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 To evaluate the efficacy,safety,and cost-effectiveness of berberine-based quadruple therapy vs.the clarithromycin-based quadruple therapy for Helicobacter pylori(H.pylori)eradication in treatment-na?ve patients.Methods This was a single-center,prospective,open-label randomized controlled trial.A total of 404 treatment-naive patients with H.pylori infection who visited the Outpatient Department of Gastroenterology,the First Medical Center of Chinese PLA General Hospital from September 2021 to May 2024 were enrolled.The patients were randomly assigned in a 1:1 ratio to two groups:berberine quadruple therapy group(berberine+amoxicillin+esomeprazole+colloidal bismuth pectin;n=202)and clarithromycin quadruple therapy group(clarithromycin+amoxicillin+esomeprazole+colloidal bismuth pectin;n=202).Both groups received a 14-day treatment course.The H.pylori eradication rate,incidence of adverse reactions,medication compliance,and treatment costs were compared between the two groups.Results By intention-to-treat(ITT)analysis,eradication rate did not differ significantly between the two groups[89.1%(180/202)in berberine quadruple therapy group vs.89.6%(181/202)in clarithromycin quadruple therapy group,P=0.872].The per-protocol(PP)analysis also showed no significant difference in the eradication rate between the two groups[90.4%(179/198)vs.91.3%(178/195),P=0.763].The incidence of adverse reactions in berberine quadruple therapy group was significantly lower than that in clarithromycin quadruple therapy group[18.2%(36/198)vs.38.5%(75/195),P<0.001].Specifically,the incidence of taste disturbance in berberine quadruple therapy group was significantly lower than that in clarithromycin quadruple therapy group(3.0%vs.15.4%,P<0.001).There was no statistically significant difference in medication compliance between the two groups[98.5%(195/198)in berberine quadruple therapy group vs.97.9%(191/195)in clarithromycin quadruple therapy group,P=0.688].The fixed direct medical cost per patient was significantly lower in berberine quadruple therapy group than that in clarithromycin quadruple therapy group(402.08 yuan vs.693.94 yuan).Conclusions The berberine-based quadruple therapy is as effective as traditional clarithromycin-based quadruple therapy for eradicating H.pylori,with the advantages of a lower incidence of adverse reactions and lower cost.It represents a safe,effective,and economical treatment option worthy of further promotion and application.

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 To establish and evaluate a nomogram prediction model for spontaneous peritonitis in HBV-related primary liver cancer.Methods A retrospective study was conducted on 1298 patients with HBV-related primary liver cancer hospitalized in the Kunming Third People's Hospital from January 2012 to December 2022.General data and serological indicators were collected,and patients were divided into infection group(n=262)and control group(n=1036)based on the occurrence of spontaneous peritonitis.Univariate and LASSO regression analyses were used to screen variables,followed by binary logistic regression to analyze the influencing factors of spontaneous peritonitis in HBV-related primary liver cancer patients,leading to the establishment of a nomogram prediction model.Finally,the Hosmer-lemeshow(H-L)goodness of fit test,receiver operating characteristic(ROC)curve,calibration curve,decision curve analysis(DCA)and clinical impact curve(CIC)were utilized to evaluate the fit degree,accuracy,calibration,and clinical practicability of the nomogram prediction model.Results Single factor analysis revealed significant differences between infection group and control group in portal vein cancer thrombus(PVTT),Child-Pugh grade,China Liver Cancer Staging(CNLC)stage,alcohol consumption history,smoking history,white blood cell count(WBC),neutrophil count(NE),hemoglobin(Hb),fibrinogen(FIB),abnormal prothrombin(PIVKA-Ⅱ),aspartate aminotransferase(AST),alanine aminotransferase(ALT),total protein(TP),prealbumin(PA),γ-glutamyltransferase(GGT),alkaline phosphatase(ALP),cholinesterase(CHE),total bile acid(TBA),total cholesterol(TC),low density lipoprotein(LDL),creatinine(Cr),HBV DNA,CD3+T cells count,CD4+T cells count,CD8+T cells count,CD4+T cells/CD8+T cells ratio,procalcitonin(PCT),serum amyloid A(SAA),interleukin-6(IL-6),high-sensitivity C-reactive protein(hs-CRP),alpha-fetoprotein(AFP),and IL-4(P＜0.05).LASSO regression analysis identified 5 variables:Child-Pugh grade,PVTT,WBC,CHE and hs-CRP.Binary logistic regression analysis indicated that Child-Pugh grade(Grade B:OR=5.780,95％CI 3.271-10.213,P＜0.001;Grade C:OR=14.818,95％CI 7.697-28.526,P＜0.001),PVTT(OR=2.893,95％CI 2.037-4.108,P＜0.001),WBC(OR=1.088,95％CI 1.031-1.148,P=0.002),and hs-CRP(OR=1.005,95％CI 1.001-1.010,P=0.026)were the independent risk factors of spontaneous peritonitis in HBV-related primary liver cancer patients.Using these 4 variables,a nomogram prediction model was constructed and evaluated.The P-value of the H-L goodness of fit test was 0.760.Moreover,the area under ROC curve(AUC)was 0.866,with a sensitivity of 0.870 and a specificity of 0.716.The average absolute error of the calibration curve is 0.022.DCA and CIC analyses demonstrated that the nomogram prediction model possessed some clinical utility.Conclusion The nomogram prediction model for spontaneous peritonitis in HBV-related primary liver cancer patients,constructed using Child-Pugh grade,PVTT,WBC and hs-CRP,exhibits a high fitting degree and accuracy,with the prediction probability highly consistent with the actual occurrence probability,and possesses certain clinical practicability.

OBJECTIVE: To investigate the effect of TBL1XR1 mutation on cell biological characteristics of diffuse large B-cell lymphoma (DLBCL). METHODS: The TBL1XR1 overexpression vector was constructed and DNA sequencing was performed to determine the mutation status. The effect of TBL1XR1 mutation on apoptosis of DLBCL cell line was detected by flow cytometry and TUNEL fluorescence assay; CCK-8 assay was used to detect the effect of TBL1XR1 mutation on cell proliferation; Transwell assay was used to detect the effect of TBL1XR1 mutation on cell migration and invasion; Western blot was used to detect the effect of TBL1XR1 mutation on the expression level of epithelial-mesenchymal transition (EMT) related proteins. RESULTS: The TBL1XR1 overexpression plasmid was successfully constructed. The in vitro experimental results showed that TBL1XR1 mutation had no significant effect on apoptosis of DLBCL cells. Compared with the control group, TBL1XR1 mutation enhanced cell proliferation, migration and invasion of DLBCL cells. TBL1XR1 gene mutation significantly increased the expression of N-cadherin protein, while the expression of E-cadherin protein decreased. CONCLUSION TBL1XR1 mutation plays a role in promoting tumor cell proliferation, migration and invasion in DLBCL. TBL1XR1 could be considered as a potential target for DLBCL therapy in future research.
Humans ; Lymphoma, Large B-Cell, Diffuse/pathology* ; Cell Proliferation ; Mutation ; Receptors, Cytoplasmic and Nuclear/genetics* ; Apoptosis ; Cell Line, Tumor ; Epithelial-Mesenchymal Transition ; Cell Movement ; Repressor Proteins/genetics* ; Nuclear Proteins/genetics* ; Cadherins/metabolism*

Objective:To investigate the current status, trends, and differences among institutions in the application of delivery room continuous positive airway pressure (DRCPAP) for very preterm infants treated in the institutions in the Chinese Neonatal Network (CHNN). Also, to explore the impact of DRCPAP on the outcomes of very preterm infants in China.Methods:A retrospective cohort study was conducted. Based on the CHNN very preterm infant cohort, very preterm infants (gestational ages ranging from 25 weeks +0 to 31 weeks +6) born in-hospital and treated in 79 tertiary neonatal intensive care units (NICUs) participating in the CHNN from 2019 to 2021 were included. The usage rates of DRCPAP in different hospitals, as well as gestational ages and years, were described. Data were analyzed using the Chi-square test (or Fisher's exact test) or t-test. A multivariate logistic regression model was established to explore the correlation between DRCPAP and clinical outcomes. Results:(1) A total of 18 048 very preterm infants were included. Among them, 3 666 (20.3%) received DRCPAP, and 14 382 (79.7%) did not. (2) The usage rate of DRCPAP for very preterm infants among different institutions is from 0.0% to 94.5%. Fourteen institutions did not use DRCPAP, and 55 institutions had a usage rate below 30%. The usage rate of DRCPAP in very preterm infants increased annually, from 13.8% (818/5 916) in 2019 to 26.0% (1 583/6 097) in 2021 ( χ2trend=122.00, P<0.001). (3) The DRCPAP group had higher rates of maternal assisted reproductive technology pregnancy, chorioamnionitis, a full course of antenatal corticosteroids, gestational diabetes, fetal distress, antenatal magnesium sulfate use, and cesarean delivery compared to the non-DRCPAP group [20.3% (744/3 665) vs. 17.6% (2 529/14 369), χ2=14.45; 23.0% (695/3 021) vs. 16.4% (1 956/11 926), χ2=72.57; 57.1% (2 090/3 660) vs. 54.3% (7 766/14 302), χ2=9.55; 23.0% (844/3 669) vs. 20.7% (2 969/14 342), χ2=9.77; 8.7% (319/3 666) vs. 7.0% (1 006/14 371), χ2=12.51; 87.1% (3 186/3 657) vs. 82.0% (11 736/14 312), χ2=81.38; 63.5% (2 327/3 664) vs. 60.7% (8 722/14 369), χ2=9.59; all P<0.05]. While the incidence of hypertensive disorders of pregnancy and the proportion of infants not using antenatal corticosteroids were lower in the non-DRCPAP group [17.1% (626/3 660) vs. 22.6% (3 183/14 084), χ2=44.70; 14.2% (520/3 661) vs. 19.7% (2 814/14 284), χ2=57.34; all P<0.05]. The DRCPAP group had lower birth weight and gestational age, higher 1 min and 5 min Apgar scores, and lower neonatal transport stabilization index scores [(1 308±314) g vs. (1 325±315) g, t=2.90; (29.5±1.7) weeks vs. (29.7±1.6) weeks, t=3.96; (7.9±1.8) scores vs. (7.6±1.9) scores, t=－9.80; (9.0±1.1) scores vs. (8.7±1.3) scores, t=－13.01; (11.0±9.4) scores vs. (13.1±9.8) scores, t=11.31; all P<0.05]. The incidence of early-onset sepsis was higher in the DRCPAP group than in the non-DRCPAP group [1.8% (68/3 578) vs. 1.3% (193/14 296), adjusted OR (95% CI): 1.417 (1.028-1.955)], while the rates of tracheal intubation within 72 hours, PS use, invasive mechanical ventilation, mortality, admission hypothermia, grade Ⅲ-Ⅳ intracranial hemorrhage, and stage≥3 ROP were all lower in the DRCPAP group compared to the non-DRCPAP group [22.6% (830/3 666) vs. 36.9% (5 310/14 374), adjusted OR (95% CI): 0.499 (0.448-0.555); 53.1% (1 947/3 666) vs. 58.3% (8 388/14 377), adjusted OR (95% CI): 0.764 (0.697-0.836); 30.1% (1 104/3 662) vs. 43.9% (6 307/14 366), adjusted OR (95% CI): 0.539 (0.487-0.595); 7.4% (274/3 666) vs. 10.6% (1 526/14 342), adjusted OR (95% CI): 0.709 (0.601-0.836); 57.5% (2 103/3 657) vs. 66.5% (9 501/14 287), adjusted OR (95% CI): 0.722 (0.661-0.788); 3.0% (101/3 366) vs. 5.9% (763/12 932)], adjusted OR (95% CI): 0.525 (0.412-0.669); 2.2% (65/2 954) vs. 3.3% (367/11 121), adjusted OR (95% CI): 0.692 (0.505-0.947); all P<0.05]. There were no statistically significant differences between the two groups in the incidence rates of BPD at a corrected gestational age of 36 weeks, patent ductus arteriosus ligation, spontaneous intestinal perforation, and pneumothorax. Conclusions:Domestically, the application of DRCPAP might be related to a decrease in the demand for invasive ventilation, the use of surfactants, and mortality, but it might not reduce the occurrence of bronchopulmonary dysplasia, at a corrected gestational age of 36 weeks. In recent years, the use of DRCPAP in very premature infants in China has increased, but the overall usage rate is still low, and there are significant differences between units, requiring continuous quality improvement.