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.

The pathogenesis of metabolic dysfunction-associated steatotic liver disease （MASLD） is fundamentally rooted in spleen deficiency and is closely associated with phlegm turbidity， damp-heat and blood stasis. Clinically， liver constraint with spleen deficiency and internal retention of damp turbidity represent the predominant traditional Chinese medicine （TCM） syndrome patterns. Researches have indicated intrinsic connections between the syndrome patterns and biological indicators such as gut microbiota and metabolic profiles. Regarding treatment， classical famous formulas， modern empirical formulas， and newly developed TCM drugs show positive effects in regulating glucose and lipid metabolism， improving insulin resistance， and alleviating metabolic inflammation， exhibiting multi-target mechanisms of action； acupuncture and other external therapies also provide adjunctive value. Nevertheless， current researches still have limitations such as the lack of high-quality clinical evidence and insufficient systematic elucidation of the uncerlying mechanisms. Future efforts should focus on conducting high-quality TCM clinical trials with hard endpoint outcomes such as hepatic histology outcomes， and utilizing modern technologies like multi-omics to elucidate TCM's mechanisms of action， thereby advancing the position of TCM as a first-line therapeutic strategy for MASLD.

Objective To analyze the effect of releasing the lower pulmonary ligament on right residual lung expansion after right upper lobe resection under different body mass index (BMI) levels. Methods The clinical data of patients who underwent thoracoscopic right upper lobe resection in the First Affiliated Hospital with Nanjing Medical University from 2021 to 2022 were retrospectively analyzed. Patients were divided into a group A (17 kg/m2＜BMI≤23 kg/m2), a group B (23 kg/m2＜BMI≤29 kg/m2) and a group C (BMI＞29 kg/m2) according to BMI. The presence of residual cavity was judged by chest X-ray at 7-10 days after operation, the degree of compensation change of the right main bronchus angle was measured, and the changes in lung volume were determined by CT three-dimensional reconstruction. Results A total of 157 patients who underwent thoracoscopic right upper lobe resection were included, including 71 males and 86 females, with an average age of (59.7±11.2) years. There were 50 patients in the group A, 75 patients in the group B, and 32 patients in the group C. In the group A, compared with those without releasing the lower pulmonary ligament, patients with releasing had a lower incidence of postoperative residual cavity (P=0.016), greater changes in bronchus angle (P<0.001), and smaller changes in lung volume (P<0.001). In the group B and C, there was no significant effect of releasing the lower pulmonary ligament on postoperative residual cavity, bronchus angle, and lung volume changes (P>0.05). Conclusion For patients with thin and long body shape and low BMI, releasing the lower pulmonary ligament is helpful to promote the expansion of the residual lung after right upper lobe resection and reduce the occurrence of postoperative residual cavity in patients.

Pulmonary fibrosis（PF） is a progressive and life-threatening disease with limited therapeutic options， highlighting the urgent need for innovative drug discovery strategies. To address this challenge， the authors propose the formula-originated rational intelligent screening&translation（FIRST）， a systematic framework for developing anti-fibrotic monomers derived from classical traditional Chinese medicine（TCM）. The strategy integrates three key dimensions， including tissue-oriented intelligent screening of active compounds， structural optimization based on drug-target spatial interactions and plant biosynthetic pathways， and cross-scale validation of drug. We further highlight its applications in discovering tissue-oriented novel drugs from clinically validated TCM， the development and mechanistic elucidation of anti-fibrotic therapeutics， as well as the clinical translation and secondary development of candidate drugs. This strategy paves the way for first-in-class， formula-derived monomeric drugs with defined structures， clarified mechanisms， and proven safety， offering a transformative avenue to meet the urgent therapeutic needs of PF and setting a new paradigm for TCM-based drug innovation.

Pulmonary fibrosis（PF） is a progressive and life-threatening disease with limited therapeutic options， highlighting the urgent need for innovative drug discovery strategies. To address this challenge， the authors propose the formula-originated rational intelligent screening&translation（FIRST）， a systematic framework for developing anti-fibrotic monomers derived from classical traditional Chinese medicine（TCM）. The strategy integrates three key dimensions， including tissue-oriented intelligent screening of active compounds， structural optimization based on drug-target spatial interactions and plant biosynthetic pathways， and cross-scale validation of drug. We further highlight its applications in discovering tissue-oriented novel drugs from clinically validated TCM， the development and mechanistic elucidation of anti-fibrotic therapeutics， as well as the clinical translation and secondary development of candidate drugs. This strategy paves the way for first-in-class， formula-derived monomeric drugs with defined structures， clarified mechanisms， and proven safety， offering a transformative avenue to meet the urgent therapeutic needs of PF and setting a new paradigm for TCM-based drug innovation.

ObjectiveIn the context of the digital transformation of education, this study aims to construct a triadic interactive teaching model for surgical animal surgery in clinical medicine using modern information technology. It explores the effectiveness of different teaching methods in improving students' practical skills, aseptic awareness, and teamwork abilities, providing a reference for the reform of clinical practice education. MethodsA quasi-experimental research design was adopted. A total of 80 students from the eight-year clinical medicine program at Nanjing University were selected, including the Class of 2020 (control group, n=40) and the Class of 2021 (experimental group, n=40). The control group received traditional teaching methods, while the experimental group implemented the "Teacher-Student-AI" triadic interactive teaching model. This model utilized a smart teaching platform for personalized pre-class preparation , as well as data-driven post-class review and feedback throughout the entire teaching process. The "assessment indicators and scoring criteria for the surgical animal surgery course" were used to evaluate teaching effectiveness, with independent samples t-tests used for statistical analysis. ResultsPre-course assessments revealed no statistically significant differences in baseline theoretical knowledge or practical skills between the two groups (P>0.05). Upon completion of the course, the experimental group achieved higher scores than the control group across three key dimensions: practical skills (47.98±1.34 vs 46.92±2.51, P=0.022), aseptic awareness (17.84±1.16 vs 16.94±2.29, P=0.029), and teamwork (16.82±1.44 vs 15.95±1.22, P=0.004). However, no statistically significant difference was observed in the scores for humane care awareness between the two groups (8.24±0.70 vs 8.16±0.53, P=0.589). ConclusionThe AI-based triadic interactive teaching model can, to some extent, address the limitations of traditional surgical animal surgery education. It plays a positive role in enhancing medical students' surgical skills, aseptic awareness, and collaborative abilities. This model facilitates the transition from traditional to personalized teaching and offers a practical framework for the digital reform of clinical practice education.

OBJECTIVE: To explore the clinical phenotypes and genetic characteristics of five children with Lamb-Shaffer syndrome (LAMSHF). METHODS: Five children with LAMSHF diagnosed at the Department of Pediatrics, the First Medical Center of Chinese PLA General Hospital from April 2021 to December 2024 were selected as study subjects. Clinical data of the children was collected. Genomic DNA was extracted from peripheral blood samples of the children and their parents. Whole exome sequencing (WES) was carried out to screen for variants. This study was approved by the Medical Ethics Committee of the Chinese PLA General Hospital (Ethics No.: S2025-411-01). RESULTS: All five children had presented with global developmental delay. Among them, two had manifestations of autism spectrum disorder, two had abnormal electroencephalogram findings, four had abnormal MRI results, and two had ocular abnormalities. WES has detected five novel variants in the SOX5 gene. Among these, c.1771G>C (p.Gly591Arg) was unreported previously. Sanger sequencing confirmed that none of the parents had carried the same variants, suggesting that they were all de novo variants. According to the guidelines from the American College of Medical Genetics and Genomics (ACMG), two nonsense variants and one missense variant were classified as pathogenic, whilst two missense variants were classified as likely pathogenic. CONCLUSION This study has clarified the correlation between the clinical phenotypes of five children with LAMSHF and variants of the SOX5 gene, which expanded the mutational spectrum of the SOX5 gene and provided a basis for the clinical diagnosis and genetic counseling.
Humans ; Male ; Female ; Phenotype ; Child, Preschool ; Child ; SOXD Transcription Factors/genetics* ; Exome Sequencing ; Mutation ; Infant

AIM To study the chemical constituents from Gymnema tingens Spreng.and their in vitro hypoglycemic activity.METHODS The 70％ethanol extract was isolated and purified by macroporous resin,silica gel,sephadex LH-20,and semi-preparative HPLC,then the structures of obtained compounds were identified by physicochemical propeties and spectral data.The in vitro hypoglycemic activity was evaluated by glucose uptake test in L6 cells.RESULTS Seventeen compounds were isolated and identified as 7-desoxyneocynapanogenin A(1),glaucogenin(2),cynatratoside A(3),atratcynoside F(4),(+)-lyoniresinol(5),(+)-lyoniresinol 3-O-α-D-rhamnopyranoside-(1→6)-β-D-glucopyranoside(6),fernandoside(7),3,4-dimethoxy-phenyl-1-O-β-D-apiofuranosyl-(1→2)-β-D-glucopyranoside(8),khaephuoside A(9),khaephuoside B(10),3,4,5-trimethoxy-phenyl-O-β-D-glucopyranoside(11),liquiritigenin(12),7,3'-dihydroxy-flavanone-4'-O-β-D-glucopyranoside(13),pinoresinol(14),syringaldehyde(15),(+)-1-hydroxy-pinoresinol-1-β-D-glucopyranoside(16),β-amyrin(17).Compounds 2-5、7、9、10、12、17 could promote the glucose uptake in L6 cells.CONCLUSION Compound 1 is a new compound,and 2-9、11-13、15-17 are isolated from this plant for the first time.Compounds 2-5、7、9、10、12、17 have good hypoglycemic activity.

Objective To develop a predictive model integrating clinical features,deep learning(DL),and radiomics based on T2-weighted imaging for prenatal assessment of postpartum hemorrhage(PPH)risk in high-risk pregnant women.Methods A total of 538 pregnant women with ultrasound-reported high-risk placenta accrete were retrospectively enrolled and divided into training,internal test,and external test cohorts.A nnUNet model was trained for automatic placental segmentation.Univariate and multivariate analyses were conducted on clinical features to identify those associated with PPH.Quantitative radiomic features were extracted from the placental region,and a random forest model was developed to predict estimated blood loss(EBL)and PPH risk.A DenseNet-based multi-task DL model was trained to predict PPH risk,EBL,and placenta previa status.Finally,a DL-radiomics ensemble(DRE)model was constructed by integrating clinical features,DL outputs,and radiomics scores.Diagnostic performance was evaluated using the area under the receiver operating characteristic curve(AUC)and DeLong test.Results The DRE model achieved AUC values of 0.874(95%CI:0.792-0.951)and 0.836(95%CI:0.648-0.974)in the internal and external test cohorts,respectively,significantly outperforming the standalone clinical,DL,and radiomics models.Incorporation of EBL regression improved the performance of the PPH classification model,with the external test AUC increasing from 0.261-0.788 to 0.836.Conclusion The DRE model integrating DL and radiomics can efficiently predict PPH risk and assist in the clinical management of high-risk pregnancies.