to obtain models M2,M3 and M4.The range of motion and the maximum Von Mises stress of the four models were compared in the six directions of forward bending,backward extension,left and right bending,and left and right rotation. RESULTS AND CONCLUSION:(1)The L3-L5 finite element model established in this study was effective,and the range of motion was within the range of previous solid studies under six motion states.(2)Compared with the M1 model,the L4-L5 lumbar spine range of motion increased with the increase of resection range in M2 with M3 and M4 models under forward bending,left and right bending and left and right rotation loading,and the difference was significant(P＜0.05).Under posterior extension loading,there was no significant difference in lumbar range of motion between M1 and M2(P＞0.05),but there was a significant difference of M1,M3 and M4(P＜0.05).(3)The range of motion of the L3-L4 lumbar spine had no significant change with the increase of bilateral facet arthrotomy(P＞0.05).(4)There was a significant difference in the maximum value of L4-L5 Von Mises between M1 and M2(P＜0.05),and there was a significant difference in the maximum value of L4-L5 Von Mises between M1 and M3,M4(P＜0.01),and the maximum value of L4-L5 lumbar von Mises increased with the increasing range of bilateral facet joint resection.Resection of more than 1/3 was particularly obvious.(5)The maximum value of Von Mises in the L3-L4 lumbar spine was increased with the increase of the resection range under forward bending,left and right bending and left and right rotation loading and the difference was significant(P＜0.05).(6)The results exhibited that the L4-L5 lumbar motion and intervertebral disc pressure increased with the increase of the excision range.Intervertebral disc pressure at L3-L4 increased with the increased extent of excision,but the lumbar range of motion was not significantly affected.In conclusion,the stability of the operative segment may be affected by the increase in the scope of facet joint resection.Although the immediate stability of adjacent segments is not affected,it may accelerate disc degeneration.

Objective To summarize the epidemiological and clinical features of infectious diseases of the central nervous system(CNS)by a single-center analysis.Methods A retrospective analysis was conducted on the data of 1247 cases of CNS infectious diseases diagnosed and treated in the First Medical Center of PLA General Hospital from 2001 to 2020.Results The data for this group of CNS infectious diseases by disease type in descending order of number of cases were viruses 743(59.6％),Mycobacterium tuberculosis 249(20.0％),other bacteria 150(12.0％),fungi 68(5.5％),parasites 18(1.4％),Treponema pallidum 18(1.4％)and rickettsia 1(0.1％).The number of cases increased by 177 cases(33.1％)in the latter 10 years compared to the previous 10 years(P<0.05).No significant difference in seasonal distribution pattern of data between disease types(P>0.05).Male to female ratio is 1.87︰1,mostly under 60 years of age.Viruses are more likely to infect students,most often at university/college level and above,farmers are overrepresented among bacteria and Mycobacterium tuberculosis,and more infections of Treponema pallidum in workers.CNS infectious diseases are characterized by fever,headache and signs of meningeal irritation,with the adductor nerve being the more commonly involved cranial nerve.Matagenomic next-generation sequencing improves clinical diagnostic capabilities.The median hospital days for CNS infectious diseases are 18.00(11.00,27.00)and median hospital costs are ￥29,500(￥16,000,￥59,200).The mortality rate from CNS infectious diseases is 1.6％.Conclusions The incidence of CNS infectious diseases is increasing last ten years,with complex clinical presentation,severe symptoms and poor prognosis.Early and accurate diagnosis and standardized clinical treatment can significantly reduce the morbidity and mortality rate and ease the burden of disease.