Objective: To develop a dual-branch deep learning framework for accurate multi-label classification of fundus diseases, addressing the key limitations of insufficient complementary feature extraction and inadequate cross-modal feature fusion in existing automated diagnostic methods. Methods: The fundus multi-label classification dataset with 12 disease categories (FMLC-12) dataset was constructed by integrating complementary samples from Ocular Disease Intelligent Recognition (ODIR) and Retinal Fundus Multi-Disease Image Dataset (RFMiD), yielding 6 936 fundus images across 12 retinal pathology categories, and the framework was validated on both FMLC-12 and ODIR. Inspired by the holistic multi-regional assessment principle of the Five Wheels theory in traditional Chinese medicine (TCM) ophthalmology, the dual-branch multi-label network (DBMNet) was developed as a novel framework integrating complementary visual feature extraction with pathological correlation modeling. The architecture employed a TransNeXt backbone within a dual-branch design: one branch processed red-green-blue (RGB) images to capture color-dependent features, such as vascular patterns and lesion morphology, while the other processed grayscale-converted images to enhance subtle textural details and contrast variations. A feature interaction module (FIM) effectively integrated the multi-scale features from both branches. Comprehensive ablation studies were conducted to evaluate the contributions of the dual-branch architecture and the FIM. The performance of DBMNet was compared against four state-of-the-art methods, including EfficientNet Ensemble, transfer learning-based convolutional neural network (CNN), BFENet, and EyeDeep-Net, using mean average precision (mAP), F1-score, and Cohen's kappa coefficient. Results: The dual-branch architecture improved mAP by 15.44 percentage points over the single-branch TransNeXt baseline, increasing from 34.41% to 44.24%, and the addition of FIM further boosted mAP to 49.85%. On FMLC-12, DBMNet achieved an mAP of 49.85%, a Cohen’s kappa coefficient of 62.14%, and an F1-score of 70.21%. Compared with BFENet (mAP: 45.42%, kappa: 46.64%, F1-score: 71.34%), DBMNet outperformed it by 4.43 percentage points in mAP and 15.50 percentage points in kappa, while BFENet achieved a marginally higher F1-score. On ODIR, DBMNet achieved an F1-score of 85.50%, comparable to state-of-the-art methods. Conclusion DBMNet effectively integrates RGB and grayscale visual modalities through a dual-branch architecture, significantly improving multi-label fundus disease classification. The framework not only addresses the issue of insufficient feature fusion in existing methods but also demonstrates outstanding performance in balancing detection across both common and rare diseases, providing a promising and clinically applicable pathway for standardized, intelligent fundus disease classification.

OBJECTIVE To systematically evaluate risk prediction models for chemotherapy-induced myelosuppression in pediatric patients with malignant tumors， evaluate their modeling strategies， key predictors， and predictive performance， and provide evidence-based references for clinical decision-making and research. METHODS A literature search was conducted across 11 databases， including CNKI， Wanfang Data， and PubMed， for relevant studies published before April 2025. Two reviewers independently performed literature screening and data extraction， and the risk of bias and applicability of the models were evaluated using the PROBAST tool. RESULTS Ultimately， seven studies were selected， of which four were English articles and three were Chinese articles， involving 12 risk prediction models. Although model discrimination was good （AUC 0.748-0.981）， only two models underwent external validation； furthermore， calibration was inadequately reported in three studies. PROBAST indicated that all models exhibited a high risk of bias， with major issues including a predominance of retrospective designs， inadequate sample representativeness， and lack of blinding. However， in terms of applicability， all models received favorable evaluations. In terms of modeling methods， most studies employed traditional logistic regression approaches to construct models， while only a minority introduced machine learning algorithms and conducted systematic comparisons among multiple algorithms. Models developed using machine learning methods significantly outperformed those constructed with traditional statistical methods. CONCLUSIONS The existing risk prediction models for myelosuppression after chemotherapy in children with malignant tumors demonstrate potential in clinical risk early warning. However， they generally suffer from design and methodological limitations， such as a predominance of retrospective single-center designs， few events per variable， opaque handling of missing data， and inconsistent reporting of model coefficients. Future studies should adopt prospective designs， incorporate machine learning with key clinical predictors， and follow TRIPOD reporting guidelines to enhance scientific rigor and clinical utility.

Based on the theory of "disease of both blood and water" in Essentials from the Golden Cabinet （《金匮要略》）， and in combination with the dynamic syndrome evolution of heart failure with preserved ejection fraction （HFpEF）， this paper systematically clarifies the pathomechanism of HFpEF， characterized by yang deficiency as the root， blood stasis as the pivotal factor and water retention as the manifestation. Accordingly， the therapeutic principles have been proposed， which are warming yang and banking up original qi to consolidate the root， activating blood and unblocking collaterals to smooth the mechanism， and promoting urination and regulating pivot to remove the branch. On this basis， a compound formula structure of "one monarch， one minister and one assistant" is established， forming an integrated intervention strategy that synergistically combines the three methods of warming yang， activating blood， and promoting urination through combined classical formulas. Zhenwu Decoction （真武汤）， which warms yang and dissolves rheum， is used to consolidate the root and directly target the source of yang deficiency， serving as the monarch； Guizhi Fuling Pills （桂枝茯苓丸）， which activates blood， promotes urination and unblocks the pivot， assists in interrupting the binding of blood stasis and water retention， serving as the minister； Tingli Dazao Xiefei Decoction （葶苈大枣泻肺汤）， which regulates qi， disperses retained fluids， and eliminates the manifestation， alleviates acute water-retention symptoms， serving as the assistant. This compound formula is warming without being drying， diuretic without being drastic， and dispels stasis without consuming blood， thereby achieving the therapeutic effects of warming yang， activating blood， and promoting urination.

Objective: To explore the effect of a computer assisted audiovisual combined intervention model on the core symptoms of children with autism spectrum disorder (ASD), in order to provide references for enriching intervention and treatment methods for ASD children. Methods: From December 2023 to March 2024, 36 ASD children aged 4-8 years were recruited from Tianjin Disabled Persons Rehabilitation Center and Xinxinyu Children s Rehabilitation Center, and were divided into a training group (22 cases) and a control group (14 cases). The training group completed a 12 week audiovisual training course (visual sessions:twice a week, for 50-60 minutes each session; auditory sessions:three times a week, for 15 minutes each time), while the control group received only conventional treatment interventions. Before and after the intervention, the core symptoms of ASD children were assessed using the Short Sensory Profile (SSP), Social Responsiveness Scale (SRS), and Repetitive Behavior Questionnaire-2 ( RBQ- 2). Eye tracking experiments were used to test preferences for social attention. Results: Before the intervention, there were no statistically significant differences in the total scores and factor scores of the SSP and RBQ-2 scales between the two groups ( t =-1.63, 0.38, both P >0.05). The SRS total score, social communication, and restricted interests and repetitive behavior factor scores of the training group (90.68±25.83, 33.36±11.80, 15.64±7.00) were significantly higher than those of the control group (72.29±19.84, 24.93±7.85, 10.21±5.67) ( t =2.27, 2.36, 2.43, all P <0.05). Children in the training group with higher social communication factor scores before the intervention scored lower than the control group at the post intervention test (simple slope=-14.17, t =-2.48, P = 0.02), while there was no statistically significant difference in post intervention scores between children with lower social communication factor scores before the intervention and the control group (simple slope=2.31, t =0.57, P >0.05). Eye tracking experiments showed that the total fixation time on geometric images decreased significantly more in the training group [ -4.56 (-11.42, 1.21)] compared to the control group [6.55 (-0.32, 16.53)] after the intervention ( Z=2.48, P <0.05). Conclusions The computer assisted audiovisual intervention model can effectively improve the core symptoms of ASD children with poorer social communication levels. The promotion of the intervention model needs to consider individual differences in ASD.

Background: and Purpose This study aimed to investigate the association between residual inflammatory risk (RIR) and vulnerable plaques using high-resolution magnetic resonance imaging (HRMRI) in symptomatic intracranial atherosclerotic stenosis (ICAS). Methods: This retrospective study included 70%–99% symptomatic ICAS patients hospitalized from January 2016 to December 2022. Patients were classified into four groups based on high-sensitivity C-reactive protein (hs-CRP) and low-density lipoprotein cholesterol (LDL-C): residual cholesterol inflammatory risk (RCIR, hs-CRP ≥3 mg/L and LDL-C ≥2.6 mmol/L), RIR (hs-CRP ≥3 mg/L and LDL-C <2.6 mmol/L), residual cholesterol risk (RCR, hs-CRP <3 mg/L and LDL-C ≥2.6 mmol/L), and no residual risk (NRR, hs-CRP <3 mg/L and LDL-C <2.6 mmol/L). Vulnerable plaque features on HRMRI included positive remodeling, diffuse distribution, intraplaque hemorrhage, and strong enhancement. Results: Among 336 included patients, 21, 60, 58, and 197 were assigned to the RCIR, RIR, RCR, and NRR groups, respectively. Patients with RCIR (adjusted odds ratio [aOR], 3.606; 95% confidence interval [CI], 1.346–9.662; P=0.011) and RIR (aOR, 3.361; 95% CI, 1.774–6.368, P<0.001) had higher risks of strong enhancement than those with NRR. Additionally, patients with RCIR (aOR, 2.965; 95% CI, 1.060–8.297; P=0.038) were more likely to have intraplaque hemorrhage compared with those with NRR. In the sensitivity analysis, RCR (aOR, 2.595; 95% CI, 1.201–5.608; P=0.015) exhibited an additional correlation with an increased risk of intraplaque hemorrhage. Conclusion In patients with symptomatic ICAS, RIR is associated with a higher risk of intraplaque hemorrhage and strong enhancement, indicating an increased vulnerability to atherosclerotic plaques.

The purpose of this paper is to explore the decoction and dosing methods of rhubarb root and rhizome in classical prescriptions and to provide a reference basis for the clinical use of rhubarb root and rhizome. By collating the relevant classical prescriptions of rhubarb root and rhizome in Shanghan Lun and Jingui Yaolüe, the relationship between its decoction and dosing methods and the syndrome was analyzed. The decoction of rhubarb root and rhizome in classical prescriptions can be divided into three categories: simultaneous decoction, decoction later, and other methods (impregnation in Mafei decoction, decoction with water from the well spring first taken in the morning, and pills). If it enters the blood level or wants to slow down, rhubarb root and rhizome should be decocted at the same time with other drugs. If it enters the qi level and wants to speed up, rhubarb root and rhizome should be decocted later. If it wants to upwardly move, rhubarb root and rhizome should be immersed in Mafei decoction. If it wants to suppress liver yang, rhubarb root and rhizome should be decocted with water from the well spring first taken in the morning. If the disease is prolonged, rhubarb root and rhizome should be taken in pill form. The dosing methods of rhubarb root and rhizome can be divided into five categories: draught, twice, three times, before meals, and unspecified. For acute and serious illnesses with excess of pathogenic qi and adequate vital qi, we choose draught. For gastrointestinal diseases, we choose to take the medicine twice. For achieving a moderate and long-lasting effect, we choose to take the medicine three times. If the disease is located in the lower part of the heart and abdomen, we choose to take it before meals. The use of rhubarb root and rhizome in clinical practice requires the selection of the appropriate decoction and dosing methods according to the location of the disease, the severity of the disease, the patient′s constitution, and the condition after taking the medicine.

Electroencephalogram (EEG) is a non-invasive, high temporal-resolution technique for monitoring brain activity. However, affected by the volume conduction effect, EEG has a low spatial resolution and is difficult to locate brain neuronal activity precisely. The surface Laplacian (SL) technique obtains the Laplacian EEG (LEEG) by estimating the second-order spatial derivative of the scalp potential. LEEG can reflect the radial current activity under the scalp, with positive values indicating current flow from the brain to the scalp (“source”) and negative values indicating current flow from the scalp to the brain (“sink”). It attenuates signals from volume conduction, effectively improving the spatial resolution of EEG, and is expected to contribute to breakthroughs in neural engineering. This paper provides a systematic overview of the principles and development of SL technology. Currently, there are two implementation paths for SL technology: current source density algorithms (CSD) and concentric ring electrodes (CRE). CSD performs the Laplace transform of the EEG signals acquired by conventional disc electrodes to indirectly estimate the LEEG. It can be mainly classified into local methods, global methods, and realistic Laplacian methods. The global method is the most commonly used approach in CSD, which can achieve more accurate estimation compared with the local method, and it does not require additional imaging equipment compared with the realistic Laplacian method. CRE employs new concentric ring electrodes instead of the traditional disc electrodes, and measures the LEEG directly by differential acquisition of the multi-ring signals. Depending on the structure, it can be divided into bipolar CRE, quasi-bipolar CRE, tripolar CRE, and multi-pole CRE. The tripolar CRE is widely used due to its optimal detection performance. While ensuring the quality of signal acquisition, the complexity of its preamplifier is relatively acceptable. Here, this paper introduces the study of the SL technique in resting rhythms, visual-related potentials, movement-related potentials, and sensorimotor rhythms. These studies demonstrate that SL technology can improve signal quality and enhance signal characteristics, confirming its potential applications in neuroscientific research, disease diagnosis, visual pathway detection, and brain-computer interfaces. CSD is frequently utilized in applications such as neuroscientific research and disease detection, where high-precision estimation of LEEG is required. And CRE tends to be used in brain-computer interfaces, that have stringent requirements for real-time data processing. Finally, this paper summarizes the strengths and weaknesses of SL technology and envisages its future development. SL technology boasts advantages such as reference independence, high spatial resolution, high temporal resolution, enhanced source connectivity analysis, and noise suppression. However, it also has shortcomings that can be further improved. Theoretically, simulation experiments should be conducted to investigate the theoretical characteristics of SL technology. For CSD methods, the algorithm needs to be optimized to improve the precision of LEEG estimation, reduce dependence on the number of channels, and decrease computational complexity and time consumption. For CRE methods, the electrodes need to be designed with appropriate structures and sizes, and the low-noise, high common-mode rejection ratio preamplifier should be developed. We hope that this paper can promote the in-depth research and wide application of SL technology.

Thoracic aortic aneurysm(TAA)significantly endangers the lives of individuals with Marfan syndrome(MFS),yet the intricacies of their biomechanical origins remain elusive.Our investigation delves into the pivotal role of hemodynamic disturbance in the pathogenesis of TAA,with a particular emphasis on the mechanistic contributions of the mammalian target of rapamycin(mTOR)signaling cascade.We un-covered that activation of the mTOR complex 1(mTORC1)within smooth muscle cells,instigated by the oscillatory wall shear stress(OSS)that stems from disturbed flow(DF),is a catalyst for TAA progression.This revelation was corroborated through both an MFS mouse model(Fbn1+/C1039G)and clinical MFS specimens.Crucially,our research demonstrates a direct linkage between the activation of the mTORC1 pathway and the intensity in OSS.Therapeutic administration of rapamycin suppresses mTORC1 activity,leading to the attenuation of aberrant SMC behavior,reduced inflammatory infiltration,and restoration of extracellular matrix integrity—collectively decelerating TAA advancement in our mouse model.These insights posit the mTORC1 axis as a strategic target for intervention,offering a novel approach to manage TAAs in MFS and potentially pave insights for current treatment paradigms.

OBJECTIVE: To explore the genetic basis for a boy affected with Cohen syndrome. METHODS: A boy admitted to Children's Hospital of Nanjing Medical University in January 2021 was selected as the study subject. Genome DNA was extracted from peripheral blood samples from the child and his parents. Whole exome sequencing (WES) was carried out. And candidate variants were verified by Sanger sequencing. This study was approved by the Medical Ethics Committee of the hospital (Ethics No.: 202106060-1). RESULTS: WES revealed that the child has harbored compound heterozygous variants of the VPS13B gene, namely c.1563+1G>A and c.3007insC (p.A1003Afs*13), which were inherited from his mother and father, respectively. Based on the guidelines from the American College of Medical Genetics and Genomics (ACMG), both variants were rates as pathogenic. The c.3007insC (p.A1003Afs*13) was unreported previously. CONCLUSION The compound heterozygous variants c.1563+1G>A and c.3007insC (p.A1003Afs*13) of the VPS13B gene probably underlay the pathogenesis of Cohen syndrome in this child. Above finding has enriched the mutational spectrum of VPS13B gene.
Humans ; Male ; Vesicular Transport Proteins/genetics* ; Intellectual Disability/genetics* ; Muscle Hypotonia/genetics* ; Microcephaly/genetics* ; Fingers/abnormalities* ; Myopia/genetics* ; Obesity/genetics* ; Developmental Disabilities/genetics* ; Mutation ; Exome Sequencing ; Child ; Heterozygote ; Retinal Degeneration

Developmental dysplasia of hip(DDH)usually occurs in children,and delayed diagnosis of DDH might lead to serious complications and influence long-term prognosis.The application of artificial intelligence(AI)in medical images helps to quantitatively individualize image data,reduce bias generated by manual analysis and achieve early and accurate diagnosis of children DDH.The research progresses of AI in imaging diagnosis of children DDH were reviewed in this article.