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.

This article reports a child with cardioaciocutaneous syndrome (CFCS) caused by a rare microdeletion of chromosome 19p13.3, and a literature review is conducted. The child had unusual facies, short stature, delayed mental and motor development, macrocephaly, and cardiac abnormalities. Whole-exome sequencing identified a 1 040 kb heterozygous deletion in the 19p13.3 region of the child, which was rated as a "pathogenic variant". This is the first case of CFCS caused by a loss-of-function mutation reported in China, which enriches the genotype characteristics of CFCS. It is imperative to enhance the understanding of CFCS in children. Early identification based on its clinical manifestations should be pursued, and genetic testing should be performed to facilitate diagnosis.
Humans ; Chromosome Deletion ; Chromosomes, Human, Pair 19/genetics* ; Ectodermal Dysplasia/genetics* ; Facies ; Failure to Thrive/genetics* ; Heart Defects, Congenital/genetics*

OBJECTIVE: To explore the clinical characteristics and prognosis of children with SIL-TAL1-positive T-cell acute lymphoblastic leukemia ( SIL-TAL1⁺ T-ALL). METHODS: The clinical data of 110 children with newly diagnosed T-ALL admitted to the pediatric department of our hospital from January 2010 to December 2018 were reviewed to compare the clinical characteristics, treatment response and prognosis between SIL-TAL1⁺ group and SIL-TAL1^-group. RESULTS: Among the 110 children with T-ALL, 25 cases (22.7%) were in the SIL-TAL1⁺ group and 85 cases (77.3%) in the SIL-TAL1^- group. The white blood cell (WBC) count in the SIL-TAL1⁺ group was significantly higher than that in the SIL-TAL1^- group (P < 0.05), while the other clinical characteristics and treatment response were not significantly different between the two groups. The 5-year overall survival (OS) rates of SIL-TAL1⁺ group and SIL-TAL1^- group were 80.0% and 75.5%, and 5-year disease-free survival (DFS) rates were 76.0% and 72.9%, respectively. There were no significant differences in OS rate and DFS rate between the two groups ( P >0.05). In children aged < 10 years, the 5-year OS rate of SIL-TAL1⁺ group and SIL-TAL1^- group was 100% and 75.1%, respectively, and the difference between the two groups was statistically significant (P < 0.05). CONCLUSION Although the WBC level is significantly higher in children with SIL-TAL1⁺ T-ALL than that in those with SIL-TAL1^- T-ALL, the treatment efficacy is similar between the two groups. In children aged < 10 years, the longterm survival rate is superior in the SIL-TAL1⁺ group.
Humans ; Precursor T-Cell Lymphoblastic Leukemia-Lymphoma/diagnosis* ; Prognosis ; Child ; Male ; Female ; Survival Rate ; T-Cell Acute Lymphocytic Leukemia Protein 1 ; Child, Preschool ; Oncogene Proteins, Fusion ; Leukocyte Count

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 efficacy and factors influencing treatment of pediatric noninfectious uveitis with Adalimumab (ADA).Methods:A retrospective clinical study. A total of 86 pediatric patients with non-infectious uveitis, diagnosed and treated with ADA at Department of Uveitis Specialist of Xi'an People's Hospital (Xi' an Fourth Hospital) from January 1, 2021 to December 31, 2023, were included in this study. The age of all patients was ≤16 years. Among them, 55 (63.95%, 55/86) patients received ADA combined with one immunosuppressive agent, 28 (32.56%, 28/86) patients received ADA combined with ≥2 immunosuppressive agents, and 3 (3.49%, 3/86) patients received ADA alone without any immunosuppressive agents. All patients underwent best-corrected visual acuity (BCVA) and optical coherence tomography (OCT) examinations. The thickness of the retinal nerve fiber layer (RNFL) in the macular region was measured using an OCT device. The cumulative treatment effectiveness rate at 12 months post-treatment was evaluated using the Kaplan-Meier survival analysis. Multivariate analysis was performed using the Cox proportional hazards regression model, and the optimal predictive model was selected based on the Bayesian information criterion. The association between different treatment regimens and various clinical outcomes was assessed.Results:Among the 86 pediatric patients, 42 were male and 44 were female, with a mean age of (10.47±3.23) years. The distribution of uveitis types was as follows: anterior uveitis in 37 cases, intermediate uveitis in 15 cases, posterior uveitis in 10 cases, and panuveitis in 24 cases. Anterior chamber cells (ACC), keratic precipitates, and synechiae were present in 66, 55, and 38 cases, respectively. The cumulative treatment effectiveness at 12 months was 85.1% [95% confidence interval ( CI) 71.9-92.2], with a median time to treatment effectiveness of 3 months. Compared with baseline, after 6 months of treatment, the BCVA, RNFL thickness ( Z=?6.323, ?8.017), and the grading of ACC and vitreous haze ( χ2= ?6.917, ?5.027) showed significant improvement, with statistically significant differences ( P<0.05). Multivariate analysis revealed that ACC (hazard ratio=22.31, 95% CI 2.43-204.68) and anterior uveitis (hazard ratio=3.88, 95% CI 2.03-7.42) were significantly associated with treatment effectiveness ( P<0.05). Patients with ACC had a median time to treatment effectiveness of 2 months, with a 12-month cumulative treatment effectiveness of 95.5% (95% CI 86.3-98.5). Patients with anterior uveitis had a median time to treatment effectiveness of 2 months, with a 12-month cumulative treatment effectiveness of 97.3% (95% CI 81.3-99.6). Patients without anterior uveitis had a median time to treatment effectiveness of 5 months, with a 12-month cumulative treatment effectiveness of 76.7% (95% CI 54.1-88.2). The cumulative recurrence risk at 12 months was 15.6% (95% CI 6.2-24.1). Conclusion:ADA is safe and effective in treating pediatric non-infectious uveitis, and ACC and anterior uveitis are associated with response rate.

Objective:To discern the association between autism-like behaviors and resilience within the ado-lescent demographic.Methods:A total of 7 063 middle school students were selected to assess ASD-like behaviors and resilience in adolescents using the Autism Spectrum Screening Questionnaire(ASSQ)as well as the Resilience Scale for Chinese Adolescent(RSCA).Subgroups bounded by P5 and P95 of the total ASSQ score,a comparative analysis of the resilience scores between these groups was executed.A correlation evaluation and linear regression a-nalysis was carried out between ASSQ and RSCA scores from all participants.Results:The RSCA scores within the high ASSQ scoring group,were inferior to those in the low scoring group.ASSQ scores were negatively correlated with RSCA scores for the full sample(Ps＜0.01);Social interaction scores on the ASSQ were negatively correlated with the five-factor RSCA scores(β=-0.23,-0.27,-0.11,-0.23,-0.37,Ps＜0.05).Conclusion:There was a negative association between autism spectrum disorder-like behaviors and resilience in adolescents,with more severe social interaction symptoms being associated with poorer resilience.

Objective:To discern the association between autism-like behaviors and resilience within the ado-lescent demographic.Methods:A total of 7 063 middle school students were selected to assess ASD-like behaviors and resilience in adolescents using the Autism Spectrum Screening Questionnaire(ASSQ)as well as the Resilience Scale for Chinese Adolescent(RSCA).Subgroups bounded by P5 and P95 of the total ASSQ score,a comparative analysis of the resilience scores between these groups was executed.A correlation evaluation and linear regression a-nalysis was carried out between ASSQ and RSCA scores from all participants.Results:The RSCA scores within the high ASSQ scoring group,were inferior to those in the low scoring group.ASSQ scores were negatively correlated with RSCA scores for the full sample(Ps＜0.01);Social interaction scores on the ASSQ were negatively correlated with the five-factor RSCA scores(β=-0.23,-0.27,-0.11,-0.23,-0.37,Ps＜0.05).Conclusion:There was a negative association between autism spectrum disorder-like behaviors and resilience in adolescents,with more severe social interaction symptoms being associated with poorer resilience.

Objective:To investigate the clinical efficacy and factors influencing treatment of pediatric noninfectious uveitis with Adalimumab (ADA).Methods:A retrospective clinical study. A total of 86 pediatric patients with non-infectious uveitis, diagnosed and treated with ADA at Department of Uveitis Specialist of Xi'an People's Hospital (Xi' an Fourth Hospital) from January 1, 2021 to December 31, 2023, were included in this study. The age of all patients was ≤16 years. Among them, 55 (63.95%, 55/86) patients received ADA combined with one immunosuppressive agent, 28 (32.56%, 28/86) patients received ADA combined with ≥2 immunosuppressive agents, and 3 (3.49%, 3/86) patients received ADA alone without any immunosuppressive agents. All patients underwent best-corrected visual acuity (BCVA) and optical coherence tomography (OCT) examinations. The thickness of the retinal nerve fiber layer (RNFL) in the macular region was measured using an OCT device. The cumulative treatment effectiveness rate at 12 months post-treatment was evaluated using the Kaplan-Meier survival analysis. Multivariate analysis was performed using the Cox proportional hazards regression model, and the optimal predictive model was selected based on the Bayesian information criterion. The association between different treatment regimens and various clinical outcomes was assessed.Results:Among the 86 pediatric patients, 42 were male and 44 were female, with a mean age of (10.47±3.23) years. The distribution of uveitis types was as follows: anterior uveitis in 37 cases, intermediate uveitis in 15 cases, posterior uveitis in 10 cases, and panuveitis in 24 cases. Anterior chamber cells (ACC), keratic precipitates, and synechiae were present in 66, 55, and 38 cases, respectively. The cumulative treatment effectiveness at 12 months was 85.1% [95% confidence interval ( CI) 71.9-92.2], with a median time to treatment effectiveness of 3 months. Compared with baseline, after 6 months of treatment, the BCVA, RNFL thickness ( Z=?6.323, ?8.017), and the grading of ACC and vitreous haze ( χ2= ?6.917, ?5.027) showed significant improvement, with statistically significant differences ( P<0.05). Multivariate analysis revealed that ACC (hazard ratio=22.31, 95% CI 2.43-204.68) and anterior uveitis (hazard ratio=3.88, 95% CI 2.03-7.42) were significantly associated with treatment effectiveness ( P<0.05). Patients with ACC had a median time to treatment effectiveness of 2 months, with a 12-month cumulative treatment effectiveness of 95.5% (95% CI 86.3-98.5). Patients with anterior uveitis had a median time to treatment effectiveness of 2 months, with a 12-month cumulative treatment effectiveness of 97.3% (95% CI 81.3-99.6). Patients without anterior uveitis had a median time to treatment effectiveness of 5 months, with a 12-month cumulative treatment effectiveness of 76.7% (95% CI 54.1-88.2). The cumulative recurrence risk at 12 months was 15.6% (95% CI 6.2-24.1). Conclusion:ADA is safe and effective in treating pediatric non-infectious uveitis, and ACC and anterior uveitis are associated with response rate.

As artificial intelligence technology rapidly advances, its deployment within the medical sector presents substantial ethical challenges. Consequently, it becomes crucial to create a standardized, transparent, and secure framework for processing medical data. This includes setting the ethical boundaries for medical artificial intelligence and safeguarding both patient rights and data integrity. This consensus governs every facet of medical data handling through artificial intelligence, encompassing data gathering, processing, storage, transmission, utilization, and sharing. Its purpose is to ensure the management of medical data adheres to ethical standards and legal requirements, while safeguarding patient privacy and data security. Concurrently, the principles of compliance with the law, patient privacy respect, patient interest protection, and safety and reliability are underscored. Key issues such as informed consent, data usage, intellectual property protection, conflict of interest, and benefit sharing are examined in depth. The enactment of this expert consensus is intended to foster the profound integration and sustainable advancement of artificial intelligence within the medical domain, while simultaneously ensuring that artificial intelligence adheres strictly to the relevant ethical norms and legal frameworks during the processing of medical data.
Artificial Intelligence/legislation & jurisprudence* ; Humans ; Consensus ; Computer Security/standards* ; Confidentiality/ethics* ; Informed Consent/ethics*