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.

AIM:To improve the success rate of experimental modeling of non-alcoholic fatty liver(NAFLD)in rats by high-fat diet through comparing three different formulations of high-fat diets in con-structing non-alcoholic fatty liver rats model,so as to provide a reliable animal model for the study of non-alcoholic fatty liver disease.METHODS:SPF-grade male SD rats were divided into four groups randomly:control group,high-fat diet group1(HFD1),high-fat diet group2(HFD2),high-fat diet group3(HFD3).Each group was given the corre-sponding feed for 8 weeks while modeling.The da-ta on general observation,body weight changes,and ingestion of the rats were recorded during the modeling period.After 8 weeks'feeding,liver ultra-sound,CT and MRI examination were performed for the rats of each group to check the status.Blood and liver samples were collected.Changes in liver function(ALT,AST),blood lipids(TC,TG,HDL-C,LDL-C),and inflammatory indexes(IL-1β,IL-6,TNF-α)were detected.The morphology of the liv-ers was observed with the naked eyes,and the liv-er index and Lee's index were calculated at the end of 8 weeks.The effects of different high-fat diets on the establishment of NAFLD model in SD rats were comprehensively evaluated by comparing the difference of above indexes among the groups.RE-SULTS:Compared with the control group,rats in the HFD1,HFD2 and HFD3 groups showed poor mental deterioration,decreased activity,severe hair loss,decreased food intake,increased body weights,and significantly increased liver index and Lee's in-dex,along with increased liver volume,blunt edge,steatosis and lipid deposition,and the trend was even more pronounced in the HFD3 group.Com-pared with the control group,the serum levels of ALT,AST,TC,TG,LDL-C,IL-1β,IL-6 and TNF-α were significantly increased,while the contents of HDL-C was significantly decreased in the HFD1,HFD2 and HFD3 group,especially in the HFD3 group.Com-pared with the control group,the B ultrasonogra-phy showed an enlarged liver with enhanced paren-chymal echo and pipe unsharpness,CT showed that the liver and spleen CT ratio decreased obvi-ously,and the MRI images showed obvious differ-ence of liver signal intensity between in/out of phase image in the HFD1,HFD2 and HFD3 group,and the most significant imaging changes was ob-served in the HFD3 group.CONCLUSION:The above three kinds of high-fat diets can establish NAFLD model in SD rats after 8 weeks'feeding,the models induced by HFD3 was better than those in-duced by the other two groups.NAFLD lesion is rel-atively serious and expected to last longer in HFD3 group,which are more suitable for investigating the underlying mechanisms of non-alcoholic fatty liver disease and development of lipid-lowering drugs.

Objective: Assessment of short-term and mid-term changes of knee cartilage in non-professional long-distance runners before and after marathon using T₂ mapping imaging. Methods: Twenty-four knee joints of 12 healthy volunteers (5 males and 7 females) who participated in the marathon were examined by 3.0 T MRI one week before the race, within 12 hours after the race and two months after the race, respectively. The age ranged from 21.0 to 37.0 years. Athletes run more than three times a week, and each time was about 30 minutes. From T₂ mapping we can obtain T₂ value of 6 cartilage subregions of knee joint. Paired t-test was used to compare the T₂ values of cartilage before and after the marathon. The comparison between the superficial and deep T₂ values of cartilage was performed by independent sample t test. Results: T₂ mapping imaging showed that the T₂ value of the superficial articular cartilage was higher than that of deep articular cartilage in pre-competition, within 12 hours after competition and 2 months follow-up, respectively (t=11.095, 10.385 and 10.102, P<0.01). The T₂ value of superficial articular cartilage decreased after the competition compared with that of pre-competition (t=2.18,P<0.05), while the T₂ value of deep articular cartilage showed no significant difference compared with that of pre-competition (t=1.832, P> 0.05). Except for the MTP-DZ, the T₂ value of the remaining cartilage subregions showed a trend of decrease within 12 hours after the competition. There were significant differences in the subregion of MFC-DZ (t=2.110,P<0.05). At the follow-up of 2 months, cartilage T₂ values of the MTP-SZ,LTP-SZ,LTP-DZ,MFC-DZ, Patella-SZ, Patella-DZ, Trochlea-SZ, Trochlea-DZ subregions showed a trend of recovery, and there was no significant difference compared with pre-competition (t=0.857,0.573, 0.146, 0.510, -0.594, -1.135, -0.812, -0.679; P>0.05). The T₂ values of the LFC-SZ and LFC-DZ were lower than those before the race, with statistical significance (t=2.378, 3.147,P<0.05). Conclusion T₂ value could reflect the change process of articular cartilage tissue composition in marathon runners from T₂ mapping imaging. Under stress, the changes of superficial T₂ value are more significant. The T₂ value of knee cartilage can gradually recover to the pre-run level 2 months after running.

Objective Assessment of short?term and mid?term changes of knee cartilage in non?professional long?distance runners before and after marathon using T2 mapping imaging. Methods Twenty?four knee joints of 12 healthy volunteers (5 males and 7 females) who participated in the marathon were examined by 3.0 T MRI one week before the race, within 12 hours after the race and two months after the race, respectively. The age ranged from 21.0 to 37.0 years. Athletes run more than three times a week, and each time was about 30 minutes. From T2 mapping we can obtain T2 value of 6 cartilage subregions of knee joint. Paired t?test was used to compare the T2 values of cartilage before and after the marathon. The comparison between the superficial and deep T2 values of cartilage was performed by independent sample t test. Results T2 mapping imaging showed that the T2 value of the superficial articular cartilage was higher than that of deep articular cartilage in pre?competition, within 12 hours after competition and 2 months follow?up, respectively (t=11.095, 10.385 and 10.102, P<0.01). The T2 value of superficial articular cartilage decreased after the competition compared with that of pre?competition (t=2.18,P<0.05), while the T2 value of deep articular cartilage showed no significant difference compared with that of pre?competition (t=1.832, P> 0.05). Except for the MTP?DZ, the T2 value of the remaining cartilage subregions showed a trend of decrease within 12 hours after the competition. There were significant differences in the subregion of MFC?DZ (t=2.110,P<0.05). At the follow?up of 2 months, cartilage T2 values of the MTP?SZ,LTP?SZ,LTP?DZ,MFC?DZ, Patella?SZ, Patella?DZ, Trochlea?SZ, Trochlea?DZ subregions showed a trend of recovery, and there was no significant difference compared with pre?competition(t=0.857,0.573, 0.146, 0.510,-0.594,-1.135,-0.812,-0.679; P>0.05). The T2 values of the LFC?SZ and LFC?DZ were lower than those before the race, with statistical significance (t=2.378, 3.147,P<0.05). Conclusion T2 value could reflect the change process of articular cartilage tissue composition in marathon runners from T2 mapping imaging. Under stress, the changes of superficial T2 value are more significant. The T2 value of knee cartilage can gradually recover to the pre?run level 2 months after running.