1.Pathogenesis Reasoning Chain-of-thought Supervision for Large Language Models: Syndrome Manifestation Recognition and Multidimensional Evaluation in Spleen-stomach Disorders
Shu-Han YANG ; Yu-Xin HU ; Xin-Yu YU ; Yu-Ying TU ; Yi-Chang ZANG ; Pan-Fei LI
Progress in Biochemistry and Biophysics 2026;53(5):1240-1263
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.
2.Pathogenesis Reasoning Chain-of-thought Supervision for Large Language Models: Syndrome Manifestation Recognition and Multidimensional Evaluation in Spleen-stomach Disorders
Shu-Han YANG ; Yu-Xin HU ; Xin-Yu YU ; Yu-Ying TU ; Yi-Chang ZANG ; Pan-Fei LI
Progress in Biochemistry and Biophysics 2026;53(5):1240-1263
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.
3.Predictive model for the risk of postoperative lung infection in esophageal cancer patients: A systematic review and meta-analysis
Yuqi LIU ; Ying ZHOU ; Ziyue SONG ; Linmei FENG ; Wenjing TU ; Qian WANG
Chinese Journal of Clinical Thoracic and Cardiovascular Surgery 2026;33(08):1290-1298
Objective To systematically evaluate the risk prediction models for postoperative pulmonary infection in patients with esophageal cancer, providing an objective basis for clinical selection and optimization of models. Methods A systematic search was conducted in Chinese and English databases such as VIP, Wanfang, CNKI, CBM, PubMed, Cochrane Library, Embase, and Web of Science for studies related to the risk prediction models of postoperative pulmonary infection in patients with esophageal cancer from the inception to September 30, 2024. The PROBAST tool was used to assess the quality of prognostic model research, and the RevMan 5.4 software was used for meta-analysis of predictive factors. Results A total of 17 articles were included, containing 26 pulmonary infection risk prediction models. The area under the receiver operating characteristic curve (AUC) ranged from 0.627 to 0.942, among which 21 models had good predictive performance (AUC>0.7). Quality assessment through the PROBAST tool revealed that all 17 articles had a high risk of bias. Meta-analysis results showed that common predictive factors for postoperative pulmonary infection in esophageal cancer included smoking history (OR=1.97), smoking index ≥200 (cigarettes-years) (OR=4.38), smoking index ≥400 (cigarettes-years) (OR=2.00), age (OR=1.27), comorbid diabetes (OR=2.13), comorbid emphysema or chronic obstructive pulmonary disease (OR=1.55), low plasma albumin levels (OR=1.17), prognostic nutritional index (OR=4.45), history of related lung diseases (OR=2.10), tumor location (OR=2.32), surgical approach (OR=2.21), operation time (OR=1.73), preoperative serum procalcitonin levels (OR=3.06), anastomotic leakage (OR=3.39), reduced forced expiratory volume in the first second/forced vital capacity ratio (OR=0.86), and hoarseness (OR=2.23). Conclusion At present, the risk prediction models for postoperative pulmonary infection in esophageal cancer are still in the stage of continuous development and optimization, and their research quality needs to be further improved. Future research can refer to the predictive factors summarized in this study based on meta-analysis, combined with clinical practice, to select appropriate methods to construct and validate the risk prediction models for postoperative pulmonary infection in esophageal cancer, thus providing early targeted preventive strategies for high-risk patients.
4.The introduction on the new standards of pharmaceutical excipients in the Chinese Pharmacopoeia 2025 Edition
CHEN Lei ; CHEN Ying ; TU Jiasheng ; LIU Yanming ; ZHENG Luxia ; ZHANG Jun ; MA Shuangcheng
Drug Standards of China 2025;26(1):058-066
According to the work goals and tasks determined by edition outline of the Chinese Pharmacopoeia 2025 Edition, the Chinese Pharmacopoeia 2025 Edition has been completed. Among them, 52 new pharmaceutical excipients monographs have been added, an increase of 15.5% compared with the 2020 Edition, and the total number has reached 387. This article focuses on the general framework and the main characteristics of the standards of pharmaceutical excipients in the Chinese Pharmacopoeia 2025 Edition, which can contribute to accurately understand and utilize the standards in Chinese Pharmacopoeia.
5.The technology of fecal microbiota transplantation and its application progress
Shuo YUAN ; Yi-fan ZHANG ; Peng GAO ; Jun LEI ; Ying-yuan LU ; Peng-fei TU ; Yong JIANG
Acta Pharmaceutica Sinica 2025;60(1):82-95
Fecal microbiota transplantation (FMT) technology originated in China during the Eastern Jin Dynasty and has rapidly developed over the past two decades, becoming a primary method for studying the causal relationship between gut microbiota and the occurrence and progression of diseases. At the same time, the therapeutic effects of FMT in the field of gastrointestinal diseases have gained widespread recognition and are gradually expanding into other disease areas. The FMT procedure is relatively complex, and there is currently no standardized method; its success is influenced by various factors, including the donor, recipient, processing of the fecal material, and the method of implantation. Given the increasingly recognized relationship between gut microbiota and various diseases, FMT has become a research hotspot in both scientific studies and clinical applications, achieving a series of significant advancements. To help researchers better understand this technology, this paper will outline the development history of FMT, summarize common operational methods in research and clinical settings, review its application progress, and look forward to future development directions.
6.Effect of preoperative anxiety on consciousness and autonomic nervous activity during propofol anesthesia
Guanqi MA ; Ying HU ; Weifeng TU ; Cui LI ; Jinping GUAN ; Junlong ZHANG
Chinese Journal of Anesthesiology 2025;45(1):42-48
Objective:To evaluate the effect of preoperative anxiety on the consciousness and autonomic nervous activity during propofol anesthesia.Methods:This study was a secondary analysis of data from the clinical trial in a prospective single-arm study. One hundred and thirty patients, aged 18-65 yr, with a body mass index of 18.5-27.9 kg/m 2, of American Society of Anesthesiologists Physical Status classification I or Ⅱ, scheduled to receive propofol anesthesia, were selected from the Second People′s Hospital of Lianyungang. The six-item of the state anxiety inventory (SAI) of the State-Trait Anxiety Inventory was used to assess the anxiety of patients 1 h before surgery. The patients were divided into 2 groups according to the cut-off value of 12: obvious anxiety symptom (SAI score >12) group (group A, n=49) and no obvious anxiety symptom (SAI score ≤12) group (group B, n=81). After admission to the operating room, the patient was required to hold a 50 ml syringe filled with water. Propofol was given by target-controlled infusion (TCI) with the target plasma concentration set at 5 μg/ml. When the effect-site concentration (Ce) of propofol increased to 3.5 μg/ml (all the patients lost consciousness), the closed-loop TCI was used to maintain BIS value between 45 and 55. The patients were monitored for 20 min after stopping the pump infusion (anesthesia recovery period). The disappearance time of verbal command, disappearance time of eyelash reflex, time of syringe dropping, recovery time of verbal command, recovery time of eyelash reflex, Ce at the recovery of verbal command, Ce at the recovery of eyelash reflex, Ce within the first 5 min of the closed-loop TCI, and consumption of propofol during anesthesia were recorded. The peripheral perfusion index, low frequency power and high frequency power of heart rate variability were recorded, and the ratio of low frequency power to high frequency power was calculated. Pearson correlation analysis was used to assess the correlation between preoperative SAI score and consciousness-related indicators, simulated Ce of propofol and consumption of propofol. Results:Compared with group B, the disappearance time of verbal command, disappearance time of eyelash reflex, and time of syringe dropping were significantly prolonged, the consumption of propofol, simulated Ce at recovery of verbal command and within the first 5 min of closed-loop TCI were increased, the peripheral perfusion index was decreased at each time point before administration and at 14-20 min of anesthesia recovery, and the low-frequency power was decreased during anesthesia maintenance in group A ( P<0.05). The SAI score was positively correlated with the disappearance time of verbal command ( r=0.220, P=0.012), time of syringe dropping ( r=0.206, P=0.029), consumption of propofol ( r=0.330, P<0.001), and the simulated Ce at the recovery of verbal command ( r=0.215, P=0.015) and simulated Ce at recovery of eyelash reflex ( r=0.207, P=0.022). Conclusions:Preoperative anxiety may lead to prolonged loss of consciousness and more marked inhibition of sympathetic nerve activity during propofol anesthesia.
7.Infant-cued behavior-based feeding pattern in preterm infants:a scoping review
Ting TU ; Hongqing GUO ; Lan HU ; Ying YUAN
Chinese Journal of Nursing 2025;60(15):1914-1921
Objective To generalize and summarize the types,content elements,outcome measures,application effects,and safety monitoring of the infant-cued behavior-based feeding,so as to provide guidance for the clinical application and research in the future.Methods The literature was retrieved from PubMed,CINAHL,CNKI,and Wanfang Data-base,with a retrieval timeframe from database establishment to August 20,2024.The included literature was sum-marized and analyzed.Results A total of 28 studies were included,of which 15 were randomized controlled trials,7 were quasi-experimental studies,and 6 were observational studies.The infant-cued behavior-based feeding pattern could be categorized as ad libitum feeding,demand feeding,semi-demand feeding,and behavioral cue-based feeding.The content elements included assessment,supporting strategies,and advancement of oral feeding.The infant-cued behavior-based feeding pattern had a positive impact on feeding and developmental outcomes,immune function,rate of complications and length of hospital stay in preterm infants.The safety monitoring consisted of heart rate,blood oxygen saturation,blood sugar,and daily weight gain.Conclusion It is beneficial to implement the infant-cued behavior-based feeding pattern in preterm infants,which leads to better clinical outcomes.To support its implementation in clinical practice,we need to validate and enhance the application value of the feeding pattern in the future by concentrating on clinically individual differences,refining feeding intervention programs,improving effect evaluation and safety monitoring.
8.Traditional Chinese Medicine in Regulating Crown-like Structures for Treatment of Obese PCOS: A Review
Wenyan TU ; Liqun YE ; Muyang LI ; Yihong YIN ; Ying SHEN ; Qidi LIN
Chinese Journal of Experimental Traditional Medical Formulae 2025;31(13):296-307
Polycystic ovary syndrome (PCOS), a common reproductive endocrine disorder in women, is one of the leading causes of ovulatory infertility in women of reproductive age. Due to its heterogeneous etiology, complex symptoms, and challenging treatment, PCOS has become a focal point of research in gynecological and reproductive medicine globally. The pathogenesis of PCOS is complex and may involve regulatory mechanisms such as inflammatory responses, oxidative stress, and cellular autophagy. Crown-like structures (CLSs) refer to pro-inflammatory microenvironments formed by macrophages engulfing adipocytes. The inflammatory disorders induced by CLSs are one of the key factors contributing to the development of PCOS and its complications. Current studies have indicated that the obese status in PCOS accelerates the formation of CLSs, and the density of CLSs can predict the progression of metabolic disorders and influence the outcomes of various metabolic diseases. Traditional Chinese Medicine (TCM) offers the unique advantages of a holistic view, four diagnostic methods, and syndrome differentiation and treatment to ameliorate the symptoms and signs of PCOS through multiple levels, pathways, and targets. Although studies on the mechanisms of metabolic diseases and CLS formation have been reported in China and abroad, there is still a lack of literature on the correlation between CLSs and PCOS, as well as reviews on TCM interventions targeting CLSs for treating this disease. Therefore, this paper summarized the correlation between obese PCOS and CLSs and reviewed recent studies on TCM interventions based on CLS formation (adipose tissue-macrophage inflammatory crosstalk) in the treatment of obese PCOS, aiming to provide new research perspectives for the prevention and treatment of PCOS using TCM.
9.Traditional Chinese Medicine in Regulating Crown-like Structures for Treatment of Obese PCOS: A Review
Wenyan TU ; Liqun YE ; Muyang LI ; Yihong YIN ; Ying SHEN ; Qidi LIN
Chinese Journal of Experimental Traditional Medical Formulae 2025;31(13):296-307
Polycystic ovary syndrome (PCOS), a common reproductive endocrine disorder in women, is one of the leading causes of ovulatory infertility in women of reproductive age. Due to its heterogeneous etiology, complex symptoms, and challenging treatment, PCOS has become a focal point of research in gynecological and reproductive medicine globally. The pathogenesis of PCOS is complex and may involve regulatory mechanisms such as inflammatory responses, oxidative stress, and cellular autophagy. Crown-like structures (CLSs) refer to pro-inflammatory microenvironments formed by macrophages engulfing adipocytes. The inflammatory disorders induced by CLSs are one of the key factors contributing to the development of PCOS and its complications. Current studies have indicated that the obese status in PCOS accelerates the formation of CLSs, and the density of CLSs can predict the progression of metabolic disorders and influence the outcomes of various metabolic diseases. Traditional Chinese Medicine (TCM) offers the unique advantages of a holistic view, four diagnostic methods, and syndrome differentiation and treatment to ameliorate the symptoms and signs of PCOS through multiple levels, pathways, and targets. Although studies on the mechanisms of metabolic diseases and CLS formation have been reported in China and abroad, there is still a lack of literature on the correlation between CLSs and PCOS, as well as reviews on TCM interventions targeting CLSs for treating this disease. Therefore, this paper summarized the correlation between obese PCOS and CLSs and reviewed recent studies on TCM interventions based on CLS formation (adipose tissue-macrophage inflammatory crosstalk) in the treatment of obese PCOS, aiming to provide new research perspectives for the prevention and treatment of PCOS using TCM.
10.Construction and empirical research of the nursing quality evaluation index system for interventional operating rooms under the context of medical alliance system
Wen ZHONG ; Yi LI ; Yanhui WANG ; Liang ZHOU ; Ying TU ; Guyu TAN
Modern Hospital 2025;25(6):873-876
Objective Based on the interventional operating room,the nursing quality evaluation index system was con-structed under the background of medical union and empirical analysis was conducted,so as to provide support and basis for accu-rate and efficient evaluation of nursing quality in interventional operating room.Methods From July to September 2024,based on evidence-based methods,the relevant literature was retrieved and evaluated,and the structure-process-outcome theory was taken as the basic framework,and the relevant nursing quality indicators were preliminarily drawn up in interventional operating rooms.In October 2024,a total of 28 experts were selected to conduct expert letter consultation(2 rounds)with Delphi method,from which the indicators were revised and finally clarified,and empirical analysis was carried out.Results There were two rounds of expert consultation in this paper.The questionnaire recovery rate of the first round was 85.71%(24/28).The Cs,Ca and Cr of the first round were 0.895,0.942 and 0.919,respectively.The recovery rate of the second round questionnaire,Cs,Ca and Cr of the experts were 100.00%(24/24),0.891,0.941 and 0.916,respectively,and Kendall's W value was within 0.088-0.301(P<0.05).The final nursing quality sensitive index system consists of 1 structure,5 process and 2 outcome inde-xes,respectively.Fifty patients undergoing interventional surgery were included before and after intervention.Empirical analysis showed that the incidence of adverse events after intervention(4.00%)was lower than that before intervention(16.00%)(P<0.05).Conclusion The sensitive index of nursing quality based on the background of medical union is practical and scientific,and can provide practical guidance for the evaluation,continuous improvement and optimization of nursing quality in interventional operating room,and reduce the incidence of adverse events.

Result Analysis
Print
Save
E-mail