OBJECTIVES: We propose a multi-feature fusion model based on manually extracted features and deep learning features from endoscopic images for grading rebleeding risk of peptic ulcers. METHODS: Based on the endoscopic appearance of peptic ulcers, color features were extracted to distinguish active bleeding (Forrest I) from non-bleeding ulcers (Forrest II and III). The edge and texture features were used to describe the morphology and appearance of the ulcers in different grades. By integrating deep features extracted from a deep learning network with manually extracted visual features, a multi-feature representation of endoscopic images was created to predict the risk of rebleeding of peptic ulcers. RESULTS: In a dataset consisting of 3573 images from 708 patients with Forrest classification, the proposed multi-feature fusion model achieved an accuracy of 74.94% in the 6-level rebleeding risk classification task, outperforming the experienced physicians who had a classification accuracy of 59.9% (P<0.05). The F1 scores of the model for identifying Forrest Ib, IIa, and III ulcers were 90.16%, 75.44%, and 77.13%, respectively, demonstrating particularly good performance of the model for Forrest Ib ulcers. Compared with the first model for peptic ulcer rebleeding classification, the proposed model had improved F1 scores by 5.8%. In the simplified 3-level risk (high-risk, low-risk, and non-endoscopic treatment) classification task, the model achieved F1 scores of 93.74%, 81.30%, and 73.59%, respectively. CONCLUSIONS The proposed multi-feature fusion model integrating deep features from CNNs with manually extracted visual features effectively improves the accuracy of rebleeding risk classification for peptic ulcers, thus providing an efficient diagnostic tool for clinical assessment of rebleeding risks of peptic ulcers.
Humans ; Deep Learning ; Peptic Ulcer ; Risk Assessment ; Peptic Ulcer Hemorrhage ; Recurrence

Adolescent depression is increasingly recognized as a serious mental health disorder with distinct clinical and molecular features. Using single-nucleus RNA sequencing, we identified cell-specific transcriptomic changes in the nucleus accumbens (NAc), particularly in astrocytes, of adolescent macaques exhibiting depressive-like behaviors. The level of diacylglycerol kinase beta was significantly reduced in neurons and glial cells of depressed macaques, while FKBP5 levels increased in glial cells. Disruption of GABAergic synapses and disruption of D-glutamine and D-glutamate metabolism were linked to depressive phenotypes in medium spiny neurons (MSNs) and subtypes of astrocytes. Communication pathways between astrocytes and D1/D2-MSNs were also disrupted, involving factors like bone morphogenetic protein-6 and Erb-B2 receptor tyrosine kinase-4. Bulk transcriptomic and proteomic analyses corroborated these findings, and FKBP5 upregulation was confirmed by qRT-PCR, western blotting, and immunofluorescence in the NAc of rats and macaques with chronic unpredictable mild stress. Our results highlight the specific roles of different cell types in adolescent depression in the NAc, offering potential targets for new antidepressant therapies.
Animals ; Nucleus Accumbens/metabolism* ; Male ; Transcriptome ; Depression/genetics* ; Astrocytes/metabolism* ; Neurons/metabolism* ; Rats

Drug development remains a critical issue in the field of biomedicine. With the rapid advancement of information technologies such as artificial intelligence (AI) and the advent of the big data era, AI-assisted drug development has become a new trend, particularly in predicting drug-target associations. To address the challenge of drug-target prediction, AI-driven models have emerged as powerful tools, offering innovative solutions by effectively extracting features from complex biological data, accurately modeling molecular interactions, and precisely predicting potential drug-target outcomes. Traditional machine learning (ML), network-based, and advanced deep learning architectures such as convolutional neural networks (CNNs), graph convolutional networks (GCNs), and transformers play a pivotal role. This review systematically compiles and evaluates AI algorithms for drug- and drug combination-target predictions, highlighting their theoretical frameworks, strengths, and limitations. CNNs effectively identify spatial patterns and molecular features critical for drug-target interactions. GCNs provide deep insights into molecular interactions via relational data, whereas transformers increase prediction accuracy by capturing complex dependencies within biological sequences. Network-based models offer a systematic perspective by integrating diverse data sources, and traditional ML efficiently handles large datasets to improve overall predictive accuracy. Collectively, these AI-driven methods are transforming drug-target predictions and advancing the development of personalized therapy. This review summarizes the application of AI in drug development, particularly in drug-target prediction, and offers recommendations on models and algorithms for researchers engaged in biomedical research. It also provides typical cases to better illustrate how AI can further accelerate development in the fields of biomedicine and drug discovery.

Drug development remains a critical issue in the field of biomedicine.With the rapid advancement of information technologies such as artificial intelligence(AI)and the advent of the big data era,AI-assisted drug development has become a new trend,particularly in predicting drug-target associations.To address the challenge of drug-target prediction,AI-driven models have emerged as powerful tools,of-fering innovative solutions by effectively extracting features from complex biological data,accurately modeling molecular interactions,and precisely predicting potential drug-target outcomes.Traditional machine learning(ML),network-based,and advanced deep learning architectures such as convolutional neural networks(CNNs),graph convolutional networks(GCNs),and transformers play a pivotal role.This review systematically compiles and evaluates AI algorithms for drug-and drug combination-target predictions,highlighting their theoretical frameworks,strengths,and limitations.CNNs effectively identify spatial patterns and molecular features critical for drug-target interactions.GCNs provide deep insights into molecular interactions via relational data,whereas transformers increase prediction accu-racy by capturing complex dependencies within biological sequences.Network-based models offer a systematic perspective by integrating diverse data sources,and traditional ML efficiently handles large datasets to improve overall predictive accuracy.Collectively,these AI-driven methods are transforming drug-target predictions and advancing the development of personalized therapy.This review summa-rizes the application of AI in drug development,particularly in drug-target prediction,and offers rec-ommendations on models and algorithms for researchers engaged in biomedical research.It also provides typical cases to better illustrate how AI can further accelerate development in the fields of biomedicine and drug discovery.

Abstract Adolescent health literacy constitutes a fundamental, economical and effective strategy for addressing their health issues and fostering healthy behaviors, while assessing health literacy plays a pivotal role in evaluating adolescents health literacy. The study systematically reviews existing adolescent health literacy assessment tools at both domestically and internationally, and analyzes them through three dimensions:structural components, applicability and scientific validity. It further examines emerging trends in the development of such tools, aiming to offer theoretical underpinnings and practical recommendations for their refinement, thereby more effectively addressing the evolving health needs of adolescents.

Objective: This study aimed to explore the root length of maxillary and mandibular anterior teeth and central incisor crown-root morphology in patients with high-angle skeletal Class Ⅱ open bite, aiming to provide a reference for clinical treatment. . Methods: This study was reviewed and approved by the Ethics Committee, and informed consent was obtained from the patients. CBCT images of eighty-one untreated patients (40 anterior open bite patients and 41 normal overbite patients) with high-angle skeletal Class Ⅱ malocclusion were selected before treatment. Dolphin software was used to study the root length of maxillary and mandibular anterior teeth and central incisor crown-root morphology, and the differences between the two groups were analyzed. Results: There was no statistical significance in the root length of maxillary lateral incisor and canine between the open bite group and the normal overbite group, significant differences were found in the root length of maxillary central incisor （11.12 ± 1.37） mm、mandibular central incisor（10.15 ± 1.09）mm, mandibular lateral incisor（11.27 ± 1.15）mm and mandibular canine（12.81 ± 1.48）mm between the open bite group and the normal overbite group（P<0.05）. On the other hand, the two groups were significantly different in crown-root morphology of the maxillary central incisor (1.10° ± 3.62° vs. 4.53° ± 2.30°, P<0.01) but not in the mandibular central incisor. Conclusion The root length of the maxillary central incisor, mandibular central incisor, mandibular lateral incisor, mandibular canine in high-angle Class Ⅱ open bite patients is shorter than that in high-angle Class Ⅱ normal overbite patients, and the long axis of the crown of the maxillary central incisor in high-angle Class Ⅱ open bite patients obviously deviates toward the labial side relative to the long axis of the root. The crown-root angle is smaller, which is beneficial to torque control or adduction movement of the anterior teeth in high-angle Class Ⅱ open bite patients.

Objective:To compare demographic characteristics,clinical characteristics,therapeutic characteris-tics and physiological indicators of patients with bipolar Ⅰ disorder and bipolar Ⅱ disorder.Methods:A total of 381 patients with bipolar disorder(BD)diagnosed by the Diagnostic and Statistical Manual of Mental Disorders 5 th Edi-tion(DSM-5)were selected,including 302 patients with BD-Ⅰ(79.27％),74 patients with BD-Ⅱ(19.42％)and 5 patients with other specific and related disorders(1.31％).Demographic and clinical characteristics were collected with self-designed clinical information questionnaire.Multivariate logistic regression and multivariate linear regres-sion analysis were used for analysis.Results:Compared with patients with BD-Ⅱ,patients with BD-Ⅰ had more risk to have psychotic features(OR=5.75,95％CI:2.82-11.76),longer disease duration,and more repeated transcra-nial magnetic therapy(OR=3.09,95％CI:1.02-9.35),higher uric acid,total cholesterol and high-density lipo-protein.BD-Ⅰ in Han nationality was more common(OR=11.50,95％CI:1.76-75.30),and had lower education level(OR=10.22,95％CI:1.16-89.77),and less family history of psychosis(OR=2.34,95％CI:1.01-5.42).Conclusion:There are significant differences between BD-Ⅰ and BD-Ⅱ in demographic and clinical charac-teristics,treatment status,and physiological indicators,which could provide clues for exploring the pathogenesis of bipolar disorder.

Background and purpose:For patients with human epidermal growth factor receptor 2(HER2)-positive metastatic breast cancer,trastuzumab treatment can prolong the overall survival and significantly improve the prognosis of patients.However,the reference original research trastuzumab(Herceptin?)is more expensive.Biosimilars have comparable efficacy and safety profiles while increasing patient access to treatment.This clinical trial aimed to evaluate the efficacy,pharmacokinetics,safety and immunogenicity of the trastuzumab biosimilar AK-HER2 compared to trastuzumab(Herceptin?)in patients with HER2-positive metastatic breast cancer.Methods:This multi-center,randomised,double-blind phase Ⅲ clinical trial was conducted in 43 subcenters in China.This study complied with the research protocol,the ethical principles stated in the Declaration of Helsinki and the quality management standards for drug clinical trials.It was approved by the hospital's medical ethics committee.The clinical trial registration agency is the State Food and Drug Administration(clinical trial approval number:2015L04224;clinical trial registration number:CTR20170516).Written informed consent was obtained from subjects before enrollment.Enrolled patients were randomly assigned to the AK-HER2 group and the control group,respectively receiving AK-HER2 or trastuzumab(initial loading dose 8 mg/kg,maintenance dose 6 mg/kg,every 3 weeks as a treatment cycle,total treatment time is 16 cycles)in combination with docetaxel(75 mg/m2,treatment duration is at least 9 cycles).The primary endpoint of this clinical trial was the objective response rate(ORR9)between the AK-HER2 group and the control group in the 9th cycle.Secondary efficacy endpoints included ORR16,disease control rate(DCR),clinical benefit rate(CBR),progression-free survival(PFS)and 1-year survival rate.In this study,100 subjects(AK-HER2 group to control group=1:1)were randomly selected for blood sample collection after the 6th cycle of medication,The collection time points were 45 minutes after infusion(the end of administration),4,8,24,72,120,168,336,and 504 hours after the end of administration.After collection,blood samples were analyzed by PK parameter set(PKPS).Other evaluation parameters included safety and immunogenicity assessment.Results:A total of 550 patients with HER2-positive metastatic breast cancer were enrolled in this clinical trial between Sep.2017 and Mar.2021.In the AK-HER2 group(n=237),129 subjects in the experimental group achieved complete response(CR)or partial response(PR),and the ORR9 was 54.4%.There were 134 subjects in the control group(n=241)who achieved CR or PR,and the ORR9 was 55.6%.The ORR9 ratio between the AK-HER2 group and the control group was 97.9%[90%confidence interval(CI):85.4%-112.2%,P=0.784],which was not statistically significant.In all secondary efficacy endpoints,no statistically significant differences were observed between the two groups.We conducted a mean ratio analysis of pharmacokinetics(PK)parameters between the AK-HER2 group and the control group,and the results suggested that the pharmacokinetic characteristics of the two drugs are similar.The incidence of treatment emergent adverse event(TEAE)leading to drug reduction or suspension during trastuzumab treatment was 3.6%(10 cases)in the AK-HER2 group and 8.1%(22 cases)in the control group.There was statistically significant difference between the two groups(P=0.027).The incidence rate was significantly lower in the AK-HER2 group than in the control group,and there was no statistically significant difference among the other groups.The differences in the positive rates of anti-drug antibodies(ADA)and neutralizing antibodies(NAB)between groups were of no statistical significance(P=0.385 and P=0.752).Conclusion:In patients with HER2-positive metastatic breast cancer,AK-HER2 was comparable to the trastuzumab(Herceptin?)in terms of drug efficacy,pharmacokinetics,safety and immunogenicity.