ObjectiveThe essence of syndrome manifestation recognition in traditional Chinese medicine (TCM) is to infer the body’s latent pathogenesis state from clinical observational information, rather than to perform simple label matching. However, previous studies have largely modeled this task as syndrome pattern classification within a fixed label space, which does not adequately reflect the cognition process of TCM syndrome differentiation centered on pathogenesis reasoning, and is also insufficient to capture the openness, semantic variability, and cross-disease reusability of syndrome manifestation expression. This study aimed to investigate whether introducing pathogenesis reasoning chain-of-thought (PR-CoT) supervision into large language models (LLMs) could improve the quality and cognitive consistency of syndrome manifestation recognition and support cross-disease transfer. MethodsSyndrome manifestation recognition was formulated as a conditional generation task under the framework of clinical observational information (X)→pathogenesis structure (Z)→syndrome pattern output (Y), where Z serves as an explicit intermediate structural variable linking the clinical evidence and syndrome judgment. Within this framework, a PR-CoT-supervised dataset for syndrome manifestation recognition was constructed based on medical case records of spleen-stomach disorders. After preprocessing, information extraction, manual proofreading, and data cleaning, the dataset comprised 4 800 training cases, 400 development cases, and 400 test cases. Each sample was annotated with a structured PR-CoT consisting of three progressive levels: clinical information summarization, comprehensive pathogenesis analysis, and syndrome pattern output. Supervised fine-tuning was conducted on open-source LLMs, with an end-to-end model serving as the baseline. Qwen3-32B was used as the primary experimental model, and Qwen3-14B as the scale comparison model. A progressive multidimensional evaluation framework was further established, comprising a structural parsing level, a semantic similarity level, and an expert blind review level. At the structural parsing level, syndrome pattern expressions were decomposed into structural elements and evaluated using Precision, Recall, F1 score, and Jaccard similarity. At the semantic similarity level, independent LLMs scored the theoretical proximity between predicted and reference syndrome patterns. At the expert blind review level, three TCM experts independently evaluated model outputs on two dimensions: syndrome differentiation consistency and terminology standardization of syndrome patterns. In addition, zero-shot cross-disease transfer evaluation was conducted on gynecological and heart-system disorder test sets. ResultsAt the structural parsing level, PR-CoT supervision did not lead to a stable improvement in the element-wise overlap of syndrome pattern structural components. Compared with the corresponding baselines, neither Qwen3-32B nor Qwen3-14B showed consistent advantages in structural matching metrics after the introduction of PR-CoT supervision. In contrast, at the semantic similarity level, PR-CoT supervision produced stable positive gains across different model scales and evaluation systems. The average semantic score of Qwen3-32B increased from 6.425 8 in the baseline model to 6.585 0 after PR-CoT supervision, and that of Qwen3-14B increased from 5.870 0 to 5.964 2. At the expert blind review level, the overall score of Qwen3-32B (PR-CoT) was 7.026 0±0.107 7, higher than 6.416 3±0.288 9 for its baseline. In zero-shot cross-disease testing, the PR-CoT model still showed advantages in semantic evaluation and expert evaluation on both gynecological and heart-system disorder test sets, indicating a certain degree of transferability. ConclusionThe benefits of PR-CoT supervision are mainly reflected in TCM semantic consistency and clinical plausibility, rather than in improved hard matching of structural elements. These findings support understanding syndrome manifestation recognition as a process of generating and expressing latent pathogenesis structures, rather than as a classification task within a traditional fixed label space. By introducing pathogenesis reasoning as an explicit intermediate structure into the modeling process and combining it with a progressive multidimensional evaluation framework, this study provides a methodological pathway for intelligent TCM syndrome differentiation that integrates theoretical alignment, interpretability, and multi-level evaluation.

ObjectiveThe essence of syndrome manifestation recognition in traditional Chinese medicine (TCM) is to infer the body’s latent pathogenesis state from clinical observational information, rather than to perform simple label matching. However, previous studies have largely modeled this task as syndrome pattern classification within a fixed label space, which does not adequately reflect the cognition process of TCM syndrome differentiation centered on pathogenesis reasoning, and is also insufficient to capture the openness, semantic variability, and cross-disease reusability of syndrome manifestation expression. This study aimed to investigate whether introducing pathogenesis reasoning chain-of-thought (PR-CoT) supervision into large language models (LLMs) could improve the quality and cognitive consistency of syndrome manifestation recognition and support cross-disease transfer. MethodsSyndrome manifestation recognition was formulated as a conditional generation task under the framework of clinical observational information (X)→pathogenesis structure (Z)→syndrome pattern output (Y), where Z serves as an explicit intermediate structural variable linking the clinical evidence and syndrome judgment. Within this framework, a PR-CoT-supervised dataset for syndrome manifestation recognition was constructed based on medical case records of spleen-stomach disorders. After preprocessing, information extraction, manual proofreading, and data cleaning, the dataset comprised 4 800 training cases, 400 development cases, and 400 test cases. Each sample was annotated with a structured PR-CoT consisting of three progressive levels: clinical information summarization, comprehensive pathogenesis analysis, and syndrome pattern output. Supervised fine-tuning was conducted on open-source LLMs, with an end-to-end model serving as the baseline. Qwen3-32B was used as the primary experimental model, and Qwen3-14B as the scale comparison model. A progressive multidimensional evaluation framework was further established, comprising a structural parsing level, a semantic similarity level, and an expert blind review level. At the structural parsing level, syndrome pattern expressions were decomposed into structural elements and evaluated using Precision, Recall, F1 score, and Jaccard similarity. At the semantic similarity level, independent LLMs scored the theoretical proximity between predicted and reference syndrome patterns. At the expert blind review level, three TCM experts independently evaluated model outputs on two dimensions: syndrome differentiation consistency and terminology standardization of syndrome patterns. In addition, zero-shot cross-disease transfer evaluation was conducted on gynecological and heart-system disorder test sets. ResultsAt the structural parsing level, PR-CoT supervision did not lead to a stable improvement in the element-wise overlap of syndrome pattern structural components. Compared with the corresponding baselines, neither Qwen3-32B nor Qwen3-14B showed consistent advantages in structural matching metrics after the introduction of PR-CoT supervision. In contrast, at the semantic similarity level, PR-CoT supervision produced stable positive gains across different model scales and evaluation systems. The average semantic score of Qwen3-32B increased from 6.425 8 in the baseline model to 6.585 0 after PR-CoT supervision, and that of Qwen3-14B increased from 5.870 0 to 5.964 2. At the expert blind review level, the overall score of Qwen3-32B (PR-CoT) was 7.026 0±0.107 7, higher than 6.416 3±0.288 9 for its baseline. In zero-shot cross-disease testing, the PR-CoT model still showed advantages in semantic evaluation and expert evaluation on both gynecological and heart-system disorder test sets, indicating a certain degree of transferability. ConclusionThe benefits of PR-CoT supervision are mainly reflected in TCM semantic consistency and clinical plausibility, rather than in improved hard matching of structural elements. These findings support understanding syndrome manifestation recognition as a process of generating and expressing latent pathogenesis structures, rather than as a classification task within a traditional fixed label space. By introducing pathogenesis reasoning as an explicit intermediate structure into the modeling process and combining it with a progressive multidimensional evaluation framework, this study provides a methodological pathway for intelligent TCM syndrome differentiation that integrates theoretical alignment, interpretability, and multi-level evaluation.

Objective To evaluate the cost–utility of lecanemab and donanemab in the treatment of early Alzheimer’s disease （AD） in China. Methods A Markov model was developed from a societal perspective to compare standard of care （SOC）, lecanemab plus SOC, and donanemab plus SOC in terms of costs, quality-adjusted life years （QALY）, and incremental cost–utility ratios （ICUR）. Deterministic and probabilistic sensitivity analyses were conducted to assess parameter uncertainty. Results Both intervention strategies increased QALY compared with SOC but were associated with substantially higher costs, and neither was cost-effective under the current willingness-to-pay threshold. Conclusion Lecanemab and donanemab may be more appropriately considered within a multi-tiered healthcare reimbursement framework in China rather than relying solely on basic medical insurance.

OBJECTIVE To investigate the mechanism of Shenling baizhu san （SLBZS） in ameliorating ulcerative colitis （UC）. METHODS Mice were randomly divided into normal group， model group， and low-， medium-， high-dose SLBZS groups （3， 6， 12 g/kg）， with 10 mice in each group. Except for the normal group， mice in other groups were given free access to 3% dextran sulfate sodium solution for 7 days to establish UC models. Simultaneously， mice in each group were intragastrically administered corresponding medicinal liquid or double-distilled water once a day for consecutive 7 days. After the last administration， the histopathological morphology of colon tissues was observed； the serum levels of inflammatory factors including tumor necrosis factor-α （TNF-α）， interleukin-1β （IL-1β） and IL-18 were detected； the total bile acid contents in liver， gallbladder and colon tissues were determined； the expression level of NOD-like receptor family pyrin domain-containing protein 3 （NLRP3） in colon tissues was tested. The expression levels of flora metabolism-related proteins [G protein-coupled receptor 41（GPR41），GPR43，farnesoid X receptor （FXR）， G protein-coupled bile acid receptor 1 （TGR5）， flavin-containing monooxygenase 3 （FMO3）， cytochrome P450 family 27 subfamily A member 1 （CYP27A1）， CYP7A1] in liver or colon tissues， as well as the proteins related to Toll-like receptor 4 （TLR4）/nuclear factor-κB （NF-κB）/NLRP3 inflammatory axis [TLR4， NF-κB， phosphorylated NF-κB （p-NF-κB）， myeloid differentiation factor 88 （MyD88）， NLRP3， apoptosis-associated speck-like protein containing a CARD （ASC）， caspase-1] in colon tissues were measured. RESULTS Compared with the model group， the colon tissue injury and inflammatory cell infiltration were alleviated in medium- and high-dose SLBZS groups. The serum levels of TNF-α， IL-1β and IL-18， total bile acid contents in liver and colon tissues， as well as the protein expression levels of FMO3 and CYP7A1 in liver tissues， TLR4， p-NF-κB， NF-κB， MyD88， NLRP3， ASC and caspase-1 in colon tissues were significantly decreased （P＜0.05）. Meanwhile， the total bile acid content in gallbladder， the protein expression levels of GPR41 and GPR43 in colon tissues， along with CYP27A1， FXR and TGR5 in liver tissues were significantly increased （P＜0.05）. CONCLUSIONS SLBZS may relieve colonic inflammatory injury and ameliorate UC by regulating the expression of flora metabolism-related proteins and inhibiting the abnormal activation of TLR4/NF-κB/NLRP3 inflammatory axis.

Objective To evaluate the cost-utility of risdiplam in the treatment of pediatric spinal muscular atrophy （SMA） and analyze changes in patients’ out-of-pocket economic burden under the medical insurance scenarios. Methods A Markov model was developed from the healthcare system perspective to compare the long-term costs and quality-adjusted life years （QALYs） of risdiplam treatment versus conventional therapy. The incremental cost-utility ratio （ICUR） was calculated. One-way sensitivity analysis and probabilistic sensitivity analysis were conducted to assess the robustness of the model. On this basis, health insurance scenarios were simulated to evaluate patients’ individual economic burden. Results Base-case analysis showed that risdiplam treatment generated additional health benefits but substantially increased medical costs. The resulting ICUR did not demonstrate a clear cost-utility advantage under the current willingness-to-pay threshold. Sensitivity analyses indicated that the model results were relatively robust, with drug costs and key utility parameters being the main drivers of economic outcomes. Health insurance scenario analysis demonstrated that increasing reimbursement levels and reducing patient copayment rates could significantly alleviate patients’ financial burden. Conclusion Risdiplam for pediatric SMA faces economic challenges under the current health insurance payment framework in China. However, optimizing health insurance reimbursement policies and developing multi-tiered healthcare security mechanisms may improve its affordability and accessibility.

OBJECTIVE To explore the clinical characteristics of the nontuberculous Mycobacteria pulmonary dis-ease(NTMPD)patients with previous pulmonary tuberculosis(PPTB)and analyze the clinical difference from the recurrence of pulmonary tuberculosis.METHODS By means of retrospective survey,the patients who were diag-nosed with NTMPD and recurrent pulmonary tuberculosis in Wuhan Pulmonary Hospital from Mar.2021 to Oct.2023 were recruited as the research subjects,a total of 395 patients with NTMPD were enrolled in the study and were divided into the PPTB-NTMPD group with 92 cases and the NPPTB-NTMPD group with 303 cases according to the history of PPTB.The baseline data,clinical symptoms,imaging findings,underlying diseases,pulmonary diseases,and species of nontuberculous Mycobacteria(NTM)were observed and compared.Totally 92 patients with recurrent pulmonary tuberculosis were randomly screened and assigned as the recurrent pulmonary tuberculo-sis group in a 1:1 ratio by matching the PPTB-NTMPD group with the gender and age.The major clinical charac-teristics were compared between the two groups.The 92 patients with PPTB-NTMPD were divided into the 1-10 years group with 40 cases,the 10-30 years group with 37 cases,and the more than 30 years group with 15 cases according to the interval between the initial diagnosis of pulmonary tuberculosis and the diagnosis of NTMPD.The major clinical characteristics were compared among the groups.RESULTS The age was(64.21±10.71)years old in the PPTB-NTMPD group,(60.26±11.83)years old in the NPPTB-NTMPD group(t=3.020,P=0.003).The proportion of patients with body mass index less than 18.5 kg/m2 was 59.78％in the PPTB-NTMPD group,41.25％in the NPPTB-NTMPD group(x2=6.155,P=0.013);the proportion of patients with cough was 77.17％in the PPTB-NTMPD group,65.68％in the NPPTB-NTMPD group(x2=4.313,P=0.038);the inci-dence of cavitary shadow was 50.00％in the PPTB-NTMPD group,35.31％in the NPPTB-NTMPD group(x2=6.414,P=0.011);the incidence of emphysema and pulmonary bullae was 29.35％in the PPTB-NTMPD group,12.87％in the NPPTB-NTMPD group(x2=13.766,P＜0.001);the incidence of chronic obstructive pulmonary disease(COPD)was 22.83％in the PPTB-NTMPD group,14.19％in the NPPTB-NTMPD group(x2=3.875,P=0.049);the incidence of damaged lung was 9.78％in the PPTB-NTMPD group,2.97％in the NPPTB-NT-MPD group(x2=7.530,P=0.014);there were significant differences.Mycobacterium intracellulare and Myco-bacterium abscessus were the predominant species of NTM in both the PPTB-NTMPD group and the NPPTB-NT-MPD group,there was no significant difference in the distribution of NTM species between the two groups of pa-tients.The incidence of patch shadow of the PPTB-NTMPD group was lower than that of the recurrent pulmonary tuberculosis group(P＜0.05),the incidence of bronchiectatic shadow of the PPTB-NTMPD group was higher than that of the recurrent pulmonary tuberculosis group(P＜0.05).There were significant differences in the age,incidence of pleural thickening and incidence of COPD among the patients with different time intervals between ini-tial diagnosis of pulmonary tuberculosis and the diagnosis of NTMPD in the PPTB-NTMPD group(P＜0.05).CONCLUSIONS The previous pulmonary tuberculosis mainly affect the body mass index less than 18.5 kg/m2 and the post-tuberculosis pulmonary diseases such as cough,pulmonary cavity,emphysema,pulmonary bullae,COPD and damaged lung of the NTMPD patients.The NTMPD patients with previous pulmonary tuberculosis are more likely to have bronchiectasia than the patients with recurrent tuberculosis.It is necessary for the clinicians to attach great importance.

Objective To evaluate the subchronic systemic toxicity of disposable plasma virus-inactivated blood transfusion sets using hydroxyethyl starch(HES)130/0.4 sodium chloride injection as an extraction medium.Methods Firstly,40 Sprague Dawley(SD)rats including 20 male and 20 female ones were seleted and randomly enrolled into a sample group and a control group by sex,with 20 ones in each group.Secondly,instead of plasma HES 130/0.4 sodium chloride injection was used to leach disposable plasma virus-inactivated blood transfusion sets to prepare the test solution by simulating clinical application such as lighting,adsorption and filtration and storage.Finally,the test solution and HES 130/0.4 sodium chloride injection were injected into the tail vein of the SD rats at a dose of 20 mL/kg for 28 d in the sample group and in the control group respectively,and the subchronic systemic toxicity of disposable plasma virus-inactivated blood transfusion sets and the feasibility of using HES 130/0.4 sodium chloride injection as the extraction medium to assess their subchronic systemic toxicity were evaluated with clinical observation,body mass monitoring,clinical pathology examination,gross necropsy and histopathology examination.Results The sample group and control group had no significant differences in mortality rates,clinical observation results,body mass,gross necropsy results,hematological and coagulation examination results and organ weight(all P>0.05);blood biochemical examinations showed the male rats in the sample group had the cholesterol(CHO)values higher while the creatinine(CR)values lower than those in the control group,with the differences being statistically significant(both P<0.05)and the two indexes within the range of the laboratory's historical reference data,and other blood biochemical indexes were not significantly different(all P>0.05);the sample group had the spleen weight-to-body mass ratios of the female rates lower significantly than those in the control group(P<0.05),and the ratios of other organ weight to body mass had significant differences(all P>0.05);histopathology examination showed slight pathological changes in liver,spleen and kidney of female rats and in spleen and kidney of male rats in the sample group,and the female and male rats in the control group had similar pathological changes found in the sample group,which might be caused by HES metabolites.Conclusion Disposable plasma virus-inactivated blood transfusion sets prove to have no significant subchronic systemic toxicity,and its feasible to use HES 130/0.4 sodium chloride injection as the extraction medium to evaluate the subchronic systemic toxicity of disposable plasma virus-inactivated blood transfusion sets.[Chinese Medical Equipment Journal,2025,46(10):29-35]

The development of artificial intelligence(AI)technology has provided new avenues for the public to access health information.While both domestic and international research and practical applications of health conversational AI have emerged,there remains a lack of consensus on the key issues surrounding their development.Based on a comprehensive review of domestic and international literature and multiple rounds of expert consultations,the study focuses on six issues:the definition,characteristics,application scenarios,evaluation elements,evaluation methods,as well as prospects and challenges of AI-driven health consultation systems,and develops the practice guideline for health conversational artificial intelligence.The findings aim to provide scientific reference for promoting the innovative application of AI technology in the healthcare field.

Decabromodiphenyl ether(BDE-209)and decabromodiphenyl ethane(DBDPE)are widely used brominated flame retardants,which have been detected in the atmosphere,water,soil,and various organisms.In this study,a method based on high-performance liquid chromatography-inductively coupled plasma-mass spectrometry(HPLC-ICP-MS)was developed for determination of BDE-209 and DBDPE in sediment.Firstly,the target compounds in the sediments were extracted by accelerated solvent extraction(ASE),and the extraction solvent was hexane/dichloromethane(1∶1,V/V).The extract was concentrated by rotary evaporation and purified by a composite silica gel column(6 g neutral silica gel,8 g acidic silica gel,and 4 g anhydrous sodium sulfate),concentrated by nitrogen blowing,and then re-dissolved with 1 mL of toluene for instrumental determination.The chromatographic separation was carried out on a TC-C18(2)column(250 mm×4.6 mm)with isocratic elution using methanol-isopropanol-water(89∶6∶5,V/V)as the mobile phase,and the samples were separated within 20 min.Further,the Br element was quantified by ICP-MS to realize the detection of the target.The results showed that the method established in this study exhibited good linearity(R2>0.999)in the range of 100-10000 ng/mL,and the limits of quantification(LOQs)of the method were 2.0 ng/g for BDE-209 and 10.0 ng/g for DBDPE,with the relative standard deviations(RSDs,n=3)lower than 10%,and the recoveries were in the acceptable range(80.9%-120.7%).The matrix effect was effectively controlled within 10%.In addition,by analyzing the actual sediment samples from Guangxi,a background point,and Taizhou,Zhejiang,a typical contaminated area,it was found that neither BDE-209 nor DBDPE was detected in the sediment from Guangxi,while the concentrations of BDE-209 and DBDPE in the sediment from Zhejiang ranged from 1591.8 to 3362.9 ng/g,which further demonstrated the applicability and reliability of the method for analyzing actual environmental samples.This study provided a strong technical support for the accurate detection of POPs in the environment.

Objective:Analyze the imaging and clinical data of cases undergoing thoracoscopic subsegmental resection for non-small cell lung cancer(NSCLC) with a diameter≤2 cm, and explore the clinical outcomes of subsegmental resection.Methods:A retrospective analysis was conducted on the clinical data of 58 patients who underwent thoracoscopic subsegmentectomy in China-Japan Friendship Hospital from January 2020 to July 2024. Three-dimensional reconstruction technology was used for surgical planning before the operation, and thoracoscopic subsegmentectomy was performed, including single lung subsegmentectomy(Group 1), multiple lung subsegmentectomy(Group 2), and combined segmentectomy and subsegmentectomy(Group 3).Results:All patients successfully completed the surgery, with 23 cases of single lung subsegmentectomy, 6 cases of multiple lung subsegmentectomy, and 29 cases of combined segmentectomy and subsegmentectomy. The median intraoperative blood loss was 30.0(20.0, 30.0)ml, the average operation time was(2.03±0.68) h, the average pathological size of the nodules was(10.53±4.45) mm, and the average postoperative tube retention was(2.55±0.92) days. There were 6 cases of postoperative complications, including pulmonary air leakage in 2 cases, cerebral embolism in 1 case, pulmonary embolism in 1 case, pulmonary infection in 1 case, and atrial fibrillation in 1 case. All patients had negative surgical margins in the postoperative pathology. Group 1 had less average intraoperative blood loss than Group 2, with statistically significant differences( P=0.027). Surgical procedures for the upper lobe of the lung mainly involve the resection of combined segments and subsegments, while those for the lower lobe primarily consist of single segmentectomy. Conclusion:Subsegmentectomy is an effective surgical approach when the nodule is small and a clear margin can be ensured, allowing for better preservation of remaining lung tissue. Bleeding during multiple subsegmentectomies is greater than that in single subsegmentectomy and combined segmentectomy with subsegmentectomy, which may be related to the more complex vascular variations in multiple subsegmentectomies.