OBJECTIVE To investigate the influence of CYP2C19 gene polymorphism on platelet function and inflammatory cytokines in elderly patients with ischemic stroke， and to analyze potential factors associated with poor prognosis. METHODS A retrospective study was conducted on elderly patients with ischemic stroke admitted to our hospital from June 2024 to June 2025， wh o underwent CYP2C19 genotype testing and received antiplatelet therapy with clopidogrel. The levels of platelet function indicators and inflammatory cytokines before and after treatment were compared among patients with different metabolic phenotypes. Based on the prognosis at 6 months post-treatment， patients were divided into poor prognosis group and good prognosis group. Univariate analysis was performed on general data， metabolic phenotype， the levels of platelet function indicators and inflammatory cytokines. Variables with P ＜0.05 and the levels of inflammatory cytokines before treatment were included in a multivariate Logistic regression analysis to identify independent risk factors for poor prognosis. Multiple linear regression was used to further analyze the relationship between metabolic phenotypes and inflammatory cytokines. RESULTS A total of 448 elderly patients with ischemic stroke were included； among them， 162 cases were normal metabolic phenotype， 218 were intermediate metabolic phenotype， and 68 were poor metabolic phenotype. No rapid or ultrarapid metabolic phenotypes were observed. After treatment， platelet aggregation rate， the levels of P-selectin and platelet activated complex-1 （PAC-1）， high-sensitivity C-reactive Protein （hs-CRP）， interleukin-1β （IL-1β）， IL-6 and tumor necrosis factor-α （TNF-α） in the normal metabolic phenotype group， intermediate metabolic phenotype group， and poor metabolic phenotype group （except for platelet aggregation rate， and the levels of P-selectin and PAC-1 in the poor metabolic phenotype group） were significantly lower than those before treatment in the same group. Moreover， the above indicators in the normal metabolic phenotype group were significantly lower than those in the intermediate and poor metabolic phenotype groups at the corresponding time， and the levels of platelet function indicators in the intermediate metabolic phenotype group were significantly lower than those in the poor metabol ic phenotype group at the corresponding time （ P ＜0.05）. Univariate and multivariate Logistic regression analyses showed that combined with hypertension， combined with diabetes mellitus， and intermediate or poor metabolic genotypes were independent risk factors for poor prognosis in elderly patients with ischemic stroke （ P ＜0.05）. Multiple linear regression analysis showed that serum levels of hs-CRP， IL-1β， IL-6 and TNF-α before treatment were significantly higher in patients with intermediate and poor metabolic genotypes compared to those with normal metabolic genotype （ P ＜0.05）， with a greater magnitude of increase in inflammatory cytokines observed in the patients with poor metabolic genotype. CONCLUSIONS The elderly ischemic stroke patients with CYP2C19 intermediate and poor metabolic genotypes have poor inhibition effect on platelet and higher levels of inflammatory cytokines than normal metabolic genotype； CYP2C19 gene polymorphism， and in combination with hypertension and diabetes， can be used as independent predictors of poor prognosis.

ObjectiveDiscriminating atypical hepatocellular carcinoma (HCC) from other malignancies in liver nodules classified as Liver Imaging Reporting and Data System category M (LR-M) remains a significant diagnostic challenge on conventional ultrasound examination. The LR-M category, originally intended to capture non-HCC malignancies, paradoxically contains up to 63% of atypical HCCs that deviate from classic enhancement patterns, leading to potential misdiagnosis and suboptimal treatment planning. While deep learning has shown promise in HCC diagnosis, most existing models rely exclusively on single-modality ultrasound, overlooking the diagnostic benefits of integrating complementary information from multiple imaging sources. To address this gap, we propose a novel attention-weighted tri-modal ultrasound network (TUS-Net) that integrates contrast-enhanced ultrasound (CEUS), B-mode ultrasound (BUS), and time-intensity curves (TICs) to improve diagnostic accuracy for these clinically challenging lesions. MethodsOur framework incorporates a three-dimensional convolutional neural network (C3D) backbone to extract spatiotemporal features from CEUS videos, capturing dynamic vascular patterns critical for lesion characterization. To effectively fuse complementary modalities, we introduce a dual-channel feature fusion module (DCFFM) that adaptively combines features from CEUS and BUS through channel-wise attention mechanisms, allowing the model to dynamically weigh the contribution of each modality based on diagnostic relevance. Additionally, we propose a temporal intensity feature fusion module (TIFFM) that leverages quantitative hemodynamic information from TICs to guide the model’s attention toward diagnostically critical temporal phases, such as arterial wash-in and portal venous washout. The model is further enhanced by automated lesion localization using YOLOX and class activation mapping for interpretability, ensuring that predictions align with clinically meaningful imaging features. ResultsEvaluated on a tri-modal ultrasound dataset comprising 161 patients with pathologically confirmed LR-M nodules (131 atypical HCC and 30 non-HCC malignancies), our model achieved an accuracy of 86.83%, a sensitivity of 92.50%, a specificity of 75.50%, and an AUC of 89.32% in screening atypical HCC. Compared to single-modality baselines, TUS-Net demonstrated superior specificity, a clinically critical metric given the higher risk associated with misclassifying non-HCC malignancies. Ablation studies confirmed the contribution of each module, with the full model outperforming both standard C3D and 3D ResNet backbones integrated with attention mechanisms. A reader study involving junior and senior radiologists further validated the clinical utility of AI assistance, showing consistent improvements in specificity and inter-reader consistency, particularly for less experienced clinicians. ConclusionThese results surpass existing benchmark models and demonstrate the potential of our approach to enhance diagnostic precision in clinically specific cases. By intelligently fusing multi-modal ultrasound data with attention-guided mechanisms, TUS-Net offers a reliable and interpretable tool that holds promise for improving the non-invasive diagnosis of atypical HCC in challenging LR-M liver nodules.

ObjectiveDiscriminating atypical hepatocellular carcinoma (HCC) from other malignancies in liver nodules classified as Liver Imaging Reporting and Data System category M (LR-M) remains a significant diagnostic challenge on conventional ultrasound examination. The LR-M category, originally intended to capture non-HCC malignancies, paradoxically contains up to 63% of atypical HCCs that deviate from classic enhancement patterns, leading to potential misdiagnosis and suboptimal treatment planning. While deep learning has shown promise in HCC diagnosis, most existing models rely exclusively on single-modality ultrasound, overlooking the diagnostic benefits of integrating complementary information from multiple imaging sources. To address this gap, we propose a novel attention-weighted tri-modal ultrasound network (TUS-Net) that integrates contrast-enhanced ultrasound (CEUS), B-mode ultrasound (BUS), and time-intensity curves (TICs) to improve diagnostic accuracy for these clinically challenging lesions. MethodsOur framework incorporates a three-dimensional convolutional neural network (C3D) backbone to extract spatiotemporal features from CEUS videos, capturing dynamic vascular patterns critical for lesion characterization. To effectively fuse complementary modalities, we introduce a dual-channel feature fusion module (DCFFM) that adaptively combines features from CEUS and BUS through channel-wise attention mechanisms, allowing the model to dynamically weigh the contribution of each modality based on diagnostic relevance. Additionally, we propose a temporal intensity feature fusion module (TIFFM) that leverages quantitative hemodynamic information from TICs to guide the model’s attention toward diagnostically critical temporal phases, such as arterial wash-in and portal venous washout. The model is further enhanced by automated lesion localization using YOLOX and class activation mapping for interpretability, ensuring that predictions align with clinically meaningful imaging features. ResultsEvaluated on a tri-modal ultrasound dataset comprising 161 patients with pathologically confirmed LR-M nodules (131 atypical HCC and 30 non-HCC malignancies), our model achieved an accuracy of 86.83%, a sensitivity of 92.50%, a specificity of 75.50%, and an AUC of 89.32% in screening atypical HCC. Compared to single-modality baselines, TUS-Net demonstrated superior specificity, a clinically critical metric given the higher risk associated with misclassifying non-HCC malignancies. Ablation studies confirmed the contribution of each module, with the full model outperforming both standard C3D and 3D ResNet backbones integrated with attention mechanisms. A reader study involving junior and senior radiologists further validated the clinical utility of AI assistance, showing consistent improvements in specificity and inter-reader consistency, particularly for less experienced clinicians. ConclusionThese results surpass existing benchmark models and demonstrate the potential of our approach to enhance diagnostic precision in clinically specific cases. By intelligently fusing multi-modal ultrasound data with attention-guided mechanisms, TUS-Net offers a reliable and interpretable tool that holds promise for improving the non-invasive diagnosis of atypical HCC in challenging LR-M liver nodules.

Background Prehospital emergency medical personnel (PEMP) are exposed to long-term high-pressure work, which can exacerbate psychological distress and impair job satisfaction and sleep quality. However, in-depth research on the interactions among these factors is lacking. Objective To assess the status of psychological distress, job satisfaction, and sleep quality of PEMP in Guangzhou and to explore the mediating role of sleep quality in the relationship between psychological distress and job satisfaction. Methods From February to May 2025, 1085 PEMP from "120" emergency network hospitals in Guangzhou were selected using convenience sampling. Data were collected via the General Information Questionnaire, Kessler Psychological Distress Scale, Minnesota Satisfaction Questionnaire, and Pittsburgh Sleep Quality Index. Statistical analyses were performed using SPSS 25.0, and The mediation model of sleep quality in linking psychological distress and job satisfaction was constructed using AMOS 28.0. The bias-corrected Bootstrap method was employed to assessed the significance of the mediating effect. Results A total of 1063 valid responses were received (97.97% valid response rate). The mean scores were: psychological distress (27.99±10.75), job satisfaction (69.45±15.84), and sleep quality (9.82±4.47). Significant differences in the three scores were found across gender, age, monthly night shift frequency, and hospital grade (P<0.05). Higher job satisfaction was linked to lower psychological distress and better sleep quality and its dimensions, while psychological distress directly correlated with poorer sleep quality (P<0.01). Sleep quality partially mediated the relationship between psychological distress and job satisfaction, with a mediating effect of −0.195, accounting for 43.62% of the total effect. Conclusion The participants report moderate psychological distress, moderate-to-high job satisfaction, and poor sleep quality. Psychological distress directly affects job satisfaction and indirectly through its impact on sleep quality. Interventions aimed at improving sleep health and mental health are essential to improve personnel well-being and work efficiency.

Background Prehospital emergency medical personnel (PEMP) are exposed to long-term high-pressure work, which can exacerbate psychological distress and impair job satisfaction and sleep quality. However, in-depth research on the interactions among these factors is lacking. Objective To assess the status of psychological distress, job satisfaction, and sleep quality of PEMP in Guangzhou and to explore the mediating role of sleep quality in the relationship between psychological distress and job satisfaction. Methods From February to May 2025, 1085 PEMP from "120" emergency network hospitals in Guangzhou were selected using convenience sampling. Data were collected via the General Information Questionnaire, Kessler Psychological Distress Scale, Minnesota Satisfaction Questionnaire, and Pittsburgh Sleep Quality Index. Statistical analyses were performed using SPSS 25.0, and The mediation model of sleep quality in linking psychological distress and job satisfaction was constructed using AMOS 28.0. The bias-corrected Bootstrap method was employed to assessed the significance of the mediating effect. Results A total of 1063 valid responses were received (97.97% valid response rate). The mean scores were: psychological distress (27.99±10.75), job satisfaction (69.45±15.84), and sleep quality (9.82±4.47). Significant differences in the three scores were found across gender, age, monthly night shift frequency, and hospital grade (P<0.05). Higher job satisfaction was linked to lower psychological distress and better sleep quality and its dimensions, while psychological distress directly correlated with poorer sleep quality (P<0.01). Sleep quality partially mediated the relationship between psychological distress and job satisfaction, with a mediating effect of −0.195, accounting for 43.62% of the total effect. Conclusion The participants report moderate psychological distress, moderate-to-high job satisfaction, and poor sleep quality. Psychological distress directly affects job satisfaction and indirectly through its impact on sleep quality. Interventions aimed at improving sleep health and mental health are essential to improve personnel well-being and work efficiency.

Objective To investigate the risk factors for postoperative recurrence and metastasis in patients with T1 stageⅢ/Ⅳ clear cell renal cell carcinoma(ccRCC), and to analyze the influence of surgical methods on prognosis, so as to provide prognostic evaluation and treatment guidance for clinical practice. Methods A retrospective analysis was conducted on 398 patients diagnosed with T1 stage Ⅲ/Ⅳ ccRCC who underwent either partial nephrectomy(PN)or radical nephrectomy(RN)at Jiangsu Province Hospital during Jan. 2008 and Sep. 2022. The patients were divided into recurrence group(n=49)and non-recurrence group(n=349)according to whether the tumor recurred during the follow-up period. The general, clinical data and prognosis were compared between the two groups. Univariate and multivariate Cox regression analyses were used to explore the factors influencing disease-free survival(DFS)and overall survival(OS)after surgery. Subgroup analysis using Kaplan-Meier curves was performed for T1a and T1b patients. Results During the median follow-up of 67(42, 98)months, the recurrence group was characterized by older age, larger tumor diameter, higher proportion of T1b stage, and more complex tumor anatomy. Univariate and multivariate Cox regression analyses showed that age ≥55 years and R.E.N.A.L. score ≥7 were independent risk factors for poorer DFS and OS(P＜0.05). Subgroup analysis revealed that PN or RN did not significantly affect DFS or OS in T1a patients and OS in T1b patients, while RN significantly improved DFS in T1b patients(P＜0.05). Conclusion Age ≥55 years and R.E.N.A.L. score ≥7 are independent risk factors for poor DFS and OS in highgrade T1 ccRCC patients. For T1b patients, RN may improve DFS.

Objective: To analyze the time-related factors influencing daily blood donation attendance at fixed donation sites in Nanjing and the heterogeneity among sites, and to provide evidence for optimizing site layout and operational management. Methods: Based on daily attendance data from 17 fixed donation sites in Nanjing for the full year of 2025, matched with contemporaneous meteorological and air quality data, a two-stage negative binomial generalized linear mixed model (NB-GLMM) was constructed. In the second stage, site-level static characteristics (number of scenic spots, daily operating hours, number of shopping malls, public transport stops, hospitals, universities, and residential buildings) were additionally incorporated. Model comparison was performed using Akaike information criterion (AIC). Significance of random effects was tested by likelihood ratio test, and site-specific random effects were extracted using empirical Bayes estimation. Robustness of the results was verified through residual diagnostics and multiple sensitivity analyses. Results: The first-stage model showed that rainfall was associated with lower attendance (IRR=0.943, P<0.001); holidays were associated with higher attendance (IRR=1.162, P<0.001); Saturdays and Sundays showed higher attendance than Mondays, with the strongest association on Sundays (IRR=1.195, P<0.001). Temperature exhibited a non-linear relationship (linear term P=0.051, quadratic term P=0.036). The joint results of the linear and quadratic terms for temperature suggested a nonlinear relationship between attendance and temperature. After introducing static characteristics in the second stage, the number of scenic spots (IRR=1.372, P=0.008) and daily operating hours (IRR=1.823, P<0.001) were found to be positively associated with attendance, while the numbers of shopping malls, public transport stops, hospitals, universities, and residential buildings showed no significant effects. Conclusion: Daily blood donation attendance at fixed sites in Nanjing is influenced by multiple factors including weather, day of the week, holidays, and site surroundings. Substantial heterogeneity exists among different sites; extended operating hours and more nearby scenic spots are associated with increased attendance.

Gastric cancer remains a highly prevalent malignancy worldwide, with surgical resection currently constituting the cornerstone of treatment aimed at improving long-term patient survival. Owing to their notable advantages, including reduced surgical trauma and accelerated postoperative recovery, minimally invasive procedures are progressively supplanting conventional open surgery and have become the mainstream approach in gastric cancer management. Concurrently, the rapid advancement of artificial intelligence (AI) technologies has enabled real-time intraoperative monitoring of surgical scenes, thereby furnishing novel technical support for adjunctive decision-making, surgical navigation, and skill assessment during gastrectomy. This article provides a systematic review of the current status of AI applications in minimally invasive gastric cancer surgery, with a particular focus on research progress pertaining to instrument recognition, surgical phase identification, delineation of normal anatomical structures, detection of metastatic foci, and early warning of intraoperative adverse events. Furthermore, we discuss the potential value of AI in enhancing surgical efficiency, ensuring patient safety, and optimizing surgical education. On this basis, we further analyze the principal challenges and inherent risks confronting current AI systems, with the aim of informing future technological innovation and facilitating clinical translation.

OBJECTIVE To establish standards for the rational use of Compound amino acid injection （14AA-SF） （hereinafter referred to as “14AA-SF”）， and evaluate its rationality and cost-effectiveness.METHODS The standards for the rational use of 14AA-SF were established based on the Delphi method. Archived medical records of patients who used 14AA-SF in the Dept. of Neurosurgery of People’s Hospital of Bozhou from January 2024 to December 2025 were randomly selected. On the basis of the established standards for the rational use， the attribute hierarchical model was applied to evaluate the use rationality of 14AA-SF in the included medical records. The decision tree model was applied to calculate the incremental cost-effectiveness ratio （ICER）， so as to evaluate the cost-effectiveness of 14AA-SF administered via single-bottle infusion versus “all-in-one” infusion， and the robustness of the cost-effectiveness evaluation results was verified through sensitivity analysis.RESULTS A total of 205 medical records were collected， among which 163 records （79.51%） were qualified with medical record score （MRS） ≥90， and 42 records （20.49%） were unqualified with MRS ＜90. The main irrational types of 14AA-SF use included unreasonable infusion method （67 case-times， accounting for 32.68%）， unreasonable amino acid dosage （62 case-times， accounting for 30.24%） and unreasonable non-protein calorie-nitrogen ratio （55 case-times， accounting for 26.83%）. Compared wtih single-bottle infusion， the incremental cost of “all-in-one” infusion for each 1% increase in effective rate was 165.40 yuan. The one-way sensitivity analysis showed that the effective rate of single-bottle infusion had the greatest impact on ICER. The probabilistic sensitivity analysis showed that under the willingness-to-pay （WTP） threshold （89 358 yuan） set in this study， the cost-effectiveness probability of “all-in-one” infusion relative to single-bottle infusion was 89.9%. When the WTP threshold exceeds 17 871.60 yuan， the “all-in-one” infusion is more cost-effective than the single-bottle infusion.CONCLUSIONS The established standards for the rational use of 14AA-SF are practical and feasible. There are phenomena of irrational use of 14AA-SF in Dept. of Neurosurgery of the hospital in terms of indications， nutritional risk screening， infusion methods and other aspects， which need to be further standardized. The “all-in-one” infusion is more cost-effective than the single-bottle infusion.

Objective To investigate the population density and seasonal fluctuations of Aedes albopictus in Guangzhou City, Guangdong Province, from 2021 to 2023, so as to provide insights into A. albopictus control and management of dengue fever. Methods The surveillance of A. albopictus density was performed in all surveillance sites assigned across all streets (townships) in Guangzhou City during the period from January to December from 2021 to 2023. The surveillance frequency was twice every half month from May to September, and once every month for the rest of a year. In each surveillance period, A. albopictus mosquito larvae were captured from indoor and outdoor small water containers in residential areas, parks, medical facilities, schools, other government sectors and social organizations, construction sites, special industries and others for mosquito species identification. Adult mosquitoes were captured using electric mosquito suction apparatus for species identification and gender classification. Adult mosquitoes and mosquito eggs were collected with mosquito and egg traps at the breeding and dwelling places of Aedes mosquitoes for identification. The mosquito oviposition index (MOI), Breteau index (BI), adult mosquito density index (ADI) and standard space index (SSI) were calculated. The A. albopictus density was classified into grades 0, 1, 2 and 3 in each surveillance site, with Grade 0 density defined eligible, and the eligible rate of A. albopictus density was calculated at all surveillance sites each year from 2021 to 2023. In addition, the changing trends in MOI, SSI, BI and ADI of A. albopictus were analyzed in Guangzhou City from 2021 to 2023. Results The eligible rates of A. albopictus density were 61.69%, 68.75% and 55.15% in surveillance sites of Guangzhou City from 2021 to 2023 (χ² = 297.712, P < 0.001), and appeared a tendency towards a reduction followed by a rise each year, which gradually reduced since January, maintained at a low level during the period between May and October, and gradually increased from November to December. The MOI, SSI, BI and ADI of A. albopictus all appeared a tendency towards a rise followed by a reduction in Guangzhou City during the period between January and December from 2021 to 2023. The BI of A. albopictus peaked in the first half of June in 2021 (4.03), the first half of July in 2022 (3.89) and the last half of August in 2023 (5.02), and the SSI of A. albopictus peaked in the last half of June in 2021 (0.93), the last half of May in 2022 (0.59), and the last half of June (0.94) and the first half of September in 2023 (1.12). In addition, the MOI of A. albopictus peaked in the first half of May in 2021 (8.64), the first half of June in 2022 (8.96), and the last half of May (10.21) and the last half of June in 2023 (10.89), and the ADI of A. albopictus peaked in the first half of June in 2021 (3.41), the last half of June in 2022 (4.06), and the first half of July in 2023 (3.61). Conclusions The density of A. albopictus is high in Guangzhou City during the period from May to October, and the risk of local outbreak caused by imported dengue fever is high. Persistent intensified surveillance of the density and seasonal fluctuation of A. albopictus is recommended and timely mosquito prevention and control is required according to the fluctuation in the A. albopictus density.