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: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 explore the changes and correlations of color， components and anti-tumor activity of Gelsemium elegans during sand-frying processing， so as to provide a reference for the toxicity-attenuating and efficacy-preserving processing as well as quality control of sand-fried G. elegans decoction pieces.METHODS Raw G.elegans and its processed products prepared by sand-frying for 15， 30， 75， 105 and 120 s were taken as test samples. The chromatic values， contents of total alkaloids and seven main active components （gelsemine， humantenine， koumine， gelsenicine， etc.） were determined. Their in vitro anti-tumor activity was investigated. The correlations between color and component contents， as well as between total alkaloid content and anti-tumor activity were analyzed.RESULTS The color of G.elegans samples generally darkened after sand-frying， and the total color difference （Δ E * ） of samples sand-fried for 75 s was greater than 20. At the initial stage of sand-frying （15-30 s）， the total alkaloid content in processed products decreased sharply compared with the raw material； at the later-stage of sand-frying （75-105 s）， the declining rate of total alkaloid content slowed down and tended to be stable. The contents of gelsenicine and humantenidine in processed products dropped drastically at 30 s of sand-frying， while the contents of koumine and sempervirine decreased substantially （by about 80%） after 75 s of sand-frying and then gradually tended to be stable. The anti-HepG2 cell proliferation activity of each processed product decreased with the extension of frying time. Correlation analysis demonstrated that color brightness and yellow-blue chromaticity were significantly positively correlated with the contents of total alkaloids and all determined active components except gelsemine （ P ＜0.05 or P ＜0.01）； red-green chromaticity was significantly positively correlated with the contents of koumine， sempervirine， gelsemivirine and humantenidine （ P ＜0.05 or P ＜0.01）； Δ E * was significantly negatively correlated with the contents of total alkaloids and all determined active components except gelsemine. Besides， total alkaloid content was significantly negatively correlated with the half maximal inhibitory concentration of the samples on HepG2 cells （ P ＜0.05 or P ＜0.01）.CONCLUSIONS With the prolongation of sand-frying time， the color of sand-fried G.elegans decoction pieces gradually turned brownish-brown， the contents of seven active components and total alkaloids decreased progressively， and its anti-HepG2 cell activity showed an overall downward trend. Sand-frying for 75 s serves as a critical turning-point for the above-mentioned changes.

ObjectiveTo analyze the evolutionary characteristics and genetic variations of the HA (hemagglutinin) and NA (neuraminidase) genes of influenza A(H1N1) viruses in Shanghai during 2024, to investigate their transmission patterns, and to evaluate their potential impact on vaccine effectiveness. MethodsFrom January to October 2024, throat swab specimens were collected from influenza like illness (ILI) patients at 4 hospitals in Shanghai. Real-time fluorescence ploymerase chain reaction (RT-PCR) was used for virus detection and isolation of H1N1 influenza viruses. Forty influenza A(H1N1) virus strains were sequenced using Illumina NovaSeq 6000 platform, followed by phylogenetic analyses, genetic distance analysis, and amino acid variation analyses of HA and NA genes. ResultsPhylogenetic tree of the HA and NA genes revealed that the 40 influenza A(H1N1) virus strains circulating in Shanghai in 2024 exhibited no significant geographic clustering, with a broad origin of strains and complex transmission chains. Genetic distance analyses demonstrated that the average intra-group genetic distances of HA and NA genes among the Shanghai strains were 0.005 1±0.000 6 and 0.004 6±0.000 6, respectively, which were comparable to or higher than those observed in global surveillance strains. Both HA and NA genes displayed frequent mutations. Compared to the 2023‒2024 and 2024‒2025 Northern Hemisphere A(H1N1) vaccine strains (WHO-recommended), the HA proteins of 40 Shanghai strains exhibited amino acid substitutions at positions 120, 137, 142, 169, 216, 223, 260, 277, 356 and 451, with critical mutations at positions 137 and 142 located within the Ca2 antigenic determinant. Furthermore, mutations in the NA protein were observed at positions 13, 50, 200, 257, 264, 339 and 382. ConclusionThe genetic background of the 2024 Shanghai influenza A(H1N1) virus strains is complex and diverse, and antigenic variation may affect vaccine effectiveness. Therefore, it is recommended to enhance genomic surveillance of influenza viruses, evaluate vaccine suitability, and implement more targeted prevention and control strategies against imported influenza viruses.

Objective To investigate the clinical application of ceftazidime/avibactam(CAZ/AVI)in lung trans-plant recipients with pulmonary infection caused by carbapenem-resistant Gram-negative bacilli(CRGNB),and ana-lyze the factors affecting the prognosis.Methods Lung transplant recipients who had CRGNB pulmonary infection and were treated with CAZ/AVI were included in the analysis.Based on 14-day clinical response,14-day microbial response,and 30-day survival status,the recipients were divided into a clinical response group and a clinical failure group,a microbial response group and a microbial failure group,as well as a survival group and a death group,re-spectively.Univariate analysis was conducted on various data from the two groups.Factors affecting therapeutic ef-ficacy and survival were included in a binary logistic regression model.Independent risk factors for CAZ/AVI anti-infective efficacy and all-cause mortality outcomes were analyzed.Results A total of 43 recipients were included.After 14-day anti-infective treatment,32 recipients(74.42％)achieved clinical response,and 30 recipients(69.77％)achieved microbial response.34 recipients(79.07％)survived 30 days after CAZ/AVI treatment.The Charlson comorbidity index(CCI),proportion of renal dysfunction,and incidence of shock in recipients in the clini-cal response group were all lower than those in the clinical failure group(P＜0.05),while the serum albumin(ALB)level was higher(P＜0.05).The incidence of shock in recipients in the microbial response group was lower than that in the microbial failure group(P＜0.05).CCI,proportion of renal dysfunction,and incidence of shock in recipients in the survival group were all lower than those in the death group(all P＜0.05),while ALB level was higher during treatment period(P＜0.05).Multivariate analysis of 14-day clinical response and 30-day survival showed that higher CCI was an independent risk factor affecting 14-day clinical response of recipients(OR=2.22,95％CI:1.07-4.63),while lower ALB levels(OR=0.72,95％CI:0.54-0.98)and higher CCI(OR=5.27,95％CI:1.18-23.58)were independent risk factors for 30-day all-cause mortality in recipients with pulmonary in-fection after lung transplant.Conclusion CAZ/AVI may be an effective drug for treating pulmonary infection caused by CRGNB in lung transplant recipients.Higher CCI is an independent risk factor for 14-day clinical failure in recipients after CAZ/AVI treatment.Lower ALB level and higher CCI are independent risk factors for increased 30-day mortality in recipients.

Objective To analyze the medication rules and clinical experience of Qin Zujie in treating virtual labor based on data mining.Methods Outpatient medical records of Qin Zujie,diagnosed with deficiency syndrome at Ethnic Medicine Characteristic Diagnosis and Treatment Center,Guangxi International Zhuang Medical Hospital from January to December 2024 were selected.A standardized database was established and processed,yielding 148 clinical prescriptions containing 171 herbal ingredients.Frequency analysis,association rule analysis,and cluster analysis were employed to examine medication rules and clinical experience.Results The 20 Chinese herbal ingredients used more than 30 times each,with the first three being Zhigancao,Baizhu,Danggui.These herbs are primarily warm in nature,neutral in temperature,and slightly cold in effect,characterized by sweet,pungent,and bitter flavors.They mainly affect the spleen meridian and function to tonify deficiency,with Qi-tonifying herbs being the most common.Through association rule analysis,the most frequently combined pairs were Baizhu-Dangshen.Cluster analysis identified three core prescription clusters.Conclusion Qin Zujie's academic thought of treating virtual labor is to supplement deficiency as the main,detoxification as the auxiliary,and treat both the symptoms and the causes.

Objective:To explore the influencing factors for pathological upgrading in prostate cancer patients with a Gleason score of 3 + 3 undergoing targeted biopsy，and to establish a nomogram prediction model.Methods:A retrospective analysis was conducted on 191 patients with localized prostate cancer diagnosed with a Gleason score of 3 + 3 through targeted biopsies at the First Affiliated Hospital of Nanjing Medical University from January 2020 to June 2024. The age of the patients was 67（61，73）years，with prostate-specific antigen（PSA）level of 7.44（5.53，10.19）ng/ml，prostate volume of 35.64（26.59，48.97）ml，and PSA density（PSAD）of 0.20（0.14，0.31）ng/ml 2. Among them，61 cases（31.94%）had a Prostate Imaging Reporting and Data System（PI-RADS）score of 3，104 cases（54.45%）had a score of 4，and 26 cases（13.61%）had a score of 5. The diameter of the main lesion was 10.75（7.86，14.00）mm. The lesions were located in the peripheral zone in 78 cases（40.84%），the transition zone in 99 cases（51.83%），and the anterior fibromuscular stroma in 14 cases（7.33%）. The lesions were found at the apex in 56 cases（29.32%），in the body in 120 cases（62.83%），and at the base in 15 cases（7.85%）. MRI revealed only one lesion with a PI-RADS score ≥ 3 in 131 cases，two suspected lesions in 43 cases，three suspected lesions in 12 cases，and four suspected lesions in 5 cases. Systematic biopsy was positive in 121 cases（63.4%）and negative in 70 cases（36.6%）. The lesions were confined to the left lobe in 63 cases（32.98%），right lobe in 68 cases（35.60%），and involved both lobes in 60 cases（31.41%）. The interval between biopsy and surgery was 9.0（7.0，14.0）days. Univariate analyses were performed using Mann-Whitney U tests or χ2 tests，and multivariate logistic regression was used to identify independent predictors of pathological upgrading. A nomogram model was constructed based on these independent predictors. The model’s discriminative ability was assessed using the area under the receiver operating characteristic（ROC）curve（AUC），and internal validation of the model’s consistency was conducted using the bootstrap resampling method. Decision curve analysis（DCA）was performed to assess clinical utility. Results:Among the 191 cases，60（31.4%）had no pathological upgrading after surgery，while 131（68.6%）showed upgrading. Univariate analysis showed that the maximum diameter of the main lesion［9.0（6.0，13.2）mm vs. 11.0（8.4，14.0）mm］，number of suspicious lesions on MRI［1.0（1.0，1.0）vs. 1.0（1.0，2.0）］，number of positive systematic biopsy cores［1.0（0，2.0）vs. 1.0（0，3.0）］，percentage of positive systematic biopsy cores［0.08（0，0.17）vs. 0.12（0，0.25）］，number of positive targeted biopsy cores［2.0（1.0，3.0）vs. 3.0（1.0，4.0）］，percentage of positive targeted biopsy cores［0.37（0.24，0.75）vs. 0.50（0.38，0.85）］，level of the index lesion，location of the index lesion，and PI-RADS score were associated with pathological upgrading（ P < 0.05）. Multivariate logistic regression analysis showed that PI-RADS score 4（ OR = 5.88，95% CI 2.41 - 14.35），number of suspicious lesions on MRI（ OR = 4.15，95% CI 1.88 - 9.17），location of the index lesion in the transition zone（ OR = 6.86，95% CI 2.81 - 16.73），and percentage of positive targeted biopsy cores（ OR = 4.37，95% CI 1.38 - 14.90）were independent risk factors for pathological upgrading（ P < 0.05）. The nomogram model constructed using these predictors had an AUC of 0.845. Internal validation using the Bootstrap method yielded an AUC value of 0.812，indicating high predictive accuracy of the model. The calibration curve indicated good calibration. Decision curve analysis showed that the threshold range for net benefit in the model was between 12% - 100%. Conclusions:The PI-RADS score 4，the number of lesions with PI-RADS ≥ 3，the location of the main lesion in the transition zone，and the percentage of positive needles in targeted biopsy are independent risk factors for pathological upgrading from Gleason score 3 + 3. The nomogram model constructed from these factors demonstrates good predictive performance and provides a reference for clinical decision-making.

Magnetic resonance angiography(MRA)is a non-invasive imaging technique used to observe blood vessels.Quantitative analysis of MRA images enables visualization of vascular pathways,condition,and blood flow dynamics,which is essential for diagnosing vascular diseases such as vascular lesions,stenosis,and occlusions.Vessel segmentation serves as the fundamental basis for quantitative vascular analysis.However,the complex morphology of vessels,difficulties in labeling,and scarcity of accurate 3D vascular annotations pose significant challenges for MRA-based vessel segmentation.A strategy of selectively occluding vessels during model training is proposed to enhance the algorithm's capacity to capture the topological structure of blood vessels,thereby improving the continuity of vessel segmentation results.Additionally,a Refine network is incorporated to refine the binary segmentation results of the segmentation network,thereby further improving segmentation accuracy.Model training and testing are carried out using 42 cases of 3D MRA data from the public MIDAS dataset.For the test set,the 3D U-Net baseline model with vessel occlusion strategy shows a β0 Error of 1.2742±0.2103 and a β1 Error of 0.3393±0.0818,respectively,which are 0.1136 and 0.0280 lower than the baseline.The model integrating vessel occlusion strategy and Refine network achieves an average Dice score of 0.7105±0.0125,which is 0.0028 higher than the baseline.These results demonstrate that the proposed method effectively improves both vascular connectivity and segmentation accuracy.

To compare gait parameters during single-task and dual-task walking in older adults, and to examine differences in dual-task costs between individuals with high versus low balance abilities under different task conditions.

To explore the dose-response correlation between appendicular skeletal muscle index (ASMI) and muscle strength in older adults.