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.

ObjectivePrimary liver cancer, predominantly hepatocellular carcinoma (HCC), is a significant global health issue, ranking as the sixth most diagnosed cancer and the third leading cause of cancer-related mortality. Accurate and early diagnosis of HCC is crucial for effective treatment, as HCC and non-HCC malignancies like intrahepatic cholangiocarcinoma (ICC) exhibit different prognoses and treatment responses. Traditional diagnostic methods, including liver biopsy and contrast-enhanced ultrasound (CEUS), face limitations in applicability and objectivity. The primary objective of this study was to develop an advanced, light-weighted classification network capable of distinguishing HCC from other non-HCC malignancies by leveraging the automatic analysis of brightness changes in CEUS images. The ultimate goal was to create a user-friendly and cost-efficient computer-aided diagnostic tool that could assist radiologists in making more accurate and efficient clinical decisions. MethodsThis retrospective study encompassed a total of 161 patients, comprising 131 diagnosed with HCC and 30 with non-HCC malignancies. To achieve accurate tumor detection, the YOLOX network was employed to identify the region of interest (ROI) on both B-mode ultrasound and CEUS images. A custom-developed algorithm was then utilized to extract brightness change curves from the tumor and adjacent liver parenchyma regions within the CEUS images. These curves provided critical data for the subsequent analysis and classification process. To analyze the extracted brightness change curves and classify the malignancies, we developed and compared several models. These included one-dimensional convolutional neural networks (1D-ResNet, 1D-ConvNeXt, and 1D-CNN), as well as traditional machine-learning methods such as support vector machine (SVM), ensemble learning (EL), k-nearest neighbor (KNN), and decision tree (DT). The diagnostic performance of each method in distinguishing HCC from non-HCC malignancies was rigorously evaluated using four key metrics: area under the receiver operating characteristic (AUC), accuracy (ACC), sensitivity (SE), and specificity (SP). ResultsThe evaluation of the machine-learning methods revealed AUC values of 0.70 for SVM, 0.56 for ensemble learning, 0.63 for KNN, and 0.72 for the decision tree. These results indicated moderate to fair performance in classifying the malignancies based on the brightness change curves. In contrast, the deep learning models demonstrated significantly higher AUCs, with 1D-ResNet achieving an AUC of 0.72, 1D-ConvNeXt reaching 0.82, and 1D-CNN obtaining the highest AUC of 0.84. Moreover, under the five-fold cross-validation scheme, the 1D-CNN model outperformed other models in both accuracy and specificity. Specifically, it achieved accuracy improvements of 3.8% to 10.0% and specificity enhancements of 6.6% to 43.3% over competing approaches. The superior performance of the 1D-CNN model highlighted its potential as a powerful tool for accurate classification. ConclusionThe 1D-CNN model proved to be the most effective in differentiating HCC from non-HCC malignancies, surpassing both traditional machine-learning methods and other deep learning models. This study successfully developed a user-friendly and cost-efficient computer-aided diagnostic solution that would significantly enhances radiologists’ diagnostic capabilities. By improving the accuracy and efficiency of clinical decision-making, this tool has the potential to positively impact patient care and outcomes. Future work may focus on further refining the model and exploring its integration with multimodal ultrasound data to maximize its accuracy and applicability.

Definitive treatment of lung cancer with radiotherapy is challenging, as respiratory motion and anatomical changes can increase the risk of severe off-target effects during radiotherapy. Online adaptive radiotherapy (ART) is an evolving approach that enables timely modification of a treatment plan during the interfraction of radiotherapy, in response to physiologic or anatomic variations, aiming to improve the dose distribution for precise targeting and delivery in lung cancer patients. The effectiveness of online ART depends on the seamless integration of multiple components: sufficient quality of linear accelerator-integrated imaging guidance, deformable image registration, automatic recontouring, and efficient quality assurance and workflow. This review summarizes the present status of online ART for lung cancer, including key technologies, as well as the challenges and areas of active research in this field.
Humans ; Lung Neoplasms/radiotherapy* ; Radiotherapy Planning, Computer-Assisted/methods*

OBJECTIVE: To elucidate how spinal manipulative therapy (SMT) exerts its analgesic effects through regulating brain function in lumbar disc herniation (LDH) patients by utilizing resting-state functional magnetic resonance imaging (rs-fMRI). METHODS: From September 2021 to September 2023, we enrolled LDH patients (LDH group, n=31) and age- and sex-matched healthy controls (HCs, n=28). LDH group underwent rs-fMRI at 2 distinct time points (TPs): prior to the initiation of SMT (TP1) and subsequent to the completion of the SMT sessions (TP2). SMT was administered once every other day for 30 min per session, totally 14 treatment sessions over a span of 4 weeks. HCs did not receive SMT treatment and underwent only one fMRI scan. Additionally, participants in LDH group completed clinical questionnaires on pain using the Visual Analog Scale (VAS) and the Japanese Orthopedic Association (JOA) score, whereas HCs did not undergo clinical scale assessments. The effects on the brain were jointly characterized using the amplitude of low-frequency fluctuations (ALFF) and regional homogeneity (ReHo). Correlation analyses were conducted between specific brain regions and clinical scales. RESULTS: Following SMT treatment, pain symptoms in LDH patients were notably alleviated and accompanied by evident activation of effects in the brain. In comparison to TP1, TP2 exhibited the most significant increase in ALFF values for Temporal_Sup_R and the most notable decrease in ALFF values for Paracentral_Lobule_L (voxelwise P<0.005; clusters >30; FDR correction). Additionally, the most substantial enhancement in ReHo values was observed for the Cuneus_R, while the most prominent reduction was noted for the Olfactory_R (voxelwise P<0.005; clusters >30; FDR correction). Moreover, a comparative analysis revealed that, in contrast to HCs, LDH patients at TP1 exhibited the most significant increase in ALFF values for Temporal_Pole_Sup_L and the most notable decrease in ALFF values for Frontal_Mid_L (voxelwise P<0.005; clusters >30; FDR correction). Furthermore, the most significant enhancement in ReHo values was observed for Postcentral_L, while the most prominent reduction was identified for ParaHippocampal_L (voxelwise P<0.005; clusters >30; FDR correction). Notably, correlation analysis with clinical scales revealed a robust positive correlation between the Cuneus_R score and the rate of change in the VAS score (r=0.9333, P<0.0001). CONCLUSIONS Long-term chronic lower back pain in patients with LDH manifests significant activation of the "AUN-DMN-S1-SAN" neural circuitry. The visual network, represented by the Cuneus_R, is highly likely to be a key brain network in which the analgesic efficacy of SMT becomes effective in treating LDH patients. (Trial registration No. NCT06277739).
Humans ; Magnetic Resonance Imaging ; Intervertebral Disc Displacement/diagnostic imaging* ; Male ; Female ; Brain/diagnostic imaging* ; Adult ; Manipulation, Spinal/methods* ; Middle Aged ; Lumbar Vertebrae/physiopathology* ; Pain Management ; Rest ; Case-Control Studies

Debates persist regarding the efficacy and safety of azvudine, particularly its real-world outcomes. This study involved patients aged ≥60 years who were admitted to 25 hospitals in mainland China with confirmed SARS-CoV-2 infection between December 1, 2022, and February 28, 2023. Efficacy outcomes were all-cause mortality during hospitalization, the proportion of patients discharged with recovery, time to nucleic acid-negative conversion (T _NANC), time to symptom improvement (T _SI), and time of hospital stay (T _HS). Safety was also assessed. Among the 5884 participants identified, 1999 received azvudine, and 1999 matched controls were included after exclusion and propensity score matching. Azvudine recipients exhibited lower all-cause mortality compared with controls in the overall population (13.3% vs. 17.1%, RR, 0.78; 95% CI, 0.67-0.90; P = 0.001) and in the severe subgroup (25.7% vs. 33.7%; RR, 0.76; 95% CI, 0.66-0.88; P < 0.001). A higher proportion of patients discharged with recovery, and a shorter T _NANC were associated with azvudine recipients, especially in the severe subgroup. The incidence of adverse events in azvudine recipients was comparable to that in the control group (2.3% vs. 1.7%, P = 0.170). In conclusion, azvudine showed efficacy and safety in older patients hospitalized with COVID-19 during the SARS-CoV-2 omicron wave in China.

Rapid and ultrasensitive detection of pathogen-associated biomarkers is vital for the early diagnosis and therapy of bacterial infections. Herein, we developed a close-packed and ordered Au@AgPt array coupled with a cascade triggering strategy for surface-enhanced Raman scattering (SERS) and colorimetric identification of the Staphylococcus aureus biomarker micrococcal nuclease (MNase) in serum samples. The trimetallic Au@AgPt nanozymes can catalyze the oxidation of 3,3',5,5'-tetramethylbenzidine (TMB) molecules to SERS-enhanced oxidized TMB (oxTMB), accompanied by the color change from colorless to blue. In the presence of S. aureus, the secreted MNase preferentially cut the nucleobase AT-rich regions of DNA sequences on magnetic beads (MBs) to release alkaline phosphatase (ALP), which subsequently mediated the oxTMB reduction for inducing the colorimetric/SERS signal fade away. Using this "on-to-off" triggering strategy, the target S. aureus can be recorded in a wide linear range with a limit of detection of 38 CFU/mL in the colorimetric mode and 6 CFU/mL in the SERS mode. Meanwhile, the MNase-mediated strategy characterized by high specificity and sensitivity successfully discriminated between patients with sepsis (n = 7) and healthy participants (n = 3), as well as monitored the prognostic progression of the disease (n = 2). Overall, benefiting from highly active and dense "hot spot" substrate, MNase-mediated cascade response strategy, and colorimetric/SERS dual-signal output, this methodology will offer a promising avenue for the early diagnosis of S. aureus infection.

OBJECTIVE: Humans are exposed to complex mixtures of environmental chemicals and other factors that can affect their health. Analysis of these mixture exposures presents several key challenges for environmental epidemiology and risk assessment, including high dimensionality, correlated exposure, and subtle individual effects. METHODS: We proposed a novel statistical approach, the generalized functional linear model (GFLM), to analyze the health effects of exposure mixtures. GFLM treats the effect of mixture exposures as a smooth function by reordering exposures based on specific mechanisms and capturing internal correlations to provide a meaningful estimation and interpretation. The robustness and efficiency was evaluated under various scenarios through extensive simulation studies. RESULTS: We applied the GFLM to two datasets from the National Health and Nutrition Examination Survey (NHANES). In the first application, we examined the effects of 37 nutrients on BMI (2011-2016 cycles). The GFLM identified a significant mixture effect, with fiber and fat emerging as the nutrients with the greatest negative and positive effects on BMI, respectively. For the second application, we investigated the association between four pre- and perfluoroalkyl substances (PFAS) and gout risk (2007-2018 cycles). Unlike traditional methods, the GFLM indicated no significant association, demonstrating its robustness to multicollinearity. CONCLUSION GFLM framework is a powerful tool for mixture exposure analysis, offering improved handling of correlated exposures and interpretable results. It demonstrates robust performance across various scenarios and real-world applications, advancing our understanding of complex environmental exposures and their health impacts on environmental epidemiology and toxicology.
Humans ; Environmental Exposure/analysis* ; Linear Models ; Nutrition Surveys ; Environmental Pollutants ; Body Mass Index

Nocardia is an emerging worldwide distributed zoonotic pathogen,and the common infected population is immu-nocompromised.This article reports a case of non-immunodeficient infected with Nocardia brevicatena/paucivorans complex,combined with literature analysis,aiming to improve scientific reference for clinical diagnosis,treatment and prevention of No-cardia.The patient was a 54-year-old male with no chronic diseases,hospitalized due to repeated low back pain and fever dis-comfort,MRI showed lumbar soft tissue and abdominal abscess.The examination of the immune function aspect was normal,conventional pathogen culture was negative.mNGS showed Nocardia brevicatena/paucivorans complex infection.The anti-in-fection regimen of SMZ was safe and effective,and no recurrence was observed after 1 year of follow-up.Although Nocardia is an opportunistic pathogen,immunocompetent people are still at risk of infection and often present with skin and soft tissue le-sions.

Objective:To investigate the clinical application value of a novel filter's retraction hook capture technique of pull-assisted method for the endovascular retrieval of conical inferior vena cava (IVC) filters whose hook attached to the wall.Methods:From January 2020 to December 2024, patients with conical filters whose hook attached to the wall admitted at Beijing Jishuitan Hospital were enrolled consecutively.Results:A total of 46 patients underwent filter retrieval using filter's retraction hook capture technique of pull-assisted method. Among these patients, 39 cases (84.8%) were successful in filter retrieval, with the penetration distance of cranial anchor vertex of 3.3(2.5, 4.4) mm, and 13 (33.3%) filters were deformed. The other 7 cases were unsuccessful, with a penetration distance of cranial anchor vertex of 5.0 (4.3, 5.0) mm, and 6 (85.7%) filters were deformed. There was a statistically significant difference between the two groups ( P<0.05). One case (2.2%) had IVC injury, one case (2.2%) experienced filter fracture, and no symptomatic pulmonary embolism occurred. Logistic regression analysis showed that filter deformation was an independent dangerous factor for filter's retraction. Conclusions:Filter's retraction hook capture technique of pull-assisted method is effective in removing conical filters whose hook attached to the wall, with no symptomatic PE occurring. This method can be considered as a new adjuvant technique for filter retrieval.