ObjectiveTo compare the performance of different automatic evaluation methods for detecting pattern hallucinations generated by large language models （LLMs） in traditional Chinese medicine （TCM） syndrome differentiation and to identify a strategy with higher overall discriminative performance against expert manual judgment. MethodsA standardized TCM syndrome-differentiation dataset containing 598 cases was constructed from case reports published in Chinese core journals indexed by Peking University Core Journals or Chinese Science Citation Database. Each of the 598 cases was submitted to five Chinese LLMs， with each model generating one syndrome-pattern output per case， yielding 2990 outputs in total. The 598 outputs generated by Qwen3-Next-80B-A3B-Thinking were independently annotated for syndrome-pattern hallucinations by two licensed TCM physicians with intermediate or higher professional titles. Disagreements were adjudicated by a third licensed TCM physician with a senior associate professional title and more than 10 years of clinical experience. The resulting consensus annotations served as the reference labels for training and validating the automated evaluation method. Based on these reference labels， three categories of automated evaluation methods were developed and compared following a progressive strategy from text-level semantic similarity， to syndrome-element structural consistency， and finally multi-feature fusion. Method A employed sentence-embedding models based on semantic similarity between pattern texts. Method B adopted a rule-based threshold method using weighted Jaccard coefficients of disease-location and disease-nature syndrome elements. Method C utilized a machine-learning classifier integrating syndrome-element matching scores， missing and redundant syndrome-element counts， and multiple semantic-similarity features. Five-fold cross-validation was used to evaluate the discriminative performance of each method against manual judgment. The optimal method was then applied to uniformly assess the pattern hallucination rates of five LLMs. ResultsIn method A， text2vec-large-chinese showed best overall performance， with an area under the receiver operating characteristic curve （AUROC）， area under the precision-recall curve （PR-AUC）， F1-score， sensitivity， and specificity of 0.849， 0.884， 0.794， 0.801， and 0.691， respectively. In method B， the corresponding metrics of the rule-based syndrome-element Jaccard threshold method were 0.848， 0.851， 0.817， 0.845， and 0.684， respectively； those of the LLM-based syndrome-element Jaccard method were 0.872， 0.906， 0.795， 0.764， and 0.780， respectively. In method C， the multilayer perceptron （MLP） classifier showed the best overall performance， with AUROC， PR-AUC， F1-score， sensitivity， and specificity of 0.933， 0.951， 0.871， 0.857， and 0.846， respectively. Under the unified evaluation framework combining DeepSeek-R1 syndrome-element extraction and MLP classifier， the estimated pattern hallucination rates of the five LLMs ranged from 56.0% to 77.4%. ConclusionAmong the three automatic evaluation methods， Method C integrating syndrome-element discrepancy features and text semantic features using an MLP classifier showed the highest overall discriminative performance against manual judgment， and is more suitable for evaluating syndrome hallucinations in LLM-based TCM syndrome differentiation tasks.

A VLi-net based cell nucleus segmentation method integrating convolutional neural networks(CNN)and Vision Transformer(ViT)is proposed to address the limitation that the U-Net with CNN as its backbone is only proficient in capturing local features and has a restricted receptive field.Firstly,to mitigate challenges such as high cost of data annotation and insufficient annotated data,text annotations are introduced to enhance the network's understanding of image information.Secondly,to improve the segmentation performance of VLi-net,ViT and CNN are combined to fully extract global and local features,with multi-receptive field convolution features incorporating into the ViT structure for effectively mitigating the issues of limited local information interaction and single feature representation in ViT.Finally,an interactive fusion module(ViFusion)is used to efficiently fuse the multi-level features from the CNN and ViT branches.Experimental results show that VLi-net achieves a Dice coefficient of 80.85%and a mean intersection over union(MIoU)of 66.83%on the MoNuSeg dataset,obtains a Dice coefficient of 80.53%and a MIoU of 67.54%on the DSB-2018 dataset,and has a Dice coefficient of 86.87%and a MIoU of 77.44%on the TNBC dataset.These findings confirm that VLi-net outperforms other methods across multiple experimental metrics.

A VLi-net based cell nucleus segmentation method integrating convolutional neural networks(CNN)and Vision Transformer(ViT)is proposed to address the limitation that the U-Net with CNN as its backbone is only proficient in capturing local features and has a restricted receptive field.Firstly,to mitigate challenges such as high cost of data annotation and insufficient annotated data,text annotations are introduced to enhance the network's understanding of image information.Secondly,to improve the segmentation performance of VLi-net,ViT and CNN are combined to fully extract global and local features,with multi-receptive field convolution features incorporating into the ViT structure for effectively mitigating the issues of limited local information interaction and single feature representation in ViT.Finally,an interactive fusion module(ViFusion)is used to efficiently fuse the multi-level features from the CNN and ViT branches.Experimental results show that VLi-net achieves a Dice coefficient of 80.85%and a mean intersection over union(MIoU)of 66.83%on the MoNuSeg dataset,obtains a Dice coefficient of 80.53%and a MIoU of 67.54%on the DSB-2018 dataset,and has a Dice coefficient of 86.87%and a MIoU of 77.44%on the TNBC dataset.These findings confirm that VLi-net outperforms other methods across multiple experimental metrics.

Objective To propose a classification model based on a pseudo-bag strategy and feature adjustment for whole slide imaging in pathology.Methods A pseudo-bag generator was constructed to divide a parent bag into 3 pseudo-bags for increasing the number of training bags.Then,a pseudo-bag learning method based on Nystr?m-based algorithm for approximating self-attention and a selective feature fusion method were employed to process the pseudo-bags.Specifically,the pseudo-bag learning method based on Nystr?m-based algorithm for approximating self-attention reduced computational complexity through an improved multi-head self-attention mechanism while deeply extracting instance features to obtain pseudo-bag classification predictions,thereby enhancing pseudo-bag classification accuracy;and the selective feature fusion method refined pseudo-bag features by filtering and extracting relevant instances.Finally,the model adjusted bag features by extracting confounding factors to avoid interference from irrelevant information and further improve classification accuracy.Results The proposed model was evaluated on two datasets(CAMELYON-16 and TCGA-NSCLC)and compared with 10 other methods,and the results demonstrated that the proposed model achieved the best performance.The proposed method reached an accuracy of 0.943 on the CAMELYON-16 dataset and 0.906 on the TCGA-NSCLC dataset.Conclusion The proposed model can significantly improve the accuracy of whole-slide pathological image classification by effectively mitigating the overfitting and avoiding interference from irrelevant information.

There is significant inter-individual variation of pharmacokinetics and pharmacody-namics in patients receiving tacrolimus(TAC)for an-ti-rejection therapy,which cause the rejection or toxic action.Based on results of therapeutic drug monitoring and pathophysiological index of trans-plant patients,the individualized dosing regimen can be designed and adjusted by using model in-formed precision dosing(MIPD).The patients'clini-cal outcome can be improved.In the consensus,the different methods of MIPD used for patients re-ceived TAC for anti-rejection therapy were intro-duced,which can be used for the designing and ad-justing doing regimen,predicting adverse drug reac-tion,improving medication adherence and econom-ics during therapy.

There is significant inter-individual variation of pharmacokinetics and pharmacody-namics in patients receiving tacrolimus(TAC)for an-ti-rejection therapy,which cause the rejection or toxic action.Based on results of therapeutic drug monitoring and pathophysiological index of trans-plant patients,the individualized dosing regimen can be designed and adjusted by using model in-formed precision dosing(MIPD).The patients'clini-cal outcome can be improved.In the consensus,the different methods of MIPD used for patients re-ceived TAC for anti-rejection therapy were intro-duced,which can be used for the designing and ad-justing doing regimen,predicting adverse drug reac-tion,improving medication adherence and econom-ics during therapy.

Objective To propose a classification model based on a pseudo-bag strategy and feature adjustment for whole slide imaging in pathology.Methods A pseudo-bag generator was constructed to divide a parent bag into 3 pseudo-bags for increasing the number of training bags.Then,a pseudo-bag learning method based on Nystr?m-based algorithm for approximating self-attention and a selective feature fusion method were employed to process the pseudo-bags.Specifically,the pseudo-bag learning method based on Nystr?m-based algorithm for approximating self-attention reduced computational complexity through an improved multi-head self-attention mechanism while deeply extracting instance features to obtain pseudo-bag classification predictions,thereby enhancing pseudo-bag classification accuracy;and the selective feature fusion method refined pseudo-bag features by filtering and extracting relevant instances.Finally,the model adjusted bag features by extracting confounding factors to avoid interference from irrelevant information and further improve classification accuracy.Results The proposed model was evaluated on two datasets(CAMELYON-16 and TCGA-NSCLC)and compared with 10 other methods,and the results demonstrated that the proposed model achieved the best performance.The proposed method reached an accuracy of 0.943 on the CAMELYON-16 dataset and 0.906 on the TCGA-NSCLC dataset.Conclusion The proposed model can significantly improve the accuracy of whole-slide pathological image classification by effectively mitigating the overfitting and avoiding interference from irrelevant information.

Immune evasion has made ovarian cancer notorious for its refractory features, making the development of immunotherapy highly appealing to ovarian cancer treatment. The immune-stimulating cytokine IL-12 exhibits excellent antitumor activities. However, IL-12 can induce IFN-γ release and subsequently upregulate PDL-1 expression on tumor cells. Therefore, the tumor-targeting folate-modified delivery system F-DPC is constructed for concurrent delivery of IL-12 encoding gene and small molecular PDL-1 inhibitor (iPDL-1) to reduce immune escape and boost anti-tumor immunity. The physicochemical characteristics, gene transfection efficiency of the F-DPC nanoparticles in ovarian cancer cells are analyzed. The immune-modulation effects of combination therapy on different immune cells are also studied. Results show that compared with non-folate-modified vector, folate-modified F-DPC can improve the targeting of ovarian cancer and enhance the transfection efficiency of pIL-12. The underlying anti-tumor mechanisms include the regulation of T cells proliferation and activation, NK activation, macrophage polarization and DC maturation. The F-DPC/pIL-12/iPDL-1 complexes have shown outstanding antitumor effects and low toxicity in peritoneal model of ovarian cancer in mice. Taken together, our work provides new insights into ovarian cancer immunotherapy. Novel F-DPC/pIL-12/iPDL-1 complexes are revealed to exert prominent anti-tumor effect by modulating tumor immune microenvironment and preventing immune escape and might be a promising treatment option for ovarian cancer treatment.

Objective:To describe the characteristics, etiology and patterns of outpatients and inpatients patients with moderate or severe valvular heart disease (VHD).Methods:This is a cross-sectional study. Outpatients and inpatients with moderate or severe VHD who underwent transthoracic echocardiography for first examination from 1 st January 2001 to 1 st January 2020 in Southwest Hospital, Army Medical University were enrolled. Data were collected from medical records and big data platform of Southwest Hospital. Characteristics of age and gender, etiology and types of VHD were descriptively analysed. Results:A total of 68 354 patients with moderate or severe VHD were enrolled. The age was 63 (50, 72) years. And 35 706 (52.24%) patients were female. (1) Age characteristics: There was similar age trend between male and female patients with moderate or severe VHD. The number of patients increased firstly and then decreased and reached its peak in the age group of 65-69 years old. The peak age of mitral stenosis patients was 45-49 years, which was earlier than that of whole patients with moderate or severe VHD. The median age of patients with bicuspid aortic valve was 42 years. (2) Gender characteristics: The proportion of tricuspid regurgitation, pulmonary regurgitation, mitral regurgitation, mitral stenosis and valve surgery in female patients with moderate or severe VHD were higher than those in male patients. The proportion of aortic regurgitation, aortic stenosis and bicuspid aortic valve in male patients with moderate or severe VHD were significantly higher than those in female patients (all P<0.05). (3) Etiology: The proportion of rheumatic VHD was 13.07% (8 934/68 354), which was higher than that of degenerative VHD (0.67% (458/68 354)). (4) Types of VHD: Tricuspid regurgitation made contribution to the largest proportion with 60.72% (41 503/68 354), followed by mitral regurgitation, aortic regurgitation, mitral stenosis, pulmonary regurgitation and aortic stenosis. Conclusions:There are certain regional characteristics in the prevalence of moderate or severe VHD in southwest China, suggesting different attention should be paid on the whole process of refined management of moderate or severe VHD.

Allergic rhinitis is a common allergic disease in children.Its pathogenesis is complex and it is difficult to achieve radi-cal cure or effective and stable long-term treatment goals.Chinese medicine has obvious advantages in preventing and treating allergic rhinitis in children due to its wide range of targets,long-lasting effects and few adverse reactions.This paper proposes that the onset of allergic rhinitis is mostly caused by the dysfunction of the lung,spleen and kidney,the external wind triggering the latent wind,and the combination of the two winds.A staged prevention and treatment strategy of Chinese medicine should be adopted,which includes dispersing external wind,suppressing latent wind,and promoting lung-qi and clearing nasal orifice during the attack period to treat its symptoms,and preventing external wind,calming down latent wind,and regulating and tonifying the lung,spleen,and kidney during the remission period to treat its root cause;meanwhile,attention should be paid to avoiding the adverse effects of congenital endowment factors and the induction of acquired environmental factors,strengthening the body's health to protect against the evil wind,preventing the transformation of existing diseases and the recurrence of allergic rhinitis in children at all stages.