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.Effects of oral non-peptidic thrombopoietin receptor agonists on hepatic enzyme in adult patients with immune thrombocytopenia:a meta-analysis
Tiantian LU ; Nan SHEN ; Suyue ZHU ; Jingjing YAN
China Pharmacy 2026;37(4):510-515
OBJECTIVE To systematically evaluate the effects of oral non-peptidic thrombopoietin receptor agonists (TPO-RAs) on hepatic enzyme in adult patients with immune thrombocytopenia. METHODS A comprehensive literature search was conducted in PubMed, Web of Science, CNKI, Wanfang database and the Chinese Medical Association Journal Full-Text Database to collect randomized controlled trials (RCTs) comparing oral non-peptidic TPO-RAs (intervention group) with placebo or conventional therapy (control group). All databases were searched from their inception to June 2025. After literature screening, data extraction and quality assessment of the included studies, meta-analysis was conducted using RevMan 5.4.1 software. RESULTS Twelve RCTs comprising 1 388 patients were included, with 971 in the intervention group and 417 in the control group. Meta-analysis results showed that there were no significant differences between the two groups in terms of the incidence of hepatic enzyme elevation[OR=1.24, 95%CI (0.77, 1.99), P =0.37 ] , the incidence of hepatic enzyme elevation in patients treated for ≥6 weeks[OR=1.21, 95%CI (0.73, 1.99), P =0.46 ] , and the incidence of severe hepatic enzyme elevation[OR=1.39, 95%CI(0.46, 4.20), P =0.55 ] . Subgroup analysis showed that there were no significant differences in the incidence of hepatic enzyme elevation between the intervention group and control group among patients using eltrombopag[OR=1.57,95%CI(0.85,2.87), P =0.15 ] , avatrombopag[OR=0.88,95%CI (0.09,8.46), P =0.91 ] , and hetrombopag[OR=1.04,95%CI(0.30,3.65), P =0.95 ] , respectively. CONCLUSIONS Oral non-peptidic TPO-RAs do not significantly increase the risk of hepatic enzyme elevation in adult patients with immune thrombocytopenia, and show an overall favorable hepatic safety profile.
3.Application advances, ethical dilemmas, and future directions of large language models in lung cancer diagnosis and treatment
Zhizhen REN ; Yufan XI ; Xu ZHU ; Yijie LUO ; Geting HUANG ; Junqiao SONG ; Xiuyuan XU ; Nan CHEN ; Qiang PU
Chinese Journal of Clinical Thoracic and Cardiovascular Surgery 2026;33(03):353-362
Lung cancer is a leading cause of cancer-related morbidity and mortality worldwide. Coupled with the substantial workload, the clinical management of lung cancer is challenged by the critical need to efficiently and accurately process increasingly complex medical information. In recent years, large language models (LLMs) technology has undergone explosive development, demonstrating unique advantages in handling complex medical data by leveraging its powerful natural language processing capabilities, and its application value in the field of lung cancer diagnosis and treatment is continuously increasing. The paper systematically analyzes that the exceptional potential of LLMs in lung cancer auxiliary diagnosis, tumor feature extraction, automatic staging, progression/outcome analysis, treatment recommendations, medical documentation generation, and patient education. However, they face critical technical and ethical challenges including inconsistent performance in complex integrated decision-making (e.g., TNM staging, personalized treatment suggestions) and "black box" opacity issues, along with dilemmas such as training data biases, model hallucinations, data privacy concerns, and cross-lingual adaptation challenges ("data colonization"). Future directions should prioritize constructing high-quality multimodal corpora specific to lung cancer, developing interpretable and compliant specialized models, and achieving seamless integration with existing clinical workflows. Through dual drivers of technological innovation and ethical standardization, LLMs should be prudently advanced for holistic lung cancer management processes, ultimately promoting efficient, standardized, and personalized diagnosis and treatment practices.
4.Effect and Mechanism of Liangyi Paste on Hepatic Lipid Deposition in Naturally Aged Mice with High-fat Diet via Cuproptosis/Oxidative Stress Pathway
Meiling ZHANG ; Yuanguang DONG ; Xiaofei SUN ; Jiaxin WANG ; Yu LIU ; Jingxuan ZHU ; Qun WANG ; Nan SONG ; Guoyuan SUI ; Lianqun JIA
Chinese Journal of Experimental Traditional Medical Formulae 2026;32(9):91-99
ObjectiveTaking the cuproptosis/oxidative stress pathway as the entry point, this study investigated the effect and mechanism of Liangyi Paste on hepatic lipid deposition in naturally aged mice fed with a high-fat diet. MethodsAfter adaptive feeding, 80 ten-week-old male C57BL/6 mice were used. Thirty of them were randomly divided into three groups (10 mice per group): The 12-month-old control group (12MCON), the 15-month-old control group (15MCON), and the 15-month-old group with a high-fat diet (15MHFD). The 12MCON and 15MCON groups were continuously fed a standard diet, while the 15MHFD group started receiving a high-fat diet at 12 months of age. Tissue samples were collected at the corresponding time points for each group. The remaining 50 mice were randomly divided into five groups (10 mice per group): the 20-month-old control group (20MCON), the model group, and the low-, medium-, and high-dose Liangyi Paste groups (2.91 , 5.82 , 11.64 g·kg-1·d-1, respectively). The 20MCON group was continuously fed a standard diet, while the other groups started receiving a high-fat diet at 15 months of age. At 18 months of age, the Liangyi Paste groups were administered the corresponding doses of Liangyi Paste by gavage, while the 20MCON and model groups were given an equal volume of saline by gavage. After 8 weeks of continuous gavage (when the mice reached 20 months of age), tissue samples were collected. Hepatic TG levels were measured using assay kits; liver histology and lipid deposition were observed via hematoxylin-eosin (HE) and oil red O staining; reactive oxygen species (ROS) were detected by enzyme-linked immunosorbent assay (ELISA); Cu2+, superoxide dismutase (SOD), and malondialdehyde (MDA) levels were measured by colorimetry; mRNA and protein expression of genes related to cuproptosis and oxidative stress pathways were analyzed by Real-time polymerase chain reaction(Real-time PCR) and Wes automated protein expression system. ResultsCompared with 12MCON, the 15MCON group showed significantly increased hepatic TG, Cu2+, ROS, and MDA levels (P<0.01), decreased SOD (P<0.01), hepatocyte swelling, and disordered arrangement. The mRNA and protein levels of ferredoxin 1 (FDX1), dihydrolipoamide S-acetyltransferase (DLAT), heat shock protein 70 (HSP70), dihydrolipoamide dehydrogenase (DLD), pyruvate dehydrogenase E1 subunit-β (PDHB), nuclear factor erythroid 2-related factor 2 (Nrf2), and peroxisome proliferator-activated receptor γ (PPARγ) were significantly elevated (P<0.05, P<0.01). Compared with 15MCON group, the 15MHFD and 20MCON groups exhibited further increases in TG, Cu2+, ROS, and MDA (P<0.01), reduced SOD (P<0.01), and aggravated hepatocyte swelling and disorder. There were increased lipid droplets with mild vacuolization in the 15MHFD group, and no significant lipid deposition was observed in the 20MCON group. FDX1, DLAT, HSP70, DLD, PDHB, Nrf2, and PPARγ mRNA and protein levels were significantly increased (P<0.05, P<0.01). Compared with 20MCON group, the model group demonstrated markedly elevated TG, Cu2+, ROS, and MDA (P<0.01), reduced SOD (P<0.01), severe hepatic steatosis, and upregulated expression of FDX1, DLAT, HSP70, DLD, PDHB, Nrf2, and PPARγ mRNA and proteins (P<0.05, P<0.01). All abnormalities were significantly reversed after Liangyi Paste treatment. ConclusionLiangyi paste can ameliorate hepatic lipid deposition in naturally aged mice with a high-fat diet by modulating the cuproptosis/oxidative stress pathway.
5.Relationship between gut microbiota and onset of depression in chronic unpredictable mild stress model rats of both sexes
Kanghong ZHU ; Yumeng GAO ; Mengxue HUANG ; Liu YANG ; Zizhan GAO ; Hao CHU ; Nan DENG ; Ling HU ; Zijian WU
Acta Laboratorium Animalis Scientia Sinica 2025;33(8):1144-1152
Objective To observe the differences in gut microbiota in chronic unpredictable mild stress(CUMS)-induced depression model rats of both sexes,and to provide experimental evidence for exploring sex differences in depression onset.Methods Thirty-two healthy SD rats were divided randomly into four groups based on sex:Male control group(Control-M),Female control group(Control-F),Male model group(Model-M),and Female model group(Model-F)(n=8 rats per group).Rats in the control groups were fed without stimulation,while rats in the model groups were stimulated using the 28 d CUMS-induced depression method.After successful modeling,fresh feces were collected from all rats for high-throughput 16S rRNA sequencing.Behavioral observations were also conducted before and after preparing the model.Results The result of sucrose-preference,open-field,and forced-swimming tests differed significantly between the control and model groups.The result of the sucrose-preference test also differed between the sexes,while there was no difference in the open-field or forced-swimming test between the sexes.The α and β diversity of the gut microbiota genera showed an upward trend in the CUMS group compared with the control group.The ratio of Firmicutes/Bacteroidetes and the richness of the Roseburia and Lachnospiraceae_NK4A136_group were decreased in male rats but showed an increasing trend in female rats.Conclusions The ratio of Firmicutes/Bacteroidetes in the gut microbiota may be a key factor affecting the difference in the onset of depression between males and females,while the Roseburia and Lachnospiraceae_NK4A136_group be potential factors in correcting the gut microbiota and improving the symptoms of depression.
6.Full genome analysis of G4P23porcine rotavirus and its pathogenicity in suckling mice and piglets
Hui DENG ; Ran TAO ; Nan HAN ; Jianxin WANG ; Xuefan SU ; Chen WANG ; Xi CHENG ; Xianyu BIAN ; Jiapeng SONG ; Xuejiao ZHU ; Xuehan ZHANG ; Hongbo XIAO ; Jinzhu ZHOU ; Bin LI
Chinese Journal of Zoonoses 2025;41(9):902-909
To perform the phylogenetic characterization of an isolated porcine rotavirus(PoRV)and investigate its pathogenicity in suckling mice and piglets.A G4P[23]genotype PoRV strain JSJR2023 was successfully isolated from the diarrheic piglet feces through propagation in MA104 cells.The viral proliferation kinetics were analyzed using TCID50 assays,followed by complete genome sequencing through Sanger sequencing platforms.Comprehensive genotyping and phylogenetic reconstruction were conducted using MEGA7.0 with maximum likelihood algorithms.Pathogenicity was assessed in the following animal models:5-day-old C57BL/6 mice and 3-day-old piglets.Multidimensional evaluation included clinical monitoring(diarrhea scoring,growth parameters),virological detection,and histopathological analysis of intestinal tissues.The virus strain JSJR2023 could replicate efficiently in MA104 cells,achieving peak titers of 107.5 TCID50/mL.Whole genome genotype analysis showed that the strain belonged to G4-P[23]-I5-R1-C1-M1-A8-N1-T1-E1-H1.Phylogenetic analysis indicated that the VP3 and NSP4 genes of JSJR2023 strain were most closedrelated to human species rotaviruses,suggesting genetic reassortment between human and porcine RV strains.The animal experiments in suckling mice showed that the JSJR2023 strain infection caused diarrhea symptoms,intestinal edema and congestion,and shedding of intestinal villus epithelial cells.The pathogenicity experiments in piglets showed that compared with the control group,the challenged group of pig-lets had severe diarrhea symptoms,accompanied by reduced appetite and listlessness.Post-mortem examination revealed that the intes-tines were significantly thinner,congested,and filled with yellow watery contents.The challenged piglets showed typical pathological changes such as thinning of the intestinal wall and shortening and shedding of intestinal villi.In conclusion,this study successfully iso-lated a human-porcine recombinant G4P[23]PoRV strain and established the infection models in suckling mice and piglets,providing important tools for investigating the pathogenic mechanism of PoRV,evaluating vaccines and developing antiviral drug.
7.Analysis of Clinical Characteristics and Prognostic Risk Factors of Mycoplasma pneumoniae Pneumonia in Adults of Different Age Groups
Jia ZHU ; Tao-mei ZHANG ; Qing-liu TAN ; Cong-hui CHEN ; Ya-nan MA
Progress in Modern Biomedicine 2025;25(15):2472-2477
Objective:To explore the clinical characteristics and prognostic risk factors of Mycoplasma pneumoniae pneumonia(MPP)in adults across different age groups,providing evidence for age-stratified management strategies.Methods:A retrospective analysis was conducted on 80 MPP patients admitted to the Respiratory Department of a hospital between January 2023 and December 2024.Patients were divided into three groups based on age:young adults(18-40 years),middle-aged adults(41-60 years),and elderly adults(≥ 61 years).Demographic features,clinical indicators,and mixed infections were analyzed.Univariate and multivariate Logistic regression were used to identify risk factors for disease severity.Results:The severity rate was significantly higher in the elderly group(41.2%)compared to the young adult group(7.1%)and middle-aged group(20.0%)(P=0.022).Elderly patients also exhibited significantly higher rates of underlying diseases(chronic lung disease:35.3%vs.3.6%in young adults),elevated inflammatory markers(C-reactive protein:68.3±19.5 mg/L vs.32.5±8.4 mg/L in young adults),mixed infections(52.9%vs.14.3%),and prolonged hospital stays(8.61±2.22 days vs.5.01±1.11 days)(P<0.05).Multivariate analysis identified age(OR=1.79 per 10 years),chronic lung disease(OR=3.25),blood urea nitrogen ≥6 mmol/L(OR=2.44),and mixed infections(OR=4.26)as independent risk factors for severe MPP(P<0.05).Conclusion:Clinical manifestations and prognoses of MPP in adults vary significantly across age groups.Elderly patients are characterized by high mixed infection rates,intense inflammatory responses,and renal function impairment,necessitating individualized monitoring and intervention strategies.
8.Effect of Guanxinning injection on myocardial infarction by regulating cardiac immunity through CCL21
Yu-xin BAI ; Ying-xue ZHANG ; Ting-ting SHI ; Si-nan ZHU ; Zhen-kun XU ; Hong WANG ; Lu CHEN
Chinese Pharmacological Bulletin 2025;41(5):960-969
Aim To investigate the mechanism of Guanxinning injection regulating cardiac immune mi-croenvironment to improve myocardial infarction in mice.Methods In this study,MI model was estab-lished by permanent ligation of left anterior descending coronary artery in mice.The mice were divided into five groups:sham operation group,model group,Guanxinning injection low dose group,Guanxinning in-jection high dose group and positive drug captopril group.Hearts were weighed,heart tissues were collect-ed,and Masson staining was used for pathological anal-ysis of heart tissues;immunofluorescence staining was used to detect apoptosis and CCL21 expression in the infarct border zone;flow cytometry was used to detect the proportion of immune cells in myocardial ischemia tissues and lymph nodes;PCR was used to detect CCL21 expression in heart and in vitro human lymphat-ic endothelial cells(HLEC).Results Compared with the model group,the low and high dose groups of Guanxinning injection significantly improved cardiac hypertrophy.Apoptosis in the border zone of myocardi-al infarction was reduced in the low and high dose groups of Guanxinning injection and captopril group.Compared with the model group,the proportion of leu-kocytes in the infarct border zone was dreduced and the proportion of CD4+T cells,Treg cells,and CD8+T cells in the mediastinal lymph nodes and infarct border zone of the heart was regulated in the low and high dose groups of Guanxinning injection;CCL21 secretion by the heart and lymphatic vessels increased.Conclu-sions Guanxinning injection can significantly improve cardiac hypertrophy and fibrosis in MI mice,reduce ap-optosis in the infarct border zone,and play a role in an-ti-myocardial ischemia injury by promoting CCL21 ex-pression in lymphatic vessels to regulate the proportion of mediastinal lymph nodes and cardiac T cells after myocardial infarction.
9.DNA methylation age prediction model based on ovarian granulosa cells
Peng LIU ; Bowen ZHU ; Yuheng LI ; Liang WANG ; Nan LIU ; Ningxia SUN
Chinese Journal of Reproduction and Contraception 2025;45(5):442-447
Objective:To develop a novel ovarian biological age prediction model based on DNA methylation for ovarian aging assessment.Methods:A prospective cohort study method was used. From March 2024 to January 2025, we collected 96 ovarian granulosa cell samples of infertility patients due to fallopian tube factors or male factors from the Reproductive Medicine Center of the Second Affiliated Hospital of Naval Medical University. Then we analyzed DNA methylation patterns across five age-associated gene regions ( ELOVL2, miR29B2C, TRIM59, KLF14 and FHL2) in a discovery cohort comprising 63 human ovarian granulosa cell samples. Targeted bisulfite sequencing was performed through PCR amplification followed by next-generation DNA sequencing. Leveraging elastic net regression analysis, we developed a predictive model incorporating 29 methylation sites that demonstrated strong age correlation. The model was subsequently validated using an independent cohort comprising 33 human ovarian granulosa cell samples. Results:The DNA methylation age prediction model based on ovarian granulosa cells showed the following results in the discovery cohort as follows: median absolute error (MAE) was 2.534 ( R=0.742, P<0.001). In the independent validation cohort, MAE was 3.019 ( R=0.729, P<0.001). Conclusion:In this study, we utilized human ovarian granulosa cells as experimental samples to develop a novel DNA methylation-based model for predicting ovarian biological age. By integrating multiple methylation sites across five age-related gene regions, this model serves as a robust indicator of ovarian aging status.
10.Construction and immune efficacy evaluation of BNeV VLPs based on VP1 protein in mice
Lu DING ; Xiangyue HUANG ; Jinbo WU ; Chaohui ZHANG ; Qing ZHU ; Chenxi ZHU ; Gu-nan DENG ; Ajia AKE ; Chunsai HE ; Yuanzhen MA ; Bin ZHANG
Chinese Journal of Veterinary Science 2025;45(3):412-419
The codon was optimized for the bovine nebovirus(BNeV)VP1 gene and the recombi-nant plasmid pFastBac-Dual-VP1 was constructed,and BNeV-VP1 virus-like particles(VLPs)were prepared using a baculovirus expression system,and identified by Western blot,indirect im-munofluorescence and electron microscopy.Successfully validated VLPs were mixed with MF59 adjuvant and CpG-ODG,and mice were immunised by intramuscular injection and evaluated for immunity effects.The results showed that the optimized CAI(codon adaptation index)of VP1 gene was 0.93 and the GC content was 60.4%.The constructed recombinant plasmid was trans-formed into DH10Bac for blue-white spot screening,and after successful verification,it was trans-fected into SF9 cells to obtain recombinant baculovirus Baculo-BNeV-VP1.BNeV virus-like parti-cles with diameters ranging from 35 to 40 nm were observed under the electron microscope,and both IFA and Western blot experiments proved that the target proteins were successfully ex-pressed and biologically active,and protein optimisation revealed that the highest protein expres-sion was found at the infectious dose MOI=0.5.Mice were immunized by intramuscular injection after 50 μg of VLPs were mixed with MF59 adjuvant and CpG-ODN.The results showed that the VLPs immunization group produced IgG antibodies 7 days after the first dose,and the antibody ti-ter increased gradually,reaching a maximum of 1∶102 400,and declined at 35 d,but still main-tained a high level;The blocking titer BT50 is up to 640,which can induce the production of BNeV VP1-specific blocking antibody in mice.In this study,the baculovirus expression system was used to express the VP1 protein of BNeV,and BNeV VLPs were successfully constructed,which could induce humoral immune response in mice,which provided a reference for the follow-up research of BNeV vaccine.

Result Analysis
Print
Save
E-mail