Objective To investigate the sputum metabolic profiles of patients with occupational coal workers' pneumoconiosis (CWP) by an untargeted metabolomics method, and to identify relevant differential metabolic pathways and potential biomarkers. Methods A total of 12 male patients with stage Ⅰ CWP were selected as the CWP group, and 16 healthy male individuals were selected as the control group, using a judgmental sampling method. Sputum metabolites of individuals in both groups were detected to perform non-targeted metabolomic analysis using the ultra-performance liquid chromatography coupled with quadrupole time-of-flight mass spectrometry. Differential metabolites (DMs) and their pathways were screened using principal component analysis, partial least squares discriminant analysis, and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis. Potential biomarkers were analyzed and identified via the receiver operating characteristic curve (ROC). Results There were apparent metabolic alterations observed in sputum of CWP patients compared with healthy controls. In the positive ion mode, a total of 42 DMs were identified in sputum from CWP patients, including 19 downregulated and 23 upregulated metabolites. In the negative ion mode, a total of 25 DMs were identified in sputum from CWP patients, including 16 downregulated and 9 upregulated metabolites. KEGG enrichment analysis of sputum from CWP patients showed that seven DMs pathways were enriched in ABC transporters, histidine metabolism, phenylalanine metabolism, arachidonic acid metabolism, linoleic acid metabolism, purine metabolism, and oxidative phosphorylation, involving 26 DMs. ROC analysis indicated that 16(R)-hydroxyarachidonic acid, pyrophosphate, and 2-hydroxyphenylacetate of these 26 DMs may serve as potential biomarkers for CWP. Conclusion Sputum metabolomic profiles were altered in CWP patients compared with healthy controls. The potential biomarkers of CWP prevention and treatment are 16(R)-hydroxyarachidonic acid, pyrophosphate, and 2-hydroxyphenylacetate.

With the rapid advancement of artificial intelligence technology, chatbots have shown great potential in the healthcare sector. From personalized health advice to chronic disease management and psychological support, chatbots have demonstrated significant advantages in improving the efficiency and quality of healthcare services. As the scope of their applications expands, the relationship between technological complexity and practical application scenarios has become increasingly intertwined, necessitating a more comprehensive evaluation of both aspects. This paper, from the perspective of he althcare applications, systematically reviews the technological pathways and development of chatbots in the medical field, providing an in-depth analysis of their performance across various medical scenarios. It thoroughly examines the advantages and limitations of chatbots, aiming to offer theoretical support for future research and propose feasible recommendations for the broader adoption of chatbot technologies in healthcare.

Objective: To investigate the status of drug resistance before anti-retroviral therapy among newly dignosed HIV/AIDS patients aged ≥50 years in Yangzhou City, Jiangsu Province, so as to provide the evidence for improving the anti-retroviral therapy effect of AIDS. Methods: HIV/AIDS patients aged ≥50 years who were newly dignosed in Yangzhou City from 2021 to 2024 and did not receive anti-retroviral therapy were selected. Basic information were collected through the Chinese Disease Prevention and Control Information System. Blood samples were collected to determine CD4+T lymphocyte (CD4 cell) counts and HIV-1 viral load. Following nucleic acid extraction, the pol gene region was amplified using reverse transcription and nested PCR, and subsequently subjected to Sanger sequencing. The resulting sequences were uploaded to the Stanford University HIV Drug Resistance Database to analyze drug resistance to non-nucleoside reverse transcriptase inhibitors (NNRTIs), protease inhibitors (PIs), and nucleoside reverse transcriptase inhibitors (NRTIs). Results: Totally 404 blood samples from HIV/AIDS patients were collected, with successful sequencing of the pol gene region in 341 cases. Among them, 253 (74.19%) were males and 88 (25.81%) were females, with a mean age of (62.48±7.60) years. A total of 152 cases (44.57%) had CD4 cell counts below 200 cells/μL, and 296 cases (86.80%) had HIV-1 viral loads exceeding 5 000 copies/mL. A total of 87 cases exhibited drug resistance-associated mutations, corresponding to a mutation rate of 25.51%. The predominant mutation site was V179, with a mutation rate of 17.01%. A total of 29 cases exhibited resistance to at least one drug, resulting in a resistance rate of 8.50%. The resistance rates to NNRTIs, PIs, and NRTIs were 5.57%, 2.93%, and 1.17%, respectively. The HIV/AIDS patients exhibited varying degrees of resistance to 13 anti-retroviral drugs, with low- or intermediate-level drug resistance being predominant. High-level drug resistance cases were observed against NNRTIs such as nevirapine and efavirenz. Conclusions The drug resistance rate before anti-retroviral therapy among newly dignosed HIV/AIDS patients aged ≥50 years in Yangzhou City was at a moderate level. The predominant resistance mutation was observed at V179 site, with NNRTIs resistance being most prevalent, primarily demonstrating low- or intermediate-level drug resistance.

OBJECTIVE: To analyze the rules of moxibustion for low back pain by ZHOU Meisheng by using data mining and knowledge graph technology. METHODS: Taking the medical cases of moxibustion for low back pain from ZHOU Meisheng's legacy manuscripts and existing works as the research objects, information on disease types, symptoms, tongue manifestations, pulse conditions, syndrome patterns, moxibustion methods and acupoints were collected. Frequency statistics and community analysis were conducted by the ancient and modern medical record cloud platform V 2.3.7, cluster analysis of high-frequency acupoints was performed by SPSS26.0, association rule analysis of high-frequency acupoints was performed by SPSS Modeler 18.0, and the generated linked data were imported into Cytoscape 3.9.1 for complex network analysis. Knowledge graph of moxibustion for low back pain by ZHOU Meisheng was constructed based on the results of data mining. The data storage and display of knowledge graph were realized through the Neo4j 3.5.25 graph database, and the Cypher query language was used for knowledge graph retrieval and discovery. RESULTS: A total of 219 medical cases were collected, involving 14 disease types, 85 related clinical symptoms, 5 related TCM syndrome types, and 6 moxibustion methods. The acupoints were mostly attributed to the governor vessel, the bladder meridian of foot-taiyang, non-meridian and non-acupoint areas. The core prescription of acupoints derived from complex network analysis included tender points, Shenshu (BL23), Jinsuo (GV8), Yinjiao (CV7), Yaoyangguan (GV3), Yanglingquan (GB34), which were largely coincides with high-frequency acupoints. Cluster analysis obtained 4 cluster combinations. Community analysis yielded 6 communities, each corresponding to different acupoints.The constructed knowledge graph contained 187 nodes and 696 relationships, by retrieving clinical elements related to low back pain, the disease-moxibustion association graph, disease-acupoint association graph, accompanying symptom-acupoint association graph and syndrome type-matching point association graph were obtained. CONCLUSION When treating low back pain with moxibustion, ZHOU Meisheng adopts the principle of promoting circulation, distinguishing diseases to determine the treatment, selecting acupoints according to the diseases, and matching points according to the symptoms.And taking tender points, Shenshu (BL23), Jinsuo (GV8), Yinjiao (CV7), Yaoyangguan (GV3), Yanglingquan (GB34) as core acupoints, combined with tenderness point selection, acupoint selection based on meridian and zangfu syndrome differentiation, "sunshine area" acupoint selection, yin-yang acupoint matching. Additionally, he skillfully employs special points, such as Zhongzhu (KI15) and ear tips, pays attention to the reform of moxibustion tools, and innovates the moxibustion techniques, using distinctive moxibustion tools and methods to treat low back pain.
Moxibustion/methods* ; Humans ; Data Mining ; Low Back Pain/history* ; Acupuncture Points ; History, Ancient ; Female ; China ; Male ; Adult ; Middle Aged

Objective:To investigate the efficacy of nafamostat mesylate in traumatic extracorporeal membrane oxygenation (ECMO) therapy.Methods:Patients admitted to the intensive care units of Children's Hospital Affiliated to Xi'an Jiaotong University and Gansu Provincial Maternal and Child Health Hospital for ECMO-assisted cardiopulmonary support due to trauma from January 2021 to December 2024 were selected as the study subjects. Based on different anticoagulation strategies, patients were divided into the nafamostat mesylate group( n=11) and the common heparin group( n=10). The general conditions of the two groups were compared. In addition, differences in various clinical indicators during the ECMO-assisted process were compared, including white blood cell count (WBC), platelet count (PLT), hemoglobin level (Hb), hematocrit (HCT), prothrombin time(PT), activated partial thromboplastin time (APTT), thrombin time(TT), fibrinogen(FIB) and D-dimer. Furthermore，the differences in the total volume of component blood transfusions, coagulation substances, complications and prognosis between the two groups were compared. Results:No statistically significant differences were observed between the two groups regarding age, gender, weight, type of ECMO support, type of trauma, presence of active bleeding, or rate of surgical intervention. There were no statistically significant differences in WBC, PLT, Hb, HCT, PT, APTT, TT, FIB, and D-dimer between the two groups of patients prior to the initiation of ECMO support (all P>0.05). Compared with the common heparin group, children in the nafamostat mesylate group had lower PT[(21±6)s vs. (27±3)s; (20±4) vs. (28±5)], APTT[(68±8)s vs. (89±12)s; (64±15)s vs. (85±21)s], TT [(25±11)s vs. (31±13)s; (24±8)s vs. (35±6)s], and D-dimer[(5.8±1.1) μg/mL vs. (11.5±5.6) μg/mL; (4.2±1.8) μg/mL vs. (14.6±2.5) μg/mL],and higher FIB[(2.1±0.5) g/L vs. (1.6 ± 0.3) g/L; (2.4 ± 0.4) g/L vs. (1.3 ± 0.6) g/L] when ECMO assisted for 24 h and 72 h,the differences were all statistically significant(all P<0.05). Compared to the nafamostat mesylate group, the common heparin group exhibited significantly higher total cumulative infusion amounts of red blood cell suspension, plasma, platelet, FIB, hemocoagulase, and thromboplastin complex during the ECMO-assisted process, and the differences were statistically significant (all P<0.05). There was no statistically significant difference between the two groups of children in terms of time to ECMO assistance, membrane lung failure, loop thrombosis, embolism, and successful withdrawal rate (all P>0.05). Conclusion:Nafamostat mesylate can effectively reduce the risk of bleeding and minimize the requirement for blood product and coagulation substance infusions during traumatic ECMO assistance.

Objective To investigate the expression levels of serum tyrosine 3-monooxygenase/tryptophan 5-monooxygenase activation protein β(14-3-3β)and clara's cell secretory protein 16(CC16)in patients with chronic obstructive pulmonary disease(COPD)complicated by respiratory failure,and their relationship with prognosis.Methods A total of 232 patients with COPD complicated with respiratory failure admitted to Yantian Hospital of Southern University of Science and Technology from April 2020 to October 2023 were enrolled in the COPD complicated with respiratory failure group.According to the severity of the disease,they were divided into mild group(n=67),moderate group(n=73)and severe group(n=92).According to the 28-day prognosis,they were divided into death group(n=73)and survival group(n=159).In addition,80 patients with simple COPD(COPD group)and 80 healthy subjects(control group)were selected at the same time.Enzyme-linked immunosorbent assay(ELISA)was used to detect the expression of serum 14-3-3β and CC16.Multivariate Logistic regression analysis was used to analyze the factors of death in patients with COPD complicated with respiratory failure.The receiver operating characteristic(ROC)curve was used to analyze the predictive value of serum 14-3-3β and CC16 expression on the death of patients with COPD complicated with respiratory failure.Results The expression of serum 14-3-3β in COPD complicated with respiratory failure group was higher than that in COPD group and control group(U=3.894,11.417),the expression of CC16 was lower than that in COPD group and control group(t=5.845,14.306),and the differences were statistically significant(all P<0.05),respectively.The expression of serum 14-3-3β in severe group was higher than that in moderate group and mild group(U=5.179,8.234),the expression of CC16 was lower than that of moderate group and mild group(t=4.090,9.281),and the differences were statistically significant(all P<0.05),respectively.The 28-day mortality rate of 232 COPD patients with respiratory failure was 31.47%(73/232).The expression of serum 14-3-3β in the death group was higher than that in the survival group,and the expression of CC16 was lower than that in the survival group,the differences were statistically significant(U/t=6.790,8.265,all P<0.05).The age of the death group was older than that of the survival group,the degree of airflow limitation and the number of acute exacerbations within 1 year were higher than those of the survival group,and the differences were statistically significant(t/χ2/U=3.895,7.202,3.360,all P<0.05).Age,severe airflow limitation,extremely severe airflow limitation,and the number of acute exacerbations within 1 year,elevated 14-3-3β were independent risk factors for death in patients with COPD complicated with respiratory failure(Wald χ2=3.914～22.668,all P<0.05),and elevated CC16 was an independent protective factor(Wald χ2=23.675,P<0.05).The area under the curve(AUC)of serum 14-3-3β combined and CC16 expression in predicting the death of patients with COPD complicated with respiratory failure which was greater than that of serum 14-3-3β and CC16 expression alone,the differences were statistically significant(Z=3.995,3.813,all P<0.01).Conclusion The increase of serum 14-3-3β expression and the decrease of CC16 expression in patients with COPD complicated by respiratory failure are closely related to the aggravation of the disease and poor prognosis.The combination of serum 14-3-3β and CC16 expression is of high value in predicting the death of patients with COPD complicated with respiratory failure.

Objective To explore the value of radiomics and deep transfer learning(DTL)based on gadoxetic acid disodium-enhanced magnetic resonance imaging(MRI)for preoperative prediction of microvascular invasion(MVI)in hepatocellular carcinoma(HCC).Methods A retrospective analysis was conducted using the MRI and clinicopathological data of 369 HCC patients who underwent surgery and had pathologically confirmed MVI at the Affiliated Hospital of Qingdao University from January 2019 to September 2024.According to the negative and positive manifestations of MVI,these patients were divided into MVI-group(n=219)and MVI+group(n=150);and they were then randomly assigned into the training set(n=258)and the test set(n=111)in a ratio of 7:3.Based on the hepatobiliary phase images,the optimal features were extracted and screened from radiomics features,DTL features,and the fusion features of the two.Nine machine learning models were constructed using 3 algorithms(random forest,multi-layer perceptron,and support vector machine,separately)and trained on radiomics features,DTL features,and the fusion features of the two.The diagnostic efficacy of each model was evaluated using receiver operating characteristic curve,and the optimal model was identified as the output model.Results Among all the constructed models,those based on fused features outperformed models using individual features.The random forest classifier model in the training set had the best performance,with an AUC of 0.998(95%CI:0.996-1.000),and was therefore selected as the output model in this study.Conclusion Radiomics and DTL models based on gadoxetic acid disodium-enhanced MRI can effectively predict the MVI in HCC.Among these,the random forest classifier model utilizing fused features in the training set exhibits the best performance.

Objective:To evaluate the relevant factors to optimize the learning curve and the impact of the basic skills of endoscopic technology on the learning curve of gasless transaxillary posterior endoscopic thyroidectomy.Methods:A retrospective analysis was performed to evaluate the clinical outcomes of 50 patients who underwent Glandular Ultrasound-Assisted (GUA) thyroid surgery by a surgeon with a background in endoscopic thyroid surgery via the thoracic-areolar approach, and 50 patients operated on by a surgeon without such experience at the Thyroid Surgery Department of Jilin University China-Japan Union Hospital from Apr. to Dec. 2023. The patients were divided into two groups: the Endoscopic Experience Group and the Non-Endoscopic Experience Group. The Cumulative Sum Control Chart (CUSUM) was applied to construct learning curves for both groups, dividing the technical exploration period from the mastery period. The analysis compared the surgical time, postoperative first-day drainage volume, number of central lymph nodes dissected rates, and postoperative complications between the two groups and across the two phases.Results:The analysis of the learning curve revealed that the inflection point of the Endoscopic Experience Group was 15, while of the Non-Endoscopic Experience Group was 18. The learning curve was divided into the technical exploration stage and the proficient mastery stage. The operative time of technical exploration stagde was significantly longer than of proficient mastery stage of both group (183.46±36.13min vs.144.40±26.14min, P<0.001; 186.89±48.91min vs.131.59±22.90min; P<0.001) . The operative time in the proficient mastery stage of the Endoscopic Experience Group was longer than that of the Non-Endoscopic Experience Group (144.40±26.15min vs. 131.59±22.90min, P<0.05) . The postoperative drainage volume in the Endoscopic Experience Group was lower than that in Non-Endoscopic Experience Group in both stages (65.40±32.48mL vs.93.22±30.67mL, 57.40±15.35mL vs.78.50±28.30mL, P<0.05) , and the postoperative drainage volume in the proficient mastery stage of the Non-Endoscopic Experience Group was significantly lower than in the technical exploration stage (93.22±30.67mL vs.78.50±28.30mL, P<0.05) .No significant differences in central lymph node dissection numbers or postoperative complications were observed between the groups at both stages. Conclusions:There is a specific learning curve in the early stage of gasless transaxillary posterior endoscopic thyroidectomy. After crossing the learning curve, the operation time is obviously shortened with the improvement of the operator's surgical technique.Having a basic understanding of endoscopic technology in the early stage can reduce the occurrence of postoperative drainage, but has a minimal impact on the learning curve.

Objective To explore the value of radiomics and deep transfer learning(DTL)based on gadoxetic acid disodium-enhanced magnetic resonance imaging(MRI)for preoperative prediction of microvascular invasion(MVI)in hepatocellular carcinoma(HCC).Methods A retrospective analysis was conducted using the MRI and clinicopathological data of 369 HCC patients who underwent surgery and had pathologically confirmed MVI at the Affiliated Hospital of Qingdao University from January 2019 to September 2024.According to the negative and positive manifestations of MVI,these patients were divided into MVI-group(n=219)and MVI+group(n=150);and they were then randomly assigned into the training set(n=258)and the test set(n=111)in a ratio of 7:3.Based on the hepatobiliary phase images,the optimal features were extracted and screened from radiomics features,DTL features,and the fusion features of the two.Nine machine learning models were constructed using 3 algorithms(random forest,multi-layer perceptron,and support vector machine,separately)and trained on radiomics features,DTL features,and the fusion features of the two.The diagnostic efficacy of each model was evaluated using receiver operating characteristic curve,and the optimal model was identified as the output model.Results Among all the constructed models,those based on fused features outperformed models using individual features.The random forest classifier model in the training set had the best performance,with an AUC of 0.998(95%CI:0.996-1.000),and was therefore selected as the output model in this study.Conclusion Radiomics and DTL models based on gadoxetic acid disodium-enhanced MRI can effectively predict the MVI in HCC.Among these,the random forest classifier model utilizing fused features in the training set exhibits the best performance.

Objective:To analyze risk factors associated with false-positive results of quantitative fecal immunochemical testing (FIT), evaluate its performance for detecting advanced colorectal neoplasia across different subgroups, and explore the optimal positivity thresholds for each subgroup.Methods:Individuals who participated in the Zhejiang Colorectal Cancer Screening Program in 2020-2021, completed questionnaire-based risk assessment and quantitative FIT for initial screening, and undertook colonoscopy for confirmed diagnosis were included in this study. The information of individuals, including demographic characteristics, lifestyles, history of diseases, and family history of colorectal cancer (CRC), was collected by using questionnaires. The diagnostic outcomes of the individuals were obtained through colonoscopy and pathological examination. Multivariate logistic regression analyses were conducted to identify factors associated with false-positive FIT results. The optimal threshold of FIT was determined based on the receiver operating characteristic (ROC) curve and 10-fold cross-validation. The effectiveness of FIT screening in different subgroups was compared using the unified threshold of 100 ng/ml or optimal positivity thresholds.Results:There were 25 874 individuals included in the analysis, with 14 694 (56.79%) having fecal hemoglobin concentrations ≥100 ng/ml. A total of 3 830 advanced adenoma cases (14.80%) and 362 CRC cases (1.40%) were identified. Age below 60 years old, females, underweight, smoking, drinking, use of nonsteroidal anti-inflammatory drugs, no family history of CRC, no history of intestinal disease, no history of hypertension, and physical inactivity were associated with an elevated risk of false-positive results in FIT ( P<0.05). Compared to the predetermined threshold of 100 ng/ml, the false positive rate (FPR) of quantitative FIT decreased from 52.3% to 37.3% in all individuals, and decreased by more than 20% in females, individuals with normal weight, smokers, and those without a history of intestinal disease when adopting the optimal threshold (all P<0.001). Conclusion:The risk of false-positive results in quantitative FIT varies across different subgroups. Adopting the optimal thresholds could improve the specificity and reduce the FPR of quantitative FIT for CRC screening.