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.Ultrashort echo time-MRI for evaluating pulmonary nodules
Shuo LIU ; Chao HAN ; Ying WANG ; Yingying CUI ; Nan MENG ; Kaixin LI ; Meiyun WANG
Chinese Journal of Medical Imaging Technology 2025;41(9):1531-1534
Objective To observe the value of ultrashort echo time(UTE)-MRI for evaluating pulmonary nodules.Methods Totally 58 patients with pulmonary nodules detected with CT were prospectively enrolled,and UTE-MRI was performed.Taken CT as the referent standard,the value of UTE-MRI for evaluating the diameter,composition,lung imaging reporting and data system(Lung-RADS)type and radiographic signs of pulmonary nodules was analyzed.Results CT detected 66 pulmonary nodules with a diameter of(11.60±5.20)mm in 58 patients,including 29 solid nodules,24 partially solid nodules and 13 ground-glass nodules.There were 14 Lung-RADS type 2,12 type 3,12 type 4A,11 type 4B and 17 type 4X lung nodules,among which 12 were found with lobulated sign,and 14 were found with spiculated sign.UTE-MRI detected 63 nodules with a diameter of(11.34±4.82)mm,including 25 solid nodules,28 partially solid nodules and 10 ground-glass nodules.There were 12 Lung-RADS type 2,12 type 3,12 type 4A,11 type 4B and 16 type 4X lung nodules.Among which 10 nodules were found with lobulated sign and 11 were found with spiculated sign.No significant difference of the measured diameters of lung nodules was found between UTE-MRI and CT(P=0.803),and the results of UTE-MRI and CT had good correlation and consistency(rs=0.953,ICC=0.946),which also had good consistency for assessing nodule composition and Lung-RADS type(Kappa=0.871,0.960).Meanwhile,no significant difference of the display rates of lobulated nor spiculated signs was noticed between UTE-MRI and CT(both P>0.05).Conclusion UTE-MRI was helpful for evaluating pulmonary nodules,with efficacy comparable to CT.
3.The clinical effect of three dimensional print guided F/B-TEVAR vascular endoluminal repair of anastomotic leaks after open surgery for type A aortic dissection
Yu ZHOU ; Yuexue HAN ; Jianhang HU ; Tao TANG ; Lili SUN ; Wendong LI ; Nan HU ; Chen LIU ; Jun SHAO ; Hao YU ; Zhao LIU
Chinese Journal of Surgery 2025;63(9):836-841
Objective:To evaluate the clinical value of three-dimensional(3D) printing-assisted fenestrated/branched endovascular aortic repair (F/B-TEVAR) for the treatment of endoleak after open surgery for type A aortic dissection.Methods:A multi-center retrospective case series analysis was conducted on 16 patients with anastomotic leakage following Stanford type A aortic dissection open repair, admitted to 12 medical centers between January 2019 and December 2023. All surgeries were led by the vascular surgery team from Department of Vascular Surgery, Nanjing Drum Tower Hospital. The study included 12 males and 4 females, with an age of (58.1±8.2) years (range: 42 to 75 years). Preoperative patient-specific 3D-printed models or 3D parametric surface topological guides were created based on aortic CT angiography data. These models assisted intraoperative external positioning of fenestration sites, combined with stent diameter selection and inner branch techniques to complete endovascular repair. Surgical procedures, complications, and clinical outcomes were evaluated, with follow-up CT imaging to assess efficacy.Results:All patients successfully underwent surgery without conversion to open repair. One patient had distal stent migration from a prior open repair, requiring intraoperative coverage of the main stent window and conversion to an in situ fenestration procedure. The mean operative time was (332.6±111.2)minutes (range: 80 to 460 minutes). No renal failure, paraplegia, or branch artery loss occurred. Postoperative follow-up ( M(IQR)) was 18(18) months(range: 6 to 36 months), with follow-up rates of 16/16 at 6 months, 10/16 at 12 months, 8/16 at 24 months, and 2/16 at 36 months. During follow-up, endoleak occurred in 3 patients, cerebral infarction in 1 patient, and death in 1 patient. The remaining patients demonstrated stable stent positioning, patent branches, and no endoleak. Conclusion:3D technology-guided F/B-TEVAR shows favorable mid-to short-term outcomes in treating anastomotic leakage after open repair of type A aortic dissection.
4.Significance of basophil levels in prognostic evaluation of intra-abdominal infection
Ming-min PANG ; Shao-hua FAN ; Mei-chen YAN ; Bao LIU ; Ju YANG ; Ya-nan LI ; Shi-han ZHANG ; Ting-yu MENG ; Tao GAO
Chinese Journal of Current Advances in General Surgery 2025;28(5):367-372
Objective:To assess the relationship between basophil levels and mortality in patients with intra-abdominal infection.Methods:Information on patients with intraperitoneal infection admitted to the intensive care unit were extracted from the MIMIC database.A time-dependent Cox regression model was used to adjust for confounders associated with 28-day mortality.Propensity score matching(PSM)was used to balance the baseline differences be-tween groups with different basophil levels,and a restricted cube chart(RCS)was used to show the relationship between basophil count and 28-day mortality in patients with intra-abdominal infection.Results:A total of 4403 patients with intra-abdominal infection were enrolled in the MIMIC database.Patients with high basophil levels have lower mortality than those with low basophil levels.There was an L-shaped curve between basophil level and 28-day mortality,with a cut-off value of 0.47×109/L.Cox regression analysis showed that basophil levels were an independent protective factor for mortal-ity in patients with intra-abdominal infection after adjusting for potential confounders(HR=0.586,95%CI:0.443-0.769).Protective factors for death at basophil levels remained after PSM adjusted for potential confounders(HR=0.628,95%CI:0.470-0.832).Conclusion:Basophil level is an independent protective factor for mortality in patients with intra-abdominal infection,and basophil levels should be dynamically monitored to better evaluate the prognosis of patients.
5.Analysis of the Funding Status and Research Focus in Field of Anti-Tumor Drugs Pharmacology in Projects Funded byNational Natural Science Foundation of China from 2015 to 2024
Yue LIU ; Nan ZHENG ; Li-wei HAN
Chinese Pharmacological Bulletin 2025;41(12):2201-2211
Aim To delineate the research hotspots in the National Natural Science Foundation of China(NS-FC)in the field of Anti-Tumor Drugs Pharmacology(application code H3505)from 2015 to 2024,and to identify common deficiencies in grant applications,with the aim of promoting high-quality development of original anti-tumor targets and innovative drug research in China.Methods We compiled NSFC data for 2015-2024 on General Program,Young Scientist Fund and Fund for Less Developed Regions in the field of Anti-Tumor Drugs Pharmacology,examining appli-cation and funding trends.A multidimensional analysis was conducted on project topics,disciplinary hot-spots,interdisciplinary integration,and the adoption of emerging technologies and methodologies.Results A total of 798 projects were funded,with direct alloca-tions of RMB 293.52 million.Funded research con-centrated on drug resistance,novel target discovery,immunotherapy,and the tumor microenvironment;key mechanisms included programmed cell death,immune checkpoints,and metabolic reprogramming;and ad-vanced approaches-microbiome and multi-omics inte-gration,organoid models,and AI-assisted target min-ing-were widely adopted.Unfunded proposals most often suffered from insufficient innovation,weak pre-liminary data,and poorly defined scientific questions.Conclusions Antitumor pharmacology in China has transitioned from nonspecific cytotoxic strategies to pre-cision,multi-modal approaches exemplified by nucleic acid drugs,immunotherapies,and multitarget inhibi-tors.Continued emphasis on original target discovery,mechanistic innovation,and interdisciplinary,technol-ogy-driven research will be critical to advancing the field.
6.Mechanism of action of Sterculiae Lychnophorae Semen against PM2.5-induced acute lung injury based on network pharmacology,molecular docking and experimental validation
Fan ZHANG ; Yi-fan DU ; Xiao-shu DENG ; Zu-feng ZHANG ; Xian-lei HAN ; Wei TIAN ; Xiu-mei LI ; Mian CHEN ; Fei LIU ; Nan WANG
Chinese Pharmacological Bulletin 2025;41(12):2362-2369
Aim To investigate the anti-acute lung injury(ALI)effect of Sterculiae Lychnophorae Semen(SLS)and its mechanism.Methods The main ac-tive components of SLS and their core targets and path-ways of action against ALI were obtained by network pharmacology methods.Subsequently,molecular doc-king technology and in vitro cellular experiments were applied for validation.Results A total of 19 core tar-gets were obtained,including HSP90AA1,CASP3,TNF,MAPK8 and MAPK14.The mechanisms may in-volve signaling pathways such as cancer,PI3K/Akt and MAPK.Molecular docking confirmed that the key targets of SLS formed a better binding activity with the relevant active ingredients.The in vitro results showed that SLS was able to protect the PM2.5-contaminated BEAS-2B cells,inhibit their NO,IL-1β and TNF-αlevels,and reduce the expression of p-p38 MAPK and p-JNK proteins.Conclusions The study successfully predicts the active ingredients,targets and signaling pathways of SLS against ALI,and in vitro experiments demonstrate that SLS might protect BEAS-2B cells from PM2.5 stimulus-induced inflammation and apoptosis by inhibiting the over-activation of p38 MAPK and JNK signaling pathways.
7.Effects of scaffolding teaching mode from the perspective of deep learning in basic nursing techniques course
Fangman CHENG ; Han LYU ; Nan SHENG ; Binqian GE ; Ying LIU ; Chunmei ZHANG ; Ping ZHOU
Chinese Journal of Modern Nursing 2025;31(32):4464-4469
Objective:To explore the effectiveness of the scaffolding teaching mode from the perspective of deep learning in basic nursing techniques course.Methods:From September 2023 to June 2024, 324 nursing students at Suzhou Vocational Health College in the 2023 cohort were selected as research subjects using convenience sampling. Three classes were designated as experimental group ( n=167), and three classes were designated as control group ( n=157). Control group received the traditional teaching mode, while experimental group received a scaffolding teaching mode from the perspective of deep learning on the basis of control group. Teaching effectiveness was evaluated using course grades, the Academic Self-Efficacy Scale, and the Measure Scale of Autonomous Learning Competencies of Nursing Undergraduates. Results:A total of 324 questionnaires were distributed and 324 valid questionnaires were collected, with a valid response rate of 100.00% (324/324). After teaching, the experimental group's nursing course grades were statistically higher than those of the control group ( P<0.05). After teaching, the experimental group's nursing students demonstrated statistically higher total scores and scores across all dimensions on the Academic Self-Efficacy Scale compared to the control group ( P<0.01). After teaching, the experimental group of nursing students scored statistically higher than the control group on the total score of the Measure Scale of Autonomous Learning Competencies of Nursing Undergraduates, as well as on the dimensions of self-management and learning cooperation ability ( P<0.01) . Conclusions:The implementation of the scaffolding teaching mode from the perspective of deep learning for basic nursing techniques course enhances vocational nursing students' academic performance, academic self-efficacy, and autonomous learning competencies.
8.Analysis of application and funded projects of National Natural Science Foundation of China in Clinical Pharmacology from 2015 to 2024
Nan ZHENG ; Yue LIU ; Li-wei HAN
The Chinese Journal of Clinical Pharmacology 2025;41(2):151-158
Objective To analyze the application and funded projects in the field of clinical pharmacology supported by the National Natural Science Foundation of China(NSFC)from 2015 to 2024,aiming to provide a reference for the application.Methods Information on the application and funded projects under the"Clinical Pharmacology(H3511)"code from 2015 to 2024 was collected from the NSFC management information system,including funding year,project name,project category,funding amount,keywords,research directions,etc.A database was established based on this information for statistical analysis.Results From 2015 to 2024,2 031 applications were received for general,young scientist,and regional scientist foundation in the clinical pharmacology research field,with 320 projects funded,accumulating a direct funding of 118.48 million yuan.The number and amount of funded projects showed a fluctuating upward trend.The diseases studied in the applications were mainly concentrated in malignant tumors,cardiovascular and cerebrovascular diseases,neuropsychiatric diseases,and metabolic diseases.The research directions of the applications and funded projects were mainly focused on personalized medication,drug adverse reactions and toxicity,and clinical biomarkers,accounting for 67.6%and 70.3%of the total number of application and funded projects,respectively.Conclusion With the emergence of new methods and technologies,clinical pharmacology has developed rapidly,and basic research in this field has increasingly attracted attention.The NSFC has always emphasized support for this field.However,there are still some issues of the applications,such as the need to strengthen clinical characteristics,uneven distribution of disease types and research directions,and unbalanced development within the discipline,which require increased attention from clinical pharmacology researchers.
9.Ultrashort echo time-MRI for evaluating pulmonary nodules
Shuo LIU ; Chao HAN ; Ying WANG ; Yingying CUI ; Nan MENG ; Kaixin LI ; Meiyun WANG
Chinese Journal of Medical Imaging Technology 2025;41(9):1531-1534
Objective To observe the value of ultrashort echo time(UTE)-MRI for evaluating pulmonary nodules.Methods Totally 58 patients with pulmonary nodules detected with CT were prospectively enrolled,and UTE-MRI was performed.Taken CT as the referent standard,the value of UTE-MRI for evaluating the diameter,composition,lung imaging reporting and data system(Lung-RADS)type and radiographic signs of pulmonary nodules was analyzed.Results CT detected 66 pulmonary nodules with a diameter of(11.60±5.20)mm in 58 patients,including 29 solid nodules,24 partially solid nodules and 13 ground-glass nodules.There were 14 Lung-RADS type 2,12 type 3,12 type 4A,11 type 4B and 17 type 4X lung nodules,among which 12 were found with lobulated sign,and 14 were found with spiculated sign.UTE-MRI detected 63 nodules with a diameter of(11.34±4.82)mm,including 25 solid nodules,28 partially solid nodules and 10 ground-glass nodules.There were 12 Lung-RADS type 2,12 type 3,12 type 4A,11 type 4B and 16 type 4X lung nodules.Among which 10 nodules were found with lobulated sign and 11 were found with spiculated sign.No significant difference of the measured diameters of lung nodules was found between UTE-MRI and CT(P=0.803),and the results of UTE-MRI and CT had good correlation and consistency(rs=0.953,ICC=0.946),which also had good consistency for assessing nodule composition and Lung-RADS type(Kappa=0.871,0.960).Meanwhile,no significant difference of the display rates of lobulated nor spiculated signs was noticed between UTE-MRI and CT(both P>0.05).Conclusion UTE-MRI was helpful for evaluating pulmonary nodules,with efficacy comparable to CT.
10.Significance of basophil levels in prognostic evaluation of intra-abdominal infection
Ming-min PANG ; Shao-hua FAN ; Mei-chen YAN ; Bao LIU ; Ju YANG ; Ya-nan LI ; Shi-han ZHANG ; Ting-yu MENG ; Tao GAO
Chinese Journal of Current Advances in General Surgery 2025;28(5):367-372
Objective:To assess the relationship between basophil levels and mortality in patients with intra-abdominal infection.Methods:Information on patients with intraperitoneal infection admitted to the intensive care unit were extracted from the MIMIC database.A time-dependent Cox regression model was used to adjust for confounders associated with 28-day mortality.Propensity score matching(PSM)was used to balance the baseline differences be-tween groups with different basophil levels,and a restricted cube chart(RCS)was used to show the relationship between basophil count and 28-day mortality in patients with intra-abdominal infection.Results:A total of 4403 patients with intra-abdominal infection were enrolled in the MIMIC database.Patients with high basophil levels have lower mortality than those with low basophil levels.There was an L-shaped curve between basophil level and 28-day mortality,with a cut-off value of 0.47×109/L.Cox regression analysis showed that basophil levels were an independent protective factor for mortal-ity in patients with intra-abdominal infection after adjusting for potential confounders(HR=0.586,95%CI:0.443-0.769).Protective factors for death at basophil levels remained after PSM adjusted for potential confounders(HR=0.628,95%CI:0.470-0.832).Conclusion:Basophil level is an independent protective factor for mortality in patients with intra-abdominal infection,and basophil levels should be dynamically monitored to better evaluate the prognosis of patients.

Result Analysis
Print
Save
E-mail