With the continuous deepening of the concept of precision diagnosis and treatment for lung cancer, how to achieve higher efficiency and accuracy in the screening, diagnosis, and treatment pathways in clinical practice has become an important issue that urgently needs to be overcome. The current clinical difficulty lies in the fact that despite continuous advancements in imaging and molecular diagnostic technologies, there are still limitations in manual efficiency and subjective experience when it comes to massive data analysis and multi-scale feature extraction. Artificial intelligence (AI), especially algorithm systems based on deep learning, is an innovative technology capable of deeply empowering medical big data. This method utilizes algorithms such as convolutional neural networks, combined with radiomics, pathomics, and multi-modal data fusion analysis, demonstrating immense potential in early precise detection and benign-malignant differentiation of pulmonary nodules, digital pathological subtype recognition and non-invasive prediction of driver genes, precise 3D surgical planning and automatic delineation of radiotherapy target volumes, as well as dynamic risk warning during follow-up. This innovative technology provides a brand-new solution for realizing intelligent and individualized lung cancer diagnosis and treatment models. This consensus, based on the latest evidence from evidence-based medicine and combined with the development trends in the AI field and real-world clinical needs, was ultimately formed by gathering the consensus opinions of multidisciplinary experts in radiology, pathology, thoracic surgery, and other fields. The main content covers the application specifications of AI in the three core scenarios of lung cancer screening, diagnosis, and treatment, the technical standards for data collection and algorithm validation, as well as the ethical and regulatory challenges faced at the current stage. It aims to clarify the applicable boundaries of AI as a clinical auxiliary decision support tool, providing scientific guidance and standardized exploration directions for peers currently engaged in or planning to carry out AI-assisted clinical diagnosis, treatment, and translation of lung cancer.

Objective:To investigate the mechanism of microRNA-597-3p ( miR-579-3p) in improving hypoxic-ischemic encephalopathy (HIE) by regulating the interleukin-1 receptor-associated kinase 1 (IRAK1)/tumor necrosis factor receptor-associated factor 6 (TRAF6)/transforming growth factor-β-activated kinase 1 (TAK1)/nuclear factor-κB (NF-κB) signaling pathway. Methods:Peripheral blood of healthy neonates ( n=5) and HIE newborns ( n=5) from June 2023 to June 2024 in the No.215 Hospital of Shaanxi Nuclear Industry were collected. The differential expression of miRNA in peripheral blood of neonates with HIE was analyzed by next-generation sequencing. The oxygen-glucose deprivation (OGD) method was used to establish the HIE cell model, which was designated as the OGD group, while cells without OGD treatment served as the control group. Rat adrenal pheochromocytoma PC12 cells were transfected with miR-579-3p mimic (mimic), mimic negative control (mimic-NC), miR-579-3p inhibitor (inhibitor), inhibitor negative control (inhibitor-NC), respectively, and divided into the mimic group, mimic-NC group, inhibitor group, inhibitor-NC group. On the basis of mimic group and inhibitor group, IRAK1 overexpression plasmid ( oeIRAK1) and IRAK1 small interfering RNA plasmid ( siIRAK1) were transfected respectively, which were divided into oeIRAK1 group and siIRAK1 group. The relative expression level of mRNA was detected by real-time reverse transcription-PCR. The absorbance ( A) value was detected by cell counting kit-8 assay. The apoptosis rate was detected by Annexin Ⅴ-fluorescein isothiocyanate/propidium iodide double staining. The relative expression of protein was detected by Western blotting. The levels of pro-inflammatory factors and anti-oxidative stress factors were detected by enzyme-linked immunosorbent assay. Bioinformatics analysis was used to predict the binding site between miR-579-3p and IRAK1, and the dual-luciferase reporter assay was performed to validate their targeting relationship. A total of 12 7-day-old SD rats were selected and randomly divided into the HIE group and the sham group by the random number table method. In the HIE group, the right common carotid artery of rats was permanently occluded, followed by hypoxia exposure, whereas rats in the sham group only underwent common carotid artery exposure without ligation or hypoxia treatment. Immunohistochemical staining was used to detect the expression of protein in brain tissue of HIE rats. The least significant difference t-test was used for comparison between the two groups, and one-way analysis of variance was used for comparison between multiple groups. Results:Compared with the healthy neonates, a total of 278 differentially expressed miRNAs were detected in the peripheral blood of neonates with HIE, among which miR-579-3p showed the most significant downregulation. The relative expression level of miR-579-3p in the mimic group (15.78±1.93) was significantly higher than that in the mimic-NC group (1.00±0.14) ( P<0.01), and in the inhibitor group (0.29±0.14) was significantly lower than that in inhibitor-NC group (1.00±0.14) ( P<0.01). The A value in the mimic group (0.89±0.09) was significantly higher than that in the mimic-NC group (0.52±0.08) ( P<0.01), and in the inhibitor group (0.30±0.05 ) was significantly lower than that in the inhibitor-NC group (0.56±0.07) ( P<0.05). The apoptosis rate of mimic group [(7.47±1.53)%] was significantly lower than that in the mimic-NC group [(30.97±3.47)%] ( P<0.05), and in the inhibitor group [(49.05±4.21)%] was significantly lower than that in the inhibitor-NC group [(35.51±3.64)%] ( P<0.01). The relative expression level of cysteine aspartic acid specific protease-3 ( Caspase-3) and B-cell lymphoma-2 (Bcl-2) associated X protein ( Bax) (1.21±0.10, 1.40±0.13), the relative expression of cleaved Caspase-3 and Bax (1.00±0.13, 1.13±0.09), the levels of interleukin (IL)-6 and tumor necrosis factor (TNF)-α [(45.15±5.14, 38.34±5.69) pg/mg] in the mimic group were significantly lower than those in the mimic-NC group [2.10±0.14, 2.37±0.16, 2.29±0.09, 2.27±0.12, (95.67±9.05, 63.99±5.24) pg/mg] (all P<0.01). The relative expression level of Bcl-2 and Claspin (1.03±0.09, 1.00±0.04), the relative expression of Bcl-2 and Claspin (1.21±0.06, 0.94±0.09), the levels of superoxide dismutase (SOD) and glutathione peroxidase (GSH-Px) [(67.65±6.86, 58.07±5.20) U/mg] in the mimic group were significantly higher than those in the mimic-NC group [(0.51±0.05, 0.52±0.02, 0.58±0.05, 0.46±0.07, 42.08±5.86, 29.80±4.85) U/mg] ( P<0.05, 0.01). The relative expression level of Caspase-3 and Bax (2.72±0.16, 2.97±0.10), the relative expression of cleaved Caspase-3 and Bax (3.25±0.17, 2.76±0.16), the levels of IL-6 and TNF-α [(122.80±11.59, 92.58±7.56) pg/mg] in the inhibitor group were significantly higher than those in the inhibitor-NC group [(1.86±0.14, 2.12±0.10, 2.35±0.15, 1.82±0.15, (88.13±8.59, 68.61±6.17) pg/mg] ( P<0.05, 0.01). The relative expression levels of Bcl-2 and Claspin mRNA (0.17±0.04, 0.20±0.06), the relative expression of Bcl-2 and Claspin proteins (0.11±0.03, 0.13±0.05), the levels of SOD and GSH-Px [(16.62±3.19, 12.01±1.92) U/mg] in the inhibitor group were significantly lower than those in the inhibitor-NC group [0.54±0.05, 0.54±0.05, 0.53±0.10, 0.45±0.07, (38.09±5.47, 30.90±3.87) U/mg] ( P<0.05, 0.01). IRAK1 was identified as a putative target of miR-579-3p. The relative luciferase activity of wild-type IRAK1 in the mimic group (0.45±0.05) was significantly lower than that in the mimic-NC group (1.00±0.08) ( P<0.01). The relative expression levels of IRAK1, TRAF6, TAK1 and NF-κB mRNA in the mimic group (0.96±0.09, 0.96±0.11, 1.34±0.16, 1.74±0.20), the relative expression of IRAK1, TRAF6, TAK1, and phosphorylated NF-κB (p-NF-κB) proteins (0.96±0.20, 1.27±0.19, 1.34±0.18, 1.16±0.19) were all lower than those in the mimic-NC group (1.96±0.17, 1.88±0.24, 2.39±0.23, 2.44±0.20, 2.33±0.22, 2.17±0.24, 2.25±0.28, 2.06±0.28) (all P<0.01). The relative expression levels of IRAK1, TRAF6, TAK1 and NF-κB mRNA in the inhibitor group (2.54±0.15, 2.61±0.13, 2.97±0.15, 2.99±0.20), the relative expression of IRAK1, TRAF6, TAK1 and p-NF-κB proteins (3.73±0.34, 3.17±0.25, 3.68±0.22, 3.29±0.20) were all higher than those in the inhibitor-NC group (1.87±0.18, 1.84±0.19, 2.15±0.24, 2.24±0.26, 2.35±0.23, 1.94±0.25, 2.05±0.27, 2.17±0.29) (all P<0.01). The A value in the oeIRAK1 group (0.66±0.07) was significantly lower than that in the mimic group (0.94±0.10) ( P<0.05), and in the siIRAK1 group (0.51±0.08) was significantly higher than that in the inhibitor group (0.23±0.05) ( P<0.05). The apoptosis rate in the oeIRAK1 group [(23.32±2.40)%] was significantly higher than that in the mimic group [(9.02±1.23)%] ( P<0.05), and in the siIRAK1 group [(28.27±2.57)%] was significantly lower than that in the inhibitor group [(48.96±4.60)%] ( P<0.01). The levels of IL-6 and TNF-α in the oeIRAK1 group [(85.10±6.98, 59.49±5.78) pg/mg] were significantly higher than those in the mimic group [(47.51±8.87, 23.65±3.75) pg/mg], and the levels of SOD and GSH-Px [(49.58±5.63, 41.31±6.21) U/mg] were significantly lower than those in the mimic group [(81.22±6.94, 62.26±5.44) U/mg] ( P<0.05, 0.01). The levels of IL-6 and TNF-α in the siIRAK1 group [(108.25±9.47, 74.87±8.71) pg/mg] were significantly lower than those in the inhibitor group [(142.65±13.88, 104.49±9.18) pg/mg], and the levels of SOD and GSH-Px [(35.87±3.69, 34.89±4.96) U/mg] were significantly higher than those in the inhibitor group [(15.55±3.70, 15.62±3.30) U/mg] ( P<0.05, 0.01). The relative expression level of miR-579-3p in the HIE group (0.47±0.11) was significantly lower than that in the sham group (1.00±0.09) ( P<0.01). The levels of IL-6 and TNF-α in the HIE group [(62.18±6.42, 68.42±4.91) pg/mg] were significantly higher than those in the sham group [(20.77±4.68, 31.16±4.95) pg/mg], and the levels of SOD and GSH-Px [(22.63±3.33, 19.07±2.86) U/mg] were significantly lower than those in the sham group [(47.89±4.58, 56.55±4.45) U/mg] (all P<0.01). The relative levels of TLR4, IRAK1, TRAF6, TAK1, NF-κB, and Caspase-3 mRNA, TLR4, IRAK1, TRAF6, TAK1, p-NF-κB, and cleaved Caspase-3 proteins in the HIE group (1.87±0.24, 2.03±0.21, 2.23±0.20, 1.85±0.18, 1.91±0.20, 2.36±0.20, 2.36±0.28, 2.16±0.28, 1.95±0.27, 2.05±0.26, 2.34±0.24, and 2.72±0.23) were all higher than those in the sham group (1.00±0.15, 1.00±0.11, 1.00±0.09, 1.00±0.05, 1.00±0.12, 1.00±0.15, 1.00±0.15, 1.00±0.21, 1.00±0.10, 1.00±0.14, 1.00±0.19, 1.00±0.16) (all P<0.01), which were consistent with the immunohistochemical staining results. Conclusions:miR-579-3p might alleviate HIE-induced neuronal damage by regulating the IRAK1/TRAF6/TAK1/NF-κB signaling pathway.

Objective To investigate the prevalence of specific IgE antibodies against Alternaria alternata and Aspergillus fumigatus in serum samples from clinical allergy patients across selected provinces in China.Methods Data on specific IgE antibodies for Alternaria A.and Aspergillus F.were collected from 20 hospital laboratories in 17 cities spanning 11 provinces.The study analyzed the levels of specific IgE and their variations across different provinces and seasons.Results A total of 27 471 cases of Alternaria A.and 32 843 cases of Aspergillus F.specific IgE data were included.The national average positive rate of Alternaria A.IgE was 10.40%,with the highest rate of 22.68%in Jiangsu and the lowest rate of 2.06%in Guangxi.For Aspergillus F.specific IgE,the average positive rate was 4.24%,with Hubei province having the highest rate(7.25%)and Hunan province the lowest(1.23%).The difference in IgE levels for both Alternaria A.and Aspergillus F.among provinces were statistically significant(H=9 955,16 993,all P<0.0001).Among patients,5.85%had Alternaria A.specific IgE levels at grade 3 or above,while only 0.57%had Aspergillus F.specific IgE levels at this level.When examining seasonal variations using data from Liaoning,Hunan and Anhui provinces,significant seasonal changes were observed for both Alternaria A.and Aspergillus F.IgE antibodies(HAlternaria A=347.6,338.0,401.3,HAspergillus F=196.6,133.7,231.7,all P<0.0001).Conclusion The sensitization to Alternaria A.and Aspergillus F.exhibits distinct geographical characteristics and vary significantly with seasons.Given the relatively high IgE levels associated with Alternaria A.,it should be given adequate clinical attention.

Objective To explore the risk factors for portal vein thrombosis(PVT)in patients with cirrhosis in the decompensated stage and construct a risk prediction model for PVT so as to improve the early diagnosis rate of decompensated liver cirrhosis PVT.Methods The clinical data of patients with cirrhosis in the decompensated stage admitted to Department of Gastroenterology,The First Affiliated Hospital of Xi'an Jiaotong University between June 2018 and June 2023 were collected and divided into cirrhosis PVT group(n=135)and cirrhosis non-PVT group(n=225)according to whether or not portal vein thrombosis was formed.We made a univariate analysis of the general data,laboratory indexes,liver function scores and imaging findings of the patients in the two groups,and indexes with statistically significant differences were included in binary Logistic regression for multifactorial analysis to screen out independent risk factors.A predictive model of binary Logistic regression was established based on the independent risk factors.The clinical data of the validation set were incorporated into the model,the accuracy of the prediction model was evaluated by receiver operating curve(ROC),and the practicability of the model was evaluated by consistency curve to complete the validation and evaluation of the constructed prediction model.Internal stability of the model was verified with Bootstrap method.Finally,R software(4.3.1)was used to draw a nomogram of the prediction model to visualize the model.Results Univariate analysis revealed statistically significant differences between patients in the PVT and non-PVT groups in the following five aspects:history of splenectomy,history of endoscopic varicose vein treatment,portal vein diameter,and neutrophil-to-lymphocyte ratio(P＜0.05).Binary Logistics regression analysis showed that a history of splenectomy(P=0.002,OR=3.012,95％CI:1.500-6.047),a history of endoscopic varicose vein treatment(P=0.001,OR=2.276,95％CI:1.400-3.698),widening of portal vein diameter(P=0.007,OR=1.942,95％CI:1.202-3.136),increased neutrophil-to-lymphocyte ratio(P=0.009,OR=1.886,95％CI:1.170-3.041),and elevated D-dimer(P＜0.001,OR=3.725,95％CI:2.149-6.485)were independent risk factors for the formation of portal vein thrombosis in patients with cirrhosis in the decompensated stage of cirrhosis chemically presented(P＜0.05).The area under the ROC curve of the predictive model and the model after internal validation was 0.760 and 0.7494,respectively.The model still had good prediction ability and accuracy in the verification set.Conclusion A history of splenectomy,history of endoscopic varicose vein treatment,widening of portal vein diameter,increased neutrophil-to-lymphocyte ratio,and elevated D-dimer concentration are independent risk factors for the formation of portal vein thrombosis in patients with decompensated cirrhosis.The Logistic prediction model and visual nomogram constructed based on the above independent risk factors have a good ability to predict the occurrence of PVT in patients with decompensated cirrhosis and have important clinical guiding significance for early screening patients with PVT in decompensated cirrhosis.

Objective To establish an efficient and stable model of pelvic inflammatory disease in rats via a non-surgical method,and to evaluate its application in pharmacodynamic testing.Methods Female Sprague-Dawley rats were divided randomly into the following groups:control group;model group with phenol for 7 d;model group with phenol for 10 d;treatment group modeled with phenol;model group with low concentration of bacteria;model group with high concentration of bacteria;and treatment group modeled with bacteria.Rats in the model and treatment groups were injected with 25%phenol gel and 2×107 or 2×108 Escherichia coli and Staphylococcus aureus mixture via a non-surgical method,to construct a rat model of pelvic inflammatory disease.Rats in the treatment groups received the Chinese patent medicine Jingangteng capsules by gavage,and rats in the control group received the same volume of solvent solution.The health status,weight changes,and uterine appearance were monitored and the uterine coefficient was calculated.Pathological changes in the uterus and fallopian tubes,endometrial thickness,and number of glands were detected by hematoxylin and eosin staining.Serum levels of interleukin(IL)-1β,IL-6,and tumor necrosis factor(TNF)-α were detected by enzyme-linked immunosorbent assay.Protein expression of the macrophage marker CD68 was detected by immunofluorescence.Expression of Toll-like receptor 4(TLR4)/nuclear factor(NF)-κB pathway-related proteins in the uterus was detected by Western blot.Results The mortality rate in the model group was only 5%.Compared with the control group,model rats showed decreased body weight,increased uterine coefficient,pathological changes in the uterus and Fallopian tubes,thinner endometrium,fewer glands,significantly higher serum levels of IL-1β,IL-6,and TNF-α and more macrophages in the uterine tissue,and activation of the TLR4/NF-κB signaling pathway.The 7 d phenol and low-concentration bacterial solution models were judged to be mild pelvic inflammatory disease models,and the 10 d phenol and high-concentration bacterial solution models were considered severe pelvic inflammatory disease models.Treatment with Jingangteng capsules relieved the pathological symptoms in the uterus and fallopian tubes,in line with the efficacy evaluation of clinical pelvic inflammatory disease.Conclusions We established rat models of pelvic inflammatory disease using phenol and a mixed bacterial solution via a non-surgical method,to simulate the different pathological states of pelvic inflammatory disease caused by different factors.These models will be suitable for evaluating drug efficacy and elucidating the pathological mechanism of pelvic inflammatory disease.

Obesity is a significant risk factor for cancer and is associated with breast cancer metastasis. Nevertheless, the mechanism by which alterations in systemic metabolism affect tumor microenvironment (TME) and consequently influence tumor metastasis remains inadequately understood. Herein, we found that perturbations in circulating metabolites induced by obesity promote metastasis-like phenotypes in breast cancer. Oleoylcarnitine (OLCarn) concentrations were elevated in the serum of obese mice and humans. Administration of exogenous OLCarn induces metastasis-like characteristics in breast cancer cells. Mechanistically, OLCarn directly interacts with the Arg176 site of adenylate cyclase 10 (ADCY10), leading to the activation of ADCY10 and enhancement of cAMP production. Mutations at Arg176 prevent OLCarn from binding to ADCY10, disrupting the ADCY10-mediated activation of cyclic adenosine monophosphate (cAMP) signaling pathway. This activation promotes transcription factor 4 (TCF4)-dependent kinesin family member C1 (KIFC1) transcription, thereby driving breast cancer metastasis. Conversely, the neutralization of both ADCY10 and KIFC1 through knockdown or pharmacological inhibition abrogates the oncogenic effects mediated by OLCarn. Hence, obesity-induced systemic environmental changes lead to the aberrant accumulation of OLCarn within the TME, making it a potential therapeutic target and biomarker for breast cancer.

Objective To evaluate the application value of nucleic acid mass spectrometry in detecting ERG11 gene mutations associated with azole resistance in Candida tropicalis(C.tropicalis),thereby providing a scientific basis for the rational clinical use of azole antifungal agents.Methods Clinical isolates of C.tropicalis were obtained from the Affiliated Hospital of Xuzhou Medical University between July 2023 and November 2024.For each isolate,specific ERG11 mutations(A395T and C461T)were analyzed using Sanger sequencing and nucleic acid mass spectrometry.Sanger sequencing was used as the reference standard to evaluate the performance of mass spectrometry in mutation detection.Results The established nucleic acid mass spectrometry method effectively identified four genotypes of ERG11(A395T and C461T),with distinct spectral peaks for each mutation and no cross-interference observed.Among 88 clinical isolates,the sensitivity,specificity,positive predictive value(PPV),and negative predictive value(NPV)for the A395T mutation were 86.7%,100%,100%,and 97.3%,respectively.For the C461T mutation,the corresponding values were 81.3%,100%,100%,and 96.0%,respectively.Kappa statistics demonstrated a high level of agreement between mass spectrometry and sequencing results.Conclusions Nucleic acid mass spectrometry exhibits high sensitivity and specificity in the detection of ERG11 mutations,enabling rapid,accurate,and adaptable high-throughput analysis.This makes it a highly effective method for identifying mutations associated with fungal resistance.

Objective To assess functional abnormalities in frontoparietal attention network of patients with traumatic brain injury(TBI),and to analyze neural mechanism of its attention disorders based on independent component analysis(ICA).Methods Totally 84 TBI patients were prospectively enrolled and divided into mild(n=33),moderate(n=27)and severe TBI groups(n=24)according to symptoms,non-contrast enhanced MRI findings,Glasgow coma scale(GCS)and Mayo craniocerebral trauma grading criteria.ICA was used to extract and analyze the optimal independent components(IC)of the dorsal attention network(DAN)and ventral attention network(VAN).Multiple linear regression analyses were performed to explore the correlations of functional connectivity(FC)of brain regions being significantly different among groups within attention network and the scores of trail making test A/B(TMT-A/B).Results Compared with control group and mild TBI group,spatial distribution of DAN(especially bilateral frontal eye field[FEF])in moderate and severe TBI groups were significantly smaller,the clusters tended to be discrete,and the intensity were significantly weakened.The right temporal parietal junction(TPJ)and right ventral frontal cortex(VFC)of VAN in control group,mild and moderate TBI groups were significantly affected,and VFC connection strength was the highest in severe TBI group.FC intensity in right VFC of TBI patients was positively correlated with TMT-A score(r=0.654,P＜0.001),which in left FEF was negatively correlated with TMT-B score(r=—0.383,P＜0.001).Conclusion FC of frontoparietal attention network was abnormal in TBI patients,which might be neural mechanism for its attention behavior disorders.

Objective:Based on the consultation records of the online medical consultation platform, the chronic kidney disease (CKD) health information needs of users are analyzed, so as to provide a decision-making basis for the development of personalized CKD health information services, management and intervention.Methods:From September 15th to December 30th, 2022, the combination of manual retrieval and web crawler was used to obtain the question records of "kidney disease" in the three online medical platforms including Good Doctor Online, Dingxiang Doctor and Chunyu Doctor. Also, the hot question records were classified according to the content of users' questions. The top 50 high-frequency keywords were calculated by word segmentation and term frequency-inverse document frequency (IF-IDF) algorithm. The Gephi software was used to visualize the co-occurrence matrix of keywords.Results:A total of 78 405 valid consultation records were collected in the obtained question content. The classification results showed that the users' CKD health information needs can be classified into five categories, including basic knowledge (etiology and risk factors, symptoms and signs, classification and grading of diagnosis, complications or comorbidities, drug effects and adverse reactions), treatment plan, social life impact (physical impact, family impact, social interaction impact, economic burden impact, work impact), self-management (self-care or home care, emergency response) and social support (family support, professional medical team support).Conclusions:The information demand for CKD based on the online medical consultation platform presents a wide range of content and diversified subjects. This suggests that targeted and personalized health information guidance should be formulated when providing CKD health information services for different subjects.

Objective:To investigate the effect of a blended teaching mode based on medical knowledge graph and artificial intelligence (AI) teaching assistant on students' learning effectiveness and systematic thinking ability in teaching of pathogens and immunology.Methods:A controlled experimental design was employed, involving 114 clinical medical students ("5+3" integrated) enrolled in 2023 at Tianjin Medical University. They were divided into experimental group ( n=57) and control group ( n=57) based on their classes. A blended online-offline teaching mode based on medical knowledge graph and AI teaching assistant was used in the experimental group, and a blended online-offline teaching mode based on the online resources of this course was used in control group. Teaching effectiveness was assessed by comparing the scores of four chapter quizzes and the final exam between the two groups, as well as by analyzing student questionnaire responses. Data were analyzed using t test and χ2 test. Results:The experimental group achieved significantly higher scores in all four chapter quizzes and the final exam compared to the control group [(78.77±19.65) vs. (69.47±22.95), (84.56±14.02) vs. (76.49±16.20), (81.89±13.60) vs. (73.13±16.52), (81.56±21.28) vs. (73.16±16.27), (69.75±13.30) vs. (64.10±14.93), all P<0.05]. The questionnaire survey showed that 74.07% students ( n=40) believed that this blended teaching model stimulated their learning interest, 74.07% ( n=40) students believed that this blended teaching model enhanced their learning initiative, 81.48% ( n=44) students believed that this blended teaching model could help them construct the relationship between the old and new knowledge and thus improved memory retention, and 81.48% students b( n=44) elieved that this blended teaching model helped them integrate and summarize knowledge and construct systematic knowledge framework to improve their learning effectiveness. Conclusions:The blended teaching mode based on medical knowledge graph and AI teaching assistant is beneficial for cultivating students' systematic thinking ability and is helpful to improve their learning efficiency and effectiveness.