Objective To evaluate the predictive value of different machine learning models constructed in a multicenter environment for potential organ donors and verify their clinical application feasibility. Methods The study included 2 000 inpatients admitted to five domestic tertiary hospitals from January 2020 to December 2023, who met the criteria for potential organ donation assessment. They were randomly divided into a training set and an internal validation set (7∶3). Another 300 similar patients admitted to the First Affiliated Hospital of Harbin Medical University from January 2024 to April 2025 were included as an external validation set. The area under the curve (AUC), sensitivity, specificity, accuracy and F1-score of three models were compared, and the consistency of the potential organ donor determination process was tested. Multivariate logistic regression analysis was used to identify predictive factors of potential organ donors. Decision curve analysis (DCA) was employed to verify the resource efficiency of each model, and the threshold interval and intervention balance point were assessed. Results Apart from age, there were no significant differences in other basic characteristics among the centers (all P>0.05). The consistency of the potential organ donor determination process among researchers in each center was good [all 95% confidence interval (CI) lower limits >0]. In the internal validation set, the XGBoost model had the best predictive performance (AUC=0.92, 95% CI 0.89-0.94) and the best calibration (P=0.441, Brier score 0.099). In the external validation set, the XGBoost model also had the best predictive performance (AUC=0.91, 95% CI 0.88-0.94), outperforming logistic regression and random forest models. Multivariate logistic regression showed that mechanical ventilation had the greatest impact (odds ratio=2.06, 95% CI 1.54-2.76, P<0.001). DCA indicated that the XGBoost model had the highest net benefit in the threshold interval of 0.2-0.6. The “treat all” strategy only had a slight advantage at extremely low thresholds. The recommended threshold interval, which balances intervention costs and clinical benefits, considers ≥50% positive predictive value (PPV) and ≤50 referrals per 100 high-risk patients. Conclusions The XGBoost model established in a multicenter environment is accurate and well-calibrated in predicting potential organ donors. Combined with DCA, it may effectively guide the timing of clinical interventions and resource allocation, providing new ideas for the assessment and management of organ donation after brain death.

ObjectiveTo investigate whether Huanglian Jiedutang （HLJD） and its major active constituents （geniposide， baicalin， and berberine） can inhibit the inflammatory response of BV2 cells under lipopolysaccharide （LPS） stimulation via the high-mobility group protein B1 （HMGB1）/Toll-like receptor 4 （TLR4）/nuclear factor-κB （NF-κB） signaling pathway， and to explore differences in therapeutic efficacy among the three monomers， their combined formula， and HLJD under equal content ratios. MethodsBV2 microglial cells were used as the primary experimental model. Cell viability was assessed using the cell counting kit-8 （CCK-8） method to examine the effects of different concentrations of dimethyl sulfoxide （DMSO， 0.8%， 0.4%， 0.2%， 0.1%， and 0.05%） on cell viability. IncuCyte was employed to monitor the growth of cells under different concentrations of HLJD （200， 100， 50， 25， 12.5， 6.25 mg·L^-1）. Nitric oxide （NO） assay was used to screen the optimal HLJD concentration. High-performance liquid chromatography （HPLC） determined the content of geniposide， baicalin， and berberine in HLJD， and experimental groups were subsequently established according to the relative proportions of these constituents. CCK-8 assay evaluated cell viability under different treatments. Enzyme-linked immunosorbent assay （ELISA） measured levels of inflammatory factors （TNF-α， IL-1β， IL-6， IL-10） in the supernatant. Flow cytometry assessed the effects of treatments on M1-type polarization of BV2 cells. Western blot determined the expression levels of HMGB1， TLR4， and NF-κB-related proteins. ResultsCompared with the blank group， DMSO at concentrations ≤0.2% did not affect cell viability within 48 h. BV2 cell growth plateaued at 24 h after treatment with 200 mg·L^-1 HLJD. Under stimulation with 2 mg·L^-1 LPS， this concentration of HLJD effectively reduced NO release， and 6 h pre-treatment had a stronger inhibitory effect on NO than direct administration. HPLC results showed that 1 mg of HLJD freeze-dried powder contained approximately 24 μg of geniposide， 15 μg of baicalin， and 30 μg of berberine. Based on these ratios， experimental groups were blank， LPS （2 mg·L^-1）， HLJD （200 mg·L^-1）， monomer combination， geniposide （4.8 mg·L^-1）， baicalin （3 mg·L^-1）， and berberine （6 mg·L^-1）. The monomer combination group consisted of all three active constituents dissolved together. LPS and HLJD or its active constituents did not affect cell viability compared with the blank group. LPS significantly increased TNF-α， IL-1β， IL-6， and IL-10 in the supernatant （P<0.01）. HLJD and its active constituents significantly reduced pro-inflammatory factors TNF-α， IL-1β， and IL-6 （P<0.05， P<0.01） while upregulating anti-inflammatory IL-10 （P<0.01）， with the monomer combination showing the strongest effect （P<0.05， P<0.01）. Compared with the blank group， LPS significantly increased the proportion of CD80⁺CD86⁺ （M1-type） BV2 cells （P<0.01）. HLJD and its constituents partially inhibited M1 polarization （P<0.05， P<0.01）， with the monomer combination exhibiting the most pronounced effect （P<0.05， P<0.01）. Compared with the blank group， LPS upregulated HMGB1， TLR4， and NF-κB-related proteins （P<0.01）， whereas HLJD and its active constituents significantly reduced their expression （P<0.05， P<0.01）， with the monomer combination having the strongest regulatory effect （P<0.05， P<0.01）. ConclusionHLJD and its major active constituents （geniposide， baicalin， berberine） can inhibit LPS-induced inflammatory responses in BV2 cells. The combination of the three active constituents demonstrates the most potent anti-inflammatory effect， significantly attenuating M1-type polarization of BV2 cells via the HMGB1/TLR4/NF-κB signaling pathway.

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.

LIU Wansu， as the foremost of the four great masters of the Jin-Yuan period， established the "theory of fire-heat'' and extended the fire-heat pathogenesis framework to the field of stroke， thereby forming the theory of ''fire-heat inducing stroke''. This achieved a paradigmatic shift in stroke etiology from ''exogenous wind inducing stroke'' to ''fire-heat inducing stroke''. This paper systematically reviews the developmental trajectory of LIU Wansu's ''fire-heat inducing stroke'' theory and explores the social background， academic origins， and core connotations of its theoretical construction. The study found that， based on the ''Nineteen Pathomechanisms'' in the Huangdi's Internal Classic （Huang Di Nei Jing） and combined with clinical practice， LIU Wansu proposed that fire-heat is the fundamental cause of stroke， and that the Six Climatic Factors and the Five Zhi-Emotions can all transform into fire. He further constructed a stratified syndrome differentiation and therapeutic system centered on clearing heat and purging fire， emphasizing differentiated treatment of exterior and interior syndromes， Six Meridians syndrome differentiation， and seasonally adjusted medication. This theory not only resolved the diagnostic and therapeutic dilemmas of febrile epidemic diseases during the Jin-Yuan period， but also exerted a profound influence on later physicians such as ZHANG Zihe and ZHU Danxi， thereby promoting the pluralistic development of stroke theory in traditional Chinese medicine （TCM）. Modern pharmacological research provides solid scientific evidence， confirming that the ''fire-heat'' pathological state is highly associated with key mechanisms such as excessive inflammatory responses， oxidative stress， and excitatory amino acid toxicity following cerebral ischemia. Heat-clearing and fire-purging prescriptions and agents， such as Huanglian Jiedu Tang and baicalin， can exert multi-target neuroprotective effects by regulating inflammatory signaling， enhancing antioxidant enzyme activity， and balancing neurotransmitters. This not only verifies the scientific basis of the ''fire-heat inducing stroke'' theory from a modern biological perspective but also provides conclusive evidence for the clinical application of heat-clearing and fire-purging therapy. LIU Wansu's ''fire-heat inducing stroke'' theory represents a major milestone in the historical understanding of stroke pathogenesis， and its academically transitional insights continue to hold core guiding value for the pattern identification and treatment of ischemic stroke today.

LIU Wansu， as the foremost of the four great masters of the Jin-Yuan period， established the "theory of fire-heat'' and extended the fire-heat pathogenesis framework to the field of stroke， thereby forming the theory of ''fire-heat inducing stroke''. This achieved a paradigmatic shift in stroke etiology from ''exogenous wind inducing stroke'' to ''fire-heat inducing stroke''. This paper systematically reviews the developmental trajectory of LIU Wansu's ''fire-heat inducing stroke'' theory and explores the social background， academic origins， and core connotations of its theoretical construction. The study found that， based on the ''Nineteen Pathomechanisms'' in the Huangdi's Internal Classic （Huang Di Nei Jing） and combined with clinical practice， LIU Wansu proposed that fire-heat is the fundamental cause of stroke， and that the Six Climatic Factors and the Five Zhi-Emotions can all transform into fire. He further constructed a stratified syndrome differentiation and therapeutic system centered on clearing heat and purging fire， emphasizing differentiated treatment of exterior and interior syndromes， Six Meridians syndrome differentiation， and seasonally adjusted medication. This theory not only resolved the diagnostic and therapeutic dilemmas of febrile epidemic diseases during the Jin-Yuan period， but also exerted a profound influence on later physicians such as ZHANG Zihe and ZHU Danxi， thereby promoting the pluralistic development of stroke theory in traditional Chinese medicine （TCM）. Modern pharmacological research provides solid scientific evidence， confirming that the ''fire-heat'' pathological state is highly associated with key mechanisms such as excessive inflammatory responses， oxidative stress， and excitatory amino acid toxicity following cerebral ischemia. Heat-clearing and fire-purging prescriptions and agents， such as Huanglian Jiedu Tang and baicalin， can exert multi-target neuroprotective effects by regulating inflammatory signaling， enhancing antioxidant enzyme activity， and balancing neurotransmitters. This not only verifies the scientific basis of the ''fire-heat inducing stroke'' theory from a modern biological perspective but also provides conclusive evidence for the clinical application of heat-clearing and fire-purging therapy. LIU Wansu's ''fire-heat inducing stroke'' theory represents a major milestone in the historical understanding of stroke pathogenesis， and its academically transitional insights continue to hold core guiding value for the pattern identification and treatment of ischemic stroke today.

ObjectiveThis paper aims to investigate and evaluate the staged efficacy and safety of the representative empirical prescription of the “Zhu Ke Jiao” theory， Qijia Rougan prescription， combined with entecavir in the treatment of hepatic fibrosis in chronic hepatitis B. MethodsA multicenter randomized controlled clinical study was conducted， and 101 patients diagnosed with chronic hepatitis B-related hepatic fibrosis （CHB-HF） who met the diagnosis and inclusion criteria were randomly assigned to an observation group （Qijia Rougan prescription + entecavir） and a control group （entecavir）. The treatment duration was 24 weeks. Liver stiffness measurement （LSM）， fibrosis-4 index （FIB-4）， portal vein diameter， hepatitis B serology， biochemical indicators， hepatic fibrosis markers in serum ［hyaluronic acid （HA）， laminin （LN）， procollagen Ⅲ peptide （PⅢP）， and type Ⅳ collagen （Ⅳ-C）］， and traditional Chinese medicine syndrome scores were used as efficacy evaluation indicators. Efficacy assessments and explorations of different staged subgroups of Qijia Rougan prescription were conducted according to LSM values based on the Metavir pathological staging standard. ResultsA total of 98 cases were included for statistical analysis， with 49 cases in the observation group and 49 in the control group. The general data of the patients in both groups were comparable. Compared with the same group before treatment， the observation group showed a significant reduction in LSM and FIB-4 （P<0.01）， as well as notable improvements in LN， Ⅳ-C， and various TCM syndrome scores （P<0.05， P<0.01）. When compared to the control group after treatment， the observation group demonstrated significant improvements in LSM， FIB-4， and various TCM syndrome score indicators （P<0.05， P<0.01）， indicating that the observation group performed better than the control group. Subgroup analysis of the regression of hepatic fibrosis stages showed that compared to the same group before treatment， the observation group had better improvement in regression of stages F2 and F3 （P<0.05）. When compared to the control group after treatment， the observation group exhibited superior improvement in regression of stage F3 （P<0.05）. No adverse events occurred in either group during the treatment period. ConclusionCompared with entecavir alone， the combination of Qijia Rougan prescription and entecavir significantly improves the degree of hepatic fibrosis and clinical TCM symptoms in patients. The optimal intervention period is primarily during stage F3， which is a potential “interception” point of the “Zhu Ke Jiao” theory.

ObjectiveTobacco-related diseases remain one of the leading preventable public health challenges worldwide and are among the primary causes of premature death. In recent years, accumulating evidence has supported the classification of nicotine addiction as a chronic brain disease, profoundly affecting both brain structure and function. Despite the urgency, effective diagnostic methods for smoking addiction remain lacking, posing significant challenges for early intervention and treatment. To address this issue and gain deeper insights into the neural mechanisms underlying nicotine dependence, this study proposes a novel graph neural network framework, termed TI-GNN. This model leverages functional magnetic resonance imaging (fMRI) data to identify complex and subtle abnormalities in brain connectivity patterns associated with smoking addiction. MethodsThe study utilizes fMRI data to construct functional connectivity matrices that represent interaction patterns among brain regions. These matrices are interpreted as graphs, where brain regions are nodes and the strength of functional connectivity between them serves as edges. The proposed TI-GNN model integrates a Transformer module to effectively capture global interactions across the entire brain network, enabling a comprehensive understanding of high-level connectivity patterns. Additionally, a spatial attention mechanism is employed to selectively focus on informative inter-regional connections while filtering out irrelevant or noisy features. This design enhances the model’s ability to learn meaningful neural representations crucial for classification tasks. A key innovation of TI-GNN lies in its built-in causal interpretation module, which aims to infer directional and potentially causal relationships among brain regions. This not only improves predictive performance but also enhances model interpretability—an essential attribute for clinical applications. The identification of causal links provides valuable insights into the neuropathological basis of addiction and contributes to the development of biologically plausible and trustworthy diagnostic tools. ResultsExperimental results demonstrate that the TI-GNN model achieves superior classification performance on the smoking addiction dataset, outperforming several state-of-the-art baseline models. Specifically, TI-GNN attains an accuracy of 0.91, an F1-score of 0.91, and a Matthews correlation coefficient (MCC) of 0.83, indicating strong robustness and reliability. Beyond performance metrics, TI-GNN identifies critical abnormal connectivity patterns in several brain regions implicated in addiction. Notably, it highlights dysregulations in the amygdala and the anterior cingulate cortex, consistent with prior clinical and neuroimaging findings. These regions are well known for their roles in emotional regulation, reward processing, and impulse control—functions that are frequently disrupted in nicotine dependence. ConclusionThe TI-GNN framework offers a powerful and interpretable tool for the objective diagnosis of smoking addiction. By integrating advanced graph learning techniques with causal inference capabilities, the model not only achieves high diagnostic accuracy but also elucidates the neurobiological underpinnings of addiction. The identification of specific abnormal brain networks and their causal interactions deepens our understanding of addiction pathophysiology and lays the groundwork for developing targeted intervention strategies and personalized treatment approaches in the future.

OBJECTIVE: To elucidate the rules of temporal and spatial variations in distal skin potential at Hegu (LI4) under different stimulation modes by extracting nonlinear characteristic parameters from array multichannel data and adopting multivariate statistical analysis. METHODS: Seven healthy subjects were selected and the surface potential at the left Quchi (LI11) was collected using 14×9 array multichannel electrodes. Using Hegu (LI4) on the left as the stimulation point, four stimulation modes were applied, i.e. being quiescent, point pressing, moxibustion, and manual needling manipulation. Electrical signals were collected for 30 s in each mode, with a 5-min interval between operations, and a sampling frequency of 16 384 Hz. The data was denoised using ensemble empirical mode decomposition (EEMD), and sample entropy (SaEn) features were extracted. Statistical analysis was conducted on these data using factor analysis and multivariate analysis of variance. RESULTS: The SaEn values of most electrode channels were higher under point pressing, moxibustion and manual needling manipulation compared with those under quiescent condition. Under manual needling manipulation, the SaEn value of the electrode channel reached the peak in the first time interval (1-5 s) and it was declining thereafter. Factor analysis showed that the specificity of activation channels was concentrated at the left Quchi (LI11) (loading capacity ≥0.90). Analysis of variance indicated that the significant differences were presented in average sample entropy (SaEn()) values of activation channels among different stimulation modes at Hegu (LI4) (P<0.001), but there was no statistically significant interaction effect between groups and time intervals (P>0.05). CONCLUSION Through nonlinear characteristic parameter extraction and multivariate statistical analysis, we have uncovered the complex temporal and spatial dynamical rules of distal skin potential at Hegu (LI4) under various stimulation modes and successfully identified the specific activation characteristics at Quchi (LI11).
Humans ; Moxibustion ; Acupuncture Points ; Male ; Adult ; Female ; Young Adult ; Acupuncture Therapy/instrumentation*

Recent advances in large models demonstrate significant prospects for transforming the field of medical imaging. These models, including large language models, large visual models, and multimodal large models, offer unprecedented capabilities in processing and interpreting complex medical data across various imaging modalities. By leveraging self-supervised pretraining on vast unlabeled datasets, cross-modal representation learning, and domain-specific medical knowledge adaptation through fine-tuning, large models can achieve higher diagnostic accuracy and more efficient workflows for key clinical tasks. This review summarizes the concepts, methods, and progress of large models in medical imaging, highlighting their potential in precision medicine. The article first outlines the integration of multimodal data under large model technologies, approaches for training large models with medical datasets, and the need for robust evaluation metrics. It then explores how large models can revolutionize applications in critical tasks such as image segmentation, disease diagnosis, personalized treatment strategies, and real-time interactive systems, thus pushing the boundaries of traditional imaging analysis. Despite their potential, the practical implementation of large models in medical imaging faces notable challenges, including the scarcity of high-quality medical data, the need for optimized perception of imaging phenotypes, safety considerations, and seamless integration with existing clinical workflows and equipment. As research progresses, the development of more efficient, interpretable, and generalizable models will be critical to ensuring their reliable deployment across diverse clinical environments. This review aims to provide insights into the current state of the field and provide directions for future research to facilitate the broader adoption of large models in clinical practice.
Humans ; Diagnostic Imaging/methods* ; Precision Medicine/methods* ; Image Processing, Computer-Assisted/methods*