Gastric variceal rupture and bleeding and hepatic encephalopathy are common and life-threatening complications in decompensated cirrhosis. As a minimally invasive interventional technique， balloon-occluded retrograde transvenous obliteration （BRTO） has made significant progress in the clinical management of gastric varices and hepatic encephalopathy in recent years. This article systematically reviews the technical principles， indications （e.g.， isolated gastric varices and refractory hepatic encephalopathy）， clinical efficacy （an acute hemostasis rate of 85% — 95%， a 1-year rebleeding rate of <15%， and an improvement rate of 60% — 80% for hepatic encephalopathy）， and safety （including complications such as renal impairment and elevated portal vein pressure） of BRTO. Meanwhile， this article discusses the advantages and disadvantages of BRTO and conventional treatment modalities （e.g.， transjugular intrahepatic portosystemic shunt and endoscopic treatment） and reviews the latest technological improvements in recent years， such as coil-assisted retrograde transvenous obliteration and plug-assisted retrograde transvenous obliteration. Future research should focus on the precision of patient selection （e.g.， stratification based on hemodynamic parameters）， the optimization of embolic materials （e.g.， application of new biodegradable embolic agents）， and the development of individualized treatment regimens， so as to improve efficacy and reduce the risk of complications.

Objective To analyze the distribution status of the end digits of standardized blood pressure measurement recordings in the clinic and the effectiveness of standardized blood pressure measurement for community hypertension screening. Methods The first visit blood pressure measurement data from the Community Health Service Center in Jing'an District, Shanghai from June 2023 to May 2024 were collected and analyzed. According to different measurement methods, the data were divided into two groups: standardized blood pressure measurement and conventional blood pressure measurement. SPSS 19.0 software was used for data analysis. The differences in the distribution balance of the end digits of blood pressure values and the detection rate of blood pressure elevation between the two different groups were analyzed. Results The frequency range of the end digits of blood pressure recorded values in the standardized pressure measurement group was 9.42% to 10.83%, and the detection rate of elevated blood pressure was 24.89%. The conventional pressure measurement group had a preference of the end digit ^"0^", and the detection rate of elevated blood pressure was only 2.12%. The results of multiple logistic regression analysis showed that gender, age, season, and different blood pressure measurement modes were all influencing factors for the detection rate of elevated blood pressure. Conclusion The standardized blood pressure measurement mode in the clinic is suitable for community hypertension screening and pressure measurement, with higher data quality than the conventional pressure measurement mode.

OBJECTIVE To construct a preoperative medication reconciliation model assisted by a localized large language model （LLM） for gastric cancer and evaluate its clinical efficacy. METHODS A total of 249 gastric cancer patients with a history of continuous medication before admission in the Gastric Surgery Department of Jiangsu Cancer Hospital were retrospectively enrolled. Patients were divided into training set （154 cases） and validation set （95 cases） based on the order of time. Based on guidelines， drug package inserts， and other evidence， a standardized medication reconcili ation process and a structured knowledge base were constructed. DeepSeek-V3 LLM was deployed privately in the hospital， combined with retrieval-augmented generation technology， to achieve automated integration of medication information， risk screening， and generation of personalized recommendations. The quality of LLM-generated recommendations was evaluated using automatic metrics （BERT Score and ROUGE-1， 2， L） and manual scoring [seven-dimensional index （7DI） ] . Spearman correlation analysis was performed to explore the correlation between automatic scores and manual scores. Cronbach’s α coefficient was used to test the internal consistency of manual scoring results. The time consumed by manual and LLM-assisted medication reconciliation was compared across tasks of different difficulty levels （simple， moderate， and high）. RESULTS A structured knowledge base covering 8 major drug categories was finally established， covering common and high-risk preoperative medication scenarios and providing structured retrieval support for the LLM. For automatic evaluation， the precision， recall， and F1-score of BERT Score were 0.783±0.033， 0.811±0.038， and 0.796±0.028， respectively. The F1-scores of ROUGE-1， ROUGE-2 and ROUGE-L were 0.566±0.067， 0.338±0.076 and 0.468±0.082， respectively. The 7DI scores from three manual raters ranged from 32.06 to 33.45. The F1-score of automatic scoring was significantly positively correlated with the 7DI score of manual scoring （maximum coefficient of determination＝0.611， P ＜0.001）， and the internal consistency of manual scoring was good （Cronbach’s α ＝ 0.876）. In terms of efficiency， LLM-assisted medication reconciliation reduced time consumption by more than 90% compared with manual reconciliation in the simple， moderate， and high-difficulty groups （ P ＜0.001）. CONCLUSIONS The medication reconciliation model constructed based on a localized LLM and structured knowledge base shows high accuracy， consistency， and clinical applicability in complex preoperative medication scenarios for gastric cancer. It can improve the efficiency of medication reconciliation and reduce potential medication risks.

Objective To construct machine learning models based on preoperative inflammatory and radiological features for the prediction of glioma grading and isocitrate dehydrogenase (IDH) mutation status, and to analyze application values of these models and identify the optimal predictive models. Methods A retrospective analysis was conducted on the data of pathologically confirmed glioma patients admitted to Tongji Hospital Affiliated to Tongji University from March 2019 to March 2023. LASSO regression was used to screen feature variables, and predictive models were constructed based on logistic regression (LR), random forest (RF), support vector machine (SVM), gradient boosting decision tree (XGBoost) and K-nearest neighbor (KNN) algorithms. The model performance was comprehensively evaluated using metrics including discrimination ability, area under the precision-recall curve (AUC), accuracy, F1 score and Brier score. The DeLong test was adopted to compare the AUC values among different models; Friedman rank-sum test was used to determine the overall performance differences of the models, with the Nemenyi test applied for multiple comparison correction. Results In the task of glioma grading prediction, the LR model achieved the highest comprehensive score (0.726), and no significant difference was observed between the LR model and the other four models; age was positively correlated with glioma grading (P=0.003). In the task of IDH mutation status prediction, the XGBoost model obtained the highest comprehensive score (0.832), which was superior to the LR (0.762, P=0.035) and KNN models (0.754, P=0.025), while no statistical differences were found between the XGBoost model and the RF or SVM models. Conclusions The LR model for glioma grading prediction and XGBoost model for IDH mutation prediction constructed based on a task-oriented strategy achieve a favorable interpretability while ensuring optimized performance, thereby providing reliable decision support for the individualized diagnosis and treatment of glioma.

Neurocritical care involves complex pathophysiological mechanisms, and its incidence is higher, injuries are more severe, and treatment is more challenging in high-altitude environments. This consensus, based on the latest domestic and international evidence-based medical data, establishes a standardized, goal-oriented framework for neurocritical care management applicable in high-altitude regions and nationwide. The consensus was developed following international standards for evidence quality assessment and underwent two rounds of Delphi expert consultation, resulting in 32 recommendation statements covering three parts: management systems, monitoring and assessment, and core strategies. Key updates include: advocating for the establishment of independent neurocritical care units and implementing precise tiered diagnosis and treatment based on the "Five Differences in Critical Care" concept; constructing a "trinity" multimodal brain monitoring system centered on cerebral blood flow, cerebral oxygenation, and brain function, emphasizing routine bedside transcranial Doppler ultrasound, cerebral oximetry, and continuous electroencephalography monitoring; shifting management strategies from mild hypothermia therapy to targeted temperature management, and defining the "446" target management pathway for the supercritical stage; emphasizing the assessment of static and dynamic cerebrovascular autoregulation functions through multimodal methods to achieve individualized optimal mean arterial pressure management; elevating cerebrospinal fluid management goals to the level of "glymphatic system" function maintenance; implementing a multidisciplinary collaborative, whole-process management model focusing on patients' long-term neurological functional outcomes; de-escalation criteria include multidimensional indicators such as recovery of brain structure, restoration of cerebrovascular autoregulation, improvement in cerebrospinal fluid dynamics, and reduction in biomarker levels; and integrating cutting-edge technologies like artificial intelligence into post-critical care management and rehabilitation planning. This consensus systematically integrates the entire process of neurocritical care management, reflecting the modern connotation of goal-oriented, dynamic, and multimodal integration in neurocritical care medicine. It aims to adapt to new trends such as deepening understanding of pathophysiological mechanisms, the integration of medicine and engineering, and the empowerment of artificial intelligence, thereby further advancing the discipline of critical care medicine.

ObjectiveThis paper aims to investigate the differential mechanisms underlying the staged therapeutic effects of Qijia Rougan formula on liver fibrosis using proteomic technology. MethodsThe staged rat model of liver fibrosis was established by subcutaneous injection of carbon tetrachloride （CCl₄） and olive oil. One hundred and four SD rats were randomized into thirteen groups：a normal group，a two-week model group，a four-week model group，a six-week model group，an eight-week model group，a two-week Qijia Rougan formula group，a four-week Qijia Rougan formula group，a six-week Qijia Rougan formula group，an eight-week Qijia Rougan formula group，a two-week compound Biejia Ruangan tablet group，a four-week Compound Biejia Ruangan Tablet group，a six-week Compound Biejia Ruangan Tablet group，and an eight-week compound Biejia Ruangan tablet group. After two weeks of drug intervention，liver tissue and abdominal aortic blood samples were collected from the rats for testing. Hematoxylin-eosin （HE） staining，Masson staining，and Picro Sirius red staining were used to observe pathological damage and collagen fiber deposition in liver tissues. Immunohistochemistry （IHC） was employed to detect the contents of fibrosis markers in liver tissues. The contents of liver function indicators in the serum were measured using a fully automated biochemical analyzer，and the levels of liver fibrosis indicators in the serum were assessed by enzyme-linked immunosorbent assay （ELISA）. Liver tissues from the normal group，each model group，and each Qijia Rougan formula group were subjected to label-free quantitative proteomic analysis to identify differential proteins among the groups，with key proteins validated by Western blot. Finally，bioinformatics analysis was performed on the differential proteins. Results（1） The staged rat model of liver fibrosis constructed with CCl₄ and olive oil showed pathological results at the 2^nd，4^th，6^th，and 8^th weeks of modeling that were consistent with the Metavir standards for the F1，F2，F3，and F4 stages. Compared with those in the normal control group，the protein expressions of α-smooth muscle actin （α-SMA） and Collagen Ⅰ were significantly increased in each stage （P<0.05）. The levels of liver function indicators in the serum，including alanine aminotransferase （ALT），aspartate aminotransferase （AST），alkaline phosphatase （ALP），direct bilirubin （DBIL），and total bilirubin （TBil） in each model group，were significantly elevated in each stage （P<0.01）. The levels of liver fibrosis indicators in the serum，including procollagen Ⅲ peptide （PⅢP），type Ⅳ collagen（Ⅳ-C），hyaluronic acid （HA），and laminin （LN） in each model group，were significantly increased in each stage （P<0.05，P<0.01）. This study successfully established a staged rat model of liver fibrosis. （2） Compared with the model groups at each stage，the administration groups showed a reduction in hepatocyte ballooning degeneration，a more orderly arrangement of hepatocytes，and a decrease of inflammatory cell infiltration. The blue-stained collagen fibers became significantly thinner and finer，with reduced and narrowed fibrous septa. The areas of collagen fibers and Picro Sirius red staining were reduced （P<0.05）. The positive areas of α-SMA and Collagen Ⅰ expression were significantly decreased （P<0.05）. The levels of ALT，AST，ALP，DBIL，and TBil in the rats of the model groups at each stage were significantly reduced （P<0.05，P<0.01）. The levels of PⅢP，Ⅳ-C，HA，and LN in the rats of the model groups at each stage were significantly decreased （P<0.05）. Among these，the improvements in all indicators were most significant in the F3 stage （P<0.01）.（3） The proteomic results show that a total of 165 differential proteins exhibit a callback trend when comparing the model groups at four stages with the normal group，and when comparing the Qijia Rougan formula group with the model group. Western blot analysis reveals that the levels of NAD（P）H：quinone oxidoreductase 1 （NQO1），mitogen-activated protein kinase 1 （MAPK1），arginase 1 （Arg1），and glutathione S-transferase α1 （GSTA1） were consistent with the proteomic results. Bioinformatics results reveal that 165 differentially expressed proteins are enriched in multiple signaling pathways. Notably，signaling pathways such as drug metabolism-cytochrome P450，arginine biosynthesis，and the peroxisome proliferator-activated receptor （PPAR） signaling pathway were found to be closely associated with liver fibrosis，suggesting that the Qijia Rougan formula may exert its staged regulatory effects on liver fibrosis by regulating these pathways. ConclusionThe Qijia Rougan formula may achieve staged regulation of liver fibrosis by regulating drug metabolism-cytochrome P450，arginine biosynthesis，and the PPAR signaling pathway.

AIM: To investigate the effects of Conbercept on various optical coherence tomography(OCT)biomarkers in patients with retinal vein occlusion-related macular edema(RVO-ME), and to analyze the correlation of these biomarker changes with visual prognosis.METHODS: Retrospective study. A total of 57 patients(57 eyes)with RVO-ME, including 25 patients(25 eyes)with central retinal vein occlusion(CRVO)and 32 patients(32 eyes)with branch retinal vein occlusion(BRVO), were enrolled in this study. All the patients received intravitreal injection of conbercept once a month, three times in total. The preoperative and postoperative best-corrected visual acuity(BCVA), and changes in OCT biomarkers, including central macular thickness(CMT), the length of disorganization of the retinal inner layers(DRIL), the number of hyperreflective dots(HRD), the area of intraretinal fluid(IRF), the area of subretinal fluid(SRF), and the length of ellipsoid zone(EZ)disruption were compared. Furthermore, the relationship of these changes with BCVA was analyzed.RESULTS:Compared with the baseline, at 3 mo post-treatment, BCVA(LogMAR)was improved, CMT was decreased, the length of DRIL was shortened, the number of HRD was reduced, the area of IRF was decreased, the area of SRF was reduced, and the length of EZ disruption was shortened(all P<0.05). Spearman correlation analysis showed that there was no correlation between the changes in CMT, the length of DRIL, the number of HRD, the area of IRF, the area of SRF and the change in BCVA before and after treatment(P>0.05). However, the change in the length of EZ disruption was positively correlated with the change in BCVA(rs=0.34, P=0.011), and the R2 value of the fitting curve between the change in the length of EZ disruption and the change in BCVA was 0.113(P=0.011). When comparing the pre- and post-treatment changes in BCVA, the length of DRIL, the number of HRD, the area of IRF, the area of SRF, and the length of EZ disruption between patients in the CRVO group and BRVO group, no significant differences were observed(all P>0.05). In contrast, a significant difference was found in the change in CMT between the two groups(P=0.002).CONCLUSION:Conbercept effectively improves multiple OCT biomarkers in patients with RVO-ME. Repair of EZ disruption is a key driver of visual recovery, and its stability may serve as a novel indicator for personalized decision-making in anti-vascular endothelial growth factor therapy.

Objective: To investigate the association between outdoor light at night (LAN) exposure and sleep quality among primary and secondary school students, and to examine the pathways of negative affects including depressive, stress and anxiety symptoms, so as to provide a theoretical basis for optimizing the school environment and enhancing the physical and mental well being of students. Methods: In December 2024, a total of 36 885 students from 154 primary and secondary schools in Suzhou, Nantong, and Changzhou were included via a stratified cluster sampling method. Sleep quality and negative affect were assessed by using the Pittsburgh Sleep Quality Index (PSQI), Center for Epidemiologic Studies Depression Scale (CES-D), Generalized Anxiety Disorder-7 (GAD-7), and Depression, Anxiety and Stress Scale-21 (DASS-21), respectively. Outdoor LAN exposure data were obtained from the national polar orbiting partnership visible infrared imaging radiometer suite nighttime light(NPP-VIIRS NTL) dataset provided by the National Earth System Science Data Center. Multivariate Logistic regression model was employed to analyze the association between LAN exposure and sleep quality across different regions, with stratification by monitoring site. Spearman correlation analysis was used to examine the relationships between LAN, negative affect, and sleep quality. The mediating role of negative affect was tested by using Model 4 of the PROCESS macro. Results: The detection rates among students were 13.95%( n =5 147) for depressive symptom, 16.72%( n =6 166) for stress symptom, and 17.49%( n =6 451) for anxiety symptom. The median outdoor LAN exposure was 28.85(19.10, 41.44)nW/(cm 2 · ( sr). After adjusting for confounders, multivariate Logistic regression analysis showed that high LAN exposure ( Q 4) was positively associated with sleep problems (urban areas: OR =1.28, 95% CI = 1.17- 1.41; rural areas: OR =1.21, 95% CI =1.07-1.36; both P <0.05). Spearman correlation analysis revealed significant positive correlations between LAN exposure, depressive symptoms, stress symptoms, anxiety symptoms, and sleep quality ( r =0.03-0.75, all P < 0.01). The mediation analysis indicated that all dimensions of negative affect significantly mediated the relationship between high LAN exposure and poor sleep quality (all P <0.01). Specifically, the indirect effects were 0.03 (95% CI =0.02-0.05) for depressive symptom, 0.05(95% CI =0.03-0.08) for stress symptom, and 0.07(95% CI =0.05-0.09) for anxiety symptom. Overall, 31.9% of the total effect was mediated by negative affect, with anxiety (14.89%) being the strongest mediator, followed by stress (10.64%) and depression(6.38%). Conclusion Reducing high levels of outdoor LAN exposure and implementing interventions targeting negative affect may contribute to improved sleep quality among primary and secondary school students.

BACKGROUND: The potential impact of pre-existing coronary artery stenosis (CAS) on right ventricular (RV) function during acute pulmonary embolism (PE) episodes remains underexplored. This study aimed to investigate the association between pre-existing CAS and RV dysfunction in patients with acute PE. METHODS: In this multicenter, case-control study, 89 cases and 176 controls matched for age were enrolled at three study centers (Peking Union Medical College Hospital, Fuwai Hospital, and the Second Affiliated Hospital of Harbin Medical University) from January 2016 to December 2020. The cases were patients with acute PE with CAS, and the controls were patients with acute PE without CAS. Coronary artery assessment was performed using coronary computed tomographic angiography. CAS was defined as ≥50% stenosis of the lumen diameter in any coronary vessel >2.0 mm in diameter. Conditional logistic regression analysis was used to evaluate the association between CAS and RV dysfunction. RESULTS: The percentages of RV dysfunction (19.1% [17/89] vs. 44.6% [78/176], P <0.001) and elevated systolic pulmonary artery pressure (sPAP) (19.3% [17/89] vs. 39.5% [68/176], P = 0.001) were significantly lower in the case group than those in the control group. In the multivariable logistic regression model, CAS was independently and negatively associated with RV dysfunction (adjusted odds ratio [OR]: 0.367; 95% confidence interval [CI]: 0.185-0.728; P = 0.004), and elevated sPAP (OR: 0.490; 95% CI: 0.252-0.980; P = 0.035), respectively. CONCLUSIONS Pre-existing CAS was significantly and negatively associated with RV dysfunction and elevated sPAP in patients with acute PE. This finding provides new insights into RV dysfunction in patients with acute PE with pre-existing CAS.
Humans ; Pulmonary Embolism/complications* ; Case-Control Studies ; Male ; Ventricular Dysfunction, Right/physiopathology* ; Female ; Middle Aged ; Aged ; Coronary Stenosis/complications* ; Logistic Models ; Adult

With the emergence of deep learning techniques based on convolutional neural networks, artificial intelligence (AI) has driven transformative developments in the field of medical image analysis. Recently, large language models (LLMs) such as ChatGPT have also started to achieve distinction in this domain. Increasing research shows the undeniable role of AI in reshaping various aspects of medical image analysis, including processes such as image enhancement, segmentation, detection in image preprocessing, and postprocessing related to medical diagnosis and prognosis in clinical settings. However, despite the significant progress in AI research, studies investigating the recent advances in AI technology in the aforementioned aspects, the changes in research hotspot trajectories, and the performance of studies in addressing key clinical challenges in this field are limited. This article provides an overview of recent advances in AI for medical image analysis and discusses the methodological profiles, advantages, disadvantages, and future trends of AI technologies.
Artificial Intelligence ; Humans ; Image Processing, Computer-Assisted/methods* ; Neural Networks, Computer ; Deep Learning ; Diagnostic Imaging/methods*