This comprehensive review systematically explores the multifaceted applications, inherent challenges, and promising future directions of artificial intelligence (AI) within the medical domain. It meticulously examines AI's specific contributions to basic medical research, disease prevention, intelligent diagnosis, treatment, rehabilitation, nursing, and health management. Furthermore, the review delves into AI's innovative practices and pivotal roles in clinical trials, hospital administration, medical education, as well as the realms of medical ethics and policy formulation. Notably, the review identifies several key challenges confronting AI in healthcare, encompassing issues such as inadequate algorithm transparency, data privacy concerns, absent regulatory standards, and incomplete risk assessment frameworks. Looking ahead, the future trajectory of AI in healthcare encompasses enhancing algorithm interpretability, propelling generative AI applications, establishing robust data-sharing mechanisms, refining regulatory policies and standards, nurturing interdisciplinary talent, fostering collaboration among industry, academia, and medical institutions, and advancing inclusive, personalized precision medicine. Emphasizing the synergy between AI and emerging technologies like 5G, big data, and cloud computing, this review anticipates a new era of intelligent collaboration and inclusive sharing in healthcare. Through a multidimensional analysis, it presents a holistic overview of AI's medical applications and development prospects, catering to researchers, practitioners, and policymakers in the healthcare sector. Ultimately, this review aims to catalyze the deep integration and innovative deployment of AI technology in healthcare, thereby driving the sustainable advancement of smart healthcare.

OBJECTIVE To establish population pharmacokinetic model of levetiracetam （Lev） for Chinese patients with post- stroke epilepsy （PSE）， and provide reference for formulating individualized dosing regimens for Lev therapy in this specific population. METHODS Blood concentration data and clinical diagnosis and treatment information of PSE patients meeting the inclusion criteria were retrospectively collected and divided into model group and validation group at an 8∶2 ratio using a random number method. Based on the model group data， a population pharmacokinetic model was developed using nonlinear mixed-effects modeling. Internal evaluation was performed through goodness-of-fit tests and bootstrap analysis， while external validation was conducted using the validation group data. RESULTS A total of 75 blood concentration measurements from 70 PSE patients were collected， with 60 measurements from 55 patients used for model development and 15 measurements from 15 patients reserved for external validation. The final model estimated a population typical value of clearance at 2.98 L/h. Estimated glomerular filtration rate， daily dose， and homocysteine level significantly influenced clearance of Lev （P＜0.01）. The model demonstrated satisfactory predictive performance， as evidenced by goodness-of-fit tests， bootstrap analysis， and external validation results. CONCLUSIONS Daily dose， estimated glomerular filtration rate， and homocysteine level are identified as significant covariates influencing Lev clearance in Chinese PSE patients. When making clinical decisions， comprehensive consideration should be given to the patient’s treatment response， physiological and pathological conditions， and the occurrence of adverse reactions， etc. The dosage of Lev should be adjusted based on the results of population pharmacokinetic model.

As the volume of medical research using large language models (LLM) surges, the need for standardized and transparent reporting standards becomes increasingly critical. In January 2025, Nature Medicine published statement titled by TRIPOD-LLM reporting guideline for studies using large language models. This represents the first comprehensive reporting framework specifically tailored for studies that develop prediction models based on LLM. It comprises a checklist with 19 main items (encompassing 50 sub-items), a flowchart, and an abstract checklist (containing 12 items). This article provides an interpretation of TRIPOD-LLM’s development methods, primary content, scope, and the specific details of its items. The goal is to help researchers, clinicians, editors, and healthcare decision-makers to deeply understand and correctly apply TRIPOD-LLM, thereby improving the quality and transparency of LLM medical research reporting and promoting the standardized and ethical integration of LLM into healthcare.

The irrational use of Chinese patent medicines （CPM） is becoming more and more prominent， which makes the demand for clinical practice guidelines of CPM gradually increase. In order to make domestic scholars understand the latest developments and existing problems of the CPM guidelines， and promote its development， this paper introduced the concept of CPM guidelines， summarized the characteristics of the two development modes， namely “taking CPM as the key” and “taking disease/syndrome as the key”， and analyzed the current methodological status of developing and reporting CPM guidelines. Based on the existed problems， three suggestions have been put forward to optimize the quality of CPM guidelines， which were clarifying the target users and scope of CPM guidelines， establishing an open and transparent mechanism of the personnel involvement and process steps， and formulating implementable and operable recommendations for the use of CPM.

ObjectTo explore the risk factors related to the intensity of post-stroke depression in patients with cerebral infarction during hospitalization in the rehabilitation department. MethodsThe hospital consultation records of cerebral infarction patients in Beijing Bo'ai Hospital from December, 2019 to February, 2023 were reviewed from the hospital information system, and those who were diagnosed as depression visited the department of psychology were selected. It was collected including general information of sexes, ages, education levels, matrimony; medical features of course, location, affected side, sensory disorders, aphasia, agrypnia, dysphagia, hand-shoulder syndrome, constipation; functioning of muscle strength and Brunnstrom stages; and scores of Mini-Mental State Examination (MMSE), National Institutes of Health Stroke Scale (NIHSS), Fugl-Meyer Assessment (FMA), Fugl-Meyer Assessment-Balance (FMA-B), modified Barthel Index (MBI) and Hamilton Depression Scale (HAMD). Patients with HAMD scores ≤ 20 were as the low group, and those > 20 were as the high group. ResultA total of 2 403 hospitalized stroke patients were included, out of which 269 patients with cerebral infarction were diagnosed as depression and visited the department of psychology; while 103 cases were in the low group and 166 cases were in the high group. The incidence of constipation was less, and the incidence of dysphagia and shoulder-hand syndrome was higher in the high group (χ² > 5.379, P < 0.05), with weaker strength of iliopsoas muscle and quadriceps muscle, earlier of Brunnstrom stage of lower extremities and hands, and worse scores of NIHSS, MMSE, FMA, FMA-B and MBI (|Z| > 2.020, t > 2.171, P < 0.05). Logistic regression showed that constipation (OR = 0.435), quadriceps muscle strength (OR = 0.782) and dysphagia (OR = 2.602) related to the intensity of post-stroke depression in convalescent patients (P < 0.05). ConclusionPost-stroke dysphagia and poor quadriceps muscle strength may exacerbate post-stroke depression; however, constipation may not.

OBJECTIVE To provide a detailed report and interpretation of the method and results for determining the weights of the technical indicators from the “multi-dimensional and multi-criteria comprehensive evaluation index system （first edition）” stated in Guideline for Multi-dimensional and Multi-criteria Comprehensive Evaluation of Chinese Patent Medicine. METHODS Normalization calculations were performed on the comprehensive weight values calculated by the analytic hierarchy process and expert weighting method to obtain the objective weights of the indicators. RESULTS The weight results of the six primary dimensions in the current comprehensive evaluation indicator system of Chinese patent medicine showed effectiveness dimension＞ safety dimension＞standard dimension＞application dimension＞scientific dimension＞economic dimension， with weight values of 0.281 0， 0.268 5， 0.195 8， 0.107 3， 0.096 1 and 0.051 3 respectively， consistent with the results of most researches currently. CONCLUSIONS The process of weight determination in this indicator system is scientifically reasonable， with clear methods and clear interpretations， and is worthy of further optimization and widespread application.

Lecanemab is a new drug used to treat early Alzheimer's disease (AD) with mild cognitive impairment or mild dementia. It is a human anti-Aβ fibril monoclonal IgG1 antibody, which is injected intravenously into the patient, through the blood-brain barrier into the brain, clearing amyloid plaque, thereby slowing the rate of cognitive decline in patients and delaying disease progression. This article reviews the pharmacological studies, clinical studies, safety and limitations of lecanemab, in order to help clinical understand the current research status and existing achievements of this drug.

Objective:To establish a 14-day prognosis model for emergency patients with acute ischemic cerebral stroke and evaluate its predictive efficacy.Methods:A prospective cohort study was conducted. Patients with acute ischemic stroke admitted to the emergency department of Beijing Bo’ai Hospital within 72 hours of onset from October 2018 to December 2020 were enrolled. Univariate and multivariate logistic regression analysis were used to screen the risk factors of poor prognosis. The ROC curve was drawn to determine the cut-off value of continuous variables and discretise data with reference to clinical practice. The corresponding scores were set up according to the β regression coefficient of each variable, and the clinical scale prediction model of short-term prognosis of acute cerebral infarction was established. Patients with ischemic stroke in the hospital from January to December 2021 were selected as the internal validation, to verify the constructed predictive model.Results:A total of 321 patients were included in the study, including 223 in the training set and 98 in the internal validation set. Multivariate logistic regression analysis showed that age, hypersensitive C-reactive protein, prealbumin (PA), infarct volume, Frailty Screening Questionnaire (FSQ) and National Institute of Health Stroke Scale (NIHSS) were independent risk factors for poor short-term prognosis of acute cerebral infarction. The total score of the clinical prediction scoring system for short-term prognosis of acute cerebral infarction in the emergency department was 15 points, including age ≥74 years (1 point), PA ≤373 mg/L (2 points), large artery atherosclerosis (1 point), cardiogenic embolism (2 points), infarct volume ≥ 2.18 cm 3 (2 points), FSQ ≥3 points (1 point), NIHSS ≥4 points (6 points); The area under the ROC curve (AUC) of the scoring system for predicting short-term poor prognosis of acute cerebral infarction was 0.927 (95% CI: 0.894-0.960). The optimal cut-off value was ≥5 points, and the sensitivity and specificity were 0.770 and 0.976, respectively. In the internal validation set, the scoring system had similar predictive value for poor outcomes (AUC=0.892, 95% CI:0.827-0.957). Conclusion:The scoring system for short-term prognosis prediction of acute ischemic cerebral infarction has good diagnostic efficacy, and could guide clinicians to judge the prognosis of emergency patients in the early stage.

Objective:To investigate the value of radiomics nomogram based on standardized pre-treatment chest enhanced CT in predicting the mutation status of epidermal growth factor receptor (EGFR) for patients with lung adenocarcinoma.Methods:A retrospective analysis was conducted on pre-treatment chest enhanced CT images and clinical data of 262 patients from the affiliated hospital of Jining Medical University with pathologically proven primary lung adenocarcinoma who received EGFR gene testing, including EGFR wild type ( n=122) and mutant type ( n=140). The patients were divided into training group ( n=183) and testing group ( n=79) according to a ratio of 7∶3 by stratified sampling method. Standardized pre-processed the images, delineated the ROI and extracted the radiomics features. Least absolute shrinkage and selection operator (LASSO) algorithm was used to reduce the dimension and select key features. The standardized radiomics model, clinical model and the combined model were established by Logistic Regression (LR) machine learning method. Calculated the Rad-score and drew the nomogram. ROC curve and Delong were used to evaluate and compare the predictive performance of different models. Results:23 standardized enhanced CT radiomics features and 4 clinical features were selected. The predictive performance of standardized radiomics model was better than that of non-standardized radiomics model [area under curve (AUC): 0.863 vs. 0.805, t=2.19, P<0.05]. The AUCs of the combined model and standardized radiomics model were higher than that of the clinical model (training group: 0.885, 0.863 vs. 0.774, t=3.57, 2.17, P<0.05; testing group: 0.873, 0.829 vs. 0.763, t=2.19, 2.02, P<0.05). The radiomics nomogram was built based on Rad-score, age, sex, smoking history and BMI. Conclusions:The combined model and standardized radiomics model could effectively predict the mutation status of EGFR gene in lung adenocarcinoma patients before treatment, providing valuable clinical insights.

Objective:To explore the association of frailty and serum C-terminal agrin fragment(CAF)with the prognosis of elderly patients with acute coronary syndrome(ACS).Methods:In this prospective cohort study, clinical data of 207 older patients with ACS between January 2020 and May 2022 were collected.Serum samples were obtained within 24 hours after enrollment to detect CAF levels.Meanwhile, the thrombolysis in myocardial infarction(TIMI)and frailty screening questionnaire(FSQ)scores were assessed on admission.Patients were followed up for major adverse cardiovascular and cerebrovascular events(MACCE)for 90 days.Multivariate logistic regression was used to analyze the influencing factors of MACCE.The receiver operating characteristic(ROC)curve was performed to evaluate the predictive ability of the FSQ score, serum CAF and their combination for MACCE.According to 90-day mortality, patients were divided into a survival group(n=176)and a death group(n=31). The Cox proportional hazards regression model was used for survival analysis.Results:The FSQ score( Z=4.412, P<0.001)and serum CAF( Z=6.702, P<0.001)in the MACCE group were higher than those in the non-MACCE group.Logistic regression analysis showed that after adjusting for age, sex, TIMI score and complete revascularization, frailty defined by FSQ( OR=1.714; 95% CI: 1.059-2.775; P=0.028)and high serum CAF( OR=1.230; 95% CI: 1.122-1.350; P<0.05)were independent risk factors for MACCE.The area under the ROC curve(AUC)of the FSQ score for predicting MACCE was 0.797(95% CI: 0.735-0.850; P<0.001), the predictive cut-off point was an FSQ score >2, and the Youden index(YI)was 0.419, yielding a sensitivity of 0.708 and a specificity of 0.711.In addition, the AUC of serum CAF for predicting MACCE was 0.766(95% CI: 0.701-0.822; P<0.001), the predictive cut-off point was >6.01 μg/L, and YI was 0.460, yielding a sensitivity of 0.750 and a specificity of 0.710.The predictive ability of FSQ combined with CAF for MACCE was higher than FSQ( Z=2.294, P=0.022)or CAF( Z=2.545, P=0.011)alone.Cox regression analysis showed that frailty defined by FSQ( HR=3.487; 95% CI: 1.329-9.153; P=0.011)was independently associated with all-cause mortality within 90 days after ACS. Conclusions:Frailty assessment and serum CAF detection can improve the risk stratification of elderly patients with ACS.