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.

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.

Tuberculosis remains the leading cause of death from a single infectious agent worldwide.China is still one of the countries with a high burden of tuberculosis.The structural and functional damage to lung tissue caused by post-tuberculosis pulmonary disease(PTLD)imposes a considerable disease burden on people.How-ever,due to the insufficient understanding of post-tuberculosis lung diseases in the past,the long-term manage-ment of this disease has been lacking for the post-tuberculosis population.This article elaborates on the epidemiol-ogy and hazards,pathogenesis,clinical manifestations,diagnostic methods,treatment and preventive measures of common PTLD,with the aim of enhancing the awareness and attention of tuberculosis physicians to PTLD.It also summarizes the limitations in the etiological mechanism,diagnosis and treatment methods of PTLD,as well as the urgency of in-depth research.

As the first defense of respiratory system,airway epi-thelial cells(AECs)play an important role in separating the re-spiratory internal and external environment.They are essential for the natural immune function.Small airway lesions are an im-portant early pathology of chronic obstructive pulmonary disease(COPD),when AECs are exposed to harmful particles or gases for a long time,the epithelial barrier is damaged,and the signa-ling pathways which involved in differentiation,repair,and in-flammatory are disordered,resulting in epithelial cell cycle stag-nation and accelerated aging.A number of studies have sugges-ted that AECs of COPD patients express high levels of aging markers,suggesting that senescence of AECs is closely related to COPD.This review discusses the potential mechanisms of AECs senescence in COPD,the impact of AECs senescence on the de-velopment and severity of the disease,and highlights potential targets for modulating cellular senescence in airway epithelium as a therapeutic approach in COPD.

Small molecules generally refer to natural compounds with a molecular weight less than 1 kD,which have significant research value in many fields such as biological science,pharmaceutical and health care,and environmental science,etc.Aptamers,as a new type of bio-recognition element,can bind specifically to small molecules and other targets with high affinity.They are single-stranded short nu-cleotide sequences selected by the systematic evolution of ligands by exponential enrichment(SELEX)technology.Compared with aptamers of large molecules,aptamers of small molecules often exhibit disad-vantages such as difficult screening,low affinity,and poor specificity.To solve these bottleneck prob-lems,the article provides some strategies of screening and identification of aptamers for small molecules.Firstly,when designing the screening plan for aptamers of small molecules,this article suggests focusing on key points such as library design,SELEX methods,amplification methods,screening process monito-ring methods,and aptamer sequence analysis methods,then choosing or designing appropriate strategies based on the actual situation.Secondly,the structure and affinity of aptamers directly affect their applica-tions.The article briefly introduces methods for analyzing the secondary and tertiary structures of aptam-ers and six common methods for affinity determination.Finally,the article discusses some problems exist-ing in screening and identification of aptamers for small molecules and looks forward to future develop-ment prospects.The whole article aims to provide research strategies for multiple steps of screening and i-dentification of aptamers for small molecules and offer references for improving the efficiency of aptamer screening and identification.

Objective To screen and analyze biomarkers of comorbidity in nonalcoholic fatty liver disease(NAFLD)and colorectal cancer(CRC)based on bioinformatics so as to explore the mechanism of comorbidity.Methods After data related to NAFLD and CRC were downloaded from the public database(GEO),gene set enrichment analysis(GSEA)was performed for all genes to screen the expression pathways,and the common differentially expressed genes of both genes were annotated by GO and KEGG functions.Subsequently,PPI protein interaction network was constructed using STRING online database to screen key genes,and transcription factor-gene regulatory network was constructed based on key genes,and biomarker genes were identified using LASSO regression.Finally,the transcription factor prediction analysis of key genes and SCD was conducted through Network Analyst database,and the transcription factor network was constructed.Results GSEA analysis revealed that signaling pathways such as TNF-α and IL-6-JAK-STAT3 were activated in NAFLD patients.The 17 common differentially expressed genes were mainly enriched in biological processes such as bivalent inorganic cation homeostasis and neutrophil chemotaxis in GO analysis.KEGG pathway analysis showed that common differentially expressed genes were mainly concentrated in IL-17 signaling pathway.The protein interaction network screened out five key genes:TREM1,FPR1,IL1RN,S100A9,and S100A8.IL1RN and SCD,identified by LASSO regression,were validated as biomarker genes using external datasets and immunohistochemistry data from the human protein atlas(HPA)database,showing high expression in CRC tissues.A transcription factor-key comorbidity gene regulatory network was constructed using Network Analyst.Conclusion The genes for comorbidities between NAFLD and colorectal cancer discovered by bioinformatics reveal the potential pathogenesis of comorbidities,and provide ideas for cancer screening and early diagnosis of colorectal cancer in NAFLD patients.

Thyrotoxic periodic paralysis(TPP) is a rare but potentially life-threatening complication of thyrotoxicosis. TPP secondary to thyrotoxicosis factitia(TF) is exceptionally uncommon. Here, we report the case of a 31-year-old female who developed TF and subsequent TPP after taking laxatives containing unknown ingredients. The patient presented with acute onset of quadriplegia. Laboratory investigations revealed severe hypokalemia(serum potassium 1.47mmol/L), thyrotoxicosis[thyroid stimulating hormone(TSH)<0.005 mIU/L, free thyroxine(FT 4) 30.84 pmol/L, free triiodothyronine(FT 3) 7.71 pmol/L], and a decreased thyroglobulin level(3.08 ng/mL). The patient′s symptoms resolved rapidly following potassium supplementation, and thyroid function gradually normalized after discontinuation of the suspected medication. This case highlights the importance of considering TF in the differential diagnosis of TPP, particularly in patients with a history of unknown medication use. It also underscores the diagnostic value of thyroglobulin measurement in identifying TF and outlines clinical strategies for the management of TF-induced TPP.

Objective:To evaluate the association between postoperative delirium (POD) and preoperative frailty in elderly patients undergoing knee or hip arthroplasty.Methods:This nested case-control study utilized medical records from elderly patients who underwent knee or hip arthroplasty under combined spinal-epidural anesthesia at Qingdao Municipal Hospital between September 2021 and May 2023. Participants were divided into 2 groups based on clinically diagnosed POD: POD group ( n=53) and non-POD group ( n=256). Univariate analysis was conducted on suspected influencing factors, and logistic regression analysis was utilized to identify the risk factors for POD. Receiver operating characteristic and clinical decision curves were plotted to evaluate the predictive performance of these risk factors for POD. Mediation analysis was performed, and a clinically applicable nomogram was constructed to achieve visual prediction of outcomes. Results:There were statistically significant differences in age, preoperative frailty, body mass index, American Society of Anesthesiologists Physical Status classification, Memorial Delirium Assessment Scale scores, and concentrations of Aβ 42, Aβ 40, phosphorylated tau protein (p-tau protein) and tau protein, Aβ 42/tau ratio and Aβ 42/p-tau ratio in cerebrospinal fluid (CSF) between non-POD group and POD group ( P<0.05). Preoperative frailty was a risk factor for POD ( P<0.05). Mediation analysis revealed that the association between preoperative frailty and POD was mediated by CSF tau protein concentrations. The area under the receiver operating characteristic curve of preoperative frailty and CSF biomarker concentrations in predicting POD was 0.974 ( P<0.05). The clinical decision curve demonstrated that the model combining the preoperative frailty and CSF biomarker concentrations predicted a higher net benefit ( P<0.05). The clinical decision curve showed that the model combining preoperative frailty and CSF biomarker concentrations predicted a higher net benefit. Conclusions:Preoperative frailty is a risk factor for POD in elderly patients undergoing knee or hip arthroplasty, and its combination with CSF biomarker concentrations can effectively predict the occurrence of POD. CSF tau concentration mediates the association between preoperative frailty and development of POD.

Thyrotoxic periodic paralysis(TPP) is a rare but potentially life-threatening complication of thyrotoxicosis. TPP secondary to thyrotoxicosis factitia(TF) is exceptionally uncommon. Here, we report the case of a 31-year-old female who developed TF and subsequent TPP after taking laxatives containing unknown ingredients. The patient presented with acute onset of quadriplegia. Laboratory investigations revealed severe hypokalemia(serum potassium 1.47mmol/L), thyrotoxicosis[thyroid stimulating hormone(TSH)<0.005 mIU/L, free thyroxine(FT 4) 30.84 pmol/L, free triiodothyronine(FT 3) 7.71 pmol/L], and a decreased thyroglobulin level(3.08 ng/mL). The patient′s symptoms resolved rapidly following potassium supplementation, and thyroid function gradually normalized after discontinuation of the suspected medication. This case highlights the importance of considering TF in the differential diagnosis of TPP, particularly in patients with a history of unknown medication use. It also underscores the diagnostic value of thyroglobulin measurement in identifying TF and outlines clinical strategies for the management of TF-induced TPP.

Objective:To explore the value of ultrasound imaging characteristics combined with clinical indicators in assessing the prognosis of patients with pancreatic ductal adenocarcinoma (PDAC).Methods:A retrospective analysis was conducted for patients who underwent pancreatic contrast-enhanced ultrasound (CEUS) from September 2017 to October 2023 at Peking Union Medical College Hospital and were diagnosed with PDAC based on pathological findings. Various parameters were recorded, including CA19-9 levels, tumor size, location, morphologic features, echogenicity, presence of internal cystic components, dilatation of the main pancreatic duct, peripheral vascular invasion, CEUS characteristics, presence or absence of liver metastasis, and treatment methods. In April 2024, patient survival information was obtained through telephone follow-up or review of medical records. Based on the results of the cox regression model analysis, a nomogram model of the risk of death was developed. The receiver operating characteristic (ROC) curves were applied to evaluate the predictive efficacy of the model. The calibration curves were plotted to evaluate the accuracy of the model, and clinical decision curves were used to evaluate the clinical benefit of the model.Results:This study included a total of 207 patients with PDAC. As of April 2024, 71 patients were alive and 136 died, with a median survival time of 14 months (95% CI: 12 -17). Multivariate analysis confirmed that the elevated CA19-9 ( HR=1.689, 95% CI: 1.102-2.588), tumor size >4 cm ( HR=1.641, 95% CI: 1.159-2.322), taller-than-wide shapes ( HR=1.450, 95% CI: 1.019-2.065), incomplete hypo-enhancement ( HR=1.618, 95% CI: 1.100-2.380), and liver metastasis ( HR=1.687, 95% CI: 1.175-2.423) were independent risk factors for survival in patients with PDAC. A nomogram model was further constructed for 6-month, 12-month and 3-year survival of patients with PDAC. The areas under the ROC curve were 0.679, 0.705 and 0.815, respectively. The calibration curves suggested that the model was more accurate, and the clinical decision curves showed that the model had a better clinical benefit. Conclusion:The combined use of ultrasound imaging characteristics and clinical indicators could effectively predict the prognosis of PDAC patients. Specifically, tumor size >4 cm, taller-than-wide shapes, incomplete hypo-enhancement, elevated CA19-9, and the presence of liver metastasis are correlated with poorer survival outcomes. The nomogram model constructed on the basis of these factors can be used to assess the survival of patients with PDAC.