Objective:To construct a "local-systemic" dual-track prediction model integrating the resistance index (RI) score of bilateral sacroiliac joints and the systemic immune-inflammation index (SII), and to evaluate its predictive efficacy for the active stage of ankylosing spondylitis (AS).Methods:A total of 205 patients with ankylosing spondylitis (AS) from the Second Hospital of Lanzhou University between April 2022 and April 2025 were retrospectively enrolled and categorized into an active group ( n=113) and a remission group ( n=92). Hematological parameters and ultrasound data were collected. The resistance index (RI) of the synovial area in bilateral sacroiliac joints was measured by Doppler ultrasound and scored as follows: RI < 0.5: 3 points; RI 0.5~0.55: 2 points; RI > 0.55: 1 point; undetectable blood flow: 0 points. A total bilateral RI score (range 0 to 6) was calculated. The systemic immune-inflammation index (SII) was derived as (neutrophils× platelets)/lymphocytes. Normality was tested for all continuous variables; normally distributed data were compared using the t-test, while non-normally distributed data were analyzed with the Mann-Whitney U test. Categorical variables were compared using the χ2 test or analysis of variance.Variable selection was performed using Lasso regression, and a multivariate logistic regression model was developed to assess predictive performance. Results:The proportion of patients with a bilateral RI total score≥5 was significantly higher in the active group compared to the remission group (50 of 113, 44.3% vs 2 of 92, 2.2%, χ2=55.63, P<0.001). Multivariate logistic regression analysis, after adjustment for confounding variables, identified the SII [ OR(95% CI)=1.01(1.00, 1.01), P<0.001], bilateral RI total score [ OR(95% CI)=1.67(1.29, 2.26), P<0.001], erythrocyte sedimentation rate [ OR(95% CI)=1.19(1.11, 1.30), P<0.001], and mean corpuscular hemoglobin concentration [ OR(95% CI)=1.09(1.03, 1.17), P<0.001] as independent risk factors for active AS. Conversely, lymphocyte count [ OR(95% CI)=0.42(0.18, 0.92), P=0.030] and globulin [ OR(95% CI)=0.89(0.80, 0.99), P=0.040] were significantly associated with protective effects. The bilateral RI total score demonstrated the strongest predictive effect, with each 1-point increase associated with a 67% elevation in the risk of active disease. ROC curve analysis indicated that the area under the curve (AUC) for predicting whether AS is in the active disease phase was 0.94 for the combined model (SII+bilateral RI total score), compared with 0.93 for the SII-alone model and 0.92 for the bilateral RI total score-alone model, demonstrating superior predictive performance of the combined model (SII+bilateral RI total score). An online prediction tool has been developed based on the combined model. Conclusion:The dual-track prediction model, which integrates local joint hemodynamic characteristics and systemic immune-inflammatory status, facilitates a multidimensional assessment of the risk of active AS and provides an objective basis for early identification.

Objective:To construct a "local-systemic" dual-track prediction model integrating the resistance index (RI) score of bilateral sacroiliac joints and the systemic immune-inflammation index (SII), and to evaluate its predictive efficacy for the active stage of ankylosing spondylitis (AS).Methods:A total of 205 patients with ankylosing spondylitis (AS) from the Second Hospital of Lanzhou University between April 2022 and April 2025 were retrospectively enrolled and categorized into an active group ( n=113) and a remission group ( n=92). Hematological parameters and ultrasound data were collected. The resistance index (RI) of the synovial area in bilateral sacroiliac joints was measured by Doppler ultrasound and scored as follows: RI < 0.5: 3 points; RI 0.5~0.55: 2 points; RI > 0.55: 1 point; undetectable blood flow: 0 points. A total bilateral RI score (range 0 to 6) was calculated. The systemic immune-inflammation index (SII) was derived as (neutrophils× platelets)/lymphocytes. Normality was tested for all continuous variables; normally distributed data were compared using the t-test, while non-normally distributed data were analyzed with the Mann-Whitney U test. Categorical variables were compared using the χ2 test or analysis of variance.Variable selection was performed using Lasso regression, and a multivariate logistic regression model was developed to assess predictive performance. Results:The proportion of patients with a bilateral RI total score≥5 was significantly higher in the active group compared to the remission group (50 of 113, 44.3% vs 2 of 92, 2.2%, χ2=55.63, P<0.001). Multivariate logistic regression analysis, after adjustment for confounding variables, identified the SII [ OR(95% CI)=1.01(1.00, 1.01), P<0.001], bilateral RI total score [ OR(95% CI)=1.67(1.29, 2.26), P<0.001], erythrocyte sedimentation rate [ OR(95% CI)=1.19(1.11, 1.30), P<0.001], and mean corpuscular hemoglobin concentration [ OR(95% CI)=1.09(1.03, 1.17), P<0.001] as independent risk factors for active AS. Conversely, lymphocyte count [ OR(95% CI)=0.42(0.18, 0.92), P=0.030] and globulin [ OR(95% CI)=0.89(0.80, 0.99), P=0.040] were significantly associated with protective effects. The bilateral RI total score demonstrated the strongest predictive effect, with each 1-point increase associated with a 67% elevation in the risk of active disease. ROC curve analysis indicated that the area under the curve (AUC) for predicting whether AS is in the active disease phase was 0.94 for the combined model (SII+bilateral RI total score), compared with 0.93 for the SII-alone model and 0.92 for the bilateral RI total score-alone model, demonstrating superior predictive performance of the combined model (SII+bilateral RI total score). An online prediction tool has been developed based on the combined model. Conclusion:The dual-track prediction model, which integrates local joint hemodynamic characteristics and systemic immune-inflammatory status, facilitates a multidimensional assessment of the risk of active AS and provides an objective basis for early identification.

Traditional Chinese medicine(TCM)is an ancient medical system distinctive and effective in treating cancer,depression,coronavirus disease 2019(COVID-19),and other diseases.However,the relatively abstract diagnostic methods of TCM lack objective measurement,and the complex mechanisms of action are difficult to comprehend,which hinders the application and internationalization of TCM.Recently,while breakthroughs have been made in utilizing methods such as network pharmacology and virtual screening for TCM research,the rise of machine learning(ML)has significantly enhanced their inte-gration with TCM.This article introduces representative methodological cases in quality control,mechanism research,diagnosis,and treatment processes of TCM,revealing the potential applications of ML technology in TCM.Furthermore,the challenges faced by ML in TCM applications are summarized,and future directions are discussed.

Idiopathic pulmonary fibrosis (IPF), a chronic interstitial lung disease, is characterized by aberrant wound healing, excessive scarring and the formation of myofibroblastic foci. Although the role of alternative splicing (AS) in the pathogenesis of organ fibrosis has garnered increasing attention, its specific contribution to pulmonary fibrosis remains incompletely understood. In this study, we identified an up-regulation of serine/arginine-rich splicing factor 7 (SRSF7) in lung fibroblasts derived from IPF patients and a bleomycin (BLM)-induced mouse model, and further characterized its functional role in both human fetal lung fibroblasts and mice. We demonstrated that enhanced expression of Srsf7 in mice spontaneously induced alveolar collagen accumulation. Mechanistically, we investigated alternative splicing events and revealed that SRSF7 modulates the alternative splicing of pyruvate kinase (PKM), leading to metabolic dysregulation and fibroblast activation. In vivo studies showed that fibroblast-specific knockout of Srsf7 in conditional knockout mice conferred resistance to bleomycin-induced pulmonary fibrosis. Importantly, through drug screening, we identified lomitapide as a novel modulator of SRSF7, which effectively mitigated experimental pulmonary fibrosis. Collectively, our findings elucidate a molecular pathway by which SRSF7 drives fibroblast metabolic dysregulation and propose a potential therapeutic strategy for pulmonary fibrosis.

Traditional Chinese medicine (TCM) is an ancient medical system distinctive and effective in treating cancer, depression, coronavirus disease 2019 (COVID-19), and other diseases. However, the relatively abstract diagnostic methods of TCM lack objective measurement, and the complex mechanisms of action are difficult to comprehend, which hinders the application and internationalization of TCM. Recently, while breakthroughs have been made in utilizing methods such as network pharmacology and virtual screening for TCM research, the rise of machine learning (ML) has significantly enhanced their integration with TCM. This article introduces representative methodological cases in quality control, mechanism research, diagnosis, and treatment processes of TCM, revealing the potential applications of ML technology in TCM. Furthermore, the challenges faced by ML in TCM applications are summarized, and future directions are discussed.

Objective A Nomogram model of cognitive frailty was constructed and validated in hospitalized older adults,providing a reference for early screening,intervention and personalized management of cognitive frailty.Methods A convenience sampling approach was employed to recruit 322 elderly inpatients from a tertiary hospital in Beijing between October 2024 and February 2025 as study participants,and data were collected using the General Information Questionnaire,the Short Form-Mini-Nutritional Assessment,the Asens Insomnia Scale,the Activity of Daily Living Rating,the Self-Rating Anxiety Scale,the Geriatric Depression Scale-15,the Social Support Rating Scale,the Frailty Phenotype scale,the Subjective Cognitive Decline Questionnaire-9,the Mini-Mental State Examination,and the Clinical Dementia Rating.Lasso-Logistic regression was used to screen the variables,R software was used to draw the nomogram model;Bootstrap method was used for internal validation.Results Lasso-Logistic regression screened 8 predictors of age,depression,anxiety,support utilization,nutritional status,literacy,physical activity,and chronic pain,with an area under the subject operating characteristic curve of 0.830(95％CI:0.787-0.873),a sensitivity of 0.764,a specificity of 0.730,an accuracy of 0.748,and a calibrated curve,Brier score,and Hosmer-Lemeshow test(P=0.774)all showed that the model fit was good.Conclusion The Lasso-Logistic regression-based nomogram model of cognitive frailty in hospitalized older adults has good predictive performance and clinical utility,and can be used as a reference for early identification and intervention of cognitive decline in hospitalized older adults.

Objective:To analyze the burden and trends of noncommunicable chronic disease(NCD) attributable to metabolic factors in China from 1990 to 2021.Methods:Data from the Global Burden of Diseases(GBD) 2021 database were utilized to describe changes in mortality and disability-adjusted life years(DALYs) of NCD in China from 1990 to 2021. Stratified analyses were conducted by age, sex, sociodemographic index(SDI), and related risk factors. Statistical analyses and predictions were conducted using the age-period-cohort model and the Nordpred model.Results:In 2021, the age-standardized mortality rate and age-standardized DALYs rate of NCD attributable to metabolic factors in China were 227.56 per 100 000 and 4 829.39 per 100 000, respectively. Their average annual percentage changes were -0.76%( P<0.001) and -0.77%( P<0.001). Overall, the burden decreased progressively with higher SDI levels. Analysis using the age-period-cohort model indicated reduced birth cohort and period effects for metabolic factor-attributable NCD, while age effects rose significantly. The minimum relative risk( RR) value was observed in the 15-19 age group( RR=0.01), and the maximum RR value occurred in the 95-99 age group( RR=996.86). The overall rising mortality trend indicated that age effects are the predominant driver at present. Projections estimate that by 2046, deaths from metabolic factor-attributable NCD in China will reach 8 189 563, with an age-standardized mortality rate of 236.95 per 100 000. Conclusions:China continues to face a substantial burden of NCD linked to metabolic factors, with older adults, males, and individuals with hypertension, diabetes, and prediabetes identified as key populations requiring targeted interventions.

Objective:To explore the focus points and emotional tendencies of cancer patients regarding medical services from their free-form comments, and to analyze the elements of perceived medical service value by patients.Methods:A tertiary general hospital in Shanghai invited patients to evaluate inpatient services through the assessment module of a patient satisfaction management platform, which included open-ended questions in the patient satisfaction survey. This study extracted free-form comment texts from cancer patients regarding medical services collected from September 2022 to August 2023 from this platform, forming a corpus of patient opinions. A hybrid method combining manual coding and natural language processing techniques was used to conduct sentiment analysis and topic modeling on the free-form comment texts. Descriptive analysis was performed on the comment themes and emotions, and logistic regression analysis was used to explore the relationship between patient negative comments and patient satisfaction and loyalty.Results:A total of 10 446 comment texts provided by cancer patients were included. The most frequently targeted objects in the comments were nurses (2 223 cases), followed by doctors (1 125 cases) and nursing assistants (162 cases). The results of sentiment analysis showed that 7 058 patients made positive comments, 1 958 patients made negative comments, 1 114 patients made neutral comments, and 316 patients made mixed comments. The results of topic analysis showed that the five most frequently mentioned themes by cancer patients were humanistic care (1 327 times), information provision, communication, and education (589 times), diet (582 times), effectiveness of diagnosis and treatment (506 times), and medical staff responsibility (489 times). Logistic regression analysis showed that negative evaluations of humanistic care ( OR=0.306, P<0.001), professional technical level ( OR=0.425, P=0.010), information provision, communication, and education ( OR=0.475, P<0.001), and response to needs ( OR=0.412, P=0.026) were significantly associated with decreased patient satisfaction. Negative evaluations of humanistic care ( OR=0.407, P<0.001), professional technical level ( OR=0.466, P=0.009), and information provision, communication, and education ( OR=0.557, P<0.001) were significantly associated with decreased patient loyalty. Conclusions:Hospital managers should value the role of patient free-form comment texts in improving medical services and make full use of natural language processing techniques for information mining. In addition, the results of this study show that patient comments mainly involve the interpersonal and functional aspects of service, especially the interpersonal interactions, which are often the " critical moments" for patients to critically evaluate hospital services during their service experience.