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.

Objective To establish autoverification rules for six routine coagulation assays(PT,APTT,TT,Fib,DD,and FDP)based on the stratification of outpatients and inpatients,in accordance with CLSI AUTO-10A,AUTO-15,and WS/T 616-2018 guide-lines,and to validate the feasibility of this stratified strategy with clinical data while optimizing verification efficiency.Methods A to-tal of 323 451 coagulation test results from Zhongshan Hospital,Fudan University in 2022 were retrospectively analyzed to define auto-verification rules involving critical values,instrument flags,logical rules,historical comparison,and numerical ranges.A stratified au-toverification system was established by applying distinct rules for outpatient and inpatient populations.Subsequently,the rules were op-timized using 87 830 coagulation test results from January to March 2024,and the consistency between autoverification and manual veri-fication was prospectively evaluated using 33 968 consecutive coagulation specimens collected in April 2024.Results A stratified au-toverification system was successfully developed,comprising a total of 53 rules.The pass rate of overall verification was 77.16％(26 210/33 968),with a true-positive rate of 19.64％(6 672/33 968),a false-positive rate of 3.20％(1 086/33 968),a true-nega-tive rate of 77.16％(26 210/33 968),and no false negatives were detected.Conclusion The proposed autoverification system signifi-cantly improved verification efficiency.The stratified design based on outpatient and inpatient populations effectively minimized the risk of false negatives,and may provide a novel approach for the further development and optimization of coagulation test autoverification.

OBJECTIVE To study the chemical constituents in the root bark of Schisandra sphenanthera and their cytotoxic activ-ities.METHODS The compounds were isolated and purified by silica,Sephadex LH-20 and semi preparative-HPLC and the chem-ical structures were identified by 1 H-NMR,13 C-NMR and MS data analysis.The cytotoxic activities of the compounds were deter-mined by MTT method.RESULTS Twenty lignans were isolated and deduced as:Matairesinol(1),2-Hydroxy-2-(3′,4′-di-hydroxyphenyl)methyl-3-(3″,4″-dimethoxyphenyl)methyl-gamma-butyrolactone(2),(+)-Nortrachelogenin(3),2-Hydroxy-2-(4′-O-β-D-glucopyranosyl-3′-hydroxyphenyl)methyl-3-(3″,4″-dimethoxyphenyl)methyl-γ-butyrolactone(4),Nortracheloside(5),Burselignan(6),(+)-Cycloolivil(7),5-Methoxy-(+)-isolariciresinol(8),(-)-Isolariciresinol 3α-O-β-D-glucopyranoside(9),(+)-9-O-β-D-Glucopyranosyl lyoniresinol(10),(-)-Secoisolariciresinol(11),Licarin A(12),Cedrusin(13),Mataires-inol 4′-O-β-D-glucopyranoside(14),Pregomisin(15),Meso-dihydroguaiaretic acid(16),7S,8R-Erythro-4,9,9′-trihydroxy-3,3′-dimethoxy-8-O-4′-neolignan-7-O-β-D-glucopyranoside(17),Gomisin M2(18),Gomisin M3(19),Pinoresinol(20).Com-pounds 1-3,12,15,16,18 and 19 showed cytotoxic activity against A549,HCT116 and SW620 cell lines with IC50 values ranging from 1.4 to 22.9 μmol·L-1.CONCLUSION Compounds 1-4,6-12,14,17-19 are isolated from the plant for the first time,com-pounds 1-3,12,15,16,18 and 19 exhibit cytotoxic activities.

Objective To compare imaging quality and semi-quantitative parameters of 18F-FDG whole-body images obtained with domestic NeuWise and Philips Ingenuity TF PET/CT scanners.Methods Thirty-four patients who underwent 18F-FDG whole-body scanning using NeuWise and Philips Ingenuity TF PET/CT systems respectively on the same day were enrolled.The imaging quality and semi-quantitative parameters of 2 kind images,also the mean standard uptake value(SUVmean)of normal tissue,the maximum standard uptake value(SUVmax),peak standard uptake value(SUVpeak),SUVmean of lesions,total lesion glycolysis(TLG)and metabolic tumor volume(MTV)were compared.Results No significant difference of imaging quality nor semi-quantitative parameters of lesions(all P＞0.05),while significant differences of SUVmean of aortic arch,liver,lumbar vertebra and spinal cord were found between 2 kind images(all P＜0.05).Strong correlations of SUVmax,SUVmean,MTV and TLG of lesions(r,=0.734-0.890,P＜0.001),and high correlation of SUVpeak(rs=0.919,P＜0.001)were found between 2 kind images.The consistency of SUVmax,SUVpeak,SUVmean,TLG and MTV at the lesion site between 2 kind images were very good to extremely good(ICC=0.891-0.986,all P＜0.001),and the differences of all above semi-quantitative parameters were within 95％confidence interval.Conclusion Imaging quality of 18F-FDG whole-body images obtained with domestic NeuWise and Philips Ingenuity TF PET/CT scanners could meet the requirements of clinical diagnosis and treatment,and semi-quantitative parameters obtained based on both images had good consistencies.

Objective To improve the underwater technique during the start phase of breaststroke swimmers through swimming flume training and assess its effectiveness in enhancing competitive performance.Methods The participants were 14 male swimmers in the short-distance Level 4 category from the Shanghai Swimming Team,and they were randomly divided into the experimental group(n=7)and control group(n=7).The experimental group underwent swimming flume start training,while the control group underwent regular pool start training,and both groups received the training twice a week for 16 weeks.Before and after the 16-week training,a 15-m breaststroke start test was conducted.Repeated measures analysis of variance and paired t-tests were used to compare the changes in kinematic parameters(time,speed,and entry angle)between and within groups.Results After 16 weeks of specialized training,the 15-m performance at the start for the experimental group and control group(F(1,12)=6.52,P＜0.05,η2=0.39)showed an interaction,with the experimental group performing better after training compared to the control group before training(P＜0.05).In the experimental group,the duration of the pull-out phase(F(1,12)=10.28,P＜0.01,η2=0.46)and the second sliding phase(F(1,12)=4.81,P＜0.05,η2=0.22)was improved;the distance of the pull-out phase(F(1,12)=4.71,P＜0.05,η2=0.21)and the second sliding phase(F(1,12)=4.90,P＜0.05,η2=0.21)was improved;the speed of the pull-out phase(F(1,12)=4.77,P＜0.05,72=0.20)was significantl improved.The within-group statistics showed that the experimental group significantly improved their exit speed(P＜0.05).The hand entry angle was optimized(P＜0.05),while changes in other joint angles were not significant.Conclusions Swimming flume training reduced the time spent in the pull-out and second gliding phases during the breaststroke start,effectively preventing the speed loss during underwater gliding.This provides an experimental evidence for enhancing start performance and optimizing training method for breaststroke swimmers.

Objective To compare imaging quality and semi-quantitative parameters of 18F-FDG whole-body images obtained with domestic NeuWise and Philips Ingenuity TF PET/CT scanners.Methods Thirty-four patients who underwent 18F-FDG whole-body scanning using NeuWise and Philips Ingenuity TF PET/CT systems respectively on the same day were enrolled.The imaging quality and semi-quantitative parameters of 2 kind images,also the mean standard uptake value(SUVmean)of normal tissue,the maximum standard uptake value(SUVmax),peak standard uptake value(SUVpeak),SUVmean of lesions,total lesion glycolysis(TLG)and metabolic tumor volume(MTV)were compared.Results No significant difference of imaging quality nor semi-quantitative parameters of lesions(all P＞0.05),while significant differences of SUVmean of aortic arch,liver,lumbar vertebra and spinal cord were found between 2 kind images(all P＜0.05).Strong correlations of SUVmax,SUVmean,MTV and TLG of lesions(r,=0.734-0.890,P＜0.001),and high correlation of SUVpeak(rs=0.919,P＜0.001)were found between 2 kind images.The consistency of SUVmax,SUVpeak,SUVmean,TLG and MTV at the lesion site between 2 kind images were very good to extremely good(ICC=0.891-0.986,all P＜0.001),and the differences of all above semi-quantitative parameters were within 95％confidence interval.Conclusion Imaging quality of 18F-FDG whole-body images obtained with domestic NeuWise and Philips Ingenuity TF PET/CT scanners could meet the requirements of clinical diagnosis and treatment,and semi-quantitative parameters obtained based on both images had good consistencies.

Objective To improve the underwater technique during the start phase of breaststroke swimmers through swimming flume training and assess its effectiveness in enhancing competitive performance.Methods The participants were 14 male swimmers in the short-distance Level 4 category from the Shanghai Swimming Team,and they were randomly divided into the experimental group(n=7)and control group(n=7).The experimental group underwent swimming flume start training,while the control group underwent regular pool start training,and both groups received the training twice a week for 16 weeks.Before and after the 16-week training,a 15-m breaststroke start test was conducted.Repeated measures analysis of variance and paired t-tests were used to compare the changes in kinematic parameters(time,speed,and entry angle)between and within groups.Results After 16 weeks of specialized training,the 15-m performance at the start for the experimental group and control group(F(1,12)=6.52,P＜0.05,η2=0.39)showed an interaction,with the experimental group performing better after training compared to the control group before training(P＜0.05).In the experimental group,the duration of the pull-out phase(F(1,12)=10.28,P＜0.01,η2=0.46)and the second sliding phase(F(1,12)=4.81,P＜0.05,η2=0.22)was improved;the distance of the pull-out phase(F(1,12)=4.71,P＜0.05,η2=0.21)and the second sliding phase(F(1,12)=4.90,P＜0.05,η2=0.21)was improved;the speed of the pull-out phase(F(1,12)=4.77,P＜0.05,72=0.20)was significantl improved.The within-group statistics showed that the experimental group significantly improved their exit speed(P＜0.05).The hand entry angle was optimized(P＜0.05),while changes in other joint angles were not significant.Conclusions Swimming flume training reduced the time spent in the pull-out and second gliding phases during the breaststroke start,effectively preventing the speed loss during underwater gliding.This provides an experimental evidence for enhancing start performance and optimizing training method for breaststroke swimmers.

Objective To establish autoverification rules for six routine coagulation assays(PT,APTT,TT,Fib,DD,and FDP)based on the stratification of outpatients and inpatients,in accordance with CLSI AUTO-10A,AUTO-15,and WS/T 616-2018 guide-lines,and to validate the feasibility of this stratified strategy with clinical data while optimizing verification efficiency.Methods A to-tal of 323 451 coagulation test results from Zhongshan Hospital,Fudan University in 2022 were retrospectively analyzed to define auto-verification rules involving critical values,instrument flags,logical rules,historical comparison,and numerical ranges.A stratified au-toverification system was established by applying distinct rules for outpatient and inpatient populations.Subsequently,the rules were op-timized using 87 830 coagulation test results from January to March 2024,and the consistency between autoverification and manual veri-fication was prospectively evaluated using 33 968 consecutive coagulation specimens collected in April 2024.Results A stratified au-toverification system was successfully developed,comprising a total of 53 rules.The pass rate of overall verification was 77.16％(26 210/33 968),with a true-positive rate of 19.64％(6 672/33 968),a false-positive rate of 3.20％(1 086/33 968),a true-nega-tive rate of 77.16％(26 210/33 968),and no false negatives were detected.Conclusion The proposed autoverification system signifi-cantly improved verification efficiency.The stratified design based on outpatient and inpatient populations effectively minimized the risk of false negatives,and may provide a novel approach for the further development and optimization of coagulation test autoverification.

The integration of science and education is not only an important strategy for promoting social progress and technological development,but also a modern form of higher education aiming at cultivating innovative talents.Conducting scientific research training for undergraduate medical students is one of the important ways to cultivate their innovative abilities and comprehensive qualities.Our team proposed a"teaching,science,and ideology trinity"teaching model to comprehensively cultivate students' scientific research comprehensive abilities under the value orientation of ideological and political education by or-ganically integrating molecular biology experimental teaching with the scientific research training of under-graduate medical students.In this teaching activity,taking the experiment of gene polymorphism as an example,our team selected students with research potential from the whole grade and divided them into 4 project groups that were instructed by 4 teachers.The students were trained in the whole process of scien-tific research,including topic selection,project writing,experimental designing,application for research ethics,and project summary.Our team has always adhered to student-contentedness of educational con-cepts to stimulate students' intrinsic motivation throughout the teaching process.Students are the design-ers and implementers of the project,and teachers are only guides and promoters of learning.After this training,students not only became familiar with the writing and implementation of scientific research pro-jects,but also improved their literature reading,experimental designing,experimental skills,and prob-lem-solving abilities.More importantly,this teaching activity also cultivated students' awareness of re-search ethics and academic moral standards.