Purpose/Significance To analyze the influencing factors of user satisfaction on mobile medical platform,and to put for-ward suggestions to improve the quality of medical service and user satisfaction.Method/Process Taking the evaluation texts of users from eight tertiary hospitals in Guangxi in the online mobile medical service platform of haodf.com as the research object,the paper uses the software of ROST content mining system to filter meaningless words,carries out emotional analysis,word frequency statistical analysis and co-occurrence matrix semantic network analysis.Result/Conclusion The degree of satisfaction and recognition of users to the mo-bile medical service platform is high.From the perspective of influencing factors and departments,it puts forward some suggestions on how to improve the satisfaction of mobile medical service users.

Acute pancreatitis （AP） is a gastrointestinal disease that requires early intervention， and when it progresses to moderate-severe AP （MSAP） or severe AP （SAP）， there will be a significant increase in the mortality rate of patients. Machine learning （ML） has achieved great success in the early prediction of AP using clinical data with the help of its powerful computational and learning capabilities. This article reviews the research advances in ML in predicting the severity， complications， and death of AP， so as to provide a theoretical basis and new insights for clinical diagnosis and treatment of AP through artificial intelligence.

OBJECTIVE: To establish a machine learning model based on extreme gradient boosting (XGBoost) algorithm for early prediction of severe acute pancreatitis (SAP), and explore its predictive efficiency. METHODS: A retrospective cohort study was conducted. The patients with acute pancreatitis (AP) who admitted to the First Affiliated Hospital of Soochow University, the Second Affiliated Hospital of Soochow University and Changshu Hospital Affiliated to Soochow University from January 1, 2020 to December 31, 2021 were enrolled. Demography information, etiology, past history, and clinical indicators and imaging data within 48 hours of admission were collected according to the medical record system and image system, and the modified CT severity index (MCTSI), Ranson score, bedside index for severity in acute pancreatitis (BISAP) and acute pancreatitis risk score (SABP) were calculated. The data sets of the First Affiliated Hospital of Soochow University and Changshu Hospital Affiliated to Soochow University were randomly divided into training set and validation set according to 8 : 2. Based on XGBoost algorithm, the SAP prediction model was constructed on the basis of hyperparameter adjustment by 5-fold cross validation and loss function. The data set of the Second Affiliated Hospital of Soochow University was served as independent test set. The predictive efficacy of the XGBoost model was evaluated by drawing the receiver operator characteristic curve (ROC curve), and compared it with the traditional AP related severity score; variable importance ranking diagram and Shapley additive explanation (SHAP) diagram were drawn to visually explain the model. RESULTS: A total of 1 183 AP patients were enrolled finally, of which 129 (10.9%) developed SAP. Among the patients from the First Affiliated Hospital of Soochow University and Changshu Hospital Affiliated to Soochow University, there were 786 patients in the training set and 197 in the validation set; 200 patients from the Second Affiliated Hospital of Soochow University were used as the test set. Analysis of all three datasets showed that patients who advanced to SAP exhibited pathological manifestation such as abnormal respiratory function, coagulation function, liver and kidney function, and lipid metabolism. Based on the XGBoost algorithm, an SAP prediction model was constructed, and ROC curve analysis showed that the accuracy for prediction of SAP reached 0.830, the area under the ROC curve (AUC) was 0.927, which was significantly improved compared with the traditional scoring systems including MCTSI, Ranson, BISAP and SABP, the accuracy was 0.610, 0.690, 0.763, 0.625, and the AUC was 0.689, 0.631, 0.875, and 0.770, respectively. The feature importance analysis based on the XGBoost model showed that the top ten items ranked by the importance of model features were admission pleural effusion (0.119), albumin (Alb, 0.049), triglycerides (TG, 0.036), Ca²⁺ (0.034), prothrombin time (PT, 0.031), systemic inflammatory response syndrome (SIRS, 0.031), C-reactive protein (CRP, 0.031), platelet count (PLT, 0.030), lactate dehydrogenase (LDH, 0.029), and alkaline phosphatase (ALP, 0.028). The above indicators were of great significance for the XGBoost model to predict SAP. The SHAP contribution analysis based on the XGBoost model showed that the risk of SAP increased significantly when patients had pleural effusion and decreased Alb. CONCLUSIONS A SAP prediction scoring system was established based on the machine automatic learning XGBoost algorithm, which can predict the SAP risk of patients within 48 hours of admission with good accuracy.
Humans ; Pancreatitis ; Acute Disease ; Retrospective Studies ; Hospitalization ; Algorithms

Objective:To explore the application of transbronchial lung cryobiopsy guided by endobronchial ultrasound sheath (EBUS-GS-TBCB) in diagnosis of nonresolving pneumonias.Methods:Sixty patients with nonresolving pneumonias from March 2019 to July 2020 in Dalian Municipal Central Hospital were selected. The patients were divided into EBUS-GS-TBCB group (31 cases) and transbronchial forcep lung biopsy guided by endobronchial ultrasound sheath(EBUS-GS-TBLB) group (29 cases) by random digits table method.Results:The diagnostic rate of nonresolving pneumonias in EBUS-GS-TBCB group was significantly higher than that in EBUS-GS-TBLB group: 87.10% (27/31) vs. 65.52% (19/29), and there was statistical difference ( χ2 = 3.90, P = 0.048). There were no statistical difference in sensitivity, specificity, accuracy, positive predictive value and negative predictive value between 2 groups ( P>0.05). There were no statistical difference inthe shortest distance from lesions to pleura, incidence of pneumothorax and incidence of bleeding between EBUS-GS-TBCB group and EBUS-GS-TBLB group: (27.42 ± 2.88) mm vs. (27.01 ± 2.37) mm, 6.45%(2/31) vs. 3.45%(1/29) and 22.58%(7/31) vs. 13.79% (4/29), P>0.05. Among the causes of nonresolving pneumonias, infectious factors accounted for 21.67% (13/60), non infectious factors accounted for 66.67% (40/60), and uncertain causes accounted for 11.67% (7/60). Conclusions:The diagnostic rate of EBUS-GS-TBCB in nonresolving pneumonias is significantly higher than EBUS-GS-TBLB, and the complications such as bleeding and pneumothorax do not increase significantly.

Objective:To analyze the risk factors of major adverse kidney events within 30 days (MAKE30) in patients with acute pancreatitis (AP).Methods:A retrospective cohort study was conducted. A total of 162 patients who were first diagnosed with AP in the First Affiliated Hospital of Soochow University from June 2019 to June 2021 and the onset time was less than 72 hours were enrolled. Patients were divided into MAKE30 group and non-MAKE30 group according to the occurrence of MAKE30 after hospitalization. MAKE30 was defined as death from any cause, new renal replacement therapy (RRT), and persistent renal insufficiency (PRD). The clinical data of the two groups at admission were compared. The independent risk factors of MAKE30 were analyzed by multivariate Logistic regression method, and a regression equation was established as a quantitative prediction model of MAKE30. Receiver operator characteristic curve (ROC curve) was drawn to analyze the prediction of the quantitative prediction model value.Results:All 162 patients were included in the final analysis, including 32 in the MAKE30 group and 130 in the non-MAKE30 group. Univariate analysis showed that compared with the non-MAKE30 group, the body mass index (BMI), the proportion of severe AP, and the acute physiology and chronic health evaluation Ⅱ (APACHEⅡ) score, the sequential organ failure assessment (SOFA) score, blood urea nitrogen (BUN), serum creatinine (SCr), C-reactive protein (CRP), HCO 3-, Cl - levels and the proportion of hyperchloremia at admission in the MAKE30 group were significantly increased. Multivariate Logistic regression analysis showed that APACHE Ⅱ score at admission [odds ratio ( OR) = 1.659, 95% confidence interval (95% CI) was 1.426-1.956, P = 0.009], SOFA score ( OR = 1.501, 95% CI was 1.236-1.840, P = 0.014) and hyperchloremia ( OR = 1.858, 95% CI was 1.564-2.231, P = 0.004) were independent risk factors for MAKE30 in AP patients. The MAKE30 regression equation was established by the above risk factors [Logit( P) = 0.063+0.525×APACHEⅡ score+0.328×SOFA score+0.895×hyperchloremia], which was used as the MAKE30 quantitative prediction model. ROC curve analysis showed that the area under the ROC curve (AUC) of the model for predicting MAKE30 was 0.846 (95% CI was 0.774-0.923, P = 0.001). The patients were divided into two subgroups with hyperchloremia (Cl -≥110 mmol/L, n = 19) and non-hyperchloremia (Cl - < 110 mmol/L, n = 143) according to the blood Cl - level at admission. The incidence of MAKE30 and acute kidney injury (AKI) in the hyperchloremia group was significantly increased (MAKE30: 68.4% vs. 13.3%, AKI: 89.5% vs. 43.4%), and the levels of BUN and SCr at admission were significantly increased [BUN (mmol/L): 9.3±2.5 vs. 5.9±1.1, SCr (μmol/L): 162.3±26.4 vs. 78.6±9.2], the total length of hospital stay and length of intensive care unit (ICU) stay were significantly longer [total length of hospital stay (days): 10.2±1.6 vs. 5.6±1.2, length of ICU stay (days): 6.2±1.0 vs. 3.1±0.6], the cumulative intravenous infusion volume increased significantly at 48 hours and 72 hours (mL: 7 235.9±1 025.3 vs. 5 659.6±956.7 at 48 hours, 11 052.6±1 659.8 vs. 7 156.9±1 052.4 at 72 hours), differences were statistically significant (all P < 0.01). Conclusions:MAKE30 can be used as an important indicator to evaluate the short-term clinical prognosis of AP patients. APACHEⅡ score, SOFA score and hyperchloremia at admission are the main risk factors. The risk model of MAKE30 based on these three indicators has good predictive performance. AP patients with hyperchloremia are at high risk of developing MAKE30, which should be highly regarded in clinical practice.

Glycosylation is one of the most important posttranslational modification (PTM) for proteins. Glycans of glycoproteins play pivotal effects in cell recognition, signal transduction, differentiation, proliferation and immigration. The sialylation, fucosylation and degree of branching are intimately related to the development and progression of various malignancies and autoimmune diseases. Both glycans as well as glycoprotein have become the hot targets for disease biomarker exploration and therapeutic interventions.

Glycosylation is a part of the structure and function of immunoglobulin G (IgG). Sialic acid is located at the end of IgG N-glycan and regulates IgG anti-inflammatory and pro-inflammatory activities by changing the binding of IgG with fragment crystallizable gamma receptors (FcγRs) and complements. Low IgG sialylation is closely related to the occurrence, development and prognosis of autoimmune diseases and shows great potential in the field of diagnosis, monitoring and treatment, and may function as a new biomarker and therapeutic target for autoimmune diseases.

Objective:To establish a lectin enzyme-linked immunosorbent assay (lectin-ELISA) for the dection of sialylated fetuin-A and to explore the clinical diagnostic value of sialylated fetuin-A in hepatocellular carcinoma (HCC).Methods:From January 2017 to December 2020, 300 HCC patients and 160 disease controls, including 36 liver cirrhosis subgroups and 124 chronic hepatitis B subgroups, were collected from Shanghai Eastern Hepatobiliary Surgery Hospital. At the same time, 100 healthy subjects were collected as healthy controls. Lectin-ELISA method for detecting sialylated fetuin A was established based on the principle that Sambucus nigra lectin (SNA) can recognize the structure of α-2, 6-linked sialic acid residues. Differences between groups were compared using t-test or analysis of variance. Logistic regression method was used to establish the multi-index joint detection model, and receiver operating characteristic curve (ROC) was used to evaluate the efficacy of single index and joint detection model in the diagnosis of HCC.Results:A lectin-ELISA method for the detection of serum Sia-fetuin A was established. The linear regression coefficient of the system was 0.978 5, and the precision evaluation and interference experiments were in line with the clinical detection requirements. Using this method to detect serum Sia-fetuin A levels in each group, the levels of HCC group, disease control group and healthy control group were 1.362±0.310, 1.199±0.370, 1.086±0.420, respectively, and the three groups decreased in turn. The areas under the curve of Sia-fetuin A, α-fetoprotein, and their combined detection models for differential diagnosis of HCC were 0.790, 0.809, and 0.860, respectively. The diagnostic model had a sensitivity of 79.3% (238/300) and a specificity of 95.0% (247/260). Among the 300 patients in the HCC group, 138 (46%) patients were negative for serum AFP (<20 μg/L), and their serum Sia-fetuin A level was 1.364±0.305. Combining the disease control group and the healthy control group into the non-Cancer group, the serum Sia-fetuin A level was 1.146±0.381. The serum level of Sia-fetuin A in AFP-negative HCC patients was higher than that in non-HCC group ( t=6.134, P<0.001). The areas under the curve of Sia-fetuin A and the combined diagnostic model for the diagnosis of AFP-negative HCC were 0.776 and 0.919, respectively. The combined diagnostic model had a sensitivity of 93.4% (129/138) and a specificity of 77.3% (201/260). Conclusion:Serum Sia-fetuin A and combined determination model can provide a new auxiliary diagnostic index for AFP-negative HCC.

Circulation biomarker detection is one of the most feasible options for disease screening and monitoring. Focusing on the biomarkers of end stage liver diseases (liver cirrhosis and hepatocellular carcinoma), this article summarized the classification of biomarkers, the exploration and translation of new biomarkers, as well as the applications of the algorithms of the biomarkers. The key points involved in both new biomarker exploration and algorithm construction were addressed. The comprehensive application of available markers, using algorithms, is strongly recommended and should be strengthened in the future for precise clinical management and high-risk predictions in diseases such as hepatocellular carcinoma.

Objective:To investigate the clinical management value of chitinase 3-like 1 protein(CHI3L1) in hepatocellular carcinoma (HCC) by studying the expression of CHI3L1 in peripheral blood, liver cancer and paired adjacent non-tumor tissues.Methods:Retrospective study. From 2013 to 2017, 405 patients with HCC in Third Affiliated Hospital of Naval Medical University were enrolled into the study. Meanwhile, 112 patients with liver cirrhosis (LC), 114 health subjects were included as disease and health controls. CHI3L1 in peripheral blood was detected by ELISA kit. Tissues array was made by collecting 90 pairs of tumor tissues and matched paracancer tissues, from HCC patients who were conformed by pathology. The expression of CHI3L1 in HCC tissues was analyzed by immunohistochemistry. Differences between independent groups were tested by Mann-Whitney U test or Kruskal Wallis H test, Pearson correlation analysis was used for analyzing the relationship between two subjects, and matched rank sum test was used for cancer tissue and adjacent tissue comparison. Results:The median (quartile) of CHI3L1 protein in LC group, HCC group and NC group was 195.8 (103.3,330.4) μg/L,118.2 (74.9,201.0) μg/L,46.8 (30.7,66.4) μg/L independently. The protein level of CHI3L1 in LC group was significantly higher than that in HCC group and health control group ( Z=5.186,12.928, P<0.001). HCC group was significantly higher than that in health control group ( Z=10.788, P<0.001). The level of CHI3L1 in HCC group was not related to whether liver cirrhosis was accompanied ( Z=-0.286, P=0.775). The level of serum CHI3L1 was positively correlated with noninvasive fibrosis markers (HA, PⅢNP, Ⅳ-C, FIB-4 index) ( r=0.202,0.159,0.299 and 0.221, P<0.05) and negatively correlated with ALB( r=-0.326, P<0.05) while positively correlated with AST and PT( r=0.138, 0.160, P<0.05). Positively correlation was observed between CHI3L1 and tumor size ( r=0.284, P<0.001). CNLC stage [CHI3L1 level in advanced group125.2(81.9,228.5)μg/L was higher than that in early group112.0(70.2,169.2)μg/L ( Z=-2.326, P=0.018)], but no correlation with microvascular invasion( Z=-1.531) and tumor capsule(χ 2=0.818, P>0.05). In 73 cases of HCC tissues, the positive rate of CHI3L1 was 78% (57/73) in cancer tissues and 83%(61/73) in paired adjacent non-tumor tissues. The staining intensity score of paracancer tissue 1.5(1.5,2.5) was higher than that of cancer tissue 1.5(1.5,2.0)( Z=-2.053, P=0.040). Conclusions:The tissue source of CHI3L1 protein in HCC includes cancer tissue and paracancerous tissue. The detection of serum CHI3L1 level is helpful to evaluate tumor load assessment and disease stratification management in HCC.