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.

OBJECTIVE: To investigate the clinical characteristics and prognosis of peripheral T-cell lymphoma, not otherwise specified (PTCL-NOS). METHODS: Clinical data of 10 patients with PTCL-NOS in Gansu Provincial Hospital from May 2016 to June 2023 were collected. The treatment outcomes were evaluated, and the factors affecting prognosis were analyzed. RESULTS: The median age of onset for the 10 patients was 60.7 (47-75) years, with 7 males and 3 females. Nine cases received chemotherapy, while one case died suddenly after diagnosis, and the median course of chemotherapy was 6.9 (1-13) courses. Assessing the efficacy, 3 patients achieved complete remission (CR) while 7 patients showed progression. Age, sex, lactate dehydrogenase (LDH) level, Ki-67 and the presence of hemophagocytic lymphohistocytosis (HLH) were not statistically correlated with CR rate ( P >0.05). Patients with IPI score 3-5, and Ann Arbor stage III-IV had statistically lower CR rates (both P <0.05). Age, B symptoms, LDH level ,hemoglobin, Ki-67 index and PLR value were not statistically correlated with overall survival (OS) time ( P >0.05). Male, platelet <150×10⁹/L, IPI score 3-5, Ann Arbor stage III-IV, presence of HLH, NLR≥4.05, and LMR <2.81 were statistically correlated with shorter OS (all P <0.05). Among the 10 patients, 3 cases have survived and are still in CR status, while 7 cases have died, with a median survival time of 7.5 (1-85) months. CONCLUSIONS Patients with IPI score 3-5 and Ann Arbor stage III-IV have low CR rate and poor prognosis. The OS of patients who are male, with platelet <150×10⁹/L, IPI score 3-5, Ann Arbor stage III-IV, complication of HLH, NLR≥4.05, and LMR <2.81 is short, and prognosis is poor.
Humans ; Lymphoma, T-Cell, Peripheral/diagnosis* ; Male ; Prognosis ; Middle Aged ; Female ; Aged

OBJECTIVE: To analyze the clinical features of acute myeloid leukemia patients with DEK-NUP214 fusion gene positive. METHODS: The DEK-NUP214 fusion gene was amplified by multi-nested PCR in 26 patients admitted to the First Affiliated Hospital of Soochow University from January 2018 to October 2023, and the disease course and post-transplant survival data were obtained by searching outpatient and inpatient medical records and telephone follow-up. RESULTS: The median follow-up time of pateints was 21.25（0.9-60.2） months. Among 26 patients with DEK-NUP214 fusion gene positive AML, 15 patients had FLT3-ITD gene mutation positive. One patient died after abandoning treatment due to non-remission of induction chemotherapy, one died due to infection, and 23 patients received allo-HSCT after achieving CR, of which one patient died within one month after transplantation due to multiple infections and one died due to severe pulmonary infection that did not respond to treatment. One patient received allo-HSCT in non-remission state and later died due to recurrence. CONCLUSION DEK-NUP214 fusion gene positive AML is a type of acute leukemia subtype with high risk and poor prognosis. Allo-HSCT treatment at the early stage of disease remission is the most effective way to improve the prognosis of patients.
Humans ; Leukemia, Myeloid, Acute/genetics* ; Poly-ADP-Ribose Binding Proteins ; Oncogene Proteins, Fusion/genetics* ; Nuclear Pore Complex Proteins/genetics* ; Oncogene Proteins/genetics* ; Chromosomal Proteins, Non-Histone/genetics* ; Male ; Female ; Adult ; Mutation ; Hematopoietic Stem Cell Transplantation ; Middle Aged

OBJECTIVE: To investigate the predictive value of the prognostic nutritional index (PNI) and systemic inflammatory response index (SIRI) for short-term efficacy and prognosis in newly treated patients with peripheral T-cell lymphoma (PTCL). METHODS: The general data, laboratory indicators, disease stage and other clinical data of 91 newly treated PTCL patients admitted to the Affiliated Hospital of Xuzhou Medical University from January 2015 to December 2023 were retrospectively analyzed. The optimal cutoff values for PNI and SIRI were determined using receiver operating characteristic (ROC) curves, and the patients were stratified into groups based on these cutoffs to compare clinical features and short-term efficacy between the different groups. Kaplan-Meier method was used to plot survival curves, and univariate and multivariate analyses were performed to identify the factors affecting overall survival (OS). RESULTS: The optimal cutoff values for PNI and SIRI were 45.30 and 1.74×10⁹/L, respectively. Patients in different PNI groups showed statistically significant differences in age, Ann Arbor stage, lactate dehydrogenase (LDH) level, international prognostic index (IPI), prognostic index for PTCL-not otherwise specified (PIT), pathological subtypes, and complete response (CR) rate (P < 0.05). PTCL patients in different SIRI groups exhibited significant differences in Ann Arbor stage, LDH level, IPI score, PIT score, and CR rate (P < 0.05). Logistic regression analysis showed that age ≥60 years old (OR =2.750), Ann Arbor stage Ⅲ-Ⅳ (OR =5.200), IPI score ≥2 (OR =7.650), low PNI (OR =3.296), and high SIRI (OR =3.130) were independent risk factors affecting treatment efficacy in PTCL patients (P < 0.05). Cox proportional hazards regression model analysis showed that low PNI and elevated β₂-microglobulin (β₂-MG) levels were independent risk factors affecting OS (P < 0.05). CONCLUSION PNI and SIRI have certain application value in evaluating short-term efficacy and prognosis in patients with PTCL. Compared with SIRI, PNI demonstrates greater predictive value for patient prognosis.
Humans ; Prognosis ; Lymphoma, T-Cell, Peripheral/therapy* ; Retrospective Studies ; Nutrition Assessment ; Male ; Female ; Middle Aged ; ROC Curve ; Inflammation

OBJECTIVE: To analyze the Epstein-Barr virus (EBV) load in different lymphocyte subsets, as well as clinical characteristics and outcomes in patients with hematologic malignancies experiencing EBV reactivation. METHODS: Peripheral blood samples from patients were collected. B, T, and NK cells were isolated sorting with magnetic beads by flow cytometry. The EBV load in each subset was quantitated by real-time quantitative polymerase chain reaction (RT-qPCR). Clinical data were colleted from electronic medical records. Survival status was followed up through outpatient visits and telephone calls. Statistical analyses were performed using SPSS 25.0. RESULTS: A total of 39 patients with hematologic malignancies were included, among whom 35 patients had undergone allogeneic hematopoietic stem cell transplantation (allo-HSCT). The median time to EBV reactivation was 4.8 months (range: 1.7-57.1 months) after allo-HSCT. EBV was detected in B, T, and NK cells in 20 patients, in B and T cells in 11 patients, and only in B cells in 4 patients. In the 35 patients, the median EBV load in B cells was 2.19×10⁴ copies/ml, significantly higher than that in T cells (4.00×10³ copies/ml, P <0.01) and NK cells (2.85×10² copies/ml, P <0.01). Rituximab (RTX) was administered for 32 patients, resulting in EBV negativity in 32 patients with a median time of 8 days (range: 2-39 days). Post-treatment analysis of 13 patients showed EBV were all negative in B, T, and NK cells. In the four non-transplant patients, the median time to EBV reactivation was 35 days (range: 1-328 days) after diagnosis of the primary disease. EBV was detected in one or two subsets of B, T, or NK cells, but not simultaneously in all three subsets. These patients received a combination chemotherapy targeting at the primary disease, with 3 patients achieving EBV negativity, and the median time to be negative was 40 days (range: 13-75 days). CONCLUSION In hematologic malignancy patients after allo-HSCT, EBV reactivation commonly involves B, T, and NK cells, with a significantly higher viral load in B cells compared to T and NK cells. Rituximab is effective for EBV clearance. In non-transplant patients, EBV reactivation is restricted to one or two lymphocyte subsets, and clearance is slower, highlighting the need for prompt anti-tumor therapy.
Humans ; Hematologic Neoplasms/virology* ; Herpesvirus 4, Human/physiology* ; Epstein-Barr Virus Infections ; Hematopoietic Stem Cell Transplantation ; Virus Activation ; Lymphocyte Subsets/virology* ; Flow Cytometry ; Killer Cells, Natural/virology* ; Male ; Female ; B-Lymphocytes/virology* ; Viral Load ; Adult ; T-Lymphocytes/virology* ; Middle Aged

OBJECTIVE: This study aimed to explore the association between body mass index (BMI) and mortality based on the 10-year population-based multicenter prospective study. METHODS: A general population-based multicenter prospective study was conducted at four sites in rural China between 2013 and 2023. Multivariate Cox proportional hazards models and restricted cubic spline analyses were used to assess the association between BMI and mortality. Stratified analyses were performed based on the individual characteristics of the participants. RESULTS: Overall, 19,107 participants with a sum of 163,095 person-years were included and 1,910 participants died. The underweight (< 18.5 kg/m ²) presented an increase in all-cause mortality (adjusted hazards ratio [ aHR] = 2.00, 95% confidence interval [ CI]: 1.66-2.41), while overweight (≥ 24.0 to < 28.0 kg/m ²) and obesity (≥ 28.0 kg/m ²) presented a decrease with an aHR of 0.61 (95% CI: 0.52-0.73) and 0.51 (95% CI: 0.37-0.70), respectively. Overweight ( aHR = 0.76, 95% CI: 0.67-0.86) and mild obesity ( aHR = 0.72, 95% CI: 0.59-0.87) had a positive impact on mortality in people older than 60 years. All-cause mortality decreased rapidly until reaching a BMI of 25.7 kg/m ² ( aHR = 0.95, 95% CI: 0.92-0.98) and increased slightly above that value, indicating a U-shaped association. The beneficial impact of being overweight on mortality was robust in most subgroups and sensitivity analyses. CONCLUSION This study provides additional evidence that overweight and mild obesity may be inversely related to the risk of death in individuals older than 60 years. Therefore, it is essential to consider age differences when formulating health and weight management strategies.
Humans ; Body Mass Index ; China/epidemiology* ; Male ; Female ; Middle Aged ; Prospective Studies ; Rural Population/statistics & numerical data* ; Aged ; Follow-Up Studies ; Adult ; Mortality ; Cause of Death ; Obesity/mortality* ; Overweight/mortality*

Objective To develop an enteral nutrition delivery device for nasal feeding patients to relieve gastrointestinal discomfort during enteral nutrition therapy and solve the problems in heating,quality preservation and constant temperature maintenance of the solution.Methods The enteral nutrition delivery device was mainly composed of a nasal feeding container,a fixing frame,a limiting trough,a heating wire,a temperature control unit and a speed control unit.The heating wire and temperature and speed control units were combined to facilitate the nasogastric solution in the container to be delivered to the patient's stomach and intestines at a constant temperature and a uniform speed.The device developed had its shell made of 304 waterproof stainless steel and the nasal feeding container made of 05-polypropylene.Results The device developed realized constant-temperature and uniform-speed delivery and auto cleaning,and could reduced effectively gastrointestinal adverse reactions during nasal feeding.Conclusion The device developed gains advantages in easy operation and autonomous feeding,and thus is worth promoting for nasal feeding patients clinically.[Chinese Medical Equipment Journal,2025,46(2):108-111]

Integrated metabolomic and lipidomic profiling,utilizing liquid chromatography coupled with high-resolution mass spectrometry(LC-HRMS),has emerged as a pivotal strategy for biomarker discovery.However,the inherent polarity disparity between metabolites and lipids complicates simultaneous analysis.To address this,a dual-stationary phase polarity-extended liquid chromatography(PELC)system was developed,which surpassed conventional one-dimensional LC(1D-LC)by enabling comprehensive coverage of both polar and non-polar compounds within a single injection.This system enhanced chromatographic resolution,peak capacity,and throughput while minimizing analytical variability.Extracellular vesicles(EVs),lipid bilayer-enclosed nanoparticles ubiquitously present in biofluids,had gained prominence as reservoirs of cancer biomarkers due to their cargo stability and pathophysiological relevance.Herein,the application of PELC-HRMS for concurrent metabolome-lipidome profiling in EVs was pioneered.A total of 193 metabolites were identified using this technique coupled with MS-DIAL software and Human Metabolome Database.Subsequently,this technique was employed to explore potential biomarkers for prostate cancer(PCa).Multivariate analysis identified 17 differentially abundant metabolites in PCa,implicating dysregulated pathways including purine metabolism,starch and sucrose metabolism,galactose metabolism,cysteine and methionine metabolism,and biosynthesis of unsaturated fatty acids.Notably,creatine(AUC=0.92)and DG 42:5(AUC=0.80)demonstrated robust diagnostic efficacy,attributable to their broad polarity ranges and EV-specific enrichment.This study established PELC as a high-fidelity platform for multi-omics integration in complex biospecimens,advancing mechanistic insights into metabolic rewiring and disease pathophysiology.