ObjectiveTo explore the potential mechanism of Xianglian Huazhuo prescription （XLHZ） in treating chronic atrophic gastritis （CAG） by regulating cell cycle and inhibiting proliferation， using bioinformatics technology and animal experiments. MethodsDifferential expressed genes （DEGs） related to CAG were screened using GEO database and GEO2R tool. Weighted gene co-expression network analysis （WGCNA） was employed to search for hub genes of CAG. These hub genes were intersected with cell cycle proliferation based on GeneCards database. Eenrichment analysis of the intersecting genes was performed to obtain signaling pathways and biological processes related to CAG. Protein protein interaction （PPI） analysis of genes was conducted using the Protein Interaction Platform （STRING） database to search the super hub gene （hub 2.0）， and animal experiments were conducted for further validation. Fourteen of 70 male Wistar rats were randomly selected as the normal group， and the remaining 56 rats were prepared by the combined modeling method of "starvation disorder+N-methyl-N-nitro-N-nitrosoguanidine （MNNG） + sodium salicylate". The successfully modeled rats were randomly divided into the model group， XLHZ-H， XLHZ-M， and XLHZ-L groups （36， 18， 9 g·kg^-1， respectively）， and Morodan group （1.4 g·kg^-1）. Each group was given corresponding intervention for 60 days. Hematoxylin-eosin （HE） staining was used to observe the histopathological changes of gastric mucosa in rats. The ultrastructure of gastric mucosal tissue cells was observed by transmission electron microscopy. The relative expression levels of TGF-β₁， Smad2 and Smad3 proteins， S/G₂/M phase marker geminin and proliferation marker MCM2 were detected by Western blot in gastric mucosal tissue， and Spearman correlation analysis was performed. ResultsA total of 15 hub 2.0 genes were identified， including TGF-β₁， suggesting the involvement of the TGF-β₁ signaling pathway in the CAG pathogenesis. Compared with the normal group， the expressions of TGF-β₁， Smad2， geminin and MCM2 proteins in the gastric mucosa tissue of the model group were increased （P<0.05）， and the expression of Smad3 protein was decreased （P<0.05）. Compared with the model group， the expressions of TGF-β₁ and geminin in the gastric mucosa were decreased in the drug groups （P<0.05）. The XLHZ-M group， XLHZ-H group and Morodan group had significantly decreased protein expression of Smad2 and MCM2 （P<0.05）. The protein expression of Smad3 was significantly increased in XLHZ-M， XLHZ-H， and Morodan groups （P<0.05）. Spearman correlation analysis showed that Smad3 was negatively correlated with other indicators， and positively correlated with other indicators （P<0.01）. ConclusionXLHZ may inhibit TGF-β₁/Smads signaling pathway， regulate cell cycle， and inhibit proliferation in the treatment of CAG.

ObjectiveTo explore the potential mechanism of Xianglian Huazhuo prescription （XLHZ） in treating chronic atrophic gastritis （CAG） by regulating cell cycle and inhibiting proliferation， using bioinformatics technology and animal experiments. MethodsDifferential expressed genes （DEGs） related to CAG were screened using GEO database and GEO2R tool. Weighted gene co-expression network analysis （WGCNA） was employed to search for hub genes of CAG. These hub genes were intersected with cell cycle proliferation based on GeneCards database. Eenrichment analysis of the intersecting genes was performed to obtain signaling pathways and biological processes related to CAG. Protein protein interaction （PPI） analysis of genes was conducted using the Protein Interaction Platform （STRING） database to search the super hub gene （hub 2.0）， and animal experiments were conducted for further validation. Fourteen of 70 male Wistar rats were randomly selected as the normal group， and the remaining 56 rats were prepared by the combined modeling method of "starvation disorder+N-methyl-N-nitro-N-nitrosoguanidine （MNNG） + sodium salicylate". The successfully modeled rats were randomly divided into the model group， XLHZ-H， XLHZ-M， and XLHZ-L groups （36， 18， 9 g·kg^-1， respectively）， and Morodan group （1.4 g·kg^-1）. Each group was given corresponding intervention for 60 days. Hematoxylin-eosin （HE） staining was used to observe the histopathological changes of gastric mucosa in rats. The ultrastructure of gastric mucosal tissue cells was observed by transmission electron microscopy. The relative expression levels of TGF-β₁， Smad2 and Smad3 proteins， S/G₂/M phase marker geminin and proliferation marker MCM2 were detected by Western blot in gastric mucosal tissue， and Spearman correlation analysis was performed. ResultsA total of 15 hub 2.0 genes were identified， including TGF-β₁， suggesting the involvement of the TGF-β₁ signaling pathway in the CAG pathogenesis. Compared with the normal group， the expressions of TGF-β₁， Smad2， geminin and MCM2 proteins in the gastric mucosa tissue of the model group were increased （P<0.05）， and the expression of Smad3 protein was decreased （P<0.05）. Compared with the model group， the expressions of TGF-β₁ and geminin in the gastric mucosa were decreased in the drug groups （P<0.05）. The XLHZ-M group， XLHZ-H group and Morodan group had significantly decreased protein expression of Smad2 and MCM2 （P<0.05）. The protein expression of Smad3 was significantly increased in XLHZ-M， XLHZ-H， and Morodan groups （P<0.05）. Spearman correlation analysis showed that Smad3 was negatively correlated with other indicators， and positively correlated with other indicators （P<0.01）. ConclusionXLHZ may inhibit TGF-β₁/Smads signaling pathway， regulate cell cycle， and inhibit proliferation in the treatment of CAG.

ObjectiveTo investigate the mechanism by which Xianglian Huazhuo prescription alleviates chronic atrophic gastritis （CAG） through modulating the nuclear factor erythroid 2-related factor 2 （Nrf2）/quinone oxidoreductase 1 （NQO1）/heme oxygenase-1 （HO-1） pathway of oxidative stress （OS） based on animal experiments. MethodsFifty rats were modeled for CAG by gavage of 2% sodium salicylate （10 mL·kg^-1）， free access to 180 mg·L^-1 N-methyl-N′-nitro-N-nitrosoguanidine （MNNG）， and periodic diet intervention （72 h cycle： 24 h fasting + 48 h free access to food）. The modeled rats were grouped by a random number method as follows： low-， medium-， and high-dose （9， 18，36 g·kg^-1， respectively） Xianglian Huazhuo prescription， Moradan， and model， with 10 rats in each group. The general state of all rats was observed. Hematoxylin-eosin （HE） staining was employed to observe the pathological changes of the gastric mucosal tissue. The ultrastructure of cells was observed under transmission electron microscopy. Biochemical methods were adopted to determine the levels of malondialdehyde （MDA） and total superoxide dismutase （T-SOD） in the rat serum. The mRNA and protein levels of Kelch-like ECH-associated protein 1 （Keap1）， Nrf2， NQO1， and HO-1 in the gastric mucosal tissue of rats from each group were measured by Real-time polymerase chain reaction （Real-time PCR） and Western blot， respectively. Reactive oxygen species （ROS） staining was employed to detect the fluorescence intensity of ROS in the gastric mucosal tissue of rats in each group. ResultsCompared with the normal group， the modeled rats were sluggish， had yellowing and sparse hair， short and dark tail， loose stool， and reduced food intake， and showed the gastric mucosa characterized by atrophy， thinning， congestive edema， declined gland density and disarrangement， and apparent inflammatory cell infiltration， irregular nucleus morphology， nuclear condensation， mitochondria with decreased number， declined density， and abnormal morphology， vacuolization， declined serum T-SOD level （P<0.05）， elevated serum MDA level （P<0.05）， downregulated mRNA and protein levels of Nrf2， NQO1， and HO-1 （P<0.05）， upregulated mRNA and protein levels of Keap1 （P<0.05）， and enhanced mean fluorescence intensity of ROS （P<0.05）. Compared with the model group， each treatment group exhibited improved general state， alleviated histopathological changes in gastric mucosa， improved ultrastructural characteristics of cells， elevated T-SOD level and declined MDA level in the serum （P<0.05）， upregulated mRNA and protein levels of Nrf2， NQO1， and HO-1， downregulated mRNA and protein levels of Keap1 （P<0.05）， and attenuated mean fluorescence intensity of ROS （P<0.05）. ConclusionXianglian Huazhuo prescription can improve the general status， ameliorate pathological changes of gastric mucosa， and repair the cell ultrastructure to varying degrees， demonstrating a definite therapeutic effect on CAG. It may restrain OS and safeguard the natural physiological function of gastric mucosa in CAG rats by regulating the Nrf2/NQO1/HO-1 signaling pathway.

ObjectiveTo investigate the mechanism by which Xianglian Huazhuo prescription alleviates chronic atrophic gastritis （CAG） through modulating the nuclear factor erythroid 2-related factor 2 （Nrf2）/quinone oxidoreductase 1 （NQO1）/heme oxygenase-1 （HO-1） pathway of oxidative stress （OS） based on animal experiments. MethodsFifty rats were modeled for CAG by gavage of 2% sodium salicylate （10 mL·kg^-1）， free access to 180 mg·L^-1 N-methyl-N′-nitro-N-nitrosoguanidine （MNNG）， and periodic diet intervention （72 h cycle： 24 h fasting + 48 h free access to food）. The modeled rats were grouped by a random number method as follows： low-， medium-， and high-dose （9， 18，36 g·kg^-1， respectively） Xianglian Huazhuo prescription， Moradan， and model， with 10 rats in each group. The general state of all rats was observed. Hematoxylin-eosin （HE） staining was employed to observe the pathological changes of the gastric mucosal tissue. The ultrastructure of cells was observed under transmission electron microscopy. Biochemical methods were adopted to determine the levels of malondialdehyde （MDA） and total superoxide dismutase （T-SOD） in the rat serum. The mRNA and protein levels of Kelch-like ECH-associated protein 1 （Keap1）， Nrf2， NQO1， and HO-1 in the gastric mucosal tissue of rats from each group were measured by Real-time polymerase chain reaction （Real-time PCR） and Western blot， respectively. Reactive oxygen species （ROS） staining was employed to detect the fluorescence intensity of ROS in the gastric mucosal tissue of rats in each group. ResultsCompared with the normal group， the modeled rats were sluggish， had yellowing and sparse hair， short and dark tail， loose stool， and reduced food intake， and showed the gastric mucosa characterized by atrophy， thinning， congestive edema， declined gland density and disarrangement， and apparent inflammatory cell infiltration， irregular nucleus morphology， nuclear condensation， mitochondria with decreased number， declined density， and abnormal morphology， vacuolization， declined serum T-SOD level （P<0.05）， elevated serum MDA level （P<0.05）， downregulated mRNA and protein levels of Nrf2， NQO1， and HO-1 （P<0.05）， upregulated mRNA and protein levels of Keap1 （P<0.05）， and enhanced mean fluorescence intensity of ROS （P<0.05）. Compared with the model group， each treatment group exhibited improved general state， alleviated histopathological changes in gastric mucosa， improved ultrastructural characteristics of cells， elevated T-SOD level and declined MDA level in the serum （P<0.05）， upregulated mRNA and protein levels of Nrf2， NQO1， and HO-1， downregulated mRNA and protein levels of Keap1 （P<0.05）， and attenuated mean fluorescence intensity of ROS （P<0.05）. ConclusionXianglian Huazhuo prescription can improve the general status， ameliorate pathological changes of gastric mucosa， and repair the cell ultrastructure to varying degrees， demonstrating a definite therapeutic effect on CAG. It may restrain OS and safeguard the natural physiological function of gastric mucosa in CAG rats by regulating the Nrf2/NQO1/HO-1 signaling pathway.

BACKGROUND:Small diameter artificial vessels are urgently needed to treat coronary artery and peripheral artery diseases in clinical practice.At present,vascular tissue engineering has become the main method for preparing small diameter artificial vessels.Selecting suitable biomaterials and cell sources is the key factor for successful construction of small diameter tissue engineered vessels. OBJECTIVE:To observe the effect of four kinds of electrospinning membrane materials on proliferation,adhesion and differentiation of bone marrow mesenchymal stem cells into vascular smooth muscle cells. METHODS:Bone marrow mesenchymal stem cells were isolated and extracted from SD rats.The bone marrow mesenchymal stem cells were inoculated separately on polycaprolactone(PCL),polycaprolactone-hyaluronic acid(PCL-HA),polycaprolactone-silk-filament proteins(PCL-SF),and polycaprolactone-gelatin(PCL-GEL)electrospinning membrane materials.After 1,3,and 7 days of culture,the cell arrangement on the material was observed under scanning electron microscope.The proliferation and adhesion of the material were observed by phalloidin staining.The mRNA expressions of CD90,Meflin,and transforming growth factor β were detected by qRT-PCR.After 7 days of induced differentiation into vascular smooth muscle cells,the mRNA expression ofɑ-smooth muscle actin on the material was detected by qRT-PCR. RESULTS AND CONCLUSION:(1)Bone marrow mesenchymal stem cells were arranged along the fibers of the four kinds of electrospinning membranes under scanning electron microscopy.(2)Phalloidin staining showed the regular distribution of bone marrow mesenchymal stem cells on the four kinds of electrospinning membranes and parallel distribution along the fiber direction.Moreover,PCL-HA,PCL-SF,and PCL-GEL electrospinning membranes were more conducive to the proliferation and adhesion of bone marrow mesenchymal stem cells than PCL electrospinning membranes.Compared with PCL-HA and PCL-GEL electrospinning membranes,PCL-SF electrospinning membranes were more conducive to the proliferation and adhesion of bone marrow mesenchymal stem cells.(3)qRT-PCR showed that the four kinds of electrospun membrane materials could maintain the mRNA expression of CD90 and Meflin in bone marrow mesenchymal stem cells,but there was no significant difference between groups(P>0.05).The mRNA expression of transforming growth factor β in PCL-HA,PCL-SF,and PCL-GEL groups was higher than that in PCL group on days 1 and 7(P<0.05),and the mRNA expression of transforming growth factor β in PCL-SF group was higher than that in the other three groups on days 3 and 7(P<0.05).The mRNA expression of transforming growth factor β in PCL-HA group was higher than that in PCL-GEL group on day 7(P<0.05).(4)qRT-PCR showed that the mRNA expression of ɑ-smooth muscle actin in PCL-SF group was higher than that in the other three groups(P<0.05),and that in PCL-HA group was higher than that in PCL group(P<0.05).(5)The results indicate that compared with PCL,PCL-HA and PCL-GEL electrospinning membranes,PCL-SF electrospinning membranes combined with bone marrow mesenchymal stem cells are more suitable for the preparation of small diameter tissue engineered vessels.

BACKGROUND:Human endometrial mesenchymal stem cells can directly repair the damaged endometrium,promote angiogenesis,and restore the morphological structure of the uterus.However,after the direct injection of stem cells into the damaged endometrium,the cell survival rate is low,the retention time is short,and the repair effect is limited.OBJECTIVE:To observe the effect of polycaprolactone-hyaluronic acid electrospinning membrane combined with human endometrial mesenchymal stem cells on endometrial injury in rats.METHODS:(1)Cell experiment:Human endometrial mesenchymal stem cells were extracted by collagenase digestion method.Polycaprolactone-hyaluronic acid electrospinning membrane was prepared by electrospinning technology.The human endometrial mesenchymal stem cells were inoculated on polystyrene culture plate and polycaprolactone-hyaluronic acid electrospinning membrane.The proliferation and adhesion of the cells were observed by DNA quantitative analysis,WST-1 cell activity test,phalloidin staining,and scanning electron microscopy.The mRNA expressions of CD90 and Meflin in electrospun membrane were detected by qRT-PCR.(2)Animal experiments:27 female SD rats in estrus were selected to establish uterine adhesion model by mechanical scratching method and randomly divided into three groups with nine rats in each group:The blank control group did not receive any treatment;the control group was implanted with polycaprolacton-hyaluronic acid electrospinning membrane;the experimental group was implanted with polycaprolacton-hyaluronic acid electrospinning membrane/human endometrial mesenchymal stem cell mesh.Samples were collected at 3,7,and 14 days after surgery.Hematoxylin-eosin staining was used to observe the morphological structure of uterus and the number of glands.qRT-PCR and immunofluorescence staining were used to observe the expression of CD31 and vascular endothelial growth factor in uterine tissue.RESULTS AND CONCLUSION:(1)Cell experiment:Compared with polystyrene culture plate,polycaprolactone-hyaluronic acid electrospinning membrane could promote the proliferation and adhesion of human endometrial mesenchymal stem cells.Polycaprolactone-hyaluronic acid electrospinning membrane supported the expression of CD90 and Meflin genes of human endometrial mesenchymal stem cells.(2)Animal experiments:Hematoxylin-eosin staining showed that polycaprolactic-hyaluronic acid electrospinning membrane/human endometrial mesenchymal stem cell patch could promote the recovery of endometrial morphological structure after injury.The endometrial thickness and number of gland on day 14 after surgery were higher than those in blank control group and control group(P<0.05).qRT-PCR and immunofluorescence staining showed that the mRNA and protein expressions of CD31 and vascular endothelial growth factor in the experimental group were higher than those in the blank control group and the control group at 7 and 14 days after surgery(P<0.05).(3)The results showed that polycaprolacton-hyaluronic acid electrospinning membrane could improve the survival rate of stem cells and prolong the contact time between stem cells and the damaged tissue,and the composite transplantation of the two could better repair the damaged endometrial tissue.

BACKGROUND:Human endometrial mesenchymal stem cells can directly repair the damaged endometrium,promote angiogenesis,and restore the morphological structure of the uterus.However,after the direct injection of stem cells into the damaged endometrium,the cell survival rate is low,the retention time is short,and the repair effect is limited.OBJECTIVE:To observe the effect of polycaprolactone-hyaluronic acid electrospinning membrane combined with human endometrial mesenchymal stem cells on endometrial injury in rats.METHODS:(1)Cell experiment:Human endometrial mesenchymal stem cells were extracted by collagenase digestion method.Polycaprolactone-hyaluronic acid electrospinning membrane was prepared by electrospinning technology.The human endometrial mesenchymal stem cells were inoculated on polystyrene culture plate and polycaprolactone-hyaluronic acid electrospinning membrane.The proliferation and adhesion of the cells were observed by DNA quantitative analysis,WST-1 cell activity test,phalloidin staining,and scanning electron microscopy.The mRNA expressions of CD90 and Meflin in electrospun membrane were detected by qRT-PCR.(2)Animal experiments:27 female SD rats in estrus were selected to establish uterine adhesion model by mechanical scratching method and randomly divided into three groups with nine rats in each group:The blank control group did not receive any treatment;the control group was implanted with polycaprolacton-hyaluronic acid electrospinning membrane;the experimental group was implanted with polycaprolacton-hyaluronic acid electrospinning membrane/human endometrial mesenchymal stem cell mesh.Samples were collected at 3,7,and 14 days after surgery.Hematoxylin-eosin staining was used to observe the morphological structure of uterus and the number of glands.qRT-PCR and immunofluorescence staining were used to observe the expression of CD31 and vascular endothelial growth factor in uterine tissue.RESULTS AND CONCLUSION:(1)Cell experiment:Compared with polystyrene culture plate,polycaprolactone-hyaluronic acid electrospinning membrane could promote the proliferation and adhesion of human endometrial mesenchymal stem cells.Polycaprolactone-hyaluronic acid electrospinning membrane supported the expression of CD90 and Meflin genes of human endometrial mesenchymal stem cells.(2)Animal experiments:Hematoxylin-eosin staining showed that polycaprolactic-hyaluronic acid electrospinning membrane/human endometrial mesenchymal stem cell patch could promote the recovery of endometrial morphological structure after injury.The endometrial thickness and number of gland on day 14 after surgery were higher than those in blank control group and control group(P<0.05).qRT-PCR and immunofluorescence staining showed that the mRNA and protein expressions of CD31 and vascular endothelial growth factor in the experimental group were higher than those in the blank control group and the control group at 7 and 14 days after surgery(P<0.05).(3)The results showed that polycaprolacton-hyaluronic acid electrospinning membrane could improve the survival rate of stem cells and prolong the contact time between stem cells and the damaged tissue,and the composite transplantation of the two could better repair the damaged endometrial tissue.

Objective To analyze the current status and frontier trends of research on the application of cognitive-motor dual task training(CMDT)in stroke. Methods Relevant literatures on the application of CMDT in stroke were retrieved from CNKI,Wanfang data,VIP,SinoMed and Web of Science Core Collection from inception to October 11,2023,and was analyzed with CiteSpace 6.2R4. Results A total of 285 articles were included with 124 in Chinese and 161 in English.The annual number of publications showed a general upward trend.United States,Canada,Netherlands,China and United Kingdom were the lead-ing countries in terms of output in English.The scholar with the most publications in Chinese was Zheng Jiejiao,and the Department of Rehabilitation Medicine in Huadong Hospital of Fudan University was the leading institu-tion for Chinese publications.Vrije University Amsterdam was the leading institution for English publications.The most frequent Chinese keywords were gait,falls,balance,cognition and postural control.The most frequent English keywords were dual-task,walking,gait,balance and cognitive-motor interference.Bursting keywords from the past two years included gait training,cognitive tasks,balance ability and cognitive-motor interference. Conclusion The researches on the application of CMDT in stroke are on the rise,with hotspots including gait training,cognitive tasks and cognitive-motor interference.The mechanisms of CMDT and the development of optimal CMDT rehabilitation protocols for stroke may be researched more in the future.

Objective To investigate the risk factors of 28-day mortality in septic patients and develop optimizing clinical prediction model based on machine learning algorithms.Methods Data from patients admitted to the department of intensive care unit(ICU)of the Affiliated Hospital of Nanjing University of Chinese Medicine from January 2019 to December 2023 were retrospectively analyzed.The data extracted included①gender,age,history of hypertension,diabetes,coronary heart disease,chronic obstructive pulmonary disease(COPD)and chronic kidney disease(CKD);②Vital signs and results of laboratory examination at admission were also collected,then acute physiology and chronic health evaluationⅡ(APACHEⅡ)score and sequential organ failure assessment(SOFA)score were calculated;③The other laboratory test results not included in APACHEⅡscore and SOFA score,such as blood lactate acid(Lac),alanine aminotransferase(AST),hemoglobin(Hb),procalcitonin(PCT),brian natriuretic peptide(BNP),C-reactive protein(CRP),activated partial thromboplastin time(APTT),D-dimer and troponin I(TNI)were also gathered.According to the 28-day survival,the patients were divided into a survival group and a death group.The difference of the clinical data and related loboratory indicators between the two groups of sepsis patients were compared.LASSO regression and Boruta algorithm were used to screen predictive variables.Models of Logistic regression(LG),neural network(NN)and light gradient boosting machine(LightGBM)were constructed.The data was divided into training set and verification set under a ratio of 7:3,and fivefold cross-validation was used to evaluate the stability of the models.Confusion matrix,receiver operator characteristic curve(ROC curve)and calibration curve were also used to assess the recognition ability and accuracy of three models.Decision curve analysis(DCA)was conducted to evaluate the models'utility in decision-making.Shapley additive explanations(SHAP)analysis was used to explain the best-performing model.Results A total of 426 patients were included in the study,of which 256 survived and 170 died.Compared with death group,the age(72.09±14.08 vs.76.88±11.32,P<0.05),COPD[11.33%(29/256)vs.20.00%(34/170)],CKD[20.31%(52/256)vs.31.77%(54/170)],Lac on admission[mmol/L:1.72(1.20,2.66)vs.2.25(1.60,3.50)],AST[U/L:32.00(18.00,59.75)vs.37.00(24.00,76.50)],CRP[mg/L:71.23(22.51,151.79)vs.87.00(37.00,173.36)],APACHEⅡscore(19.96±6.55 vs.22.83±6.92)and SOFA score[7(5,10)vs.9(5,12)]in surrial group were significantly decreased,the difference were statistically significant(all P<0.05).Age,APACHEⅡscore,Lac,PCT and CRP were revealed as independent predictors of 28-day mortality in sepsis by LASSO regression and Boruta algorithm,the above 5 variables were incorporated into the LG,NN and LightGBM models,and the five-fold cross-validation showed that the LightGBM model had the best stability.The confusion matrix,ROC curve and calibration curves of the 3 models were plotted,and the results showed that the F1 score of the 3 models were 0.61,0.63 and 0.74,respectively;area under the curve(AUC)was 0.68,0.74 and 0.87,respectively;the Log Loss was 0.62,0.41 and 0.34,respectively;and the Brier scores were 0.22,0.13 and 0.09,respectively,indicating that LightGBM model was optimal.DCA showed that LightGBM model had the greatest clinical net benefit.SHAP showed that the predicted results were in good agreement with the actual results.Conclusion The LightGBM model exhibited the best performance in predicting 28-day mortality in septic patients and has the potential to help clinicians identify high-risk patients and guide clinical decision-making.

Objective To investigate the risk factors of 28-day mortality in septic patients and develop optimizing clinical prediction model based on machine learning algorithms.Methods Data from patients admitted to the department of intensive care unit(ICU)of the Affiliated Hospital of Nanjing University of Chinese Medicine from January 2019 to December 2023 were retrospectively analyzed.The data extracted included①gender,age,history of hypertension,diabetes,coronary heart disease,chronic obstructive pulmonary disease(COPD)and chronic kidney disease(CKD);②Vital signs and results of laboratory examination at admission were also collected,then acute physiology and chronic health evaluationⅡ(APACHEⅡ)score and sequential organ failure assessment(SOFA)score were calculated;③The other laboratory test results not included in APACHEⅡscore and SOFA score,such as blood lactate acid(Lac),alanine aminotransferase(AST),hemoglobin(Hb),procalcitonin(PCT),brian natriuretic peptide(BNP),C-reactive protein(CRP),activated partial thromboplastin time(APTT),D-dimer and troponin I(TNI)were also gathered.According to the 28-day survival,the patients were divided into a survival group and a death group.The difference of the clinical data and related loboratory indicators between the two groups of sepsis patients were compared.LASSO regression and Boruta algorithm were used to screen predictive variables.Models of Logistic regression(LG),neural network(NN)and light gradient boosting machine(LightGBM)were constructed.The data was divided into training set and verification set under a ratio of 7:3,and fivefold cross-validation was used to evaluate the stability of the models.Confusion matrix,receiver operator characteristic curve(ROC curve)and calibration curve were also used to assess the recognition ability and accuracy of three models.Decision curve analysis(DCA)was conducted to evaluate the models'utility in decision-making.Shapley additive explanations(SHAP)analysis was used to explain the best-performing model.Results A total of 426 patients were included in the study,of which 256 survived and 170 died.Compared with death group,the age(72.09±14.08 vs.76.88±11.32,P<0.05),COPD[11.33%(29/256)vs.20.00%(34/170)],CKD[20.31%(52/256)vs.31.77%(54/170)],Lac on admission[mmol/L:1.72(1.20,2.66)vs.2.25(1.60,3.50)],AST[U/L:32.00(18.00,59.75)vs.37.00(24.00,76.50)],CRP[mg/L:71.23(22.51,151.79)vs.87.00(37.00,173.36)],APACHEⅡscore(19.96±6.55 vs.22.83±6.92)and SOFA score[7(5,10)vs.9(5,12)]in surrial group were significantly decreased,the difference were statistically significant(all P<0.05).Age,APACHEⅡscore,Lac,PCT and CRP were revealed as independent predictors of 28-day mortality in sepsis by LASSO regression and Boruta algorithm,the above 5 variables were incorporated into the LG,NN and LightGBM models,and the five-fold cross-validation showed that the LightGBM model had the best stability.The confusion matrix,ROC curve and calibration curves of the 3 models were plotted,and the results showed that the F1 score of the 3 models were 0.61,0.63 and 0.74,respectively;area under the curve(AUC)was 0.68,0.74 and 0.87,respectively;the Log Loss was 0.62,0.41 and 0.34,respectively;and the Brier scores were 0.22,0.13 and 0.09,respectively,indicating that LightGBM model was optimal.DCA showed that LightGBM model had the greatest clinical net benefit.SHAP showed that the predicted results were in good agreement with the actual results.Conclusion The LightGBM model exhibited the best performance in predicting 28-day mortality in septic patients and has the potential to help clinicians identify high-risk patients and guide clinical decision-making.