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 investigate anxiety symptoms and associated factors in relocated elderly residents of new townships,and to provide evidence for prevention interventions.Methods:A cross-sectional study was conduc-ted in relocated elderly residents in new townships of three urban areas in Shandong Province from 2021 to 2023.The study instruments included Ascertain Dementia-8,Generalized Anxiety Disorder-7,self-administered de-mographic characteristics information questionnaire.Multivariate analysis of factors associated with anxiety symp-toms in elderly residents was performed using multivariate logistic regression.Results:The prevalence rate of mild anxiety symptoms was 5.8％,and the rate of moderate-to-severe anxiety symptoms was 1.3％in 3 313 resi-dents.Multivariate analysis found that self-assessed general psychological condition(OR=0.52),good family envi-ronment(OR=0.34),no self-perceived cognitive impairment(OR=0.31),no chronic diseases(OR=0.42),and only one chronic disease(OR=0.61)were protective factors for mild anxiety symptoms,and very good dietary structure(OR=2.15)and fair dietary structure(OR=2.39)were risk factors for those.Very good family environ-ment(OR=0.11)and average family environment(OR=0.16),and no self-perceived cognitive impairment(OR=0.14)were protective factors for moderate-to-severe anxiety symptoms,and 0-3 years(OR=3.24)and 4-6 years(OR=3.28)of relocation were risk factors for those.Conclusion:Family environment,dietary structure,and duration since relocation are key factors associated with anxiety symptoms among relocated elderly residents in new townships.Targeted interventions should be implemented to address their mental health needs.

The key link to colorectal cancer progression is that epigenetic derangements promote reprogramming of glucose metabolism.Based on many studies on the relationship between spleen being the origin of acquired constitution and epigenetics,spleen qi dispersing essence and glucose metabolism,combined with the extensive experience of traditional Chinese medicine in the treatment of colorectal cancer and the accumulation of researches on glucose metabolism,it is believed that the"epigenetic-reprogramming of glucose metabolism"pathology of colorectal cancer is the microscopic essence of the key pathogenesis of"dysfunction of spleen transforming essences".The"promoting spleen and dispersing essences"method means that"promoting spleen"to reverse epigenetic derangements and"dispersing essences"to regulate reprogramming of glucose metabolism was put forward to treat colorectal cancer.Look forward to enrich the scientific connotation of the pathogenesis of"dysfunction of spleen transforming essences",new ideas and potential empirical basis for traditional Chinese medicine intervention in colorectal cancer were provided.

Three evaluation methods were recommended for the key properties of the intravascular catheter lubricating coating such as stability,lubricity and integrity,including insoluble particle test method,friction test procedure and appearance detection method.Fifteen batches of microcatheters produced by different manufacturers were selected for testing to clarify the three methods in test principle,step,result,characteristic.References were provided for the design,production,evaluation and regulation of intravascular catheters with lubricant coatings.[Chinese Medical Equipment Journal,2025,46(1):66-72]

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.

Objective:The influencing factors of intraoperative hypoxemia in patients with benign central airway stenosis were investigated by machine learning algorithm, and the prediction model of hypoxemia was constructed and verified.Methods:A case-control study was used in this study. The clinical data of 650 patients with benign central airway stenosis who who received surgical treatment in the First Affiliated Hospital of PLA Naval Medical University from June 2022 to April 2024 were retrospectively analyzed. And they were divided into a training set ( n=455) and a test set ( n=195) according to 7:3. The training set was used for establishing Logistic regression model and conducting internal verification, and the test set was used for external verification. The least absolute shrinkage and selection operator (LASSO) regression and Boruta algorithm were used to select the factors affecting intraoperative hypoxemia in patients with benign central airway stenosis. A Logistic regression prediction model was constructed, and the model was evaluated using area under the receiver operating characteristic curve (AUC), decision curve analysis (DCA) and calibration curve. Shapley additive interpretation (SHAP) were used to analyze the importance of influencing factors. Results:Among 650 patients, 279 were males and 371 were females, aged (37.86 ± 8.82) years. Nine feature variables were screened by LASSO regression, while 7 feature variables were screened by Boruta algorithm, the intersection of the two was operation time, complications, degree of airway stenosis, thermal ablation therapy, balloon dilation, and airway stent, respectively, based on this, a logistic regression prediction model was constructed.The AUC values of the training set, validation set and test set of the model were 0.928 (95% CI 0.903-0.954), 0.922 (95% CI 0.843-0.995) and 0.919 (95% CI 0.872-0.965), respectively. The calibration curve showed that the predicted results of the model were in good agreement with the actual results, and the DCA curve showed that the model had clinical application value. SHAP analysis showed that the importance of variables affecting intraoperative hypoxemia in benign central airway stenosis patients was ranked as operation time, thermal ablation therapy, degree of airway stenosis, comorbidification, balloon dilation, and airway stent. Conclusions:The Logistic regression prediction model of intraoperative hypoxemia built based on machine learning algorithm has good prediction efficiency, which is helpful to early identification of risk groups and prevention of hypoxemia.

Objective To evaluate the efficacy of the manual cyclotorsion compensation method based on the obser-vation screen in the correction of with-the-rule astigmatism during small incision lenticule extraction(SMILE).Methods This prospective study enrolled 40 patients who underwent SMILE and another 40 patients who underwent Q-value-guided femtosecond laser-assisted laser in situ keratomileusis(FS-LASIK)at the Ophthalmology Center of the Second Affiliated Hospital of Nanchang University from April to September 2024.Through the random number method,one eye of each pa-tient who underwent SMILE was assigned to the manual cyclotorsion compensation group(the cyclotorsion group,40 eyes),while the contralateral eye was assigned to the non-cyclotorsion group(the non-cyclotorsion group,40 eyes).Meanwhile,one eye of each patient who underwent FS-LASIK was randomly selected for the FS-LASIK group(40 eyes).The age,uncorrected distance visual acuity(UDVA),spherical power,cylindrical power,spherical equivalent(SE),and corrected distance visual acuity(CDVA)were recorded before surgery and 3 months after surgery,respectively.Astigma-tism was subjected to Alpins vector analysis,and the indicators for assessing astigmatism included target-induced astigma-tism(TIA),surgically-induced astigmatism(SIA),difference vector(DV),correction index(CI),success index(IOS),angle of error(AE),and absolute value of angle of error(|AE|).Results Before surgery,there was no significant difference in age,spherical power,cylindrical power,SE,and CDVA among the cyclotorsion group,the non-cyclotorsion group,and the FS-LASIK group(all P＞0.05).At 3 months after surgery,the cyclotorsion group showed better UDVA and lower cylindrical power than the non-cyclotorsion group,with statistically significant differences(both P＜0.05);howev-er,there was no significant difference in CDVA,spherical power,and SE between the two groups(all P＞0.05).Besides,there was no significant difference in UDVA,CDVA,or refractive parameters between the cyclotorsion and FS-LASIK groups(all P＞0.05).The Alpins vector analysis of astigmatism 3 months after surgery revealed better SIA,CI,IOS,and|AE|in the cyclotorsion group compared with the non-cyclotorsion group,with statistically significant differences(all P＜0.05);however,there was no significant difference in TIA,DV,and AE between the two groups(all P＞0.05).No signifi-cant differences were found between the cyclotorsion and FS-LASIK groups in any astigmatism vector parameter(all P＞0.05).The linear regression analysis results indicated a high linear correlation between TIA and SIA in all groups.Conclu-sion The manual cyclotorsion compensation method based on the observation screen in the correction of with-the-rule astigmatism during SMILE is comparable to Q-value-guided FS-LASIK and superior to the conventional central tear film marking method in SMILE.

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.

Objective:To investigate the relationship between lipid levels and cognitive impairment in the elderly Chinese population using prospective cohort data.Methods:Based on the China-PAR (Prediction for Atherosclerotic Cardiovascular Disease Risk in China) cohort, this study included 24 380 individuals aged ≥60 years who participated in the cognitive function follow-up survey from 2018 to 2019. Cognitive function was assessed using the Mini-Mental State Examination (MMSE), with cognitive impairment defined according to different educational levels: MMSE ≤17 for illiterate individuals, MMSE ≤20 for those with primary education and MMSE ≤24 for those with secondary education or above. Multivariable linear regression and logistic regression models were employed to examine the associations between six baseline lipid indicators and cognitive scores, as well as cognitive impairment. Additionally, restricted cubic splines were used to explore the exposure-dose relationship between lipid levels and cognitive function.Results:The study population had a median follow-up time of 11.6 years, with a baseline age of (59.7±6.8) years. Among the participants, 9 510 (39.0%) were males, and the mean MMSE score was 24.7±6.8. A total of 3 887 individuals (15.9%) were identified as cognitively impaired. The results of multivariable linear regression and logistic regression indicated that total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C) and non-high-density lipoprotein cholesterol (non-HDL-C) levels were not only significantly positively associated with cognitive scores but also significantly associated with a lower risk of cognitive impairment. Each 1 mmol/L increase in these lipid levels corresponded to β values (95% CI) of 0.267 (0.173-0.361), 0.385(0.271-0.499) and 0.331(0.231-0.431), respectively. Each 1 mmol/L increase in these lipid levels corresponded to odds ratio ( OR) (95% CI) values of 0.915 (0.876-0.956), 0.875 (0.830-0.923) and 0.886 (0.848-0.927), respectively. The dose-response curve demonstrated that the negative association was primarily observed within the guideline-recommended optimal lipid level range. Specifically, when LDL-C was less than 3.4 mmol/L and non-HDL-C was less than 4.1 mmol/L, the corresponding OR (95% CI) values were 0.859 (0.796-0.926) and 0.876 (0.818-0.939). Conclusion:Lipid levels exhibit a certain linear negative association with cognitive impairment in elderly Chinese adults, with LDL-C and non-HDL-C demonstrating a stronger effect, particularly within the guideline-recommended optimal range.