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.

ObjectiveThe widespread adoption of portable fundus cameras for primary care and community screening is hindered by limitations in current autofocus(AF) technologies. Image-based methods relying on sharpness evaluation require iterative searches, resulting in slow convergence, while projection-based techniques are susceptible to optical artifacts and calibration errors. To address these challenges, this study introduces a novel AF system based on direct wavefront sensing, designed to deliver simultaneous high speed, high precision, and operational robustness within the compact form factor essential for portable ophthalmic devices. MethodsOur approach fundamentally reimagines the AF process by directly measuring the ocular wavefront aberration. We developed a custom portable fundus camera integrating a miniaturized Shack-Hartmann wavefront sensor (SHWS) into the optical path. An 850 nm laser diode projects a point source onto the retina via oblique illumination to minimize corneal reflections. Light scattered from this spot carries the eye’s refractive error through the imaging optics and is directed to the SHWS, positioned at a plane optically conjugate to the primary color CMOS imaging sensor. A microlens array within the SHWS samples the incident wavefront, generating a pattern of focal spots on a CCD. Real-time centroid analysis of these spots provides a map of local wavefront slopes. These measurements are processed through a singular value decomposition (SVD) algorithm to fit a Zernike polynomial basis set, enabling real-time reconstruction of the wavefront phase. The defocus component (S) is extracted from the second-order Zernike coefficients, providing a direct, quantitative measure of the refractive error in diopters. This value serves as a precise error signal in a closed-loop control system, which commands a voice-coil actuated focusing lens to its null position in a single, deterministic step, eliminating the need for iterative search algorithms. ResultsComprehensive evaluation demonstrated the system’s high performance. Testing on a calibrated model eye (OEMI-7) established a highly linear relationship between the computed defocus S and the focusing lens position across a ±20 Diopter (D) compensation range, achievable within a 5 mm mechanical travel. The system achieved a focusing precision of 0.08 D, corresponding to an 18-fold improvement over a conventional projection spot-size method tested under identical conditions. The total focus acquisition time, encompassing wavefront measurement, computation, and lens actuation, averaged under 0.5 s. Clinical validation with 25 human volunteers (50 eyes, refractive range -15 D to +10 D) confirmed practical efficacy. The wavefront-sensing AF succeeded in 92% of attempts with a mean time of 0.5 s, substantially outperforming a projection-based benchmark which achieved only a 32% success rate with an average time of 4.25 s. The system provided instantaneous directional guidance and maintained stability during minor ocular movements. Objective assessment of image quality, via amplitude contrast of retinal vasculature, showed consistent and significant enhancement following AF correction across the entire tested diopter range. ConclusionThis work successfully implements and validates a direct wavefront-sensing autofocus paradigm for portable fundus cameras. By directly quantifying and compensating for the optical defocus aberration, this method bypasses the fundamental limitations of image-processing and projection-based techniques, enabling rapid, precise, and deterministic diopter compensation. The developed system delivers an exceptional combination of a wide operational range (±20 D), high accuracy (0.08 D), fast convergence (0.5 s), and a compact physical footprint. This technology provides a practical and high-performance focusing solution capable of enhancing the reliability, throughput, and diagnostic utility of portable retinal imaging in large-scale screening applications. Future efforts will be directed towards system cost optimization and performance adaptation for diverse ocular conditions.

Objective To design an intelligent monitoring platform for adverse drug reactions(ADRs)to solve the problems of the traditional ADR monitoring mode.Methods The ADR intelligent monitoring platform was designed based on artificial intelligence and big data technologies,which was developed with Browser/Server(B/S)architecture,C#programming language and.NET development tool.There were five functional modules involved in the platform for ADR knowledge base,monitoring rule setting,intelligent monitoring,report management and statistical analysis.Results The platform realized the full-process management of ADR intelligent monitoring,reporting,review and statistical analysis,which enhanced the ADR report in quantity,quality and timeliness.Conclusion The platform contributes to improving the monitoring of ADR and patient medication safety.[Chinese Medical Equipment Journal,2025,46(9):33-38]

Objective To design an intelligent monitoring platform for adverse drug reactions(ADRs)to solve the problems of the traditional ADR monitoring mode.Methods The ADR intelligent monitoring platform was designed based on artificial intelligence and big data technologies,which was developed with Browser/Server(B/S)architecture,C#programming language and.NET development tool.There were five functional modules involved in the platform for ADR knowledge base,monitoring rule setting,intelligent monitoring,report management and statistical analysis.Results The platform realized the full-process management of ADR intelligent monitoring,reporting,review and statistical analysis,which enhanced the ADR report in quantity,quality and timeliness.Conclusion The platform contributes to improving the monitoring of ADR and patient medication safety.[Chinese Medical Equipment Journal,2025,46(9):33-38]

Objective:This study aims to investigate the predictive performance of serum levels of white blood cell count(WBC),D-dimer(D-D)and N-terminal pro-brain natriuretic peptide(NT-proBNP)for major adverse cardiovascular events(MACE)in elderly patients with acute ST-segment elevation myocardial infarction(ASTE-MI).Methods:A total of 70 elderly patients with ASTEMI treated in Bozhou People's Hospital between April 2020 and May 2023 were prospectively selected as observation group.Incidence of MACE during 1-year follow-up were recorded,another 50 patients with unstable angina pectoris treated in our hospital simultaneously were selected as control group.Serum levels of WBC,D-D and NT-proBNP were compared among above groups.The receiver operating characteristic(ROC)curve was used to evaluate the predictive value of serum WBC,D-D and NT-proBNP for MACE in elderly patients with ASTEMI.A nomogram was established,and calibration curve and deci-sion curve analysis(DCA)were used to evaluate the performance of model.Results:A total of 40 cases(59.70％)experienced MACE during one-year follow-up.Compared to those in control group,patients in observation group had significant higher serum WBC[(11.43±1.98)×109/Lvs.(6.30±1.99)× 109/L],D-D[(0.91±0.20)mg/L vs.(0.47±0.18)mg/L]andNT-proBNP[(192.31±63.19)pg/ml vs.(114.05±22.79)pg/ml](P＜0.001 all).Compared to participants without MACE,those with MACE had significantly higher serum WBC[(13.33±1.90)× 109/L vs.(10.27±0.98)× 109/L],D-D[(1.11±0.25)mg/L vs.(0.87±0.21)mg/L]and NT-proBNP[(238.73±50.22)pg/ml vs.(150.70±39.16)pg/ml](P＜0.001 all).ROC analysis showed that the ar-ea under the curve(AUC)of the combined detection of serum WBC(AUC=0.791,95％CI 0.677～0.879),D-D(AUC=0.767,95％CI 0.650～0.859)and NT-proBNP(AUC=0.733,95％CI 0.614～0.832)was 0.916(95％CI 0.825～0.969),which was significantly higher than those of single detections(Z=2.386,4.953,3.190,P=0.017,0.004,＜0.001).The total score of the nomogram model constructed based on the levels of WBC,D-D and NT-proBNP ranged from 70 to 126 points.The predicted incidence was basically consistent with the actual in-cidence.For the internal verification of the model,the AUC of ROC curve of the training set and the validation set was 0.863 and 0.926 respectively.The DCA curve was located above the critical curve,indicating that the model had a net benefit and good clinical effectiveness.Conclusion:Serum WBC,D-D and NT-proBNP significantly el-evated in elderly patients with ASTEMI.The combined detection of serum WBC,D-D and NT-proBNP levels has good predictive value for MACE in these patients.

With the rapid development of generative artificial intelligence(GAI)technologies,their widespread application in academic research and writing is continuously expanding the boundaries of sci-entific inquiry.However,this trend has also raised a series of ethical and regulatory challenges,inclu-ding issues related to authorship,content authenticity,citation accuracy,and accountability.In light of the growing involvement of AI in generating academic content,establishing an open,controllable,and trustworthy ethical governance framework has become a key task for safeguarding research integrity and maintaining trust within the academic community.This expert consensus outlines ethical requirements across key stages of AI-assisted academic writing-including topic selection,data management,citation practices,and authorship attribution.It aims to clarify the boundaries and ethical obligations surrounding AI use in academic writing,ensuring that technological tools enhance efficiency without compromising in-tegrity.The goal is to provide guidance and institutional support for building a responsible and sustainable research ecosystem.

ObjectiveThe aim of this study was to investigate the prophylactic effects of caloric restriction (CR) on lipopolysaccharide (LPS)-induced septic cardiomyopathy (SCM) and to elucidate the mechanisms underlying the cardioprotective actions of CR. This research aims to provide innovative strategies and theoretical support for the prevention of SCM. MethodsA total of forty-eight 8-week-old male C57BL/6 mice, weighing between 20-25 g, were randomly assigned to 4 distinct groups, each consisting of 12 mice. The groups were designated as follows: CON (control), LPS, CR, and CR+LPS. Prior to the initiation of the CR protocol, the CR and CR+LPS groups underwent a 2-week acclimatization period during which individual food consumption was measured. The initial week of CR intervention was set at 80% of the baseline intake, followed by a reduction to 60% for the subsequent 5 weeks. After 6-week CR intervention, all 4 groups received an intraperitoneal injection of either normal saline or LPS (10 mg/kg). Twelve hours post-injection, heart function was assessed, and subsequently, heart and blood samples were collected. Serum inflammatory markers were quantified using enzyme-linked immunosorbent assay (ELISA). The serum myocardial enzyme spectrum was analyzed using an automated biochemical instrument. Myocardial tissue sections underwent hematoxylin and eosin (HE) staining and immunofluorescence (IF) staining. Western blot analysis was used to detect the expression of protein in myocardial tissue, including inflammatory markers (TNF-α, IL-9, IL-18), oxidative stress markers (iNOS, SOD2), pro-apoptotic markers (Bax/Bcl-2 ratio, CASP3), and SIRT3/SIRT6. ResultsTwelve hours after LPS injection, there was a significant decrease in ejection fraction (EF) and fractional shortening (FS) ratios, along with a notable increase in left ventricular end-systolic diameter (LVESD). Morphological and serum indicators (AST, LDH, CK, and CK-MB) indicated that LPS injection could induce myocardial structural disorders and myocardial injury. Furthermore, 6-week CR effectively prevented the myocardial injury. LPS injection also significantly increased the circulating inflammatory levels (IL-1β, TNF-α) in mice. IF and Western blot analyses revealed that LPS injection significantly up-regulating the expression of inflammatory-related proteins (TNF-α, IL-9, IL-18), oxidative stress-related proteins (iNOS, SOD2) and apoptotic proteins (Bax/Bcl-2 ratio, CASP3) in myocardial tissue. 6-week CR intervention significantly reduced circulating inflammatory levels and downregulated the expression of inflammatory, oxidative stress-related proteins and pro-apoptotic level in myocardial tissue. Additionally, LPS injection significantly downregulated the expression of SIRT3 and SIRT6 proteins in myocardial tissue, and CR intervention could restore the expression of SIRT3 proteins. ConclusionA 6-week CR could prevent LPS-induced septic cardiomyopathy, including cardiac function decline, myocardial structural damage, inflammation, oxidative stress, and apoptosis. The mechanism may be associated with the regulation of SIRT3 expression in myocardial tissue.

Objective To explore an experimental protocol for differentiating human-induced pluripotent stem cells(iPSCs)into highly pure midbrain dopaminergic(DA)neurons.Methods By optimizing a blend of small molecules and recombinant human growth factors,iPSCs were induced to differentiate into ventral midbrain floor plate DA progenitor cells and subsequently into mature substantia nigra pars compacta DA neurons.Throughout the differentiation process,Real-time PCR and immunofluorescent staining were utilized as a method for quality assessment.Results iPSCs firstly differentiate into dopaminergic precursor cells,and then gradually differentiate into DA neurons expressing tyrosine hydroxylase(TH).Conclusion The protocol successfully yields approximately high purity tyrosine hydroxylase-positive(TH+)DA neurons.This differentiation technique offers an effective cellular model for studying the physiological mechanisms and pathogenesis of Parkinson's disease,providing valuable insights for future research and potential therapeutic strategies.

Facing of mounting resource constraints and rising demands for personalization in medical education,regional anatomy teaching urgently requires transformation.In this paper,we focus on the regional anatomy of the thoracic wall,in order to explore a novel AI-driven teaching paradigm.Anchored in the core principle of"virtual-real integration with cadaveric dissection as the cornerstone,"the paradigm redefines educational objective and constructs an intelligent,closed-loop teaching model integrating students,computers,and instructors.Leveraging the robust support of digital intelligence(e.g.,DeepSeek),this paradigm incorporates interactive method including group collaboration,branching instruction,and gamified assessments.It achieves a comprehensive intelligent transformation of the entire teaching process-from goal setting and plan customization to activity implementation,task completion,outcome exchange,multidimensional evaluation,and reflective iteration.This new paradigm centers on medical students and leverages digital intelligence to activate deep personalized learning potential.It seamlessly integrates fundamental anatomical knowledge with clinical scenarios(e.g.,key anatomy in breast cancer surgery,flap design in breast reconstruction),and significantly enhances clinical decision-making abilities,scientific research and innovative thinking,as well as medical humanistic literacy,paving a new path for intelligent medical education.