ObjectiveTo develop a predictive model and application for spontaneous passage of common bile duct stones using automated machine learning algorithms given the complexity of treatment decision-making for patients with common bile duct stones， and to reduce unnecessary endoscopic retrograde cholangiopancreatography （ERCP） procedures. MethodsA retrospective analysis was performed for the data of 835 patients who were scheduled for ERCP after a confirmed diagnosis of common bile duct stones based on imaging techniques in Changshu First People’s Hospital （dataset 1） and Changshu Traditional Chinese Medicine Hospital （dataset 2）. The dataset 1 was used for the training and internal validation of the machine learning model and the development of an application， and the dataset 2 was used for external testing. A total of 22 potential predictive variables were included for the establishment and internal validation of the LASSO regression model and various automated machine learning models. The area under the receiver operating characteristic curve （AUC）， sensitivity， specificity， and accuracy were used to assess the performance of models and identify the best model. Feature importance plots， force plots， and SHAP plots were used to interpret the model. The Python Dash library and the best model were used to develop a web application， and external testing was conducted using the dataset 2. The Kolmogorov-Smirnov test was used to examine whether the data were normally distributed， and the Mann-Whitney U test was used for comparison between two groups， while the chi-square test or the Fisher’s exact test was used for comparison of categorical data between groups. ResultsAmong the 835 patients included in the study， 152 （18.20%） experienced spontaneous stone passage. The LASSO model achieved an AUC of 0.875 in the training set （n=588） and 0.864 in the validation set （n=171）， and the top five predictive factors in terms of importance were solitary common bile duct stones， non-dilated common bile duct， diameter of common bile duct stones， a reduction in serum alkaline phosphatase （ALP）， and a reduction in gamma-glutamyl transpeptidase （GGT）. A total of 55 models were established using automated machine learning， among which the gradient boosting machine （GBM） model had the best performance， with an AUC of 0.891 （95% confidence interval： 0.859‍ ‍—‍ ‍0.927）， outperforming the extreme randomized tree mode， the deep learning model， the generalized linear model， and the distributed random forest model. The GBM model had an accuracy of 0.855， a sensitivity of 0.846， and a specificity of 0.857 in the test set （n=76）. The variable importance analysis showed that five factors had important influence on the prediction of spontaneous stone passage， i.e.， were solitary common bile duct stones， non-dilated common bile duct， a stone diameter of <8 mm， a reduction in serum ALP， and a reduction in GGT. The SHAP analysis of the GBM model showed a significant increase in the probability of spontaneous stone passage in patients with solitary common bile duct stones， non-dilated common bile duct， a stone diameter of <8 mm， and a reduction in serum ALP or GGT. ConclusionThe GBM model and application developed using automated machine learning algorithms exhibit excellent predictive performance and user-friendliness in predicting spontaneous stone passage in patients with common bile duct stones. This application can help avoid unnecessary ERCP procedures， thereby reducing surgical risks and healthcare costs.

Objective·To compare the performance of various deep learning methods in the segmentation and classification of colorectal polyp endoscopic images,and identify the most effective approach.Methods·Four colorectal polyp datasets were collected from three hospitals,encompassing 1 534 static images and 15 videos.All samples were pathologically validated and categorized into two types:serrated lesions and adenomatous polyps.Polygonal annotations were performed by using the LabelMe tool,and the annotated results were converted into integer mask formats.These data were utilized to train various architectures of deep neural networks,including convolutional neural network(CNN),Transformers,and their fusion,aiming to develop an effective semantic segmentation model.Multiple performance indicators for automatic diagnosis of colon polyps by different architecture models were compared,including mIoU,aAcc,mAcc,mDice,mFscore,mPrecision and mRecall.Results·Four different architectures of semantic segmentation models were developed,including two deep CNN architectures(Fast-SCNN and DeepLabV3plus),one Transformer architecture(Segformer),and one hybrid architecture(KNet).In a comprehensive performance evaluation of 291 test images,KNet achieved the highest mIoU of 84.59％,significantly surpassing Fast-SCNN(75.32％),DeepLabV3plus(78.63％),and Segformer(80.17％).Across the categories of"background","serrated lesions"and"adenomatous polyps",KNet's intersection over union(IoU)were 98.91％,74.12％,and 80.73％,respectively,all exceeding other models.Additionally,KNet performed excellently in key performance metrics,with aAcc,mAcc,mDice,mFscore,and mRecall reaching 98.59％,91.24％,91.31％,91.31％,and 91.24％,respectively,all superior to other models.Although its mPrecision of 91.46％was not the most outstanding,KNet's overall performance remained leading.In inference testing on 80 external test images,KNet maintained an mIoU of 81.53％,demonstrating strong generalization capabilities.Conclusion·The semantic segmentation model of colorectal polyp endoscopic images constructed by deep neural network based on KNet hybrid architecture,exhibits superior predictive performance,demonstrating its potential as an efficient tool for detecting colorectal polyps.

Objective To develop deep learning models for colonoscopy quality control using various deep learning architectures,and to delve into the decision-making mechanisms.Methods The colonoscopy images were selected from two datasets separately constructed by the HyperKvasir and Changshu Hospital Affiliated to Soochow University,encompassing intestines of varying degrees of cleanliness,polyps,and cecums.After image preprocessing and enhancement,transfer learning was carried out using the pre-trained models based on convolutional neural network(CNN)and Transformer.The model training adopted cross-entropy loss functions and Adam optimizer,and simultaneously implemented learning rate scheduling.To enhance model transparency,a thorough interpretability analysis was conducted using Grad-CAM,Guided Grad-CAM,and SHAP.The final model was converted to ONNX format and deployed on various equipment terminals to achieve real-time colonoscopy quality control.Results In a dataset of 3 831 colonoscopy images,EfficientNet model outperformed the other models on the test set,achieving an accuracy of 0.992 which was higher than those of the other models based on CNN(DenseNet121,ResNet50,VGG19)and Transformer(ViT,Swin,CvT),with a precision,recall rate,and F1 score of 0.991,0.989,and 0.990.On an external test set of 358 images,EfficientNet model had an average AUC,precision,and recall rate of 0.996,0.948,and 0.952,respectively.Although EfficientNet model is high-performing,some misjudgments still occurred.Interpretability analysis highlighted key image areas affecting decision-making.In addition,EfficientNet model was successfully converted to ONNX format and deployed on multiple platforms and devices,and it ensured real-time colonoscopy quality control with an inference speed of over 60 frames per second.Conclusion Among the 7 models developed for colonoscopy quality control based on CNN and Transformer,EfficientNet demonstrated exemplary performance across all categories and is deployed for real-time predictions on multiple terminals,aiming to provide patients with better medical care.

It is well known that Tokyo University Hospital (TUH) is one of the most famous medical centers around the world in Japan. Among several departments in TUH, the Department of Hepatobiliary-Pancreatic Surgery and Artificial Organ and Transplantation Division are representative predominant divisions, which have high-quality diagnostic and treatment service systems being at the leading position in the world. The authors visit the Department of Hepatobiliary-Pancreatic Surgery and Artificial Organ and Transplantation Division in 2022. When studying and working in TUH, the authors make a multi-angle comparisons of the development status and technical levels of hepatobiliary and pancreatic surgery among TUH, western countries and China. It aims to share and exchange with fellow surgeons at home and abroad, and bring benefits or enlightenment to the professional advancement of them.

Objective:To evaluate the oncologic outcomes of different laparoscopic radical hysterectomy.Methods:From January 2011 to December 2014, the laparoscopic operation cases of cervical cancer at stage Ⅰb1, Ⅰb2, Ⅱa1 and Ⅱa2, including the histologic subtypes of squamous-cell carcinoma, adenocarcinoma and adenosquamous carcinoma, were collected in five clinical centers. The data were divided into two groups according to the surgical procedures, that is, modified laparoscopic-vaginal radical hysterectomy (mLVRH) and total laparoscopic radical hysterectomy (TLRH). The overall survival rate (OS), disease-free survival rate (DFS) at 5 years were retrospectively analyzed in this study.Results:There were 674 cases in total, including 377 cases of mLVRH, 297 cases of TLRH. (1) The OS at 5 years: the mLVRH was 96.1% and the TLRH was 92.0%, and the mLVRH was higher than that of TLRH （ P=0.010）. Stratify analysis， including stage of disease （Ⅰb1 and Ⅱa1）， histologic subtypes （squamous-cell carcinoma, adenocarcinoma）， lymph node metastasis， revealed that, ① Stage of disease: in stage Ⅰb1, the OS at five years of mLVRH was higher than that in TLRH group （98.6% vs 93.6%, P=0.012）. In stage Ⅱa1, there was significant difference between the two groups, the OS at five years of mLVRH and TLRH were 93.6% and 77.6% （ P=0.007）. ② Histologic subtypes: for the OS at five years of squamous-cell carcinoma, mLVRH and TLRH were 96.1% and 92.3%, and there was significant difference （ P=0.046）； for adenocarcinoma, the OS at five years were 91.0% and 88.6%， and there was no difference between two groups （ P=0.230）. ③ Lymph node metastasis： the mLVRH and TLRH with lymph node metastasis, the OS at five years were 98.6% and 96.4%; the mLVRH and TLRH without lymph node metastasis, the OS at five years were 89.3% and 80.8%. There were no significant differences between the two groups,respectively ( P=0.156， P=0.093）. (2) The DFS at 5 years: there was no significant difference between mLVRH and TLRH (94.1% vs 90.9%, P=0.220). Stratify analysis for stage of disease, the mLVRH group was higher than that in the TLRH group in stage Ⅰb1 (97.0% vs 92.8%, P=0.039). However, for stage Ⅱa1, there was no significant difference between mLVRH and TLRH group （88.2% vs 75.8%, P=0.074）. Conclusions:The results of this retrospective study indicated that different laparoscopy surgical procedures had diverse oncologic outcomes. The OS at 5 years of the mLVRH is superior to the TLRH. The DFS at 5 years in Ⅰb1 stage, the mLVRH is higher than the TLRH. Therefore, the modified laparoscopy is still an alternative surgery for early cervical cancer patients when following the principle of no-tumor-exposure.

Objective To compare the efficacy,safety and tolerability between bromfenac sodium ophthalmic solution 0.1％ eye drop and loteprednol etabonate 0.5％ eye drop after femtosecond laser in situ keratomileusis (FS-LASIK).Methods A prospective randomized controlled trail was performed.Seventy-six patients who had been undergone FS-LASIK in Daping Hospital were enrolled from December 2013 to December 2014.They were divided into two groups randomly by using the random number table method.The bromfenac sodium group included 38 patients administered 0.1％ bromfenac sodium ophthalmic solution,and the loteprednol etabonate group comprised 38 patients treated with 0.5％ loteprednol etabonate.Patients in both groups initially administered tobramycin-dexamethasone four times a day after FS-LASIK for 1 week.Then the patients in the bromfenac sodium group were treated with bromfenac sodium twice a day until 1 month postoperatively.While the patients from the loteprednol etabonate group were treated with loteprednol etabonate twice a day until 1 month after surgery.Visual acuity,refraction,intraocular pressure and ocular manifestation were examined preoperative and 1 day,1 week,1 month,3 months,6 months after FS-LASIK.This study was approved by the medical ethics Committee of Daping Hospital.Results No significant difference was found in the rate of uncorrected visual acuity (UCVA) which achieved to preoperative best corrected visual acuity (BCVA) between the two groups at each time point after operation (x2 =8.000,P =0.613).There was no difference in spherical equivalent (SE) between the two groups postoperatively (Fgroup =1.137,P =0.288),but a significant difference was found between the different time points (Ftime =29.771,P =0.000).Postoperative SE in the two groups tended to have mild hyperopia in the early days,and then gradually tended to " zero" or mild myopia.There was a significant difference with intraocular pressure among different time points (Ftime =87.519,P =0.000),but no significant difference of intraocular pressure was found between the groups (Fgroup =1.989,P =0.161).One month after surgery,the intraocular pressure were higher than 20 mmHg in two eyes in the loteprednol etabonate group,and two eyes were 5 mmHg higher than preoperative intraocular pressure.In bromfenac group,no one was higher than 5 mmHg compared with the preoperative intraocular pressure,and the pressures of all the eyes were lower than 18 mmHg.All the patients in bromfenac group were no drug withdrawal.But three patients were forced to guide drug withdrawal due to high intraocular pressure in loteprednol group.Conclusions Bromfenac sodium is safe,effective and well tolerated after FS-LASIK.

Background Corneal biomechanical properties is important in the safety assessment of corneal refractive surgery.Corvis is a new device for measuring corneal biomechanics properties.Objective This study was to observe the correlation among corneal thickness, Corvis intraocular pressure and corneal biomechanical properties with Corvis.Methods A prospective observational study was performed.One hundred and fifty eyes of 75 patients with corneal thickness from 501 μm to 590 μm were divided into three groups according to the corneal thickness:low corneal thickness group (corneal thickness range from 501 μm to 530 μm), middle corneal thickness group (corneal thickness range from 531 μm to 560 μm) ,and high corneal thickness group (corneal thickness range from 561 μm to 590 μm);and 50 eyes of 25 patients for each group.The difference of intraocular pressure, corneal thickness and deformation amplitude (DA) among the three groups were analyzed by one-way ANOVA and the correlation among the groups were analyzed by liner regression.Results The DA in the low corneal thickness group and middle corneal thickness group were significantly higher than that in the high corneal thickness group (P ＜ 0.05).The intraocular pressure was statistically different among the 3 groups (F =9.98, P＜0.05).DA was negatively correlated with intraocular pressure and corneal thickness (r=-0.84,-0.33;both at P＜0.01), with the linear regression DA =1.69-0.04×IOP (F=366.19, t=-19.14,P＜0.01).Conclusions Corneal thickness cannot simply represent the corneal biomechanical properties in the safety assessment of corneal refractive surgery,IOP should be considered.

OBJECTIVETo construct dengue virus-specific full-length fully human antibody libraries using mammalian cell surface display technique.

METHODSTotal RNA was extracted from peripheral blood mononuclear cells (PBMCs) from convalescent patients with dengue fever. The reservoirs of the light chain and heavy chain variable regions (LCκ and VH) of the antibody genes were amplified by RT-PCR and inserted into the vector pDGB-HC-TM separately to construct the light chain and heavy chain libraries. The library DNAs were transfected into CHO cells and the expression of full-length fully human antibodies on the surface of CHO cells was analyzed by flow cytometry.

RESULTSUsing 1.2 µg of the total RNA isolated from the PBMCs as the template, the LCκ and VH were amplified and the full-length fully human antibody mammalian display libraries were constructed. The kappa light chain gene library had a size of 1.45×10(4) and the heavy chain gene library had a size of 1.8×10(5). Sequence analysis showed that 8 out of the 10 light chain clones and 7 out of the 10 heavy chain clones randomly picked up from the constructed libraries contained correct open reading frames. FACS analysis demonstrated that all the 15 clones with correct open reading frames expressed full-length antibodies, which could be detected on CHO cell surfaces. After co-transfection of the heavy chain and light chain gene libraries into CHO cells, the expression of full-length antibodies on CHO cell surfaces could be detected by FACS analysis with an expressible diversity of the antibody library reaching 1.46×10(9) [(1.45×10(4)×80%)×(1.8×10(5)×70%)].

CONCLUSIONUsing 1.2 µg of total RNA as template, the LCκ and VH full-length fully human antibody libraries against dengue virus have been successfully constructed with an expressible diversity of 10(9).

Animals ; Antibodies, Viral ; CHO Cells ; Cell Surface Display Techniques ; Cricetinae ; Cricetulus ; Dengue Virus ; immunology ; Gene Library ; Genetic Vectors ; Humans ; Immunoglobulin Heavy Chains ; immunology ; Transfection

OBJECTIVETo explore the association of fish intake with the risk of renal cancer.

METHODSPubMed, Embase, CNKI and CA databases were searched for case-control studies or cohort studies examining the relationship between fish or fish products intake and renal cancer. Heterogeneity among the selected studies was assessed using I2 score, and the publication bias was assessed using funnel plots.

RESULTSSeventeen articles were included in the analysis with a heterogeneity across the studies (P=0.003, I(2)=52.3%). A random-effects model was used to generate the pooled risk ratio (RR) and 95% confidence interval (CI), and no statistically significant association was found between the risk of renal cancer and fish intake (RR=0.90; 95% CI, 0.78-1.02). In subgroup analysis, no evidence was found that the study design, study region or publication date influenced the results; but in the gender subgroup analysis, fish intake we found to decrease the risk of renal cancer in men but not in women.

CONCLUSIONThe results of meta-analysis do not support an association between fish intake and a lowered risk of renal cancer.

Carcinoma, Renal Cell ; etiology ; Diet ; Female ; Fish Products ; Humans ; Kidney Neoplasms ; etiology ; Male ; Odds Ratio ; Risk Factors

OBJECTIVETo construct full-length human bladder cancer-specific antibody libraries for efficient display of full-length antibodies on the surface of mammalian cells.

METHODSThe total RNA was isolated from peripheral blood mononuclear cells from patients with bladder cancer. The repertoires of IgG1 heavy chain variable region (VH) and Kappa light chain were amplified by RT-PCR using specific primers. The antibody genes were inserted into the vector pDGB-HC-TM to construct the bladder-cancer-specific antibody libraries of heavy chains and light chains. Ten clones from each library were randomly picked for gene sequencing and transient transfection into FCHO cells to analyze antibody display on mammalian cell surface by flow cytometry after staining with corresponding fluorescent labeled antibodies.

RESULTSThe libraries of bladder-cancer-specific antibody heavy chain (IgG1) and light chain (LCk) were successfully constructed. Seven out of the 10 clones randomly selected from the heavy chain library and 9 out of the 10 clones from the light chain library showed correct open reading frame, coding for 7 unique VH and 9 unique LCk. The combinatory library size reached 3.32×10(11).

CONCLUSIONWe have successfully constructed a full-length human bladder-cancer-specific antibody library with a combinatory diversity of 3.32×10(11) based on mammalian display technology, which can be used for screening monoclonal antibodies against bladder-cancer-associated antigens.

Amino Acid Sequence ; Animals ; Antibodies ; genetics ; Cell Surface Display Techniques ; Gene Library ; Humans ; Immunoglobulin Heavy Chains ; genetics ; Immunoglobulin kappa-Chains ; genetics ; Peptide Library ; Urinary Bladder Neoplasms ; genetics ; immunology