1.Clinical Value of A Arterial and Portal CT-Based Deep Learning Model for the Differential Diagnosis of Benign and Malignant Pancreatic Cystic Lesions
Boyu CHEN ; Wenyi DENG ; Fuze CONG ; Huadan XUE
Medical Journal of Peking Union Medical College Hospital 2026;17(4):963-975
To develop a deep learning model based on arteriovenous dual-phase CT imaging features and evaluate its diagnostic value in differentiating benign from malignant pancreatic cystic lesions(PCLs). Preoperative contrast-enhanced CT images of patients with histopathologically confirmed PCLs at Peking Union Medical College Hospital from June 1, 2014 to May 31, 2023 were retrospectively collected. The data were randomly partitioned at the lesion level into training, validation, and test sets in a 3∶1∶1 ratio. CT images were preprocessed and regions of interest were delineated. Using postoperative pathological results as the reference standard, five deep learning models(ResNet50, DenseNet121, ResNeXt50, EfficientNet-b5, and MobileNetV2) were constructed to extract arteriovenous dual-phase CT imaging features for binary classification of PCLs. The optimal model was selected based on the area under the curve(AUC), accuracy, sensitivity, and specificity. Further comparative analyses were conducted against single-phase CT-based deep learning models, conventional radiomics models, and radiologist interpretations to comprehensively evaluate the performance of the arteriovenous dual-phase CT-based deep learning model in the differential diagnosis of PCLs. A total of 480 patients with 485 lesions(206 malignant and 279 benign) were ultimately enrolled. The training, validation, and test sets comprised 291, 97, and 97 lesions, corresponding to 288, 96, and 96 patients, respectively. In the validation set, the ResNeXt50 model based on arteriovenous dual-phase CT features achieved an AUC of 0.837(95% CI: 0.748-0.915), an accuracy of 77.32%(95% CI: 67.70%-85.21%), a sensitivity of 80.49%(95% CI: 65.13%-91.18%), and a specificity of 75.00%(95% CI: 61.63%-85.61%). In the test set, the corresponding values were 0.822(95% CI: 0.737-0.904), 73.20%(95% CI: 63.24%-81.68%), 82.93%(95% CI: 67.94%-92.85%), and 66.07%(95% CI: 52.19%-78.19%), demonstrating overall superior performance. Calibration curves indicated that the predicted probabilities of the model were generally consistent with observed outcomes, albeit with certain shortcomings in probability calibration. Decision curve analysis demonstrated that, overall, the use of this model for clinical decision-making conferred a net benefit to patients. Compared with single-phase CT-based models and conventional radiomics models, the ResNeXt50 model incorporating arteriovenous dual-phase CT features exhibited superior overall performance, and its diagnostic performance was comparable to that of radiologists, with good consistency. The deep learning model based on arteriovenous dual-phase CT imaging features demonstrates diagnostic value in differentiating benign from malignant PCLs and may serve as an adjunctive reference for preoperative clinical assessment. However, its stability and generalizability warrant further validation in larger cohorts and with external datasets.
2.Prediction of Mismatch Repair Deficiency Status in Endometrial Cancer Using Multiparametric MRI Radiomics and Deep Learning: A Multimodal Model with Preliminary Validation
Liru WANG ; Shangying YANG ; Boyu CHEN ; Fuze CONG ; Xinran LI ; Xinyu LIU ; Huadan XUE ; Zhengyu JIN ; Yang XIANG ; Yonglan HE ; Yuan LI
Medical Journal of Peking Union Medical College Hospital 2026;17(4):976-984
To explore the clinical value of a multimodal predictive model based on multiparametric magnetic resonance imaging(MRI) radiomics combined with deep learning(DL) features for the preoperative noninvasive assessment of mismatch repair-deficient(MMRd) status in endometrial cancer(EC). Patients diagnosed with EC at Peking Union Medical College Hospital from January 2015 to December 2021 were retrospectively enrolled and randomly divided into a training set and a validation set at a ratio of 8∶2. Relevant clinical data were collected, and radiomics features and DL features were extracted from preoperative contrast-enhanced T1-weighted imaging(CE-T1WI), fat-suppressed T2-weighted imaging(fs-T2WI), and diffusion-weighted imaging(DWI) sequences. High-dimensional feature selection and dimensionality reduction were performed sequentially using the recursive feature elimination(RFE) algorithm to generate a radiomics score(Rad-score) and a deep learning score(DL-score), respectively. Multivariate logistic regression was utilized to construct a clinical model, a pure radiomics model, a clinical-radiomics model, and an integrated multimodal model incorporating clinical indicators, Rad-score, and DL-score. Model performance was assessed and compared using area under receiver operating characteristic curve(AUC) and DeLong test. A total of 509 patients were enrolled in this study, comprising 413 in the training cohort and 96 in the validation cohort. Independent predictors: Multivariate analysis indicated that preoperative fasting blood glucose level, histological grade, lymph node metastasis status, Rad-score, and DL-score were all independent significant predictors of MMRd status in EC patients. The integrated multimodal model demonstrated optimal predictive performance with an AUC of 0.699(95% CI: 0.635-0.763) in the training set, which was superior to the clinical model(AUC=0.629, 95% CI: 0.561-0.697) and the pure radiomics model(AUC=0.641, 95% CI: 0.575-0.706). In the validation set, the integrated model maintained good generalizability, achieving an AUC of 0.655(95% CI: 0.535-0.775), and its diagnostic efficacy was higher than that of the clinical model(AUC=0.578, 95% CI: 0.450-0.705) and the pure radiomics model(AUC=0.611, 95% CI: 0.488-0.734). According to the DeLong test, the incorporation of DL features resulted in the clinicalradiomicsdeep learning model performing better than both the clinicalonly model( The initially developed clinical-radiomics-deep learning model exhibits a certain predictive potential for the MMRd status in patients with EC. The inclusion of DL features may help complement the limitations of traditional evaluations, offering a preliminary radiological reference for preoperative non-invasive screening. However, given the current diagnostic performance, its overall accuracy and clinical generalizability warrant further validation in multi-center, large-sample external cohort studies.
3.An equation for calculating the osmolarity of adult total nutrient admixture
Xiaoqin LANG ; Boyu LU ; Xue BAI ; Jiaqiang XU ; Bin ZHAO ; Dan MEI
Chinese Journal of Clinical Nutrition 2020;28(2):87-92
Objective:To establish a predictive equation for the osmolarity of parenteral nutritional prescription in China.Methods:From July 2019 to September 2019, 2 480 individualized samples of 328 different parenteral nutritional prescriptions (adult) of Peking Union Medical College Hospital were collected, and the osmolarity of parenteral nutritional solution samples was determined by a freezing point reduction method. Pearson χ2 test and a multiple linear regression analysis were utilized to establish a prediction equation for the osmolality of parenteral nutritional solution. Results:The average osmolarity of parenteral nutritional prescription was (1 164.20 ± 252.59) mOsm/kg, and the best fitting equation was (9.66A+ 7.88G+ 3.52F+ 36.4Na+ 27.55K+ 3.38P+ 7.46W-250)/V.Conclusion:The osmolarity is determined accurately and effectively by the fitting equation, which provide a benefit reference for the formulation, review and selection of clinical parenteral nutrition prescription especially in China.
4.Pharmacokinetics of gene recombined angiogenesis inhibitor Kringle 5 in vivo using 131I specific markers and SPECT/CT$
Ge YAN ; Danrong YANG ; Yan YU ; Jianjun XUE ; Yifan JIA ; Xuanzi SUN ; Boyu WANG ; Zewei ZHAO ; Maode WANG
Journal of Pharmaceutical Analysis 2015;5(5):313-317
The previous pharmacokinetic methods can be only limited to drug analysis in vitro, which provide less information on the distribution and metabolismof drugs, and limit the interpretation and assessment of pharmacokinetics, the determination of metabolic principles, and evaluation of treatment effect. The objective of the study was to investigate the pharmacokinetic characteristics of gene recombination angiogenesis inhibitor Kringle 5 in vivo. The SPECT/CT and specific 131I-Kringle 5 marked by Iodogen method were both applied to explore the pharmacokinetic characteristics of 131I-Kringle 5 in vivo, and to investigate the dynamic distributions of 131I-Kringle 5 in target organs. Labeling recombinant angio-genesis inhibitor Kringle 5 using 131I with longer half-life and imaging in vivo using SPECT instead of PET, could overcome the limitations of previous methods. When the doses of 131I-Kringle 5 were 10.0, 7.5 and 5.0 g/kg, respectively, the two-compartment open models can be determined within all the metabolic process in vivo. There were no significant differences in t1/2α, t1/2β, apparent volume of distribution and CL between those three levels. The ratio of AUC(0 ? 1) among three different groups of 10.0, 7.5 and 5.0 g/kg was 2.56:1.44:1.0, which was close to the ratio (2:1.5:1.0). It could be clear that in the range of 5.0–10.0 g/kg, Kringle 5 was characterized by the first-order pharmacokinetics. Approximately 30 min after 131I-Kringle 5 was injected, 131I-Kringle 5 could be observed to concentrate in the heart, kidneys, liver and other organs by means of planar imaging and tomography. After 1 h of being injected, more radionuclide retained in the bladder, but not in intestinal. It could be concluded that 131I-Kringle 5 is mainly excreted through the kidneys. About 2 h after the injection of 131I-Kringle 5, the radionuclide in the heart, kidneys, liver and other organs was gradually reduced, while more radionuclide was concentrated in the bladder. The radionuclide was completely metabolized within 24 h, and the distribution of radioactivity in rats was similar to normal levels. In our study, the specific marker 131I-Kringle 5 and SPECT/CT were suc-cessfully used to explore pharmacokinetic characteristics of Kringle 5 in rats. The study could provide a new evaluation platform of the specific, in vivo and real-time functional imaging and pharmacokinetics for the clinical application of 131I-Kringle 5.
5.Research progress in Treg/Th17 imbalance in persistent HBV infection
Journal of Clinical Hepatology 2015;31(4):603-606
Persistent infection with hepatitis B virus (HBV)is associated with host immune response.CD4 +T cells play an important role in HBV-specific immune response.The restoration of HBV-specific T-cell response after antiviral therapy using nucleoside and nucleo-tide analogues is also associated with CD4 +T cells.In recent years,two new subsets of CD4 +T cells,namely regulatory T cells (Tregs) and T helper 17 cells (Th17 cells),have been identified and shown to be related to disease progression and liver damage in patients with persistent HBV infection.Here we primarily summarized the differentiation and function of Tregs and Th17 cells and reviewed the interaction between the two types of cells in persistent infection and their changes after clinical antiviral therapy.We hope it will be helpful to clinical immunotherapy and prognostic assessment.
6.Research progress in chronic kidney disease with hepatitis C virus infection
Hua JIANG ; Min ZHOU ; Boyu XUE
Journal of Clinical Hepatology 2015;31(11):1912-1915
Patients with chronic kidney disease (CKD) are susceptible to hepatitis C virus (HCV) infection due to hemodialysis and kidney transplantation. Despite the progress in treatment and new medicines for hepatitis C, the efficacy, tolerability, and safety of medicines in CKD patients remain unclear. This article reviews the current situation of HCV infection in CKD patients, methods for evaluating liver condition in HCV-CKD patients, and the latest progress in HCV-CKD treatment. More clinical practice is needed to verify the selection and dosage of medicines for HCV-CKD.
7.Effect of Qishen-Qinggan decoction on proliferation and apoptosis of human hepatocellular carcinoma cellline SMMC-7721
Lingfeng WAN ; Boyu XUE ; Zhangpu LIU ; Ming SHAO
International Journal of Traditional Chinese Medicine 2013;35(12):1079-1081
Objective To investigate effect of Qishen-Qinggan decoction on proliferation and apoptosis of human hepatocellular carcinoma cellline SMMC-7721.Methods SMMC-7721 cells were cultivated in vitro.Logarithmic growth phase cells were divided into a drug intervention group and a control group.SMMC-7721 cells were treated with Qishen-Qinggan decoction of 0.135、0.27、0.54、1.08、2.16 g/ml respectively for 12、24、48 hours in the drug intervention group while the control group remained untreated.The inhibition rate of SMMC-7721 cells were detected by MTT assay,cell apoptotic rate were measured by flow cytometry analysis.Results Qishen-Qinggan decoction of 0.135,0.27,0.54,1.08,2.16 g/ml had a significantly inhibitory effect on SMMC-7721 cells in a dose and time dependent manner.OD values of 12 hours were 0.89±0.05,0.85±0.05,0.80±0.06,0.78± 0.02,0.69±0.07,OD values of 24 hours were 0.77±0.07,0.74±0.07,0.59±0.07,0.50±0.09,0.39±0.08,OD values of 48 hours were 0.78±0.05,0.61±0.08,0.44±0.10,0.39±0.08,0.34±0.07 respectively.Each drug intervention group had significant difference compared with control group.Qishen-Qinggan decoction of 0.27,0.54 g/ml could induce apoptosis of SMMC-7721 were 11.19 ± 2.23,15.69 ± 2.51,compared with control group of 1.41 ± 0.22.Apoptotic rates of Qishen-Qinggan decoction of 1.08 g/ml had extremely significant difference with the control group (41.83 ± 7.11 vs 1.41 ± 0.22).Conclusion Qishen-Qinggan decoction could inhibit the proliferation of SMMC-7721 cells probably by inducing cell apoptosis.
8.Application of TCM constitution theory on non-alcoholic fatty liver disease
Lingfeng WAN ; Xiuqin WANG ; Boyu XUE
International Journal of Traditional Chinese Medicine 2013;(3):237-239
Objective To discuss clinical application of theory of TCM constitution on non-alcoholic fatty liver disease (NAFLD).Methods Literature of nearly 15 years was summarized to analyze the development of Chinese constitution theory and its application of on typing and treatment of NAFLD.Results Main types of constitution in patients with NAFLD were phlegm-dampness,deficiency of Qi and damp-heat.On tiis basis,treatment based syndrome differentiation of TCM could effectively prevent and treat NAFLD.Conclusion TCM constitution theory provide research ideas for NAFLD.
Result Analysis
Print
Save
E-mail