OBJECTIVE:Based on different algorithms of machine learning,the prediction model of lumbar disc herniation has become a trend and hot spot in the development of precision medicine.However,there is limited evidence on the reporting quality and methodological quality of prediction models of lumbar disc herniation outcomes using machine learning.This article is aimed to explore the performance of machine learning algorithms in predicting the prognosis of lumbar disc herniation by comprehensively analyzing the report quality and risk of bias of previous studies that developed and validated prognosis prediction models based on machine learning through a comprehensive literature search,in order to explore the performance of machine learning algorithms in predicting the prognosis of lumbar disc herniation.METHODS:The databases of CNKI,WanFang,VIP,SinOMED,PubMed,Web of Science,Embase,and The Cochrane Library were searched by computer.Studies on the use of machine learning to develop(and/or validate)prognostic prediction models for lumbar disc herniation were collected from the inception of the database to December 31,2023.Two researchers independently screened the literature,extracted data,and assessed the risk of bias of the included studies.The reporting quality and risk of bias of the included studies were assessed by the Multivariable Transparent Reporting of Predictive Models(TRIPOD)statement and the Predictive Model Risk of Bias Assessment Tool(PROBAST).The results of the evaluation were analyzed using descriptive statistics and visual charts.RESULTS:(1)A total of 23 articles were included,and the TRIPOD compliance of each study ranged from 11％to 87％,with a median compliance of 54％.The quality of reporting of titles,detailed descriptions of treatment measures,blinding of predictors,handling of missing data,details of risk stratification,specific procedures for enrollment,model interpretation,and model performance was mostly poor,with TRIPOD adherence rates ranging from 4％to 35％.(2)Of all included studies,61％had a high risk of bias and 39％had an unclear overall risk of bias.The area under the curve,accuracy,sensitivity and specificity were used to evaluate the performance of the model.The areas under the curve of 20 models were reported,ranging from 0.561 to 0.999.Three models reported the accuracy of the model,ranging from 82.07％to 89.65％.(3)Among all included studies,the statistical analysis domain was most often assessed as having a high risk of bias,mainly due to the small number of valid samples,the selection of predictors based on univariate analysis and the lack of calibration and discrimination assessment of the model in the study.CONCLUSION:These results indicate that machine learning can achieve good predictive ability in the development and validation of prognostic models for lumbar disc herniation.The commonly used algorithms include regression algorithm,support vector machine,decision tree,random forest,artificial neural network,naive Bayes and other algorithms.Reasonable algorithms combined with clinical practice can improve the accuracy of prognosis prediction of lumbar disc herniation.However,the reporting and methodological quality of prognosis prediction models based on machine learning are poor,the prediction performance of different models varies greatly,and the generalization and extrapolation of research models are unclear.There is an urgent need to improve the design,implementation and reporting of such studies.To promote the application of machine learning in the clinical practice of lumbar disc herniation prediction models,it is necessary to comprehensively consider various predictors related to the prognosis of the disease before modeling,and strictly follow the relevant standards of PROBAST tool during modeling.

Objective: To analyze the school tuberculosis (TB) outbreaks and preventive treatment in Hefei from 2022 to 2024, so as to provide reference for TB prevention and control in schools. Methods: Data were collected on all school based TB outbreaks occurring during 2022-2024 in Hefei, defined as ≥2 epidemiologically linked TB cases within the same school during a single semester. Statistical analyses were performed using the Chi square test. Results: Close contacts exhibited significantly higher TB incidence (2.88%) and latent mycobacterium tuberculosis infection (LTBI) rates (13.80%) in the school TB outbreaks, compared to non close contacts (0.12% and 2.63%, respectively). Among close contacts, secondary school students showed lower TB incidence (0.48%) and LTBI prevalence (3.42%) than both primary school or younger children (0.68%, 6.95%) and college students ( 0.78% , 6.50%), with statistically significant differences ( χ 2=360.91, 6.37; 791.71, 102.03, all P <0.05). The proportion of LTBI individuals recommended for preventive therapy was higher in primary school or younger groups (98.59%) than in secondary (95.25%) or college students (86.34%) ( χ 2=25.86, P <0.01). However, among those recommended, close contacts had higher uptake (85.82%) and completion rates (87.25%) of preventive therapy than non close contacts (69.63% and 70.57%); similarly, secondary school students demonstrated higher uptake (91.21%) and completion rates (86.45%) compared to primary school or younger (88.57%, 83.87%) and college students (57.28%, 64.08%) ( χ 2=30.52, 26.72; 125.17, 38.84, all P <0.01). Subsequent TB incidence among LTBI close contacts (13.30%) and among those who did not complete preventive therapy (22.73%) were significantly higher than among non close contacts (2.80%, 2.41%), respectively ( χ 2=32.19, 13.87, both P <0.05). Conclusions In school TB outbreaks, close contacts face higher LTBI prevalence and subsequent TB risk than non close contacts. College students show notably low adherence to preventive therapy. It is necessary to take targeted measures to improve the compliance of preventive measures among students.

OBJECTIVE:Based on different algorithms of machine learning,the prediction model of lumbar disc herniation has become a trend and hot spot in the development of precision medicine.However,there is limited evidence on the reporting quality and methodological quality of prediction models of lumbar disc herniation outcomes using machine learning.This article is aimed to explore the performance of machine learning algorithms in predicting the prognosis of lumbar disc herniation by comprehensively analyzing the report quality and risk of bias of previous studies that developed and validated prognosis prediction models based on machine learning through a comprehensive literature search,in order to explore the performance of machine learning algorithms in predicting the prognosis of lumbar disc herniation.METHODS:The databases of CNKI,WanFang,VIP,SinOMED,PubMed,Web of Science,Embase,and The Cochrane Library were searched by computer.Studies on the use of machine learning to develop(and/or validate)prognostic prediction models for lumbar disc herniation were collected from the inception of the database to December 31,2023.Two researchers independently screened the literature,extracted data,and assessed the risk of bias of the included studies.The reporting quality and risk of bias of the included studies were assessed by the Multivariable Transparent Reporting of Predictive Models(TRIPOD)statement and the Predictive Model Risk of Bias Assessment Tool(PROBAST).The results of the evaluation were analyzed using descriptive statistics and visual charts.RESULTS:(1)A total of 23 articles were included,and the TRIPOD compliance of each study ranged from 11％to 87％,with a median compliance of 54％.The quality of reporting of titles,detailed descriptions of treatment measures,blinding of predictors,handling of missing data,details of risk stratification,specific procedures for enrollment,model interpretation,and model performance was mostly poor,with TRIPOD adherence rates ranging from 4％to 35％.(2)Of all included studies,61％had a high risk of bias and 39％had an unclear overall risk of bias.The area under the curve,accuracy,sensitivity and specificity were used to evaluate the performance of the model.The areas under the curve of 20 models were reported,ranging from 0.561 to 0.999.Three models reported the accuracy of the model,ranging from 82.07％to 89.65％.(3)Among all included studies,the statistical analysis domain was most often assessed as having a high risk of bias,mainly due to the small number of valid samples,the selection of predictors based on univariate analysis and the lack of calibration and discrimination assessment of the model in the study.CONCLUSION:These results indicate that machine learning can achieve good predictive ability in the development and validation of prognostic models for lumbar disc herniation.The commonly used algorithms include regression algorithm,support vector machine,decision tree,random forest,artificial neural network,naive Bayes and other algorithms.Reasonable algorithms combined with clinical practice can improve the accuracy of prognosis prediction of lumbar disc herniation.However,the reporting and methodological quality of prognosis prediction models based on machine learning are poor,the prediction performance of different models varies greatly,and the generalization and extrapolation of research models are unclear.There is an urgent need to improve the design,implementation and reporting of such studies.To promote the application of machine learning in the clinical practice of lumbar disc herniation prediction models,it is necessary to comprehensively consider various predictors related to the prognosis of the disease before modeling,and strictly follow the relevant standards of PROBAST tool during modeling.

Background Air pollution remains a critical public health issue, with persistent exposure to air pollutants continuing to pose significant health risks. Currently, research investigating the association between air pollution and myocardial infarction mortality in Shenzhen remains inadequate. Objective To quantitatively assess the association between air pollutants and myocardial infarction mortality in residents. Methods Based on the mortality surveillance system of Shenzhen Center for Disease Control and Prevention, we conducted a time-stratified case-crossover study of 10089 permanent residents who died from myocardial infarction in Shenzhen between 2013 and 2022. Using residential address information, we obtained individual-level exposure data for air pollutants from the China High Air Pollutants dataset and meteorological factors from the China Meteorological Administration Land Data Assimilation System. A time-stratified case-crossover study design was employed to construct a conditional logistic regression model to assess the association between short-term exposure to air pollutants and myocardial infarction mortality. The exposure-response relationship was visualized, and a comprehensive health risk assessment was performed. Results This study included a total of 10 089 cases of myocardial infarction deaths in Shenzhen. The median [interquartile range (IQR)] concentrations of fine particulate matter (PM_2.5), inhalable particulate matter (PM₁₀), sulfur dioxide (SO₂), nitrogen dioxide (NO₂), carbon monoxide (CO), and ozone (O₃) in Shenzhen from 2013 to 2022 were 24.59 (20.19) μg·m⁻³, 42.85 (28.42) μg·m⁻³, 8.53 (3.39) μg·m⁻³, 29.47 (13.56) μg·m⁻³, 0.77 (0.27) mg·m⁻³, and 86.53 (51.39) μg·m⁻³, respectively. The moving average concentrations of PM_2.5 and PM₁₀ over the current day and the previous 2 days (lag02) showed the highest risk of myocardial infarction mortality, with an odds ratio (OR) and 95% confidence interval (95%CI) of 1.004 (1.001, 1.007) and 1.004 (1.002, 1.006), respectively. At lag05, SO₂ demonstrated the strongest association with the risk of myocardial infarction mortality, with an OR (95%CI) of 1.042 (1.019, 1.065). The exposure-response relationships of PM_2.5, PM₁₀, and SO₂ with myocardial infarction mortality were approximately linear, whereas NO₂ exhibited a nonlinear relationship. At lag05, the highest risk of myocardial infarction mortality associated with NO₂ exposure was observed when the NO₂ concentration was ≤21.92 μg·m⁻³, with an OR (95% CI) of 1.024 (1.003, 1.046). The health risk assessment indicated that, using the WHO Air Quality Guidelines as the reference, the local PM_2.5 and NO₂ exposure led to an excess mortality of 398 and 298 cases, respectively. The sensitivity analysis revealed that after adjusting for O₃ and restricting the study period to 2013—2019, the effect estimates for PM_2.5, PM₁₀, NO₂, and SO₂ on myocardial infarction mortality slightly increased. Conclusion Short-term exposure to air pollutants, including PM_2.5, PM₁₀, NO₂, and SO₂, could increase the risk of myocardial infarction mortality. This study provides important scientific evidence for environmental health risk assessment and environmental management in Shenzhen.

Background/Aims: Metabolic dysfunction-associated steatohepatitis (MASH) is a significant risk factor for gallstone formation, but mechanisms underlying MASH-related gallstone formation remain unclear. Golgi membrane protein 1 (GOLM1) participates in hepatic cholesterol metabolism and is upregulated in MASH. Here, we aimed to explore the role of GOLM1 in MASH-related gallstone formation. Methods: The UK Biobank cohort was used for etiological analysis. GOLM1 knockout (GOLM1-/-) and wild-type (WT) mice were fed with a high-fat diet (HFD). Livers were excised for histology and immunohistochemistry analysis. Gallbladders were collected to calculate incidence of cholesterol gallstones (CGSs). Biles were collected for biliary lipid analysis. HepG2 cells were used to explore underlying mechanisms. Human liver samples were used for clinical validation. Results: MASH patients had a greater risk of cholelithiasis. All HFD-fed mice developed MASH, and the incidence of gallstones was 16.7% and 75.0% in GOLM1-/- and WT mice, respectively. GOLM1-/- decreased biliary cholesterol concentration and output. In vivo and in vitro assays confirmed that GOLM1 facilitated cholesterol efflux through upregulating ATP binding cassette transporter subfamily G member 5 (ABCG5). Mechanistically, GOLM1 translocated into nucleus to promote osteopontin (OPN) transcription, thus stimulating ABCG5-mediated cholesterol efflux. Moreover, GOLM1 was upregulated by interleukin-1β (IL-1β) in a dose-dependent manner. Finally, we confirmed that IL-1β, GOLM1, OPN, and ABCG5 were enhanced in livers of MASH patients with CGSs. Conclusions In MASH livers, upregulation of GOLM1 by IL-1β increases ABCG5-mediated cholesterol efflux in an OPN-dependent manner, promoting CGS formation. GOLM1 has the potential to be a molecular hub interconnecting MASH and CGSs.

Aim To investigate the protective effect of liraglutide(LIRA)on acetaminophen(APAP)-in-duced hepatotoxicity at the in vivo level and to reveal the underlying mechanism.Methods Forty SPF grade male C57BL/6J mice were randomly divided into the Control,LIRA(200 μg·kg-1),APAP(500 mg·kg-1),LIRA+APAP,LIRA+APAP+3-methylade-nine(3-MA,30 mg·kg-1)groups,with eight mice in each group.The mice were administered for three con-secutive days,and the materials were taken after 24 h.The general condition and body weight of mice in each group were recorded,and liver morphology was ob-served.Serum ALT and AST levels,as well as SOD ac-tivity,MDA,and GSH content in liver homogenates,were measured using biochemical assay kits.The levels of inflammatory cytokines IL-6,TNF-α,and IL-1β in serum were detected by ELISA.Liver pathological changes were assessed by HE staining,while mitochon-drial and autophagosome structures in liver tissues were observed using transmission electron microscopy.The number of PCNA-positive cells in liver tissues was e-valuated using immunohistochemical staining.The pro-tein expression levels of LC3Ⅱ,p62,Bax,Bcl-2,PC-NA,and CyclinD1 in liver tissues were determined by Western blot.Results LIRA pretreatment can im-prove the general condition of mice with acetamino-phen-induced liver injury(AILI),reduce serum ALT and AST levels,and effectively ameliorate the appear-ance and morphology of the liver as well as the patho-logical damage to liver tissue.Simultaneously,the lev-els of inflammatory cytokines IL-6,TNF-α,and IL-1βare significantly decreased;SOD activity and GSH con-tent are significantly increased,while MDA content is significantly reduced.Transmission electron microsco-py observations reveal the presence of numerous auto-phagosomes in the cytoplasm of liver tissue.Immuno-histochemical staining results indicate a significant in-crease in the number of PCNA-positive cells.Further-more,the expression of LC3Ⅱ,Bcl-2,PCNA,and Cy-clinD1 proteins in liver tissue is significantly upregulat-ed,while the expression of p62 and Bax proteins is significantly downregulated.However,after interven-tion with the autophagy inhibitor 3-MA,the aforemen-tioned protective effects of LIRA are significantly.Conclusions LIRA pretreatment can significantly im-prove liver injury in AILI mice.Its protective mecha-nism may be related to enhancing autophagy in hepato-cytes,thereby reducing oxidative stress,inflammatory response and apoptosis in liver of AILI mice.

Objective:In the context of China's cardiovascular disease(CVD)high-risk population screening and intervention project,this study systematically evaluates the applicability of the EQ-5D-5L and SF-6Dv2 instruments among individuals at high risk of CVD.Methods:Convergent validity was assessed using Spearman's correlation coefficient.Measurement agreement was evaluated through intraclass correlation coefficients(ICC)and Bland-Altman plots.Factors influencing utility differences were explored using multiple linear regression analysis.Kruskal-Wallis test and t-test were used to examine discriminant validity.Sensitivity was compared by effect size(ES),relative efficiency(RE),and the area under the receiver operating characteristic curve(ROC-AUC).Floor and ceiling effects were also compared.Results:Among 5,415 individuals at high risk of CVD,the two instruments showed moderate overall correlation and acceptable convergent validity,but dimension-specific correlations were weak,and measurement consistency was low(ICC=0.367).Both instruments effectively distinguished different health states,yet the SF-6Dv2 demonstrated superior sensitivity and a milder ceiling effect.Conclusion:When measuring the health utility value of CVD patients,scale selection should be cautious,especially for high-risk groups,and SF-6Dv2 is more appropriate.

Objective:To investigate the outcomes of natural cycle (NC) and modified natural cycle (MNC) assisted reproductive technology (ART) in patients with diminished ovarian reserve (DOR), and to provide a scientific basis for individualized treatment strategies for DOR patients.Methods:A retrospective cohort analysis was performed on the clinical data of DOR patients who underwent ART at the Center for Reproductive Medicine of the Department of Obstetrics and Gynecology, Peking University Third Hospital from January 1, 2015 to December 31, 2023. Patients were divided into the NC group ( n=801) and the MNC group ( n=385) based on their treatment protocol. The primary outcomes were cycle cancellation rate and oocyte retrieval rate. Secondary outcomes included clinical pregnancy rate and live birth rate per fresh embryo transfer cycle and frozen-thawed embryo transfer cycle, cumulative pregnancy rate and cumulative live birth rate per started cycle and per transfer cycle, as well as laboratory parameters such as the number of retrieved oocytes, the number of two pronuclei (2PN) fertilized oocytes, the number of transferable embryos, and transferable embryo formation rate. Further, multivariate logistic regression was used to analyze the impact of the treatment protocol on pregnancy and live birth outcomes. Results:There were no statistically significant differences between the NC and MNC groups in terms of general characteristics such as age, body mass index, and baseline hormone levels (all P>0.05). The cycle cancellation rate was significantly higher in the NC group [19.10% (153/801)] than in the MNC group [10.65% (41/385), P<0.001], and the oocyte retrieval rate was significantly lower in the NC group [66.31% (431/650)] than in the MNC group [74.86% (259/346), P=0.005]. The number of retrieved oocytes [1 (0,1)], the number of 2PN fertilized oocytes [1 (0,1)], and the number of transferable embryos [0 (0, 1)] were also significantly lower in the NC group than in the MNC group [1 (1, 2), P<0.001; 1 (1, 1), P<0.001; 0 (0, 1), P<0.001]. However, there were no statistically significant differences in 2PN fertilization rate and transferable embryo formation rate between the NC and MNC groups (all P>0.05). In both fresh embryo transfer cycles and frozen-thawed embryo transfer cycles, there were no statistically significant differences in clinical pregnancy rate and live birth rate between the NC and MNC groups (all P>0.05). The cumulative pregnancy rate per started cycle and transfer cycle, the cumulative live birth rate per started cycle and per transfer cycle were also not significantly different between the NC and MNC groups (all P>0.05). Multivariate logistic analysis showed no significant association between NC and clinical pregnancy or live birth compared with MNC. Conclusion:While MNC to some extent reduced the cycle cancellation rate and improved oocyte retrieval rates compared with NC, it did not ultimately improve pregnancy outcomes in DOR patients.

Objective To compare the accuracy and clinical application value of contrast-induced fractional flow reserve(cFFR),resting full-cycle ratio(RFR),and quantitative flow ratio(QFR)in evaluating coronary artery borderline lesions,using fractional flow reserve(FFR)as the gold standard.Methods A retrospective study was conducted including 143 patients with 143 lesions who underwent coronary angiography(CAG)and were tested for cFFR,RFR,QFR,and FFR at Peking University Third Hospital from September 2020 to January 2022.Clinical data,CAG lesion anatomical data,and target vessel cFFR,RFR,QFR,and FFR measurements were collected.The correlation and diagnostic concordance of cFFR,RFR,QFR,and FFR were analyzed.Results The mean age of the 143 patients was 66(58,71)years,with 90(62.9％)being male.Among them,60(42.0％)patients were diagnosed with unstable angina,and 115(80.4％)target lesions were located in the left anterior descending artery.Correlation analysis showed that RFR,QFR,and cFFR were all significantly correlated with FFR(r=0.956,r=0.861,r=0.751,P＜0.001).Bland-Altman analysis demonstrated high agreement between cFFR,RFR,QFR,and FFR.Receiver operating characteristics(ROC)curve analysis showed that the area under the curve(AUC)for RFR corresponding to FFR ≤0.80 was 0.92(95％CI 0.87-0.96),for QFR was 0.89(95％CI 0.83-0.95),and for cFFR was 0.96(95％CI 0.94-0.99).Conclusions cFFR,RFR,and QFR show high concordance with FFR,with similar diagnostic consistency between the three methods and FFR.

This study was aimed at analyzing the biotypes and genetic diversity characteristics of five non-major Brucella species,to provide a scientific basis for understanding the species diversity of Brucella and strengthening pathogen monitoring and control.According to the biotypes(species,hosts,isolation locations,and time)and MLVA-16 genotypes(MLVA-16 lo-cus data,MLVA-11 genotypes)of five non-major pathogenic Brucella in the international MLVA database,we used Bionu-merics 8.0 software and PHYLOVIZ2.0 online software to analyze the geographical origin and genetic diversity characteristics of strains.A total of 227 strains were studied,including 121 Brucella ceti,47 B.pinnipedialis,37 Brucella ovis,11 B.mi-croti,and Brucella neotomae.The greatest host diversity was observed for B.ceti,followed by B.pinnipedialis and B.mi-croti.B.ceti was distributed in European and South American countries;B.pinnipedialiswas distributed in Europe;and B.microti.was distributed in the Czech Republic,Austria,and Hungary in Central Europe.B.ovis was widely distributed in Af-rica,Argentina,Australia,Brazil,Greece,the United States,Spain,and France.The MLVA-11 genotypes of different types of Brucella showed high polymorphism and large differences,thus suggesting that the strains have different geographical ori-gins.MST analysis indicated that the studied strains were divided into four branches(BCⅠ-Ⅳ),among which B.ceti was di-vided into two different branches(BC-Ⅰ and BC-Ⅱ),the strains of other types formed different branches(or sub-branches),and the strains of different types showed clear regional and dominant host characteristics.Genetic correlation analysis of strains of the Brucella genus revealed that non-major pathogenic Brucella had clear genetic,distribution,and host spectrum differ-ences with respect to four classical pathogenic Brucella species.Five non-major pathogenic Brucella strains presented unique genetic evolutionary patterns,geographical distributions,and host tropism characteristics,thereby providing new insight for understanding the biological and genetic diversity of those Brucella strains.