1.Explainable Machine Learning Model for Predicting Prognosis in Patients with Malignant Tumors Complicated by Acute Respiratory Failure: Based on the eICU Collaborative Research Database in the United States
Zihan NAN ; Linan HAN ; Suwei LI ; Ziyi ZHU ; Qinqin ZHU ; Yan DUAN ; Xiaoting WANG ; Lixia LIU
Medical Journal of Peking Union Medical College Hospital 2026;17(1):98-108
To develop and validate a model for predicting intensive care unit (ICU) mortality risk in patients with malignant tumors complicated by acute respiratory failure (ARF) based on an explainable machine learning framework. Clinical data of patients with malignant tumors and ARF were extracted from the eICU Collaborative Research Database in the United States, including demographic characteristics, comorbidities, vital signs, laboratory test indicators, and major interventions within the first 24 hours after ICU admission.The study outcome was ICU death.Enrolled patients were randomly divided into a training set and a validation set at a ratio of 7:3.Predictor variables were selected using least absolute shrinkage and selection operator (LASSO) regression.Five machine learning algorithms-extreme gradient boosting (XGBoost), support vector machine (SVM), Logistic regression, multilayer perceptron (MLP), and C5.0 Decision Tree-were employed to construct predictive models.Model performance was evaluated based on the area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, and other metrics.The optimal model was further interpreted using the Shapley additive explanations (SHAP) algorithm. A total of 3196 patients with malignant tumors complicated by ARF were included.The training set comprised 2, 261 patients and the validation set 935 patients; 683 patients died during ICU stay, while 2513 survived.LASSO regression ultimately selected 12 variables closely associated with patient ICU outcomes, including sepsis comorbidity, use of vasoactive drugs, and within the first 24 hours after ICU admission: minimum mean arterial pressure, maximum heart rate, maximum respiratory rate, minimum oxygen saturation, minimum serum bicarbonate, minimum blood urea nitrogen, maximum white blood cell count, maximum mean corpuscular volume, maximum serum potassium, and maximum blood glucose.After model evaluation, the XGBoost model demonstrated the best performance.The AUCs for predicting ICU mortality risk in the training and validation sets were 0.940 and 0.763, respectively; accuracy was 88.3% and 81.2%;sensitivity was 98.5% and 95.9%.Its predictive performance also remained optimal in sensitivity analyses.SHAP analysis indicated that the top five variables contributing to the model's predictions were minimum oxygen saturation, minimum serum bicarbonate, minimum mean arterial pressure, use of vasoactive drugs, and maximum white blood cell count. This study successfully developed a mortality risk prediction model for ICU patients with malignant tumors complicated by ARF based on a large-scale dataset and performed explainability analysis.The model aids clinicians in early identification of high-risk patients and implementing individualized interventions.
2.Gut microbiota and their metabolites in hemodialysis patients.
Junxia DU ; Xiaolin ZHAO ; Xiaonan DING ; Qinqin REN ; Haoran WANG ; Qiuxia HAN ; Chenwen SONG ; Xiaochen WANG ; Dong ZHANG ; Hanyu ZHU
Chinese Medical Journal 2025;138(4):502-504
3.Jasurolignoside from Ilex pubescens exerts a therapeutic effect on acute lung injury in vitro and in vivo by binding to TLR4.
Shan HAN ; Chi Teng VONG ; Jia HE ; Qinqin WANG ; Qiumei FAN ; Siyuan LI ; Jilang LI ; Min LIAO ; Shilin YANG ; Renyikun YUAN ; Hongwei GAO
Chinese Journal of Natural Medicines (English Ed.) 2025;23(9):1058-1068
Acute lung injury (ALI) is a severe disease caused by viral infection that triggers an uncontrolled inflammatory response. This study investigated the capacity of jasurolignoside (JO), a natural compound, to bind to Toll-like receptor 4 (TLR4) and treat ALI. The anti-inflammatory properties of JO were evaluated in vitro through Western blotting, enzyme-linked immunosorbent assay (ELISA), immunofluorescence staining, and co-immunoprecipitation. The investigation utilized a lipopolysaccharide (LPS)-induced ALI animal model to examine the therapeutic efficacy and mechanism of JO in vivo. JO attenuated inflammatory symptoms in infected cells and tissues by modulating the NOD-like receptor family pyrin domain containing protein 3 (NLRP3) inflammasome and the nuclear factor κB (NF-κB)/mitogen-activated protein kinase (MAPK) pathway. Molecular docking simulations revealed JO binding to TLR4 active sites, confirmed by cellular thermal shift assay. Surface plasmon resonance (SPR) demonstrated direct interaction between JO and TLR4 with a Kd value of 35.1 μmol·L-1. Moreover, JO inhibited tumor necrosis factor α (TNF-α), interleukin-1β (IL-1β), and IL-6 secretion and reduced leukocyte, neutrophil, lymphocyte, and macrophage infiltration in ALI-affected mice. JO also enhanced lung function and reduced ALI-related mortality. Immunohistochemical staining demonstrated JO's ability to suppress TLR4 expression in ALI-affected mouse lung tissue. This study establishes that JO can bind to TLR4 and effectively treat ALI, indicating its potential as a therapeutic agent for clinical applications.
Toll-Like Receptor 4/chemistry*
;
Animals
;
Acute Lung Injury/chemically induced*
;
Mice
;
Humans
;
Ilex/chemistry*
;
Molecular Docking Simulation
;
Male
;
NF-kappa B/immunology*
;
Mice, Inbred C57BL
;
NLR Family, Pyrin Domain-Containing 3 Protein/immunology*
;
Tumor Necrosis Factor-alpha/genetics*
;
Interleukin-1beta/genetics*
;
RAW 264.7 Cells
;
Disease Models, Animal
4.Effect of HEDIS-based phased health education on clinical nursing in infertile patients treated with assisted reproductive technology
Xiangling HAN ; Yang WANG ; Qing ZHANG ; Yanju QIN ; Yanmin LUAN ; Qinqin ZHANG
Journal of Navy Medicine 2025;46(6):614-619
Objective To investigate the application of the Health Effective Data and Information System(HEDIS)-based phased health education in infertile patients treated with assisted reproductive technology(ART).Methods A total of 120 infertile patients who were admitted to The First Affiliated Hospital of Naval Medical University from March 2023 to September 2023 were consecutively selected and randomly assigned to observation group or control group at a ratio of 1:1 using a random number table.During the ART treatment period,the control group was given conventional nursing care,while the observation group was given HEDIS based phased health education for nursing intervention.The negative emotion score,shame score,self-management ability score,and nursing satisfaction were compared between the two groups.Results After intervention,the Hamilton Anxiety Scale(HAMA)score and Hamilton Depression Scale(HAMD)score,and self-rating scale(ISS)scores of the observation group were lower than those of the control group(P<0.05).The degree of nursing satisfaction of the observation group was higher than that of the control group(P<0.05).Conclusion The HEDIS-based phased health education can alleviate the negative emotions of anxiety and depression,reduce the sense of shame,and enhance nursing satisfaction in infertile patients during ART treatment.
5.Genetic characterization of coxsackievirus A6 in Mianyang city in 2022
Yue GONG ; Kexin ZONG ; Qian CHENG ; Lulu WANG ; Renhe ZHU ; Qin LUO ; Yanhai WANG ; Guoyong MEI ; Jun HAN ; Qinqin SONG
Chinese Journal of Microbiology and Immunology 2024;44(5):414-422
Objective:To investigate the etiology of hand, foot and mouth disease (HFMD) in Mianyang city in 2022, and to analyze the genetic characteristics of coxsackievirus A6.Methods:Pharyngeal swabs were collected from patients with HFMD in Mianyang city in 2022. Real-time fluorescence quantitative PCR was used to detect enteroviruses in the samples. Part of the VP1 gene in enterovirus-positive samples was amplified by nested PCR using enterovirus typing primers to further identify the viral types. The VP1 coding region of all CVA6-positive samples and the whole genome of some samples were amplified and sequenced by PCR. The endemic strain in Mianyang city was analyzed for phylogeny, gene homology, amino acid variation and genetic recombination.Results:A total of 151 pharyngeal swabs were collected, and 104 enterovirus-positive samples were detected by real-time fluorescence quantitative PCR, with an overall detection rate of 68.88% (104/151). The typing results showed that there were 77 cases of CVA6 infection, with a positive rate of 50.99% (77/151). The full-length VP1 genes of 77 CVA6 strains were amplified, sequenced, and successfully spliced, and phylogenetic analysis showed that all 77 strains were of the D3 genotype. There were multiple amino acid variant sites in the prevalent strains in Mianyang city compared with the reference strain. Twenty whole genome sequences were amplified, sequenced, and successfully spliced, and homology analysis showed that the nucleotide homology between the 20 positive sequences ranged from 97.0% to 99.9%. Phylogenetic tree and recombination analysis showed that no recombination occurred in the coding regions of the epidemic strains in this study.Conclusions:The predominant pathogen causing HFMD in Mianyang city in 2022 is CVA6 D3 subtype, which is consistent with the national epidemic in 2022.
6.Virus spectrum and epidemic characteristics of hospitalized children with acute respiratory tract infections in Guilin area, 2021 to 2022
Renhe ZHU ; Hu LONG ; Rundong CAO ; Lulu WANG ; Juan SONG ; Qinqin SONG ; Guoyong MEI ; Zhiqiang XIA ; Jun HAN ; Chen GAO
Chinese Journal of Experimental and Clinical Virology 2024;38(1):43-48
Objective:To understand the composition of respiratory virus spectrum in hospitalized children with acute respiratory infections in Guilin City from 2021 to 2022, and to explore the epidemic characteristics of different respiratory viruses.Methods:Information of a total of 638 hospitalized children with acute respiratory infections (ARI) syndrome under the age of 14 years in Guilin city, Guangxi from September 2021 to October 2022 was collected as research data. The collected pharyngeal swabs from pediatric patients were subjected to real-time fluorescence reverse transcription polymerase chain reaction for nucleic acid testing, screening for 15 common respiratory viruses, and analyzing the virus spectrum characteristics of hospitalized pediatric patients with acute respiratory infections.Results:Among the 638 specimens, 365 were tested positive for at least one virus, with a positive rate of 57.21% (365/638). The detection rate for two or more viruses was 12.85% (82/638). There were 12 types of viruses detected, and the difference in virus detection rates among different seasons was statistically significant (P<0.002). The virus positive detection rate of samples from different age groups was the highest in the 0-2-year old group (40.66%), followed by the 3-5-year old group (34.80%) and the 6-year and older group (24.54%).Conclusions:During the period from September 2021 to October 2022, all 12 respiratory viruses were prevalent in Guilin area, with obvious summer peak as characteristics. It is suggested that corresponding prevention and control measures should be focused on pathogens that may cause acute respiratory infections in children during the local summer to prevent and control the spread and prevalence of respiratory infections.
7.Characterization of rhinovirus in patients with respiratory tract infection in Mianyang, 2021—2022
Yue GONG ; Ming PAN ; Guo CHEN ; Qinqin SONG ; Yanhai WANG ; Chen GAO ; Zhiqiang XIA ; Lulu WANG ; Qian CHENG ; Kexin ZONG ; Jun HAN
Chinese Journal of Experimental and Clinical Virology 2024;38(3):279-285
Objective:To understand the prevalence and characteristics of Rhinovirus (HRV) infection in influenza-like Illness (ILIs) patients in Mianyang, Sichuan province, China.Methods:Throat swabs were collected from patients of ILIs in sentinel hospitals in Mianyang during 2021—2022. Real-time fluorescence quantitative PCR was used to detect 16 common pathogens. The VP4/VP2 coding region genes of HRV positive samples were amplified by nest PCR. The phylogeny, consistency and amino acid variation of different serotypes were analyzed and compared with reference sequences from GenBank database.Results:A total of 332 ILIs′ samples were collected with a virus detection rate of 58.73% (195/332) in Mianyang. Among them, 23 samples (23/332) were HRV-positive, and 18 VP4/VP2 sequences of HRV strains were successfully amplified. It was found that 13 HRV serotypes were detected in ILIs samples in Mianyang, which belonged to three genotypes, namely HRV-A (12 strains), HRV-B (5 strains) and HRV-C (1 strain).Conclusions:HRV was one of the pathogens of ILIs cases in Mianyang during 2021—2022, with HRV-A types as the dominant strains.
8.Establishment of a method for rescuing coxsackievirus B3 by transfecting RNA
Mei LI ; Xinling WANG ; Qinqin SONG ; Miaomiao CHI ; Jun HAN ; Juan SONG
Chinese Journal of Experimental and Clinical Virology 2024;38(4):468-473
Objective:To establish a method via transfection of RNA to rescue coxsackievirus B3 B3 (CVB3).Methods:The efficiency of CVB3 genomic RNA extraction from three nucleic acid extraction reagents, Qiagen 57704, Qiagen 52904, and Trizol, and the transfection efficiency of viral RNA with two transfection reagents (Lipofectamine 2000 and Lipofectamine 3000) were compared. The efficiency of CVB3 rescue in Vero cells and HEK293T cells to determine the optimal conditions for rescuing CVB3.Results:The number of phagolysosomes for virus rescue by Qiagen 57704, Qiagen 52904, and Tizol kit extracted RNA was 13.33±1.53, 150±15.00, and 1.67±0.58, respectively, and there was a statistically significant difference in the efficiency of the three method of extracting CVB3 RNA to rescue the viral RNA ( F=268.920, P<0.001); The number of phage spots formed by Lipofectamine3000 and Lipofectamine2000 transfected RNA was 74.50±3.00 and 22.00±5.00, respectively, and the difference was statistically significant ( P<0.01); Qiagen 52904 reagent extracted CVB3 nucleic acid more efficiently than Qiagen 57704 and Trizol reagents; the transfection efficiency of transfection reagent Lipofectamine 3000 was 3 times more than than that of Lipofectamine 2000, and the efficiency of virus rescue of CVB3 in HEK293T cell culture was higher than that of HeLa and Vero cells, and the copy numbers of the three kinds of viruses rescuing the virus were 6.09×10 7±8.00×10 5, 5.18×10 3±6.17×10 2 and 0, the difference was statistically significant ( F=17 383.644, P<0.001), and it was also found that the efficiency of virus rescue could be improved by multiple elution when extracting RNA. Conclusions:In this study, we successfully established the method of transfecting RNA to rescue CVB3, which can effectively improve the efficiency of virus rescue by choosing Qiagen 52904 nucleic acid extraction kit, increasing the number of elution, using Lipofectamine 3000 transfection reagent, and transfection of HEK293T cells.
9.Imbalance of Th17/Treg induced by coxsackievirus B3 leads to viral pancreatitis in mice
Ruifang WANG ; Kexin ZONG ; Qinqin SONG ; Dong XIA ; Haijun DU ; Mi LIU ; Zhiqiang XIA ; Jun HAN ; Juan SONG
Chinese Journal of Experimental and Clinical Virology 2023;37(2):128-134
Objective:A mouse model of pancreatitis induced by coxsackievirus B3 (CVB3) was established. The pathological change of pancreas and the infiltration of Th17/Treg cells were observed.Methods:The BALB/c mice were inoculated intraperitoneally with CVB3 to induce acute viral pancreatitis model. Then the pathological changes of pancreas were observed by HE staining; the viral RNA load and relative expression of cytokines (IFN-γ, IL-6 and IL-17) mRNA were detected by q-PCR; the proportion of infiltrated CD45 + CD3 + T cells, CD4 + and CD8 + T cells, Th17 and Treg cells in the pancreas was determined by flow cytometry. Results:Three days after CVB3 infection, the viral RNA load in pancreas was the highest (0.96±0.18) and gradually decreased with prolongation of infection. Compared with the 3 dpi group, the viral RNA load in pancreas was decreased (0.96±0.18 vs. 0.62±0.14) at 7 dpi, but there was no statistically significant difference. In addition, the infiltration of immune cell in pancreas increased significantly after 7dpi and the pathological score >2. The percent of infiltrated Th17 cells (1.05±0.21 vs. 22.13±5.79) and Treg cells (3.11±0.78 vs. 8.25±1.30) among CD4 + T cells significantly increased after infection (P<0.05), and the Th17/Treg also increased (P<0.01). Compared with the control group, the relative mRNA expression of IFN-γ (1.05±0.23 vs. 672.6±47.67), IL-6 (1.00±0.38 vs. 68.28±4.57), and IL-17 (1.01±0.11 vs. 54.15±7.94) in pancreas increased at 7 days after CVB3 infection ( P<0.01). Conclusions:The infiltration of Th17/Treg cells and the expression of related cytokines related cytokines IL-6 and IL-17 mRNA were upregulated in pancreas, which promoted the process of CVB3-induced pancreatitis.
10.Analysis and prediction of the association between respiratory syncytial virus infection and meteorological factors in Xuzhou city from 2015 to 2021
Rundong CAO ; Dong XIA ; Qinqin SONG ; Juan SONG ; Zhiqiang XIA ; Mi LIU ; Haijun DU ; Renhe ZHU ; Jun HAN ; Chen GAO
Chinese Journal of Experimental and Clinical Virology 2023;37(2):152-158
Objective:To analyze the association of respiratory syncytial virus infection with meteorological factors and to predict and explain the trends.Methods:Data of cases with severe acute respiratory infections in hospitalized children in Xuzhou City were collected from 2015-2021. Respiratory syncytial virus (RSV) was detected by real-time fluorescence polymerase chain reaction. The result were statistically analyzed using SPSS 26.0 software, including constructing a negative binomial regression model to explore meteorological factors that impact RSV detection and a multivariate time series model to predict its epidemiological trend from 2020 to 2021.Results:A total of 1 663 samples of children with severe acute respiratory infections were collected from 2015to 2021, of which 218 (13.1%) were positive for RSV. Seasonal effects on RSV detection were evident: there was a 1-year cycle with a peak in winter (December-February) and a trough in summer (June-August). The negative binomial regression analysis showed that monthly mean temperature, monthly mean relative humidity, and monthly total sunshine hours may influenced RSV detection. The prediction result of the time series model with sunshine hours as the covariate showed that the prediction was better for 2020, and the actual values were close to the predicted values. The expected trends in 2021 were consistent, but the actual values were higher than predicted.Conclusions:Monthly mean temperature, monthly mean relative humidity, and monthly total sunshine hours may influence RSV detection in the Xuzhou region.A prediction model can be built using data from 2015-2019, where deviations in the predicted values for 2021, reflecting that disease prevalence is multifactorial correlated, suggest a possible rise in RSV prevalence in the future.

Result Analysis
Print
Save
E-mail