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.

Despite significant advances in the field of critical care medicine over the past three decades, veno-arterial extracorporeal membrane oxygenation (V-A ECMO) remains the primary temporary mechanical circulatory support modality for patients with acute severe circulatory failure. With the accumulation of clinical experience and the increasing maturity of operational techniques in V-A ECMO, its technical management—particularly hemodynamic management—has become a key factor influencing patient outcomes. To further improve patient survival, the Chinese Critical Care Ultrasound Study Group, in collaboration with the Hemodynamic Therapy of Critical Care Collaborative Group and the Critical Care Medicine Branch of the China International Exchange and Promotive Association for Medical and Health Care, organized experts in critical care medicine to develop the Consensus on Hemodynamic Management in Adult Veno-Arterial Extracorporeal Membrane Oxygenation (2026 Edition). Based on years of clinical experience and the latest evidence-based medical evidence, this consensus formulates 15 systematic recommendations covering the entire process of V-A ECMO, including initiation, management during support, and weaning, aiming to promote standardized application of hemodynamic management in V-A ECMO.

Objectives: . Idiopathic sudden sensorineural hearing loss (ISSNHL) is a significant cause of hearing loss. Intratympanic steroid injection (ITSI) is commonly used as an initial or salvage treatment; however, the lack of a standardized treatment protocol has resulted in variability in clinical practice. In addition, no efficient prediction model currently exists to support personalized management. Therefore, this study aimed to develop tailored management strategies for ISSNHL using a machine-learning model. Methods: . This retrospective multicenter cohort study was conducted between January 2015 and December 2020, with data analysis performed between January 2021 and March 2024. Patients were selected based on the International Classification of Diseases, 10th Revision criteria for ISSNHL, along with relevant medication and procedure codes. Patients with pure-tone audiogram results not meeting ISSNHL criteria, better initial hearing in the affected ear, an identifiable etiology, no post-treatment audiogram, or delayed treatment (>6 weeks) were excluded. We included 770 patients diagnosed with ISSNHL who received ITSI. The primary outcome was the area under the receiver operating characteristic curve for prediction performance. Recovery status was determined using the last pure-tone audiogram. Modeling was conducted on the Quanta for Medical Care AI platform using five machine-learning algorithms and a nested cross-validation framework, in which feature selection and hyperparameter tuning were performed in the inner folds and model performance was evaluated in the outer folds. Results: . A random forest classifier outperformed the other models in predicting hearing outcomes, achieving an area under the receiver operating characteristic curve of 0.788. Time to ITSI was the most influential treatment-related factor, with ITSI administered within 10 days of hearing loss being associated with better outcomes. This model can be used to provide personalized prognostic estimates under different treatment protocols. Conclusion . The machine-learning-based prediction model facilitates personalized treatment strategies and timely treatment adjustments for ISSNHL, thereby optimizing the likelihood of complete recovery.

To achieve rapid and visual detection of Plasmodium vivax,a detection method based on recombinase polymerase amplification(RPA)technology and lateral flow dipstick(LFD)was established and evaluated.Targeting the conserved sequence of the P.vivax 18S rRNA gene(GenBank:DQ660817.1)as the target sequence,primers and probes were designed with Primer Premier 5,and the P.vivax recombinant plasmid(pUCPv)was constructed as the standard.A sensitive and specific RPA-LFD-based rapid visual detection method for P.vivax nucleic acids was established.The plasmid standard was serially diluted 10-fold to concentrations of 1×103,1×102,1×101,1×10?,and 1×10?1 copies/μL for sensitivity testing.To evaluate specificity,whole blood DNA samples from patients infected with Plasmodium falciparum,Plasmodium malariae,Plasmodium ovale,or Leishmania donovani,as well as healthy participants,were tested by RPA-LFD.Additionally,The assay′s accuracy was evaluated by testing whole blood DNA samples from 24 confirmed P.vivax-infected patients.This study successfully established a sensitive,specific,and rapid visual RPA-LFD method for detecting P.vivax nucleic acids.The assay can complete P.vivax detection within 20 minutes under isothermal conditions at 39 ℃,achieving a sensitivity of 1 copy/μL.There is no significant cross reaction with parasites such as other Plasmodium species and L.donovani,and the specificity is 100%.All 24 DNA samples from confirmed P.vivax patients were detected,showing a 100%detection rate.The developed RPA-LFD assay exhibits excellent sensitivity and specificity,requires only simple heating equipment,and is user-friendly.This rapid visual detection method is particularly suitable for P.vivax screening in low-resource settings.

Objective:To evaluate the safety and efficacy of immunoadsorption with protein A column(PA-IA) in refractory anti-N-methyl-D-aspartate receptor (NMDAR) encephalitis in children.Methods:The clinical data of 8 children with anti-NMDAR encephalitis who failed to receive first-line treatment from January 2022 to June 2023 in PICU of Xi'an Children's Hospital were retrospectively collected.Clinical features,modified Rankin score (mRS),anti-NMDAR antibody titers in serum and cerebrospinal fluid(CSF),cells count in CSF,protein level in CSF,serum IgG,complement C3,C4 level were analyzed before and after immunoadsorption therapy. All adverse events that occurred during treatment were recorded in detail.Result:Among the 8 children,there were 5 males and 3 females,aged 5 to 12 years old,and each patient was immunized 6 to 10 times,and 75 times of PA-IA treatment were performed. The mRS score decreased in 8 children after PA-IA treatment. Blood IgG level,CSF cell number,protein level in CSF,blood complement C3 and C4 levels,and antibody titers related in blood and CSF were all lower than those before PA-IA treatment. The differences were all statistically significant ( P<0.001).During 75 times of treatment,only 1 case of membrane rupture occurred during treatment,the treatment was stopped and continued after the plasma separator was replaced,with no other adverse reactions. Conclusion:PA-IA could effectively improve the clinical symptoms and neurological function of children with refractory anti-NMDAR encephalitis with good safety,but its exact efficacy in children with anti-NMDAR encephalitis needs to be verified by prospective studies with larger samples.

Objective To investigate the effect of different decompression methods on the efficacy of the unilateral biportal endoscopy(UBE)in the treatment of severe lumbar spinal stenosis(LSS).Methods A total of 203 elderly patients with LSS who underwent UBE treatment in our hospital from September 2021 to June 2024 were selected and divided into the unilateral group(102 cases)and the bilateral group(101 cases)according to different decompression methods.The unilateral group underwent unilateral decompression during operation,while the bilateral group underwent bilateral decompression(with bilateral decompression through the same incision and unilateral approach)during operation.The perioperative conditions,anatomical image parameters,visual analogue scale(VAS)scores,Oswestry disability index(ODI),and complications were compared between the two groups.The rehabilitation progress and complications of patients with different Charlson comorbidity index(CCI)values were compared.Results The operation time of patients in the bilateral group was longer than that in the unilateral group(P<0.05).The dural sac area and lumbar spinal canal area of patients in the bilateral group were larger than those in the unilateral group immediately after operation and 3 months after operation(P<0.05).The VAS scores and ODI 7 days and 3 months after operation of patients in the bilateral group were lower than those in the unilateral group(P<0.05).There was no statistically significant difference in the total incidence of complications between the two groups of patients(P>0.05).The first time of getting out of bed after operation and hospital stay in patients with CCI value>2 scores were longer than those in patients with CCI value≤2 scores(P<0.05).Conclusion Compared with unilateral decompression,bilateral decompression for elderly patients with severe LSS takes a longer surgical time,but it can better restore lumbar shape,alleviate pain,improve lumbar function,and has higher safety,and the milder the patient's condition,the faster the postoperative recovery.

Objective:To analyze the efficacy of transjugular intrahepatic portosystemic shunt (TIPS) combined with main splenic artery embolization in the treatment of patients with portal hypertension upper gastrointestinal bleeding complicated with extensive portal vein thrombosis (PVT).Methods:This study was a prospective, single-center, open-label, single-arm clinical trial. In the first phase, 81 patients with portal hypertension upper gastrointestinal bleeding who were admitted to Beijing Shijitan Hospital, Capital Medical University from January 2018 to December 2018 were consecutively enrolled, including 57 males and 24 females, with the age of (51.3±10.4) years. During TIPS surgery, the pressure of the portal vein before and after the balloon blocking the splenic artery was measured to clarify the contribution of the splenic artery to portal hypertension. In the second stage, from January 2019 to December 2022, 104 patients with portal hypertension upper gastrointestinal bleeding complicated with extensive PVT were re-enrolled, including 71 males and 33 females, with the age of (50.9±12.5) years. TIPS combined with main splenic artery embolization was performed, and portal vein pressure was measured before and after embolization. Follow up on the postoperative esophageal and gastric varices of the patients in the second stage.Results:The portal vein pressures before and after the first stage of balloon occlusion of the splenic artery were (35.2±8.4) mmHg (1 mmHg=0.133 kPa) and (24.2±6.3) mmHg, respectively. The pressure after occlusion was lower than that before occlusion, and the difference was statistically significant ( t=10.54, P<0.001). The portal vein pressures before and after the second stage embolization were (36.1±9.5) mmHg and (21.1±4.7) mmHg respectively. The pressure after embolization was lower than that before embolization, and the difference was statistically significant ( t=13.47, P<0.001). In the second stage, among the 104 patients, the proportion of those whose varicose veins disappeared or improved 6 months after the operation was 43.3%(45/104) and 51.0%(53/104), respectively. There were no patients with aggravation or rebleeding due to rupture. One year later, 8 patients (7.7%) had aggravated or ruptured esophageal and gastric varices with bleeding. Two years later, 12 patients (11.5%) had aggravated or bleeding. Conclusion:TIPS combined with main splenic artery embolization can effectively reduce the portal vein pressure in patients with portal hypertension upper gastrointestinal bleeding complicated with extensive PVT, improve the degree of esophageal and gastric varices, and reduce the risk of gastrointestinal bleeding.

A mounting body of research suggests that circRNAs significantly contribute to the develop-ment of myocardial fibrosis.The microarray results of human circular RNA expression profile indicated that circHERC4_041 expression increased in the myocardium of patients with heart failure,RT-qPCR a-nalysis confirmed that the myocardial expression level of circHERC4_041 in individuals with heart failure were considerably elevated compared to that in healthy organ donors.Fluorescence in situ hybridization(FISH)confirmed that circHERC4_041 was abundant in the cytoplasm of human cardiomyocyte AC16.Overexpression of circHERC4_041 in mouse myocardial fibroblasts(mCFs)mediated by adenovirus in-hibited the expression of fibrosis-related proteins in mCFs.Experiments involving cell proliferation,wound healing,and Transwell assays demonstrated that overexpression of circHERC4_041 suppressed the growth and mobility of mCFs(P＜0.001).Sequence analysis results suggested that circHERC4_041 con-tains potential ribosome entry sequence(IRES)and open reading frame(ORF).Western blot confirmed that circHERC4_041 could translate the 516 amino acid HERC4-516aa protein,which was mainly located in the cytoplasm of the cell.Cell functional experiments confirmed that circHERC4_041 inhibited the fi-brotic phenotype of mCFs by specifically translating HERC4-516aa(P＜0.05).The specific interaction between HERC4-516aa and transglutaminase 2(TGM2)was confirmed by IP-MS screening and Co-IP i-dentification.Further results found that the degradation of TGM2 was promoted through proteasome path-way.The overexpression of TGM2 in mCFs facilitated by adenoviral vectors could counteract the suppres-sive effects of HERC4-516aa on the fibrotic phenotype of mCFs.Therefore,this study confirmed that the HERC4-516aa protein translated by circHERC4_041 can specifically bind to TGM2 to inhibit the fibrotic phenotype of myocardial fibroblasts.

Aim To investigate the effect of linarin on improving cognitive behavior of APP/PS1 mice,and to explore the therapeutic effect of linarin on A β deposi-tion and neuroinflammation and its correlation.Meth-ods APP/PS1 transgenic mice were randomly divid-ed into the model group,high-dose group,medium-dose group,low-dose group and positive control group.C57BL/6J mice were set as the normal group.Morris water maze was used to evaluate the learning and mem-ory abilities of mice.TUNEL staining was used to de-tect the apoptosis of neurons in the CA1 region of mice.IHC was used to detect the expression levels of Aβ42 and GFAP.Western blot was used to detect the expression levels of BACE1 and PS-1.Results Com-pared with the normal group,mice of the model group showed lower NCP,shorter target quadrant travel,less target quadrant residence time percentage(all P＜0.01),higher apoptosis rate of neurons in the CA1 re-gion(P＜0.01),significantly higher protein expres-sion levels of A β42 and GFAP(all P＜0.01),and significantly higher protein expression levels of BACE1 and PS-1(all P＜0.01).Compared with the model group,the medium-dose group,high-dose group and positive control group showed higher NCP,longer tar-get quadrant travel,more target quadrant residence time percentage(all P＜0.05),lower apoptosis rate of neurons in the CA1 region(P＜0.01),significantly lower protein expression levels of A β42 and GFAP(all P＜0.01),and significantly lower protein expression levels of BACE1 and PS-1(all P＜0.01).Conclu-sions Linarin can inhibit two key enzymes to reduce the decomposition of APP and the generation of A β42,thereby inhibiting the activation of astrocytes,allevia-ting neuroinflammation,improving the core pathologi-cal features of AD,and thus significantly improving learning and memory impairment in APP/PS1 mice.