BACKGROUND:The sports medicine community has widely called for the use of machine learning technology to efficiently process the huge and complicated sports data resources,and construct intelligent sports injury prediction models,enabling accurate early warning of sports injuries.It is of great significance to comprehensively summarize and review such research results so as to grasp the direction of early warning model improvement and to guide the construction of sports injury prediction models in China. OBJECTIVE:To systematically review and analyze relevant research on sports injury prediction models based on machine learning technology,thereby providing references for the development of sports injury prediction models in China. METHODS:Literature search was conducted on CNKI,Web of Science and EBSCO databases,which mainly searched for literature related to machine learning techniques and sports injuries.Finally,61 articles related to sports injury prediction models were included for analysis. RESULTS AND CONCLUSION:(1)In terms of external risk feature indicators,there is a lack of competition scenario indicators,and the inclusion of related feature indicators needs to be further improved to further enrich the dimensions of the dataset for model training.In addition,the inclusion feature weighting methods of the sports injury prediction model are mainly based on filtering methods and the use of embedding and wrapping weighting methods needs to be strengthened in order to enhance the analysis of the interaction effects of multiple risk factors.(2)In terms of model body training,supervised learning algorithms become the mainstream choice.Such algorithms have higher requirements for the completeness of sample labeling information,and the application scenarios are easily limited.Therefore,the application of unsupervised and semi-supervised algorithms can be increased in the later stage.(3)In terms of model performance evaluation and optimization,the current studies mainly adopt two verification methods:HoldOut crossover and k-crossover.The range of AUC values is(0.76±0.12),the range of sensitivity is(75.92±11.03)%,the range of specificity is(0.03±4.54)%,the range of F1 score is(80.60±10.63)%,the range of accuracy is(69.96±13.10)%,and the range of precision is(70±14.71)%.Data augmentation and feature optimization are the most common model optimization operations.The accuracy and precision of the current sports injury prediction model are about 70%,and the early warning effect is good.However,the model optimization operation is relatively single,and data augmentation methods are often used to improve model performance.Further adjustments to the model algorithm and hyperparameters are needed to further improve model performance.(4)In terms of model feature extraction,most of the internal risk profile indicators included are mainly based on anthropometrics,training load,years of training,and injury history,but there is a lack of sports recovery and physical function indicators.

Objective:To analyze the prognostic factors of patients with liver cirrhosis and portal vein thrombosis (PVT), and to construct a prognostic prediction model based on machine learning methods.Methods:The clinical data of 388 patients with liver cirrhosis and PVT admitted to the Fifth Medical Center of PLA General Hospital from January 2022 to April 2024 were retrospectively collected and analyzed, including 243 males and 145 females, aged (56.9±10.9) years. A total of 388 patients were randomly divided into the training set ( n=310) and the testing set ( n=78) in a 4∶1 ratio. The Boruta algorithm was used to screen the key features in the training set, and then four machine learning algorithms, including random forest, support vector machine, generalized linear model and Bayesian, were used to establish a survival prediction model. Model performance was evaluated by the receiver operating characteristic (ROC) curves of the test set and the training set. The patients were followed up for 1 year for survival. Sort the importance of features based on the SHAP value. Results:There were 250 patients (80.6%) who survived and 60 (19.4%) who died. The model for end-stage liver disease score, total bilirubin, serum creatinine, prothrombin time, international normalized ratio, D-dimer, white blood cell count, severe ascites ratio, and Child-Pugh grade C ratio of liver function in the death group were higher than those in the survival group, and the red blood cell count and hematocrit were lower than those in the survival group, and the differences were statistically significant (all P<0.05). The areas under the ROC curve for predicting survival by random forest, support vector machine, generalized linear model and Bayesian model were 0.92, 0.78, 0.81 and 0.71 in the training set, and the area under the ROC curve in the testing set were 0.81, 0.72, 0.67 and 0.68, respectively. Random forest had the best prediction performance, with an accuracy of 81.7%, a sensitivity of 84.6%, and a specificity of 76.9% in the testing set. In the analysis of the importance of characteristic parameters of the random forest model, total bilirubin, red blood cells, hematocrit, serum creatinine, ascites classification, etc. had a relatively high contribution to the model. Conclusion:In the survival prediction model of patients with liver cirrhosis and PVT based on machine learning algorithm, the random forest model had high prediction performance, and total bilirubin may be the most important factor affecting the survival prognosis of patients.

AIM:To evaluate the application effec-tiveness of a Bayesian mixture model based on the principal stratum strategy for estimating the com-plier average causal effect(CACE)in clinical trials with non-compliance.METHODS:Using a non-infe-riority randomized controlled trial investigating a novel drug for primary type 2 diabetes mellitus(non-inferiority margin:-0.4)as a case study,the primary analysis applied a Bayesian mixture model under the monotonicity assumption to estimate CACE of between-group differences in glycated he-moglobin(HbA1c)changes within the compliant stratum,followed by non-inferiority testing.Sensi-tivity analyses included a Bayesian mixture model relaxing the monotonicity assumption and compar-ing results with per-protocol set(PPS)analysis.RE-SULTS:In the primary analysis,the posterior mean of CACE for HbA1c change in the compliant stratum was 0.081％,with a one-sided 97.5％credible inter-val lower bound of-0.124,exceeding the non-infe-riority margin(-0.4％),supporting the non-inferiori-ty efficacy of the novel drug in the compliant stra-tum(P(H1|Data)=1).Consistent findings were ob-served in PPS analyses(estimated effect:0.136％;one-sided 97.5％credible interval lower bound:-0.069％),further validating methodological robust-ness.CONCLUSION:In clinical trials with noncom-pliance as an intercurrent event,the Bayesian mix-ture model under the principal stratum strategy ef-fectively adjusts for compliance-related bias and yields conservative,robust estimates of causal ef-fects,supporting its value in efficacy evaluation un-der complex compliance scenarios.

Objective:To investigate the efficacy of radiotherapy in patients with advanced esophageal cancer receiving first-line chemoimmunotherapy.Methods:A retrospective analysis was conducted on the data of 137 patients with Stage Ⅳ esophageal squamous cell carcinoma (ESCC) treated at our hospital from January 2018 to May 2023. These patients were divided into two groups: a group treated with first-line chemoimmunotherapy combined with radiotherapy (chemoimmunotherapy + radiotherapy group, n = 43) and a group treated with only chemoimmunotherapy ( n = 94). Inverse probability of treatment weighting (IPTW) was applied to balance baseline characteristics between the groups. With overall survival (OS) and progression-free survival (PFS) as study endpoints, the survival data were analyzed using the Kaplan-Meier method, the log-rank test, and the Cox regression method. Results:Before calibration, the chemoimmunotherapy + radiotherapy group significantly outperformed the sole chemoimmunotherapy group in median PFS (13.6 months vs. 7.0 months; HR: 0.501, 95% CI: 0.309-0.811, P = 0.005). After calibration using the COX proportional-hazards model for age, gender, Eastern Cooperative Oncology Group (ECOG) performance status, smoking history, T/N/M stage, and tumor location, the chemoimmunotherapy + radiotherapy group still had significant advantages in PFS (14.7 months vs. 7.0 months; HR: 0.441, 95% CI: 0.261-0.745, P = 0.002). IPTW analysis further confirmed this trend (13.9 months vs. 7.0 months; HR: 0.492, 95% CI: 0.304-0.795, P < 0.001). Specifically, the median OS of the chemoimmunotherapy + radiotherapy group demonstrated significant improvement in all analyses: pre-calibration (29.5 months vs. 18.0 months; HR: 0.507, 95% CI: 0.297-0.867, P = 0.013), after calibration using the Cox model (27.5 months vs. 16.7 months; HR: 0.470, 95% CI: 0.266-0.830, P = 0.009), and after calibration using IPTW (29.5 months vs. 16.9 months; HR: 0.448, 95% CI: 0.262-0.764, P < 0.001). Conclusions:The combination of radiotherapy and first-line chemoimmunotherapy can significantly improve survival outcomes of patients with advanced ESCC, suggesting its potential as a standard treatment strategy.

Objective lt aims to identify key issues in the current informatization development of internal control over revenue and expenditure in public hospitals.The findings are intended to serve as a reference for deepening this informatization effort.Methods Following the steps and requirements of content analysis method,it involved a semi-quantitative analysis of policies and expert interviews to establish an analytical framework.Two researchers uti-lized NVivo 15 software to analyze the policy and articles.The analysis results were quantified statistically,and key is-sues were summarized.Results lt developed a content analysis framework comprising 5 primary categories and 17 secondary categories.Based on this framework,326 analytical units extracted from 81 articles were categorized and statistically analyzed.Key problems corresponding to each secondary category were summarized.Conclusion To ad-dress five key issues,it proposes a five-dimensional solution including:building a digital foundation to integrate busi-ness systems;establishing a Master Data Management platform to dismantle data silos;developing flexible ap-proval mechanismsfor medical emergencies;deploying an Al risk control hub for precise payment interception,and implementing zero-trust architecture to enhance e-bill anti-counterfeiting-ultimately forming a business-finance-in-tegrated,data-and-Al-driven,proactively secured internal control system covering budget formulation to fund su-pervision.

Galectin-9 (Gal-9), a member of the β-galactoside-binding lectin family, is widely expressed in various tissues and cells. It can specifically bind to multiple glycoprotein receptors, including the receptors of Tim-3, CD44, 4-1BB/CD137, and Dectin-1, thereby regulating the activity of immune cells and participating in crucial physiological and pathological processes such as immune regulation and tumor development. Given its role in immunomodulation, Gal-9 is considered a potential target for immunotherapy, showing promising prospects in the treatment of various diseases, including autoimmune disorders, transplantation rejection, pregnancy complications, inflammation, infection, and cancer. This review summarizes the biological effects mediated by Gal-9 upon binding to its receptors, which may help to explore the potential application value of Gal-9 in clinical diagnosis and therapy.

To explore the application of sequential analysis in post-market safety dynamic surveillance of vaccines. Under the dynamic monitoring data of vaccines post-market approval, this research introduces the fundamental principles of maximizing sequential probability ratio test (MaxSPRT) and Bayesian sequential analysis, employing R software. Through an example of dynamic safety monitoring data of vaccines post-market approval, we analyze using the MaxSPRT and Bayesian sequential analysis. The MaxSPRT identified a safety signal in week 4 ( P<0.05), while Bayesian sequential analysis indicated that the 95% highest density interval for the RR value at week 4 is 1.13-3.27, suggesting the first appearance of a safety signal at week 4. The MaxSPRT and Bayesian sequential analysis effectively leverage continuously accumulating dynamic monitoring data, thereby serving as a valuable method for post-market safety surveillance of vaccines.

Due to its synergistic effects and reduced side effects, combination therapy has become an important strategy for treating complex diseases. In traditional Chinese medicine (TCM), the "monarch, minister, assistant, envoy" compatibilities theory provides a systematic framework for drug compatibility and has guided the formation of a large number of classic formulas. However, due to the complex compositions and diverse mechanisms of action of TCM, it is difficult to comprehensively reveal its potential synergistic patterns using traditional methods. Synergistic prediction based on molecular compatibility theory provides new ideas for identifying combinations of active compounds in TCM. Compared to resource-intensive traditional experimental methods, artificial intelligence possesses the ability to mine synergistic patterns from multi-omics and structural data, providing an efficient means for modeling and optimizing TCM combinations. This paper systematically reviews the application progress of AI in the synergistic prediction of TCM active compounds and explores the challenges and prospects of its application in modeling combination relationships, thereby contributing to the modernization of TCM theory and methodological innovation.
Artificial Intelligence ; Medicine, Chinese Traditional/methods* ; Drugs, Chinese Herbal/pharmacology* ; Humans ; Drug Synergism

Objective To analyze the differences in clinical indicators between malignant insulinoma and benign in-sulinoma,in order to provide diagnostic and therapeutic insights for the early detection and diagnosis of malignant insulinoma.Methods A retrospective analysis was conducted in patients diagnosed and treated for insulinoma at Peking Union Medical College Hospital from January 2018 to June 2022.Among them,10 cases were diagnosed as malignant insulinoma.Twenty cases of benign insulinoma patients matched for age,sex,and body mass index(BMI),were randomly selected.Statistical analysis was performed to compare the differences between malignant and benign insulinomas.Results 1)Compared to benign insulinoma,malignant insulinoma showed significantly ele-vated C-peptide(CP)and C-peptide to glucose ratio(CPGlu)during hypoglycemia(blood glucose＜3.0 mmol/L)[6.04(3.40,6.76)vs 1.68(1.39,2.47)ng/mL,P＜0.05),2.25(1.12,3.58)vs 0.74(0.54,1.54),P＜0.05].The tumor diameter(DIA)was larger(1.9±0.6 vs 1.4±0.3 cm,P＜0.05),and the insulin level at 300 minutes(INS300)during the 5-hour oral glucose tolerance test(5 h OGTT)was significantly elevated(30.47±5.67 vs 9.67±3.32)μIU/mL,P＜0.01).Levels of blood tumor markers AFP,CEA,and CA724 were also increased(P＜0.05).2)Correlation analysis indicated that CP,CPGlu,DIA,INS300,AFP,CEA,and CA724 were positively correlated with malignant insulinoma during hypoglycemia.3)The ROC curve analysis suggested that the optimal cut-off points for distinguishing malignant from benign insulinomas were CP 2.49 ng/mL,CPGlu 1.31,DIA 1.85 cm,and INS300 20.22 μIU/mL,respectively.Conclusions In clinical practice,if an insulinoma patient has a CP level higher than 2.49 ng/mL and a tumor diameter larger than 1.9 cm during hypoglycemia,the possibility of malignant insulinoma should be considered,warranting further examinations and enhanced follow-up.Persistent elevation of AFP,CEA,and CA724 may indicate malignant insulinoma.

ObjectiveTo investigate the role of silent information regulatory protein （SIRT6）/hypoxia-inducible factor-1α （HIF-1α） pathway in regulating the reprogramming of glucose metabolism in ulcerative colitis （UC） and the mechanism of intervention of Shaoyaotang. MethodsForty-eight c57bL/6 mice were randomly divided into a blank group， a model group， a Mesalazine group （0.42 g·kg^-1）， a Shaoyaotang group （31.08 g·kg^-1）， an inhibitor group （OSS-128167， 50 mg·kg^-1）， and an inhibitor + Shaoyaotang group （50 mg·kg^-1 OSS-128167 + 31.08 g·kg^-1 Shaoyaotang）. A UC model was established by the administration of 2.5% dextran sulfate sodium （DSS） solution for mice in other groups for 7 d， except for the blank group. The mice in each group were treated with saline， Mesalazine， Shaoyaotang， inhibitor， and inhibitor + Shaoyaotang， respectively， for 7 d. The mice were necropsied 24 h after the last administration of the drug. The blood was collected from the orbital region， and colon tissue was taken. Hematoxylin-eosin （HE） staining was used to observe the pathological changes in colon tissue. Enzyme-linked immunosorbent assay （ELISA） was employed to detect serum interleukin （IL）-10， IL-17， and IL-6 levels. A biochemical method was used to detect glucose and lactate dehydrogenase A （LDHA） levels. Immunohistochemistry （IHC） was employed to detect IL-22 and transforming growth factor-β₁ （TGF-β₁） levels in colon tissue， and Western blot and real-time fluorescence quantitative polymerase chain reaction （Real-time PCR） were used to detect relative protein and mRNA expressions of SIRT6， HIF-1α， and LDHA. ResultsCompared with those of the blank group， disease activity index （DAI） scores of mice in the model group and inhibitor group were significantly increased （P<0.01）. The length of colon tissue was significantly shortened， and colon tissue was congested and eroded. The pathohistological scores were significantly increased （P<0.01）. The levels of serum inflammatory factors IL-17 and IL-6 were significantly elevated， and the levels of IL-10 were significantly decreased （P<0.01）. The protein expressions of IL-22 and TGF-β₁ were significantly reduced in colon tissue （P<0.01）. The relative protein and mRNA expressions of SIRT6 were significantly decreased （P<0.01）， and the relative protein and mRNA expressions of HIF-1α and LDHA and the contents of glucose and lactate were significantly elevated （P<0.01）. The level of inflammation in the colon of the mice in the inhibitor group was more severe than that in the model group （P<0.01）. Compared with the model group， the Mesalazine group， the Shaoyaotang group， and the inhibitor + Shaoyaotang group showed reduced colonic injury， significant decrease in serum IL-17 and IL-6， significant increase in IL-10 （P<0.01）， significant increase in the protein expressions of IL-22 and TGF-β₁ in colon tissue （P<0.01）， significant increase in the protein expressions of SIRT6 and the relative mRNA expressions （P<0.01）， and significant reduction in the protein expressions of HIF-1α and LDHA， the relative mRNA expressions， and the contents of glucose and lactate （P<0.01）. Compared with those in the Shaoyaotang group， the serum IL-17 and IL-6 were significantly increased， and IL-10 was significantly decreased in the inhibitor + Shaoyaotang group （P<0.01）. The protein expressions of IL-22 and TGF-β₁ in colon tissue were significantly decreased （P<0.01）. The expressions of SIRT6 protein and the relative mRNA expressions were significantly decreased （P<0.01）. The protein expressions of HIF-1α and LDHA， the relative mRNA expressions， and the contents of glucose and lactate were significantly elevated （P<0.01）. However， the difference between the Shaoyaotang group and the Mesalazine group was not significant. ConclusionShaoyaotang can effectively treat DSS-induced mice with UC through the SIRT6/HIF-1α pathway， and its mechanism of action may be related to the regulation of the SIRT6/HIF-1α pathway and glucose metabolism reprogramming and the inhibition of glycolysis.