To develop and validate a model for predicting intensive care unit (ICU) mortality risk in patients with malignant tumors complicated by acute respiratory failure (ARF) based on an explainable machine learning framework.

OBJECTIVE To explore the potential mechanism by which drug-containing serum of Dianxianqing granules （DXQ） inhibits microglial ferroptosis. METHODS Male SD rats were given normal saline and Dianxianqing granules solution via intragastric administration to prepare normal serum and DXQ， respectively. Mice microglia BV2 cells were collected and successfully transfected with a negative control small interfering RNA （si-NC）， and then they were included in the si-NC group and cultured under normal conditions. Cells successfully transfected with small interfering RNA targeting glutathione peroxidase 4 （GPX4） （si-GPX4） were divided into the si-GPX4 group， the CsA group （treated with 1 μmol/L cyclosporine A）， and the DXQ- L， DXQ-M and DXQ-H groups （treated with 5%， 7% and 10% DXQ， respectively）. These groups were subsequently treated with their corresponding drug solutions and ferroptosis inducer Erastin （10 μmol/L）. The intracellular levels of total iron ions， glutathione （GSH）， reactive oxygen species （ROS）， and the expression of mitochondrial superoxide were determined in each group after 48 h of treatment. Additionally， mitochondrial membrane potential， the opening degree of mitochondrial permeability transition pore （MPTP）， and mRNA expressions of GPX4 and cyclophilin D （CypD） were detected. Furthermore， the expressions of ferroptosis-related proteins[GPX4， transferrin receptor 1 （TfR1） and ferritin heavy chain 1 （FTH1）]， as well as MPTP-related proteins [adenine nucleotide translocator （ANT）， cytochrome C （CytC）， mitochondrial calcium uniporter （MCU） and CypD] were assessed. RESULTS Compared with si-NC group， the levels of total iron ions and ROS， the expression level of mitochondrial superoxide， the opening degree of MPTP， protein and its mRNA expressions of CypD as well as protein expressions of TfR1 and MCU were increased or up-regulated significantly （P＜0.01）； however， GSH content， mitochondrial membrane potential， protein and mRNA expressions of GPX4， and protein expressions of FTH1， ANT and CytC were decreased or down-regulated significantly （P＜0.01）. Compared with the si-GPX4 group， the cells in the DXQ-M， DXQ-H groups showed a general improvement in the above quantitative indicators （P＜0.01 or P＜0.05）. CONCLUSIONS DXQ can enhance antioxidant capacity by activating the GSH/GPX4 pathway， regulate the expressions of TfR1 and FTH1 protein to correct iron ion homeostasis， inhibit excessive opening of MPTP to improve mitochondrial function， and ultimately suppress microglial ferroptosis.

OBJECTIVE To evaluate the efficacy and safety of immune checkpoint inhibitors （ICIs） as first-line therapy for advanced gastric cancer. METHODS PubMed， Web of Science， Embase， The Cochrane Library， Wanfang Data， CNKI， and VIP databases were searched to collect phase Ⅲ clinical randomized controlled trials （RCTs） on ICIs as first-line therapy for advanced gastric cancer， as well as abstracts from relevant oncology academic conferences. The search period spanned from database inception to June 1， 2025. After screening literature， extracting data， and assessing quality， a network meta-analysis was performed using R software version 4.3.2. RESULTS A total of 8 studies involving 7 801 patients were included. Network meta-analysis results showed that， in terms of efficacy， compared with chemotherapy （Chemo）， SHR-1701_Chemo， Cadonilimab_Chemo， Sintilimab_Chemo， Pembrolizumab_Chemo， and Tislelizumab_Chemo significantly prolonged median overall survival （OS） and median progression free survival （PFS） in patients （P＜0.05）； whereas Nivolumab_Chemo only significantly improved median PFS （P＜0.05）. Surface under the cumulative ranking curve （SUCRA） results indicated that the top 2 interventions for median OS were SHR-1701_Chemo and Cadonilimab_Chemo； for PFS， the top 2 were Cadonilimab_Chemo and SHR-1701_Chemo. For patients with combined positive score （CPS） ≥5 points for programmed death-ligand 1 （PD-L1）， Cadonilimab_Chemo and SHR- 1701_Chemo also demonstrated the optimal OS and PFS benefits （P＜0.05）. Regarding safety， there were no statistically significant differences among the interventions in the incidence of any adverse events （AEs） or grade ≥3 AEs （P＞0.05）. The SUCRA ranking for the incidence of any AEs showed the top 2 were SHR-1701_Chemo and Chemo； for grade ≥3 AEs， the top 2 were Chemo and Sugemalimab_Chemo. CONCLUSIONS For patients with advanced gastric cancer， Cadonilimab_Chemo and SHR-1701_Chemo demonstrate the best benefits in terms of OS and PFS， with their advantages remaining clear in patients with PD-L1 CPS≥5 points. In terms of safety， the risk of developing any AEs and grade ≥3 AEs is relatively lowest with Chemo.

ObjectiveA middle cerebral artery occlusion (MCAO) mouse model is established by electrocoagulation of the middle cerebral artery. The study examines the mechanism by which exosomes (EXO) derived from human amniotic mesenchymal stem cells (hAMSCs) improve ischemic stroke and regulate neural ferroptosis-related injury. MethodsThirty-two SPF-grade male C57BL/6J mice aged 6 - 8 weeks were randomly divided into four groups (n=8 per group): sham group (Sham), model group (MCAO), MCAO plus normal saline group (MCAO+NaCl), and MCAO plus exosome group (MCAO+EXO). The mouse MCAO model was established by electrocoagulation of the middle cerebral artery. Mice in the Sham group underwent exposure of the middle cerebral artery without electrocoagulation. Twenty-four hours before MCAO induction, mice in the MCAO+EXO group received a tail vein injection of 100 μL of exosomes derived from the culture supernatant of hAMSCs at a concentration of 9.5×10¹¹ particles/mL. Mice in the MCAO+NaCl group were injected with an equal volume of normal saline via the tail vein. Twenty-four hours after model establishment, neurological deficits were evaluated using the Longa neurological deficit scoring system. Cerebral infarct volume was assessed by 2,3,5-triphenyltetrazolium chloride (TTC) staining. Hematoxylin and eosin (HE) staining was performed to evaluate morphological changes of neurons in the ischemic brain regions. The contents of ferrous iron (Fe²⁺), malondialdehyde (MDA), total glutathione (total GSH), oxidized glutathione (GSSG), and reduced glutathione (GSH) in the infarct core and peri-infarct regions were determined using microcolorimetric assays to evaluate differences among groups. The mRNA expression levels of ferroptosis-related factors, including nuclear factor erythroid 2-related factor 2 (NRF2), solute carrier family 7 member 11 (SLC7A11), and glutathione peroxidase 4 (GPX4) in the infarct core and peri-infarct regions were measured by real-time quantitative PCR. Protein expression levels of NRF2, SLC7A11, and GPX4 in the infarct and peri-infarct regions of each group were analyzed by Western blotting. ResultsCompared with the MCAO group, the Longa neurological deficit score was significantly reduced in the MCAO+EXO group (P<0.01). Prominent cerebral infarction was observed in the MCAO group, whereas the infarct volume ratio was markedly decreased in the MCAO+EXO group compared with the MCAO group (P<0.001). Histopathological analysis revealed that mice in the MCAO group exhibited obvious neuronal damage, including cytoplasmic vacuolar degeneration, nuclear pyknosis and fragmentation, unclear nuclear structure, and disorganized neuronal arrangement, compared with the Sham group. In contrast, neurons in the MCAO+EXO group showed relatively preserved morphology, with intact cellular structures and large, regular nuclei located centrally within the cells. Biochemical analysis demonstrated that Fe²⁺ and MDA levels in the infarct core and peri-infarct regions were significantly increased in the MCAO group compared with the Sham group (P<0.001). These levels were significantly reduced in the MCAO+EXO group compared with the MCAO group (P<0.01). In addition, total glutathione (total GSH), oxidized glutathione (GSSG), and reduced glutathione (GSH) levels were markedly decreased in the MCAO group relative to the Sham group (P<0.01). Compared with the MCAO group, the MCAO+EXO group exhibited significantly increased levels of total GSH and GSH (P<0.001), while no significant change was observed in GSSG levels (P>0.05). Furthermore, both mRNA and protein expression levels of nuclear factor erythroid 2-related factor 2 (NRF2), solute carrier family 7 member 11 (SLC7A11), and glutathione peroxidase 4 (GPX4) were significantly downregulated in the MCAO group compared with the Sham group (P<0.01, P<0.001). In contrast, both mRNA and protein expression levels of NRF2, SLC7A11, and GPX4 were significantly upregulated in the MCAO+EXO group compared with the MCAO group (P<0.05). ConclusionIn the mouse MCAO model, tail vein injection of exosomes derived from hAMSCs can improve motor function, reduce infarct area, protect neuronal cell morphology, and reduce the degree of nerve injury. Exosomes may exert a protective effect by activating the NRF2/SLC7A11/GPX4 pathway and reducing ferroptosis in neuronal cells of MCAO model mice.

ObjectiveHeadspace solid-phase microextraction-gas chromatography-mass spectrometry（HS-SPME-GC-MS） and GC-triple quadrupole MS（GC-QqQ-MS） in combination with non-targeted and targeted metabolomics were employed to systematically analyze the chemical composition differences of Xihuangwan prepared with natural musk and artificial musk， and establish an identification system for them. MethodsThe volatile components of 9 batches of Xihuangwan samples from 8 manufacturers were analyzed by HS-SPME-GC-MS non-targeted metabolomics， and identified by comparing their MS data with the National Institute of Standards and Technology（NIST） spectral library. Orthogonal partial least squares-discriminant analysis（OPLS-DA） was used to identify differential volatile components of Xihuangwan prepared with natural musk and artificial musk. Additionally， GC-QqQ-MS targeted metabolomics was applied to quantify the levels of α-pinene， β-elemene， muscone， dehydroepiandrosterone， bornyl acetate， and octyl acetate in 27 batches of samples from 9 manufacturers. Cluster analysis， principal component analysis（PCA）， and partial least squares-discriminant analysis（PLS-DA） were conducted to further explore the differences in volatile components between Xihuangwan samples prepared with natural musk and artificial musk. ResultsNon-targeted metabolomics identified 291 volatile compounds in Xihuangwan， including alkanes， esters， alkanes， alcohols， ketones， naphthalenes and others. OPLS-DA analysis revealed distinct separation between Xihuangwan samples containing artificial musk（A1， C1， D1， E1， F1， G1， I1） and those containing natural musk（H1， H3）. A total of 30 differential metabolites were identified. The relative contents of these 30 differential metabolites were visualized using a radar chart， revealing significant differences in the levels of octanol， borneol acetate and muscone. Cluster analysis and PCA results from targeted metabolomics indicated that Xihuangwan could be classified into two distinct groups：one composed of natural musk（H1， H3） and the other of artificial musk， sample H2. PLS-DA identified muscone， octyl acetate， and dehydroepiandrosterone as key differential volatile components. Although no significant difference was observed in the content of octyl acetate between the two groups， statistically significant differences were found for muscone and dehydroepiandrosterone（P<0.05）. ConclusionMuscone and dehydroepiandrosterone can be used for the differentiation of Xihuangwan samples containing natural musk from those containing artificial musk. This study systematically and comprehensively analyzed the differences in the types and contents of major volatile components in Xihuangwan prepared with natural musk and artificial musk， providing a scientific basis for quality evaluation and control of Xihuangwan.

ObjectiveHeadspace solid-phase microextraction-gas chromatography-mass spectrometry（HS-SPME-GC-MS） and GC-triple quadrupole MS（GC-QqQ-MS） in combination with non-targeted and targeted metabolomics were employed to systematically analyze the chemical composition differences of Xihuangwan prepared with natural musk and artificial musk， and establish an identification system for them. MethodsThe volatile components of 9 batches of Xihuangwan samples from 8 manufacturers were analyzed by HS-SPME-GC-MS non-targeted metabolomics， and identified by comparing their MS data with the National Institute of Standards and Technology（NIST） spectral library. Orthogonal partial least squares-discriminant analysis（OPLS-DA） was used to identify differential volatile components of Xihuangwan prepared with natural musk and artificial musk. Additionally， GC-QqQ-MS targeted metabolomics was applied to quantify the levels of α-pinene， β-elemene， muscone， dehydroepiandrosterone， bornyl acetate， and octyl acetate in 27 batches of samples from 9 manufacturers. Cluster analysis， principal component analysis（PCA）， and partial least squares-discriminant analysis（PLS-DA） were conducted to further explore the differences in volatile components between Xihuangwan samples prepared with natural musk and artificial musk. ResultsNon-targeted metabolomics identified 291 volatile compounds in Xihuangwan， including alkanes， esters， alkanes， alcohols， ketones， naphthalenes and others. OPLS-DA analysis revealed distinct separation between Xihuangwan samples containing artificial musk（A1， C1， D1， E1， F1， G1， I1） and those containing natural musk（H1， H3）. A total of 30 differential metabolites were identified. The relative contents of these 30 differential metabolites were visualized using a radar chart， revealing significant differences in the levels of octanol， borneol acetate and muscone. Cluster analysis and PCA results from targeted metabolomics indicated that Xihuangwan could be classified into two distinct groups：one composed of natural musk（H1， H3） and the other of artificial musk， sample H2. PLS-DA identified muscone， octyl acetate， and dehydroepiandrosterone as key differential volatile components. Although no significant difference was observed in the content of octyl acetate between the two groups， statistically significant differences were found for muscone and dehydroepiandrosterone（P<0.05）. ConclusionMuscone and dehydroepiandrosterone can be used for the differentiation of Xihuangwan samples containing natural musk from those containing artificial musk. This study systematically and comprehensively analyzed the differences in the types and contents of major volatile components in Xihuangwan prepared with natural musk and artificial musk， providing a scientific basis for quality evaluation and control of Xihuangwan.

Objective To explore whether there is a causal relationship between intestinal flora and esophageal cancer. Methods Summary statistics of intestinal flora and esophageal cancer were obtained from the Genome-wide Association Studies (GWAS) database. Five methods, including inverse variance weighted (IVW), weighted median estimation, Mendelian randomization (MR)-Egger regression, single mode, and weighted mode, were used for analysis, with IVW as the main analysis method. Sensitivity analysis was used to evaluate the reliability of MR results. Results In the IVW method, Oxalobacteraceae [OR=1.001, 95%CI (1.000, 1.002), P=0.023], Faecalibacterium [OR=1.001, 95%CI (1.000, 1.002), P=0.028], Senegalimassilia [OR=1.002, 95%CI (1.000, 1.003), P=0.006] and Veillonella [OR=1.001, 95%CI (1.000, 1.002), P=0.018] were positively correlated with esophageal cancer, while Burkholderiales [OR=0.999, 95%CI (0.998, 1.000), P=0.002], Eubacterium oxidoreducens [OR=0.998, 95%CI (0.997, 0.999), P=0.038], Romboutsia [OR=0.999, 95%CI (0.998, 1.000), P=0.048] and Turicibacter [OR=0.998, 95%CI (0.997, 0.999), P=0.013] were negatively correlated with esophageal cancer. Sensitivity analysis showed no evidence of heterogeneity, horizontal pleiotropy and reverse causality. Conclusion Oxalobacteraceae, Faecalibacterium, Senegalimassilia and Veillonella increase the risk of esophageal cancer, while Burkholderiales, Eubacterium oxidoreducens, Romboutsia and Turicibacter decrease the risk of esophageal cancer. Further studies are needed to explore how these bacteria affect the progression of esophageal cancer.

Qualitative and quantitative analysis methods for chemical components of different processed products of Corni Fructus were established to systematically characterize and identify these components, and the content of the main differential components was determined. The chemical components of different processed products of Corni Fructus were collected using ultra-high performance liquid chromatography-quadrupole time-of-flight tandem mass spectrometry(UPLC-Q-TOF-MS). Through analysis of self-built databases, literature, and reference standards, a total of 93 components were obtained, including 19 iridoids, 15 flavonoids, 16 organic acids, eight triterpenoids, eight tannins, four amino acids, two polysaccharides, five olefins, and 16 other compounds. Additionally, by using multivariate statistical methods, the differential components between different processed products of Corni Fructus were screened under the conditions of VIP>1.0 and FC<0.5 or FC>2.0 and P<0.05. The PCA and OPLS-DA results showed differences in the chemical components between different processed products of Corni Fructus. A total of 21 differential components were screened, including tartaric acid, morroniside, and rutin. On this basis, ultra-high performance liquid chromatography-triple quadrupole tandem mass spectrometry(UPLC-QqQ-MS/MS) was used to determine the content of 10 main common differential components, including gallic acid, morroniside, ursolic acid, loganin, swertiamarin, rutin, 5-hydroxymethylfurfural, cornuside Ⅰ, quercetin, and oleanolic acid. The above 10 components showed a good linear relationship within the determined concentration range, with the precision, stability, repeatability, and sample recovery rate all meeting the requirements. Compared with that in Corni Fructus, the content of iridoid glycosides in wine-prepared Corni Fructus and wine-and honey-prepared Corni Fructus decreased, while the content of gallic acid, rutin, quercetin, 5-hydroxymethylfurfural, ursolic acid, and oleanolic acid increased. Compared with wine-prepared Corni Fructus, wine-and honey-prepared Corni Fructus showed varying degrees of increase in all other components, except for a slight decrease in gallic acid content. In summary, this study clarified the influence of different processing methods on the chemical components of Corni Fructus, providing a theoretical basis for the scientific connotation, overall quality evaluation, and clinically rational application of Corni Fructus processing in the future.
Tandem Mass Spectrometry/methods* ; Chromatography, High Pressure Liquid/methods* ; Cornus/chemistry* ; Drugs, Chinese Herbal/chemistry* ; Fruit/chemistry*

With the continuous advancement of medical research, it is increasingly recognized that the human body functions as a highly coordinated complex system, and the development of diseases often involves intricate interactions among multiple subsystems, including organs, tissues, and cells. Conventional pharmacological research, which primarily focuses on isolated subsystems, tends to emphasize direct interactions between drugs and the molecular targets in diseased organs. However, this approach often falls short in addressing the multifaceted challenges posed by complex diseases such as metabolic disorders, autoimmune diseases, cancers, and aging. In recent years, inter-organ cross-talk and its role in diseases progression, as well as cross-organ effects of drug intervention, have gained significant attention. This has highlighted the potential for treating complex diseases through holistic regulation of multiple organs. Traditional Chinese medicine(TCM) has long embraced a holistic and systemic approach for treatment, with concepts such as the interdependence and mutual restraint of the five Zang organs, the interconnection of Zang organs and Fu organs, treating the Zang organ diseases by regulating the Fu organs, treating the child organ diseases to cure the parent organs, and treating upper organ diseases by regulating lower organs. These concepts provide valuable insights into exploring the pathways and molecular mechanisms underlying inter-organ cross-talk. Building on our previous work on indirect actions of TCM, this paper introduces the concept of indirect pharmacology mediated by intermediate substances, as a new extension of classical pharmacology. This approach aims to offer new perspectives and innovative ideas for understanding inter-organ cross-talk and discovering cross-organ therapeutic strategies.
Humans ; Medicine, Chinese Traditional ; Drugs, Chinese Herbal/pharmacology* ; Animals

Medical visual question answering (MVQA) plays a crucial role in the fields of computer-aided diagnosis and telemedicine. Due to the limited size and uneven annotation quality of the MVQA datasets, most existing methods rely on additional datasets for pre-training and use discriminant formulas to predict answers from a predefined set of labels. This approach makes the model prone to overfitting in low resource domains. To cope with the above problems, we propose an image-aware generative MVQA method based on image caption prompts. Firstly, we combine a dual visual feature extractor with a progressive bilinear attention interaction module to extract multi-level image features. Secondly, we propose an image caption prompt method to guide the model to better understand the image information. Finally, the image-aware generative model is used to generate answers. Experimental results show that our proposed method outperforms existing models on the MVQA task, realizing efficient visual feature extraction, as well as flexible and accurate answer outputs with small computational costs in low-resource domains. It is of great significance for achieving personalized precision medicine, reducing medical burden, and improving medical diagnosis efficiency.
Humans ; Image Processing, Computer-Assisted/methods* ; Diagnosis, Computer-Assisted/methods* ; Algorithms ; Telemedicine