Objective:To develop and evaluate a medical visual question answering (VQA) model for neonatal lung ultrasound (LUS) images to enhance intelligent auxiliary diagnosis of neonatal pulmonary diseases.Methods:Using data from neonates admitted to Beijing Obstetrics and Gynecology Hospital, Capital Medical University (January 2023 to December 2024), an image-question-answer dataset comprising 251 LUS images was constructed [43 pneumonia (17.1%), 42 neonatal respiratory distress syndrome (16.7%), 83 transient tachypnea (33.1%), and 83 normal (33.1%) images] with a four-tier medical question-answer framework. Building upon the Qwen2.5-VL-7B base model and integrating LoRA fine-tuning with chain-of-thought prompting, we developed the NLUS-VQA model to enhance visual-language semantic alignment and enable stepwise clinical reasoning, achieving efficient small-sample adaptation. Model performance was comprehensively assessed through natural language generation metrics (BLEU-4, ROUGE-1/2/L), qualitative evaluation of characteristic recognition, and clinical consistency analysis.Results:(1) Quantitative evaluation demonstrated that NLUS-VQA achieved scores of 22.38 (BLEU-4), 48.26 (ROUGE-1), 22.40 (ROUGE-2), and 37.20 (ROUGE-L), representing significant improvements over baseline models. (2) Qualitatively, the model exhibited strong performance in identifying lung consolidation, coalescent B-lines, and snowflake signs, with its chain-of-thought strategy enhancing clinical interpretability and answer accuracy. (3) Clinically, NLUS-VQA achieved a Cohen's Kappa coefficient of 0.78 and diagnostic accuracy of 80.8% (21/26), indicating substantial agreement with clinical experts.Conclusion:The NLUS-VQA model demonstrates robust interpretability in recognizing key sonographic patterns (e.g. lung consolidation, confluent B-lines, and snowflake signs), providing a scalable framework for small-sample medical image analysis, though diagnostic performance on complex conditions remains limited by dataset scale and minority class representation.

Objective:To develop and evaluate a medical visual question answering (VQA) model for neonatal lung ultrasound (LUS) images to enhance intelligent auxiliary diagnosis of neonatal pulmonary diseases.Methods:Using data from neonates admitted to Beijing Obstetrics and Gynecology Hospital, Capital Medical University (January 2023 to December 2024), an image-question-answer dataset comprising 251 LUS images was constructed [43 pneumonia (17.1%), 42 neonatal respiratory distress syndrome (16.7%), 83 transient tachypnea (33.1%), and 83 normal (33.1%) images] with a four-tier medical question-answer framework. Building upon the Qwen2.5-VL-7B base model and integrating LoRA fine-tuning with chain-of-thought prompting, we developed the NLUS-VQA model to enhance visual-language semantic alignment and enable stepwise clinical reasoning, achieving efficient small-sample adaptation. Model performance was comprehensively assessed through natural language generation metrics (BLEU-4, ROUGE-1/2/L), qualitative evaluation of characteristic recognition, and clinical consistency analysis.Results:(1) Quantitative evaluation demonstrated that NLUS-VQA achieved scores of 22.38 (BLEU-4), 48.26 (ROUGE-1), 22.40 (ROUGE-2), and 37.20 (ROUGE-L), representing significant improvements over baseline models. (2) Qualitatively, the model exhibited strong performance in identifying lung consolidation, coalescent B-lines, and snowflake signs, with its chain-of-thought strategy enhancing clinical interpretability and answer accuracy. (3) Clinically, NLUS-VQA achieved a Cohen's Kappa coefficient of 0.78 and diagnostic accuracy of 80.8% (21/26), indicating substantial agreement with clinical experts.Conclusion:The NLUS-VQA model demonstrates robust interpretability in recognizing key sonographic patterns (e.g. lung consolidation, confluent B-lines, and snowflake signs), providing a scalable framework for small-sample medical image analysis, though diagnostic performance on complex conditions remains limited by dataset scale and minority class representation.

Objective:To assess the effect of intra-aortic balloon pump (IABP) on in-hospital mortality in patients with cardiac arrest undergoing extracorporeal cardiopulmonary resuscitation (ECPR).Methods:A retrospective study was performed on 696 patients with intra-hospital cardiac arrest undergoing ECPR from Samsung Medical Center in Korea between January 2004 and December 2013. According to whether IABP was used, the patients were divided into ECPR group and ECPR+IABP group. Cox regression and propensity score matching (PSM) were used to examine the correlation between IABP usage and in-hospital mortality, and standardized mean difference ( SMD) was used to check the degree of PSM. Survival analysis of in-hospital mortality was performed by the Kaplan-Meier method, and further analyzed by the Log-Rank test. Using the propensity score as weights, multiple regression model and inverse probability weighting (IPW) model were used for sensitivity analysis. In-hospital mortality, extracorporeal membrane oxygenation (ECMO) withdrawal success rate and neurological function prognosis were compared between the two groups. Results:A total of 199 patients with cardiac arrest undergoing ECPR were included, including 120 males and 79 females, and the average age was (60.0±16.8) years. Thirty-one patients (15.6%) were treated with ECPR and IABP, and 168 patients (84.4%) only received ECPR. The total hospitalized mortality was 68.8% (137/199). The 1 : 1 nearest neighbor matching algorithm was performed with the 0.2 caliper value. The following variables were selected to generate propensity scores, including age, gender, race, marital status, insurance, admission type, service unit, heart rate, mean arterial pressure, respiratory rate, pulse oxygen saturation, white blood cell count. After the propensity score matching, 24 pairs of patients were successfully matched, with the average age of (63.0±12.8) years, including 31 males and 17 females. The in-hospital mortality was 72.6% (122/168) and 48.4% (15/31) in the ECPR group and the ECPR+IABP group [hazard ratio ( HR) = 0.48, 95% confidence interval (95% CI) was 0.28-0.82, P = 0.007]. Multiple regression model, adjusted propensity score, PSM and IPW model showed that the in-hospital mortality in the ECPR+IABP group was significantly lower compared with the ECPR group ( HR = 0.44, 0.50, 0.16 and 0.49, respectively, 95% CI were 0.24-0.79, 0.28-0.91, 0.06-0.39 and 0.31-0.77, all P < 0.05). The combined application of IABP could improve the ECMO withdrawal success rate [odds ratio ( OR) = 8.95, 95% CI was 2.72-29.38, P < 0.001] and neurological prognosis ( OR = 4.06, 95% CI was 1.33-12.40, P = 0.014) in adult cardiac arrest patients. Conclusion:In patients with cardiac arrest using ECPR, the combination of IABP was independently associated with lower in-hospital mortality, higher ECMO withdrawal success rate and better neurological prognosis.

Bivalves are species-rich mollusks with prominent protective roles in coastal ecosystems.Across these ancient lineages,colony-founding larvae anchor themselves either by byssus produc-tion or by cemented attachment.The latter mode of sessile life is strongly molded by left-right shell asymmetry during larval development of Ostreoida oysters such as Crassostrea hongkongensis.Here,we sequenced the genome of C.hongkongensis in high resolution and compared it to reference bivalve genomes to unveil genomic determinants driving cemented attachment and shell asymmetry.Importantly,loss of the homeobox gene Antennapedia(Antp)and broad expansion of lineage-specific extracellular gene families are implicated in a shift from byssal to cemented attachment in bivalves.Comparative transcriptomic analysis shows a conspicuous divergence between left-right asymmetrical C.hongkongensis and symmetrical Pinctada fucata in their expression profiles.Especially,a couple of orthologous transcription factor genes and lineage-specific shell-related gene families including that encoding tyrosinases are elevated,and may cooperatively govern asymmet-rical shell formation in Ostreoida oysters.

Objective:To establish a nomogram prediction model for the prognosis of patients with septic cardiomyopathy (SCM) based on afterload-corrected cardiac performance (ACP), in order to identify septic patients with poor outcomes and treatment.Methods:The data of patients admitted to the department of critical medicine of the Second Affiliated Hospital of Guangzhou Medical University from June 2016 to June 2019 were analyzed. All patients were monitored by pulse indication continuous cardiac output (PiCCO) monitor more than 24 hours and diagnosed as SCM with ACP less than 80%. The predictors of 30-day death risk of SCM patients were screened by univariate Cox regression analysis. Multivariate Cox regression analysis was used to establish the prediction model for 30-day death risk of SCM patients, which was displayed by the nomogram. Finally, the discrimination and calibration of the model were analyzed by receiver operator characteristic curve (ROC curve) and consistency index (C-index).Results:A total of 102 patients with SCM were included and the 30-day mortality was 60.8% (62 cases). Among 102 patients with SCM, 57 patients (55.9%) had mild impairment of cardiac function (60%≤ACP < 80%), and the 30-day mortality was 43.9% (25/57); 39 patients (38.2%) had moderate impairment of cardiac function (40%≤ACP < 60%), and the 30-day mortality was 79.5% (31/39); 6 patients (5.9%) had severe impairment of cardiac function (ACP < 40%), and the 30-day mortality was 100% (6/6). There was significantly difference in mortality among the three groups (χ 2 = 24.156, P < 0.001). The potential risk factors for 30-day death of SCM patients screened by univariate Cox regression analysis were included in multivariate Cox regression analysis. The results showed that the independent risk factors for 30-day death of SCM patients were acute physiology and chronic health evaluation Ⅱ [APACHEⅡ, risk ratio ( HR) = 1.031, 95% confidence interval (95% CI) was 1.002-1.061, P = 0.039], vasoactive inotropic score (VIS, HR = 1.003, 95% CI was 1.001-1.005, P = 0.012), continuous renal replacement therapy (CRRT; HR = 2.106, 95% CI was 1.089-4.072, P = 0.027), and ACP ( HR = 0.952, 95% CI was 0.928-0.977, P < 0.001). The nomogram model was established based on the above independent risk factors and age, and the area under the curve (AUC) was 0.865 (95% CI was 0.795-0.935), P < 0.001; C-index was 0.797 (95% CI was 0.747-0.847), P > 0.05. Conclusions:The nomogram model based on age, APACHEⅡ score, VIS score, CRRT and ACP has a certain clinical reference significance for the prediction of 30-day mortality of SCM patients. The discrimination and calibration are good, however, further verification is needed.

Objective:To investigate the relationship between 1-hour lactate (1 h Lac) and 30-day mortality in critical care patients in intensive care unit (ICU).Methods:A retrospective, observational cohort study was performed with adult critical patients (age ≥ 16 years old) having lactate records within 1 hour after ICU admission from Medical Information Mart for Intensive Care-Ⅲ database (MIMIC-Ⅲ). According to the 1 h Lac level, the patients were divided into three groups: < 2 mmol/L, 2-4 mmol/L, and > 4 mmol/L groups. The baseline characteristics were analyzed. Multivariable Logistic regression analysis was performed to assess the association between 1 h Lac and 30-day mortality. The receiver operating characteristic (ROC) curve was used to analyze the predictive value of 1 h Lac for 30-day mortality, and Kaplan-Meier survival curve was performed according to the best cut-off value. In addition, sensitivity analysis was carried out for each classification variable.Results:A total of 3 969 ICU patients were included, with 673 died in 30 days, and the total mortality was 16.95%. There were 1 664, 1 588, 717 patients in Lac < 2 mmol/L, 2-4 mmol/L and > 4 mmol/L group, respectively. There were significant differences in age, ICU duration, ICU type, heart rate, leukocyte count, hemoglobin, creatinine, sequential organ failure score (SOFA), ventilator application, vasoactive drug use and main diagnosis among the three groups. Multivariable Logistic regression analysis showed that a 1 mmol/L increment in Lac was associated with 0.24 times higher risk of 30-day mortality [odds ratio ( OR) = 1.24, 95% confidence interval (95% CI) was 1.19-1.29, P < 0.000 1]. ROC curve analysis showed that the area under ROC curve (AUC) of 1 h Lac for predicting 30-day mortality of severe patients was 0.694 (95% CI was 0.669-0.718). The cut-off value was 3.35 mmol/L with sensitivity of 0.499 and specificity of 0.779, whilst positive likelihood ratio was 2.260, and negative likelihood ratio was 0.643. According to the cut-off value of 1 h Lac, the patients were divided into high lactate group (≥ 3.35 mmol/L) and low lactate group (< 3.35 mmol/L). In the two subgroups, 30-day mortality was 31.58% (336/1 064) and 11.60% (337/2 905), respectively. The Kaplan-Meier survival curve showed that the 30-day cumulative survival rate of high lactate group was significantly lower than that of low lactate group (Log-rank test: χ 2 = 247.72, P < 0.000 1). Multiple Logistic regression analysis showed that the 30-day mortality rate of high lactate group was 2.34 times that the level of low lactate group ( OR = 2.34, 95% CI was 1.90-2.88, P < 0.000 1), after the adjustment of age, time of admission, type of ICU, hemoglobin, leukocyte count, use of vasopressor, use of ventilator and main diagnosis of patients. Stratified analysis showed that the relationship between 1 h Lac and 30-day mortality was stable. Conclusions:1 h Lac is associated with 30-day mortality in critical care patients. The risk of death was significantly increased in critically ill patients with 1 h Lac higher than 3.35 mmol/L.

AIM: To evaluate the effect of inhibiting ubiquitin-specific protease 14 (USPl4) activity on oxidative stress induced by H2O2 of H9c2 cells.METHODS: The H9c2 cells were incubated with H2O2 at 25 μmol/L for 2 h to establish the oxidative stress injury model.The cells were divided into control group, H2O2 group, IU1 group (25 μmol/L or 50 μmol/L) and IU1+ H2O2 group.The H9c2 cells activity was measured by MTS assay.The level of intracellular reactive oxygen species (ROS) and cell survival rate were analyzed by flow cytometry assay.The changes of the mitogen-activated protein kinase (MAPK) family related proteins were detected by Western blot.RESULTS: Compared with control group, the cell activity and the viability rate in H2O2 group were decreased (P<0.05), while the intracellular ROS, the protein levels of Bax/Bcl-2, P53, p-ERK1/2, p-JNK and p-P38 were increased (P<0.05).Compared with H2O2 group, the cell activity and the viability rate of the H9c2 cells in IU1+H2O2 group were increased (P<0.05), while the intracellular ROS, the protein levels of Bax/Bcl-2, P53, p-ERK1/2, p-JNK and p-P38 were decreased (P<0.05).CONCLUSION: Inhibition of USPl4 activity reduces the oxidative stress injury of the H9c2 cells.The mechanism may be related to inhibition of the MAPK signaling and down-regulation of apoptosis related proteins.

Ojective To determine the risk factors and the clinical distribution of multiple drug resistant bacteria in stroke- associated pneumonia (SAP) patients with multiple drug-resistant bacterial infections from in-tensive care unit, providing guidance for clinical treatment of SAP. Methods A retrospective study was de-signed to analyze the clinical data of the SAP patients from January 2012 to December 2015. Univariate analysis and multivariable regression analysis were taken for risk factors of MDR infections , and investigated the distribu-tion and drug resistance of MDR. Results There were 183 SAP patients, of which 131patients (71.6%) had MDR infection. There are 193 MDR strains in the 131 patients , the first 5 MDR strains were Acinetobacter bau-mannii, Pseudomonas aeruginosa, Klebsiella pneumoniae, Staphylococcus aureus and Escherichia coli. MDR in-fection is highly associated with sever unconsciousness , time of stay in ICU longer than 7 days , ventilator time longer than 7 days and late-onset pneumonia and the difference was significant (P < 0.05). Conclusion SAP patients with MDR bacterial infections are in association with the following risk factors: sever unconsciousness , time of stay in ICU longer than 7 days, ventilator time longer than 7 days and late-onset pneumonia. The rate of MDR bacterial infections was high , and these MDR strains were widely different degrees of resistance to many antibiotics.

Objective To observe the effects of Ulinastatin on pulmonary vascular endothelium permeability and respiratory function in patients with extrapulmonary acute respiratory distress syndrome(ARDS exp). Methods The data of 39 patients with ARDS exp were retrospectively analyzed.According to whether treated with Ulinastatin or not, all patients were divided into Ulinastatin group(n = 21)and control group(n = 18); The level of extravascular lung water index (EVLWI), pulmonary vascular permeability index(PVPI) and respiratory function were measured before and after the treatment. Results The mortality rate of Ulinastatin group was lower than that of control group (28.6% vs 38.9%). The time of decreasing EVLWI, PVPI and improving PaO2/FiO and respiratory function in Ulinastatin group was shorter than that of control group, and the effect was superior. Conclusion Ulinastatin could reduce EVLWI and PVPI, improve pulmonary compliance and oxygenation, and reduce mortality rate in patients with extrapulmonary acute respiratory distress.

OBJECTIVETo assess the effects of continuous blood purification (CBP) on extravascular lung water and respiratory function in patients with extrapulmonary acute respiratory distress syndrome (ARDSexp).

METHODSThe data of 31 patients with ARDSexp admitted in our department were retrospectively analyzed.Sixteen of the patients received CBP, and the other 15 patients did not (control group). The level of extravascular lung water index (EVLWI), pulmonary vascular permeability index (PVPI), and respiratory function were measured before and after CPB.

RESULTSThe mortality rate was significantly lower in CBP group than in the control group (12.5% vs 33.3%, P<0.05). The patients in CPB group showed markedly earlier and significantly greater improvements in EVLWI, PVPI, PaO2/FiO, and respiratory function than the control patients (P<0.05).

CONCLUSIONCBP can reduce EVLWI and PVPI, improve pulmonary compliance and oxygenation, and reduce mortality rate in patients with ARDSexp.