Radiotherapy is a crucial treatment modality for head and neck tumors. However, while effectively killing tumor cells, it significantly disrupts the homeostasis of the oral microecology, which is closely associated with various complications such as radiation-induced oral mucositis. Literature review indicates that as radiotherapy doses accumulate and treatment durations extend, the richness and diversity of the oral microbiota show a declining trend, with the genus Streptococcus decreasing most markedly. In contrast, radiotherapy selectively promotes the proliferation of bacterial phyla such as Proteobacteria and Bacteroidetes, which are rich in opportunistic pathogens. Mechanistically, radiotherapy activates the nuclear factor-kappa B pathway, triggering chronic inflammation and oxidative stress, damaging the epithelial barrier, suppressing local immunity, and causing damage to organs such as the salivary glands. It can also induce systemic diseases via the oral-gut axis, forming a multi-level, interconnected pathogenic network. In terms of interventions, treatment strategies including probiotics and prebiotics have shown promising efficacy against side effects such as radiation-induced oral mucositis. Saliva-based oral microbiota transplantation is an emerging strategy that is expected to become widely utilized for restoring oral microecological balance. Existing interventions provide preliminary pathways for clinical practice, but this field still faces several key scientific questions. The association between oral microecology and systemic diseases remains largely correlative, lacking causal evidence. Furthermore, critical parameters for oral microbiota transplantation, such as donor screening criteria, transplantation protocols, and long-term safety, are not yet well-defined. Therefore, future research should focus on conducting large-scale clinical trials to establish standardized protocols and safety evaluation systems for oral microecological interventions, and explore combined treatment therapies such as probiotics, prebiotics, and microbiota transplantation to advance the development of personalized precision modulation. These will enable more effective management of radiotherapy-induced oral microecological dysbiosis and improve treatment outcomes and quality of life for patients with head and neck tumors.

摘要 目的 近年来多项研究提示瘦素可能在帕金森病（PD）患者神经元细胞中发挥保护作用，但相关结果尚不一致，因此本研究采用Meta分析方法系统评估瘦素在PD患者中的作用。方法 在PubMed、Embase、ISI Web of Science和中国国家知识基础设施（CNKI）数据库中系统性检索2020年3月底前的文献研究。为了评估总体效果，采用固定效应模型计算合并效应量、标准化均差（SMD）和95%CI。随后进行了亚组分析、敏感性分析和发表偏倚分析，以揭示纳入研究的稳定性和可信度。结果 本研究纳入了7项研究，包括167例帕金森病患者和141名健康对照者。本研究发现，帕金森病患者的血清瘦素水平显著低于健康对照者（SMD=-0.328，95%CI -0.560~-0.096，P=0.006）。亚组分析显示，白种人PD患者的血清瘦素水平较低（SMD=-0.301，95%CI -0.566~-0.036，P=0.026），而亚洲患者的血清瘦素水平差异无统计学意义（SMD=-0.414，95%CI -0.891~0.063，P=0.089）。敏感性分析显示总体效应量的稳定性。通过观察漏斗图，未发现明显的不对称，Begg检验也未发现显著的发表偏倚。结论 本Meta分析结果表明PD患者血清瘦素水平较健康人群降低，提示血清瘦素可能参与PD的病理过程，其具体机制尚待进一步研究阐明。 Abstract Objective Many studies in recent years have shown that leptin may exert a protective effect on neurons in patients with Parkinson disease（PD），but related research findings remain inconsistent. Therefore， this article systematically reviews the role of leptin in PD patients through a meta-analysis. Methods PubMed，Embase，ISI Web of Science， and CNKI were searched for related articles published up to March 2020. In order to assess the overall effect， the fixed effects model was used to calculate pooled effect size， standardized mean difference （SMD）， and 95% confidence interval（CI）. Subgroup analysis， sensitivity analysis， and publication bias analysis were performed to assess the stability and reliability of the studies included in the meta-analysis. Results A total of 7 studies were included in the meta-analysis according to strict inclusion criteria， involving 167 PD patients and 141 healthy controls. The results showed that the PD patients had a significantly lower serum level of leptin than the healthy controls （SMD=-0.328，95%CI -0.560 to -0.096， P=0.006）. The subgroup analysis showed that Caucasian PD patients had a relatively low serum level of leptin （SMD=-0.301， 95%CI -0.566 to -0.036，P=0.026）， while there was no significant difference in the serum level of leptin among Asian patients （SMD=-0.414， 95%CI -0.891 to 0.063， P=0.089）.The sensitivity analysis showed the stability of overall effect size. The funnel plots showed no significant asymmetry， and no obvious publication bias was observed by the Begg test. Conclusion This meta-analysis shows that PD patients have a lower serum level of leptin than the healthy population， suggesting that serum leptin might be involved in the pathological process of PD， and further studies are needed to clarify the specific mechanism.
Leptin ; Adipokines

This paper summarizes Professor WANG Xixing's clinical experience in treating immune checkpoint inhibitor-related pneumonitis （CIP） based on the theory of "cough attributed to the five zang （脏） organs". Cough is a common predominant symptom of CIP. According to the theory of "cough attributed to the five zang organs"， drug toxicity triggers cancer toxin， leading to disharmony among the five zang organs， and then lung failing to diffuse and govern descent as the core pathogenesis. Therefore， treatment should focus on harmonizing the five zang organs to restore the normal function of lung qi to diffuse and govern descent. In clinical practice， CIP can be classified into four syndrome patterns， including lung yin depletion， deficiency of both the lung and the spleen with phlegm-dampness， liver fire harassing the lung， and lung-kidney yin deficiency. Correspondingly， Chaimai Jinluo Runfei Decoction （柴麦金络润肺汤） is used to nourish yin and moisten the lung； Qigui Peitu Huayin Decoction （芪桂培土化饮汤） is used to fortify the spleen and tonify the lung， resolve dampness and dispel phlegm； Chaidan Shuyu Runjin Decoction （柴丹疏郁润金汤） is used to drain liver and clear the lung； and Dimai Jinshui Xiangsheng Decoction （地脉金水相生汤） is used to nourish the kidney and moisten the lung.

ObjectiveTo explore the effect of the medical-community linkage model on activities of daily living, psychological status and motor function of stroke patients in the community. MethodsA total of 60 stroke patients admitted to two community health service centers and their affiliated stations in Fengtai District, Beijing, from January, 2024 to August, 2025 were enrolled and randomly divided into control group (n = 30) and intervention group (n = 30). The control group received routine medicine, dietary care and rehabilitation management, while the intervention group underwent rehabilitation with the medical-community linkage model, for twelve weeks. They were assessed with modified Barthel Index (MBI), Hamilton Anxiety Scale (HAMA), Hamilton Depression Scale (HAMD) and Fugl-Meyer Assessment (FMA) before and after intervention. ResultsAfter intervention, the MBI, HAMA, HAMD and FMA scores of patients improved in both groups (|t| > 5.599, P < 0.001), and improved more in the intervention group than in the control group (P < 0.05), except MBI. The HAMA and HAMD scores of family members decreased in both groups (|t| > 10.333, P < 0.001), and decreased more in the intervention group than in the control group (t > 5.681, P < 0.001). ConclusionThe medical-community linkage model can further improve the motor function of stroke patients in community, as well as the psychological status of both patients and their family members.

Objective To understand the antibody levels of diphtheria and tetanus among healthy population in Shanghai Pudong New Area, and to provide a scientific basis for improving the vaccine immunization strategy. Methods Random sampling was used to select healthy people of all ages in 16 communities in Shanghai Pudong New Area from 2018 to 2024, and serum samples were collected and tested for serum anti-diphtheria and tetanus toxin IgG antibodies by enzyme-linked immunosorbent assay (ELISA) method to analyze the antibody positivity rate (≥0.1 IU/ml) and the geometric mean concentration (GMC) of antibodies. Results A total of 3 312 serum samples were included, with a male-to-female ratio of 0.76:1, and 53.77% were local residents. The seropositivity rates and geometric mean concentrations (GMC) of both diphtheria and tetanus antibodies generally declined with increasing age, but exhibited a transient rebound in the 7y-. A total of 1 175 individuals (35.48%) were seropositive for diphtheria, with a GMC of 0.054 IU/mL. For tetanus, 988 individuals (29.83%) were seropositive, with a GMC of 0.033 IU/mL. Significant differences in seropositivity rates (χ²_diphtheria=950.005,χ²_tetanus=1 324.393) and GMC (H_diphtheria=1027.160,H_tetanus=1 142.007) were observed among different age groups (P<0.001). Significant differences in seropositivity rates (χ²_diphtheria=950.005,χ²_tetanus=1324.393) and GMC (H_diphtheria=1027.160,H_tetanus=1142.007) were also found across different years (P<0.001). Conclusion The prevalence of diphtheria and tetanus antibodies in the healthy population of Pudong New Area is relatively low, particularly among adults over 20 years of age with inadequate immunization. This underscores the need to reinforce the National Immunization Program (NIP) vaccine specifications for children under 6 years of age and implement an immunization strategy for adolescents or adults against diphtheria and tetanus.

With the rapid development of the biopharmaceutical field, the efficient and simultaneous extraction of multiple biological components from biological samples has become a critical process for advancing scientific research. The ability to simultaneously extract various molecular components such as metabolites, DNA, RNA, and proteins is pivotal for multi-omics studies, which aim to comprehensively understand the molecular mechanisms of biological systems. Traditional methods often extract these components separately, leading to challenges such as sample loss, time consumption, contamination, and inconsistencies across different data types. In contrast, simultaneous extraction techniques address these issues by maintaining the consistency of each biological component’s physiological state, improving data reliability and facilitating integration across omic platforms. This review systematically summarizes recent advances in simultaneous extraction technologies, focusing on methods such as methanol/chloroform extraction, TRIzol reagent extraction, and modified Folch extraction, which have shown significant promise in improving the efficiency and integrity of biological sample preparation. These methods offer various advantages, such as reduced sample volume requirements, decreased contamination risk, and enhanced extraction consistency, which are crucial for studies involving small sample sizes or precious clinical specimens. Among these, methanol/chloroform extraction stands out for its simplicity, low cost, and ability to extract a wide range of biological molecules. However, it does face limitations, such as its inefficiency in extracting lipids and potential RNA contamination. On the other hand, the TRIzol reagent method has become a widely adopted technique due to its ability to simultaneously isolate RNA, proteins, and metabolites from the same sample. Despite its effectiveness, the TRIzol method has limitations in RNA quality, especially when handling complex samples or those with high protein content. Modified Folch extraction, which combines liquid-liquid extraction with commercial kits, offers a highly efficient way to extract polar metabolites, lipids, RNA, DNA, and proteins from small tissue samples. This method has proven advantageous in terms of extraction yield, especially for challenging or rare samples, although it requires precise handling to avoid cross-contamination between phases. The integration of automated platforms, microfluidics, and high-throughput systems is another exciting avenue for improving simultaneous extraction. Automation facilitates large-scale, reproducible sample processing with minimal human error, while microfluidics provides high precision in sample handling and enables real-time monitoring of extraction efficiency. These innovations not only enhance the speed and reproducibility of sample preparation but also open new possibilities for single-cell analysis, where sample volumes are often limited, and extraction efficiency is critical. In addition to the technical aspects, the review also highlights the importance of optimizing extraction protocols for specific sample types, such as clinical tissues, plants, and microorganisms. For example, the challenge of extracting multiple components from cancer tissues, where sample degradation and contamination risks are high, can be mitigated by carefully selecting extraction reagents and minimizing sample handling steps. Similarly, in plant studies, where metabolite diversity is vast, the simultaneous extraction methods must be optimized to account for the unique composition of plant tissues, which often include complex secondary metabolites and cell wall components. Looking forward, the development of more efficient and standardized simultaneous extraction methods will be crucial for advancing multi-omics research. There is a growing need for protocols that can be tailored to specific research needs, ensuring both reproducibility and flexibility in diverse applications. Additionally, combining these extraction methods with high-resolution analytical techniques such as mass spectrometry and next-generation sequencing will further enhance the potential of multi-omics studies to provide comprehensive insights into biological systems. As these technologies continue to evolve, their application in personalized medicine, environmental research, and agriculture holds great promise for addressing critical scientific challenges. In conclusion, while simultaneous extraction technologies have made significant strides, several challenges remain in optimizing extraction efficiency, ensuring reproducibility, and reducing costs. Future research should focus on refining extraction protocols, developing innovative extraction reagents, and expanding the scope of these methods to cater to a broader range of biological samples. Ultimately, the continued integration of these advanced techniques will revolutionize the way biological samples are prepared, analyzed, and understood in the context of multi-omics research.

ObjectiveDiscriminating atypical hepatocellular carcinoma (HCC) from other malignancies in liver nodules classified as Liver Imaging Reporting and Data System category M (LR-M) remains a significant diagnostic challenge on conventional ultrasound examination. The LR-M category, originally intended to capture non-HCC malignancies, paradoxically contains up to 63% of atypical HCCs that deviate from classic enhancement patterns, leading to potential misdiagnosis and suboptimal treatment planning. While deep learning has shown promise in HCC diagnosis, most existing models rely exclusively on single-modality ultrasound, overlooking the diagnostic benefits of integrating complementary information from multiple imaging sources. To address this gap, we propose a novel attention-weighted tri-modal ultrasound network (TUS-Net) that integrates contrast-enhanced ultrasound (CEUS), B-mode ultrasound (BUS), and time-intensity curves (TICs) to improve diagnostic accuracy for these clinically challenging lesions. MethodsOur framework incorporates a three-dimensional convolutional neural network (C3D) backbone to extract spatiotemporal features from CEUS videos, capturing dynamic vascular patterns critical for lesion characterization. To effectively fuse complementary modalities, we introduce a dual-channel feature fusion module (DCFFM) that adaptively combines features from CEUS and BUS through channel-wise attention mechanisms, allowing the model to dynamically weigh the contribution of each modality based on diagnostic relevance. Additionally, we propose a temporal intensity feature fusion module (TIFFM) that leverages quantitative hemodynamic information from TICs to guide the model’s attention toward diagnostically critical temporal phases, such as arterial wash-in and portal venous washout. The model is further enhanced by automated lesion localization using YOLOX and class activation mapping for interpretability, ensuring that predictions align with clinically meaningful imaging features. ResultsEvaluated on a tri-modal ultrasound dataset comprising 161 patients with pathologically confirmed LR-M nodules (131 atypical HCC and 30 non-HCC malignancies), our model achieved an accuracy of 86.83%, a sensitivity of 92.50%, a specificity of 75.50%, and an AUC of 89.32% in screening atypical HCC. Compared to single-modality baselines, TUS-Net demonstrated superior specificity, a clinically critical metric given the higher risk associated with misclassifying non-HCC malignancies. Ablation studies confirmed the contribution of each module, with the full model outperforming both standard C3D and 3D ResNet backbones integrated with attention mechanisms. A reader study involving junior and senior radiologists further validated the clinical utility of AI assistance, showing consistent improvements in specificity and inter-reader consistency, particularly for less experienced clinicians. ConclusionThese results surpass existing benchmark models and demonstrate the potential of our approach to enhance diagnostic precision in clinically specific cases. By intelligently fusing multi-modal ultrasound data with attention-guided mechanisms, TUS-Net offers a reliable and interpretable tool that holds promise for improving the non-invasive diagnosis of atypical HCC in challenging LR-M liver nodules.

With the rapid development of the biopharmaceutical field, the efficient and simultaneous extraction of multiple biological components from biological samples has become a critical process for advancing scientific research. The ability to simultaneously extract various molecular components such as metabolites, DNA, RNA, and proteins is pivotal for multi-omics studies, which aim to comprehensively understand the molecular mechanisms of biological systems. Traditional methods often extract these components separately, leading to challenges such as sample loss, time consumption, contamination, and inconsistencies across different data types. In contrast, simultaneous extraction techniques address these issues by maintaining the consistency of each biological component’s physiological state, improving data reliability and facilitating integration across omic platforms. This review systematically summarizes recent advances in simultaneous extraction technologies, focusing on methods such as methanol/chloroform extraction, TRIzol reagent extraction, and modified Folch extraction, which have shown significant promise in improving the efficiency and integrity of biological sample preparation. These methods offer various advantages, such as reduced sample volume requirements, decreased contamination risk, and enhanced extraction consistency, which are crucial for studies involving small sample sizes or precious clinical specimens. Among these, methanol/chloroform extraction stands out for its simplicity, low cost, and ability to extract a wide range of biological molecules. However, it does face limitations, such as its inefficiency in extracting lipids and potential RNA contamination. On the other hand, the TRIzol reagent method has become a widely adopted technique due to its ability to simultaneously isolate RNA, proteins, and metabolites from the same sample. Despite its effectiveness, the TRIzol method has limitations in RNA quality, especially when handling complex samples or those with high protein content. Modified Folch extraction, which combines liquid-liquid extraction with commercial kits, offers a highly efficient way to extract polar metabolites, lipids, RNA, DNA, and proteins from small tissue samples. This method has proven advantageous in terms of extraction yield, especially for challenging or rare samples, although it requires precise handling to avoid cross-contamination between phases. The integration of automated platforms, microfluidics, and high-throughput systems is another exciting avenue for improving simultaneous extraction. Automation facilitates large-scale, reproducible sample processing with minimal human error, while microfluidics provides high precision in sample handling and enables real-time monitoring of extraction efficiency. These innovations not only enhance the speed and reproducibility of sample preparation but also open new possibilities for single-cell analysis, where sample volumes are often limited, and extraction efficiency is critical. In addition to the technical aspects, the review also highlights the importance of optimizing extraction protocols for specific sample types, such as clinical tissues, plants, and microorganisms. For example, the challenge of extracting multiple components from cancer tissues, where sample degradation and contamination risks are high, can be mitigated by carefully selecting extraction reagents and minimizing sample handling steps. Similarly, in plant studies, where metabolite diversity is vast, the simultaneous extraction methods must be optimized to account for the unique composition of plant tissues, which often include complex secondary metabolites and cell wall components. Looking forward, the development of more efficient and standardized simultaneous extraction methods will be crucial for advancing multi-omics research. There is a growing need for protocols that can be tailored to specific research needs, ensuring both reproducibility and flexibility in diverse applications. Additionally, combining these extraction methods with high-resolution analytical techniques such as mass spectrometry and next-generation sequencing will further enhance the potential of multi-omics studies to provide comprehensive insights into biological systems. As these technologies continue to evolve, their application in personalized medicine, environmental research, and agriculture holds great promise for addressing critical scientific challenges. In conclusion, while simultaneous extraction technologies have made significant strides, several challenges remain in optimizing extraction efficiency, ensuring reproducibility, and reducing costs. Future research should focus on refining extraction protocols, developing innovative extraction reagents, and expanding the scope of these methods to cater to a broader range of biological samples. Ultimately, the continued integration of these advanced techniques will revolutionize the way biological samples are prepared, analyzed, and understood in the context of multi-omics research.

ObjectiveDiscriminating atypical hepatocellular carcinoma (HCC) from other malignancies in liver nodules classified as Liver Imaging Reporting and Data System category M (LR-M) remains a significant diagnostic challenge on conventional ultrasound examination. The LR-M category, originally intended to capture non-HCC malignancies, paradoxically contains up to 63% of atypical HCCs that deviate from classic enhancement patterns, leading to potential misdiagnosis and suboptimal treatment planning. While deep learning has shown promise in HCC diagnosis, most existing models rely exclusively on single-modality ultrasound, overlooking the diagnostic benefits of integrating complementary information from multiple imaging sources. To address this gap, we propose a novel attention-weighted tri-modal ultrasound network (TUS-Net) that integrates contrast-enhanced ultrasound (CEUS), B-mode ultrasound (BUS), and time-intensity curves (TICs) to improve diagnostic accuracy for these clinically challenging lesions. MethodsOur framework incorporates a three-dimensional convolutional neural network (C3D) backbone to extract spatiotemporal features from CEUS videos, capturing dynamic vascular patterns critical for lesion characterization. To effectively fuse complementary modalities, we introduce a dual-channel feature fusion module (DCFFM) that adaptively combines features from CEUS and BUS through channel-wise attention mechanisms, allowing the model to dynamically weigh the contribution of each modality based on diagnostic relevance. Additionally, we propose a temporal intensity feature fusion module (TIFFM) that leverages quantitative hemodynamic information from TICs to guide the model’s attention toward diagnostically critical temporal phases, such as arterial wash-in and portal venous washout. The model is further enhanced by automated lesion localization using YOLOX and class activation mapping for interpretability, ensuring that predictions align with clinically meaningful imaging features. ResultsEvaluated on a tri-modal ultrasound dataset comprising 161 patients with pathologically confirmed LR-M nodules (131 atypical HCC and 30 non-HCC malignancies), our model achieved an accuracy of 86.83%, a sensitivity of 92.50%, a specificity of 75.50%, and an AUC of 89.32% in screening atypical HCC. Compared to single-modality baselines, TUS-Net demonstrated superior specificity, a clinically critical metric given the higher risk associated with misclassifying non-HCC malignancies. Ablation studies confirmed the contribution of each module, with the full model outperforming both standard C3D and 3D ResNet backbones integrated with attention mechanisms. A reader study involving junior and senior radiologists further validated the clinical utility of AI assistance, showing consistent improvements in specificity and inter-reader consistency, particularly for less experienced clinicians. ConclusionThese results surpass existing benchmark models and demonstrate the potential of our approach to enhance diagnostic precision in clinically specific cases. By intelligently fusing multi-modal ultrasound data with attention-guided mechanisms, TUS-Net offers a reliable and interpretable tool that holds promise for improving the non-invasive diagnosis of atypical HCC in challenging LR-M liver nodules.

ObjectiveTo investigate the efficacy of selective internal radiation therapy （SIRT） in patients with unresectable hepatocellular carcinoma， and to provide a reference for the selection of clinical treatment regimens. MethodsA retrospective analysis was performed for the clinical data of 73 patients with unresectable hepatocellular carcinoma who received yttrium-90 microsphere SIRT in Eastern Hepatobiliary Surgery Hospital from May 1， 2023 to September 1， 2024. According to tumor characteristics， physical status， liver reserve function， laboratory tests， and SIRT treatment strategy， the patients were divided into radiation segmentectomy group with 9 patients， conversion therapy group with 47 patients， and palliative treatment group with 17 patients. Based on the results of postoperative follow-up， modified Response Evaluation Criteria in Solid Tumors were used to assess radiographic images. A one-way analysis of variance was used for comparison of normally distributed continuous data between three groups， and the chi-square test was used for comparison of categorical data between three groups； the Logistic regression model was used to perform the multivariate analysis. ResultsThere was a significant difference in postoperative outcome between the radiation segmentectomy group， the conversion therapy group， and the palliative treatment group （χ²=30.060， P<0.001）. The disease control rate was 100.0% （9/9） in the radiation segmentectomy group， 83.0% （39/47） in the conversion therapy group， and 29.4% （5/17） in the palliative treatment group， with a significant difference between the three groups （χ²=19.575， P<0.001）， and there was also a significant difference in objective response rate between the three groups （χ²=17.749， P<0.001）. The multivariate Logistic regression analysis showed that the number of tumors （odds ratio ［OR］=0.085， 95% confidence interval ［CI］： 0.008 — 0.906， P=0.041） and combined targeted immunotherapy （OR=18.808， 95%CI： 1.704 — 207.616， P=0.017） were independent influencing factors for achieving complete response. ConclusionThe number of tumors is an independent influencing factor for the efficacy of SIRT and is an important basis for selecting different treatment goals. SIRT combined with targeted immunotherapy may achieve better efficacy.