1.Prediction of risk of in-hospital death in patients with chronic heart failure complicated by lung infections using interpretable machine learning
Caiyu SHEN ; Shuai WANG ; Ruiying ZHOU ; Yuhe WANG ; Qin GAO ; Xingzhi CHEN ; Shu YANG
Journal of Southern Medical University 2024;44(6):1141-1148
Objective To predict the risk of in-hospital death in patients with chronic heart failure(CHF)complicated by lung infections using interpretable machine learning.Methods The clinical data of 1415 patients diagnosed with CHF complicated by lung infections were obtained from the MIMIC-IV database.According to the pathogen type,the patients were categorized into bacterial pneumonia and non-bacterial pneumonia groups,and their risks of in-hospital death were compared using Kaplan-Meier survival curves.Univariate analysis and LASSO regression were used to select the features for constructing LR,AdaBoost,XGBoost,and LightGBM models,and their performance was compared in terms of accuracy,precision,F1 value,and AUC.External validation of the models was performed using the data from eICU-CRD database.SHAP algorithm was applied for interpretive analysis of XGBoost model.Results Among the 4 constructed models,the XGBoost model showed the highest accuracy and F1 value for predicting the risk of in-hospital death in CHF patients with lung infections in the training set.In the external test set,the XGBoost model had an AUC of 0.691(95%CI:0.654-0.720)in bacterial pneumonia group and an AUC of 0.725(95%CI:0.577-0.782)in non-bacterial pneumonia group,and showed better predictive ability and stability than the other models.Conclusion The overall performance of the XGBoost model is superior to the other 3 models for predicting the risk of in-hospital death in CHF patients with lung infections.The SHAP algorithm provides a clear interpretation of the model to facilitate decision-making in clinical settings.
2.Prediction of risk of in-hospital death in patients with chronic heart failure complicated by lung infections using interpretable machine learning
Caiyu SHEN ; Shuai WANG ; Ruiying ZHOU ; Yuhe WANG ; Qin GAO ; Xingzhi CHEN ; Shu YANG
Journal of Southern Medical University 2024;44(6):1141-1148
Objective To predict the risk of in-hospital death in patients with chronic heart failure(CHF)complicated by lung infections using interpretable machine learning.Methods The clinical data of 1415 patients diagnosed with CHF complicated by lung infections were obtained from the MIMIC-IV database.According to the pathogen type,the patients were categorized into bacterial pneumonia and non-bacterial pneumonia groups,and their risks of in-hospital death were compared using Kaplan-Meier survival curves.Univariate analysis and LASSO regression were used to select the features for constructing LR,AdaBoost,XGBoost,and LightGBM models,and their performance was compared in terms of accuracy,precision,F1 value,and AUC.External validation of the models was performed using the data from eICU-CRD database.SHAP algorithm was applied for interpretive analysis of XGBoost model.Results Among the 4 constructed models,the XGBoost model showed the highest accuracy and F1 value for predicting the risk of in-hospital death in CHF patients with lung infections in the training set.In the external test set,the XGBoost model had an AUC of 0.691(95%CI:0.654-0.720)in bacterial pneumonia group and an AUC of 0.725(95%CI:0.577-0.782)in non-bacterial pneumonia group,and showed better predictive ability and stability than the other models.Conclusion The overall performance of the XGBoost model is superior to the other 3 models for predicting the risk of in-hospital death in CHF patients with lung infections.The SHAP algorithm provides a clear interpretation of the model to facilitate decision-making in clinical settings.
3.Comparisons in the changes of clinical characteristics and cerebrospinal fluid cytokine profiles between varicella-zoster virus meningitis/encephalitis and other central nervous system infections.
Caiyu MA ; Yuying LU ; Qinghua ZHANG ; Han CHEN ; Qingxia ZHANG ; Hao HU ; Zhi SONG ; Ru CHEN ; Ding LIU
Journal of Central South University(Medical Sciences) 2022;47(10):1345-1354
OBJECTIVES:
Varicella-zoster virus (VZV) is one of the most common etiologies of viral meningitis/encephalitis. The early clinical manifestations and cerebrospinal fluid (CSF) changes of VZV meningitis/encephalitis lack specificity, and it is easy to be misdiagnosed as other viral encephalitides or tuberculous meningitis. This study aims to investigate whether the clinical characteristics, CSF analysis findings, and CSF cytokine levels could distinguish VZV meningitis/encephalitis from central nervous system (CNS) herpes simplex virus (HSV) and Mycobacterium tuberculosis (MTB) infections.
METHODS:
The medical records from 157 CNS infections, including 49 HSV (45 HSV-1, 4 HSV-2), 55 VZV, and 53 MTB infections between January 2018 and June 2021 in the Cytology Laboratory, Department of Neurology, Third Xiangya Hospital of Central South University were retrospectively reviewed. The data of 3 groups included demographic characteristics, laboratory results, radiographic findings, and outcomes. The levels of 12 cytokines (IL-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12p70, IL-17, IFN-γ, IFN-α, and TNF-α) in the CSF of 68 patients (13 HSV, 22 VZV, and 33 MTB infection cases) were quantified. Clinical and laboratory data were compared among the 3 groups.
RESULTS:
The most common clinical manifestations in the 3 groups were fever, headache, vomiting, and neck stiffness. The clinical manifestations of HSV and VZV CNS disease were similar, although fever and altered consciousness were less common in the VZV group than those in the HSV and MTB groups (63.6% vs 87.8% vs 96.2%, P<0.001, and 14.5% vs 26.5% vs 47.2%, P=0.004, respectively). Seven patients (7/55, 12.7%) presented cutaneous zoster in the VZV group. CSF leukocyte count was significantly higher in the VZV group (230×106 cells/mL) and MTB groups (276×106 cells/mL) than that in the HSV group (87×106 cells/mL, P=0.002). CSF protein level was significantly higher in the VZV than that in the HSV group (1 034 mg/L vs 694 mg/L, P=0.011) but lower than that in the MTB group (1 744 mg/L, P<0.001). IL-6 (VZV vs HSV vs MTB: 2 855.93 pg/mL vs 2 128.26 pg/mL vs 354.77 pg/mL, P=0.029) and IL-8 (VZV vs HSV vs MTB: 4 001.46 pg/mL vs 1 578.11 pg/mL vs 1 023.25 pg/mL, P=0.046) levels were significantly different among the 3 groups and were elevated in the VZV group.Post hoc analysis revealed that IL-6 and IL-8 were significantly higher in the VZV group than those in the MTB group (P=0.002 and P=0.035, respectively), but not in the HSV group (P>0.05).
CONCLUSIONS
VZV meningitis/encephalitis presents with CSF hypercellularity and proteinemia, challenging the classical view of CSF profiles in viral encephalitis. CSF IL-6 and IL-8 levels are elevated in patients with VZV meningitis/encephalitis, indicating a more intense inflammatory response in these patients.
Humans
;
Central Nervous System Infections
;
Encephalitis
;
Encephalitis, Varicella Zoster/diagnosis*
;
Encephalitis, Viral/diagnosis*
;
Herpesvirus 3, Human
;
Interleukin-6
;
Interleukin-8
;
Meningitis
;
Retrospective Studies
4.Analysis of clinical characteristics and prognosis of 4 264 patients with asymptomatic and mild novel coronavirus infections in Shanghai
Xiangru XU ; Ding SUN ; Min CAO ; Wen ZHANG ; Yuting PU ; Caiyu CHEN ; Yuting SUN ; Shuang ZHOU ; Bangjiang FANG
Chinese Critical Care Medicine 2022;34(5):449-453
Objective:To analyze the clinical characteristics and prognosis of patients infected with novel coronavirus Omicron variant in Shanghai, as to provide a reference for epidemic prevention, clinical diagnosis, and treatment.Methods:Altogether 4 264 novel coronavirus Omicron variant-infected patients with positive results of nucleic acid admitted to Shanghai New International Expo Center N3 Mobile Cabin Hospital from April 2 to May 7, 2022, were included. The demographic and baseline clinical characteristics, treatment strategy, prognosis, and different factors affecting the length of hospital stay were analyzed.Results:A total of 4 264 novel coronavirus variant Omicron-infected cases were collected, including 3 111 cases (73.0%) asymptomatic infections and 1 153 cases (27.0%) mild infections. The overall median age was 45 (33, 55) years old with a range from 2 years old to 81 years old. The male to female ratio was 1.37∶1. Altogether 3 305 cases (77.5%) had been vaccinated, of which 3 166 cases completed more than 2 doses. The upper respiratory tract symptoms such as cough and expectoration were the most common clinical manifestations of these infected patients. During the course of the disease, patients with asymptomatic infection were mainly treated with traditional Chinese medicine (TCM, 55.1%) and clinical observation (36.8%), and those with mild infection were mainly treated with TCM (42.2%) or integrated Chinese and Western medicine (30.4%). All patients were cured and discharged. The overall median length of hospital stay and the negative conversion time of nucleic acid were 9 (6, 10) days and 8 (5, 9) days, respectively. Compared with the asymptomatic infected patients, the hospitalization duration and the nucleic acid negative conversion time of the mildly infected patients were slightly longer [days: 10 (8, 11) vs. 9 (5, 10); 8 (6, 10) vs. 7 (4, 9), both P < 0.001]. Multiple linear regression analysis showed that the increasing age and mild infection were associated with longer hospitalization duration, and the treatment of TCM or integrated Chinese and Western medicine was associated with shortened length of hospital stay (all P < 0.05). Conclusions:The current novel coronavirus Omicron variant epidemic in Shanghai mainly caused asymptomatic and mild infections. The young and middle-aged population had a relatively high infection rate. The upper respiratory tract symptoms such as cough and expectoration were the most common clinical symptoms. Elderly and confirmed patients had prolonged hospitalization duration, while for patients receiving TCM treatment, the hospitalization duration was shortened.
5.Defining a critical period in calvarial development for Hedgehog pathway antagonist-induced frontal bone dysplasia in mice.
Yuanjing JIANG ; Shixian ZHANG ; Chuanqing MAO ; Yongzhen LAI ; Di WU ; Hu ZHAO ; Caiyu LIAO ; Weihui CHEN
International Journal of Oral Science 2019;11(1):3-3
The Hedgehog (Hh) signalling pathway is essential for cellular proliferation and differentiation during embryonic development. Gain and loss of function of Hh signalling are known to result in an array of craniofacial malformations. To determine the critical period for Hh pathway antagonist-induced frontal bone hypoplasia, we examined patterns of dysmorphology caused by Hh signalling inhibition. Pregnant mice received a single oral administration of Hh signalling inhibitor GDC-0449 at 100 mg•kg or 150 mg•kg body weight at preselected time points between embryonic days (E)8.5 and 12.5. The optimal teratogenic concentration of GDC-0449 was determined to be 150 mg•kg. Exposure between E9.5 and E10.5 induced frontal bone dysplasia, micrognathia and limb defects, with administration at E10.5 producing the most pronounced effects. This model showed decreased ossification of the frontal bone with downregulation of Hh signalling. The osteoid thickness of the frontal bone was significantly reduced. The amount of neural crest-derived frontal bone primordium was reduced after GDC-0449 exposure owing to a decreased rate of cell proliferation and increased cell death.
Administration, Oral
;
Anilides
;
pharmacology
;
Animals
;
Bone Diseases, Developmental
;
chemically induced
;
Cell Proliferation
;
drug effects
;
physiology
;
Female
;
Frontal Bone
;
abnormalities
;
Hedgehog Proteins
;
antagonists & inhibitors
;
Limb Deformities, Congenital
;
chemically induced
;
Mice
;
Micrognathism
;
chemically induced
;
Osteogenesis
;
drug effects
;
Pregnancy
;
Pyridines
;
pharmacology
;
Signal Transduction
;
drug effects
6. Impact of oxidative stress on renal dopamine D1 receptor dysfunction in offspring of diabetic rat dams
Hao LUO ; Na WANG ; Caiyu CHEN ; Xiaoli LUO ; Hongyong WANG ; Chunyu ZENG
Chinese Journal of Cardiology 2019;47(5):393-398
Objective:
To explore the effects of oxidative stress on renal dopamine D1 receptor dysfunction in offspring of diabetic rat dams.
Methods:
The pregnant Sprague Dawley (SD) rats (
7.Defining a critical period in calvarial development for Hedgehog pathway antagonist-induced frontal bone dysplasia in mice
Jiang YUANJING ; Zhang SHIXIAN ; Mao CHUANQING ; Lai YONGZHEN ; Wu DI ; Zhao HU ; Liao CAIYU ; Chen WEIHUI
International Journal of Oral Science 2019;11(1):58-69
The Hedgehog (Hh) signalling pathway is essential for cellular proliferation and differentiation during embryonic development.Gain and loss of function of Hh signalling are known to result in an array of craniofacial malformations.To determine the critical period for Hh pathway antagonist-induced frontal bone hypoplasia,we examined patterns of dysmorphology caused by Hh signalling inhibition.Pregnant mice received a single oral administration of Hh signalling inhibitor GDC-0449 at 100 or 150 mg·kg-1 body weight at preselected time points between embryonic days (E)8.5 and 12.5.The optimal teratogenic concentration of GDC-0449 was determined to be 150 mg·kg-1.Exposure between E9.5 and E10.5 induced frontal bone dysplasia,micrognathia and limb defects,with administration at E10.5 producing the most pronounced effects.This model showed decreased ossification of the frontal bone with downregulation of Hh signalling.The osteoid thickness of the frontal bone was significantly reduced.The amount of neural crest-derived frontal bone primordium was reduced after GDC-0449 exposure owing to a decreased rate of cell proliferation and increased cell death.
8.Pathogenic characteristic and distribution of Yersinia enterocolitica in Citellus dauricus plague focuses,Inner Mongolia
Li DONG ; Huixia YU ; Caiyu CHEN ; Lixin WANG ; Huabin WANG ; Huaiqi JING ; Xin WANG
Chinese Journal of Zoonoses 2017;33(3):256-259
In order to investigate the distribution of Yersinia enterocolitica in Citellus dauricus plague focuses in Inner Mongolia,three different ecological environ/ments were chosen as the sampling area.Feces,tongue roots throat swabs,and intestinal contents of rodent,livestock,and poultry were separately collected,and different Y.enterocolitica strains were isolated,and identified.PCR analysis was conducted to detect the toxicity genes of Y.enterocolitica.Statiscal analysis was performed by chisquare test.Of the 3 260 samples,65 Y.enterocolitica strains were isolated and the overall detection rate was 1.99%.To include O ∶ 3/3,O ∶ 5/1A,O ∶ 4/4 serum biological type,the pathogenic strain of serotype O ∶ 3 and biological typt 3 carryinq toxicity genes ail,ystA,VirF yadA and rfbc was isolated from pigs in Citellus dauricus plague focuses,Inner Mongolia are the major carrier of pathogenic Y.enterocolitica distributed in three different ecological environment,and distributed mainly in agricultural area.
9. Effect and mechanism of silibinin on the inhibition of ALK positive NSCLC cells by sensitizing crizotinib
Caiyu LIN ; Conghua LU ; Yonghong PAN ; Lin JIAO ; Hengyi CHEN ; Li LI ; Yong HE
Chinese Journal of Oncology 2017;39(9):650-656
Objective:
To explore the synergistic effect of silibinin combined with crizotinib on anaplastic lymphoma kinase positive (ALK+ ) non-small cell lung cancer (NSCLC) cells and its mechanism.
Methods:
H2228 and H3122 cells were treated with silibinin, crizotinib alone or in combination. Cell proliferation was measured by 3-(4, 5-Dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) assay and colony formation assay. Migration or invasion ability was tested by wound healing assay or transwell assay, respectively. Expressions of E-Cadherin and vimentin protein were examined by immunofluorescence staining. The protein expressions of ALK, p-ALK, E-Cadherin and Vimentin were detected by western blotting.The anti-cancer effect of silibinin combined with crizotinib
10.Regulation of human G-coupled protein kinase 4 A142V on AT1 receptor in rat vascular smooth muscle cells
Kun DENG ; Li LIU ; Caiyu CHEN ; Ken CHEN ; Wei WANG ; Yongqiao ZHOU ; Duofen HE ; Chunyu ZENG
Chongqing Medicine 2013;(33):3977-3979
Objective To study the effect of human G-coupled protein kinase 4(GRK4) A142V overexpression on angiotensin Ⅱ1 type(AT1 ) receptor and its-mediated proliferation of rat vascular smooth muscle cells .Methods We constructed a lentiviral vec-tor carrying human GRK4-EGFP gene and observed its expression in A10 cells .Expression of AT1 receptor were determined by im-munoblotting ,GRK4 activity were checked by spectrophotometry ;the linkage between GRK4 and AT1 receptor were determined by co-immunoprecipitation .[3 H] thymidine incorporation was used to detect changes of cell proliferation .Results As compared with the control cells ,A142V-transfected cells had higher GRK4 activity and higher AT1 receptor expression ;there was linkage between GRK4 and AT1 receptor ,the co-immunoprecipitation levels were lower in A142V cells .The basal levels of VSMC proliferation was higher in A142V cells ,Ang Ⅱ increased VSMC proliferation to a greater extent in A 142V cells .Conclusion GRK4 A142V ,via in-creasing GRK4 activity ,increases AT1 receptor expression and function in vascular smooth muscle cell proliferation .

Result Analysis
Print
Save
E-mail