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 Zuogui Pill (左归丸) Decoction on Cerebral Perfusion and Corticospinal Tract Remodeling in Patients With Ischemic Stroke of the Liver-Kidney Yin Deficiency Syndrome
Xinyu XU ; Dan WU ; Xiaohui YAN ; Hao HUA ; Siqi MEI ; Zihan ZHOU ; Jinglin ZHAO ; Shiyu HE ; Xiaoyu FEI ; Yangfan LIU ; Yue SU ; Li LI ; Wenlei LI
Journal of Traditional Chinese Medicine 2026;67(21):2302-2310
ObjectiveTo evaluate the clinical efficacy of Zuogui Pill (左归丸,ZGP) decoction on neurological function, motor function, and activities of daily living in patients with ischemic stroke (IS) of liver-kidney yin deficiency syndrome, and to assess its potential mechanisms from the perspectives of cerebral perfusion and corticospinal tract (CST) remodeling. MethodsA total of 108 patients with IS of liver-kidney yin deficiency syndrome and onset time ≤14 days were enrolled and randomly divided into a control group and a treatment group, with 54 cases in each group. The control group received conventional western medicine, while the treatment group received ZGP decoction (one dose per day) in addition to the conventional treatment. Both groups were treated for 90 days. The National Institutes of Health Stroke Scale (NIHSS), modified Rankin Scale (mRS), modified Barthel Index (MBI), Fugl-Meyer Assessment (FMA), and Chinese medicine syndrome score (CMS) were evaluated before treatment and after (90±3) days of treatment. Three-dimensional arterial spin labeling (3D-ASL) and diffusion tensor imaging (DTI) were performed to obtain ipsilesional cerebral blood flow (CBF) in the unaffected and affected hemispheres, relative CBF (rCBF), and fractional anisotropy (FA) of the CST. Nodal analysis of FA values along the CST was also conducted. Correlation analysis was used to assess the relationships between post-treatment cerebral perfusion parameters (CBF and rCBF), and FA values with clinical neurological function scores. ResultsAfter treatment, both groups showed significantly decreased NIHSS, mRS and CMS scores, together with significantly increased FMA and MBI scores (P<0.01). Intergroup comparison showed that NIHSS, mRS, and CMS scores in the treatment group were significantly lower than those in the control group , while MBI and FMA scores were significantly higher (P<0.01). The 3D-ASL results revealed that CBF and rCBF values in the affected hemisphere significantly increased in both groups after treatment, and the treatment group had significantly higher values (P<0.01). DTI showed that the FA value of the CST on the affected side in the treatment group was significantly higher than that in the control group after treatment (P<0.05). Nodal analysis along the CST further demonstrated that the continuity of FA values at lesion-related nodes 45~99 in the treatment group was significantly better than that in the control group (P<0.01), with significant increases at nodes 65~71 and 76~81 compared with baseline (P<0.05 or P<0.01), whereas the control group showed a decreasing trend at nodes 80~93 (P<0.05). Correlation analysis indicated that CBF and rCBF values in the affected hemisphere were negatively correlated with NIHSS and mRS scores and positively correlated with MBI score (P<0.05). FA value of the CST on the affected side, both before and after treatment, were negatively correlated with NIHSS score and positively correlated with FMA score (P<0.05 or P<0.01). ConclusionZGP decoction can effectively improve neurological function, motor function, and activities of daily living in IS patients with liver-kidney yin deficiency syndrome, and its mechanism may be related to enhanced cerebral perfusion in the affected hemisphere and corticospinal tract structural remodeling.
3.Application of Digital Nutrition Technologies in Adult Weight Management
Zihan XIONG ; Ning FAN ; Yue ZHAO ; Yuanyuan YAN ; Hui WANG
Journal of Sichuan University (Medical Sciences) 2025;56(4):900-906
Obesity and overweight remain major global public health challenges.In this paper,we examine current applications of digital nutrition technologies in weight management using the scoping review framework of Arksey and O'Malley.Visual recognition technologies have evolved from early-stage food image recognition into more advanced systems,although the vast diversity of food types continues to pose challenges for accuracy and generalizability.The combination of wearable devices and non-wearable sensors has increased the diversity of collected data and improved user comfort.Data integration and analysis tools,such as digital virtual twins,support personalized interventions by integrating multidimensional data.In addition,emerging applications have demonstrated encouraging clinical outcomes.However,existing studies have limitations,such as small sample sizes,high levels of information bias,and low evidence quality.Barriers,such as limited accessibility and high initial costs,further constrain the scalability of digital nutrition tools.Future studies should focus on expanding sample sizes,improving population diversity,and developing effective assessment methods and adjustments for specific target populations.Privacy and data security concerns must also be addressed to ensure safe implementation.Overall,digital nutrition technologies offer a promising approach to weight management,but continuous efforts in research and cost reduction are needed to fully realize their potential in weight management and to promote health for all.
4.Establishment of amachine learning-based precision recruitment method at the county level
Xiaoyan FU ; Zihan ZHANG ; Fang ZHAO ; Chunlan ZHOU ; Wenbiao LIANG ; Cheng YU ; Yingzhi YAN ; Wei SI ; Weibin TAN ; Hui XUE
Chinese Journal of Blood Transfusion 2025;38(12):1752-1758
Objective: To establish a machine learning-based precision blood donor recruitment model at the county level and assess its generalizability and applicability. Methods: A retrospective study was conducted using blood donation and SMS recruitment data from the Taicang Branch of the Suzhou Blood Center between 2019 and 2024. Multiple machine learning algorithms were employed, including extreme gradient boosting, support vector machine, k-nearest neighbor, logistic regression, decision tree, random forest, and multilayer perceptron. These were combined with techniques such as synthetic minority oversampling, undersampling, and cost-sensitive learning (using MFE and MSFE loss functions). Model parameters were optimized through grid search to identify the best-performing model. Results: In a prospective comparative study against conventional methods, the machine learning models increased the recruitment success rate among high-willingness donors by an average of 129.15%, and the recruitment efficiency per SMS improved by 125.02% compared with the traditional method. Under full-scale SMS sending, the recruitment rate per SMS increased by 42.61%, and SMS sending efficiency improved by 31.77%, significantly enhancing recruitment performance. Conclusion: This study represents the first application of a machine learning-based precision donor recruitment model at the county-level in China. The precise recruitment framework not only improves recruitment efficiency and reduces recruitment costs but also demonstrates strong scalability and generalizability. It provides a scientific and feasible intelligent pathway to ensure the safety and sustainability of the blood supply.
5.Effects of Peiyuan Yishen Antai Decoction on Cell Senescence and Decidualization of Aged Mice with Recurrent Spontaneous Abortion Based on P53/P21 Signaling Pathway
Chang SHU ; Yanfeng LIU ; Fei YAN ; Nan DENG ; Fanhui JIN ; Linlin GUO ; Zihan XU
Chinese Journal of Information on Traditional Chinese Medicine 2025;32(1):91-97
Objective To explore the effects of Peiyuan Yishen Antai Decoction on cell senescence and decidualization of aged mice with recurrent spontaneous abortion(RSA)based on P53/P21 signaling pathway;To discuss its mechanism in the treatment of RSA.Methods 8-9 month old CBA/J female mice were randomly divided into aging model group(normal saline 0.3 mL/d),Chinese materia medica group(18.72 g/kg)and dydrogesterone group(2.6 mg/kg).6-8 week old CBA/J female mice were assigned to a reproductive age blank group(normal saline 0.3 mL/d),with five mice in each group.After administering the corresponding solutions by gavage for 2 weeks,the reproductive age blank group mice were housed in cages with BALB/c male mice,while the remaining groups were housed in cages with DBA/2 male mice.After confirming the presence of vaginal plug,gavage was continued until the 9th day of pregnancy for sampling.Mice embryo loss rate was observed,HE staining was used to observe the morphology of uterine decidua,immunohistochemistry was used to detect the localization and expression of aging related protein Pi6INK4a in decidua tissue,RT-PCR and Western blot were used to detect the mRNA and protein expressions of P53,P21,insulin-like growth factor binding protein 1(IGFBP-1)and prolactin(PRL)in decidual tissue.Results Compared with the reproductive age blank group,the embryo loss rate of the aging model group significantly increased(P<0.05),the irregular endometrial cells were vacuolated,the number of glands and blood vessels was reduced,the expression of P16INK4a in decidual tissue significantly increased(P<0.05),and the expressions of P53,P21 mRNA and protein significantly increased(P<0.05),the mRNA and protein expressions of IGFBP-1 and PRL significantly decreased(P<0.05).Compared with the aging model group,the Chinese materia medica group and the dydrogesterone group showed an decreasing trend in embryo loss rate in mice,the arrangement of cells in the decidua tissue was relatively neat,and the number of blood vessels and glands increased,the expression of P16INK4a in decidua tissue was significantly decreased(P<0.05),the mRNA and protein expressions of P53 and P21 were significantly decreased(P<0.05),while the mRNA and protein expressions of IGFBP-1 and PRL significantly increased(P<0.05).Conclusion Peiyuan Yishen Antai Decoction can improve the aging status of endometrial stromal cells in aged RSA model mice by regulating the P53/P21 signaling pathway,promote decidualization process,and thus exert a protective effect on pregnancy.
6.Mechanism analysis of platelet activation induced by V. vulnificus hemolysin.
Yan WANG ; Zihan FENG ; Yaru WANG ; Shiqing LI ; Xin CHEN ; Jinglin WANG ; Yuan YUAN
Chinese Journal of Cellular and Molecular Immunology 2025;41(2):134-142
Objective To evaluate whether Vibrio vulnificus secreted exotoxin-hemolysin (VVH) can activate platelet, an important blood immune cell, and to explore the possible molecular mechanism of platelet activation by VVH. Methods Transcriptomics and immunohistochemistry were used to analyze whether Vibrio vulnificus infection caused platelet activation in mice. Then, flow cytometry was used to identify whether VVH was the main stimulator of platelet activation. Naturally expressed VVH toxin was purified and prepared. The effects of extracellular and intracellular Ca2+ signal inhibitors on VVH activated platelets were evaluated by flow cytometry and Western blotting. The immune activation effect of VVH in the early stage of Vibrio vulnificus infection was analyzed in vivo. Results VVH was the main stimulator of platelet activation in Vibrio vulnificus culture supernatant. Natural VVH can induce the increase of P-selectin (CD62P) on platelet surface, the formation of platelet-neutrophil complex (PNC), and the release of platelet microvesicles. The activation mechanism may be related to the VVH pore-dependent Ca2+-calmodulin (CaM) -myosin light chain kinase (MLCK) signaling pathway, which led to the release of platelet alpha particles and cascade activation of platelets. In a mouse model of ALD infected by Vibrio vulnificus gavage, VVH was strongly associated with platelet activation. Conclusion This study shows that VVH is an important platelet activating molecule in the early stage of Vibrio vulnificus infection, and its induction of platelet activation may be related to the pathogenic process.
Animals
;
Platelet Activation/drug effects*
;
Hemolysin Proteins/pharmacology*
;
Vibrio vulnificus/metabolism*
;
Mice
;
Blood Platelets/drug effects*
;
Vibrio Infections/immunology*
;
P-Selectin/metabolism*
;
Bacterial Proteins
;
Female
7.Structures and functions of the MICOS: Pathogenesis and therapeutic implications in Alzheimer's disease.
Zihan WANG ; Kaige ZHANG ; Minghao HUANG ; Dehao SHANG ; Xiaomin HE ; Zhou WU ; Xu YAN ; Xinwen ZHANG
Acta Pharmaceutica Sinica B 2025;15(6):2966-2984
Mitochondrial dysfunction is a critical factor in the pathogenesis of Alzheimer's disease (AD). The mitochondrial contact site and cristae organizing system (MICOS) plays a pivotal role in shaping the inner mitochondrial membrane, forming cristae junctions and establishing interaction sites between the inner and outer mitochondrial membranes and thereby serving as a cornerstone of mitochondrial structure and function. In the past decade, MICOS abnormalities have been extensively linked to AD pathogenesis. In particular, dysregulated expression of MICOS subunits and mutations in MICOS-related genes have been identified in AD, often in association with hallmark pathological features such as amyloid-β plaque accumulation, neurofibrillary tangle formation, and neuronal apoptosis. Furthermore, MICOS subunits interact with several etiologically relevant proteins, significantly influencing AD progression. The intricate crosstalk between these proteins and MICOS subunits underscores the relevance of MICOS dysfunction in AD. Therapeutic strategies targeting MICOS subunits or their interacting proteins may offer novel approaches for AD treatment. In the present review, we introduce current understanding of MICOS structures and functions, highlight MICOS pathogenesis in AD, and summarize the available MICOS-targeting drugs potentially useful for AD.
8.Delivery of Sophora flavescens Ait. using a dissolving microneedle enables enhanced psoriasis treatment
Zihan Zhou ; Jie Zhang ; Yiwen Chen ; Bingbing Wang ; Ping Hou ; Zifan Ding ; Luzheng Zhang ; Jianlin Wang ; Nailiang Yang ; Cong Yan
Journal of Traditional Chinese Medical Sciences 2025;2025(2):277-286
ObjectiveTo assess the efficiency of a Sophora flavescens Ait (S. flavescens, Ku Shen)-soluble microneedle (SFA-MN) for improving skin lesion symptoms in mice with psoriasis.MethodsSFA-MNs were prepared using a two-mold molding process with 20% w/v polyvinylpyrrolidone and 15% w/v polyvinyl alcohol. The SFA-MNs were assessed for morphology, mechanical properties, in vitro dissolution, identification of components, and skin lesion improvement in imiquimod-induced psoriasis mice.ResultsThe SFA-MNs demonstrated good mechanical properties for efficiently penetrating the dermis, facilitating efficient drug delivery. Furthermore, they effectively inhibited mast cell levels in the dorsal lesion area of psoriasis mice and reduced the expression of the T-lymphocyte factor cluster of differentiation 3 and tumor necrosis factor-α. In addition, this system alleviated skin inflammation, splenic swelling, and thymic atrophy in the psoriasis-like mouse model. Seven major components were detected from SFA-MNs by comparison of the mass-to-nucleus ratios (m/z) of the secondary fragments N-methylcytisine, 5α, 9α-dihydroxymatrine, sophoramine, matrine, oxysophocarpine, oxymatrine, and kushenol O.ConclusionThe drug delivery strategy combining traditional herbal S. flavescens with soluble microneedle technology provides more targeted and effective immune regulation for treating psoriasis-like mice models, enabling enhanced therapeutic effects compared with the control group.
9.Impact of lymphovascular space invasion on prognosis and adjuvant therapy in early-stage endometrial carcinoma
Zihan YAN ; Yunlong SHENG ; Xiaorong HOU ; Ke HU ; Fuquan ZHANG
Chinese Journal of Radiation Oncology 2025;34(5):502-507
Endometrial carcinoma (EC) is one of the most common gynecologic malignancies, with the majority of patients diagnosed at an early stage. Surgery is the primary treatment modality, and postoperative risk stratification plays a key role in guiding adjuvant therapy decisions. Lymphovascular space invasion (LVSI) is an important pathological feature that affects the prognosis of patients and serves as one of the key factors in risk stratification. This review summarizes the prognostic significance of LVSI in patients with early-stage EC and its impact on the selection of adjuvant therapies, aiming to provide new insights for exploring precise and individualized diagnosis and treatment.
10.Construction of device-related pressure injury prevention program in orthopedics patients based on evidence-based and SSKINS clinical management model
Zihan HE ; Huiwen WANG ; Rong YAN ; Suyun LI ; Yunfang LIU ; Yali WAN ; Rong PI ; Fang LIU ; Yan WANG
Chinese Journal of Practical Nursing 2025;41(8):593-600
Objective:To construct a device-related pressure injury prevention program for orthopedic patients based on evidence-based and SSKINS (S: skin, S: surface, K: keep moving, I: incontinence, N: nutrition, S: self-care) clinical management model, and to provide reference for clinical nursing practice.Methods:In June 2023, a research team was set up to retrieve the literature related to device-related pressure injury in orthopedic patients, and the evidence was extracted and screened with reference to the SSKINS clinical management model, and the first draft of the protocol was formulated based on clinical practice. The final plan was determined through two rounds of expert consultations from October to December 2023.Results:The response rate of the two rounds of expert correspondence was 100%(20/20). The expert authority coefficients were 0.897 and 0.907, the overall Kendall coordination coefficients of the two rounds were 0.116 and 0.136 respectively, and the Kendall coordination coefficients of all levels were 0.104-0.213 (all P<0.001). The final protocol contains 6 first-level items (assessment and examination, the use and adjustment of device, prophylactic dressings, skin protection, nutritional support, health education) and 31 second-level items. Conclusions:The proposed orthopedic device-related pressure injury prevention program is scientific and practical, and can provide a reference for the prevention and management of clinical medical staff.


Result Analysis
Print
Save
E-mail