This paper outlines the development of artificial intelligence （AI） and its applications in traditional Chinese medicine （TCM） research， and elucidates the roles and advantages of large language models， knowledge graphs， and natural language processing in advancing syndrome identification， prescription generation， and mechanism exploration. Using spleen-stomach diseases as an example， it demonstrates the empowering effects of AI in classical literature mining， precise clinical syndrome differentiation， efficacy and safety prediction， and intelligent education， highlighting an upgraded research paradigm that evolves from data-driven and knowledge-driven approaches to intelligence-driven models. To address challenges related to privacy protection and regulatory compliance in cross-institutional data collaboration， a "trusted federated evidence-based analysis platform for TCM spleen-stomach diseases" is proposed， integrating blockchain-based smart contracts， federated learning， and secure multi-party computation. The deep integration of AI with privacy-preserving computing is reshaping research and clinical practice in TCM spleen-stomach diseases， providing feasible pathways and a technical framework for building a high-quality， trustworthy TCM big-data ecosystem and achieving precision syndrome differentiation.

ObjectiveTo investigate the triglyceride-glucose (TyG) index and the level of serum homocysteine (Hcy) in association with the incidence of stroke in type 2 diabetes mellitus (T2DM) patients. MethodsBased on the chronic disease risk factor surveillance cohort in Pudong New Area, Shanghai, excluding those with stroke in baseline survey, T2DM patients who joined the cohort from January 2016 to October 2020 were selected as the research subjects. During the follow-up period, a total of 318 new-onset ischemic stroke patients were selected as the case group, and a total of 318 individuals matched by gender without stroke were selected as the control group. The Cox proportional hazards regression model was used to adjust for confounding factors and explore the serum TyG index and the Hcy biochemical indicator in association with the risk of stroke. ResultsThe Cox proportional hazards regression results showed that after adjusting for confounding factors, the risk of stroke in T2DM patients with 10 μmol·L⁻¹-¹ and Hcy>15 μmol·L⁻¹ was 1.532 times (95%CI: 1.017‒2.309 times, P=0.041) and 1.738 times (95%CI: 1.119‒2.699 times, P=0.014) higher respectively, compared to T2DM patients with Hcy≤10 μmol·L⁻¹. T2DM patients with TyG>9.074 had 1.396 times higher risk of stroke (95%CI: 1.091‒1.787, P=0.008) compared to those with TyG≤9.074. The study found that urea and α1-antitrypsin (α1-AT) were independent risk factors for the incidence of stroke in T2DM patients. For every unit increase in urea andα1-AT, the risk of stroke in T2DM patients increased by 233.6% (HR=3.336, 95%CI: 1.588‒7.009, P=0.001) and 11.3% (HR=1.113, 95%CI: 1.028‒1.205, P=0.008), respectively. Cumulative risk curves showed that Hcy>10 μmol·L⁻¹ and TyG>9.074 accelerated the time to stroke occurrence in T2DM patients. ConclusionElevated levels of Hcy>10 μmol·L⁻¹ and a higher TyG index are risk factors for stroke in T2DM patients. Effective interventions for Hcy and TyG should be conducted for diabetic patients to reduce the incidence of stroke.

Objective To compare the prognostic differences between simple and complex segmentectomies. Methods We conducted a retrospective cohort analysis of patients with solid pulmonary nodules (≤2 cm) who underwent segmentectomy. Recurrence-free survival (RFS) and local recurrence rates were evaluated. Results We included57 patients undergoing complex segmentectomy and 53 patients undergoing simple segmentectomy. Among patients who did not receive adjuvant therapy, those in the complex group had a significantly lower five-year RFS than those in the simple group (69.86% vs. 85.97%, P=0.04). Furthermore, the local recurrence rate was significantly higher in the complex group (18.75% vs. 4.65%, P=0.003) than in the simple group. Conclusion For solid pulmonary nodules (≤2 cm), complex segmentectomy is associated with inferior local control and worse RFS than simple segmentectomy.

The diagnosis of Meige's syndrome predominantly relies on the clinical assessment by physicians.Given the complexity and similarity of its symptoms to other neurological disorders,the diagnosis is crucial for both doctors and patients.Herein a detection dataset for Meige's syndrome is compiled from video recordings of 31 patients,and an automated diagnostic system for Meige's syndrome(MS-Net)applicable to untrimmed videos is developed.The system utilizes RetinaNet and UNet3+to construct temporal detection and segmentation branches for multi-scale feature extraction and temporal segmentation,obtains probability vectors for detection windows and the probability of disease onset per frame via the decoding of temporal detection and segmentation branches,and finally generates a refined probability for each window by processing the probability predictions from both branches using a multi-layer perceptron.The model performance is optimized using additional loss functions and data augmentation techniques,operating on features interpretable by clinical physicians.MS-Net can assist in the diagnosis of Meige's syndrome,improving the accuracy,convenience,and efficiency of the early diagnosis.The comparison of MS-Net with other state-of-the-art networks indicates that MS-Net achieves comparable performance in terms of average precision while utilizing interpretable features required in clinical practice.

The diagnosis of Meige's syndrome predominantly relies on the clinical assessment by physicians.Given the complexity and similarity of its symptoms to other neurological disorders,the diagnosis is crucial for both doctors and patients.Herein a detection dataset for Meige's syndrome is compiled from video recordings of 31 patients,and an automated diagnostic system for Meige's syndrome(MS-Net)applicable to untrimmed videos is developed.The system utilizes RetinaNet and UNet3+to construct temporal detection and segmentation branches for multi-scale feature extraction and temporal segmentation,obtains probability vectors for detection windows and the probability of disease onset per frame via the decoding of temporal detection and segmentation branches,and finally generates a refined probability for each window by processing the probability predictions from both branches using a multi-layer perceptron.The model performance is optimized using additional loss functions and data augmentation techniques,operating on features interpretable by clinical physicians.MS-Net can assist in the diagnosis of Meige's syndrome,improving the accuracy,convenience,and efficiency of the early diagnosis.The comparison of MS-Net with other state-of-the-art networks indicates that MS-Net achieves comparable performance in terms of average precision while utilizing interpretable features required in clinical practice.

Total hip arthroplasty(THA)is an effective treatment for elderly femoral neck fractures,mid-to late-stage femoral head necrosis,and end-stage hip osteoarthritis.However,serious complications such as aseptic loosening of the prosthesis,peripheral fractures,and dislocation of the prosthesis still exist following THA,which makes the selection of the appropriate hip prosthesis type and placement position before THA an important challenge for surgeons.Currently,the commonly used preoperative planning methods for THA mainly rely on static images from two-dimensional(2D)X-ray or three-dimensional(3D)computed tomography(CT),which fail to adequately consider the hip joint in weight-bearing as well as motion,lumbar-hip joint changes,and prosthetic impingement during motion.Recently,the dual fluoroscopic imaging system,as a new in-vivo,dynamic radiological imaging technology,provides comprehensive and accurate dynamic 3D data for THA preoperative planning.However,the technical process and expert consensus on preoperative 3D planning of THA using a dual fluoroscopic imaging system have not yet been established,which affects the promotion and application of this technology.In light of the above,national orthopaedic experts and related professional representatives discussed and proposed seven consensus issues,and the'expert recommendation rate'and'strong recommendation rate'were obtained through a questionnaire survey on the recommendations of the participating experts.This consensus aims to provide guidance and reference for the standardised application of preoperative 3D planning of THA using the dual fluoroscopic imaging system.

Objective:To explore the value of conventional magnetic resonance imaging (MRI), diffusion weighted imaging (DWI), diffusion kurtosis imaging (DKI) sequence and pathological examination in predicting the efficacy of neoadjuvant chemotherapy (NAC) in advanced breast cancer.Methods:The clinical data of 65 cases of advanced breast cancer with NAC confirmed by pathology in the Affiliated Hospital of Jining Medical University from March 2022 to May 2023 were retrospectively analyzed, including 20 cases in the pathological complete remission (pCR) group and 45 cases in the non pCR group; All patients underwent routine MRI, DWI, DKI examinations and pathological analysis. The clinical pathological data, routine MRI features, apparent diffusion coefficient (ADC) values, mean kurtosis coefficient (MK), and mean diffusion coefficient (MD) between the two groups were analyzed; We compared the differences in various parameters between two groups and plotted receiver operating characteristic (ROC) curves to compare their diagnostic efficacy of NAC in breast cancer.Results:There were significant differences in molecular typing, estrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor 2 (Her-2) and Ki-67 between pCR group and non pCR group (all P<0.05). In pCR group, Her-2 overexpression type and triple negative breast cancer (TNBC) type breast cancer were more common. ER and PR were mostly negative, Her-2 was mostly positive, and Ki 67 was mainly positive. The difference in tumor T2WI signal between the pCR group and the non pCR group was statistically significant ( P<0.05), with the pCR group showing mostly moderate/low T2WI signal. The ADC and MD values of the pCR group were lower than those of the non pCR group, while the MK value of the pCR group was higher than that of the non pCR group, and the differences were statistically significant (all P<0.001). The area under the ROC curve (AUC) for predicting the efficacy of NAC using a clinical pathological model was 0.829, which was higher than the AUC of molecular subtypes, ER, PR, Her-2, and Ki-67 ( Z=3.008, 2.697, 2.815, 2.131, 2.376, all P<0.05); The AUC of the DKI+ DWI predicting the efficacy of NAC was 0.934, which was higher than that of the DWI single sequence model, and the difference in type was statistically significant ( Z=2.396, P=0.017). The diagnostic efficacy of the DKI+ DWI model was higher than that of the single parameter ADC, MD, and MK, and the differences were statistically significant ( Z=2.396, 2.219, 2.161, all P<0.05); The AUC of the combined imaging and pathology model was 0.983, and its diagnostic efficacy was higher than that of the conventional MRI feature model, pathology model, DWI model, and DKI model, with statistically significant differences ( Z=5.877, 2.961, 3.240, 2.264, all P<0.05). Conclusions:The results of pathology, conventional MRI, DWI and DKI parameters of pCR and non pCR breast cancer patients are significantly different, and the combined model is better than the single model in predicting the efficacy of NAC.

Given that low-dose computed tomography significantly amplifies image noise due to the mitigation of radiation exposure,which degrades image quality and lowers the precision of clinical diagnoses,a novel model incorporating convolutional neural network and Transformer is established,in which an intra-patch feature extraction module is used to effectively preserve tiny details in the image.A double attention Transformer is constructed by incorporating a multiple-input channel attention module into the self-attention for tackling the problem of incorrect restoration of texture details during denoising using Swin Transformer.AAPM dataset is used for testing,and the results demonstrate that the proposed algorithm not only surpasses the existing algorithms in denoising performance,but also excels in preserving tiny details in the image.

An improved U-shaped multi-lesion segmentation model,namely dense dilated attention pyramid UNet(DDAPNet),is proposed to overcome the difficulty in learning multi-scale features and address the issue of blurry boundaries in diabetic retinopathy(DR)segmentation task.DR images are treated with Patch processing to enhance the model's ability to capture local lesion features.After backbone feature extraction,a redesigned dense dilated attention pyramid module is introduced to expand the receptive field and address the issue of blurry lesion boundaries;and simultaneously,pyramid split attention module is used for feature enhancement;and then,the features output by the two modules are fused.Additionally,an improved residual attention module is embedded within skip connections to reduce interference from shallow redundant information.The joint validation on DDR dataset and real dataset from a specific hospital shows that compared with the original model,DDAPNet model improves the Dice similarity coefficient for segmentations of microaneurysms,hemorrhages,soft exudates and hard exudates by 4.31%,2.52%,3.39%and 4.29%,respectively,and increases mean intersection over union by 1.80%,2.24%,4.28%and 1.98%,respectively.The proposed model makes the segmentation of lesion edges smoother and more continuous,notably enhancing the segmentation performance for conditions like soft exudates in retinal lesions.

Objective To observe the value of CT radiomics nomogram for predicting Ki-67 expression of thymus epithelial tumors.Methods Totally 163 patients with thymus epithelial tumor,including 114 patients in training set and 49 patients in validation set were retrospectively enrolled.The patients were further divided into low expression(＜50％)and high expression(≥50％)subgroups according to Ki-67 index.Multivariate logistic regression analysis was performed to screen independent predicting factors of Ki-67 expression in thymus epithelial tumors,and clinical-CT model was constructed.The optimal radiomics features were extracted and screened based on chest plain and venous phase enhanced CT images,respectively.Then radiomics modelplain and radiomics modelenhanced were constructed,and Radscoreplain and Radscoreenhanced were calculated,respectively.The nomogram model was constructed based on clinical-CT model,Radscoreplain and Radscoreenhanced.Receiver operating characteristic curves were drawn,and the area under the curves(AUC)were calculated to evaluate the efficacy of each model for predicting Ki-67 expression of thymus epithelial tumors.Results Patient's gender and enhanced CT value of lesion were both independent predicting factors of Ki-67 expression in thymus epithelial tumors(both P＜0.05).The AUC of clinical-CT model,radiomics modelplain,radiomics modelenhanced and nomogram model for predicting Ki-67 expression was 0.736,0.814,0.836 and 0.857 in training set,which was 0.746,0.746,0.750 and 0.799 in validation set,respectively.Conclusion CT radiomics nomogram could be used to predict Ki-6 7 expression of thymus epithelial tumors.