1.Explainable machine learning model for predicting septic shock in critically sepsis patients based on coagulation indexes: A multicenter cohort study.
Qing-Bo ZENG ; En-Lan PENG ; Ye ZHOU ; Qing-Wei LIN ; Lin-Cui ZHONG ; Long-Ping HE ; Nian-Qing ZHANG ; Jing-Chun SONG
Chinese Journal of Traumatology 2025;28(6):404-411
PURPOSE:
Septic shock is associated with high mortality and poor outcomes among sepsis patients with coagulopathy. Although traditional statistical methods or machine learning (ML) algorithms have been proposed to predict septic shock, these potential approaches have never been systematically compared. The present work aimed to develop and compare models to predict septic shock among patients with sepsis.
METHODS:
It is a retrospective cohort study based on 484 patients with sepsis who were admitted to our intensive care units between May 2018 and November 2022. Patients from the 908th Hospital of Chinese PLA Logistical Support Force and Nanchang Hongdu Hospital of Traditional Chinese Medicine were respectively allocated to training (n=311) and validation (n=173) sets. All clinical and laboratory data of sepsis patients characterized by comprehensive coagulation indexes were collected. We developed 5 models based on ML algorithms and 1 model based on a traditional statistical method to predict septic shock in the training cohort. The performance of all models was assessed using the area under the receiver operating characteristic curve and calibration plots. Decision curve analysis was used to evaluate the net benefit of the models. The validation set was applied to verify the predictive accuracy of the models. This study also used Shapley additive explanations method to assess variable importance and explain the prediction made by a ML algorithm.
RESULTS:
Among all patients, 37.2% experienced septic shock. The characteristic curves of the 6 models ranged from 0.833 to 0.962 and 0.630 to 0.744 in the training and validation sets, respectively. The model with the best prediction performance was based on the support vector machine (SVM) algorithm, which was constructed by age, tissue plasminogen activator-inhibitor complex, prothrombin time, international normalized ratio, white blood cells, and platelet counts. The SVM model showed good calibration and discrimination and a greater net benefit in decision curve analysis.
CONCLUSION
The SVM algorithm may be superior to other ML and traditional statistical algorithms for predicting septic shock. Physicians can better understand the reliability of the predictive model by Shapley additive explanations value analysis.
Humans
;
Shock, Septic/blood*
;
Machine Learning
;
Male
;
Female
;
Retrospective Studies
;
Middle Aged
;
Aged
;
Sepsis/complications*
;
ROC Curve
;
Cohort Studies
;
Adult
;
Intensive Care Units
;
Algorithms
;
Blood Coagulation
;
Critical Illness
2.The Association of Polymorphisms Drug Metabolism and Transport of Imatinib Related Gene with Severe Hematology Adverse Effects in Chronic Myeloid Leukemia Patients.
Wen-Jing ZHOU ; Nian WANG ; Li LIN ; Li-Juan WU ; Yuan-Xin YE
Journal of Experimental Hematology 2025;33(2):344-351
OBJECTIVE:
To screen the genetic risk factors related to severe hematology adverse effects (AEs) in patients with chronic myeloid leukemia (CML) treated with imatinib (IM), and explore the correlation of single nucleotide polymorphisms (SNPs) in IM drug metabolism and transport pathway gene polymorphism with the risk of severe hematology AEs.
METHODS:
172 newly diagnosed Chinese Han patients in CML chronic phase (CML-CP) treated with IM were included and divided into severe hematology AEs group and non-severe hematology AEs group. The demographic characteristics and laboratory test results were compared between the two groups. 11 gene SNP sites in the included subjects were genotyped using SNaPshot multiplex SNPs technique.
RESULTS:
Compared with non-severe hematology AEs group, the severe hematology AEs group had higher white blood cell (WBC) and EOS% (both P < 0.05), but lower hemoglobin (Hb) and hematocrit (HCT) (both P < 0.01). For rs1045642 of ABCB1 gene, there were significant differences in the distribution of allele frequency and genotype frequency of this loci between severe hematology AEs group and non-severe hematology AEs group (both P < 0.05). Carriers of rs1045642 mutation allele A had an increased risk of severe hematology AEs (OR =2.09, 95% CI : 1.24-3.55, P =0.005). There was a significant difference in the distribution of NR1I2 gene rs3814055 genotype between severe hematology AEs group and non-severe hematology AEs group (P < 0.05). The additive model and recessive model of ABCB1 gene rs1045642 and the recessive model of NR1I2 gene rs3814055 were associated with the increased risk of severe hematology AEs (OR =2.14, 3.28, 5.54, all P < 0.05).
CONCLUSION
Peripheral blood WBC, EOS%, Hb and HCT in patients with newly diagnosed CML-CP are all related to the risk of severe hematology AEs. ABCB1 gene rs1045642 and NR1I2 gene rs3814055 related to the metabolism and transport pathway of IM are associated with severe hematology AEs after IM treatment in CML-CP patients, and they may be potential molecular markers to predict the risk of severe hematology AEs of CML patients treated by IM.
Humans
;
Imatinib Mesylate
;
Leukemia, Myelogenous, Chronic, BCR-ABL Positive/genetics*
;
Polymorphism, Single Nucleotide
;
Genotype
;
ATP Binding Cassette Transporter, Subfamily B
;
Gene Frequency
;
Female
;
Male
;
Middle Aged
;
Adult
;
Asian People
3.Effect of new labor process on early postpartum pelvic floor muscle strength
Yu WANG ; Xinnan HOU ; Lei ZHANG ; Songwen NIAN ; Ruilin GUO ; Bingbing XIAO ; Xiaoqing WANG ; Xiaoxiao WANG ; Ye LU
Chinese Journal of Perinatal Medicine 2024;27(6):499-503
Objective:To explore the changes in early postpartum pelvic floor muscle strength following the implementation of the new labor process.Methods:This retrospective cohort study selected 1 834 primiparous women with singleton, full-term pregnancies who delivered at Peking University First Hospital from February 2011 to March 2016 and had a pelvic floor re-examination 6-8 weeks postpartum. Out of these, 738 cases who followed the old labor process before 2014 were categorized as the old process group, and 1 096 cases who followed the new labor process after 2014 were categorized as the new process group. Basic data, childbirth information, and postpartum pelvic floor muscle strength of the two groups were compared. Data were statistically analyzed using t-test, Chi-square test, Mann-Whitney U test, Wilcoxon rank-sum test, and ordered multicategory logistic regression to assess the impact of the new and old labor process and other factors on pelvic floor muscle strength. Results:The total duration of labor, as well as the duration of the first, second, and third stages of labor, were longer in the new process group than in the old process group [549.0 min (360.0-768.0 min) vs. 482.5 min (328.0-635.0 min), 465.0 min (297.5-672.5 min) vs. 420.0 min (285.0-555.0 min), 42.0 min (24.0-74.0 min) vs. 27.0 min (18.0-45.0 min), with Z-value of-5.72,-3.95, and-9.28, all P<0.05). The rates of vaginal delivery and labor analgesia were higher in the new process group [72.1% (790/1 096) vs. 67.2% (496/738), χ2=7.41; 67.4% (739/1 096) vs. 53.4% (394/738), χ2=36.82; both P<0.05]. There were no statistically significant differences in the comparison of Class Ⅰ and Class Ⅱ muscle strength grades between the two groups (all P>0.05). Conclusion:There was no significant decline in early postpartum pelvic floor muscle strength following the implementation of the new labor process standards.
4.Application of "Flipped Classroom" in comprehensive experiment teaching of Digital Signal Processing for biomedical engineering specialty
Linqiong SANG ; Li WANG ; Yongjian NIAN ; Liang QIAO ; Jingna ZHANG ; Ye ZHANG ; Pengyue LI ; Qiannan WANG ; Mingguo QIU
Chinese Journal of Medical Education Research 2023;22(2):212-215
"Flipped Classroom" is a new kind of "student-centered" teaching model, which can give full play to the advantages of both sides of teaching and learning. According to this teaching model, we redesigned the teaching process, in which the students studied by themselves and built their own knowledge system. Moreover, each of them took part in three stages of experimental design including digital signal collection, analysis and processing in groups. Results have shown that this model can fully stimulate students' learning interest, not only helps students to deepen understanding of digital signal processing theory knowledge, but also strengthen the ability of autonomous learning and team collaboration. The teaching model maybe have certain reference function in comprehensive experiment teaching of Digital Signal Processing course for biomedical engineering specialty.
5.Thoughts on the course construction of Brain Functional Imaging for postgraduates majoring in biomedical engineering
Jingna ZHANG ; Ye ZHANG ; Yongjian NIAN ; Liang QIAO ; Li WANG ; Linqiong SANG ; Qiannan WANG ; Pengyue LI ; Mingguo QIU
Chinese Journal of Medical Education Research 2023;22(6):873-876
Based on the summary and reflection of the existing course construction content, this paper redesigns the course teaching to Brain Functional Imaging and forms a new curriculum construction scheme, including improving the professional teachers' ability of the teaching team by means of "internal training" + "external introduction", building an online and offline integrated teaching mode by combining online teaching resources such as course website with offline teaching (such as literature guidance, classroom discussion, comprehensive experimental design, etc.), and designing comprehensive experiments related to Electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS), which is driven by scientific research projects, based on the scientific research experimental platform of the teaching and research department. This construction scheme is of great significance for improving the teaching quality of the course, stimulating the learning interest of graduate students, and cultivating the comprehensive application and practical innovation ability of graduate students' brain imaging technology. And it also provides the reference for the further construction of the course and teaching reform in the future.
6.The efficacy of radiotherapy based combined therapy for unresectable locally invasive bladder cancer and its associated factors analysis.
Si Jin ZHONG ; Jun Jun GAO ; Ping TANG ; Yue Ping LIU ; Shu Lian WANG ; Hui FANG ; Jing Ping QIU ; Yong Wen SONG ; Bo CHEN ; Shu Nan QI ; Yuan TANG ; Ning Ning LU ; Hao JING ; Yi Rui ZHAI ; Ai Ping ZHOU ; Xin Gang BI ; Jian Hui MA ; Chang Ling LI ; Yong ZHANG ; Jian Zhong SHOU ; Nian Zeng XING ; Ye Xiong LI
Chinese Journal of Oncology 2023;45(2):175-181
Objective: Retrospective analysis of the efficacy and influencing factors of bladder preservation integrated therapy for unresectable invasive bladder cancer confined to the pelvis was done, also including the bladder function preservation and adverse effects analysis. Methods: Sixty-nine patients with unresectable locally invasive bladder cancer who received radiotherapy-based combination therapy from March 1999 to December 2021 at our hospital were selected. Among them, 42 patients received concurrent chemoradiotherapy, 32 underwent neoadjuvant chemotherapyand 43 with transurethral resection of bladder tumors (TURBT) prior to radiotherapy. The late adverse effect of radiotherapy, preservation of bladder function, replase and metastasis and survival were followed-up. Cox proportional hazards models were applied for the multifactorial analysis. Results: The median age was 69 years. There were 63 cases (91.3%) of uroepithelial carcinoma, 64 of stage Ⅲ and 4 of stage Ⅳ. The median duration of follow-up was 76 months. There were 7 grade 2 late genito urinary toxicities, 2 grade 2 gastrointestinal toxicities, no grade 3 or higher adverse events occurred. All patients maintained normal bladder function, except for 8 cases who lost bladder function due to uncontrolled tumor in the bladder. Seventeen cases recurred locally. There were 11 cases in the concurrent chemoradiotherapy group with a local recurrence rate of 26.2% (11/42) and 6 cases in the non-concurrent chemoradiotherapy group with a local recurrence rate of 22.2% (6/27), and the difference in local recurrence rate between the two groups was not statistically significant (P=0.709). There were 23 cases of distant metastasis (including 2 cases of local recurrence with distant metastasis), including 10 cases in the concurrent chemoradiotherapy group with a distant metastasis rate of 23.8% (10/42) and 13 cases in the non-concurrent chemoradiotherapy group with a distant metastasis rate of 48.1% (13/27), and the distant metastasis rate in the non-concurrent chemoradiotherapy group was higher than that in the concurrent chemoradiotherapy group (P=0.036). The median 5-year overall survival (OS) time was 59 months and the OS rate was 47.8%. The 5-year progression-free survival (PFS) time was 20 months and the PFS rate was 34.4%. The 5-year OS rates of concurrent and non-concurrent chemoradiotherapy group were 62.9% and 27.6% (P<0.001), and 5-year PFS rates were 45.4% and 20.0%, respectively (P=0.022). The 5-year OS rates of with or without neoadjuvant chemotherapy were 78.4% and 30.1% (P=0.002), and the 5-year PFS rates were 49.1% and 25.1% (P=0.087), respectively. The 5-year OS rates with or without TURBT before radiotherapy were 45.5% and 51.9% (P=0.233) and the 5-year PFS rates were 30.8% and 39.9% (P=0.198), respectively. Multivariate Cox regression analysis results showed that the clinical stage (HR=0.422, 95% CI: 0.205-0.869) was independent prognostic factor for PFS of invasive bladder cancer. The multivariate analysis showed that clinical stages (HR=0.278, 95% CI: 0.114-0.678), concurrent chemoradiotherapy (HR=0.391, 95% CI: 0.165-0.930), neoadjuvant chemotherapy (HR=0.188, 95% CI: 0.058-0.611), and recurrences (HR=10.855, 95% CI: 3.655-32.638) were independent prognostic factors for OS of invasive bladder cancer. Conclusion: Unresectable localized invasive bladder cancer can achieve satisfactory long-term outcomes with bladder-preserving combination therapy based on radiotherapy, most patients can retain normal bladder function with acceptable late adverse effects and improved survival particularly evident in patients with early, concurrent chemoradiotherapy and neoadjuvant chemotherapy.
Humans
;
Aged
;
Treatment Outcome
;
Retrospective Studies
;
Combined Modality Therapy
;
Chemoradiotherapy/methods*
;
Urinary Bladder Neoplasms/radiotherapy*
;
Antineoplastic Combined Chemotherapy Protocols/therapeutic use*
;
Neoplasm Staging
7.Effects of propiconazole on physiological and biochemical properties of Panax notoginseng and dietary risk assessment.
Zi-Xiu ZHENG ; Li-Sha QIU ; Kai ZHENG ; Lan-Ping GUO ; Xiu-Ming CUI ; Hong-Juan NIAN ; Ying-Cai LI ; Shao-Jun HUANG ; Ye YANG
China Journal of Chinese Materia Medica 2023;48(5):1203-1211
To study the residue and dietary risk of propiconazole in Panax notoginseng and the effects on physiological and bioche-mical properties of P. notoginseng, we conducted foliar spraying of propiconazole on P. notoginseng in pot experiments. The physiolo-gical and biochemical properties studied included leaf damage, osmoregulatory substance content, antioxidant enzyme system, non-enzymatic system, and saponin content in the main root. The results showed that at the same application concentration, the residual amount of propiconazole in each part of P. notoginseng increased with the increase in the times of application and decreased with the extension of harvest interval. After one-time application of propiconazole according to the recommended dose(132 g·hm~(-2)) for P. ginseng, the half-life was 11.37-13.67 days. After 1-2 times of application in P. notoginseng, propiconazole had a low risk of dietary intake and safety threat to the population. The propiconazole treatment at the recommended concentration and above significantly increased the malondialdehyde(MDA) content, relative conductivity, and osmoregulatory substances and caused the accumulation of reactive oxygen species in P. notoginseng leaves. The propiconazole treatment at half(66 g·hm~(-2)) of the recommended dose for P. ginseng significantly increased the activities of superoxide dismutase(SOD), peroxidase(POD), and catalase(CAT) in P. notoginseng leaves. The propiconazole treatment at 132 g·hm~(-2) above inhibited the activities of glutathione reductase(GR) and glutathione S-transferase(GST), thereby reducing glutathione(GSH) content. Proconazole treatment changed the proportion of 5 main saponins in the main root of P. notoginseng. The treatment with 66 g·hm~(-2) propiconazole promoted the accumulation of saponins, while that with 132 g·hm~(-2) and above propiconazole significantly inhibited the accumulation of saponins. In summary, using propiconazole at 132 g·hm~(-2) to prevent and treat P. notoginseng diseases will cause stress on P. notoginseng, while propiconazole treatment at 66 g·hm~(-2) will not cause stress on P. notoginseng but promote the accumulation of saponins. The effect of propiconazole on P. notoginseng diseases remains to be studied.
Panax notoginseng/chemistry*
;
Panax
;
Antioxidants/pharmacology*
;
Saponins/pharmacology*
;
Glutathione
;
Risk Assessment
8.Preliminary study on the regulation of acute myeloid leukemia by FLT3 gene expression.
Sishi TANG ; Yanhong ZHOU ; Wenjing ZHOU ; Nian WANG ; Binwu YING ; Yuanxin YE
Chinese Journal of Medical Genetics 2023;40(9):1113-1117
OBJECTIVE:
To assess the influence of FLT3 expression on the prognosis of patients with acute myeloid leukemia (AML) by cell experiment and clinical data analysis.
METHODS:
Models for FLT3 over-expression and interference-expression in AML cells were constructed. The level of BAK gene expression and its protein product was determined, along with the proliferation and apoptosis of leukemia cells. FLT3 gene expression and FLT3-ITD variant were determined among patients with newly diagnosed AML.
RESULTS:
Compared with the interference-expression group, the level of BAK gene expression and its protein in FLT3 over-expression AML cells was significantly lower (P < 0.001), which also showed significantly faster proliferation (P < 0.001) and lower rate of apoptosis (P < 0.001). The expression level of FLT3 gene among patients with newly diagnosed AML was also significantly higher compared with the healthy controls (P < 0.001). The FLT3 gene expression of FLT3-ITD positive AML patients was higher than that of FLT3-WT patients (P = 0.002). Survival analysis showed that AML patients with high FLT3 expression in the medium-risk group had a lower complete remission rate and overall survival rate compared with those with a low FLT3 expression (P < 0.001).
CONCLUSION
Over-expression of FLT3 may influence the course of AML by promoting the proliferation of leukemia cells and inhibiting their apoptosis, which in turn may affect the prognosis of patients and serve as a negative prognostic factor for AML.
Humans
;
Apoptosis/genetics*
;
Data Analysis
;
Leukemia, Myeloid, Acute/genetics*
;
Gene Expression
;
fms-Like Tyrosine Kinase 3/genetics*
9.Molecular mechanism of Ganoderma against gastric cancer based on network pharmacology and experimental test.
Jia-Yi ZHONG ; Hai-Bing CHEN ; Da-Zeng YE ; Zheng-Jun DENG ; Jia-Jia SHAO ; Jia-Wei HAN ; Jun-Hui YUAN ; Nian-Ying DENG
China Journal of Chinese Materia Medica 2022;47(1):203-223
This study aims to explore the molecular mechanism of Ganoderma against gastric cancer based on network pharmacology, molecular docking, and cell experiment. The active components and targets of Ganoderma were retrieved from Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform(TCMSP), and gastric cancer-related targets from GeneCards and Online Mendelian Inheritance in Man(OMIM). The protein-protein interaction(PPI) network of the common targets was constructed with STRING, followed by Gene Ontology(GO) term enrichment and Kyoto Encyclopedia of Genes and Genomes(KEGG) pathway enrichment analysis of the common genes based on Bioconductor and R language. The medicinal-disease-component-target network and medicinal-disease-component-target-pathway network were established by Cytoscape. Molecular docking was performed between β-sitosterol(the key component in Ganoderma) and the top 15 targets in the PPI network. Cell experiment was performed to verify the findings. A total of 14 active components and 28 targets of Ganoderma were retrieved, and the medicinal and the disease shared 25 targets, including caspase-3(CASP3), caspase-8(CASP8), caspase-9(CASP9), and B-cell lymphoma-2(BCL2). The common targets involved 72 signaling pathways and apoptosis and p53 signaling pathway may play a crucial role in the effect of Ganoderma against gastric cancer. β-sitosterol had strong binding activity to the top 15 targets in the PPI network. The in vitro cell experiment demonstrated that β-sitosterol inhibited gastric cancer AGS cell proliferation by inducing cell apoptosis and cell cycle arrest in the S phase, which might be related to the regulation of the p53 pathway. This study shows the multi-component, multi-target, and multi-pathway characteristics of Ganoderma against gastric cancer, which lays a scientific basis for further research on the molecular mechanism.
Ganoderma
;
Humans
;
Medicine, Chinese Traditional
;
Molecular Docking Simulation
;
Network Pharmacology
;
Stomach Neoplasms/genetics*
10.Impact of regional positive lymph node ratio of gallbladder carcinoma on prognosis.
Ye Ming ZHOU ; Hong Chao MI ; Wei JIANG ; Yin JIANG ; Yong Fei HUA ; Chang Jiang LU ; Chun Nian WANG ; Cai De LU
Chinese Journal of Oncology 2022;44(11):1202-1207
Objective: To study the impact of regional positive lymph node ratio (LNR) on prognosis of patients with gallbladder carcinoma. Methods: The clinicopathological and survival data of 53 patients with gallbladder carcinoma who underwent radical resection with regional lymph node metastasis in Ningbo University Affiliated Lihuili Hospital from May 2012 to December 2020 were collected, and receiver operating characteristic curve (ROC) was used to determine the optimal cut-off value of LNR for predicting postoperative survival status in patients with gallbladder carcinoma. According to the critical value, the patients were divided into low LNR group and high LNR group. The clinicopathological features and prognosis of the two groups were compared. Log rank test was used for univariate analysis of prognostic factors in patients with gallbladder carcinoma, and Cox proportional hazards model was used for multivariate analysis. Results: A total of 417 regional lymph nodes were dissected in 53 patients, of which 144 lymph nodes were positive, with a positive rate of 34.5%. The optimal cut-off value of LNR for predicting postoperative survival status of patients with gallbladder carcinoma was 0.33. According to this cut-off value, patients were divided into low LNR group (LNR≤0.33, 28 cases) and high LNR group (LNR>0.33, 25 cases). The recurrence rates were 64.3% (18/28) and 88.0 % (22/25) in low LNR group and high LNR group, respectively. The median recurrence-free survival (RFS) was 8 and 7 months, respectively (P=0.032). In the low LNR group, the 1-, 3-, and 5-year survival rates were 56.2%, 38.4%, and 32.0%, respectively, and the median overall survival (OS) was 16 months. In the high LNR group, the 1-, 3-, and 5-year survival rates were 37.9%, 5.4%, and 0, respectively, and the median OS was 9 months. The postoperative survival rate of patients in the low LNR group was better than that in the high LNR group (P=0.008). Univariate analysis showed that LNR was even associated with RFS and OS in patients with gallbladder carcinoma (P<0.05). Multivariate analysis showed that LNR>0.33 was an independent risk factor for postoperative RFS (HR=1.977, 95% CI: 1.045-3.740), but not for OS (HR=1.561, 95% CI: 0.685-3.553). Conclusion: On the basis of clearing a sufficient number of regional lymph nodes, patients with gallbladder carcinoma with regional LNR>0.33 are more likely to relapse after operation, but the predictive value of LNR>0.33 OS is insufficient.
Humans
;
Lymph Node Ratio
;
Gallbladder Neoplasms/pathology*
;
Lymph Node Excision
;
Lymphatic Metastasis/pathology*
;
Neoplasm Staging
;
Retrospective Studies
;
Lymph Nodes/pathology*
;
Prognosis

Result Analysis
Print
Save
E-mail